Philology and Criticism

i Philology and Criticism ii CULTURAL, HISTORICAL AND TEXTUAL STUDIES OF SOUTH ASIAN RELIGIONS The volumes featured in the Anthem Cultural, Historical and Textual Studies of South Asian Religions series are the expression of an international community of scholars committed to the reshaping of the ield of textual and historical studies of religions. The volumes in this series examine practice, ritual and other textual religious products, crossing diferent area studies and time frames. Featuring a vast range of interpretive perspectives, this innovative series aims to enhance the way we look at religious traditions. Series Editor Federico Squarcini, University of Florence, Italy Editorial Board Piero Capelli, Ca’ Foscari University of Venice, Italy Vincent Eltschinger, ICIHA, Austrian Academy of Sciences, Austria Christoph Emmrich, University of Toronto, Canada James Fitzgerald, Brown University, USA Jonardon Ganeri, University of Sussex, UK Barbara A. Holdrege, University of California, Santa Barbara, USA Sheldon Pollock, Columbia University, USA Karin Preisendanz, University of Vienna, Austria Alessandro Saggioro, Sapienza University of Rome, Italy Cristina Scherrer-Schaub, University of Lausanne and EPHE, France Romila Thapar, Jawaharlal Nehru University, India Ananya Vajpeyi, University of Massachusetts, Boston, USA Marco Ventura, University of Siena, Italy Vincenzo Vergiani, University of Cambridge, UK iii Philology and Criticism A Guide to Mahābhārata Textual Criticism Vishwa Adluri and Joydeep Bagchee iv Anthem Press An imprint of Wimbledon Publishing Company www.anthempress.com This edition irst published in UK and USA 2018 by ANTHEM PRESS 75–76 Blackfriars Road, London SE1 8HA, UK or PO Box 9779, London SW19 7ZG, UK and 244 Madison Ave #116, New York, NY 10016, USA Copyright © Vishwa Adluri and Joydeep Bagchee 2018 The author asserts the moral right to be identiied as the author of this work. All rights reserved. This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License. Without limiting the rights under copyright reserved above, no part of this publication may be reproduced, stored or introduced into a retrieval system, or transmitted, in any form or by any means (electronic, mechanical, photocopying, recording or otherwise), without the prior written permission of both the copyright owner and the above publisher of this book. British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library. ISBN-13: 978-1-78308-576-7 (Hbk) ISBN-10: 1-78308-576-2 (Hbk) This title is also available as an e-book. v nā rā yaṇaṃ namaskṛtya naraṃ caiva narottamam | devīṃ sarasvatīṃ caiva tato jayamudīrayet || Dedicated to Vishnu Sitaram Sukthankar vi vi Philology […] has become the modern form of criticism. —Michel Foucault, The Order of Things vi ix CONTENTS List of Illustrations xi Foreword xv Preface xvii Acknowledgments xxi Prologue xxiii Chapter Summaries 1 Introduction Ad Fontes, Non Ultra Fontes! 11 About This Book 11 Why a Critical Edition? 11 What Is a Critical Edition? 12 How to Interpret the Critical Edition 17 Conclusion 20 Chapter One Arguments for a Hyperarchetypal Inference 45 The Normative Redaction Hypothesis 45 Normative Redaction, Archetype and Original 46 Criticism: Higher and Lower 49 The Argument from Spread and the Argument from Resilience 54 The Argument from Empty Reference 67 The Argument from Loss 75 Chapter Two Reconstructing the Source of Contamination 119 Understanding “Contamination” 119 Contamination: Hyperarchetypal and Extra-stemmatic 120 Identifying the Source of Contamination 122 The Argument from Uncertainty 126 The Argument from Oral Source 131 The Argument from (Postulated) Antiquity and the Argument from Ideology 140 Chapter Three Confusions Regarding Classiication 163 Classiication: Typological and Genealogical 163 Determining Filiation 164 Eliminating Witnesses 167 x x PHILOLOGY AND CRITICISM The Argument from Brevity and the Argument from False Premises 169 The Argument from a Misapprehension Concerning Classiication (Schriftartprämisse) 182 The Argument from Extensive Contamination 209 The Argument from Independent Recensions 247 The Argument from Expertise 269 Conclusion: Textual Criticism and Indology 319 Epilogue 339 Appendices 343 1. The Volumes of the Critical Edition 343 2. Editions Besides the Critical Edition 345 3. English Translations of the Mahā bhā rata 347 4. How to Use the Critical Apparatus 351 5. How Editors Reconstructed the Reading of the Archetype 355 6. How to Cite the Mahā bhā rata 357 7. The Extent of the Mahā bhā rata’s Books 359 8. The 18 Parvans and 100 Upaparvans of the Mahā bhā rata 361 9. The Arrangement of the Parvans in the Southern Recension 367 10. Other Narrative Divisions 379 11. Sukthankar’s Table of the Manuscripts Collated for the Ā diparvan 383 12. Extent of the Ś ā radā Codex for the Ā diparvan 385 13. Abbreviations and Diacritical Signs Used in the Critical Edition 387 14. Abbreviated Concordance of the Principal Editions of the Mahā bhā rata 389 15. Stemmata for the Diferent Parvans of the Mahā bhā rata 393 16. Commentaries on the Mahā bhā rata 397 17. Commentaries on the Bhagavadgītā 425 18. The Use of Venn Diagrams to Depict Manuscript Relationships 429 Glossary 481 Annotated Bibliography 493 Index 525 xi ILLUSTRATIONS Figures 1 The two options of a philology oriented toward the text and a philology oriented toward the witness 13 2 The constituted text along with its critical apparatus: understanding what one is reading 17 3 The part stands in for the whole 23 4 The birth and death of manuscripts 25 5 Textual tree of Ā diparvan versions, illustrating the stemmatic relationships 46 6 The “real” stemma 48 7 Maas’s hypothetical stemma, illustrating the distinction between hyparchetype, archetype and original 51 8 Flores’s argument from the spread of errors, and Bigger’s normative redaction hypothesis 58 9 Bigger’s argument from the resilience of tradition 61 10 The stemma as a minimal architecture 64 11 Our abstract stemma 65 12 Making the archetype and the normative redaction coincide 66 13 Bigger’s argument from “empty reference” 72 14 Bigger’s “prehistory of the normative redaction” 122 15 Reconstructing the source of contamination 123 16 Extra-stemmatic contamination into an extant witness 127 17 The consequences of assuming extra-stemmatic contamination 128 18 Extra-stemmatic contamination into a direct descendant of the archetype: S as an example 128 19 The consequences of assuming extra-stemmatic contamination into S 129 20 S as the original oral epic 129 21 Extra-stemmatic contamination into a direct descendant of the archetype: N as an example 130 22 The consequences of assuming extra-stemmatic contamination into N 130 23 N as the original oral epic 130 24 The original oral epic as the source of N 133 25 Contamination via an oral source 134 26 Recentiores non deteriores 135 27 Recentiores deteriores 137 xi xii PHILOLOGY AND CRITICISM 28 Transmission via an oral source and the inevitability of a written intermediary 138 29 Von Büren’s stemma of the manuscript tradition of Isidore’s Etymologies 142 30 Hyperarchetypal contamination, extra-stemmatic contamination and the resilience of tradition 143 31 Eliminatio 168 32 Grünendahl reproduces Sukthankar’s stemma 184 33 Sukthankar’s list of manuscripts forming the critical apparatus 187 34 Sukthankar’s stemma reversed 180 degrees around a central axis 191 35 Reversed stemma with the subrecensions in turn reversed around a central axis 192 36 Lüders’s list of the manuscripts collated for his sample critical edition 194 37 Treating each manuscript as an independent witness 196 38 Groups versus individual witnesses 198 39 Mapping the relationship of manuscripts within a group to each other 199 40 The evolution of northern Brā hmī 207 41 The evolution of southern Brā hmī 207 42 Agreement between independent versions 211 43 Stemma lectionum of verse 1.29.5 212 44 Alternative stemma lectionum of verse 1.29.5 212 45 Contamination, undermining the assumption of independence 213 46 Contamination, the real nature of the relationships in our stemma 214 47 Contamination from and into the central subrecension 215 48 The interpolated passage 321* 216 49 D as the source of the interpolated passage 321* 217 50 γ as the source of the interpolated passage 321* 218 51 (Non)contamination of Ñ4 and D2.5 with K 220 52 Constituting groups on the basis of additional passages missing from manuscripts 222 53 Identifying a core K group on the basis of missing additional passages 223 54 Using the absence of interpolations to reine the classiication of manuscripts 225 55 The descent of S in Grünendahl’s classiication 227 56 The true position of S in Grünendahl’s classiication 228 57 The order of interpolations 229 58 Unrelated manuscripts on the same stemma 231 59 The expansion of the text according to Grünendahl 232 60 Open and closed branches of the tradition 233 61 Brushing aside the dead ends 234 62 Hypothetical stemma with K as the archetype 237 63 Understanding Grünendahl’s model for reconstructing archetypes 239 64 Stemma lectionum of verses 1.1.1A and 1.1.2 240 xi ILLUSTRATIONS xiii 65 The expansion of the text according to Grünendahl 244 66 Abolishing the distinction between group and version 245 67 The fragmentation of the K group into K1 and K0.2–6 according to Grünendahl 246 68 Collapsing Ś1 and K1 into a single version 247 69 Grünendahl explains “contamination” 256 70 Reconstructing the reading of 1.1.1A 356 71 P. P. S. Sastri’s scheme of adhyāyas from his critical edition of the southern recension 371 72 P. P. S. Sastri’s scheme of ślokas from his critical edition of the southern recension 372 73–77 Parvan divisions and śloka counts of the Kumbhakonam recension 373 78 A. C. Burnell’s stemma 393 79 V. S. Sukthankar’s “Pedigree of Ā diparvan versions” 394 80 V. S. Sukthankar’s stemma of the Ā raṇyakaparvan manuscripts 394 81 R. Vira’s stemma of the Virā ṭaparvan manuscripts 395 82 S. K. De’s stemma of the Udyogaparvan manuscripts 395 83 P. L. Vaidya’s stemma of the Harivaṃśā manuscripts 396 84 Brockington’s “Venn diagram” 436 85 A “Venn diagram” depicting sets whose members are individual manuscripts 439 86 The corrected “Venn diagram” 440 87 A “Venn diagram” of sets that contain manuscripts of the S and NE groups 441 88 A “Venn diagram” of sets containing manuscripts of the S and NE groups 442 89 A modiied version of Brockington’s original “Venn diagram” 442 90 The corrected “Venn diagram” 443 91 Expressing relations between two orders of sets 444 92 The source of the error 446 93 Mapping interpolations using a “Venn diagram” 447 Tables 1 Grünendahl’s list of interpolations in the northern recension 170 2 How Grünendahl imagines the process of constitution 174 3 Grünendahl’s table of the distribution of interpolations in the northern recension 177 4 Grünendahl’s error in constituting the text 179 5 The text as the true basis of classiication 205 6 K contamination and the fragmentation of the Mahā bhā rata tradition 238 7 How Grünendahl imagines the classiication of manuscripts 250 8 Grünendahl establishes “iliation” 252 xvi xiv PHILOLOGY AND CRITICISM 9 Grünendahl establishes that K3 is not contaminated with the passages with which it is not contaminated (hence, the editor has failed to establish the contamination of K3) 254 10 Grünendahl demonstrates that ν and S cannot be against γ (for at least one γ manuscript always agrees with them) 261 11 Sukthankar considers the agreement of the versions 265 xv FOREWORD Philology and Criticism is the irst book of its kind. Incisive in its analysis, this book undertakes a rigorous defense of the Mahā bhā rata critical edition. Following a prologue and an introduction, this book is divided into three chapters. Each chapter states a problem and discusses key concepts and principles in textual criticism pertaining to it. Thereafter, the authors guide the reader through a history of responses to the problem. Each response is posed as an argument (via citing the critic who raised it). The authors address each argument individually in a separate section. In each section, they consider whether the argument can be defended from some perspective. Once they establish that the argument is untenable, they state their conclusion. In this way, they systematically work through contemporary criticisms of the critical edition, focusing primarily on Andreas Bigger’s and Reinhold Grünendahl’s work. The irst chapter addresses the view that the constituted text of the critical edition reconstructs merely a late stage of the transmission. Although several scholars advocate this thesis (James L.  Fitzgerald, for instance, thinks the critical edition reconstructs a “Gupta-era archetype,” which he elsewhere calls a “written Sanskrit text” of the epic), the authors focus on the thesis’s author: Andreas Bigger. Bigger holds that the critical edition merely reconstructs a text he calls the “normative redaction” of the Mahā bhā rata, supposedly the result of “a uniform redaction” of the epic undertaken during its irst transcription from a luid oral tradition. Adluri and Bagchee demonstrate the circularity of this claim. The second chapter addresses the underlying assumption of Bigger’s work, though it also broadens the scope to include other Mahā bhā rata critics. The authors show that Bigger’s thesis appears plausible only because scholars assume an oral epic pre- ceded the written Sanskrit Mahā bhā rata. The authors demonstrate that their arguments are not stemmatic and hence do not hold. The chapter concludes with a discussion of the so-called analytic approach to the epic. The authors argue that this approach is pre- mised on an uncritical view of Indian history, whose origins they outlined in their book The Nay Science: A History of German Indology. The third chapter, the longest in this book, addresses a perplexing problem:  How were the Mahā bhā rata manuscripts classiied? Were they classiied by script as Grünendahl argues? The authors convincingly demon- strate that they were not. As they show, script cannot play a role in classiication for it is an external marker. This chapter also addresses Grünendahl’s claim that extensive con- tamination makes a critical edition of the Mahā bhā rata unachievable as well as his claim that focusing on a regional recension would have led to a better edition. The conclusion provides a summation not only of this book but also of the authors’ irst book. It presents a serious challenge to contemporary Sanskrit philology inasmuch as it relies on opinion rather than argument. xvi xvi PHILOLOGY AND CRITICISM While I  disagree with the authors in some respects (most notably on the southern recension’s place in the Mahā bhā rata tradition), Philology and Criticism will stimulate debate. It poses a major challenge to scholars who have made unguarded statements about the Mahā bhā rata’s origins in an oral tradition. As I  have argued since 2001 (Rethinking the Mahābhārata: A Reader’s Guide to the Education of the Dharma King), the notion of an oral epic is a myth. The Mahā bhā rata is clearly a written text. In all likelihood, a small group of Brahmans created it. T. P. Mahadevan identiies these Brahmans with the his- torical Pū rvaśikha Brahmans. In spite of the text’s expansion and changes, no evidence exists that it was ever transmitted orally. Philology and Criticism conclusively vindicates this view. As I demonstrated in my review of Fitzgerald’s translation of Books 11 and 12 of the Mahā bhā rata (in the Journal of the American Oriental Society), those who resurrect the oral hypothesis do so for ideological reasons rather than because persuasive historical evidence exists. Relying on nineteenth-century views of the epic (whose assumptions Adluri and Bagchee criticized in their irst book), they overlook the fact that the archetype presupposes a written transmission. Adluri and Bagchee have staked their position, and scholars in the future will have to account for their view in some way. I hardly need add that Philology and Criticism is essential reading for Mahā bhā rata scholars. This book signiicantly advances our knowledge of the critical edition. It is an essential reference work—not least because of its appendices, which enable scholars to consult details of the edition without carrying around all 19 volumes of the edition. Philology and Criticism also addresses a major lacuna in Mahā bhā rata studies today. It is the irst work to explain what the Mahā bhā rata critical edition is, how it was created, what its merits are and why criticisms of the edition are frequently based on insuicient knowl- edge of the principles involved. Though aimed at an advanced audience, the analysis is clear and systematic, and the arguments can be followed by anyone who takes the time and trouble (and perhaps uses paper and a pencil). In that sense it is not a diicult book to read, though its scope is breathtaking. Few today in Mahā bhā rata studies have such a thorough grasp of the critical edition or are as qualiied to speak to the issues of textuality, orality, the manuscript tradition, what can be reconstructed and what can be shown with philological methods. In my assessment, the authors present a cogent inter- pretation of the critical edition. Their clariication of its overarching project is brilliant and makes a lasting contribution to the ield. Alf Hiltebeitel Washington, DC xvi PREFACE This book is a guidebook. That means it is intended for use.1 The reader must make use of the tools presented herein to test for herself the validity and rigor of the arguments of contemporary Mahā bhā rata scholars. This is all the more necessary as hardly a ield of scholarship existing today is as rife with competing and contradictory theories as con- temporary Mahā bhā rata studies. This book does not claim to be a comprehensive guide to textual criticism (for that the reader will have to read one of the classic manuals on the subject such as M. L. West’s Textual Criticism and Editorial Technique or Giorgio Pasquali’s Storia della tradizione e critica del testo),2 but it does claim to bring to Mahā bhā rata criticism the sort of reason, precision, clarity of thought and beauty that is a hallmark of textual criticism in other ields. Regrettably, Mahā bhā rata studies has long been a stronghold of neo-Aryanist ideology, anti-Semitism and Romantic fantasy. Thus, this book has had to be equal parts guidebook and polemical essay. As such, it owes as much to the tradition of Nietzsche as to that of West. There is no doubt in our minds that the great editors of classical and medieval texts made enduring contributions to the study of texts and, in particular, to the canons of method that enable us to expunge centuries of error and dross and come closer to the authors’ original texts. There is also no doubt in our minds that these techniques are also applicable (with the necessary riders and adjustments) to the study of Indian texts. As Pasquali wrote (and Paolo Trovato now airms), “I, at least, cannot imagine that the original, say, of a Chinese or Bantu text could be constructed from copies or any other testimony, in sum, from its tradition, otherwise than on the basis of Maas’s considerations and the rules he laid down.”3 In fact, there is an entire tradition of Indian editors (not only V. S. Sukthankar, to whom this book is dedicated, but also S. K. Belvalkar, P. L. Vaidya and others4) who could rightfully take their place alongside the great editors in the classics. But (and we are no less convinced of this than of the appli- cability of textual criticism to the Mahā bhā rata) textual criticism cannot and may not be used to promote ideological agendas. Andreas Bigger writes, “That I make the critical edition of the Mahā bhā rata the foundation of my work does not mean that I approve of the text constituted by the editors in all respects. [Reinhold] Grünendahl and others have demonstrated that in the domain of lower criticism of the Mahā bhā rata the inal word is yet to be spoken. The results of this criticism of the ‘irst critical edition’—for one may not forget that Sukthankar quite consciously referred to it as such—will low into this work [Bigger means his book] and, where the evidence forces itself upon us, will be expanded upon.”5 The reader will frequently encounter such statements in the work of Mahā bhā rata authors quoted in this book. She will ind examples of circular reasoning, conclusions that do not follow from their premises, arguments ad baculum and appeals to xvi xviii PHILOLOGY AND CRITICISM inappropriate authority. And she will ind that the only two constants among these theses, theories and dogmatic positions are that Indians are not capable of reading their texts critically and that priests are corrupt and mendacious.6 This failure to engage with the theory of textual criticism has extracted a heavy price. It has meant that Mahā bhā rata scholars have not kept abreast of recent developments in textual criticism, whether in the areas of hermeneutics, literary criticism, structuralism, post-Lachmannian theories of critical editing or the study of variantistica—in short, that entire ield that is today denoted by the term “ecdotics” and encompasses the study of textual cultures in the widest sense.7 It has meant that scholarship on the epic, even considered as arising out of and responding to the documentary impulse,8 has failed to contribute in any meaningful way to a history of the text. Against this intellectually stunted and resentment-driven science, Sukthankar’s genius stands out all the brighter. Notes 1 Technical terms, especially on their irst occurrence, are placed in italics to draw the reader’s attention to the relevant glossary entry. An earlier glossary may be found in S.  M. Katre, Introduction to Indian Textual Criticism. With Appendix II by P. K. Gode, Deccan College Handbook Series 5 (Pune: Deccan College, 1954), 90–99 (“A Glossary of Some Important Terms Used in Textual Criticism”). Sanskrit equivalents may be found in Venkatesh Laxman Joshi, ed., Prauḍha Manoramā with Commentary Śabdaratna, Deccan College Monograph Series 31 (Poona: Deccan College, 1966), 331–52. 2 See the bibliography for further works on the subject; any of the many works on textual criti- cism listed in this book may be read with great proit. Paolo Trovato, Everything You Always Wanted to Know about Lachmann’s Method: A Non-Standard Handbook of Genealogical Textual Criticism in the Age of Post-Structuralism, Cladistics, and Copy-Text (Padua: Libreriauniversitaria, 2014) supersedes the earlier manuals and is now the deinitive resource. 3 Paul Maas, Critica del testo, trans. Nello Martinelli (Florence: Le Monnier, 1972), v, cited and translated in Trovato, Everything You Always Wanted to Know about Lachmann’s Method, 74. Trovato’s comment can be found in the preface to his book (ibid., 21). 4 See Appendix 1 of this volume for a complete list of editors involved in the critical edition pro- ject and also of the volumes and their dates of publication. 5 Andreas Bigger, Balarāma im Mahābhārata: Seine Darstellung im Rahmen des Textes und seiner Entwicklung (Wiesbaden: Harrassowitz, 1998), 15. All translations from German sources are ours. 6 For a discussion of the racial, anti-Semitic and anticlerical resentments that drove this scholar- ship, see our The Nay Science: A History of German Indology (New York: Oxford University Press, 2014). See especially chapters 2 and 4. 7 The somewhat recherché term “ecdotics” will perhaps be unfamiliar to English-speaking readers, but it has been a widespread synonym for textual criticism in Romance languages since Dom Henri Quentin introduced it in his Essais de critique textuelle (Ecdotique) (Paris: Ricard, 1926). For a discussion of the term’s use in Italian textual criticism, see Paola Pugliatti, “Textual Perspectives in Italy: From Pasquali’s Historicism to the Challenge of ‘Variantistica’ (and Beyond),” Text 11 (1998): 164–69. Of course, here we are interested in the term not as a synonym for textual criticism but in its wider sense, where it has come to mean any study that is not limited “to the ways and methods of the traditional critical edition,” but “include[s] all the elements which mark the entire movement of a text from the author to the readers (or users), provided that these elements are viewed from a perspective of editions, ancient or modern, destined for study and for a typographic, digital reading, or under the aspect of whatever third possibility.” This is xi PREFACE xix how the term is deined, for example, in the Foreword of the journal Ecdotica (http://ecdotica. org/index.php?option=com_content&task=view&id=16&Itemid=42). 8 Sheehan expresses the principle best:  “The idea of the textual unconscious was key to the documentary impulse. By divorcing the physical features of the manuscript from its literary content, and by using these physical features to historicize the manuscript, both Mabillon and Montfaucon successfully removed the question of literary content from the domain of serious scholarship. In a sense, they operated within that wider shift from ‘gentlemanly humanism’ to a ‘professionalized philology’ that we have already seen in the English letters in the early part of the eighteenth century. For those on the modern side of the querelles des anciens et des modernes, like the Maurist brothers, scholarship should not be distracted by the idle pleasures of aesthetic judgment. Nor should it be moved by the particular arguments made in the texts it analyzed. Rather, it should invent nonliterary techniques (of which paleography was one) for evaluating documents.” Jonathan Sheehan, The Enlightenment Bible:  Translation, Scholarship, Culture (Princeton, NJ:  Princeton University Press, 2005), 102. In the ield of classical and medieval textual editing, the documentary impulse may have led to signiicant insights. In the ield of Mahā bhā rata studies, however, beyond the superb work done by the editors in the Mahā bhā rata critical edition, the documentary impulse has at most been cited as justiication for not reading the text. It has not led to any meaningful consideration of the physical aspects of the text or the tradition. x xxi ACKNOWLEDGMENTS We owe, irst and foremost, an enormous debt of gratitude to our families for their patience, love and encouragement. We fondly remember Suguna and Indrasena Adluri and Sumati Bhandarkar, who would have loved to read this book. We were fortunate to have many sources for advice for this book. We would like to recognize them here: Michael D. Reeve for reading an early draft; Paolo Trovato for his enthusiastic cham- pionship of our work and his invitation to teach in the summer school on textual criti- cism at the University of Ferrara; Patrick Olivelle for encouraging us to write this book and recommending it to the publisher; Paola Pugliatti for being an invaluable resource on developments in Italian textual criticism; John Lenz for conversations on Nietzsche and philology; Alf Hiltebeitel for his scholarship and for writing the foreword; Bruce M. Sullivan for stimulating discussions about the Mahābhārata; Madhavi Kolhatkar for her profound knowledge of the Vedic tradition; Aruna Bagchee, Edward P. Butler, Matt Newman and Paolo Alberto Celentano for their assistance with translations; and fac- ulty at the Sanskrit Department at Pune University for inviting us to lecture and share our research. The section on the development of Indian scripts is indebted to Saraju Rath at IIAS, Leiden. Jahnavi Bidnur graciously shared her work on the Mahābhārata commentators. We also thank all participants of the textual criticism workshop at the Bhandarkar Oriental Research Institute, Pune. We could not have asked for better students. To our friends and well-wishers Srinivas and Anu Udumudi, Hari Kiran and Padmaja Vadlamani, Aditi Banerjee, Edward P. Butler, Arbogast Schmitt, John Lenz, Peter K. J. Park, Brooks Schramm, Kirsi I. Stjerna, Alice Crary, Robert Yelle, Graham M. Schweig, Susan Ginsburg and Thomas Komarek: thank you, your interest keeps us writing. Omar Dahbour’s support was vital to continuing this work during a critical tran- sition. Thank you to Ami Naramor, Vincent Rajan, and the entire production team at Anthem for producing such a stunning book. Vishwa thanks Swami Prabuddhananda Sarasvati for rigorously instructing him in the traditions of Mīmāṃsa and Vedānta for over two decades. Through him, the philosophical, logical and text-commentarial tra- dition of Ādi Śaṅkarācārya and his successors in the Sringeri Sharada Peetham and Kailash Ashrama became accessible. Finally, this book, which was written over a 10-year period with neither public research grants nor university funding, could not have been completed without Joachim Eichner’s steadfast inancial support. Kalyan Viswanathan, Krishnan and Indu Ramaswamy and Deepanshu and Silvana Bagchee have been stal- wart friends and contributors. Open Access publication was enabled by the Sanatana Dharma Foundation’s ASHEERVADA initiative. Last but not least, Ushakant and Irma Thakkar—your philanthropy, your public service, and your respect for the humanities are an unfailing source of inspiration for us. xxi Sukthankar’s irst (?) attempt at drawing up a stemma codicum of the Ādiparvan versions, dated September 24, 1925. Source: Reproduced from Katre, “Vishnu Sitaram Sukthankar and His Contribution to Indology,” 485. xxii PROLOGUE saṃ gacchadhvaṃ saṃ vadadhvaṃ saṃ vo manāṃsi jānatām | devā bhāgaṃ yathā pūr ve saṃjānānā upāsate || — Ṛg Veda 1.191.2 Czech artist František Kupka (1871–1957)1 painted Le Premier Pas or The First Step, reproduced on the cover of this book.2 The painting on the cover and this essay accom- panying it provide a useful avenue for understanding certain concepts:  origins and archetypes, singularity and plurality, emanation and transmission and the parallelism of cosmology and art. These themes are essential to understanding the conceptually diicult task V. S. Sukthankar set for himself when he created the design for the critical edition of the Mahā bhā rata. Therefore, before we look at the technicalities of a critical edition, we wish to irst explore the logic and artistry of creating stemmata through a related medium: abstract art.3 The choice of abstract art to illustrate the concept of a critical edition may seem strange at irst, yet it is also obvious when one considers that, like abstract art, stem- mata are idealized representations of relationships that have no basis in matter.4 Hence, when we approach stemmatics from the perspective of abstract art, we gain a new per- spective on textual criticism—one that goes beyond the standard presentations of this ield.5 The movement known as “abstractionism” itself originated in the early part of the twentieth century in response to a speciic concern: artists wished to free themselves from the constraints of having to represent something.6 Abstract art and stemmatics thus both respond to a similar problem:  the igurative representation of abstract relations that nonetheless permit us to intuit certain features of reality––features that possibly go beyond what we can intuit with our senses.7 By collating various manuscripts, identifying the textual coherences and harmonies, arranging them according to the logic of emanation and carefully distinguishing the original from the archetype, Sukthankar created an intellectual organization that does justice to the complicated architecture and reception of the Sanskrit epic. His work it- tingly transcends the crude mechanical models created by the Indologists whose work we analyze in this book.8 Trained as a mathematician, with a keen appreciation for the subtle nuances of ideas contained in the text, Sukthankar culled the many extant manuscripts into a single pyramidal architecture.9 Scientists who appreciate genetic relationships, as well as artists who understand how plural elements can be meaningfully organized, will no doubt appreciate his creation.10 Both Sukthankar and Kupka wanted to move beyond the fetishism of facts to an engagement with truth. If Sukthankar mathematically, philosophically and aesthetically xvi xxiv PHILOLOGY AND CRITICISM transcended philology while doing it full justice, Kupka did something similar with art. He wished to transcend the formal and material dimensions of painting by making it self-consciously intellectual, mathematical and spiritual.11 In 1892, Kupka moved from Prague to the Vienna Academy, where he “rea[d] avidly: particularly Greek and German philosophers:  Plato, Aristotle, Paracelsus, Kant, Schopenhauer, Nietzsche and German romantics. Also he rea[d] extensively on astronomy, astrology, Theosophy.”12 Whatever the contemporary evaluation of theosophy as a discipline, it was an early protestation against the gross materialism and hemorrhage of meaning inaugurated by modernity.13 Artists rather than philosophers and social scientists reacted immediately to these trends— Kupka foremost among them.14 As with Sukthankar, no opposition appears between the spiritual and the scientiic in Kupka—a distinction that itself has its root in the obsessive dichotomy between faith and reason endemic to Christianity.15 Margit Rowell notes: Kupka’s most fundamental premises—that nature had a spiritual reality determined by inal causes, that the hidden laws of this reality are present in all of nature’s manifestations including man and the artist’s function is to make visible these laws, not by copying nature but by  creating a parallel order—spring from Goethe’s aesthetic. […] Through a better understanding of natural causes, rhythms, structures and progressions, he hoped to develop a parallel vision, order and language. His interest in physiology, biology and astronomy therefore had its roots in mystical thought. By extension, he paid acute attention to his own sense impressions and evoked coenesthesis as a form of access to higher knowledge. Through a close observation of his own body’s rhythms, reactions to stimuli, sense perceptions, emo- tional responses, he attempted to develop a sixth sense, an extrasensory receptivity which he believed led to a state of superconsciousness.16 These methods and intellectual eforts led Kupka to describe the artist’s relationship to inner visions as follows—a description that provides a model for visualizing not only the intra-textual setup of the Sanskrit epic as containing additional temporal dimensions clariied by avataraṇa or descent but also the pan-Indic relationships that Sukthankar explored in the manuscript tradition: In our inner visions, the diferent fragments which loat in our heads are incoherently situated in space. Even in remembered so-called representative images of organic complexes, they are so strangely situated that the painter […] who would wish to project them would have to go even beyond the fourth dimension. Some parts penetrate each other; others seem completely detached, disconnected from the organism to which they are supposed to belong. The same is true of purely subjective visions where often only fragments, plexuses of forms, or colors are given. Before we can seize them and set them down, we must draw lines between them and establish a structural coherence.17 Kupka thus saw in the cosmic rhythms and repetitions a truth that the artist experienced in his visions, and it was the artist’s task to go beyond representing the objects given to the senses and rather depict the intellectual perception of the connections between the fragmentarily given sense data. Concomitant with this purely intellectual approach was a spiritual orientation, which included self-cultivation and a refusal to accept the crude empiricism of modernity. xv PROLOGUE xxv The creative solution Kupka adopted is not too diferent from the poetic solutions found in the epic. The frame narratives and the descent of the characters in the Mahā bhā rata represent a self-conscious repetition that organizes itself by organizing space: descent of gods and titans, or repetition in vertical space, which in turn organizes the ield of action (the battleield), the ield of transmutation (sacriice) and the ield of recognition (the forest). By extension, it is these repetitions and rhythms that guide the expansion and proliferation of the manuscripts, not mechanical and extrinsic “contamina- tion.” Every interpolation thus clariies the text by providing a new chromatic variation, and Sukthankar’s task can be seen as one of cataloguing and arranging the cosmos of manuscripts in the overall intellectual composition of his enterprise. Kupka, on his part, was aware of this intellectual-spiritual-artistic conceptual constellation. It was the spiri- tual worldview of many artists of his day. Maurice Tuchman summarizes this worldview as follows: The universe is a single, living substance; mind and matter also are one; all things evolve in dialectical opposition, thus the universe comprises paired opposites (male–female, light–dark, vertical–horizontal, positive–negative); everything corresponds in a universal analogy, with things above as they are below; imagination is real; and self-realization can come by illumina- tion, accident, or an induced state; the epiphany is suggested by heat, ire or light.18 It is hard to remember that Tuchman is speaking of the early abstraction movement in art, not about the Mahā bhā rata. It is even harder to believe that in the annals of Mahā bhā rata scholarship, not one scholar understands its creative elements so succinctly.19 The artist, it seems, is the very teacher of the epic. These insights ind their inest expression in The First Step, a painting Kupka executed during 1909–13.20 The painting contains a luminous black background, which evokes a pregnant darkness full of potential, and not merely a blackness of absence. The orga- nization of the various circles creates a map of space and dimensionality within the background, and thus demonstrates that the background is not non-being. The painting itself is a harmonious variation of a single form—the circle—echoed in its appearance or disappearance (in red), its concrete manifestation and endurance (blue and white) and inally its repetition and multiplication (a circumference composed of blue and red circles). The three processes of evocation, manifestation and multiplication create a com- plex sense of movement. Kupka was experimenting with motion at the time he painted The First Step, as were the Futurists. But the movement Kupka depicts is not mere physical movement, but a complex one of pulsations in existence. Its cosmological meaning was not lost on later commentators. Roger Lipsey comments on his Disks of Newton, a series of works to which The First Step was a prelude: Kupka’s transformation of color theory diagrams into a rotating, complex, genuinely spir- ited evocation of cosmos and light represents the high point of what might be called the naïve phase of his work, a phase of mobile search without the hardening that often occurs when answers are attained or, on the contrary, doubt gains the upper hand. The image moves freely and glows, conveying sensations of ease and pleasure. It is, as much as any painting, an Orphic work of strong poetic appeal; sunny and conident, pitched to the scale of the cosmos xvi xxvi PHILOLOGY AND CRITICISM and approachable. Kupka intersects here and is generally thought to precede Delaunay. Their paintings projecting the humble color wheel out into the cosmos constitute, to my mind, an undeniable manifestation of the spiritual in art. Perhaps neither closely reasoned nor metaphys- ically elaborate, they are nonetheless a celebration of cosmos that can leave few untouched.21 Besides the cosmological and spiritual meanings of the painting, Kupka endowed it with a critical evaluation of art. Kupka himself spoke of “a realm of rhythms and signs” to shed light on his art: We have to try […] to separate two incompatible elements, that is to say, the imitative work which today is superluous, from art itself. This is a realm of rhythms and signs too abstract to be captured easily and which form the leitmotif of all compositions, the basic arabesque, a kind of framework which the painters […] as of old ill with a vocabulary of forms taken from nature. If we sacriiced the intruding element we would of course have to face the danger of talking in an unusual language. Yet there is a kind of pictorial geometry of thought, the only possible one, which forces the painter to lie less. And that is what I am trying to achieve.22 Kupka’s painting thus illustrates not only a skill, or allegiance to a movement, or incremental innovation (invenzione) or design (disegno). Art can embody thinking, and precisely the kind of thinking that—while relected in a historical object—transcends history. In Kupka’s painting, art likewise transcends the universe and its coming- into-being and passing away. It does not abolish, but preserves the manifestation and repetitions of the universe. Each of the blue and red “instantiations” of the one con- cept, the circle, is diferent. These diferences are preserved, and yet their perfect procession and repetition and also their interaction (see the green ghostly circles) add to an overall sense of continuity in the cosmic order of the circle. The clariication of these existential movements, and the constant presence of a singular reality (here the circle) in the manifold “lies less” than the representation of a single concrete object. The painting discloses a profound truth: the truth of mimesis, the “lesser lie” that the truer existence is never what something is historically, but always a play of paradigms that transcend it.23 The reason for using this image by Kupka as the cover of a book on the study of Mahā bhā rata manuscripts is simple:  the Mahā bhā rata is a literary creation; it is art. In its materiality, it is of course created within history, but in its intellectual efort it transcends it. Both the painting and epic exploit the contextualization of the macrocosm with the microcosm to break free from the “literalism” of both. The First Step and the Mahā bhā rata are essentially “cosmological” works. Tuchman notes that The First Step is “a painting whose imagery is rooted in astrology and pure abstraction. The painting may be interpreted as a diagram of the heavens and as a nonrepresentational, antidirectional image referring to ininity and evoking a belief that one’s inner world is truly linked to the cosmos.”24 Like Kupka’s work, the epic is a cosmological work executed as a series of echoes: the intra-textual author Vyā sa’s conceptualization of the epic on the slopes of the axis mundi, Mount Meru; his teaching it to his students in an academic setting; one student’s (Vaiśaṃpā yana’s) repetition at the horriic scene of the sacriicial immolation of snakes; and the bard Ugraśravas’s (literally, the “he of the awesome voice”) recounting of the narrative xxivi PROLOGUE xxvii in the sylvan and peaceful assemblage of sages in the Naimiṣa Forest. The text itself presents these repetitions. Of another order are the repetitions of vignettes and motifs and messages in the various sectarian bibles: the Purā ṇas. A. K. Ramanujan ofers the best statement of the mimetic self-consciousness and the inbuilt mechanisms for trans- mission of the epic.25 Kupka helps us visualize Ramanujan’s insight, one that states that repetition and modulation of repeating elements is itself the structure of the epic: I’d suggest that the central structuring principle of the epic is a certain kind of repetition. One might say that repetition or replication is the central principle of any structuring. What occurs only once does not allow us to talk of structure. Einmal ist keinmal—it’s as if what happens once does not happen at all. Students of narrative like Propp, Levi-Strauss, Dumezil, and J.  Hillis Miller have made this idea a commonplace. Indian artworks, like the Hindu temple, or the decads (pattu) of Tamil classical or bhakti poetry, of the rā gas of Karnatak music, are built on the principle of interacting structures of repetition and elaboration and variation. Not only are there repetitive phrases, similes, and formulaic descriptions that the students of oral poetics (Parry, Lord, et al.) have taught us to recognize, but incidents, scenes, settings, and especially relationships are repeated.26 Kupka’s art self-consciously creates by using repeating patterns. For instance, in 1921 he painted the Hindu Motif, consisting of repetitions and modulations, abstractly recreating the architectural logic of the Hindu temple.27 This work paradigmatically illustrates his interest in Indian thought as well as his ability to recognize the repeating, abstract and symbolic qualities of Indian art. So much for the “external,” that is, formal aspects of mimesis as concerns the text. “Internally,” that is, with regard to the narrative and content, the mimetic nature assumes cosmological attributes. The text is presented as if it is a “his- tory” but the universe presented in this “history” is itself a mimetic object. The author enters the text and procreates the characters. Besides this literary duplication, there is a cosmological one: all the characters in the world described in the epic are “descended” from certain prototypes: gods and titans. The idea of mimesis plays a crucial and enduring role in the Mahā bhā rata and in the Indian textual tradition. Brahmā , the creator god, always creates the universe according to a paradigm, symbolically “given” to him by the One Being, called Nā rā yaṇa in the epic. The universe is always an artefact, created and recreated, endlessly in cycles. Coming-to-be and passing away is the ultimate indicator of the mimetic nature of our perceived and lived reality. And the epic is careful to present this repetitive cycle, rather than a naïve linear history: one that takes luxing time as a permanent framework. It is precisely by overcoming history that the epic “lies less.” Likewise, ideas of rebirth, lack of ultimacy of phenomenal reality and the soteriological presence of Being are ubiquitous elements with which all Indian philosophical systems grapple. Lurking behind the issue of any witness text of the epic are the usual problems ger- mane to all ancient texts, for example the Nibelungenlied from the German tradition or Euripides’s Iphigenia in Aulis in the Greek. These include textual variations and insertions and emendations, bequeathing to philology the task of coping with multiplicity in the textual tradition. But the Mahā bhā rata seems to anticipate and absorb these issues into its very composition. The question of a “lost” original is trumped not only by the various xxivi xxviii PHILOLOGY AND CRITICISM “versions” narrated within the text, but more seriously, any historical event is also divested of originality (the characters of gods and titans are merely enacting roles). In fact, the universe itself is a mimetic process, hardly a static object. Originality does not belong in the universe; it remains a feature of “unfallen” Being (Brahman). To seek either an original event (his- tory) or an original narrative (text) violates the epic’s understanding of itself. Those who seek an “original” Mahā bhā rata (as opposed to the original of any other text) are not like the blind men who variously represent the elephant as a snake or a pillar or a wall with respect to its various parts. They are the fools searching for a barren woman’s son. Sukthankar therefore carefully distinguished “older” from more developed forms of the texts, and discovered not an “original” but an archetype. The “archetype” in Kupka’s painting is not any particular circle but the concept circle, which is essentially abstract, and which “lies less.” The plural depictions of circles and their variations are essen- tial to the recognition of the concept. Similarly, the plural witness texts are “recogniz- ably” the Mahā bhā rata with respect to the archetype recovered in its critical edition. Sukthankar’s “critical” project negotiates between a method that prefers a fetish original to an actual text and the text’s obsessive disavowal of the category “original” in its lit- erary and its philosophical vision. Any great philologist can recover a most “ancient” text, but Sukthankar’s stemmatic arrangement of a plurality of texts as an astrolabe is the work of a philosophical and artistic genius.28 The critical edition does not replace the witness texts; it makes us more conident in appreciating them, and seeing them as singular/plural.29 Unfortunately, few have seen these abstract yet “less untrue” dimensions of the Mahā bhā rata critical edition project. This is not surprising. Mahā bhā rata scholarship has been ravaged by the crudest sort of butchers, untrained in philosophy and aesthetics and lacking the minimal sentience required to distinguish history from iction. Therefore the need for this book, which serves to remind scholars of the brilliance and rigor of the critical edition scholars’ work, and which hopes to teach the scholars of the future to appreciate the critical edition as a creative project of great subtlety, abstraction and truth that guides the thinker in the textual universe of the itihāsa purāṇa. Notes 1 Kupka, a pioneer of abstraction in art along with Piet Mondrian, Vasily Kandinsky and Kazimir Malevich, is less well known today than his peers. Yet he was one of the most impor- tant igures for the development of the movement known as “abstractionism.” See Ludmila Vachtová, Frank Kupka, Pioneer of Abstract Art, trans. Zdeněk Lederer, with an introduction by J. P. Hodin (New York: McGraw-Hill, 1968) and William R. Everdell, The First Moderns: Proiles in the Origins of Twentieth-Century Thought (Chicago and London:  University of Chicago Press, 1997). Kupka was also an inluential theoretician of twentieth-century art, expressing his views widely in articles and magazine interviews. Leighten considers his main work, La Création dans les arts plastiques (1912), “the central text of anarchist aesthetics in the modernist period.” Patricia Leighten, The Liberation of Painting:  Modernism and Anarchism in Avant-Guerre Paris (Chicago and London: University of Chicago, 2013), 15. Many of Kupka’s other writings can be found in the one-volume complete edition of untitled articles written between 1932 and 1936 for the journal Abstraction, création, art non-iguratif (Paris):  František Kupka, Abstraction, création, art non- iguratif, 1932–36 (New York: Arno Press, 1968). See also Pierre Brullé and Marketa Theinhardt, xi PROLOGUE xxix “Painting Despite Everything: František Kupka on Creation in the Plastic Arts,” in Painting the Universe: František Kupka, Pioneer in Abstraction, ed. Dorothy Kosinski and Jaroslav Anděl (Ostildern- Ruit, Germany: Verlag Gerd Hatje, 1997), 151–77 for a discussion. 2 Margit Rowell, ed., František Kupka 1871–1957:  A Retrospective (New  York:  Solomon R.  Guggenheim Foundation, 1975). This is the catalog for the exhibition Margit Rowell curated, for which Meda Mládek served as the consultant. Containing articles on the painter, his context and his contribution and a discussion of the formal and metaphysical aspects of his work, this volume remains the best resource on Kupka available in English. 3 A stemma (plural: stemmata) is a visual representation of genealogical relationships that takes the form of lines drawn between manuscripts and subfamilies of a work. It is also known as a textual tree or a genealogical tree and usually represents the descent of manuscript copies or apographs from their sources though it may also be used to diagram other sorts of relationships (for example, contamination between two manuscripts). 4 That is to say, the relationships themselves (for example, that a scribe A copied a manuscript a from source b) do not have any basis in matter; not that the relationships do not exist or that there is no material basis (manuscripts, etc.) for positing these relationships. 5 See, for instance, Otto Stählin, Editionstechnik, Completely Revised Second Edition (Leipzig:  B. G. Teubner, 1914); Louis Havet, Manuel de critique verbale appliquée aux textes Latins (Paris: Libraire Hachette, 1911); and Paul Maas, Textkritik (Leipzig:  B. G.  Teubner, 1927), with successive editions. (All references in this work are to the 4th edition of 1960.) 6 See Leah Dickerman, Inventing Abstraction, 1910–1925: How a Radical Idea Changed Modern Art (New York: Museum of Modern Art, 2013). 7 Meecham and Sheldon note that “years before the irst photographs of the earth from space, Kupka was painting what he believed to be ‘visions’ of the cosmos. Although Kupka never claimed that his ‘inner visions’ were any more than fragments which ‘loat in our heads,’ he believed that his clairvoyant vision lent him a transcendence which enabled him to survey the cosmos.” Pam Meecham and Julie Sheldon, Modern Art:  A Critical Introduction, 2nd edn. (Oxford: Routledge, 2005), 56–57. It is as such a visionary, who had a total, synoptic vision of the Mahā bhā rata tradition in his head, that we shall try to present Sukthankar here—and defend him against his critics. 8 Of course, we might have studied other Indologists. Extending our analysis back in time, we might have looked at the work of critics Edward W. Hopkins and Adolf Holtzmann Jr. or at Hermann Oldenberg. Likewise, we might have extended our analysis forward in time to study the work of James L. Fitzgerald or Georg von Simson or any other member of the so-called analytic school. The reason we did not do so is that their work has already been subject to a critique in our earlier book The Nay Science: A History of German Indology (New York: Oxford University Press, 2014). 9 Sukthankar received a BA degree in mathematics from the University of Cambridge in 1906; we do not know why in 1911 he went to Berlin to study philology, but he must also have con- tinued the association with Cambridge, for he received an MA from that university the following year. See S. M. Katre, “Vishnu Sitaram Sukthankar and His Contribution to Indology,” in V. S. Sukthankar Memorial Edition, vol. 2: Analecta, ed. P. K. Gode (Bombay: Karnatak Publishing House, 1945), 464. Incidentally, Katre also arrives at substantially the same assessment: “The scientiic training which Sukthankar received at Cambridge while preparing himself for the Mathematical Tripos, stood him in good stead during his Berlin days. Although he took up Indian Philology and Philosophy as his main branch of study, this Mathematical training pre- pared him for a scientiic outlook on matters literary or historical, and there was no study or investigation which he considered was low enough for a scholar if it led to proper utilisation of the material available.” Ibid., 465–66. 10 For an articulation of this insight within philology, see Sean Alexander Gurd, Iphigenias at Aulis:  Textual Multiplicity, Radical Philology (Ithaca, NY:  Cornell University Press, 2005), 188. Gurd’s “goal is to assess the realities involved in the multiple productions of a classical text x xxx PHILOLOGY AND CRITICISM so as to facilitate a literary philology alive to the fact of plurality. I call this a radical philology.” Ibid., 72. Gurd succinctly lays out various theoretical perspectives (Augustine, Marx, Foucault, Derrida, Goldhill, Finley, Page, Diggle, etc.) to show how the valorization of a singular original is a misguided fetishism. He recommends a more complex approach. “My central proposition is that critical texts are singular plural—that every single edition models and relects a plurality of other versions and variants—and that this singular plurality of the critical edition constitutes its sense.” Ibid., x. Gurd distinguishes the “core of textual criticism” from “the fetishism of the critical text” by analysis that “must oppose variability to stability, plurality to unity, and a con- crete to a nostalgic idealism.” Ibid., 35. 11 Theosophy provided artists of his generation an avenue whereby they could challenge the narrow deinitions of rationality and the dehumanizing materialism that were part of the Enlightenment’s legacy. For example, all four pioneers of modern abstract painting were inluenced by theosophy, but the list is quite extensive. Through the channels of philosophy opened up by Schopenhauer and Nietzsche, and through channels of art opened up by theos- ophy, a living dialogue of ideas progressed, in contrast to the obsession with historical facts and the pseudoscience of Indologie being forged in Germany. Their forensic science (who, why or what killed the Kṣatriya Urepos) remains one of the most spectacular blemishes on the human sciences to date. 12 Meda Mládek and Margit Rowell, “Chronology,” in František Kupka 1871–1957: A Retrospective (New York: Solomon R. Guggenheim Foundation, 1975), 306. 13 Increasingly in the 1930s and into the 1940s, mystical and occult beliefs came under suspicion because of their political associations, which were clear and well known. The Nazi theory of Aryan supremacy, for example, was indebted to various versions of theosophy, such as theozoology, which pertains to birth by electric shock into the astral ether, and ariosophy, which fuses ideas of karma, the ether and sun worship with idolatry of Aryan ancestry. “Adolf Hitler’s conidant Otto Wagener explained to Hitler the nineteenth-century occult writer Karl von Reichenbach’s theory of Odic force, according to which ‘every human being has an unknown source of power that produces rays. These not only inhabit the body but also radiate from it, so that a person is surrounded by something like a ield laden with this Odic force.’ Hitler immedi- ately applied these ideas to the potential reviviication of society by ‘the invisible strength which is transferred from them [storm-trooper divisions] to us like an aura.’ No doubt the perception of a link between alternate belief systems and fascism made critics and historians in these decades reluctant to confront the spiritual associations of abstract art. To use the word spiritual in the late 1930s and 1940s, as Richard Pousette-Dart recently acknowledged, was near-heresy and dangerous to an artist’s career.” Maurice Tuchman, The Spiritual in Art:  Abstract Painting 1890–1985 (New York: Aberville Publishers, 1986), 18. Albeit for diferent reasons, even today, Indology remains in the grip of political paranoia whenever spiritual aspects of Hinduism are presented. It is virtually taboo for scholars of Hinduism to engage in its theology, ontology or ethics except with great afectations of “critical” distance. 14 This was true not only in Europe but also in America. “The historian T.  J. Jackson Lears has recently argued that anti-modernism is the central notion unifying the leading American thinkers from the transcendentalists through Walt Whitman and William James. Modernism was regarded as something to be fought because it was synonymous with the loss of inner spiri- tual values. [William] James emphasized that the only way to attain true supremacy and higher consciousness was by losing oneself, by breaking down the conines of personality, and he pointed to the ‘immense elation and freedom as the outlines of conining selfhood melt down.’ In The Varieties of Religious Experience (1902) James stated, ‘Our normal waking consciousness as we call it, is but one special type of consciousness, whilst all about it, parted from it by the ilm- iest of screens, there lie potential forms of consciousness entirely diferent.’ He acknowledged that sensory, symbolic elements could ‘play an enormous part in mysticism.’ ” Ibid., 34. 15 While the Indological philologist earned his bread and butter by laboriously deconstructing texts to separate material “fact” from spiritual insight, the artist had already moved beyond this xxi PROLOGUE xxxi distinction, sensing that it was problematic and ultimately untenable. These echoes were felt as deeply as in Nietzsche’s The Gay Science, which declared the death of God (section 108) and as far away as Tahiti, where the French-born Paul Gauguin wrote: “The Word remains. Nothing of this Word is dead. The Vedas, Brahma, Buddha, Moses, Israel, Greek philosophy, Confucius, the Gospel, all exist. […] From a religious point of view, the Catholic Church no longer exists. It is now too late to save it.” Paul Gaugin, Gaugin’s Intimate Journals, trans. Van Wyck Brooks (Mineola, NY: Dover Publishing Company, 1997), 79. 16 Margit Rowell, “František Kupka:  A Metaphysics of Abstraction,” in František Kupka 1871– 1957: A Retrospective (New York: Solomon R. Guggenheim Foundation, 1975), 76. 17 Kupka, Manuscript II, 28, cited in Rowell, ed., František Kupka 1871–1957: A Retrospective, 77. 18 Tuchman, The Spiritual in Art, 19. 19 The banal, self-serving, self-congratulatory and unscientiic theories of the Indologists regarding the Mahā bhā rata have been suiciently discussed in our earlier book The Nay Science. Against the contributions of a genius such as Kupka these tired ruminations appear even more facile and pointless. 20 The Museum of Modern Art in New York acquired this painting in 1956 and it remains there to this day: The First Step, MoMA no. 562, 1956; 83.2 x 129.6 cm. 21 Roger Lipsey, The Spiritual in Twentieth Century Art (Mineola, NY:  Dover Publications, 2004), 101, 103. 22 Kupka, cited in ibid., 103. 23 Here also, referring to Gurd is useful. In his book, he makes “an attempt to characterize textual criticism as a ield deined by multiplicity and variation. It also contains a series of attempts to attend carefully and heedfully to each of its objects. Thus it shares its ambition with every other philological project. But if philology consists for some in training the vision ‘to see a whole landscape in a bean,’ I have tried to see each critical version pulsing with the rich plurality of many others: ‘the universe in a grain of sand.’ ” Gurd, Iphigenias at Aulis, ix. 24 See Tuchman, The Spiritual in Art, 36. Tuchman continues: “Years earlier Kupka had written of a mystical experience in which ‘it seemed I was observing the earth from the outside. I was in great empty space and saw the planets rolling quietly.’ ” 25 A. K. Ramanujan, “Repetition in the Mahā bhā rata,” in Essays on the Mahābhārata, ed. Arvind Sharma (Delhi: Motilal Banarsidass, 2007), 419–43. 26 Ibid., 421–22. 27 This painting, also known as Graduated Red, was completed between 1919 and 1923. It is cur- rently in the collection of the Musée National d’Art Moderne, Centre Georges Pompidou, Paris, France. It features on the cover of our edited volume Vishwa Adluri and Joydeep Bagchee, eds., Argument and Design: The Unity of the Mahābhārata (Leiden and Boston, MA: Brill, 2016). 28 Sukthankar was aware of these aspects of his project. He explicitly notes that “the Mahā bhā rata is the whole of the epic tradition: the entire Critical Apparatus. Its separation into the consti- tuted text and the critical notes is only a static representation of a constantly changing epic text—a representation made for the purpose of visualizing, studying and analyzing the pan- orama of the more grand and less grand thought  movements that have crystallized in the shape of the texts handed down to us in our Mahā bhā rata manuscripts.” V. S. Sukthankar, “Prolegomena,” in The Ad̄ iparvan for the First Time Critically Edited, vol. 1 (Poona: Bhandarkar Oriental Research Institute, 1933), cii (Sukthankar’s italics). German Indologists, no less than Sukthankar’s Indian detractors, have therefore misunderstood him when they suggest that Sukthankar reduced the Mahā bhā rata to and/or extracted a core text. Actually, he preserved all its available versions, creating the superset of Mahā bhā ratas and thus a text embodying, more than ever, its claim:  yadihāsti tadanyatra yannehāsti na tatkvacit (“whatever is here […] that is found elsewhere. But what is not here is nowhere else”; Mahā bhā rata 1.56.33cd and 18.50.38cd). 29 The critical edition does not eliminate the need to look carefully at witness texts. These serve again and again to reine and gloss over the more “archaic” material recovered by archetype. xxxii newgenprepdf xxxii PHILOLOGY AND CRITICISM To give one example, Hudson’s spectacularly erroneous reading of the Mahā bhā rata (see Emily T. Hudson, Disorienting Dharma: The Aesthetics of Sufering in the Mahā bhā rata [New York: Oxford University Press, 2013]) could easily have been prevented had she looked at the vulgate, which contains the Kaṇikanīti, an insertion of 230 lines in the vulgate that is moved to the appendices (App. 1, no. 81) in the critical edition. See Vishwa Adluri, “Ethics and Hermeneutics in the Mahā bhā rata,” review of Disorienting Dharma: The Aesthetics of Sufering in the Mahā bhā rata, by Emily Hudson, International Journal of Hindu Studies 20, no. 3 (2016): 385–92 and, in still greater detail, Vishwa Adluri, “Hindu Studies in a Christian, Secular Academy,” International Journal of Dharma Studies 5, no. 1 (2017), doi:10.1186/s40613-016-0037-5. 1 CHAPTER SUMMARIES Introduction: Ad Fontes, Non Ultra Fontes! About This Book The aim of this book and its connection with our irst book; the central problem confronting Mahā bhā rata studies Why a Critical Edition? Why a critical text is required and what problem it attempts to solve What Is a Critical Edition? A description of the critical edition:  its components, how it reduces the plurality of readings to one and what the status of the resultant text, the constituted text, is. Three misconceptions about the critical edition: (1) it is eclectic, (2) it is not a text and (3) it can be replaced by a text with an apparatus of variants How to Interpret the Critical Edition The text reconstructed in the critical edition is the archetype of the tradition, deined as the latest common ancestor of the manuscripts examined for that edition. This sense of archetype should not be confused with the archetype as an especially authoritative or unique exemplar, for our stemma is merely hypothetical and models only a part of the historical reality—the part that is either preserved in or can be reconstructed from our manuscripts. Conclusion The hypothetical ancestor of our manuscripts was probably one of several exemplars in existence at the time. It is solely by chance that only its descendants, rather than those of other manuscripts, survived, resulting in its apparently unique position in the history of the text. This uniqueness is only apparent: it is a consequence of the fact that our stemma models only a part of the history of the text. From the apparently unique nature of the archetype, we may not conclude there was an actual reduction in the number of exemplars at the time. A fortiori all theories that attempt to explain the reduction in terms of the actions of putative “redactors” at the time are false. The idea of a conscious redac- tion (of oral epic materials) arises only because some scholars do not know how to read the stemma correctly. 2 2 PHILOLOGY AND CRITICISM Chapter One: Arguments for a Hyperarchetypal Inference The Normative Redaction Hypothesis The constituted text is not the archetype of the tradition but merely a “normative redaction,” deined as “a redaction that had a normative efect and overgrew all other versions.” The critical edition reconstructs “a text that was a historical fact at a certain period in time,” but precisely because it is not the archetype, it should not hinder us from exploring “the prehistory of the normative redaction.” Above all, we should consider “passages rejected from the constituted text,” as they could be evidence of a “parallel transmission” of the Mahā bhā rata. Normative Redaction, Archetype and Original Textual criticism allows us to reconstruct the archetype of the tradition, which represents a constriction in the tradition attributable to a “normative redaction.” This reconstructed archetype, however, only gives us access to the oicial Brahmanic text resulting from the redaction of an earlier oral tradition. It neither accurately models the contents of the tradition nor can it be seen as a copy of the original, since the tradition was plural above the archetype and a single original never existed. Criticism: Higher and Lower The redeinition of the constituted text as a normative redaction rather than an arche- type permits us to reconstruct earlier stages of the tradition using “higher criticism.” In contrast to textual criticism, which is a rigorous and mechanical procedure that begins with the manuscript evidence and attempts to infer the manuscripts’ likely sources based on shared errors of transcription, higher criticism uses subjective, a priori criteria to identify certain passages as older than others and therefore as part of the “genuine” epic tradition. The Argument from Spread and the Argument from Resilience Textual criticism only permits us to reconstruct the source of the irst branching. However, this ignores the possibility that the tradition was plural before the apex. The reconstruc- tion of the archetype is thus an error, arising from the fallacious assumption that all readings derive from a unique text (argument from spread). Our reconstructed archetype is based on readings that passed through the constriction between the two cones, whereas readings that fell directly from the upper cone to the lower without passing through the apex could be older (argument from resilience). Refutation of the argument: The argument from spread fails because it overlooks that the shape of the tradition before the archetype is irrelevant for the reconstruction of the archetype, which remains ex hypothesi the oldest ancestor of the extant witnesses that can be reconstructed. In fact, the archetype makes no claims about whether the tradition spread away from the original or all readings fell through one particular manuscript. 3 CHAPTER SUMMARIES 3 It only claims that all our manuscripts are descended, however remotely, from this one source. Likewise, the argument from resilience fails because it erroneously infers the exis- tence of a real constriction in the tradition from the archetype’s apparently unique posi- tion in the stemma and further argues that some readings at least could have escaped the archetype’s constricting efect. The Argument from Empty Reference The critical edition contains references to events, whose description no existing passages meet. However, these “empty references” must once have had a referent. If suitable descriptions occur in the apparatus, we may assume that they were the original referents, contained in a version of the Mahā bhā rata distinct from the normative redaction. The critical edition discards these passages as unique to one branch of the tradition, but this is false because they are actually older than the normative redaction it reconstructs. The normative redaction eliminated the passages, but they later reentered the tradition, thus occurring in some manuscripts only. Restoring the referent to the constituted text restores the text to a state before the normative redaction. Refutation of the argument:  The argument sufers from the fallacy of proof by assertion, because it is tantamount to asserting that a passage A is old because it is con- sidered old. Its fallacious nature is not immediately apparent because, rather than state it simply, Bigger asserts that A is actually Rʹ, the referent of passage R, whereas the reading R → Rʹ reconstructs an original R → Rʹ; as the surviving referent of this original refer- ence Rʹ is thus old. The equivocation between Rʹ as a passage in the extant witnesses thought to correspond to a reading R (that is, Rʹ1) and Rʹ as the referent of an original R (that is, Rʹ2) permits him to “infer” Rʹ’s antiquity, even though, actually, he has only asserted it. The empty reference plays no role in demonstrating A’s antiquity; it merely conceals the illegitimate shift from Rʹ1 to Rʹ2. The Argument from Loss The argument from loss entails the claim that certain passages, though contained in the Mahābhārata’s hypothetical “normative redaction,” were lost from part of the subsequent manuscript tradition. They are therefore not found in the constituted text, although, as typical of Brahmanic concerns, they should have been included in it. Refutation of the argument: The argument from loss replaces a critical principle— agreement between independent families—with a subjective and prejudicial principle for reconstruction. It uses this dogmatic principle—passages considered “Brahmanic” should be restored to the constituted text, even if they lack manuscript support—to assimilate the constituted text to a hypothetical Brahmanic redaction. It then uses this restored text, in turn, as evidence for the reality of the Brahmanic redaction. Given its circularity, the argument should persuade no one. That it does illustrates the truth of the observation that if “the circle that has been constructed [is] large and confusing, […] the logical mistake goes unseen.” 4 4 PHILOLOGY AND CRITICISM Chapter Two: Reconstructing the Source of Contamination Understanding “Contamination” Mahābhārata critics understand “contamination” not as textual critics do (for whom it refers to the addition of readings from a second source besides the one the scribe copied), but as the interpolation of Brahmanic elements into an original Ksạ triya epic. They thus interpret contamination to mean a kind of pollution (namely, with the Brahmans’ religious ideology). Contamination: Hyperarchetypal and Extra-stemmatic This section distinguishes between extra-stemmatic contamination, intra-stemmatic contamination and hyperarchetypal contamination. We examine Bigger’s view that the hypothetical Kṣatriya epic underwent contamination from Brahmanic sources prior to the formation of the archetype (hyperarchetypal contamination) and that the resulting Brahmanic text then itself underwent contamination beneath the archetype with remnants of the Kṣatriya epic that survived either in folk traditions or in the poets’ memory (extra-stemmatic contamination). Identifying the Source of Contamination This section contrasts the Mahābhārata critics’ arbitrary and tendentious arguments for Brahmanic “contamination” with objective criteria for identifying the source of con- tamination. We show that the minimum condition to identify interpolations is a stemma of the form 3+1, where three manuscripts descend from the contaminated source and a fourth descends from an independent source. If any two manuscripts descended from the contaminated source agree against the manuscript descended from an independent source, then their reading is the reading of the source of contamination. The Argument from Uncertainty Mahā bhā rata critics often try to undermine the stemma by positing contamination from a nonextant oral source (extra-stemmatic contamination). They claim that as our stemma does not accurately represent historical reality, we cannot deinitively exclude earlier, nonextant sources. Refutation of the argument:  The argument from uncertainty fails because intro- ducing uncertainty into a system afects all outcomes equally. Introducing uncertainty into the stemma does not make it likelier that a Kṣatriya epic existed. It only appears to favor the Kṣatriya epic because the critics assume its existence in advance and that they can know its contents without a stemmatic reconstruction. The Argument from Oral Source If the oral tradition survived past the Brahmans’ seizure and destruction of the original epic, elements from it could have trickled back into the manuscript tradition. This could 5 CHAPTER SUMMARIES 5 have occurred if a scribe either recollected an oral version or possessed a transcript of it. In that case, we are justiied in thinking that some of the readings in our manuscripts (and mutatis mutandis the constituted text) are older than others, and some at least as old as the oral epic. Refutation of the argument: The argument does not address the question of how we can identify the alleged remnants of the oral epic. Every manuscript contains readings of varying antiquity but, without the stemmatic method, we cannot sift between them. The argument also does not address the problem that the alleged remnants could be later innovations or inaccurate recollections. Indeed, in a long chain of transmitters such as an oral tradition presumes it is especially likely that the readings underwent deterioration. The analytic critics fail to perceive the problem because they focus on passages rather than readings, and do not ask how, even if we grant that the passages were original, their readings could have been transmitted unchanged. The Argument from (Postulated) Antiquity and the Argument from Ideology This concluding section examines two related arguments, the argument from (postulated) antiquity and the argument from ideology. The irst refers to the Mahā bhā rata critics’ habit of declaring that certain passages are old because they are postulated as old. The second refers to their ideological reasons for doing so, namely, proving the baleful nature of Brahman domination for India. Refutation of the argument: Neither argument is actually an argument. The argu- ment from (postulated) antiquity merely asserts the antiquity of certain passages without demonstrating it. The argument from ideology elevates the German scholars’ anti- Semitic prejudices to a irst principle and uses this, in turn, to sustain the kind of circular, counterfactual and logically fallacious arguments we have seen. These prejudices hold the key to understanding the German scholars’ work, speciically their insistence on a tendentious Brahmanic redaction of an earlier heroic epic. Chapter Three: Confusions Regarding Classiication Classiication: Typological and Genealogical Many criticisms of the Mahā bhā rata critical edition arise from a misunderstanding about how editors classify manuscripts, how they determine iliation and how they reconstruct— via eliminating either the readings of witnesses or entire witnesses—the archetype. Crucially, arguments for diferent classes of manuscripts (or entirely new “recensions”) are based on the erroneous assumption that because a group of manuscripts lacks certain passages characteristic of other manuscripts, the former constitute a family, even though the dichotomy is merely typological, not genealogical, and as yet no iliation has been established between the manuscripts so separated. 6 6 PHILOLOGY AND CRITICISM Determining Filiation Filiation can be established only through shared errors, which permit us to identify two manuscripts as more closely related than others of that family (all of which will contain the same text, but not the same errors, which are unique to this branch of the tradition). It is hence incorrect to establish iliation on the basis of shared readings, as they identify the two manuscripts only as members of the family chosen for study (manuscripts of the Mahā bhā rata), but do not permit us to deine them as a speciic branch of that family— manuscripts descended from the irst source of the error or errors. Eliminating Witnesses A comparison of the constituted text with variant readings or passages rejected from it will not permit us to establish an earlier stage of the transmission. At most, we might establish that a reading rejected by the editor is preferable and restore it in our text, but this text will still be, by deinition, the archetype and not a supra-archetype of the kind critics think they can produce by randomly selecting passages from the appendices (on the grounds of their presumed “Kṣatriya” origins). The Argument from Brevity and the Argument from False Premises In creating the critical edition, Sukthankar followed the evidence of the shortest witness, the Ś ā radā codex, and reprinted the readings of this manuscript as his constituted text. However, as the Ś ā radā codex was not the shortest witness (for the Nepā lī manuscript NAK 5/ 356 is shorter by 63 passages), the critical edition must be begun anew. Refutation of the argument: This conclusion would be justiied only if the editor actu- ally reprinted the readings of the Ś ā radā codex as his constituted text. In point of fact, he compared manuscripts of diferent classes to reconstruct the reading of the archetype. He showed in several cases that the Ś ā radā codex itself contained interpolations and additions. This means that what is decisive is not the length of the Ś ā radā codex over the Nepā lī manuscript NAK 5/356, but whether passages in the constituted text exist, which are not attested in the latter, something Grünendahl (who raises the objection) has not shown. The Argument from a Misapprehension Concerning Classiication ( Schriftartprämisse) The Mahā bhā rata manuscripts were classiied into versions on the basis of their script, according to a principle Lüders irst articulated in 1908 (the script is characteristic of the version). However, this classiication is erroneous, inasmuch as the script proved not characteristic of the version and neither the variant readings nor the additional passages were consistently characteristic of speciic versions. Refutation of the argument: The Mahā bhā rata manuscripts were classiied not by their script, but by their text. The term version is a synonym for a manuscript’s text. If the editors nonetheless looked to the script as a irst, extrinsic indicator of a manuscript’s 7 CHAPTER SUMMARIES 7 text, this is because, in practice, it was often a reliable guide to the text contained in a manuscript. The editors were not blind to the circumstance that sometimes a manu- script, contrary to their expectation of it, could contain the text of another version. They often reassigned manuscripts assumed to belong to one version on the basis of their scripts to another (whose members were all in a diferent script). If the editors had merely looked to the script they could neither have drawn up a stemma nor reconstructed the reading of the archetype. The Argument from Extensive Contamination The widespread presence of contamination between Mahā bhā rata manuscripts (evinced by the existence of interpolated passages across recensions) makes a satisfactory classiication impossible. Refutation of the argument:  In practice, no editor found contamination such a problem that he could not establish the manuscripts’ true iliation. Though present, con- tamination was restricted largely to the additional passages, which, indeed, were often transmitted horizontally. The manuscripts’ true iliation almost always became obvious in terms of the signiicant errors in the text. Grünendahl errs because he considers only the interpolated passages rather than readings. His understanding of contamination is also peculiar: according to him, the absence of interpolation is also a kind of “contami- nation” (namely, non-contamination), for how (he reasons) could two manuscripts agree in featuring the same text if not for the fact that one of them “contaminated” the other? The answer is simple: the manuscripts contain the identical text, because the same text was handed down to them and no inluence of the one upon the other needs to be pre- sumed to explain the circumstance that they lack the same interpolations. Interpolations, except when interpolated, tend not to be present, so that no special circumstance is required to explain their absence in a manuscript. The Argument from Independent Recensions Since widespread contamination between Mahā bhā rata manuscripts makes it impossible to draw up a stemma, the editors should have concentrated on a regional recension and produced an edition using one manuscript, preferably the shortest, as the base. Refutation of the argument: The argument overlooks the fact that, even if the editors had chosen to produce a critical edition of one of the regional recensions, they could not have done so without considering its place in the overall tradition. Further, if contam- ination makes it impossible to produce a critical edition of the Mahā bhā rata tradition as a whole, this is especially true of a critical edition of a regional recension, since con- tamination was most pronounced between the manuscripts of a recension. Grünendahl’s proposed edition (based on the Nepā lī manuscript NAK 5/356) would be not a critical edition but a pseudo-Bédierist edition. He could neither draw up a satisfactory stemma nor proceed in cases of crux or where the Nepā lī manuscripts contained inconsistent readings. His edition opens the loodgates for conjectural criticism, which is all he seeks to do. 8 8 PHILOLOGY AND CRITICISM The Argument from Expertise The argument from expertise refers to the replacement of argument with expert tes- timony. The mere fact of citation, without an attempt at paraphrase much less clar- iication, serves to validate an author’s views. Rightly speaking, the argument from expertise is not an argument at all. It merely relects the institutionalized prejudice in favor of certain authors and positions. It reveals the broken system of preferment at Mahā bhā rata studies’ core. The scholars entrusted with discriminating between correct and incorrect scholarship have proven incapable of distinguishing objective research from ideological views. 9 10 Dr. V. S. Sukthankar (1887–1943), scholar, mathematician, Sanskritist, editor and the archi- tect of the Mahābhārata critical edition. Source: Reproduced from Katre, “Vishnu Sitaram Sukthankar and His Contribution to Indology,” 462. 1 INTRODUCTION: AD FONTES, NON ULTRA FONTES! Many believe stemmata to be accurate depictions of the historical vicissitudes of transmission, but it is the few who believe that this is not the case who are right. —Paolo Trovato, Everything You Always Wanted to Know about Lachmann’s Method About This Book The aim of this book and its connection with our irst book; the central problem confronting Mahābhārata studies The Mahā bhā rata critical edition, which was begun in 1931 and completed in 1966, is now more than ive decades old. This is an appropriate moment to revisit the history of this edition and its reception. It is also an appropriate moment to create the scholarly tools and auxiliary materials required to use this edition, something the edition’s creators envisaged but never carried out in the past.1 Philology and Criticism addresses this need by discussing criticisms and misconceptions of the critical edition. It also carries forward the criticism of German Mahā bhā rata studies undertaken in The Nay Science, whose unscientiic and ideologically motivated theories of the Mahā bhā rata pose the greatest single obstacle to the correct reception, interpretation and use of the edition.2 As will become clear, contem- porary objections to the critical edition are essentially restatements of theoretical positions German critics staked out in the nineteenth century.3 Addressing these objections, there- fore, requires us to understand the original reasons for these claims, as well as to grasp the reasons scholars continue to resurrect them. Why a Critical Edition? Why a critical text is required and what problem it attempts to solve To a scribe working in the thirteenth century, the question of which Mahā bhā rata man- uscript to copy would have been an easy one to answer: as a rule, he would not have had access to more than a few manuscripts, mostly descended from the same source, and so, barring a few variants, he would have had little to choose from among them. He might have chosen the one most easily available to him, or the most complete, or the best pre- served, and produced a copy, inserting the variants of other manuscripts (if he chose to consult them at all) in the margins or perhaps on separate sheets (the śodhapatra, as they 12 12 PHILOLOGY AND CRITICISM are called). Occasionally, a scribe working in one of the centers of Hindu intellectual life (like the seventeenth-century commentator Nīlakaṇṭha) might have had access to a larger number of manuscripts, indeed, to an overview of the manuscript tradition from across the subcontinent. In that case, the scribe, being more concerned with preserving the tra- dition and regarding himself as a part of a living tradition entitled to make selections and to emend and rearrange the texts, might have selected the best narratives, creating a composite manuscript that would, in turn, have generated further copies, establishing a new tradition. In time, new verses might have entered the tradition through the inevi- table processes of exegesis and commentary or new episodes might have been composed for places where narrative lacunae were felt, but this augmentation would not have been considered a problem: it was evidence of the tradition’s vitality, of its ability to renew and translate itself for a new circle of readers each time. The irst printed editions of the Mahā bhā rata continued this practice by reprinting the available manuscripts.4 Their editors were not concerned with establishing a critical text. They saw themselves as part of a living tradition and their sole concern was to make available in the new medium what previously existed only in the manuscript tradition. Besides employing teams of learned śas̄ tris or paṇḍits to check and possibly emend the source manuscript, they were not overly concerned with the text’s accuracy or with rec- onciling variations between their edition and the other printed editions. A new set of concerns emerged in the late nineteenth century, as the Mahā bhā rata became an object of specialist concern: What is “the Mahā bhā rata”? What was the oldest form of the text? Between two competing versions, which one must be judged more authentic? It was partly to resolve these questions and partly to bear out their own theories about an original epic (the so-called Urepos) that calls for a single, sci- entiically validated text arose.5 As Moriz Winternitz, an Austrian Indologist and a leading advocate for a Mahā bhā rata critical edition, declared, “What we really need, and what seems to me to be the sine quâ non for historical and critical researches regarding the text of the Mahabharata, is a critical edition which should neither satisfy the people of Northern India nor those of the Dekkhan, but which should satisfy the wants of Sanskrit scholarship. I repeat what I said at the last Congress in Paris, that ‘a critical edition of the Mahabharata made by European scholars according to the principles followed in editing any other important text, is wanted as the only sound basis for all Mahabharata studies—nay, for all studies connected with the epic litera- ture of India.’ ”6 What Is a Critical Edition? A description of the critical edition: its components, how it reduces the plurality of readings to one, and what the status of the resultant text, the constituted text, is. Three misconceptions about the critical edition: (1) it is eclectic, (2) it is not a text and (3) it can be replaced by a text with an apparatus of variants. A critical edition circumvents the problem of multiple versions of a text by subjecting the diferent readings of the available manuscripts (called variae lectiones or variant 13 INTRODUCTION 13 readings) to a rigorously hierarchical procedure:  readings thought more likely to represent the reading of the ancestor of all manuscripts examined for the edition (known as the archetypus or archetype) are printed above the line as the critically consti- tuted text (constitutio textus), while the remaining variants (which hence are considered “corruptions” of the former) are printed below the line (where they form the apparatus criticus or critical apparatus of the edition). A critical edition thus creates an overview of the entire tradition, assigning the available readings a speciic place either above or below the line depending on how archaic they can claim to be. Reading the stemma codicum (the genealogical tree of manuscripts) from top to bottom provides an overview of everything the tradition contains—not only the inferred (that is, hypothetical) texts posited as having existed on the basis of the available manuscript evidence but also the actual, physically extant manuscripts that remain our only source of evidence for the tradition. A critical edition thus represents the easiest and most elegant way to arrange the available information. As Gianfranco Contini observes, “a critical edition is, like any other scientiic act, a mere working hypothesis, the most satisfactory, namely, the most economic one, and one which proves apt to connect a system of data.”7 This view has replaced the earlier tendency to regard a critical edition as a facsimile of an existing text, that is, either the text of the author’s hand, the so-called autograph, or a copy of this text, that is, an apograph that was the irst source of the surviving tradition. This tendency became unsustainable not only because to maintain that the reconstruction corresponds in all its particulars to an actual text requires a leap of faith but also because of the pos- itive evidence that not all elements of the constituted text are of the same antiquity. As V. S. Sukthankar observes, the Mahā bhā rata critical edition “precisely like every other edition […] is a mosaic of old and new matter. That is to say, in an average adhyaya of this edition (as of any other edition) we may read a stanza of the second century BC followed by one written in the second century AD. Sometimes the gap will occur in the middle of a line, precisely as in every other edition.”8 In fact, every line or every word of the genealogical-reconstructive edition or the reconstructive edition, as we shall refer to it henceforth, occurs at a point on a continuum that we, along with Lino Leonardi, might describe as anchored between the text and its extant manuscripts that attest to or witness to the text (hence called witnesses). In each case, a line either “[goes] back to a stage of the transmission as close as possible to the original text” or it “[represents] one or more outcomes of the process of transmission materially attested in the surviving tradition.”9 See Figure 1. a b c ... n text witness Figure 1 The two options of a philology oriented toward the text and a philology oriented toward the witness Source: Modiied from Leonardi, “Il testo come ipotesi,” 7. 14 14 PHILOLOGY AND CRITICISM A critical edition is thus the living image of the text’s diachronic history, with each of its elements corresponding to a deinite stage of the text. This will be either the archetype of the tradition in the cases that we can reduce the plurality of readings to one, or one of its descendants in the cases that we can unify only some branches of the tradition and encounter a crux (that is, two competing readings, either of which could be original). In all cases, the critical edition will exhibit greater historical continuity (pro- vided the editor has read the evidence correctly) than any of the competing editions, simply because of its concern with accurately modeling the history of the text. Thus, the text will be continued for the most part at the same level, namely, at the level of the archetype, except where the editor could not determine the reading that is the ancestor of all the others, at which point it will descend to the level of one of its hyparchetypes.10 But even in respect of these passages the constituted text is less eclectic than any of its witnesses, since it prints the next most proximate reading, rather than permitting the fall to continue uncontrollably through centuries. As Michele Barbi notes, “It is not a matter of treating as certain what is only probable, or forcing everyone to accept our conclusions; but when one has illustrated to scholars the state of things and the reasons for and against a given conclusion, has one not done what science allows and demands? […] Why should we give up considering each case for itself, and why should we not be allowed to substitute our cautious and reasoned judgment for that of a tran- scriber whose judgment […] we do not see the reasons for and cannot measure the extent of ?”11 Understanding a critical edition in this more sophisticated way also permits us to clarify some pervasive misconceptions of the Mahā bhā rata critical edition. The critical edition is not an artiicial text, a “Frankenstein’s monster, pieced together from various scraps of diferent bodies,” as some suggest.12 It is, rather, a particular arrangement of textual materials (as every edition is) undertaken to expunge centuries of scribal error and variation, and to provide as close an approximation of the original text as possible.13 It is furthermore a rigorously scientiic text in that it follows a rational logic and that each of its steps is clearly documented. Contrary to the charge that it creates a new text, one that lacks either an organic community or continuity with the tradition, every line of the reconstructive edition is validated by the tradition.14 Recall that each line of the consti- tuted text occurs at a point on a continuum deined by the two ends of the text and its witnesses. If the editor can resolve the diferent manuscript readings and derive them from the presumed original, the resultant text is that of the author’s hand or one as close to it as possible. It is thus better and more accurate than any of the surviving variants.15 If the editor cannot do so, the reading he prints is nonetheless materially attested in one of the witnesses. More likely, it will be attested in several since it represents the consensus of at least one branch of the manuscript tradition. The reading is also better than the alternatives, which the editor after careful evaluation discarded. In both cases, the con- stituted text represents a better text than available in any of its witnesses.16 A verse found in a Telugu manuscript, for instance, has only that manuscript to attest to it. In contrast, every line of the critical edition of the Mahā bhā rata is of greater antiquity than any of the surviving witnesses. And while the constituted text of a critical edition is eclectic in 15 INTRODUCTION 15 the sense that it preserves readings of difering antiquity, it is no more so than any other edition. The argument can be made that it is less so because, as Sebastiano Timpanaro describes, “to this random and irrational eclecticism [of the tradition],” “we must oppose our choice, which is based on rational judgment and therefore is not eclectic in the pejo- rative sense.”17 The charge of Frankensteinism is thus misguided. In contrast to the transmitted texts of the tradition, which are composites of the author’s intention and what the copyist understood or chose to write down in the speciic case (and this over centuries), the crit- ical edition ofers a more legible and scientiic text.18 In his activity, the editor seeks to eliminate conlated sources (manuscripts copied from two or more sources, which are truly hybrid editions), he disentangles readings of doubtful or difering antiquity that have been commingled through the scribes’ activity, and he attempts to put in place, as much as is possible, materials of similar antiquity. He does so, furthermore, using all the knowledge and judgment about the tradition at his disposal.19 And while his edition is an interpre- tation of the text, it is no less naïve to imagine a copyist faced no interpretive choices when transcribing his exemplar than it is to maintain “we can and must edit [recensere] without interpreting” (Lachmann). In the absence of a critical procedure, the choice of one version over another, no matter how justiiable in terms of the text’s contents, must remain arbitrary.20 The emphasis on eclecticism thus obscures the real question: which is the more scientiic and more legible text and which is the truer image of the tradition’s diachronic reality? Inverting Doniger’s metaphor, we could rather say the editors of the Bhandarkar Institute were like doctors who took of bandages and laid bare the patient’s sores, bringing to light what had been concealed and restoring what had been incorrectly joined or separated.21 A second pervasive misconception that is the mirror image of the irst is that the Mahā bhā rata critical edition is merely an “Arbeitsinstrument” or working tool.22 This mis- conception, found most often in the work of the German Mahā bhā rata critics whose Protestant makeup disposes them against the very idea of a text,23 arises from a misun- derstanding regarding the nature of the editor’s activity.24 To the critic, it may appear as though textual criticism’s primary purpose is merely to provide “rational procedures for advancing the most likely hypothesis about what the original was like and how, in rough outline, it was transmitted down to its preserved witnesses.”25 But textual criticism also seeks to provide “an edition of an ancient text that is accessible to the modern reader and, at the same time, conforms to the intention of its author.”26 Without this aspect, the editor’s activity would be diminished. It would be reduced to proposing hypotheses, without contributing in any way to the text’s preservation. It would lead to the rad- ical separation of the critic’s activity from the text’s readers, a possibility that no one can seriously contemplate.27 As Leonardi reminds us, “The concept of a critical edition [also] implies a responsibility to propose a text that, while meeting the requirements of science and elucidating the manuscript tradition, is not merely accessible to specialists, but recovers a work of the past for a contemporary public to read; one that does not reproduce a document, but interprets the tradition as a whole as a means to transfer its textual reality into something that is readable today.”28 What was said earlier about the 16 16 PHILOLOGY AND CRITICISM critical edition ofering the most scientiic, the most legible and the most transparent text (in terms of its diachronic history) also applies here: to say the critical edition ofers the best possible text implies a concurrent responsibility to read the constituted text and not merely regard it as an object of specialist investigations. A text that cannot be read is not a text. A third misconception consists in thinking the Mahā bhā rata critical edition merely provides a text of the Mahābhā rata along with an apparatus of variants. As we have already dealt with the misconception that the critical edition provides just a text, we focus here on this statement’s second half. This misconception is evi- dent in Sylvain Lévi’s suggestion to Sukthankar, made in a 1929 review of the irst fascicule of the critical edition, that he could just print the text of the vulgate (the edition of Nīlakaṇṭha, called the vulgate, because it is the most widespread) along with a list of the variants in other manuscripts.29 The problem with Lévi’s proposal, which he apparently advocated to reduce the scope of editorial judgment (that is, the use of subjective iudicium), is twofold.30 First, even if the editor is not supposed to select the best manuscript (the bon manuscrit, as it is called) and is supposed to merely reprint the vulgate rather than edit it, the proposal does not avoid the problem of choosing between variants. Which version of the vulgate should the editor reprint? As Sukthankar observes, the printed editions are inferior to the manuscripts of the vulgate.31 While the editor could, in theory, correct the printed editions using the manuscript sources, such an approach is hardly warranted, as Sukthankar notes. “It would, however, hardly repay, now, the trouble to reedit, from manuscripts, the version of Nīlakaṇṭha, as there are far better versions that could be edited instead, for instance, the Kaśmīrī.”32 Correcting the printed editions would, furthermore, reintroduce the necessity of making editorial choices.33 Second, and more signiicant, Lévi seriously underestimates the value of a critical edition. A critical edition, as we saw, is not just a text with an apparatus of variants. It is, rather, the living image of the text’s diachronic history. By considering on what manuscripts a reading is based, the reader can intuit not only the attestation for it but also its relative antiquity and authenticity. Inversely, by seeing how far back a variant or an episode can be traced, she can infer its date or even its geographic origin. A critical edition represents the editor’s best understanding of the tradition, based on his years-long digestion of the manuscript evidence. Granted an average reader may never read a critical edition in this nuanced way. She may never, for instance, appreciate that when a reading is attested in both the northern and southern recensions of the Mahā bhā rata, it is unambiguously the reading of the archetype. But the edition itself contains much more information than a mere list of variants would: it contains information on the families to which these variants belong, the classiication and genealogical relation- ship of manuscripts and the codicological weight of individual variants. In contrast, a mere list of variants—without any attempt to organize the information, that is, to classify the manuscripts into families, to group variants and to subordinate some of them to others—contains no information beyond the mere fact of the existence of variants.34 It can hardly be considered a substitute for the true critical edition.35 See Figure 2. 17 INTRODUCTION 17 θ N S ν γ σ ε Ś K Ñ V B D T G M K5, D10.11, T2, G1.2.4–6 (G3.7 part.), M1–4 K2, D10–12, Cd, T1.2, G1–7, M1–4 K0.3 (part), K1.4–6, V1, B, D (part) a b c ... n text witnesses Figure 2 The constituted text along with its critical apparatus: understanding what one is reading Note: The text of the critical edition represents the text of the archetype, as best reconstructed by the editor on the basis of the available manuscript evidence. It is thus running text and is both more continuous and less eclectic than the other editions. Only in places where the editor is uncertain of the reading (indicated by a wavy line under the text) is the text discontinuous, based not on the consensus of the tradition but only a part of it. In the example, verses 20 and 21cd are unambiguously the reading of the archetype; a part of verse 21ab is marked as uncertain and represents, in this case, the reading of only some manuscripts. A look at the critical apparatus suices for the reader to know exactly on which manuscripts the text is based and thus the stage of the tradition reconstructed or represented in the text before her. Reading a critical edition, then, is a matter of constant triangulation between three elements: the constituted text, the stemma and the critical apparatus. Additionally, a reader may wish to keep in mind implicitly a diagram such as Leonardi’s line, which allows her to represent in space the successive transformations of the text. If a reader knows how to use these elements, she will never be in any doubt about what she is reading (and will never make the mistake of thinking her text is “pieced together from various scraps of diferent bodies”). How to Interpret the Critical Edition The text reconstructed in the critical edition is the archetype of the tradition, deined as the latest common ancestor of the manuscripts examined for that edition. This sense of archetype should not be confused with the archetype as an especially authoritative or unique exemplar, for our stemma is merely hypothetical and models only a part of the historical reality—the part that is either preserved in or can be reconstructed from our manuscripts. 18 18 PHILOLOGY AND CRITICISM Having examined three misconceptions of the Mahā bhā rata critical edition in the pre- ceding section—namely: (1) the critical edition is eclectic; (2) the critical edition is not a text; and (3) the critical edition can be replaced by any other text with an apparatus of variants—let us turn to its interpretation. The correct interpretation of the consti- tuted text is that it is the archetype of the extant tradition, that is to say, the latest common ancestor of the extant manuscripts that can be reconstructed based on their evidence and therefore also the earliest form of the text that can be reconstructed based on these manuscripts.36 The extant witnesses will have older ancestors in common (for example, the source of the latest common ancestor and its source in turn and so on), but nothing can be said about these texts because, once we have interpreted all the evidence and drawn the nec- essary inferences, no data remain for us to arrange. We have accounted for all the variant readings in the manuscripts (eliminating, for instance, the errors found only in individual branches of the tradition or in individual manuscripts), and so what we have is the ear- liest and most authoritative text we can reconstruct.37 Thus, while we might assume that some of the readings in our archetype are themselves corruptions, we could not show that they are.38 There is no reason to interpret the constituted text of the Mahā bhā rata critical edition as anything other than what Sukthankar declares it, “the oldest form of the text which it is possible to reach, on the basis of the manuscript material available.”39 This sense of archetype, which we call the logical sense, must be distinguished from a second sense, which we refer to as the material or historical sense. As Timpanaro shows, before the term archetype acquired the meaning it has today in textual criticism,40 it had an earlier meaning. “Scholars used to think that the Humanists (like the ancients before them; cf. Cicero, Ad Att. 16.3.1) meant by the term archetypum or codex archetypus only the ‘oicial text’ checked by the author and intended to be published afterward in further copies.”41 Trovato discusses how this older sense of archetype, which he calls archetype1 and glosses as “oicial text, prepared by its author for publication,” is sometimes crossed with the irst sense (archetype2) to yield a third sense of the term: the archetype as “an especially authoritative exemplar, or as the result of a sudden and inexplicable bottleneck in the ancient and medieval tradition whereby only one copy survived.”42 Even highly experi- enced critics sometimes make this mistake. Hermann Kantorowicz writes that: [I]t is possible, and often necessary, to distinguish from the original an “archetype,” with a “textual history,” often centuries long, separated from the original, and which was saved by chance, or because it carried an erudite version, until it came down, alone or with only a few companions, to a time closer to our own.43 Giorgio Pasquali also states: Lachmann founded his method on the assumption that the tradition of every author always and in every case harked back to a single exemplar that was already disigured by errors and lacunas, which he called the archetype. No one doubts that this is mostly the case. […] On careful consideration, it must appear unlikely that in every case only one exemplar of each surviving work had been saved in the Middle Ages, whether Western or Byzantine, while all the others had perished with the fall of ancient civilization.44 19 INTRODUCTION 19 The error lies, as Trovato observes, in confusing “two conceptual spheres that we need to keep distinct: the history of the tradition, with its unattainable real trees, that is, the often very rich ensemble of mss. that historically existed, including those that disappeared without leaving traces, and textual criticism, with its very tangible although perfectible stem- mata codicum, based on the few mss. that have come down to us.”45 The ambiguity in the meaning of archetype, meaning once a hypothetical ancestor of the surviving manuscripts and once an actual manuscript that alone survived the reduction in the number of exemplars in the “medieval” period, is responsible for much of the confusion in the interpretation of the critical edition.46 This ambiguity also underlies an erroneous interpretation of the Mahā bhā rata critical edition, as we ind it in the work of Georg von Simson.47 Observing that the Mahā bhā rata “cannot be regarded as the direct transcription of oral poetry,” von Simson asks, “How then is the emergence of our text to be understood?”48 He suggests “these people [the Mahā bhā rata’s hypothetical redactors], after they had put together a irst, rough version of the text [from the oral materials in circulation], continued to work on it and inserted additional passages, which we today recognize as interpolations.”49 Then he adds: If, however, we proceed from the assumption the redactions had before them older versions of the Mahā bhā rata, then we must reckon with the fact that these older versions were already revised in various ways and that interpolations entered into them before they were evaluated for our Mahā bhā rata [the critical edition]. Further, we may assume that even after the completion of the inal redaction, for a certain period of time until the inal separation of the north and south Indian recensions, interpolations entered into the text, and which hence could not be eliminated using the editorial principles of the Poona edition [the critical edition]. The history of development of our Mahā bhā rata text can then be outlined as follows: (1) In the beginning there was […] the short epic transmitted as an oral, improvised composition in the tradition. Here it is meaningless to inquire into the original text and the author: the poem was realized by several epic bards in constantly changing form; bard and author were identical in this stage. From this period arise many of the formulaic expressions of our Mahā bhā rata, the technique of the adhyā ya introductions and conclusions, the schematic development of the small battle-scenes, etc. (2) There followed a period of written ixation of the text [schriftliche Fixierung des Textes]; indeed, we can also assume with certainty that diferent versions of the poem were written down at diferent times [zu verschiedenen Zeiten verschiedene Fassungen des Gedichts aufgezeichnet wurden], of which remains can still be traced in our text. (3) Finally, an individual diaskeuast or a committee of diaskeuasts, compilators, or redactors took up the task of forging together from the diferent written versions in circulation [verschiedene umlaufende schriftliche Fassungen] a great epic intended to exceed all previous versions in both extent and comprehensivity. The result was a text one can characterize as the goal of the Poona edition, a goal that could, of course, only be partially attained due to the inadequacy of the written transmission.50 Von Simson commits the very error Trovato cautions against. From the fact that the lines in our stemma converge at the apex, giving us a codex unicus, he concludes the constriction in the tradition must have been real, and he uses this to open up the tradition again above the archetype, giving the tradition the form of an hourglass. He explains the absence of any other descendants from the pre-archetypal tradition in terms of the inluence or the 20 20 PHILOLOGY AND CRITICISM dominance of the archetype, which he calls “the inal redaction” and attributes to the redactors’ desire to ofer “as comprehensive a text as possible.”51 Unfortunately, the constriction in the tradition is only apparent. As Trovato observes: Tracing a tradition back to an archetype dating, say, from the fourth century, does not at all mean that “in antiquity” (or in the Middle Ages, or in the early modern period) a single witness of our text was preserved, or a single copy that was authoritative for one reason or another. What it means is that the witnesses available today do not allow modern philologists to trace their way any further back than a given manuscript (usually lost), often far removed from the original, and sometimes datable with fairly reasonable approximation.52 From the apparently unique nature of the archetype,53 we may not conclude that only one exemplar of the text was in circulation, nor that this exemplar represents a “inal redac- tion.”54 Von Simson is clearly interpreting archetype in the second of our two senses listed earlier, that is, as “an especially authoritative exemplar, or as the result of a sudden and inexplicable bottleneck in the ancient and medieval tradition whereby only one copy sur- vived.”55 But no evidence exists that such a constriction occurred and hence also no evidence appears of a concerted inal redaction responsible for reducing the number of copies of the work to one. Von Simson’s error consists of not realizing “the archetype of the stemma has nothing to do with the history of the tradition (oicial copies, if any; copies commissioned for circulation by the author himself, etc.), but only with the ensemble of manuscripts that happen to be available today, used by the philologist in the stage of recensio.” This is why Trovato recommends textual critics “use the word archetype [only] to designate the point in the stemma beyond which the surviving tradition does not allow them to reach.”56 Conclusion The hypothetical ancestor of our manuscripts was probably one of several exemplars in existence at the time. It is solely by chance that only its descendants, rather than those of other manuscripts, survived, resulting in its apparently unique position in the history of the text. This uniqueness, however, is only apparent: it is a conse- quence of the fact that our stemma models only a part of the history of the text. From the apparently unique nature of the archetype, we may not conclude there was an actual reduction in the number of exemplars at the time. A fortiori all theories that attempt to explain the reduction in terms of the actions of putative “redactors” at the time are false. The idea of a conscious redaction (of oral epic materials) arises only because some scholars do not know how to read the stemma correctly. Von Simson is not simply confused about the meaning of archetype; he also intentionally exploits the ambiguity in its meaning to contest the editor’s choices. From the fact that the number of variants decreases as we approach the archetype, he concludes the archetypal variants must be the result of careful selection, and thus opts for poly- genesis of variants above the archetype. The constituted text is retained as the arche- type, although it no longer refers to archetype either in the technical sense (the latest common ancestor of the extant manuscripts that can be reconstructed on the basis of their evidence) or in the historical sense (the oicial text of the author’s hand), but is understood to mean “the late work of relatively untalented compilers.”57 Bizarrely, von Simson holds on to the aim of a reconstruction of the archetype,58 even though 21 INTRODUCTION 21 his suggestion undermines the logical relationships on which the reconstruction is pre- mised, and proposes to resurrect an earlier stage of the tradition through a subjective evaluation of variants. According to him, Since the older Bhā rata or Mahā bhā rata versions were not suppressed by the gigantic epic all at once but only bit by bit, some of the regionally attested manuscript versions could go back to their inluence. The hypothesis of an Ur-text is not signiicantly impaired by this assumption; but our attitude towards the readings will quite likely change, since they, in many cases, could easily be older than the Ur-text envisaged by the critical editor.59 Scholars will recognize this passage as the source not only of Andreas Bigger’s “nor- mative redaction” hypothesis or his hypothesis of a “parallel transmission” of the Mahā bhā rata but also of his idea that some of the readings rejected from the constituted text could actually have a greater claim to antiquity than the archetype.60 This passage is also the source of Reinhold Grünendahl’s thesis that the text “was transcribed directly from the oral tradition in a particular script.”61 His claim that “the text could have been ixed in a script from the oral tradition repeatedly and in completely diferent periods [der Text kann mehrmals und zu ganz verschiedenen Zeiten aus der mündlichen Tradition heraus ixiert worden sein]”62 practically repeats verbatim von Simson’s idea that “diferent versions of the poem were written down at diferent times [zu verschiedenen Zeiten verschiedene Fassungen des Gedichts aufgezeichnet wurden].” Finally, we may cite the work of James L. Fitzgerald, who cites, without crediting, von Simson’s thesis: The Pune text gives us an approximation of what must have been a very prestigious and important written Sanskrit text that eclipsed prior versions of the Mahābhārata both oral and written, though probably it did not eliminate them altogether.63 Fitzgerald adds: The critically established text is in part a “lowest common denominator,” and while Sukthankar was right to posit that everything in this “Gupta archetype” (not his designation) was faithfully transmitted everywhere (thus everything not found everywhere derives from some source outside the Gupta archetype), that putative fact does not mean that any and all textual elements not deriving from the Gupta archetype are posterior developments. In all likelihood there were prestigious written redactions of a Sanskrit Mahā bhā rata prior to that of the Gupta era as well as major and minor oral traditions. Many elements of such traditions not included in the “oicial” Gupta era redaction no doubt found their way into many, or even all, of the particular manuscript traditions through the normal processes of conlation.64 In spite of the widespread acceptance of von Simson’s thesis, it is unambiguously false. The critical edition does not provide any support for the thesis of “older Bhā rata or Mahā bhā rata versions.” No justiication exists for thinking the text reconstructed in it is an especially prominent exemplar, the product of an intentional redaction, compilation of diferent narratives, replacement of oral versions and the like.65 As we have seen, these interpretations are false, arising from a misconception regarding the nature of stemmatic 2 22 PHILOLOGY AND CRITICISM reconstruction, namely, that because the archetype occupies a prominent position on our stemma, it must also have been an especially prominent exemplar for the tradi- tion.66 In truth, the archetype may have been quite unexceptional from the perspective of the tradition. It is only by chance that its descendants, rather than those of the other manuscripts in circulation at the time, survived. Further, in the absence of stemmatic arguments, how does von Simson know some readings “could easily be older than the Ur-text envisaged by the critical editor”? The short answer is: he does not. His selection of “older” readings relies on a priori ideas of an older epic (for example, the war books or scenes rich in words for conlict are older than other sections).67 The suggestion that “the older Bhā rata or Mahā bhā rata versions” would have been “suppressed by the gigantic epic [not] all at once but only bit by bit” ofers a super- icially plausible model of how these allegedly older readings could have survived past the constriction (or, as Bigger calls it, the “normalization”) in the tradition. But as no evidence exists that such a redaction took place and the idea of a constriction is itself erroneous, the thesis’s probative value is nil. The idea rests on a misunderstanding concerning the nature of stemmatic reconstruction, coupled with creative theorizing. Likewise, the suggestion that “some of the regionally attested manuscript versions could go back to their inluence [of putative older oral versions of the epic]” is a feint, intended to provide a supericially plau- sible model of how readings rejected from the constituted text as less archaic could be older than the archetype. In reality, the stemmatic arguments for these readings as corruptions is extremely strong, and anyone who wishes to contest them must do so: either (1) on the basis of a diferent interpretation of the manuscript evidence, which von Simson has not provided; or (2) by rejecting stemmatic arguments altogether, which von Simson has also not done. His thesis is a transparent attempt to hold on, seemingly, to stemmatic recon- struction, while advancing arguments for the antiquity of certain readings that are actually non-stemmatic in nature, that is to say, his work is critical in name only. At this point von Simson might interject that, although he was mistaken about the constriction, the fact that several exemplars of the Mahā bhā rata were in existence at the time of the archetype makes it especially likely that the critical edition represents just one version of the work. It could therefore well be the case that the archetype was an author- itative exemplar, speciically redacted by the Brahmans and thereafter copied assiduously so that it replaced all other exemplars. Since our stemma models only a part of the his- torical reality, we would be hasty to assume this version is an accurate representative of the broader tradition. It may model just that part the Brahmans cared to preserve, and hence he is not wrong to seek to discern the outlines of a heroic Aryan, Kṣatriya tradition that preceded the Brahmanic one and was opposed to it. The problem with this argument is threefold: 1. Von Simson’s argument only shifts the problem from above the archetype to either adjacent to it or below it. It does not overcome the central diiculty that, in the absence of manuscripts or positive historical information, we have no reason to assume the existence of a heroic oral epic. If we have no evidence for a reduction of a plural epic tradition above the archetype into it, we also have no evidence for such a reduction alongside or shortly after it. 23 INTRODUCTION 23 2. Note also that the fact that we do not possess all exemplars ever produced does not mean our stemma is not an accurate representation of reality. As Timpanaro observes, “if it were possible to trace out the genealogical tree of all the manuscripts of a given text that really existed (what Fourquet and Castellani call ‘the real tree’), then this would almost always turn out to be much richer than the stemma we end up reconstructing on the basis of shared corruptions.” But “for the purposes of recensio this causes no problems: our simpliied stemmas function just as well for reconstructing the reading of the archetype as they would if we were able to trace out the ‘real stemmas.’ ”68 Although our hypothetical stemma models only part of the reality, and our reconstruction would be richer if we had access to more manuscripts (and it would take us back to an earlier stage of the tradition if we had access to older manuscripts or manuscripts related to our witnesses through remote ancestors), the text we restore would still be recognizably a text of this work.69 It would not take us back to a diferent work altogether. At most we would be able to identify some more latent errors in our archetype or ind that features we thought our witnesses owed to it are owed to a still earlier ancestor. But we would not ind our archetype is exceptional in any way. Consider, for example, the simpliied stemma in Figure 3. In the example, ω was copied at θ, δ, ε and ι. θ generated further copies at α, B and γ; α generated copies at A and β, ε at η and ζ and ι at κ and λ. Δ was destroyed before it produced further copies; A and B survived, but β, γ, η, ζ and κ and λ did not. Our stemma would reconstruct only a part of the historical reality (θ–A–B). The fact that we do not have access to α makes no diference to the reconstruc- tion because, as Trovato observes, “the lines of stemmata, like geometrical lines, are composed of ininite points,”70 that is to say, there could have been ininitely many intermediate copies between θ and A  (or α and A) without this making any diference to the reconstruction. Access to β would also not help us, since the errors unique to it are worthless for reconstructing the reading of the arche- type and those it shares with A could be owed just as well to α as to θ. If we had access to γ, we could be more certain of reconstructing θ, but we could still not ω ε θ δ ι η α κ λ A β B Υ ζ Figure 3 The part stands in for the whole 24 24 PHILOLOGY AND CRITICISM reconstruct an earlier stage of the tradition. Access to δ, ε or ι (or, if ε and ι are destroyed, to η, ζ, κ and λ, since we could restore the reading of ε from η and ζ and of ι from κ and λ) would permit us to reconstruct ω, and in this case we would be reconstructing an earlier stage of the tradition. But ω remains the archetype and the text we obtain, though older, would not be radically diferent, since θ, δ, ε and ι are all copies of the same work.71 Although the versions more likely to survive are those of successful editions (that is, the so-called vulgatae) or those that produce the most descendants, unless we had a reason for thinking the text had undergone a drastic change in the time it was copied from ω to δ (or ε or ι) and the time it was copied from ω to θ, we have no reason for thinking δ, ε and ι represent Kṣatriya versions of the epic and θ the Brahmanic version of it. As for the suggestion that θ might itself be the “Brahmanic redaction,” what sense does it then make to assert that it is a copy of ω? Θ would be a new work, and far from possessing just the archetype of the surviving manuscripts, we would have the irst source of the entire tradition, whether surviving or lost, that is, the original.72 Unless we can prove θ is a completely diferent text (which requires another exemplar for comparison, that is, δ, ε, ι or one of their descendants), we are justiied in thinking the part of the tradition we have (AB) stands for the whole.73 It would be unparsimonious to think otherwise.74 3. More serious, von Simson has not understood the way manuscripts are created and destroyed at all. He thinks some exceptional event such as the destruction of the heroic Kṣatriya culture and Brahmans’ concerted eforts to assert their own reli- gious ideology is required in reducing the number of copies available to one. But in truth, the process is less spectacular. As Vincenzo Guidi and Paolo Trovato show, at almost all rates of decimation, the probability of a one-branched tradition is higher than the alternatives.75 Although their analysis depends on the tradition studied, they conclude that the calculations also apply to the study of real trees.76 Indeed, Trovato notes: Whatever the decimation rates used in our experiments, it is intuitively evident that, given any real tree, with 2, 3, or exceptionally even 5 or 6 branches, the chances of survival for any of its primary branches (which we can call real families) will be: a. high if it is a family crowded with witnesses (and therefore in many cases with further branches of its own), b. low if it is a family with few members, c. very low, or non-existent, if it is a family composed of a single manuscript.77 Thus all that is required for other, hypothetically existent branches of the Mahā bhā rata to have died out is for the number of copies in these branches to fall below a certain threshold—without the necessity of assuming the copies contained signiicantly diferent versions of the epic. For the tradition of Sannazaro’s Arcadia, Guidi and Trovato show that it suiced “for the decimation rate to be above 51.6% for there to be a high probability of ending up with a two-branched stemma and the extinction of the archetype.”78 Given the vicissitudes of manuscript transmission, it is quite plausible that all our surviving manuscripts are descended, directly or indirectly, from a single source, without this source, for that 25 INTRODUCTION 25 12 10 15 16 13 12 13 18 21 32 27 31 22 28 23 25 941 CE 1144 CE 13 21 13 32 40 41 37 21 27 32 45 22 90 105 38 40 41 34 36 45 92 61 63 39 37 47 91 96 69 43 38 95 100 49 101 86 1287 CE End of process Figure 4 The birth and death of manuscripts Source: Reproduced from Trovato, Everything You Always Wanted to Know about Lachmann’s Method, 139. reason, being the sole manuscript in existence at the time,79 and without it being a special Brahman-dominated version, “relect[ive] of the[ir] view […] and […] to a great extent driven by their interests and ideologically prejudiced.”80 Figure 4 illustrates the process of transmission of a work and clariies why the consti- tuted text need not be as late as is often thought. At 941 CE, three copies (10, 12 and 13) of the work survive and the tradition consists of two branches. At 1144 CE, 12 is lost, but not before it set forth the tradition (of which again copies 15, 16, 18, 23 and 25 are lost; only 22 survives). Copy 13 is about to disap- pear, but has generated ive descendants. At 1287 CE, the last surviving manuscript of the left branch (22) is about to disappear; without it, the original cannot be reconstructed anymore. Nonetheless the surviving witnesses permit us to reconstruct 13, which as their 26 26 PHILOLOGY AND CRITICISM latest common ancestor is the archetype. If a part of the tradition descended from 13 was damaged (for example, 90–92), this would not matter, because we could still infer the existence of 21 from 96 and 105 (though not their intermediate archetypes), and so reconstruct 13. If, on the contrary, 96 and 105 were destroyed as well, we could only reconstruct 32, which in this case would be the archetype. Note that we do not need to possess all the manuscripts ever produced to reconstruct the archetype. If 22 or any of its descendants survived, we could still reconstruct the original itself: 22 is more valuable to us than any number of descendants of 13 (provided we have at least one). This example, though adapted from its original context,81 illustrates the danger of thinking the Mahā bhā rata critical edition reconstructs a “Gupta-era text” (Fitzgerald) and the like. The truth is we do not know, because the complete tradition is almost never spread out before our eyes like this. 91 and 101 are very late copies, separated by seven generations from the original, yet their agreement suices to give us the text of 13, removed just two generations from the original. All that counts is that our manuscripts model as wide a spectrum of the tradition as possible. One good manuscript deriving from higher in the hierarchy brings us closer to the original than several lower down in the stemma, and the lateness of our manuscripts should not be used as an argument for the lack of originality of the text reconstructed in the critical edition. In the example given earlier, a late sixteenth- or seventeenth-century copy of 22, pre- suming it survived until then, was more valuable than fourteenth- or ifteenth-century copies of 32. And since we do not know which manuscripts our witnesses are copies of, and we have no way of estimating their distance from the original,82 we should place much greater faith in the fact that we have some excellent manuscripts, whose text may derive from eighth- or ninth-century CE sources, or even earlier ones, and that we appear not to have lost any signiicant part of the tradition.83 If Sukthankar’s testimony—“there is nothing to suggest that our Mahā bhā rata manuscripts have sufered any serious loss at any time. There never was any lack of manuscripts, many of which were preserved carefully in temples, and which must have been copied repeatedly, for the enhancement of merits. There is no evidence of any break in the tradition at any time or any place, within the conines of India at least. The probable inference is that our manuscripts con- tain all that was there originally to hand down, and more”84—is correct, then we have reason to think the Mahā bhā rata as reconstructed in the critical edition is essentially the work as it always existed:  as we saw in the aforementioned example, all that matters is we have one exemplar from the left branch of the tradition for us to reconstruct the orig- inal. In other words, what is decisive is not the loss of individual manuscripts but the annihilation of entire branches. As long as even one descendant of the branch survives, its essential information is preserved and this means, provided none of the higher-level branches has been fully decimated, that we can reconstruct a very early stage of the transmission. If, as Sukthankar suggests, there was continuous copying of manuscripts, it is entirely plausible that even higher-level branches of the Mahā bhā rata tradition have been preserved (note that it is not the preservation of lower-level branches that counts). It does not matter that these branches have been preserved through frequent copying because the fact that our witnesses are three or four generations removed from the source matters less than the fact that, when overlaid over each other, so to speak, 27 INTRODUCTION 27 they give us exceptional resolution of the original.85 There is no reason for thinking the text we possess is anything other than the Mahā bhā rata as it always was: a work of phi- losophy, law, cosmology and didactics. Von Simson’s attempted “text-critical investigations” (textkritische Untersuchungen) into the Mahā bhā rata mark a turn away from the methodological advances of the nineteenth century—the introduction of a systematic recensio and replacement of emendatio ope ingenii with emendatio ope codicum.86 They actually set the ield of textual criticism back 200 years, to a stage before the discovery that the archetype can be reconstructed following a mathe- matical and universally replicable procedure. In place of objective rules and calculations of probability for deriving the reading of the ancestor of a given set of manuscripts, von Simson’s work advocates a return to an arbitrary and subjective Quellenkritik, where readings are selected at random because they conform to a preconceived notion of the original epic. As we demon- strated in The Nay Science, the arguments for defending the greater antiquity of some readings or passages over others are always the same four: (1) the bardic hypothesis, (2) the Ksạ triya hypothesis, (3) the war narrative hypothesis and (4) the Brahmanic hypothesis.87 The redef- inition of the constituted text of the critical edition as a conscious redaction of earlier oral epic materials permits a recourse to a well-established German tradition of interpreting the Mahā bhā rata as evidence of the enslavement of the Ksạ triyas by the Brahmans and the down- fall of a heroic Aryan people.88 We cannot even argue that von Simson is advocating a return to the greater use of subjective iudicium, the greater use of the editor’s subjective judgment of the authenticity and quality of readings, because this presupposes a familiarity with criteria such as the usus scribendi (the scribe’s habitual style) and lectio diicilior (the principle that the more diicult or the more obscure reading is preferable), neither of which is applicable here because: (1) von Simson does not think a single author existed and cannot be familiar with his style; and (2) he does not know which the lectio diicilior is since he is not comparing two readings, one of which represents the banalization of the other, but two readings, one of which is supposed to represent the “oicial” choice of the Mahā bhā rata redactors and the other the “original” content of the bardic narration. In fact, the only criterion applicable is the critic’s sense of what may or may not have been appropriate to a heroic culture—a crite- rion already applied in the work of Adolf Holtzmann Jr.89 As the unacknowledged source of almost all modern Mahā bhā rata criticism after the completion of the critical edition, von Simson’s work has had a massive inluence on the edition’s reception. Yet it is completely false. It has given scholars a way to recast or to completely circumvent the evidence of the critical edition. As the earliest stage of the text that can be reconstructed using scientiic means, the latter posed a major challenge to the Mahā bhā rata critics’ theories of a heroic original epic, with later Brahmanic and bhakti interpolations. The critical edition ofered no support for the distinction of “epic” from “pseudo-epic” elements. It also did not justify scholars’ claims that parts of the constituted text of the Mahā bhā rata could be earlier than others, or their attempted “layering” of the epic. It did not support their contention that the Bhagavadgītā was not originally part of the epic or that books such as the postwar Śā nti and Anuśā sana parvans were added after the formation of the original epic. There was also no evidence to support the thesis of three historical versions of the epic corresponding to its three narrations—the narration 28 28 PHILOLOGY AND CRITICISM by Ugraśravas to Ś aunaka, by Vaiśaṃpā yana to Janamejaya and by Vyā sa to his disciples. Even though it rejected some of the most popular narratives in the Mahā bhā rata tra- dition as later insertions, the text the critical edition produced was much closer to the traditional reception of the Indian epic as a body of inspired literature than to the German critics’ assertions. The Mahā bhā rata critics had hoped for a critical edition as the best means for undermining the authority of the textual tradition,90 and the Bhandarkar editors had countered with an edition bearing out the traditional reception of the epic. If the critics were to survive, they would have to redeine the critical edition’s evidence in a way conducive to their speculative views, and this explains the contemporary signiicance of the- ories that the Mahā bhā rata critical edition captures but a late stage—speciically the stage of its irst transcription from an older, luid oral epic tradition—of its transmission.91 Von Simson’s thesis ofered the critics a way to claim they were cognizant of the Mahā bhā rata critical edition and took its evidence into account in their work, while continuing the kind of subjective, unscientiic and only allegedly historical reconstructions they had previously undertaken.92 As the basis of almost all claims about the Mahā bhā rata tradition in the con- temporary period, the redeinition of the critical edition as either a “inal redaction” or as a “normative redaction” (hereafter, we shall use Bigger’s term for it) is thus key to under- standing contemporary Mahā bhā rata scholarship. In the next chapter, we therefore take a closer look at it. Notes 1 See P.  L. Vaidya’s “Postscript” to the introduction to the Ś ā ntiparvan (S. K.  Belvalkar, The Śāntiparvan:  Introduction [Poona:  Bhandarkar Oriental Research Institute,  1966], ccxlviii). Belvalkar intended to furnish his uninished introduction with a description of the Mokṣadharmaparvan’s contents before his death. 2 Nineteenth-century German Indologists placed their trust in an oral Aryan epic, the so-called Urepos. They argued this epic was the property of heroic Kṣatriya warriors, and they blamed the Brahmans for turning it into a legalistic document concerning ritual, worship of ancestors and theology. Although these views were anti-Semitic in inspiration and in intent (for Brahmans read: priestly or Semitic; for heroic read: Aryan and Germanic), they still taint contemporary Mahā bhā rata studies. For a discussion, see our The Nay Science:  A History of German Indology (New York: Oxford University Press, 2014), chapters 1–2. 3 Thus, for instance, Fitzgerald, when he writes: “Between 400 and circa 50 BCE: the continued circulation of that old narrative within a new oral Pā ṇḍava epic, in which ive entirely new, semidivine heroes were injected into the Bharata dynasty (depicted as incapable of sustaining itself now) and then sufered exile from it, and, inally, allied with the old Pañcā la rivals, over- threw the Kuru-Bharatas and established a new Bharata reign. This hypothesis thus agrees, in part, with A. Holtzmann Junior’s thesis about a reversal of the polarity of the heroes and villains (see Holtzmann, 1892),” and “Diferent versions of the ‘inversion’ theory were debated at length by various authors pointed to in the last note. My historical focus on the epic’s history is quite diferent from that of these earlier scholars, though clearly I think the ‘inversionists’ were on the right track.” James L. Fitzgerald, “Mahā bhā rata,” in Brill’s Encyclopedia of Hinduism, ed. Knut A. Jacobsen, Helene Basu, Angelika Malinar and Vasudha Narayanan (Leiden: E. J. Brill, 2010); Brill Online, 2015; http://referenceworks.brillonline.com/entries/brill-s-ency- clopedia-of-hinduism/mahabharata-BEHCOM_2020040 (accessed September 23, 2015) and James L. Fitzgerald, “No Contest between Memory and Invention:  The Invention of the Pā ṇḍava Heroes of the Mahā bhā rata,” in Epic and History, ed. David Konstan and Kurt A. 29 INTRODUCTION 29 Raalaub (Malden, MA: Wiley-Blackwell, 2010), 119, n. 18. But in less explicit ways as well, contemporary Mahā bhā rata studies largely reprise nineteenth-century views:  for instance, when scholars insist the epic’s origins lie in a heroic, oral bardic tradition. In spite of the fact that no evidence exists for this view and scholars have not found a way to combine textual crit- icism, which presumes the mechanical transcription of written exemplars, with theories of oral transmission, they still defend the oral origins hypothesis because of the unparalleled license it gives them to make claims about the Mahā bhā rata. The entire problem of Mahā bhā rata studies thus revolves around this one question: can scholars demonstrate the existence of an oral epic in noncircular, non-self-referential or non-question-begging ways? 4 The Mahábhárata, an Epic Poem Written by the Celebrated Veda Vyása Rishi, 4 vols. (Calcutta: Education Committee’s Press/Baptist Mission Press, 1834–39); Atmaram Khadilkar, ed., Mahābhārata with Nı̄lakaṇt ̣ha’s Commentary (Bombay:  Ganpat Krishnaji’s Press, 1863); Vasudev Balacharya Ainapure, ed., The Mahābhārata with the Commentary Bhāvadı̄pa of Nıl̄ akaṇtḥ a, 6  vols. (Bombay:  Gopal Narayan, 1901); T. R. Krishnacharya and T. R. Vyasacharya, eds., Sriman Mahabharatam:  A New Edition Mainly Based on the South Indian Texts with Footnotes and Readings, 19  vols. (Bombay:  Javaji Dadaji’s “Nirnaya-sagar” Press, 1906–14); P. P.  S. Sastri, ed., The Mahābhārata (Southern Recension): Critically Edited by P. P. S. Sastri (Madras: Ramaswamy Sastrulu, 1931–33); Ramachandra Kinjawadekar, ed., Mahābhāratam with the Commentary of Nı̄lakaṇtḥ a, 6 vols. (Pune: Citrashala Press, 1929–36). Two earlier editions, the Madras edition of 1855 and the Sarfojirajapuram edition of 1896, seem not to have survived. 5 See Winternitz, expressing this precise anxiety in the bulletin of the 12th International Congress of Orientalists:  “This much, however, is clear even from the few extracts from South-Indian MSS.  hitherto published, that the text of the Mahabharata as found in our Bombay and Calcutta editions is an utterly insuicient basis for critical researches concerning the life of the great Hindu Epic, and that the text on which all Mahabharata studies have hitherto been founded, is not the text but only one of the texts of the Mahabharata.” Moriz Winternitz, “A Proposal for the Formation of a Sanskrit Epic Texts Society, to Be Laid before the Indian Section of the XII International Congress of Orientalists, Held at Rome, in October 1899,” XIIme Congrès International des Orientalistes, Bulletin no. 3 (1899): 46 (Winternitz’s italics). 6 Ibid., 47 (Winternitz’s italics) and see also Moriz Winternitz, “Promemoria über die Nothwendigkeit einer kritischen Ausgabe des Mahā bhā rata, insbesondere der südindischen Recension,” Almanach der Kaiserlichen Akademie der Wissenschaften 51 (1901):  207:  “I therefore already declared at the Paris Congress that a critical edition of the Mahābhārata organized by Western scholars according to the principles of philological criticism valid in Europe is the conditio sine qua non of all Mahābhārata research” (Winternitz’s italics). 7 Gianfranco Contini, “Ricordo di Joseph Bedier,” in Esercizi di lettura sopra autori contemporanei con un’appendice su testi non contemporanei (Turin: Einaudi, 1974), 369 (Pugliatti’s translation). 8 V. S. Sukthankar, “Prolegomena,” in The Ad̄ iparvan for the First Time Critically Edited, vol. 1 (Poona: Bhandarkar Oriental Research Institute, 1933), ciii. 9 Lino Leonardi, “Il testo come ipotesi (critica del manoscritto-base),” Medioevo romanzo 35 (2011): 7. 10 Sukthankar’s earlier comment (see n. 8 of this Introduction) therefore must be relativized. The critical edition is a “mosaic of old and new matter,” but to a lesser degree than the competing editions, because of the editor’s concern with printing as archaic and as consistent a text as possible. In fact, except for the places in which the text is forced to descend to the level of one of its hyparchetypes (subarchetypes or archetype of only one branch of the tradition)—indicated in the critical edition by a wavy line—we have no reason for thinking the text is not continuous. 11 Michele Barbi, La nuova ilologia e l’edizione dei nostri scrittori da Dante al Manzoni (Florence: Sansoni, 1938), xxii–xxiii. 12 Doniger’s expression; see Wendy Doniger, “How to Escape the Curse,” review of The Mahabharata, translated by John Smith, London Review of Books 31, no. 19 (2009):  17–18, 30 30 PHILOLOGY AND CRITICISM www.lrb.co.uk/v31/n19/wendy-doniger/how-to-escape-the-curse (accessed October 13, 2015). The complete passage reads: “There are several recensions of the Mahā bhā rata, each preserved and cherished by a particular community. The critical edition, by contrast, is like Frankenstein’s monster, pieced together from various scraps of diferent bodies; its only com- munity is that of the Pune scholars, the Frankensteins. Moreover, it left out a great deal of material that the Indian literary and religious traditions have continued to draw on, such as the passage in which Vyasa dictates the entire text to the elephant-headed god Ganesha.” 13 These materials are not taken from the various manuscripts in existence. They are intel- lectually intuited as belonging to their common source, based on well-understood rules and calculations of probability. For instance, if we have a tripartite stemma consisting of three extant manuscripts A, B and C deriving from a lost, common source α and two manuscripts read m whereas the third reads n and m and n are not insigniicant variants, what is the like- lihood two manuscripts simultaneously arrived at m whereas the third preserved the orig- inal reading n? In restoring m to the constituted text, the editor does not “piece together” the edition from m and similar passages:  rather, he intuits with his mind’s eye what the extant manuscripts’ common source must have contained, such that it can give rise to the observed variants or such that the observed variation between the witnesses can be satisfac- torily explained. 14 See preceding note. “Validated by the tradition” does not mean the line is taken from the manuscript or manuscripts that contain it. Rather, the line is included in the constituted text because it is logically understood as a feature of the archetype, and this logical understanding is based on the tradition, although it then goes beyond it. 15 Sukthankar is aware of both aspects. See his “Prolegomena,” lxxxvi: “Our objective can only be to reconstruct the oldest form of the text which it is possible to reach, on the basis of the manuscript mate- rial available”; and ibid., ciii: “It only claims to be the most ancient one according to the direct line of transmission, purer than the others in so far as it is free from the obvious errors of copying and spurious additions” (all italics Sukthankar’s). 16 Note that adopting a particular recension of the Mahā bhā rata such as the Telugu recension, as Doniger suggests, will not avoid the problem. First, the Telugu recension is a composite not only in the sense that it is a mixture of good readings with scribal changes, obvious errors and so on, but also because it includes much material from the neighboring Grantha recension. The Telugu recension was found to be the most eclectic recension of the Mahā bhā rata. It contained material not only from the Grantha but also from Devanā garī, Bengā lī and other northern sources. Second, which of the available Telugu manuscripts should we use? The Bhandarkar editors collected and examined 28 manuscripts of the Telugu recension, of which three were included in the critical apparatus. Each contained a diferent mixture of readings and episodes. Which should we take as representative of the “Telugu recension”? The task of reducing the available readings to one cannot be avoided. Third, even if we restricted our- selves to a single exemplar (for example, the manuscript T1), we would not solve the problem of “[leaving] out a great deal of material that the Indian literary and religious traditions have continued to draw on.” For this exemplar, like any other Mahā bhā rata manuscript, represents just a selection of the material found in the complete tradition. It will also leave out certain narratives. The problem can only be solved if we create an edition containing or representing all Mahā bhā rata manuscripts, but how else to present this information and evaluate narratives for their authenticity than by following a strictly hierarchical procedure, that is, by subordi- nating the more recent readings to the more archaic? 17 Sebastiano Timpanaro, The Genesis of Lachmann’s Method, ed. and trans. Glenn W. Most (Chicago: University of Chicago Press, 2005), 159, n. 3. 18 As Whitehead and Pickford also note. See Frederick Whitehead and Cedric E.  Pickford, “The Introduction to the Lai de l’Ombre:  Sixty Years Later,” Romania 94 (1973):  153:  “It is exaggerated to represent a manuscript as preserving a state of the text that has enjoyed an 31 INTRODUCTION 31 authentic existence, since what we are dealing with is in most cases something as composite as any reconstituted critical text produced by a scholar. A manuscript generally presents us with several strata of alteration and behind many readings there is a complex history of change.” 19 Giorgio Pasquali, Storia della tradizione e critica del testo (Florence:  Le Monnier, 1952), 123, expresses it beautifully: “The better the judge of […] a reading knows the language and habits of the ages that transmitted his reading, or that may have coined it, the sounder his judgment will be. The best critic of a Greek text handed down by Byzantine tradition will be the one who, besides being a perfect Hellenist, is also a perfect Byzantinist. The best editor of a Latin author transmitted in Medieval or post-Medieval mss. will be the one who is as knowledgeable about the Middle Ages and Humanism as he is about the author and his language and times. Such a critic is an ideal that no one can perfectly incarnate, but that it is everyone’s duty to strive to come near to” (Trovato’s translation). 20 Bieler makes the same point: “Some scholars have gone so far as to forgo any attempt at a classiication of manuscripts; they prefer to single out one manuscript that is on the whole more satisfactory than the rest and follow this codex optimus except where it is defective or intelligible; there, and there only, they would have recourse to other manuscripts or to emendations at their own risk. This practice has no less a champion than Professor Bédier. […] To be sure, it does away with a great deal of trouble—but is it really safer or more objective? Unless an extant manuscript is demonstrably the original or at least the archetype (in which event it would be treated as unique no matter what method the editor adopts), the choice for a basic text of one manuscript rather than another rests on the same subjective considerations as does the old-fashioned stemma codicum; and if this textus unius codicis is safe from the fallacies of recensio, it has admittedly no determinable relation to the original—in other words, it is deliberately uncritical. The inevitable improvements upon the basic manuscript must be arbitrary, even if the editor has tact and taste; where these indeinable qualities are lacking, we may be prepared for the worst. […] In the last analysis this is the abdication of reason in a sphere where it was least expected.” Ludwig Bieler, The Grammarian’s Craft: An Introduction to Textual Criticism, 3rd edn. (New York: The Catholic Classical Association, 1960), 18–19. 21 The metaphor is F. A. Wolf ’s. The complete passage reads as follows: “A true, continuous, and systematic recension difers greatly from this frivolous and desultory method. In the latter we want only to cure indiscriminately the wounds that are conspicuous or are revealed by some manuscript or other. We pass over more [readings] which are good and passable as regards sense, but no better than the worst as regards authority. But a true recension, attended by the full complement of useful instruments, seeks out the author’s handiwork at every point. It examines in order the witnesses for every reading, not only for those that are suspect. It changes, only for the most serious reasons, readings that all of these approve. It accepts, only when they are supported by witnesses, others that are worthy in themselves of the author and accurate and elegant in their form. Not uncommonly, then, when the witnesses require it, a true recension replaces attractive readings with less attractive ones. It takes of bandages and lays bare the sores. Finally, it cures not only manifest ills, as bad doctors do, but hidden ones too.” F. A. Wolf, Prolegomena to Homer, trans. with an introduction and notes by A. Grafton, G. W. Most and J. E. G. Zetzel (Princeton, NJ: Princeton University Press, 1985), 43–44. 22 The term is von Hinüber’s. He voices the underlying objection thus:  “Looking back, the moment it [the critical edition] was planned turned out to be exceptionally fortuitous, as the methodological advances that have been achieved in the recent period would have prevented this edition [from being realized,] for its stated goal is to reconstruct the Ur-Mahā bhā rata, a task of which we know in the meantime that it is completely unattainable. The text we read in the ‘critical’ edition never existed in this form. What has arisen is not the Ur-text, which could not have arisen at all, but a completely new normalized recension. […] It is 32 32 PHILOLOGY AND CRITICISM beyond question that they [the editors] have presented researchers with an extremely useful working tool [Arbeitsinstrument], provided that it is used keeping in view its own methodo- logical and historical conditions.” Oskar von Hinüber, “Der vernachlässigte Wortlaut. Die Problematik der Herausgabe buddhistischer Sanskrit-Texte, Zur Überlieferung, Kritik und Edition alter und neuerer Text,” in Kleine Schriften, Teil I, ed. Harry Falk and Walter Slaje (Wiesbaden: Harrassowitz, 2008), 436. 23 More precisely, since even German Indologists cannot fully dispense with a text, what they really reject is the idea of a pregiven text, that is to say, a text that they have not given themselves and hence could act as a check on their interpretations. Their eforts are thus directed against the very idea of a transmitted text, be it the vulgate or the critical edition. Although this eleva- tion of individual subjectivity over the text has complex roots, one source is clearly Luther’s opposition to tradition as an independent source of revelation and, even more so, his valo- rization of the historical context over other kinds of exegetic concerns. Luther’s statement, “We do not illuminate the history (rem gestam) by the mysteries of Scripture, but [we illumi- nate] the mysteries of Scripture by history, that means: [we illuminate] the Old Testament by the Gospel and not vice versa,” could stand as the irst principle of the Indologists’ phi- lology. Martin Luther, Commentary on Psalm 22, cited and translated in Siegfried Raeder, “The Exegetical and Hermeneutical Work of Martin Luther,” in Hebrew Bible/Old Testament: The History of Its Interpretation, vol. 2: From the Renaissance to the Enlightenment, ed. Magne Saebø (Göttingen: Vandenhoeck & Ruprecht, 2008), 377–78. But it will take the dissolution of the link between the text and divine inspiration in the transformation of the principle of scriptura sacra est verbum dei (the holy Bible is the Word of God) into the principle scriptura sacra continet verbum dei (the holy Bible contains the Word of God) in neo-Protestant theology to inally lib- erate them from the dependence on not only the tradition but also the text. The primary fetter to such a historicist hermeneutic is, of course, any form of traditional authority, including the text itself insofar as it constitutes a source of authority heteronomous to their own, whereas the primary obstacle to reading the text “historically” is the circumstance that the text interprets itself. Hence their insistence on removing all parts of the text that explicitly state its purpose including the Ādiparvan, the Bhagavadgītā and the Nārāyaṇīya. 24 We ind the same objection in a letter addressed to one of us, warning against the error of reading the constituted text as running text. “The critical edition is no more than a conve- nient working tool [Arbeitsinstrument], with whose aid one can survey what is (hopefully) a rep- resentative selection of (hopefully) important manuscripts.” Michael Hahn, letter to Vishwa Adluri, February 10, 2010, 4. But the constituted text is running text and can be read as such. It was Sukthankar’s explicit intent that it be so. Against German critic Walter Ruben, who recommended an indiscriminate use of the wavy line (indicating uncertainty), Sukthankar objected: “even a constituted text is after all a text, meant to be read like any other text, and not a chart of the aberrations of careless copyists of the last two millennia.” V. S. Sukthankar, “Epic Studies III: Dr. Ruben on the Critical Edition of the Mahā bhā rata,” Annals of the Bhandarkar Oriental Institute 11, no. 3 (1930): 280. 25 Alfredo Stussi, Introduzione agli studi di ilologia italiana, 4th edn. (Bologna: il Mulino, 2011), 116. 26 Lorenzo Renzi and Alvise Andreose, Manuale di linguistica e ilologia romanza (Bologna: il Mulino, 2009), 268. 27 Arrowsmith makes the same point. Philologists have been busy “keeping the texts pure and uncontaminated” but without asking “for what and to whom.” William Arrowsmith, “The Shame of the Graduate Schools: A Plea for a New American Scholar,” Arion: A Journal of Humanities and the Classics, third series, 2, no. 2/3 (1992–93): 173 (Arrowsmith’s italics). The philologist’s irst concern must be to interpose himself as little as possible between the author and his readers. 28 Leonardi, “Il testo come ipotesi,” 6. 29 See Sylvain Lévi, review of The Mahābhārata. For the First Time Critically Edited, by Vishnu S. Sukthankar, Journal Asiatique 215 (1929): 347: “Si j’osais me permettre une suggestion dans ce domaine, je conseillerais à l’éditeur de renoncer, par pitié par nous, à la part même du 3 INTRODUCTION 33 travail qui lui tient le plus à coeur et qui apporte à son esprit le plus de satisfaction, la recon- struction de l’Ur-Mahā bhā rata comme il se plait à dire, d’accepter la Vulgate—autrément dit l’édition de Nīlakaṇṭha, par exemple—comme point de départ, et de nous livrer au plus tôt le dépouillement, comme il est fourni par ses notes, des manuscrits décrits et classés selon l’excellente méthode qu’il a adoptée.” 30 Lévi’s proposal goes beyond Joseph Bédier’s. Whereas Bédier renounced Lachmannism only for editing the bon manuscrit, Lévi requires the editor to renounce any attempt at interpreting the manuscript material, including the identiication of the best manuscript. The editor is not supposed to edit at all; he should merely reprint the vulgate. At most, his editorial activity extends to collating manuscripts and collecting (not even ordering) the variants. Here Lachmann’s dictum, “we can and must edit [recensere] without interpreting,” would truly be ful- illed, but hardly in a way the elder scholar could have desired. 31 “The (printed) editions of Nīlakaṇṭha’s version leave much to be desired. They have arbitrarily changed many of the readings and added a certain number of lines which are not found in the Nīlakaṇṭha manuscripts hitherto examined.” Sukthankar, “Prolegomena,” lxvii. 32 Ibid., lxviii. 33 Lévi could argue the editor should select one manuscript of the vulgate and stick rigorously to it. But even here, he could not avoid providing a justiication of why this manuscript over another, unless he wished to be perfectly arbitrary. A threshold of evaluative activity exists below which an editor may not go. A critical edition can be made mechanical and free of subjectivity to a great extent: it cannot be automated completely. 34 The same error occurs in Georg von Simson, Mahābhārata: Die Große Erzählung von den Bhāratas (Berlin: Verlag der Weltreligionen, 2011), 688, when von Simson asserts: “The value of the critical edition consists above all in the fact that it ofers a wealth of readings in its critical apparatus and appendix that, in many cases, ease the task of interpretation, and that a not insigniicant number of individual verses, groups of verses or even longer textual passages that appear only in individual manuscripts or only in a few manuscripts have become recog- nizable beyond reproach as late additions.” Von Simson seriously underestimates the value of a critical edition when he thinks its primary function is to supply the reader with variant readings “that lighten the task of interpretation.” The variants reprinted in the apparatus are precisely those the editor rejected as of doubtful authenticity. Restoring them to the consti- tuted text whenever the reader faces interpretive challenges undoes the point of establishing a critical text. Von Simson’s understanding is closer to Lévi’s, for whom the apparatus represented a mere spectrum of possibilities, from which anyone could choose. Von Simson recommends this approach because his work consists primarily of identifying elements of the “oral epic,” and hence he considers it important to portray the critical apparatus as no more than a collection of readings of equal authority (and some of greater authority than the constituted text). 35 Practically also, such a list would be less useful than Lévi imagines. In a true critical apparatus, if the manuscripts a, b, c, d and e belong to the family A and for a given verse all of them feature the variant x, the editor does not need to list them separately. He can merely note A x or A: x, depending on his conventions. This principle holds true even if one or more of them vary. In that case, he might note A (exc. b) x; b y or A (exc. b, e) x; b y; e z or A (d as in text) x. This approach, of course, relies on his having completed a preliminary recensio or study of the manuscript materials so that he knows a, b, c, d and e are descendants of a common source. In contrast, Lévi’s proposed list will merely contain the raw data from the initial collations (the comparisons of the manuscripts), without the all-important preliminary digestion the editor will have done for the reader. Since Lévi rejects genealogical analysis, this would be an apparatus without an accompanying stemma and hence almost completely useless. It would inform the reader that variants exist, but she would not know what the value of those variants is, what the genetic relationship between them (or between the witnesses in which they occur) is and whether it is legitimate to replace the printed text with a certain variant. 34 34 PHILOLOGY AND CRITICISM 36 Reeve and Trovato ofer similar formulations: Reeve:  “a. (archetype of speciic witnesses) latest common ancestor; b.  (archetype of a work, or without qualiication) latest common ancestor of the known witnesses; known witness when common ancestor of the rest.” Michael D. Reeve, “Archetypes,” in Manuscripts and Methods: Essays on Editing and Transmission (Roma: Edizioni di Storia e Letteratura, 2011), 117. Trovato:  “The archetype (archetype2) [is a text that] can be reconstructed logically—with varying degrees of accuracy—on the basis of the number of conjunctive errors that are common to all its descendants; that is, by subtracting from the total number of errors attested in the tradition all separative errors that are exclusive to each family of witnesses, and those characterizing individual witnesses.” Paolo Trovato, Everything You Always Wanted to Know about Lachmann’s Method: A Non-Standard Handbook of Genealogical Textual Criticism in the Age of Post-Structuralism, Cladistics, and Copy-Text (Padua: Libreriauniversitaria, 2014), 64. 37 Barring divinatio (conjecture), of course. But for reasons that will become clear, neither does much scope for divinatio in the Mahā bhā rata exist nor are the Mahā bhā rata critics’ attempts to reconstruct an earlier stage of the transmission based on divinatio, at least as understood in textual criticism. 38 Note that no new arrangement of the data will permit us to reach an earlier stage of the tradition. We could, at most, restore some other variant or variants to the constituted text (moving ones we previously thought were original to the critical apparatus), but this text would once again be, by deinition, the archetype, and therefore the earliest form of the text that can be reconstructed based on the available evidence. Note also that from the fact that the scribe committed error z in moving from the archetype to one of its descendants, we cannot assume that he, or another scribe, must likewise have committed error y in moving from the ancestor of the archetype to the archetype. This would be an error of induction. And while testimonia such as fragments or archaeological and other historical evidence may cause us to revise some of our judgments, they will never justify us in editing our text drastically to it those testimonia, for testimonia by their very nature are uncertain and could refer to a version of this text, an abridgment or another text altogether. This is the error we ind in Dieter Schlinglof, “The Oldest Extant Parvan-List of the Mahā bhā rata,” Journal of the American Oriental Society 89 (1969): 334–38 and the literature based on his work. See John L. Brockington, “The Spitzer Manuscript and the Mahā bhā rata,” in From Turfan to Ajanta: Festschrift for Dieter Schlinglof on the Occasion of His Eightieth Birthday, ed. Eli Franco and Monika Zin (Lumbini: Lumbini International Research Institute, 2010), 75–87. 39 Sukthankar, “Prolegomena,” lxxxvi (Sukthankar’s italics). 40 That is, as a manuscript, detected through reconstruction, that stands at the apex of the stemma and from which the branching of the tradition began. 41 Timpanaro, The Genesis of Lachmann’s Method, 50. 42 Trovato, Everything You Always Wanted to Know about Lachmann’s Method, 64. 43 Hermann Kantorowicz, Einführung in die Textkritik:  Systematische Darstellung der textkritischen Grundsätze für Philologen und Juristen (Leipzig: Dieterich, 1921), 13 (italics added). 44 Pasquali, Storia della tradizione e critica del testo, 15, cited and translated in Trovato, Everything You Always Wanted to Know about Lachmann’s Method, 65 (Trovato’s italics). 45 Trovato, Everything You Always Wanted to Know about Lachmann’s Method, 65. 46 We have no reason to assume this except in an ininitesimal percentage of texts. In the majority of cases that critics assumed such a reduction, it can be shown that they are confusing the archetype in the irst sense with the second. Even Timpanaro occasionally commits this error, as when he writes: “Some years ago, Courtney wrote a curious essay where he postulates a fourth-century archetype for all the Virgilian codices that have come down to us. […] He appears to be aware of the unlikelihood that a single copy of the Virgilian text was preserved at a certain point in antiquity,” not noticing that the confusion he attributes to Courtney is 35 INTRODUCTION 35 his own. Sebastiano Timpanaro, Per la storia della ilologia virgiliana antica (Rome:  Salerno Editrice, 1984), 181, cited and translated in Trovato, Everything You Always Wanted to Know about Lachmann’s Method, 65. Without other evidence (for example, the testimony of ancient scribes or historiographers), we cannot know whether all the copies of a text were reduced at some point to one. Only in cases where we possess only a single exemplar of a text can we assume such a thing, but to call this witness an archetype of itself would be an unusual use of the term. 47 Georg von Simson, “Altindische epische Schlachtbeschreibung, Untersuchungen zu Kompositionstechnik und Entstehungsgeschichte der Bücher VI bis IX des Mahā bhā rata” (Habilitationsschrift, University of Göttingen, 1974). 48 Ibid., 282. 49 Ibid., 283. 50 Ibid., 283–84. We have given the German because these phrases recur almost verbatim in the work of Bigger and Grünendahl. Bigger: “zu irgendeiner Zeit in schriftlicher Form ixiert wurde”; “die einzig schriftliche Fassung”; “Verschiedene Abschreiber fügten jedoch […] aus anderen (jüngeren?) schriftlichen Fassungen—im Verlaufe der Zeit Passagen aus anderen Versionen in die Abschrift ein.” Andreas Bigger, Balarāma im Mahābhārata: Seine Darstellung im Rahmen des Textes und seiner Entwicklung (Wiesbaden: Otto Harrassowitz, 1998), 14 and 111. Grünendahl: “schriftlich in einer bestimmten Schrift ixiert worden”; “mehrmals und zu ganz verschiedenen Zeiten […] schriftlich ixiert worden sein.” Reinhold Grünendahl, “Zur Klassiizierung von Mahā bhā rata- Handschriften,” in Studien zur Indologie und Buddhismuskunde, ed. Reinhold Grünendahl, Jens Uwe- Hartmann and Petra Kiefer-Pülz, Indica et Tibetica 22 (Bonn:  Indica et Tibetica Verlag, 1993), 128. 51 Von Simson, “Altindische epische Schlachtbeschreibung,” 282. 52 Trovato, Everything You Always Wanted to Know about Lachmann’s Method, 66. 53 Apparent only from the perspective of the tradition, for which there would have been many copies, and not from our perspective, from which the archetype is really unique. Perhaps corresponding to the distinction between the several senses of archetype we should introduce a distinction between a text that is a codex unicus for us and one that is a codex unicus simply or without qualiication. 54 It is irrelevant whether this supposed redaction is called a “inal redaction” (von Simson), a “normative redaction” (Bigger) or a “Gupta redaction” (Fitzgerald): the error is the same. All three authors commit the error of identifying this supposed exemplar with the constituted text reconstructed in the critical edition, Fitzgerald most egregiously in Brill’s Encyclopedia, when he introduces and discusses the critical edition under the heading:  “The Written Sanskrit Mahā bhā rata: The Gupta-Era Text Reconstructed from the Manuscript Tradition.” 55 Trovato, Everything You Always Wanted to Know about Lachmann’s Method, 64. 56 Ibid., 66. 57 Von Simson, “Altindische epische Schlachtbeschreibung,” 285. 58 See ibid.: “All these features [errors in recollection, introduction of synonymous expressions, dis- placement of verses, etc.], which, together with the efects of the contamination of manuscripts, have made the creation of the critical edition so diicult appear to me, however, to exclude nei- ther the hypothesis of an archetype of the extant manuscripts nor that of an Ur-text, to which these would have to be traced back. One must only be clear about the fact that this Ur-text was the late work of relatively untalented compilers. […] If there ever existed an Indian Homer, then his work is not our Mahā bhā rata but an earlier version of the epic, now forever lost to us.” 59 Ibid., 285–86. 60 Bigger, Balarāma im Mahābhārata; and see also Andreas Bigger, “The Normative Redaction of the Mahābhārata:  Possibilities and Limitations of a Working Hypothesis,” in Stages and Transitions:  Temporal and Historical Frameworks in Epic and Purāṇic Literature. Proceedings of the Second Dubrovnik International Conference on the Sanskrit Epics and Purāṇas, ed. Mary Brockington (Zagreb: Croatian Academy of Sciences and Arts, 2002), 21–33. 36 36 PHILOLOGY AND CRITICISM 61 Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 128. 62 Ibid. 63 James L. Fitzgerald, “The Mahā bhā rata,” in The Hindu World, ed. Sushil Mittal and Gene Thursby (New York: Routledge, 2004), 70. 64 Ibid. (italics added). And see also Fitzgerald, “Mahā bhā rata,” where Fitzgerald, combining von Simson’s and Bigger’s hypotheses, presents this account: “The text established by Sukthankar and his colleagues is an approximation of a written redaction of the text that became norma- tive (Bigger 1998; Bigger 2002), which seems to have served as the actual written source of most of the subsequent manuscript tradition, but which evidently did not function as the sole source of the text of every parvan.” 65 Philipp Maas thinks “the inal redaction(s) of the MBh,” which, he argues, “would be the archetype(s) of all manuscripts,” are “the result of a single authorial/redactorial intention.” He adds: “It is, however, improbable, in my view, that the comprehensive literary material was created anew for this occasion. I would rather assume that a lot of material (written and/or oral) was reused, rearranged, and adapted to the intentions that guided the com- position of the MBh [these ‘intentions’ can apparently be grasped in an act of immediate intellectual intuition, an intuitus originarius].” Philipp A.  Maas, message to Vishwa Adluri, online session on Academia.edu, October 10, 2015. Why exactly he thinks the “inal redac- tion” is the “archetype” of all manuscripts he does not explain: surely the archetype is the latest common ancestor of the surviving manuscripts, and there is no reason it should be the “inal redaction.” Then again, the text of the “inal redaction,” assuming one occurred, could have existed in several manuscripts. Which of these was the archetype? Maas evidently views transmission not in terms of mechanical transcription of written exemplars, but as an oral epic tradition and its “Brahmanic” redaction or redactions. He uses archetype precisely in the sense Trovato cautions against, that is, as “an especially authoritative exemplar, or as the result of a sudden and inexplicable bottleneck in the ancient and medieval tradition whereby only one copy survived.” Note also that if archetype is to have a deinable meaning, it can only exist in the singular: there is no such thing as the “archetypes” of a tradition. Likewise, the expression “inal redactions” is logically incoherent: if there were several, nei- ther can be inal. 66 These claims are false for a second reason, namely, von Simson’s arguments for an oral epic are merely suggestive and do not prove the existence of such an epic. As in his article “Die Einschaltung der Bhagavadgı̄tā im Bhı̄ṣmaparvan des Mahābhārata,” Indo-Iranian Journal 11, no. 3 (1968/69): 159–74, he uses the fact that questions are not answered immediately or in the most obvious way to impute the work of later “interpolators,” and from this he concludes that the Mahā bhā rata, puriied of these passages, would correspond roughly to the original oral bardic epic. The argument is doubly lawed, because: 1. Digression and a resumption of the original theme following a digression are not evidence for interpolation: the Mahā bhā rata has a recursive style, and the repetition of groups of words or phrases is not restricted to “interpolators.” Von Simson’s analysis assumes that because the narrator resumes the narration with the same or similar words after a digres- sion, the intervening passage must be an insertion and the repetition evidence of a hasty attempt to smooth over the transition. 2. Even if von Simson could show these passages are interpolations, this would not prove the existence of an earlier oral epic. Von Simson merely assumes one of the characteristics of the oral tradition is its brevity, its focus on the main combatants and its emphasis on warfare. But these could also be features of a written composition. His arguments also do not succeed in eradicating or altering the story’s basic dimensions: its placement within a mimetic universe, the yuga framework and its representation as another stage in the perennial conlict between the gods and the titans. The following is a typical example of his analysis: Noting that Saṃjaya does not respond directly to Dhṛtarā ṣṭra’s “precise 37 INTRODUCTION 37 question” (ke ’rakṣan dakṣiṇaṃ cakram ācār yasya mahātmanaḥ | ke cottaram arakṣanta nighnataḥ śātravān raṇe ||, Mahā bhā rata 7.129.3) but instead describes how some Pā ṇḍava heroes attacked Droṇa and how the nighttime battle commenced in general terms, he suggests this is evidence the following passages are an insertion into the original text. He further notes that a few lines later Dhṛtarā ṣṭra once again interrupts Saṃjaya to ask about the combatants’ reactions to Droṇa’s appearance on the ield and for further details (tasmin praviṣtẹ durdharṣe sṛñjayān amitaujasi | amṛṣyamāṇe saṃrabdhe kā vo ‘bhūd vai matis tadā ||, 130.1). His conclusion is as follows: “Verse 130.5 of Dhṛtarā ṣṭra’s question appears to be connected directly with 129.3[:] Whereas the inal pāda of this verse corresponds word for word with 129.3d, the diaskeuast attempts in another place to vary: he contrasts kathaṃ mṛtyum upeyivān in 129.4 with a better chosen kathaṃ … pañcatvam upajagmivān in 130.7. In the description of battle, death is spoken of often enough. However, when in spite of con- sulting the Pratīka-Index, the word pañcatva is attested at the beginning of a pāda only out- side of the battle, except for this one case, then we can consider this a deinite sign for the latest origin of our passage.” Von Simson, “Altindische epische Schlachtbeschreibung,” 208. The literal repetition amounts to four words (nighnataḥ , śātravān, raṇe and kathaṃ), three of which occur ive times in the Mahā bhā rata (the phrase nighnataḥ śātravān alone occurs seven times and the singular form nighnataḥ śatrūn another seven) and are common enough words in a description of battle, and the fourth is one of the commonest interrogatives in the Mahā bhā rata. There is nothing surprising about Saṃjaya initially responding with a general description of the battle: this is a common feature of the Mahā bhā rata, and von Simson exaggerates the discrepancy when he suggests Dhṛtarā ṣṭra is forced to restate his question. 67 Von Simson’s 1974 Habilitationsschrift is rich in examples of subjective reconstruction. He argues in the introduction that “if traces of the old, orally improvising heroic poetry are to be found anywhere, then it is here [in the battle scenes].” He continues, “Precisely the wealth of small duel-scenes ofered the possibility of distinguishing the prosaic and the typ- ical [elements] from the unique, the essayed original [elements], those leading to the artiicial poetry that was conceived of in written form. The origins of the Mbh out of an orally impro- vising epic poetry manifests itself in a wealth of details also demonstrable in the epic poetry of other peoples. […] That the description of battle contains older and more recent sections can be deinitively demonstrated in at least some places.” Von Simson, “Altindische epische Schlachtbeschreibung,” 9. The methods he applies to this end are “statistical examinations of vocabulary and metrical analyses” and “[the identiication of] repeated groups of verses” (also occasionally a more nebulous criterion of “style”). Ibid., 10. Using these criteria he eliminates several verses, even though the criteria are subjective and the argument is circular, and his claim to “continue the text-critical work [already done by Sukthankar] on the basis of the Poona edition [the critical edition]” (ibid., 7) amounts to a confusion of lower and higher criticism. 68 Timpanaro, The Genesis of Lachmann’s Method, 164. 69 Something similar can be observed in the efort to sequence all the genes of the human in the Human Genome Project. Even though every human carries a diferent combination of DNA bases or “letters,” sequencing the genes of just a few sample individuals (in actual fact, more than 70 percent of the reference genome came from a single male donor) still provides us with an accurate picture of the human genome. If we added samples, we would obtain a more granular picture (which is being done in the project to develop a haplotype map of the human genome), but we would not suddenly arrive at the chimpanzee genome. We would deepen our understanding of individual variation and the range of our analysis, but because the individuals studied are members of the same family, we would still ind that the genome initially sequenced accurately represents the whole. 70 Trovato, Everything You Always Wanted to Know about Lachmann’s Method, 148. 38 38 PHILOLOGY AND CRITICISM 71 In Figure 3, although we possess only the part θ–A–B (θ reconstructed on the basis of AB), θ–A–B nonetheless stands for the whole because, as in a hologram, every part contains the whole. Θ will give us the work x, which is contained in all of its exemplars without ever being identical to any of them, including ω. It may give us that work with variations—as do all the exemplars: even in ω, where x comes as close to being realized as possible, ω is never fully iden- tical with x, the work in mente, even before we consider the possibility of scribal errors, authorial revisions and the like—but it gives us that work nonetheless. 72 Could von Simson accept θ is the original, a work in ieri, the apex not only of our hypothetical stemma but also of the historical tree (the arbre réel) and yet insist we must draw a dotted line into θ from some point still higher than it, indicating contamination from an earlier source? He could, but only if he had another manuscript for comparison, that is, either an apograph of this source or a quotation from it or even testimony about it. But the Mahā bhā rata cannot itself be both an example of a contaminated text and evidence for the source of contamination. 73 The following illustration may make things clearer. Let our nine regional recensions of the Mahā bhā rata stand for higher-level branches, descending through one or two generations from the original. If eight of these branches were lost through decimation and only the Grantha branch was preserved, generating the descendants G1–7, we could reconstruct their archetype G through comparing their readings. We could identify some interpolations and correct some corruptions, though perhaps not many, as the archetype is close in time to its witnesses (the archetype for this reason would also be more diicult to identify). If we had access to the Ś ā radā codex, we could reconstruct not just G, but also the common ancestor of the Ś ā radā and Telugu manuscripts, by deinition the archetype of the entire Mahā bhā rata tradition. We would ind this text is much shorter than G (the Grantha recension has a strong tendency to inlation) and we could restore many good readings (Ś ā radā is a better tradition). But we would not ind Ś ā radā was a Kṣatriya epic, whereas G contained a Brahmanic text. At most, we would note a tendency to elaboration, speciication of the proper rites and rituals and more extensive phalaśrutis in G. But no recension among our nine recensions would lack entire books (for example, the Ś ā nti or the Anuśā sana parvans) of the Mahā bhā rata. The idea that if we could just ind another recension, at the same level in our stemma as the archetype, we would suddenly see that the Mahā bhā rata was originally a Kṣatriya epic is an illusion, no doubt arising from the earnest desire to possess such an epic. 74 It is also hardly credible that θ represented such a major innovation over ω yet no one noticed the diference or that the other exemplars, which would now be not just better or worse copies of the same work but completely diferent works than θ, were not conserved. The reasons for a line’s extinction are varied and include natural catastrophes, unfavorable economic and social conditions and so on, but one of them is clearly that successful editions that meet popular demand lessen the incentive to continue copying earlier models. In this case, since θ is not just an edition of ω but a completely diferent work, we have no reason for thinking ω would not continue to be preserved from one of its other descendants. Von Simson’s argument assumes just that blend of malevolence and destructiveness he attributes to the Brahmans, the proof of which is supposed to be that the original has not survived! 75 Vincenzo Guidi and Paolo Trovato, “Sugli stemmi bipartiti. Decimazione, assimmetria e calcolo delle probabilitità,” Filologia italiana 1 (2004): 9–48. 76 Guidi and Trovato restrict themselves to the traditions of printed editions, which have the advantage that the entire tradition, as well as the exact dates of creation and loss of exemplars, is known. 77 Trovato, Everything You Always Wanted to Know about Lachmann’s Method, 145. 78 Ibid. 79 To be very clear, we are discussing two diferent things here: the irst is whether the archetype was the sole manuscript in existence at the time it was created or at the time copies were made from it (it was not); the second is whether it is possible that branches of the tradition deriving 39 INTRODUCTION 39 from other exemplars in existence became extinct over time, leaving this branch as the sole one in existence (it is highly probable, as Guidi and Trovato’s work and the experience of other editors shows). Thus von Simson is wrong on multiple counts: irst, he thinks the archetype was the only text in existence at the time; then, he imagines a special event was required to ensure it was the only text in existence at the time; and inally, he also confuses the gradual decimation of exemplars over time below the one exemplar that happened to generate copies that survived, leading to this branch of the tradition becoming the sole branch we can reconstruct, with the extinction of the Ksạ triya tradition all at once and alongside or just before the archetype (barring those “surviving” elements, of course, that he thinks he can detect with higher critical means in the Mahā bhā rata). 80 Von Simson, Mahābhārata: Die Große Erzählung von den Bhāratas, 591. 81 See the note to the diagram. The original is from Michael P. Weitzman, “Computer Simulation of the Development of Manuscript Traditions,” Bulletin of the Association for Literary and Linguistic Computing 10, no. 2 (1982): 55–59, but we could not examine this work. Weitzman’s convention is to designate lost exemplars by circles; dotted circles represent exemplars about to disappear. Weitzman and Trovato both feature the diagrams in order to make a point about the frequency of two-branched stemmata (namely, given random manuscript “death” a two-branched stemma is the likeliest outcome). Our concern is not the likelihood of bipartite stemmata but the constantly changing position or the constantly changing identiication of the archetype in the stemma. 82 Spencer and Howe argue we can “estimate […] the actual number of changes given the observed number of changes, using a mathematical model for copying errors.” Matthew Spencer and Christopher J. Howe, “Estimating Distance between Manuscripts Based on Copying Errors,” Literary and Linguistic Computing 16, no. 4 (2001): 467. Their work is based on deriving a relationship between observed and actual errors, but permits us neither to date manuscripts nor to know the number of intermediaries separating an exemplar from a given source. Distance, in their model, apparently refers to “the average number of changes per location (word)” and is expressed literally as the length of the lines in the stemma. Ibid., 467–68. 83 The oldest dated manuscript of the Ā diparvan is the Nepā lī manuscript Ñ3 dated 1511 CE. Sukthankar estimates the Ś ā radā codex S1́ is “three or four centuries old.” Sukthankar, “Prolegomena,” xlvii. This would place it around the sixteenth century CE. The manuscript’s text appears very old: it lacks many interpolations that characterize the other manuscripts; it is free of their innovations; the language is archaic. Assuming birch-bark manuscripts last 300–400 years before they deteriorate and an upper limit for the age of palm-leaf manuscripts is 800–1,000 years, the scribe of the Ś ā radā codex could have had access to a twelfth-century CE exemplar. If its scribe in turn had access to an old exemplar, we could be looking at a text perhaps ive or six generations removed from the autograph. The classic study on Indian writing materials is A. F. Rudolf Hoernle, “An Epigraphical Note on Palm-Leaf, Paper and Birch-Bark,” Journal of the Asiatic Society of Bengal LXIX, II (1900): 93–134. See also D. C. Sircar, Indian Epigraphy (Delhi: Motilal Banarsidass, 1965). Stefan Baums, “Gandhā ran Scrolls: Rediscovering an Ancient Manuscript Type,” in Manuscript Cultures:  Mapping the Field, ed. Jörg Quenzer, Dmitry Bondarev and Jan-Ulrich Sobisch (Berlin, Munich and Boston, MA: de Gruyter, 2014), 183–226, describes birch-bark manuscripts dating to the second century BCE, but these appear to be exceptional cases. 84 Sukthankar, “Prolegomena,” xcv (Sukthankar’s italics). 85 As always, frequent copying of manuscripts is the best remedy against loss of the tradition and what frequent copying introduces in terms of errors it more than makes up in terms of preserving the tradition. A lower signal-to-noise ratio—that is, a higher level of noise—is pref- erable to no signal at all. 86 Recensio refers to the systematic review or examination of the available manuscripts (or at least a large selection of them). Emendatio ope ingenii is emendation with the help of native wit as opposed to emendation with the help of manuscripts (emendatio ope codicum). 87 Adluri and Bagchee, The Nay Science, 79–83. 40 40 PHILOLOGY AND CRITICISM 88 See von Simson’s introduction to his abridged translation—Mahābhārata:  Die Große Erzählung von den Bhāratas. After stating, “Researchers are still divided about the genesis of the Sanskrit version, which dates back possibly to the middle of the irst century BC but was probably brought into the form we know today only centuries later,” he writes: “as with the Trojan War depicted in the Homeric Iliad, in the center of the Mahābhārata is a war we cannot locate politi- cally and historically.” Ibid., 583 and 585. Then he continues: “The Brahman authors and redactors to whom we owe the epic’s latest revisions sought to legitimize with all the rhetorical means [at their disposal] a social structure in which they could claim the dominant role. Already since the late Vedic period the battle for supremacy between the Kṣatriyas and Brahmans was a constant topic. […] The picture we encounter here is by no means a balanced one, but rather, relects the view of the epic’s Brahmanic revisionists, and this view is to a great extent driven by their interests and ideologically prejudiced. The interests of the Brahmans, that is, the class interests of the priests and intellectuals, are clearly expressed above all in Books III, XII and XIII where there was rich occasion for reminding the Kṣatriyas, that is, the members of the warrior class, of their duties and for warning them by means of suitable narrative examples of the punishments that threaten for overstepping bounds. Such reprimands can even be put in the mouths of Kṣatriyas, as is the case, for example, of Kṛṣṇa in the ‘Bhagavadgītā ’ or Bhīṣma in Books XII and XII—probably they attained even greater signiicance in this way.” Ibid., 591. See also ibid., 603: “Such textual passages, which retard [the action] and invite one to relect [on it] and provide a commentary on the main narrative from moral and philosophical perspectives, could hardly have belonged to the oldest poem, which appears to have known only an articulation of the plot into ‘the quarrel,’ ‘the forest-exile,’ and ‘the war,’ as is appro- priate to a heroic poem”; ibid., 638: “He [Vyā sa] is a Brahman, priest, seer and ascetic—his introduction as the author of the Mahābhārata thus probably irst occurred in the period when the Brahmans had taken over the poem from the sū tas, the epic bards”; ibid., 639: “The great battle, however, in which we may see the core of the old heroic poem, is narrated by the char- ioteer (sūta) Saṃjaya to his master, the blind King Dhṛtarā ṣṭra”; ibid., 640: “the ixation of the text in writing was, in fact, preceded by a phase in which epic materials were transmitted by orally improvising epic bards, whom we can thus identify with the sū tas”; ibid., 641: “Although indications of early precursors of the Mahābhārata exist that were possibly transmitted only orally [and were] perhaps closer to ballads than to an epic, [precursors] that may have emerged in the milieu of the sū tas which was closer to the Kṣatriyas, the text we possess places so much emphasis on the Brahmans’ superiority over the Kṣatriyas (founded on their spirituality and ascetic lifestyle) that one can infer an extensive revision by the Brahmans”; “Here [in the ifth book] the talk is still of the glory of the warriors, which may have been the genuine topic of the old bardic poem”; ibid., 643: “The Brahmans’ animosity toward the warrior nobility expressed here indicates just how far the Mahābhārata had traveled from its origins in a bardic poem that probably had a much closer connection with the Kṣatriyas. The repeated destruction of the Kṣatriyas occurs as a punishment for their arrogance and their insuicient respect toward the Brahmans. It is thus not a glorious victory as in the old heroic poem”; ibid., 644: “certainly, this was the main objective of the Brahmanic revisionists of the heroic poem transmitted of old, namely, to impel the Kṣatriyas to accept the Brahmanic legal order”; and ibid., 645: “there once existed diferent versions of the Bhā rata poem in competition with another, be they oral or be they written, which were at some point combined into a [single] text by the Brahmans’ redactional activity.” And see also ibid., 632 and 664 for references to “Indo-Germanic” and “Aryan,” respectively as characterizing the epic’s earliest stage. 89 See Adluri and Bagchee, The Nay Science, chapter  2. The key passage from Holtzmann Jr. is: “We thus ind some similarities in the conditions of the most ancient Indian and the most ancient Germanic periods; here as there we encounter a battle-lusty warrior race with all the seeds of culture alongside a cruel crudeness of passion. Whether here one must suppose [the epic is] a recollection of a primordial age during which they lived together or a further development 41 INTRODUCTION 41 that ran parallel under similar existential conditions, I cannot decide here. [But] it is certain the Mahâbhârata has preserved traces for us that reach back to a very early period of Indian antiquity; that therein recollections of an Indo-Germanic primordial age have been preserved, is, at least so far, at best plausible.” Adolf Holtzmann Jr., Zur Geschichte und Kritik des Mahābhārata (Kiel: C. F. Haeseler, 1892), 51 (italics added). 90 The relevant passages appear in Winternitz’s article “The Mahabharata,” The Visva-Bharati Quarterly 1 (1924): 343–47. See especially ibid., 343: “[I]f I wished to show you the diference between the Indian and the Western attitude of mind in studying things Indian, I could not do better than showing you what the Mahábhárata is to us, why we study it and how we study it”; “Every Indian is justly proud of the Mahábhárata, and every Indian probably knows some- thing of the story and the characters of the great Epic. But I venture to doubt whether many Indians know what the Mahábhárata really is and all that it contains in its hundred thousand slokas. For, as I have said, it is rather a whole literature than a single poem”; “Mahábhárata is an abbreviation of Mahábhárat-ákhyána, and means ‘the great story of the battle of the Bháratas.’ The Bháratas are mentioned already in the Rigveda as a warlike tribe, and in the Bráhmanas we irst meet with Bharata, the son of Duhsanta and Sakuntalá, who is considered as the ancestor of the royal family of the Bháratas”; ibid., 344: “In consequence of some family quarrel in the royal dynasty of the Kauravas, there arose a great and bloody war, in which the old family of the Kauravas or Bháratas was almost entirely annihilated. Although we know of this war only from the Mahábhárata, and not from any other sources, we shall have to look upon it as most probably a historical event. The story of this battle was sung in ballads, and out of these ballads some great poet of name unknown created a great historic epic of the battle of Kurukshetra. This old heroic poem forms the kernel of the Mahábhárata”; ibid., 345–46: “To the present day this gigantic work in spite of all the divergent elements which have entered into it, is generally considered in India as one uniform poem, composed by the venerable Rishi, Krishna Dvaipáyana, or Vyása, who is also credited with the arrangement of the four Vedas and the authorship of the Puránas. (This is about the same as if one were to believe that the whole of Sanskrit literature from Kalidasa to Jayadeva was composed by one man.)”; ibid., 346: “These statements [regarding the epic’s multiple narrations] prove that, even in India, in spite of the pious belief in the authorship of Vyása for the whole poem, some recollection was yet retained of the fact that the Mahábhárata had gradually grown from an originally smaller poem to its present size. Of this fact there can not be the least doubt that our Mahábhárata, as we have it now before us, is a very diferent work from the original epic poem of the battle of the Bharatas”; and ibid., 347: “The critical historian of Indian literature will not join in this praise of the Mahábhárata. He will not see in it as a whole, a work of art at all, but a literary monstrosity. The Mahábhárata, as we have it now, is a vast compilation of the most heterog- enous [sic] matter, a very jungle of poetry and learning in which the most beautiful lowers of poetry grow side by side with weeds of no beauty at all; and the profoundest wisdom is found by the side of the silliest rubbish” (Winternitz’s italics throughout). 91 Referring to the text as “the epic as we have it” is another way to achieve the same end. If the expression merely refers to the text we have, it is unobjectionable, for the epic we have is obvi- ously, though trivially so, the epic as we have it. Actually, however, what Fitzgerald (who uses the expression) intends therewith is the presumed contrast with another version, diferent from the one we presently have. Here is how he glosses it: “By ‘the epic as we have it’ and ‘text of the Mahābhārata’ I mean the written redaction of a Sanskrit text of the MBh that was composed and promulgated sometime around the time of the Gupta empire. This text was approxi- mately recovered in the unsuccessful efort to arrive at a critical edition.” James L. Fitzgerald, “The Rā ma Jā madagnya ‘Thread’ of the Mahā bhā rata: A New Survey of Rā ma-Jā madagnya in the Pune Text,” in Stages and Transitions: Temporal and Historical Frameworks in Epic and Purāṇic Literature. Proceedings of the Second Dubrovnik International Conference on the Sanskrit Epics and Purāṇas, ed. Mary Brockington (Zagreb: Croatian Academy of Sciences and Arts, 2002), 89, n. 1. We 42 42 PHILOLOGY AND CRITICISM discuss the problems with thinking the critical edition represents a “written redaction” and with the idea that an archetype must have been specially “composed and promulgated” in the next chapter. 92 This is neatly illustrated by Fitzgerald’s comments in the introduction to his translation of the Ś ā ntiparvan. Contrasting himself with Madeleine Biardeau, he writes, “Readers may see, however, that I am, at least in principle, much more concerned than she is with matters of history. Where Biardeau sees in the Mahābhārata some instance of a uniied epic-purāṇic cos- mology and theology, I see it as situated in particular circumstances, as a pragmatic utterance which certain agents used to some advantage. I also see it as having a diachronic history—that is, as containing within it various later developments of some of its own earlier formulations.” James L. Fitzgerald, “Introduction [to the Book of Peace],” in James L. Fitzgerald, trans., The Mahābhārata, vol. 7:11. The Book of the Women; 12. The Book of Peace, Part One (Chicago: University of Chicago Press, 2004), 127, n. 198. The problem is: a study of the text’s diachrony must begin with the archetype and what develops from it as the history of the text and not earlier. This history is captured in the critical edition, which thus presents the only objective basis for its study. The diachronic history of the text is precisely what the stemma and apparatus of variants record. To seek to study it prior to the text’s irst recorded emergence (or what can be reconstructed as such) is the study not of diachrony but of pseudochrony. In that it exceeds the domain of what can be said, it is not even pseudochrony but simply nonsense. 43 4 The irst editorial board of the Mahābhārata critical edition Source:  Reproduced from the frontispiece of the Prospectus of a Critical and Illustrated Edition of the Mahā bhā rata, India’s Great National Epic. 45 Chapter One ARGUMENTS FOR A HYPERARCHETYPAL INFERENCE Our only source is the manuscripts themselves, and therefore, in the inal analysis, these stemmata. We therefore do not have the right to repudiate their evidence on the pretext that it appears absurd to us.1 —Robert Marichal, “La critique des textes” The Normative Redaction Hypothesis The constituted text is not the archetype of the tradition but merely a “normative redaction,” deined as “a redaction that had a normative efect and overgrew all other versions.” The critical edition reconstructs “a text that was a historical fact at a certain period in time,” but precisely because it is not the archetype, it should not hinder us from exploring “the prehistory of the normative redaction.” Above all, we should consider “passages rejected from the constituted text,” as they could be evidence of a “parallel transmission” of the Mahābhārata. We commence our review of post-critical edition Mahā bhā rata scholarship with a look at Andreas Bigger’s work. Although the thesis that the critical edition reconstructs merely the Mahā bhā rata’s “inal redaction” rather than its archetype is not new—as we have seen, it originated in 1974 with Georg von Simson2—its contemporary renaissance can be traced almost exclusively to Bigger,3 speciically his hypothesis that the constituted text reconstructs the “normative redaction” of the Mahā bhā rata, deined as “a uniform redaction that was ixed in a written form at some time and to which further texts and streams were added by later copyists.”4 The characteristic feature of this redaction, how- ever, and the feature that lends it its name, is that it “had a normative efect and overgrew all other versions.”5 As an example of such a version, Bigger cites Nīlakaṇṭha’s seventeenth-century vul- gate edition. He argues, “if in the seventeenth century a commentator could compile a text of the Mahābhārata that disseminated itself in certain circles and thereby doubtless suppressed other versions, it is also conceivable that a much earlier redaction attained such authority that it dominated the entire written transmission.”6 Apparently, the oicial text of the Mahābhārata in the normative redaction decimated other versions in exis- tence, both oral and written, thus enabling a new Brahmanically authorized version— and, concurrently, a Brahmanic vision of society—to take hold.7 The salient feature of the normative redaction, however, and the reason Bigger advances the hypothesis, is that it permits him to reinterpret verses found only in some manuscripts as remnants of an earlier oral epic tradition. Although rejected from the constituted text as of doubtful authenticity, Bigger argues, “from the perspective of a theory of a normative redaction […], some of them represent a parallel transmission. The normative redaction admittedly 46 46 PHILOLOGY AND CRITICISM suppressed the other versions in these cases. But some copyists incorporated parts of these versions in their copy of the normative redaction.”8 Identifying and collecting these passages thus not only permits him to reconstruct an earlier stage of the transmission than what is represented in the critical edition (the so-called prehistory of the normative redaction);9 it also undermines the critical edition’s claims to accuracy and authenticity. If Bigger’s thesis is correct, it will reveal, as Oskar von Hinüber postulates, that “what has arisen [in the critical edition] is not the Ur-text, which could not have arisen at all, but a completely new normalized recension [eine ganz neue normalisierte Rezension].”10 Ultimately, the normative redaction hypothesis validates the German Mahā bhā rata critics’ suspicion that “much of what [we] would like to count among the characteristic content of the poetry, proves ultimately the inferior, derivative work of later copyists [Schreibermachwerk],” thereby reinstating their authority.11 Hence their interest in embracing it. Normative Redaction, Archetype and Original Textual criticism allows us to reconstruct the archetype of the tradition, which represents a constriction in the tradition attributable to a “normative redaction.” This reconstructed archetype, however, only gives us access to the oicial Brahmanic text resulting from the redaction of an earlier oral tradition. It neither accurately models the contents of the tradition nor can it be seen as a copy of the original, since the tradition was plural above the archetype and a single original never existed. Before we consider Bigger’s normative redaction hypothesis in greater depth, let us irst understand why the stemma codicum indicates a narrowing—but only apparently so— of the tradition. Consider V. S. Sukthankar’s stemma of the Ā diparvan (Figure 5).12 Vyāsa’s Bhārata Ur-Mahābhārata N S ν γ σ ε Śārada K Nepālı̄ Maithilı̄ Bengali Devanāgarı̄ Telugu Grantha Malayālam Ś1 K0–6 N˜ 1–3 V1 B1–6 Da1.2, Dn1–3, T1.2 G1–6 M1–5 D1–7 Figure 5 Textual tree of Ādiparvan versions, illustrating the stemmatic relationships 47 ARGUMENTS FOR A HYPERARCHETYPAL INFERENCE 47 The Mahā bhā rata editors used the common-error method, so called because it is based on the insight that only shared errors prove manuscripts related, to identify the genealogical relationships between the witnesses.13 Once these relationships were identiied and the manuscripts it into a stemma, they used simple rules to determine the reading of their probable ancestor, known as the archetype. Because all extant manuscripts were descended from a single source, by eliminating innovations or secondary readings unique to a speciic branch of the tradition, the editors could arrive at a codex unicus, the unique manuscript that represented their best conjecture of the text of this source. The circumstance that the editors reduced the multiple readings in evidence to a single one for every line of the Mahā bhā rata, thus creating a unique text, explains the apparent narrowing of the tradi- tion as we ascend toward the archetype. We must remember, however, that this narrowing is only apparent. In reality, the archetype was not the sole source in existence at the time. It just so happens that its descendants rather than those of other manuscripts survived, giving it its apparently unique position in the stemma. If we wished to represent the “true” state of afairs, our stemma would probably look like Figure 6. From this diagram, it should be immediately clear that the archetype was nei- ther the sole exemplar in existence at the time nor does it represent a constriction of the tradition.14 It is therefore incorrect to assume that the tradition was plural above the archetype, but underwent a reduction at the time of the archetype, as the many versions in existence were assimilated or standardized into one oicial copy. Yet Bigger commits this precise error. From the fact that the archetype occupies the vertex of the stemma, he concludes that the tradition was reduced to a single exem- plar, and uses this fact to open up the tradition again above the archetype. Let us read his own words: The MBh, which presents itself in the manuscripts, goes back to a uniform redaction that was ixed in a written form at some time and to which further texts and streams [sic] were added by later copyists. In that case a critical edition makes sense, since it represents an attempt to reconstruct this normative redaction.15 Concurrent with this redeinition of the archetype as a “normative redaction” (and the identiication of this so-called redaction itself with a real, historically existing text),16 Bigger also argues for reconsidering the relationship of the archetype and the original. Rather than view the archetype as editors have traditionally done—that is, as a copy, however remote, descended from the original so that if we can reconstruct it, we will come as close as possible to the author’s text17—he argues that it represents a deliberate revision of the original. Indeed, this revision is so extensive that it represents a break in the transmission and the original epic survives only as fragments that the normative redaction could not erase.18 Bigger explains: At the time of its creation, the normative redaction was not the sole version of the Mahā bhā rata. Rather, there existed other versions parallel to it, which, however, were suppressed in the course of time by the normative redaction. It is conceivable that, in its time, the normative redaction was the sole written version of the Mahā bhā rata, which allowed  48 newgenrtpdf K 55 13 56 57 72 73 1 14 46 47 71 9 12 15 70 78 79 48 49 50 51 52 53 2 22 74 80 10 11 40 43 44 54 75 81 82 3 4 16 25 26 27 76 77 83 84 85 5 17 18 19 39 41 61 20 28 29 30 42 45 58 63 21 23 24 59 60 62 64 65 66 6 7 8 31 32 33 34 35 36 37 38 67 68 Figure 6 The “real” stemma 49 ARGUMENTS FOR A HYPERARCHETYPAL INFERENCE 49 it to dominate the written tradition of this work. In the course of time, however, diferent copyists inserted passages from other versions—partly from direct recollection, partly from other (younger?) written versions—into the transcript.19 If we imagine the tradition as two cones placed one on top of the other apex to apex, we come closest to understanding Bigger’s view of the tradition. Above their common vertex, the tradition was oral and plural. Rather than a single authorial text, several versions existed. The normative redaction interrupted this variable tradition and created, instead, a single text. It not only led to a reduction in the number of copies or versions in circulation but also exercised a normative efect on their contents in the sense that, henceforth, only the authorized version (or versions in conformity with it) was preserved. Repeated copying of this new version caused the tradition once again to spread beneath the archetype, but the relationship of the archetype to the original is no longer that of a copy and its source. Rather, the archetype represents a new tradition connected only tenuously to the earlier one through elements accidentally preserved in the archetype. Proceeding in reverse, we irst encounter several manuscript versions of the Mahā bhā rata. As we successively eliminate their unique readings, the tradi- tion keeps narrowing until we reach their common source, the archetype. However, because this work is neither the irst origin of the tradition nor a faithful reproduc- tion of the preceding tradition, the critical edition does not provide us an accurate insight into the Mahā bhā rata.20 It merely occupies a prominent place on the stemma because it reconstructs a deliberate redaction that reduced the number of copies in existence. The Mahā bhā rata editors erred in privileging this work, and, in at least one instance, for similar reasons as its hypothetical architects, the Brahmans.21 Using “the methods of higher textual criticism,” we can and must go beyond its text to at least a partial original by identifying remnants of the older oral tradition surviving either in the archetype or through other means.22 As “the most complete collection of Mahā bhā rata versions,” the critical edition presents an invaluable aid in this task, but no more than that.23 Criticism: Higher and Lower The redeinition of the constituted text as a normative redaction rather than an archetype permits us to recon- struct earlier stages of the tradition using “higher criticism.” In contrast to textual criticism, which is a rigorous and mechanical procedure that begins with the manuscript evidence and attempts to infer the manuscripts’ likely sources based on shared errors of transcription, higher criticism uses subjective, a priori criteria to identify cer- tain passages as older than others and therefore as part of the “genuine” epic tradition. Although every single one of Bigger’s contentions is false—the critical edition does not attempt to reconstruct the Mahā bhā rata’s “normative redaction”; there is no evidence that such a redaction existed; it is erroneous to assume that the archetype was the sole exemplar in existence at the time; the apparently unique nature of the archetype must not be confused with a real reduction in the number of exemplars in circulation; and it is illicit to use this hypothetical reduction to open up the tradition once again above the archetype, giving the tradition a funnel shape—we nevertheless take a closer look at his 50 50 PHILOLOGY AND CRITICISM work, as it permits us to delineate a truly critical approach from one that merely appears critical. We shall focus on the distinction between historical criticism, sometimes called higher criticism in contradistinction to lower criticism and identiied with the reconstruction of stages before the archetype,24 and textual criticism—the stemmatic method proper, which culminates with the reconstruction of the archetype.25 Textual criticism proceeds from the assumption that our manuscripts are copies of an original, whether extant or not, and seeks on the basis of the genealogical relationships between them to identify their descent from it. If the original is no longer extant, these relationships can be used to conjecture the reading of the archetype, the (latest) common ancestor from which all extant manuscripts are descended, and to recon- struct this archetype by eliminating the innovations unique to one branch of the tradi- tion or to speciic manuscripts. We begin with the witnesses and attempt, on the basis of their evidence, to reconstruct the text from which they derive (the ancestor or source). For example, if we have three witnesses A, B and C arranged in a tripartite stemma of the form α / | \ A BC and A and B ofer the reading x and C ofers the reading y, it is likelier that A and B pre- serve the original reading and C contains the innovation, provided neither A nor B is the source of the other’s reading (horizontal transmission or contamination) and x is not such that A and B could have arrived at it independently (simultaneous innovation, polygenetic innovation or polygenesis of errors). Fitting manuscripts into a stemma thus presents editors a mechan- ical way of choosing between the available variants for a variation place rather than using subjective criteria such as sense or style.26 Above all, it lets us identify latent errors, that is, readings that, due to their unexceptional nature, might have escaped our notice as innovations over the original in the absence of the stemmatic method. Gianfranco Contini describes stemmata as “an objective and mechanical tool, invented to sort out, in the irst instance, the quarrel between variants that are in themselves equally acceptable: [...] variants that are pronounced erroneous serve to brand as erroneous those that have remained equally acceptable.”27 Paul Maas irst formalized the principles by which editors make these choices in his classic Textkritik. He used the hypothetical stemma shown in Figure 7 to illustrate the basic principles. We do not review these principles here, but it is important to note the special status of α in his stemma. Maas noted that α, as “the source from which the irst branching occurred,” is called the “archetype” and that “the text of this archetype is free of all errors that arose after the branching; it is thus closer to the original than the text of all witnesses. If we thus succeed in securing this text, the constitutio is signiicantly advanced.”28 Although Maas conceded that in three of the four possible outcomes—“recensio thus leads either to an extant codex unicus or to an archetype that can be reconstructed with certainty or to two variant texts that are either extant or can be reconstructed and ensure the text of the archetype only when they agree but not when they vary”—we must examine the 51 ARGUMENTS FOR A HYPERARCHETYPAL INFERENCE 51 (Original) α (Archetypus) β (Hyparchetypus) γ (Hyparchetypus) (K) δ A B C (D) E ε F H G J Figure  7 Maas’s hypothetical stemma, illustrating the distinction between hyparchetype, archetype and original Source: Modiied from Maas, Textkritik, 7. transmission to see if it is original, reconstruction typically aims not at an original, but at an archetype.29 Thus he noted: In this examination, the transmission proves to be the best conceivable, or equivalent to others we can conceive, or worse than others we can conceive but still tenable, or untenable. In the irst of these four cases the transmission must be considered original, in the last one as corrupt, in the two middle ones one may or must doubt. […] If the archetype of an entire work proves completely free of corruptions, it can be the original, that is, the branching could have begun with the original itself. I know of no major work of a classical author in which this possi- bility must be considered.30 Maas allowed a limited scope for divinatio (conjecture), through which the editor can attempt to improve the text and thus in some respects come closer to the irst source of the entire tradition,31 but he remained skeptical of historical criticism: The methods of historical source criticism are closely related. But whereas the literary trans- mission leads back to an original essentially identical with all witnesses in that it is likewise 52 52 PHILOLOGY AND CRITICISM a manuscript, at the beginning of the historical transmission stands an event that by its very nature resists being given a written form and is already colored or falsiied by the irst witness, indeed often intentionally so. Whereas the literary art work’s organic nature, which we expe- rience as necessary in every element, can survive millennia without severe damage, espe- cially in a culture subject to its efect, often only the roughest outline of the historical event transcends doubt, and often not even this. […] But nowhere will the path be as clear, the goal as attainable as in textual criticism of the classical authors.32 Whereas textual criticism typically culminates with the reconstruction of the arche- type, the German Mahā bhā rata critics saw the critical edition as merely preliminary in their search for the hypothetical original epic. They argued that it represented a beginning or a base from which they could reconstruct, using criteria such as style, hypothetical concerns and social and ideological features, the heroic oral epic posited as the Mahā bhā rata’s distant and lost source. Here is how Bigger encapsulates the distinction: Lower criticism alone demonstrates a passage as not belonging to the normative redaction. Higher criticism cannot change anything regarding this inding. But it can help interpret this passage and locate it text-historically. One could say lower textual criticism is the scenery on a stage, which fundamentally structures the space. In contrast, higher textual criticism is [like] the lighting, which admittedly cannot alter the scenery and its structure, but can enlarge and intensify its efect through casting the appropriate light on it.33 We later examine Bigger’s arguments from higher criticism for reconstructing an ear- lier stage of the transmission than the critical edition. But note the redeinition of tex- tual criticism in terms of lower criticism as merely a propaedeutic to higher criticism. Textual criticism does not “alone demonstrat[e] a passage as not belonging to the norma- tive redaction.” It gives us a reconstructed archetype that is the latest common ancestor of all extant manuscripts and, barring its errors of transcription, a descendant of the author’s copy of the work. It tells us about the reading of this archetype, rather than merely determining which passages belonged to it.34 It does not say anything about their belonging to a “normative redaction,” because no evidence exists for such a redaction and it is prejudicial to redeine the critical edition thus. Higher criticism not only “cannot change anything regarding this inding”; it also cannot “help interpret this passage and locate it text-historically.”35 As we saw in the introduction, absent extrinsic historical evi- dence only lower criticism permits us to evaluate the antiquity of readings. Indeed, all Bigger means by “interpret” is that we can determine whether a given passage relects a Kṣatriya or a Brahmanic worldview. In contrast, we saw in the introduction that a study of the text’s diachrony must begin with the archetype and what develops from it as the history of the text and not earlier. The idea that, beyond the reconstructed archetype, we can identify a textual history for certain passages is a misconception, arising from the German critics’ faith that they can identify remnants of the oral epic based on subjective and a priori criteria. Insofar as it seeks to demonstrate that readings rejected from the constituted text are older than the archetype, higher criticism is not even a method, but merely a bombastic name for their prejudices. Sukthankar succinctly summarizes these prejudices: 53 ARGUMENTS FOR A HYPERARCHETYPAL INFERENCE 53 A careful analysis of the poem from this view-point reveals the fact that in its present form at least, the work has a radical defect in so far as it consists fundamentally of two mutually incompatible elements, namely, a certain “epic nucleus” and an extensive and undigested mass of didactic-episodical matter, elements which are but loosely hinged together and which form moreover an unbalanced combination. The irst element, the epic nucleus, is naturally the older component and is presumably based on an historical reality, which is pre- served in a highly distorted and tendentious form but which retains nevertheless certain gen- uine archaic features in fossilized condition such as polyandry and levirate, which latter are of immense interest and importance for the study of Indian ethnology and prehistoric antiq- uity. The nucleus mentioned above was now unfortunately used—or rather misused—by wily priests, tedious moralists and dogmatizing lawyers as a convenient peg on which to hang their didactic discourses and sacerdotal legends, which have naturally no organic connection with the epic nucleus. This nucleus of the epic, a Kṣatriya tale of love and war, does possess a sort of unity, which is lacking entirely in the other element, the priestly episodes and the moralizing discourses, which latter by themselves, loosened from their moorings, would neatly and automatically fall apart. The epic story is in part at least a fairly well-constructed narrative, worthy of our attention, and produces the impression of having been yet more virile—a real “human document”—before it was distorted in the process of assimilation with the moralistic pabulum and legal claptrap of a grasping and degenerate priesthood.36 His review of the German scholars’ so-called higher criticism ironically mocks the German “savants”: The Mahā bhā rata is in short a veritable chaos, containing some good and much useless matter. It is a great pity that a ine heroic poem, which, may even be found to contain precious germs of ancient Indian history, should have been thus ruined by its careless custodians. But it is not quite beyond redemption. A skillful surgical operation—technically called “Higher Criticism”—could still disentangle the submerged “epic core” from the adventitious matter— known to textual critics as “Interpolation”—in which it lies embedded. The Mahā bhā rata Problem thus reduces itself to the discovery of criteria which will enable us to analyse the poem and to dissect out the “epic nucleus” from the spurious additions with which it is deeply incrusted. This is the “Analytical Theory” of the origin and the character of the Mahā bhā rata, which was espoused by the majority of the Western critics of the Great Epic of India, chief among them being Lassen, Weber, Ludwig, Sörensen, Hopkins and Winternitz.37 We do not trace further criticisms of the German scholars’ so-called higher criticism. It suices to note that Bigger’s attempted reinterpretation of the constituted text as the Mahā bhā rata’s “normative redaction” is not disinterested.38 It actively undermines the relationships on which the critical edition is based for a racially and anti-Semitically charged hypothesis of a Brahmanic takeover of a Kṣatriya epic.39 The normative redac- tion difers from both the archetype and the original in that it refers neither to an appar- ently unique source from which all surviving witnesses are descended nor to the irst source of the entire tradition or a copy of this irst source. Once this hypothesis is granted, the stemmatic reconstruction cannot stand. The critical edition becomes pre- cisely what Bigger and von Simson declare it: evidence of Brahmanic mischief. In the next three sections, we examine Bigger’s arguments for a Kṣatriya layer. We show that none of them meets standards for objective, dispassionate analysis. Indeed, not one is text-critical. 54 54 PHILOLOGY AND CRITICISM The Argument from Spread and the Argument from Resilience Textual criticism only permits us to reconstruct the source of the irst branching. However, this ignores the possibility that the tradition was plural before the apex. The reconstruction of the archetype is thus an error, arising from the fallacious assumption that all readings derive from a unique text (argument from spread). Our reconstructed archetype is based on readings that passed through the constriction between the two cones, whereas readings that fell directly from the upper cone to the lower without passing through the apex could be older (argu- ment from resilience). Refutation of the argument: The argument from spread fails because it overlooks that the shape of the tradition before the archetype is irrelevant for the reconstruction of the archetype, which remains ex hypothesi the oldest ancestor of the extant witnesses that can be reconstructed. In fact, the archetype makes no claims about whether the tradition spread away from the original or all readings fell through one particular manuscript. It only claims that all our manuscripts are descended, however remotely, from this one source. Likewise, the argu- ment from resilience fails because it erroneously infers the existence of a real constriction in the tradition from the archetype’s apparently unique position in the stemma and further argues that some readings at least could have escaped the archetype’s constricting efect. Neither von Simson’s nor Bigger’s claims are tenable on stemmatic grounds, but to grasp the precise source of their error we must irst understand exactly how they redeine the critical edition. As we have seen, neither thinks the critical edition reconstructs a text descended more or less immediately from the irst source of the entire tradition. Likewise, they do not think the critical edition represents the archetype in the sense of the latest common ancestor to which the extant witnesses owe their text. In one sense, the witnesses owe their text to it, because they are clearly descended from it, but, in another sense, they do not owe their text to it, because some of their readings—speciically, the passages they identify as “Kṣatriya” passages—derive from an earlier, oral epic tradition. Von Simson and Bigger use this circumstance to claim that, whereas the archetype—or, as they call it, the inal or the normative redaction40—exercised a constricting efect on the tradition, controlling the number of Kṣatriya elements that lowed past this constric- tion to its descendants, it is the irst source of only some of their readings. Crucially, it is not the source of the readings that lowed through it. Reconstructing the archetype therefore gives us access neither to the tradition before it nor to the tradition after it, since we owe only a part of this tradition to it. In fact, the only moment in the text’s history the archetype accurately models is the moment of its inception—precisely the moment the redaction occurred.41 Hence its limited value. Mutatis mutandis the critical edition is useful only as a base for further investigations. It recovers a certain stage of the transmis- sion, permitting us to identify everything included at this point in the text, but we must be prepared to unfold the tradition both before and after it, identifying, on one hand, everything eliminated from it and, on the other, everything composite in it suggesting remote origins. Although unexceptionable from the perspective of their theory of editing, several problems exist with von Simson and Bigger’s interpretation of the critical edition. Let us address them individually. First, the view that the archetype is not the source of all the readings it contains, since some of them originated with its ancestors in a chain of manuscripts leading back to the original is true, but irrelevant for the deinition of the archetype, which remains the latest common ancestor to which the extant witnesses 5 ARGUMENTS FOR A HYPERARCHETYPAL INFERENCE 55 owe their text. Second, the idea that the archetype’s apparently unique position in the stemma derives from its constricting efect on the tradition is a misconception. As we saw, the narrowing of the tradition as we ascend toward the archetype is only apparent. The impression of a constriction only arises because von Simson and Bigger open up the tradition before and after the archetype, giving it the form of an hourglass, but not at the archetype. Third, insofar as they are aware of it, they misinterpret Maas’s deinition of the archetype as “the source from which the irst branching occurred.”42 They think it correct provided it is applied only to the tradition after the archetype, that is to say, the archetype is the source from which the irst branching of the later tradition occurred, but because it is not the source from which the irst branching simpliciter occurred—which, if it occurred at all, would have begun with the original or the moment of its irst copying— it is erroneous to treat it as the archetype. The archetype is the irst starting from below or it is the irst before the extant tradition, but because editors reconstruct only the tra- dition from below, they overlook its perfectly provisional position. As Michael D. Reeve has already addressed all three fallacies in an article on reconstructing archetypes, let us look at his work.43 Reeve examines Enrico Flores’s argument for treating the archetype not as the source of conjunctive errors common to the entire surviving tradition, but as a manuscript that editors reconstruct by accident because they approach the tradition from below.44 He irst notes Flores’s problematic deinition of the archetype: Before tackling the fallacy itself, I must say how Flores understands “archetype.” He begins by citing Maas’s deinition, “the exemplar from which the irst split began” but takes it out of context. […] In context, […] “the irst split” means the earliest of the splits revealed by analysis of the extant witnesses; but “in reality,” says Flores, “the irst split or splits, a fact apparently all too often forgotten, began from the moment when a text was written down and started to circulate and spread in a number of copies.” He therefore understands Maas’s “irst” as the irst from below, or in other words the irst reached in the process of reconstruction from the extant witnesses; but since in this process many points of convergence may be reached before the archetype, he has to gloss “the irst” as “the irst before the extant tradition.” […] Equipped with this lopsided distinction between the irst from above and the irst from below, Flores goes on to allege that scholars have deined “archetype” either from above as “the oldest witness, a lost one, common to the whole manuscript tradition of a text” (presumably he means all the manuscripts that have ever existed) or “the lost witness from which descend more or less imme- diately all the oldest witnesses that survive and the rest of the derivative manuscript tradition.” The second deinition, as he says, is Maas’s; but whose is the irst? Quite apart from its poor wording (since a witness cannot be common to a tradition), what it deines is the original, unless the original is preserved or only its descendants count as witnesses. When Flores says that for Maas the two deinitions are equivalent […], how is he taking this irst one?45 Reeve resumes: Already, then, Flores’s remarks about the meaning of “archetype” betray misapprehension and confusion. Confusion persists when he tries to prove Maas’s notion of “archetype” inco- herent, and it is in these attempts that he repeats in various forms a fallacious argument about conjunctive errors. […] In a later passage, Flores sets out to show that the origin of 56 56 PHILOLOGY AND CRITICISM conjunctive errors can be pushed back ever higher in the transmission, so high indeed that it may lie in the original. […] To that end he constructs an example in which three witnesses, ABC, share signiicant errors, xyz, “which take one back to the archetype α.” α A B C x, y, z x, y, z x, y, z For the moment he is using ‘archetype’ in a sense that he attributes to his opponents, “witness without which ABC would not have existed”; and from the example he argues that α is not “the only copy that survived, without which ABC would not exist.” Actually the conclusion follows immediately from the deinition because any ancestor of α, for instance the original, would satisfy it no less than α. Before I  examine this argument, therefore, the deinition must be tightened so that α is the latest witness without which ABC could not have existed. That is essentially how Maas deined “archetype” and surely how must people deine it; and Flores too, since he is discussing one reconstructed witness, namely “the archetype α,” and not a line of reconstructed witnesses, presumably had that deini- tion in mind. His argument runs as follows: since all the ancestors of α as far back as a copy of the original could have had the error z, it may well be that ABC derive in various ways from that copy without any need to postulate an archetype. By deinition, however, B (say) cannot have inherited any of the errors xyz from an ancestor of α down a line of tradition that did not pass through α; and if ABC derive in various ways from a copy of the original but not all through any later witness, then that copy of the original was α.46 So far, von Simson’s and Bigger’s views resemble Flores’s in that they think the identii- cation of the archetype is an error arising from the fact that editors reconstruct the tra- dition from below. Like Flores, they think what the critical edition reconstructs cannot be the source of the irst branching, because several such existed. More precisely, the critical edition cannot do so because, given the tradition’s oral nature, a irst branching never existed: what the critical edition therefore reconstructs is the irst branching of the extant tradition.47 More important, they understand archetype, like Flores, to mean “the oldest witness, a lost one, common to the whole manuscript tradition of a text,” that is, the original, albeit of only the manuscript tradition. The second stage of their argument again resembles Flores’s in that they seek to show that the origin of certain readings can be pushed back ever higher in the transmission, so high indeed that it may lie in the original, except, whereas in Flores’s case xyz referred to conjunctive errors, for von Simson and Bigger, they represent remnants of the oral epic predating its redaction. Bigger writes: [T]he passages rejected from the constituted text are by no means insigniicant. From the perspective of a theory of a normative redaction that was reconstructed by the critical edition, some of them represent a parallel transmission. The normative redaction admittedly 57 ARGUMENTS FOR A HYPERARCHETYPAL INFERENCE 57 suppressed the other versions in these cases. But some copyists incorporated parts of these versions in their copy of the normative redaction.48 Note the parallel. For Bigger as for Flores, since all the archetype’s ancestors as far back as a copy of the original could have had the reading z (a remnant of the Kṣatriya epic), ABC could derive in various ways from that copy—in fact, he explicitly notes, “some copyists incorporated parts of these versions in their copy of the normative redaction”— without any need to postulate an archetype. With the exception of one possibility we must examine later—z could have entered the witnesses without passing through the archetype—the argument is exactly the same and Reeve’s objection holds. By deini- tion, the witnesses cannot have inherited any of the readings xyz from an ancestor of the archetype down a line of tradition that did not pass through it; and if ABC derive in various ways from a copy of the original but not all through any later witness, then that copy of the original is the archetype. Let us look at Flores’s third argument for redeining the archetype, for it presents the closest parallel to Bigger’s work. Reeve writes: Flores’ third version of the fallacy about conjunctive errors occurs earlier in his book than the other two. […] Though hypothetical again, xyz this time are not errors but manuscripts, and from an archetype θ they inherit the errors abc. Some of these errors, however, are older than others. The error a arose in the original and passed on to all its descendants, θ included. The error b arose in a copy of the original called α and passed to β and any other copies of α that there might have been; it too reached θ. “Ma θ” (and to this sentence he adds an exclama- tion mark) “non è la somma delle serie complete” [is not the sum of the complete series] of all the manuscripts copied from the original, from α, from β, and so on. It derives ultimately from α, because it has the error b, but is wrongly imagined to be “l’unico sopravvissuto di una serie sterminata di atti di copia” [the unique survivor of an endless series of acts of copying]. If we move up the stemma from θ to α, we shall keep inding errors common to the whole tradition, but their number will gradually fall until in the original either none remain or only those that arose in it. The archetype, therefore, since it is always constructed “dal basso e dalla tr. ms. conservatasi, come portatore degli errori comuni (o congiuntivi)” [from the bottom or from the manuscripts that have been preserved, as the bearer of the common (or conjunctive) errors], will always be a copy “apparentemente unico” [apparently unique], and he spaces out “apparentemente” for emphasis. He goes on to call θ a constriction, the bottom of one funnel and the top of another inverted below it: the errors abc drop to the bottom of the upper funnel and by dispersing in the lower funnel reach xyz. The problem, he says, is that the con- striction does not exist, because the errors in the upper funnel do not collect at the bottom but disperse. The apparent uniqueness of the archetype therefore results from the reductive logic of reconstructing a unique copy in reality non-existent “as such” the further one moves up from the extant manuscripts to the original. “That is the real explanation.”49 To understand Flores’s argument, consider the hypothetical stemma in Figure 8. The error a arises in the original ω and descends via α, β and θ to the witnesses XYZ. The error b likewise arises in a descendant of ω called α and descends via β and θ to the witnesses XYZ as well as to any other descendants of α that might have existed. The error c arises in θ and descends to the witnesses XYZ. Because we reconstruct the tradition “dal basso” 58 58 PHILOLOGY AND CRITICISM ω a α a, b first spread β θ (normative redaction) second spread X Y Z a, b, c Figure  8 Flores’s argument from the spread of errors, and Bigger’s normative redaction hypothesis (from the bottom), XYZ appear to owe the errors abc to θ, which thus appears to be the sum of the complete series of α, β and so on, even though it is only one of α’s descendants. Indeed, although “it ultimately descends from α, it is falsely imagined to be the unique survivor of an endless series of acts of copying.”50 Its apparently unique position can be explained by the fact that the errors ab and any others that might have arisen before the archetype appear to pass through it, even though they do not actually drop into the bottom of the upper funnel as the dashed lines suggest, but rather were transmitted also to α’s descendants and to whatever descendants of the other manuscripts before θ might have existed—a fact we cannot see because we begin with the extant witnesses and these other descendants did not survive. Stemmatic reconstruction thus always leads us to reconstruct an apparently unique copy—the archetype—even though the constriction is not real and this copy does not “really” exist, because it is not the source of all the errors attributed to it. Flores’s error can be dismissed easily, because since at least the nineteenth century editors have recognized that a reconstructed archetype does not imply a constriction or a bottleneck. Rather, as we have noted, the archetype is deined as a manuscript, detected through reconstruction, that stands at the apex of the stemma and from which the branching of the tradition began. In fact, if the descendants of the other manuscripts had survived, their common ancestor would in turn have been the archetype and we would reconstruct this manuscript regardless of whether we thought the tradition had an even greater spread. In other words, all Flores has done is posit a constriction in order to argue it did not exist, and all this to demonstrate that the archetype was not the sole manuscript in existence! The error arises from forgetting the deinition of the archetype, which has nothing to do with the real number of exemplars in existence, but merely speciies the point at which stemmatic reconstruction comes to an end. As Reeve notes: 59 ARGUMENTS FOR A HYPERARCHETYPAL INFERENCE 59 [T]he argument contains two fallacies, of which the one that I have been discussing is the irst. It occurs in his [Flores’s] remarks about moving up from the extant manuscripts to the original. First one meets the archetype θ, which had the errors abc, but as one moves up through β and α to the original, c and b fall away until only a remains. The argument seems to be like this: like θ, its ancestors β and α are reconstructed from below and therefore qualify for the name “archetype,” from which it follows that there was no single archetype θ. The fallacy, elementary again despite its complex setting, lies in not sticking to the original deinition of “archetype,” namely “latest common ancestor of xyz” or “latest witness to which xyz owe the errors abc,” but instead allowing β and α to count as archetypes even though ex hypothesi the oldest ancestor of xyz that can be reconstructed is θ, a descendant of β and α.51 Likewise, the second fallacy arises from forgetting that the shape of our stemma is unre- lated to the shape of the real, historical stemma, and if there was a reduction in the number of copies, then it is all the likelier that our manuscripts derive from a single source: The second fallacy in the argument occurs when he [Flores] introduces the image of the two funnels, which has been given memorable expression elsewhere […]: It is probably true to say that the classical tradition as it expands and contracts in its course from Antiquity to the end of the Renaissance does conform to a basic pattern. In its crude and essential form it appears to the imagination to follow the traditional lines of the hour- glass, which funnels down to a narrow middle and then bellies out again, or the simpliied shape in which the female form is often represented—broad shoulders, tiny waist, full skirt. The vital statistics of the igure will vary considerably from text to text; but these diverse patterns, when superimposed one upon the other, should still produce a dominant type. The slender waist is the most permanent feature, for the Dark Ages so constricted the low of classical learning that for a time it was universally reduced to a trickle. Flores transfers the image from history to stemmatic logic. For him θ could have been “of the 9th century or the 3rd–5th,” and what matters to him is its deinition as a single manu- script, something narrower, that is, than the network of its earlier or later relatives. Now it may be that if one were to represent schematically, with the original at the top, the network of all the manuscripts that had ever had any of the errors abc up to the production of θ, it would look more like a pyramid than a funnel; but why should the extant manuscripts not all derive from a single point in the pyramid, especially if, by the time that the earliest of them was produced, attrition had turned the lower section of the pyramid into something more like a funnel? When Flores exclaims that θ is not the sum of all the manuscripts that ever had any of the errors abc, who does he suppose holds that it is? Or is he demanding it should be?52 In either case, at no point does our stemma make a claim about the “real” shape of the tradition, whether pyramidal or funnel shaped. It simply makes the claim that our extant witnesses all derive from a single point in the tradition, and if more witnesses existed that did not derive from this point they would all derive from some other—higher—point on the stemma and this would then be the archetype. Once we understand that the shape the critic posits for the tradition is irrelevant to our reconstruction of the archetype, we can also address von Simson’s and Bigger’s criticisms. Both scholars think that the circumstance that the reconstruction culminates 60 60 PHILOLOGY AND CRITICISM with the codex unicus of the archetype is evidence that only one manuscript existed at the time. This manuscript, they reason, can only be of the Brahmanic version and the reduction in the number of exemplars in existence the result of a Brahmanic redaction of the epic’s oral versions. Whereas Sukthankar began from the extant witnesses and identiied the shared readings of independent branches of the tradition as the readings of their common ancestor, they argue that this approach leads only to a reconstructed ancestor that is the source of the irst branching of the extant tradition—that is, of Flores’s second spread. Actually, however, the editor should have kept in mind that the tradition before this ancestor was not single but multiple, since θ’s ancestors—that is, α, β and so on—also produced copies—Flores’s irst spread. The reconstruction “is a constriction […] at whose base the common errors a, b, c […] collect, and whence they fall from above so that we ind them in the surviving witnesses x y z.” Actually, however, “the constriction does not exist because the errors that fall from above do not collect at the base of the funnel, but fall dispersing in a radial pattern.”53 It is only here that the analogy between Flores’s and Bigger’s arguments breaks down, because, whereas Flores thinks the constriction is only apparent and uses this to argue against the historical reality of an archetype, Bigger thinks the constriction was real and uses this to redeine the archetype as a normative redaction. In his view, the stemmatic method only takes us back to the archetype, but, using the “methods […] of higher criticism,”54 we can demonstrate the existence of a “parallel tradition” of the Mahā bhā rata.55 Consider Figure 9. Abcdefghi are textual passages. The original ω contains a. A  copy of the original α contains abg, while other copies contain defghi (only one additional copy is shown at β). The creation of a normative redaction at θ artiicially constricts the tradition and only def make it past θ. The passages ghi are lost, but abc reenter the manuscript tradition via a parallel transmission. The surviving witnesses XYZ contain abcdef but not ghi, which are lost for good. Using the stemmatic method, the editor reconstructs the archetype θ in the belief that it accurately represents the tradition before it. Actually, however, θ is not the archetype but a conscious redaction; the editor restores def but not abc thinking they are later than def, even though θ is, as Flores notes, not the sum of the complete series of manuscripts copied from the original. But whereas, for Flores, θ only appears the sum of the complete series of manuscripts copied from the original because the descendants of α, β and so on did not produce any surviving witnesses, for Bigger, θ is not the sum of the complete series of manuscripts copied from the original because α, β and so on contained a diferent text than θ and produced descendants, albeit ones whose existence must irst be identiied in the surviving manuscripts.56 If the editor can identify these remnants of the earlier oral tradition in the larger manuscript tradition, he can reconstruct a partial text αʹ that includes abc (but not g, which is lost for good) and presumably also ω, the irst source of the entire tradition. We refer to this text as the supra-archetype in the following. Both Bigger and Flores employ an argument we might call the “argument from the spread of errors” to deny the existence of a single archetype. Both interpret the arche- type as a constriction in the tradition rather than as Maas deines it—that is, as “the source text with which the irst branching began”—and both use this confusion about the deinition to deny that all our manuscripts descend from a single archetype. Like 61 ARGUMENTS FOR A HYPERARCHETYPAL INFERENCE 61 a ω Original a b g α Copy d β Copy e c h f i Normative θ redaction Witnesses a b f e d c XYZ Figure 9 Bigger’s argument from the resilience of tradition Flores, Bigger assumes that the tradition consists of two “spreads.” His normative redac- tion corresponds to θ in Flores’s scheme. Like θ, the normative redaction exercises a constricting efect on the tradition. The increase in versions may begin with the orig- inal ω if multiple copies of the original existed or after the original if only one copy of the original existed, but because we construct the archetype “dal basso,” it appears as though all these readings passed through the archetype, when some at least could have descended through other means. Note that what we earlier called the irst spread is irrel- evant for Bigger, as he does not assume a single origin: for him, what is decisive is not the spread away from ω but the narrowing of the tradition into θ. Thus whereas Flores’s concern is that the errors abc could have arisen in the ancestors of the archetype θ (in ω, α, β and so on) before passing through it, Bigger argues that abc, which for him are not errors but “parallel versions” of the Mahā bhā rata, may not have passed through θ at all and could still be older than the material that passed through it. The argument from spread fails, as Reeve shows, because it continuously shifts the deinition of archetype, illicitly extending it to sources that can be posited, but for which no evidence exists, behind the archetype. Even if we posit a series of manuscripts α, β and so on before the archetype and suppose that the errors abc arose with the original or with one of its copies, all these errors would pass through θ such that α and β and whatever other sources we posited between ω and θ fall out of consideration. As the “latest witness [recte: source] to which xyz owe the errors abc,” θ remains the archetype. 62 62 PHILOLOGY AND CRITICISM Likewise, the argument from resilience fails because it illicitly redeines the archetype to mean the irst source of the extant tradition where, instead of meaning all that exists, extant tradition is taken to mean: that part of the tradition the Brahmans did not erase. Like Flores, von Simson and Bigger interpret the archetype to mean not the latest ancestor without which the surviving tradition could not exist, but an exemplar editors for some reason single out in a long line of witnesses that can be reconstructed. Like Flores, they attempt to transfer an understanding of history to stemmatic logic and, again like Flores, what matters to them is that the inal or the normative redaction is a single manuscript (Bigger calls it the “sole written version […] in its time”), narrower than the tradition before or after it.57 But as Reeve shows, what is decisive is not whether the dissemination of all manuscripts containing the errors or readings abc from the original to the archetype traced a pyra- midal or a funnel shape, but whether the extant manuscripts could have derived from a single point in the pyramid. In fact, a narrowing of the tradition such as Flores and Bigger posit makes it especially likely this was the case. But whereas Flores’s error arises from a simple albeit common confusion about the archetype, von Simson and Bigger use this confusion to posit a perverse and purposeful Brahmanic redaction of an earlier heroic Kṣatriya epic. They argue that this redaction decimated the other versions of the Mahā bhā rata in existence (speciically, those in conlict with the Brahmans’ ideology), and they claim that these versions represented the older and genuine epic tradition. They agree with Flores that the identiication of the archetype is an error, but not because the tradition is wider than our stemma indicates— in fact, they think, moving backward in time, that the stemma replicates the tradition precisely until the moment the normative redaction occurred. In contrast to Flores, who attacks the stemma because he thinks it does not depict the historical reality accurately, they criticize it because they think it does not represent a section of the tradition, equiva- lent in all its parts, but the whole of that tradition.58 In other words, they reject the distinction textual critics make between the arbre réel and the arbre généalogique,59 and argue that the interpretation of the critical edition as an archetype is false.60 For them, the stemma is not an idealized representation of a reality necessarily more complex such that it can stand in for any part or, maximally, the whole. Rather, it is a literal depiction of the epic’s travails after it entered the Brahmans’ hands. As the living image of the epic’s defacement after the Brahmans gained control over it, the critical edition is a necessary step for tracing this history of collusion, but no more than that.61 Despite their seeming self-evidence, von Simson’s and Bigger’s theories are easily refuted. First, the stemma is not an image of historical reality. As Ben Salemans argues, It would be wrong to interpret a stemma as an exact historical picture of the history of the deliverance of the versions. One should bear in mind that a stemma is a minimal picture relating only to the text versions that still exist. Thus, a stemma can only be considered as a hypothesis about (a part of) the historic reality. On and around the lines of descent, we can imagine lost manuscripts whose contents are unknown.62 Second, the stemma reconstructs the historical vicissitudes of transmission only in general terms. It does not let us pinpoint speciic events or contingencies because of the paucity 63 ARGUMENTS FOR A HYPERARCHETYPAL INFERENCE 63 of our data points. It can, for example, tell us that all our manuscripts are descendants of a common source; it cannot tell us exactly when this source was created or who its author or scribe was. If we imagine that the 1,319 manuscripts in the stemma in Figure 6 repre- sent all the manuscripts that ever existed and manuscripts 1–85 our extant witnesses—the groups Ś ā radā through Malayā lam—then Sukthankar’s stemma represents the minimal architecture necessary to connect the manuscripts based on what we observe about their shared errors of transcription. Performing eliminatio codicum descriptorum and extending the lines from our reconstructed hyparchetypes backward to their hypothetical source, we would arrive at the stemma in Figure 10.63 The posited hyperarchetypes are nodes from which the transmission is assumed to have branched,64 but they do not resemble a historical exemplar in all its particulars. It is therefore false to identify a reconstructed archetype with a historically existing source and this itself with a Brahmanic redaction.65 Third, the constriction does not exist, as we know. If we wished to trace the real tree, which would include not only many more exemplars before and after the archetype but also alongside it, we would arrive at either no shape at all or a meaningless rectangle—our sheet of paper.66 Real trees trace complex shapes.67 Although it is tempting to relate their shape to the trans- mission, we must remember that what they depict is a two-dimensional representation of a reality that unfolded in space and time.68 In Figure 10, we retained the positions of the main nodes from the “real” tree. Actually, since we do not often draw stem- mata using real trees as our template, our stemma would be more abstract as shown in Figure 11. Fourth, even if we had access to the real tree and a part of it depicted a severe dec- imation of the tradition, we could not so easily conclude that this was due to an ideo- logically motivated redaction. The causes of manuscript loss are complex. We could not say anything about a sudden reduction in the number of exemplars in existence (ire? destruction of a library? abandonment of a workshop?) without corroborating histor- ical information. The German scholars do not know who the patrons were, what their tastes were or who the scribes were. They know nothing about the conditions of copying or even the main centers of copying. Yet they postulate a generic “Brahmanism” as the explanation for all manner of textual decay. The only reason their theses appear compel- ling is the biased and question-begging redeinition of the critical edition as a normative redaction.69 But, ifth, even if we grant the German critics that the tradition traced exactly the shape they postulate—a cinched skirt with the inal or normative redaction occurring at the exact point the sides are drawn in—how do they know that the critical edition reconstructs this source? Bigger claims that “the critical edition […] represents an attempt to lay hold of this normative redaction.”70 Johannes Bronkhorst does not think “the written archetype of the surviving manuscripts” reconstructed in the critical edition is “identical with the irst written version of the Mahā bhā rata,”71 but only because he wishes to postulate a irst written version completely devoid of Brahmanic philosophy.72 In his view, the archetype was more likely a popular commentarial version, whose “pres- tige” caused it “to overshadow all others.”73 Either way, he thinks the text reconstructed in the critical edition represents a consciously redacted archetype. “[T]here are good  64 newgenrtpdf W O N P S F M 55 xV R 13 E J L 56 57 Q 72 U 14 D 46 47 71 T 15 I 78 70 79 K 74 80 C 22 H 40 43 44 81 16 B 25 26 27 G 18 19 A 39 41 23 24 Figure 10 The stemma as a minimal architecture 65 ARGUMENTS FOR A HYPERARCHETYPAL INFERENCE 65 θ W O S N P F M 55 R V 13 E J 56 57 L Q 73 U 72 14 46 D I 71 47 81 T 79 K 70 74 C 15 22 H 78 80 40 43 44 25 26 27 G 16 B 39 41 18 19 A 23 24 Figure 11 Our abstract stemma reasons to think that the core of the Mahā bhā rata of the critical edition was composed and written down at some time during the two centuries preceding the Common Era. Parts were subsequently added until approximately the time of the Guptas, when the archetype of our critical text was established.”74 Apart from the infelicitousness of the expression “the archetype of our critical text,” we now know that the idea that an arche- type is “established” is a fallacy. The archetype is a manuscript identiied “by chance” as the latest common ancestor of the extant witnesses: it does not represent either a more authoritative version or one assiduously copied.75 It just so happens that its descendants, rather than those of other manuscripts, survived. It is therefore false to claim that its text represents “a Brahmanical reworking of earlier material” and the like.76 It is especially misleading to argue that the critical edition reconstructs just this exact source, as though through some magical leap it would reach exactly this point—the assumed constriction— in the tradition. Consider Figure 12. Even if we grant the German critics that a real constriction occurred, whether called the inal redaction (von Simson), the normative redaction (Bigger) or the bottleneck (Bronkhorst), how do they know the critical edition reconstructs the source of this con- striction? The critical edition reconstructs an archetype that occurs somewhere in the tra- dition, but whether it occurs higher up or lower down in the tradition will depend on the manuscripts that survived.77 As the apex of the stemma, the archetype could be located at any point on the real tree. When the German scholars speak of a constriction in the 6 66 PHILOLOGY AND CRITICISM ω ω Original Original θ Archetype Normative Normative redaction redaction N S θ Archetype Witnesses K V B D T G M Witnesses Figure 12 Making the archetype and the normative redaction coincide tradition or the “irst written version” of the Mahā bhā rata, however, it is the apex of the real tree they mean. The only condition under which the two—that is, the apex of the small triangle representing the part of the tradition represented in our stemma and the apex of the larger triangle representing the Brahmanic tradition following the hypothetical establishment of an “oicial text”—could coincide is if no part of the tradition was ever lost. In that case, the archetype of the tradition would be identical with the original, and what the critical edition reconstructs would not be an archetype identiied “by chance” during recensio, but the archetype as the source from which all manuscripts that ever existed descended. This is represented on the diagram’s right: the dotted triangle is displaced minimally to show that two triangles exist.78 In other words, what the German scholars have done is take our stemma, which represents a part of the tradition only, and identiied it with the whole of that tradi- tion. In their view, no part of the Brahmanic tradition was ever lost such that when we take the extant witnesses and attempt to reconstruct their lost source, the manu- script we arrive at is not a random exemplar located somewhere lower down on the tree, but the irst source of the entire tradition. Bigger’s identiication of the critical edition with a hypothetical normative redaction holds on this condition and this condition only. From the German scholars’ perspective, the critical edition’s sole law is that it overlooks the fact that this work is not the irst source of the Mahābhārata tout court, but only its irst Brahmanic revision, and they seek to ameliorate this by positing a second triangle that is the reverse of the irst. Like Flores, they think they can thereby ill in the “vast expanse between archetype and the original text.”79 The only problem with this thesis is: all their arguments for a Brahmanic redaction are a priori, based on what they think the work is.80 No evidence exists for such a redaction, and the only reason it appears plausible is that the German scholars have redeined the critical edition as a Brahmanic redaction. 67 ARGUMENTS FOR A HYPERARCHETYPAL INFERENCE 67 The evident circularity of this procedure did not hinder them from drawing “evidence” from the critical edition of this hypothetical redaction. On the contrary, proceeding from the premise that the archetype reconstructed in the critical edition represents a nor- mative redaction, they identiied strata, interpolations, further redactions and so on— conclusions that “logically” followed from their redeinition of the critical edition as a Brahmanic redaction.81 This circularity, rather than any arguments, explains the seeming self-evidence of their work. In their unchaste desire to see a Brahmanic redaction at work in the text, the German critics do not realize that they themselves created its reality! Despite the German critics’ complicated analyses, their error actually relects a simple misunderstanding: a failure to keep separate “two conceptual spheres that we need to keep distinct: the history of the tradition, with its unattainable real trees, that is, the often very rich ensemble of mss. that historically existed, including those that disappeared without leaving traces, and textual criticism, with its very tangible although perfectible stem- mata codicum, based on the few mss. that have come down to us.”82 When the German critics use the term archetype, they mean it in the irst of Trovato’s two senses: archetype1, the “oicial text, prepared by its author for publication,” but they then confuse this with the critical edition, which is an archetype in the second of Trovato’s two senses: archetype2, a manuscript that “can be reconstructed logically—with varying degrees of accuracy—on the basis of the number of conjunctive errors that are common to all its descendants.”83 From this identiication of the irst sense of archetype with the second one, they then conclude “a sudden and inexplicable bottleneck in the ancient and medieval tradition whereby only one copy survived,”84 and identify this bottleneck not with a chance reduc- tion in the number of exemplars as has been postulated for the classical tradition,85 but a willful and malicious Brahmanic “takeover” of an earlier, heroic oral epic tradition.86 It is but a short step thence to establish a “prehistory of the normative redaction”87—a prehistory that, according to the German scholars, cannot but be the history of the Brahman domination of India.88 Once we remember that no evidence exists for such a redaction and that the con- stituted text is the earliest text that can be reconstructed using objective, noncircular methods, their work loses its aura of unimpeachability. Since we have no evidence of a Brahmanic “takeover” of the epic beyond what the German critics erroneously deduced from the apparent constriction in the tradition at the archetype, it is best to interpret the critical edition as what it is: a reconstructed archetype, where archetype simply means the latest common source from which all extant manuscripts more or less immediately derive. In the next section, we examine Bigger’s arguments for claiming that abc are older than θ. The Argument from Empty Reference The critical edition contains references to events whose description no existing passages meet. However, these “empty references” must once have had a referent. If suitable descriptions occur in the apparatus, we may assume that they were the original referents, contained in a version of the Mahābhārata distinct from the nor- mative redaction. The critical edition discards these passages as unique to one branch of the tradition, but this is false because they are actually older than the normative redaction it reconstructs. The normative redaction eliminated the passages, but they later reentered the tradition, thus occurring in some manuscripts only. Restoring the referent to the constituted text restores the text to a state before the normative redaction. 68 68 PHILOLOGY AND CRITICISM Refutation of the argument: The argument sufers from the fallacy of proof by assertion, because it is tan- tamount to asserting that a passage A is old because it is considered old. Its fallacious nature is not immediately apparent because, rather than state it simply, Bigger asserts that A is actually Rʹ, the referent of passage R, whereas the reading R → Rʹ reconstructs an original R → Rʹ; as the surviving referent of this original refer- ence Rʹ is thus old. The equivocation between Rʹ as a passage in the extant witnesses thought to correspond to a reading R (that is, Rʹ1) and Rʹ as the referent of an original R (that is, Rʹ2) permits him to “infer” Rʹ’s antiquity, even though, actually, he has only asserted it. The empty reference plays no role in demonstrating A’s antiquity; it merely conceals the illegitimate shift from Rʹ1 to Rʹ2. Let us return to Bigger’s thesis. As we saw, he interprets the archetype’s apparently unique position in the stemma as a real constriction and, from this, infers a Brahmanic “redaction” of an earlier oral epic tradition. More important, he argues that some readings could have escaped the archetype’s constricting efect if they fell directly from the upper cone to the lower without passing through the archetype. These readings, which he interprets as evidence of the Kṣatriya tradition’s resilience, are his basis for reconstructing a partial supra-archetype θʹ, which, he argues, best approximates the con- tent and form of the earlier oral epic. In chapter four, Bigger explains how he intends to identify these readings using “empty references.” He introduces his notion of empty references in the Mahā bhā rata thus: In my assessment, parallel versions of the MBh can be demonstrated by one means only: namely, if references to them exist in the normative redaction without this reference being satisied in the normative redaction itself (that is, a so-called empty reference). How do such empty references enter the normative redaction? The likeliest explanation is that a redactor borrowed a passage from another version without being aware this passage contained a reference to another passage contained in the other version but not his own version of the text.89 He continues: But it is also possible that such a reference was not considered an internal (that is, text- immanent) reference by its composer (or the redactor who inserted it) but an external (that is, text-transcendent) reference, that is to say, he [the composer] was aware that the passage he referred to was not present in his text but it was so familiar to his readers that he could be con- ident that the reference would be recognized as such. We thus have two types of references, the text-immanent reference (which is never empty), which refers to another passage in the extant MBh, and the text-transcendent (which is always empty), which refers beyond the extant MBh. It is important that in the course of time a text-transcendent reference can become a text-immanent reference in that a later redactor considered it necessary to insert the external passage that is referred to into the text.90 If we wish to formalize these relationships, we can express them as follows: A: R → Rʹ and R→( )ʹ B: R → Rʹ and R→( )ʹ  and  R→( ) 69 ARGUMENTS FOR A HYPERARCHETYPAL INFERENCE 69 where R is a reference in a source A, Rʹ is its referent, B is a target text or copy and the empty parentheses indicate the absence of a referent. We adopt the following convention to distinguish between Bigger’s two hypothetical situations—intentional and accidental absence of referent: in the former case, we insert a prime symbol after the empty paren- theses to indicate that the author was aware a referent could be expected but chose not to include it; for the latter, we use the empty parentheses alone. The source A by deinition cannot contain an empty reference; it can only contain either a complete reference (R→ Rʹ) or a text-transcendent reference (R→ ( )ʹ).91 In contrast, the target text B can borrow either the reference and referent (text-immanent reference), or the reference and absence of referent (text-transcendent reference) or the referent can be lost during copying (empty reference). This last situation can again be of two types: either the referent was consid- ered not worth mentioning (that is, a text-immanent reference became a text-transcendent reference) or it was lost during copying (empty reference in the genuine sense). In the former case, the loss of referent is intentional and does not permit us to infer “iliation.”92 Conversely, an originally text-transcendent reference in the source can become a text- immanent reference in the target text if its copyist or redactor composes a new referent to satisfy the perceived absence of a referent. We can express this situation as follows: R → (Rʹʹ)ʹ where (Rʹʹ)ʹ indicates that a text-transcendent reference has become a text-immanent reference; Rʹʹ is a pseudo-referent. Finally, since a text-immanent reference itself can become a text-transcendent reference, as discussed earlier, we introduce the convention of inserting a double prime symbol after the empty parentheses to indicate intentional loss of referent. Our complete set of equations looks as follows: A: R → Rʹ and R→( )ʹ B: R → Rʹ and R→( )ʹ  and  R→( ) R → ( )ʹʹ or R→(R ʹʹ)ʹ  Note that of the ive possible outcomes of copying—the target borrows the complete ref- erence, it borrows the text-transcendent reference, a text-immanent reference becomes a text-transcendent reference, a text-transcendent reference becomes a text-immanent reference and empty reference—only the last is “genealogically” relevant from Bigger’s perspective. He writes: This [the fact that the source contains text-transcendent references, text-immanent references can become text-transcendent references and text-transcendent references can become text- immanent references] signiicantly limits the value of the empty reference for the identiica- tion of parallel versions. An empty reference always refers to an extant text, but we do not know if this text was a part of a MBh-version at the time or if it was transmitted at that time in an entirely diferent context. In my view, we can only speak of a parallel version of the MBh with a degree of plausibility, when this text is a case of another version or an addition to the main thread of the narrative of the MBh.93 70 70 PHILOLOGY AND CRITICISM As examples, he cites Mahā bhā rata 3.79 and 4.44: In the normative redaction, Subhadrā ’s abduction is frequently referred to when the aim is to praise Arjuna’s deeds. However, two of these passages must strike the reader, because they speak of events not familiar to him from the normative redaction: in 3.79, we ind our- selves with the Pā ṇḍavas in their forest exile. Arjuna has departed to obtain divine weapons in the Himā layas. The remaining brothers express their longing for him in that each praises Arjuna’s deeds and virtues. On this occasion Sahadeva says Arjuna defeated the Yā davas in battle and abducted Subhadrā with Kṛṣṇa’s approval. As we know, an important element of the Subhadrā episode, as we are so far familiar with it, is that the Yā davas do not pursue Arjuna and engage him in battle. Kṛpa goes even further in 4.44 when he rebukes Karṇa’s boasting: he says about Arjuna that he abducted Subhadrā all alone and challenged Kṛṣṇa to a chariot duel. We hear nothing about this either in the normative redaction of the MBh.94 We already saw that the thesis of a normative redaction is false. Since the constriction does not exist, all claims of readings that survived beyond the “Brahmanic” redaction are a for- tiori false. Indeed, since we have no evidence for such a redaction beyond what the German critics assert, the thesis is not only false; it is also prejudicial. It attests to nothing as much as their latent anti-Brahmanism. Since we have already addressed this prejudice’s roots else- where, we do not pursue the question here.95 Let us instead consider Bigger’s arguments for asserting that Mahā bhā rata 3.79 and 4.44 are older than the archetype. He writes: However, the picture changes when we look at the passages removed during the constitution of the critical edition: all manuscripts listed under the siglum S expand the Subhadrā episode by numerous passages. A corrector of D4 follows them in that he adds these passages in the margins and on separate pages of his manuscript. The two most important expansions can be found in App., 1, no. 114 and no. 115, which I will summarize in the following. Here also the story begins with the Raivataka festival in 1.211 during which Arjuna falls in love with Subhadrā . But he does not abduct her right away; rather, he uses a trick in App., 1, no. 114 which is inserted after 1.212.1: clad as a yati (an ascetic) he approaches the Yā davas. They receive him as a guest. Baladeva (against Kṛṣṇa’s advice, who is in the know, but cleverly transfers responsibility to his brother) houses him in the women’s quarters (App. 1, no. 114.1– 57). Subhadrā is already in love with Arjuna from his mere description. She thinks she recognizes the yati and asks him to speak about Arjuna. Arjuna reveals himself. Subhadrā grows lovesick. Devakī, who notices this and suspects that it is Arjuna in the yati’s guise, conides in Vasudeva. Vasudeva along with a few Yā davas (but without Baladeva and Uddhava) decides that they should marry. A  festival for Mahā deva is to be held on an island before the wedding. The Yā davas depart. Subhadrā remains to take care of Arjuna. He sees the time for abduction has come (58–221). He proposes marriage in the Gandharva style to Subhadrā . Subhadrā hesitates. Thereupon the gods and the family (except for Baladeva) come and perform the nuptials. Kṛṣṇa urges them to depart, but ofers them Rukmiṇī’s house for the irst night. The two enjoy the night (222–322). The next day Arjuna advises Subhadrā to equip the king’s chariot as though in play. When he is ready, she takes the reins in her hands. He mounts the chariot and they drive out of the city together. Some citizens cheer, whereas others are enraged. Arjuna shoots arrows at banners and other standards but does not injure anyone (323–395). A small army under Vipṛthu’s command is gathered at Raivataka. With Subhadrā as his charioteer, Arjuna charges him. Vipṛthu is 71 ARGUMENTS FOR A HYPERARCHETYPAL INFERENCE 71 defeated but makes peace with Arjuna as he is also in the know. Arjuna and Subhadrā change to Kṛṣṇa’s chariot and lee toward Indraprastha (396–462). The story continues with 1.212 until 1.213.12ab: the wrathful Yā davas want to pursue Arjuna, but Kṛṣṇa advises against it. In App. 1, no. 115 which is inserted thereafter, the Yā davas nonetheless follow Arjuna. He hears the din and wants to ight them. Subhadrā begs him to spare her relatives. The pair then continue their light. The Yā davas meanwhile encounter Vipṛthu’s defeated troops and abandon pursuit (1–45). The episode of Draupadī’s jealousy toward Subhadrā follows; it like- wise difers (46–63), but this doesn’t concern us here. S’s insertions present a version of the Subhadrā episode in harmony with 3.79.27. Since the Subhadrā episode belongs to the main thread of the MBh narrative, I consider it likely that we really have a parallel version to 1.211–213. This parallel version was inserted at a later date in the normative redaction. A closer look, however, lets us still clearly recognize it as a case of insertion: thus Arjuna arrives twice at the Yā davas. Although according to App. 1, no. 114 Subhadrā was abducted directly from the palace, in 1.212.9 her guardians hasten back to Dvā rakā to give alarm. Likewise, the Yā davas who returned to the island after the wedding (App. 1., no. 114.305) are back in Dvā rakā in 1.212 with a suddenness that astonishes the reader. This suggests that two versions of the same narrative were con- founded. The presence of an empty reference in App. 1, no. 113.1–3 illustrates that the narrative contained in S was not transmitted in its entirety. For Arjuna here recalls Gada’s kathā about Subhadrā ’s beauty and ponders on how to win her. However, such a narration by Gada is found in none of the versions collected in the critical edition. There is, furthermore, evidence that a third narrative of Subhadrā ’s abduction existed. Verse 4.44.6 cited earlier provides the impulse for it. It mentions a chariot battle between Kṛṣṇa and Arjuna. But this is not found in either of the two existing versions.96 As impressive as this display of his knowledge of the Mahā bhā rata is, Bigger’s arguments for Mahā bhā rata 3.79 and 4.44 as empty references are fundamentally unsound. As we observed earlier, the circumstance that a reference lacks a referent is not in itself conclu- sive, since the passage can be either of the type R → ( )ʹ or of the type R → ( )ʹʹ—that is, the reference could have been absent in the source itself or the copyist could have inten- tionally omitted the referent. Further, Bigger does not contemplate the possibility that the appendix passages nos. 114 and 115 could be innovations, composed to satisfy the perceived absence of referent—that is, that they could be pseudo-referents, even though the passages were clearly created with the intent of legitimizing Arjuna’s abduction of Subhadrā and providing a suitable genealogy for the future prince Abhimanyu.97 Indeed, even if the source contained a fulilled text-immanent reference (A:  R → Rʹ), as Bigger avers, it does not follow that Raʹ and Rbʹ are identical:  Rbʹ could still be a pseudo-referent. Bigger evidently thinks nos. 114 and 115 are remnants of an older oral epic tradition because they meet his expectations of what such a Kṣatriya epic could have looked like.98 According to him, when his eponymous Brahmans created the normative redaction with the intent of suppressing the Kṣatriya epic, they eliminated these passages as not in keeping with their Brahmanic ideology, but overlooked the existence of other passages that referred to them (Mahā bhā rata 3.79 and 4.44). This created an anomaly in the text that could only be resolved once the eliminated passages reentered the tradition—either 72 72 PHILOLOGY AND CRITICISM R -> R' Source text R -> R' Copy Normative redaction Archetype R E Descended from R D R -> R' F Contain a mix of readings from R -> R' G and (the recombined reference) R' B Descended from Figure 13 Bigger’s argument from “empty reference” because the Brahmans could not completely erase the scribes’ memories of the older epic or because a written version of it survived despite the Brahmans’ eforts. Thus he notes, “[i]n the course of time […] diferent copyists inserted passages from other versions— partly from direct recollection, partly from other (younger?) written versions—into the transcript.”99 In his view, Sukthankar erred in excluding these passages from the consti- tuted text because he overlooked the possibility that they could be older than the arche- type, even though not attested in all its descendants—the evidence for this being that they are not features of the archetype and hence must have been eliminated from the text during the formation of the normative redaction. Figure 13 makes his argument clear. The source α contains both the reference R and the referent Rʹ. The complete ref- erence is transmitted to a copy β and all its descendants. The Brahmans responsible for the normative redaction eliminate Rʹ as a Kṣatriya passage, but overlook its reference R, which is thus retained in their text as an empty reference. The tradition developing from the normative redaction thus includes R but not Rʹ. Rʹ meanwhile survives due to a par- allel transmission, that is, one not under the Brahmans’ control. It reenters the tradition at B, which represents either a manuscript or some other kind of record of the oral epic. It is thence transmitted to F and G, which are descended from the normative redaction, but not E and D (also descended from the normative redaction). Stemmatic reconstruc- tion will include R in the constituted text as it is found in all manuscripts examined for the edition (DEFG), but not Rʹ, thinking it an innovation in the manuscripts that feature 73 ARGUMENTS FOR A HYPERARCHETYPAL INFERENCE 73 it (FG), but this, Bigger argues, is false: Rʹ is actually older than any of the passages in the constituted text. It is as old as R, and if we collect all such passages we can reconstruct a partial supra-archetype αʹ that gives us access to an earlier stage of the tradition than what is represented in the critical edition. The problem with this argument is that it does not actually ofer us a means to select passages. Despite Bigger’s pretense, Rʹ is not actually identiied on the basis of a com- parison with other manuscripts. Rʹ denotes a pair of fragmentary passages (App.  1, nos. 114 and 115) he identiies in other manuscripts on “higher text-critical” grounds as remnants of the older Kṣatriya epic. The idea that we can reconstruct an earlier stage of the tradition based on a comparison between manuscripts descended from the normative redaction (DE), manuscripts descended from a parallel transmission (B) and manuscripts representing a conlated tradition (FG) is a misconception, because B does not exist and DEFG are all descended from the archetype! B is hypothesized to explain the survival of α’s readings, but actually the evidence for B is exactly as tenuous as for α. In other words, all Bigger has done is posit a manuscript in the lower cone to infer the existence of a source in the upper cone, and this as a means of demonstrating α’s reality! But if he wished to assert α’s existence, he could have done so without the detour via the ghost manuscript B.  B in his analysis is a double of α:  it only serves the purpose of making α more tangible, so to speak. Actually, however, B’s existence is predicated on α—an older oral epic tradition existed—and α’s existence is predi- cated on B—a part of the older epic tradition survived, thus permitting us to assert its existence—so his argument is circular. The same passages—App. 1, nos. 114 and 115—are used both times to make the existence of an oral epic tradition plausible: once as passages allegedly found in the upper cone and once as passages found in the lower cone. When we recall that no evidence exists for App. 1, nos. 114 and 115’s occurrence in the upper cone, or rather, the “evidence” for their occurrence in the upper cone is that they are found in the lower cone and considered examples of the Kṣatriya epic, it is clear that the argument sufers from petitio principii. App. 1, nos. 114 and 115 are introduced once as B, a remnant of the older oral epic tradition, and once as α, the source of which B is a remnant. But since the source of which B is supposedly a rem- nant is the oral epic tradition—that is to say, source is a synonym for oral epic tradition—all Bigger has done is postulate the oral epic tradition’s existence. The only reason the banality of this claim is not evident is he moves α down and back up so fast that the reader doesn’t realize it is the same source both times: α/B, a hypothetical manuscript brought down to the lower cone to create the impression of support for a source in the upper cone. Once we realize that neither α nor B exist and B, in particular, denotes not a manu- script but readings Bigger identiies in the archetype’s descendants, the suggestion that some readings rejected from the constituted text could derive from an older source falls out of consideration as unveriiable. Nevertheless, could Bigger not be right? Could our appendix passages 114 and 115 not be Rʹ, the lost referent of an original reading R → Rʹ? Here it is important to disam- biguate between two distinct questions one might be asking. On one hand, the question can be understood to mean: “Could our appendix passages 114 and 115 not be Rʹ, a ref- erent corresponding to R?” to which the answer is “possible but unlikely” since we have 74 74 PHILOLOGY AND CRITICISM no evidence that R refers to Rʹ, or that it requires a referent satisfying it. R was posited as a reference to make plausible 114 and 115’s status as the missing component of an orig- inal R → Rʹ. On the other, it can be understood to mean: “Could our appendix passages 114 and 115 not be descended from an original reading R → Rʹ?” to which the answer is “no” because ex hypothesi our reading R → Rʹ is not descended from R → Rʹ, but from the combination of R with Rʹ. We have no evidence for an original reading R → Rʹ other than the combination of R with Rʹ, a combination we have no evidence for other than its component parts, both of which occur in the lower part of our stemma. With one exception R and Rʹ do not even occur in the same manuscripts in our stemma.100 In other words, all Bigger has done is take two readings not necessarily related, com- bined them and then postulated that his combination corresponds to an original com- bination whose elements underwent dissociation due to a normative redaction, and this as a means of demonstrating the normative redaction’s existence! His entire “argu- ment” consists in irst assigning to 114 and 115 the designation Rʹ and then equivocating between two meanings of Rʹ: once as passages found in some extant witnesses and once as the referent of a hypothetical original reading R, and using the former as evidence for the latter. The so-called empty reference’s sole function is to enable the equivocation between Rʹ1 and Rʹ2, through which the existence of Rʹ2 and mutatis mutandis of the oral epic tradi- tion can irst be made plausible. When we recall that we only have our appendix passages 114 and 115 and that we have no evidence that they existed as the referent of an original reading R → Rʹ apart from the fact that Bigger postulates that they represent the lost referent of R → Rʹ, one half of which descended through the normative redaction and one half through other means, the circularity of his argument is apparent.101 Whatever plausibility his argument has derives from this inherent circularity. The empty reference does not strengthen it an iota. Indeed, it cannot because to identify App. 1, nos. 114–15 as the missing referents of an original R → Rʹ is to beg the question, which precisely concerns their authenticity. The error, elementary despite its complex setting, consists once again in equivocating between two senses of archetype, using it once to mean the “oicial text, prepared by its author for publication” and once to mean a manuscript that “can be reconstructed logically—with varying degrees of accuracy—on the basis of the number of conjunc- tive errors that are common to all its descendants.”102 From the fact that the archetype in this latter sense does not contain the passages Bigger concludes that the archetype in the irst sense must also not have contained them and, from this, he further concludes that they must have been removed even though no evidence exists that they existed in the irst place. Since he thinks the only people who could have had an interest in removing them were the “Brahmans,” he sees himself justiied in assuming a Brahmanic redaction of an earlier oral Ksạ triya epic and concludes—quite logically from his perspective, as it were—that this text, which corresponds in all its particulars to that text, must be the normative redaction. Actually, however, we have only one text—the constituted text of the critical edition—so the reference both times is to the same text, except Bigger lets it count once as an extant witness and once as a text reconstructed on its basis. Since a witness cannot be a witness of itself (except in the trivial sense that it attests to its own contents), the evidentiary value of his procedure is nil. In fact, for all his talk of scientiicity, all Bigger has done is select 75 ARGUMENTS FOR A HYPERARCHETYPAL INFERENCE 75 passages at random and combine them into a new text—an inane procedure justiiable only if what results is not any arbitrary combination but the “original.”103 The only reason this inanity is not apparent to him is that the German critics have gotten tight not only on “Criticism” but also on their own loquaciousness.104 As self-evident as Bigger’s suggestion may appear that Sahadeva’s reference to Arjuna’s victory over the Yā davas at Mahā bhā rata 3.79.26–27 cannot hang into nothingness but must be based on an actual description of the event in the text, neither the manuscript evidence nor the stemma supports the thesis. As the sole passages in the Mahā bhā rata answering to this description, App.  1, nos. 114–15 appear to be logical candidates for the reference, but they are unmistakably late. The only reason Bigger advances them is that he neglects their actual contents, which are clearly Brahmanic, for a fantasy about the Ksạ triya epic.105 Despite its rapturous reception among a section of Mahā bhā rata scholars, his work recalls rather Nietzsche’s comment: “he, a poetasting criminal!”106 In the next section, we examine Bigger’s arguments for thinking some passages “contained in the normative redaction” could have been “lost in some manuscripts in the course of transmission.”107 The Argument from Loss The argument from loss entails the claim that certain passages, though contained in the Mahābhārata’s hypothet- ical “normative redaction,” were lost from part of the subsequent manuscript tradition. They are therefore not found in the constituted text, although, as typical of Brahmanic concerns, they should have been included in it. Refutation of the argument:  The argument from loss replaces a critical principle—agreement between independent families—with a subjective and prejudicial principle for reconstruction. It uses this dogmatic prin- ciple—passages considered “Brahmanic” should be restored to the constituted text, even if they lack manuscript support—to assimilate the constituted text to a hypothetical Brahmanic redaction. It then uses this restored text, in turn, as evidence for the reality of the Brahmanic redaction. Given its circularity, the argument should per- suade no one. That it does illustrates the truth of the observation that if “the circle that has been constructed [is] large and confusing, […] the logical mistake goes unseen.” In the previous section, we saw how Bigger identiied the constituted text with a hypothet- ical “normative redaction” and, from the absence of certain passages from the constituted text, concluded that they were also absent in the normative redaction. In the second stage, he expands his attack on the critical edition, although the argument is now the exact inverse of the former one. Whereas he earlier argued that the critical edition reconstructs a histor- ical redaction of the Mahā bhā rata in all its particulars and used this to make claims about the hypothetical redactors’ activity, he now reverses direction and argues that, although the critical edition reconstructs the Mahā bhā rata’s “normative redaction,” it does not replicate it in all its particulars. Speciically, it excises passages not attested in all major recensions as likely insertions in the manuscripts containing them, even though, according to him, these passages were originally a part of the normative redaction and were only subsequently lost in one branch of the tradition. His exact argument is as follows: Before I tackle the individual recensions, I must address several passages that are so wide- spread that they cannot be assigned to a speciic version. We already found one passage 76 76 PHILOLOGY AND CRITICISM to which this applies in chapter  4.1:  [Ā diparvan] no.  80. [Sabhā parvan] no.  1 and [Droṇaparvan] no. 5 join this passage. It is striking that the transmission of these passages is not uniform. Beside [Droṇaparvan] no. 5 no other star- and appendix-passages of this book are attested in the same manuscripts. [Ā diparvan] no. 80 goes together with nos. 63, 75, 81, and 85; [Sabhā parvan] no. 1 [goes together] with nos. 3, 4, 13, and 23. Such passages can be interpreted in two ways: on one hand, we can assume these passages were very popular and hence transmitted across diferent manuscript traditions. On the other hand, they could be cases of passages contained in the normative redaction but lost in the course of transmission in certain manuscripts. However, this latter possibility would throw not only Sukthankar’s text-critical approach but also Grünendahl’s out of the window, since both apply the prin- ciple “short equals old” or do not question it. However, this possibility cannot be completely excluded, since the MBh contains several proofs that individual manuscripts omitted some verses. If this occurred very early in the history of transmission, would it not be conceivable that an entire tradition lost a passage?108 Bigger continues: Before I  address this question further, I  would like to consider these passages’ content. I already treated App. 1, no. 80 above (p. 116). The passage describes Yudhiṣṭhira as yuvarāja and the Pā ṇḍavas with their capabilities. They increasingly begin to intimidate Dhṛtarā ṣṭra. In tabular form, the passages from the irst book, which have the same attestation, have the following contents: no. 63 Story of Duḥ śalā ’s birth. no. 75 The child Aśvatthā ma drinks rice water in the belief that it is milk. no. 81 Kaṇikanīti (Dhṛtarā ṣṭra’s instruction by his minister Kaṇika). no. 85 The Pā ṇḍavas cross the Ganges after escaping from the lacquer-house with Vidura’s aid. In terms of their content, all these passages are of secondary signiicance for the MBh’s further course. We cannot identify any plausible reason why a later redactor should have con- sidered these passages worthy of addition. It appears even less plausible that these passages attained such popularity that they could spread all across India. However, these are arguments from higher textual criticism, whose application to lower textual criticism I already refused myself in the introduction (p. 19). There is, however, yet another argument: All these passages are not attested in the manuscripts Ś1, K0–3 and Ñ4. Of these, Ś1 and K0–3 are related with each other. Ñ4 appears not to belong to this group on the basis of the Schriftartenprämisse [sic]. However, this has forfeited its unimpeachability with Grünendahl’s discussion, as I showed in chapter 4.2.1. Ñ4 therefore could be related with Ś1 and K0–3 and constitute a recension together with them. Thereby there is a possibility that the textual passages men- tioned earlier could have been lost in this recension [the hypothetical recension comprising Ś1, K0–3 and Ñ4], that is, that they were originally contained in the normative redaction. We would thereby overturn the previous maxim “short equals old” but would possibly come closer to the normative redaction’s actual form [tatsächliche Gestalt].109 In the critical edition, passages 63, 75, 80, 81 and 85 are all relegated to the critical appa- ratus (in a separate appendix) as additions to the text of the archetype. The evidence that they are additions is unambiguous. Not only are they not found in the best Ā diparvan 7 ARGUMENTS FOR A HYPERARCHETYPAL INFERENCE 77 manuscripts, Ś1 and K0–3, but the insertion of three of them in the manuscripts that contain them is inconsistent.110 Sukthankar comments on these passages as follows: Important omissions which distinguish ν (really only Ś1 K0–3) from all other manuscripts are these: (i) the adhyā ya giving a naive account of the birth of Duḥ ṣā la (Bom. adhy. 116), which uncommonly looks like being an afterthought (App. I, No. 63); (ii) a passage of about 25 lines describing how Droṇa’s son Aśvatthā man is given lour mixed with water, which he drinks in the belief that it is milk (App. I, No. 75); (iii) an adhy. (Bom. adhy. 139), in which there is an incidental allusion to the installation of Yudhiṣṭhira as Yuvarā ja, and which is repetitious and incoherent (App. I, No. 80); (iv) the so-called polity of Kaṇi(ṅ)ka, Kaṇi(ṅ)kanītī (Bom. adhy. 140), which is a replica (nat- urally with many additions, omissions and variant readings) of the advice given by Bhāradvāja (apparently a gotra name of this very individual) to Ś atruṁjaya, and duly communicated by Bhīṣma to Yudhiṣṭhira in the Ś ā nti (App. I, No. 81); (v) the crossing of the Ganges by the Pā ṇḍavas (Bom. adhy. 149), a superluous adhyāya, which only serves to confound the already confused geography of the narrative (App. I, No. 85). These ive passages are found in all manuscripts collated except Ś1 K0–3, but it is worthy of note that even apart from their omission in ν, the documentary evidence with regard to at least two of them‚ is confused and unsatisfactory. No. v (crossing of the Ganges) is inserted in diferent groups of manuscripts at diferent points of the text. In No. iv (Kaṇikanīti), on the other hand, most of the Southern manuscripts repeat, after the interpolation, the immedi- ately preceding portion of the original, apparently in order to restore the context broken by the intrusion of extraneous matter.111 K2–6 were subsequently found contaminated with either the central recension γ or with the southern recension S.112 K5, furthermore, was incomplete, ending after 1.3.152, whereas K6 was collated only until the end of adhyāya 2 (K5 was collated only until 1.2.40).113 The ν group thus really comprised only Ś1 and K0–4. Only one member of this recension (K4) contained the passages, and that member was heavily contaminated with a southern source.114 Sukthankar therefore considered it likelier that nos. 63, 75, 80, 81 and 85 entered the tradition at a later stage than that an entire recension (ν, com- prising the manuscripts Ś1 and K0–3) omitted the passages. His reasons for considering the passages additions were threefold: (1) the manuscript evidence for them was unsatis- factory; (2) their content attested to their secondary character; and (3) most important, his observation that Mahā bhā rata scribes tended to conserve as much of the transmitted text as possible.115 There was no evidence that the passages were haplographic omissions in the manuscripts that lacked them. Mechanical damage to their common source could likewise be ruled out. The subsequent discovery of a Nepā lī manuscript in the collection of the Rajaguru of Nepal conirmed Sukthankar’s decision.116 As he noted: Hitherto, it may be observed, the data of the Sā radā and the “K” versions had mainly to vouchsafe for the puriication of the epic text attempted in the Critical Edition. Now, the omissions of the constituted text are supported from a new source and from an entirely diferent version, the Nepā lī. Our MS. is in fact surprisingly free from most of the matter pronounced 78 78 PHILOLOGY AND CRITICISM to be spurious on the evidence of the then available MSS. of diferent versions. The longer insertions given in Appendix I are conspicuous by their absence in this MS. Here is a list of the long passages which are missing in the new MS. No. 1: The Brahmā -Gaṇeśa complex; inserted at diferent points in K4–6 Dn Dr D2– 12.14 S, i. e., om. in its totality only in K0–3 D1. No. 13: A passage of six lines ins. in K (except K1) V1 (marg.) Da Dn D1–7 G1.2.4.5. No. 14: The Purā ṇic story of Aruṇa, who is made to act as the charioteer of the Sun; ins. in K4(marg.) Ñ V1 B D (except D5; D2 on suppl. fol.) T1 G1.2.4.5. No. 41: Names of the hundred sons of Dhṛtarā ṣṭra; ins. in K3.4 Ñ2.3 V1 B D (except D5). No. 42: Story of the birth of Abhimanyu; ins. in K3.4 Ñ V1 B D (except D5). No. 43: Story of the birth of Karṇa; ins. in K4 Ñ V1B D (except D5). No. 61: Details of the marriage of Pā ṇḍu and Mā drī, subst. for 1.105.4–7, in K4 Ñ B D. No. 63: Story of the birth of Duḥ śalā ; ins. in all MSS. except Ś1 K0–3. No. 71–72: Details of the picnic (jalakrı̄ḍā); ins. in K4 Ñ B D. No. 75: Story of piṣt ̣odaka given to Droṇa’s son; ins. in all MSS. except Ś1 K0–3. No. 76: Arjuna’s pre-eminence among the pupils of Droṇa; ins. in K4 Ñ B D T2 G1.2.4.5. No. 78:  Description of the conquest and annexation of Drupada’s kingdom by the Pā ṇḍavas exacted by Ā cā rya Droṇa as gurudakṣiṇā; ins. in K4 Da1 Dn D2(supp. fol. sec m.).4.5 S. No. 80: Installation of Yudhiṣṭhira as heir-apparent; ins. in all MSS. except Ś1 K0–3. No. 81: Kaṇikanīti; ins. in all MSS. except Ś1 K0–3. No. 85: Story of the boat sent by Vidura for the use of the Pā ṇḍavas; ins. in all MSS. except Ś1 K0–3. No. 118: Ś vetaki-episode; ins. in all N MSS. Cf. also passage No. 120 ins. in T2 G3.4, and No. 121 ins. in Ś1 K0.1–4.117 The evidence that these passages are additions to the manuscript tradition is therefore unambiguous. Yet rather than acknowledge the evidence, Bigger widens the scope of his argument to encompass the Mahā bhā rata’s second book also. He writes: Can one say something similar about the other passages listed above? [Sabhā parvan] App. 1, no. 148 reports how Kṛṣṇa returns to Dvā rakā after the Khā ṇḍava episode. His reception there (among others, by Baladeva) as well as the deeds of the Pā ṇḍavas who remained behind are described. In addition to this passage, there are other passages in Book 2 that manifest a similar state of evidence. However, the situation is by far not as uni- form as for the passages in Book 1 just treated (more on this below). I irst present the content of the passages briely: no. 3 Expansion of the description of Kubera’s palace. no. 4 How Yudhiṣṭhira and his brothers administer their kingdom. no. 13 Expansion of Sahadeva’s deeds during the campaign of world conquest. no. 23 Expansion of the list of kings, whom Ś iśupā la would prefer to Kṛṣṇa. As was the case in Book 1, in terms of their content, these passages do not have any spe- cial signiicance for the further course of the MBh and it does not appear plausible why they should have become so popular. Thus, we are once again confronted with the question of whether these are passages that were lost from the normative redaction in speciic recensions. ́ K0–4, Ñ1 and D6. This close state of evidence (which, App. 1, no. 2 is not attested in S1, however, only corresponds in one case with the state of evidence for the passages found in Book 1) could indicate that here also a passage was lost from the normative redaction. However, on 79 ARGUMENTS FOR A HYPERARCHETYPAL INFERENCE 79 the other hand, T, G (without G3) and M share the passage and furthermore omit two half- verses. App. 1, no. 1 shares this peculiarity with App. 1, no. 3. In both cases, the transmission of the passage in the S manuscripts is so confused that I consider it unlikely that these passages belonged to the normative redaction. This inding is conirmed by the state of evidence for the remaining passages: alongside App. 1, no. 1, only App. 1, no, 23 is not attested in D6. The latter is also missing in D4 but is, on the other hand, contained in K4. The relationship of the manuscripts is therefore less uniform than in the case of the passages from Book 1. In my opinion, at most App. 1, nos. 4 and 13 can be considered losses from the normative redaction.118 It is clear why Bigger cites nos. 4 and 13 as examples of “losses from the normative redaction.” The two passages are found in all northern recension manuscripts with the exception of the Ś-K group and the sole Nepā lī manuscript collated for this parvan Ñ1. They thus exhibit the same “state of evidence” as his examples from Book 1.119 Bigger evidently thinks the two passages strengthen his reason for thinking the Ś, K and Ñ manuscripts “constitute a recension together.” If so, then, like his examples from the Ā diparvan, nos. 4 and 13 could be examples of passages “lost in this recension [his hypo- thetical recension comprising Ś1, K1–4 and Ñ1].” He would then have two examples (one set from each book) of passages that were “originally contained in the normative redaction” but were lost from the subsequent tradition, thus bolstering his claim that restoring certain passages to the constituted text would “overturn the previous maxim ‘short equals old’ but would possibly come closer to the normative redaction’s actual form.”120 Unfortunately, the claim is nonsense. First, as noted, the evidence for the passages being additions is unambiguous. Merely for the last of the two passages Bigger cites (no.  13), the manuscript evidence reveals that its insertion was inconsistent and the passage kept shifting. Thus, whereas the majority of the northern manuscripts (that is, V1, B, Dn and D1–3.5.6) along with some southern manuscripts (T1, G1.3–5 and M) insert the passage after 2.28.10, D4 inserts it after 2.28.9, while G2.6 inserts it after 2.28.9a. The passage is inserted in the margin in B1. D4 omits stanza 10 and the irst line of the passage by haplography, while G2.6 omits verses 9b onward up to the irst half of the irst line of the passage (also by haplography). The second passage’s inser- tion is more consistent but it is inserted in the margin in B1. Second, the passages’ contents themselves attest to their secondary nature. Third, Bigger’s arguments against the passages being additions to the tradition are erroneous. He writes that “In terms of their content, all these passages are of secondary signiicance for the MBh’s fur- ther course” and adds: “We cannot identify any plausible reason why a later redactor should have considered these passages worthy of addition.” But this is to misunder- stand the nature of manuscript transmission. Additions tend, by their very nature, to be “of secondary signiicance.” They need not directly afect the text’s “further course” for scribes to include them. Neither do we need to look for a special reason why scribes might have “considered these passages worthy of addition.” The history of Mahā bhā rata transmission demonstrates that scribes continuously added materials to their inherited text—materials we would consider superluous, irrelevant and even contradictory.121 80 80 PHILOLOGY AND CRITICISM Likewise, it is erroneous to think passages must attain “popularity” to “spread all across India.” The rate of difusion of innovations (which is what insertions or additions to the text are) is simply a function of the rate at which manuscripts are copied, how often they are compared and the scribes’ own practice in respect of innovations (that is, how likely they are to copy the innovations they encounter). If manuscripts are copied frequently, innovations will difuse rapidly. If manuscripts are compared frequently, innovations will difuse through not only vertical but also horizontal transmission. If scribes tend to copy everything they encounter, even “unpopular” innovations will be conserved, or, rather, their popularity or lack thereof is irrelevant for their dissemina- tion. It is erroneous to draw conclusions about an innovation’s popularity from the fact that it is widespread. The innovation could simply have occurred higher up in the tra- dition. Alternatively, it could have occurred in a manuscript that either produced many descendants or produced descendants that survived, whereas those of other branches, which did not contain the innovation, did not. We must remember that the manuscripts we possess are but a small percentage of all manuscripts ever produced. The circum- stance that we frequently encounter a particular innovation in our manuscripts does not necessarily mean that it was also common in the wider tradition (that is, the part that did not survive).122 Bigger’s arguments for nos. 4 and 13 as “losses from the normative redaction” are not only against good sense but also against the editor’s reasoned judgment. Like the Ā diparvan passages, which he sourced from the editor’s introduction, his examples for the Sabhā parvan are all taken from Franklin Edgerton’s introduction to this book. Edgerton speciically cites nos. 1, 2, 3, 4, 13, 19 and 23 as evidence that Ś and K constitute an inde- pendent group against the Nepā lī, Bengā lī and Devanā garī manuscripts (his “Eastern” recension) and the southern manuscripts.123 On the question of whether passages absent from one recension could be omissions in that recension rather than additions in the recensions containing them, he is even more emphatic than Sukthankar: On one extremely important methodological point, I  have gradually come, in the course of my work, to a position which entirely agrees with Sukthankar’s procedure, but goes slightly further than his theory as stated xciv f. This concerns the “plus or minus” passages; passages (long or short) which are entirely missing (with no equivalent substitutes) in one or more recensions. Sukthankar relegates such passages, almost without exception, to his Notes or Appendix, and does not admit them to the text, even when all recensions but one contain them. Yet he says (xcv) that they are “not necessarily spurious. There might be a hun- dred good reasons why [they] are missing in a particular recension.” Well, I suppose there “might” conceivably be such reasons. As far as Book 2 is concerned, I  can only say that I very gravely doubt the existence of such cases. I have come to believe that any passage, long or short, which is missing in any recension or important group of MSS. as a whole, must be very seriously suspected of being a secondary insertion. For the Mbh., I should now hesitate long before including any such case in the edited text. This was not a preconceived notion. Indeed I started with a quite diferent attitude. It is only long and careful study which has inally forced me to this conclusion (which, I take the liberty of saying, I believe is at least not far from Sukthankar’s own view at present). I should now go as far as to assert that prob- ably not one of some ifty MSS. I have studied for Book 2, nor any of their genealogical 81 ARGUMENTS FOR A HYPERARCHETYPAL INFERENCE 81 ancestors, ever deliberately or intentionally omitted a single line of the text; and that such acci- dental omissions as occurred were, in every recension represented by several MSS., almost invariably made good in at least some MSS. which we possess of the recension in question. Mistakes will happen, of course. Doubtless few MSS.  are wholly free from accidental omissions (often, but by no means always, explainable by haplography or the like). But in the Sabhā parvan, very rarely will these accidental lacunae be found in all the MSS. of any recension (we must count out the Nepalese, for which in this book we have only a single MS.). And it appears that no scribe, no redactor, ever knowingly sacriiced a single line which he found in his original. Not even if he found something which seemed to him incompre- hensible, inconsistent with the context, irreligious, or immoral. In that case he sometimes changed the text, doubtless feeling that a mistake had been made and that he was only doing the pious work of restoring what must have been the original intention. But there is certainly not a shred of evidence for a single deliberate omission: and I do not believe it ever took place. In general, therefore, I  feel rather more conident than Sukthankar seems to have felt in editing Book 1, that we are following the original text in omitting, generally speaking every line that is missing in any recension as a whole (excepting only Ñ because it is represented by a single MS., and in one MS. anything can happen). In other words, I believe that in gen- eral these are not “omissions” in the recension that lacks them, but secondary insertions in those that contain them, even though these latter may be what we must regard as generally independent versions.124 This leaves Bigger one inal argument for claiming that the passages under consideration could be “losses from the normative redaction.” Although no better than the rest, for clarity’s sake, we also address it here. After raising objections to the editors’ decisions on grounds of “plausibility” and “popularity,” Bigger sets these objections aside, claiming: “these are arguments from higher textual criticism, whose application to lower textual criticism I already refused myself in the introduction.” Then, noting, “There is, however, yet another argument,” he introduces his inal and, in his view, conclusive argu- ment for considering the passages “losses.”125 He writes: All these passages are not attested in the manuscripts S1, ́ K0–3 and Ñ4. Of these, S1́ and K0–3 are related with each other. Ñ4 appears not to belong to this group on the basis of the Schriftartenprämisse [sic]. However, this has forfeited its unimpeachability with Grünendahl’s discussion, as I showed in chapter 4.2.1. Ñ4 therefore could be related with Ś1 and K0–3 and constitute a recension together with them. Thereby there is a possibility that the textual passages mentioned earlier could have been lost in this recension [the hypothetical recension comprising Ś1, K0–3 and Ñ4], that is, that they were originally contained in the normative redaction. We would thereby overturn the previous maxim “short equals old” but would pos- sibly come closer to the normative redaction’s actual form [tatsächliche Gestalt].126 This argument presents four problems. First, Ñ4 is not related with the Ś and K manuscripts, as an analysis of their readings demonstrates.127 Second, the circum- stance that two manuscripts lack the same interpolations is not an argument for their being related (that is, more closely than through the common ancestor they share with all other manuscripts—by deinition, the archetype). Third, the circumstance that passages “could have been lost in this recension” (emphasis added) is not an argument 82 82 PHILOLOGY AND CRITICISM for their actually being lost, especially since the existence of this recension is only hypo- thetical and premised solely on the fact that the manuscripts in question do not feature the same passages. Indeed, the “recension” does not exist, because Ñ4’s shared errors of transcription reveal it is descended from ε, the common ancestor of not only the other Nepā lī manuscripts but also the Maithili and Bengā lī manuscripts, and it does not have an ancestor in common with Ś and K more recent than N!128 Fourth, the claim that Ñ4, Ś1 and K0–3 “could […] constitute a recension together” is circular because their assignment to this hypothetical recension presumes that they owe the passages’ absence to “loss” from a common ancestor but this is the very issue in question. Let us call these four problems the practical, the theoretical or methodological, the grammatical and the logical objections. We address each of them in turn. Sukthankar’s collations of the manuscript in his 1939 article “The Oldest Extant Ms. of the Ā diparvan” clearly demonstrate that “Ñ4” (the siglum that Bigger, following Reinhold Grünendahl, assigns this manuscript) is related to the other Nepā lī manuscripts. Their shared errors of transcription establish their descent from a common source they share with each other but with no other group of manuscripts.129 Ñ4’s sole resemblance to the Ś and K manuscripts is that, like them, its text is fairly free of longer interpolations.130 But the circumstance that two manuscripts retain a largely original text in respect of the additions is not an argument for their closer relation, since they could owe these absences to the archetype or to the original itself. Bigger evidently thinks that Ś1, K0–3 and Ñ4 “constitute a recension together” because he thinks the circumstance that they lack the same passages makes them alike, but this is to misunderstand how editors establish iliation.131 The concept of iliation refers exclusively to the signiicant errors shared between manuscripts, which serve to identify them as members of a subfamily. In contrast, readings thought likely to be those of the original are irrelevant for determining iliation.132 When Bigger now claims that “there is a possibility that the textual passages mentioned earlier could have been lost in this recension,” and further concludes: “We would thereby over- turn the previous maxim ‘short equals old’ but would possibly come closer to the nor- mative redaction’s actual form,” he moves from an expression of probability to one of certainty, even though he has not yet shown that the passages were lost or that Ś1, K0–3 and Ñ4 have a more recent ancestor in common than the one they share with the remaining northern manuscripts (by deinition, N). The only reason this illicit shift from a conditional statement to a categorical assertion is not apparent is the retention of the subjunctive mood for both. The fourth objection is the most damaging to Bigger’s case but it is also the most dif- icult to demonstrate because it is where he has concealed his entire “argument.” Let us begin with his simplest formulation of the argument: Such passages can be interpreted in two ways: on one hand, we can assume these passages were very popular and hence transmitted across diferent manuscript traditions. On the other hand, they could be cases of passages contained in the normative redaction but lost in the course of transmission in certain manuscripts. However, this latter possibility would throw not only Sukthankar’s text-critical approach but also Grünendahl’s out of the window, since both apply the principle “short equals old” or do not question it. However, this possibility cannot be completely excluded, since the MBh contains several proofs that individual manuscripts 83 ARGUMENTS FOR A HYPERARCHETYPAL INFERENCE 83 omitted some verses. If this occurred very early in the history of transmission, would it not be conceivable that an entire tradition lost a passage?133 The argument is circular because it states that the passages could be omissions from some manuscripts if lost from some manuscripts early in the history of transmission. The argu- ment certainly “proves” the conclusion but only trivially so, since the conclusion adds no new information beyond that already contained in the premises. Unlike the editors, who provided noncircular arguments—the scribes’ observed practice, frequency of manu- script comparison, absence of haplography, inconsistent insertion (including one mar- ginal occurrence) and the circumstance that sometimes the same verses repeated after the insertion as occurred before it—against the passages being omissions in the manuscripts that lack them, Bigger’s argument does not add to the sum of our knowledge. It merely states that the passages could be losses if they are losses. The reason the argument’s circularity is not immediately apparent is that Bigger does not state so directly. Like most circular arguments, “the presumption that is at the heart of the fallacy is buried in the verbiage of the premises, sometimes obscured by confusing or unrecognized synonyms.”134 Thus, “lost” is a synonym for “omitted,” while “in the course of transmission” is a synonym for “in the history of transmission.” The antecedent of “such passages,” “several passages that are so widespread that they cannot be assigned to a speciic version,” is a meaningless piece of verbiage. Actually, all Bigger means by it is: appendix passages found in manuscripts of both recensions. Since the appendix passages are by deinition passages omitted in some manuscripts, all Bigger has said so far is that passages omitted in some manuscripts (his “such passages” or his “passages that […] cannot be assigned to a speciic version”) could be “losses.” If we recall that the basic question confronting the editors was “are passages found only in some recensions omissions in the manuscripts that lack them or additions in the rest?” all Bigger has done is answered the question with a lat: “yes, they are omissions.” The additional qualiication that these losses could have “occurred very early in the history of transmission” adds nothing to the question’s resolution. It only attests to Bigger’s penchant for transferring problems concerning the extant manuscripts higher up the stemma (see previous section). In this case, the transferal achieves nothing because no one asserts that the losses must be recent, and its sole function is to ensure that “the circle that has been constructed [is] large and confusing, and thus the logical mistake goes unseen.”135 Bigger’s second formulation of the argument is exactly as circular as the irst. He writes: All these passages are not attested in the manuscripts Ś1, K0–3 and Ñ4. Of these, Ś1 and K0–3 are related with each other. Ñ4 appears not to belong to this group on the basis of the Schriftartenprämisse [sic]. However, this has forfeited its unimpeachability with Grünendahl’s discussion, as I showed in chapter 4.2.1. Ñ4 therefore could be related with Ś1 and K0–3 and constitute a recension together with them. Thereby there is a possibility that the textual passages mentioned earlier could have been lost in this recension [the hypothetical recension comprising Ś1, K0–3 and Ñ4], that is, that they were originally contained in the normative redaction. We would thereby overturn the previous maxim “short equals old” but would pos- sibly come closer to the normative redaction’s actual form [tatsächliche Gestalt].136 84 84 PHILOLOGY AND CRITICISM Initially, the argument appears plausible. The passages under discussion could have been lost from the manuscripts that lack them, and in that case they must have been features of the archetype, since this is implicit in the concept of “loss.” However, as yet, Bigger has merely articulated a hypothesis. He has not answered the question of whether the passages were, in fact, lost. Moreover, the hypothesis is tautological: obviously for the passages to be “lost” they must have originally been contained in the tradition, and vice versa. Not only is the hypothesis tautological; the argument itself is circular, since the statement “Thereby there is a possibility that the textual passages mentioned ear- lier could have been lost in this recension, that is, that they were originally contained in the normative redaction” functions both as the argument’s premise(s) and conclu- sion. Disguised by the ambiguous “that is” the statement could be making one of two claims:  (1) the passages could have been lost if they were originally contained in the archetype and (2) the passages could have been lost and, if so, they must have been originally contained in the archetype. It is precisely this ambiguity or, rather, the clandestine shift between the irst sense, where the statement functions as a pre- mise, and the second, where it functions as a conclusion (in fact, establishing the con- clusion Bigger seeks, namely, that the passages were in the archetype), that lends the argument its persuasive force. Actually, the argument consists of a pair of premises, since (2)  can be rewritten as “The passages must have been originally contained in the archetype if they were lost” and Bigger’s real conclusion is not “the passages could have been lost and, if so, they must have been originally contained in the archetype” but simply “the passages were contained in the archetype.” We can formalize the argument as: Premise (1):  p ⊃ q (If they were originally contained in the archetype, the passages could have been lost.) Premise (2): q ⊃ p (If they were lost, the passages must have been originally contained in the archetype.) Conclusion: ∴ p, q (The passages were originally contained in the archetype and they were lost.) The fallacy is now evident: (1) and (2) are mutually implicating statements, and hence logically equivalent (p ⊃ q ≡ q ⊃ p). The conclusion, moreover, is already contained in the premises, since it asserts p and q, the antecedent and consequent of the conditional prop- osition p ⊃ q (and also the consequent and antecedent of the conditional proposition q ⊃ p). The argument thus begs the question. Indeed, it is not just circular; it is viciously cir- cular since the conclusion does not expand beyond the premises and Bigger at no time demonstrates the truth of p. He merely says that if the passages were originally contained in the archetype, they were contained in the archetype—which is true but not much of an argument. Given its vacuity, how could this argument convince anyone? We cite four possible explanations: 1. The ambiguous “that is,” which obscured the logical relation between the statement’s two parts and did not let them appear in their function as antecedent and consequent of a conditional proposition. The relationship of the antecedent and conse- quent can, moreover, be read in two ways depending on how we construe “that is” (that 85 ARGUMENTS FOR A HYPERARCHETYPAL INFERENCE 85 is, either as p ⊃ q or as q ⊃ p). Alternatively, the obscure phrase does not express a condi- tional relationship but a statement of fact. In that case, p and q are not the antecedent and consequent of a conditional proposition but mere assertions (p, q), and the statement represents the conclusion of Bigger’s argument. The same statement could thus function as both premise(s) and conclusion. 2. The argument’s question-begging nature, in which one of the conclusions (p) was buried in the premise p ⊃ q and tautologically validated through the second conditional proposition q ⊃ p. The argument was, moreover, circular, since every time p is true q will likewise be true, as a truth table will demonstrate—and vice versa. The second proposi- tion is thus actually superluous (since p was already assumed true) but Bigger needs it to suggest that p was derived as the conclusion of a valid inference (q ⊃ p; q, ∴ p) (and likewise for q: p ⊃ q; p, ∴ q).137 He can now assert the conclusion p, q, even though he has demon- strated neither p nor q. And since “anteriorly present in” is already entailed in the concept of “lost,” efectively, Bigger has done no more than assert p (p = q). 3. The equivocation between normative redaction and archetype, which endowed the former concept with material reality, even though we possess only one text—the consti- tuted text of the critical edition; by deinition, the archetype of the tradition—and hence both times the reference can only be to the latter. 4. The statement that the passages “could have been lost in this recension,” which has the efect of: (1) begging the question and (2) widening the deinitional circle, so that we do not see that the argument really only makes a claim about the passages’ absence from Ś1, K0–3 and Ñ4, the only part of Bigger’s argument not in dispute. Concerning (1): The central question at issue is whether the passages are features of the archetype (in which case their absence in some manuscripts must be attributed to loss) or later additions to the tradition so to assert that the passages “could have been lost” in no way answers this question. It merely begs it. Concerning (2): When Bigger says “lost in this recension” the reference can only be to the passages’ absence in the manuscripts that do not contain them. But the equivocation between “loss” and “absence” allows him to suggest that Ś1, K0–3 and Ñ4 form a recension, thus allowing him to shift the locus of the passages’ “loss” (that is, absence) from the manuscripts themselves (in which case the claim’s vacuity would be self-evident) to a hypothetical ancestor of the recension, from which—hypothetically, since he never explicitly says so—they were lost. The argument is thus exactly circular as the larger argument in which it is embedded. It takes the form: Premise (1):  p ⊃ q (If the passages were lost, S1, ́ K0–3 and Ñ4 could constitute a recension.) Premise (2): q ⊃ p (If Ś1, K0–3 and Ñ4 constitute a recension, the passages were lost.) Conclusion: ∴ q, p (Ś1, K0–3 and Ñ4 constitute a recension and the passages were lost.) The reason the argument’s circularity is not evident is that Bigger does not explicitly state the premise that the passages were lost from a common ancestor of the recension comprising Ś1, K0–3 and Ñ4 or a common ancestor that gave rise to the recension comprising Ś1, K0–3 and Ñ4, even though the existence of this hypothetical ancestor is implied in the concept of “loss.” Rewritten correctly, (1) actually says: “If the passages 86 86 PHILOLOGY AND CRITICISM were lost [from a common ancestor of Ś1, K0–3 and Ñ4], Ś1, K0–3 and Ñ4 could constitute a recension,” and now the fallacy is clear. Implicit in the unstated pre- mise that Ś1, K0–3 and Ñ4 have a common ancestor more recent than the one they share with the remaining northern manuscripts (by deinition, N) is a further claim, namely, that Ś1, K0–3 and Ñ4 constitute a recension, and thus, disguised by its obscure form, what the argument actually says is that if the manuscripts constitute a recension, they could constitute a recension. Just as we found that Bigger’s main argument was not just question-beginning but had buried one of its conclusions in one of its premises, so also this subsidiary argument—the conditional proposition p ⊃ q—is not just question-begging but buries its consequent in the compound statement that serves as its antecedent. Given this (complex) antecedent, the conse- quent must be true. One inal comment is in order before we lay Bigger’s inal argument to rest. Even granted its circularity, how could it appear to so many that Ś1, K0–3 and Ñ4 could, as Bigger says, “constitute a recension”? Two possible explanations arise: either these people understood the concept of iliation and how recensions are identiied as little as Bigger (as our examination of Grünendahl’s and John Brockington’s work demonstrates) or they were misled by the erroneous nature of Bigger’s argument for Ś1, K0–3 and Ñ4 as a recension. As the last remaining logical trick in Bigger’s arsenal, it deserves a closer look. The key part of the argument reads as follows: All these passages are not attested in the manuscripts Ś1, K0–3 and Ñ4. Of these, Ś1 and K0–3 are related with each other. Ñ4 appears not to belong to this group on the basis of the Schriftartenprämisse [sic]. However, this has forfeited its unimpeachability with Grünendahl’s discussion, as I showed in chapter 4.2.1. Ñ4 therefore could be related with Ś1 and K0–3 and constitute a recension together with them.138 If we remove the superluous reference to Grünendahl (which serves only to obscure the sentence structure), the statement reads:  “All these passages are not attested in the manuscripts Ś1, K0–3 and Ñ4. [...] Ñ4 therefore could be related with Ś1 and K0–3 and constitute a recension together with them.” Depending on how we read the ambiguous connective, the proposition that Ś1, K0–3 and Ñ4 constitute a recension can be derived in two ways: (1) As the apodosis of a simple conditional proposition: If the passages are absent in Ś1, K0–3 and Ñ4, then Ś1, K0–3 and Ñ4 constitute a recension. (2) As the conclusion of a mixed hypothetical syllogism with a conditional irst pre- mise and a categorical second premise: If Ś1, K0–3 and Ñ4 constitute a recension, then the passages will be absent in Ś1, K0–3 and Ñ4. The passages are absent in Ś1, K0–3 and Ñ4. Therefore Ś1, K0–3 and Ñ4 constitute a recension. Unlike the earlier examples, where Bigger relied on the synonymy of “loss” to make the hypothesis probable, here he sticks to the formulation “not attested” or, as we would 87 ARGUMENTS FOR A HYPERARCHETYPAL INFERENCE 87 say, “absent.” The two versions of the argument therefore do not turn on an equivoca- tion. Their error must be located elsewhere. Concerning (1): The conditional proposition p ⊃ q is false, because it is possible for q to be false, while p is true. In fact, for every value of p, q will be false, and thus the conjunc- tion p ⋅ ~q will always be true. Since p ⊃ q is logically equivalent to ~(p ⋅ ~q), this is equiv- alent to saying that p ⊃ q will always be false. Concerning (2): The argument commits the fallacy of airming the consequent, that is to say, its categorical premise airms the consequent of its conditional irst premise. The argument thus has the form p ⊃ q; q, ∴ p. As no valid inference can be drawn about the antecedent of a conditional proposition from the truth of its consequent, the argument is false. Bigger’s arguments for a normative redaction are so circular, question-begging, erro- neous and prejudicial that no one should have cited them. The surprise is that so many did.139 While it may appear that he provided a model for how the passages could have been lost (by attributing their “loss” to a common ancestor), the model remains only hypo- thetical. It does not demonstrate that the passages were lost any more than it demonstrates that the manuscripts in question had a common ancestor. In fact, both claims, as we know, are counterfactual. The argument only appeared persuasive due to its inherent and mul- tiple circularity. This raises the question(s): Why entangle oneself in logical knots to prove the passages were lost from the tradition? Why argue for their loss against all evidence? What is at stake for Bigger in this issue? To answer these questions we must briely return to our previous discussion. Recall that in the previous section Bigger identiied the constituted text with a hypo- thetical “normative redaction” and, from the absence of certain passages in the former, concluded that they must also have been absent in the latter. He then inferred from this absence that the passages must have been removed from the normative redaction and, ultimately, cited this “removal” as evidence of the normative redaction’s historical reality, even though this “removal” was only apparent rather than demonstrated. At that stage, Bigger was only interested in showing that the Brahmans eliminated certain passages from the oral epic. In contrast, he now seeks to show that the Brahmans not only elimi- nated passages from the oral epic but also added passages to it, and here the constituted text can no longer serve him as evidence because he seeks examples of passages added to the tradition. Although in one sense everything that is not explicitly a remnant of the hypothetical Kṣatriya tradition that preceded the Brahmanic redaction is an addition to the tradition (and therefore he did not need to seek examples of passages beyond the archetype), Bigger interprets this demand to mean passages whose superluity over the archetype’s text is evident. The only passages that satisfy this criterion are the appendix passages. When restored to the constituted text, their nonorganic nature is palpable (which is, of course, why they were removed from it in the irst place). As later additions to the archetype, these passages are also markedly Brahmanic. They thus eminently satisfy Bigger’s criteria. The problem is: though excellent examples of the kind of materials the Brahmans allegedly added to the oral epic, the passages are not actually part of the constituted text. As passages lacking uniform attestation, they were considered additions to the tradition after the archetype and hence moved to the critical apparatus, whereas Bigger 8 88 PHILOLOGY AND CRITICISM needs to show they were added to it during its formation. His solution is to propose that the passages were contained in the archetype (which he identiies with his hypo- thetical “normative redaction”) but were “lost in the course of transmission in certain manuscripts.”140 This explains their absence in some recensions, while simultaneously permitting him to restore them to the constituted text, as his theory of a redaction requires. Thus, whereas the editors removed all such passages to an appendix, Bigger argues in a second step that some of them (speciically, nos. 63, 75, 80, 81 and 85 from the Ā diparvan and nos. 4 and 13 from the Sabhā parvan) should have been retained as at least as old as the rest of the “normative redaction.”141 As in the irst stage of his argument, when he randomly restored passages to the constituted text, here also he randomly restores passages to it except, whereas he earlier justiied their restoration on the grounds that they were passages the Brahmans removed during their redaction of an earlier oral epic, he now justiies restoring them on the grounds that they are passages the Brahmans added during their redaction. What he fails to realize thereby is that no evidence exists that the Brahmans either added passages to or removed passages from an earlier oral epic, and the only person making changes to the text is he himself ! Either way, the Brahmans cannot win. If parts that Bigger thinks belonged to the ear- lier Kṣatriya epic are not in the text, he blames the Brahmans for removing them. But if parts that he thinks they added to the text during their redaction are not in the text, he attributes their absence to accidental loss and still blames the Brahmans for adding them. Bigger’s work is the most egregious example of German anti-Brahmanism, but it is not the sole one. The thesis of a redaction, whether called inal, normative or Brahmanic, ultimately reaches back to Adolf Holtzmann Jr.142 In contrast, the idea of a Brahmanic “takeover” of an earlier Kṣatriya tradition dates back to Christian Lassen, where its origins were clearly racist.143 German Mahā bhā rata critics argued sophistically and dis- honestly for a critical edition.144 They were never interested in a secure text.145 Rather, they feigned interest in textual criticism because only thus could they sustain the illusion of objective inquiries and of binding procedures and results. Everyone in the ield oper- ated under an as if: write as if the Kṣatriya epic existed; as if the Brahmans had corrupted it; as if a redaction occurred. From Holtzmann to Bigger, Indology unfolded within this as if.146 Degrees were granted based not on the quality of evidence or arguments, but on the extent to which students conformed to this as if.147 Scholars were cited based on the extent to which they assimilated themselves to this as if.148 Corresponding to this as if of the Mahā bhā rata tradition was a second as if: write as if the professor was infallible; write as if a genuine intellectual tradition of Indology existed; write as if the German critics were beyond criticism.149 Arguing like the Protestant theologian Johann Jerusalem, who wrote: “My experience is my proof ” (meine Erfahrung ist mein Beweis), the German Indologists needed no proof of what the Brahmans did beyond their experience of the work.150 Arguments were superluous because they did not seek to demonstrate anything. At best, arguments had a rhetorical value in that they conirmed the basic experience of the work or provided a means, in communal experience, to return again and again to the basic precept of Brahmanic corruption. 89 ARGUMENTS FOR A HYPERARCHETYPAL INFERENCE 89 Notes 1 “Notre seule source, ce sont les manuscrits eux-mêmes, donc, en dernière analyse, les stemma. Nous n’avons donc pas le droit de répudier leur témoignage sous le prétexte qu’il nous paraît absurde.” Robert Marichal, “La critique des textes,” in L’histoire et ses méthodes, ed. Charles Samaran (Paris: Éditions Gallimard, 1961), 1285. 2 See the introduction to this work. In support of his claim “The MBh, which presents itself in the manuscripts, goes back to a uniform redaction that was ixed in a written form at some time and to which further texts and streams [sic] were added by later copyists. In that case a critical edition makes sense, since it represents an attempt to reconstruct this normative redaction,” Bigger cites ive sources:  Georg von Simson, “Altindische epische Schlachtbeschreibung, Untersuchungen zu Kompositionstechnik und Entstehungsgeschichte der Bücher VI bis IX des Mahā bhā rata” (Habilitationsschrift, University of Göttingen, 1974), 281–86; Franklin Edgerton, “Introduction,” in The Sabhāparvan for the First Time Critically Edited (Poona: Bhandarkar Oriental Research Institute, 1944), xxxvi; Peter Schreiner, “Die Hymnen des Viṣṇupurā nạ:  Materialien zur Textanalyse des Viṣṇupurā nạ” (Habilitationsschrift, University of Tübingen, 1980), 11–13; Ian Proudfoot, Ahiṃsā and a Mahābhārata Story: The Development of the Story of Tulādhāra in the Mahābhārata in Connection with Non-violence, Cow Protection and Sacriice, Asian Studies Monographs, n.s., 9 (Canberra: Faculty of Asian Studies, Australian National University, 1987), 37–39; and Mahesh M.  Mehta, “The Mahā bhā rata:  A Study of the Critical Edition with Special Reference to the Suparṇā khyā na of the Ā diparvan:  Part II,” Bhāratıȳ a Vidyā 32 (1972):  3–72. Bigger writes: “This was already postulated by Georg von Simson (1974), p. 281–286. However, in contrast to him I do not believe that the version represented in the Critical Edition was the irst comprehensive MBh. Rather, it is the youngest and therefore still attainable normative redaction. Cp. also EDGERTON:  CE 2, p.  XXXVI, SCHREINER (1980), p.  11–13, PROUDFOOT (1987), p. 37–39 und MEHTA (1972), p. 64f. (this work was unfortunately not completely accessible to me. Cp. therefore also the synopsis in MEHTA (1965), p. 64– 67.).” Andreas Bigger, Balarāma im Mahābhārata: Seine Darstellung im Rahmen des Textes und seiner Entwicklung (Wiesbaden: Otto Harrassowitz, 1998), 14 and 14, n. 76. Contrary to Bigger’s suggestion, neither Edgerton nor Proudfoot support his thesis. Proudfoot, for example, notes: “The intensity of variation revealed by the Poona collation is not a mark of oral transmission; nor does it indicate that the archetype reconstructed by the Poona editors was not a written text. The very notion that extant readings can be collated and compared in order to reveal an older reading, and that these older readings should combine into a coherent text, both make sense only if all the collated manuscripts have sprung from an archetypal text which had a ixed written form. In his introduction to the Sabhā parvan, Edgerton has unequivocally accepted this view:  ‘… with due allowance for many minor uncertainties, and for errors in editing, [the reconstruc- tion] is a text … which once existed, and from which all MSS. of the work known to us are directly descended. … I believe irmly in the one-time real existence of this text.’ Since the whole conception of the Critical Edition rests four-square on these assumptions it is curious that Sukthankar and Belvalkar have not felt able to concur fully with Edgerton’s irm statement. Sukthankar emphasizes the ‘luidity’ of the Mahā bhā rata tradition, going so far as to say that the two principal recensions are ‘in the inal analysis, independent copies of an orally transmitted text.’ Belvalkar seems to agree. If this were true, a stemmatically-based critical treatment of the whole tradition would be a travesty of reality, and the presumptions under- lying the Critical Edition project would be made invalid. However Sukthankar’s reluctance to subscribe to Edgerton’s position is due to his misunderstanding of the nature of the oral tradition. The deference of Sukthankar and others to oral transmission is probably attrib- utable in part to the high value placed upon committing to memory and reciting from memory in the Indian learned tradition. At the back of every researcher’s mind is the knowl- edge that the Vedas have been accurately preserved orally over three millennia. But the Ś ruti 90 90 PHILOLOGY AND CRITICISM ofers no good analogy for the epic tradition. With the uncreated Ś ruti whose every syllable has power and which lies at the heart of brā hmaṇic orthodoxy, the high premium placed on absolutely accurate transmission led not only to the development of special mnemonic techniques but also to a copious ancillary literature in written form which helped to stabilize the tradition. The Mahā bhā rata has lacked the incentive, the prestige, and the apparatus for such a controlled oral transmission. So, while Sukthankar may have thought it attractive to use orality as an explanation for the Mahābhārata’s untamed text tradition, in fact evidence deducible from the text tra- dition itself overwhelmingly points to an archetype which had written form. The technical evidence of orally-transmitted epic traditions reveals that oral recitation does not and cannot aim to reproduce an archetype exactly. Rather, in every such oral recital, the bard recomposes a relex of an epic prototype. Each performance is a unique and unrepeatable event; each bard has an individual style and interprets the prototype individually. The prototype is a disembodied concept which lives vestigially in the collective memory of the audience and with sharper deinition in the mind of the bard. The prototype is incarnated only in the bard’s expressions of it. Thus, in recital, there is no question of striving for precision in reproducing the exact form of the prototype, for it has none. In the absence of a ixed form of the tradition, there is no absolute standard against which to measure the variety of relexes which ind expression. Only the transmitter who handles written material has the means to identify errors (viz. departures from a standard) and the liberty to cross-check, to reconsider, and to make corrections in his text. The ephemeral nature of the spoken word means that for oral transmitters imprecision and variation are not and cannot be an issue. It is under- standable then that oral relexes of a prototype vary considerably both broadly and in detail. It is therefore nonsense to imagine it possible to reconstruct a prototype ‘text’ by comparing records of its oral relexes. Not only has a prototype no form in this sense, but the relexes would be so divergent that comparison of the kind required would not be feasible. […] It is thus beyond question that in the Mahābhārata collation we have to deal with a manuscript tradition descending from a written archetype. But this is not to say that the development of the Mahā bhā rata text tradition has not been afected by the environ- ment of a semi-literate culture. In this milieu, committing to memory and reciting from memory have been an important part of scholarship. A  cluster of elements is involved, centring upon the personal relationship between the guru and his students, which goes with oral instruction and requires that the student be a receptacle for knowledge rather than a critical or innovative recipient. These elements are at least in part products and perpetuators of the scarcity of written material before printing. Such a milieu will have tended to impart to a written tradition the characteristic ‘luidity’ which has led Sukthankar and others to waver on the question of oral transmission. The tradition becomes susceptible to interpolation: scholars who have inter- nalized the epic idiom and metre will have been capable of giving shape to new material as required. The tradition tends to show intensive cross-infection between texts:  a scholar’s memory and public recitations constitute channels of infection which do not involve direct contact between manuscripts. Under such conditions the tradition loses some of the predict- ability and conservatism of traditions transmitted merely by copyists; but its instability is an attribute not of any text but of the overall development of the tradition. Each manuscript text, including the archetype, has been just as ixed as writing can make it. There is a world of diference between a scholar committing to memory and reproducing extensive parts of the Mahābhārata as recorded in manuscripts, and a recital of the tradition by an illiterate bard.” Proudfoot, Ahiṃsā and a Mahābhārata Story, 37–39 (Sukthankar’s italics his own; all other italics added). Edgerton’s comments are discussed later; Schreiner’s views deserve a separate treatment. 3 Besides the scholars cited in this chapter and the last, see also Tamar C. Reich, “Ends and Closures in the Mahābhārata,” International Journal of Hindu Studies 15, no. 1, The Mahābhā rata— Perspectives on its Ends and Endings (2011): 12 and Johannes Bronkhorst, “Archetypes and Bottlenecks: Relections on the Textual History of the Mahā bhā rata,” in Pūr vāparaprajñābhinand am: East and West, Past and Present. Indological and Other Essays in Honour of Klaus Karttunen, ed. Bertil 91 ARGUMENTS FOR A HYPERARCHETYPAL INFERENCE 91 Tikkanen and Albion M. Butters, Studia Orientalia 110 (Helsinki: Finnish Oriental Society, 2011), 43, n. 5 and 51, n. 19. Parts of this article, albeit with additional references to Bigger, have been repurposed in Johannes Bronkhorst, How the Brahmins Won: From Alexander to the Guptas (Leiden: Brill, 2016). See especially 64, n. 185, 88, n. 246 and 96, n. 265. 4 Bigger, Balarāma im Mahābhārata, 14. 5 Ibid., 15. 6 Ibid. The claim is erroneous because, while vulgatae such as Nīlakaṇṭha’s edition were popular and widely copied (the majority of Mahā bhā rata manuscripts examined for the critical edition were of this edition), they had little appreciable efect on the manuscript tradition. Certainly, there is no evidence that Nīlakaṇṭha’s edition ever “suppressed” other editions. Scribes appear to have copied whatever versions existed and, above all, their locally available version(s). The vulgate’s inluence is seen primarily in horizontal transmission of new narratives, but this holds for almost every version. It did not lead to an appreciable decline in exemplars of other versions, as can be seen from the fact that our reconstructed archetype is not of Nīlakaṇṭha’s edition, as Bigger’s thesis, were it correct, requires, but of a much earlier text, which is, in fact, much closer to the text of the Ś ā radā version. 7 For the origins of this thesis, see Chapter 2, “The Argument from (Postulated) Antiquity and the Argument from Ideology.” Like much else in contemporary Mahā bhā rata studies the thesis’s origins ultimately go back to the work of racist and anti-Semite Christian Lassen (1800–76), where it served the function of airming a heroic Indo-Germanic identity. 8 Bigger, Balarāma im Mahābhārata, 18. 9 Ibid., 14. 10 Oskar von Hinüber, “Der vernachlässigte Wortlaut, Die Problematik der Herausgabe buddhistischer Sanskrit-Texte, Zur Überlieferung, Kritik und Edition alter und neuerer Text,” in Kleine Schriften, Teil I, ed. Harry Falk and Walter Slaje (Wiesbaden: Harrassowitz, 2008), 436. 11 Hermann Oldenberg, Das Mahābhārata:  Seine Entstehung, sein Inhalt, seine Form (Göttingen: Vandenhoek & Ruprecht, 1922), 2. Compare Holtzmann Sr.: “The ancient epic legends of the Indians are, indeed, contained in the Mahā bhā rata, a not very old composite creation [Machwerk], but not in their complete extent, not in their pure form, but repeatedly revised, mutilated, expanded, [and] defaced.” Adolf Holtzmann Sr., Indische Sagen. Zweite verbesserte Aulage in zwei Bänden, vol. 1 (Stuttgart: Adolph Krabbe, 1854), vii. 12 Sukthankar’s original stemma is reproduced as Figure 79 in Appendix 15. This stemma is mod- iied to clarify the stemmatic relationships. 13 Reeve traces the principle’s origin in Michael D.  Reeve, “Shared Innovations, Dichotomies, and Evolution,” in Manuscripts and Methods: Essays on Editing and Transmission (Rome: Edizioni di Storia e Letteratura, 2011), 55–103. He inds that “the irst explicit formulation remains Paul Lejay’s in 1888.” Ibid., 68. Froger provides a succinct account of the principle: “The method known as the ‘common error method’ reconstructs the genealogy of the manuscripts by taking as the sign of a common origin, not just the common characteristics (internal or external), nor the common readings (good or bad), but only the erroneous readings. Its governing principle is thus, ‘common errors entail a common origin.’ ” Jacques Froger, La critique des textes et son automatisation (Paris: Dunod, 1968), 41. Dembowski rightly notes: “Lachmannian classiication was based on the crucial idea of common error, so much so that the method has often (and rightly) been called ‘the common error method.’ ” Peter F. Dembowski, “The ‘French’ Tradition of Textual Philology and Its Relevance to the Editing of Medieval Texts,” Modern Philology 90, no. 4 (1993): 514. 14 This stemma features 1,319 manuscripts generated at random descending in three families from α, β and γ. As in Figure 4 in the introduction, enclosed circles indicate lost exemplars. In con- trast to Figure 4, however, only the surviving manuscripts (1–85) are numbered. Additionally, seven manuscripts or nodes—δ, ε, ζ, η, ι, κ and λ—are given special names because of the potential for the contamination at these points to mislead us in our inferences. The primary 92 92 PHILOLOGY AND CRITICISM families (that is, those descending from α, β and γ) were colored blue, red and indigo, respec- tively. Intermediate shades were chosen for the subfamilies descending from composite sources (that is, sources descending from two primary families, a primary family and a subfamily or two subfamilies). The eventual archetype θ (whose status is evident on performing eliminatio codicum descriptorum) is not the sole manuscript in existence at the time. It is also not the source of the largest number of copies or the largest family (which remains γ), as is often erroneously thought. While the traditions of vulgatae, which generate large numbers of copies, are more likely to survive, this does not justify us in automatically assuming that the archetype must have been a vulgate version. If a single manuscript from a diferent branch of the tradition sur- vived, the archetype would be an ancestor of the vulgate(s). Vice versa, if for some reason the descendants of vulgate texts situated higher in the stemma ceased to be copied, the vulgates would have no appreciable efect on the tradition. 15 Bigger, Balarāma im Mahābhārata, 14. 16 See ibid., 15–16: “Setting out from the hypothesis of a normative redaction […] I have in the CE [critical edition] of the Mahā bhā rata a text that was a historical fact at a speciic time. When this time precisely was cannot be determined at present. In my opinion, our knowledge at present does not suice for a precise dating of the normative redaction. However, this does not hinder me from seeing in the text constituted in the CE a text that was a historical fact at a point in time unknown to me.” 17 See Paul Maas, Textkritik, 4th edn. (Leipzig: B. G. Teubner, 1960), 6: “ The source with which the irst branching began is called the archetype. Its text is free from all errors that arose after the branching and hence closer to the original than the text of any witness. If we can reconstruct its text successfully, the constitutio is advanced signiicantly.” 18 It is crucial that the revision be extensive, but not complete, for, as Bigger acknowledges, this renders the search for remnants of the oral epic tradition impossible. See Bigger, Balarāma im Mahābhārata, 98:  “Finally one could also assume that the normative redaction subjected the entire Mahā bhā rata to a comprehensive revision. In this case, we could bring our work here to an end, since we could then no longer say something deinite about earlier strata.” If there is a single hermeneutic principle guiding German Mahā bhā rata scholarship, it is this: the Brahmanic redactors succeeded in defacing the original heroic epic greatly, but they could not erase it completely. That the search for the original is no less chimerical in the case of a nonex- istent revision than in the case of a complete revision apparently does not strike them. 19 Ibid., 111. 20 Bigger sees his work as “conirming” a basic suspicion of the German critics since Lassen irst articulated it in 1837, namely, that the Mahā bhā rata underwent a “diaskeusis.” See Christian Lassen, “Beiträge zur Kunde des indischen Alterthums aus dem Mahâbhârata I: Allgemeines über das Mahābhārata,” Zeitschrift für die Kunde des Morgenlandes 1 (1837):  77–78:  “For us, that which manifests most clearly in the legends concerning Vyā sa is simultaneously the most important, namely, that the Mahā bhā rata underwent a diaskeusis.” Insofar as Bigger’s work does not prove but merely assumes this hypothesis, its scientiic value is nil. 21 See Georg von Simson, “Die Einschaltung der Bhagavadgı̄tā im Bhı̄ṣmaparvan des Mahābhārata,” Indo-Iranian Journal 11, no. 3  (1968/69):  159–74, speciically his claim that the reason “the editor [Belvalkar] did not see or did not want to see the passage [is] it lies a good twenty-ive adhyā yas back—among them the most important ones of the entire Mahā bhā rata for the devout Hindu.” Ibid., 161. Von Simson’s idea of a common Brahmanic pattern of falsiication extending from antiquity to the most recent period of the critical edition is addressed in Vishwa Adluri and Joydeep Bagchee, The Nay Science: A History of German Indology (New York: Oxford University Press, 2014), 289–91. 22 Bigger, Balarāma im Mahābhārata, 17. Compare ibid., 17 (“the methods applied are those of higher textual criticism”), 19 (“higher textual criticism”), 69 (“an investigation of the MBh based on higher textual criticism”), 70 (“the special problems […] parts of books 1, 3, 12 and 13 present to higher textual criticism”), 83 (“results of higher textual criticism”), 119–20 93 ARGUMENTS FOR A HYPERARCHETYPAL INFERENCE 93 (“higher textual criticism has not yet pressed into these areas”), 121 (“arguments of higher tex- tual criticism”) and 123 (“the criteria of higher criticism”). 23 Ibid., 15. 24 Soulen and Soulen note: “Lower Criticism is an unhappy term, now of infrequent par- lance, characterizing TEXTUAL CRITICISM in contrast to so-called higher criticism, i.e., all other forms of BIBLICAL CRITICISM. The term has fallen into disuse because of its pejorative sound coupled with the increasing acknowledgment that textual criticism is both important and complex.” Richard N.  Soulen and R.  Kendall Soulen, Handbook of Biblical Criticism, 4th edn. (Louisville, KY: Westminster John Knox Press, 2011), 121 (capitalization and block text in original). We agree with them that it is an unfortunate term, but we use it for heuristic reasons. Actually, we should relegate everything the so-called Mahā bhā rata critics have done (excepting, of course, Sukthankar and his team) to the unhappy reaches of “higher anti-Semitism.” For the expression, and its deinition, see Solomon Schechter, “Higher Criticism—Higher Anti-Semitism,” in Seminary Address and Other Papers (Cincinnati, OH: Ark Publishing, 1915), 36–37. 25 Trovato speaks of “staunch ‘reconstructionists,’ who believe that the task of a scientiic edition is not merely to transcribe a manuscript, or to reconstruct the archetype of surviving manuscripts, but to use the archetype as a point of departure, using all available means—linguistic, stylistic or metrical information, historical data, etc.—to try to come as close as possible to the lost original, detecting and correcting, as far as possible, but always as rationally and transparently as possible, the errors shared by surviving copies.” Paolo Trovato, Everything You Always Wanted to Know about Lachmann’s Method:  A Non-Standard Handbook of Genealogical Textual Criticism in the Age of Post-Structuralism, Cladistics, and Copy-Text (Padua: Libreriauniversitaria, 2014), 15. These conditions are simply not given in the case of the Mahā bhā rata, a text whose earliest version is dated to the second century AD. The irony is that those who insist we can reconstruct the “lost original” simultaneously maintain that it was not one but several, that it had not one author but a tradition of “epic bards” and that these bards did not reproduce a single text but freely com- posed episodes as they pleased. The criterion of style in the sense that Trovato means (that is, as an author’s distinctive stamp) simply does not apply here. When Mahā bhā rata critics speak of “stylistic analysis” they mean parts identiied as characteristic of the oral, bardic tradition, the latter itself identiied on the basis of these parts. The circularity of this conclusion evidently evades them. 26 As Most also airms: “Lachmann’s method is mechanical, both in the sense that it must pre- suppose the unthinking transcription of manuscripts if it is to be applied to them and in the sense that the determination of relations of iliation is achieved on the basis of simple rules and calculations of probability. Ideally, choices of manuscripts and of readings based on this method will be rational in that they will depend not on the taste of the individual scholar but on objective evidence that can be mathematized and evaluated.” Glenn W.  Most, “Editor’s Introduction,” in Sebastiano Timpanaro, The Genesis of Lachmann’s Method, ed. and trans. Glenn W.  Most (Chicago:  University of Chicago Press, 2005), 10. Mechanical does not mean the method is unthinking or care is not required in evaluating the manuscripts. It merely means that once the editor has it his manuscripts into a stemma and explained his principles, he will attempt as far as possible not to deviate from those principles. 27 Gianfranco Contini, “La ‘Vita’ francese ‘di Sant’Alessio’ e l’arte di pubblicare i testi antichi (1970),” in Frammenti di ilologia romanza: Scritti di ecdotica e linguistica (1932–1989), ed. Giancarlo Breschi, vol. 2 (Florence: Edizioni del Galluzo, 2007), 961. Eklund is good on the stemma’s role in the latter: “A very serious criticism of the stemmatic method is that it is a single, gigantic vicious circle. […] This objection may seem well founded but it may, nevertheless be rejected. […] As I have just pointed out, the stemma is built up by means of the manifest errors but then it is not these same manifest errors which are corrected by means of the stemma. […] No, the stemma is used to enable a choice to be made as soon as latent errors are encountered.” Sten Eklund, “The Traditional or the Stemmatic Editorial Technique?” Eranos 104 (2006–2007): 12 94 94 PHILOLOGY AND CRITICISM (italics in original). Both passages are cited (Contini also translated) in Trovato, Everything You Always Wanted to Know about Lachmann’s Method, 59. 28 The special status of the archetype is highlighted by the fact that Maas notes, “another name is not available for it. One should therefore not refer to other intermediate members between the original and extant witnesses as the archetype, irrespective of how important they may be under [the] circumstances.” Maas, Textkritik, 6 (emphasis in original). 29 Ibid., 9 (italics added). 30 Ibid., 10 (italics added). 31 See ibid., 10:  “If the transmission proves corrupt, one must attempt to correct it through divinatio. This attempt either leads to an evident emendation or to multiple equally satisfactory conjectures or to the insight that a correction [Heilung] through divinatio cannot be hoped for (crux).” 32 Ibid., 15 (italics in original). 33 Bigger, Balarāma im Mahābhārata, 19. 34 An archetype, understood in the strict sense, only indicates that a reading is later than it—that is, it is a corruption of the original or a more original reading. Bigger, however, means something else with “not belonging to it.” In his view, since lower criticism can only tell us whether the arche- type contained a passage but cannot exclude its existence before the archetype, “belonging to it” must be interpreted literally as to it and it alone. This is Flores’s precise anxiety: the archetype is “non […] la somma delle serie complete” (not the sum of the complete series). Bigger, however, uses it to draw a most perverse conclusion: if the archetype is not the sum of the complete series preceding it, that series could have contained anything, including a Ksạ triya epic. 35 The idea of “text-historical” (textgeschichtliche) as opposed to “text-critical” (textkritische) investigations derives from Paul Hacker’s article “Zur Methode der geschichtlichen Erforschung der anonymen Sanskritliteratur des Hinduismus. Vortrag gehalten auf dem XV. Deutschen Orientalistentag Göttingen 1961,” Zeitschrift der Deutschen Morgenländischen Gesellschaft 111, no. 2 (1961): 483–92. Despite the similarity of names, the two are unrelated. Despite, also, the name textgeschichtlich, the method is unrelated with objective historical investigations based on empir- ical evidence. Indeed, it explicitly rejects the latter. Here is how Hacker deines this specious method:  “From such changes (I mean:  inversions of the text, expansions, interpolations and even individual word variants) we can at times practically read of intellectual-histor- ical processes. And since for the most part we lack direct historical evidence, textual history [Textgeschichte] or, speaking more generally, the method of comparing multiple transmissions is often the sole scientiic means of knowledge for the historical processes. The history of religion of Hinduism in its diferent branches—history of myths, of cult, of religious ethics and laws, of piety—but also the history of philosophy in some of its branches can no longer be carried out scientiically without the use of this method.” Ibid., 489 (italics in original). It is important to note that when Hacker speaks of “textual history,” he means something other than what is meant, for instance, in the Italian school of textual criticism: “a focus on the history of the tradition; a habit of taking into consideration the existence of authorial variations and of multiple versions; a mistrust of evaluating manuscripts merely for their antiquity and apparent quality; use of geographical criteria, and so forth.” Alberto Varvaro, “The ‘New Philology’ from an Italian Perspective,” trans. Marcello Cherchi, Text 12 (1999): 51. Rather, he means the precise opposite: a history constructed whole cloth out of the texts alone. For Hacker’s rejec- tion of any contact with historical methods, see Hacker, “Zur Methode der geschichtlichen Erforschung der anonymen Sanskritliteratur des Hinduismus,” 492: “In all these eforts at a history of Hinduism, the method should, when one sets out from the texts, initially be restricted to philology, that is to say, ethnological, anthropological, archaeological and similar relections should, at irst, be kept strictly apart. Only once we have obtained results can the dialogue with the other historical sciences—however, important and necessary—become truly fruitful.” Oddly enough, although he claims a “philological” pedigree for his method, Hacker also rejects a “merely external, more or less mechanical textual criticism […]. Inner criteria of an intellectual-historical 95 ARGUMENTS FOR A HYPERARCHETYPAL INFERENCE 95 nature must accompany it hand in hand.” Ibid., 488 (all italics Hacker’s). Likewise, when Bigger says that “higher criticism […] can help interpret this passage and locate it text-historically,” he means neither objective historical investigations into the language of the authors and the milieu in which it was read and disseminated nor the abstract reconstruction of the tradi- tion presented in a Lachmannian stemma. Rather, he means the text’s decomposition into its alleged component parts using the “inner criteria” of the Brahmans’ perceived ideology, the Kṣatriyas’ likely interests and similar tendentious arguments. To call these criteria “criteria of an intellectual-historical nature” is, of course, an abuse of that term, possibly only in a milieu where the sole intellectual-historical project of the past centuries has been to prove the evils of “priestly domination.” 36 Vishnu S. Sukthankar, On the Meaning of the Mahābhārata (Bombay: Asiatic Society, 1957), 10. 37 Ibid., 10–11. See also ibid., 30–31 for Sukthankar’s rejection of the search for the lost oral epic: “Higher Criticism would have us search for the lost ‘epic nucleus,’ which is apparently some- thing immensely worth possessing. With that end in view it proceeds by the method of athetizing certain lines, passages, chapters, or even whole books. These are spurious, and all the rest is the work of one great poet. This method has been applied to comparatively more recent and also much simpler works, about whose historical context we happen to be better informed and where it would be much more legitimate. Yet even in these cases it has, as is well known, completely broken down. Very little relection is needed to convince one that a mere process of stripping of what we regard as spurious will not automatically leave us with the pure and unalloyed ‘original.’ As we analyse the poem back towards its source, it proves to have not one source but many. What shall we do then? We know nothing about the hypothetical ‘nucleus.’ Moreover the nucleus we may discover in our analytical adventures is likely to prove to be not the ‘original’ we are looking for, but merely a projection of our own feelings. On the other hand we have got the poem, about which there is no doubt, and we may be able to puzzle out a good deal about its meaning, its inner meaning, if we tried. Let us then focus our thoughts upon that and try to understand it as best as we can. I believe we shall ind in the poem itself something far greater and nobler than the lost paradise of the primitive Ksạ triya tale of love and war, for which the Western savants have been vainly searching and which the Indian people had long outgrown and discarded.” 38 In fact, paraphrasing von Simson, we could say, “The picture we encounter here is by no means a balanced one, but rather, relects the view of the epic’s German critics, and this view is to a great extent driven by their interests and ideologically prejudiced. The interests of the German Indologists, that is, the class interests of the priests and intellectuals, are clearly expressed, above all, in writings from the nineteenth and twentieth centuries.” The quotation is from Georg von Simson, Mahābhārata: Die Große Erzählung von den Bhāratas (Berlin: Verlag der Weltreligionen, 2011), 591. 39 See Adluri and Bagchee, The Nay Science, chapters 1–2. 40 Bigger advocates “normative” over “inal,” noting: “in contrast to him [von Simson] I do not believe that the version represented in the Critical Edition was the irst comprehensive MBh. Rather, it is the youngest and therefore still attainable normative redaction.” Bigger, Balarāma im Mahābhārata, 14, n. 76. Insofar as von Simson himself does not think the revision of the epic ceased with this redaction this is a distinction without a diference. 41 The coincidence between the constituted text and the hypothetical redaction explains the apparent self-evidence of von Simson’s and Bigger’s claims. It is also the reason few scholars detected the problems with their work. Von Simson and Bigger do not demonstrate the con- stituted text is a redaction; they merely declare it so. Once the critical edition is redeined thus, all their conclusions follow with strict necessity. Who will show that a deinitional circle exists here? 42 Maas, Textkritik, 6. 43 Michael D. Reeve, “Reconstructing Archetypes:  A New Proposal and an Old Fallacy,” in Hesperos:  Studies in Ancient Greek Poetry Presented to M.  L. West on His Seventieth Birthday, ed. P. J. Finglass, C. Collard and N. J. Richardson (Oxford: Oxford University Press, 2007), 326–40. 96 96 PHILOLOGY AND CRITICISM 44 Enrico Flores, Elementi critici di critica del testo ed epistemologia (Naples: Lofredo, 1998). 45 Reeve, “Reconstructing Archetypes,” 331–32. 46 Ibid., 333–34 (italics in original). 47 Von Simson and Bigger mean something else than editors when they recognize the archetype as the source of the irst branching of the extant tradition. Whereas editors typically mean a manuscript detected during recensio from which all extant witnesses can be shown to derive, von Simson and Bigger interpret extant tradition to mean the tradition after the Brahmanic redac- tion, the whole of which is assumed to have survived even if some manuscripts did not survive, since in contrast to the preceding tradition it did not undergo conscious decimation. In other words, what the critical edition reconstructs is the irst Brahmanic version of the text. The branching began from this source not just in the sense that all extant witnesses can be traced back to it but that all manuscripts ever were copied from it or from its descendants. 48 Bigger, Balarāma im Mahābhārata, 18. 49 Reeve, “Reconstructing Archetypes,” 335–36 (italics added). 50 “But θ is not the sum of the complete series α, β, γ, etc! Θ stems ultimately from α, because it carries the error b, but it is mistakenly considered to be the sole survivor of an endless series of acts of copying (the librarians’ workshops of antiquity and the reproduction of the text in countless copies). Moving up from the bottom, we keep inding errors common to the entire tradition of the manuscript, but in progressively smaller numbers: in theory reaching zero errors in the original or only those errors that arose in the original. The archetype, therefore, because it is always constructed from the bottom or from the manuscripts that have been preserved, as the bearer of the common (or conjunctive) errors will always be an apparently unique copy.” Flores, Elementi critici di critica del testo ed epistemologia, 54. 51 Reeve, “Reconstructing Archetypes,” 336. 52 Ibid., 336–37. 53 “Θ, whether of the ninth century or of the third to the fourth century, is a bottleneck, the bottom of one funnel and the vertex of a second inverted one [below it] where, [moving] from the bottom up, all the common errors a, b and c (which are all earlier than the ninth century) are collected and, from the top, precipitate all those which we ind in the surviving witnesses XYZ. The aporia consists in this: the bottleneck does not exist, since the errors falling from the top do not coalesce at the bottom of the funnel but fall dispersing radially.” Flores, Elementi critici di critica del testo ed epistemologia, 54. 54 Bigger, Balarāma im Mahābhārata, 17. See also ibid., 19, 68, 70, 83, 119, 121 and 123. 55 Ibid., 18. 56 The notion of a “descendant” whose existence must irst be identiied in the surviving manuscripts requires clariication, since a descendant is not normally found in another man- uscript, which would be contrary to its very meaning. What Bigger means is that some of the passages found in the witnesses XYZ are not descended from the ancestors they are hypoth- esized to have descended from, that is, ancestors of XYZ extending in a line as far back as θ. Rather, they are older than θ, having descended directly from α and other copies that preceded θ, without passing through the alleged constriction at θ. 57 Bigger, Balarāma im Mahābhārata, 111. 58 The stemma is best understood as a hypothetical architecture that connects the surviving points in the tradition bearing in mind their intrinsic relationships. Although only partial in this sense, the stemma nevertheless stands for the whole the way every part of a hologram contains the image of the complete object. The stemma presents a good guide to the tradition, even parts we cannot access. Without evidence to the contrary, we should assume that the remainder of the tradition, the part we cannot reconstruct, unfolded more or less similarly, since manuscript copying presents only a limited number of options. Editors consider stemmata representative in this sense. 59 “[T]he ‘stemma’ […] is not, strictly speaking, the whole genealogical tree for the entire tradition of the text. It is a schema which we hope not to be unfaithful to the real genealogical tree, but in which all the ramiications may not be retraced. […] In addition, the schematic character of the stemma springs from the absence of a rather large number of intermediary copies, lost over 97 ARGUMENTS FOR A HYPERARCHETYPAL INFERENCE 97 the centuries, and because of these missing copies we shall never be able to know with absolute certitude whether a manuscript is a son of another, or its nephew, or grandson.” Pierre Marie de Contenson, “Principles, Methods and Problems of the Critical Edition of the Works of Saint Thomas as Presented in the ‘Leonine Edition’,” Tijdschrift voor Filosoie 36, no. 2 (1974): 356. 60 Compare von Simson, who argues: “All these features [errors in recollection, introduction of synonymous expressions, displacement of verses, etc.], which, together with the efects of the contamination of manuscripts, have made the creation of the critical edition so diicult appear to me, however, to exclude neither the hypothesis of an archetype of the extant manuscripts nor that of an Ur-text, to which these must be traced back. One must only be clear about the fact that this Ur-text was the late work of relatively untalented compilers. […] If there ever existed an Indian Homer, then his work is not our Mahā bhā rata but an earlier version of the epic, now forever lost to us.” Von Simson, “Altindische epische Schlachtbeschreibung,” 285. Bigger does not use the term archetype. Note that although von Simson grants an archetype, he does not mean it in the sense editors do. Rather, he interprets archetype precisely to mean a second work. For him, the archetype is not a more or less faithful copy descended through many acts of individual copying from the original, but “the late work of relatively untalented compilers.” As a reconstruction of this work, the critical edition is an archetype. It is not an archetype in the sense we mean and in the sense that editors use the term. 61 Bigger acknowledges that “such an investigation has much about it that is hypothetical,” but he insists that “as long as we do not ind a manuscript of the Mahā bhā rata from the irst century BCE, we are dependent on such a procedure if we wish to advance beyond the boundary the text constituted in the critical edition places on us.” Bigger, Balarāma im Mahābhārata, 70. 62 Benedictus Johannes Paulus Salemans, “Building Stemmas with the Computer in a Claddistic, Non-Lachmannian Way:  The Case of Fourteen Versions of ‘Lanseloet van Denemerken’ ” (PhD Diss., Katholieke Universiteit Nijmegen, 2000), 14. 63 This stemma is only hypothetical as the contamination at δ, ε, ζ, η, ι, κ and λ could potentially mislead us in our inferences, but it makes an important point. The shape of our stemma is not an accurate guide to the shape of the tradition but it is an accurate guide—indeed, the sole one—to its contents. The Indologists erred in drawing the opposite conclusion. They thought that the narrowing of the stemma toward its apex evidenced a reduction in the number of exemplars as we approach the archetype but, for this very reason, the reconstructed arche- type was not indicative of the tradition’s contents given the “manifest” decimation of all other copies. Who will explain to them that this argument is not only false but also circular? Even if we granted that a decimation occurred, this would not establish that the lost exemplars contained a text signiicantly diferent from the archetype. The only condition under which this conclu- sion is justiied is if we assumed the existence of an alternative tradition, one diferent from the one eventually conserved and, moreover, that someone had an interest in erasing one tradition and conserving the other (since chance would not explain it). In other words, we must assume Brahmanic corruption to “prove” its existence, which was, of course, the German scholars’ aim from start to inish. 64 Much would depend on how we read the evidence of the manuscripts. For instance, without looking at the manuscripts we could not know whether P was a descendant of β with occasional borrowing from another tradition or a descendant of the tradition descending from ε with only occasional allegiance to β. In the latter case, we would have connected P with F through ε and ε itself with O through δ. Alternatively, if the borrowing into ε from the tradition descending from δ was insigniicant, we might think P and F formed a separate subfamily (the family of α) and seek to connect them with β through their common ancestor ω, the irst source of the entire tradition (though, again, it is doubtful we would realize this). The value of a theoretical exercise such as the one presented in these igures is not that it provides deinite answers but that it stimulates relection on contamination. 65 Could the Indologists cite Figure 10 to argue that, whereas the constituted text reconstructs θ, a manuscript in the family of β, there are two other families about which we know nothing, and hence they are justiied in thinking that θ reconstructs the Brahmanic version of the text, whereas 98 98 PHILOLOGY AND CRITICISM either α or γ (or α and γ) could represent the Ksạ triya version, which has not descended to us? Given their perverse attachment to the thesis of Brahmanic corruption of an earlier Ksạ triya epic, we are certain that they will. But four problems arise. First, it is false that we know nothing about the other two families because we have access to β and nothing in the stemma tells us that the irst two families were radically diferent from the third. Second, the stemma reveals that no attempt was made to replace the families of α and γ with θ, as their thesis of a motivated Brahmanic redaction requires. It is only accidental that θ’s descendants rather than α’s or γ’s survived. Third, θ’s readings are actually a good guide to the tradition because θ is an unexcep- tional copy, indistinguishable (except in the matter of its idiosyncratic scribal errors and minor additions) from thousands of similar copies created or in existence at the same time. It is only chance that it, rather than one of its peers, became the archetype. Fourth, the stemma features an example of extra-stemmatic contamination to illustrate that we could not simply claim con- tamination from a hypothetical survivor of the Ksạ triya epic. In the enclosed region on the right, extra-stemmatic contamination from ι is postulated to explain the unexplained good readings of 74, 78, 79, 80 and 81. The manuscripts’ shared good readings, which they share neither with any of the members of the other groups nor with 70, 71 and 72 with whom they otherwise share some errors, suggesting a common ancestor at S, leads us to posit extra-stemmatic contami- nation into a common ancestor V. The Indologists have failed to demonstrate a similar situa- tion in a single instance. All their examples of “contamination” consist of identifying an alleged “Ksạ triya” passage in an isolated exemplar. Most important, this stemma illustrates that the fam- ilies of α and γ survived until quite late, which, by the Indologists’ theory of how Brahmanism became the dominant cultural and intellectual force in India following the Brahmans’ takeover of the Ksạ triya epic, should mean that Indian culture retained its heroic Ksạ triya outlook until quite recently. We should expect to see these manuscripts’ inluence on Indian culture, but the Indologists themselves hold that the Ksạ triya element in Indian culture died out quite early, indeed, as early as the Brahmanic redaction of the Ksạ triya epic. Any stemma in which the destruction of all other exemplars is not concurrent with the creation of the archetype is thus incommensurable with their view that India’s downfall commences with the rise of Brahmanism. 66 The number of copies typically increases over time, thus suggesting a generic triangle as the most common shape for stemmata. But we also know of traditions that were decimated over time, leaving a single manuscript. Their stemmata would trace an inverted triangle. We could not draw these stemmata. But this precisely illustrates the danger of making inferences about a tradition from the shape of its stemma. On the increase in copies over time, see Most, “Editor’s Introduction,” 7–8: “It is only a guess, but probably a good one, that for most of the history of human culture, the normal situation was one that began with a single exemplar to be copied (the source text) and ended up, as result and usually as purpose, with more than one copy of the text (the source text plus the target text, or multiple target texts): transmission usually entailed multiplication. And given that the proce- dure was performed neither by machines nor by gods but by humans, and that humans err, trans- mission always entailed error, and multiplication of copies usually entailed proliferation of errors.” 67 Real trees not only trace complex shapes but, because they are an abstract representation, they can also be manipulated along any vector. Figures 6 and 10 are intended to show that real trees do not have a distinct shape, but even they do not depict the “real” tree, which does not exist. They are themselves abstract representations, albeit more complex ones. We could expand or com- press them. We could move diferent groups around. We could extend individual lines to create highly abstract shapes or we could apply fractal logic to create complex, repeating patterns. Convention dictates that the archetype be located at the top, but, in fact, we could move it to any apex. In the inal analysis, a stemma is only a formal notation for depicting manuscript relationships: it is not the image of their transmission, as is often erroneously thought. 68 Pasquali thought that without a corroborating history of the text, stemmata remained entirely abstract. His sense of the importance of history, knowledge of the language and immersion in the author’s or period’s style is beautifully summarized in this quote:  “Every language is what it is and produces the efect it produces, solely due to its history. Philological inquiries 9 ARGUMENTS FOR A HYPERARCHETYPAL INFERENCE 99 are necessary for me in the irst place in order to understand the words and phrases [of the language], and not merely their color or style. Language is like a stream of water carrying the taste of both the rock from where it sprang and the land over which it lowed.” Giorgio Pasquali, “Arte Allusiva,” in Stravaganze quarte e supreme (Venice: Pozza, 1951), 11. 69 Von Simson and Bigger’s work recalls Nietzsche’s metaphor of hiding and locating a camel: “If someone hides something behind a bush, looks for it in the same place and then inds it there, his seeking and inding is nothing much to boast about; but this is exactly how things are as far as the seeking and inding of ‘truth’ within the territory of reason is concerned. If I create the deinition of a mammal and then, having inspected a camel, declare, ‘Behold, a mammal,’ then a truth has certainly been brought to light, but it is of limited value, by which I mean that it is anthropomorphic through and through and contains not a single point which could be said to be ‘true in itself,’ really and in a generally valid sense, regardless of mankind.” Friedrich Nietzsche, “On Truth and Lying in a Non-moral Sense,” trans. Ronald Speirs, in The Birth of Tragedy and Other Writings, ed. Raymond Geuss and Ronald Speirs (Cambridge: Cambridge University Press, 1999), 147. In this case, the deinition is the bush, and only those who cannot see through it will be impressed with their ideas. 70 Bigger, Balarāma im Mahābhārata, 14. 71 Johannes Bronkhorst, Greater Magadha: Studies in the Culture of Early India (Leiden and Boston, MA: Brill, 2007), 95. 72 See ibid., 96: “The assumption of an archetype that is diferent from the irst written version is attractive in the case of the Mahā bhā rata. This text contains many portions—e.g. the Bhagavadgītā , the Anugītā , the Anuśā sanaparvan, etc.—which are most easily understood as later additions to an older text.” 73 “However, one manuscript, or a small number of them, may attain a position of prestige which causes it (or them) to overshadow all others. Something like this can happen when the irst or most important commentary is written. The commentator may use just one version of the text, perhaps the only one he is acquainted with, or the one he likes best.” Ibid., 95–96. 74 Ibid., 71. 75 Bronkhorst is thus wrong when he writes: “If the commentary becomes well-known, subse- quent readers and copyists may prefer that version of the text to all others. This is one way in which one version of a text may replace all others, and become the archetype of all the manuscripts available many centuries later. This may not however be the only way how this can happen. Manuscripts preserved in major libraries or centres of learning may be copied more often than others, and for this reason become authoritative. Whatever the exact reason in each case, it is important to note that it can and does happen that the manuscript tradition of a text passes through a bottleneck, not necessarily in the sense that there is only one manuscript left at that time, but rather that just one manuscript becomes the ancestor of all those that sur- vive at a given later point in time. The result is that a manuscript that is far removed in time from the original may become the archetype of all those that survive later on.” Ibid., 96. In Figure 10, θ, which eventually becomes the archetype, is not the most copied; α and γ generate far more copies—γ’s descendants constitute a whopping 27.5  percent of our stemma—yet their traditions ultimately die out. Although only hypothetical—we do not have access to real trees—this stemma illustrates a truism that cannot be repeated often enough: the archetype is a manuscript identiied “by chance” during the recensio. No one can say at the beginning of a tradition which source has the greatest likelihood of becoming the archetype. 76 “The present version of the text, which may be a Brahmanical reworking of earlier material, appears to have had as one of its main purposes to teach kings how to behave in accordance with Brahmanical expectations. The need for such an ideological statement, scholars point out, was strongly felt during the aftermath of the Mauryan empire, whose rulers, as we have seen, did not observe the rules of Brahmanical society. The irst Brahmanical reworking of earlier material, and the irst writing down of the Mahā bhā rata, may therefore have taken place during the period in which the memory of the Mauryan empire was still strong.” Ibid., 95. 10 100 PHILOLOGY AND CRITICISM 77 Sukthankar notes: “the constituted text cannot be accurately dated, nor labelled as pertaining to any particular place or personality.” In his view, the constituted text cannot claim great antiquity, a judgment we think he might now revise. See V. S. Sukthankar, “Prolegomena,” in The Ad̄ iparvan for the First Time Critically Edited, vol. 1 (Poona: Bhandarkar Oriental Research Institute, 1933), ciii. 78 Could the German Mahā bhā rata critics use this stemma to claim that as our stemma traces only a very small part of the historical reality, the archetype must occur quite low down, thus increasing the likelihood of a Brahmanic redaction? We can make four objections: (1) even with a few survivors we can return quite far back in the tradition; (2) the hypothetical Brahman redactors would have to know they had to redact just this source, which is impossible because at the time there was no way to know that this exemplar rather than any other would produce descendants that survived; (3) a “Brahmanic” text is attested at least as far back as the eighth century and possibly earlier; (4) without corroborating evidence, no reason exists to assume a sudden deviation in the text. Although additions and changes can occur to a tradition over time, they are typically due to scribal error, insertions and commentarial glosses. Nowhere in the world has a complete “inversion” in a text’s sympathies and content been observed as is posited for the Mahā bhā rata. The Indologists’ scientism, provincialism, bigotry and incapacity to think outside their hackneyed categories and interpretive schema would have provoked Gaugin’s scorn: “Brains unitted for intellectual quests, having no consciousness of life save as eating and drinking, with no real aim except to obey a rule, covered with a mantle of hypocrisy that is worn with contempt by other male virgins.” Paul Gaugin, Gaugin’s Intimate Journals, trans. Van Wyck Brooks (Mineola, NY: Dover Publishing Company, 1997), 94. 79 Flores, Elementi critici di critica del testo ed epistemologia, 50. 80 Bronkhorst provides the best evidence of this. He writes: “The Mahā bhā rata, as it has reached us, is clearly a Brahmanical text, which misses few occasions to preach a Brahmanical vision of the world. The role and the duties of kings, in particular, receive ample attention. This is hardly surprising in a text whose main narrative tells the story of a war between kings who disputed each others’ claims to kingship. The present version of the text, which may be a Brahmanical reworking of earlier material, appears to have had as one of its main purposes to teach kings how to behave in accordance with Brahmanical expectations. The need for such an ideological statement, scholars point out, was strongly felt during the aftermath of the Mauryan empire, whose rulers, as we have seen, did not observe the rules of Brahmanical society. The irst Brahmanical reworking of earlier material, and the irst writing down of the Mahā bhā rata, may therefore have taken place during the period in which the memory of the Mauryan empire was still strong. Some scholars go one step further and point out that the Mahā bhā rata emphasizes that kings should be Kṣatriyas.” Bronkhorst, Greater Magadha, 94–95. 81 Oliver Hellwig, “Stratifying the Mahābhārata: The Textual Position of the Bhagavadgītā,” Indo- Iranian Journal 60, no. 2 (2017): 132–69 applies quantitative methods (a Latent Dirichlet Allocation) to the Bhīsm ̣ aparvan to evaluate von Simson’s theories of the Bhagavadgītā’s exceptional position. He concludes that “the proposed method […] substantiate[s] the textual structure that von Simson has proposed for the BhG and its embedding in the BhīP,” but is careful neither to attribute the text to Brahmans nor to make claims about hypothetical redactors (or their activity) and, on the whole, is extremely close about whether the Bhagavadgītā is an insertion in the Mahābhārata. Ibid., 165. In response to a question Hellwig clariied that, beyond noting statistically signiicant linguistic diferences between the Bhagavadgītā and its surroundings, the method makes no statement about the Bhagavadgītā’s “secondary” nature. Compare ibid., 134: “I would like to emphasize that these indings do not imply any statement about the sequence of events that led to the composition of the BhīP or of the Mbh. They can be reconciled [both] with theories that postulate a relatively short duration of composition (Hiltebeitel, 2001; Witzel, 2005; Mahadevan, 2008) [and] with tem- porally more extended models such as proposed by Fitzgerald (2006), because the algorithm does not contain a temporal component.” 82 Trovato, Everything You Always Wanted to Know about Lachmann’s Method, 65. 83 Ibid., 64. 10 ARGUMENTS FOR A HYPERARCHETYPAL INFERENCE 101 84 Ibid. 85 See Leighton Durham Reynolds, Texts and Transmission:  A Survey of the Latin Classics (Oxford: Clarendon Press, 1983), xiii. Reeve cites the passage in the quotation above. L. D. Reynolds and N. G. Wilson, Scribes and Scholars:  A Guide to the Transmission of Greek and Latin Literature (Oxford: Clarendon Press, 1991), 60 note that “after the destruction of the previous centuries many texts survived in one copy only.” They speak of “the period from roughly 550 to 750” as “one of almost unrelieved gloom for the Latin classics on the continent.” Ibid., 85. But the causes of destruction were complex. Although Christianity certainly played a role— through disinterest, oicial neglect and, in some cases, censorship—no evidence exists for a concerted program of decimation as is sometimes asserted. 86 This thesis’s origins ultimately go back to Lassen, “Beiträge zur Kunde des indischen Alterthums aus dem Mahâbhârata I,” 61–86 and Holtzmann Sr., Indische Sagen. See especially Lassen’s comments: “The collection was primarily intended for the warrior caste. […] The Râmâyana and Mahâbhârata really form the literature of the Kṣatriya and the doctrine of the gods and religious view presented therein is not the purely priestly […], but rather, the form the religious doctrine attained in the mind of the warrior caste. […] [T]he Indian priestly caste had a clear understanding of the appropriate means […] for securing it lasting dominance over the minds of the remaining people. It therefore cannot surprise us that we ind epic poetry used as an instrument of an addiction to spiritual domination. […] There was no means as appropriate for raising the warrior caste in the priestly spirit than linking instruction about religious and social laws to the narratives that already enjoyed universal favor. That such an intention was realized in the diaskeusis of the Mahâbhârata appears clear to me from the large number and the extent of the didactic portions.” Lassen, “Beiträge zur Kunde des indischen Alterthums aus dem Mahâbhârata I,” 85–86. Although called a “diaskeusis” rather than a “redaction,” the thesis is the same. For references to “diaskeuasts,” see ibid., 63, 64, 68, 75 and 81, n. 18. 87 Bigger, Balarāma im Mahābhārata, 14. 88 The error is simple enough, but it has had grave consequences. It has wasted the time and minds of a generation of scholars. Naama Shalom, Re-ending the Mahābhārata: The Rejection of Dharma in the Sanskrit Epic (Albany: State University of New York Press, 2017) cites Bigger thus: “Moreover, as is suggested by the scholar Andreas Bigger, whenever MBh scholars rely on the C.Ed., they necessarily adopt the fact that the epic had a textual archetype at some point in history, which this edition represents. Since the manuscripts of the MBh appear to stem from a single source-text, Bigger further argues that the written Sanskrit MBh (i.e., the MBh-saṃhitā of 100,000 verses) was ixed at some point in history and thus came to dominate the entire manuscript tradition of the epic. For this reason, Bigger names this text the ‘normative redaction’ of the MBh, as follows: German philology has coined the word Endredaction (inal redaction) for the archetype of the MBh. I prefer ‘normative redaction’ for the following reason: it does not suggest that this redaction was the last one that ever occurred in the history of the MBh. Rather, it is most probably one among others. What makes it so important is the fact that this redac- tion came to dominate the whole written tradition of the Sanskrit MBh; it became norma- tive for this part of the transmission. […] We may now assume, as a working hypothesis, that the text of the Critical Edition of the MBh is a reconstruction of this normative redaction. […] We then have before us a text which was, at a certain time, a historical reality. […] The normative redaction is thus the earliest text we can possibly establish. If we assume that there was a normative redaction, an archetype that can be reconstructed, we must at least assume that this text, even if orally transmitted, was already ixed. It is therefore no longer a loating oral tradition […] but a ixed oral text, similar to the Vedic corpus, though not so rigidly ixed. It is possible that our normative redaction had the form of a ixed oral text and there is no way of disproving it, but I still think it is more plausible to assume that the normative redaction was the irst written MBh. This may also explain why it was possible for one redaction to have come to dominate the whole manuscript-tradition.” 102 102 PHILOLOGY AND CRITICISM Ibid., 119–20 and see also 212, n.  38. All references are to Andreas Bigger, “The Normative Redaction of the Mahābhārata:  Possibilities and Limitations of a Working Hypothesis,” in Stages and Transitions:  Temporal and Historical Frameworks in Epic and Purānị c Literature. Proceedings of the Second Dubrovnik International Conference on the Sanskrit Epics and Purānạ s, ed. Mary Brockington (Zagreb: Croatian Academy of Sciences and Arts, 2002), 21–33 (Shalom cites from pages 19–24). 89 Bigger, Balarāma im Mahābhārata, 111. 90 Ibid., 111–12. 91 Bigger’s suggestion that “purely theoretically, an empty reference could also be completely empty, that is, refer to no adequate text. This, however, will be the case only very rarely if at all” complicates the picture. Ibid., 112, n. 3. We are unclear about what a “completely empty” reference would look like. Does Bigger imagine every statement in the text must have a matching description? If we grant empty references in the source, their identiication in the target text only becomes more diicult. 92 That is, in the peculiar sense that Bigger understands this term. Actually, we could not infer il- iation even in the latter case, since no way exists to determine which manuscript the damaged copy stems from, except that Bigger thinks he can identify the “source” based on the fact that it must contain an intact reference (or rather, since the source also does not exist, he thinks he can hypothetically reconstruct it by recombining the dissociated elements of the reference). 93 Ibid., 112. 94 Ibid. 95 See Vishwa Adluri and Joydeep Bagchee, “Jews and Hindus in Indology,” paper uploaded to Academia.edu, www.academia.edu/30937643/Jews_and_Hindus_in_Indology (accessed April 14, 2017). 96 Bigger, Balarāma im Mahābhārata, 113–14. 97 Compare Sukthankar, “Prolegomena,” xliv (Sukthankar’s italics):  “In passage No. 114 of App. I, Subhadrā and Arjuna were likewise secretly and hastily married in the presence of gods, ṛṣis‚ and elders, while Balarāma was away from the scene, to legitimize Abhimanyu (the father of the famous Parikṣit and grandfather of Janamejaya, to whom the epic was narrated.” 98 See Bigger, Balarāma im Mahābhārata, 107: “The Bhā rgava layer represents one of the earliest Brahmanic layers in the developmental history of the MBh. However, for these Brahmans the Bhā rgavas were not necessarily their own family, but rather the Brahmans par excellence, the preeminent family. If the Bhā rgava layer is really the earliest Brahmanic layer, the oldest Balarā ma layer must be attributed to the period before it. Here it is most certainly not the Brahmans but the Kṣatriyas—at least as the audience and patrons—who were the bearers of the transmission. This accords with the view that Balarā ma in the central passages is presented as a representative of Kṣatriya dharma. That the Brahmans had little use for such a Kṣatriya is conirmed by the fact that the Nā rā yaṇīya layers, which became inluential at a time when the transmission of the MBh already lay in the Brahmans’ hands, had little to add to the igure of Balarā ma beside Saṃkarṣaṇa. In my opinion, 1.211–213 also belongs among the Balarā ma passages that preceded the Brahamanic layers. The problems that are discussed there belong unambiguously in Kṣatriya circles.” 99 Ibid., 111. 100 The appendix passages 114 and 115 occur in the southern recension and one Devanā garī manuscript (D4); 115 is additionally inserted in K4. The southern manuscripts insert 114 after 1.212.1cd (T3 and G2.4 insert it after 7ab); D4 inserts it on a supplementary folio in a second hand after 2ab. They insert 115 after 1.213.12ab; K4 inserts it in the margin in a second hand; D4 inserts it on a supplementary folio in a second hand. Bigger evidently claims that “the picture changes when we look at the passages removed during the constitution of the critical edition: all manuscripts listed under the siglum S expand the Subhadrā episode by numerous passages. […] The two most important expansions can be found in App., 1, nos. 114 and no. 115,” because he thinks that, taken together, the passages evince the existence of a “continuous” text—at least in the manuscripts that preserved the “parallel version.” See 103 ARGUMENTS FOR A HYPERARCHETYPAL INFERENCE 103 ibid., 113. But this is false because the siglum S does not encompass the same manuscripts both times. The manuscripts comprising S for each of the three parvans are listed next: Ā diparvan Ā raṇyakaparvan Virā ṭaparvan T1 = Melkote, Yadugiri Yatiraj T1 = Lahore, D. A. T1 = Melkote, Yadugiri Math Library MS (without V. College Library, Yatiraj Mahal Library, number). no. 3908. no. 155. T2 = Tanjore, Saraswathi T2 = Tanjore, Saraswathi T2 = Tanjore, Saraswathi Mahal Library, no. 11865. Mahal Library, Mahal Library, T3 = Tanjore, Saraswathi no. 11809. no. 11809 (8245). Mahal Library, no. 11809. G1 = Melkote, Yadugiri Yatiraj G1 = Melkote, Yadugiri G1 = MS lent by Rao Math Library, MS (without Yatiraj Math Library Bahadur Professor number). MS (without number). Dr. S. Krishnaswami Aiyangar, MA, PhD, F.A.S.B., Madras. G2 = Melkote, Yadugiri Yatiraj G2 = Poona, BORI, Mbh. G2 = Melkote, Yadugiri Math Library, MS (without Collection, no. 53 Yatiraj Math Library MS number). (= Institute’s Collection, (without number). G3 = Tanjore, Saraswathi no. 266). G3 = Poona, BORI, Mbh. Mahal Library, no. 11823. G3 = Tanjore, Saraswathi Collection, no. 54. G4 = Tanjore, Saraswathi Mahal Library, Mahal Library, no. 11838. no. 11839. G5 = Tanjore, Saraswathi G4 = Pudukottah State Mahal Library, no. 11851. Library, no. 322. G6 = Tanjore, Saraswathi Mahal Library, no. 11860. G7 = Melkote, Yadugiri Yatiraj Math Library MS (without number). M1 = MS belonging to Chief M1 = MS (without number) M1 = Trivandrum, Palace of Idappali, Cochin. belonging to Ponnokkoṭṭu Library, no. 377 (in M2 = Cochin, State Mana Nambudiripad, Devanā garī transcript). Library, no. 5. Alwaye, Travancore. M3 = Cochin, State Library, M2 = Malabar, Poomulli M2 = Trivandrum, Palace no. 1. Dated Kollam 1013 Mana Library, no. 299. Library, no. 378 (in (ca. 1838 CE). Devanā garī transcript). M4 = MS belonging to M3 = Poona, BORI, Mbh. Kallenkara Pisharam of Collection, no. 33. Cochin. M4 = Cochin, State Library, M5 = Cochin no. 14. Dated Collam (Jayantamangalam); in 1006 (ca. 1831 CE). property of the Paliyam M5 = MS belonging to Family. Mankavu Padinnare M6 = Malabar (Nareri Mana); Kovilakam of Calicut. in private possession. Dated Kollam 1013 (ca. M7 = Cochin (Avanạ pparambu 1838 CE). Mana); in private possession. M8 = Malabar Poomulli Mana Library, no. 297. Dated Kollam 1017 (ca. 1842 CE). 104 104 PHILOLOGY AND CRITICISM As the table indicates, only one Telugu manuscript (no. 11809 in the collection of the Saraswathi Mahal Library, Tanjore) is continuous across all three parvans, and thus can contain Mahā bhā rata 3.79 and 4.44 and the appendix passages 114 and 115. One other Telugu manuscript (no. 155 in the collection of the Yadugiri Yatiraj Mahal Library, Melkote) was apparently continuous (according to the Virā ṭaparvan’s editor, Raghu Vira, “The complete MS. contains the irst ive parvans of the Mahā bhā rata”), but it was not used for the Ā di and Ā raṇyakaparvans, so that we cannot know if it contained App. 1, nos. 114 and 115. See Raghu Vira, “Introduction,” ̣ in The Virātaparvan for the First Time Critically Edited (Poona:  Bhandarkar Oriental Research Institute, 1936), viii. The manuscript bearing the siglum G2 for the Ā diparvan is also con- tinuous (Sukthankar notes, “The MS.  contains the irst 4 parvans:  Ā di, Sabhā , Araṇya and Virā ṭa, written probably by the same hand”; Sukthankar, “Prolegomena,” xxi). It was used for the Ā di and Virā ṭa parvans (Raghu Vira: “The MS. contains the irst four parvans of the Mahā bhā rata”; Vira, “Introduction,” ix), but not the Ā raṇyakaparvan. Sukthankar notes of G1 of the Ā raṇyakaparvan that “The manuscript contains the irst four parvans and a part of the ifth,” but it was used for neither the Ā di nor the Virā ṭa parvans. V. S. Sukthankar, “Introduction,” in The Ar̄ aṇyakaparvan for the First Time Critically Edited, vol. 1 (Poona: Bhandarkar Oriental Research Institute, 1942), ix. Thus even if we grant Bigger’s claim that App. 1, nos. 114–15 are Rʹ corresponding to an original R in Mahā bhā rata 3.79, we would get a relation- ship R → Rʹ only in manuscript no. 11809; for the rest App. 1, nos. 114–15 would not satisfy Sahadeva’s reference unless we granted that he went looking for other manuscripts. 101 For all its complexity—an original reference R → Rʹ that underwent dissociation into R and Rʹ, the referent of which was eliminated but nevertheless survived so that ultimately both parts R and Rʹ descended—the empty reference serves one and one purpose only. It ensures that no part of the original epic is lost, thus enabling Bigger to reconstruct it. If the older epic were well and truly lost, it could not be reconstructed. But if it survived, he would have nothing to do. Hence the pretense of a “lost” element. Not only does this let him validate German prejudices about a Kṣatriya epic lost due to the Brahmans; he can now also “show” that it was not lost after all since fragments survive in the Brahmanic Mahā bhā rata. From his perspective, we only need recognize these fragments and we shall miraculously recover the “lost” epic. 102 Trovato, Everything You Always Wanted to Know about Lachmann’s Method, 64. 103 Bigger, Balarāma im Mahābhārata, 147:  “We thus see that the evidence from the MBh does not conirm the majority of current theories about the emergence and development of the igure of Balarā ma, or they refute it. Interestingly, Balarā ma gains features in the course of his development that science [Wissenschaft] accepts as originally belonging to his igure. This gives rise to the suspicion that science [Wissenschaft] has projected features found in younger texts uncritically into the past. My investigation of the textual history of the MBh clearly demonstrates the questionableness of such an approach.” 104 On Criticism see Lassen, “Beiträge zur Kunde des indischen Alterthums aus dem Mahâbhârata I,” 72 (“For the Criticism of the ancient Indian epic poetry, which is preserved for us in the Râmâyana and the Mahâbhârata alone, I wish to extract the following axioms from what was previously said. The epic narration was originally orally transmitted and this transmis- sion was in the hands of the priestly caste.”); Holtzmann Sr., Indische Sagen, x (“Thus I will- ingly concede that not all features in my replica of the Indian epic will concur with the model, when Criticism at some point has succeeded in reconstructing the latter in its genuine beauty.”) and xiv–xv (“But nonetheless even this poem is far removed from presenting itself to us in its original purity: it has been subject to manifold changes and expansions, and there is need, in respects great and small, of protracted critical work before the Sanskrit text attains a form in which it deserves to be translated.”) and Adolf Holtzmann Jr., Zur Geschichte und Kritik des Mahābhārata (Kiel: C. F. Haeseler, 1892), 9 (“in the inner Criticism of the Mahābhārata lies a problem that is destined to proitably employ many future generations of scholars”). And see also Oldenberg, Das Mahābhārata, 2 (“it still lies in the distant future that this most variably 105 ARGUMENTS FOR A HYPERARCHETYPAL INFERENCE 105 transmitted text of this monstrous work will be puriied and secured by the philological art on which it has a claim”). 105 Sukthankar considers App. 1, no. 114 an example of an addition in the southern recen- sion “[m]ade apparently with the object of correcting the laxity of sexual relations implied in the old narrative.” Sukthankar, “Prolegomena,” xliii. Otherwise also the passage is an example of the southern recension’s rather baroque character. Sukthankar notes: “These additions (respectively omissions) and verbal variants sometimes go to such a length that, at times, there emerges in the end an entirely diferent story. Compare, for instance, the two versions of the highly popular episode ‘Rape of Subhadrā ’ (Subhadrā haraṇa) in adhy. 211–212 of our edition and passage No. 114 of App. I (comprising over 460 lines!). We ind that the Southern version of this story is enriched with many entirely novel and startling features, such as Arjuna’s masquerading as a peripatetic monk (yati), or his ierce battle with the Yā dava forces led by Vipṛthu, which he, of course, routs, alone and unaided, or rather merely with the help of his newly acquired, valiant and resourceful wife, who acts as his charioteer!” Ibid., xxxiv. 106 See Friedrich Nietzsche, Anti-Education:  On the Future of Our Educational Institutions, ed. Paul Reitter and Chad Wellmon, trans. Damion Searls (New York: New York Review of Books, 2016), 44: “One of them writes poetry, and is clever enough to look up words in Hesychius’s dictionary: He is convinced at once that his calling is to modernize Aeschylus, and then he inds people gullible enough to claim that he and Aeschylus are congenial—he, a poetasting criminal!” 107 Bigger, Balarāma im Mahābhārata, 120. 108 Ibid., 120–21. 109 Ibid., 121. 110 All manuscripts except Ś1 and K0–3 (V1 missing for no. 63) insert no. 63 after adhyāya 107 and no. 75 after 1.122.31ab. Matters are more complex regarding the remaining three passages. Whereas Ñ, V1, B, Da, Dn and D1.2 insert no. 80 after adhyāya 128; K4, D4.5 and S insert an additional passage (no. 79) and insert no. 80 thereafter. Likewise, whereas K4, Ñ, V1, B, D and M5–8 (the latter four manuscripts omit lines 49–103) insert no. 81 after no. 80 and T and G2.4.5 do so after no. 83; G1.3.6 insert it after adhyāya 130. M3 omits lines 49–103 and inserts the passage after 1.129.1. Finally, K4, Ñ, V1, B, D and M insert no. 85 after adhyāya 136; T and G1 do so after 1.137.17. 111 Sukthankar, “Prolegomena,” li–lii (italics added). 112 Ibid., l–li: “The K version, though comprising manuscripts akin to each other and clearly distinguishable from those of the Devanā garī version, is by no means—as is natural—quite homogeneous. Only K0.1 represent the version K in a comparatively pure form, while the remaining manuscripts of the group (i.e. K2–6) are really nothing more than misch-codices, being conlated either with γ or with S. On the other hand, just owing to this conlation, some of the other composite Devanā garī manuscripts (particularly D2.5) have so many features in common with K, that they may as well be separated from D and classed under K. The con- tamination of K3–6 with γ is illustrated by the following passages: No. 14 of App. I (found in K4 marg., and Ñ V1 B D); No. 41(in K3.4 and Ñ2.3 V1 B D except D5); No. 42–43 (in K3.4 and Ñ V1 B D except D5). K4 includes passage No. 61 (of App. I) and 1131*, like Ñ B D. The contamination of K2–4 with γ is illustrated by 116*, 119*, 122*, 124*, 125*, 128*, 132*, 137*, 139*, 142* 143*, 144* 145*, 151*, 157*, 160*, 162*, 166*, 167*, 168*, 172*, 173*, 189*, 190*, 191*, 221*‚ 228*, 245*, 281*, 305*, 354*, 372*, 405*, 416*, 417*, 438*, 487*, 490*, 523*, 536*, 564*, 692*, 694*, 824*, 1000*, 1035*, etc., etc. The contamination of K4–6 with S is exempliied by the following among other facts. K5.6 contain 22*, K4.6 25*, K4 49*, K4 (suppl. fol.) passage No. 55 (125 lines) and No. 100 (118 lines), of App. I: all of these are Southern passages. K4–6‚ moreover, contain the Brahmā episode in adhy. 1 (a slippery passage, which migrates from place to place), while K4.6 have found place even for the venerable elephant-headed Gaṇeśa‚ who is unquestionably a late Northern intruder. In 106 106 PHILOLOGY AND CRITICISM K4 these interpolations are written out on separate folios (called here śodhapatra), and inserted at appropriate places, which shows the interpolations on the high road to recognition as gen- uine parts of the Mahā bhā rata.” 113 “This MS. [K5] is incomplete, ending with 1.3.152. It was collated at the Visvabharati, up to 1.2.40, and was then reported to be missing.” “Collated up to the end of adhy. 2 only [said of K6].” Ibid., xiii (Sukthankar’s italics). 114 For Sukthankar’s comments on K4, see ibid., xii: “A carelessly written complete MS., with ळ for ल throughout, which is a Southern trait; written by one hand, but preserved in the Collection in two bundles numbered 565 and 566. Supplementary folios at 2, 114, 150, 151, 205 include certain long passages (some from Southern sources), copied by the same hand; notable among them being the Brahmā -Ganẹ śa interpolation, whose point of insertion is indicated by a small mark made in the body of the text, and the marginal remark atra śodhapatramekaṃ (cf. v.l. 1.1.53). There are some excerpts in margins, intended as glosses. Marginal additions of lines and stanzas are frequent only in the irst 35 folios, afterwards few and far between. Corrections are made with yellow pigment. Colophons frequently contain adhyā ya names, subparvan names, but no adhyā ya number. The copyist was Ganẹ śa‚ son of Trimbaka.” 115 His speciic reasons, outlined in the “Prolegomena,” lii–liii, are as follows: “It was remarked above that ν is the shortest of the extant versions of the Ā di. Let us examine, without bias, this feature of ν. Those passages that are lacking in ν, in comparison with the other versions, cannot all be omissions in ν, whether accidental or intentional. They cannot be intentional omissions, notwithstanding that these missing passages are mostly of inferior character, intrinsically worthless, repetitious, superluous, or inally such as scholars have already (even before the discovery of this version) marked as likely interpolations. For, this Ś ā radā (Kaśmīrī) version of the Ā di is not an abstract or an adaptation. It claims to be the unabridged text itself, in all its fullness, and I see no suicient reason to doubt the a priori presumption that it is not an abridged version. The explanation that primarily with the very object of excising what seems to us to be superluous or repetitious matter, an abridgement might have been intentionally made in the past by some Kaśmīrī redactor or a syndicate of redactors, would be a grotesque distortion of Indian literary and religious tradition. No one in the past found the epic text too long. Far from it. It was perhaps not long enough. Taking away some- thing from the received text of the Mahā bhā rata and passing it of as the original work is a thing categorically diferent from adding something to it. To add small details here and there, embellishing and amplifying the original, would be merely a gentle and lowly ser- vice ad majorem gloriam dei. Even long pieces may sometimes be added, if they are actually found in other Mahā bhā rata manuscripts; and occasionally, even if they are not found in the current manuscripts, provided there is at least oral tradition to support their claims. No doubt the received text contained diiculties and obscurities and repetitions. But they would be merely due to corruptions of the text; the diiculties could be solved and the purpose of the repetitions explained by a really learned Pandit, who knows and understands everything. That the omissions cannot be the result of a preconceived plan to shorten or to improve the text, follows further from two other facts:  irstly, enough digressions and superluities still remain in ν, which would have all been swept away in pursuance of the alleged plan; and, secondly, ν has its own interpolations, albeit they are few in number and short in extent, such as 349* (in K V1 Da D2.4), 451* (K D2), 516* (K except K3 Dn D1), 565* (K except K2), 1499* (Ś1 K D5), 1735* (Ś1 K1 only), 1855* (Ś1 K Ñ1), 2077* (Ś1 K except K2 and Ñ1.3 V1 D2.5), etc. While these so-called ‘omissions’ cannot be all intentional, they can also not be all accidental. The text is continuous and complete in itself. It has no apparent lacunae, as it surely would have had, if the omissions had been due to fortuitous loss or destruction of some intermediate folios of a parent manuscript. It may further be pointed out that many of the apparent ‘omissions’ of ν, in relation to γ or the Vulgate (i.e. Nīlakaṇṭha’s text) are conirmed by the rival recension, the Southern recension; e.g. the Gaṇeśa episode (App. I, No. 1), or the 107 ARGUMENTS FOR A HYPERARCHETYPAL INFERENCE 107 anticipation of the list of the hundred sons of Dhṛtarā sṭ̣ ra (No. 41), or again the story of the birth of Abhimanyu (No. 42), or inally the anticipation of the story of the birth of Karṇa (No. 43) in the Saṁbhavaparvan. These passages are omitted in S no less than in K. In these instances, moreover, the intrinsic probability is wholly on the side of those manuscripts that lack these accretions. It is unnecessary to dilate on the Gaṇeśa episode, which, on the face of it, is a later addition, and which has been dealt with so often by diferent critics. As for the two passages, Nos. 42–43 of App. I, it is suicient to observe that the adhyā ya in which they occur is meant to be a mere list of the dramatis personae, in which each actor in the great drama is identiied as the incarnation of some god, goddess, or titan, taking this or that part in one momentous phase of an all-embracing cosmic movement. The adhyā ya being originally a mere (metrical) list (as it is in the constituted text and the Southern recension), such stories as the account of the birth of Abhimanyu and Karṇa are wholly out of place here, and could not possibly have belonged to the original scheme of the adhyā ya. The contrary supposition only stultiies the original writer, making him out to be an irresponsible lunatic, scarcely a desirable conclusion from the orthodox viewpoint. Likewise many of the apparent omissions in ν in relation to the Southern recension are conirmed by other Northern versions; e.g. the anticipation of the birth of Kṛṣṇā and Dhṛṣṭadyumna (App. I, No. 79), or the Nā ḷā yaṇī epi- sode (No. 100), or the account of a battle between the Kauravas and the Pā ṇḍavas (No. 103), which are peculiar to S. In other words, these ‘omissions’ are documented by the whole of N” (all italics Sukthankar’s). 116 This is the manuscript Sukthankar described as “the oldest extant ms. of the Ā diparvan” and provided full collations of in his 1939 article (see next note for citation). It is now pre- served as NAK 5/356 in the National Archives, Kathmandu. It is also known as A28-5 for its reel number and available in scanned form from the Staatsbibliothek zu Berlin. For details see:  http://134.100.29.17/wiki/A_28-5_Mahā bhā rata. Bigger, following Grünendahl, refers to it as “Ñ4.” 117 V. S. Sukthankar, “Epic Studies VII: The Oldest Extant Ms. of the Ā diparvan,” Annals of the Bhandarkar Oriental Institute 19 (1938–39): 203–4. 118 Bigger, Balarāma im Mahābhārata, 122. 119 Edgerton’s note on the two passages reads: “After 2.12.8, V1 B (B1 marg.) D S ins. (cf. 2.30.1– 6)” (for no. 4) and “V1 B (B1 marg.) Dn D1–3.5.6 T1 G1.3–5 M ins. after 2.28.10: D4 (which om. hapl. stanza 10 and line 1 of the foll. passage), ins. after 9: G2.6 (which both om. hapl. from 9b up to the prior half of line 1 of the foll. passage) ins. after 9a” (for no. 13). Franklin Edgerton, The Sabhāparvan for the First Time Critically Edited (Poona:  Bhandarkar Oriental Research Institute, 1944), 368 and 376. The Sabhā parvan critical apparatus comprises 30 manuscripts: Ś1, K1–4, Ñ1, V1, B1–6, Dn1–2, D1–6, T1, G1–6 and M1–2. Passages 4 and 13 occur in all manuscripts except Ś1, K1–4 and Ñ1. 120 Bigger, Balarāma im Mahābhārata, 121. 121 Sukthankar is persuasive on this issue. He writes: “There has been an extraordinary reluc- tance among scholars to face the fact that the Mahā bhā rata manuscripts may contain and do contain quantities of spurious matter. But there is now no excuse for such recalcitrance. The critical apparatus of this edition contains a unique record of hundreds of lines which are evi- dently and unquestionably spurious. Here is a list of passages from our Appendix, each found in one manuscript only: App. I, No. 2 (in K6 marg.: containing 4 lines); No. 4 (K3: 14 lines); No. 5 (B4: 23 lines); No. 7 (G1: 4 lines); No. 16 (K4: 9 lines); No. 25 (D5: 4 lines); No. 26 (B4: 6 lines); No. 31 (K4: 27 lines); No. 34 (K4: 6 lines); No. 44 (D2: 24 lines); No. 49–50 (Da1: 21 lines); No. 66 (D4: 47 lines); No. 70 (G1: 8 lines); No. 74 (B1: 9 lines); No. 94 (D4: 31 lines); No. 98 (D4: 50 lines); etc., etc. These are passages from the Appendix alone, to which many of them have been relegated on account of either their length or their irrelevancy; but the foot- notes contain hundreds, nay thousands, of lines of precisely the same character. Then there are also lines which are found in only two or three manuscripts, of which I have counted some 108 108 PHILOLOGY AND CRITICISM 300 instances. A number of new additions have been now given by Professor Sastri, who has examined other Telugu and Grantha manuscripts for his edition of the Ā di in the Southern recension. And I am fully persuaded that if we examine yet other manuscripts, we shall still ind fresh passages which had never been seen or heard of before. No sane person would maintain that these are all original passages lost in all manuscripts except the few late and inferior manuscripts in which they happen to occur. It is not always easy, as has already been remarked, to prove that these ‘additional’ passages are interpolations. The epic metre is easy to imitate; the epic grammar is lexible; the epic style is nondescript. The additional lines are generally fashioned with skill, and itted in with cunning. The following interpolated stanzas, by a poet aspiring after higher things, in fancy metre and classical style are rather exceptional: 1859*: bhīma uvā ca | re bhū bhujo yadi bhuvollasitaṁ na kiṁ cit tat kiṁ spṛhā jani sutā ṁ prati pā rṣatasya | jajñe spṛhā tha katham ā gatam ā gataṁ vā prā ṇā dhike dhanuṣi tat katham ā graho ’bhū t || kasya droṇo dhanuṣi na guruḥ svasti devavratā ya mandā bhyā saḥ kurupatir ayaṁ śrīsamutthair vilā saiḥ | re karṇā dyā ḥ śṛṇuta madhurā ṁ brā hmaṇasyā śu vā ṇīṁ rā dhā yantraṁ racayatu punar viddham apy astv aviddham || An interesting instance of a passage which is betrayed by its contents is an extravaganza in some Grantha manuscripts. This bizarre interpolation describes among other things, with circumstantial detail, the marriage of Parā śara and Satyavatī (alias Matsyagandhā ). At this ceremony, the shades of the ancestors of both the bride and the bridegroom are invoked, all the details of a regular Hindu marital rite are minutely observed, and the marriage is sol- emnized in the presence of Vasiṣṭha, Yā jñavalkya and other great Ṛṣis living in the Naimiṣa forest, with the distribution of baksheesh to Brahmins. It is an interesting speculation whether credulity can go so far as to regard even such passages as an authentic part of the original Mahā bhā rata or Bhā rata of Vyā sa‚ just because the passage is found in some Mahā bhā rata manuscripts. The footnotes contain a rare selection of passages that are either palpably absurd, sometimes contradicting the immediate context, or else have little connection with the context in which they lie embedded: quotations, glosses, fanciful additions of details, the jetsam and lotsam of Mahā bhā rata poesie. These bewildering luctuations in the text are quite unique, being peculiar to the Mahā bhā rata. They are not found in the manuscripts of the Vedic literature or in those of grammatical, philosophical, or rhetorical texts or of the works of the classical poets and dramatists. This only proves the Mahā bhā rata was peculiarly liable to inlation and elaboration.” Sukthankar, “Prolegomena,” c–ci. 122 Bigger evidently confuses two senses of popular, namely, frequently encountered and commonly liked or approved and, from the circumstance that the passages occur frequently, draws the erroneous conclusion that they must, therefore, also have been popular in the latter sense. The conclu- sion is doubly erroneous because scribes tended to copy everything in their sources without discriminating as to their popularity (indeed, they could not have known at the time whether the lines or episode they were copying were “popular” because they would most likely not have had access to more than two sources and it is unlikely the passage, if it had already spread by horizontal transmission, would be found in three or four sources but not one, unless the former were from one region and the latter from another). 123 “The following are found in E1 and S generally, not in W (= Ś K), but also not in Ñ1: No. 1 of App. I (but S omits some parts and transposes others); 41*; 42*; 52* (found in K4); No. 2 of App. I (found in K4); 60*; 62*; 103* (= 105* in S; in G1.6 only the irst line); 108*; No. 3 of App. I (in S mostly in diferent places from VBD); 117*; 118*; 121* and 122* (in S in slightly diferent place from VBD); 125*; 127* (not in B6); 132*; 133*; 134*; 145* (not in B1); No. 4 109 ARGUMENTS FOR A HYPERARCHETYPAL INFERENCE 109 of App. I (in B1 only margin!); 196*; 197* (in S in diferent place from VBD); 297* (not in B6); No. 13 of App. I; 302* (= lines 1–4 of No. 14 of App. I); No. 19 of App. I (in K4; not in D4.6; same insertion in expanded form as No. 20 of App. I in S); 329*, lines 1–2 (in K4, not in D4; K1.3 have a diferent secondary insertion); No. 23 of App. I (in K4; not in D4.6).” Edgerton, “Introduction,” xlvi. 124 Ibid., xxxiv–xxxv (all italics Edgerton’s). The reference is to Sukthankar, “Prolegomena,” xcv–xcvi. The complete passage reads: “This leads us to the question of ‘additional’ passages in general. Our attitude with regard to them is quite clear, in my opinion. The irst and fore- most source of our knowledge as to what the Mahā bhā rata comprises, is and must remain the manuscript evidence itself. For example, the question—which seems to trouble a great many people, judging by the inquiries on the point received at the Institute—whether the Uttaragītā , Gajendramokṣa and Anusmṛti are parts of the Mahā bhā rata, must be answered by the manuscripts themselves. If none of our manuscripts contain these passages, it is prima facie evidence that they are not parts of the Mahā bhā rata. There is nothing to suggest that our Mahā bhā rata manuscripts have sufered any serious loss at any time. There never was any lack of manuscripts, many of which were preserved carefully in temples, and which must have been copied repeatedly, for the enhancement of merits. There is no evidence of any break in the tradition at any time or any place, within the conines of India at least. The probable inference is that our manuscripts contain all that was there originally to hand down, and more. What late writers and commentators have said about passages not found in our manuscripts is always a matter of secondary importance; it cannot ipso facto nullify or override the primary evidence of manuscripts. Such extrinsic testimony has only local or personal value; it can always be rebutted by the evidence of the Mahā bhā rata manuscripts. Likewise, whether an episode, adhyā ya, passage, stanza or line may be regarded as belonging to the Mahā bhā rata or not must primarily depend upon whether the manuscripts contain it. Extrinsic evidence, in so far as it is valid, will principally hold good only for the period or locality to which it belongs. Intrinsic evidence may be considered; but, being of a subjec- tive character, it must be used with caution. Our primary evidence being the manuscripts themselves, we are bound to view with suspicion, as a matter of principle, any part of the text which is found only in one recension, or only in a portion of our critical apparatus. Therefore, the evidence for such passages as are contained only in one manuscript, or a small group of manuscripts or versions, or even in a whole recension must be pronounced to be defective. Consequently, all lines belonging to one recension only, and a fortiori such as pertain to a combination of manuscripts amounting to less than a recension, for which there is nothing corresponding in the other recension and which are not absolutely necessary for the context—all lines, in short, with a defective title—have been placed in the footnotes or the Appendix, pending further inquiry regarding their credentials. Such passages are not all necessarily spurious. There might be a hundred good reasons why the questionable passages are missing in a particular recension or version. It might conceivably be, for instance, that the shorter recension represents (as a certain scholar has said) ‘a mutilated and hastily put together composition of the Middle Indian Redactors, who could not lay their hands on all manuscripts of the Mahā bhā rata.’ The shorter version might again be, theoretically, a consciously abridged or expurgated version. Or, more simply, the omission might be due to mere oversight of some scribe who had quite unintentionally omitted the defaulting passage and this mistake of the irst scribe had been perpetuated by the other copyists. And so on and so forth. But all these are mere possibilities. All these reasons in general and particular must be adduced and proved, or at least made probable, in any given case. Moreover, the manuscripts clearly show that there has been in progress, through centuries, constant com- parison of manuscripts. In view of this circumstance, the explanation that the omission of a passage in a whole version might be due to a scribe’s omission loses much of its force. Omission is as much a fact in Mahābhārata textual tradition as addition. And it is fair to demand of a person who alleges the authenticity of such one-recension passages why the rival recension does not 10 110 PHILOLOGY AND CRITICISM contain it. The general condemnation of a recension or version that it is mutilated, merely on the ground that it lacks certain passages that are found in a rival recension or version, is entirely meaningless; for the argument might easily be reversed, so that the controversy will resolve merely into mutual vituperation. What I mean is this. From the fact that one of the recensions, say N, does not contain a certain passage or a certain set of passages found in another, say S, it is illogical to argue that N is a mutilated version; because such an argument can with equal cogency be applied to S, in regard to certain other passages that are missing in S but found in N. The point is so important and at the same time so diicult to grasp that I shall endeavour to make my meaning clearer with the help of a concrete illustration. My contention is this. From the fact that the Southern recension contains, say, the Nā ḷā yaṇī (App. I, No. 100), which is missing in the Northern recension, it would be illogical to argue that the Northern recension is defective or mutilated; because one can, with equal cogency, seek to establish the mutilation or defection of the Southern recension by pointing, say, to the Gaṇeśa passage, which is found only in certain Northern manuscripts and is entirely missing in the Southern manuscripts. The argument could have been employed with greater semblance of reason and plausibility, had there been only a mere plus or minus on either side, but is entirely without cogency in the present instance where there are both additions and omissions on both sides” (all italics Sukthankar’s). 125 Bigger, Balarāma im Mahābhārata, 121. 126 Ibid. 127 There is no manuscript called Ñ4 in Sukthankar’s critical apparatus. The reference is to the manuscript NAK 5/356 mentioned earlier. Grünendahl and Bigger refer to this manuscript as Ñ4 to imply proximity with the manuscripts Ñ1–3, which were actually used in the critical edition of the Ā diparvan. In the following, as a shorthand and to avoid confusion, we retain their usage. 128 Bigger evidently thinks Ś1, K0–3 and Ñ4 “could […] constitute a recension together” because he thinks the fact that these six manuscripts do not feature the same passages is evidence for their closer relation, but he overlooks the fact that if N, their latest common ancestor, did not contain them, they also would not contain them. Even if we assume, as Bigger evidently does, that the passages were contained in the so-called normative redaction, this still does not prove Ś1, K0–3 and Ñ4 are closely related because the passages’ absence in these manuscripts could be due to their absence in N, whereas those descendants of N that do feature them could have copied them from another source. 129 “Our MS.  shows in point of readings the greatest agreement with mss. of the Nepā lī group: quite naturally, as it is written in the same script and belongs to the same provincial version. In particular, its agreements with Ñ3, the best and the oldest MS.  of the Nepā lī version belonging to our apparatus, are unique and striking. The total number of variant readings in our MS.  amounts approximately to a little more than 2500. Out of these, its unique agreements with Ñ3 (or in a few places along with one or two other MSS.) against all other MSS. count over 500. This shows that our MS. deinitely belongs to the Nepā lī version, and is not a copy of a MS. of the Ś ā radā or ‘K’ version—a very important point—though it has marked ainities in many a place with the latter group.” Sukthankar, “The Oldest Extant MS. of the Ā diparvan,” 207 (Sukthankar’s italics). Thereafter Sukthankar appends a list of 60 readings, chosen at random, of the manuscript’s unique agreements with Ñ3. 130 “Here it is necessary to emphasize the curious circumstance that while the new MS. is on a par with Ñ3 in respect of individual readings, it is far superior to Ñ3 in so far that it lacks about ninety per cent of the insertions which Ñ3 has in common with the Vulgate, that is to say, the Bengali–Devanā garī group. It thus gives its welcome support to the Ś ā radā –K group and our Critical Edition as regards that large mass of passages in which these texts are lacking.” Ibid., 208–9 (Sukthankar’s italics). 131 Bigger could perhaps argue that mechanical damage to their common ancestor N explains the passages’ absence in Ś1, K0–3 and Ñ4, whereas those of N’s descendants that 1 ARGUMENTS FOR A HYPERARCHETYPAL INFERENCE 111 feature them copied them back in from another source (these manuscripts in this respect pre- serving the “original” reading, albeit reconstructed or restored after damage to their source). But this explanation is not consistent with the manuscript evidence. K4, Ñ1–3, V1, B, D and M5–8 insert App. 1, no. 81 after App. 1, no. 80; the T manuscripts and G2.4.5 do so after App. 1, no. 83, whereas G1.3.6 does so after adhyāya 130 and M3 does so after 1.129.1. M3, moreover, like M5–8 omits lines 49–103. Likewise, K4, Ñ1–3, V1, B, D and M insert App. 1, no. 85 after adhyāya 136; whereas T and G insert it after 1.137.17. If the passages were original, as Bigger contends, we would expect them to occur in the same place in all the manuscripts that feature them. Even if Bigger were to argue that the fact that K4, Ñ1–3, V1, B and D insert the passages roughly in agreement with the M manuscripts suggests that M was the source of their reconstructed text, he could not explain the variance between the G manuscripts and thus between the G manuscripts and T (in respect of App. 1, no. 81) and between the T and G manuscripts and M (in respect of App. 1, no. 85). Surely, if they all owe their reading to their common source S and, via S, to the archetype θ, they must all fea- ture the passages in the same place? Moreover, if K4, Ñ1–3, V1, B and D owe the restored reading App. 1, no. 81 (after damage to their source N) to M, they should follow the latter’s southern spelling of Kaṇika as Kaṇiṅka but they consistently feature the northern spelling. They should also feature M’s additional insertions (for example, the six lines inserted in all southern manuscripts except T1 and G3 after line 7 and the one-line insertion found in all southern manuscripts after line 177 of App. 1, no. 81) and omit the lines it omits (for example, lines 10–11, omitted in all S manuscripts except T1; line 16, omitted in G3 and M6–8; lines 21–27, omitted in M3.5; lines 21–28, omitted in M6–8; lines 36–42, omitted in M; lines 49–103, omitted in M5–8; line 171, omitted in all S manuscripts except T1; line 184, omitted in T, G2.4.5 and M; lines 188–89, omitted in S except T1 [and Ñ3 and D5]; and line 190, omitted in all S manuscripts except T1). Bigger has not read the appendix materials suiciently carefully. 132 Reeve states the point brilliantly: “The principle about shared errors is merely an adapta- tion of the broader principle to the easily overlooked and seldom controversial fact that one family has already been assembled: the family selected for study. Why is it a fallacy to apply the broader principle within that family? Textual critics usually reply that original readings may survive anywhere in a tradition, which is true but not obviously to the point. It is a fallacy precisely because the principle has already been used for deining that family. Alternatively, it is not a fallacy as long as one remembers that the families now to be deined are sections of that family.” Reeve, “Shared Innovations, Dichotomies, and Evolution,” 59 (Reeve’s italics). 133 Bigger, Balarāma im Mahābhārata, 120–21. 134 Irving M. Copi, Carl Cohen and Kenneth MacMahon, Introduction to Logic, 14th edn. (London and New York: Routledge, 2016), 137. 135 Ibid. 136 Bigger, Balarāma im Mahābhārata, 121. 137 Unlike the fallacy of airming the consequent, which is a formal fallacy, that is, an argument whose invalidity is unrelated to the propositions’ material contents, but simply a consequence of the fact that the conclusion is unrelated to the premises, a circular argument is a valid argument. In a circular argument, the conclusion follows from the premises but trivially so. Bigger’s argument is therefore valid, but it does not expand our knowledge of the tradition. Crucially, it cannot settle the issue of whether the passages are additions in the manuscripts that contain them or omissions in those that lack them. 138 Ibid. 139 The most egregious is Christopher R. Austin, “Vedic Myth and Ritual in the Mahābhārata: A Critical Study of the Mahāprasthānika- and Svargārohaṇa Parvans” (PhD diss., McMaster University, 2007). We cite a selection of the most problematic claims: “I embrace a concep- tion of the Mahābhārata as a text heterogeneous in content and authorship, but consolidated in a comprehensive act of editing which produced a written form of the text by around the 12 112 PHILOLOGY AND CRITICISM 4th or 5th century CE. The present-day Critical Edition of the Bhandarkar Institute represents an approximation of this early written text. This form of the Mahābhārata was authoritative for later tradition, and as such may be referred to as the ‘normative redac- tion.’ The ixing of the text at this point was not an act of creation ex nihilo, but one of redaction: forms of the text had hitherto existed as loating oral compositions/ and almost certainly in earlier written forms as well. The heterogeneity of this normative redaction (which, again, is now available to us more or less in the form of the Critical Edition) is borne out in much of the work of analytic scholarship on the basic text of the Critical Edition, and allows us to infer that the process of growth by interpolation and accretion, which we know occurred subsequent to the creation of the normative redaction, likely occurred in some degree prior to the creation of this written archetype as well.” Ibid., 13–14. “James Fitzgerald refers to this form of the text as the ‘written archetype,’ and writes: ‘The efort to establish a critical edition of the Sanskrit text of the Mahābhārata … revealed that a single Sanskrit version of the “Mahābhārata,” ixed in writing, was at the base of the entire manuscript tradition of the Sanskrit Mahābhārata. […] [The Critical Edition’s manuscripts] point conclusively to a single written “text” of a Mahābhārata at some point in the ancestry of these manuscripts […] The critical edition of Poona is the closest approximation to the archetype behind the manuscript tradition we will ever get.’ In the accompanying notes, Fitzgerald adds: ‘The amount of unity, both petty and general, that exists among the MBh manuscripts […] can be explained only on the assumption of a ixed text antecedent to those manuscripts, an archetype. For the variations which exist can be explained as later, particular innovations resulting from various dynamic factors in the tradition, while the unity cannot be explained, generally, as parallel independent invention.’ Fitzgerald, ‘India’s Fifth Veda,’ 152–53. On the whole I embrace Fitzgerald’s conception of this earliest infer- able form of the text, but prefer Bigger’s term ‘normative redaction’ (see below) since the term ‘archetype’ may tend to suggest a simple and homogenous text in contradistinction to a later complex and developed one, and this is not the case with the text of the Critical Edition.” Ibid., 14, n. 32. “I borrow the phrase ‘normative redaction’ from Andreas Bigger [...], who articulates a position similar to that expressed by Fitzgerald, but with important qualiications. Bigger characterizes the early written form of the text as the ‘normative redaction,’ likening it to ‘a kind of screen-shot,’ (20) emphasizing thereby the luidity of the text both prior and subsequent to the creation of what Fitzgerald refers to as the ‘written archetype.’ ” Ibid., 14, n. 33. “Hence I do not conceive of this early form of the written Mahābhārata—the written archetype or normative redaction—as the product of unitary authorship per se. That is to say, I do not believe that the material constituting the written archetype came about as an actual literary creation by design of a single writer. But it does seem likely, given the integrated and cohesive nature of the Critical Edition text, that the normative redaction was prepared in writing and organized in quite a comprehensive act of editing and compilation. Hence, whatever the provenance, authorship and age of the materials brought together in the course of this editing, the normative redaction presents them in a continuous and integrated framework that would appear to make of the Mahābhārata a self-contained piece of literature. Consequently, the position I take which is deinitive for the work of this dissertation is to acknowledge and accommodate a degree of heterogeneity of the poem’s materials, while recognizing the important integrating efect of redaction which the written archetype ixed, preserved, and made authoritative for later manuscript traditions of the Mahābhārata. Once again, it is on the basis of the present day Critical Edition of the poem that I and other scholars infer the existence and character of such a written archetype or normative redaction, and in this dissertation I will build hypoth- eses about the normative redaction based principally upon a close reading of the Bhandarkar Critical Edition text.” Ibid., 15. “Above I  have made passing reference to some of the elements which bespeak the poem’s unity and integrity as a piece of literature: the coherent 13 ARGUMENTS FOR A HYPERARCHETYPAL INFERENCE 113 narrative, consistency in the rendering of the story’s characters, structural symmetry and repetition of motifs across the poem. I would argue that such elements provide us with some examples of the normative redaction’s cohesiveness—a cohesiveness that, as I  have just proposed, is likely attributable to a process of systematic redaction or editing. But in addition, I would argue that the two narrative frames within which the entire Mahābhārata story is presented also provide us with an indication of the integrating and consolidating efect of redaction. These frames are established at the very beginning of the poem, are carried through consistently across the entire text, and closed at the epic’s termination. Moreover, they point to an aspect of the normative redaction which is central to this disser- tation:  the structuring efect of Vedic ritual upon the Mahābhārata’s narrative.” Ibid. Fitzgerald’s claims are equally problematic: “The written Sanskrit Mahābhārata was ixed and promulgated in Northern India between about 300 and 450 CE, that is, about the time of the Gupta empire, and that text became, de fact, an almost normative redaction for written copies of the Sanskrit text (Bigger 1998:  13–19; Fitzgerald 1985:  126–28). […] There is broad agreement that the particular precursors of this text, that is, some kind of ‘Bhā rata’ epic, came into existence and began developing sometime after 400 BCE. There is also broad agreement, though not unanimity (see Hiltebeitel 1999c), that the Mahābhārata has antecedents of some kind in older Indo-Ā ryan, oral bardic literature and perhaps even in some more ancient Indo-European bardic songs about warriors and wars.” James L. Fitzgerald, “The Mahā bhā rata,” in The Hindu World, ed. Sushil Mittal and Gene Thursby (New York: Routledge, 2004), 52–53. And see also James L. Fitzgerald, “Making Yudhiṣṭhira the King: The Dialectics and Politics of Violence in the Mahābhārata,” Rocznik Orientalistyczny LIV, no. 1 (2001): 68–69: “By that term [Mahā bhā rata] I mean a written, Sanskrit text that ‘precipitated out’ of wider, mainly oral, traditions of epic and didactic poetry. I believe this written Sanskrit text was provoked by the rise of the Nandas and the Mauryas, and partic- ularly by the ‘dharma-campaign’ of Aśoka Maurya. I believe it was completed through a deliberate authorial and redactorial efort sometime during or shortly after the times of the Brahmin dynasties of the Śuṅgas and the Kā ṇvas; that is, after the middle of the second century B. C. and before the end of the irst century B. C., though perhaps even as late as sometime in the irst century A. D. This written text then became a major new element operating alongside of and interacting with the oral traditions that preceded it and which certainly persisted after its creation. I believe this written Mahābhārata may have been sys- tematically expanded one or more times between this tendentious, post-Mauryan redaction and 400 CE, thus complicating the traditions of Bhārata, and, or, Mahābhārata further. In addition, during this period there were, no doubt, also additions and excisions in all branches of the manuscript tradition, additions that were neither artistic nor systematic (that is, particular keepers of given manuscripts inserted explanations and clariications, passages which they thought appropriate to transmit as part of the MBh for one reason or another, and sometimes even whole episodes; or they cut away passages or episodes they thought inappropriate). Many of these non-artistic ‘improvements’ were then preserved when the afected manuscripts were copied. At some point around the time of the Gupta Empire […] another written Sanskrit text of the Mahābhārata was created and promulgated out of this complex tradition and this ‘Gupta text’ became, de facto, the normative written version of the text, a version that served as the ultimate archetype of all later Sanskrit manuscripts of the Mahābhārata throughout India for the next 1500 years. This text appears to have absorbed or otherwise eliminated all or most other written versions, though some later variations in the manuscript tradition may represent survivals from pre-normative traditions, written or oral. While demonstrating the existence of this [our italics] archetype, Sukthankar’s efort to establish a critical edition of the MBh on the basis of the extant manuscripts proved untenable, in the end, to retrieve this [our italics] archetype, though he and his colleagues went ahead and gave us a conjectural—though very valuable in my 14 114 PHILOLOGY AND CRITICISM judgment—approximation of it.” The passage includes a reference to two of Bigger’s works: “See my article India’s Fifth Veda, op. cit., and A. Bigger, Balarāma im Mahā bhā rata, [...] pp.  13–19 and A.  Bigger, The Normative Redaction of the Mahā bhā rata:  Possibilities and Limitations of a Working Hypothesis.” Ibid., 69, n. 17. 140 Bigger, Balarāma im Mahābhārata, 120. 141 This appendix is part of the critical apparatus. It forms an enlarged critical apparatus, consisting of insertions too long to include in the regular critical apparatus and a secondary critical apparatus consisting of variants to these passages. 142 “On the whole both editions, C and B [Calcutta and Bombay], do not difer signii- cantly:  they belong to one family and are based on a common foundation of a rigorously conducted, completed redaction.” Adolf Holtzmann Jr., Das Mahābhārata nach der nordindischen Recension (Kiel: C. F. Haeseler, 1894), 9 (Holtzmann’s italics). Hopkins thereafter introduces the term “inal redaction” (according to him, “the place and time of the inal redaction” was east of the “Holy Land” “as early as 400, perhaps 300, CE”) and Kirste then translates Hopkins’s “inal redaction” into the German “Schlußredaktion.” See Edward W. Hopkins, “The Bhā rata and the Great Bhā rata,” The American Journal of Philology 19, no. 1 (1898): 20– 22 and J. Kirste, “Zur Mahā bhā ratafrage,” Wiener Zeitschrift für die Kunde des Morgenlandes 14 (1900): 222–23. 143 Lassen, “Beiträge zur Kunde des indischen Alterthums aus dem Mahâbhârata I,” 61–86. See earlier for Lassen’s references to a “diaskeusis.” For Lassen’s comments about the Mahā bhā rata “as an instrument of an addiction to spiritual domination,” see ibid., 86. And for his comments about the Mahā bhā rata as containing “references to the two races that fought each other in Indian prehistory: the originally native black [race] and the Sanskrit- speaking, light-skinned [race] that had immigrated from the north, whose western racial relatives are, even now, successfully ighting a similar battle with similar superiority over the red races of America,” see ibid., 75 (Lassen’s italics). But a connection between the Brahmanic takeover and the inversion of racial hierarchies will not be made until Holtzmann Jr. See Adluri and Bagchee, The Nay Science, chapter 2. For Lassen, the priestly, “semitic” element is dominant. 144 Winternitz’s comments reveal the tactical aims that the call for a critical edition served. In Moriz Winternitz, “Genesis des Mahā bhā rata,” Wiener Zeitschrift für die Kunde des Morgenlandes 14 (1900): 61 he notes: “I, however, believe that as long as a halfway critical edition of the Mahā bhā rata does not exist, as long as we cannot say at least with some certainty whether a verse or a passage occurs in all recensions or not, we have a duty to examine every indi- vidual śloka for its authenticity before we use it for historical-critical purposes” (all italics in original). Winternitz raised these objections against Joseph Dahlmann, who argued for the Mahā bhā rata’s unity. But Winternitz misunderstood the unity that Dahlmann was positing for the Mahā bhā rata: not a historical unity but the unity of the received text or the unity of the literary context, within which the text is read. The suspension of all intellective activity vis-à-vis Hinduism insofar as it is based on the epic is carried further in Moriz Winternitz, “Promemoria über die Nothwendigkeit einer kritischen Ausgabe des Mahā bhā rata, insbesondere der südindischen Recension,” Almanach der Kaiserlichen Akademie der Wissenschaften 51 (1901):  206–7. There he argues:  “It is clear from the excerpts of the manuscripts that I published that, for all historical and critical investigations concerning the ancient Indian epic, the so- called vulgates and the diferent editions printed in India are completely and utterly unsatisfactory. For the diferences between the printed texts and the text of the south Indian manuscripts is so great that we cannot say with certitude that any passage belongs to the epic as long as we do not know if it also occurs in the south Indian text. Therefore all hypotheses about the ancient Indian epic’s age and emergence, about the genuineness or lack thereof of individual portions, about the greater or lesser age of individual sections thereof will be suspended in the air so long as we do not know what kind of text the diferent manuscripts ofer. All 15 ARGUMENTS FOR A HYPERARCHETYPAL INFERENCE 115 investigations about the epic’s language and style lack a proper foundation so long as a reli- able text is unavailable. All research into mythology, religion, philosophy and law on the basis of the epic is without foundation so long as we cannot with certitude say that a passage cited for some myth, for some religious or philosophical doctrine or for some legal principle is also really found in the diferent recensions” (all italics Winternitz’s). Could Winternitz have anticipated that someday a passage would be “really found” in all recensions, and German Indologists would still argue for removing it, declaring it a Brahmanic “interpolation”? 145 Winternitz argues disingenuously against Dahlmann:  “What else have critics until now done except proceed from the Mahā bhā rata as it currently exists and sought, where possible, to eliminate whatever did not organically cohere with the epic—the reference here is not to an imaginary Ur-epic but to the epic poem contained in our Mahā bhā rata.” Winternitz, “Genesis des Mahā bhā rata,” 53 (Winternitz’s italics). But insofar as Mahā bhā rata critics eliminated “whatever did not organically cohere with the epic” according to their expectations, they precisely were not setting out from the text “as it currently exists.” The prepossession for an “Ur-epic” is very much in operation. In fact, neither the receptus (Dahlmann’s text) nor the constitutio textus (the text of most post-critical edition scholars) should be considered deinitive or have value for us. In the German Mahā bhā rata scholars’ eyes, granting the texts coher- ence is the irst sin (and interpretation a close second). 146 For examples of this as if in operation, see Holtzmann Jr., Zur Geschichte und Kritik des Mahābhārata, 51 (“I cannot decide whether we must here assume a recollection of a commonly lived primordial age or further developments that occurred in parallel under similar living conditions. But it is certain that the Mahā bhā rata has preserved traces for us that reach back to a hoary period of Indian antiquity; that recollections from an Indo- Germanic primordial age are preserved therein can until now at best be made plausible.”) and Hermann Weller, “Who Were the Bhriguids,” Annals of the Bhandarkar Oriental Research Institute 18, no. 3 (1937): 296–98 (“I have always had the impression that the original heroic poem could not have become an Encyclopaedia Brahmanica, a book of Brahmanic conception of the world (Weltanschauung), unless and until that particular psychological trend of India which we epitomize under the name Hinduism had become so vigorous as to overwhelm and vanquish the theosophic outlook of Vedic India. This victory might perhaps have been achieved about the fourth century B. C. The Greek Megasthenes already had before him a hinduized India. Also in the textual history the Great Epic, the decisive step had been taken by that time: the heroic poem had been expanded by additions coloured by the Brahmanic conception of the world (Weltanschauung). But to be sure the epic had not yet attained its pre- sent dimensions. This metamorphosis took place, as has now been shown by Sukthankar, chiely under the inluence of the Bhriguids. Now everybody will admit that the Vedic Age stood more under the inluence of the Aryan than of the Pre-Aryan-Indid tribes, but that the latter in course of time became more and more dominant, and that along with them came to the forefront gods characteristic of the Indian Middle Ages. By the above-men- tioned Pre-Aryan Indid tribes, I  understand representatives of the most eastern branch of the racial stock stretching from the Mediterranean zone via Arabia to Western India, comprising Mediterranean, Oriental and Indid elements. Now I ask: Did this widely ram- iied Brahmin group, calling itself the Bhriguids, belong perhaps to this primeval race, the race which in some epoch of hoary antiquity had caused the earlier inhabitants of India to retreat into mountain fastnesses, the race of the real Gracil-Indid people who (according to Eickstedt) are represented even today in the purest state in Hindustan (Doab) and in the Deccan, that highly gifted race which constituted the ruling power in Mohenjo Daro, and which later produced eminent philosophers? To be sure even in Rigvedic times all Brahmins were not pure Aryans (Nordic); even at that time there must have been pure Indid and Aryan-Indid representatives of this caste. And the further the Aryans spread over India, the greater must have been the number of non-Nordic Brahmins. What has been traditionally 16 116 PHILOLOGY AND CRITICISM handed down to us about the Bhriguids in our Mahā bhā rata contains undoubtedly his- torical reminiscences. Even when they had partly formed alliances with the older military aristocracy by matrimonial ties, at bottom they were inimically disposed towards the ruling class; and oft-repeated legend of Parasu-Rā ma, who exterminated all warriors (Ksatriyas), is evidence of a real tragic conlict in a hoary past. The contradictory traditions to the efect that the later warrior-caste either was the result of marriages between Ksatriya widows and Brahmins or else was propagated by concealed survivors of the Ksatriya caste who had escaped destruction—these conlicting traditions can perhaps be reconciled on the supposi- tion that both eventualities had taken place.”) Bigger himself provides the emblem of this as if logic when he writes: “Just as we can only postulate the occurrence of a normative redac- tion so also we remain dependent on hypotheses concerning the written MBh’s subsequent development. The normative redaction spread gradually across all India. Did it thereby suppress other written traditions? Or did it irst cause other traditions to be written down? At any rate, as the critical edition shows, it is certain later additions encroached on the nor- mative redaction. Insofar as they did not belong to the parallel versions just discussed, these additions were either composed by the scribes themselves, written down from memory (that is, an oral text was written down) or copied from another source. Only the last option can be demonstrated provided the source from which this textual passage was taken is still pre- served.” Bigger, Balarāma im Mahābhārata, 118. 147 A letter to one of the authors provides a good example: “But you do not at all address the question of what the text of the hypothetical redactors is. That of the critical edition? You cannot possibly assume that. As every serious colleague will conirm for you, the critical edition is no more than a convenient working tool [Arbeitsinstrument] with whose assistance one can easily survey a (hopefully) representative selection of manuscripts of (hopefully) impor- tant manuscripts.” Michael Hahn, letter to Vishwa Adluri, February 10, 2010, 4. Apparently, Hahn was unaware that the assumption that the critical edition’s text is identical to the hypo- thetical redactors’ text had already been made and considered lawless scholarship by his colleagues by this time. 148 Again, an email to one of the authors provides a good example: “Dear Dr. Adluri, it does not come as a surprise to hear that. I know the German scene all too good [sic] to be surprised. And to tell you the truth: I was much astonished to come to know that my colleague Hahn acts as your supervisor. I appreciate [very] much that you invited me despite all your troubles with ‘German Indology’ (which in fact should not be reduced to Hahn and Hanneder though I esteem them), for which I feel awfully sorry. There is an (obviously young) colleague of us [sic], Dr. Chris Austin, who has published (at least) one very good paper on the ‘ritual redac- tion’ of the Mahabharata. Do you know it? Yours, Thomas.” Thomas Oberlies, email to Vishwa Adluri, January 4, 2010. The scholar recommended is Christopher R. Austin, whose deference to German ideas of the critical edition as a “normative redaction” we have already seen. The reference is to Christopher R. Austin, “The Sārasvata Yātsattra in Mahā bhā rata 17 and 18,” International Journal of Hindu Studies 12, no. 3 (2008): 283–308. Among the German literature Austin cites is Thomas Oberlies, “Arjunas Himmelreise und die Tīrthayā tra der Pā ṇḍavas:  Zur Struktur des Tīrthayā trā parvan des Mahā bhā rata,” Acta Orientalia 56 (1995): 106–24; Thomas Oberlies, “Die Ratschläge des Sehers Nā rada: Ritual an und unter der Oberläche des Mahā bhā rata,” in New Methods in the Research of Epic/Neue Methoden der Epenforschung, ed. Hildegard L. C. Tristram (Tübingen: Gunter Narr Verlag, 1998), 125–41; Bigger, Balarāma im Mahābhārata; Andreas Bigger, “Wege und Umwege zum Himmel. Die Pilgerfahrten im Mahā bhā rata,” Journal Asiatique 289, no. 1 (2001): 147–66; Heino Gehrts, Mahābhārata: Das Geschehen und seine Bedeutung (Bonn: Bouvier Verlag, 1975) and one reference to a source in a German journal: Herman Tieken, “The Mahābhārata after the Great Battle,” Wiener Zeitschrift für die Kunde Südasiens 48 (2004): 5–46. Clearly, suicient to render it a work of scholarship. 17 ARGUMENTS FOR A HYPERARCHETYPAL INFERENCE 117 149 Hahn expresses the temerity of questioning the gray eminences of German Indology thus: “How could he [Vishwa Adluri], a newcomer in the ield, believe that he has a better insight into the MBh (that he never read in the original language) than a true giant like Oldenberg?” Michael Hahn, email to Greg Bailey, Monday, January 19, 2009. 150 Johann Friedrich Wilhelm Jerusalem, Zweyte Sammlung einiger Predigten (Braunschweig: Rudolph Schroders Erben, 1752), 306. The phrase repeats as a leitmotif throughout his writing career. For more sources, see Karl Aner, Die Theologie der Lessingzeit (Halle a.d. Saale: G. Olms, 1929). 18 What the scribe said: stock phrases Mahābhārata copyists adopted to indicate their concern with idelity to the source Source: Reproduced from Vaidya, “Introduction [to the Harivaṃśa],” xxiv–xxv. The phrases translate, in order, as: 1. I have put in the words as I saw [them]. 2. The good [-hearted] ones should forgive the mistake committed by a slip of pen. 3. I have written as I saw the book [that is, as I saw in the book]. Be it connected or disconnected, relevant or irrelevant, it is not my fault. 4. If there is any letter missing or more or less [etc.] the intelligent ones should forgive it. Who does not commit a mistake? 5. A dot, bad handwriting, a misplaced visarga, broken marks, wrong split of the word [lit., a fault in breaking a word] are all due to hurry or haste in writing and without intention. The good-minded ones should forgive such things. 6. If there is in this treatise, a letter missing, loss of morals or wrong reading, it has all occurred due to folly. The forgiving, intelligent ones should get it corrected. 7. Whatever I have written here incorrectly due to the fault of the mirror and due to confusion of the intellect or mind, all that should be corrected by the noble ones. Generally, the eyes of the writer fail or get confused. 8. If there is anything less or in excess due to the sluggishness of eyes, speed of hand or lack of clarity of the alphabets, it should be corrected by the intelligent ones. 19 Chapter Two RECONSTRUCTING THE SOURCE OF CONTAMINATION He [the editor] should ensure, by all the means provided by his science and his ingenuity, that he discovers therein the constants of the “vertical” tradition. It is at this price, and at this price only, that he will be able to deter- mine the reality, and eventually the amplitude, of a “horizontal” transmission.1 —Jean Irigoin, “Quelques rélexions sur le concept d’archétype” Understanding “Contamination” Mahābhārata critics understand “contamination” not as textual critics do (for whom it refers to the addition of readings from a second source beside the one the scribe copied), but as the interpolation of Brahmanic elements into an original Kṣatriya epic. They thus interpret contamination to mean a kind of pollution (namely, with the Brahmans’ religious ideology). From the preceding discussion it may seem that we ignore contamination, but we are extremely sensitive to it. In fact, as we demonstrate in this chapter, the problems in the critics’ work arise because they are not concerned with legitimate instances of contami- nation—instances where it frustrates attempts to draw up stemmata or makes us hesitant about our judgments—but use the term in a peculiar way. As the Mahā bhā rata critics interpret the term, contamination refers not to the phenomenon that scribes occasionally combined readings from diferent sources, but to a Brahmanic takeover or “redaction” of an original oral epic. The critics think the epic was thus “contaminated” with the Brahmans’ religious ideology. But the critics also posit a second type of contamination, namely, the survival of earlier oral epic elements in folk memory, which then reentered the written, Brahmanic tradition.2 In their opinion, collecting these elements, which appear stemmatically later but are actually historically earlier, lets us restore a partial supra-archetype of the oral epic.3 Alongside this work of recovery, they argue for a dia- chronic layering or stratiication (diachronische Schichtung) of the constituted text using cri- teria such as content, meter4 and “style”5 to identify and eventually eliminate elements perceived as “Brahmanic” from the latter.6 Contrary to what the critics think, however, the circumstance that we can create a version of the Mahā bhā rata corresponding to our prejudices about it does not mean that such a version actually existed. It merely reveals these prejudices as tasteless and uneducated.7 Where it contravenes basic philological principles, the Indologists’ so-called analytic scholarship contributes neither to our understanding of the text nor to our under- standing of the tradition.8 In continuation of our analysis, in this chapter we examine their arguments for identifying an earlier stage of the tradition than what is present in 120 120 PHILOLOGY AND CRITICISM the critical edition, while simultaneously developing objective criteria for reconstructing the source of contamination. Contamination: Hyperarchetypal and Extra-stemmatic This section distinguishes between extra-stemmatic contamination, intra-stemmatic contamination and hyperarchetypal contamination. We examine Bigger’s view that the hypothetical Kṣatriya epic underwent contamination from Brahmanic sources prior to the formation of the archetype (hyperarchetypal contamina- tion) and that the resulting Brahmanic text then itself underwent contamination beneath the archetype with remnants of the Kṣatriya epic that survived either in folk traditions or in the poets’ memory (extra-stemmatic contamination). Sebastiano Timpanaro distinguishes two types of contamination: 1. Extra-stemmatic contamination, deined as “contamination deriving from manuscripts that do not form part of the tradition that has survived more or less completely,” and 2. Intra-stemmatic contamination, that is, “relations of collation between surviving manuscripts which form part of the stemma we can reconstruct.”9 The second type concerns relations between extant sources and hence does not feature in criticisms of the critical edition from a hyperarchetypal perspective. In contrast, extra- stemmatic contamination often features, especially in the form that an older oral epic tradition existed alongside the Mahā bhā rata. Extra-stemmatic contamination, in turn, can be of two types: 1a. Contamination into an ancestor of the archetype, and 1b. Contamination into an extant source from a no longer extant source. We refer to the former as “hyperarchetypal contamination” in the following in contradis- tinction to the latter (“extra-stemmatic contamination” in the strict sense).10 As we have seen, the critics argue that elements of the older oral epic could have survived past its alleged Brahmanic redaction in two ways: 1. Although signiicant, the redaction was not complete. Episodes or narratives from the earlier Kṣatriya stage were thus unintentionally preserved. 2. Alternately, if the Brahmans’ inluence was limited to the major manuscript traditions, the older epic tradition could have survived at the margins of Brahmanic society, and occasionally reintroduced Kṣatriya elements into the epic. The irst view is a form of hyperarchetypal contamination (contamination of the orig- inal epic with a Brahmanic source prior to the formation of the archetype). In contrast, the second view invokes extra-stemmatic contamination to explain the occurrence of 12 RECONSTRUCTING THE SOURCE OF CONTAMINATION 121 allegedly oral materials in the Mahā bhā rata. Andreas  Bigger’s work may serve as an example. Concluding his analysis of Balarā ma in the Mahā bhā rata, Bigger writes, If one wants to summarize the insights […] gained in this work, the following rough scheme ofers itself for the diachronic stratiication [Schichtung] of the results: 1. The oldest stage (of which we cannot deinitively say if it belonged to the Mahā bhā rata all along or from what point onward it was inserted). 2. Additions before the Pā ñcarā tra layer. 3. Additions in or after the Pā ñcarā tra layer. With this stage the normative redaction was reached. 4. Developments after the normative redaction:  a. southern recension. b.  northern recension.11 In contrast to the irst stage, when “not the Brahmans but the Kṣatriyas were the bearers of the transmission—at least as the audience and as patrons,”12 the next three stages relect the rise of the Brahmans as the Mahā bhā rata’s custodians.13 Bigger characterizes the corresponding changes to the text as follows: “In the second stage, […] Baladeva is revealed as a strict follower of Kṣatriya dharma. In a second step, what is really a religious dharma becomes predominant since Baladeva embarks on a pil- grimage (tı̄rthayātrā) as an alternative to the Kurukṣetra battle. Baladeva thereby airms dharma, but Kṣatriya dharma is supplemented by one the Brahmins wish to see of the Kṣatriyas.”14 “In the third stage, the igure of Saṅkarṣaṇa, one of the four vyūhas, who is also called Ś eṣa, is associated with Baladeva. […] As a result of this assimilation, the idea emerges that Ś eṣa [was] incarnated as Baladeva. With this stage, the nor- mative redaction is attained, as reconstructed in the critical edition. In the following period, Baladeva undergoes a more or less independent development in the northern and southern recension.”15 Figure 14 clariies Bigger’s hypothetical four-stage evolu- tion of the text. The original ω generates copies at α, β, γ and δ. These texts undergo continuous interpolation from Brahmanic sources, divided into pre-Pā ñcarā tra, Pā ñcarā tra and post-Pā ñcarā tra sources (P1, P2 and P3 in the stemma). As a result, the archetype θ is no longer a Kṣatriya epic but a Brahmanic work. Elements of the original either could have survived past the text’s redaction (resilient transmission) or been reintroduced into the tradition beneath the archetype (recombinant transmission). Θ generates copies at ε, ζ, η and ι, but they are copies of the new Brahmanic work (the change in the angle of the line relects the change in custodianship). The copies undergo further contamination with either Brahmanic sources (B) or remnants of the Kṣatriya tradition (K). Whereas the con- tamination of the Kṣatriya epic with pre-Pā ñcarā tra, Pā ñcarā tra and post-Pā ñcarā tra sources occurs above the archetype and hence represents hyperarchetypal contamina- tion, the contamination of the Brahmanic text is a case of extra-stemmatic contamina- tion (contamination from no longer extant sources). Using the stemmatic method, we would reconstruct the archetype θ, but this is deicient. The stemmatic method wrongly eliminates older Kṣatriya passages that reentered the tradition beneath the archetype. Furthermore, it overlooks that the archetype θ is itself a composite work, the result 12 122 PHILOLOGY AND CRITICISM ω K atriya α P1 Brahmanic contamination of β the K atriya epic P2 (hyperarchetypal γ contamination) P3 K atriya and Brahmanic δ Extra-stemmatic Change in θ Normative redaction contamination transmission ε B Contamination ζ into the K Brahmanic text η Brahmanic B ι Witnesses Figure 14 Bigger’s “prehistory of the normative redaction” of multiple redactions, containing a mixture of pre-Pā ñcarā tra, Pā ñcarā tra and post- Pā ñcarā tra interpolations into an original Kṣatriya source. Identifying the Source of Contamination This section contrasts the Mahābhārata critics’ arbitrary and tendentious arguments for Brahmanic “contami- nation” with objective criteria for identifying the source of contamination. We show that the minimum condition to identify interpolations is a stemma of the form 3+1, where three manuscripts descend from the contaminated source and a fourth descends from an independent source. If any two manuscripts descended from the contami- nated source agree against the manuscript descended from an independent source, then their reading is the reading of the source of contamination. The Mahā bhā rata critics’ arguments for rejecting or including speciic passages in the constituted text are not stemmatic.16 Typically, their sole criterion for considering cer- tain passages “interpolations” is their a priori assumptions about the text’s history.17 By way of contrast, let us examine the conditions that must be met to reconstruct the source of contamination objectively. We take as our example Lachmann’s 1817 review of 123 RECONSTRUCTING THE SOURCE OF CONTAMINATION 123 Friedrich Heinrich von der Hagen’s edition of the Nibelungenlied. Timpanaro summarizes Lachmann’s review as follows:18 In July 1817 Lachmann published a long review of Friedrich Heinrich von der Hagen’s edition of Der Nibelungen Lied (Hagen 1816) and Georg Friedrich Benecke’s edition of Bonerius’ Der Edel Stein (Benecke 1816). Lachmann distinguished two redactions in the manuscript tradition of the Nibelungen Lied: a shorter and more genuine one contained in the manuscript he called B; and another, longer and heavily interpolated one, represented by the manuscripts GEM. According to Lachmann, both redactions have reached us disigured by corruptions and sec- ondary interpolations; but while the irst one cannot be reconstructed in its original form until another manuscript, a brother of B, is discovered, the second one can be reconstructed by comparing GEM. According to Lachmann, such a comparison reveals that the ancestor of GEM was still fairly free of interpolations in the text written by the irst hand but that a second hand inlicted many changes and arbitrary additions upon the original text. Each copyist of GEM reproduced now the readings of the irst hand, now that of the second hand, and also inter- polated on his own.19 Lachmann does not draw up a stemma, but Timpanaro ofers the one shown in Figure 15. Ω indicates the oldest redaction, φ the irst hand of the ancestor of GEM, and φ2 the interpolations by the second hand in the same manuscript.20 Lachmann adds that the editor’s task is to identify the interpolations in the ancestor of GEM, in other words, to reconstruct the readings of φ2.21 He provides the following rules, in which the signs > and < signify “better than” and “worse than” just as in mathematics they signify “greater than” and “less than”:22 1. Three manuscripts out of four outvote one every time. 2. When any two agree, BG < EM (that is, where B agrees with G, the unanimous reading of E and M is preferable), GE > BM, GM > BE. 3. Where three readings exist, BG < E—M (against the reading shared by B and G, the two others in E and M are preferable), G—E > BM, G—M > BE; on the other hand, EM = B—G (the agreement of E and M leads to no secure decision against the two readings of B and G), BM = G—E, BE = G—M.23 4. When all four disagree, the original reading is just as uncertain.24 ω B φ ( φ 2) G E M Figure 15 Reconstructing the source of contamination Source: Reproduced from Timpanaro, The Genesis of Lachmann’s Method, 140. 124 124 PHILOLOGY AND CRITICISM Of the four rules, the irst is obviously false: the agreement of BGE against M or of BGM against E or BEM against G gives us the reading of φ and not φ2. On the other hand, if the three manuscripts that agree are GEM (against B), it is not necessary that they repro- duce the reading of φ2. B could have innovated with respect to ω so that the original reading of ω (and φ) is reproduced in GEM or φ could have innovated with respect to ω so that GEM difer from B but do not contain the reading of φ2. The irst rule, then, is useless for reconstructing the reading of φ2. The second rule, in contrast, is correct. When two readings exist, each one attested by two manuscripts, the reading of B and one apograph of φ was already present in φ and ω, while the other two apographs of φ reproduce the variant of the second hand (φ2).25 The emended version of the third rule is also correct, but does not help us identify the reading of φ2. If three readings exist, the agreement of one of the three apographs of φ with B clearly yields the reading φ (and hence of ω). On the other hand, the readings of the two remaining apographs are equally likely to reproduce the reading of φ2 or they could both be innovations with respect to φ2. In either case, we cannot reconstruct. Further, if two apographs of φ agree, while the third and B reproduce isolated readings, we have no grounds for preferring the shared reading of the two apographs: either the shared reading of the two apographs or the isolated reading of the third apograph can repro- duce the reading of φ2. Finally, the fourth rule is correct and self-evident. Since the irst rule is false and the fourth rule, though correct, is tautological, they can be eliminated. The second and third are more interesting for determining the conditions necessary to identify the source of contamination. As Timpanaro, citing Fritz Bornmann, notes, the two rules are formally parallel. “Both rules mean in substance that a reading attested by one or two apographs of φ can represent φ2 as long as such a reading is not found in B too.”26 This minimum condition does not get us very far. It merely clariies that if a reading is attested in a manuscript not descended from the contaminated manuscript, its reading cannot be the reading of the source of contamination (φ2), but must be the reading of the irst source of the entire tradition (ω) and have descended thence via φ to our witnesses. The third rule builds on this condition. It clariies that if two apographs con- tain one reading and the third another and neither reading is attested in a manuscript not descended from the contaminated manuscript, then either reading could be the reading of the source of contamination φ2. If, on the other hand, the manuscript not descended from the contaminated manuscript and one of the apographs share a reading, while the other two contain isolated readings, we can exclude the shared reading, but we have no grounds for preferring one of the isolated readings. Only the second rule allows us to unambigu- ously identify the reading of the source of contamination: if two readings exist, one of which is attested in a manuscript not descended from the contaminated manuscript and in a descendant of the contaminated manuscript, then the other is the reading of φ2. From the foregoing discussion, we see that B, the manuscript not descended from the source of contamination, has a crucial role in identifying the source of contamina- tion, as it represents an independent control. If we did not have at least one manuscript not descended from the source of contamination, it would not matter whether all three apographs GEM agreed or only two or none. In none of these three cases could we exclude certain readings. Hence, the most basic rule refers to B. The presence of at least 125 RECONSTRUCTING THE SOURCE OF CONTAMINATION 125 one manuscript not descended from the source of contamination is a necessary but insuf- icient condition for identifying the source of contamination. Two further conditions must be met for us to identify the source of contamination. First, the number of possible readings cannot exceed two. As the third rule shows, if three readings exist, the reading of the source of contamination cannot be unambigu- ously identiied. Second, the minimum number of apographs of the contaminated man- uscript is three, for if only two apographs exist and one of them shares a reading with the manuscript not descended from the source of contamination (B), we could not be certain the reading of the other was not an innovation in copying from φ. On the other hand, if only two apographs exist and both share the same reading but difer from the reading of B, we could not be sure that the shared reading was not an innovation in copying from ω or, alternatively, that B did not innovate in copying from ω. That is to say, only if we can be sure that the reading of ω was transmitted correctly to φ (which practically will mean inding it in one of its descendants) and we have a second reading attested in at least two apographs (if it occurs only in one, it could be an independent innovation), can we be certain of identifying the reading of the source of contamination. Hence, we arrived at the second rule: only when two apographs of the contaminated manuscript agree against a third and when this third agrees with an independent witness is it certain that the reading of the two apographs is the reading of the source of contamination. The problem can be expressed more simply using the concept of a bipartite stemma. Since a bipartite stemma does not let us mechanically exclude one of the variants (for either of the readings might be correct), we require a tripartite stemma.27 But since a tri- partite stemma alone will not let us identify the reading of the source of contamination (if all three apographs agree, this could be the reading of the original, or, if two agree against one, this could be the reading of the original and the isolated reading an inde- pendent innovation in that apograph), we require an independent control: a manuscript not descended from the contaminated manuscript. The circumstance that these two manuscripts constitute a bipartite stemma is irrelevant, as we are not interested in iden- tifying the reading of their archetype. On the other hand, if φ were the archetype and variants were entered into it, we could not identify the source of contamination. Even if the descendants of φ reproduced an eclectic mix of readings, now following the irst hand, now the second, we could not identify the source of contamination. Although the reading of two against one would always give us the reading of the archetype, we could not know whether this was the reading of φ or of φ2. In fact, it is doubtful we could draw up a stemma at all.28 On the other hand, if one scribe in copying from it reproduced all its variants, but the other two scribes ignored them all, we could not identify the source of contamination. We would consider the two copies that share errors descendants of a common ancestor and accordingly redraw our lines to yield a bipartite stemma. Finally, it may seem as though to identify contamination all we need is four manuscripts descended from a common ancestor, of which each pair reproduces a reading, but this is incorrect. While it is certain one of the two readings is φ’s and the other φ2’s, we could not know which was which. Hence, a quadripartite stemma would not enable us to determine the reading of the source of contamination. The condition is rather of the form 3+1, where the number 3 indicates a tripartite stemma and the number 1 an 126 126 PHILOLOGY AND CRITICISM independent manuscript linked to the tripartite stemma via a common ancestor (this may or may not be the irst source of the entire tradition). Timpanaro’s stemma in fact represents the minimum conditions to identify the source of contamination. As we shall see, these conditions are seldom if ever met in Mahā bhā rata studies. The Argument from Uncertainty Mahābhārata critics often try to undermine the stemma by positing contamination from a nonextant oral source (extra-stemmatic contamination). They claim that as our stemma does not accurately represent historical reality, we cannot deinitively exclude earlier, nonextant sources. Refutation of the argument:  The argument from uncertainty fails because introducing uncertainty into a system afects all outcomes equally. Introducing uncertainty into the stemma does not make it likelier that a Kṣatriya epic existed. It appears to favor the Kṣatriya epic only because the critics assume its existence in advance and that they can know its contents without a stemmatic reconstruction. In logic, the argument from uncertainty refers to the fallacy of rejecting well-grounded inductive truths because they are merely inductive. For instance, someone could argue that as only inductive proof for event a exists, he is justiied in maintaining ~a, even though the probability for a, written as p(a), approaches near certainty: p(a) ≈ 1. The critic thus invokes uncertainty to make space for a speciic outcome, even though the empirical evidence is against its occurrence. Likewise, the analytic critics often invoke uncertainty to defend their view that the Mahā bhā rata developed from an oral epic. Typically, they do so by positing extra- stemmatic contamination with an oral source, arguing that as we cannot deinitively exclude an oral precursor of the Mahā bhā rata, they are justiied in considering certain elements in the text oral in origin. They thus dispute the scientiically and inductively more probable conclusion for the less probable one. The problem is: introducing extra- stemmatic contamination into the stemma, whether at the level of one of the witnesses or at the level of a subarchetype, does not make the existence of a Kṣatriya epic like- lier. On the contrary, it will almost always lead to misleading results. Consider Bigger’s argument that the interpolations found in γ and S are original, whereas the passages were lost in ν. Although we rejected this argument in the previous chapter, Bigger could claim that γ and S contain the correct reading if he invoked extra-stemmatic contamination. He could argue that the scribe of D, besides copying from γ, kept an eye on θʹ and borrowed some readings from it. If θʹ was then lost,29 the true stemma would appear as in Figure 16. Θʹ represents the scribes’ “memory” (Gedächtnis), whereas D is a mixture of readings “transcribed” (niedergeschrieben) from memory and others inherited from γ.30 Bigger could thus resurrect his claim that the critical edition only reconstructs a certain stage of the transmission, but it does not contain the oldest readings. Θʹ could represent an older copy than θ or a purer copy of the original ω (θ could be a contaminated apograph of the orig- inal combining elements of the Kṣatriya epic with elements from a Brahmanic source— ωʹ in the stemma; the dashed line indicates contamination of θ from ωʹ). D’s readings are therefore preferable given its access to θʹ. It follows the purer line ω—θʹ as compared 127 RECONSTRUCTING THE SOURCE OF CONTAMINATION 127 ω′ (Brahmanic source) ω (Ks·atriya epic) θ θ′ N S ν γ σ ε Śārada K Nepālı̄ Maithilı̄ Bengali Devanāgarı¯ Telugu Grantha Malayālam Ś1 K0–6 Ñ1–3 V1 B1–6 Da1.2, Dn1–3, T1.2 G1–6 M1–5 D1–7 Figure 16 Extra-stemmatic contamination into an extant witness with Ś, which follows ν—N. Passages rejected from the constituted text because they were found only in some manuscripts could thus be older than the archetype. Bigger could argue that focusing on these passages lets him reconstruct the original (at least in part, since not all θʹ’s passages are transmitted to D). The “methods of higher criticism” prove superior, as they correct for the stemmatic method’s overly rigid procedure. We would observe that D occasionally avoids errors common to the rest (errors that occurred in θ and were transmitted to the rest via NS), and that γ sometimes agrees with N and S, sometimes with D.31 We would construct the stemma shown in Figure 17. We would discard γ as a contaminated manuscript ofering nothing not found in the other sources, and we would consider D’s peculiar readings as likely as those of θ to be the archetype’s.32 Insofar as they are drawn from θʹ, this would be correct, but insofar as they were errors made by γ or in copying from γ, it would be false.33 Discarding γ, we would reconstruct the text of N and S and also θ (our former archetype). Sukthankar’s stemma only takes us this far, since he considered D’s peculiar readings additions or corruptions in it, and hence consigned them to the apparatus. In contrast, following Bigger’s hypothesis, we would consider D a text on par with the reconstructed text of θ. As this stemma is bipartite, either set of readings could be the original. But since D is an extant witness, whereas θ is a conjectured source, we would consider D’s readings more authoritative.34 We would reconstruct a Kṣatriya epic following D’s readings.35 128 128 PHILOLOGY AND CRITICISM ω′ (Brahmanic source) ω (Ks·atriya epic) θ D N S ν γ σ ε Śārada K Nepālı̄ Maithilı̄ Bengali Telugu Grantha Malayālam Ś1 K0–6 Ñ1–3 V1 B1–6 T1.2 G1–6 M1–5 Figure 17 The consequences of assuming extra-stemmatic contamination ω′ ω θ θ′ N S Figure 18 Extra-stemmatic contamination into a direct descendant of the archetype: S as an example The outcome is the same if we assume contamination into one of the archetype’s direct descendants (Figure 18). Once again, we would observe that S occasionally avoids errors common to the rest (errors found in θ and its descendants), and that θ and S sometimes agree and sometimes do not. We would construct the stemma seen in Figure 19. Since we have no reason to posit the existence of an intermediate source between N and S, we would eliminate θ and adjust our dashed lines to obtain the stemma seen in Figure 20.36 We would conclude that either S had undergone contamination with θʹ or θ or N had undergone contamination with ωʹ, but we could not decide between the two. In both 129 RECONSTRUCTING THE SOURCE OF CONTAMINATION 129 ω′ S θ N Figure 19 The consequences of assuming extra-stemmatic contamination into S ω′ S N Figure 20 S as the original oral epic cases, however, the argument leads to a reductio ad absurdum. In the irst case (Figure 18), we would think that elements in S not shared with N were inherited from a descen- dant of the Kṣatriya epic, even though S’s distinctive elements are characteristically Brahmanic. In the second case (Figures  19 and 20), we would consider S itself the Kṣatriya source. Even if we assumed that the former stemma represented the true state of afairs, we could reconstruct neither θʹ (S includes only some of its readings) nor ω. In the second example, we would reconstruct S from its witnesses T, G and M, and think that the passages unique to θ (or N) derived from another source (ωʹ), but we could not reconstruct this text either (N includes only some of ωʹ’s readings). Either way, we could not reconstruct an older text than the archetype θ, whether the Kṣatriya epic ω or the Brahmanic source ωʹ. The southern recension is a later, more inlated tradition than the northern.37 It is also considered more Brahmanic.38 The example thus illustrates the absurd consequences of claiming contamination from a nonextant oral source into S.  But what if we con- sidered the pair θʹ and N instead of θʹ and S? Since the northern recension is alleg- edly closer to the historical core devoid of Brahmanic elements,39 Bigger could argue that N represents a contaminated source containing a mixture of older oral elements and elements inherited from the normative redaction (our archetype θ). Although itself “contaminated,” N represents a purer source than S, which can be disregarded for reconstructing the archetype θ. We can restrict ourselves to examining the elements in N inherited from θʹ, the copy of the Kṣatriya epic ω. To understand the problems with this second hypothesis, let us again construct a pair of stemmata. Suppose that the scribe of N, besides copying θ, kept an eye on θʹ and borrowed some readings from it. If θʹ was then lost, the true stemma would appear as seen in Figure 21. N contains a mixture of readings from θ and θʹ, where θ is a contaminated copy of the Ksạ triya epic ω (contamination with a Brahmanic text ωʹ) and θʹ is either an oral source or elements of the Ksạ triya epic preserved via the poets’ memories. Stemmatic recon- struction would retain the readings attested in S and N. In cases of crux, that is, when S 130 130 PHILOLOGY AND CRITICISM ω ω′ θ′ θ N S Figure 21 Extra-stemmatic contamination into a direct descendant of the archetype: N as an example N ω′ θ S Figure 22 The consequences of assuming extra-stemmatic contamination into N N ω′ S Figure 23 N as the original oral epic and N difer and a third reading explaining the other two is not available, the editor would resort to other criteria. In contrast, Bigger would argue that N consistently contains better readings (those inherited from θʹ). We would observe that N sometimes avoids errors common to the rest (errors found in θ and its descendants), and that θ and N sometimes agree and sometimes do not. We would construct the stemma seen in Figure 22. Once again, we would eliminate θ and adjust our dashed lines to yield the stemma seen in Figure 23. We would discard S as a contaminated manuscript ofering a mix of readings from N and ωʹ, and we would consider N’s peculiar readings the archetype’s (or, rather, the hypothetical Kṣatriya epic’s). Insofar as they are drawn from θʹ, this would be correct, but insofar as they were errors made by N, or in copying from θ, it would be false.40 Presuming we could exclude the alternative, does this stemma let us reconstruct an older text than the archetype θ? Four arguments speak against it: 1. This stemma discards S’s evidence, even though Sukthankar found it to be an inde- pendent tradition rather than a copy of N. 13 RECONSTRUCTING THE SOURCE OF CONTAMINATION 131 2. S’s evidence is essential to reconstruct N. Sukthankar considered S’s independent agreement with ν against γ the strongest argument for the reading of the archetype when ν and γ difered.41 3. The tradition is reduced to two branches, ν and γ. Since we no longer have a mechanical way of choosing between variants, and Sukthankar thought “the Śā radā (Kaśmīrī) version is certainly the best Northern version, and probably, taken as a whole, the best extant version of the Ā di,”42 we would often prefer the readings of Ś or Ś and K over γ. N will thus often be identical with Ś. Not only will it not difer signiicantly from Ś, it also cannot claim great antiquity. 4. Considered as the original epic, N is later than the so-called normative redac- tion θ (allegedly the result of contamination of the Kṣatriya epic with Brahmanic elements).43 Even so, this stemma does not let us reconstruct the original epic. This stemma is based on the observation that N sometimes agrees with S and sometimes difers from it. We do not know if this is because N was contaminated with the “oral” source θʹ, a copy of the Kṣatriya epic ω, or because either θ or S was contaminated with a Brahmanic source—ωʹ. For Bigger, however, the distinction is irrelevant. From his perspective, the reading of the northern recension is preferable every time N and S disagree, since it represents the purer transmission. But Sukthankar followed this very principle in reconstructing the archetype! He notes, “When the two recensions have alternate readings neither of which can have come from the other and which have equal intrinsic merit (N:  S), I  have, for the sake of consistency and with a view to avoiding unnecessary and indiscriminate fusion of versions, adopted, as a stopgap, the reading of N.”44 Adopting N’s readings consistently thus will not lead to an older stage than the constituted text, which prints N’s unique readings whenever N and S disagree. It merely robs us of the ability to check N’s readings (or its witnesses’ readings) against S. Even if we assume that this stemma is more accurate than Sukthankar’s, it does not let us recon- struct an earlier stage of transmission than what is contained in the critical edition, since this requires that we identify not N’s unique readings, but its super-unique readings— readings not just unique to N but unique even among its readings. Contamination of either kind, whether extra-stemmatic or intra-stemmatic, is a threat to any stemma. As Martin L.  West notes, “If contamination is present in more than a slight degree, it will be found that no stemmatic hypothesis is satisfactory.”45 Bigger, however, inten- tionally introduces uncertainty into the stemma to make space for a Kṣatriya epic. The attempt fails because introducing uncertainty into a system makes all outcomes equally unlikely.46 Once we introduce uncertainty, it is not restricted to one exemplar, but travels through the stemma to all the exemplars linked with the contaminated source. Introducing uncertainty into the stemma may undermine the critical edition, but it will not yield a Kṣatriya epic. The Argument from Oral Source If the oral tradition survived past the Brahmans’ seizure and destruction of the original epic, elements from it could have trickled back into the manuscript tradition. This could have occurred if a scribe either recollected an 132 132 PHILOLOGY AND CRITICISM oral version or possessed a transcript of it. In that case, we are justiied in thinking that some of the readings in our manuscripts (and mutatis mutandis the constituted text) are older than others, and some are at least as old as the oral epic. Refutation of the argument: the argument does not address the question of how we can identify the alleged remnants of the oral epic. Every manuscript contains readings of varying antiquity but, without the stemmatic method, we cannot sift between them. The argument also does not address the problem that the alleged remnants could be later innovations or inaccurate recollections. Indeed, in a long chain of transmitters such as an oral tradition presumes, it is especially likely that the readings underwent deterioration. The analytic critics fail to perceive the problem because they focus on passages rather than readings, and do not ask how, even if we grant that the passages were original, their readings could have been transmitted unchanged. In the previous section, we have seen that positing contamination from a nonextant oral source did not permit us to reconstruct an earlier stage of the tradition because of contamination’s potential to mislead us in our inferences. But what if the oral tradition survived until a very recent date? Would that make a diference? Would it permit us to identify elements of the hypothetical Kṣatriya epic consistently? It appears as though all we need do to overcome the objections of the previous section is to localize contamination to a late stage of the tradition. If contamination can be localized to an extant witness, for instance, and we can show that only speciic readings derive from the hypothetical oral source (the source of contamination), we need no longer fear our stemma shifting due to contamination. This is also Bigger’s solution. He writes: Most likely, there existed a ixed oral transmission before the normative redaction was com- piled, even though it may be that it did not cover a single text, but only parts of it. This oral tradition did not disappear at the same moment as the irst mss. of the normative redaction came into being. I would expect it to continue for many centuries parallel to the normative redaction. One may also assume that some of the passages I was able to identify in my thesis derive from another strand of the oral tradition and were introduced into the written tra- dition at a later time. The minor variants noted by Sukthankar, on the other hand, can be explained as having been produced by a scribe, who, though copying a ms., also knew the MBh by heart and therefore tended to mix what he read with what he remembered.47 Consider the stemma in Figure 24. The oral epic generates “copies” at α, β and δ. It is transcribed for the irst time at N, which generates further copies at ν and γ (unlike α, β and δ, ν and γ are physical copies of a written source). Finally, the revised Brahmanic text also reaches southern India, where it gives rise to S and its descendants. Alongside the two written traditions of the Mahā bhā rata, the northern and the southern, we thus also have an oral tradition that survived “for many centuries parallel to the normative redaction.” Unlike the former, however, it does not appear on our stemma because of its intangible quality. But we would be wrong to discount it for this reason. Its efects are felt in the visible portion of our stemma because some manuscripts occasionally contain “readings” that they can owe only to the Kṣatriya epic. Contamination occurs not only into the archetype of the written tradition (hyperarchetypal contamination) and between the manuscripts on our stemma (intra-stemmatic contamination) but also from sources not entered into our stemma (extra-stemmatic contamination). 13 RECONSTRUCTING THE SOURCE OF CONTAMINATION 133 ω (K atriya epic) α N β δ ω′ (Brahmanic source) S ν γ σ ε rada K Nep l Maithil Bengali Devan gar Telugu Grantha Malay lam 1 V1 T1.2 Figure 24 The original oral epic as the source of N But unlike our previous examples, these sources are not not listed on our stemma because they disappeared long ago. Rather, they are contemporary with the surviving witnesses. They are not listed because, other than their efects, which are real and can be detected in the surviving witnesses, they left no record. Their efects, however, evince the existence of a “parallel transmission” or “parallel tradition” alongside our written tradition.48 Initially, the argument appears identical to the previous one. However, Bigger could specify that he does not mean contamination in the sense that a scribe combined readings from two sources into a lost ancestor of the extant witnesses. Rather, what he envisions is that the scribe of D4 followed the text of γ (or a source between γ and D4) con- sistently, but, in one place, following chapter  1.211, he remembered an older version of the narrative and wrote this down instead. If, as Bigger says, “diferent copyists inserted passages from other versions—partly from direct recollection, partly from other (younger?) written versions—into the transcript,”49 albeit in recent memory, the stemma need not shift. The insertion of 114 and 115 into D4 is a unique example of an oral version transcribed from memory. It does not afect the rest of the tradition but, in this one case, we are justiied in preferring D4’s lectiones singulares (singular readings). The stemma in Figure 25 clariies the situation. The oral epic generates “copies” at θ and θʹ. Θ, however, already represents a “redac- tion” of the text. This oicial Brahmanic redaction includes only a small measure of ω’s readings (namely, those the Brahmans either overlooked or found indispensable), but because it gave rise to copies at N and S and they, in turn, gave rise to descendants that survived, it became the archetype. Actually, however, it is not the archetype but a “nor- mative redaction” and the true archetype (in fact, the original) is not θ but ω. We cannot 134 134 PHILOLOGY AND CRITICISM ω′ (Brahmanic source) ω (Ks·atriya epic) θ θ′ N S ν γ σ ε Śārada K Nepālı̄ Maithilı̄ Bengali Devanāgarı̄ Telugu Grantha Malayālam Ś1 K0–6 Ñ1–3 V1 B1–6 Da1.2, Dn1–3, T1.2 G1–6 M1–5 D1–7 Figure 25 Contamination via an oral source reconstruct ω because most of its readings were lost in creating θ, but we do have a valuable witness for its (genuine) readings in D because of its access to θʹ. The problem with this thesis is twofold. First, we have no way of identifying the readings D owes to θʹ rather than θ. Doubtless, the critics will propose subjective cri- teria such as style or perceived conformity with the ideal of a heroic epic. But even if we grant this, a second problem remains. Recall that the readings D inherits from θʹ are preferable to those it inherits from γ because they allegedly follow the more direct trans- mission. In contrast to D’s other readings, which descended through multiple individual acts of copying (not to mention the deliberate defacement they sufered at the Brahmans’ hands), the appendix passages 114 and 115 derive from a more or less direct ancestor of ω. But what is the scribe’s source? On the stemma θʹ appears as an ancient source available until very recently to the scribe of D. But actually, if the scribe’s “source” is his recollec- tion, θʹ is recent rather than old. A recent source is, of course, not necessarily bad. In his Storia della tradizione e critica del testo, Giorgio Pasquali showed that, contrary to the prejudice against the recentiores (the more recent manuscripts), later exemplars sometimes preserved better readings than the oldest surviving manuscripts because of their editors’ access to good old manuscripts (which did not survive).50 He coined the phrase recentiores non deteriores (recent, not necessarily worse) to express the circumstance that good readings are not 135 RECONSTRUCTING THE SOURCE OF CONTAMINATION 135 restricted to the meliores (the manuscripts judged the best) but can occur in any part of the tradition. In West’s words, “A propensity to emendation, so far from discrediting a manu- script may be symptomatic of an interest in the text that also prompted the consultation of out-of-the-way copies, like the use of λ by the late manuscript A in the stemma on page 44. Conversely very old copies such as papyri sometimes disappoint expectations by giving a worse text than the medieval tradition instead of a better one.”51 West presents the stemma shown in Figure 26. Λ, though a late copy, contains old readings because of its scribe’s access to γ. A is contaminated with λ, but, far from being a disadvantage, this actually means that it contains better readings than it otherwise would (readings it inherits from B or M). We would therefore be wrong to reject A’s individual good readings, even though they lack support from the other manuscripts. Initially, Bigger’s argument appears similar. Like A’s individual good readings, the appendix passages 114 and 115 are preferable because of the scribe’s access to θʹ. They should not have been rejected from the constituted text, even though the attestation for them was inconsistent. Contrary to Bigger’s claim, however, θʹ’s status is far from clear. Although placed in the closest proximity to ω in our stemma, its position is actu- ally uncertain. If the scribe inserted 114 and 115 from memory, θʹ is not ancient but contemporary with D. It should occur signiicantly lower in our stemma. A long chain α β γ δ ζ ε η θ M D g b O B N G H K C F λ E I L A Figure 26 Recentiores non deteriores Source: Reproduced from West, Textual Criticism and Editorial Technique, 44. 136 136 PHILOLOGY AND CRITICISM of intermediate sources separates it from ω. Even if it ultimately derives from ω, what is the proof that it was transmitted accurately? Moreover, lacking a physical copy, the scribe is at the mercy of his memory. He cannot know if what he wrote corresponded to what he heard. Even if he thought he transcribed it faithfully, he cannot know this. Any additions he made should therefore be treated as his own idiosyncratic changes. Alternatively, θʹ could represent an ancient source, a “copy” made from ω soon after its text was established. This is indeed Bigger’s solution. He writes: “These additions […] were either composed by the scribe himself or transcribed from memory (that is, an oral text became a written text) or copied from another source.”52 If θʹ is an oral source, how- ever, how could a scribe from the eighteenth century (the approximate date of D4, the manuscript containing the appendix passages 114 and 115) have access to it? The only way he could know it is if he heard it from a narrator, who heard it from another, who in turn heard it from another and so on. That means, however, that his source, although recent (in fact, as recent as his hearing of it) is not ancient. It should be treated no difer- ently from anything else he might have been told because neither he nor his source can know that what he heard corresponds to the original reading.53 In West’s example, the principle recentiores non deteriores applies because λ is a man- uscript. As such, it can be both recent and archaic. As a material object, it is recent, whereas, as a text, it is archaic. λ could have been copied from γ shortly after the latter itself was made and survived for a long time so that A’s scribe had access to it or γ itself survived for a long time and λ was made from it closer in time to A. In either case, the rel- evant manuscript’s nature as a physical object lets it bridge the expanse of time between its source and the copy made from it. Θʹ, however, is not a manuscript but a “copy” of the original oral epic. As such, it has descended through ininitely many acts of narration to D’s scribe (or his informant). Which of these versions should we designate as θʹ? Even if we decide that only the irst act of oral transmission in this extended series deserves the name θʹ, we cannot circumvent the problem. The scribe’s source will now no longer be θʹ but a remote descendant of it called θʹʹ, and we will still face the problem of how to explain the unclariied (manifold) transmission between θʹ and θʹʹ (Figure 27). This stemma illustrates the vast gray space that exists between the scribe’s recollec- tion or copying of the passages at θʹʹ and their origin θʹ in the tradition. This space is only imperfectly illed in by ambiguating on the meaning of “source.” It does not suice to note that the scribe’s source is his recollection or a “parallel” oral tradition existing behind our stemma because the real question is: can we reconstruct the chain of trans- mission all the way to the irst source of the surviving tradition? If not, then the assertion that these hypothetical sources contain good old readings (readings they owe to ω) is per- fectly arbitrary.54 The readings could have undergone a thousand changes during trans- mission. We could also just assert that they are ancient because we say they are ancient.55 One way out of this impasse is for Bigger to posit an intermediate written source. He could argue that if the oral tradition underwent transcription at a suiciently early date— say, around the time the canon was sealed and a “ixed oral text” emerged56—a written text of the original epic could have survived alongside the Brahmanic Mahā bhā rata. Rather than recall a passage from an oral source, D’s scribe could have kept an eye on this transcript and entered variants from it into his copy of γ. Alternatively, Bigger could argue 137 RECONSTRUCTING THE SOURCE OF CONTAMINATION 137 ω′ (Brahmanic source) ω (Ks·atriya epic) θ θ′ Unclarified (manifold) transmission N S ν γ σ θ′′ ε Śārada K Nepālı̄ Maithilı̄ Bengali Devanāgarı̄ Telugu Grantha Malayālam Ś1 K0–6 Ñ1–3 V1 B1–6 Da1.2,Dn1–3, T1.2 G1–6 M1–5 D1–7 Figure 27 Recentiores deteriores that if a scribe wrote down the oral epic as he heard it and this text was then transmitted and was lost only recently, the scribe’s recollection would be recent. We would have a recent source θʹ with access to archaic readings via the intermediate manuscript δ, which explains D’s good readings. We can represent the two alternatives as seen in Figure 28. Both arguments insert a written intermediary between the oral and manuscript traditions to overcome the problem that an oral source can either be ancient (in which case it is likely corrupt since it has a long transmission) or it can be recent (in which case it need not be archaic and should therefore be treated like any other insertion). The line on the right represents the transcription of the oral epic and the subsequent contami- nation of D with δ, the immediate source of the appendix passages. The line on the left represents transcription of the oral epic and the subsequent contamination of D with an oral source θʹʹ. Δ is only the indirect source of the passages. Whereas the dashed line on the right represents the break between θʹ and its irst transcription at δ, the dashed line on the left represents the break between the loss of δ and the recording of its variants in D at θʹʹ. Inserting a written intermediary, however, only apparently solves the problem. Consider Bigger’s irst suggestion. If D’s scribe had access to δ, the transcript of the oral epic, why would he consistently follow the later and worse source γ and insert only two passages from δ in his text? Would he not, rather, have copied the latter? Moreover, if δ survived until recently, why is its inluence not felt in other manuscripts? Indeed, if the normative redaction allegedly suppressed all other versions of the epic, how could this 138 138 PHILOLOGY AND CRITICISM ω′ (Brahmanic source) ω (Ks·atriya epic) Unclarified θ θ′ (manifold) transmission δ N S Unclarified (manifold) transmission ν γ δ σ ε θ′′ Śārada K Nepālı̄ Maithilı̄ Bengali Devanāgarı̄ Telugu Grantha Malayālam Ś1 K0–6 Ñ1–3 V1 B1–6 Da1.2, Dn1–3, T1.2 G1–6 M1–5 D1–7 Figure 28 Transmission via an oral source and the inevitability of a written intermediary one exemplar or one of its descendants survive in the eighteenth century? And if a copy survived in the eighteenth century, why did it disappear just then? The alternative that 114 and 115 are remnants of an oral tradition that underwent tran- scription and, after this source was lost, returned to being oral elements thus appears prefer- able. In this case, the passages could be both recent and archaic and the objection recentiores deteriores would no longer hold. This is Bigger’s precise claim. He replaces δ with a “ixed oral text,” but in all other respects the argument is identical. First, a “loating oral tradition” was codiied and established as a text. Then it was transmitted via a “ixed oral transmission.” The “oral tradition did not disappear at the same moment as the irst mss. of the norma- tive redaction came into being.” Rather, it likely “continue[d] for many centuries parallel to the written tradition.” But once this oral transmission ceased, the “ixed oral text” (δ in our stemma) returned to an abstract tradition in that “a scribe, though copying a ms., also knew the MBh by heart and therefore tended to mix what he read with what he remembered.”57 (The dashed line between δ and θʹʹ in the diagram indicates this break.) Finally, its elements were incorporated into D at θʹʹ, explaining the former’s singular readings, which evince the existence of an alternative, Ksạ triya version of the epic. Bigger does not clarify how many centuries later the oral transmission died out, but from the fact that the version the scribe of D heard cannot be far removed from δ (otherwise, all the problems we encountered in the previous stage of the argument will resurface), we can assume that it existed until at least the sixteenth century. In that case, it is plausible that the scribe of D had access to a version of the Mahā bhā rata removed, at most, one or two generations from ω (albeit with a break, since his source was not δ itself but his or his informant’s recollection of it). If some of D’s variants such as 114 and 115 derive from this copy, they would be both recent and archaic. 139 RECONSTRUCTING THE SOURCE OF CONTAMINATION 139 But far from solving the problem, this worsens it. For starters, if the scribe “knew the Mahā bhā rata by heart,” why did he not simply produce a copy of this version? Why would a scribe who knew an older version of the epic copy a later and more corrupt source? Is it credible that, rather than correct this source’s corruptions, he merely chose to add two passages from memory? Further, how can we know that the manuscript the scribe copied, γ, was not better than the lost exemplar δ given that he clearly preferred to copy γ? Indeed, if he knew it by heart, δ was not lost at all! In other words, Bigger is postulating a text in reality nonextant to explain why a text allegedly lost appears extant; and this is meant as a way of reducing the gap between θʹʹ and θʹ! Either way, we cannot be certain that δ represents the original epic. Not only do we have no evidence that δ’s scribe (or, rather, codiier, since in Bigger’s scheme it represents an “oral text”) did not innovate with respect to his oral source; we also have no evidence that the readings D allegedly owes to θʹ actually derive from δ. Bigger asserts that they do, but that is because he does not ask himself how D’s scribe could have access to δ, how θʹ’s readings could have been transmitted unchanged to θʹʹ and what reason we have for thinking they were transmitted unchanged. In fact, he does not consider “readings” at all in the sense that editors do. His preferred “readings” are not variants but passages he considers characteristic of a Kṣatriya epic. Even if they had undergone a myriad of changes during transmission, he would consistently prefer them over the archetype’s readings for no other reason than that they appear to him “not Brahmanic.” The sole criterion he applies is lectio heroicior praeferenda est (the more heroic reading is preferable),58 and he does not care whether the “heroic” reading has degenerated over time, if it is a recent, that is to say, ersatz creation or, indeed, if it represents an authentic tradition.59 He neither examines the text of the alleged heroic epic (for instance, to ascertain whether its Sanskrit is from the sixteenth, the twelfth or the third century) nor does he evaluate its variants (there are none: the lectio heroica is a lectio singularis). As long as it satisies the criterion of “a cruel crudeness of passion” he accepts it into his reconstruction.60 Scholars interested in a scientiic textual criticism will apply more stringent criteria. They will consider it insuicient to identify tales of derring-do—abduction of women, ight scenes and ribald merry-making—Bigger’s sole criterion for considering a passage “ancient.” They will not mistake their expectations of a rough-hewn warrior age for genuine authenticity.61 Even if they suspected that some passages were “Kṣatriya” in origin, they will distinguish between their readings and their contents. They will not allow their ideas about the “classical” ideal of epic to mislead them.62 Most important, they will ask if readings suspected of deriving from the hypothetical Kṣatriya epic are necessarily good and how they can be both ancient and attested. As we saw, an oral source like that Bigger posits cannot simultaneously meet the traditional criteria for preferring readings, antiquity and attestation.63 A reading can either be ancient (in which case its accuracy is doubtful, given the long transmission between original and descendant) or it can be attested (in which case its antiquity is doubtful). A good reading from the former perspec- tive will always be bad from the latter and vice versa. This makes the question of why the German critics repeatedly prioritized hypothetical oral sources over the manuscript tradition all the more puzzling. 140 140 PHILOLOGY AND CRITICISM The Argument from (Postulated) Antiquity and the Argument from Ideology This concluding section examines two related arguments, the argument from (postulated) antiquity and the argument from ideology. The irst refers to the Mahābhārata critics’ habit of declaring that certain passages are old because they are postulated as old. The second refers to their ideological reasons for doing so, namely, proving the baleful nature of Brahman domination for India. Refutation of the argument: Neither argument is actually an argument. The argument from (postulated) antiquity merely asserts the antiquity of certain passages without demonstrating it. The argument from ideology elevates the German scholars’ anti-Semitic prejudices to a irst principle and uses this, in turn, to sustain the kind of circular, counterfactual and logically fallacious arguments we have seen. These prejudices hold the key to understanding the German scholars’ work, speciically their insistence on a tendentious Brahmanic redaction of an earlier heroic epic. Throughout this chapter, we have tried to rehabilitate the hypothesis of contamination from an oral source. We irst deined and clariied three types of contamination. We then examined ways to objectively reconstruct the source of contamination. We next studied the problems with asserting contamination from a nonextant oral source and showed that it only caused the stemma to shift in unpredictable ways. Finally, we tried to see whether, if we kept everything else the same and allowed contamination from an oral source only into a recent copy, the hypothesis could work. In every case, the hypothesis proved untenable. Even after we made the ultimate concession—that we knew not only which manuscript had undergone contamination but also which two interpolations it owed to the hypothetical oral epic—we still found the hypothesis indefensible. It must therefore be abandoned. If the hypothesis of contamination from an oral source cannot be defended, the hypothesis of a Brahmanic redaction of an earlier oral epic likewise becomes unsus- tainable.64 This afects not only Bigger’s so-called normative redaction but also his so-called prehistory of the normative redaction as pre-Pā ñcarā tra, Pā ñcarā tra and post- Pā ñcarā tra redactors allegedly interpolated large quantities of their theological doctrines into an earlier Kṣatriya epic. Since there is no longer any reason for assuming the exis- tence of this epic, we must likewise abandon this thesis. A  critical clariication of the critics’ arguments for it revealed that they had, in every instance, assumed its existence as the basic condition for their reconstructions and textual histories rather than demon- strating it.65 These reconstructions, moreover, amounted to massive interventions in the tradition. Nowhere in the history of scholarly editing have interventions been made on this scale in the name of “critical” scholarship.66 We must suspect that the Mahā bhā rata critics’ aim was not the restoration and preservation of the text, not even of its “best” readings,67 but a campaign of erasure, indoctrination and conversion.68 Whereas historical investigations are necessary and useful, what the Mahā bhā rata critics provided in the name of a history of the text—or, rather, a Textgeschichte, which is very diferent69—was not an objective history but a ictional narrative about the Brahmans’ baleful inluence over India. Its roots in German Protestantism are now evi- dent. Its anti-Semitic sources, motivation and intent are no longer debatable.70 Redeining the Mahā bhā rata critical edition as a “normative redaction” in question-begging ways 14 RECONSTRUCTING THE SOURCE OF CONTAMINATION 141 does not advance the case for a Brahmanic redaction of an earlier Kṣatriya epic.71 As the German scholars’ last hope for reviving their racist, sectarian and xenophobic narratives, its overcoming marks a crossroads for Mahā bhā rata studies.72 As we saw, the circumstance that editors can reconstruct an archetype does not mean that this text (the codex unicus of the archetype) was the sole one in existence in its time. Even if someone wanted to expunge the tradition, he could not have known that he had to redact this exemplar rather than any other, since he could not have known that it rather than any other would produce descendants that survived. We could perhaps argue that the Brahmans redacted all exemplars ever in existence, so that regardless of which exemplar we reconstruct as our archetype we will end up with a Brahmanic version of the text. A conscious decimation of all sources that ever existed such as the analytic critics posit for the Mahā bhā rata, however, has never been observed. Neither would it be possible in ancient times, lacking centralized repositories and electronic databases, to locate all exemplars in existence. Even if the argu- ment is that the Brahmans need only have destroyed suicient numbers of exemplars of the Kṣatriya epic to ensure that their version, rather than the competing Kṣatriya version, survived, we could not explain how they located these exemplars.73 Moreover, if they destroyed exemplars individually there was not one redaction but several unless we wish to maintain that the destruction was separate from the act of redaction. How did they collect the exemplars? Which one did they select as the base for their redaction? Were they indiferent to the choice or was the version selected already a proto-Brahmanic epic? The alleged exemplars, moreover, are not manuscripts but refer to oral narratives. How then could the Brahmans have decimated them? Even if the argument is that the Mahā bhā rata critics, lacking a knowledge of the stemmatic method, wished to focus on the history of the text, they have not succeeded in providing a history of the tradition. A history of the text is, of course, a necessary and desirable auxiliary to the abstract reconstruction in the stemma. We recognize its importance. But it must begin with the text after its irst emergence (or what can be reconstructed as such). A  good example is the work of Veronika von Büren, who researches the manuscript tradition of Isidore’s Etymologies: The notice on Receswinthus does not therefore seem to originate from any of the eighteenth- century witnesses. The oldest original occurrences are provided by D and H, both following the French tradition. Bischof proposed localizing the origin of D in the region of Murbach. For H he proposed an origin in western France on account of the style of writing, reminiscent of that of Tours, and pushed back its date, previously ixed at the start of the century. The spelling of H is full of Visigothic symptoms such as uncertainty in the use of the initial h (habel for Abel, 1.VII; XVII. 10.1 ortus for hortus), hesitation between K and Ch (kam for Cham, 1. VII), b–p (scribtum, 1. VII), b–v (XVI.25.1 iubat corr. iuvat), e–I, q–c (locuntur for loquuntur, 1. VIII), o–u (Epicoros for Epicurus, 1.VIII), i–g (II.2.67 Jasius for Gasius). A good deal of these irregularities have been corrected soon after the copy, sometimes even by the copyist. They must therefore come from the model, like the syllabiication of words, especially in the titles (for example XVI.25 de pon de ribus), that one inds in the Visigothic evidence AKLM. This model also permitted the confusion r–s (II.2.64. ab Iscule H ab Hercule H2), which is again the case with Visigothic manuscripts of the eighteenth century. For his text, H is sometimes in agreement with KLM, but more often with D and the group L*Bby, thus the testimony of Theodulf. In 142 142 PHILOLOGY AND CRITICISM the chapter on rhetoric igures of the second book, which H and D possess with the group of Theodulf, HD are consistent with each other, with some rare exceptions, and join almost sys- tematically the testimony of Theodulf. The decoration of H, the initials outlined in green or yellow, are reminiscent of manuscripts at Fleury. I think that H made use of the manuscripts that Theodulf had gathered to design his Isidorian edition. Very likely he had it copied in one of his scriptoria, Fleury or Micy S. mesmin, on an Italian model. I have been able to establish that he served subsequently in a sciptorium of Hincmar of Reims. One therefore can state that the basic version from the end of the 5th book is the one transmitted by the Italian witnesses (and T) and that all the textual additions are linked to the activity of Theodulf. It is possible that he is even at the origin of the epitome of this chronicle, whose foreign character with respect to the composition of the Etymologies already irritated Porzig and subsequently Reydellet.74 Von Büren’s arguments are not stemmatic but they are empirical, being based on her analysis of the copyists’ diferent styles and usages (the usus scribendi). Her work contrasts with the German Mahā bhā rata critics’, who neither know nor care about manuscript culture and scribal practices in ancient India, yet invoke the Mahā bhā rata’s apparent het- erogeneity to discount its textual integrity.75 One could just as well suspect the authenticity of Isidore’s Etymologies, a work no less diverse in its contents.76 We are aware that specu- lation about Isidore’s sources exists. But here also, a diference exists between classical or romance philologists and the Mahā bhā rata critics. We only need compare, for instance, the work of a scholar such as Jacques Fontaine with that of John L. Brockington (discussed ‘Codex inemendatus’ d’lsidore edition de Braulio I partie en III livres L* codices wisigoths de De natura rerum Chronicon l’ltalie du nord 750 Travaux pour le LG (Théodulf et autres) Bern AA 90.23 l LG S. Gall T 800 D H a Reg lat 310 C Familie espagnole C2 Edition de Théodulf B b eY Avant 820 d y f F l v Bern 36 Figure 29 Von Büren’s stemma of the manuscript tradition of Isidore’s Etymologies Source: Reproduced from von Büren, “La place du manuscrit Ambr. L 99 sup. dans la transmission des Étymologies d’Isidore de Séville,” 43. 143 RECONSTRUCTING THE SOURCE OF CONTAMINATION 143 in Appendix 18) to dispel the illusion that the Mahā bhā rata critics actually identiied the sources the Mahā bhā rata’s alleged composers used.77 In fact, contrary to their claims, the Mahābhārata critics have not considered the work’s sources at all—its reuse and trans- formation of Vedic ritual and philosophy, its borrowing from Purāṇic cosmologies, its references to the dharma sūtras and ́āstras, its citation of all the known branches of knowl- edge and its relation to the Sanskrit traditions of kāvya and alaṅkāráāstra.78 Rather, by obsessively focusing on the so-called Kṣatriya epic, they overlooked valuable testimony in the work itself about its sources, its reception and its proper interpretation.79 As Bigger’s work demonstrates, the German Indologists disputed the Mahābhārata’s form even after a critical text became available. Even though the stemma provides the only objective basis for studying the text’s history (and the edition itself is, at present, our only source of knowledge about the tradition), the German scholars chose to discount its evidence for a fetish history of an original oral epic and its Brahmanic “contamination.” We conclude by presenting two contrasting igures—Von Büren’s stemma of the manuscript tradition of Isidore’s Etymologies (Figure 29) and a inal stemma (Figure 30) that summarizes all the arguments the analytic critics have made for their view of the Mahābhārata. Original K atriya/ Bardic ω P1 Hyperarchetypal β contamination Brahmanic P2 γ P3 δ Extra-stemmatic contamination Archetype θ (alleged normative redaction) N S ν γ σ ε K V B D T G M Witnesses Figure  30 Hyperarchetypal contamination, extra-stemmatic contamination and the resil- ience of tradition 14 144 PHILOLOGY AND CRITICISM Figure 30 combines elements of Bigger’s so-called prehistory of the normative redac- tion with Sukthankar’s stemma. It indicates the transmission prior to the reconstructed archetype as Bigger imagines it. The two upper cones overlaid over each other represent the Kṣatriya and the Brahmanic traditions, respectively. Over time, as elements are inter- polated into the Kṣatriya epic from Brahmanic sources, the former declined until at θ it ceases to exist. The elements l, m and n survived past the alleged normative redaction’s constricting inluence and reentered the tradition below the archetype (recombinant transmission). Whereas some elements survived past the alleged normative redaction’s constricting inluence and reentered the tradition below the archetype (recombinant transmission), others were retained in the redacted text itself and thence transmitted to the witnesses (resilient transmission). After the occurrence of a normative redaction at θ, the tradition again diverged. (The narrowing of the upper cones into the norma- tive redaction therefore does not represent the same thing in both cases. In the case of the Kṣatriya epic, it represents the gradual decline of this tradition. In the case of the Brahmanic work, it represents a standardization of texts rather than a decline.) These two igures—van Büren’s stemma and our stemma—illustrate better than words ever could the diference between conscientious historical and philological research and empty and untenable speculation. They simultaneously illustrate the direction in which a responsible Mahā bhā rata criticism must now develop. Notes 1 “[…] il [l’éditeur] doit s’eforcer, par tous les moyens dont disposent sa science et son ingéniosité, de retrouver, par dessous, les constantes de la tradition ‘verticale’. C’est à ce prix, et à ce prix seulement, qu’il pourra déterminer la réalité, et éventuellement l’ampleur, d’une transmission horizontale.” Jean Irigoin, “Quelques rélexions sur le concept d’archétype,” Revue d’histoire des textes 7 (1977): 242–43. 2 For references to a written tradition as opposed to an oral one, see James L. Fitzgerald, “The Many Voices of the Mahā bhā rata,” review of Rethinking the Mahābhārata: A Reader’s Guide to the Education of the Dharma King, by Alf Hiltebeitel, Journal of the American Oriental Society 123, no. 4 (2003): 803–18 and James L. Fitzgerald, “Mahābhārata,” in The Hindu World, ed. Sushil Mittal and Gene Thursby (New York and London: Routledge, 2004), 52–74. 3 Mahā bhā rata critics appear unaware of Maas’s deinition of the archetype as “the source text with which the irst branching began.” Paul Maas, Textkritik, 4th edn. (Leipzig:  B. G. Teubner, 1960), 6. Because they erroneously think a critical edition is made solely by elim- inating interpolations, they think no reason exists why this process should not continue ad ininitum. Austin referees Bigger’s work thus: “Important here in his [Bigger’s] framing of the ‘M0’ moment is the looming presence of the subsequent M+N materials, which establish the patterns of lux and inlation of the normative redaction over time. For Bigger it is thus only natural to seek, now through higher text-critical tools, these same patterns of lux and inla- tion within the M0 text itself.” Christopher Austin, “Evaluating the Critical Edition of the Mahā bhā rata:  Inferential Mileage and the Apparatus Materials,” Journal of Vaishnava Studies 19, no. 2 (2011): 76. In his view, nothing prevents us from positing an ancestor of M0 (the archetype) on analogy with the relation of the archetype to its witnesses (his so-called M+N moments). “If we have nothing available to take us back beyond M0—no means to trace with any certainty the processes leading to and producing the archetype—we may look to the evi- dence of M+N, which may leave us with the suspicion that the M0 text represented by the CE will, upon inspection, reveal some of the same kinds of patterns of addition and accretion as 145 RECONSTRUCTING THE SOURCE OF CONTAMINATION 145 do the manuscripts from which it was constituted. According to this view, to resist the possibility of M–N moments is to turn the ‘screen shot’ of the CE into a kind of Ur-MBh. Perhaps most importantly, it may be suggested that the evidence of M+N makes the refusal to consider the existence of M–N unreasonable.” Ibid., 77. The problem with this argument is that if we can posit an ancestor of θ called θʹ on analogy with the relation of θ to its witnesses (θʹ: θ :: θ: Ś, K, etc.), we can also posit an ancestor of θʹ on analogy with the relation of θʹ to θ (θʹʹ: θʹ :: θʹ: θ). Thereafter we could keep positing ancestors as follows:  θʹʹʹ:  θʹʹ ::  θʹʹ:  θʹ, θʹʹʹʹ:  θʹʹʹ ::  θʹʹʹ:  θʹʹ, θʹʹʹʹʹ: θʹʹʹʹ :: θʹʹʹʹ: θʹʹʹ and so on. 4 See Mary Carroll Smith, “The Mahā bhā rata’s Core,” Journal of the American Oriental Society 95, no.  3 (1975):  480–81. See also James L.  Fitzgerald, “Triṣtụ bh Passages of the Mahābhārata,” Epic Undertakings:  Papers of the 12th World Sanskrit Conference, vol. 2., ed. Robert P.  Goldman and Muneo Tokunaga (Delhi:  Motilal Banarsidass Publishers, 2009), 95–117, and James L. Fitzgerald, “Toward a Database of the Non-Anuṣṭubh Verses in the Mahā bhā rata,” in Epics, Khilas, and Purāṇas: Continuities and Ruptures. Proceedings of the Third Dubrovnik International Conference on the Sanskrit Epics and Purāṇas, ed. Petteri Koskikallio (Zagreb: Croatian Academy of Sciences and Arts, 2005), 137–48. 5 We put “style” in scare quotes because, obviously, the Mahā bhā rata critics mean something else by the term than editors, for whom it refers to either an author’s style (deducible from multiple works known to be written by him) or to a scribe’s style (again, deducible from his manuscripts and peculiarities of handwriting or dialect). The Mahā bhā rata critics do not think a single author existed. They know nothing about the language or regional dialects at the time. They have made no study of the scribal tradition. Rather, when they use the word style, what they mean is the distinction between Brahmanic and heroic poetry, a distinction they make on the basis of the contents of this poetry (or, rather, since it all occurs in the Mahā bhā rata, the contents of the parts assumed Brahmanic, etc.) 6 Even Van Buitenen, though aware of “the antibrahminism of Western scholarship of the last century,” falls into the trap of positing a “Brahmanization.” He distinguishes four stages of the epic:  “The Central Story,” “The Fuzzy Edges,” “The Second Perimeter” (the stage of “inept mythiication”) and “The Third Perimeter” (identiied with “Brahminization”). For his comments, see J. A. B. van Buitenen, “The Mahā bhā rata: Introduction,” in J. A. B. van Buitenen, trans., The Mahābhārata, vol. 1:  The Book of the Beginning (Chicago:  University of Chicago Press, 1973), xiii–xxiii. 7 Besides the obvious prejudices in Mahā bhā rata studies such as the prejudice against “Brahmanism” rooted in German anti-Semitism, the prepossession in favor of a heroic, bardic epic rooted in German romanticism and the prejudice that all “higher” cultural achievements in ancient India (including the parts of the epic considered genuine, heroic, rational, free-spirited, etc.) were owed to the Aryans or Indo-Germans, epistemic prejudices also exist about the correct interpretation of texts, the economy of meaning within which texts emerge, the ontological status of language and the value of philosophy and ethics. We speciically mean these here. 8 The so-called analytic school traces back to the work of Hopkins. See Edward W. Hopkins, “The Bhā rata and the Great Bhā rata,” The American Journal of Philology 19, no.  1 (1898):  1–24 and Edward W. Hopkins, The Great Epic of India: Character and Origin of the Mahābhārata (New York: Charles Scribner’s Sons, 1901). Simon Brodbeck, “Analytical and Synthetic Approaches in Light of the Critical Edition of the Mahā bhā rata and the Harivaṃsá ,” Journal of Vaishnava Studies 19, no. 2 (2011): 223–50 provides an overview of the two terms and their associated schools. 9 Timpanaro’s deinitions in Sebastiano Timpanaro, The Genesis of Lachmann’s Method, ed. and trans. Glenn W. Most (Chicago: University of Chicago Press, 2005), 179. Reeve argues that R. D. Dawe, The Collation and Investigation of Manuscripts of Aeschylus (Cambridge: Cambridge University Press, 1964), 159–60 used the term before Timpanaro, but does not seem to have invented it. Michael D.  Reeve, “Stemmatic Method:  ‘Qualcosa che non funziona?’ ” in The Role of the Book in Medieval Culture:  Proceedings of the Oxford International Symposium, September 146 146 PHILOLOGY AND CRITICISM 26–October 1, 1982, ed. P. Ganz, Bibliologia 3 (Turnhout: Brepols, 1986), 66, no. 20. See ibid. for Reeve’s criticisms of Timpanaro’s deinition of extra-stemmatic contamination, which leaves him “wondering how anything can be extrastemmatic in the stemma of a whole tradi- tion” (Reeve’s italics). Reeve’s criticism does not concern, as Trovato thinks, “the consequences that extra-stemmatic contamination has, according to Timpanaro, on bipartition, and not the fact […] that a medieval or Renaissance copyist could easily draw on early witnesses which have disappeared without leaving surviving copies.” Paolo Trovato, Everything You Always Wanted to Know about Lachmann’s Method: A Non-Standard Handbook of Genealogical Textual Criticism in the Age of Post-Structuralism, Cladistics, and Copy-Text (Padua: Libreriauniversitaria, 2014), 137. Rather, as Reeve clariied in a draft version shared with the authors, it concerns “a purely logical objec- tion to calling anything ‘extrastemmatic’ in the stemma of a whole tradition: ‘from outside’ what stemma, then? One can always draw up a stemma of some witnesses and leave the rest out of it” (Reeve’s italics). Trovato’s “extra-archetypal” is therefore preferable. Reeve’s objection does not apply here, as the analytic critics do not mean contamination from a part of the stemma outside our hypothetical stemma (that is, the stemma we draw up) but from a source outside even the “real” stemma—indeed, a source so distinct from the epic tradition that we must posit a sinister Brahmanic takeover of the latter as the reason they became associated. See also the next note. 10 Despite the similarity of name, hyperarchetypal contamination difers from what Trovato calls “extra-archetypal contamination.” The latter refers to the circumstance that “a ms. that descends to some extent from a known exemplar contains high-quality variants not found in any known ms. or subfamily,” in which case we assume that its scribe simultaneously drew on a second source from a branch lying outside the tradition descended from the archetype, a branch that ultimately did not survive—hence the name extra-archetypal contamination. If the con- tamination occurred from a manuscript from a branch of the tradition, one or more of whose members survived, then their common ancestor—that is, the ancestor they share with surviving manuscripts of other branches—would be the archetype, and the contamination would be intra-stemmatic. Trovato’s extra-archetypal contamination is, in fact, a synonym for Timpanaro’s extra- stemmatic contamination, preferable not only because it clariies that the contamination occurred from a branch that descends outside of the archetype or from a higher-order ancestor than the archetype but also because it avoids the misleading suggestion that contamination could occur from outside the stemma, which it, of course, cannot (hence Reeve’s criticism cited in the previous note). What Timpanaro, strictly speaking, means is that it occurs from outside the stemma of the surviving manuscripts, that is, the stemma that we draw up. In contrast, hyperarchetypal contamination refers to the circumstance that the Mahā bhā rata critics consider the archetype itself a composite of two diferent sources—a Kṣatriya and a Brahmanic tradition— on no other ground except that they think they can identify the perceived class interests of the two groups at work in diferent parts of the text. The contamination thus occurs during the for- mation of the archetype, that is, from above the archetype into it, as the preix hyper emphasizes. It difers from extra-archetypal contamination in that contamination does not occur from a branch outside the tradition descended from the archetype into an extant source but from a source unrelated to the tradition into the archetype, a source that is, for this reason, really outside the stemma. Hence, Timpanaro’s extra-stemmatic contamination, had it not already been claimed and had Trovato’s extra-archetypal contamination established itself instead, would have been more appropriate. A corresponding term does not exist in classical philology, probably because, while we know of extant archetypes that reveal the inluence of two hands, no one has suggested that a reconstructed archetype is a mix of two traditions. For Trovato’s discussion of extra-stem- matic contamination, see Trovato, Everything You Always Wanted to Know about Lachmann’s Method, 134–38 and see especially ibid., 137–38 for his justiication for retaining Timpanaro’s term. The quotation is from page 134. See also Mortimer Chambers, review of Thucydidis Historiae, Vol. I: Libri I–II, by Carolum Hude and Otto Luschnat, Classical Philology 51, no. 1 (1956): 44, to our knowledge the earliest known reference to an “extra-archetypal” manuscript or source. 147 RECONSTRUCTING THE SOURCE OF CONTAMINATION 147 Trovato incorrectly attributes it to Edward N. O’ Neil, review of Juvénal, les baladins et les rétiaires d’après le manuscrit d’Oxford, by Jean Colin, Classical Philology 51, no. 1 (1956): 42–44. 11 Andreas Bigger, Balarāma im Mahābhārata:  Seine Darstellung im Rahmen des Textes und seiner Entwicklung (Wiesbaden: Otto Harrassowitz, 1998), 145. In a footnote, Bigger clariies that by “all along” (schon immer) he “of course, does not mean that the Mahā bhā rata existed since the beginning of time.” Rather, he “means the textual corpus resulting from the irst redaction of the Mahā bhā rata.” Ibid., 145, n. 1. “Textual corpus” suggests a written work, but elsewhere he argues that the Mahā bhā rata was irst “transcribed” during the normative redaction. He also refers to the normative redaction as the “sole written version.” See ibid., 111 and 118. We therefore assume that he means a body of oral retellings. 12 Ibid., 107. Bigger cites four sources in support:  Moriz Winternitz, Geschichte der indischen Litteratur, vol. 1, 2nd edn. (Leipzig:  C. F.  Amelangs, 1909), 266; D. D. Kosambi, “The Autochthonous Element in the Mahā bhā rata,” Journal of the American Oriental Society 84, no. 1 (1964): 43; Van Buitenen, “The Mahā bhā rata: Introduction,” xxi and Klaus Mylius, Geschichte der altindischen Literatur (Bern: Scherz, 1988), 91 and 105. Kosambi writes that “the MBh began as a “Kuru lament, changed into a Pandu song of victory; absorbed Naga myths during acculturation; acquired fresh episodes from the Mathuran Kṛṣṇa saga; and was brahminised to some extent by the Kaśyapas, but overwhelmingly by the Bhṛgus into its present form.” Kosambi, “Autochthonous Element,” 43. Mylius claims that “in contrast to Vedic literature, epic literature emerged primarily in Kṣatriya circles,” whereas the later “bridge narrations” “are not always borne by the anti-ascetic, active attitude of the Kṣatriyas but have been occasionally composed and inserted into the epic by the Brahmans.” Mylius, Geschichte der altindischen Literatur, 91 and 105. Winternitz provides the strongest support for Bigger’s Brahmanization hypothesis. His comments are examined later. In contrast, Van Buitenen provides only qualiied support for Bigger’s views. He notes, “It would be easy at this point to be traduced into believing that a diferent mentality has taken over the grand old baronial tradition, and to identify that mentality with that of the brahmins. But it is hard to revive the antibrahminism of Western scholarship of the last century, since all the evidence points to a necessary symbiosis of brahmin and baron; as our text has it, ‘The baronage is in league with brahmindom, and brahmindom is allied with the baronage.’ Perhaps we should rather think that the baronial-bardic tradition, out of which the epic grew, was expiring and absorbed into another tradition of wandering reciters of brahmin-type lore.” Van Buitenen, “The Mahā bhā rata: Introduction,” xxi. 13 See Bigger, Balarāma im Mahābhārata, 107: “The Bhā rgava layer represents one of the earliest Brahmanic layers in the developmental history of the MBh. However, for these Brahmans the Bhā rgava were not necessarily their own family, but rather the Brahmans par excellence, the preeminent family. If the Bhā rgava layer is really the earliest Brahmanic layer, the oldest Balarā ma layer must be attributed to the period before it. […] That the Brahmans had little use for such a Kṣatriya [as Balarā ma] is conirmed by the fact that the Nā rā yaṇīya layers, which became inluential at a time when the transmission of the MBh already lay in the Brahmans’ hands, had little to add to the igure of Balarā ma beside Saṃkarṣaṇa. […] In my opinion, the Baladeva tı̄rthayātra stands on the threshold or, rather, already on the other side [of this change]. Its religiosity, which speciically emphasizes a reverential attitude vis-à-vis the Brahmans, makes it likely that the redactors responsible for this passage were Brahmans. This is possibly one of the oldest passages in the MBh, which trace to a Brahmanic inluence.” 14 Ibid., 145. 15 Ibid., 145–46. 16 Reeve points out that stemmatists can reconstruct an archetype, but not much above it. They can posit an ancestor that had particular sections or particular leaves or was written in a par- ticular script, but whether they would be doing so qua stemmatists is debatable. We could hypothetically argue that before the archetype x and y must have happened, for the following reasons. But any argument of this type must be noncircular and independently veriiable. 148 148 PHILOLOGY AND CRITICISM Historical sources must exist—testimonia, citations, documentary evidence and the like—that verify it. Finally, it may appear that the analytic critics’ argument is an inductive argument of the kind “the surviving witnesses were interpolated, so it is likely that the archetype or its ancestors were too.” The problem is that their argument is precisely not inductive. All the empir- ical examples of interpolation are—and they themselves not only grant but underscore this fact— examples of Brahmanic interpolations into a Brahmanic work, whereas they posit Brahmanic interpolations into a Kṣatriya epic, for which no proof has yet been found. The argument is, at best, hypothetical, while its proof is viciously circular. 17 Angelika Malinar, The Bhagavadgı̄tā: Doctrines and Contexts (Cambridge: Cambridge University Press, 2007) features the term interpolation 36 times, whereas layer occurs 45 times. Vishwa Adluri, review of The Bhagavadgı̄tā: Doctrines and Contexts, by Angelika Malinar, History of Religions 50, no. 1 (2010):  102–07 notes Malinar’s ininitely malleable identiication of layers. For a critique of earlier schemes for layers, see Vishwa Adluri and Joydeep Bagchee, “Paradigm Lost:  The Application of the Historical-Critical Method to the Bhagavadgītā ,” International Journal of Hindu Studies 20, no. 2 (2016): 199–301. 18 The following discussion is based on Timpanaro’s. His text was modiied and integrated with the running text. Lachmann’s original comments are cited in the notes where possible. 19 Timpanaro, The Genesis of Lachmann’s Method, 139. 20 Ibid., 140. 21 Ibid. The original reads:  “Diese Änderungen, welche bald dieser, bald jener Schreiber übersehen, und jeder mit neuen vermehrt hat, herauszuinden, das ist die Aufgabe des Herausgebers.” Karl Lachmann, “Der Nibelungen Lied, herausg. von v.  d. Hagen. Breslau 1816,” in Kleinere Schriften zur deutschen Philologie, vol. 1, ed. Karl Müllenhof (Berlin: G. Reimer, 1876), 87. 22 Timpanaro, The Genesis of Lachmann’s Method, 140. 23 Accepting Bornmann’s emendations. See Fritz Bornmann, “Sui criteri di ‘recensio’ meccanica enunciati da Lachmann nel 1817,” Rivista di Letterature Moderne e Comparate 15 (1962): 45–53. Bormann’s emendations are discussed in Timpanaro, The Genesis of Lachmann’s Method, 142. This discussion is not reproduced here, as the details of Lachmann’s method are less important than the observation that a minimum number of manuscripts arranged in a minimal architec- ture is needed to reconstruct the source of contamination. 24 Timpanaro, The Genesis of Lachmann’s Method, 140. The presentation follows Timpanaro’s discussion with the aforesaid emendation. In the original, Lachmann’s rules read: “(1) Drey Handschriften unter unseren vieren überstimmen alle Mal eine. (2)  Wo je zwey überein stimmen ist BG < EM (d.h. in Stellen, wo B mit G übereinstimmt, die einstimmige Lesart von E und M vorzuziehen), GE > BM, GM > BE. (3) Wo drey Lesarten sind, da ist BG < E—M (die Lesart, welche BG gemeinschaftlich haben, die beiden andern in E und M vorzuziehen), GE > B—M, GM > B—E; hingegen EM = B—G (die Übereinstimmung von E und M führt gegen die zwey Lesarten von B und G zu keiner Entscheidung), BM = G—E, BE = G—M. (4) Eben so ungewiss bleibt die ursprüngliche Lesart, wo alle vier uneinig sind.” Lachmann, “Der Nibelungen Lied, herausg. von v. d. Hagen. Breslau 1816,” 87. 25 Timpanaro, The Genesis of Lachmann’s Method, 141. 26 Ibid., 143. 27 Unless we can show both readings are corruptions of an original. In the Mahā bhā rata, the two hyparchetypes difered not only in readings but also in entire accounts, so this cannot apply. 28 Reeve writes, “if each [scribe] behaves eclectically as in your hypothetical example, it will be hard to draw up a stemma at all, because there will be errors shared by each of the three pos- sible pairs where the other ms. has a true reading that couldn’t have been conjectured from the shared one. Whether it then makes sense to speak of an archetype at all [would] depend […] on the nature of such errors as all three share. Missing chunks of text or large transpositions that disrupt the low are the kind of thing that leads editors of classical texts and doubtless 149 RECONSTRUCTING THE SOURCE OF CONTAMINATION 149 others to postulate an archetype even if they can’t draw a satisfactory stemma below it.” In his opinion, whether one “one can tell where the contamination happened—in the archetype or somewhere below it—[will, ultimately,] depend on how the pairs behave, and the argument may well turn on probabilistic assumptions about the habits of scribes.” Personal communica- tion, e-mail message to authors, January 1, 2011. 29 The following discussion draws on M. L. West, Textual Criticism and Editorial Technique (Stuttgart: B. G. Teubner, 1973). West speaks merely of errors, whereas the Mahā bhā rata editors had to deal both with errors and with extensive divergences between texts. We consider this limitation later. 30 Bigger, Balarāma im Mahābhārata, 118. 31 In West’s analysis, B is an extant copy, whereas γ is an inferred subarchetype. Nonetheless, assuming that we can reconstruct ε from Ñ, V and B and assuming that we inferred γ as ε’s source, contamination of D would seriously complicate our inferences about γ’s relationship to D. 32 The argument is, of course, only hypothetical. No good foundation exists for comparing extant witnesses with inferred sources until a stemma is established, but this shows just how arbitrary the analytic critics’ claims about sources and transmission are without a rigorous stemma. Because contamination makes it impossible to draw up a stemma with conidence, it erases the distinction between more probable inferences and less probable ones. This is, of course, the Mahā bhā rata critics’ real aim. By vitiating a perfectly valid stemma and arguing from hypotheticals, they replace serious discussion with arbitrary speculation. Compare Bigger’s use of “it is conceiv- able” (denkbar wäre), “possibly” (möglicherweise), “probably” (wahrscheinlich) and “perhaps” (vielleicht). For citations see, in order, ibid., 15, 35, n. 76, 47, 55, n. 150, 67, 99, 117, 121, 141 and 148; 7, n. 44, 51, n. 132, 60, 61, n. 180, 62, n. 185, 63, 72, 75, 94, 98, 100, n. 111, 107, 121, 126, 130, 134 and 139; 33, 38, 39, 53, 63, 66, 70, 72, 74, 75, 84, 88, 89, n. 58, 92, 96, 97, 98, 102 (also 102, n. 118), 104, n. 126, 105, 106, 107, 111, 112, 114, 117, 118, 119, 122, 123, 126, 129, 131 (also 131, n. 76), 132 (also 132, n. 81), 136, 139, 141, 143, 147 and 149; and 10, 31, 48, 60 (also 60, n. 176), 64, n. 190, 70, n. 5, 76, n. 32, 84, 85, 93, 97, 103, 110, 106 and 124. 33 West does not distinguish between extra-stemmatic and intra-stemmatic contamination, but since A in his scheme is a lost manuscript, Timpanaro’s distinction applies. It lets us distinguish contamination between extant witnesses from contamination from sources not in our stemma. The circumstance that θʹ is an oral source rather than a written exemplar is irrelevant for the argument. 34 This is especially true if we considered θ a contaminated exemplar (contamination with the Brahmanic source ωʹ). In this case, we would disregard θ’s readings, and D would be identical with ω. 35 We, of course, know that D represents a late and inferior group of manuscripts descended from γ. The diiculty is therefore merely hypothetical. We would reject contamination from an older source θʹ. But it nevertheless shows that, as long as we do not know the actual transmission, contamination will lead to misleading inferences. 36 Theoretically, several such intermediate sources could have existed. A stemma, however, only records the nodes from which two or more lines branch because without a minimum of two manuscripts or manuscript (sub)families for comparison, we cannot infer the existence of a (common) ancestor. We thus posit the minimum number of intermediate nodes necessary to connect a family of manuscripts based on their observed diferences and agreements, even though, most likely, signiicantly more manuscripts existed between the nodes than recorded. “The lines of stemmata, like geometric points, are composed of ininite points.” Trovato, Everything You Always Wanted to Know about Lachmann’s Method, 148. 37 Sukthankar notes, “S is the ultimate source from which all versions of the Southern recen- sion are, directly or indirectly, derived and which is appreciably longer than N, and far more elaborate (textus ornatior).” V. S. Sukthankar, “Prolegomena,” in The Ad̄ iparvan for the First Time Critically Edited, vol. 1 (Poona: Bhandarkar Oriental Research Institute, 1933), xxxi. See also ibid., xxxvi: “The Southern recension of the Ā di at least is thus not merely longer, but also 150 150 PHILOLOGY AND CRITICISM fuller, more exuberant, more ornate than the Northern. It may therefore be itly styled, in rela- tion to the Northern, the textus ornatior.” 38 Sukthankar identiies seven types of changes in the southern manuscripts over the northern manuscripts in the “Prolegomena”: 1. “Variants of isolated words or phrases, (a) unimportant and (b) important.” 2. “Larger variants between continuous passages, as a whole, the total extent remaining approximately the same.” 3. “Expansion of the text in S without materially altering the nature of the contents or the course of the narrative.” 4. “Southern additions which alter the purport of the fable as narrated in the Northern recension.” 5. “Additions in S, due to the illing out of lacunae (real or imaginary).” 6. “Multiplication of ights and battlescenes.” 7. “Omissions in S, as compared with N.” Ibid., xxxvii–xlvii. Under the third category, he lists six types of changes, including:  “additions in S, due to the explicit mention of the observance of the correct and complete Brahmanic ritual and ceremonial on the proper occasions”; “expansion in S of existing scenes by the addition of speeches or detailed descriptions and by other digressions”; “additions of little ethical, moral and sententious maxims, to which S, permeated as it is by a conscious didactic purpose, is particularly partial”; and “additional stanzas in S with, perhaps, a certain amount of sexual appeal, bearing the taint of later decadence.” Ibid., xl and xlii–xliii. The increasing com- plexity of ritual is usually cited as evidence for the southern recension’s Brahmanic character, but actually, all the changes, including the expansion of the battle scenes, are characteristic of the Mahā bhā rata’s later development. They cannot be cited as evidence for an original Kṣatriya epic. Bigger’s suggestion that “[Mahā bhā rata] 1.211–13 likewise belong among the Balarā ma passages that preceded the Brahmanic layers. The problems discussed there unequivocally belong in Kṣatriya circles” (Bigger, Balarāma im Mahābhārata, 107) illustrates just how persistently the myth of a “primitive Kṣatriya tale of love and war” (Sukthankar) endures. Scholars appear willing to discount tangible proof that the Mahā bhā rata’s Kṣatriyas are its own invention for a reality no less imaginary (that is, their view that these Kṣatriyas once really existed). Simon Brodbeck, review of Strı̄: Women in Epic Mahābhārata, by Kevin McGrath, Indo- Iranian Journal 53, no. 1 (2010): 91 perceptively notes: “If McGrath’s self-inlicted problems of textual deinition seem rather perverse, one might be tempted to speculate more broadly in search of an explanation. The reason projects such as McGrath’s can proceed—even after the production of the Poona critical edition—has to do, I suspect, with a kind of socio-existential desire that formed a context for the early heyday of European Indology and has been fossilised within the discipline ever since.” 39 Scholars have tried to peg the argument that the northern recension is closer to the original story on Sukthankar’s comment that the southern recension is characterized “by its precision, schematization, and thoroughly practical outlook. Compared with it, the Northern recension is dis- tinctly vague, unsystematic, sometimes even inconsequent, more like a story rather naïvely narrated, as we ind in actual experience.” Sukthankar, “Prolegomena,” xxxvi (Sukthankar’s italics). But Sukthankar never once speaks of an original story. He also does not use the term Kṣatriya in the “Prolegomena.” Certainly, Brahmanism is articulating itself ever more richly in the epic, but this does not mean that its earlier stages were, for their lack of development, Kṣatriya rather than Brahmanic. See Alf Hiltebeitel, Nonviolence in the Mahābhārata:  Śiva’s Summa on Ṛṣidharma and the Gleaners of Kurukṣetra (London and New York: Routledge, 2016), 46, 146 and 104, n. 19. 40 Actually, matters are not so clear. Since θ itself is a mix of Kṣatriya and Brahmanic sources (ω and ωʹ), N could have inherited its “Kṣatriya” elements from θ. The probability, however, is slight, since according to Bigger, θ is a Kṣatriya text only to a trivial degree. In contrast, 15 RECONSTRUCTING THE SOURCE OF CONTAMINATION 151 the likelihood that N inherited Kṣatriya elements from θʹ is greater. This is, of course, the reason Bigger introduces it into the stemma. Unfortunately, no evidence exists to support either the claim that the Kṣatriya tradition survived to reenter from an alternative source θʹ or the claim that the Brahmans overwrote θ. Indeed, the sole purpose of introducing θʹ into the stemma (apart from avowedly “increasing” the chances of N being a Kṣatriya epic) is to make the latter claim appear plausible. In reality, θ is neither Brahmanic nor non-Brahmanic. It only appears “Brahmanic” by the postulated contrast with another source, which allegedly descended independent of the Brahmanic tradition from the Kṣatriya epic. Once this latter source is admitted, it appears inevitable that we should discount θ’s evidence for it. In efect, Bigger has made us reject the only source we possess (our reconstructed archetype) for a non- existent source. 41 See Sukthankar, “Prolegomena,” lv:  “In my opinion, therefore, this fact of the concord between ν and S in small details, coupled with the almost entire lack of agreement as regards the additions peculiar to ν or S, is the strongest argument imaginable for the independence of these two versions, and consequently for the primitive character of their concordant readings. It is needless to point out that this is a factor of supreme importance for the reconstruction of the original” (Sukthankar’s italics). 42 Ibid., lvi. 43 Besides not being very old, the reconstructed archetype also will not ofer a good text. Sukthankar notes that the Ś ā rā da codex itself contains “corruptions and interpolations” that “must be carefully corrected and controlled with the help of the other versions, particularly of those of the rival recension.” Ibid., lxi. If we eliminated S, we could no longer correct these problems. Our reconstruction of N will ofer a worse text than Sukthankar’s reconstructed archetype. 44 Ibid., xci. 45 West, Textual Criticism and Editorial Technique, 36. In dealing with an “open” recension, West recommends eliminating manuscripts that resist stemmatic analysis and whose readings are less important for the reconstruction of the archetype, and focusing instead on manuscripts that do descend “directly enough from an archetype for it to be reconstructed.” Ibid., 38. 46 The idea that we can introduce uncertainty into the stemma to make space for an oral epic is a fallacy. It is like claiming that if we shuled four jokers into a pack we are likelier to draw an ace than before. We would think someone who claims this either a fool or a cheat. Since introducing uncertainty into a system afects all outcomes equally, it cannot increase the chances of some outcomes over others. We can only inluence the likelihood of drawing an ace if we shuled a large number of aces into the deck, that is to say, if we stacked the deck. 47 Andreas Bigger, “The Normative Redaction of the Mahābhārata: Possibilities and Limitations of a Working Hypothesis,” in Stages and Transitions: Temporal and Historical Frameworks in Epic and Purāṇic Literature. Proceedings of the Second Dubrovnik International Conference on the Sanskrit Epics and Purāṇas, ed. Mary Brockington (Zagreb: Croatian Academy of Sciences and Arts, 2002), 24. 48 For references to “parallel transmission,” see Bigger, Balarāma im Mahābhārata, 18 (also 18, n. 98), 106 (also 106, n. 132), 117 and 149. For references to “parallel tradition,” see ibid., 158 and 164. And see also ibid., 108 (“versions that ran parallel” and” “parallel-text”), 111 (“par- allel versions”), 112 (“a parallel version”), 114 (“a parallel version” and “this parallel version”), 115, n. 14 (“both parallel versions”), 116 (“genuine parallel version”), 117 (“parallel version,” also: “parallel versions,”), 117, n. 26 (“parallel versions”), 118 (“parallel versions”), 124 (“the parallel version” and “southern parallel version”), 131 (“parallel version”), 132 (“parallel version”) and 165 (“parallel version”). 49 Ibid., 111. 50 Giorgio Pasquali, Storia della tradizione e critico del testo (Florence: Le Monnier, 1952). 152 152 PHILOLOGY AND CRITICISM 51 West, Textual Criticism and Editorial Technique, 50. 52 Bigger, Balarāma im Mahābhārata, 118. 53 More precisely, contamination from an alleged oral source should be treated like the scribe’s own innovations, since we neither have grounds to assume its existence nor could we distin- guish such contamination (if it occurred) from his own innovations. Although Bigger argues that the scribe could have heard this version, we cannot know this and hence have no reason to assume it. Otherwise, any one could assert of any reading that it was not the scribe’s own innovation but an “ancient” reading heard from a secret tradition of illuminati dedicated to preserving good readings. 54 In the examples Timpanaro cites, extra-stemmatic contamination is posited to explain good readings found in an extant source, readings it can owe neither to the primary source it was copied from nor to any other extant source (from whatever family) nor to a nonextant member of one of these families. In this case, we posit that its good readings must have descended from a diferent branch of the tradition altogether, that is, a branch connected with the others through an ancestor of the archetype. Compare his precise explanation of the conditions under which we may posit extra-stemmatic contamination: “There are correct readings at which no medieval copyist-philologist (in certain cases not even the best modern philologist) could arrive conjecturally. A more serious danger consists in the possibility that a copyist, for example, of the α branch […] might have healed errors or illed lacunas not by conjecture and not even by checking one of the other witnesses that have survived to our day, but by collating a manuscript of a completely diferent branch or tradition which was later lost.” Timpanaro, The Genesis of Lachmann’s Method, 179. In contrast, Bigger cites extra-stemmatic contamination (although he does not call it that) as a justiication for the goodness of the passages he selects. They are not good because of their inherent quality (which he never evaluates) but because they are posited as not having descended from the Brahmanic tradition. The argument is not only circular; it is perverse. We shall therefore, henceforth, refer to it as his circular, prejudicial and false hypothesis of the contamination of a tendentious Brahmanic redaction of an earlier oral Kṣatriya epic with the hypothetical remnants of this fantasy epic, reserving extra-stemmatic contamination for legitimate instances of a phe- nomenon well known in textual criticism. 55 This is “the Argument from (Postulated) Antiquity” or, more simply, “the Argument from Assertion.” See the next section. 56 See Bigger, “The Normative Redaction of the Mahābhārata,” 24: “If we assume that there was a normative redaction, an archetype that can be reconstructed, we must at least assume that this text, even if orally transmitted, was already ixed. It is therefore no longer a loating oral tradition as we know it from many other cultures in which epics are transmitted orally, but a ixed oral text, similar to the Vedic corpus, though not so rigidly ixed. It is possible that our normative redaction had the form of a ixed oral text and there is no way of disproving it, but I still think it is more plausible to assume that the normative redaction was the irst written MBh. This may also explain why it was possible for one redaction to have come to dominate the whole manuscript-tradition.” 57 Ibid. 58 Actually, Bigger’s criterion is not “the more heroic reading is preferable” but “what is imagined as the more heroic reading is preferable.” As we saw, he has no way of knowing that the reading he identiies as the original is identical with the reading of the putative heroic epic—that is to say, that it descended unchanged from the latter. It could be a composite of the author/bard’s original intention and what the scribes understood or heard or transcribed in each individual case. From Bigger’s perspective, however, the distinction is irrelevant. Rather than understand reading as editors do (that is, as a lexical or semantic unit that, for this reason, is sensitive to even the slightest change), he understands it as a complete mise-en-scène, a set of ideological tropes and/or motifs he thinks characteristic either of Kṣatriyahood or of Brahmanism and thus serve to distinguish the former from the latter. The putative heroic epic’s readings can thus 153 RECONSTRUCTING THE SOURCE OF CONTAMINATION 153 undergo as much deterioration as they like; it will not alter the fact that they indicate a diferent strand and stratum of Indian thought and history than the alleged lectiones brahmanicae. 59 As “the oldest form of the text which it is possible to reach, on the basis of the manuscript material available” (Sukthankar, “Prolegomena,” lxxxvi; Sukthankar’s italics), the constituted text should be ipso facto the most Kṣatriya version of the text if the history of India has been a history of progres- sive Brahmanization, as the Indologists argue. In that case, the Indologists, who are dissatisied with anything except the purest specimen of the Kṣatriya epic, should accept this text. Even if we grant them that it represents the outcome of a Brahmanic redaction, they should accept it as the closest thing to the Kṣatriya epic we shall ever possess, since we have no standard of comparison outside of the Mahā bhā rata’s later textual history and this history reveals a pro- gressive Brahmanization. Hence the need to discover within this history, that is, the history of the Mahā bhā rata’s subsequent transformations—a history that is recorded in the critical appa- ratus, surviving fragments of the Kṣatriya tradition. Bigger goes to the only place someone who rejects the constituted text can—its critical apparatus—hoping to ind what he did not in the former: a relection of what he thought the Mahā bhā rata would be. 60 The criterion is Holtzmann Jr.’s, who expresses it thus: “We thus ind some similarities in the conditions of the most ancient Indian and the most ancient Germanic periods; here, as in the latter, we encounter a battle-lusty warrior race with all the seeds of culture alongside a cruel crudeness of passion. Whether here one must suppose [that the epic is] a recollection of a pri- mordial age during which they lived together or a further development that ran parallel under similar existential conditions, I  cannot decide here. [But] it is certain that the Mahābhārata has preserved traces for us that reach back to a very early period of Indian antiquity; that therein recollections of an Indo-Germanic primordial age have been preserved, is, so far at least, at best plausible.” Adolf Holtzmann Jr., Zur Geschichte und Kritik des Mahābhārata (Kiel: C. F. Haeseler, 1892), 51. 61 Compare ibid., 45–46 (Holtzmann’s emphasis): “Here we must, in the irst place, emphasize the thoroughly warlike worldview that characterizes the soul of the old parts of the epic. […] Instead of the elegiac softness, the resignation, being tired of life, of later Indian literature the raw warrior-like air of the old Germanic North blows against us here. If we were ever to succeed in determining the oldest cultural stage of the Indian race attainable to research and to dissolve away almost by means of a chemical process all inluences of the Brahmanism that is gradually developing […] we would ind conditions before us only a little diferent from those described by Tacitus as unique to the ancient Germans. But even in its contemporary ruined form the Mahābhārata often delivers us the best commentary on Germania. Here we read of the passion for gambling of the Germans, of how they wagered possessions and property, wife and child, [and] inally even themselves.” And compare also Moriz Winternitz, Geschichte der indischen Litteratur, vol. 3 (Leipzig: C. F. Amelangs, 1920), 1–2 (italics added): “The ancient heroic poetry that we saw in the Mahā bhā rata and the Rā mā yaṇa was also doubtless originally ‘courtly’ poetry. The bards (sū tas), who were the transmitters of this ancient heroic poetry, lived at the princes’ courts where they praised them in song. But they also accompanied them into battle to hymn the warriors’ heroic deeds as they saw them irsthand. These court bards were closer to the warriors than the learned Brahmans. They were often the warriors’ charioteers in their military campaigns and participated in the warrior lifestyle. It was still a rough age as these heroic poems arose, an age of battle and crude customs where the hunt, dicing and contests were among the princes’ favored occupations. And only during revelries, festivals, victory celebrations and sacriicial feasts did the bards recite their ballads. Only later, as court life became more reined did ever more learned poets take the bards’ place, poets who had been raised in the Brahmans’ schools and competed with them in learning.” 62 Winternitz, Geschichte der indischen Litteratur, vol. 1, 261–64 ofers the best testimony for how a narrow, anachronistic understanding of “epic” limited Western scholars’ appreci- ation of the Mahā bhā rata. He writes:  “Alongside this itihā sa literature […] there must 154 154 PHILOLOGY AND CRITICISM also have existed genuine epic poems, heroic poems [Heldengesänge], and perhaps also cycles of epic ballads in the ancient period. After all, the two epics that survived in our times, the Mahābhārata and the Rāmāyaṇa, are only the sediment of a long period of epic compo- sition that preceded them. Long before these two epics existed as such, bards must have sung songs about the great battle of nations  [Völkerschlacht] that forms the subject of the Mahā bhā rata and about the deeds of Rā ma, the hero of the Rā mā yaṇa. Moreover, it is inconceivable that the battles of the Kauravas and Pā ṇḍavas and the exploits of Rā ma con- stituted the sole subjects of poetry. Doubtless, many other heroes and great events in some of the other dynastic lines were also hymned. Not all of this ancient heroic poetry, whose existence we must assume, disappeared without a trace. Some of it survived in our two epics in ruins and fragments. […] The epic poems must have arisen in the circles of these bards who, at any rate, must have been very close to the warrior caste [Kriegerstande]. However, what we now know of as the national epics of the Indians—the Mahābhārata and the Rāmāyaṇa— are not the old heroic ballads that the royal bards and traveling minstrels of ancient India sang, ballads that great poets or, at least, skilled, poetically talented compilers reworked into uniform compositions. Rather, they are collections of poems of disparate contents and inconsistent value that arose in the course of centuries as a result of ongoing insertions and changes. Naturally, old heroic ballads form the core of both works, but in large measure the more religious itihā sa literature was incorporated into them and the Mahā bhā rata, in particular, has almost completely lost its character of an epic because of the insertion of extensive religious teachings. In truth, we can only speak of the Mahā bhā rata as an ‘epic’ and as a ‘poem’ to a very limited degree. Yes, in a certain sense the Mahā bhā rata is not a poetic creation at all but rather a complete literature [eine ganze Litteratur]” (emphasis in orig- inal). Note how Winternitz’s expectations completely shape his reception of the text (must have, inconceivable, doubtless, must assume, naturally, etc.). 63 See Lieve Teugels, “Textual Criticism of Late Rabbinic Midrashim: The Example of Aggadat Bereshit,” in Recent Developments in Textual Criticism:  New Testament, Other Early Christian and Jewish Literature, ed. Wim Weren and Dietrich-Alex Koch (Assen: Van Gorcum, 2003), 209– 10: “The traditional paradigms of textual criticism of classical and biblical literature do not apply to most rabbinic works. First, textual criticism works best on authored works. Second, in textual criticism as it is usually conceived, diferent sorts of (mechanical) errors are used to determine the relation between the witnesses and consequently to make a stemma that should eventually lead to the reconstruction of a text as close as possible to the original. Both conditions are very questionable as to many rabbinic texts. As to the irst point, almost all rabbinic works are collective enterprises, accumulations of traditional materials, compiled, arranged and edited by subsequent authors-editors. As to the second, the co-existence of oral and written traditions in rabbinic literature opposes the idea of an ‘original text,’ that is, where it regards early and classical rabbinic works.” Although Teugels is speaking of the rab- binic tradition, which, in the Indologists’ minds, ofers the closest parallel to the Brahmanic tradition, his criticisms also apply to their hypothetical oral, bardic tradition. The Indologists cannot simultaneously claim that they do textual criticism and seek an oral epic. In truth, while both conlation and expansion occurred, the Mahā bhā rata tradition was extremely conservative, as Sukthankar records: “Taking away something from the received text of the Mahā bhā rata and passing it of as the original work is a thing categorically diferent from adding something to it. To add small details here and there, embellishing and amplifying the original, would be merely a gentle and lowly service ad majorem gloriam dei. Even long pieces may sometimes be added, if they are actually found in other Mahā bhā rata manuscripts; and occasionally, even if they are not found in the current manuscripts, provided there is at least oral tradition to support their claims.” Sukthankar, “Prolegomena,” lii. No evidence exists that the Brahmans, at any time, removed anything once it had acquired scriptural sanction. The idea that the Brahmans savagely and vindictively removed from the text everything reminiscent 15 RECONSTRUCTING THE SOURCE OF CONTAMINATION 155 of former Kṣatriya supremacy is a fantasy emerging from the German scholars’ Protestant anticlericalism. See the next but one note. 64 Contamination from an oral source is required to posit an oral epic that is the source of that contamination. Once the existence of this source is granted, a Brahmanic redaction is required to explain why we ind no evidence for it (beyond the passages asserted to be remnants of it). If contamination from an oral source cannot be demonstrated, we have reason neither to assume an oral epic nor its destruction at the Brahmans’ hands. 65 A reconstruction of the thesis’s origins is beyond the scope of the present book. For complete sources and translations, see Vishwa Adluri and Joydeep Bagchee, The Nay Science: A History of German Indology (New York: Oxford University Press, 2014), chapters 1–2. We mention some key names here, however. The irst person to assert that the Brahmans’ aim was “to turn the heroic book [Heldenbuch] into a tool of indoctrination about the gods, laws and duties” was Christian Lassen. Lassen added, “It is in the nature of their [social] position and can also be deduced from various provisions of their legal code that the Indian priestly caste had a clear awareness of the means suitable for securing it lasting dominance over the minds of the remaining folk. It therefore cannot surprise us that we ind epic poetry being used as a tool of an addiction to spiritual domination. […] There was no more suitable means of raising the warrior caste in the priestly spirit than linking instruction about religious and social laws to the narratives that already enjoyed general approval. That such an intention was realized in the diaskeusis of the Mahâbhârata is clear to me.” Christian Lassen, “Beiträge zur Kunde des indischen Alterthums aus dem Mahâbhârata I: Allgemeines über das Mahābhārata,” Zeitschrift für die Kunde des Morgenlandes 1 (1837): 83 and 86. Thereafter, Goldstücker, reviewing Lassen’s Indische Alterthumskunde, 4 vols. (Bonn and Leipzig: H. B. Koenig, 1847–61), gave the thesis canonical status. “The groundwork of the poem, as mentioned before, is the great war between two rival families of the same kin; it occupies the contents of about 24,000 verses. This, however, was overlaid with episodical matter of the most heterogeneous kind. […] Nor was this merely [a] matter of accident in the sense in which such a term might vaguely be used. A record of the greatest martial event of ancient India would have emphatically been claimed as the property of the second or military caste, the Kshattriyas. […] But such an exaltation of kingly splendour and of the importance of the military caste, would as naturally threaten to depress that of the irst or Brahmanical caste. Brahmans, therefore, would endeavour to become the arrangers of the national epos; and as the keepers of the ancestral lore, as the spiritual teachers and guides, as priestly diplomatists, too, they would easily succeed in subjecting it to their censorship. […] It became thus the aim of the Brâhmanas to transform the original legend of the great war into a testimony to the superiority of their caste over that of the Kshattriyas. And this aim was efected not only by the manner in which the chief story was told, but also by adding to the narrative all such matter as would show that the position and might of a Kshattriya depends on the divine nature and favour of the Brâhmana caste.” Theodor Goldstücker, “Hindu Epic Poetry: The Mahâbhârata,” The Westminster Review, n.s., 33 (1868): 388–89. The thesis survived into the twentieth century and shaped plans for a critical edition. Winternitz, one of the collaborators on the Mahā bhā rata critical edition, wrote:  “We know that literary activity in ancient India lay most often in the hands of the priests, the Brahmans. We saw how they brahmanized the old, popular magical chants of the Atharvaveda and how they confounded the alien, indeed, antagonistic philos- ophy of the Upanisạ ds with their priestly wisdom. The more beloved and popular the heroic ballads became, the greater the Brahmans’ interest in gaining power over this epic poetry must have been. They knew how to confound this originally purely worldly poetry with their own religious compositions and their worthless theological-priestly knowledge. The result was that legends about the gods, mythological narratives of Brahmanic origin but also, to a great extent, didactic segments concerning Brahmanic philosophy and ethics and Brahmanic law entered, were incorporated into, the Mahā bhā rata. For this priestly caste, the popular epic provided a welcome means to disseminate their own teachings and thus to strengthen and to consolidate their power and 156 156 PHILOLOGY AND CRITICISM their inluence. They were the ones who inserted all the numerous sagas and legends (itihā sas) that recount the wonders of the famous seers of the primordial age, the Ṛsị s, the ancestors of the Brahmans—how through sacriice and asceticism they gained terrifying power not only over humans but also over the gods and how, when angered, they could bring down lords and nobles, indeed, even divine kings. But the Mahā bhā rata was too much a book of the people, too much the property of other social groups, especially of the warrior caste that it could really ever become a Brahmanic work or the property of some Vedic school. Indeed, it was not the Veda-knowing, learned Brahmans who participated in expanding the Mahā bhā rata—hence the strikingly poor knowledge about the true Brahmanic theology and science of sacriice—but the Purohitas, priests who, like the sū tas (bards), were in the kings’ service and thus came into closer contact with the epic poetry. This less learned priestly class was the one, too, that later served as temple priests at the famous cult centers and pilgrimage sites most often dedicated to the gods Visn ̣ or Śiva and preoccupied themselves with cultivating the local sagas associated ̣ u with such places as well as with the legends concerning the gods Visn ̣ and Śiva. As we saw, ̣ u this transpired mainly in the Purā nạ s but also in the Mahā bhā rata into which numerous local sagas and myths about Visṇ ụ and Śiva composed in the style of the Purānạ s as well as purānạ -like cosmologies, geographic lists, and genealogies found their way.” Winternitz, Geschichte der indischen Litteratur, vol. 1, 265–66 (all emphasis Winternitz’s). 66 Compare the contributions in David C.  Greetham, ed., Scholarly Editing:  A Guide to Research (New York: Modern Language Association of America, 1995). 67 As Contini observes, “The ‘best’ manuscript is never a completely ‘good’ witness: the irreg- ular quantitative distribution of its ‘goodness’ or certainty, for which only imperfect graphic representations have been suggested to date, is the foundation of the composite edition, which is obviously the symbolic projection of a tension towards a comparatively even level of cer- tainty against the axiologic discontinuity ofered by ‘reality’ when this is critically questioned.” Gianfranco Contini, “La critica testuale come studio di strutture (1971),” in Frammenti di ilologia romanza: Scritti di ecdotica e linguistica (1932–1989), vol. 1, ed. Giancarlo Breschi (Florence: Edizioni del Galluzo, 2007), 69 (Pugliatti’s translation). Identifying and sourcing “better” readings from other parts of the tradition is therefore a perfectly legitimate endeavor. One can also, where necessary, contest editorial choices. But as we saw in the previous section, the Indologists did not actually argue from readings. Rather, they merely identiied passages they considered “Kṣatriya” in origin. They neither asked whether their readings were good as readings nor how, assuming that the passages were original, they (that is, their readings) could have avoided deterioration. Neither the goodness of a reading nor its attestation from the tradition should be an argument against its potentially “Brahmanic” character. 68 As we surmised in Adluri and Bagchee, The Nay Science, 391 (italics in original): “German Indology thus sees itself not merely as learning and collecting objective knowledge as in the case of natural sciences. It is more: the German Indologist directly intervenes in history and changes it. On the one hand, Indians are to be shown to be at the mercy of the tyranny of their misshapen, lecherous, and fantastical gods. Even more urgently, they are to be shown as being subject to the tyranny of priestly authority. Texts need to be puriied of Brahmanical interpolations and metaphysical speculations. Freedom on a political and cultural level, in this idiom, can only be secured when the task of securing texts coincides with purifying them. Thus, we have seen in the third chapter the profuse enthusiasm with which German Indologists vied with each other to ind excuses to dissect the Gītā . If only Brahmanical and bhakti aspects were removed, Indians, as brethren in the world humanitarian project, could stand up as good, free Kṣatriyas, interpreted here to mean as good Prussian soldiers. Or very nearly so.” 69 As clariied in Chapter  1, Textgeschichte is inaccurately translated as “history of the text.” It refers, rather, to an a priori racial and anti-Semitic history created from the texts to conirm a Protestant view of history. 157 RECONSTRUCTING THE SOURCE OF CONTAMINATION 157 70 We already saw that the thesis of an original heroic epic originated with Christian Lassen, the father of racial anti-Semitism. In later years, the thesis played a crucial role in his ideas about the “Semitic” and the “Indo-Germanic” races. Lassen writes:  “History is evidence that the Semites lacked the harmonious balance of all psychic powers through which the Indo-Germans became preeminent. […] The Semite cannot separate the relationship of the world to man in general from the relationship of the world to his own ‘I.’ He cannot represent ideas in the mind in pure objectivity. His way of looking at things is subjective and egotistical. His poetry is lyrical [and] hence subjective. His spirit expresses its joy and its pain, its love and its hatred, its admiration and its scorn. […] Even if he expands his horizon it is only to represent his tribe as an individual over against other tribes. […] He is unsuccessful at [creating] epic because here the ‘I’ of the poet recedes before the object [and] even less at dramatic works, which demand that the poet shed his personality even more completely. [In contrast,] the Indo-Germans possess, alongside the lyrical, also the other genres of poetry. They alone produced a national drama. They alone produced the great heroic poems [Heldengedichte] that relect the great deeds of antiquity handed down in the legend in glori- ied form, that present the entire worldview of the spirit of a people [Volksgeistes] to us and are present as the result of the poetic efort of an entire people. The Semite is lacking in the material of the epic, but not the saga, which he poetically ornaments and develops but does not combine into larger cycles and, instead, preserves in his memory as primordial history.” Christian Lassen, Indische Alterthumskunde, vol. 1: Geographie und die älteste Geschichte (Bonn: H. B. Koenig, 1847), 414–15. 71 See Bigger, “The Normative Redaction of the Mahābhārata,” 20: “The assumption that there was an archetype, a single text from which all mss available have descended, is a conditio sine qua non for any critical edition. It forms the basis of the Critical Edition of the MBh. In taking the Critical Edition for what it is meant to be, we are accepting (silently or openly) the historical existence of an archetype. Let me now take one step further: if we assume that the archetype is a historical reality, what is this reality we are talking about? German philology has coined the word Endredaktion (inal redaction) for the archetype of the MBh. I prefer ‘normative redac- tion’ for the following reason: it does not suggest that this redaction was the last one that ever occurred in the history of the MBh. Rather, it is most probably one among others. What makes it so important is the fact that this redaction came to dominate the whole written tradition of the Sanskrit MBh; it became normative for this part of the transmission. Thus, ‘normative redaction’ is a itting name. We may now assume, as a working hypothesis, that the text of the Critical Edition of the MBh is a reconstruction of this normative redaction. […] So far, I have outlined a working hypothesis whose kernel—the assumption that there was an archetype— formed the basis of the Critical Edition of the MBh, even if the people working on it were not always aware of this fact.” 72 See ibid. for Bigger’s attempted redeinition of the critical edition:  “The critical edition is just a kind of screenshot, a frozen moment in a loating tradition. That this tradition kept on loating, even after the normative redaction, can be seen from the many *- and appendix- passages which are found in the mss. The normative redaction is thus the earliest text we can possibly establish with the methods of lower textual criticism, and it therefore provides the starting point for higher textual criticism, the object of which is to explore the textual history of the MBh before the normative redaction. It is, in the terminology of this volume, a stage from which we can try to reconstruct further stages before and after it, as well as the transitions between them.” 73 One solution is that the Brahmans identiied and eliminated the most prominent exemplars, the ones most likely to generate rich traditions. But how could they have known this in advance? It thus seems that the Brahmans must have eliminated all exemplars in existence without exception, for any single survivor could have become the ancestor of a rich assemblage of descendants. 158 158 PHILOLOGY AND CRITICISM 74 Veronika von Büren, “La place du manuscrit Ambr. L 99 sup. dans la transmission des Étymologies d’Isidore de Séville,” in Nuove ricerche su codici in scrittura latina dell’Ambrosiana, Atti del Convegno Milano, 6–7 ottobre 2005, ed. Mirella Ferrari and Marco Navoni (Milan: Vita e Pensiero, 2007), 43–44. 75 A good example is Winternitz. A comparison of his “philology” with Von Büren’s is instruc- tive. Unlike Von Büren, who argues from observed facts about the scribes’ practice, Winternitz uses the epic’s diversity of contents to argue for a multiplicity of authors. And whereas Von Büren seeks to reconstruct the transmission of the Etymologies, Winternitz focuses on an a priori and merely hypothetical history of development of the epic as it allegedly arose from Vedic, pre-Vedic and Brahmanic sources. His work is thus, at best, a contribution to bib- lical criticism and its methods of source criticism (Quellenkritik) rather than an objective and secular textual criticism. “Thus, even what we can call the ‘genuine epic’ [eigentliche Epos] is, as it has come down to us, assuredly not the work of a single poet. Even this ‘core’ of the Mahā bhā rata is no longer the ancient heroic poem. Rather, the latter is contained in it in highly diluted form. We saw that a horrendous quantity of the most disparate compositions grew up around this core: heroic ballads from diferent circles of legends, Brahmanic mythic and legendary compositions, ascetic poetry and doctrinal poems of all kinds comprising the simplest ethical maxims to comprehensive philosophical poems, codiied books of law and entire Purā nạs. Whoever wishes to believe, along with the orthodox Indians or with Dahlmann, that our Mahā bhā rata is the work of a single individual, must [also] assume that this man was simultaneously a great poet and a miserable bungler, a wise man and an idiot, a brilliant artist and a laughable pedant, to say nothing of the fact that this miracle worker would have to be a knower and adherent of the most contrary religious views and the most contradic- tory philosophical doctrines. Likewise, as concerns language, style and meter the Mahā bhā rata’s constituent parts do not exhibit any uniformity. We can only speak in general terms of ‘epic Sanskrit’ as the language of the popular epics. In reality, the language of the epic is older in parts, that is, closer to ancient Indian of the Vedic prose compositions, than in others. […] As concerns style, we can only say in general terms that the epic is as yet far from the so-called ‘Kā vya style,’ that is, the style of later poetry, which is characterized by an excessive use of ornaments (alamkā ras). But passages that already suggest this Kā vya style are not lacking. Along with them, we ind portions composed in the naïve style of the old Itihā sas such as are narrated in the Brā hmaṇas and Upaniṣads, whereas in numerous other passages the careless Purā ṇa style dominates. As concerns meter, the śloka that developed out of the old anuṣṭubh is the epic meter par excellence. But older and later forms exist of this śloka and all of them are present in the Mahā bhā rata. Further, our epic also contains old prose sections, whose prose is occasionally rhythmic, alternating with strophes. Likewise, both the older form, which still resembles the Vedic meter, and more recent forms of the triṣṭubh meter occur frequently in the Mahā bhā rata, though the śloka is nearly 20 times more common. Indeed, we also encounter individual examples of the artistic verse forms of classical Sanskrit poetry already in the Mahā bhā rata. Thus everything indicates that the Mahā bhā rata does not have a uniform origin but consists of older and more recent sections belonging to diferent centuries. Content and form equally conirm that some parts of the Mahā bhā rata trace back to the period of the Veda, whereas others must be contemporary with the late productions of Purā ṇic literature.” Winternitz, Geschichte der indischen Litteratur, vol. 1, 392–94 (all emphasis Winternitz’s). 76 The translators of the irst complete English translation of Isidore’s Etymologies note, “To get an idea of what a seventh-century Irish monk, or a lecturer at a cathedral school in the eleventh century, or an Italian poetry of the fourteenth century, or a lexicographer of the sixteenth century [that is to say, almost anyone except a German professor from the nineteenth century] could learn from the Etymologies, one might pick a bit of lore from each 159 RECONSTRUCTING THE SOURCE OF CONTAMINATION 159 of the twenty books in which the work has come own to us,” and then they list the following examples: • Caesar Augustus used a code in which he replaced each letter with the following letter of the alphabet, b for a, etc. (I.xxv.2). • Plato divided physics into four categories: arithmetic, geometry, music and astronomy (II. xxiv.4). • The term cymbal derives from the Greek words for “with” and “dancing,” sun and bala (III. xxii.12). • A physician needs to know the Seven Liberal Arts of Grammar, Rhetoric, Dialectic, Arithmetic, Geometry, Music, and Astronomy (IV.xiii.1–4). • In ancient times execution by sword was preferred as speedier (V.xxvii.35). • Architects use green Carystean marble to panel libraries, because the green refreshes weary eyes (VI.xi.2). • Esau had three names, meaning “red” (for the stew he made), “bloody” (for his complexion) and “hairy” (VII.vi.33–34). • Aristotle says that Zoroaster, the irst magician, composed 2 million verses (VIII.ix.1). • A soldier (miles) is so-called because once there were 1,000 (mille) in one troop (IX.iii.32). • The word for a garrulous person (garrulus) derives from the name of the constantly chattering bird the jackdaw (graculus) (X.114). • In the womb, the knees (genua) are pressed against the face, and help to form the eye sockets (genae); hence their name (XI.i.108). • The ibis purges itself by spewing water into its anus with its beak (XII.vii.33). • Because of its brightness, lightning reaches the eyes before thunder reaches the ears (XIII. viii.2). • Gaul is so named from the whiteness of its people, for “milk” in Greek is gala (XIV.iv.25). • Minerva is “Athena” in Greek; she is reputed to be the inventor of many arts because var- ious arts, and philosophy itself, consider the city of Athens their temple (XV.i.44). • Amber is not the sap of the poplar, but of pine, because when burned it smells like pine pitch (XVI.viii.6). • An altar was dedicated in Roma to Stercutus, who brought the technique of dunging (stercorare) ields to Italy (XVII.i.3). • The battering ram takes its name “ram” from its character, because it butts walls (XVIII. xi.1). • The women of Arabia and Mesopotamia wear the veil called theristrum even today as a protection from heat (XIX.xxv.6). • Wine (vinum) is so called because it replenishes the veins (vena) with blood (XX.ii.2). The Etymologies of Isidore of Seville, trans. Stephen A.  Barney, W.  J. Lewis, J.  A. Beach and Oliver Berghof (Cambridge:  Cambridge University Press, 2006), 3.  We shudder to think what an Indologist such as Winternitz would have done with the Etymologies. Most likely, he would have decided that because these things are not considered knowledge in his time, they could not be knowledge and therefore one individual could impossibly have compiled them. 77 See Jacques Fontaine, Isidore de Séville: genése et originalité de la culture hispanique au temps des Wisigoths (Turnhout: Brepols, 2000). See also the collected essays in Jacques Fontaine, Tradition et actualité chez Isidore de Séville (London: Variorum, 1988). 78 This is why in Appendix 17 we turn to the commentarial tradition. Unlike the German Indologists, the indigenous commentators were always aware of the intellectual universe within which the Mahā bhā rata locates itself. They not only took its claim to encyclopedic wisdom seriously; they also forged intertextual connections between the Mahā bhā rata and other works. The Indologists dismissed the commentators’ views as the opinion of “orthodox” (strenggläubig) Indians and claimed they were replacing it with historically accurate scholarship, but, in fact, 160 160 PHILOLOGY AND CRITICISM it was precisely the Indians who were historical and the Germans who were motivated by a religious, that is to say, iconoclastic frenzy. 79 In service of their Lutheran anti-Judaic and anti-rabbinic narratives, the German scholars not only disregarded the work itself; they also disregarded everything we know about the Indian tra- dition. Contrary to their claim about a separate heroic and/or bardic tradition alongside the Brahmanic tradition, literary transmission in India has always been in the Brahmans’ hands. Moreover, contrary to their charge of reckless innovation, the Brahmanic tradition has always been remarkably conservative. It represents a single unbroken tradition of exegesis and com- mentary. The Indologists’ insistence on an “epic” tradition in contradistinction to the Vedic is especially puzzling when one recalls that the Mahā bhā rata’s self-proclaimed title, itihāsa purāṇa, is a Vedic genre, occurring for the irst time in Atharvaveda 15.6.11. The Brā hmaṇas pro- vide for the narration of itihāsa and purāṇa narratives between certain rites (see, for example, Ś atapatha Brā hmaṇa 11.5.6–7 and 14.5.4 and 10). Chā ndogya Upaniṣad 7.1.2, 7.1.4, 7.2.1 and 7.7.1 and Bṛhadā raṇyaka Upaniṣad 2.4.10 also refer to itihāsapurāṇa. Chā ndogya Upaniṣad 7.1.2 speciically adds that the itihāsapurāṇa is pañcama veda, the ifth Veda. 16  162 newgenrtpdf The irst few folios of the oldest extant manuscript of the Ādiparvan (formerly in the collection of the Rajaguru Pandit Hemraj and now in the National Archives Kathmandu; accession number NAK 5/356) Source: Reproductions courtesy of the Staatsbibliothek, Berlin (A 28–5 in the digital repository of the Staatsbibliothek) 163 Chapter Three CONFUSIONS REGARDING CLASSIFICATION The editorial formula of adopting the base ms. unless there are manifest errors actually leads to a reconstructive edition, but without the application to this reconstruction of a method capable of dealing with the dynamics of variants and accounting for them in the edition.1 —Lino Leonardi, “Il testo come ipotesi” Classiication: Typological and Genealogical Many criticisms of the Mahābhārata critical edition arise from a misunderstanding about how editors clas- sify manuscripts, how they determine iliation and how they reconstruct—via eliminating either the readings of witnesses or entire witnesses—the archetype. Crucially, arguments for diferent classes of manuscripts (or entirely new “recensions”) are based on the erroneous assumption that because a group of manuscripts lacks certain passages characteristic of other manuscripts, the former constitute a family, even though the dichotomy is merely typological, not genealogical, and as yet no iliation has been established between the manuscripts so separated. The problem is not simply that, misled by their conidence in their ability to recover an “original,” German Mahābhā rata scholars extended textual criticism beyond its legit- imate domain. Rather, they had not understood its basic principles, namely, that the editor must start with the manuscript evidence and cannot make inferences beyond what the evidence warrants. In their eagerness to participate in the “critical” undertaking now under way—a reconstruction of the Mahā bhā rata that would refute Indian views of the text as an encyclopedic work of ethics, law and philosophy2—they neglected to familiarize themselves with the basic steps of textual criticism. Indeed, as we see in this chapter, they failed to grasp even “the fundamental principle of textual criticism,”3 the principle by which iliation is determined. Recall that the crux of Andreas Bigger’s argument for the critical edition as a “norma- tive redaction” was his thesis that “Ñ4 […] could be related with Ś1 and K0–3 and con- stitute a recension together with them.” In that case, he argued, “there is a possibility that the textual passages mentioned earlier [App. 1, nos. 63, 75, 80, 81 and 85] were originally contained in the normative redaction,” but “lost in this recension.”4 Bigger concluded that this “would overturn the previous maxim ‘short equals old,’ ” but it would permit us to “come closer to the normative redaction’s actual form.”5 Bigger’s reason for hypoth- esizing that Ñ4 constituted a “recension” with Ś1 and K0–3 was that “all these passages are not attested in the manuscripts Ś1, K0–3 and Ñ4.”6 He conceded that Ñ4 “appears not to belong to this group on the basis of the Schriftartenprämisse [sic].” However, he 164 164 PHILOLOGY AND CRITICISM continued, “this [principle] has lost its unimpeachability with Grünendahl’s discussion”7 (the reference is to Reinhold Grünendahl’s “Zur Klassiizierung von Mahā bhā rata- Handschriften”), and from this he concluded, as we saw earlier, that Ñ4 could be related with Ś1 and K0–3. Bigger fails to realize that this dichotomy is “merely typological, not genealogical.”8 As yet, he has only sorted the manuscripts into two groups—an interpo- lated and a non-interpolated group—but “in the genealogical sense he has not yet shown that the uninterpolated manuscripts form an ordo, classis, or familia.”9 He has not shown that the manuscripts that do not feature the interpolation are related through a relation- ship of descent. Further, he does not realize that the absence of the same passages from two or more manuscripts is not evidence of their closer relation,10 since this can only be established by considering their shared corruptions rather than readings.11 Jacques Froger writes: The method known as the “common error method” reconstructs the genealogy of the manuscripts by taking as the sign of a common origin, not just the common characteristics (internal or external), nor the common readings (good or bad), but only the erroneous readings. Its governing principle is thus, “common errors entail a common origin.”12 For Bigger to fulill his claim that Ś1, K0–3 and Ñ4 constitute a recension, he must show they have common errors, which permits us to infer one or more, more or less mediate ancestors, but he does not address this at all. Rather, his sole argument for hypothesizing that the six manuscripts have a common ancestor more recent than either N or θ is, as he says, that certain passages are “not attested” in all of them,13 even though this criterion neither permits him to establish iliation nor—once ili- ation has been established on the basis of shared corruptions—is it of any use in identifying “recensions.” Indeed, the term itself is used incorrectly, since, strictly speaking, what he has is not a recension but a class of manuscripts identiied on the basis of not featuring a set of passages found in the remaining manuscripts. In other words, he has committed the error of calling “β everything that in reality is merely ‘non-α.’ ”14 As this problem goes to the heart of the attempt to “place the classiication of the [Mahā bhā rata] manuscript tradition on new foundations,”15 we focus at length on it in this chapter. We irst show why the presence of shared corruptions is the “only truly probative evidence”16 for iliation, before discussing how, using this principle, an editor eliminates either individual readings or entire manuscripts to reach the archetype. The remainder of this chapter deals with several confusions regarding classiication found in the work of Grünendahl.17 Determining Filiation Filiation can be established only through shared errors, which permit us to identify two manuscripts as more closely related than others of that family (all of which will contain the same text, but not the same errors, which are unique to this branch of the tradition). It is hence incorrect to establish iliation on the basis of shared readings, as they identify the two manuscripts only as members of the family chosen for study (manuscripts 165 CONFUSIONS REGARDING CLASSIFICATION 165 of the Mahābhārata), but do not permit us to deine them as a speciic branch of that family—manuscripts descended from the irst source of the error or errors. Even though the principle that iliation can only be established using shared innovations is the axiomatic foundation of textual criticism,18 it is frequently misunderstood in prac- tice.19 In the irst chapter, we noted that Lachmann’s method is a mechanical process in that it relies on “simple rules and calculations of probability”20 to determine the relations of iliation between manuscripts. As Peter F. Dembowski notes, In the case of a work extant in several manuscripts, the editor strove to establish a rigorous classiication. Ideally, this classement des manuscrits would disclose the generic relationships between the manuscripts as well as their relationships to the hypothetical original (O). Lachmannian classiication was based on the crucial idea of common error, so much so that the method has often (and rightly) been called ‘the common error method.’ Simply put, this method assumes that two or more manuscripts will not contain the same error unless they descend from the same ‘ancestor.’ The manuscripts that share the common error are said to constitute the same family.21 The basic idea behind the method is the “assumption that the errors are, by deinition, corruptions of a ‘good’ ancestor—that is to say, manuscript iliation became apparent.”22 The notion that manuscript iliation becomes apparent when one looks at the inherited errors of manuscripts—these “fossils” of the process of textual transcription, as Paul Maas calls them—underpins the critic’s practice. Since no scribe would intentionally copy errors from a source text (although he will copy passages from it), if his manuscript contains all of the former’s errors, plus at least one more unique error, then we can infer that the irst manuscript must have been his source.23 Crucially, this means iliation cannot be demon- strated directly from the fact that the manuscript also contains these passages found in a second, for the same passages could also have occurred in a third, or a fourth and so on. Rather, to establish iliation, we must demonstrate that this manuscript could not contain these errors unless it had this source (for the errors, especially if they are signiicant, are likely unique to this source). Let us consider some examples. In the section titled “Leitfehler” (errores signiicativi) of his Textkritik, Maas noted: “The dependence of a witness from another cannot, as a rule, be directly demonstrated but only via ruling out independence.”24 Of the two types of errores signiicativi he listed, the so-called Trennfehler (errores separativi) acquire especial signiicance, since a witness’s inde- pendence against another can only be established if the irst (A) contains an error not found in the second (B) and this error could not have been corrected through conjecture in the period that separates the irst from the second.25 However, this only establishes B’s independence against A, since A could be descended from B. Consequently, to establish mutual independence, we must ind a corresponding separative error in B. In contrast, the circumstance that A contains errors not found in B does not justify the conclusion that these errors existed in a hypothetical source text α. To establish this, we require a second witness B that also includes them. If we can demonstrate that it is unlikely that A and B arrived at the errors independently—that is, that the errors are Bindefehler or errores 16 166 PHILOLOGY AND CRITICISM conjunctivi—and neither is A descended from B nor B from A (by establishing separative errors in both), we can assume that A and B inherited them from a common ancestor α. These relationships can be expressed graphically as follows: (a) A (b) B (c) α (lost archetype) | | /\ B A A B Type (a) can be excluded only if we ind a separative error in A against B, while type (b) can be excluded only if we ind a separative error in B against A. If we ind a sepa- rative error in A against B and one in B against A and at least one conjunctive error in both, then and only then is type (c) demonstrated. These diagrams illustrate how pre- cisely mathematical textual criticism is. Although inferential, the method is rigorous, permitting not just any inference but only valid ones—inferences that must be true because all other alternatives can be excluded.26 Why can a common text or common readings not establish iliation? The answer can be expressed in terms of the conlicting aims of the two operations or, rather, of the redundancy of the principle of common readings after its irst application: Froger seems to have been translating “identity of reading implies identity of origin,” but he was wrong to call it an “illogisme,” because witnesses that share true readings owe them to the original just as much as witnesses that share errors owe them to a corrupt descendant of the original. Not only, in fact, is the principle sound, but it does an important piece of work not done by the narrower principle about “la communauté des fautes”: it identiies the family whose members require classiication, the family of witnesses to a particular text. […] The principle about shared errors is merely an adaptation of the broader principle to the easily overlooked and seldom controversial fact that one family has already been assembled:  the family selected for study. Why is it a fallacy to apply the broader principle within that family? Textual critics usually reply that original readings may survive anywhere in a tradition, which is true but not obviously to the point. It is a fallacy precisely because the principle has already been used for deining that family. Alternatively, it is not a fallacy as long as one remembers that the families now to be deined are sections of that family.27 In our case, this means after we apply the principle the irst time to establish that Ś1, K1–6, Ñ1–4 and so forth are members of the family of Mahā bhā rata manuscripts, we cannot apply it again to deine more speciic families. For this, we require the speciic readings that originated in just this branch of the family, which, from the perspective of the archetype, can only be corruptions of the original or errors. Thus, we would have to examine innovations that arose in speciic branches of the tradition (for example, in Ś or K or Ñ) and, from these, we might conclude that Ś and K are related through a more recent ancestor (ν) or that Ś, K and Ñ are related through a more remote ancestor (N).28 The circumstance that some manuscripts in these groups do not share certain features does not prove them related,29 because to the extent that these features are innovations in other branches of the tradition, they say nothing about Ś, K and Ñ other than that they are conservative and to the extent that these features are innovations in these 167 CONFUSIONS REGARDING CLASSIFICATION 167 branches—hypothetically accepting Bigger’s theory that Ś1, K0–3 and Ñ4 dropped the interpolation—they could have occurred independently. Eliminating Witnesses A comparison of the constituted text with variant readings or passages rejected from it will not permit us to establish an earlier stage of the transmission. At most, we might establish that a reading rejected by the editor is preferable and restore it in our text, but this text will still be, by deinition, the archetype and not a supra- archetype of the kind critics think they can produce by randomly selecting passages from the appendices (on the grounds of their presumed “Kṣatriya” origins). Even if we accept Bigger’s argument that Ś1, K0–3 and Ñ4 together constitute a recen- sion, it is not clear what this argument achieves. For, in any case, as the editor proceeds up the stemma, eliminating witnesses or individual readings and unifying the branches of the tradition, he would have to make a decision between the reading of this branch (which does not contain the appendix passages 63, 75, 80, 81 and 85)  and the other branch (which does contain them). Bigger would presumably opt for the inlationary tradition’s reading, since he argues that “there is a possibility that the textual passages mentioned earlier could have been lost in this recension,”30 but this argument sufers from petitio principii. The real question is: how is the editor to adjudicate between two sets of readings, given that he does not know which of them is correct? The critical edition eliminates these passages as interpolations from the constituted text on the presumption that scribes added material but never consciously omitted any- thing. Bigger’s argument relies not only on restoring material from the critical apparatus to the constituted text but also on comparisons between this material and the consti- tuted text to make inferences about the heroic oral epic. This is misguided because while we can compare the readings of one or more manuscripts and use this to challenge the editor’s inferences, no comparison of the constituted text with one of the witnesses used to constitute it will return us to an earlier state of the tradition. It is easy to under- stand why the inference is erroneous: if a comparison of the archetype θ with one of its witnesses led us to a supra-archetype θ′, then the comparison of the supra-archetype θ′ with one of its witnesses or with the archetype θ itself, which in this case acts as one of the witnesses for θ′, would lead us to a supra-supra-archetype θ′′, and so on. The principle that one cannot consider the archetype as a text on par with its witnesses is not new: it is a special instance of eliminatio codicum descriptorum. In the most concise formulation of the principle by Maas it states “that a witness is worthless (that is, worthless as a witness), when it derives purely from an extant source text or one that can be reconstructed without its help. Should one succeed in demonstrating this of one witness (cf. § 8), the witness must be eliminated (eliminatio codicum descriptorum).”31 To explain the principle, Michael D.  Reeve ofers a simpliied bipartite stemma αCH with one subarchetype γ. In this simpliied stemma (Figure 31), CEH have “no authority against α.” However, “if […] someone were to ask […] ‘can any of CEH be eliminated?’, the answer would be not ‘yes, all of them, because they all descend from α,’ but ‘no’ or ‘only if we have α’ or ‘it depends on the readings.’ ”32 To illustrate these four instances, 168 168 PHILOLOGY AND CRITICISM α α γ γ C E H C E H appellatur nominatur vocatur appellatur vocatur vocatur α α γ γ C E H C E H appellatur appellatur vocatur appellatur appellatur appellatur Figure 31 Eliminatio Source: Reproduced from Reeve, “Reconstructing Archetypes: A New Proposal and an Old Fallacy,” 328. Reeve proposes four stemmata: in the irst, C has the reading appellatur, and E and H the readings nominatur and vocatur, respectively; in the second, C reads appellatur and E and H both read vocatur; in the third, C and E read appellatur, but H reads vocatur; and in the fourth, CEH all read appellatur. Of these four, in the irst “there is no purely stemmatic way of determining what reading γ or α had” (we have three distinct readings, any of which may be correct), with the consequence that “no elimination is possible, whether of a reading or a witness.” In the second, “γ plainly had vocatur, so that elimination is possible, whether of EH or of their reading vocatur (vocatur—that is, as the reading of EH, not as a reading alto- gether).” In other words, if both branches of the bipartite stemma EH have the identical reading, the reading of their source text can be considered secure (vocatur). In this case, γ is secured and EH can be eliminated.33 However, this does not mean we can eliminate vocatur as a reading, since we now have a bipartite stemma Cγ, where one branch has the reading appellatur and the other vocatur. As hyparchetypes, it is impossible to decide between the two on stemmatic grounds. Once one grasps the principle of eliminatio codicum descriptorum, it is easy to follow the remaining diagrams. In the third of Reeve’s examples, α and γ “plainly had appellatur,” “so that elimination is again possible, not just of H or its reading vocatur but also of CγE or their reading vocatur.” The agreement between C and E tells us that the original reading in α must have been appellatur and that this reading descended via γ to E but not to H, which contains the erroneous reading vocatur. Hence, both the witness H and its reading vocatur can be eliminated (eliminatio lectionum singularium). But since the text of α also now counts as secure (its reading must be appellatur), CγE can be eliminated. Finally, in the fourth and last example, “α and γ even more plainly had appellatur,”34 so we can eliminate both the manuscripts CγEH and their reading appellatur. 169 CONFUSIONS REGARDING CLASSIFICATION 169 Ascent up a stemma thus often, though not always, lets us eliminate witnesses. The point is not so much to establish the conditions under which eliminatio codicum descriptorum is possible as to show that the comparison of a reconstructed archetype with its elimi- nated witnesses is not going to tell us more about that archetype. It certainly is not going to permit us to go beyond the archetype. Descent back down a stemma once we have reconstructed the archetype is not going to tell us more about it, and if we compare the reading of α in Reeve’s third example with H, we will not be any the wiser for it. It would be erroneous to claim that vocatur was the better reading all along, but this is exactly what Bigger’s argument (and by extension the arguments of those who claim to ind evidence for a heroic Kṣatriya epic via comparing the constituted text with its critical apparatus) amounts to. In the remainder of this chapter, we look at how these two errors—a confu- sion about iliation and a confusion about eliminatio—have led to calls to place the critical edition “on new foundations” in the work of Grünendahl.35 The Argument from Brevity and the Argument from False Premises In creating the critical edition, Sukthankar followed the evidence of the shortest witness, the Sā́ radā codex, and reprinted the readings of this manuscript as his constituted text. However, as the Sā́ radā codex was not the shortest witness (for the Nepālı̄ manuscript Ñ4 is shorter by 63 passages), the critical edition must be begun anew. Refutation of the argument: This conclusion would only be justiied if the editor actually reprinted the readings of the Śāradā codex as his constituted text. In point of fact, he compared manuscripts of diferent classes to reconstruct the reading of the archetype. He showed in several cases that the Śāradā codex itself contained interpolations and additions. This means that what is decisive is not the length of the Śāradā codex over the Nepālı̄ manuscript NAK 5/356, but whether passages in the constituted text exist, which are not attested in the latter, something Grünendahl (who raises the objection) has not shown. Grünendahl’s irst line of attack on the stemma comprises a pair of related arguments that we call the argument from brevity and the argument from false premises. It is not so easy to state the irst, for it sufers from the fallacy of complex question. At its simplest, the argument from brevity can be stated as the view that the editor ought to base his recon- struction on the shortest witness. Grünendahl’s formulation of the argument is more complex. On one hand, he implies that this principle was adopted for the Mahā bhā rata critical edition. On the other, he argues that it was incorrectly applied, for the witness thought the shortest (Ś1) proved not to be the shortest. The argument from brevity thus entails two separate claims, the irst of which rests on a misapprehension regarding tex- tual criticism (reconstruction should be based on stemmatic logic), and the second of which is demonstrably false (the critical edition is not based on the shortest manuscript). Consequently, Grünendahl’s third claim (that the principle was incorrectly applied in the critical edition) is also false. In fact, the only part of his argument that is true is that Ś1 is not the shortest known witness, but as Sukthankar was the irst to point this out in 1939, it is unclear what Grünendahl hopes to achieve.36 For the sake of clarity, in the following we reserve the designation “the argument from brevity” for the irst part of Grünendahl’s tripartite argument, the view that the editor ought to base his reconstruction on the shortest witness. In contrast, we desig- nate the second and third parts—the assumption that this was the principle adopted 170 170 PHILOLOGY AND CRITICISM for the Mahā bhā rata critical edition and the assumption that this principle was incor- rectly applied in the critical edition—as the “argument from false premises” (if the prin- ciple was not adopted, the question of its correct or incorrect application is moot; hence the third part also rests on false premises). It is important to disambiguate between the diferent stages of the argument for, as we see in what follows, the persuasiveness of Grünendahl’s case rests on his confusion of the two. His overarching conclusion that the Mahā bhā rata critical edition must be begun anew using the Newā rī manuscript Ñ4 as the base appears persuasive to him only because of his misperception that the constituted text is based on the shortest witness. Had he understood that the constituted text is based on stemmatic logic, he would have seen that the critical edition already controlled for the issues he raises. The discovery of a shorter manuscript will not signiicantly alter the critical edition because the constituted text is already shorter than its shortest witnesses. Precisely because it is shorter than any known witnesses, it permits us to identify the expansions in these texts. The argument from brevity irst occurs on page  107 of Grünendahl’s article. He states: “among the manuscripts used [for the critical edition], the manuscript S1́ takes on a preeminent role for Sukthankar.”37 He writes that “he [Sukthankar] justiied this less in terms of its quality than in terms of its textual extent”38 and as evidence presents Table 1. Initially, it is not clear what this table is supposed to demonstrate. Grünendahl explains that in the irst row he lists what he calls the “pure textual additions” of the northern recension,39 an expression he clariies as follows: “By pure textual additions I mean in the following those star and appendix passages, which represent a factual increase of the text.” He glosses this peculiar formulation as follows: “[passages] which thus do not merely ofer another line or another verse in place of a line or a verse of the constituted text or insert parts of the constituted text in another position.”40 According to his calculations, the northern recension “exhibits 1021 pure textual additions,”41 and he further breaks them down into Table 1 Grünendahl’s list of interpolations in the northern recension Ñ1 Ñ2 Ñ3 Ñ4 Ś1 K0 K1 K2 K3 K4 total 316 422 267 87 (?) 250 150 201 277 518 Ś1 available 237 317 210 75 105 178 112 149 193 378 V1 B1 B2 B3 B4 B5 B6 379 380 318 408 387 408 404 270 282 225 303 286 299 300 Da1 Da2 Dn1 Dn2 Dn3 D1 D2 D3 D4 D5 D6 D7 416 393 492 494 501 446 441 430 573 446 348 335 309 291 356 359 364 324 330 253 399 360 232 231 Source: Reproduced from Grünendahl, “Zur Klassiizerung von Mahābhārata-Handschriften,” 108–9. 17 CONFUSIONS REGARDING CLASSIFICATION 171 “957 star passages listed in the footnotes and 64 textual passages listed in appendix I.”42 Already, it is not clear whether “pure” is meant to qualify “textual additions” or some other expression of which it is a component, for surely it is one thing to speak of “pure textual additions” and another to speak of “pure textual additions of manuscript x.”43 The latter, if it is to have a meaning at all, must refer to the net expansion of manuscript x—that is, the pure textual additions in the irst sense minus whatever the manuscript omits of the constituted text, but Grünendahl does not consider this at all. He takes the number of additions as representative of the text’s size, but this is to assume that all our manuscripts reproduce perfectly the text of the archetype, which is a gross oversimpliication.44 He also treats all additions as identical for the purposes of calculation even though the shortest star passages are single-line insertions, whereas the longest appendix passage is 462 lines (App. I, 114, inserted on a separate folio in D4, the only northern manuscript to feature it). Clearly, then, treating all additions as representing an equivalent “factual increase in the size of the text” is misleading and will produce erroneous results. We do not even need to know the length of every passage to show this. Let us grant the view that every addition counts as identical. We may assume a standard unit for this, such as the number 1 (each addition represents an insertion of one line). In Grünendahl’s table, an expansion of 430 lines in D3 stands opposed to an expansion of 573 lines in D4, which is the same as saying that D4 is longer than D3 by 143 lines. We know of at least one expansion (App. 1, 114) that it is not an insertion of one line. Substituting the true igures for this passage we arrive at a igure of 1,034 for the expansion of D4 (573 – 1 + 462) and the diference between D4 and D3 is now not 143 lines, but 604 lines. Of course, if we were to carry out the calculation consistently, D3 may have longer insertions where D4 has shorter ones so that the igures for the two might once again approach each other. But this already shows that treating additions as identical for the purposes of cal- culation is a careless way of doing things: it misleads us not only about the true extent of the witness texts but also about the true extent of interpolation. In Grünendahl’s table, a less interpolated manuscript could easily appear more interpolated.45 The principle that the shorter reading is preferable (brevior lectio potior) is well known in textual criticism, but it cannot be applied dogmatically. Whereas additional sentences and phrases in manuscripts are as likely the additions of copyists as the author’s work, this does not always hold. The additional passages may go back to a second edition of the author’s work.46 In any case, as Martin L. West reminds us, “The quality of a man- uscript can only be established by reading it. And when an opinion has been formed on the quality of a manuscript, it can be used as a criterion only when other criteria give no clear answer. […] Each variant must be judged on its merits before the balance can be drawn and a collective verdict passed.”47 It is absurd to treat the shortest witness as unfailingly accurate in each case. It may have additions of its own, as Sukthankar demonstrated of the Ś ā radā codex.48 Further, in asserting the argument from brevity, Grünendahl overlooks a crucial point. The principle brevior lectio potior is part of the canon of principles of the textual critic, but it is not the whole of that canon. It cannot replace the need for recensio. Only the latter will permit us conidence in our judgments, because we are now looking at the witness, even the shortest, in the context of the entire tradition. We are not simply deducing from its relative shortness that it has the correct text but using its 172 172 PHILOLOGY AND CRITICISM length as one of the criteria in that comparison to assess its relative merit.49 Grünendahl goes astray because he focuses exclusively on the question of “interpolations” and poses this question in purely quantitative terms.50 But, in fact, it is not what is correctly transmitted (treating interpolations as part of the text transmitted from the ancestor to the witnesses) but the unique errors (errores signiicativi) each scribe introduces that permit us to deduce the interrelationship between manuscripts. Reeve notes that “there are still people who build stemmata not just on insigniicant errors but on agreement in the truth,”51 but Grünendahl’s use of interpolations to estab- lish families is even more problematic. He has yet to show that basing the reconstruction on Ñ4 rather than Ś1 would lead to a signiicant alteration of the archetype. Doubtless he thinks it would, but this is because he applies a false standard. According to him, Sukthankar premised his reconstruction on the Ś ā radā codex and he takes this princi- pally, if not solely, to mean that Sukthankar looked to this text when deciding whether to include or exclude interpolations. According to his calculations, Ś1 has a surplus of 63 “pure textual additions” over Ñ4, which fact he uses to insinuate that if the reconstruc- tion had been based on Ñ4, the archetype would have been shorter by this amount. We have already seen that the calculations are based on poor math: if all or even a signii- cant portion of Ñ4’s insertions were longer than those of Ś1, the conclusion would not hold. But the more serious problem concerns Grünendahl’s misguided belief that he has shown something signiicant by showing that Ñ4 is shorter than Ś1 by a large amount, namely, that this diference implies a similar reduction in the length of the archetype. A look at the second part of his argument is revealing. After introducing the table, he writes: “Because of the fragmentary nature of Ś1, its total number of interpolations cannot be identiied: 329 additions of the northern recension correspond to missing parts of this manuscript.”52 To circumvent the problem, he introduces the second row of his table. This row, he says, only takes into consideration the 692 textual additions that corre- spond to the extant parts of Ś1. He inds that 105 of these interpolations also occur in Ś1 and, of these 105, 14 interpolations “ ‘do not also occur in all versions’ [of the northern recension!].”53 “However, it is remarkable that except for Ś1, all 14 additions are found only in K1. This conirms not only the close connection between Ś1 and K1 as noted by Sukthankar […] but also, I hold, justiies us in the assumption that the cell containing a question mark in the list above, namely, the total number of textual additions in Ś1— including those in the missing parts of the manuscript—is probably in the same range as that of K1.”54 From this, Grünendahl concludes, “even so the text available in Ś1 is signiicantly shorter than that in most of the other manuscripts.” “However,”—and this is decisive for him—“the title of ‘shortest known version’ unambiguously belongs to another manuscript, namely, Ñ4.” From this discussion and from a second table (which we discuss later on) it is clear what Grünendahl hopes to show from these igures. He takes the additions in witness texts as indicative of their excess length over the constituted text. He begins with the fact that the critical edition lists 1,021 additions (whether or not we call them “pure” here is irrelevant for the argument) and then reduces these to the 692 additions that he can check against the Ś ā radā codex. He then distributes these 629 additions across the 29 columns of his table (some repeatedly, since they are shared across witnesses) to arrive at 173 CONFUSIONS REGARDING CLASSIFICATION 173 the number of passages each witness has in excess of the constituted text. From the fact that Ñ4 has 87 additional passages as compared with Ś1, which has 150, he concludes that Ñ4 is shorter than Ś1. So far he is on safe ground (ignoring for the moment the problem with his identiication of number of additions with extent of addition), but he also has not said anything that was not common knowledge. The function of these inexact and misleading calculations cannot be to demonstrate Ñ4’s brevity. Rather, from Grünendahl’s perspective, they are signiicant because they imply that the diference in length between Ś1 and Ñ4 signiicantly afects the constituted text. Thus, he repeats twice, “To be sure, Sukthankar himself conceded this [that Ñ4 is the shorter manuscript] in his article, however he did not draw the consequences that would have resulted for his editorial principles from this fact.”55 “He himself [Sukthankar] later corrected his main argument, namely, that it [the Kaśmīrī version] was the shortest version, in favor of Ñ4—however, without drawing the consequences that resulted from this fact for his classiication of the manuscripts. For, when this claim can no longer be maintained, his entire demonstration built upon it is beside the point.”56 What these “consequences” are can be easily gleaned. From Grünendahl’s table, it is clear that he imagines that since Ñ4 is shorter than Ś1 by 63 insertions the constituted text must contain an excess of 63 insertions over Ñ4. But this is not true! It would only be true if in reconstructing the archetype Sukthankar had actually followed the argument from brevity. In that case, he would have followed what he thought was the shortest text—as Grünendahl claims he did, albeit incorrectly, since Ś1 is not the shortest manuscript. In reality, of course, Sukthankar’s reconstruction is based on the stemma. In other words, he has compared diferent manuscripts and manuscript classes to infer the reading of their archetype. His method is additive rather than subtractive: he compares manuscripts to see what is common to them and how readings were transmitted and corrupted over time. He thus takes the reading of the archetype—the one that explains the others— rather than taking one manuscript as his ideal and stripping away everything from the others not contained in the former. Contrary to Grünendahl’s assertions, the constituted text does not follow any one manuscript, shortest or otherwise. Grünendahl ought to have seen this: his own table shows that Ś1 has additions over the constituted text! That means what is decisive is not the diference in length of Ñ4 over Ś1 (–63 insertions), but the diference in length of Ñ4 over the constituted text. This is something Grünendahl has not calculated.57 He has not shown that Ñ4 is shorter than the consti- tuted text, which would imply that certain passages have been included in the latter even though they are not attested in one of its witnesses. He has in fact demonstrated the opposite. His table clearly shows that Ñ4 is longer by 87 insertions than the constituted text. The table also shows that the constituted text, as the archetypal text attested in independent witnesses, is shorter than all known witnesses of the tradition—as it must be if it is older than the latter and we know that the Mahā bhā rata was expanded over centuries.58 It is easy to see that the real source of these elementary errors is Grünendahl’s misplaced faith in the argument from brevity. A  simple comparison of four groups— interpolations unique to Ś1, interpolations unique to Ś1 and K1, interpolations in Ś1 and K1 with other witnesses and interpolations in N without Ś1 and K159—shows how Grünendahl imagines a reconstruction based exclusively on the principle brevior lectio 174 174 PHILOLOGY AND CRITICISM Table 2 How Grünendahl imagines the process of constitution Ś1 Ś1 and K1 Ś1 and K1 N (unique (unique (with other (without ́1 and K1) interpolations) interpolations) witnesses) *374 *1735 App. 1, no. 331 App. 1, no. 63 *2161 App. 1, no. 58 App. 1, no. 75 App. 1, no. 118 App. 1, no. 80 App. 1, no. 81 App. 1, no. 85 Note: 1. Ś1 is missing for this passage. potior might look. For the sake of simplicity, we restrict ourselves to the appendix passages, though the argument is no diferent if we include the star passages as well.60 We have four columns as shown in Table 2. Following Grünendahl’s understanding of critical reconstruction, we would include the passage *374 as it is present in the shortest text, which we have adopted as the norm for the reconstruction. We would also include the passages *1735 and *2161, since they too occur in Ś1 and the fact that they also occur in K1 is either incidental or at most ofers further conirmation for them. We would also include the three appendix passages nos. 33, 58 and 118, even though they are found in other northern recension manuscripts.61 In fact, as can be seen from the table, once we invoke the argument from brevity, we can exclude only one group of interpolations: those absent from Ś1 and K1. The problem with the argument from brevity, then, is that it does not yield a crit- ical principle for reconstruction, or, if it does, then only a very weak one, since it only permits us to exclude the interpolations of the fourth column but not those of the irst three. Basing a reconstruction on the shortest known witness is not a critical procedure, for it presumes that the shortest witness contains the most authoritative text. This, how- ever, need not be the case for several reasons.62 Even if we consider the shortest version more authoritative than the others, this applies only to its total length and not to its indi- vidual readings. Grünendahl goes astray because he presumes that the shortest version’s readings have not sufered any degradation. From his perspective, it suices for the editor to reprint this witness’s text (perhaps correcting obvious errors), but this is because he focuses exclusively on interpolations rather than the readings. He does not see that it is easier to notice and control for interpolations, especially the obvious interpolations found in the Mahā bhā rata tradition, than it is to identify the changes to the reading of the arche- type. From the nonspecialist’s perspective, the former appear far more signiicant. They leap immediately to his eye, whereas to notice the presence of a likely hiatus that troubled later copyists and led to various attempts at emendation requires a greater facility with the language (as well as a thorough knowledge of the principles of textual criticism). From the textual critic’s perspective, however, the small diferences in readings accu- mulated over the course of centuries are more signiicant. They permit him to deduce the genealogical relationships between manuscripts. If two manuscripts have two large interpolations or one has a larger interpolation than the other, we cannot say anything 175 CONFUSIONS REGARDING CLASSIFICATION 175 about their relationship simply by looking at the interpolations. Even if one manuscript did not have the interpolation, we could not declare it the source of the interpolated one. If the interpolation is obvious, the scribe of the shorter manuscript may have chosen to ignore it. Alternatively, the interpolation may have been present only in a diferent branch of the tradition, so that our scribe was unaware of it. Without a stemma, we would not even be able to identify the interpolation as an interpolation: the larger man- uscript may go back to a second edition of the text made by the author or the “interpo- lation” may have been on a separate folio that was lost before our scribe made his copy. Without looking at the passage in context and assessing the balance of probabilities, we could never determine the true state of afairs. Already, Grünendahl’s exclusive reliance on interpolation or, rather, lack of interpo- lation as a criterion for choosing the best witness displays great uncertainty regarding textual criticism. The problem is further compounded when he imputes the same mis- guided view to Sukthankar and, furthermore, criticizes him for not correctly applying the principle brevior lectio potior. He argues: “Since Sukthankar argues exclusively in terms of quantity, he can really only elevate Ś1 to the norm in a quantitative respect. This would correspond to his conviction that one—manifestly in contrast to extent!—may not constitute the text on the basis of individual manuscripts, indeed not even [on the basis of] individual versions or recensions.”63 A few pages later, he repeats: “It is true of the constitutive manuscripts [sic] in general what was already noted regarding Ś1: Sukthankar reaches his decision in favor of Ś1 and the K group not on the basis of qualitative but rather on the basis of quantitative aspects.”64 And again: “ν alone appears in general not to ofer a suicient qualitative foundation for the constitution of the text.”65 Yet, as we have seen, the only person who looks at manuscripts in a purely quantita- tive respect is Grünendahl. He is the one who coins the peculiar expression “pure textual additions,” and he is the one who thinks that it is legitimate to look at additions only in terms of whether they represent a “factual increase of the text.” He is the one who sets aside the passages’ placement, whether they replace or repeat a part of the archetypal text, and whether they alter the order of its passages.66 He is the one who thinks that all insertions, regardless of their length, can be treated as the same for purposes of comparison. He is the one who creates a misleading table where the number, not extent, of interpolations is supposed to indicate a witness’s length. He is the one who does not look at readings in con- text, trying to determine whether one can be derived from the other, but advocates a recon- struction based on the shortest witness. He is the one who thinks the argument from brevity ofers an adequate theoretical foundation for a critical edition and he is the one who thinks that the discovery of a shorter witness must make a signiicant diference to the constituted text. The notion that Sukthankar reconstructed the archetype on the basis of the extent—in contrast to the content—of a single manuscript is Grünendahl’s own delusion. A closer look at Grünendahl’s article shows that he has misunderstood a key phrase in Sukthankar’s discussion of the Ś ā radā codex. Describing the advantages of this man- uscript, Sukthankar writes: While it is the shortest extant version, it is a demonstrable fact that it contains relatively little matter that is not found, at the same time, in all other versions of both recensions. It is clear, 176 176 PHILOLOGY AND CRITICISM therefore, that it must contain, relatively, less spurious matter than any other known version. That is precisely the main reason why it is taken as the norm for this edition.67 Grünendahl repeats this exact sentence, which he appends to his discussion of why the Ś ā radā manuscript was chosen as the textus simplicior. He writes: “On the basis of his esti- mate [of the length of the manuscript] he [Sukthankar] identiies S1́ as the ‘shortest known version,’ as the textus simplicior that contains relatively little ‘that is not also found in all other versions of both recensions.’ ” Thereafter, he presents his table of textual additions in the manuscripts of the northern recension and, on the basis of these calculations, arrives at the conclusion that “of these additions [Grünendahl means the 692 listed in the second row] S1́ features 105, among them 14 whole additions, about which one can say along with Sukthankar that they occur in S1́ but ‘not also in all other versions’ ” and then notes parenthetically: “of the northern recension!” and he adds an exclamation for emphasis. But while he repeats the phrase “not also in all other versions,” he completely inverts its meaning. In Sukthankar’s text, the phrase is inserted to underscore that Ś1, though signiicantly shorter than the other witnesses, is not an abridgement or a diferent branch of the tradition. From the fact that the text of Ś1 coincides to a great extent with that found in the other manuscripts, Sukthankar concludes that it must contain “relatively less spurious matter.” This is, of course, why it was taken as the norm for this edition (that is, because of its signiicant coincidence with the tradition, yet its comparatively pure state). Grünendahl repeats the statement almost verbatim and the irst time he cor- rectly understands the force of the qualiication. He notes that Ś1 was chosen as the textus simplicior on the grounds that it “contains relatively little ‘that is not also found in all other versions of both recensions.’ ” He correctly construes “not also found in all other versions of both recensions” as referring to the fact that Ś1 has few unique additions of its own. Yet, the second time, he completely misunderstands the force of the qualiica- tion: this time he applies it to the interpolations and says that we can say of them “along with Sukthankar that they occur in Ś1 but ‘not also in all other versions.’ ” On the basis of this misunderstanding, he now presents Table 3. This table shows the 14 interpolations Grünendahl identiies as found in Ś1 but “not also in all other versions.” In the irst column, he lists the 14 passages. Thereafter, he enters a “1” in the column if the passage is found in the respective manuscript, leaving the column blank if it is not. If the manuscript is incomplete and the section containing the passage is missing, he enters a “0” to indicate its absence. At the bottom, he lists the total for each manuscript. Grünendahl does not comment on the table other than to note that “they [‘[the] 14 additions, about which one can say along with Sukthankar that they occur in Ś1 but “not also in all other versions” ’] are compiled in the list on p. 111. It is remarkable thereby that, with the exception of Ś1, all 14 additions are attested only in K1.”68 From the pre- ceding it is clear what he thinks to show with this table. Because he thinks that Sukthankar followed the argument from brevity in constituting the archetype, he reads Sukthankar’s comment that the Ś ā radā codex contains “relatively little matter that is not found, at the same time, in all other versions of both recensions” as an assertion to the efect that the Ś ā radā codex’s additions are attested in all other versions. It is on this presupposition and this 17 CONFUSIONS REGARDING CLASSIFICATION 177 Table 3 Grünendahl’s table of the distribution of interpolations in the northern recension Ñ1 Ñ2 Ñ3 Ñ4 Ś1 K0 K1 K2 K3 K4 V1 B3 B5 B6 1735* 1 1 2161* 1 1 2080* 1 1 1772* 1 1 1 1 1 1 1 1 2081* 1 1 1 1 1 1 1 App. I, 121 1 1 1 1 1301* 1 1 1 1 1 0 1 1 2077* 1 1 1 1 1 1 1 1 2118* 1 1 1 1 1 1 1 1 1 969* 1 1 1 1 1 1 1528* 1 1 1 1 1 1 1 1 1 1855* 1 1 1 1 1 1 1 1 1499* 1 1 1 1 1 1 961* 1 1 1 1 1 1 1 1 Totals: 6 4 1 1 14 9 14 8 8 9 3 3 2 4 Da1 Da2 Dn1 Dn2 Dn3 D1 D2 D4 D5 1735* 2161* 2080* 1 1 1772* 1 1 1 1 1 2081* 1 1 App. I, 121 1301* 1 1 1 1 1 1 1 2077* 1 1 2118* 1 1 1 1 1 969* 1 1 1 1 1 1 1 1528* 1 1 1 1 1855* 1499* 1 961* Totals: 1 1 4 4 4 2 6 4 9 T1 T2 G1 G2 G3 G4 G5 G6 M5 1735* 2161* 0 2080* 0 1772* 2081* 0 App. I, 121 0 1301* 2077* 0 2118* 0 1 969* Table 3 (Continued) 178 178 PHILOLOGY AND CRITICISM Table 3 (Cont.) T1 T2 G1 G2 G3 G4 G5 G6 M5 1528* 1855* 1499* 961* 1 1 1 1 1 1 1 1 1 Totals: 1 1 1 2 1 1 1 1 1 Source: Modiied from Grünendahl, “Zur Klassiizerung von Mahābhārata-Handschriften,” 110. presupposition only that Sukthankar would have been justiied in constituting his text on its basis, for in that case, Ś1’s additions are additions in the tradition as a whole as well and may, therefore, legitimately be included in the constituted text. Little wonder, then, that Grünendahl thinks he has a fatal objection to Sukthankar, when he now presents a table (Table 3) demonstrating incontrovertibly that the Ś ā radā codex’s additions are not universally attested in the tradition. As he puts it, the Ś ā radā codex contains “14 whole interpolations, about which one can say along with Sukthankar that they occur in Ś1 but ‘not also in all other versions.’ ” According to him, this is a devastating objection to basing the constituted text on Ś1, for, as his table shows, Ś1 has additions that are at last partly unique to it, and this means its text (which was already shown to be not the shortest) cannot reliably be followed in constituting the archetype. Grünendahl has evidently been misled by the negation in “relatively little matter that is not found, at the same time, in all other versions of both recensions.” Sukthankar’s point was not that Ś1’s additions were found in all other manuscripts. Rather, he was emphasizing the fact that, with few exceptions (his “relatively little matter”), Ś1’s text was identical with that of all the other versions. To express this state of afairs, he chose the phrase “not found […] in all other versions.” From his perspective, the important point was that the archetypal text of Ś1 is attested in all other versions. Grünendahl, however, interprets him as saying that what is decisive is that the additions are attested in all other versions, and he constructs a complex table to show that this is not so. However, it is irrelevant for an addition whether it is found in only one manuscript or in only some. Nothing is gained for an addition’s lack of originality to demonstrate that it is not found in all of the other versions and yet, this is precisely what Grünendahl does. For the same reason, he thinks this table now casts doubt on the advisability of constituting the archetype using Ś1 as a base, since Ś1 has additions not shared with all the other manuscripts. But the table would be an objection to Sukthankar only if Grünendahl had shown that Sukthankar included certain passages in the constituted text on Ś1’s authority, even though they are not attested in other witnesses. But ex hypothesi the passages are not included in the constituted text: they are identiied as interpolations in Ś1, which is to say, they are additions to the archetype reconstructed in the critical edition. That we can identify these passages as interpolations at all is due to the fact that Sukthankar by following the stemmatic method could successfully reconstruct the text of the archetype. All Grünendahl has done with this table is take passages identiied as interpolations in Ś1 and show that they are not found in all other versions of the northern recension, and all 179 CONFUSIONS REGARDING CLASSIFICATION 179 this in order to argue for their status as interpolations! The only thing that follows from this pretentious table is that Ś1 contains some passages not found in other versions but he could have shown this more easily had he simply listed the interpolations unique to Ś1 (or, since there is only one, the interpolations unique to Ś1 and K1). His table is neither a comprehensive list of passages attested in Ś1 but not in other manuscripts nor a list of passages attested in Ś1 and other manuscripts (the former would be the unique additions in Ś1; the latter the non-unique additions in Ś1). The table merely shows that Ś1 has some passages not shared by all the other manuscripts, but as this has no further bearing on the reconstruction we can replace it with a simpler table (Table 4), showing that: 1. Ś1 is not free from unique interpolations (column 1). 2. Ś1 and K1 are not free from unique interpolations either (column 2). 3. Ś1 and K1 sometimes share interpolations with other witnesses or with the northern recension as a whole (column 3). 4. However, Ś1 and K1 are also free of some interpolations found in the northern recension as a whole, which is the reason for giving their reading greater credence (column 4). This fourth point is crucial. It is the reason for Sukthankar’s greater conidence in Ś1 and K1. Grünendahl should have attacked this point, if he wished to overturn Sukthankar’s edition, but he does not address it at all. Instead, his table is of the form “Ś1 and K1 without N as a whole.” We can append it as a ifth column to Table 4 where its redun- dancy is obvious. We have chosen to expose the laws in Grünendahl’s argument through care- fully reproducing his calculations, but he could also have looked at Sukthankar’s Table 4 Grünendahl’s error in constituting the text ́ S1 ́ and K1 S1 ́ and K1 S1 N ́ and K1 S1 (unique (unique (with other (without Ś1 (without N interpolations) interpolations) witnesses) and K1) as a whole) *374 *1735 App. 1, no. 331 App. 1, no. 63 1735* *2161 App. 1, no. 58 App. 1, no. 75 2161* App. 1, no. 118 App. 1, no. 80 2080* App. 1, no. 81 1772* App. 1, no. 85 2081* App. 1, no. 121 1301* 2077* 2118* 969* 1528* 1855* 1499* 961* Note: 1. Ś1 contains a lacuna here. 180 180 PHILOLOGY AND CRITICISM “Prolegomena.” Had he read Sukthankar’s explanation of his principles, he would have seen that his (Grünendahl’s) argument is based on false premises, above all, the erroneous assumption that the constituted text is based on the shortest manuscript. Sukthankar is aware of the problem with basing the constituted text on one manuscript. “By following any manuscript—even the oldest and the best—we shall be authenticating just that arbi- trary mixture of versions which it is the express aim of this method to avoid!”69 He explic- itly cautions against following any one manuscript uncritically: “The optimistic view that any extant manuscript, however old and trustworthy, of some favoured version or recen- sion, could give us, with a few additions and alterations, the text of Vyā sa’s Bhā rata or Mahā bhā rata is the index of a naive mentality and does not need any elaborate refuta- tion.”70 Sukthankar emphasizes that the Ś ā radā codex also cannot be followed uncritically. He notes that although “an unbiased comparative survey of the diferent versions leads one to the conclusion that the Ś ā radā (Kaśmīrī) version is certainly the best Northern version, and probably, taken as a whole, the best extant version of the Ā di,” it is “not by any means, entirely free from corruptions and interpolations.” “These must be carefully corrected and controlled with the help of the other versions, particularly of those of the rival recension.”71 He notes several such passages, proving that his reconstruction is based not on Ś ā radā readings alone but on a comparison of manuscripts of all versions.72 Sukthankar knew that stemmatic reconstruction will give him a text older than any of the existing versions. Contrary to Grünendahl’s belief that he followed Ś1’s reading uncrit- ically, thinking it the shortest, he notes the advantages of the inlated southern recension: It should thus seem that the inidelities of the Southern recension are conined mainly to a tendency to inlation and elaboration. In parts unafected by this tendency, it is likely to prove, on the whole, purer, more conservative and more archaic than even the best Northern version. The Southern variants, therefore, deserve the closest attention and most sympathetic study.73 Grünendahl’s assertion that Sukthankar relied primarily on a quantitative metric for reconstruction is a red herring, arising from his own confusion of quantitative with qual- itative criteria.74 Sukthankar’s method, as his explicit statements in the “Prolegomena” demonstrate, is stemmatic and in keeping with established canons of textual criticism. Describing the method to be followed in the section “What is then possible?,” he notes: With that end in view, we must examine as many manuscripts—and above all as many classes of manuscripts—as possible, and group them into families. We must try to ascertain and eval- uate the tradition of each family, eschewing late and worthless material. We may then consider the relation of these traditions in regard to the variae lectiones, and the genuine and spurious parts of the text. Beyond that, we have to content ourselves with selecting the readings [that are] apparently the earliest and choosing that form of the text which commends itself by its documentary probability and intrinsic merit, recording again most carefully the variants, and the additions and omissions.75 In contrast, Grünendahl does away with the stemma and instead introduces the spurious criterion of “extent” (Umfang) to determine the best manuscript. It is on this basis that he 18 CONFUSIONS REGARDING CLASSIFICATION 181 declares that the reconstruction should have been based on Ñ4. In his encomium of this manuscript’s virtues, he does not mention the quality of its readings even once. Instead, he repeatedly emphasizes that Ñ4 must be given preference because it is the shortest man- uscript. When he ignores the placement of additional passages, their intrinsic probability and the transition from the text to the interpolation, he actually does what he accuses the editor of doing: he reduces interpolation to a purely quantitative problem. Grünendahl states, “To be sure, Sukthankar himself conceded this [that Ñ4 is the shorter manuscript] in his article, however he did not draw the consequences that would have resulted for his editorial principles from this fact,”76 but he overlooks a crucial point: the fact that Sukthankar was aware that Ñ4 was shorter than Ś1 and yet did not radically alter his edition suggests that it was not based on the argument from brevity. Likewise, when he suggests, “He himself [Sukthankar] later corrected his main argument, namely, that it [the Kámīrī version] was the shortest version, in favor of Ñ4—however, without drawing the consequences that resulted from this fact for his classiication of the manuscripts. For, when this claim can no longer be maintained, his entire demonstration built upon it is beside the point,”77 he should have realized that the fact that Sukthankar did not do so is a sign that the classiication—contrary to what Grünendahl thinks—was not based on length. Grünendahl introduces the specious theory that Sukthankar thought it acceptable to constitute the text on the basis of the extent of one manuscript, in contrast to constituting it on the readings of one manuscript (cf. “This would correspond to his [Sukthankar’s] con- viction that one—manifestly in contrast to extent!—may not constitute the text on the basis of individual manuscripts, indeed not even [on the basis of] individual versions or recensions”78) not realizing that if Sukthankar refused to countenance the latter, he could hardly have done so of the former. Had Sukthankar really—as Grünendahl claims—made one text the “basis” of his edition, Ś1 could not have additions over the constituted text.79 The circumstance that Sukthankar can identify passages in Ś1 itself as additions shows that his approach was stemmatic and based on, as he puts it, the “eclectic but cautious utilization of all manuscript classes.”80 Grünendahl has been misled by Sukthankar’s statement that he made “the Ś ā radā version the norm to follow.”81 He takes this to mean that Sukthankar thinks that the Ś ā radā codex, by virtue of being the shortest, contains the oldest text of the Mahā bhā rata and that all the editor must do is edit this manuscript, treating it as the bon manuscrit, and he will obtain the archetype. He thereby ignores Sukthankar’s qualiication that the Ś ā radā version is by no means “entirely free from corruptions and interpolations” and that these corruptions and interpolations “must be carefully corrected and controlled with the help of the other versions, particularly of those of the rival recension.”82 And while Sukthankar thought the ν recension ofered the oldest text of the Mahā bhā rata, even this statement requires qualiication:  it ofers the oldest text relative to the other extant witnesses, not the oldest text simpliciter. In fact, via comparing it with other versions, especially the rival southern recension, which though inlated preserves several archaic readings, the editor can arrive at a text older than that contained in Ś1: the text of the archetype.83 The evidence thus far conirms that the critical edition is not based on the criterion of length. Sukthankar’s method is stemmatic: he compared readings in each individual 182 182 PHILOLOGY AND CRITICISM case to arrive at the reading of the archetype. There is only one circumstance in which he mechanically adopted the reading of Ś1 and K without an intrinsic justiication. Here also he notes its exceptional nature.84 If Grünendahl nevertheless insists that Sukthankar followed the argument from brevity, it is because of the role it plays in his justiication for beginning the critical edition anew.85 From his perspective, the argument allows him to make global claims about the critical edition. It allows him to claim that the discovery of a shorter manuscript (Ñ4) implies that the Mahā bhā rata must be re-edited using Ñ4 as the base.86 He no longer needs to study hundreds of manuscripts or evaluate thousands of readings from the con- stituted text. Rather, through an arithmetic sleight-of-hand, whose confusion we already uncovered, he can give the impression that the constituted text would undergo a major revision on substituting Ñ4 for Ś1. In reality, it would not, since each reading is individ- ually justiied. Even in cases where Ñ4 and Ś1 difered, this would imply a change to the archetype only where Sukthankar, encountering cross-recensional agreement and unable to derive the reading of the archetype from either reading, reprinted the reading of Ś1 and K in the constituted text.87 Finally, the argument from brevity has a further advan- tage in that it permits Grünendahl to ignore manuscript quality, or rather, to reduce the question of manuscript quality to one of quantity. He can now claim that the sole reason Ś1 was adopted as the norm for the edition was that it was the shortest known text. Once he does this, the qualitative role the Ś ā radā codex played in the constitution of the text is no longer evident. The way is open to question Sukthankar’s manuscript classiication. The Argument from a Misapprehension Concerning Classiication (Schriftartprämisse) The Mahābhārata manuscripts were classiied into versions on the basis of their script, according to a principle Lüders irst articulated in 1908 (the script is characteristic of the version). However, this classiication is erro- neous, inasmuch as the script proved not characteristic of the version and neither the variant readings nor the additional passages were consistently characteristic of speciic versions. Refutation of the argument: The Mahābhārata manuscripts were classiied not by their script, but by their text. The term version is a synonym for a manuscript’s text. If the editors nonetheless looked to the script as a irst, extrinsic indicator of a manuscript’s text, this is because, in practice, it was often a reliable guide to the text contained in a manuscript. The editors were not blind to the circumstance that sometimes a manuscript, contrary to their expectation of it, could contain the text of another version. They often reassigned manuscripts assumed to belong to one version on the basis of their scripts to another (whose members were all in a diferent script). If the editors had merely looked to the script, they could neither have drawn up a stemma nor reconstructed the reading of the archetype. The argument from brevity and the argument from false premises are only the irst part of Grünendahl’s attack on the stemma. In this irst part, he focused on the basis, or, rather, what he thought was the basis, of the constituted text. In the second part, he turns to the basis of the stemma itself, targeting what he thinks is its principle of classiication. As this second part of the argument is also based on a misunderstanding of Sukthankar’s work, we call it the argument from a misapprehension concerning classiication (or Schriftartprämisse 183 CONFUSIONS REGARDING CLASSIFICATION 183 for short, to use Grünendahl’s term for his misinterpretation). As this misapprehension is at the heart of Grünendahl’s criticisms of the Mahā bhā rata critical edition, we look at it in detail. Grünendahl introduces the term Schriftartprämisse on page 104 of his article. He does not deine the term, but merely introduces it with the words: “on the basis of this the- oretical foundation, which I  in the following would like to characterize by the term Schriftartprämisse, Sukthankar then groups the manuscripts included for his edition of the Ā diparvan into ‘recensions’ and ‘versions’ in the Prolegomena published in the 7th fascicule (of 1933).”88 According to Grünendahl, this principle—that is, the practice of classifying manuscripts on the basis of their script—has a long tradition in Mahā bhā rata scholar- ship. He traces its origins back to Heinrich Lüders’s 1901 study of the Mahā bhā rata’s south Indian recension, where, according to him, Lüders articulated the underlying prin- ciple for the irst time:89 Almost everywhere we encounter diferent recensions of a work in India, we ind that these recensions are speciic to a particular area. In part, this is due to the fact that the Brahmans, who were responsible for the written transmission of the literature, were typically not familiar with foreign alphabets. In this manner, the transmission of a work was gradually isolated to a region. It is therefore quite possible that manuscripts in Telugu, Malayalam, and other types of scripts would ofer a diferent text than the Grantha manuscripts from the Tamil region.90 Grünendahl asserts that thereafter Lüders irst applied the principle in his 1908 Druckprobe (a sample critical edition of the Mahā bhā rata). He claims that whereas Lüders did not explicitly formulate the Schriftartprämisse, its inluence is evident in his grouping of the manuscripts: To my knowledge, it is nowhere explicitly stated what conclusions Lüders arrived at on the basis of his material. His Druckprobe of 1908 only gives indirect information regarding his views, namely, in the form of the list of manuscripts appended to the end [p.  18], which is divided into Kashmīr Version / Bengālı̄ Version / Nāgarı̄ Version by Arjunamiśra / Nāgarı ̄ Mixed Version / Nāgarı̄ Version by Nı ̄lakanṭ̣ ha/ Southern [Grantha and Telugu] Version. At the root of this classiication lies the implicit understanding that the manuscripts may be classiied according to their script.91 According to Grünendahl, the principle gained general acceptance with its adoption in the Mahā bhā rata critical edition. He notes: “V. S. Sukthankar adopted it [Lüders’s practice of classifying manuscripts by script] in his Foreword to the 1.  Fascicule of the Ā diparvan and arranged the manuscripts used by him accordingly,”92 and Lüders, in his review of the irst and second fascicules of the Ā diparvan, gave it his explicit seal of approval: “It is generally observed that, in India, for works that have been transmitted in diferent versions, the script is characteristic of the version. Lacking familiarity with for- eign scripts, the copyists would copy the manuscripts in the alphabets they were familiar with. However, the boundaries drawn by the script are in part blurred.”93 Thereafter, Grünendahl claims, the principle gained canonical status once Sukthankar grouped “on the basis of this theoretical foundation […] the manuscripts included for his edition of 184 184 PHILOLOGY AND CRITICISM the Ā diparvan into ‘recensions’ and ‘versions,’ ”94 representing them in the stemma seen in Figure 32. This stemma is familiar to us from our discussion of Sukthankar’s edition. It is the stemma presented in his “Prolegomena” to the Ā diparvan and it is based, as we have seen, on a continuous attempt to understand and reine the manuscripts’ relationships to each other based on the details that emerged in the recensio. Grünendahl presents it at the end of his discussion of manuscript classiication to suggest that it was drawn up immediately after the classiication and solely on the basis of the latter. He describes Sukthankar’s creation of the stemma as follows: Sukthankar here distinguishes two recensions, the northern and southern, and nine versions that are mainly constituted according to paleographic criteria. Their arrangement in the stemma manifestly follows a geographic logic, namely, in the case of the northern recen- sion from northwest to east (Ś ā radā → Bengali) and in the case of the southern recension from north to south (Telugu → Malayā lam). Between these two blocks stand the Devanā garī manuscripts; these are once again distributed geographically:  apart from the Devanā garī manuscripts of the so-called K version, which Sukthankar separates out and places along- side the constitutive Ś ā radā manuscript [sic], he distinguishes ive sub-versions, namely, the versions of the four commentators and the so-called composite version.95 From the preceding discussion, we know that Grünendahl does not make a distinction between the classiication of manuscripts and the constitution of the text. His discussion of Lüders’s Druckprobe shows that he thinks that what was decisive about Lüders’s edition was the arrangement of manuscripts into groups, based not on their text but on their script.96 He now goes further: he extends this reasoning to the stemma. Thus he claims that “on the basis of this theoretical foundation” “Sukthankar group[ed] the manuscripts included for his edition of the Ā diparvan into ‘recensions’ and ‘versions.’ ”97 From Vy rata Ur-Mah bh rata N S rada K Nep l Maithil Bengali Devan gar Telugu Grantha Malay lam (other than K) Figure 32 Grünendahl reproduces Sukthankar’s stemma Source: Reproduced from Sukthankar, “Prolegomena,” xxx. 185 CONFUSIONS REGARDING CLASSIFICATION 185 Grünendahl’s perspective, there is continuity between the classiication of manuscripts and drawing up the stemma. He sees the two as identical for, in his view, all Sukthankar had to do and in fact did was to list the versions he identiied from left to right in order of geographic sequence and then draw connecting lines between them. In his mind, no distinction exists between ordering the individual witnesses or witness groups along a line (which is merely a pragmatic choice and thus part of “presentation”98) and assigning the various hyparchetypes and the archetype (the latter two identiied on the basis of recensio) a place within the stemma. Since he does not work with manuscripts, all three steps—classiication or, more precisely, choice of sigla, recensio and constitutio— appear the same to him. According to him, an editor takes manuscripts, assigns them to groups based on script, lists the scripts from left to right, posits recensions and versions (constituted, he opines, “mainly […] according to paleographic criteria”) and so arrives at a stemma. What he thereby overlooks is that this would at most give Sukthankar a matrix of manuscripts arranged from “northwest to east” and from “north to south,” but not a stemma. The editor could not know how he is supposed to draw up the lines between the manuscripts or manuscript groups represented by sigla on the sheet before him. As yet, he has only identiied these groups but has no insight into their relationships. Only the latter permits him to draw up a stemma, for a stemma, in the inal analysis, is only a visual representation of the editor’s understanding of the relations of ili- ation between manuscripts. It is only on the basis of these relations of iliation, as they become clear to him in and through the recensio, that he will have an under- standing of the hyparchetypes and subrecensions from which the individual versions are descended. Sukthankar does not “distinguish” “two recensions, the northern and southern”; he is led to assume their existence based on what the witnesses tell him. No more than he distinguishes two recensions according to geographic orientation does he constitute a “version” on the basis of paleographic criteria: the term version refers to—in fact, it is a synonym for—the text contained in a manuscript. Grünendahl’s error is twofold:  (1) on one hand, he has not grasped the true basis for the classiication, which is the text itself (as it must be, if we are to reconstruct the archetype); (2) on the other, he has confused the script as an initial—and extrinsic—indicator for the version of the text contained in a given manuscript with script as the basis for constituting a version. We take a closer look at both misunderstandings here, as they are instructive regarding both the editorial process and the errors that manifest if one is unfamiliar with the principles of textual criticism. We begin with a discussion of the true basis of classiication. Classiication, Choice of Sigla, Elimination of Manuscripts and Construction of a Stemma Although we have referred throughout to recensio and analysis of the relations of ilia- tion between manuscripts, we are yet to understand exactly how they work in concert to enable an editor to reconstruct the archetype. In Chapter 1, we introduced the idea of 186 186 PHILOLOGY AND CRITICISM a stemma codicum without discussing how an editor arrives at a stemma. Contrary to what one might think, an editor does not propose a stemma: he accepts the stemma as it emerges in and through the recensio. There is a certain reciprocity between the editor’s initial assumptions regarding the tradition (obviously, he must have some idea about the tradition if he is to sort and arrange the manuscript material) and the stemma on which he inally settles. However, an editor cannot simply draw up a stemma and then attempt to it the manuscript material into it. If the proposed stemma is artiicial, the manuscripts will contradict it at every turn: the editor will soon realize the tradition could not have developed as he assumed. Even for a stemma based on a preliminary recensio, he will often ind minor reinements and modiications to the stemma are necessary as the recensio unfolds. This can be seen from Sukthankar’s edition of the Ā diparvan, where he drew and redrew stemmata until he arrived at one that satisied him as the one that best accounted for the observed facts.99 The best stemma is one for which the evidence grows richer as the reconstruc- tion proceeds. Because the relationship between recensio and the stemma is a her- meneutic relationship—an editor proposes a stemma based on the indings of his initial recensio, but then in turn looks to the stemma while conducting a more extensive recensio100—the Lachmannian method is extremely robust in practice. In the best cases, stemma and recensio reinforce each other, leading to a high degree of conidence in the reconstruction. Assuming the manuscript material has been collected and the editor has made a preliminary study of it, the irst stage is to organize the material. This is the stage West refers to as “presentation.”101 The editor will choose sigla and present an account of the manuscripts he used. At this stage, he will only focus on laying out his material and a discussion of ainities between manuscripts will often, in practice, be more convenient after he has introduced his system of nomenclature. Sukthankar’s list of manuscripts forming the critical apparatus of the Ā diparvan ofers a good example (Figure 33).102 The principle Sukthankar adopts in designating manuscripts is self-explanatory: each manuscript is denoted using the capitalized irst letter of the respective script, while the diferent manuscripts of one script are distinguished using the subscript cardinal num- bers 1 … n. In cases where there are subversions in the same script, Sukthankar appends a second subscript letter or number (Da1 for the irst of two Devanā garī manuscripts of Arjunamiśra; Da2 for the second and so on). Even though these sigla bear an obvious reference to the script of the version, the manuscripts are not grouped on the basis of script. This can be seen both from the fact that we could have replaced them with others; for instance, the series of letters from a to i or from r to z with no change in meaning (the sigla are only markers) and from the fact that manuscripts are often—in contradiction to their script—placed under diferent groups when their text warrants it. For instance, the K manuscripts are a group of Devanā garī manuscripts separated out from the remaining Devanā garī manuscripts because of their ainity to the Ś ā radā codex. Likewise, of the three Ś ā radā manuscripts examined for the edition only one contained a Ś ā radā text. The other two were found to be recent manuscripts, written in modern Ś ā radā script and containing Nīlakaṇṭha’s text.103 Sukthankar also notes that among the manuscripts he 187 CONFUSIONS REGARDING CLASSIFICATION 187 examined for the critical edition he found a Devanā garī copy of the Grantha version.104 And although most of the manuscripts of the Devanā garī version were in Devanā garī script (explaining the use of the siglum D), this is not universally true: the Devanā garī version can be found in almost any script. Sukthankar speciically mentions the case of Ś ā radā , Bengā lī, Telugu and Grantha copies of the Devanā garī version.105 Figure 33 Sukthankar’s list of manuscripts forming the critical apparatus Source: Reproduced from Sukthankar, “Prolegomena,” vii–ix. 18 188 PHILOLOGY AND CRITICISM Figure 33 (Continued) 189 CONFUSIONS REGARDING CLASSIFICATION 189 Regardless of what Grünendahl claims, there is no justiication for thinking that script was the criterion of classiication. At this stage, the editor has merely introduced his system of nomenclature. If that system bears a reference to the predominant (but not sole) script of that group as their most visible feature, this does not mean that he has grouped them according to their script. The latter must take place based on the actual textual ainities between the manuscripts, which the editor can only determine once he studies the man- uscript evidence. Thus far, he has only informed the reader about the sigla he has used to identify the manuscripts and possibly (as is the case in Sukthankar’s “Detailed Account of the Manuscripts”) about their relative quality. The editor has as yet not told the reader anything about their relations of iliation. Under discussions of manuscript quality, he may introduce further considerations such as his decision to treat certain manuscripts as a class or to exclude others he considers unimportant for the purposes of constitution.106 This can only occur once he has some familiarity with the texts contained in the manuscripts. Only on the basis of a preliminary inspection can he identify manuscripts as being of such and such a type, of falling under such and such a version and so on. As he begins to read and compare the manuscripts, he will notice that their text is not always identical: some manuscripts will ofer one text and the others another. These diferences may range from small albeit signif- icant changes in orthography to extensive revisions to the text. In the latter case, the editor will assume that he is dealing with diferent versions of the work according to the principle “tantopere inter se diferunt, ut paene quot codices, tot textus esse dicere possis” (“They difer so much among themselves, that [as a result] you could say that there are almost as many texts as there are codices”).107 Accordingly, he will endeavor to group the manuscripts in a way that best relects their variation in content and he will do this regardless of whether they have the identical script or whether he has an ideal exemplar to deine their identity.108 In fact, the manuscripts themselves, or, rather, their contents will tell him whether they belong together. Let us assume that the editor has given his manuscripts a name and, based on their supericial ainities, he has also placed them into certain groups. As yet he only has the beginnings of a stemma and not the stemma itself. Nothing at this stage permits him to propose or the reader to intuit the relationship between manuscripts. The stemma presumes a great deal more information than either of these stages ofers us. In fact, only when the editor undertakes a full-scale recensio will he begin noticing ainities between manuscripts, and only this will permit him to unify the lines of tradition and, over time, to arrive at a stemma. Let us take Sukthankar’s edition of the Ā diparvan as an example. Sukthankar starts out by selecting the sigla Ś, K, Ñ, V, B, D, T, G and M as markers for the diferent manuscripts or manuscript groups. These letters, as we have noted, refer to the scripts of the diferent manuscripts, but it would be erroneous to think that the manuscripts are therefore classiied on the basis of their script. Thereafter, the editor may also decide to neglect for the purposes of constitution certain manuscripts or that ceteris paribus certain manuscripts can be treated as substitutes (for example, M6–8, which replaced M1.2.4 after adhyāya 53). He has thus decided to exclude for the purpose of stemmatization their relationship to each other and to treat them as descendants of a hypothetical ancestor M, whose existence can be inferred on the basis of the similarity of their text. Thus far he has only chosen his sigla and entered them in a single row on the sheet before him. He has not yet begun to identify the relationship of the groups to each 190 190 PHILOLOGY AND CRITICISM other. For the purpose of demonstration, we shall assume he has nine manuscripts lined up from 1 to 9. He now assigns individual markers to these manuscripts using the letters Ś, K, Ñ, V, B, D, T, G and M. He thus has the following array: | | | | | | | | | S′ K  N V B D T G M This array does not yet tell him anything about the relationships between the manuscripts. It only tells him that he has lined them up from left to right and that he has assigned a unique marker to each of them. We have followed Sukthankar’s convention of using cap- ital letters deriving from the irst letter of the underlying script (in all except two cases, Maithilī [= V] and Nepā lī [= Ñ]), but we could also have employed any other sequence of letters such as abcdefghi. The sigla need only be unique: they specify neither an intrinsic relationship between sign and the object so designated nor how this object will relate to other objects in the series. In order to bring these markers (or their underlying objects) into a relation, two more steps are required. The array only tells us that the markers are unique; it does not as yet say anything about the objects so designated. Thus, we must make a further assumption, namely, that the objects so designated are also unique, so that to each unique marker corresponds a unique object. Using our previous convention, we might express this as follows: |S′  |V | B | D | T |G | M | |K |N Just as the irst array only told us that each object was designated using a unique marker, the second line merely assumes that each of the objects so designated is unique. In prac- tice, there will be some variation, since contamination cannot be ruled out. Thus, Ñ may sometimes feature B’s reading, sometimes K’s and sometimes D’s (and this will occur throughout its length), but as long as this variation is not serious enough to undermine the basic distinctions we have drawn, it will not be a problem.109 Returning to our example, we have assigned a label to each of our manuscripts and we have also ascertained that contamination between them is not so signiicant as to undermine any inferences as to genealogical iliation. On their own these two principles are insuicient for drawing up a stemma. They merely tell us that each text has been assigned a unique designation (its respective siglum) and that each text is assumed to be suiciently unique for the purposes of recensio. But they do not as yet tell us anything about the relations of iliation between the texts. For that, we must look at the texts themselves. Only once we begin to look at the latter will the lines begin to bend this way and that. We can represent this as follows: / \ / | \ \ / \ \ S′ K N  V B D T G M (The beginnings of a stemma) And it is only when we carry this out consistently across all our groups, unifying lines as they approach each other and positing intermediate archetypes, that we will ultimately 19 CONFUSIONS REGARDING CLASSIFICATION 191 attain a stemma. For instance, during the recensio, we might ind that Ś and K are often alike, whereas Ñ, V and B share certain features with each other. In that case, we might extend our lines backward, positing intermediate hyparchetypes for the two groups—ν for Ś and K and ε for Ñ, V and B.110 Progressing in this manner, we would begin to gradually unify the tradition, arriving ultimately at the ancestor from which all our witnesses are descended. In this example, we assumed that the editor proceeded from left to right in laying out his manuscripts, but he could have followed any other order. Our reason for preferring this sequence is that it corresponds to Sukthankar’s praxis and thus permits us to better understand how his stemma in particular was developed. But even in Sukthankar’s case, the sequence of manuscript groups from Ś through M is immaterial. He begins with the northernmost group, but—contrary to Grünendahl’s contention—he does not lay out his manuscripts according to geographic logic. The order he follows is that of their prob- able sequence, with the Ś-K group breaking of irst to form the ν subrecension, the γ subrecension thereafter giving rise to the D manuscripts and, via an intermediary ε, to the ÑVB group. To the right, he similarly lays out the witnesses of the southern recension, moving in a circuit from the T-G group, which broke of from S via the σ subrecension, to M, probably the earliest subgroup to break of from S. The manuscripts’ order is due as much to the logic of presentation (as far as possible no lines should cross) as it is to sequence. A simple experiment demonstrates that the order of the groups to which Grünendahl attaches such importance is irrelevant: we might easily reverse our stemma from right to left and again within the subrecensions from right to left with no loss of meaning (Figures 34 and 35). Vy rata Ur-Mah bh rata S N γ ν σ ε Telugu Grantha Malay lam rada K Nep l Maithil Bengali Devan gar T1.2 1 V1 Figure 34 Sukthankar’s stemma reversed 180 degrees around a central axis 192 192 PHILOLOGY AND CRITICISM Vy rata Ur-Mah bh rata S N γ ν σ ε Malay lam Telugu Grantha Nep l Maithil Bengali Devan gar rada K T1.2 V1 1 Figure 35 Reversed stemma with the subrecensions in turn reversed around a central axis Both charts demonstrate that the versions’ orientation, whether from north to south or from south to north, is irrelevant. The second chart additionally shows that the versions’ east–west orientation is irrelevant. We could also insert a third chart, which displaces the Malayā lam toward Nepali or rotates the ε subrecension around its axis to put Bengā lī next to Grantha. In all of these cases, the essential relationship of the manu- script groups to each other (for example, that Ñ, V, B and D break of from N or that the shortest lines of transmission are S—M and N—ν—Ś(K)) remains the same. Clearly, the manuscripts are not oriented according to geographic considerations. If Sukthankar nonetheless followed a discernable geographic logic, this is because this is how most of us, Grünendahl included, read a map—that is, from left to right and from north to south. But to assume from this that Sukthankar followed geographic considerations in organizing his stemma—indeed, that he, as Grünendahl says, “is at pains to let his classiication of the manuscripts appear ‘natural’ ”—is to misunderstand the irst step of preparing an edition: presentation.111 Content as the Real Basis for Classiication, Descent from Ancestors, Ideal Types and Divergence from the Norm Grünendahl’s errors are elementary, but he could have avoided them in a simple way: he could have read Lüders and Sukthankar. Both scholars are clear that the real basis for grouping manuscripts must be their text, and they do so in the very passages Grünendahl cites. For instance, in the 1901 text Grünendahl cites as the earliest source 193 CONFUSIONS REGARDING CLASSIFICATION 193 articulating the so-called Schriftartprämisse, Lüders explicitly distinguishes between the script and the text: Almost everywhere we encounter diferent recensions of a work in India, we ind that these recensions are speciic to a particular area. In part, this is due to the fact that the Brahmans, who were responsible for the written transmission of the literature, were typically not familiar with foreign alphabets. In this manner, the transmission of a work was gradually isolated to a region. It is therefore quite possible that manuscripts in Telugu, Malayalam, and other types of scripts would ofer a diferent text than the Grantha manuscripts from the Tamil region.112 Grünendahl cites this passage, but completely misunderstands the point. Lüders is not trying to distinguish versions by script: he already has diferent versions before him. In fact, he indicates this in the very irst line (“Almost everywhere we encounter diferent recensions of a work in India”). Grünendahl does not notice it and instead claims that “Lüders’s thesis [regarding the reciprocity between script and version] proved not true insofar as none of the numerous variants in his Druckprobe actually corresponded to one of his ‘versions’ dis- tinguished according to its script. This did not lead to any fundamental doubts about the correctness of the thesis; rather it was its fate to have a great inluence on Indological tex- tual criticism in general and on the critical edition of the Mahā bhā rata in particular.”113 Grünendahl also overlooks the fact that Lüders’s central concern is not with the scripts in themselves. Rather, Lüders pays attention to the scripts only insofar as they are an index for changes to the text, changes signiicant enough to justify speaking of a version of the text. Thus, he notes: “It is therefore quite possible that manuscripts in Telugu, Malayalam, and other types of scripts would ofer a diferent text than the Grantha manuscripts from the Tamil region.”114 Grünendahl overlooks not only the reference to the “diferent text[s]” but also the force of the “possible”:  Lüders is merely stating that it is possible that the manuscripts in Telugu, Malayalam and so on might ofer a diferent text than the Grantha manuscripts; whether they actually do so can only be determined by looking at the text. More than any other passage, this one ofers the clearest refutation of the so-called Schriftartprämisse. It explicitly states that the text contained in a manuscript is the real basis of the classiication. Contrary to Grünendahl’s assertion, Lüders did not group the manuscripts depending on whether they were written in Telugu, Malayalam or Grantha script. Rather, he grouped them depending on whether the text they contained approximated one or the other of the known versions of the Mahā bhā rata. And if he identiied these groups using the names of the scripts predominantly associated with those versions, this was because the script often proved the best indicator of the version contained in a manuscript. This can be clearly seen from Lüders’s appendix at the end of his 1907 Druckprobe (Figure 36). Contrary to Grünendahl’s assertion that “at the root of this classiication lies the implicit understanding that the manuscripts may be classiied according to their script,”115 this list shows that the script was not the basis for classiication. Lüders separates out a Kaśmīrī version even though no such script exists. Moreover, he splits up his 20 Devanā garī manuscripts into four groups, while retaining the same siglum (N) for all of them! The Grantha and Telugu manuscripts, though difering in terms of their script, are listed as one version (Lüders’s southern version). The real basis of classiication thus must be the text of the manuscripts. 194 194 PHILOLOGY AND CRITICISM Figure 36 Lüders’s list of the manuscripts collated for his sample critical edition Source: Reproduced from Lüders, Druckprobe einer kritischen Ausgabe des Mahābhārata, 18. Similar problems afect Grünendahl’s discussion of Sukthankar. He cites the latter’s remarks from the section titled “Classiication of Manuscripts,” but deliberately inverts their order to suggest that Sukthankar irst turned to the script as the deining charac- teristic of a manuscript and only thereafter separated out versions. In actual fact, what Sukthankar says is: The manuscript material is divided naturally into recensions by the scripts in which they are written. Corresponding to the two main types of Indian scripts, Northern and Southern, we get two main recensions of the epic. Each of these recensions is again divided into a number of sub-recensions, which I have called “versions,” corresponding to the diferent provincial 195 CONFUSIONS REGARDING CLASSIFICATION 195 scripts in which these texts are written. This principium divisionis is not as arbitrary as it might at irst sight appear. The supericial diference of scripts corresponds, as a matter of fact, to deep underlying textual diferences. [He continues:] It is common experience in India that when we have a work handed down in diferent versions, the script is invariably characteristic of the versions. The reason for this concomitance between script and version appears to be that the scribes, being as a rule not conversant with any script but that of their own particular province, could copy only manuscripts written in their special provincial scripts, exception being made only in favour of the Devanā garī, which was a sort of “vulgar” script, widely used and understood in India.116 This passage clearly shows that Sukthankar irst noted the existence of diferent recensions of the work and only thereafter proceeded to the script as the probable reason for the work’s fragmentation into distinct versions. In fact, he proceeded exactly as Lüders before him had done. Unnoticed by Grünendahl, he also reveals his principle for choosing names or sigla for his groups: he says that he has called the diferent recensions he encountered “ ‘versions,’ corresponding to the diferent provincial scripts in which these texts are written”—hence a Ś ā radā or a Kaśmīr version, a Nepālī version and so on. Rather than distinguish versions according to script, Sukthankar identiies versions and then designates them using their most visible characteristic: the predominant script in use for that version. Grünendahl inverts the order of the passages and introduces Sukthankar’s remarks on the prevalent association of certain scripts with certain versions (the lines beginning “It is common experience in India that when we have a work handed down in diferent versions, the script is invariably characteristic of the versions”) with the words: “Sukthankar him- self formulates the Schriftartprämisse as follows (Prolegomena, p.  vii).”117 Not only does he thus place the consequent before the antecedent but he also presents Sukthankar’s remarks in a less lattering light. He introduces the irst half of the paragraph (the lines beginning “The manuscript material is divided naturally into recensions by the scripts in which they are written”) (which he cites second) with the words: “Sukthankar considers the scripts the ‘natural’ criterion for the distinction of recensions (according to the two main types:  north Indian and south Indian scripts) and subrecensions, which he calls versions (Prolegomena, p. vii)”118 and he footnotes the word “natural” with the words: “Sukthankar is at pains to make his classiication of the manuscripts appear ‘natural.’ ”119 Sukthankar explicitly refers to the text as the real basis of classiication in a second passage. He writes: With regard to the versions described above, it must be frankly admitted that they do not, by any means, form watertight compartments. The isolectional boundaries, as is natural, do not coincide, but are independent of each other; in other words, the textual peculiarities, which are, in inal analysis, the real basis of our classiication, never have, as a matter of fact, an identical area of distribution.120 Grünendahl is aware of this passage for he cites it.121 He once again dismisses it. He introduces it with the words:  “in another place, however, Sukthankar restricts the Schriftartprämisse and leads them over into an argument that makes these textual 196 196 PHILOLOGY AND CRITICISM peculiarities (‘deep underlying textual diferences’) the real basis of his classiication (Prolegomena, p. lxxv)”122 and thereafter writes: “Accordingly, the textual peculiarities also do not ofer a secure theoretical foundation for a classiication of manuscripts?”123 Obsessed with discrediting Sukthankar, Grünendahl uses all his powers of skepticism and irony to cast doubt on Sukthankar’s classiication of the manuscripts. Yet what he overlooks in this remark is that if, as he says, “the textual peculiarities also do not ofer a secure theoretical foundation for a classiication of manuscripts,” no classiication would be possible. If we cannot classify manuscripts by their text, we cannot produce a critical edition. Grünendahl’s confusion of the text with its manuscript becomes especially manifest from his discussion of the Ś ā radā and Kaśmīr versions. He questions Sukthankar’s assumption of a Ś ā radā version, which he correctly identiies as the cornerstone of Sukthankar’s edition. Thus, he cites Sukthankar’s opinion that the Ś ā radā manuscript Ś1 “seems to be the only extant genuine representative of the old version of Kaśmīr,”124 but argues that “Precisely because there is no further representative of this supposed ‘old version of Kashmir,’ the existence of such a version, based on the material of one manuscript of the Ā diparvan, cannot initially be considered demonstrated, especially since external indications for this are sparingly forthcoming.”125 Contrary to Grünendahl’s claim, a single manuscript is suicient to assume the existence of a version of the text when that manuscript difers signiicantly from the others. If we ind that the text has characteristics not shared by other manuscripts or vice versa, we would assume that it is not derived from the known branches of the tra- dition and we would therefore draw a new line representing this branch of the tradi- tion. A single manuscript is suicient to posit a version, even though its relation to the others may not be immediately apparent. For instance, if we were to ind a manuscript x (its script is irrelevant for assigning it a place in the stemma) that cannot be descended from any of our known versions (the nine groups from Ś through M), we would enter it at the bottom of our stemma to one side and await the results of the recensio to deter- mine its relation to the other groups in the stemma. In fact, this is the normal state of afairs: unless an editor had evidence to the contrary, he would treat all manuscripts as unique exemplars in his stemma, entering a unique siglum for them. Sebastiano Timpanaro’s stemma for Lachmann’s edition of the Nibelungenlied, which we saw ear- lier, may serve us here as an example (Figure 37). ω B φ (φ2 ) G E M Figure 37 Treating each manuscript as an independent witness Source: Reproduced from Timpanaro, The Genesis of Lachmann’s Method, 140. 197 CONFUSIONS REGARDING CLASSIFICATION 197 Though Timpanaro’s rather than Lachmann’s, this stemma clearly shows that although Lachmann possessed only a single exemplar of manuscript B, he neverthe- less treated it as an independent witness, as any editor would do. An editor would have no reason to eliminate it from his stemma unless he had reason to believe it was a codex descriptus (a manuscript derived from an extant ancestor or one where the reading of the ancestor can be reconstructed without its help).126 It is possible that during compar- ison, he might ind that this lone manuscript’s readings were not very good or that the other manuscripts did not support them, but this would be a case of eliminatio lectionum singularium rather than eliminatio codicum descriptorum. The editor might thus choose the readings of his other manuscripts over this lone one, but he would have to justify his choices in each case. He could not simply eliminate the witness, especially one as impor- tant as the Ś ā radā codex.127 Grünendahl has been misled by the fact that, in the Mahā bhā rata critical edition, the editors for the most part were dealing not with single manuscripts but with groups of manuscripts whose internal relationships resisted analysis and that therefore had to be treated as members of a class. From this, he concludes that as Sukthankar did not possess a group of manuscripts for the Ś ā radā version, this means he had insuicient grounds for assuming such a version. What Grünendahl overlooks thereby is that the Mahā bhā rata is not a typical case. The normal state of afairs is single manuscripts or manuscripts whose relationship to each other can be stemmatized, and these would be entered as indepen- dent witnesses in our stemma. West’s stemma may serve as an example (Figure 38). This stemma features 14 individual witnesses, A, B, C, D, E, F, G, H, I, K, L, M, N and O, related via the (no longer extant) hyparchetypes β, γ, δ, ε, ζ, η and θ. The rela- tionship of nine manuscripts, A, B, C, D, E, F and M, N, O, can be easily clariied and they are therefore treated as independent witnesses. In contrast, the manuscripts G, H, K, I and L form a group whose structure resists analysis. They are therefore treated as a unit (represented by g in the stemma).128 In contrast, in the Mahā bhā rata, only a minority of manuscripts (Ś1, V1 and, for the Sabhā parvan, Ñ1) had clearly identiiable relationships.129 More often, the editors found that the relationship of the manuscripts within a group to each other could not be stemmatized. They therefore proceeded as West does in this stemma and treated them as members of a class. This fact should not mislead us into thinking that the evidence for these versions is, for this reason, stronger. The manuscript groups B, D, T, G, M and so on are analogous to the group g in West’s stemma. The fact that the individual manuscripts comprising these groups (B1–6, D1–14, T1–3, G1–7 and M1–8) are treated as a unit for purposes of stemmatization represents a limitation in our powers of analysis and is not an argument for preferring the reading of one or more of these groups to that of the individual witnesses. In claiming that there is less reason to assume the existence of a Ś ā radā version than for B, D, T, G, M and so on, because in the former case we possess only a single man- uscript, Grünendahl has mistaken an exceptional state of afairs for the normal state. His argument is analogous to claiming that in the stemma on the next page, we ought to take the evidence of G, H, K, I  and L more seriously than that of the individual manuscripts A, D and M or the hyparchetype b because the former form a group.130 In 198 198 PHILOLOGY AND CRITICISM α β γ δ ζ ε η θ M D g b O B N G H K C F λ E I L A Figure 38 Groups versus individual witnesses Source: Modiied from West, Textual Criticism and Editorial Technique, 44. contrast, he would argue that as no further evidence for A exists than A itself in West’s stemma, we should be skeptical about whether A really exists. In reality, A is an invalu- able aid for reconstructing the reading of the archetype because of its access to λ, so for us to deprive ourselves of A on the grounds that it is the sole example of a manuscript descended from MBλ is to needlessly handicap ourselves. Grünendahl’s argument against taking Ś1’s evidence seriously thus does not work. He confuses the exception with the rule. Manuscripts of a version must be taken as representatives of that version even when there are no other representatives of that version. In contrast, the case where we have multiple representatives of a version represents a limit case of this prin- ciple, where we treat multiple manuscripts as imperfect instantiations of their archetype. A  look at S.  K. Belvalkar’s stemma of the Bhīṣmaparvan (to our knowledge the sole attempt to map the relationship of the manuscripts within the various groups to each other) clariies this (Figure 39). Belvalkar’s stemma illustrates that there is no diference between treating the Malayā lam manuscripts as representatives of an ideal archetype—their hypothetical ancestor M—or as individual witnesses. Although he undertakes to analyze the manuscripts’ relationship 19 CONFUSIONS REGARDING CLASSIFICATION 199 Figure 39 Mapping the relationship of manuscripts within a group to each other Source: Reproduced from Belvalkar, “Introduction [to the Bhīṣmaparvan],” cxv. to each other more closely (for instance, he hypothesizes that the Malayā lam tradition would have split up into two branches:  a purer tradition represented by M3 and M1 and a Ś ā radā -inluenced tradition he calls MŚ,131 and that the latter would in turn have given rise to two branches: M2 and M4 and the lone manuscript M5132), his conclusions are no diferent than if he treated all the Malayā lam manuscripts as descendants of a hypothetical ancestor M (and therefore as more or less perfect representatives of its reading).133 For our purpose, what is important is that when Sukthankar and the other editors entered the sigla B, D, T, G, M and so on into their respective stemmata, they were using them as shorthand for groups, the relationships of whose members could not be further analyzed. Like Belvalkar, the editors could have expanded on the relationship between the members of a group but, as this stemma shows, M is in the inal analysis only an abbreviation for several individual manuscripts and not itself a (super)manuscript. When Grünendahl now alleges that “because there is no further representative of this  supposed ‘old version of Kashmir,’ the existence of such a version, based on the material of one manuscript of the Ā diparvan, cannot initially be considered demon- strated,” this is tantamount to claiming that because Ś1 (but no other manuscript) gives testimony  for  itself in quiet and precise terms, we should reject its testimony for the cacophony of voices represented by B, D, T, G or M. His other arguments against the existence of a Ś ā radā version are no better. He writes that “the inding in other parvans shows rather the opposite [of Sukthankar’s assumption that the Ś ā radā codex represented 20 200 PHILOLOGY AND CRITICISM the old version of Kaśmīr]:  in Mahā bhā rata 4, two of the three Ś ā radā manuscripts available were not included, because they proved to be ‘recent copies of Nīlakaṇṭha’s text and therefore as Ś ā radā codices without value’ (Raghu Vira, Mahā bhā rata 4, Introduction, p. ix f.)—a case that manifestly was not anticipated by the spiritual fathers of the Schriftartprämisse.”134 Sukthankar’s “Prolegomena” shows that, far from not anticipating this case, Sukthankar explicitly mentioned two paper manuscripts written in “modern Ś ā radā characters” and containing the text of Nīlakaṇṭha’s commentary.135 In fact, they are the same manuscripts Raghu Vira later mentions: Grünendahl has overlooked the citation.136 Sukthankar was thus aware that the existence of multiple copies written in Ś ā radā script neither added to nor detracted from the evidence for a Ś ā radā version, especially since this version was identiied not on the basis of its script but its text and since the justiication for a version neither increases nor decreases with the number of (promiscuous) copies of that version. The confusion is Grünendahl’s. On one hand, he thinks that an editor ought not take a manuscript into consideration unless he has multiple copies in the same script and, on the other, he thinks that if an editor has multiple copies in the same script but they are not identical he ought not take any of them into consideration. The editor ought to: only he ought to irst assign them under their correct archetypes—the Ś ā radā text under the Ś ā radā recension and the Devanā garī text under the Devanā garī recension. It is not so easy to forgive Grünendahl’s next argument against Sukthankar’s assumption of a Śā radā version, for the intent to mislead is deliberate. Grünendahl writes that “inter- esting in this context [the lack of—from his perspective—suicient Ś ā radā exemplars to justify the assumption of a Kaśmīrī version] is the following remark by F. Edgerton,” and he then quotes Franklin Edgerton: “Genuine old śāradā writings are now not so easy to ind; what are ofered as such often turn out to be worthless modern copies of works imported into Kashmir from the south.”137 Grünendahl does not inform the reader that Edgerton agrees with Sukthankar about the existence of a Kaśmīrī version: One of the most valuable of Sukthankar’s results is his establishment for the irst time of a “Kaśmīrī” recension of the epic, represented, to be sure, among the manuscripts here col- lated, only by devanāgarı̄ transcripts; no manuscripts in the native Kashmirian śar̄ adā alphabet are included. [There follows the comment about genuine old Ś ā radā writings being rare. Then Edgerton continues:] It is to be hoped that in some way the materials for the Kaśmīrī recension may be augmented by some original śar̄ adā texts.138 Edgerton penned these lines in a review of the irst fascicule of the Ā diparvan—that is, at a time when the Sā́ radā codex was not yet discovered. He thus underscores the correctness of identifying a Kaśmīrī version even without a Ś ā radā exemplar for comparison—the exact opposite of what Grünendahl’s quotation suggests. Moreover, rather than expressing a reservation regarding Sukthankar’s use of the Ś ā radā codex as Grünendahl implies, the passage expresses Edgerton’s hope that the constitution of the Kaśmīrī version (of which he notes: “For it seems that Sukthankar is quite right in regarding this as on the whole the oldest and best recension now known”139) would be conirmed by the discovery of a gen- uine old S ́ ā radā manuscript—as it indeed was from the second fascicule onward (1928). 201 CONFUSIONS REGARDING CLASSIFICATION 201 The passage beginning “Genuine old śar̄ adā writings” is thus a justiication and defense of Sukthankar’s constitution of a Kaśmīrī recension in the absence of a Sā́ radā manuscript. It is not a criticism, raised after the fact, of his use and identiication as such of a Ś ā radā manuscript even though “what are ofered as such [‘genuine old Ś ā radā writings’] often turn out to be worthless modern copies of works imported into Kashmir from the south.” Yet this is how Grünendahl, by citing Edgerton’s comment out of context, presents it. Edgerton held Sukthankar’s work in the highest regard.140 In several reviews he expressed his agreement with his approach,141 highlighting especially his acumen in iden- tifying the Ś ā radā (Kámīrī) version as the best.142 In his review of the second and third fascicules of the Ā diparvan he noted: Fortunately there is a complete Devanā garī ms. of Ā di (K1) which Sukthankar shows to have been a copy of a codex very closely allied to S1; ́ and there are also several other mss. in Devanā garī (the “K” group), which likewise belong to the Kashmirian recension. Probably the most brilliant result of Sukthankar’s work is his convincing demonstration that this Kashmirian recension, of which virtually nothing was previously known, is the oldest and best version of the epic now extant. No one who studies the critical apparatus will doubt this fact.143 The next review further buttressed this judgment: “The ever-increasing number of cases in which passages suspicious on other grounds are omitted in K conirms the growing con- viction that K is, on the whole, probably the best recension we have, and that Sukthankar is well justiied in giving it exceptional weight.”144 Grünendahl does not cite this background. On the contrary, he remarks on the passage from Edgerton he quotes as follows: “This too was a case that Sukthankar ought to have brought into accord with his premises,” and he follows it up with the observa- tion, “Among the manuscripts of the Ā diparvan K4 can serve in some respects here as an example.”145 The reference to K4 is footnoted with the words: “see Sukthankar’s remark regarding this, Prolegomena, p. xii.” Grünendahl has in mind the fact that in Sukthankar’s “Detailed Account of the Manuscripts” K4 is listed as probably a southern exemplar of a northern manuscript. The relevant passage reads as follows: A carelessly written complete MS., with ळ for ल, throughout, which is a Southern trait […] Supplementary folios at 2, 114, 150, 151, 205 include certain long passages (some from Southern sources), copied by the same hand; notable among them being the Brahmā -Gaṇeśa interpolation, whose point of insertion is indicated by a small mark made in the body of the text, and the marginal remark atra śodhapatramekaṃ (cf. v. 1. 1. 1. 53).146 Grünendahl seizes upon this admission (namely, that the K manuscript appears to have been written in the south) as evidence against Sukthankar’s assumption of the Kaśmīrī origins of the Ś-K version. He writes that “The indicated connection with the south is also of some interest for aspects to be dealt with later.”147 Among the criticisms he advances against Sukthankar’s constitution of a Kaśmīrī version is the objection that “although Sukthankar’s choice of the siglum ‘K’ manifestly aims at the association with Kashmir, he does not, to my knowledge, call the ‘K-version’ explicitly ‘Kashmiri’ in his Prolegomena. Only in his edition of the Ā ranỵ akaparvan does he irst risk a more deinitive statement regarding this topic, 20 202 PHILOLOGY AND CRITICISM setting out from a similar situation (Mbh 3, Introduction, p. xxi).”148 He then cites Sukthankar’s observation from the introduction to the Ā ranỵ akaparvan that “The manuscripts com- prising this version [K] must in part be merely Devanā garī copies of Ś ā radā originals and may in part represent the version of a province or a region adjoining the Ś ā radā zone, which it has not been possible to localize more precisely.”149 “However,” Grünendahl notes, “as regards the K manuscripts of the Ā diparvan, Sukthankar only mentions the geographic origin in the case of one manuscript, namely, K3, and this originates from Gujarat!” and he adds an exclamation for emphasis. “Other than this,” he continues, “the description of the K manuscripts, when measured against the signiicance accruing to them in his assessment, [is] rather scanty.”150 Grünendahl concludes: As was the case with the arrangement of the versions […], here too the striving for a geo- graphic arrangement was manifestly in the foreground [of Sukthankar’s concerns], without indications for a special relationship of the K manuscripts to Kashmir being recognizable. Generally, one must note that the editors of the individual parvans constituted the “K” version independently of the geographic origin of the respective manuscripts. This is evident, among other things, from the fact that in other parvans the “K” manuscripts could also be from Bengal or the Telugu region.151 But as with Grünendahl’s irst argument—his criticisms of the assumption of a Ś ā radā version—the confusion here is also his own. Sukthankar does not label the K manuscripts “K” because he thinks they originated from Kaśmīr. Rather, the siglum K denotes the proximity of their text to the version he calls “the Sā radā (or Kaśmīrī) Version (Ś).”152 In other words, the attribution of the group of Devanā garī manuscripts labeled K to a version allied to Ś occurs not on the basis of their geographic origin, but their text. Grünendahl misses the passage, but Sukthankar explicitly calls this group of manuscripts a “Devanā garī Group allied to the (Ś ā radā or) Kámīrī Version (K),”153 and he later underscores the diference between identifying their text with the Kámīrī version and their physical origin with Kámīr: “K is a speciic Devanā garī version allied to the Ś ā radā (or Kámīrī) version (sharply distinguished from other Devanā garī versions) […] [whose] [e]xact provenance […] is unknown.”154 The reason he adopts the siglum K is not, as Grünendahl claims, that he wishes to evoke “the association with Kashmir,” as though these manuscripts would all be from Kámīr (Sukthankar is aware that K4 was probably written in the south and that K3 originated from Gujarat: he explicitly points these facts out and Grünendahl merely repeats him), but that the siglum K is an appropriate marker of their proximity to the two manuscripts whose origins could be localized to Kámīr: 1. The Ś ā radā codex, written in “clear Ś ā radā characters (of perhaps the 16th or 17th century) [on] [b]irchbark (bhū rjapatra),” with which the K manuscripts are evidently related,155 and 2. K1, a manuscript found the closest in terms of its text to Ś1. Regarding this second manuscript, Sukthankar writes:  “Even the outward form and getup of this MS. are suggestive of Kaśmīrī origin. The lines of writing, as in Ś ā radā (bhū rjapatra) MSS. run parallel to the narrow side of the folio. The signatures in the 203 CONFUSIONS REGARDING CLASSIFICATION 203 margin are like those found in Kaśmīrī books.”156 The manifestly Kaśmīrī origins of these two manuscripts, which are his main witnesses for a distinctively Kaśmīrī recen- sion of the text, is suicient reason for Sukthankar to hereafter designate the Devanā garī manuscripts most closely related to them the “K” manuscripts (the second manuscript itself acquires the siglum K1). Note that the attribution of the siglum K1 to the latter does not refer to its Kaśmīrī origins, for in that case the siglum K would have one meaning in the case of K0.2–6 and another in the case of K1. Grünendahl also errs when he claims that Sukthankar “irst risk[s] a more deini- tive statement regarding this topic [the Kaśmīrī origins of the K manuscripts]” in the Ā raṇyakaparvan for when Sukthankar says that the K manuscripts “must in part be merely Devanā garī copies of Ś ā radā originals and may in part represent the version of a province or a region adjoining the Ś ā radā zone, which it has not been possible to localize more precisely,”157 he is not saying that the manuscripts arose in a region near the Ś ā radā zone (which is how Grünendahl interprets him) but that their text or their version must have emerged in a region adjoining the Ś ā radā zone. Thereafter it is possible that over the long centuries of manuscript copying, this “K” version migrated from its place of origin across India, especially if it was written in Devanā garī, which would have made it easy for its text to be disseminated across India. We would now have the situation that southern and Gujarati copies of a text exist, whose similarity nonetheless indicates a common origin and whose resemblance to the Ś ā radā version suggests an origin some- where close to the Ś ā radā region (logically, since their production presupposes suicient Ś ā radā sources and this is likeliest in the region where the Ś ā radā script was in use). Grünendahl’s objection to Sukthankar’s use of K is thus based on a similar misun- derstanding as his objection to Sukthankar’s use of sigla deriving from the irst letters of the other scripts. He thinks the choice of siglum has an innate bearing on the reconstruc- tion and hence raises objections against the use of particular sigla for particular groups, not realizing that this is largely a pragmatic decision. The reconstruction would not be afected if we were to call the K manuscripts D manuscripts with a superscript K (DK), as Edgerton in fact suggested.158 Sukthankar’s identiication of the K manuscripts with Kaśmīr is based on their text being of the Kaśmīrī type. The latter identiication is made on the basis of the Kaśmīrī origins of Ś1.159 In the event, the origins of some of the manuscripts (and not just the texts) proved Kaśmīrī, conirming Sukthankar’s conjecture that the version’s origins should be sought near the Ś ā radā region.160 When Grünendahl charges that a “striving for a geographic arrangement was manifestly in the foreground [of Sukthankar’s concerns], without indications for a special relationship of the K manuscripts to Kashmir being recogniz- able,” he confuses two things: a special relationship of the K manuscripts to the Kaśmīrī version and a special relationship of the K manuscripts to the province of Kaśmīr. In disregarding this distinction he goes astray. He accuses Sukthankar of selecting the siglum K in order to suggest an association with Kaśmīr, not realizing that this is true—and explicitly so—only of the association with a Kaśmīr version. Sukthankar does not at any point identify Kaśmīr as these manuscripts’ place of origin or production, and with good reason. He is aware that the fact that their text is of the Kaśmīrī type means little for their Kaśmīrī origin, since Devanā garī copies can be from any part of India.161 And while he 204 204 PHILOLOGY AND CRITICISM did not have access to K manuscripts in a script other than Devanā garī, his principles allow for this possibility, as transpired in the Bhīṣmaparvan when Belvalkar discovered a Bengā lī manuscript containing a Kaśmīrī text.162 Grünendahl cites the precise passage, yet overlooks its signiicance:  if “the editors of the individual parvans constituted the ‘K’ version independently of the geographic origin of the respective manuscripts,” then this was because geographic origin played no role in identifying the version! Every one of Sukthankar’s successors followed him in distinguishing between the text and the physical reality of a manuscript, so the confu- sion is Grünendahl’s alone.163 As with the Schriftartprämisse, where he advocated a solution (not to classify manuscripts on the basis of script) to a problem (manuscripts cannot be classiied on the basis of script) that occurred to no one else before him, here also his solution (do not identify the siglum K with Kaśmīr) is applicable to a problem (the K manuscripts are not all from Kaśmīr) only he experiences. Possibly, Grünendahl has been misled by Sukthankar’s use of sigla deriving from the initial letters of the script (or, in Lüders’s case, by the use of the titles in the Kaśmīr version, Bengā lī version, Nā garī version by Arjunamiśra and so on) to designate the Mahā bhā rata’s versions. To the novice who encounters Lüders’s or Sukthankar’s editions for the irst time it must seem self-evident that titles are essentially determined by the script. Grünendahl takes these sigla or titles as evidence that Sukthankar and Lüders classiied the manuscripts according to their script, not realizing that their real reason for using these sigla or titles was that the script of a manuscript was the most accurate indi- cator of its contents and, if in practice a manuscript deviated from our expectation of it, this would not be a problem for we would simply reassign it under its correct archetype.164 Likewise, if a group of manuscripts assumed to belong under a version because of their common script proved unlike the other members of that group (for example, the Devanā garī version of Arjunamiśra, which difered from the other Devanā garī versions), we would simply move it into its own group and assign it its own name.165 In all of these cases, what is decisive is not the system of nomenclature used, which as we saw in the preceding subsection is a matter of convention, but the way a manuscript behaves, which can only be established in the recensio. Whether a Grantha manuscript is actually a Grantha manuscript—that is to say, whether it demonstrates the speciic features of the Grantha version—is a function if its text.166 It cannot be deduced from external characteristics such as its script and this can be seen from the circumstance that some Ś ā radā manuscripts proved to contain the Devanā garī text or a Devanā garī man- uscript proved to contain the Grantha text. In contrast, had either Lüders or Sukthankar attempted to classify the manuscripts according to an extrinsic and artiicial criterion such as script, the textual evidence would have contradicted them at every turn.167 Table 5 gives a partial overview (further examples may be found by looking at the remaining parvans) of the manuscripts in the irst six parvans classiied—against the expectations their script evoked—with their true agnates.168 In each case, the editor’s comments, even if cited earlier, are provided in full in the footnote to the manuscript. If there is nonetheless congruity between the script and the version, this must be sought in the way—as Lüders and Sukthankar both hypothesized— manuscripts descended from their sources and underwent diferentiation over time. 205 CONFUSIONS REGARDING CLASSIFICATION 205 Table 5 The text as the true basis of classiication Parvan Manuscript(s) Sigla Script Expected True employed classiication classiication Ādiparvan Raghunath Temple Not collated Śā radā Ś D Library, nos. for the (text of the 3712–32 and critical vulgate) nos. 3951–79 edition1 Sabhā parvan Melkote, Yadugiri T2–42 Telugu T BD Yatiraj Math, (text of the E no. 155; type) Madras, Govt. MSS. Library, no. 1922; Madras, Govt. MSS. Library, no. 1923 Nepal Darbar V23 Newā rī Ñ V Library, no. 1 947 Adyar Library, DG1 and Devanā garī D G XI C 42 and DG24 XXXVI G 13 Virā ṭaparvan Raghunath Temple Not collated Ś ā radā Ś D Library, nos. for the (text of the 3712–32 and critical vulgate) nos. 3951–79 edition5 Udyogaparvan Tanjore, Not collated Devanā garī D G Saraswathi for the Mahal Library critical no. 1250 and edition6 no. 1297 Pudukottah State Not collated Devanā garī D S8 Collection, for the Devanā garī MS. critical (without number) edition7 Bhīṣmaparvan Dacca, University K49 Bengā lī B K Library, no. 669 Madras, Adyar K710 Devanā garī D K and S Library, No. Xc7 Harivaṃśa Unnamed Not collated Nandinā gari Nandinā gari T for the group critical (no siglum edition11 established) Scindia Institute, Not collated neo-Ś ā radā Ś? D Ujjain for the (text of the critical vulgate) edition12 Notes: 1. Ā diparvan: “[P]aper manuscripts, written in modern Ś ā radā characters, with Nīlakaṇṭha’s commen- tary […] They represent probably the Nīlakaṇṭha version.” Sukthankar, “Prolegomena,” v, n. 1. 206 206 PHILOLOGY AND CRITICISM Table 5 (Cont.) 2. Sabhā parvan:  T2:  “Fundamentally the text belongs to the ‘E’ (B and D) version, with all that version’s insertions, but it has been extensively contaminated by the insertion of many additional passages from a Southern source. […] Yet contaminatory inluence of the S version in original stanzas is rare.” T3: “The basic text agrees in general quite closely with T2. This MS. also contains some Southern insertions, but far fewer than T2.” T4: “Also fundamentally an ‘E’ (B and D) text, with all the insertions characteristic of that version, and with readings of detail agreeing even more regularly therewith than is the case with T2 and T3.” Edgerton, “Introduction,” xx–xxi. 3. V2: “Despite the alphabet [Newā rī], it obviously belongs to the Maithilī, not the Nepalese, recension and goes quite closely in general with V1.” Ibid., xviii. 4. DG1: “Devanā garī, but text of the Grantha recension; clearly a recent copy of a Grantha original.” 5. DG2: “Devanā garī, but text of the Grantha recension, like the preceding. […] This and the pre- ceding were fully collated, but they are not included in the Critical Apparatus. They are normal MSS. of the G recension, despite the use of the Devanā garī alphabet.” Ibid., xxii. 6. Virā ṭaparvan: “These [manuscripts] unfortunately turned out to be recent copies of Nīlakaṇṭha’s text, and hence of no value as Ś ā radā codices.” Vira, “Introduction,” x. 7. Udyogaparvan: Sigla not speciied: “[These manuscripts are] merely copies of Grantha MSS. in Devanā garī, no. 6 [in De’s list = no. 1297] being, in addition, a composite paper MS. written by more than one hand.” 8. Siglum not speciied: “[This manuscript is] a typical Devanā garī misch-codex which, being copied probably in the South, contains such a large number of Southern passages that it may be regarded almost as a Southern MS.” De, “Introduction,” xii. 9. Bhīsm ̣ aparvan: K4: “This is the irst time that a MS. written in Bengali characters has been classiied with K, which normally designates Devanā garī copies from a Ś ā radā original. But, à priori, there is no reason why direct copies from a Ś ā radā original should not be made in Bengali characters. The text would in course of transmission be contaminated with the Bengali version, just as, in the other case, it would be with the Devanā garī version. But the MS.  must certainly be classiied as K if it possesses suicient and distinctive characteristics of that group.” Belvalkar, “Introduction [to the Bhīsm ̣ aparvan],” xxix. 10. K7: Not part of the regular critical apparatus, but readings from it are occasionally mentioned and commented on in the Critical Notes to this parvan. Note that the siglum (K7) refers “only to the irst part of the MS., the second part being altogether ignored.” Ibid., lv. This manuscript presents a curious example: it is written in Devanā garī characters, but it is “a composite MS. consisting of two parts, separately paged, giving folios 43 +164. […] The handwriting of both parts appears similar, but the texts given by them difer, the irst part being distinctly Kashmirian, while the second, mainly Southern.” Ibid. This manuscript perfectly illustrates that the true basis of classiication must be the text contained in it. 11. Harivaṃśa: Siglum not speciied: “One manuscript in Nandinā gari script was collated, but it was found to give an identical text with that in Telugu script and was, therefore, not taken into account.” Vaidya, “Introduction [to the Harivaṃśa],” x. 12. Siglum not speciied: “It [the manuscript] is written in neo-Ś ā radā script on paper. The copyist was no doubt a Kashmir Brā hmaṇa, probably on a visit to Banaras, and seems to have prepared this Ms. there as a pastime. The text of the Harivaṃśa in it is the inlated text almost identical with the Vulgate, and what is important to note is that there is a commentary of Nīlakaṇṭha Caturdhara on the top as well as at the bottom of the page. It requires no argument to say that this Ms. was prepared much later than 1690 CE, that is, after Nīlakaṇṭha’s commentary on the whole of the Mahā bhā rata was completed. As this Ms. in neo-Ś ā radā script was prepared so late, and as it contained Nīlakaṇṭha’s commentary, its value as a Ms. of Ś ā radā version is nil.” Ibid., xvii. 207 CONFUSIONS REGARDING CLASSIFICATION 207 Northern Branch 3rd BCE Brāhmı̄ 2-1 BCE Northern Brāhmı̄ Southern Brāhmı̄ (Gupta alphabet, 4th CE) (Drāvid.i alphabet, 4th CE) 4th CE 5th CE Western Gupta Eastern Gupta 7–8 Śāradā Nāgarı̄ (8th CE) (7th CE) Devanāgarı̄ Proto-Southern-Nāgarı̄ Proto-Eastern-Nāgarı̄ 10–11 Newāri Proto-Bengali (11th CE) 13 (11th CE) Southern-Nāgarı̄ 14–15 (13th–15th CE) Bengali Maithili Oriya 16 Nandināgarı̄ Figure 40 The evolution of northern Brāhmī Source: Reproduced from Rath, “The Oriya Script,” 55. Southern Scripts, 250 BCE – 1600 CE 500BCE BRĀHMĪ 250BCE 1st CE Gupta (N) Cave 350 Cerā Cālukya Ven̄gi 650 Proto – Grāntha Vatteluttu W. Cālukya E. Cālukya Old Javanese (Kawi) ·· 1000 Transitional 1300 Old Tulu-Mal. Middle Gr. Hala-Kannad·a Old Telugu Tulu MalayāIam · Grantha Tamil Kannad·a Telugu Modern Javanese 1600 Figure 41 The evolution of southern Brāhmī Source: Modiied from Rath, “Varieties of Grantha Script,” 189. 208 208 PHILOLOGY AND CRITICISM Two charts (Figures 40 and 41)—one of the evolution of northern Brā hmī and the other of the evolution of southern Brā hmī—make this clear.169 In the irst chart, we see how the Brā hmī script developed into two agnates, northern Brā hmī and southern Brā hmī. Assuming that the archetypal Mahā bhā rata manuscript was Sanskrit written in Brā hmī script, when this manuscript was copied around the irst century BCE, it would likely have been copied into either northern Brā hmī or southern Brā hmī. If multiple examples existed and they were copied sev- eral times, they would likely have given rise to northern and southern copies of the text. Assuming further that the southern exemplars then underwent extensive revision (as they in fact did), the southern Brā hmī manuscripts would now constitute an inde- pendent branch of the tradition. If scribes preferentially copied manuscripts from ancestor scripts to descendant scripts (for example, from Proto-Bengā lī into Bengā lī or from Proto-Grantha into Grantha and so on), these inherited diferences would be consistently transmitted. We would now ind that the texts in the successor scripts of northern Brā hmī were distinct from the texts in the successor scripts of southern Brā hmī, though we would not for this reason diferentiate the manuscripts on the basis of script. Rather, the text remains the basis of classiication, but we now ind a relation between the script of a manuscript and its text. This relation would not be perfect, of course. Paralleling the formation of regional recensions of the text, some contamination would occur between versions, blurring the diferences between them. Nonetheless, over time the practice of copying manuscripts in the same script (or from an ancestral script) would reinforce regional variations in the text, and over time we might expect distinct recensions to emerge, each corresponding to the script in use in its local area. Ignoring the occasional oddity such as a Ś ā radā copy of a Devanā garī text (which would not be a problem since we would recognize the manuscript’s true descent upon examination), we could now use script as a shorthand or a substitute for the text of a version. This is not to say that we would follow the script blindly. In the inal analysis, the text alone is a guide to a manuscript’s provenance. The script of a manuscript gives us a good idea of its version and we would use this information to conduct an initial survey of the manuscripts. It is of course possible that during the recensio we might ind signiicant var- iation between the manuscripts. This would imply that our third assumption—namely, that scribes tended preferentially to copy manuscripts from ancestor scripts to descendant scripts—was false.170 In practice, Sukthankar did ind several such manuscripts, which he called “misch-codices of small trustworthiness and of no special value for critical purposes” and therefore eliminated after a preliminary recensio.171 However, as long as we can clearly trace the relationship of the better manuscripts, we shall still be able to infer the reading of their archetype.172 Manuscripts will difer across their lengths, shifting in the direction of one or the other more or less proximate neighbor. Thus, a K manuscript may concur with Ś for most of its length, but in some chapter or book might suddenly resemble the D group. Its classiication as K is independent of this variation (the sigla are only heuristic markers), so the fact that it is inconsistent for some part of its length is not a major objection. No one claims that individual manuscripts perfectly incarnate their ancestors; the only decisive consideration is whether they approximate their ideal types 209 CONFUSIONS REGARDING CLASSIFICATION 209 suiciently to justify treating them as members of a class. Some variation from the norm is inevitable. Where this variation is of signiicance—that is, in the recensio—the editor can be expected to note it in his critical apparatus.173 The stemma’s function, after all, is only to formalize and illustrate the central relationships. It cannot record every exception to the rule. Grünendahl’s confusion arises from contemplating the stemma in isolation— that is, from a failure to realize that a stemma is always presented along with a critical apparatus and must be read in conjunction with the latter: the stemma gives the rule; the critical apparatus notes all the exceptions to the rule. The argument from a misapprehension of classiication does not work. Grünendahl confuses the script as a principle of nomenclature with the script as a principle of classiication and both again with the script as a principle of reconstruction. The error he attributes to Lüders and Sukthankar (they only considered the script in classifying manuscripts) is his own: he is the only one who thinks that knowing the scripts of diferent manuscripts is suicient information not only to classify them but also to draw up a stemma. At most, the argument may imply a strong skepticism about the possibility of drawing up stemma in light of extensive contamination. But, as we see in what follows, this argument is also based on a misapprehension. The Argument from Extensive Contamination The widespread presence of contamination between Mahābhārata manuscripts (evinced by the existence of interpolated passages across recensions) makes a satisfactory classiication impossible. Refutation of the argument: in practice, no editor found contamination such a problem that he could not establish the manuscripts’ true iliation. Though present, contamination was restricted largely to the existence of the additional passages, which, indeed, were often transmitted horizontally. The manuscripts’ true ili- ation almost always became obvious in terms of the signiicant errors in the text. Grünendahl errs because he considers only the interpolated passages rather than readings. His understanding of contamination is also peculiar:  according to him, the absence of interpolation is also a kind of “contamination” (namely, non- contamination), for how (he reasons) could two manuscripts agree in featuring the same text if not for the fact that one of them “contaminated” the other? The answer is simple: the manuscripts contain the identical text, because the same text was handed down to them and no inluence of the one upon the other needs to be presumed to explain the circumstance that they lack the same interpolations. Interpolations, except when interpolated, tend not to be present, so that no special circumstance is required to explain their absence in a manuscript. In the preceding section, we saw that if the so-called Schriftartprämisse has any meaning it is that of a generalized skepticism about the value of dividing up the manuscript material into groups, given the presence of contamination. In contrast, Grünendahl’s wider claim, namely, that the critical edition is based on lawed theoretical premises since it uses script as the criterion of classiication, proved untrue since neither Lüders nor Sukthankar classiied the manuscripts by script. A closer look at Grünendahl’s article reveals that this is in fact what he means. In spite of the misleading term Schriftartprämisse and the equally misguided discussion of the untenability of this premise, Grünendahl’s discussion comes down to the fact that he thinks an assignment of manuscripts to diferent groups on the basis of their text is impossible because their text overlaps in many instances. Thus, immediately after 210 210 PHILOLOGY AND CRITICISM introducing Lüders’s description of the rationale for classiication (which he misleadingly introduces with the words: “Regarding the script as a criterion for the classiication of Mahā bhā rata manuscripts, Lüders already expressed himself in his study on the so- called Grantha recension”174), he states: “Lüders’s thesis proved not true insofar as none of the numerous variants in his Druckprobe actually proved to correspond to one of his versions distinguished on the basis of the script.”175 Thereafter he returns to the theme after several pages in which he discusses his understanding of Sukthankar’s application of the principle. He writes: Sukthankar’s application of these barely comprehensible theoretical presuppositions will be clariied using the example of his recensio of the hyparchetype ν. Precisely because of the exemplary value that is granted his Prolegomena for textual criticism it is worthwhile to look at his comments regarding this branch of the tradition in more detail, and this all the more as he grants central signiicance to the two versions derived from ν (Ś ā radā and K) for his con- stitution of the text.176 Before we look at Grünendahl’s criticisms of Sukthankar’s ν recension, let us irst under- stand contamination. As we saw in the preceding section, at heart Grünendahl’s objections to the critical edition concern an anxiety about contamination. The main issue, after all, is not whether K1 is really the manuscript K1 (as we noted, the system of nomencla- ture employed is a matter of convention, and we could also call our K manuscripts Z manuscripts with no change in our edition),177 but whether it really behaves like a K manuscript over its length. It is possible that after initially behaving like a K manuscript it could later switch to behaving like a D manuscript for the remainder of its length. Here there really would be a problem, because we would continue to assign weight to K1’s agreement with Ś1 (or M or any other an independent version) when in fact it is now behaving like a D manuscript and thus is no longer an independent witness. It has been contaminated or conlated and these sections of the text are no longer descended from a K or a Ś ā radā archetype (that is, ultimately from ν) but from a Devanā garī source. We would now be comparing not the branches of the tradition we think we are comparing but two other branches, and since we cannot know what source a scribe really transcribed for given sections of the text, we would be deceived in our inferences. The stemma in Figure 42 clariies the problem. This stemma indicates the reading of the diferent versions for verse 1.29.5 in the crit- ical edition. In the constituted text, the verse reads: adhaś cakrasya caivātra dı̄ptānalasamadyutı̄ | vidyujjihvau mahāghorau dı ̄ptāsyau dı ̄ptalocanau ||.178 The word ghorau in the inal pada (verse half) is marked uncertain, because the witnesses contain diferent readings:  Ś and K read ghorau (as in the constituted text); Ñ, V, B and D with the exception of D2 and D5 read vı̄ryau; while the Malayā lam recension breaks up into two groups of manuscripts, M2–4 reading mahākāyau whereas M1.5 reads mahāghorau. The Telugu and Grantha recensions are split: T1 features the central recension’s reading, while T2 features that of the Malayā lam group M2–4; likewise, G2 features the central recension’s reading, while G1.4–6 features that of the Malyā lam group M2–4. The one exception is G3, which concurs with the smaller Malayā lam group M1.5 and Ś -K. The constituted text is based 21 CONFUSIONS REGARDING CLASSIFICATION 211 θ N S ν γ σ ε rada K Nep l Maithil Bengali Devan gar Telugu Grantha Malay lam 1 V1 T1.2 M1.5 mah ghorau mah v ryau (as in constituted text) (except D2.5; T2 mah k yau) mah ghorau (as in constituted text) mah k yau (G2 mah v ryau; G3 mah ghorau) Figure 42 Agreement between independent versions on the agreement of Ś and K with M1.5, additionally supported by D2.5 and G3. The editor’s reasoning is easy to reconstruct: the central recension is highly contaminated as are the Telugu and Grantha manuscripts, albeit Grantha to a lesser degree than Telugu. In contrast, Ś and K are relatively pure, and their reading is conirmed by the Malayā lam manuscripts M1.5. Thus, against the possibility that the γ recension retained the correct reading (mahāvı̄ryau) and it was corrupted in S-́K and in M1.5 to mahāghorau, the editor must weigh the possibility that the correct reading (mahāghorau) was retained in S-́K with M1.5, whereas it was lost in the γ recension, either in γ itself or in one of its descendants and thence transmitted via contamination to the other manuscripts in this group.179 Since Ś—M contamination is less likely,180 Sukthankar considers the agreement of S-́K with M the strongest argument for this being the reading of the archetype,181 which is addition- ally supported by the fact that the reading mahāghorau is also found in the manuscripts D2, D5 and G3. Our stemma lectionum would look like Figure 43. Mahāghorau was probably corrupted to mahāvı̄ryau; inding mahāviryau (“greatly vig- orous,” “greatly energetic” or “greatly valorous”) not to make much sense, another copyist emended it to mahākāyau (“large bodied”). Alternately, an original mahāghorau in the source could have given rise to both mahāvı̄ryau and mahākāyau by emendation. Our stemma lectionum would look like Figure 44. If contamination was present, these inferences would not hold. In that case, it is possible that the central recension in fact retained the correct reading, mahāvı̄ryau being transmitted from the archetype via N and S to Ñ, V, B, D, T1 and G2, while mahāghorau 21 212 PHILOLOGY AND CRITICISM adhaś cakrasya caivātra dı̄ptānalasamadyutı̄ | vidyujjihvau mahāghorau dı̄ptāsyau dı̄ptalocanau || ŚK M1.5 D2.5 G3 vı̄ryau Ñ V B D1.3–4.6.7 T1 G2 kāyau T2 G1.4–6 M2–4 Figure 43 Stemma lectionum of verse 1.29.5 adhaś cakrasya caivātra dı̄ptānalasamadyutı̄ | vidyujjihvau mahāghorau dı̄ptāsyau dı̄ptalocanau || ŚK M1.5 D2.5 G3 Ñ V B D1.3–4.6.7 T1 G2 vı̄ryau kāyau T2 G1.4–6 M2–4 Figure 44 Alternative stemma lectionum of verse 1.29.5 is a corruption originating in the ν recension, transmitted via contamination to the M group (either to M as a whole, with M2–4 later emending this to mahākāyau, or only to M1.5). The stemma in Figure 45 clariies the situation. We would now be deceived in our inferences, because we would assume that Ś, K and M are independent versions (and therefore take their agreement as an argument for the reading of the archetype) when, in fact, M has ceased to be an independent witness for Ś and K. The stemma in Figure 46 illustrates the true state of relationships (in respect of this one verse). Mahāghorau is inherited (via contamination) from ν so that in respect of this one verse, M1.5 are actually apographs of ν. Consequently, when we try to reconstruct the reading of the archetype and rely on the evidence of Ś, K and M, we are in fact only relying on the evidence of the ν recension and the reading we reconstruct as the reading of the archetype is only that of the ν hyparchetype. It is in this sense that contamination, if present to more than a minor degree, can mislead us about the true relationship of the manuscripts. A manuscript classiied as M1 and so on, and therefore as a descendant of the hyparchetype M, could be a descendant of ν. This is not to say that its classiication by its script is wrong (script, as we observed, plays no role in the classiication) but that its classiication as a descendant of M (and ultimately of S) is wrong: its text is not inherited from the sources it is assumed to be inherited from, and consequently its speciic features are not features of its presumed source (M) but rather of its true source (ν). Note that renaming the manuscript with a diferent designation does not solve the problem: the real problem concerns not its name but its place in the stemma.182 In this case, our stemma would only relect the real nature of the relationships between the manuscripts if we reassigned M1.5 under the ν recension. Although potentially a problem, this type of contamination is easier to control for in practice. As Reeve notes, “It strikes me as unlikely […] that scribes would be so consistent in incorporating or ignoring variants as to disguise the nature of the arche- type completely. My general impression of contamination, and I believe Alberti’s too, 213 CONFUSIONS REGARDING CLASSIFICATION 213 θ N S ν γ σ ε rada K Nep l Maithil Bengali Devan gar Telugu Grantha Malay lam 1 V1 T1.2 M1.5 mah ghorau mah v ryau (as in constituted text) (except D2.5; T2 mah k yau) mah ghorau (as in constituted text) mah k yau (G2 mah v ryau; G3 mah ghorau) Figure 45 Contamination, undermining the assumption of independence is that it has the efect not of falsifying stemma but of frustrating attempts at drawing them up.”183 For instance, if a scribe followed one source for a certain length, but then changed sources, we would notice this, provided we had other copies of his source.184 Thus, Sukthankar notes of the Grantha version that “for the beginning of the Ā di, we get temporarily, the sub-groups G1–3 and G4–6, but soon the coniguration changes to G1.2.4.5 versus G3.6. The latter group (G3.6) represents the purer Southern tradi- tion, agreeing with M against the other Southern manuscripts, whereas the four MSS. G1.2.4.5 are not merely heavily interpolated but stand palpably under Northern inlu- ence.”185 Over the length of a manuscript, then, it would be very diicult for a man- uscript to conceal its true lineage. Thus, while there are the “misch-codices”—that is, manuscripts in whose copying the scribe has combined two sources—these are of little value for the reconstruction.186 They can be identiied as such and the editor will likely not assign their evidence weight in his reconstruction, if he chooses to note their variants at all.187 In contrast, contamination of the less promiscuous manuscripts, if present, is likely to be restricted to the insertion of passages or occasional comparisons, leading to the recording of variants as marginal glosses. It is unlikely that a scribe would have followed his source throughout and in just one instance incorporated the reading of a diferent recension, undermining our assumption of independent corroboration in just this one instance. Applied to the present case, this means, since M1.5 consistently follow M, we are justiied in thinking that their agreement with Ś and K is evidence of an original inheritance and taking this reading to be the reading of the archetype. 214 214 PHILOLOGY AND CRITICISM θ N S ν γ σ ε rada K Nep l Maithil Bengali Devan gar Telugu Grantha Malay lam M1.5 1 V1 T1.2 mah ghorau (as in constituted text) mah ghorau mah v ryau (as in constituted text) (except D2.5; T2 mah k yau) mah k yau (G2 mah v ryau; G3 mah ghorau) Figure 46 Contamination, the real nature of the relationships in our stemma Contamination is thus less of a problem in practice than it appears in the abstract. An editor can for the most part account for it, the more so as contamination in the Mahā bhā rata tradition was restricted rather to the tendency to inlation than to com- bination of readings.188 Even highly inlated versions such as the Grantha demonstrate their true iliation in terms of their minor readings. The real problem arises from the fact that Grünendahl and the other German critics do not look at the readings of manuscripts but try to establish iliation on the basis of insertions (or their absence). This leads to a distorted view of the tradition, because iliation, as we saw at the beginning of this chapter, is established on the basis of the unique errors a scribe introduces into his manuscript and not on agreement in truth or the addition of passages, which represents another kind of agreement in truth.189 Sukthankar is very clear that the real basis for iliation is, as he puts it, the agreement in “petty verbal details.” He writes: One notable feature of ν to which I must now draw attention is its frequent agreement with S against γ, especially in the matter of isolated and even unimportant readings, scattered throughout this parvan. I shall cite a few (out of the hundreds of possible) instances to exem- plify this interesting and important characteristic of ν. The readings of inlated manuscripts, which serve only to confuse the issue, have been ignored; the references are, as usual, to the adhyā ya and śloka. [A list of 30 readings follows that shows ν and S agree against γ, and Sukthankar then concludes:] Such extensive agreements in petty verbal details must necessarily be, in the main, an original inheritance, and could never be, in their totality, the result of contam- ination or conlation, as one may vaguely imagine they are; because to achieve them would 215 CONFUSIONS REGARDING CLASSIFICATION 215 necessitate more expenditure of energy than an ancient Indian redactor or reciter or com- mentator of the epic would bargain for.190 Because the stemma is based on the comparison of thousands of such minor variants, in practice we are unlikely to be deceived about the true relationship between manuscripts. If a scribe occasionally incorporated the readings of a diferent recension, we could probably still identify his true source and merely note that there is occasional contam- ination from another recension, though if we did not suspect contamination we would continue assigning signiicance to their agreement.191 The true stemma would then be as in Figure 47.192 Such occasional contamination is doubtless present, transmitted mainly via the com- parison of manuscripts, but it is largely restricted to the central manuscript groups and T and G.193 It ought not deceive us as to the real nature of the relationships in our stemma. Neither ought it to prevent us from relying on those manuscripts whose evi- dence is relatively clear to reconstruct the reaching of the archetype.194 We shall call this irst type of contamination corruption, because it originates when the true reading is lost due to corruption in one or more manuscripts and the new reading is transmitted to other manuscripts via comparison. We must now consider a second type of contamination, which concerns not contam- ination of readings but transmission of interpolations. Since this form of contamination concerns additions to the archetype, we refer to it as inlation rather than contamination. The stemma in Figure 48 makes clear the relationships. θ N S ν γ σ ε rada K Nep l Maithil Bengali Devan gar Telugu Grantha Malay lam 1 V1 T1.2 Figure 47 Contamination from and into the central subrecension 216 216 PHILOLOGY AND CRITICISM θ N S ν γ 321* σ ε rada K Nep l Maithil Bengali Devan gar Telugu Grantha Malay lam 1 V1 T1.2 except except except D2.5 T2 Figure 48 The interpolated passage 321* The passage 321* is present in all manuscripts of the central subrecension (with the exception of D2.5) and in T1. It is not present in the manuscripts of the S and ν recensions (with the exception of T1). It thus marks out this group of manuscripts against the latter two recensions. Since it is an insertion, the passage does not serve to establish il- iation, either between the manuscripts of this group or between those of the other group (that is, those that do not feature it). The passage could have arisen in one of the witnesses of the γ recension and then been transmitted to the others (horizontal transmission). Alternatively, it could have arisen in γ itself and then been transmitted to its descendants (vertical transmission). Since we are speaking of an additional passage, which is a kind of agreement in truth, we cannot determine the true state of afairs. The passage could also have arisen in the lone manuscript T1 and thence been passed along to the groups Ñ, V, B and D. It would be incorrect for us to assume, on the basis of this passage, a relation- ship of descent either between one of the witness texts (say, D) and the others or between γ and its hyparchetypes and versions if we had no other evidence to establish this. The two stemmata in Figures 49 and 50 clarify the situation. The reason we may not assume either of these relationships is not simply that we do not know which of them depicts the true state of afairs. We may not assume them because agreement in truth is not a suicient basis on which to assume iliation. If two manuscripts had an insertion, either could have gotten it from the other or they both could have gotten it from a third source. We could never determine their iliation unless 217 CONFUSIONS REGARDING CLASSIFICATION 217 θ N S ν γ σ ε 321* rada K D Grantha Malay lam 1 Nep l Maithil Bengali Telugu V1 T1.2 except except except D2.5 T2 Figure 49 D as the source of the interpolated passage 321* there were errors in transcription that revealed the true relationship, and we would need several such errors over the length of a text to draw reliable conclusions.195 Thus, while interpolation may let us distinguish between the two groups of manuscripts—the inlated and the non-inlated—it does not permit us to conclude a closer relationship between the members of either group of manuscripts. We could neither determine that Ñ, V, B and D, having the same interpolation, are genetically related nor that Ś, K, G and M, not having this interpolation, are genetically related. We could also not conclude that Ś, K and M, having a relatively pure text (one free of this particular interpolation), have a common source.196 If iliation cannot be established on the basis of the presence of insertions, it can be established even less on the basis of the absence of insertions, yet this is precisely what Grünendahl attempts.197 To understand his logic, we must irst understand a third type of “contamination.” We call it non-contamination since it concerns not so much the con- tamination of one manuscript by another but the absence of contamination of one manuscript by another (which Grünendahl thinks is evidence that the non-contaminated manuscript was under the inluence of a third—non-inlated—manuscript). To under- stand what he means, let us consider an example. Grünendahl argues that as the manuscripts D2.5 and Ñ4 are missing certain interpolations otherwise characteristic of their respective groups D and Ñ, this means these manuscripts were under the inluence 218 218 PHILOLOGY AND CRITICISM θ N S ν γ 321* σ ε rada K Nep l Maithil Bengali Devan gar Telugu Grantha Malay lam 1 V1 T1.2 except except except D2.5 T2 Figure 50 γ as the source of the interpolated passage 321* of the K group (which also lacks these interpolations). According to him, the absence of interpolations is evidence of D2.5 and Ñ4’s relative purity, and they must therefore constitute a separate group from the inlated manuscripts. Further, he reasons that as they do not feature certain interpolations against the practice of their presumed groups or archetypes albeit in agreement with the practice of the K group, this is evidence for a genetic relationship between them and the K group. Consequently, he argues that D2.5 and Ñ4 not only resemble K in respect of not featuring these interpolations but that they should also be entered into the stemma along with or under K.  The starting point for his argument is the observation that “Sukthankar thereafter presents 25 readings (‘also selected at random’; Prolegomena, p. l) that Ś1 is supposed to have in common with the K version as a whole in contrast to ‘all other manuscripts.’ ”198 Grünendahl claims that the fact that “these manuscripts [the K group] do not after all distinguish themselves from all other manuscripts is shown by the exceptions, D2.5, which he himself [Sukthankar] repeatedly lists.”199 According to him, this shows D2.5 are related to K, and he attempts to bolster this claim with their non-contamination (against their archetype D).200 Speciically, he notes that “to be sure, Sukthankar concedes that D2.5 could have ‘with advantage’ also been classiied with the K manuscripts (Prolegomena, p. li, lxxi), but why he ultimately decides against doing so is not quite com- prehensible in light of the repeated and conspicuous correspondence of D2 and espe- cially D5—against all manuscripts!—with the hyparchetype ν.”201 Grünendahl argues 219 CONFUSIONS REGARDING CLASSIFICATION 219 that is “all the more the case [all the more incomprehensible] as the inding of the readings [that is, that D2.5 share some of K’s] is explicitly conirmed by the textual additions.”202 He does not mean D2.5 might share some of the K group’s additions. As we saw, iliation cannot be established on the basis of insertions, since insertions are a form of agreement in truth. Further, since additional passages were often transmitted horizontally via a comparison of manuscripts, they are worthless for determining ver- tical transmission or descent. Grünendahl does not seek to establish iliation on the basis of insertions. Rather, he thinks he can establish it on the basis of the absence of insertions. Thus, after asserting that this is “all the more the case [all the more incompre- hensible] as the inding of the readings [that is, that D2.5 share some of K’s readings] is explicitly conirmed by the textual additions,”203 he notes: If, for instance, one considers the total number of textual additions in the manuscripts of the northern recension on p. 108f., it is striking that the number [of additions] in D2.5 lies below that of K4, which however Sukthankar nonetheless included in his K group.204 As before, when he attempted to classify manuscripts on the basis of their length, here also he thinks that D2.5’s length (more precisely, the fact that the two manuscripts contain fewer insertions than K4, since the number of insertions is a poor indicator of length) is an argument for including them in the K version. He goes even further: Only, the quantity of textual additions cannot be the criterion; I  consider the kind of commonalities that the D2.5 and K manuscripts manifest in this respect more important. Here [are] some examples.205 Grünendahl cites three examples to demonstrate the relationship of D2.5 to K. He notes that the passages 321*, 317* and 1861* are all, as he says, “not present in D2.5 K.”206 He further notes that these passages are also not found in some Nepā lī manuscripts and in the southern recension. As the entire argument turns on this claim, let us look at his presentation of the evidence: 321* not present in D2.5 K [and Ñ4; southern recension] 317* not present in D2.5 K [and Ñ3.4; southern recension] 1861* not present in D2.5, K, Ś 1 [and Ñ1.4, D1; southern recension] Source: Reproduced verbatim from Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 117. Note that the argument is not based on insertions in the text of D2.5 and K, but rather on their absence from D2.5 and K. Grünendahl himself underscores this fact: Here the proximity of D2.5 to the K group manifests precisely not in terms of shared textual additions that are present but in terms of textual additions that are not present, which in my opinion lends especial weight to the passages mentioned.207 We have seen that iliation cannot be established on the basis of the absence of speciic markers, but this is exactly what Grünendahl attempts. According to him, the absence of certain passages in manuscripts, especially when it occurs against their archetypes, is 20 220 PHILOLOGY AND CRITICISM θ N S 321* ν γ 317* 1861* σ ε rada K Nep l Maithil Bengali Devan gar Telugu Grantha Malay lam 1 V1 T1.2 D2.5 Figure 51 (Non)contamination of Ñ4 and D2.5 with K a positive argument for reclassifying them—precisely under the manuscripts that also do not feature these passages. The stemma in Figure 51 illustrates his reasoning. Θ—K represents the pure line of descent; θ—γ and θ—D the contaminated line of descent; 321*, 317* and 1861* are a trio of passages found in the contaminated central subrecension; Ñ4 and D2.5 agree with K against their archetypes and hence (argues Grünendahl) must have stood under the K group’s inluence. Evidently, Grünendahl thinks non-contamination is also a kind of contamination: if Ñ4 and D2.5 do not include the passages 321*, 317* and 1861*, they can only have followed K’s example in not doing so. Consequently, he reasons, their true archetypes cannot be Ñ and D (or ε and γ) but instead must be K.  Ñ4 and D2.5 are “contami- nated” in a negative sense, that is to say, where their archetypes feature the three interpo- lated passages, their scribes kept an eye on the text of K, saw that it did not feature the passages, and accordingly adopted its reading. This, he concludes, proves that Ñ4 and D2.5 were “contaminated” with K.208 Although the argument is lawed—there are many ways in which these manuscripts may have chosen not to feature the passages, the simplest being that they followed the reading of their presumed archetype and the presence of these passages in the other exemplars is due to later contamination—we shall continue with Grünendahl’s arguments and only later consider the problems with his view of contamination (namely, 21 CONFUSIONS REGARDING CLASSIFICATION 221 that the absence of contamination can also be thought of as a kind of “contamina- tion”). Grünendahl’s next argument concerns the relationship of D2.5 to a subgroup of K manuscripts, speciically K0–3. He writes: If one excludes K4 because of its previously indicated special position among the K manuscripts and concentrates on the relationship between D2.5 and K0–3, then this inding is conirmed by the following evidence.209 As with his argument for a close relationship between D2.5 and K, here also he focuses on absent passages. Speciically, he introduces two passages, 1214* and 221*, which he thinks demonstrate the closer relationship of D2.5 to K0–3. 1214* not present in D2.5, K0–3, Ś 1 [and Ñ4; southern recension] 221* not present in D2.5, K0–3 [and Ñ3.4; southern recension] Source:  Reproduced verbatim from Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 117. Additionally, he notes that “the close relationship of the group K0–3 specially with D5 is clear from numerous other passages: (1) not present in D5, K0–3, Ś1 [and Ñ4]: 1346*, 1444*, 1517*; (2)  not present in D5, K0–3, Ś1 [and Ñ4, southern recension (!)]:  424*, 1202*, 1344*, 1345*, 1363*, 1450*, 1458*, 1494*, 1546*, 1821*.”210 Once again, the argument is based not on insertions in D2.5 and K0–3, but, rather, on the absence of these insertions. We have already noted that the absence of a certain characteristic cannot be the basis for asserting a relationship between two species or two things. If iliation cannot be established on the basis of insertions, it can be established even less on the basis of the absence of insertions.211 It thus seems that Grünendahl is simply confused about the diference between insertion and the absence of insertion. However, his next comment reveals that he intends his readers to take the distinction literally. He writes: “As above, it is textual additions not pre- sent that indicate at the extensive commonalities between D5 and K0–3”212 and follows this up with the comment: “On the other hand, it should not be left unmentioned that D5 also has commonalities with K4, however with the diference that these show themselves in pre- sent textual additions, for example, 435*, 687*, 1715*, 1745*, 1781*, 1863* and 2090*.”213 Grünendahl is thus aware of the distinction between insertion and the absence of inser- tion.214 Yet he insists that the latter justiies him in grouping certain manuscripts together. According to him, the true nature of our stemma should be as seen in Figure 52. We already saw that neither of Grünendahl’s arguments concerning contamination works. First, contamination of D2.5 with ν is not an argument for including D2.5 under ν. The fact that conlated manuscripts contaminated with ν exist is not an argument against taking the evidence of the K manuscripts, especially those that reproduce the reading of their archetype with greater idelity, seriously. The existence of conlated specimens does not and cannot afect our ability to reconstruct the reading of the arche- type, as though we should not take the evidence of good old manuscripts seriously because late and inferior copies of them also exist. Grünendahl’s argument is a non sequitur. The only reason for not taking a manuscript’s evidence seriously is if it is an 2 222 PHILOLOGY AND CRITICISM N Characterized by inclusion of 321*, 317*, 1861* Characterized by noninclusion of 321*, 317*, 1861* ν γ ε rada Expanded K Nep l Maithil Bengali Devan gar 1 group V1 Figure 52 Constituting groups on the basis of additional passages missing from manuscripts inferior or composite copy, not because other copyists produced inferior or composite copies of its source. Indeed, by Grünendahl’s logic, we could not even rely on extant archetypes, since copyists have produced degenerate copies of them! Further, even if we were to classify D2.5 with K, this would not have a substantial bearing on the con- stituted text. It would change the critical apparatus because, in all instances where the editor noted that the K manuscripts gave a consistent reading, he would now have to note that K, except K7 and K8 (our old D2 and D5), gave this reading. Likewise, in all instances where he noted that the D manuscripts featured a unique reading, he would now have to note that D, as well as the two K manuscripts K7 and K8, feature this reading. But other than this nothing is gained by this change. The editor would still rely on the K group as the better and less conlated group and he would permit himself only to be irritated very slightly by K7 and K8’s recidivism. How to classify manuscripts, then, is as much a function of pragmatic considerations as it is of substantive ones and we could well imagine a situation where an editor labeled a southern manuscript K but then kept noting that it was the exception to the group. Grünendahl’s second argument for reconsidering the relationship of K to D2.5, namely, on the basis of shared absences rather than shared contaminated readings, is no better. He argues that the fact that D2.5, like K, maintained the purer reading (though in respect of three interpolated passages and only these interpolated passages!) is evidence of their close relationship. But the only valid inference from the absence of an interpolated passage in a manuscript is that it is not related either by descent or by contamination to a manuscript in which this interpolation appears and not, for instance, that the manuscript in which the interpolation is absent is related to another manuscript in which this passage is also absent.215 We cannot infer a relationship between two manuscripts from the fact 23 CONFUSIONS REGARDING CLASSIFICATION 223 that they both lack the same insertion.216 But this is precisely Grünendahl’s argument. He underscores that he is arguing not from insertions but from their absence:  “Here the proximity of D2.5 to the K group manifests not in common passages that are present but in textual passages that are not present, which in my opinion lends especial weight to the passages mentioned.”217 By this logic, we could also posit a relationship between the manuscripts of any two unrelated works, since they by deinition lack the text of some third composition. Let us return to our stemma (Figure 46). The correct way to read this stemma is that 321* is an interpolation conined to the γ recension. It is found in all Bengā lī manuscripts, all Devanā garī manuscripts, Maithilī and all Nepālī manuscripts with the exception of Ñ4. This does not mean that the manuscripts that lack this interpolation are more closely related, just as if two individuals caught a rare virus we might suspect that they had been associating but from the fact that two other individuals did not have this rare virus, we could never conclude that they had been associating. Grünendahl misconstrues the evidence of the stemma and concludes that since D2.5, like K, contain a text of the purer type, this means they are related (Figures 49 and 50). Having used the absence of three interpolations to posit a closer relationship between D2.5 and K, he now extends this logic to demonstrate a closer relationship between D2.5 and K0–3 using two more interpolations (1214* and 221*). As before, he notes that these passages are absent in these groups (D2.5 and K0–3). He claims that as there are even more absences in common between D2.5 and the narrower K group K0–3, they must be especially closely related.218 The argument is like this (see Figure 53). K is a noninlationary group comprised of the manuscripts K0–4, D2.5 and Ñ4, whose speciic characteristic is that none features the interpolated passages 321*, 317* and N Characterized by inclusion of 321*, 317*, 1861* Characterized by noninclusion of 321*, 317*, 1861* rada Expanded K Nep l Maithil Bengali Devan gar 1 group V1 Characterized by noninclusion Characterized by inclusion of of 1214* and 221* 1214* and 221* K4 Figure 53 Identifying a core K group on the basis of missing additional passages 24 224 PHILOLOGY AND CRITICISM 1861*. K itself breaks up into two subgroups: a core K group comprised of K0–3, which are the purer branch of the K version, and the lone manuscript K4, which represents the inlationary branch of the K version. The two groups (according to Grünendahl) are dis- tinguished in respect of the two interpolations 1214* and 221*. Thus, while K0–3 do not feature these interpolations, K4 resembles the inlationary γ recension in including them. As before, Ñ4 and D2.5 distinguish themselves from their archetypes in not featuring the interpolations 1214* and 221* but in this case they agree only with K0–3—against K4—in doing so. It is clear what Grünendahl hopes to demonstrate with this argument. According to him, these passages show that the manuscripts Ñ4 and D2.5 are not just closer to the K version than to their supposed archetypes; they stand in the closest proximity to the pure K group, K0–3. And since the script is anyway not a legitimate basis for classiication (he thinks), the assignment of the manuscripts to the versions must be reconsidered. As he puts it, “A greater signiicance accrues to the indicated commonalities between D5 and K0–3. These stand in a manifest contradiction to Sukthankar’s argument, which aims at displacing D2.5 into proximity with the conlated manuscripts K3–6 and thus to keep the ‘pure form’ of the K version, represented by K0.1, from every taint of contamination.”219 Why exactly Grünendahl thinks that by demonstrating that other manuscripts also do not feature the insertions that K0.1 do not feature, this introduces the “taint of con- tamination” into them is unclear: K0.1 would indeed possess the “taint of contamina- tion” as he so vividly describes it, if they had been contaminated, but in this case he has at most shown that D2.5 are also free from the “taint of contamination” (at least in respect of these two passages). It takes an unusual intellect to conclude that because K0.1 and D2.5 are simultaneously free of the “taint of contamination,” they must nonetheless, for being identical in this respect, possess the “taint of contamination.” One might as well conclude that because two individuals are virgins at the same time, they are nonetheless soiled, for how else could they be identical in this respect if not through intercourse with each other. As evidence that Ñ4 and D5 are closer to K0–3 than even K4 (which at least nom- inally is part of the K group), Grünendahl next lists 13 passages. He introduces them with the words: “The close relationship of the group K0–3 to D5 in particular is clear from numerous other passages: (1) not present in D5, K0–3, Ś1 [and Ñ4]: 1346*, 1444*, 1517*; (2)  not present in D5, K0–3, Ś1 [and Ñ4, southern recension (!)]:  424*, 1202*, 1344*, 1345*, 1363*, 1450*, 1458*, 1494*, 1546*, 1821*.”220 Once again Grünendahl emphasizes that he is building a stemma, not on the basis of errors in transcription or insertions, but rather on the absence of insertions. He notes: “As above it is the additions that are absent that hint at far reaching commonalities between D5 and K0–3.”221 He thus uses the absence of common features to continuously reine his classiication. In the irst stage, he noted that the K group and Ñ4 and D2.5 were linked by the fact of not containing the interpolations 321*, 317* and 1861*. Thereafter he noted that this rela- tionship is reinforced in the case of K0–3 by the fact that neither this core K group nor the aforementioned manuscripts from the Nepā lī and Devanā garī groups contained the interpolated passages 1214* and 221*. In the third stage, he now shows that Ñ4 and D5 constitute the innermost core of this group in that they most closely approach the 25 CONFUSIONS REGARDING CLASSIFICATION 225 N Characterized by inclusion of 321*, 317*, 1861* Characterized by noninclusion of 321*, 317*, 1861* ν γ ε rada Expanded K Nep l Maithil Bengali Devan gar 1 group V1 Characterized by noninclusion Characterized by inclusion of of 1214* and 221* 1214* and 221* Characterized by noninclusion of 1346*, 1444*, 1517*, 424*, 1202*, 1344*, 1345*, K4 Characterized by inclusion of 1346*, 1444*, 1363*, 1450*, 1458*, 1494*, 1546*, 1821* 1517*, 424*, 1202*, 1344*, 1345*, 1363*, D2 1450*, 1458*, 1494*, 1546*, 1821* Figure 54 Using the absence of interpolations to reine the classiication of manuscripts noninlationary tradition represented by Ś1 and K0–3. The argument would be some- thing like that presented in Figure 54. This stemma ofers the clearest indication yet of Grünendahl’s approach. As before, he uses a selection of interpolations to separate the Mahā bhā rata manuscripts into purer and less pure groups. Via continuously discarding the inlated manuscripts, he thinks he can arrive at the purest group of manuscripts, which he interprets to mean not the general correctness and antiquity of their readings, but the absence of interpolations. In the process, he breaks up the original groups Sukthankar established on the basis of the manuscripts’ true iliation and redistributes their members into new groups. Thus, his “K” group is no longer comprised of the manuscripts descended from Ś ā radā sources (ultimately, ν), but of manuscripts all of whose text approaches a certain norm in terms of not containing certain passages. How can this establish iliation? The short answer is that it cannot, because there are many ways in which two manuscripts might not have a certain interpolation. They might not have this interpolation because they are both descended from separate lines of the tradition neither of which featured the interpolation, or they might not have the inter- polation because they both did not come into contact with the source of the interpola- tion, or they might not have the interpolation because, although their respective scribes were aware that certain manuscripts contained the additional verse, they nonetheless chose not to include it. In fact, Grünendahl’s own list of passages that are, as he puts it, “not present in” D2.5 shows this: the passages 321*, 317*, 1861*, 1214*, 221*, 424*, 1202*, 1344*, 1345*, 1363*, 1450*, 1458*, 1494*, 1546*, 1821* are also absent from the southern recension, yet no one is suggesting that the southern recension should be 26 226 PHILOLOGY AND CRITICISM included in the K group! For convenience, the following table lists all the passages that Grünendahl cites as evidence that D2.5 is related with K. 321* not present in D2.5 K [and Ñ4; southern recension] 317* not present in D2.5 K [and Ñ3.4; southern recension] 1861* not present in D2.5, K, Ś 1 [and Ñ1.4, D1; southern recension] 1214* not present in D2.5, K0–3, Ś 1 [and Ñ4; southern recension] 221* not present in D2.5, K0–3 [and Ñ3.4; southern recension] 1346* not present in D5, K0–3, Ś 1 [and Ñ4] 1444* not present in D5, K0–3, Ś 1 [and Ñ4] 1517* not present in D5, K0–3, Ś 1 [and Ñ4] 424* not present in D5, K0–3, Ś 1 [and Ñ4, southern recension (!)] 1202* not present in D5, K0–3, Ś 1 [and Ñ4, southern recension (!)] 1344* not present in D5, K0–3, Ś 1 [and Ñ4, southern recension (!)] 1345* not present in D5, K0–3, Ś 1 [and Ñ4, southern recension (!)] 1363* not present in D5, K0–3, Ś 1 [and Ñ4, southern recension (!)] 1450* not present in D5, K0–3, Ś 1 [and Ñ4, southern recension (!)] 1458* not present in D5, K0–3, Ś 1 [and Ñ4, southern recension (!)] 1494* not present in D5, K0–3, Ś 1 [and Ñ4, southern recension (!)] 1546* not present in D5, K0–3, Ś 1 [and Ñ4, southern recension (!)] 1821* not present in D5, K0–3, Ś 1 [and Ñ4, southern recension (!)] Source: Reproduced verbatim from Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 117 (all italics Grünendahl's; square brackets and exclamation marks in the original). Grünendahl inserts an exclamation mark after “southern recension” in the last 10 examples as though this is remarkable, but there is nothing surprising about the cir- cumstance that the interpolations are absent in the southern recension:  it shows that interpolations, unless added, will be absent from the text. This will be true even of distant versions. Only someone who thinks the absence of interpolation is evidence for an orig- inal relationship between manuscripts will be surprised to ind that the southern recen- sion also lacks interpolations absent from Ś and K.222 Grünendahl cannot distinguish between an agreement in truth that indicates iliation through the archetype—that is to say, that Ś, K and S contain the same basic text if we ignore corruptions and inlation— and an agreement in truth that is evidence of a later and more speciic inluence, for instance, if two manuscripts do not feature the same interpolation, they can only do so if one was under the other’s “contaminating” inluence.223 If he was consistent, his stemma would look like Figure 55. N breaks up into two groups: the noninlationary recensions ν and S characterized by the absence of 321*, 317* and 1861*; and the inlationary γ recension characterized by the presence of these interpolations. Against their archetypes, Ñ4 and D2.5 do not fea- ture these interpolations and therefore are removed from their archetypes and reassigned under the ν recension, which they approximate in this respect. ν again breaks up into two groups: the noninlationary recensions S, Ś1 and K (the latter expanded to include Ñ4 and D2.5) characterized by the absence of 1214* and 221* and the inlationary K4 recension—comprised of a single manuscript—characterized by the presence of 27 CONFUSIONS REGARDING CLASSIFICATION 227 N Characterized by inclusion of 321*, 317*, 1861* Characterized by noninclusion of 321*, 317*, 1861* ν γ ε rada Expanded K Nep l Maithil Bengali Devan gar 1 group V1 Characterized by noninclusion Characterized by inclusion of of 1214* and 221* 1214* and 221* K4 Characterized by noninclusion of Characterized by inclusion of 424*, 1202*, 1344*, 1345*, 1363*, 424*, 1202*, 1344*, 1345*, 1363*, 1450*, 1458*, 1494*, 1546*, 1821* 1450*, 1458*, 1494*, 1546*, 1821* Characterized by noninclusion Characterized by inclusion of of 1346*, 1444*, 1517* 1346*, 1444*, 1517* D2 S Telugu Grantha Malay lam T1.2 Figure 55 The descent of S in Grünendahl’s classiication these interpolations. S, Ś1, Ñ4 and D2.5 represent a noninlationary tradition that now contrasts with the inlationary central recension γ and the lone K manuscript, K4, standing under the former’s inluence in respect of featuring 1214* and 221*. In the next step, this noninlationary tradition again breaks up into two groups: the noninla- tionary recensions S (the southern recension), Ś1, and a shrinking K characterized by the absence of 424*, 1202*, 1344*, 1345*, 1363*, 1450*, 1458*, 1494*, 1546* and 1821* and the inlationary D2 recension—comprised of a single manuscript—characterized by the presence of these interpolations. In the inal step, S turns around and comes to stand under D2’s inluence in that both groups are characterized by the inclusion of 1346*, 1444* and 1517*. 28 228 PHILOLOGY AND CRITICISM Clearly, this stemma does not represent the true relationships among the manuscripts, because S is manifestly not a descendant of N. But this is something that we could only establish on the basis of the manuscripts’ readings. Since Grünendahl does not consider the readings but builds his stemma using the absence or presence of speciic interpolations, nothing in his stemma would tell us that S represents a separate branch of the tradition from ν or Ś ā radā . We might construct the stemma as in Figure 56. S agrees with Ś and K throughout in not featuring the interpolations that these groups do not feature. Only at the end does it show traces of “contamination” with D2 in the peculiar sense that Grünendahl understands it, that is to say, when two manuscripts either feature or do not feature an interpolation, one of them must have been contaminated by the other to feature or not to feature the interpolation. Against its archetype, it features N Characterized by inclusion of 321*, 317*, 1861* Characterized by noninclusion of 321*, 317*, 1861* ν γ ε rada Expanded K Nep l Maithil Bengali Devan gar 1 group V1 Characterized by noninclusion Characterized by inclusion of of 1214* and 221* 1214* and 221* K4 Characterized by noninclusion of Characterized by inclusion of 424*, 1202*, 1344*, 1345*, 1363*, 424*, 1202*, 1344*, 1345*, 1363*, 1450*, 1458*, 1494*, 1546*, 1821* 1450*, 1458*, 1494*, 1546*, 1821* Characterized by noninclusion Characterized by inclusion of of 1346*, 1444*, 1517* 1346*, 1444*, 1517* D2 S Telugu Grantha Malay lam T1.2 Figure 56 The true position of S in Grünendahl’s classiication 29 CONFUSIONS REGARDING CLASSIFICATION 229 the three interpolations 1346*, 1444* and 1517* and therefore is removed from it and reassigned under the D2 recension, which it approximates in this respect. Grünendahl’s argument is palpably false, but it has a further law: the position of the groups in the stemma is dependent on the order in which we take the interpolations. If we irst took the interpolations 1346*, 1444* and 1517* that S has in common with D2, we would move S under D2, but then on inding that, against its archetype D2, S agrees with Ś and K in not featuring the interpolations 424*, 1202*, 1344*, 1345*, 1363*, 1450*, 1458*, 1494*, 1546* and 1821*, we would now bring it back under Ś and K. Our stemma would look like Figure 57. N Characterized by inclusion of 321*, 317*, 1861* Characterized by noninclusion of 321*, 317*, 1861* ν γ ε rada Expanded K Nep l Maithil Bengali Devan gar 1 group V1 Characterized by noninclusion Characterized by inclusion of of 1214* and 221* 1214* and 221* K4 Characterized by noninclusion Characterized by inclusion of of 1346*, 1444*, 1517* 1346*, 1444*, 1517* Characterized by noninclusion of 424*, 1202*, 1344*, 1345*, 1363*, 1450*, 1458*, 1494*, 1546*, 1821* D2 Characterized by inclusion of 424*, 1202*, 1344*, 1345*, 1363*, 1450*, 1458*, 1494*, 1546*, 1821* S Telugu Grantha Malay lam T1.2 Figure 57 The order of interpolations 230 230 PHILOLOGY AND CRITICISM We have already discussed why the absence or presence of interpolations cannot be a guide to iliation, so these stemmata are purely theoretical exercises. The simplest and truest relation between the manuscripts is represented by the stemma in Figure 46: the reason both N and S do not contain the three passages 321*, 317* and 1868* (and so on for the other groups and S) is that these passages are insertions that arose later in one of the subrecensions that broke of from N. Nonetheless, these theoretical exercises serve a valuable function, because they let us intuit something about Grünendahl’s understanding of textual criticism. Grünendahl imagines contamination as a kind of abstract power exercised over the manuscripts. Thus, when two manuscripts A and B do not feature an interpolation, he does not realize that this reveals nothing about their relationship. Rather, he thinks that the only way B could not feature an interpolation is if A, which also does not have the interpolation, exercised some kind of inluence on it. Hence, he concludes that B should now be grouped with A, if not as an apograph of A, at least as belonging with an expanded version called “A.” From his perspective, the fact that Ñ4 and D2.5 do not feature 321*, 317* and 1861* is not simply evidence that they retained the original text or sequence at this point and that their similarity to K in this respect is evidence of an orig- inal inheritance. No, it must be evidence of the manuscripts coming under K’s inluence, for how else (he reasons) could they have avoided including the interpolations against the practice of their archetypes. Thus, he concludes, Ñ4 and D2.5 should be separated out from the Ñ and D groups (ultimately the recension γ) and placed with Ś and K (the ν recension) and because of contamination Sukthankar’s groups no longer hold. We already saw that Grünendahl’s argument regarding classiication was not really about classiication, because the real issue is not whether the sigla for the various groups (whether derived from the initial letters of the scripts or something else) are correct, but whether manuscripts can be expected to correspond to the expectations of their groups and that means whether they can be expected to be independent witnesses in the cases in which they are assumed to be independent.224 Since the one factor that will cause manuscripts not to be independent witnesses is if their readings are inherited not from the branch of the tradition they are assumed to represent but from a neighboring branch, this is equivalent to saying that the real problem concerns contamination. Grünendahl’s inlated claims about the untenability of the so-called Schriftartprämisse thus amount to the anxiety that vertical transmission plays a lesser role in the tradition than horizontal.225 Yet his way of exploring this question is peculiar. Rather than examine contamination of manuscript readings, he focuses on interpolated passages. Here also his procedure is peculiar—he looks at instances when manuscripts agree in not featuring interpolations and from this concludes that they must be related. This is not an argument for the rela- tionship of two manuscripts, as a simple experiment demonstrates. The simplest way two manuscripts will agree in not featuring a given interpolation is if one of them is not a Mahā bhā rata manuscript (Figure 58). For any given interpolation not found in Ś and K, the Rā mā yaṇa will not contain that interpolation. Using Grünendahl’s logic, we would establish a relationship between Ś and K and the hypothetical Rā mā yaṇa manuscript through the archetype even though the text of the Rā mā yaṇa is not descended from the archetype θ. Grünendahl’s argument for the relationship of manuscripts would only work if we were to assume that the text of the archetype is handed down unchanged from 231 CONFUSIONS REGARDING CLASSIFICATION 231 θ N S Characterized by noninclusion of 321*, Characterized by inclusion of 321*, 317*, 317*, 1861*, 1214*, 221*, 1346*, 1444*, 1861*, 1214*, 221*, 1346*, 1444*, 1517*, 1517*, 424*, 1202*, 1344*, 1345*, 1363*, 424*, 1202*, 1344*, 1345*, 1363*, 1450*, 1450*, 1458*, 1494*, 1546*, 1821* 1458*, 1494*, 1546*, 1821* ν γ σ ε R m ya a rada K Nep l Maithil Bengali Devan gar Telugu Grantha Malay lam 1 V1 T1.2 Figure 58 Unrelated manuscripts on the same stemma manuscript to manuscript, that is to say, if every apograph were to perfectly incarnate its source. In that case, the only diference between manuscripts would be the presence or absence of interpolations, and hence “iliation” of a kind could be established through this criterion.226 This is in fact how Grünendahl understands the process of manuscript copying. For him, manuscript transmission is comprised of two processes and two processes only. On one hand, he thinks that the text of the archetype is handed down fully formed, like Athena emerging from the head of Zeus. On the other, he thinks that additional passages are then added to this core text, producing the regional recensions and individual witnesses. Further, since manuscript copying in his view is not just mechanical but produces facsimile copies of the sources, the only way interpolations could supervene on the core text is through contamination. In contrast, the absence of interpolations becomes an argument for “contamination” of another kind, namely, non-contamination, for if one branch of the tradition has ceased producing facsimile copies or, rather, produces facsimile copies of an inlated text, the only way individual manuscripts of that tradition could escape the tendency to produce facsimile copies of this inlated version is if they came under the inluence of another branch of the tradition, that is to say, if they started producing fac- simile copies of the text handed down in this other branch of the tradition. The stemma in Figure 59 clariies Grünendahl’s understanding of the Mahā bhā rata tradition. The text of θ descends to N, where it undergoes expansion. It then descends to ν, where it again undergoes expansion. It then descends to Ś and K relatively unchanged, since Grünendahl thinks these recensions are really but the text of ν. The text of N also descends to γ, where it undergoes signiicant expansion. It thence descends to the witness manuscripts of the central recension, except Ñ4 and D2.5, which do not feature this expanded text and whose text therefore must be that of the noninlationary K recension. Accordingly, we may posit K “contamination.” Conversely, when we wish to reconstruct 23 232 PHILOLOGY AND CRITICISM θ N S ν γ σ ε rada K Nep l Maithil Bengali Devan gar Telugu Grantha Malay lam 1 V1 T1.2 except except D2.5 Figure 59 The expansion of the text according to Grünendahl the archetype, all we must do is proceed in reverse order, following the least inlated line of the tradition. First we start with the shortest text of the Mahā bhā rata (Ñ4 not Ś1, as Grünendahl reminds us), then we proceed up the stemma, progressively discarding the additional passages of this version as well. How exactly we are to recognize the additional passages as additions, Grünendahl does not explain. Finally, we take this text, puriied of its additions, as the text of the archetype. This process, of course, only works because, in Grünendahl’s mind, the tradition itself provides a guide to the correct line of descent. Each time a manuscript is copied, it is either copied pure, in which case it sets forth the archetype and continues the tradition, or it is copied impure, in which case it does not set forth the archetype and dead-ends the tradition. The Mahā bhā rata tradition is thus composed of open and closed branches, and all the editor must do is follow the longest open branch available to him and he will reach the archetype. The stemma in Figure 60 clariies the transmission process as Grünendahl understands it. Manuscript θ is copied at N and also at S, where it undergoes inlation, dead-ending the tradition. N is copied at ν and γ, where it undergoes inlation, both times dead-ending the tradition, but its text is transmitted to Ñ4. ν and γ are again copied (at Ś and K and 23 CONFUSIONS REGARDING CLASSIFICATION 233 θ N S ν γ σ ε rada K Nep l Maithil Bengali Devan gar Telugu Grantha Malay lam 1 V1 T1.2 Figure 60 Open and closed branches of the tradition ε and D, respectively) but each time it is not the archetype that is set forth but the new inlated traditions. Likewise, S is copied at σ and M, but these are also not apographs of the archetype but of the inlated tradition S represents. Hence only Ñ4 preserves the original tradition. The tradition is composed of open and closed branches and Ñ4—N— θ as the longest open line will yield the text of the archetype. Of course, Ñ4 itself has undergone some expansion but these expansions can be imagined as closed branches of the tradition branching of from Ñ4 so that all we need to do is brush aside these branches and hold on to the core Ñ4 and we can recover the text of the archetype (Figure 61). Grünendahl thus has not understood the simplest and most elementary principle of textual criticism, one that is the basis of all of the editor’s further operations:  how to determine iliation. He thinks that all he must do is select the shortest manuscript and discard its obvious interpolations, and he will obtain the text of the archetype. He does not realize that this procedure will not give him a critical text or even a critical text of the Nepā lī version, for the latter too would require a comparison of manuscripts and, above all, not of their interpolations, as he thinks, but of their readings. As we saw, ilia- tion is determined on the basis of the signiicant errors scribes commit in transcribing their sources. It cannot be determined using extrinsic characteristics such as the presence or absence of interpolations. Yet, as Grünendahl uses the terms “relationship,” “trans- mission,” “inluence,” “line of the tradition” and so on, it is clear that he views tex- tual transmission in terms of an original correspondence (that manifests always in the 234 234 PHILOLOGY AND CRITICISM θ N S ν γ σ ε rada K Nep l Maithil Bengali Devan gar Telugu Grantha Malay lam 1 V1 T1.2 Figure 61 Brushing aside the dead ends shared absence of certain interpolations and is a sign of relationship through the source and, ultimately, through the archetype) and a secondary one (that is evidence of one— inlated—manuscript coming under the inluence of another). He explicitly declares: I think one will probably have to distinguish the relationships between K0 and K1.2 which reach far back into the transmission history from the textual form in which K0 presents itself today. It seems to me that just as inluential as the text tradition of K1, etc. for the contempo- rary form was the inluence of another line of the tradition, which is linked to the “contami- nated” K manuscript par excellence K4.227 On one hand, there is the handing down of the text, which, according to Grünendahl, occurs like a perfect incarnation. Every apograph reproduces its source perfectly: there is no corruption. This fact is responsible for the relationships between K0 and K1.2, which Grünendahl ascribes to an earlier stage of “the transmission history” and which, to remind ourselves, consist in the shared absence of certain interpolations. On the other, there is contamination, which Grünendahl always interprets as the addition of materials to the manuscript from an external source. That is why he asserts that the “relationships between K0 and K1.2” that “reach far back into the transmission history” must be distinguished from the textual form in which K0 presents itself today.228 This 235 CONFUSIONS REGARDING CLASSIFICATION 235 form is not determined by the textual tradition of K1, but by “the inluence of another line of the tradition […] K4.”229 In contrast to the relationships between K0 and K1.2, which, Grünendahl argues, manifest in the shared absence of interpolations, K4’s inlu- ence on K0 manifests in the shared presence of interpolations.230 Grünendahl evidently does not realize that even without K4’s inluence, and even if K0 and K1 were apographs of the same source, they would still be distinct manuscripts. Even if the same scribe produced K0 and K1, they would still not be identical. Manuscript copying is a manual process and hence subject to human error. Grünendahl seems to think that scribes were not copying manuscripts by hand and thus introducing hundreds of minute errors, but that they were producing facsimile copies and adding passages to them. Consequently, he posits an identity between K0 and K1 (they are facsimile copies) and attributes their diferentiation to K4. How else could facsimile copies difer if not through an external inluence? Moreover, he declares: “The link between K0 and K4 manifests clearly in a signiicant number of shared textual additions. […] I consider the hypothesis that these present additions are text-historically a more recent characteristic of K0 than the previously mentioned characteristics which are not present in K1 as well more than likely.”231 The supposition that “present additions” must be more recent than additions that are “not present” only makes sense if we assume that the latter—which manifest in the fact that two manuscripts share the same text—are evidence of an original inheritance, whereas the former are a sign of later contamination.232 It is on this model and this model of textual criticism alone that Grünendahl’s further assertions hold. Once we grasp Grünendahl’s model of textual criticism, we can also understand the reasons for his remaining errors. As in the irst part of his article, where he attempted to ana- lyze the textual form of K0 into two traditions or two layers, the errors in the second part also arise from his belief that iliation can be established not on the basis of shared errors, but on the basis of agreements in truth. In fact, his errors are a direct consequence of his aversion to looking at readings. We focus on two of Grünendahl’s claims: (1) that Sukthankar erred in identifying his K group as a separate class on the basis of their common readings and, (2) that Sukthankar should have assimilated his K1 manuscript to the S ́ ā radā codex S1. ́ According to Grünendahl, Sukthankar’s decision to identify his K group as a sepa- rate class was false insofar as their readings proved later not to be supported by the other manuscripts. By implication, Sukthankar should have based his constitution of the K version on readings that were accepted in the constituted text as well. The K manuscripts, we have noted, are a group of Devanā garī manuscripts, whose text demonstrates features in common with the Ś ā radā codex, leading Sukthankar to conclude that they were copies of a Ś ā radā source. Sukthankar separated these manuscripts from the remaining Devanā garī manuscripts (which were mainly either misch-codices or highly contami- nated manuscripts of the vulgate) and assigned them their own group (K). These two groups (Ś and K) were the grounds for his assumption of a ν hyparchetype, “the lost archetype of the NorthWestern group.”233 As the Ś and K manuscripts were observed not to have the D manuscripts’ defects and their text was frequently purer than that of the other groups, in cases where the northern manuscripts were opposed to the southern or no pattern of agreement could be discerned, Sukthankar relied on their evidence to reconstruct the reading of the archetype. 236 236 PHILOLOGY AND CRITICISM Grünendahl is aware of this. He notes that Sukthankar grants “central signiicance to the two versions derived from ν (Ś ā radā and K) for his constitution of the text”234 and hence elects to target this premise. He does so in two stages, irst questioning Sukthankar’s basis for constituting a K version on the grounds that he lacked a Ś ā radā exemplar for many stretches of the text that would justify constituting such a version, and then arguing that the K version might anyway have been assimilated to the Ś ā radā version.235 The starting point for his criticism is his observation that Sukthankar justiies his identiica- tion of the K version, albeit only in part,236 with 35 readings, “ ‘selected at random.’ ”237 Grünendahl rejects this argument. Noting that it is “striking that the Ś ā radā manuscript S1, ́ which is fragmentary but nonetheless very signiicant for the evaluation of the ‘K version’ [Grünendahl’s quotation marks], was only available for 3 of these 35 passages,” he argues that “with respect to his [Sukthankar’s] further argument, it would have been preferable ́ 238 had the editor’s ‘random’ selection alighted on passages that were also present in S1.” Grünendahl’s skepticism regarding this choice of readings is understandable given that he thinks the editor is only justiied in identifying a K version if he possesses a Ś ā radā exemplar.239 However, the argument is misguided. Sukthankar’s aim in this section is not to demonstrate the K manuscripts’ ainity to the Ś ā radā version. Rather, he is only interested in showing their relationship to each other. From his perspective, it is irrelevant whether he can demonstrate that their readings are also the readings of Ś1. Even if his choice alighted on sections for which the Ś ā radā codex was available, this would not have afected his general conclusion. It would only have reinforced his impression that the K manuscripts are closely allied with the Ś ā radā version, which he demonstrates next. Thus, following his list of 35 readings demonstrating that the K manuscripts con- stitute a distinct group, he presents 15 readings that “document” “the ainity between Ś1 and K1”240 and thereafter 25 readings that show that “Ś1 and K stand […] together against all other manuscripts (barring, of course, conlated specimens).”241 Grünendahl ignores the evidence of the irst list and dismisses the second. According to him, the list does not prove the K manuscripts’ close relationship since their readings are frequently shared with D2 and D5.242 We have already seen the problems with this latter argument, which constitutes the crux of Grünendahl’s argument from extensive contamination. We therefore focus on a second criticism he raises of Sukthankar’s list of 35 readings illus- trating the close relationship between the K group manuscripts. He writes: “[Even] if one ignores smaller inconsistencies in the list, one must note that none of the 35 readings of the K manuscripts has been included in the constituted text.” He further argues, “It can hardly have been reconcilable with Sukthankar’s editorial principles to grant such poorly attested readings such signiicance, unless it be that a greater authenticity could be claimed for them than for the others,” but this “was clearly not the case here.”243 Had Grünendahl relected on this statement, he would have realized the absurdity of expecting the editor to use readings attested in the archetype to identify his versions. At this stage, the editor is only interested in identifying the relationships between manuscripts and he will use all available information to this end. The question of which readings will be included in the constituted text is irrelevant at this stage—indeed, it cannot even arise. For instance, if we seek to identify the southern manuscripts as a group, we will make a list of their shared readings, even if it later turns out that none of them belongs in our constituted text (for 237 CONFUSIONS REGARDING CLASSIFICATION 237 we have chosen to give preference to the reading of the northern recension whenever the two recensions conlict). The latter question is irrelevant for the identiication and evaluation of manuscripts as members of a group. By Grünendahl’s logic, we could not use these southern readings. In fact, we could not identify a southern group except for readings of the southern group that also enter into the constituted text.244 Grünendahl is efectively asking Sukthankar to identify a set of readings that identify the K manuscripts as a unique group (readings not shared with the other groups), but these readings should also be accepted into the constituted text as the readings of the archetype. This would only be possible if our constituted text was based exclusively on the K group’s readings. Then, every time the other groups difered from K in their reading, we would assume that these readings are either not derived from K or are corruptions of the reading of K, but in no case would we undertake a comparison between the groups to determine the reading of their presumed archetype. This is equivalent to treating K itself as the archetype (or, rather, treating the reading shared by the largest number of K manuscripts as that of the archetype and the idiosyncratic readings of individual K manuscripts as variants) and the other manuscripts either as descendants of K or, where their readings cannot be derived from K, as a diferent tradition altogether. Since the only thing these latter manuscripts share with each other is the readings inherited from K and since in most cases we can reconstruct K’s reading without their assistance (its reading is the one shared by the largest number of K manuscripts or the one that best explains those of the remaining manuscripts of this group), this means these manuscripts are worthless for reconstructing the archetype. We would have the stemma as indicated in Figure 62. θ (= K) N S ν γ σ ε rada K Nep l Maithil Bengali Devan gar Telugu Grantha Malay lam 1 V1 T1.2 Figure 62 Hypothetical stemma with K as the archetype 238 238 PHILOLOGY AND CRITICISM Table 6 K contamination and the fragmentation of the Mahā bhā rata tradition K Ñ V B D T G M archetype archetype archetype archetype archetype archetype archetype archetype (that is, θ) K0–6 Ñ1–3 V1 B1–6 Da1.2, T1–2 G1–6 M1–5 Dn1–3, D1–7 In this stemma, the text of the archetype descends unchanged to K (this is the sole condition under which all the readings of K can also be the readings of the archetype). In contrast, all the other manuscripts are treated as composites of K and a diferent tradition altogether (speciically, the expansion the text underwent in this branch of the tradition). We could also have included dotted lines into each of these hyparchetypes or witnesses, but as it makes no diference whether the source of contamination is external or is gen- erated within the respective manuscript, this plays no role. Since these manuscripts are no longer useful for reconstructing the archetype (they are at most useful as testimonia, whose agreement with K will provide additional conirmation for the reading of the archetype), Grünendahl has efectively reduced Sukthankar’s stemma to just one branch of the tradition, the line θ—K. The task of reconstructing the reading of the K archetype (θ) is now a separate task from reconstructing the reading of the archetype of the other groups, presuming they even had a common archetype. Since their common features can now be explained in terms of contamination with K, we would not even include them in the same stemma. There would be no need to assume a common ancestor, since con- tamination with K would account for the majority of their shared readings.245 We would have the situation outlined in Table 6. Although in the preceding diagram, we included all the other manuscripts in the same stemma, treating them as descendants of the archetype θ and using lines with a small bar at their base in place of dotted lines to indicate contamination,246 actually each manuscript group now has its own archetype and since the expansion in that archetype is as important for its text as the common inheritance from θ/K, there is no reason to treat them as descendants of the archetype of K. In fact every manuscript group can now be entered as a distinct line on our stemma, connected to the K arche- type only through a dotted line. Conversely, if we wanted to reconstruct the archetype of these manuscripts, we would proceed up this line, discard obvious interpolations and end up with the (expanded) text of this version. In fact, since the fragmentation of the tradition is attributed in this model not to corruption but to expansion of the text, to arrive at the archetype of all these versions (θ), we could take any archetype, discard its speciic readings (they can be treated as expansions of the archetype θ, just as the diferent manuscripts were treated as expansions of their respective arche- type), and we would arrive at the common archetype of all these manuscripts.247 The stemma in Figure 63 clariies how reconstruction would proceed on this model of textual criticism. 239 CONFUSIONS REGARDING CLASSIFICATION 239 θ N S ν γ σ ε rada K Nep l Maithil Bengali Devan gar Telugu Grantha Malay lam 1 V1 T1.2 Figure 63 Understanding Grünendahl’s model for reconstructing archetypes K is on this scenario the shortest text and thus incarnates the readings of the arche- type in pure form. To reconstruct the archetype, we would simply follow the longest unbroken line K—N—θ. As all the other branches of the tradition are also linked to the archetype (since the pure text also descends unchanged to them, though it then undergoes expansion), theoretically we could also begin from any other manuscript group. For instance, beginning from D, we would discard its expansions and reach the text of γ. We would then discard γ’s expansions and reach the text of N and hence θ. Any manuscript on this model of textual criticism, not just Ñ4 or K, would lead to the archetype. The source of the error is clearly Grünendahl’s peculiar model of textual criticism. Earlier we saw that he imagines the evolution of the tradition in terms of two processes and two processes only—descent of the text like a perfect incarnation and expansion— and now we see that only on this understanding of textual criticism would K’s readings also be the readings of the archetype. In fact, the readings of every branch of the tradi- tion will be the readings of its speciic archetype—the expanded text handed down in this branch of the tradition. From Grünendahl’s perspective, the fact that K’s readings are not included in the constituted text is a serious failing because, except for the expansions, his model presumes the doubling of the text of K (and of every other version) and the archetype. He thereby overlooks that by asking that K’s readings be doubly attested (once 240 240 PHILOLOGY AND CRITICISM as K’s readings and once as the constituted text’s readings), he has rendered the other manuscripts’ readings superluous. If K’s readings are the readings of the archetype, why do we need the other manuscripts? And if K’s readings are the readings of the arche- type, how can K be a version of the archetype, given that the tradition’s fragmentation into versions is due to the fact that each scribe introduces unique errors (errores signiicativi) while producing his apograph? Since a version distinguishes itself from the archetype (and from all other versions descended from it) in respect of the unique errors introduced and transmitted in that branch of the tradition, the demand that the unique readings that identify a version as a version also be attested in the archetype is doubly mistaken. They can either have been introduced in the other versions via contamination (in which case they will not serve to distinguish versions) or they can be unique to K (in which case they will likely not be the readings of the archetype). There is only one situation in which Grünendahl’s demand can be met, namely, if all other versions underwent changes and the true reading was transmitted in this branch of the tradition alone. In this case, this branch would distinguish itself from all others and preserve the reading of the archetype, but this is an exceptional case, which arose in the critical edition only due to the peculiar conditions of the Mahā bhā rata tradition.248 In any case, it has become clear that even if this model of textual criticism supports Grünendahl’s claim that K’s readings should also be attested in the archetype, it does not strengthen the evidence for them. As the preceding stemma demonstrates, the text of K is identical with that of θ. Grünendahl’s demand that K’s readings also be the readings of the constituted text is not only impossible in practice but also self-defeating. If the constituted text is by deinition based on K’s readings, the demand adds nothing to the original demand that K have a unique reading against the other manuscript groups. We lomahars·an·aputra ugraśravāh· sūtah· paurān·iko naimis·āran·ye śaunakasya kulapater dvādaśavārs·ika satre K V1 B D lomahars·an·apada D10 lomahars·an·aputrastu naimis·āran·yavāsinah· K3 romahars·an·aputra ugraśravāh· S samāsı̄nān abhyagacchad brahmars·ı̄n sam·sitavratān vinayāvanato bhūtvā kadā cit sūtanandanah· K S om· sukhāsı̄nan̄ K1 mahars·ı̄n K (exe. K1) brahmars·ı̄n […] vinayāvanato D6–8 mahars·i D14 sukhāsı̄nān V 1 B D (D 13 mss.) abhyagamat D6 śam·sita° K0.3.5 °nandat D8 śam·śita° K4 śam·sita° Dr3 .r4 D1.14 śam·śita° D2.4 (marg.) Figure 64 Stemma lectionum of verses 1.1.1A and 1.1.2 241 CONFUSIONS REGARDING CLASSIFICATION 241 can illustrate this with a simple example—the stemma lectionum for verses 1.1.1A and 1.1.2 of the Mahā bhā rata is as shown in Figure 64. In the irst verse, the K manuscripts (except K3) give a unanimous reading (lomaharṣaṇaputra ugraśravāḥ ). This reading has been accepted into the constituted text, but it cannot demonstrate that the K manuscripts form a distinct group since it is also attested in V1, B and D (except D10). In the second verse, irst hemistich, the K manuscripts (except K1) again give us a unanimous reading (samāsın̄ ān abhyagacchad). Once again, this reading has been accepted into the constituted text, but it cannot dem- onstrate the K manuscripts as a distinct group since its inclusion in the constituted text is based on cross-recensional agreement (K and S having one reading against K1, V1, B and D). In the second verse, second hemistich, the K manuscripts (except K1) give a unanimous reading (maharṣın̄ ), and this time the reading is indeed unique to this group, but it has not been included in the constituted text (the editor giving preference to the agreement of K1 with S and, additionally, V1, B and D). If we assume that in the last of these cases (second verse, second hemistich), the reading of the K group was accepted as the reading of the constituted text, it is clear that Grünendahl’s demand does not strengthen the evidence for these readings. K’s readings remain K’s readings. Contrary to Grünendahl’s expectation, their inclusion in the constituted text is no longer evidence of greater attestation, since this inclusion is based solely on their occurrence in the K manuscripts. Grünendahl is misled by his suspicion that everything in the constituted text possesses greater authority. He thinks that the editor ought to have only used readings included in the latter when drawing up a list of readings unique to K but overlooks that this condi- tion can only be fulilled if the two are identical. But in that case, the condition is super- luous: if the K readings are included in the constituted text on the authority of the K manuscripts, then it adds nothing to their authority to say that they are also the readings of the constituted text. The situation remains the same in both the actual case and our hypothetical case:  K and ~K give distinct readings. The only change is that whereas in the irst instance we ruled that we prefer the readings of ~K, in this case we have opted to adopt the reading of K for the constituted text. The attestation for K remains the same: K. There is no diference between identifying a K group based only on the readings unique to K that also made their way into the constituted text and constituting a K group based only on the readings unique to K irrespective of whether they made their way into the constituted text or not.249 Grünendahl’s second claim—namely, that Sukthankar should have assimilated his K1 manuscript to the Ś ā radā codex—is also a direct consequence of this peculiar model of textual criticism. As we saw, he does not establish iliation on the basis of descent but on the basis of length. For him, it is not the inherited errors of transcrip- tion that are decisive but the convergence in extent between manuscripts.250 The latter holds the key to their “real” relationship, which, for Grünendahl, cannot be one of iliation (as we have seen, there is strictly speaking no distinction between sources and their descendants in his model of textual criticism), but only one of duplication. Thus, if two manuscripts approximate each other in their extent, Grünendahl thinks that they should be classiied together, since (he reasons) they must reproduce a text of a similar 24 242 PHILOLOGY AND CRITICISM type (that is, of the non-expansive rather than expansive type). From his perspective, it is thus inexplicable that Sukthankar entered K1 into his stemma as an independent witness rather than assimilating it to Ś1. He notes that “for Sukthankar, it is beyond doubt that the source of K1 must have been a Ś ā radā manuscript. Via reference to pronounced distinctions (loc. cit.) he shows that Ś1 however does not enter into question [as the source].”251 Grünendahl challenges Sukthankar’s decision to list K1 separately because, as he puts it, “it must be emphasized that K1 approximated Ś1 more in terms of extent and textual form than any other K manuscript. In those textual sections for which Ś1 is not available owing to its poor state of preservation, K1 distinguishes itself from the other K manuscripts through noticeably many unique readings.”252 As earlier, he invokes the convergence in length between the two manuscripts, which he thinks is a decisive argument for their relationship. That he simultaneously criticizes Sukthankar for arguing “less in terms of its quality than in terms of its textual extent”253 is typical of his tendency to attribute his confusions to others. Noting that K1 is “closer to the Ś ā radā manuscript Ś1 than the rest of the Devanā garī manuscripts comprised in the K version,” he rhetorically raises the question: What could then have prevented Sukthankar from including Ś1 and K1 in one “version” or group? I believe this must be attributed less to the diferences that are doubtless present between the two manuscripts than to the Schriftartprämisse, which he grants absolute priority here, even though he, as was said, has no doubt that K1 is directly descended from a Ś ā radā manuscript.254 From Grünendahl’s perspective, of course, there are no arguments against including Ś1 and K1 in one group. He explicitly asserts: Even if one wants to hold on to the deinition of the versions in terms of the script, one can, without surrendering one’s principles, in the case of such a clear state of afairs assign the Devanā garī manuscript K1 to that “version” to which the source in all probability belongs (here:  Ś ā radā ). That he [Sukthankar] precisely does not do this has the advantage for his argumentation that he gains an independent—at least at irst sight—witness for the textus simplicior (Prolegomena, p. l).255 We already saw that neither Lüders nor Sukthankar classiied manuscripts by their script so there can be no question of “grant[ing] absolute priority” “to the Schriftartprämisse here.” Since the versions are deined in terms of their text and not their script, there can also be no question of “hold[ing] on to the deinition of the versions in terms of the script.” If Grünendahl nevertheless insists that Sukthankar tried to hold on to his principle by classifying K1 with the other Devanā garī manuscripts of its K group even though, in terms of extent, it is closer to Ś1 to “gain” “an independent […] witness for the textus simplicior,” it is because he does not understand how iliation is established. That he can simultaneously impute intellectual dishonesty to Sukthankar for claiming two—inde- pendent—versions where he had only one is a further enticement for someone who cannot forgive Sukthankar for succeeding as a textual critic where he failed.256 The 243 CONFUSIONS REGARDING CLASSIFICATION 243 Ś ā radā and K manuscripts are related, but they are related through the hyparchetype ν. In fact, ν was deined as “the lost archetype of the NorthWestern group,”257 meaning: the latest ancestor Ś1 and K1 have in common. Thus, K1, though a Devanāgarī copy of a Ś ā radā original, cannot be assimilated with Ś1 under a hypothetical Ś ā radā arche- type. If we look at these two groups or versions in isolation, it is clear why K1 cannot be treated as a member of a hypothetical Ś ā radā group. In Sukthankar’s stemma, Ś1 and K are related through a hypothetical ancestor ν. K is an abbreviation for a group of manuscripts K0 through K6 whose relationship cannot be further analyzed and that are therefore treated as more or less perfect representatives of a hypothetical ancestor K. Sukthankar thought that the manuscripts of the K group were copies of Ś ā radā sources but they were not directly descended from the Ś ā radā codex Ś1 and hence entered them into his stemma as a separate branch of the tradition.258 Since these manuscripts were probably copied from Ś ā radā sources, it is likely that ν, their common source with Ś1, was a Ś ā radā manuscript. We could replace the siglum ν with the name “Ś ā radā ,” but the meaning of this term has changed: it no longer refers to the version of the text contained in the Ś ā radā codex Ś1 but to the hypothetical ancestor of Ś1 and the K group. The relationship of Ś1 to K remains the same: they are distinct branches of the tradition, whose latest common ancestor is the hyparchetype “Ś ā radā ” (our former ν). Our change has merely emphasized the fact that both branches transmit a text of the “Ś ā radā ” type, though in two slightly diferent versions, which we can label Ś1 and Ś2 (= K). Grünendahl’s argument is more complex: he seeks to subsume the K manuscripts not under a Ś ā radā archetype (our former ν, albeit now relabeled “Ś ā radā ” to reflect its probable script) but under the Śāradā version itself. This means he must either treat Ś1 (by definition the Ś ā radā version) as K1’s ancestor, which Sukthankar found was not the case, or he must redefine “Ś ā radā ” such that it refers neither to Ś1 nor to the hypothetical archetype of all Ś ā radā manuscripts (Sukthankar’s ν) but to a group of manuscripts that contain a similar text but whose relationship to each other cannot be further analyzed. Grünendahl does not realize this for, in his model of textual criticism, no difference between a group and a version exists. As his comment “What could then have prevented Sukthankar from including Ś1 and K1 in one ‘version’ or group?” shows, he treats the terms as synonymous and this is because in his model of textual criticism there is no individuation between manuscripts during copying. The apographs of a source reproduce its text perfectly in all respects (provided there is no contamination), so that, whereas we speak of the manuscripts of a group, which approximate the text of their version to various degrees and the text of which must be inferred from their features, this distinction (between a group and a version) no longer has any meaning for him. The version is no longer an ideal that must be inferred from the group characteristics: it is present in each of the manuscripts of the group. Manuscripts no longer need to be grouped because their relationship cannot be further analyzed. Rather, the term group now refers to a set of manuscripts all of which contain the identical text and are there- fore superimposed on each other. This text is the text of the version so that the concept of a group and the concept of a version coincide.259 To understand what 24 244 PHILOLOGY AND CRITICISM θ N S ν γ σ ε rada K Nep l Maithil Bengali Devan gar Telugu Grantha Malay lam 1 V1 T1.2 except except D2.5 Figure 65 The expansion of the text according to Grünendahl Grünendahl means, let us look again at Figure  53. We reproduce the stemma in Figure 65 for convenience. We restricted ourselves earlier to the descent of the archetype to the individual groups, but this model of textual criticism has a further consequence. If the text descends unchanged from ν to Ś and K, every manuscript in the K group will contain the exact same text (recall that for Grünendahl no corruption occurs during copying:  every apograph reproduces its source perfectly). Not only will every manuscript within the K group coincide; they will also coincide as a group with the K version, abolishing the distinction between the group and the version. The stemma will look like  the one in Figure 66. The manuscripts within the K group will coincide not only with each other and with the K version. Since the text ν descends unchanged to Ś1 as well as to K, they will also coincide with Ś1. This explains why Grünendahl considers it inexpli- cable that Sukthankar entered K1 into his stemma as an independent witness. He insinuates that Sukthankar’s aim was to gain, as he puts it, “an independent—at 245 CONFUSIONS REGARDING CLASSIFICATION 245 θ N S ν γ σ ε rada K Nep l Maithil Bengali Devan gar Telugu Grantha Malay lam 1 V1 T1.2 K0 K1 K2 K3 K4 K5 K6 Figure 66 Abolishing the distinction between group and version Note: Ñ4 and D2.5 are not shown on this stemma as irrelevant to this stage of the argument. least at irst sight—witness for the textus simplicior,” but this is because he does not see that K1 is an independent witness. It is so not only or not primarily in the sense that Grünendahl takes it—that is, a witness capable of corroborating Ś1—but in the sense that, as a manuscript whose relationship to Ś1 though not to the other K manuscripts can be stemmatized, it must be entered individually into the stemma. The stemma presented earlier is only hypothetical, since Grünendahl distinguishes between K1 and the other K manuscripts, but it illustrates the point. His model of textual criti- cism does not allow for the individuation that occurs when manuscripts are copied so that all copies of a source (provided no contamination occurred) are facsimile copies of it and duplicates of each other. If we wish to really grasp his argument for assimilating K1 to Ś1 we can replace this stemma with the two stemmata shown in Figures 67 and 68. In the irst step, Grünendahl breaks up the K group into two:  a non-expansive tradition, whose text is inherited from ν (and is thus identical with that of Ś1) and an expansive tradition, whose text is inherited from the central subrecension γ. 246 246 PHILOLOGY AND CRITICISM θ N S ν γ σ ε rada Nep l Maithil Bengali Devan gar Telugu Grantha Malay lam 1 K V1 T1.2 K1 K0 K2 K3 K4 K5 K6 Figure 67 The fragmentation of the K group into K1 and K0.2–6 according to Grünendahl The manuscripts K0 and K2–6 are distributed into various subgroups, with D2.5 among them, but Grünendahl’s criterion remains the number of insertions—that is, whether a manuscript belongs to an expansive or a non-expansive tradition. K2–6 are “grouped” not because their internal relationships are unclear, but because they are all of the same type.260 In a second step, noting that the texts of Ś1 and K1 coincide, he replaces them with a pair of coincident manuscripts as shown on the stemma in Figure 68. Finally, as there is no longer any distinction between Ś1 and K1 in his scheme, he collapses the two branches into one and replaces the overlapping manuscripts with a single one (not shown on this diagram). Grünendahl is therefore right when he plain- tively inquires, “What could then have prevented Sukthankar from including Ś1 and K1 in one ‘version’ or group?” There is nothing in this model of textual criticism to prevent Sukthankar from including Ś1 and K1 in one version, only it is not the model Sukthankar used. It is also not the model any intelligent critic would use. Only Grünendahl, with his peculiar aversion to studying readings, thinks that insertions are a criterion either for selecting the manuscript to edit or for classifying manuscripts or for constituting versions. 247 CONFUSIONS REGARDING CLASSIFICATION 247 θ N S ν γ σ ε rada Nep l Maithil Bengali Devan gar Telugu Grantha Malay lam K1 1 V1 T1.2 1K1 K0 K2 K3 K4 K5 K6 Figure 68 Collapsing Ś1 and K1 into a single version Note: The manuscripts are partially overlapped and not coincident to illustrate that they are overlaid. The overlapping circles should be imagined as a three-dimensional representation: K1 is placed directly behind Ś1, and the whole viewed at an angle. The Argument from Independent Recensions Since widespread contamination between Mahābhārata manuscripts makes it impossible to draw up a stemma, the editors should have concentrated on a regional recension and produced an edition using one manuscript, preferably the shortest, as the base. Refutation of the argument: The argument overlooks the fact that, even if the editors had chosen to pro- duce a critical edition of one of the regional recensions, they could not have done so without considering its place in the overall tradition. Further, if contamination makes it impossible to produce a critical edition of the Mahābhārata tradition as a whole, this is especially true of a critical edition of a regional recension, since con- tamination was most pronounced between the manuscripts of a recension. Grünendahl’s proposed edition (based on the Nepālı̄ manuscript NAK 5/356) would be not a critical edition but a pseudo-Bédierist edition. He could neither draw up a satisfactory stemma nor proceed in cases of crux or where the Nepālı̄ manuscripts contained inconsistent readings. His edition opens the loodgates for conjectural criticism, which is all he really seeks to do. Grünendahl’s fourth and inal argument against Sukthankar’s stemma is one we call the argument from independent recensions. In this part of his article, Grünendahl introduces no new principles but concentrates on applying those already introduced to further undermine 248 248 PHILOLOGY AND CRITICISM Sukthankar’s classiication of manuscripts. As these arguments also cast light on his pro- cedure for building stemmata, we also consider them here. The key contention Grünendahl raises in this second part is:  as the Mahā bhā rata never existed as a single text, the codex unicus that the critical edition seeks to recon- struct, the search for such a text must be abandoned.261 Instead, attention must focus on reconstructing one of the epic’s regional recensions—above all, the Nepā lī recension, which he advocates as the best recension. There are two problems with this claim: 1. It is counterfactual, since Sukthankar and all his successors had no doubt that all our Mahā bhā rata manuscripts were descended from a single source—by deinition, the archetype of the tradition.262 Thus, whatever one thinks about the tradition prior to the archetype—it was oral, it was luid, it was manifold (and we have seen all these claims in the work of Bigger, where they led him into needless confusions)—the fact remains that genealogical reconstruction demonstrates that there was an archetype and that all our witnesses are copies, however distant, of this source. 2. Any reconstruction of a regional recension that did not consider the complete Mahā bhā rata tradition would lead to arbitrary solutions as regards the choice of the manuscript to edit, the readings to reprint and also the “improvements” intro- duced into the text. Such an edition would be neither a Lachmannian edition nor a Bédierist edition. At best,263 it would be a pseudo-Bédierist edition—that is, an edition that consists in “the transcription of one of the manuscripts carrying the work, with corrections [occurring] in variable measure.”264 This edition difers from the true Bédierist edition, which is an edition “aspiring to absolute respect for the reality of a single manuscript, after an exhaustive analysis of the varia lectio, and that [is] rig- orously consistent with this principle,”265 in that it produces a text “that [is] neither the faithful and synchronic reproduction of a single witness nor the reconstruction of a former stage of the textual tradition.”266 From the perspective of someone with no experience in dealing with a complex tradition, this return to a “pre-scientiic practice” (Leonardi) may appear appealing, but it is absolutely unacceptable to those who understand editing in terms of the three requirements Leonardi speaks of (scientiicity, an eye for diachrony and legibility).267 We later consider the problems with what has been termed “criticism of the base man- uscript” (Leonardi). Before we do so, let us irst consider Grünendahl’s arguments for the Newā rī manuscript Ñ4. He makes two claims: irst, the Newā rī group as a whole is superior to the Kaśmīrī group; second, Ñ4 in particular is superior to K1, the best K manuscript. We look at each of these in turn. Grünendahl’s irst argument for the Nepā lī group’s superiority is that K1, which Sukthankar considered the best K manuscript and a copy of a Ś ā radā source, approximates the Devanā garī manuscript D5, as several common absences attest. His precise argument is as follows: “The similarities between D5 and K1—in Sukthankar’s assessment, the ‘purest’ representative of the ‘K-version’—are also remarkable, especially as these manifest in the not being present [that is, the absence] of textual additions attested in almost all other manuscripts of the northern recension.”268 As evidence, he lists the following passages: 249 CONFUSIONS REGARDING CLASSIFICATION 249 296* not present in D5, K1 [and Ñ4; southern recension] 564* not present in D5, K1, K2 [and Ñ1.4; southern recension] 396* not present in D5, K1, Ś 1 [and Ñ2.4; G3.6, T2, M] 259* not present in D5, K1, K0 [and Ñ3.4; southern recension] the sole example of a textual addition that is present: 2080* present in D5, K1, Ś 1, D2 (!) Source: Reproduced verbatim from Grünendahl, “Zur Klassiizierung von Mahā bhā rata- Handschriften,” 120. We have already seen that this argument does not hold. The absence of interpola- tion, even if the same interpolations are absent, in two manuscripts proves nothing for their closer relationship, because this absence is a feature of the original. Yet Grünendahl thinks that the circumstance that K1 and D5 both lack four passages found in most of the northern manuscripts is a telling argument for their relationship. He claims: Precisely because this is a case, above all, of not present textual additions, these similarities, in my opinion, cannot be explained away, as Sukthankar attempts to do, with the fact that they entered into the “composite manuscripts” D5 (and D2) due to contamination (conlation). This may serve as an explanation for present textual additions but hardly for not present [textual additions]. Hence, in my understanding, the above inding only permits the conclusion that D5 (and probably also D2) were, in their core, very close to K0.1(4), even though this [core], as is also the case for K0 and especially for K4, appears to have been overlaid by the later inluences of other branches of the tradition.269 This passage reveals the extent of Grünendahl’s confusion. On one hand, he imagines that if two manuscripts do not contain the same interpolation, they must be related; on the other, he thinks that, in contrast to the presence of interpolations, which can be attributed to contamination, the absence of interpolations cannot. Further, he reasons that if the simultaneous absence of an interpolation in two manuscripts cannot be attributed to contamination, it must mean that “in their core” these manuscripts were very close to each other. In contrast, the diferences between them must be explained as the result of later additions, for how else (he reasons) would they difer, if not due to an external inluence? He thereby overlooks the fact that “secondary relationship, that is[,] descent from a common source more recent than the ‘original’ (the original of all our known MSS.) […] can only be made plausible on the ground of common agreement in secondary features. Any amount of inter-agreement in features which are, or may plausibly be argued to be, original features will prove nothing; for they may on that assumption be directly inherited from the source of all our MSS.”270 The simul- taneous absence of 296*, 564*, 396* and 259* from K1 and D5 only means these passages were not contained in the archetype. It does not mean the two manuscripts are related through a more recent ancestor. Grünendahl obviously imagines a situation as shown in Table 7, where manuscripts are grouped depending on whether they contain these passages. The correct way to read the table is that the passages are characteristic of the central recension manuscripts (though not all of them) and that they are interpolations into the 250 250 PHILOLOGY AND CRITICISM Table 7 How Grünendahl imagines the classiication of manuscripts Contain the passages Do not contain Also do not contain the passages the passages 296* K0.2–4, Ñ1–3, B, Da, Dn, D1–4.6–7 K1, D5 Ñ4 and S 564* K0.3–4, Ñ2.3, B, Da, Dn. D1–4.6–7 K1.2, D5 Ñ1.4 and S 396* K0.2–4, Ñ1.3, B, Da, Dn. D1–4.6–7, Ś 1, K1, D5 Ñ2.4, G3.6, T2, M G1.2.4.5, T1 259* K2–4, Ñ1.2, B, Da, Dn. D1–4.6–7 K0.1, D5 Ñ3.4 and S manuscripts in which they are found. K1, D5 and the southern recension manuscripts in this case preserve the correct text, while the presence of the passages in the manuscripts of the ν recension that contain them must be attributed to contamination with the central recension. Grünendahl’s interpretation difers: from the fact that the passages are missing in K1 and D5, he argues that K1 and D5 must be closely related and that, as this is a case of absence in both manuscripts, the similarity cannot be attributed to contamination as Sukthankar thought.271 Further, he argues that the passages’ absence in K1 and D5 must be attributed to an earlier stage of the textual transmission before the respective manuscripts were “overlaid” with other interpolations and, from this, he concludes that D5 “and prob- ably also D2” must, originally, “in their core” have been “very close to K0.1(4).”272 The argument is supericially appealing, but erroneous. As we have established in the preceding section, the absence of speciic interpolations does not prove manuscripts related. From the fact that K1 and D5 (as well as Ñ4) lack the same passages, we cannot conclude that they are more closely related, because another manuscript not contaminated with the central recension would also not contain the passages. That these passages are also missing in the southern recension should have alerted Grünendahl to this fact. Moreover, the cat- egorization into the interpolated manuscripts and the non-interpolated manuscripts is merely typological, not genealogical. Grünendahl has only grouped his manuscripts into two groups: the meliores and the deteriores (the better and the worse manuscripts). He has not ́ K0.1, D5 and Ñ4 constitute a family in the genealogical sense. shown that the meliores S1, He obviously thinks that they do, but this is because he overestimates the role contami- nation played. According to him, since the insertions common to K1 and D5 are a case of “not present,” that is to say, absent textual additions, they cannot have entered into the manuscripts via contamination. As contamination is ruled out as the source of the resem- blance, he concludes that they must then have been original features of both manuscripts and, from this, he posits a recent common ancestor. He is right, except that he overlooks that there is a third way the passages could be absent from all three manuscripts without being owed either to “contamination” (in the peculiar sense that he understands it, that is, as “non-contamination”) or to a more recent ancestor—namely, if they were absent from the archetype. There is no reason to posit a more recent ancestor than either θ or N to explain the passages’ simultaneous absence in K0.1 and D5.273 Grünendahl’s second argument for the “contamination” of the K group is no better. He notes that Sukthankar separated out K0 and K1 from the remaining K manuscripts K2–6 as they represented a purer tradition than the latter, and argues that, contrary to 251 CONFUSIONS REGARDING CLASSIFICATION 251 Sukthankar’s division, “K2 appears not as closely related with any other K manuscript as with K1.” He further notes that “purely supericially, this was already clear from the list on p. 108 [the reference is to his article; the table is reproduced as Table 1], where, among the K manuscripts, the total number of textual additions in K2 came closest to the number [of textual additions] in K1.”274 We have already seen that the total number of interpolations in a manuscript is a poor guide to its iliation, so we ignore this part of the argument.275 Instead, let us look at the second part of Grünendahl’s argument for the relationship of K2 and K1. As evidence of K2’s proximity to K1, Grünendahl cites the fact that “if one now considers those peculiarities in the textual additions that K2 has in common only with one other K manuscript, then here also is its proximity to K1 evident. Of the 28 textual additions of this nature, K2 has four in common with K0 (1800*, 1962*, 1418*, 1934*), two with K3 (462*, 2135*), ive with K4 (1393*, App I, 20, 1673*, 905*, 1152*) and no less than seventeen with K1,” and he thereafter lists the following passages: 1436* not present in K2, K1, Ś 1 [and the southern recension] 304* not present in K2, K1 [and Ñ4; southern recension] 536* not present in K2, K1 [and Ñ1.4; southern recension] 1422* not present in K2, K1, Ś 1 [and Ñ4; southern recension] 1142*.3 not present in K2, K1, Ś 1 [and Ñ4; southern recension] 1266* not present in K2, K1, Ś 1 [and Ñ4; southern recension] 523* not present in K2, K1, D5 [and Ñ4; southern recension] 324* not present in K2, K1, D5 [and Ñ4; southern recension] 327* not present in K2, K1, D5 [and Ñ4; southern recension] 295* not present in K2, K1, D2.5 [and Ñ4; southern recension] 518* not present in K2, K1 [and Ñ1.3.4; southern recension] 558* not present in K2, K1, Ś 1, D5 [and Ñ1.4; southern recension] 274* not present in K2, K1, D5 [and Ñ1–4; southern recension] (see, further: 285*, 564* (see above), 1559*, 1561*, 1778*) Source: Reproduced verbatim from Grünendahl, “Zur Klassiizierung von Mahā bhā rata- Handschriften,” 121. Grünendahl glosses this list of passages with the observation that “in contrast to the similarities with the other K manuscripts, in the case of those between K2 and K1 it is a case of not present textual additions, which, in my opinion, permits us to once again infer a relationship between these two manuscripts that reaches far back.”276 It is clear what he thinks to demonstrate with this list. According to him, if we consider only the interpolations that serve to unite K2 with one other K manuscript at a time, we have the situation as outlined in Table 8. In Grünendahl’s view, this list of paired interpolations not only demonstrates that K2 has the largest number of interpolations in common with K1, but also that the interpolations that unite these manuscripts are of a qualitatively diferent kind. In contrast to the interpolations K2 shares with K0, K3 and K4, which are a case of “present” interpolations (which he thinks can be attributed to contamination), the interpolations common to K2 and K1 are a case of “not present” or absent interpolations (which he thinks indicate an original ainity between the manuscripts going back to an earlier stage of their transmission history). 25 252 PHILOLOGY AND CRITICISM Table 8 Grünendahl establishes “iliation” K2–K0 K2–K3 K2–K4 K2–K1 1800* 462* 1393* 1436* (missing) 1962* 2135* App. 1, 20 304* (missing) 1418* 1673* 536* (missing) 1934* 905* 1422* (missing) 1152* 1142*.3 (missing) 1266* (missing) 523* (missing) 324* (missing) 327* (missing) 295* (missing) 518* (missing) 558* (missing) 274* (missing) 285* (missing) 564* (missing) 1559* (missing) 1561* (missing) 1778* (missing) Unfortunately, this argument is also lawed. First, the presence or absence of spe- ciic interpolations is not an argument for the closer relationship of two manuscripts. The absence of passages especially does not prove anything, since the manuscripts could be from diferent branches of the tradition or even manuscripts of diferent works. Second, the claim to relationship is not reinforced by citing large numbers of such passages.277 Third, nothing is gained by specifying that the list of interpolations should be restricted to those found between K2 and one other K manuscript at a time. Grünendahl speciies this condition because he thinks that by considering only a pair of manuscripts at a time he can better isolate the source of contamination. Thus, he reasons that if the interpolation is present only in K2 and K0, then either K2 is the source of the passage or K0 is the source of the passage or they had a common ancestor, but in either case, an external source is ruled out. He thereby overlooks the fact that both K2 and K0 could have gained the passage from a third manuscript no longer extant. The restriction is hence pointless.278 Fourth, as we saw, the idea that as the similar- ities between K2 and K1  “are a case of not present textual additions,” this permits us to “infer a relationship between these two manuscripts that reaches far back”279 is Grünendahl’s delusion, arising from his belief that manuscript copying produces fac- simile copies of the text and that interpolations can supervene on the core text only through contamination (in other words, contamination is responsible for the diferen- tiation of manuscripts).280 We later look at this issue more closely, as it holds the key to all of Grünendahl’s other errors. For now, let us consider his third and inal argument for the Newā rī group’s superiority over the Kaśmīrī. Grünendahl transitions to this next argument with the words: “now to the discussion of the ‘contaminated’ manuscripts K3–6 (Prolegomena, p. li).”281 He claims that “the standard 253 CONFUSIONS REGARDING CLASSIFICATION 253 for contamination” of the remaining K manuscripts should be “Sukthankar’s ‘pure K-version,’ primarily represented by K1,”282 and he then cites Sukthankar’s view that K3–6 appear contaminated with the γ recension. Grünendahl argues that Sukthankar’s arguments for K3–6’s contamination with γ are untenable for several reasons, which he details.283 These objections are irrelevant for the main argument. We therefore ignore them.284 Instead, we focus on his main objection, which he states as follows: One might now expect the editor to preferentially demonstrate the contamination of K3–6 in terms of textual editions whose not being present [Nichtvorhandensein] is characteristic of the hyparchetype ν. This, however, is only true with caveats for the appendix passages listed by him [Sukthankar] as the following overview demonstrates: App. I: 14 not present in K0–3, D5 [and Ñ4; T2, G3.6, M] 41 not present in K0–2, D5 [and Ñ1.4; southern recension] 42 not present in K0–2, D5 [and Ñ4; southern recension] 43 not present in K0–2, D5 [and Ñ4; southern recension] 61 not present in K0–3 [and Ñ4; southern recension] [Grünendahl concludes:] Since, as was noted, K5.6 are not available for these sections, with these passages, one can at best make the contamination of K4 (and, with caveats, of K3) plausible. However, I attach greater signiicance to the observation that the not being present [Nichtvorhandensein] of these passages is manifestly not restricted to the “core group” of the K manuscripts and hence is also not characteristic of it. Regarding App. I: 14, one could also say that the presence of this passage is not even characteristic of γ, the hypothetical source of contamination, unless one wanted to assume that the south Indian manuscripts, which also feature the passages, are also contaminated.285 To understand why Grünendahl thinks this list demonstrates K3–6 are not contaminated (or at least that the evidence for their contamination is inconclusive), we must irst under- stand how he interprets contamination. A  manuscript is usually called contaminated when it inherits its readings not from its hypothetical ancestor but in some measure from a second source. In that case, the assumption is that the scribe either kept an eye on a second source while copying or that he compared his copy with another when inished and entered some of the latter’s readings into his text. Grünendahl’s understanding of contamination is more complicated. According to him, the contamination of manuscript A is not demonstrated by listing the passages it shares with the manuscripts of another recension. Rather, he argues that “the standard for contamination” should be the purest representatives of the group that includes A. We might think that all Grünendahl means thereby is that the editor must demonstrate that the passages under consideration must be interpolations in A as they are not found in the better manuscripts of its group and hence could not have been a feature of their common archetype. But this would be incor- rect. Grünendahl interprets the demand that the editor should adopt, as his standard for contamination, the purest representatives of the group that includes A, to mean that the editor should demonstrate A’s contamination in terms of interpolations characteristically absent from the hyparchetype as represented by the purest members of the group. If the 254 254 PHILOLOGY AND CRITICISM Table 9 Grünendahl establishes that K3 is not contaminated with the passages with which it is not contaminated (hence, the editor has failed to establish the contamination of K3) Insertion Found in Also not found in Not found in App. 1, no. 14 K4 (marg.) K3 K0–2 App. 1, no. 41 K3.4 Ś 1, K0–2 App. 1, no. 42 K3.4 K0–2 App. 1, no. 43 K4 K3 K0–2 App. 1, no. 61 K4 K3 Ś 1, K0–2 interpolations absent from the latter are simultaneously not not absent from A, this means their absence, or, as he puts it, the “not being present of these passages is […] not restricted to the ‘core group’ […] and hence is also not characteristic of it.” Table 9 makes the argument clearer. The appendix passages 14, 41, 42, 43 and 61 are found in most northern recension manuscripts, with the exception of the K group. App.  1, no.  14 is found in Ñ, V1, B and D with the exception of D5 (D2 on a supplementary folio). It is also found in some southern recension manuscripts, speciically the inferior representatives of their groups T1 and G1.2.4.5.286 App. 1, no. 41 is found in Ñ2.3, V1, B and D, with the exception of D5. App. 1, nos. 42 and 43 are found in Ñ, V1, B and D (again with the exception of D5), while App. 1, no. 61 is found only in Ñ, B and D (V1 is missing for this section). Against their group’s practice, the passages are also found in two K manuscripts, namely, K3 and K4. K4 includes all ive interpolations (no. 14 as a marginal insertion), while K3 includes two (nos. 41 and 42). The correct way to read this table thus would be that K3 and K4 contain several passages not found in the better members of their group and hence that they were probably contaminated with the manuscripts containing these passages. This is, in fact, Sukthankar’s reasoning: he observes that “the contamination of K3–6 with γ is illustrated by the following passages” and then lists App. 1, nos. 14, 41–43 and 61, and several smaller insertions that suggest contamination of K3–6.287 Grünendahl’s interpretation difers. He irst declares that the editor should have “preferentially demonstrate[d] the contamination of K3–6 in terms of textual editions whose not being present is characteristic of the hyparchetype ν.” Thereafter, from the fact that, in three cases, the passages’ absence is not restricted to the ifth column (K0–2) but also extends to the third (K3), he concludes that, as “the not being present of these passages is manifestly not restricted to the ‘core group’ of the K manuscripts,” the editor has failed to demonstrate the contamination of K3–6.288 This is an original interpretation, because, parsed carefully, the statement does not say anything beyond that K3 is not contaminated with the passages with which it is not contaminated. But no one claims this! Sukthankar lists App. 1, nos. 14, 41, 42, 43 and 61 as examples of passages from the central recension contaminated in K4, and he lists App. 1, nos. 41 and 42 as examples of two passages additionally found in K3. All that Grünendahl has done is taken the passages Sukthankar cites as evidence of K4’s contamination with γ, cited the fact that K3 is not contaminated with these passages, and, from this, concluded that, as their absence is not restricted to K0–2 but is also a feature of K3, Sukthankar has “at best” (allenfalls) demonstrated the contamination of K4.289 25 CONFUSIONS REGARDING CLASSIFICATION 255 As there is no great art in demonstrating that manuscripts are not contaminated with the passages with which they are not contaminated, it is not clear what Grünendahl hopes to demonstrate with this argument. Not only for K3 but also for any random manuscript, he could identify passages with which the manuscript is not contaminated. For any con- taminated manuscript A, which contains the interpolations abcdef due to horizontal trans- mission, there will be at least one interpolation x that it does not contain. We could always argue that the contamination of A must be demonstrated in terms of x and, from the fact that A does not contain x, conclude that A is not contaminated. But this would be a bizarre procedure! Contamination of manuscripts must be demonstrated in terms of the passages with which they are contaminated. No one claims that contaminated manuscripts contain all the readings with which they could possibly be contaminated. It suices to show that a manuscript contains several readings it could not have inherited from its ancestor to demon- strate contamination. The outrageousness of Grünendahl’s demand that the standard for contamination should be the readings absent in the contaminated manuscript is not imme- diately evident because he does not state it so clearly. Rather, he opts for the strange formu- lation: the contamination of manuscripts should be “preferentially demonstrate[d] […] in terms of textual additions whose not being present is characteristic of ” the better members of their group. According to him, if the absence of these passages is not restricted to the better members, but is also a feature of the contaminated manuscript, this demonstrates that the latter is not contaminated. In other words, what Grünendahl is really asking is that manu- script A should contain all the interpolations its group α lacks, for if A also does not contain just one of these, he will rule that the absence of these interpolations is not restricted to α and, from this, conclude that A cannot be contaminated. The absurdity of this demand becomes manifest when one realizes that for A to ful- ill it, the scribe of A  must know which interpolations are absent in α (and all of the interpolations absent in α) so that he could be certain to include all of them in A, that is to say, he must have a complete overview of α! But since he cannot know which interpolations are absent in α without irst knowing which interpolations exist in general, this means he must also have a complete overview of the entire Mahābhārata tradition. In other words, he must conduct a systematic recensio! This is an absurd expectation of any scribe, since he is only interested in copying the best manuscript available to him or the one he was commis- sioned to copy. He is not interested in undertaking a pan-Indian search for manuscripts on the of chance that in the future a critic might deine “contamination” so as to ensure that no manuscript will ever satisfy the conditions. We can set aside Grünendahl’s deinition of “contamination.” On this deinition of contamination, if the entire tradition contained 100 interpolations numbered 1 through 100, and α contained interpolations 1–25, the manuscript A, to be considered “contaminated,” must contain all of the 75 remaining interpolations 26–100. This would not be a contaminated manuscript but a hypercon- taminated manuscript! Grünendahl might counter this criticism by pointing out that he only holds that A should contain the interpolations whose absence is characteristic of α in the sense of being a unique feature of this group, and not all interpolations in general. But this does not make the argument stronger. Consider the following situation: suppose that the total number of interpolations in the Mahā bhā rata tradition is 100 and these interpolations are numbered 256 256 PHILOLOGY AND CRITICISM Kc v γ S 1–25 26–50 76–100 51–75 Kc 51–75 Figure 69 Grünendahl explains “contamination” 1 through 100. Numbers 1–25 represent the interpolations found in ν, 26–50 those found in γ, 51–75 those found in both γ and S and 76–100 those found in S alone (Figure 69). Grünendahl might now argue that his view is not that the contaminated K manuscript Kc must contain all the interpolations not found in ν (26–100), but only those whose absence is characteristic of ν (51–75).290 In his view, Kc represents a contaminated manuscript if and only if it contains all the interpolations characteristically absent in ν. If even one of them is not found in Kc, this would imply that this interpolation’s “not being present is […] not restricted to the ‘core group’ […] and hence is also not characteristic of it,” with the con- sequence that the contamination of Kc is not demonstrated. But this interpretation does not help matters. The scribe could only know that he had to interpolate passages 51–75 into Kc if he knew these passages—and just these—were absent only in ν, which once again means he would have had to undertake a complete recensio of the tradition.291 The condition is thus absurd. It is also absurd for a second reason: Kc could contain a large number of interpolations from γ and S, whose absence is not a characteristic feature of ν alone (the interpolations are also absent in either γ or S, represented on the diagram by the dotted lines), but, by Grünendahl’s logic, we would ignore these passages. We would not consider Kc contam- inated if it contained a large number or even all of these passages. This is an unjustiied and illegitimate narrowing of the criteria for contamination, because Kc is no less con- taminated for containing the interpolations 26–50 or 76–100. Efectively, Grünendahl has narrowed the list of relevant instances for contamination to just the passages 51–75. He further insists that Kc must not contain just some of these passages but it must con- tain all of them to be considered contaminated. This is an absurd deinition, because a manuscript’s contamination must be shown in terms of the readings with which it is con- taminated and not in terms of readings with which it is not contaminated. Grünendahl’s assertion that the editor should have “preferentially demonstrate[d] the contamina- tion of K3–6 in terms of textual editions whose not being present is characteristic of the hyparchetype ν” efectively states that, even if K3 was contaminated with large numbers of passages found only in γ or only in S and even if it was contaminated with 24 of the 25 passages found in both γ and S, he would still not accept it was contaminated, as long as he could identify one passage (the 25th of our list) whose absence was not a unique feature of ν.292 This interpretation is not only arbitrary; it also leads to the unacceptable consequence that every manuscript would be simultaneously both contaminated and non-contaminated.293 257 CONFUSIONS REGARDING CLASSIFICATION 257 Over the next few pages, Grünendahl turns to demonstrating Ñ4’s superiority over K1, though many of the arguments in favor of this manuscript must be gleaned from his criticisms of other manuscripts or manuscript groups. His irst argument—actually, a complex of arguments—is as follows: Matters are the same with the remaining appendix passages listed by Sukthankar. The irst 25 of this list (116* to 119*) cannot be checked against any Ñ manuscript, which is why any conclusions drawn from this inding once again cannot hold for the “Ñ version” assigned to the hyparchetype γ. In addition, I wish to note that six of these passages at any rate are also not found in D2, whereby its proximity to the K manuscripts is again conirmed. The inding is even clearer for D14, a manuscript that Sukthankar—in spite of its general correctness (see Prolegomena, p. xx, lxxi)—does not consider further: it contains only one of these irst 25 textual additions (132*) and hence could have, with justiication, been assigned to the K manuscripts. Furthermore, the inclusion of 151* in the list is problematic, since two out of the total of four verses of this passage are not present in an entire series of γ manuscripts.294 We already noted that Sukthankar’s list is a general tally. He does not claim every interpolation is found in every manuscript in the list, so Grünendahl’s irst objection is irrelevant. Sukthankar is in any case not interested in drawing any conclusions from these passages for the Ñ manuscripts: his contention is that they prove contamination in the inferior K manuscripts. The idea is Grünendahl’s delusion.295 The absence of these passages from some of the γ recension manuscripts only means that, in these cases, these manuscripts escaped contamination. It does not mean their “proximity to the K manuscripts is again conirmed,” because—as we have seen—the fact that two manuscripts lack the same passage is not an argument for their iliation. All we can conclude from the fact that D2 lacks 6 of the 46 (and D14 2 of the irst 25296) passages Sukthankar cites as evidence of K3–6’s contamination with γ is that D2 and D14 escaped contamination with these passages, not that they escaped contamination in general and certainly not that they are unrelated to the other D manuscripts. If we look at Sukthankar’s statement concerning D14, we see Grünendahl is wrong here also. Sukthankar does not not consider D14 “in spite of its general correctness” further: he speciically mentions that D1–14 were discarded after the second adhyāya, as “misch-codices of small trustworthiness and of no special value for critical purposes.”297 He also notes that “D8–12.14 are palpably under Southern inluence, as is evidenced by their containing one or the other of the following typical Southern insertions: 18*, 21*, 22*, 24*, 32*, 42*, 45*, 48*, 49*, 56*, 80*, 81*, 89*, 114* 117*, 138*, 149*, 170*” and of D14 that its text “is a complex. It contains some old readings such as are preserved only in the Kámīrī manuscripts, but also an extraordinarily large number of individual readings not found elsewhere (cf. 1.1.50, 63; 2.101, etc.). At the same time, it is contam- inated from some Southern source, perhaps the Malayā lam version!”298 Grünendahl has obviously been misled by Sukthankar’s comments in his “Detailed Account of the Manuscripts,” where he notes that D14 is “carefully written, has very few corrections, which are made by use of yellow pigment and a few marginal additions.”299 Grünendahl glosses this remark with “general correctness” (generelle Korrektheit), but that a manuscript is written correctly says nothing about the correctness of the text contained in it. An 258 258 PHILOLOGY AND CRITICISM excellent text can be contained in a poorly written manuscript, whereas many well- written and beautifully illuminated codices from the more recent period exist, which editors recognize as useless for reconstructing the archetype.300 Grünendahl’s error once again arises from a failure to distinguish between a manuscript’s physical form and its text. By now, this error has occurred so often that we must conclude that it is not a simple oversight: the incoherence is in his mind.301 Grünendahl next focuses on the inal 21 passages of Sukthankar’s list: With these passages, Sukthankar wishes to illustrate the contamination of K2–4.6 with the hypearchetype γ. However, his argument can hardly convince us, because K6 is not collated for the last mentioned passages and, furthermore, K2 features only 8 of the total 46 textual additions; what is more, nine of these additions are also not attested in K3. But the non-K manuscripts also do not present a uniform picture. Once again, D5 manifests diverse similar- ities with what are, in Sukthankar’s view, the non-contaminated manuscripts.302 We have already seen that these claims are not just false but deliberately misleading.303 We therefore move on to Grünendahl’s inal argument in this section: However, I consider far more signiicant the inding in the case of the Ñ manuscripts: only Ñ2 features all 21 of the above-mentioned textual passages, Ñ1 has 13, Ñ3 only 10 and Ñ4, though not considered in the edition, none! Ñ4 also does not contain any of the “important omissions, which distinguish ν (really only Ś1, K0–3) from all other manuscripts” (App. I: 63, 75, 80, 81, 85) mentioned next. As these appendix passages are in fact present in all of the other manuscripts examined by Sukthankar for the edition, they—along with the smaller additions in the preceding list—should have been the touchstone for evaluating the manu- script Ñ4, which he only came to know of later; the status of this manuscript in Sukthankar’s manuscript classiication should have, irst and foremost, been determined in terms of their being present or not being present.304 Grünendahl argues that the criterion for evaluating Ñ4’s place in the stemma should have been the 21 shorter insertions Sukthankar cited as evidence of K3–6’s contami- nation with γ and the ive longer passages that distinguish S1́ and K0–3 from all other manuscripts.305 In his view, as these passages collectively distinguish the ν recension from γ, the editor should have adopted them as the standard for determining whether Ñ4 belongs with the manuscripts of the latter. It is easy to see why he might favor this interpretation. It provides him an easy means to determine “iliation.” Compared with the alternative of poring over scores of manuscripts and determining their true iliation reading by reading, the idea that iliation can be determined using a few passages—and that too obvious interpolations such as these 26 insertions—is more attractive. Yet for all its advantages (speed, lack of ambiguity and apparent eiciency), the argument is nonsense: iliation must be determined in terms of shared corruptions; it cannot be determined in terms of the “being present or not being present” (an ihrem Vorhanden- bzw. Nichtvorhandensein) of speciic interpolations.306 Grünendahl apparently thinks that as these passages are present in all manuscripts except Ś1 and K0–3, this provides him a decisive argument for claiming that Ñ4 belongs with the former. He seems to think that as the passages are found in all the other Mahā bhā rata 259 CONFUSIONS REGARDING CLASSIFICATION 259 manuscripts, their absence in Ñ4 cannot be attributed to some other branch of the tradition. Rather, it suggests that Ñ4, Ś1 and K0–3 had a common archetype and, as this archetype can only be ν, he concludes that Ñ4 must be an apograph of a Ś ā radā source. The argument is nonsense: Ñ4 does not contain the passages because they are insertions in the manuscripts that contain them. Whereas these insertions spread by contamination to all the manuscripts of the tradition except Ñ4, Ś1 and K0–3, this does not mean that these six manuscripts owe their absence either to each other or to a more recent common archetype than N. If the passages were absent in θ and the text descended to N, they will also be absent in N. If the text descended thereafter to ν and γ, they will still be absent in ν and γ. In other words, Ś1, K0–3 and Ñ4 owe their absence to their respective archetypes and the fact that all six manuscripts simulta- neously lack the passages is only coincidental. Likewise, the fact that Ñ4 does not contain any of the 21 shorter insertions Sukthankar cited as evidence of K3–6’s contamination with γ is not an argument for the iliation of Ñ4 with S1́ and K0–3. It is not even evidence of the absence of contam- ination of Ñ4 with the remaining γ manuscripts, because to show this we must draw up a list of all the interpolations common to Ñ4 and the remaining γ manuscripts and not just a list of those characteristic of the γ manuscripts that also occur in some inferior K manuscripts. In other words, the circumstance that Ñ4 does not contain any of the passages illustrative of K3–6’s contamination with γ only means it was not contami- nated with these passages, not that it was not contaminated with γ and least of all that, because it approximates K0–2 in this respect, it should be classiied with them.307 The circumstance that ν itself has “important additions” (the appendix passage no. 121), to which Grünendahl refers next, does not mean that this recension’s evidence can be discounted for Grünendahl’s hypothetical Nepā lī subrecension.308 No one claims ν (or even its best representatives S1́ and K0.1) are completely free of interpolations: were this the case, the editor could simply have reprinted the text of one of these witnesses.309 The point of creating a critical edition is to arrive, through purifying the text of its inferior or non-original readings, at a text better than that contained in any extant witness.310 If we were to expect perfect idelity to the original from our witnesses and discarded one of them every time it failed to satisfy this expectation, we would not get far. We could then not use Ñ4 either because, although Ñ4 lacks several passages found in S1́ and K0.1,311 it is also not completely free of interpolations.312 Grünendahl’s inal argument concerns the way (he thinks) Sukthankar relied on the agreement of ν and S to constitute the text, even though ν and S do not form a closed group against γ. As this argument illustrates the extent of his confusion regarding manu- script transmission, let us look at it here. He writes: Now as regards the frequent agreement between ν and the southern recension against γ, which, according to him [Sukthankar], also manifests in isolated and minor readings. The list of 30 relevant readings (Prolegomena, p. liv) notes only a few agreements between ν and the southern recension with other manuscripts of the northern recension, namely, three with Ñ1 and one with V1. Otherwise, the list gives the impression wished for by the editor [namely,] that ν and the southern recension keep to themselves.313 260 260 PHILOLOGY AND CRITICISM Grünendahl refers to Sukthankar’s list of 30 readings demonstrating the “frequent agreement of ν with S against γ, especially in the matter of isolated and even unimpor- tant readings.”314 But in his view, this list cannot be considered evidence of the agreement of ν with S against γ, since the Ñ manuscripts not infrequently feature—or, as he puts it, “align themselves with” (sich anschließen)—the reading of ν and S. Let us irst look at his arguments: However, this does not in any way correspond to the real inding, as an examination of the passages cited using the critical apparatus reveals. The Ñ manuscripts, which he [Sukthankar] consistently assigns to γ, already demonstrate signiicant variation in the list. If one leaves the irst three passages from the irst adhyā ya out of consideration, initially seven passages with incomplete references stand out: a part of the Ñ manuscripts is assigned to γ; for the rest, no account is given. An examination of the critical apparatus reveals that in all cases one of the Ñ manuscripts also features the allegedly exclusive reading of ν and the southern recension. Additionally, there are nine passages where, according to the list all the Ñ manuscripts feature the reading of γ, but, according to the critical apparatus, individual manuscripts align them- selves with the reading of ν and the southern recension. In the case of four further passages, both according to the list and according to the critical apparatus, all Ñ manuscripts feature the reading of γ; however, an examination of Ñ3 revealed that this manuscript aligns itself with ν and the southern recension here. An examination of Ñ4 additionally showed that this manuscript features the reading of ν and the southern recension in all twenty of the above mentioned cases and, over and above that, in 1.138a, 1.144c and 187.20a. Finally, in three passages, Sukthankar notes similarities between ν and Ñ4.315 He adds: Summing up, of the 27 passages listed (without those from adhyā ya 1), there remain only three where ν and the southern recension agree without at least one Ñ manuscript aligning with them. Ñ4 in particular is not in any way deicient in this respect to the manuscripts favored by Sukthankar and should have, if one will, have been assigned to the same group, unless it be that what one judged an “original inheritance” in the case of ν, one wanted in the case of Ñ4 (and the other Ñ manuscripts) to ascribe to contamination. This too was a point, regarding which one might have expected a clariication in Sukthankar’s study of this manuscript!316 Before we evaluate Grünendahl’s speciic charges, let us irst understand his principal objection to Sukthankar’s assessment of the three recensions’ relation. Sukthankar presents a list of 30 passages to illustrate the fact that often when the manuscripts difer in their reading, ν and S agree against γ, that is to say, ν and S feature one reading and γ another. Given this divergence in reading, Sukthankar thought the reading attested in ν and S was likely the reading of the archetype, given that contamination between ν and S was less likely than between either ν and γ or γ and S. His aim was not to exclude γ (what would he gain from doing so?), but to show that when the recensions presented conlicting readings, the readings attested in γ were more likely corruptions, given that the alternative that ν and S arrived at the same reading through independent innovation was unlikely.317 Grünendahl misunderstands the passage. As he reads it, the emphasis is placed on “against γ”—that is to say, he thinks that Sukthankar’s aim was to exclude γ.318 In 261 CONFUSIONS REGARDING CLASSIFICATION 261 Table 10 Grünendahl demonstrates that ν and S cannot be against γ (for at least one γ manu- script always agrees with them) Sukthankar’s list Grünendahl’s list Verse Mss. giving Mss. giving another Mss. of the γ recension that also feature the one reading “allegedly exclusive reading” of ν with S Grünendahl’s initial seven examples: Twenty examples looked up in Ñ4: 39.16 Ś 1 K S Ñ1.2 V1 B D Ñ3 (slight Ñ4 variation) 41.17 Ś 1 K S Ñ1.2 V1 B D Ñ3 Ñ4 100.6 Ś 1 K S Ñ1.2 V1 B D (mostly) Ñ3 Ñ4 119.8 Ś 1 K S Ñ1.2 B D Ñ3 Ñ4 138.17 KS Ñ2.3 V1 B D Ñ1 Ñ4 159.20 Ś 1 K S Ñ2 V1 B D Ñ1.3 Ñ4 182.9 Ś 1 K S Ñ2.3 V1 B D Ñ1 Ñ4 Grünendahl’s next nine examples: 20.2 KS others Ñ3 Ñ4 34.13 Ś 1 K S Ñ V1 B D (mostly) Ñ3 Ñ4 37.5 Ś 1 K S Ñ V1 B D (mostly) Ñ1.3 Ñ4 64.10 KS Ñ V1 B D (mostly) Ñ1.3 Ñ4 68.14 KS Ñ V1 B D (mostly) Ñ3 Ñ4 76.22 KS Ñ B D (mostly) Ñ3 Ñ4 77.4 KS Ñ B D (mostly) Ñ3 Ñ4 143.38 Ś 1 K S Ñ V1 B D Ñ1.3 Ñ4 176.5 KS Ñ V1 B D (mostly) Ñ1.3 Ñ4 Four examples looked up in Ñ3 21.10 KS Ñ V1 B D (mostly) Ñ3 Ñ4 26.9 KS Ñ V1 B D (mostly) Ñ3 Ñ4 141.4 Ś 1 K S Ñ V1 B D (mostly) Ñ3 Ñ4 142.18 Ś 1 K S Ñ V1 B D (mostly) Ñ3 Ñ4 Additional examples looked up in Ñ4: 1.138 K V1 S BD Ñ4 1.144 KS BD Ñ4 1.208 KS V1 B D (missing) 187.20 KS Ñ V1 B D Ñ4 Three examples for which Sukthankar notes similarities between ν and Ñ1 196.4 Ś 1 K Ñ1 S Ñ2.3 V1 B D Ñ1 199.12 Ś 1 K Ñ1 S Ñ2.3 B D Ñ1 199.19 Ś 1 K Ñ1 S Ñ2.3 V1 B D Ñ1 his view, the point of Sukthankar’s list is to show ν and S constitute a closed group—or, as he puts it, “keep to themselves” (unter sich bleiben)—vis-à-vis γ. And since ν and S emphat- ically do not “keep to themselves,” for at least one of the Ñ manuscripts also features the identical reading in all but three examples, he thinks he has a devastating objection to Sukthankar (besides demonstrating that Sukthankar cannot be trusted as an editor for downplaying the Ñ manuscripts).319 As evidence, he constructs a list of 27 examples where the reading of ν and S is also attested in at least one Ñ manuscript (Table 10). 26 262 PHILOLOGY AND CRITICISM Grünendahl cites these 27 passages as examples that contradict Sukthankar’s “assertion” that ν and S constitute a closed group vis-à-vis γ, yet he overlooks that it is not surprising that at least some manuscripts of the γ recension preserved the correct reading.320 In fact, we would be surprised if none of them did, for this would mean that the γ recension was so completely overwritten that not a single original reading survived in it—an unbelievable state of afairs. In other words, all Grünendahl has done with this list is to show that, in at least 27 cases, Sukthankar’s constitution of the text inds conir- mation from γ! The list does not in any way advance our understanding of the relation of the recensions, because all Grünendahl has done is take Sukthankar’s list and, in the seven examples where Sukthankar noted Ñ1.2 (four times), Ñ2.3, Ñ2 and Ñ2.3 have a corrupt reading, he has pointed out that Ñ3 (four times), Ñ1, Ñ1.3 and Ñ1 have the correct reading! Likewise, in the eight examples where Sukthankar either noted “others” or that the Ñ group “mostly” have a corrupt reading, he has pointed out that Ñ3 (twice), Ñ1.3 (twice), Ñ3 (three times) and Ñ1.3 have the correct reading.321 In an additional four, he has consulted Sukthankar’s collation of Ñ4 in “The Oldest Extant Ms. of the Ā diparvan,” and, from the fact that Sukthankar lists no variants for these verses, con- cluded that Ñ4’s reading must be identical with the constituted text.322 Finally, in three examples where Sukthankar noted that Ñ2.3 has a corrupt reading, he has pointed out that Ñ1 has the correct reading. As there is nothing spectacular in inferring, from the fact that Sukthankar notes that two of the manuscripts of the group have a corrupt reading, that the third, therefore, must have the correct one, the value of Grünendahl’s list is zero. Grünendahl has not understood the problem the genealogical method was evolved to solve. If all three recensions had the same reading, there would be no doubt about the reading the editor had to reprint: there would be no variants and hence no scope for interpretation. The problem arises when there is a divergence among the readings and no one reading is obviously correct. In that case, which reading should the editor reprint? Is it the one found in most of the γ manuscripts alone? Is it the one found in most of the ν manuscripts alone? Is it the one found in most of the S manuscripts alone? Is it the one found in most of the ν and γ manuscripts but without S? Is it the one found in most of the γ and S manuscripts but without N? Or is it the reading found in most of the ν and S manuscripts but without γ? Of these alternatives, the last one is the most intuitive given the geographic separation between ν and S. The point is not to show that ν and S have one reading and γ another or—as Grünendahl allusively puts it—that the “allegedly exclusive reading” (angeblich exklusive Lesart) of ν and the southern recension is really exclusive to ν and S, as though we gain something by excluding Ñ from the best groups. Rather, the point is what should the editor do given that ν and S frequently agree and that, when they do so, they often do so against γ. Further, as already noted, there is nothing surprising about the fact that at least some γ manuscripts, against the practice of their recension, preserve the correct reading: as the reading of the archetype, it will have descended to them as well in addition to ν and S. The circumstance that they feature the same reading as ν and S is not an argument against the latter’s reading being correct. If anything, it is further support for their reading being the correct one. If we examine Grünendahl’s list again, we see that, when he cites these 27 passages as examples where one or more Ñ manuscripts also feature the reading of ν and S, he has at 263 CONFUSIONS REGARDING CLASSIFICATION 263 most provided additional conirmation for the constituted text.323 He has not demonstrated that Ñ1–4 contain diferent readings from ν, and, least of all, that they contain better readings than ν.324 In fact, since Ñ1–4 conirm ν’s readings in all of the examples he cites, he has not shown how basing a critical edition on the former will alter the constituted text. This problem resurfaces in his inal set of examples. In continuation of his argument that Ñ4 “is not in any way deicient in this respect to the manuscripts favored by Sukthankar […],” he notes: Following the presentation of the independence and original connection of ν and S, Sukthankar inally arrives at the conclusion that the text of ν in general is such as to evoke conidence (Prolegomena, p. lv). It preserves some archaisms and some lectio diicilior [sic], albeit in corrupted form; in other manuscripts these have often been replaced by modern forms and paraphrases. Here he thus no longer argues, as at irst, in terms of quantity (few textual additions, etc.) but in terms of the quality of ν, even though none of the unique readings identiied by him earlier (see earlier concerning Prolegomena, p. xlviii–l) could have contributed to the [text’s] constitution.325 The idea that an editor either argues or can argue “in terms of quantity” is Grünendahl’s delusion. In contrast to Sukthankar, whose assessment is based on reading the manuscripts, Grünendahl considers the number of interpolations a criterion for preferring certain manuscripts. There is not one reference to readings in his article, even though it alleg- edly draws on a careful examination of Ñ3 and Ñ4. Twenty-ive years later, he is yet to publish his collations. All his arguments are based on data Sukthankar provided—only, where Sukthankar arrived at one interpretation, he asserts the contrary. We also have seen that the idea that Sukthankar was not justiied in identifying a K group because none of the unique readings that distinguished this group “could have contributed to the [text’s] constitution” is Grünendahl’s delusion: if an editor’s aim is to demonstrate the distinc- tiveness of certain manuscripts for containing unique readings, he will do so irrespec- tive of whether these readings also enter the constituted text. Since the critical edition rejects readings unique to one branch of the tradition as, in all probability, corruptions, it is almost certain that they will not enter the constituted text. If the readings “contrib- uted to the [text’s] constitution,” they would do so precisely by virtue of the fact that they are not unique to K, but also attested in a separate branch of the tradition. Grünendahl’s next argument for Ñ4’s superiority underscores his confusion. He writes: As evidence [of the fact that ν preserves the archaic readings, where other manuscripts have replaced them by modern forms] he [Sukthankar] lists six passages. Here the evidence of the critical apparatus for them, supplemented by the inding for Ñ4: 2.144b: abhibhoḥ in K3.5 [and M1.3.5] 2.177b: ślokāgraṃ in K0.2.3(corrupt).5 [and T2, M] 2.189c: ślokāgraṃ in K0–3.5 [and D14, southern recension (except G2)] 10.6c:  kāmayā in K0.2.3.5 [and Ñ3.4 B1.3.5, Da, D (except D4), southern recension] 98.13a: kanyasa in Ś 1 (corrupt), K0 [and Ñ1.3.4] 98.18c: samudge in Ś 1 [and Ñ1.3.4] Source: Reproduced verbatim from Grünendahl, “Zur Klassiizierung von Mahā bhā rata- Handschriften,” 127. 264 264 PHILOLOGY AND CRITICISM Grünendahl inds it “remarkable,” on one hand, that “by no means all—and often not even the most important—manuscripts assigned to ν feature the ‘archaic’ reading”; on the other, that “for those [readings] that they [the better K manuscripts] feature no exclu- sivity can be claimed; ν alone appears in general not to ofer a qualitatively suicient base for the constitution of the text.”326 The irst argument is incoherent. If K3.5 or K0.2.3.5 contain the lectio diicilior, then that means that they are the best manuscripts here. The editor will not reprint an inferior reading contained in K1, because he found it a good manuscript in general. An editor who proceeded so rigidly would not produce a critical edition, but an edition with one manuscript as base. Further, the fact that K1 is the best K manuscript does not preclude it from having corruptions of its own—as Sukthankar explicitly noted.327 Grünendahl also overlooks the fact that the worse K manuscripts are considered infe- rior because, in Sukthankar’s words, they are “misch-codices conlated either with γ or with S,”328 but that they are nonetheless K manuscripts. In other words, in parts unafected by contamination, they can preserve good readings. Only someone who has not under- stood the principle that sometimes the inferior manuscripts can preserve the true reading (and hence are invaluable for reconstructing the archetype) would discard the deteriores so lightly. In four of the cases Sukthankar cites (1.2.144b, 1.2.177b, 1.98.13a and 1.98.18c), the manuscripts have a mixture of readings.329 No clear pattern of agreement can be discerned. Thus, the choice had to be for the diicult reading.330 The editor did not mark any of these cases as uncertain, showing that he was conident that he restored the original reading, which explained the corruptions. Only Grünendahl, lacking a knowl- edge of Sanskrit, seeks to undermine the edition on the specious grounds that the best manuscripts of the K group do not contain the readings. The more serious problem concerns Grünendahl’s notion that he has raised an objec- tion to Sukthankar by pointing out that “For those [readings] that they [the better K manuscripts] feature no exclusivity can be claimed.” Why does Grünendahl think the readings of the K manuscripts must be “exclusive” to them? As we saw, there is nothing surprising about the circumstance that the reading of the K manuscripts are conirmed by one or more manuscripts of one of the other groups: it only means that the reading of the archetype was also transmitted in this branch of the tradition. The fact that the Ñ manuscripts conirm some of the K manuscripts’ archaisms is an argument neither against K nor for subsuming Ñ under K. Least of all is it an argument for a new edition based on the Ñ manuscripts. As we observed, the agreement between the K and Ñ groups undermines Grünendahl’s assumption of a signiicant diference between his proposed edition based on the Ñ manuscripts and Sukthankar’s: for every reading that K and Ñ concur, there will be no diference between an edition based on the former and one based on the latter. Incidentally, it bears repeating that Sukthankar’s edition is not based on the K manuscripts. It is based, as he puts it, on the “eclectic but cautious utilization of all manuscripts classes.”331 Only Grünendahl, because he imagines the critical edition is “based” on a spe- ciic manuscript, thinks it is an objection to the edition if he can show that the reading of the manuscript selected as a “base” is not “exclusive” to it—that is to say, that other manuscripts also have the same reading for, in his opinion, this means there were other candidates for the “base” and hence Sukthankar’s choice was a partial one. 265 CONFUSIONS REGARDING CLASSIFICATION 265 Thus, after presenting his list, he comments: “ν alone appears in general not to ofer a qualitatively suicient base for the constitution of the text,” and he continues: “The exam- ination of the readings of 166.23 contrasted by Sukthankar in columns (Prolegomena, p. lvi) leads to the same result: the left column cites under the ‘vulgata’ manuscripts Ñ2 as the sole Ñ manuscript. The remaining Ñ manuscripts are not listed in any of the three columns. However, from the critical apparatus to 166.23 it is clear that Ñ1.3 (and Ñ4) also feature the reading of ν.”332 Grünendahl refers to Sukthankar’s table for verse 1.166.23, which lists three variants from the vulgate and the northern and southern recensions. Sukthankar presents this table (Table 11) as evidence that often ν alone preserves the original reading. His comment on this table is as follows: Obviously, the stumbling block was antargataṃ of the constituted text, which is a lect. dif.; here it means “forgotten,” a meaning cited in our dictionaries generally as an uncommon meaning given only by Indian lexicographers! Unless one here assumes ν to be original, it is impossible to explain this divergence of ν, γ and S, both of which give a possible though weak sense.333 Grünendahl does not dispute Sukthankar’s assessment of the lectio diicilior, as he lacks the necessary linguistic skill, knowledge of the scribes’ conventions (usus scribendi) and famil- iarity with the Mahā bhā rata tradition. Rather, he attacks the assumption that “only” ν preserves the correct reading.334 From the fact that the apparatus notes that Ñ2 features the reading of the vulgate (1782*, which replaces 1.166.23 in the vulgate manuscripts), he infers that Ñ3 and Ñ4 feature the reading of ν.335 As this means that the reading antargataṃ tu tadrājñaḥ , etc., is no longer “exclusive” to ν, he sees his suspicion that “ν alone appears in general not to ofer a qualitatively suicient base for the constitution of the text” conirmed.336 Contrary to Grünendahl’s view, however, the critical edition is not “based” on one manuscript or recension, much less on it “alone.” Sukthankar’s argument is unafected if the same reading also occurs in some other manuscripts, because he is arguing from the lectio diicilior that the reading preserved in ν (antargataṃ) probably occasioned later scribes’ emendations (some turned it into antaḥ puraṃ and others, unable to construe the verse, replaced it entirely). The circumstance that some Ñ manuscripts also retain this reading only makes this interpretation likelier. As he neither claims “exclusivity” for ν nor makes the mistake of thinking that the correct readings must be “exclusive” to one branch of the tradition for him to justify editing that branch of the tradition (which he emphatically did not do), Ñ’s apparent “recidivism” does not pose a problem for him. Table 11 Sukthankar considers the agreement of the versions K2 Ñ2 V1 B D = Vulgate ν = Text Southern Recension tato rāja parikramya antargataṃ tu tadrājñaḥ antaḥ puraṃ gato rājā yathākāmaṃ yathāsukham | tadā brahmaṇabhāṣitam | śrutvā brāhmaṇabhāṣitam | nivṛtto’ntaḥ puraṃ pārtha so’ntaḥ puraṃ praviśyātha so’ntaḥ puraṃ praviśyātha praviveśa mahāmanāḥ || saṃviveśa narādhipaḥ | na sasmāra narādhipaḥ | Source: Sukthankar, “Prolegomena,” lvi. 26 266 PHILOLOGY AND CRITICISM The reading antargataṃ tu tadrājñaḥ , etc., which originated with the archetype, could have descended to Ś and K and to Ñ4 via their respective archetypes without positing either that Ś and K got it from Ñ4 or Ñ4 got it from Ś and K or that both got it from a common archetype (according to Grünendahl, ν). Grünendahl thinks that the circumstance that Ñ3 and Ñ4 also feature the reading of ν demonstrates the untenability of an edition “based” on the Ś-K manuscripts, and more speciically, on Ś1. Because he is unclear about the distinction between a critical edition and an edition with one manuscript as base, and because this is what he would have done in Sukthankar’s place, he thinks the editor does no more than identify one manuscript as the base and reprint its readings with random “improvements.”337 From his perspective, the circumstance that Ñ3 and Ñ4 also share the readings of Ś1, the manuscript allegedly selected as the base, constitutes a major objection to Sukthankar’s edition, since it shows that Ñ3 and Ñ4 could just as well have served as a base. Ñ4 in particular not only features Ś1’s “good” readings but also resembles the K manuscripts in not featuring the interpolations that the latter do not feature. It is, moreover, shorter than Ś1. Hence he concludes that Ñ4 should have replaced Ś1 in the edition, since Sukthankar’s “own ‘inductive and prag- matic veriication’ is […] not only saddled by weaknesses in diverse details but he him- self also later corrected his main argument for the preeminent status of the ‘Kashmiri version,’ namely, that it is the shortest, in favor of Ñ4—without, however, drawing the consequences resulting therefrom for his edition.”338 “For,” Grünendahl adds, “when this claim cannot be upheld, his entire demonstration based on it is redundant.”339 Citing “abstract” considerations,340 he argues that, as the Mahā bhā rata never existed as a single text but only as several independent recensions where a constant exchange of mate- rial between recensions occurred until a tradition stabilized in each script,341 the focus must shift from reconstructing an archetype (which never existed) to editing one of the Mahā bhā rata’s regional recensions.342 In this connection, he again emphasizes that Ñ4 ofers a superior “base”: it is not only the shortest manuscript; it also preserves the most archaisms, is free of contamination, unrelated to the other Ñ manuscripts (!) and manifests independent agreement with other recensions that indicate a connection through the archetype.343 These reasons, coupled with his access to Ñ4 (and his brilliance in proving, he thinks, Sukthankar’s erroneous choice of base), lead him to propose a new critical edition of the Mahā bhā rata under his leadership. Grünendahl has clearly not grasped the concept of a critical edition. He thinks all an editor need do for a critical edition is identify the manuscript that will serve as the “base” (namely, the shortest manuscript, which preserves archaisms and is free of contamina- tion and yet exhibits agreement with other recensions) and then reprint its text. In his opinion, a systematic recensio is superluous because once the editor identiies his “base,” he will consistently defer to it (as the shortest, it preserves the most original text). At most, he may occasionally consult other manuscripts, especially those in the same script, since agreement with members of its own “family,” according to Grünendahl, is a conclu- sive argument for their common reading being the reading of their source. Where the manuscripts diverge, he chooses the reading attested in the majority or, possibly, the one found in the best manuscript (which Grünendahl equates with the shortest). If the editor considers the other manuscripts at all, he does so only to note that the other branches 267 CONFUSIONS REGARDING CLASSIFICATION 267 are more inlated or that the reading he has identiied as the original is also attested in them. As his primary goal is to reduce the complexity of the tradition, he will neglect most variants and report only the most interesting ones in the critical apparatus but, on the whole, he will exclude as many manuscript groups from consideration as pos- sible. Since Grünendahl is not interested in genealogical reconstruction of the arche- type—he considers it superluous, since, in his view, the shortest manuscript unfailingly preserves the reading of the archetype—he thinks the editor can forego a stemma.344 At most, he may draw up tabular columns such as those in Tables 1, 2 and 3, grouping the manuscripts into two groups—the contaminated and the non-contaminated—and using this classiication in place of the stemma.345 The problem with this approach is evident. Even though Grünendahl thinks his edition would be superior to Sukthankar’s, especially as it would avoid false inferences due to contamination, in truth he will be unable to identify the correct reading in most cases. Every time the Ñ manuscripts diverge, he will be unable to identify the correct variant. He might print the reading attested in the majority or the reading corroborated in one of the other recensions, but neither will ensure it is the correct one.346 Even if all four Ñ manuscripts were unanimous in their reading, he could still be misled, since all four could be corrupt. Nor could he rely on certain Nepā lī manuscripts always agreeing, because the groups constantly change. Even if he found two Ñ manuscripts consistently agreed against the other two, how would he know this was due to independent agreement rather than contamination? In cases of crux, he could not forgo the evidence of the other manuscript groups, but to correctly evaluate this, he would need to know how they relate to the Nepā lī group, which means he once again could not avoid a systematic recensio. Grünendahl might argue—as he, indeed, does—that he can abjure the require- ment of recensio because, as the shortest Nepā lī manuscript, Ñ4 serves as a “base” and he merely needs to list Ñ1–3’s variants in the critical apparatus. But this does not solve the problem. Although his proposed edition of the Nepā lī tradition would not be a true critical edition (not even of this one recension), it is also not a true Bédierist edition, that is, an edition “aspiring to absolute respect for the reality of a single manuscript, after an exhaustive analysis of the varia lectio, and that [is] rigorously consistent with this prin- ciple.”347 Rather, what it aspires to is more correctly classiied under the category of “base manuscript criticism,” that is, an edition in which “the manuscript one preaches loyalty to is exactly that, the ‘base,’ to which corrections and modiications are added, often on no other criterion than the presumed evidence of their necessity, and without trying to account for the tradition’s development.”348 There are two problems with this approach, as Leonardi notes: 1. On the ontological plane, the text thus produced is not deined even roughly on the diachronic axis running from the original to the witness. The material attractiveness of the base ms. leads to a conservative edition, but the possibility of correcting it more or less freely produces a text that stands in an ambiguous and heterogeneous position, in an indistinct hinterland of the base ms., outside the two hypothetical alternatives, the two “truths” outlined above [viz., that of philology oriented to the text and that of philology oriented to the witness]. 268 268 PHILOLOGY AND CRITICISM 2. On the phenomenological plane, too, the fact that the choice of the base ms. and the control mss. and the procedures for correcting the base ms. are not illuminated by recensio often leads to contradictory solutions, and even misleading ones as regards com- prehension of the base ms. itself, which does not ind its proper place in the tradition.349 Grünendahl’s proposed edition thus cannot serve as a guide even to the Nepā lī tradi- tion. It can explain neither the Nepā lī manuscripts’ position in the wider Mahā bhā rata tradition nor their development. It cannot tell us what stage of the (Nepā lī) tradition it reconstructs. It is unclear that the other manuscripts’ readings, where they difer, can be derived from Ñ4’s reading, since Sukthankar and the other editors found that extensive contamination between manuscripts of the regional recensions made it impossible to reduce their readings to one.350 Although it claims to provide an edition of this one man- uscript (Ñ4), in practice, unless Grünendahl reprinted its readings consistently (that is, even in cases of obvious corruption and nonsense), it cannot avoid mixing readings from heterogeneous sources, often on no other principle than that the manuscripts containing them are in the same script! The sole alternative, namely, that the editor makes emendations as they occur to him, represents a step behind the methodological advances of the nineteenth century—the introduction of a systematic recensio and the principle of emendatio ope codicum. From the perspective of someone with no experience in dealing with a complex tradition, this return to a “pre-scientiic practice” (Leonardi) may appear appealing, but it is abso- lutely unacceptable to those who understand editing in terms of the three requirements Leonardi speaks of (scientiicity, an eye for diachrony and legibility).351 Grünendahl’s decision to “abjure all interpretive requirements of the edited text”352 may seem appealing in an age in which, as Leonardi puts it, “the authentic Bédierist inheritance has […] exhausted its momentum.”353 But it poses serious challenges. As Leonardi notes, “The fact that the choice of the base ms. and the control mss. and the procedures for correcting the base ms. are not illuminated by recensio often leads to con- tradictory solutions, and even misleading ones as regards comprehension of the base ms. itself, which does not ind its proper place in the tradition.”354 Grünendahl’s advo- cacy of the base manuscript, coupled with his partisan and misleading criticisms of the Mahā bhā rata critical edition, actually obscures what we know about the Mahā bhā rata tra- dition. It has led to the erroneous perception that an encompassing Mahā bhā rata tradi- tion never existed; that we can only speak of regional traditions of the Mahā bhā rata in isolation and that studying witnesses from other recensions or the relations between these recensions cannot shed light on the speciic one currently under consideration. Going beyond Leonardi, we can say that the real danger of the pseudo-Bédierist movement is that, if left unchecked, it leads to the loss of the information that the application of the genealogical-reconstructive method to a tradition provides.355 It is not only the base man- uscript that “does not ind its proper place in the tradition.” Rather, the tradition itself, in a manner of speaking, does not ind its proper place in the tradition once skepticism about the genealogical-reconstructive method is admitted.356 Once the pseudo-Bédierist critic— on the pretext of providing an edition more faithful to a single manuscript—succeeds in fragmenting the tradition, the single-manuscript edition appears not only plausible but preferable. Indeed, it appears the sole possible solution. This is likely why German 269 CONFUSIONS REGARDING CLASSIFICATION 269 Indologists, most of whom, like Grünendahl, have only a weak grasp of textual criticism, favor it.357 Not only does it airm their prejudice that a coherent textual tradition never existed; it also conceals their deicits in undertaking a true critical edition. The Argument from Expertise The argument from expertise refers to the replacement of argument with expert testimony. The mere fact of citation, without an attempt at paraphrase much less clariication, serves to validate an author’s views. Rightly speaking, the argument from expertise is not an argument at all. It merely relects the institutionalized prejudice in favor of certain authors and positions. It reveals the broken system of preferment at Mahābhārata studies’ core. The scholars entrusted with discriminating between correct and incorrect scholarship have proven incapable of distinguishing objective research from ideological views. In three chapters now, we have seen how, although they consistently belabored the language of criticism and critical methods, the German Indologists did not grasp the concept of tex- tual criticism.358 Rather, what they classiied under that name was the attempt to separate a hypothetical Ksạ triya text from Brahmanic interpolations, allegedly inserted into the text during the Mahā bhā rata’s “redaction.” This method, insofar as it is subjective, a priori and circular, does not deserve the name “critical.” It merely evinces the German scholars’ anti-Brahmanism or anti-Semitism, which they thought justiied them in intervening in the textual traditions of other cultures, be they “corrupted” by midrashim or mı̄māmṣ ā.359 The larger issue at stake concerns how these erroneous ideas could be transmitted and admitted into the canon of legitimate scholarship. Grünendahl’s errors were so la- grant that no one with the slightest acquaintance with textual criticism would have found his arguments convincing. Anyone familiar with the basics of textual criticism knows that agreement in correct readings (that is, those considered most likely the reading of the source) does not establish iliation and agreement in the absence of readings even less so.360 Yet several Mahā bhā rata scholars, Indologists and critical philologists found Grünendahl’s arguments not only compelling but also worth endorsing. This section reviews some of their comments before drawing some conclusions about how the argu- ment from expertise has replaced legitimate research and discovery in Mahā bhā rata studies with the mere appearance of critical scholarship. We begin with Michael Witzel, who commends Grünendahl for “going beyond the facile division into various script traditions.”361 Witzel judges the Mahā bhā rata critical edition a “failure,” due to the “ultimately futile” nature of the attempt at “establishing a stemma.”362 He criticizes, via a reference to Grünendahl, “the currently—still—fashion- able ‘critical’ approach to philology as ‘the cultural technology of colonial rule,’ ”363 but does not see that his account of the Mahā bhā rata’s “history” is a priori and philologi- cally indefensible, and would make a critical edition, such as he claims to defend, impos- sible.364 In fact, where it claims a “crystallization” of the epic, it is nothing but a revival of Bigger’s “normative redaction” hypothesis, albeit using a diferent metaphor. Witzel is not alone in this regard. His colleagues Georg von Simson, Johannes Bronkhorst and Walter Slaje likewise cite Grünendahl approvingly. Von Simson notes that “Within these main recensions [of the Mahā bhā rata], versions that can be categorized according to the diferent Indian scripts were formed. However, the criterion of script is 270 270 PHILOLOGY AND CRITICISM by no means reliable; one can observe manifold reciprocal inluences between the regional transmissions.”365 The passage is footnoted with a reference to Grünendahl.366 Slaje argues, “Though I agree with Grünendahl’s (1993) criticism of the ‘Schriftartprämisse,’ I  hesitate to accept its general application to all sorts of texts. Grünendahl developed his arguments by textual criticism of the epics, targeting on the principles Sukthankar adhered to as an editor of the Poona Critical Edition of the Mahā bhā rata in particular. His results, however, should be limited to those texts that evolved in more than only one region, as it was normally the case with the so-called epic-purā ṇic ‘literatures.’ ”367 Bronkhorst endorses Grünendahl’s other criticisms of Sukthankar:  “It is true that Sukthankar subsequently obtained and described (1938) a manuscript from Nepal cov- ering (only) the Ādiparvan that presents an even shorter version of that portion of the Mahābhārata. It is equally true that Grünendahl (1993) has criticized Sukthankar’s edi- torial principles, most notably his attempt to associate a number of manuscripts (those covered by the letter K) with the Ś ā radā manuscripts.”368 We have already noted Bigger’s views of Grünendahl at length, so we do not repeat them here.369 But the work of Oskar von Hinüber deserves mention, both because of his claims regarding Grünendahl’s contribution and because of his extreme skepticism about the Mahā bhā rata critical edition. In von Hinüber’s view, the “synchronous diver- sity of scripts did not lead to the formation of recensions of the Mahā bhā rata, which were restricted to or concentrated in speciic script areas. Important new investigations by Reinhold Grünendahl suggest this and evidence from other domains of Sanskrit liter- ature conirm his indings.” Like Grünendahl and Bigger, von Hinüber cites “the inlu- ence of the oral tradition” as leading “to a similar dissolution of the text as is known was the case for the Homeric papyri” and claims “textual criticism ought to actually sur- render before the Mahā bhā rata.”370 He refers to both Grünendahl’s “Zur Klassiizierung von Mahā bhā rata-Handschriften” and his “Zur Textkritik des Nā rā yaṇīya.”371 In the attached note, von Hinüber erroneously repeats Grünendahl’s view that the Mahā bhā rata critical edition “is based […] on the idea, developed by H. Lüders, of the coincidence of the script and the recension, something that, as R. Grünendahl has shown, cannot be upheld [n.  14]” and adds:  “V. S.  Sukthankar (1887–1943), the irst general editor of the Mahā bhā rata, already expressed his doubts about the coincidence of the script and the recension in the Prolegomena, p.  vii, which are also reprinted in Sukthankar [n. 16], p. 17.”372 Continuing to propound the erroneous view that the critical edition reconstructs the Mahā bhā rata’s “normative recension,” Von Hinüber argues: “What has arisen is not the Ur-text, which could not have arisen at all, but a completely new nor- malized recension.” In his view, “The editors of the Mahā bhā rata are so to speak the late descendants of Aristarchus or Nīlakaṇṭha,” a manifestly false claim since, whereas Aristarchus attempts to reconstruct Homer’s text on the basis of distinguishing his language from that of the poets he terms the neōteroi (“the newer ones”) and Nīlakaṇṭha’s edition is a compendium following the principles of comprehensivity and Vedā ntic con- sistency, the Mahā bhā rata critical edition is a neo-Lachmannian edition that aims to reconstruct the archetype of the tradition.373 Von Hinüber’s claims reveal a surprising lack of familiarity with the Mahābhā rata tra- dition. More important, they appear grounded neither in his knowledge of the tradition 271 CONFUSIONS REGARDING CLASSIFICATION 271 nor even in a general proiciency in textual criticism. Rather, they appear motivated solely by an interest in delegitimizing the critical edition, of which he claims “its goals and methods were determined by the state of knowledge of the previous [that is, the nineteenth] century.”374 Von Hinüber’s comments reveal a discipline that has no other purpose than to reproduce its own often erroneous scholarship through asserting its can- onicity, as when he directs John Brockington’s attention to “the important contribution by R. Grünendahl, Zur Klassiizierung von Mahā bhā rata-Handschriften [...], who with good reasons disputes the correlation of script and recension in the Mahā bhā rata trans- mission already questioned by V. S. Sukthankar.”375 James L. Fitzgerald, the general editor of the Chicago translation of the Mahā bhā rata, exhibits the same reliance on Grünendahl. He presents Grünendahl’s views thus: V. S.  Sukthankar conducted his basic survey and classiication of available manuscripts on the basis of a general (not absolute) “script-premise” (the term is R. Grünendahl’s [Grünendahl, 1993a]), that H. Lüders had earlier articulated (Grünendahl, 1993a, 101–102; Sukthankar, 1933, vi–vii). This premise posited that the several regional scripts of India con- stituted relatively isolated manuscript traditions with relatively little contamination between or among them. The Sanskrit Mahābhārata critical-edition project made use of manuscripts in eight distinct Indian scripts, ive of the North—Sharada, Devanagari, Bengali, Nepali, and Maithili—and three of the South—Grantha, Telugu, and Malayalam.376 Fitzgerald adds: The most searching and particular criticisms of V. S. Sukthankar’s editorial eforts came from F. Edgerton (1944) and R. Grünendahl (1993a). These two scholars examined closely Sukthankar’s segmentation and classiication of the manuscript tradition, and some of Sukthankar’s basic commitments have been refuted as a result. There are three major criticisms of V. S. Sukthankar’s thinking that emerge from their work. First, the script-premise is, as a general description of the working of the manuscript tradition, not warranted. Then R. Grünendahl (1993a) showed further that V. S. Sukthankar’s arguments regarding ν were shaky and inconsistent. But while ofering valuable judgments about the manuscripts used by V. S.  Sukthankar for his edition of the Ā diparvan and for the formulation of his principles, neither of these critiques of V. S. Sukthankar’s judgments vitiate the practical result of V. S. Sukthankar’s reconstructed text. The efect would be quite diferent had V. S. Sukthankar argued that ν was in fact the arche- type behind the entire extant written tradition and based his reconstruction on that assumption. A third important criticism made by R. Grünendahl was to point out that a Nepali manuscript one or two hundred years older than any previously available to V. S. Sukthankar was shorter than even the Kashmiri texts of ν. V. S. Sukthankar became aware of this manuscript too late to make practical use of it (Sukthankar, 1938), but, as R. Grünendahl correctly points out, he failed to consider the full implications of this fact for his overall theory of the tradition.377 This view is repeated in the “General Introduction” to the Chicago translation, thus transmitting the error to a younger generation of scholars: For a recent discussion of some of the limitations and problems of Sukthankar’s editorial practices, see Reinhold Grünendahl’s “Zur Klassiizierung von Mahā bhā rata-Handschriften.” 27 272 PHILOLOGY AND CRITICISM Grünendahl’s study identiied a number of problems and inconsistencies in Sukthankar’s edi- torial approach, but it does not bring any telling argument against the remarkable results— primarily in terms of excellent “diicult readings”—yielded by Sukthankar’s policy of using the Ś ā radā tradition, and especially the coincidence of the Ś ā radā and Malayā li traditions, as a touchstone.378 Sheldon Pollock likewise repeats Grünendahl’s reservation about using scripts as a clas- siicatory principle. He writes: Most scholars assume that writing styles and manuscript traditions formed closed systems: Given the regional exclusivity of scripts—or what is taken to be their exclusivity— Sanskrit literary texts are said to have developed versions peculiar to writing traditions, and hence recensions tended to become regionalized. […] There is some truth to this conventional view, but it needs important qualiication. Scripts in precolonial South Asia seem to have represented as little a barrier to supralocal communication as regional languages.379 Pollock repeats this assessment in The Language of the Gods in the World of Men: Moreover, although manuscripts are classiiable into regional traditions, regionalism here pertains entirely to diferences in local writing systems. No regionalization of any conse- quence, in point of dialect or in the particulars of material or social or even religious life, can be detected in any recension or version. To the degree that regional versions can be said to exist at all […] they mark distinctions without diferences. That is, while the text was undoubtedly expanded or contracted in one recension or another, this was unaccompanied by the least hint of localization. In short, the unmistakable impression given by hundreds of medieval manuscripts copied time and again for centuries on end is that the Mahābhārata, just like Sanskrit itself, existed in a quasi-universal transregional space and spoke across this space in a truly homogenous voice.380 Notably, both passages refer to Grünendahl’s work: “Further doubts about the ‘writing- system premise’ that underlies epic text-criticism and the reality of regional recensions have been raised by Grünendahl 1993”381 and “Grünendahl 1993 importantly reconsiders the logic and reality of ‘regional recensions’ of the MBh.”382 These citations are problem- atic inasmuch as Pollock proclaims himself the exponent of a new “critical philology.”383 If we accept Pollock’s equation of philology with biography, they raise grave doubts about Pollock’s qualiications as a philologist.384 Albeit “critical philology” “with its global history, its conceptual pluralism, and the massive, kaleidoscopic archive of methods and interpretations in the making sense of texts to which it gives access” is supposed to be “the premier site for its exposition and demonstration [sc. of the ‘foundational principles of the twenty-irst century human sciences’],”385 it cannot function without a familiarity with the traditional tools of the editor and critic.386 Pollock adds: “if the rich materials from traditional India are ever to be included in this new philological discipline, we Indologists have our work cut out for us,” but this promised induction will only succeed to the extent that the Indologists master the basic philological skills.387 273 CONFUSIONS REGARDING CLASSIFICATION 273 Finally, two other scholars do not cite Grünendahl’s 1993 article, but feature the same mistake. The irst is Alf Hiltebeitel, who relies on Grünendahl’s summary of his 1993 article in the volume Nārāyaṇı ̄ya-Studien: Now in considering these manuscripts, it is necessary to bear in mind Grünendahl’s impor- tant reservations about V.  S. Sukthankar’s Schriftartprämisse:  the latter’s grounding premise as the irst Critical Edition General Editor “that a kind of script constitutes a ‘version’ ” (Grünendahl 1997: 30). Grünendahl shows that versions often overlaps scripts, and that an “M version” in particular could not simply be characterized by its “purity” relative to other (Telugu and Grantha) Southern Recension scripts (33) akin to the again-alleged purity that Sukthankar attributed to the “Ś ā radā version” of the Northern recension (31).388 The second is Brockington, who explains the Mahā bhā rata critical edition thus: In his Prolegomena to the irst volume Sukthankar declared that the aim of the Critical Edition was to reconstruct “the oldest form of the text which it is possible to establish on the basis of the manuscript material available” (p. lxxxvi). To this end, a large number of manuscripts were collated (using Nīlakaṇṭha’s texts as the vulgate), as well as several printed editions, and the result has been the collecting of a far more complete record of the variant readings and the regional traditions of the epic than was in practice available to any scholar previously. The manuscripts were classiied irstly into the Northern and Southern recension and then, on the basis of the scripts used, into versions.389 Though Brockington appears to have committed the error of thinking the manuscripts were classiied by script independent of Grünendahl (as noted earlier, he does not cite Grünendahl), the error is no less serious for that reason. Although a cursory inspection—for example, checking whether the editors actually classiied manuscripts by script—would have revealed the error in Grünendahl’s thesis, not one scholar undertook this step. All the scholars cited failed to note that there is a distinction between a manuscript’s text and its script. Despite the fact that they work primarily with the Mahā bhā rata in transliteration, they did not notice that the text can be transposed into a diferent script while still preserving the essential diferences between versions. Not one scholar provided a clariication or restatement of Grünendahl’s thesis, much less subjected it to a careful, dispassionate analysis. The experts’ failure to detect the problems with Grünendahl’s work raises serious questions about the viability of so-called critical Mahā bhā rata studies. It points to an institutional crisis, because expert testimony, paradigmatically manifest as citation of one’s peers, has replaced the need for demonstration. Not one scholar could paraphrase, much less provide an argument for Grünendahl’s work. Rather, the mere fact that they cited Grünendahl’s work functioned as an endorsement of it, even if good sense, the authority of genuine scholars and tradition spoke against it. On one hand, the Mahā bhā rata scholars cre- ated an untenable distinction between the critic—Western-trained, historical and allegedly equipped with philological expertise—and the lay reader. On the other, they themselves lack the necessary philological expertise to subject each other’s claims to an evaluation. Rather, the criterion for valid scholarship has become: Who said it? Does he possess the correct pedigree? Does he enjoy his peers’ conidence? Have other experts cited him? 274 274 PHILOLOGY AND CRITICISM Instead of its traditional sense of “expert skill or knowledge in a particular ield,” expertise has acquired the derivative meaning of “expert opinion or commentary.” It no longer refers to an ability to make authoritative pronouncements grounded in insight, technical knowledge or experience. Rather, expertise refers to the ability to declare cer- tain ideas valid merely because they exhibit the institutional features of scholarship—for instance, that a German scholar rather than an Indian made these claims. By choosing expertise over arguments and ailiation over understanding, the Mahā bhā rata critics have ill served not only their readers but also themselves. They have understood almost nothing of this text.390 They have replaced a serious engage- ment with philosophical issues with an empty formalism. They have not contributed in any way to the life of this text.391 As Sukthankar noted, their learned debates and dilemmas, their afectations of critical distance appear jejune before a text that the average Indian grasps intuitively.392 Once we set aside the Mahā bhā rata critics’ ersatz tradition, only two avenues remain for studying the Mahā bhā rata. The irst is a return to the commentarial tradition, which provides not only valuable insight into the text’s earliest state but also a guide to its history of reception.393 The second is a literary and philosophical interpretation, such as Madeleine Biardeau developed. Rejecting the German scholars’ problematic theses of an Aryan epic, Biardeau approached the Mahā bhā rata as “the principal—and undoubtedly the most ancient—of all monuments to bhakti.”394 Her work, rooted in her appreciation of Brahmanic philos- ophy and theology, ofers a promising avenue for reading the epic freed of the ballast of historicism and the German Indologists’ pseudocritical approaches.395 Beyond the myth of a scientiic Indology and a critical Sanskrit philology, the Mahā bhā rata awaits its readers as the “thought entire” (mataṃ kṛtsnaṃ, Mahā bhā rata 1.55.2 and 1.56.12) of the seer Kṛṣṇa Dvaipā yana Vyā sa. Notes 1 Lino Leonardi, “Il testo come ipotesi (critica del manoscritto-base),” Medioevo romanzo 35 (2011): 26 (Trovato’s translation). 2 See Winternitz’s comments in Moriz Winternitz, “The Mahabharata,” The Visva-Bharati Quarterly 1 (1924):  343–59 (cited in full in the introduction). Winternitz proposed a critical edition of the Mahā bhā rata as the only “basis for such higher criticism and for any attempts at the reconstruction of the old poem” (ibid., 347) as part of this project of a general enlighten- ment of the Indian mind. 3 Jacques Froger, La critique des textes et son automatisation (Paris: Dunod, 1968), 41. 4 Andreas Bigger, Balarāma im Mahābhārata: Seine Darstellung im Rahmen des Textes und seiner Entwicklung (Wiesbaden: Otto Harrassowitz, 1998), 121. 5 Ibid. 6 Ibid. (Bigger’s italics). 7 Ibid. 8 Michael D. Reeve, “Shared Innovations, Dichotomies, and Evolution,” in Manuscripts and Methods:  Essays on Editing and Transmission (Rome:  Edizioni di Storia e Letteratura, 2011), 56. Timpanaro prefers “axiological,” since most editors distinguished between the “good” (meliores) and the “bad” (deteriores) manuscripts. Sebastiano Timpanaro, The Genesis of Lachmann’s Method, ed. and trans. Glenn W. Most (Chicago: University of Chicago Press, 2005), 93, 98, 101, 174–75. 9 Reeve, “Shared Innovations, Dichotomies, and Evolution,” 56. 275 CONFUSIONS REGARDING CLASSIFICATION 275 10 At most it establishes that they are both descended from the archetype and that, in this respect at least, they preserve the latter’s reading. 11 There are many ways the manuscripts could agree in not featuring the interpolation: (1) they could be descended from three independent recensions, none of which featured the interpola- tion; (2) two of them could be descended from recensions that did not feature the interpolation and the third from a recension that did, but whose scribe did not copy it; (3) alternatively, they might not feature the interpolation because they are all descended from the archetype, which by deinition did not feature the interpolation, but this does not justify us in assuming that they constitute a “recension” as they have no ancestor in common later than the archetype, which they share with all other surviving manuscripts of the work in question. Bigger is unclear about the diference between a trait that is shared because it is (simultaneously) present in many exemplars of a work and a trait that is shared because it has been inherited. Only the latter is genealogically relevant. 12 Froger, La critique des textes et son automatisation, 41. 13 The real reason the six manuscripts do not contain the passages in question is that they were likely not part of the archetype and the manuscripts, having escaped contamination with the inlated tradition, all preserve the archetype’s text. Bigger’s argument amounts to the claim that all six manuscripts are descended from θ. But as we have already established that they are members of the family selected for study—the family of Mahā bhā rata manuscripts descended from the archetype θ—the circumstance that they lack the same passages does not bring us further. For similar reasons, N cannot be their unique common ancestor. Whereas all six manuscripts are descended from N, so are K4–6, Ñ1–3, V, B and D. What we seek, rather, is the common ancestor of just these six manuscripts, that is, something that deines them as a section of the wider family. 14 Timpanaro, The Genesis of Lachmann’s Method, 175. Timpanaro considers this error a basic confusion, implicated also in the failure to distinguish between axiological and genealogical classiication: “Another cause of erroneous bipartite classiications is partially connected with the preceding one [the tendency to group manuscripts into two groups, the meliores and the deteriores] but is more strictly derived from a logical mistake: the tendency to identify one class of manuscripts α on the basis of shared characteristics and then to call β everything that in reality is merely ‘non-α.’ There is a danger of falling into a similar error not only in textual criticism but wherever classiications need to be made: Aristotle already fought against it in zoology. If a certain number of shared corruptions deines a family of manuscripts, the lack of those corruptions does not deine another family: so after having identiied a family α it will be necessary to see whether the other manuscripts are connected by shared innovations in their turn, or whether instead they constitute diferent groups, or whether, as is also possible, they are so contaminated that their derivation from one or more subarchetypes cannot be detected.” 15 Reinhold Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” in Studien zur Indologie und Buddhismuskunde, ed. Reinhold Grünendahl,  Jens Uwe-Hartmann and Petra Kiefer-Pülz, Indica et Tibetica 22 (Bonn: Indica et Tibetica Verlag, 1993), 130. 16 Timpanaro, The Genesis of Lachmann’s Method, 65. 17 Bigger’s idea that “not [being] attested in” is an argument for iliation is borrowed from Grünendahl’s 1993 article, which elevates the idea—though fallacious—to a basic principle. Hence our focus on Grünendahl rather than Bigger in this chapter. 18 Froger traces the principle back to 1872 when Gaston Paris irst systematically distin- guished between shared corruptions and shared readings. Paul Lejay explicitly formulated the underlying principle in 1888 as follows:  “M. S.  essaye d’établir que ce ms., du XIIe s., appartient à la meilleure classe désignée par Z et que les mss. connus de cette classe, B H b, forment une sous-famille distincte de M. Dans le détail, la démonstration n’est pas très nette. M. S. commence par démontrer la seconde partie de la proposition (p. 17); mais, dans sa liste de variantes, il introduit de bonnes leçons de B H b qui ne prouvent rien. Si, en efet, B H b 276 276 PHILOLOGY AND CRITICISM ont une bonne leçon contre une faute ou plutôt une innovation de M, cela ne peut prouver seulement que le copiste de M, comme tout autre, a ses fautes personelles.” Paul Lejay, “Revue critique d’histoire et du littérature,” N. S. XXVI (1888): 281–83. Reeve, in response to d’Arco Silvio Avalle’s question “who irst saw that only shared innovations prove witnesses related?” retraces Froger’s steps and discusses why the alternative (that is, that shared readings prove witnesses related) does not hold. Reeve, “Shared Innovations, Dichotomies, and Evolution,” 55–103. 19 For a list of examples, see Timpanaro, The Genesis of Lachmann’s Method and see also Reeve, “Shared Innovations, Dichotomies, and Evolution.” 20 Glenn W. Most, “Editor’s Introduction,” in Sebastiano Timpanaro, The Genesis of Lachmann’s Method, ed. and trans. Glenn W. Most (Chicago: University of Chicago Press, 2005), 10. 21 Peter F. Dembowski, “The ‘French’ Tradition of Textual Philology and Its Relevance to the Editing of Medieval Texts,” Modern Philology 90, no. 4 (1993): 514. 22 Ibid. 23 This expresses the minimum condition, though ideally we should have several such errors, since scribes often corrected their source’s errors. However, since a scribe can only notice (and so correct) a percentage of the errors in his source, the true state of afairs should become apparent. 24 Paul Maas, Textkritik, 4th edn. (Leipzig:  B. G.  Teubner, 1960), 26. Likewise, in genetics the fact that someone possesses the genes for a hand does not permit us to conclude that he is descended from another person who also has the genes for a hand, since every human possesses the genes that program a hand. Thus, not the shared material, but the shared errors of tran- scription are decisive: if both individuals had a rare genetic defect, then and only then would we be justiied in positing a relationship (provided they could not have acquired the defect independent of each other). 25 A “signiicant error” does not mean that the error is especially meaningful, but, rather, that it cannot be such that two scribes made it independently. The error is signiicant if it signiies something for the two manuscripts’ iliation. Contrariwise, insigniicant errors can be signif- icant from the perspective of meaning, but they are still called insigniicant errors if they are worthless for determining iliation. 26 Compare Austin’s argument for Bigger’s normative redaction: “it is reasonable to build, on the basis of available evidence, inferences about things for which we have no evidence. Or, to articulate this claim in terms I have proposed, it may be argued that it is reasonable to infer that the same processes and patterns of textual growth which followed M0 (i.e. the many M+N moments documented in the CE apparatus) also preceded M0; just as the M0 text was established reductively from M+N by the Bhandarkar editors through lower textual criticism, hypotheses about M-N can be established from M0 through higher textual criti- cism.” Christopher Austin, “Evaluating the Critical Edition of the Mahābhārata: Inferential Mileage and the Apparatus Materials.” Journal of Vaishnava Studies 19, no.  2 (2011):  75. The problem is that the critical edition is already an inferred text. Its editors concluded their work not because they were apathetic or lazy, but because they exhausted all the licit inferences. All material supports only a inite number of inferences: inferences made thereafter are illicit. Hence the subtle shift in method:  since the irst method has ceased to yield results, Austin now invokes “higher textual criticism.” The similarity of the two names—higher and lower criticism—should not blind us to the fact that they are incom- mensurable. Austin disingenuously suggests that the various “M-N” moments established through higher textual criticism are analogous to the “M0” established from the “M+N” moments, even though the former are not established using a scientiic method much less the same method. 27 Reeve, “Shared Innovations, Dichotomies, and Evolution,” 59. Timpanaro expresses a similar idea in The Genesis of Lachmann’s Method, 175. 27 CONFUSIONS REGARDING CLASSIFICATION 277 28 The examples Bigger chooses (the appendix passages 63, 75, 80, 81 and 85) do not meet this criterion because they are interpolations, which can be transmitted horizontally. Further, he cannot show that the interpolation arose in a speciic branch of the tradition (for example, in γ with the consequence that all manuscripts containing them are descendants of γ). Since nothing can be concluded about the iliation of a group of manuscripts sharing common readings, not even that they form a family in the narrower sense, and the additional passages can be thought of as a kind of “common reading,” Bigger’s examples are genealogically irrel- evant. We could assert that γ is the source of Ñ1–3, V, B and D’s passages 63, 75, 80, 81 and 85 only if they contained a unique set of errors and the passages in Ñ1–3, V, B, D reproduced those errors, since it is unlikely that another manuscript would have contained the exact same errors. If we cannot establish iliation using shared readings, we can do so even less on the basis of a shared absence of readings, since the manuscripts that lack a characteristic do not consti- tute a true class, much less a genealogical family. 29 J. B. Hall, “Claudian, De Bello Gildonico,” review of Claudii Claudiani, De Bello Gildonico, ed. E. M. Olechowska, The Classical Review 33, no. 2 (1983): 203–07 discusses a similar problem with Olechowska’s division of the manuscripts of De Bello Gildonico into two families, α and β, based on common variants. 30 Bigger, Balarāma im Mahābhārata, 121. 31 For a discussion of Maas’s principle, see Elio Montanari, La critica del testo secondo Paul Maas: testo e commento (Florence: SISMEL edizioni del Galluzzo, 2003) and, for a response to Montanari, Michael D. Reeve, “Reconstructing Archetypes: A New Proposal and an Old Fallacy,” in Hesperos: Studies in Ancient Greek Poetry Presented to M. L. West on His Seventieth Birthday, ed. P. J. Finglass, C. Collard and N. J. Richardson (Oxford: Oxford University Press, 2007), 326–40. 32 Reeve, “Reconstructing Archetypes,” 329. 33 Taking Maas’s “ohne seine Hilfe” not to mean without its help tout court but no longer depen- dent on its help once reconstructed. 34 Reeve, “Reconstructing Archetypes,” 329. 35 Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 130. 36 Vishnu S. Sukthankar, “Epic Studies VII: The Oldest Extant Ms. of the Ā diparvan,” Annals of the Bhandarkar Oriental Institute 19 (1938–39): 201–62; see especially 205: “These signiicant omissions, major and minor, prove that the text of our MS. is shorter than that of all other MSS. hith- erto known” (all italics Sukthankar’s). 37 Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 107. 38 Ibid., 107. 39 Ibid., 108 (Grünendahl’s italics). 40 Ibid., 108, n. 29 (all italics Grünendahl’s). 41 Ibid., 108. 42 Ibid., 108, n. 31. 43 Grünendahl does not clarify how he calculates the “pure textual additions” of the northern recension. Does he include all additions found in one or more northern manuscripts, or does the deinition exclude additions shared with one or more southern manuscripts? If the latter, then the calculation is erroneous, for a text is no less expansive for sharing its additions with a manuscript of the rival recension. We count 43 appendix passages for the northern recension alone; 92 if we include passages shared with the southern recension. In contrast, Grünendahl appears to arrive at 64 passages by excluding those found only in a single northern manuscript when found in the southern recension as a whole. In that case, the count is 64, but we could just as well exclude passages found in two northern manuscripts when found in the southern recen- sion as a whole, since both are as likely to be contaminated with a southern source. Moreover, why should we include passages found only in a single northern manuscript, since they are also not characteristic of the northern recension as a whole? 278 278 PHILOLOGY AND CRITICISM 44 Grünendahl actually thinks this is the case, as we show in the section “The Argument from Extensive Contamination.” 45 This would be the case if a manuscript had many small interpolations as compared with another with fewer but more extensive interpolations. In Grünendahl’s table, the former would have a higher number than the latter. If we relied on his table for our knowledge of manuscripts, we would treat the latter as the purer manuscript. In general, it is erroneous to treat interpolations purely quantitatively: they must be examined in context. 46 See L. D. Reynolds and N. G. Wilson, Scribes and Scholars: A Guide to the Transmission of Greek and Latin Literature (Oxford: Clarendon Press, 1991), 214–16, on the problem of additional passages in one part of a manuscript tradition. 47 M. L. West, Textual Criticism and Editorial Technique (Stuttgart: B. G. Teubner, 1973), 50. 48 Given the Ś ā radā codex’s fragmentary nature, we cannot identify the total number of insertions unique to it. Sukthankar’s critical edition of the Ā diparvan lists only one such passage (*374 inserted after 1.31.11). A more accurate guide is to consider insertions in the ν recension (Ś and K) and especially in Ś 1 and K1, which have only two unique insertions (*1735 and *2161). In spite of the contamination of the inferior K manuscripts, the ν recension is the shortest known recension. 49 Reeve writes that he is currently writing an article on sections of a twelfth lorilegium fuller than the quotations taken from a related source by Sedulius Scottus in the ninth century; the question arises whether the additional material is a later interpolation. But he also notes that there are witnesses to the lorilegium that make a selection and so have a shorter but plainly unauthorial text. A witness’s quality cannot be determined by considering its length alone: only the quality of its readings permits us to arrive at a conclusion regarding its place in the stemma. Only then can we know whether it is a good guide and to what extent we ought to follow it. 50 Grünendahl does not look at the passages in context: he does not consider whether they intro- duce new themes or merely repeat or restate something said earlier; whether they were inserted to expand on an existing detail or to introduce new narratives or episodes; whether their inser- tion creates a hiatus and so on. Beyond calculating their relative frequency, he does not even consider the pattern of their occurrence in diferent recensions: do they occur in the better manuscripts or only in the more expansive groups; could they have been transmitted by con- tamination or are they more likely inherited from a common ancestor. Beyond the fact that they are listed as “interpolations” in Sukthankar’s apparatus, he is not interested in their status. His sole concern is to demonstrate the superiority of the Nepā lī manuscripts as concerns the number of interpolations. That he censures Sukthankar for arguing “exclusively in terms of quantity” is typical of his tendency to attribute his errors to others. 51 Michael D. Reeve, “Stemmatic Method: ‘Qualcosa che non funziona?’ ” in The Role of the Book in Medieval Culture:  Proceedings of the Oxford International Symposium, September 26–October 1, 1982, ed. P. Ganz. Bibliologia 3 (Turnhout: Brepols, 1986), 69, n. 36. 52 Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 108. 53 Ibid. 54 Ibid., 108–9. 55 Ibid., 109. 56 Ibid., 128. 57 Unlike interpolations, which the critical edition lists and where all one needs do is count the passages, calculating the length of the constituted text takes time and requires a careful study. Like Bigger, Grünendahl uses material the critical edition already provides him, makes some quick, inexact calculations and soon he can announce a new “theory” of critical editing (cf. Bigger’s claim that his work “overturn[s]” the “current maxim ‘short equals old,’ ” but pos- sibly permits us to approximate the “normative redaction’s actual form” more closely; Bigger, Balarāma im Mahābhārata, 121). 279 CONFUSIONS REGARDING CLASSIFICATION 279 58 Sukthankar notes that the manuscript in the collection of the Rajaguru (Grünendahl’s Ñ4) “though […] appreciably shorter than any other MS. of the Ā di so far known,” is not “abso- lutely free from inlation and contamination. This MS.  has in fact just a few insertions not countenanced by the Critical Edition.” Sukthankar, “The Oldest Extant Manuscript of the Ā diparvan,” 205, and see ibid., 206, for the list of insertions. In all, the manuscript contains three longer passages (App. 1, nos. 1, 12, 33 and 58) and 90 shorter one-line insertions (93*, 220*, 288*, 291*, 476*, 478*, 479*, 485*, 494*, 495*, 549*, 601*, 678*, 708*, 709*, 710*, 750*, 868*, 873*, 875*, 879*, 903*, 906*, 909*, 962*, 999*, 1004*, 1013*, 1017*, 1018*, 1022*, 1068*, 1087*, 1096*, 1104*, 1110*, 1120*, 1124*, 1141*, 1142*, 1144*, 1146*, 1148*, 1155*, 1162*, 1166*, 1173*, 1178*, 1186*, 1196*, 1212*, 1219*, 1224*, 1230*, 1240*, 1305*, 1309*, 1357*, 1430*, 1436*, 1463*, 1470*, 1623*, 1644*, 1665*, 1698*, 1786*, 1789*, 1796*, 1812*, 1815*, 1818*, 1855*, 1882*, 1883*, 1919*, 1920*, 1921*, 1922*, 1951*, 1953*, 1960*, 1975*, 2016*, 2087*, 2097*). 59 We could also have added a column for Ś and K (ν). However, in many cases, the ν recension follows the readings of S1 and K1, as the other K manuscripts are often conlated. 60 The only star passages listed are the previously mentioned *374 (the sole example of an inter- polation unique to Ś 1) and *1735 and *2161 (the sole examples of interpolations unique to Ś 1 and K1). 61 Ś 1 has a lacuna here, so it is not clear whether we would include no. 33 as we have no way of proving it existed in the shortest witness. But we might assume that it did given Ś 1’s proximity to K1 and therefore include it in our constituted text. 62 Silvia Rizzo discusses examples in her review of Il lapsus freudiano. Psicanalisi e critica testuale, by Sebastiano Timpanaro, Rivista di ilologia e di istruzione classica (1977): 102–5, but her examples are on a smaller scale. She believes that if the editor must choose between “two equally possible readings, one of which has a few additional words not necessary for the sense,” he ought not exaggerate the potential for interpolation, but should “take into account the [scribes’] tendency to eliminate superluous words.” Ibid., 105. 63 Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 109 (italics in original). 64 Ibid., 115. 65 Ibid., 127. 66 See ibid., 108, n. 29. Grünendahl glosses the expression “factual increase [in the size] of the text” with: “[passages,] which thus do not merely ofer another line or another verse in place of a line or a verse of the constituted text or insert parts of the constituted text in another position” (all italics in original). 67 V. S. Sukthankar, “Prolegomena,” in The Ad̄ iparvan for the First Time Critically Edited, vol. 1 (Poona: Bhandarkar Oriental Research Institute, 1933), xlvii. 68 Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 108. 69 Sukthankar, “Prolegomena,” lxxxiv. 70 Ibid., xcvii. 71 Ibid., lxi. 72 “Of important additions in ν, I can recall only one, that of an adhyā ya of 42 lines, at the very end of the Ā di (added probably as an Appendix), which is a variant, abbreviated version of the Ś vetaki interpolation. It is found only in Ś 1 K0.1.4 and therefore cannot even be said to be characteristic of the whole of ν (App. I, no. 121). Instances of small additions are nos. 349*, 449*, 451*, 516*, 565*, etc., found in K with or without some Devanā garī manuscripts; while 969*, 1855*, 2077*, etc., are found in Ś 1K, with or without some Devanā garī manuscripts: all these passages are missing in B [and] S.” Ibid., lii (Sukthankar’s italics). 73 Ibid., xlvi–xlvii. 74 Even the terms quantitative and qualitative are misnomers, arising from Grünendahl’s habit of counting interpolations rather than studying readings. How is a manuscript’s length any less of a qualitative criterion in evaluating it than its antiquity? How is the fact that a manuscript 280 280 PHILOLOGY AND CRITICISM is heavily interpolated any more a quantitative criterion—ignoring the absurd procedure whereby one counts interpolations rather than examine them—than its readings? 75 Ibid., lxxxvi. 76 Ibid., Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 109. 77 Ibid., 128. 78 Ibid., 109 (italics in original). 79 Ibid., 109. 80 Sukthankar, “Prolegomena,” lxxxvi (italics in original). 81 Ibid., xlvii. 82 Sukthankar, “Prolegomena,” lxi. Grünendahl could also have referred to Sukthankar’s article “More Text-Critical Notes,” which defends the constituted text against Winternitz’s objections. Sukthankar explicitly notes:  “I have taken the Ś ā radā MS.  only as the norm for my edition; I have not undertaken to reproduce its text verbatim.” Vishnu S. Sukthankar, “Epic Studies IV. More Text-Critical Notes,” Annals of the Bhandarkar Oriental Research Institute 16, no. 1/2 (1934– 35): 97. He lists several examples that illustrate that, in several cases, he rejected the Ś ā radā codex’s reading. 83 See Sukthankar, “Prolegomena,” xlvi–xlvii. 84 Sukthankar notes that in instances of “ ‘double’ agreement [‘agreement between two or more groups of each recension (N1  =  S1 and N2  =  S2)’]” “one of the agreements must, gener- ally speaking, be accidental, since both can hardly be original.” In this case, “either may be adopted, if they have equal intrinsic merit.” However, “owing to the much greater correctness and reliability of Ś 1 K” and “other things being equal,” he adopts the reading of Ś 1 and K in these cases. Ibid., xci. 85 See Grünendahl’s conclusion: “From what has been said until now certain consequences have become clear for the editorial principles presented by Sukthankar. When the Schriftartprämisse lose their place in the tool-box of Indological textual criticism, the classiication of manuscripts must be placed on new foundations. It appears logical to focus initially on the manuscripts of individual regions. In particular, the relationship of the Newari manuscripts to each other, their position in the manuscript tradition of Nepal as well as of the Mahā bhā rata as a whole deserve a closer investigation. The conditions for this are given by the work of the Nepal-German Manuscript Preservation Project.” Grünendahl, “Zur Klassiizierung von Mahā bhā rata- Handschriften,” 130. See also the next note. 86 “In particular, Sukthankar’s classiication is placed in question by Ñ4: (1) it is shorter than the textus simplicior constituted by him [sic], without standing in a demonstrable relationship to it; (2) it fulills many of Sukthankar’s criteria for his constitution of the hyparchetype ν discussed above; (3) it does not show any traces of contamination; (4) contrary to Sukthankar’s claim it stands in a closer relationship to only one other Newari manuscript, namely, Ñ3; (5) according to Sukthankar’s information, it manifests conformities with the southern recension that, when they cannot be traced back to contamination, are possibly to be ‘traced back to an original connection via the lost archetype.’ ” Ibid., 129. 87 Sukthankar considers four types of agreement—N = S (text = NS), N1 = S (text = N1S), N = S1 (text  =  NS1) and Nx  =  Sx (where x represents the series of cardinal numbers 1, 2, 3, etc.) (text = Ś 1K)—and one type of disagreement: N ≠ S (text = N). Only in one of these ive cases does he adopt the reading of Ś 1 and K without invoking extrinsic justiication. In other words, even if we followed Grünendahl in his erroneous assumption that, qua shorter, Ñ4 should be the base for the critical edition, we would replace the reading of Ś 1 and K with Ñ4’s reading only in this case. Here also Sukthankar does not justify adopting their reading on the ground that they form the shortest group. Rather, he explicitly notes that their readings are adopted in these cases owing to their “greater correctness and reliability.” Sukthankar, “Prolegomena,” xci (italics added). 88 Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 104. The reference is to Sukthankar’s statement that “in the case of works that are transmitted in diferent versions, the script is characteristic of the version.” Sukthankar, “Prolegomena,” vii. 281 CONFUSIONS REGARDING CLASSIFICATION 281 89 Heinrich Lüders, Über die Grantharecension des Mahābhārata, Abhandlungen der Königlichen Gesellschaft der Wissenschaften zu Göttingen, Phil.-Hist. Klasse, N.  F.  vol. 4, no.  6 (Berlin: Weidmannsche Buchhandlung, 1901). 90 Ibid., 3–4 (Lüders’s italics). 91 Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 102 (italics in original). 92 Ibid., 103. 93 Heinrich Lüders, review of The Mahābhārata for the First Time Critically Edited, by Vishnu S. Sukthankar, Deutsche Literaturzeitung 24 (1929): 1140. 94 Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 104. 95 Ibid., 105. 96 Heinrich Lüders, Druckprobe einer kritischen Ausgabe des Mahābhārata (Leipzig: Drugulin, 1908). 97 Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 104. 98 West assigns this name to the stage following presentation (West, Textual Criticism and Editorial Technique, 72–102). It involves selection of prefatory material, choice of sigla and the layout of the edition and the text and critical apparatus. 99 The image is reprinted as the frontispiece to this book. 100 Contini rightly proposes, “a critical edition is, like any other scientiic act, a mere working hypothesis, the most satisfactory, namely, the most economic one, and one which proves apt to connect a system of data.” Gianfranco Contini, “Ricordo di Joseph Bedier,” in Esercizi di lettura sopra autori contemporanei con un’appendice su testi non contemporanei (Turin: Einaudi, 1974), 369. Leonardi develops this idea when he writes, “I hope these three coordinates (scientiic accuracy, diachrony, readability) make it clear that referring to a critical text as a hypothesis doesn’t mean at all justifying the skepticism of contemporary philology, but in fact quite the opposite. The fact that this text originates from a hypothesis doesn’t imply that every hypothesis is valid; on the contrary, it demands a defendable justiication to support its validity.” Leonardi, “Il testo come ipotesi,” 6. 101 West, Textual Criticism and Editorial Technique, 72. 102 For more examples, see the lists of manuscripts forming the critical apparatus for the diferent parvans in the respective volumes of the Mahā bhā rata critical edition. 103 Sukthankar, “Prolegomena,” v. Grünendahl appears unaware of this passage. He writes: “The inding in other parvans shows rather the opposite:  in Mahā bhā rata 4, two of the three Ś ā radā manuscripts available were not included, because they proved to be ‘recent copies of Nīlakaṇtha’s text and therefore as Ś ā radā codices without value’ (Raghu Vira, Mahā bhā rata 4, Introduction, p. ix f.)—this also was a case that manifestly was not anticipated by the spiri- tual fathers of the Schriftartprämisse.” Grünendahl, “Zur Klassiizierung von Mahā bhā rata- Handschriften,” 109–10. Sukthankar’s footnote (Sukthankar, “Prolegomena,” v, n. 1) shows that they are the same manuscripts that Raghu Vira later cites in his edition: Grünendahl has overlooked the citation. Hence his confusion that the discovery of Ś ā radā copies of a Devanā garī text is a fatal objection to the Schriftartprämisse:  he has formulated the premise independent not only of the manuscript evidence but also of the editor’s views. 104 Sukthankar, “Prolegomena,” vii. 105 Ibid. 106 West states the underlying principle as follows: “If the older manuscripts can be itted into a stemma, the promiscuity of the younger ones should be easy to establish and may be unimpor- tant. If the relationship of the older manuscripts resists analysis, it may still be possible to identify sub-groups whose structure can be stemmatized.” West, Textual Criticism and Editorial Technique, 15. 107 Johann Gottfried Ludvig Kosegarten, Pantschatantrum, sive, Quinquepartitum de moribus exponens: ex codicibus manuscriptis (Bonn: H. B. Koenig, 1848), vi. 108 This explains why Lüders could identify a Kaśmīrī version of the Mahā bhā rata, even though he did not possess an example in the Ś ā radā script. Even though the manuscripts he iden- tiied as the Kaśmīrī version were in Devanā garī script, their common features justiied the assumption of a common ancestor. Sukthankar acknowledges Lüders’s perspicuity:  “The 28 282 PHILOLOGY AND CRITICISM Kaśmīrī character of K1 was already fully recognized by Professor Lüders, who had uti- lized it in the preparation of his specimen of a critical edition of the epic mentioned above, although he had no genuine representative of the Kaśmīrī or Ś ā radā version to compare it with.” Sukthankar, “Prolegomena,” xlix. Grünendahl thus understands neither the prin- ciple that an editor can identify manuscripts as of a certain type or descended from a common ancestor even in the absence of that ancestor nor Sukthankar’s comment when he (Grünendahl) remarks that “When he [Sukthankar] remarks in the Prolegomena (p. xlix) that Lüders ‘already recognized the Kaśmīrī character of K1, […] although he did not have a genuine representative of the Kaśmīrī or Ś ā radā version for comparison,’ this holds no less for his own constitution of the K-version in the Foreword.” Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 112. Sukthankar does not censure Lüders for consti- tuting a K version in the absence of an exemplar; rather, he praises him for recognizing the existence of this version even in the absence of an exemplar. Likewise, he does not need an extant exemplar to posit a Kaśmīrī version. Grünendahl attributes his misunderstanding to Sukthankar. 109 In the Mahā bhā rata, we initially get two subgroups in the Grantha version, G1–3 and G4–6, which soon changes to G1.2.4.5 versus G3.6, that is to say, G1 (and G2) exhibit ainities with G4–5 rather than with the purer southern group G3–6. See Sukthankar, “Prolegomena,” lxxii. These changes suggest a change of exemplar, but if the editor is sensitive to them and the reader is aware of the change in weight assigned to the manuscript (which he can do if he reads the editor’s account of manuscripts and studies the table of manuscripts collated for diferent portions of the text), they should not undermine conidence in the edition. 110 Although this method allows for a localization of the archetypes (Sukthankar notes that ν is the “lost archetype of the North-Western group”; γ is the “intermediate (inlated) source” of the “central sub-recension […] (comprising the Eastern and Western groups) [and] occupying a position intermediate between the North-Western and the Southern groups”; ε is the “lost archetype of the Eastern group”; while σ is the “lost archetype” of TG, the manuscripts from the southeastern corner of the Indian peninsula; Sukthankar, “Prolegomena,” xxx), there is a diference between identifying a broad area of inluence depending on the hyparchetypes’ position in the stemma and the traditions they engendered and simply positing “recensions” based on an a priori scheme of classiication, as Grünendahl suggests Sukthankar does. Grünendahl’s error arises from confusing the consequent and the antecedent of the argu- ment: Sukthankar does not posit a northern and a southern recension because he thinks the two families of scripts—north Indian and south Indian—had a common script as ancestor, but because he observes that the manuscripts in the two families behave as though they had a common manuscript as ancestor. In this case, one cannot hypothesize a simple oversight: the incoherence is in Grünendahl’s mind. 111 Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 106, n. 24. 112 Lüders, Über die Grantharecension des Mahābhārata, 3–4 (Lüders’s italics). 113 Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 103. Neither the idea nor the criticism is new. The idea that the Mahā bhā rata manuscripts were classiied by script occurs in Walter Ruben’s 1930 article “Schwierigkeiten der Textkritik des Mahā bhā rata,” Acta Orientalia 8 (1930): 240–56. Ruben claimed that “Sukthankar, inspired by Lüders (Abh. d.  Gött. Ges. d.  W. 1901, p.  3) groups the mss. according to their scripts in ‘versions’; he groups multiple mss. of the same script in ‘groups.’ ” However, he argued, “this division may not be taken all too seriously for textual criticism, since the mss. cannot always be consis- tently pressed into the same groups, since individual groups of diferent versions (of the same recension) go together just as often and since, inally, the mss. (or groups) break out of the framework of their recension more or less often.” Ibid., 242. Grünendahl cites the passage verbatim, additionally emphasizing the words “diferent” (verschieden) and “same” (dieselbe), in defense of his claim that “regarding Sukthankar’s grouping of the mss., a reservation is 283 CONFUSIONS REGARDING CLASSIFICATION 283 already expressed in Ruben’s essay, which he later, in another place, formulates more clearly.” Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 103–4 (the quote is cited in n.  15). He also cites Ruben’s remark that “even in the case of the Mahā bhā rata, one cannot group the mss. according to their scripts in groups, as Sukthankar has attempted to do.” Ibid., 104 (the reference is to Walter Ruben, Studien zur Textgeschichte des Rāmāyanạ [Stuttgart: W. Kohlhammer, 1936], 1–2). 114 Lüders, Über die Grantharecension des Mahābhārata, 4 (italics added). 115 Grünendahl, “Zur Klassiizierung von Mahā bhā rata Handschriften,” 102. 116 Sukthankar, “Prolegomena,” vii. 117 Grünendahl, “Zur Klassiizierung von Mahā bhā rata Handschriften,” 106. 118 Ibid., 106. 119 Ibid., 106, n. 24. 120 Sukthankar, “Prolegomena,” lxxv. 121 See Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 107. He footnotes this passage with the words: “Ruben had already expressed himself in very sim- ilar terms in his discussion” (ibid., 107, n. 28), but appears to be aware neither that Ruben’s comment was based on a misreading of Sukthankar’s discussion of the manuscripts nor that Sukthankar had already addressed Ruben’s confusion in Vishnu S. Sukthankar, “Epic Studies III: Dr. Ruben on the Critical Edition of the Mahā bhā rata,” Annals of the Bhandarkar Oriental Institute 11, no. 3 (1930): 259–83. In this brief, 22-page article, Sukthankar destroys Ruben’s argument. He shows that Ruben is a poseur, who lectures others on textual criti- cism, but himself does not understand its principles. “R.  might have mercifully spared us this learned prooemium on the aims and methods of Classical Philology and their applica- tion to the problem of the Mahā bhā rata textual criticism. The same thing has been said by others before R., more simply but with ample clearness and emphasis.” Ibid., 261. He refutes Ruben’s contention that the manuscripts were grouped by script: “This conclusion is supported by numerous other agreements throughout the Ā diparvan. That the three new Malayā lam MSS. M6. 7. 8 again go back to the same original follows not merely from the numerous readings these MSS. have in common, but conspicuously from one particular mis- take where they repeat inconsequentially, at the same point, a fragment of a stanza (1. 85. 25):  pūjayantı̄ha loke nāsādhavaḥ . Further the four Grantha MSS. G1. 2.  4. 5.  […] must go back to a not very distant common original. This group has not only numerous readings in common, but it contains quite a considerable number of interpolations peculiar to itself.” Ibid., 265 (Sukthankar’s italics) and see also ibid., 264–65. At the end of this overview of his classiication, Sukthankar notes: “It will thus be seen that these generalizations of R. are absolutely unwarranted. They appear to have been introduced by R. merely for the satis- faction of having made some pointed and efective little statement.” Ibid., 265. Sukthankar also addresses Ruben’s contentions regarding the G version (see Ruben, “Schwierigkeiten der Textkritik des Mahā bhā rata,” 242, n. 1): he notes that Ruben overlooks both his comments regarding the Grantha manuscripts’ conlated nature (“all Grantha manuscripts are prob- ably contaminated (directly or indirectly) from Northern sources in diferent degrees”; Sukthankar, “Prolegomena,” lxxii; italics in original) and his comments regarding the secondary nature of the commonalities between G1.2.4.5 and N (“Now is this a case of contamination of the four MSS. G1.2.4.5 from a Northern source; or are the common passages a remnant of the lost archetype, which were somehow lost in the remaining manuscripts of the Southern recen- sion? There is apparent agreement here between independent versions. But is this agreement original? […] The said manuscripts contain a Northern stanza (belonging to manuscripts of class T)—a mere string of attributes of Garuḍa—wedged in at a place where it can be construed neither with what precedes nor with what follows. This proves incontrovertibly that these four manuscripts G1.2.4.5 have been compared with some Northern manuscripts, and makes it highly probable that the other doubtful stanzas, which they have in common 284 284 PHILOLOGY AND CRITICISM with the Northern recension, have crept into their text in the same surreptitious way”; ibid., lxxxi). Leaving the reader in no doubt of his low opinion of Ruben as a textual critic (“There are in the Ā diparvan far more complicated passages than this:  this is one of the simpler conlations. I shudder at the thought of what R. would have done if he had to handle one of the really diicult cases of conlation”; Sukthankar, “Dr. Ruben on the Critical Edition of the Mahā bhā rata,” 271), he shows that Ruben is not even consistent in his argument: on one hand, he criticizes Sukthankar for treating the agreement of K and S as an argument for the reading of the archetype; on the other, he concurs with Lüders that the agreement of B with G (Lüders’s S) is an argument for the reading of the archetype. See Ruben, “Schwierigkeiten der Textkritik des Mahā bhā rata,” 247–48. But, as Sukthankar notes, “as a matter of documentary probability, there is no diference whatsoever between the con- sensus of K and S against B on the one hand and the consensus of B and G against D on the other.” Sukthankar, “Dr. Ruben on the Critical Edition of the Mahā bhā rata,” 276. Lüders also expressed approval of Sukthankar’s procedure in his review of the irst fascicule: “The south Indian manuscripts agree with none of the other northern versions as closely as they do with the Kashmiri version. Furthermore, they have none of the manifest additions of the southern recension. The geographic distance between the two groups of manuscripts practically rules out the direct descent of the southern recension from the Kashmiri version. One must therefore assume that in cases where the two agree in their variants, we have the reading of the oldest text that can be attained. I had already arrived at the time I was com- pleting my sample edition at this evaluation of the relationship between the Kashmiri and the southern manuscripts, which Sukthankar now [also] argues for.” Lüders, review of The Mahābhārata for the First Time Critically Edited, 1141. Sukthankar concludes with a penetrating summary of Ruben’s shortcomings (Sukthankar, “Dr. Ruben on the Critical Edition of the Mahā bhā rata,” 282–83). His inal judgment is withering: “Ruben’s paper is, in general, quite intelligent but not at all important for the Mahā bhā rata textual criticism.” Ibid., 283. 122 Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 107. 123 Ibid. (the question mark is Grünendahl’s insertion). 124 Sukthankar, “Prolegomena,” xlvii. Grünendahl deletes the words “seems to be” and Sukthankar’s observation that the manuscript “may be three or four centuries old.” In gen- eral, he makes Sukthankar’s conclusions more categorical and less considered than they are. 125 Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 109. Grünendahl’s superciliousness is evident, but the confusion is entirely his own. Both Lüders and Sukthankar are aware that the presence of contamination complicates the identiication of regional versions. But they do not allow it to confuse them regarding the underlying transmission. As Reeve notes, the real problem with contamination is not that it “has the efect […] of falsifying stemma but [that it] frustrat[es] attempts at drawing them up.” Reeve, “Stemmatic Method,” 67. Lüders and Sukthankar focus on irst identifying characteristic features of the regional versions before they address contamination. In contrast, Grünendahl begins with contam- ination and argues abstractly from it against drawing up a stemma. He can do so because he is not interested in a critical edition. Rather than follow the principle primum recensere (do the recensio irst), he raises the specter of contamination to undermine Sukthankar’s edition. He could have beneited from Reeve’s observation: “Scholars [who] believe contamination and interpolation were so common that stemmatic method seldom or never works […] are biting the hand that fed them, because unless one can actually see scribes contaminating and interpolating under one’s nose it is only by applying stemmatic method that one can detect contamination and interpolation.” Ibid., 65. Apparently, Grünendahl had such a good view of the scribes at work he does not require stemmatic method at all. 126 Reeve’s deinition. 127 In certain circumstances it may be advantageous to reduce the number of witnesses, for example, if the editor faces a large number of manuscripts. But here also he cannot proceed 285 CONFUSIONS REGARDING CLASSIFICATION 285 arbitrarily. He must irst show that the selected witnesses suice to give an account of the tra- dition. West provides the following account of the steps an editor must follow: “1. Whenever the manuscripts are at variance, make a note of the reading or readings that seem to be ancient (true, and not found by conjecture, or else attested by an ancient source unavailable to the scribes) and the manuscripts in which it or they appear. 2. Any manuscript that is the sole carrier of such readings is obviously indispensable. Adopt it. 3.  Remove from the list all the readings for which the manuscripts just adopted may serve as sources (not just the readings that appear only in them). 4. See which manuscript contains the largest number of the remaining readings. Adopt it. Remove from the list the readings it contains. 5.  Repeat (4) until every ancient reading is accounted for.” West, Textual Criticism and Editorial Technique, 43 (West’s italics). Unlike the elimination of the codices descripti, this elimination has a purely practical aim: in contrast to the former, where the eliminated witnesses cannot tell us anything new, here it is possible that the witnesses we set aside contain more information about the tra- dition, only we have decided to simplify our task by selecting only the manuscripts that suice for an overview of the tradition. Applying West’s steps, we can immediately identify the error with Grünendahl’s proposed elimination of the Kaśmīrī version: most of the ancient and good readings in the Ā diparvan were from the Ś ā radā codex. In any elimination of witnesses on pragmatic grounds, the Ś ā radā codex would be eliminated last (in fact, it would not be elimi- nated). At the latest by the second step, we would realize it was indispensable and adopt it as one of the manuscripts for our edition. 128 Ibid., 44. 129 For example, the Ś ā radā version and the Maithilī version, of which Sukthankar had only one exemplar (V1). In contrast, in the Sabhā parvan, three versions are represented by lone exemplars:  the Ś ā radā (represented by Ś 1), the Nepā lī (Ñ1) and Maithilī (V1; this is not the same V1 as in the Ā diparvan). By Grünendahl’s logic, the editor of the Sabhā parvan should not have constituted a Nepā lī version! Grünendahl does not address this issue, because it undermines his attempt to prove that the Nepā lī manuscripts are the true bearers of the Mahā bhā rata tradition. At the time, he worked for the Nepal-German Manuscript Preservation Project and evidently hoped his access to the Nepā lī manuscripts would secure him a position as editor of a new Mahā bhā rata critical edition. 130 The hyparchetype b can be reconstructed using BCEF and therefore replaces these manuscripts in the stemma (eliminatio codicum descriptorum). Its case difers from g. 131 There is no manuscript called MŚ :  this siglum is Belvalkar’s designation for the ancestor of M2, M4 and M5. The subscript letters denote “the versions with which the major version is secondarily related” (in this case Ś ā radā ). S. K. Belvalkar, “Introduction,” in The Bhīṣmaparvan for the First Time Critically Edited (Poona: Bhandarkar Oriental Research Institute, 1947), cxiv. 132 For Belvalkar’s views on the “Malayā lam version,” see ibid., cxi–cxiii. 133 We could replace the manuscripts M1–5 and the hypothetical hyparchetype MŚ with a single group M with no diference to the constitution of our archetype. Belvalkar notes that “M3–5 seem to have undergone, in parts, conlation with some Late-Northern MSS. [either D5 or D6 or their ancestor in his stemma]. […] M2.4 [, in contrast,] are the rebels of the version: they often igure as the exceptions to the list of omissions and additions found in the entire version (p. xlix–l). These are also exactly the MSS. which show the Ś K inluence (p. li–lii), which is also shown, to a smaller extent, by M5 (p. liii).” Ibid., cxi. Belvalkar’s comments suggest that M1 is the purest of the M manuscripts and that its agreement with Ś 1 provides the strongest argument for their reading being the reading of the archetype. 134 Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 109–10. 135 See Sukthankar, “Prolegomena,” v, n. 1. 136 See ibid. for the source of the manuscripts as the Raghunatha Temple Library and Sukthankar’s comment that “they represent probably the Nīlakaṇṭha version.” Sukthankar 286 286 PHILOLOGY AND CRITICISM lists the manuscripts’ numbers as 3712–32 and 3951–79 (numbered according to Aurel Stein’s catalogue of 1894). He also refers to the manuscripts two pages later as copies of the “Nīlakaṇṭha (Devanā garī) version written in Ś ā radā ” (the manuscript numbers are the same; in the footnote he adds Jammu as the location). Ibid., vii. Raghu Vira does not provide man- uscript numbers. He refers to them only as “the Ś ā radā MSS. deposited in the Raghunath Temple Library, Jammu (Kashmir).” Raghu Vira, “Introduction,” in The Virātạ parvan for the First Time Critically Edited (Poona: Bhandarkar Oriental Research Institute, 1936), ix. 137 Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 111. The reference is to Franklin Edgerton, review of The Mahābhārata, for the First Time Critically Edited, by V. S. Sukthankar, Text of 1.1.1 to 1.2.233, Journal of the American Oriental Society 48 (1928): 187. 138 Edgerton, review of The Mahābhārata, for the First Time Critically Edited, Text of 1.1.1 to 1.2.233, 187. 139 Ibid. 140 See Franklin Edgerton, “Tribute from the West,” Annals of the Bhandarkar Oriental Research Institute 24, no.  1/2 (1943):  136:  “I have just received the news of the death of Dr.  V.  S. Sukthankar. It is not only a very grave personal loss to me; I counted him one of my best friends, and I had come to feel a very deep respect and even afection for him as a man. But the loss to scholarship is immeasurable, and, naturally, far more important. I am appalled at the thought that it will now be necessary to entrust the Mahā bhā rata edition to others. Few persons now living are as well gifted by nature as he was with the peculiar combination of intellectual qualities needed for this work. And literally not one has had the experience which he had, and which is second in importance only to that native ability. He had arrived at a point where so many things had become almost automatic for him, like second nature; things which even those of us who have helped in the edition cannot control as he did, though we may have painfully struggled towards an approximation of a few of them. Now, just when he could have exploited to the full this unique combination of knowledge and experience— jñānam savijñānam—he is cut of in the midst of it.” 141 Edgerton disagrees with Sukthankar on very few points. The distinction of the Mahā bhā rata tradition into three recensions—the northwestern, northeastern and southern—is one. 142 In a review of the succeeding fascicules, Edgerton explicitly noted the addition of the Ś ā radā codex to the critical apparatus: “Of overwhelming importance is the unique ancient Ś ā radā ms. which he calls Ś 1.” Franklin Edgerton, review of The Mahābhārata, for the First Time Critically Edited, Ā diparvan: fascicules 2 and 3, by V. S. Sukthankar, Journal of the American Oriental Society 49 (1929): 283. In his next review, he again championed Sukthankar’s identiication of the Kaśmīrī version. See Franklin Edgerton, review of The Mahābhārata, for the First Time Critically Edited, Ā diparvan: fascicule 5, by V. S. Sukthankar, Journal of the American Oriental Society 52, no. 3 (1932): 252–55. His last review (of the seventh and last fascicule in 1936) included the following concluding assessment: “It seems to me now scarcely possible for an open-minded and well-informed Sanskritist to question, any longer, the tremendous value of the under- taking, nor the brilliant success of the performance. (Few such persons, I  may add, have questioned these matters even previously.)” Franklin Edgerton, review of The Mahābhārata, for the First Time Critically Edited, by V. S. Sukthankar, Ā diparvan: fascicule 7, Journal of the American Oriental Society 56, no. 3 (1936): 361. Grünendahl apparently thinks he is more intelligent than Sukthankar, Edgerton, Lüders and nine other Mahā bhā rata editors and countless reviewers combined. 143 Edgerton, review of The Mahābhārata, for the First Time Critically Edited, Ā diparvan: fascicules 2 and 3, 283. Grünendahl does doubt this fact, but this is, as we know by now, because he does not read critical apparatuses. 144 Edgerton, review of The Mahābhārata, for the First Time Critically Edited, Ā diparvan: fascicule 5, 254. 145 Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 111. 287 CONFUSIONS REGARDING CLASSIFICATION 287 146 Sukthankar, “Prolegomena,” xii. 147 Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 111. 148 Ibid., 113. 149 V. S. Sukthankar, “Introduction,” in The Ar̄ aṇyakaparvan for the First Time Critically Edited, vol. 1 (Poona: Bhandarkar Oriental Research Institute, 1942), xxi. 150 Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 113. 151 Ibid., 113 (italics added). 152 Sukthankar, “Prolegomena,” viii. 153 Ibid., viii. 154 Ibid., xxx (italics added). 155 Ibid., x. See also Sukthankar’s more detailed discussion of the Ś ā radā codex in the “Editorial Note” to the third fascicule of the Ā diparvan (the codex was already added to the critical apparatus in the second fascicule from adhyāya 26 onward):  “This birchbark (bhū rjapatra) codex measures 12 in. by 9 1/2 in. and comprises 339 folios. The characters are Old Ś ā radā , of perhaps the sixteenth or seventeenth century. The lines of writing, as is usual in Ś ā radā manuscripts, run parallel to the narrow side of the leaf.” V. S. Sukthankar, “Editorial Note (2),” in The Mahābhārata for the First Time Critically Edited, Ad̄ iparvan: Fascicule 3 (Poona: Bhandarkar Oriental Research Institute, 1929), i. 156 Sukthankar, “Prolegomena,” xi. 157 Sukthankar, “Introduction,” xxi (all italics added). 158 Franklin Edgerton, “Introduction,” in The Sabhāparvan for the First Time Critically Edited (Poona: Bhandarkar Oriental Research Institute, 1944), xliii. 159 See Sukthankar, “Introduction,” xix, n. 1 where Sukthankar’s comments extend beyond his earlier observations concerning the script and writing material to the Kaśmīrī character of its text: “The Kaśmīrī character of the text of our Ś ā radā manuscript, which is suggested by the script (Ś ā radā ) and the writing material (bhūr japatra), is conirmed not only by the introductory mantra nārāyaṇaṃ namaskṛtvā, etc.; but also by the regular omission of the word uvāca in such references as vaiśaṃpāyana uvāca; and further by such Kaśmīrī spellings of names like cārudoṣṇa (for cārudeṣṇa) in 3.17.22, corresponding to sudoṣṇā (for sudeṣṇā) in the Virā ṭa; cf. Raghu Vira, Introduction to the Virā ṭaparvan, p. xi. The Bhā ratamañjarī has cārudoṣṇa in 3.159 and sudoṣṇā in 4.36, etc. That is a real Kaśmīrī trait.” 160 For instance, the Ā diparvan manuscript K1, of which Sukthankar notes, “Even the out- ward form and get-up of this MS. are suggestive of Kaśmīrī origin. The lines of writing, as in Ś ā radā (bhū rjapatra) MSS. run parallel to the narrow side of the folio. The signatures in the margin are like those found in Kaśmīrī books. The numerous clerical errors, which disigure every page, betray the writer to be a professional scribe, not thoroughly familiar with the awkward Śāradā script [...].” Sukthankar, “Prolegomena,” xi. Likewise, he notes of the Ā raṇyakaparvan manuscripts K1 and K2 that “the handwriting of the original is Kashmirian in style” and that the “name of the scribe appears as: udairām liṣāyataṃ miśrajı̄ / śrı̄ sadānandjı̄; in the Udyogaparvan, the scribe’s name is given as Kā śmīrīya Sadā nanda,” respec- tively. Sukthankar, “Introduction,” iii. Raghu Vira notes of the Virā ṭaparvan manuscripts K1 and K2 that they feature “Devanagarī characters of the Kaśmīrī type. Orthographically one of the most corrupt as it has been copied by a scribe not well versed in the ligatures of the Ś ā radā script. K1, together with K2, writes uniformly sudoṣṇā for sudeṣṇā—an ancient Kaśmirian hyper-sanskritization, also found in Kṣemendra’s Bhā ratamañjarī.” Vira, “Introduction,” iv (Vira’s italics). Belvalkar notes of the Bhīṣmaparvan manuscripts K0 and K1 that they feature “Devanā garī characters of the Kashmirian type.” Belvalkar, “Introduction [to the Bhīṣmaparvan],” xx and xxii. He adds:  “From a study of the various post-colophon data found at the end of the parvans, we learn that a Kashmirian Pandit, Miśra Sadā nanda by name, engaged ive scribes. […] Presumably owing to the circumstance that scribes more familiar with the Devanā garī than with the Sā radā are apt to be misled by the similarity of 28 288 PHILOLOGY AND CRITICISM the Ś ā radā u, ta, na, ra, ma, śa, etc., most Devanā garī copies from a Ś ā radā original betray a persistent tendency towards haplographical omissions. In K2, we have noticed at least 60 such cases.” Ibid., xxiv–xxv. 161 See Sukthankar, “Prolegomena,” vii: “The reason for this concomitance between script and version appears to be that the scribes, being as a rule not conversant with any script but that of their own particular province, could copy only manuscripts written in their special provin- cial scripts, exception being made only in favour of the Devanā garī, which was a sort of a ‘vulgar’ script, widely used and understood in India” and ibid., lxii–lxiii: “It seems more nat- ural to regard, as already observed, the Devanā garī as a sort of ‘vulgar’ script (like the Latin, in Europe), the script understood by the savants all over India, into which many of the local versions were, from time to time, transcribed, a circumstance which facilitated contamination and conlation.” 162 See Belvalkar, “Introduction [to the Bhīṣmaparvan],” xxix:  “This is the irst time that a MS.  written in Bengali characters has been classiied with K, which normally designates Devanā garī copies from a Ś ā radā original. But, à priori, there is no reason why direct copies from a Ś ā radā original should not be made in Bengali characters. The text would in course of transmission be contaminated with the Bengali version, just as, in the other case, it would be with the Devanā garī version. But the MS. must certainly be classiied as K if it possesses suicient and distinctive characteristics of that group.” Belvalkar gives no hint of a con- lict between Sukthankar’s editorial practice and  his. Patrick Olivelle likewise notes in his critical edition of the Mānavadharmaśāstra: “The ST is represented by the following mss.: GMd1 TMd3 TMd4 GMd5 GMy, Tr1 MTr4 MTr5 MTr6. Most mss. written in the southern scripts of Grantha, Telugu, and Malayalam appear to fall under the ST, although there are exceptions. Tr1, which is written in Devanāgarī, follows the ST; it is probably a Devanāgarī copy of an original in a southern script. The ms. MTr3, on the other hand, although written in the Malayalam script, follows NT; it is probably a Malayalam copy of a northern orig- inal.” Patrick Olivelle, ed., Manu’s Code of Law: A Critical Edition and Translation of the Mānava- Dharmaśāstra (New York: Oxford University Press, 2005). Apparently, the only person uncertain about how manuscripts are classiied is Grünendahl himself. In contrast, anyone who has ever worked with manuscripts appreciates the distinction between a manuscript’s script and its text. 163 This is true even of P. L. Vaidya, who misleadingly states in his introduction to the critical edition of the Harivaṃśa that “the Mss. should be divided or classiied, according to their age and script, into recensions and versions, script-wise classiication supplying to the editor clues for diferent versions into which the text existed.” Vaidya nonetheless notes that “one manu- script in Nandinā gari script was collated, but it was found to give an identical text with that in Telugu script and was, therefore, not taken into account” and likewise that the manuscript in neo-Ś ā radā script from the Scindia Institute, Ujjain contained “an inlated text almost iden- tical with the Vulgate.” P. L. Vaidya, “Introduction,” in The Harivaṃśa for the First Time Critically Edited, vol. 1 (Pune: Bhandarkar Oriental Research Institute, 1969), xxxv–xxxvi, x and xvii. Thus, even when editors were mistaken in theory about the principle of classiication, they carried it out correctly in practice. 164 This occurred in the critical edition of the Sabhā parvan. The editor, Franklin Edgerton, identiied three manuscripts in Telugu script (T2 = Melkote, Yadugiri Yatiraj Math, no. 155. Telugu, palm leaves; T3  =  Madras, Govt. MSS. Library, no.  1922, Telugu, palm leaves; and T4 = Madras, Govt. MSS. Library, no. 1923, Telugu, palm leaves) as belonging to his “ ‘E’ (B and D) version.” Regarding their classiication, he wrote: “These MSS. T2–4 were completely collated, and their collations have been considered by the editor with great care. After much thought, it was decided not to include them in the Critical Apparatus. It is evident that to group them with T1, the only other Telugu MS. we have, would be highly misleading. T1 is on the whole a normal Southern MS., agreeing with G and M, especially closely with 289 CONFUSIONS REGARDING CLASSIFICATION 289 G. It is true that it has some few Northern readings and interpolations; but this is true of some G MSS. also, and I am not sure that it is commoner in T1 than in them. With T2–4, it is just the other way around. It is perfectly obvious that they are based on a Northern text; specif- ically, on an Eastern one, agreeing with B and D. They also contain Southern readings and (especially T2.3) interpolations; but there can be no doubt that it is these which are intrusions, whereas in T1 it is the occasional Northern features which are secondary intrusions in its basi- cally Southern text.” Edgerton, “Introduction,” xxi (Edgerton’s italics). The only person in any doubt about the true basis of classiication is Grünendahl. 165 Edgerton found this was the case for the Sabhā parvan. In contrast to Sukthankar, who thought the Ñ, V and B manuscripts formed one group and the D another, Edgerton found that “In Book 2 […] neither V nor B nor D MSS. seem to constitute a unit, genetically speaking.” “So far as I can see, they could all be lumped together as one version, which I should call E1, within which only two minor sub-varieties of little importance (Dn and D1.2) are discern- able.” Edgerton, “Introduction,” xliii. The E1 recension (= VBD) thereafter replaced in most cases the individual versions for Edgerton. For other examples of manuscripts separated out from their presumed agnates and listed as independent groups, see Table 5. 166 If further conirmation were needed that a manuscript’s script is irrelevant to its place in the stemma, we have an interesting example from the fourth volume of the critical edition. The volume’s editor, Raghu Vira, notes that collations of two manuscripts (M1 and M2— Trivandrum, Palace Library, nos. 377 and 378) were made from Devanā garī transcripts of the originals the Darbar had provided. In other words, the editor never saw the manuscripts in Malayā lam script, and yet he assigned them their correct place in the tradition! Vira, “Introduction,” ix and see also ibid., ii. 167 Grünendahl’s error is understandable. Because he has not worked with the Mahā bhā rata manuscripts (and, in fact, never created a stemma), he cannot grasp their relation. He advances an a priori theory of classiication that no one except he himself subscribes to. This theory’s origins likely lie in a supericial reading of Sukthankar’s “Prolegomena” for, in his 1982 dissertation, he labors to imitate its diction and style albeit without grasping its logic. When he criticizes Sukthankar in his 1993 article for classifying the manuscripts by script, he does not see that his criticism applies not to Sukthankar but only to his misunderstanding of Sukthankar’s principles. For in his 1982 edition of the Viṣṇudharma, Grünendahl indeed classiies the manuscripts by script. He identiies two traditions—a Nepā lī and a Devanā garī tradition—of which the former breaks apart into two groups. His assignment of manuscripts to the three groups evidently follows their script, for he notes that the manuscripts are so con- lated that it is impossible to distinguish them otherwise. 168 This table does not list K manuscripts in Devanā garī script, though numerous examples exist. However, it is worth noting that four of the ive manuscripts of the Bhagavadgītā in Ś ā radā script, Ś 2–Ś 6, contained slightly diferent texts—the versions underlying the commentaries of Abhinavagupta (Ś 3), Rā jā naka Rā makaṇṭha (Ś 4 and Ś 5) and Ā nandavardhana (Ś 6). The editor identiied them as such. See Belvalkar, “Introduction [to the Bhīṣmaparvan],” xiii. 169 We thank Saraju Rath for clarifying these charts. For the source, see Saraju Rath, “The Oriya Script: Origin, Development and Sources,” Heritage India 5, no. 2 (2012): 55 and “Varieties of Grantha Script: The Date and Place of Origins of Manuscripts,” in Aspects of Manuscript Culture in South India, ed. Saraju Rath (Leiden: Brill, 2012), 189. 170 Scribes could have copied entire exemplars from an unrelated script (for example, a Ś ā radā manuscript into Grantha) or simply selections (for example, using a Proto-Grantha exemplar for one part of the text and a Ś ā radā exemplar thereafter). But such changes of exemplar would be relatively rare and easily detectable. Comparison of manuscripts in diferent scripts is more problematic. This is where contamination might seriously mislead us. 171 Sukthankar, “Prolegomena,” v. 290 290 PHILOLOGY AND CRITICISM 172 As West also notes. See West, Textual Criticism and Editorial Technique, 15:  “If the older manuscripts can be itted into a stemma, the promiscuity of the younger ones should be easy to establish and may be unimportant.” 173 Grünendahl’s confusion is due to the fact that he has no experience with Mahā bhā rata manuscripts in spite of claiming to have undertaken a study of the Nepā lī manuscript Ñ4. Like Ruben, on whose work he bases his case, he commits avoidable errors, because he fails to appreciate either the tradition’s complexity or the constants with which the editor works. Sukthankar’s comments on Ruben as an editor are apposite: “Ruben’s long disquisition on the diiculties of the Mahā bhā rata textual criticism boils down to this: only that portion of the text which is documented by both recensions, in identical terms, is tolerably certain; the rest is doubtful in varying degrees, there being no criterion whatsoever which can enable us to dis- criminate with complete conidence between the variants; even the concordance between K and S (against B D) is not conclusive evidence of the originality of the common reading. This is a more or less obvious standpoint, though partly erroneous and distinctly timid and conser- vative. It was hardly necessary for the ‘Referent’ to write so much in order to say so little. His attitude of dogmatic doubt has caused him to make unwittingly some wild exaggerations and unwarranted generalizations as I have shown above. His perfunctory study of the manuscript evidence has led him occasionally to make gross blunders in the estimation of the relationship of the diferent manuscript groups and versions, some of which have been pointed out above. Many of the diiculties of the Mahā bhā rata textual criticism of which he speaks in his paper are due to his failure to understand the modus operandi of the redactors and copyists of the epic and his lack of insight into the character of the diferent manuscripts—an insight which can be acquired only after a long familiarity with the manuscripts themselves and a close and patient study of their tendencies and idiosyncrasies.” Sukthankar, “Dr. Ruben on the Critical Edition of the Mahā bhā rata,” 282–83. 174 Grünendahl, “Zur Klassiizierung von Mahā bhā rata Handschriften,” 103. 175 Ibid. And see also ibid., 129: “Lüders’s hypothesis cited above that the versions were ‘grad- ually isolated to a speciic region’ is, in my opinion, even merely theoretically not quite believable. Further, one can oppose against it that it was hardly conirmed in praxis. On one hand, major diferences manifest between the manuscripts of a ‘version’; on the other, the correspondences between manuscripts of diferent ‘versions’ is too obvious that the classiication according to their script can really convince us.” When Grünendahl speaks of “correspondences between manuscripts of diferent ‘versions,’ ” he does not mean their shared errors (which could be attributed to contamination), but that they share the correct reading, and from this agreement he concludes that the “classiication according to their script can[not] really convince us.” He thereby overlooks that there is no reason the correct reading cannot occur in two branches of the tradition! For instance, when ν and S agree, this does not mean that their classiication into ν and S was false, but only that their reading is the reading of the archetype! This confusion, which is his main reason for rejecting the critical edition, occupies us in the next section (“The Argument from Independent Recensions”). 176 Ibid., 107. 177 Other than the obvious one that wherever the apparatus formerly read “K” it would now read “Z.” 178 The manuscripts collated for this section of the text were Ś 1, K0–4, Ñ1–3, V1, B1–5, Da, Dn, D1–7, T1.2, G1–6 and M1–5. K therefore refers to only K0–4 and not K0–6 as elsewhere. 179 If mahāghorau were the corruption, the editor would have to explain how it was transmitted horizontally to just these three manuscripts and M1.5. It is thus more likely the reading of the archetype was transmitted via N and S to Ś , K and M, but was lost in the central recen- sion and σ with the exception of D2.5 and G3. The alternative that Ś , K or M1.5 gained the reading from each other or that they arrived at the same reading independently is less plausible. 291 CONFUSIONS REGARDING CLASSIFICATION 291 180 The Ś ā radā script was unfamiliar outside the region. It also ceased to evolve after the eighth century CE. Given their geographic separation, Sukthankar thus thought Ś 1—M contam- ination was less likely, though not impossible. See Sukthankar, “Prolegomena,” lxxiv: “This [Malayā lam] version has several striking agreements with Ś 1, a fact all the more impres- sive, because M, a Southern version, hails from the province at the opposite end of India from the province of Ś 1, a Northern version.” Sukthankar found M the least contaminated of the southern manuscript groups:  “This is the version of Malabar, the Southernmost extremity of India. It is, in my opinion, the best Southern version. It is not only largely free from the interpolations of σ (= TG), but appears to be also less inluenced by N than σ, wherein lies its importance for us.” Ibid., lxxiii (Sukthankar’s italics). 181 See ibid., lv: “In my opinion, therefore, this fact of the concord between ν and S in small details, coupled with the almost entire lack of agreement as regards the additions peculiar to ν or S is the strongest argument imaginable for the independence of these two versions, and consequently for the primitive character of their concordant readings. It is needless to point out that this is a factor of supreme importance for the reconstruction of the original.” Although he extends the argument to the southern recension, the Telugu and Grantha are contaminated with the northern recension so that, in practice, the agreement of Ś and K with M had greater probative value for Sukthankar (see preceding note). 182 This example concisely illustrates that Grünendahl’s so-called Schriftartprämisse can play no role in the reconstruction. The readings of both manuscripts have been transcribed into a third, unrelated script—Roman, though it could also be Devanā garī—and yet the essential diferences— those relevant for a reconstruction—are retained. In fact, we could convert all of our witnesses to a standard script and still undertake the reconstruction. We are interested in changes at the level of semantic and syntactical units, and it is by comparing these (for example, is the correct reading –ghorau, –vı̄ryau or  –kāyau?) that we seek to reconstruct the archetype. Grünendahl errs because he does not grasp that the script is merely a conventional system of signs—the bearer of meaning, not the meaning itself. 183 Reeve, “Stemmatic Method,” 67. 184 This only holds if a scribe combines sources that left behind other witnesses. It would not hold if his source left behind no other witnesses, and we also did not possess other exemplars of its version. If we possessed the copy in which the conlation occurred, there could be signs such as a change of hand or ink to indicate that the text was copied at diferent times. But if this copy were copied again, the apograph would contain no indication that its source was conlated. We could not know that what appeared to be a continuous text was actually an amalgam of two sources. Reeve notes, “it always needs to be proved that a manuscript derives from the source throughout,” but in practice we must assume this kind of conlation was widespread, especially as the practice appears to have been to copy the Mahā bhā rata a parvan or several parvans at a time. Michael D.  Reeve, “Eliminatio codicum descriptorum:  A Methodological Problem,” in Editing Greek and Latin Texts, ed. J. N. Grant (New York: AMS Press, 1989), 32. 185 Sukthankar, “Prolegomena,” lxxii. 186 K4 ofers valuable insight into the formation of these exemplars. Sukthankar notes: “Much light is thrown on the origin of these misch-codices by the MS. K4, a manuscript belonging to the Bombay Government Collection deposited at the Institute. In this manuscript we ind long extracts from other cognate versions (such as γ) as also from the Southern recension, written out on separate folios and inserted at appropriate places in the body of the manu- script, with the words atra śodhapatramekaṃ written on the margin of the original folio, near the place where the passage is to be interpolated. Should this manuscript happen to be copied again and should the copyist insert the passage at the place indicated by the previous scribe, the interpolation would become an integral part of the new text which is externally absolutely indistinguishable from the rest of the text.” Ibid., xcv (Sukthankar’s italics). 29 292 PHILOLOGY AND CRITICISM 187 This was the case for D8–14, which were discontinued after the second adhyāya. 188 Mahā bhā rata scribes appear to have compared manuscripts primarily to ensure no narrative or passage found in other exemplars was lost rather than to record variant readings. Sukthankar notes, “[A]n ancient Indian scribe, redactor or even commentator, not to speak of the common reciter (pā ṭhaka)—if I read aright Indian literary history—was not perturbed in the least by a little diference in wording or in sequence, especially if the variant did not give an appreciably better, or appreciably worse sense. The enormous and complicated crit- ical apparatus assembled here, moreover, can leave us in no doubt as to the attitude of the custodians of the epic tradition towards paltry verbal details: it was that of total indiference. Addition or omission of passages is, I may add, a variation of an entirely diferent order. If a reciter or commentator came across, in another manuscript, an additional passage, there was every chance of his copying it down somewhere, either in the margin of his own copy, or on a supplementary folio; for there would be, in his mind, always present the possibility that the passage in question was some part of the original that his own manuscript had unac- countably lost. How else, forsooth, could the passage get into the other manuscript?” Ibid., lv (Sukthankar’s italics) and see also ibid.,  lxxxi:  “Here one notices above all the anxiety that nothing that was by any chance found in the Mahā bhā rata manuscript should be lost. Everything was carefully preserved, assembled in a picturesque disarray.” 189 Even if two manuscripts contain the identical interpolation, we can infer iliation only if one retains the other’s errors and introduces several new signiicant errors. 190 Ibid., li (Sukthankar’s italics). 191 It is not always obvious whether a shared reading is owed to the archetype or to horizontal transmission, although editors have evolved several criteria (geographic separation, likeli- hood of contamination and the nature of change) to determine this. Edgerton’s introduc- tion features an extended discussion: “Common secondary insertions found in E and S, but not in W, are more numerous. The great majority are not found in Ñ, however, so that they are really common only to E1 and S. Since I think it can hardly be doubted that E is a real unit […] it seems that we must assume that these insertions have spread by contamination (rather than inheritance from a common archetype).” Edgerton, “Introduction,” xlv–xlvi. “In my opinion, the foregoing lists are not numerous enough to make at all likely secondary genetic connections between E, or E1 and S.  Nor are secondary readings of detail at all common. Agreements between these versions, then, must be either inheritances from the original, or due to contamination (or perhaps, in some minor points of phraseology, inde- pendent changes due to accident).” Ibid., xlvi. “Insertions common to W and E, but lacking in S, are less numerous than those common to S and E1, though much more numerous than those common to S and the whole of E (including Ñ). Some in the following list, it will be noted, are lacking in some Northern, particularly K MSS., so that it is doubtful whether they could properly be counted as tending to show closer genetic relations between W (Ś K) and E (ÑVBD) as against S; in such cases it would seem more likely that the W MSS. which con- tain the passage have been contaminated from an E source. Similarly, some insertions here listed are not found in Ñ1, suggesting that they did not belong to the original of E (or ‘N’), and could have spread by contamination to the other E MSS., or perhaps were inserted in E1.” Ibid., xlvi–xlvii. He arrives at a similar conclusion as Sukthankar: “Study of the variant readings in detail leads to no other conclusion. When the generality of W and E (that is, what in our Critical Apparatus is called N) agrees against S, there is no reason to doubt that in the vast majority of cases W and E (or N) preserve the original, and S is secondary. The same is true when W and S agree against E, and when E and S agree against W. General agreement between any two of the three main streams of the tradition creates a presumption of originality. I say, a presumption: not absolute certainty. But rarely can I discern any reason to regard as original a reading of any of the three groups—W, E, or S—when most or all of both the other groups agree against it.” Edgerton, “Introduction,” xlvii (Edgerton’s italics). 293 CONFUSIONS REGARDING CLASSIFICATION 293 192 For reasons of clarity and legibility—the line from ν or γ crosses over many lines—this stemma does not show contamination into G even though Sukthankar notes that “All Grantha manuscripts are probably contaminated (directly or indirectly) from Northern sources in diferent degrees.” Sukthankar, “Prolegomena,” lxxii (Sukthankar’s italics). It is unclear whether he means contamination from Ñ, V, B and D, which are northern recension manuscripts, or only Ś and K, as γ is referred to as the central subrecension. 193 The central subrecension γ occupies a large part of Sukthankar’s relections in the “Prolegomena.” He notes that “The view that ν and γ may stand in genetic relation to each other does not receive much support from the facts of the case. Neither ν nor γ can be derived from the other. Each possesses original features that the other lacks, as is evidenced by their alternate agreement with S, even in the matter of petty verbal details. All these coincidences need not, of course, be original. Some could be indeed secondary changes, made independently in the same direction; others again may possibly be explained as the result of contamination. There will remain still an obstinate residue of agreements between ν and S, or between γ and S, that must be set down as the expression of the ultimate connection of the respective con- cordant versions through the lost original source. Contamination between ν and γ, owing to the contiguity of the areas in which the respective versions were current, was inevitable, and must, in any case, be assumed to have existed; on the other hand, contamination between γ and S cannot be altogether denied.” Ibid., lvii (all italics Sukthankar’s). 194 “In my opinion, therefore, this fact of the concord between ν and S in small details, coupled with the almost entire lack of agreement as regards the additions peculiar to ν or S is the strongest argument imaginable for the independence of these two versions, and consequently for the primitive character of their concordant readings. It is needless to point out that this is a factor of supreme importance for the reconstruction of the original.” Ibid., lv (Sukthankar’s italics). 195 As editors recognize, it is impossible to draw up stemmata for very short texts. The only exception is if these texts are parts of larger collections. 196 A manuscript of another work such as the Rā mā yaṇa would also be free of this interpola- tion, but its absence does not justify the conclusion that the two manuscripts have a common source. 197 The argument is like claiming that two species that lack the same marker must be related. It is clearly erroneous. Sparrows and robins lack webbed feet, but they are not members of the same family. 198 Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 116. 199 Ibid. 200 This irst argument is manifestly false. Sukthankar’s list of 25 readings illustrates the K manuscripts’ proximity to the Ś ā radā codex. On the basis of this list, he posits that the Ś ā radā and K manuscripts descended from a common hyparchetype he calls ν. The Ś -K (ν) recension’s purity neither afects nor is afected by the other groups’ impurity. D2.5 are occasionally contaminated with ν. But as ν is the source of contamination, the circumstance that they share some of its readings is an argument neither for ν nor against ν. But this is pre- cisely what Grünendahl asserts. According to him, “in light of the repeated and conspicuous correspondence of D2 and especially D5—against all manuscripts!—with the hyparchetype ν,” Sukthankar’s decision not to assign these two D manuscripts to the K group is “not quite comprehensible.” Ibid., 117 (Grünendahl’s italics). The fact that D2.5 are contaminated manuscripts where the source of contamination is ν or one of its descendants is not an argu- ment for including them under either K or ν. At most, we would draw a dotted line from ν to D2.5 indicating contamination. In any case, since this does not afect the reading of the archetype, their inclusion under K makes no diference. At most, it would afect the critical apparatus. Where we noted that K has a uniform reading, we would now note that K, except D2.5, has this reading. Vice versa, where we noted that D has a uniform reading against 294 294 PHILOLOGY AND CRITICISM the other groups, we would now note that D has a uniform reading along with the two “K manuscripts” D2 and D5. 201 Ibid., 117. 202 Ibid. 203 Ibid. 204 Ibid. 205 Ibid. 206 Ibid. 207 Ibid. (all italics Grünendahl’s). 208 Grünendahl’s peculiar understanding of “contamination” as an inluence on another man- uscript even if that this manuscript also does not feature the interpolations that the irst does not feature also explains his curious deinition of “purity.” According to him, the fact that D2.5 also do not feature passages that K0.1 do not feature is evidence not only of K0.1’s inlu- ence on D2.5 but also of K0.1’s impure nature. His reasoning is as follows: “A greater signif- icance accrues to the indicated commonalities between D5 and K0–3; they stand in manifest contradiction to Sukthankar’s argument, which aims at displacing D2.5 into proximity to the conlated mss. K3–6 and thus to preserve the ‘pure form’ of the K version, represented by K0.1, from every taint of contamination.” Ibid., 118 (Grünendahl’s italics). The phrase “indicated commonalities” refers to the passages missing in D2.5 and K0–3: 1214* and 221*, and (with respect to D5 only) 1346*, 1444*, 1517*, 424*, 1202*, 1344*, 1345*, 1363*, 1450*, 1458*, 1494*, 1546* and 1821*. In his view, since D2.5 also do not feature these interpolations, they must have been under K0.1’s “inluence” for how else, he reasons, could they have chosen not to feature them if not for the inluence of the other manuscripts that also do not feature them? Consequently, K0.1 cannot be “pure” as Sukthankar thought, for they have clearly been fraternizing with other manuscripts. This is a new deinition of “purity,” which stipulates that to be “pure” a manuscript must be free of contamination and, furthermore, that no manuscript may resemble it in this respect. If even a single manuscript also does not contain the interpolations it does not contain, it must be considered impure. The deinition is inane, because a chaste person is no less chaste for the fact that several chaste individuals, others who resemble him or her in this respect, exist. Grünendahl’s deinition could only be satisied if except for this exemplar every other manuscript contained every interpolation. As this is impossible, it efectively states that to the extent that a manuscript is pure it is also simulta- neously impure! Grünendahl has repealed the principle of non-contradiction. 209 Ibid., 117. 210 Ibid. (all italics Grünendahl’s). Insertions in square brackets are Grünendahl’s. 211 A simple experiment suices to demonstrate this: if manuscripts A and B contain an addi- tional passage x and we have no further information about them, their iliation cannot be determined. A may have gotten the passage from B or B from A or both from a third source. It is correspondingly more diicult to establish iliation if neither contains the additional passage. If we could posit a relationship between two manuscripts based on the fact that one does not contain a certain passage, we could establish indiscriminately many relationships between indiscriminately many manuscripts. The Apollodorus manuscript Paris, BNF, Graecus 2722 would be related to the Galen manuscript Paris, BNF Graecus 2267 (neither contain the text of Plutarch’s Lives); the Diogenes manuscript Paris Graecus 1759 would be related to the Isocrates manuscript Vatican Graecus 936 (neither contains the text of Thucydides’s History of the Peloponnesian War) and so on. 212 Ibid., 117. 213 Ibid., 118. These passages, which Grünendahl calls “present textual additions,” are interpolations in D5 and K4. They are therefore by deinition not features of the archetype. Consequently, their occurrence in D5 and K4 does not prove them related. It only shows con- tamination, without permitting us to infer that D5 was the source and K4 the contaminated 295 CONFUSIONS REGARDING CLASSIFICATION 295 manuscript, or vice versa. We cannot prove relationship based on accidental characteristics supervening on the manuscripts. Grünendahl’s argument would only hold if he could dem- onstrate an intrinsic feature of these manuscripts, which caused them to attract these speciic interpolations, that no other manuscript possessed this feature and that this feature does not occur accidentally, but is inherited. 214 Grünendahl notes of the “present textual additions” common to D5 and K4: “it is therefore diicult to determine whether these commonalities are to be traced back to mutual inluence of the two lines of tradition ending in D5 and K4 or to the contamination of these two with a third unknown [exemplar].” Ibid., 118. He rejects this possibility for the “not present textual additions” on the grounds that in the “ case […] of not present textual additions, these simi- larities […] cannot be explained away […] with the fact that they entered into the ‘composite manuscripts’ D5 (and D2) due to contamination (conlation). This may serve as an explanation for present textual additions but hardly for not present [textual additions].” Ibid., 120. But he nevertheless thinks that if D5 and D2 resemble the K manuscripts in not featuring certain interpolations, this must be due to the K group’s “inluence” on them. Why he does not con- sider the possibility that their common ancestor N and, above all, the archetype could be responsible for this omission is a mystery. Apparently, we can conidently exclude external “inluence” in the latter cases. 215 The two manuscripts are unrelated in the literal sense that we cannot posit any relation at all rather than in the sense that we can conclude that one of them is not descended from or not contaminated with the other. 216 If the passage was not an insertion but part of the text, we might conclude that both manuscripts were descended from a common source in which the text was missing. Here also we could not establish the speciic nature of the relationship without other evidence, for instance, that one contained all the errors of the other, plus at least one more or that, in addition to containing the same lacunae, each one made at least one new error. But these inferences are based on errors in transcription rather than the presence or absence of passages. 217 Ibid., 117. 218 Grünendahl not only considers the absence of interpolation more signiicant the greater the number of interpolations two manuscripts lack in common; he also considers this absence more signiicant the greater the number of manuscripts that feature these interpolations. He writes: “We can show something similar of Sukthankar’s ‘core group’ K0.1 expanded to include K2 [in other words, for the manuscripts K0–2],” and he adds: “29 textual passages are absent in this group of three manuscripts which are present in the majority of the northern recension manuscripts.” Ibid., 118. He evidently views contamination as a kind of epidemic such that the more individuals succumb to it, the more closely related those resistant to it must be. But this is not how contamination works. Since iliation is not established on the basis of contamination, it is irrelevant whether all or only some of the manuscripts outside the “group”—using the term heuristically, since we have not established the uninterpolated manuscripts form a family in the genealogical sense—contain the interpolation. Grünendahl thereafter introduces two comparative degrees of the principle. He reasons that if all and not just some of the interpolations whose absence is characteristic of a group are absent from a manuscript and if the other members of its family reinforce this absence, then we have strong grounds for assigning the manuscript to this group. “Except in K0.1.2 they [the 29 interpolations] are consistently absent only in Ñ4 and Ś 1 (insofar as it is preserved), otherwise they are absent ifteen times in Ñ3, ten times in Ñ1 (of this seven times both in Ñ1 and Ñ3, two times including Ñ2), as well as eight times in D5!” Ibid., 118–19. His logic appears to be that the absence of all the passages in question in Ñ4 makes it just like Ś1 and K0–2, and, since they are also frequently absent in Ñ1 and Ñ3 and seven times in all three manuscripts Ñ1, Ñ3 and Ñ4, this makes the Ñ group just like the K group. The argument is nonsense: if the absence 296 296 PHILOLOGY AND CRITICISM of interpolation cannot establish iliation, the absence of many interpolations simultaneously cannot intensify the iliation, irrespective of whether all or only some of the interpolations are absent from the manuscript, even if some are absent from all the manuscripts of its group. 219 Ibid., 118. The phrase “indicated commonalities” refers to the fact of not possessing certain characteristics. 220 Ibid., 117 (all italics Grünendahl’s). Insertions in square brackets are Grünendahl’s. 221 Ibid. 222 Grünendahl has not understood the basic diference between shared readings owed to innovations in a recent ancestor, which establish iliation between two witnesses, and shared readings that are evidence of an original inheritance and only prove iliation through the archetype. The circumstance that Ś , K and S lack these passages is evidence of an original inheritance through the archetype, but not of a closer relation. Their absence from Ñ4 only conirms Sukthankar’s intuition that ν often preserved the correct reading, when γ and S had an inlated text. 223 All cases of the latter are cases of the former, since the latter case is merely hypothetical. When K and D2.5 agree in not featuring an interpolation, this does not mean that D2.5 were inluenced by K in this respect. It only means that the passage is missing from the archetype and its text has been handed down, at least in this respect, correctly to both K and D2.5. Their agreement is coincidental in the sense that it does not permit us to infer a closer rela- tionship between them. 224 In contrast to Sukthankar, who thought that, “with the possible exception of the Śāradā (Kasmı̄rı̄) version, which appears to have been protected by its largely unintelligible script and by the diiculties of access to the province, all versions are indiscriminately conlated” (Sukthankar, “Prolegomena,” lxxxii; italics in original), Edgerton felt “no version, no single manuscripts even, can be guaranteed to be absolutely ‘independent’ of any other version or manuscript,” including the Ś ā radā (Kaśmīrī) version. Edgerton, “Introduction,” xxxiii and see ibid., xxxv, n. 3: “I doubt if he [Sukthankar] would now think this qualiication necessary. I, at least, think I can deinitely prove that Ś too has some clear contaminations and ‘conlations.’ ” If Edgerton nevertheless felt conident about recovering an archetype, this was because he distinguished between horizontal transmission and the vertical constants of the tradition. Thus, noting that “what seem to be certainly ‘secondary’ agreements can be found between practically any two manuscripts. The origin of many such can easily be understood from the frequency with which alternative variants, in many manuscripts, are written above or below the lines or in the margins (the next copy would produce the “secondary” agreement!),” he asked, “May not this be due to errors in our classiication of MSS. and versions? If all MSS. of what we have called two ‘independent’ versions agree on a secondary reading, particularly an insertion, does that not raise the presumption that these versions go back to a common secondary archetype?” Ibid., xxxiii and xxxv, n. 3 (italics in original). He ofers a rule of thumb as a guide: “The answer is that such instances can be found between any two extant versions; so that we should arrive at the absurdity that each recension is necessarily related to every other. It is simply a question of more or less. When apparently secondary agreements are very numerous, we assume secondary relationship. When they are relatively sporadic, we assume contamination. Precise igures for variant readings cannot be given—hardly even for insertions.” Ibid., xxxv, n. 3. At any rate, he was conident that “There is certainly no reason to assume secondary relations between W (Ś K) and S.” Ibid., xlv. 225 This a legitimate concern, but it can only be apparent once the editor succeeds in itting the manuscripts into a stemma. As Irigoin notes: “He [the editor] should ensure, by all the means provided by his science and his ingenuity, that he discovers therein the constants of the ‘vertical’ tradition. It is at this price, and at this price only, that he will be able to determine the reality, and eventually the amplitude, of a ‘horizontal’ transmission.” 297 CONFUSIONS REGARDING CLASSIFICATION 297 Jean Irigoin, “Quelques rélexions sur le concept d’archétype,” Revue d’histoire des textes 7 (1977): 242–43. 226 This is not iliation as a textual critic understands it. It does not demonstrate descent. But it would be the sole criterion in the absence of unique errors introduced while copying manuscripts. 227 Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 119 (Grünendahl’s italics). 228 Ibid. 229 Ibid. 230 The pleonastic “shared presence of interpolations” is only necessary because Grünendahl thinks the absence of interpolations is also a criterion for iliation and hence contrasts “pre- sent” interpolations with “not present” ones. 231 Ibid., 119 (italics in original). 232 We do not mean to assert that a scribe would either irst copy the interpolations or irst choose which interpolations he wished to copy before copying his source. In this sense, Grünendahl’s assertion is true, though trivially so: the inherited text must precede the interpolations added to it. What we mean is that this conclusion only follows on the basis of his model of textual criticism: it does not hold if one grasps that this is no way of determining iliation or, indeed, of drawing up stemmata. 233 Sukthankar, “Prolegomena,” xxx. 234 Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 107. 235 Apparently, this would hold for sections for which a Ś ā radā exemplar was unavailable, even though in the irst part of his argument Grünendahl argued that “constituting” a K version without a Ś ā radā exemplar is inadmissible. The contradiction evidently eludes him. 236 Sukthankar’s examples preview only a selection of his observations while editing the man- uscript. Like other armchair critics of the Mahā bhā rata critical edition, Grünendahl reads the “Prolegomena” and cross-checks the examples given there against the apparatus, but examines neither a signiicant extent of the constituted text nor the manuscripts to verify that Sukthankar’s conclusions are borne out. He thinks that to refute Sukthankar it suices to contradict the examples in the “Prolegomena.” But the real test of his work must be whether it provides a more plausible account of the manuscript evidence, which challenge he has, of course, not confronted as yet. 237 Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 114. The reference is to Sukthankar’s “Prolegomena,” xlviii. 238 Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 114–15. 239 See ibid., 112: “Since the determination of the K manuscripts [as K manuscripts] accordingly theoretically presupposes a ‘Ś ā radā version,’ in the parvans for which no Ś ā radā manuscripts was available a K version could in fact not be deined.” 240 Sukthankar, “Prolegomena,” l. 241 Ibid., li. 242 For a discussion of the irst list, see the preceding section. The second list is discussed earlier in this section under Grünendahl’s reasons for positing D2.5 contamination with K. 243 Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 115. 244 Apart from the problem that at this stage we cannot know which these readings are (unless we irst classify the manuscript material we cannot reconstruct the reading of the archetype), the argument is nonsense. If we were to consider only readings that entered the constituted text, we would be restricted to the readings common to the northern and southern recensions. The constituted text features the reading attested in both recensions; in cases of discrepancy, it defers to the northern reading, but here we seek to identify the southern manuscripts as a distinct group! 298 298 PHILOLOGY AND CRITICISM 245 Except when those readings were not in K. In this case, we would assume contamination between neighboring groups. If the contamination was not between neighboring groups, we would rethink our view of the tradition and possibly restore our original stemma—that is, Sukthankar’s stemma of the Ā diparvan. The example illustrates the diiculty of trying to do textual criticism in the abstract, as Grünendahl attempts, rather than beginning with the manuscript evidence. 246 Each line with the small bar at its base indicates that a new text is being copied. These lines replace the dotted lines that conventionally denote contamination, because the contamination takes the form of expansion within the manuscript itself: there is no need to posit an external source. 247 The lines representing the diferent traditions are thus both linked to the archetype and not linked. They are not linked insofar as a new text is propagated henceforward in this branch of the tradition; they are linked insofar as this text consists of an unchanged original (in Grünendahl’s model there are no corruptions) and its expansions, so that merely discarding the latter permits a return to the archetype from any branch of the stemma (and not just K). 248 What is the evidence that K alone contains the correct reading when it was lost in all others? Sukthankar thought that the editor should examine “as many manuscripts—and above all as many classes of manuscripts—as possible.” Sukthankar, “Prolegomena,” lxxxvi (Sukthankar’s italics). He accepted the reading “documented by the largest number of (what prima facie appear to be) more or less independent versions” as the reading of the archetype. If no clear pattern emerged— one pair of versions featuring one reading and another pair another—he gave credence to the reading of Ś 1 and K on account of their “much greater correctness and reliability.” Ibid., xci (all italics Sukthankar’s). Only in this last case can Grünendahl’s demand be satisied, but since Ś 1 and K’s reading is included in the constituted text on their authority alone, asking for conirmation from the constituted text is pointless. 249 The argument that the editor should not have relied on readings not attested in the consti- tuted text while identifying the K group is misguided, but we can appreciate why Grünendahl raises it. From his perspective, it seems that the editor based his “constitution” of the group on readings he ultimately discarded, and Grünendahl thinks this cannot be right. As he notes, “it can hardly have been reconcilable with Sukthankar’s editorial principles to grant such poorly attested readings such signiicance.” Grünendahl, “Zur Klassiizierung von Mahā bhā rata- Handschriften,” 115. From his perspective, his insistence that the K group should be identi- ied only on the basis of readings included in the constituted text only makes the process more rigorous. He thereby overlooks that the reason the constituted text is occasionally based on the evidence of the K group alone is due to a unique limitation: if the editor could not iden- tify the reading of the archetype and had two readings of “equal intrinsic merit,” he adopted the reading of Ś and K owing to their “much greater correctness and reliability.” Sukthankar, “Prolegomena,” xci. This does not mean that these instances are better attested. They are not, for their inclusion in the constituted text is based solely on the evidence of these two manuscript groups. 250 For Grünendahl, the relationship of one manuscript to another can only be of original and facsimile rather than source and apograph. As he interprets manuscript transmission, witnesses are not related by descent to their sources. Rather, manuscript copying either produces a facsimile edition identical to its source in every respect (providing no inlation occurred) or a new, inlated text. The copy is therefore related to its source not as its descen- dant but as its peer. 251 Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 116. 252 Ibid. 253 Ibid., 107; and see ibid., 109: “Since Sukthankar argues exclusively in terms of quantity, he can really only elevate Ś 1 to the norm in a quantitative respect. This would correspond to his conviction that one—manifestly in contrast to extent!—may not constitute the text on the basis of individual manuscripts, indeed not even individual versions or recensions” (all italics in original). 29 CONFUSIONS REGARDING CLASSIFICATION 299 254 Ibid., 116. 255 Ibid. 256 For Grünendahl’s comments about Sukthankar’s (and other scholars’) failure to adequately recognize the work of German scholars, see ibid., 102:  “The preliminary work from the European side made available by the association of the [scientiic] academies did not receive any meaningful recognition just as the 230 manuscripts preserved outside India that were identiied under Lüders were barely made use of ”; ibid., 102, n.  5:  “V. S.  Sukthankar’s Prolegomena to his edition of the Ā diparvan (p. If.) only provides a very sketchy impression of the work already done and the reasons for its interruption, and he only speaks very generally of an ‘occasional use’ of the collations provided by the association of the academies (p. vi)”; ibid., 102, n. 6: “This preliminary work is not mentioned in J. Dunham: ‘Manuscripts used in the Critical Edition of the Mahā bhā rata: A Survey and Discussion,’ Essays on the Mahābhārata, Leiden 1991, pp. 1–18 (Brill’s Indological Library, 1), contrary to the title less a ‘discussion’ than a restatement of what is already suiciently well known from the Prolegomena and the Introductions by the individual editors of the BORI-edition.”; and ibid., 103: “As W. Ruben rightly supposed in his essay on fascicules 1–3 of the Ā diparvan, Sukthankar’s grouping of the manuscripts ‘according to their script in “versions” ’ (in the Foreword to the irst fasci- cule, p.  iiif.) was inspired by Lüders, even though his preliminary work is not mentioned here.” He would doubtless prefer a situation like the one, he claims, French critic Hippolyte Taine described in 1864, when he explained, “without any sense of rivalry, but with a clear understanding of what philology is about,” “why contemporary studies of Sanskrit, Persian, Greek, Latin, and Bible exegesis—in short:  philological disciplines based on historical principles—had ‘their heart and centre in Germany.’ ” Reinhold Grünendahl, “History in the Making: On Sheldon Pollock’s ‘NS Indology’ and Vishwa Adluri’s ‘Pride and Prejudice,’” International Journal of Hindu Studies 16, no. 2 (2012): 225–26. To judge by his own work, how- ever, the skills of the present generation of German scholars are not as impressive. Neither is the idea that the Bhandarkar editors were unwilling to recognize German contributions accu- rate. Vaidya recollects his irst visit to Europe: “Dr. Sukthankar, the irst General Editor of this scheme of the critical edition of the Mahā bhā rata, and myself visited Europe in 1931 to attend the International Congress of Orientalists held in Leyden. During the period of three months when we toured together, we held talks on his plan of entrusting diferent parvans to diferent scholars. He wanted for this scheme the co-operation of Indian as well as European and American scholars. It is a well-known fact that the Sabhā parvan was irst assigned to the late Dr. M. Winternitz who was chosen for that very parvan by the International Association of Academies. Both Dr.  Sukthankar and myself called on him at his residence in Prague to persuade him to come to India for some months for editing that parvan, but his age and health would not allow him to undertake the voyage. Similarly, it was Dr. Sukthankar’s idea to assign the Karṇaparvan to Dr.  Lüders. Dr.  Sukthankar spent a few weeks in Berlin in October 1931 for this purpose. Dr. Lüders also would not leave Germany on account of his academic commitments there and undertake the work.” P. L. Vaidya, “Introduction,” in The Karṇaparvan for the First Time Critically Edited (Poona: Bhandarkar Oriental Research Institute, 1954), xl–xli. Sukthankar includes a fond and respectful recollection of Moriz Winternitz in Vishnu S. Sukthankar, “In Memoriam:  Professor Moritz Winternitz (1863–1937),” Annals of the Bhandarkar Oriental Research Institute 18, no. 3 (1937): 313–20. He notes that Lüders and Winternitz persuaded the XVIII International Congress of Orientalists in 1928 to place the existing collations and the funds collected for the Mahā bhā rata critical edition at the Indian scholars’ disposal. “Professor Winternitz was not merely an Honorary Member of the Institute, but also an active member of the Mahā bhā rata Editorial Board, and of the Mahā bhā rata Board of Referees.” Ibid., 313. He recalls that Winternitz declared that “after a careful examination of the irst fascicule of the new edition, he was convinced that this edition would be the edition that is wanted and that he had in mind when thirty-one years ago 30 300 PHILOLOGY AND CRITICISM he urged the necessity of a critical edition of the Mahā bhā rata.” Ibid., 317 (italics in original). See also the bibliography for Western scholars’ encouraging reviews of the Mahā bhā rata critical edition. The only person who views textual criticism through a nationalist lens is Grünendahl. 257 Sukthankar, “Prolegomena,” xxx. 258 Sukthankar had no reason to treat Ś 1 and K1 as members of a group, since their relation- ship could be stemmatized (as descendants of a common archetype ν). He considered the possibility that Ś 1 was the ancestor of the K manuscripts and rejected it. Their manifest diferences meant they were not manuscripts of the same version, albeit with a common ancestor. 259 As we saw Grünendahl only has a tenuous grasp of why manuscripts are grouped. He writes, “Precisely because there is no further representative of this supposed ‘old version of Kashmir,’ the existence of such a version [Ś ā radā  or  Kámīrī ], based on the material of one manu- script of the Ā diparvan, cannot initially be considered demonstrated.” Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 109 (Grünendahl’s italics). Apparently, he thinks a version based on a large group is more certain than one based on a single manuscript, since there are more witnesses for it, not realizing that a version refers to a speciic form of the text, whereas a group refers to a subfamily, the relation of whose members resists analysis. He thinks grouping manuscripts is an index of their closeness, when it is an index that their relationship is unclear. His proposed solution, subsuming Ś 1 and K1 to one group, therefore will not express their greater commonality—greater, that is, than he thinks Sukthankar was willing to recognize—but at most underscore the editor’s inability to determine their rela- tionship. Since the editor was not in doubt about their relationship, Grünendahl’s proposal efectively discards known facts for a speculative consideration. 260 Their internal relationship never poses a problem for Grünendahl, because he is not inter- ested in their readings, but only their extent. 261 Grünendahl thinks that a critical edition is only justiied if it completely and accurately succeeds in reconstructing a historical archetype, but this is not the sole purpose of a critical edition. In fact, few scholars still think critical editions replicate historical exemplars in all their particulars. The more sophisticated view is that the critical edition is an arrangement of the available data, a hypothesis that presents the diachrony of the tradition. 262 Edgerton comments, “Much has been said of the ‘luidity’ of Mbh. tradition. In a sense the term is justiied. But it is dangerous to use it without deinition, and I miss this in the writings of many who use it. What do they mean? (a) If they mean that there has been a vast deal of inter-inluence and contamination between various MSS and recensions, I have already said I agree. But none of these inluences seems, as far as I can see, to show evidence of coming from any version not descended from our ‘original’ text (however expanded or otherwise sec- ondarily altered). (b) Or if what is meant by ‘luidity’ is that, before the establishment of this text, the ancestor of all our MSS, there were already diferent versions of the Mbh. stories, again I agree. I have cited evidence from our text (as Sukthankar did from the Ā di) to show that this text itself contains signs of such variant versions. The late and obviously secondary attempts to patch them up, to which reference was made, are to me further evidence that all our MSS go back to this text, inconsistencies and all. Whatever other versions existed before it are now, apparently, lost forever, except as they are dimly relected in the manner described in our text itself. But this text itself is nothing ‘luid’! To be sure we must at present, and doubtless forever, remain ignorant about its many details. But we should not confuse our ignorance with ‘luidity’ of the text itself. That, to put the matter in a nutshell, seems to me precisely what those critics are doing who call the critically edited Mbh. an ‘imaginary’ thing. Because we, the editors, honestly admit that we cannot be sure just what form it had in many details, they jump to the conclusion that it never had any form. I, on the contrary, have no doubt at all that every line of the text had once a deinite, precise form, even though we are now frequently uncertain 301 CONFUSIONS REGARDING CLASSIFICATION 301 about just what that form was. It is not an indeinite ‘literature’ we are dealing with, but a def- inite literary composition.” Edgerton, “Introduction,” xxxvi–xxxvii (all italics in original). 263 “At best,” because it is uncertain whether such an edition could be completed at all. Certainly, Grünendahl’s accomplishments in this area do not inspire conidence:  25  years since he published his criticisms of Sukthankar, his irst collations are yet to appear. 264 Leonardi, “Il testo come ipotesi,” 9. 265 Ibid., 8. 266 Cited from the abstract to the article. 267 Ibid., 9. 268 Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 120. 269 Ibid., 120–21. 270 Edgerton, “Introduction,” xxxviii (Edgerton’s italics). 271 Contrary to Grünendahl’s claim, Sukthankar did not “attempt” to “explain away” the “not present textual additions” in D2 and D5 as due to contamination. He cited contamination to explain the occasional commonalities between the K manuscripts and D5 even though D5 was not descended from ν. He could not have cited contamination to explain the absence of certain interpolations, for the simple reason that it probably never occurred to him that anyone could establish iliation on the basis of the absence of interpolation. 272 Grünendahl, “Zur Klassiizierung von Mahābhārata-Handschriften,” 120–21. 273 As the latest common ancestor of K0.1, D5, Ñ4 and S, θ is the ultimate source to which all three recensions owe the passages’ absence. As the latest common ancestor of K0.1, D5 and Ñ4, N is the latest source to which the northern recension manuscripts owe the passages’ absence. 274 Ibid., 121. 275 Classifying manuscripts by number of interpolations is like grouping children into “families” by the number of their illings. Just because it can be done, does not mean it should be done. 276 Ibid., 121 (italics in original). 277 Large numbers of interpolations only mean the two manuscripts are equally contaminated. If we found they consistently had the same interpolations, we might suspect the contamina- tion was into a common ancestor, but whether they are really related must be established in terms of shared corruptions. In contrast, large numbers of absent interpolations—a possi- bility we must address, since Grünendahl also argues from shared absences to iliation—only means both manuscripts are equally free of contamination. It does not mean one of the contaminated manuscripts belongs with the other, either in a group or as an apograph of a common hyparchetype. 278 The restriction is also pointless in another sense: it is not an accurate indicator of contamina- tion even between pairs of manuscripts. If our aim was to show the extent of contamination of K2 with some other K manuscript, we would draw up a list of all the interpolations they have in common and not just those they have in common with only each other. Although we could not rule out a third manuscript as the source of contamination, the problem is not solved by restricting ourselves to only a pair of K manuscripts at a time. A simple example suices to show this: let K3’s scribe compose 80 additional passages and K4’s scribe, two. Let the scribe of K2’s ancestor have access to both K3 and K4. Anxious not to lose any verse that might possibly be original, he copied all the additional verses into his manuscript, either in the margins or on separate sheets. If this copy were copied again and the scribe inserted the additional passages into the text, we could no longer distinguish the original from the additions. Let us further assume that the scribes of K0 and K1 also had access to K3 and copied its additional passages, either some or all of them, but between the two of them at any rate all of them. K2 is thus heavily contaminated with K3 and only slightly with K4. Yet, by Grünendahl’s logic, we would ignore the interpolations K2 shares with K3, as they are not unique to these manuscripts—or, as he puts it, K2 does not have these passages “in 302 302 PHILOLOGY AND CRITICISM common only with one other K manuscript”—and focus on the two passages K2 shares with K4, which are not contained in any other K manuscript, since K0.1 escaped contamination with K4. We would conclude that K2 is closer to K4 than any other manuscript, even though K4’s inluence on K2 is slight. We would also ignore the interpolations in K3—since they are shared with more than one K manuscript, we would not include them in our list at all—and we would conclude that K3 was the purest manuscript even though it is the source of most of the interpolations. If K2 was itself a source of new interpolations (say, 40) and 30 of them were transmitted to K0 and K1, while the remaining 10 were transmitted to K4, we would ignore the contamination of K0 and K1 since the passages are found in both, and conclude, from the fact that K4 has 10 passages in common with K2, that these two manuscripts were especially closely related. This would be erroneous, because, in fact, K2 is much more closely related with K0 and K1 (“related” in the sense of being contaminated with and not in the sense of being iliated, which is impossible to establish). 279 Ibid., 121. 280 This confusion illustrates how little Grünendahl understands manuscript transmission. The primary source of diference between manuscripts is not contamination, but the signiicant errors made each time they are copied. Contamination has rather the efect of homogenizing the tradition. 281 Ibid., 121. 282 Ibid. 283 Namely, the hyparchetype γ was “not as yet constituted at this stage of his discussion”; K5.6 “were only available for a portion of the listed passages”; “none of the Ñ manuscripts were collationed for the irst two adhyā yas […] so that conclusions based on evidence from these sections can a priori not hold for the Newari manuscripts”; and “Ś 1 also was not available for these sections of the text.” Ibid., 121–22. 284 Grünendahl’s objections are false. When Sukthankar says that “the contamination of K3–6 with γ is illustrated by the following passages,” γ stands for the manuscripts of that recension, not the hyparchetype γ itself. It is impossible to demonstrate that contamination took place with the hyparchetype itself—as Sukthankar knew. Further, the fact that some manuscripts of this recension such as D5 do not contain the passages is evidence that they were not features of the hyparchetype γ. Even if Sukthankar meant the hyparchetype itself, the idea that he cannot refer to it as it has not yet been “constituted” is mistaken. An editor need not reconstruct all of the hyparchetypes he posits during the recensio on the way to the arche- type. The reason he cannot is that it entails a circle: in order to reconstruct hyparchetype α he must refer to hyparchetype β, but he cannot refer to it—at least, by Grünendahl’s logic—without irst reconstructing it, but in order to reconstruct it, he must refer to α, but he cannot refer to it without irst reconstructing it and so on. In contrast, Grünendahl’s second claim is a ib: K5.6 were not “only available for a portion of the listed passages.” They were discarded from the critical apparatus after the irst two adhyāyas as of little value in establishing the text, as Sukthankar explicitly notes. The list of passages therefore self- evidently only applies to K5.6 for the irst two adhyāyas. Why Grünendahl feels obliged to point this out is unclear. Once an editor has shown that manuscripts are contaminated—so contaminated that he sees no point in retaining them in the critical apparatus—there is no reason he should continue giving examples of their contamination. Grünendahl raises this objection because it casts doubt on the probity of Sukthankar’s list. The same thing applies to his third claim, albeit in reverse: collations of Ñ1 and Ñ2 began only after the irst two adhyāyas; collations of Ñ3 after the 13th. Obviously, examples that pertain to parts of the text for which they were not available do not apply to them. But no one claims that every interpolation is found in every manuscript. When Sukthankar says that these passages illustrate K3–6’s contamination with the manuscripts of the γ recension, he simply means these interpolations are characteristic of the latter; not they are found in all manuscripts 30 CONFUSIONS REGARDING CLASSIFICATION 303 of the latter. It suices to show that the manuscripts on the right frequently feature the interpolations of the manuscripts on the left, without either every manuscript on the right or every manuscript on the left containing every interpolation. Finally, Ś 1 need not be available for this section of the text for the editor to identify a K type. As we saw, this is Grünendahl’s delusion. Since we addressed this confusion in the preceding section, we do not discuss it again. 285 Ibid. (all italics Grünendahl’s). Contrary to Grünendahl’s claim, Sukthankar explicitly notes that G1.2.4.5 and T1 are the inferior manuscripts of their respective groups, as several northern interpolations demonstrate. “T1 is one of the extremely few Southern manuscripts which contain the (Northern) salutational stanza nārāyaṇaṃ namaskṛtya, etc. For the Northern element in the make-up of T1, cf. 29*, 30*, 96*, 97*, 98*, 106*, etc. As compared with T1, T2 shows a purer Southern tradition and has distinct leanings towards the Grantha version.” “The latter group (G3.6) represents the purer Southern tradition, agreeing with M against the other Southern manuscripts, whereas the four MSS. G1.2.4.5 are not merely heavily interpolated but stand palpably under Northern inluence.” Sukthankar, “Prolegomena,” lxxii. “Take, for instance, the case of the subgroup G1.2.4.5 of the Grantha version. In opposition to other manuscripts belonging to the same recension and even the same version, G1.2.4.5 contain, as shown above, an astonishingly large number of passages which are found otherwise only in some inferior manuscripts of the Northern recension. Now is this a case of contamination of the four MSS. G1.2.4.5 from a Northern source; or are the common passages a remnant of the lost archetype, which were somehow lost in the remaining manuscripts of the Southern recension? There is apparent agreement here between independent versions. But is this agreement original? The clumsy interpolator of a remote ancestor of G.1.2.4.5 happens to have supplied us with the means of answering these questions. He has left behind, quite unintentionally, an impress of his ‘inger prints,’ so to say, by which we can easily and conidently trace him and examine his handiwork. The said manuscripts contain a Northern stanza (belonging to manuscripts of class γ)—a mere string of attributes of Garuḍa—wedged in at a place where it can be construed neither with what precedes nor with what follows. This proves incontrovertibly that these four manuscripts G1.2.4.5 have been compared with some Northern manuscripts, and makes it highly prob- able that the other doubtful stanzas, which they have in common with the Northern recen- sion, have crept into their text in the same surreptitious way.” Ibid., lxxxi (Sukthankar’s italics). As App. 1, no. 14 is an insertion into the Garuḍa narrative, we have every reason for thinking it a northern insertion into the manuscripts of the southern recension that feature it rather than vice versa. 286 For Sukthankar’s comments on these manuscripts, see preceding note. 287 The reference is to 116*, 119*, 122*, 124*, 125*, 128*, 132*, 137*, 139*, 142*, 143*, 144*, 145*, 151*, 157*, 160*, 162*, 166*, 167*, 168*, 172*, 173*, 189*, 190*, 191*, 221*‚ 228*, 245*, 281*, 305*, 354*, 372*, 405*, 416*, 417*, 438*, 487*, 490*, 523*, 536*, 564*, 692*, 694*, 824*, 1000* and 1035*. Sukthankar introduces them with the words: “The contamina- tion of K2–4.6 with γ is illustrated by […],” illustrating the list’s general nature. It is neither necessary that every K manuscript contain every interpolation nor necessary that every inter- polation appear in every γ manuscript. Ibid., li. 288 Grünendahl has at most shown that the absence of App. 1, nos. 14, 43 and 61 is not restricted to K0–2 but also extends to K3. In contrast, he has not shown anything concerning K4–6. But the unwarranted generalization irst lends his argument persuasive force. 289 Grünendahl could simply have noted that App. 1, nos. 14, 43 and 61 are interpolated only in K4, or he could have noted that he did not think Sukthankar had suicient examples of K3’s contamination with γ. But as we saw (see preceding note), the argument relies for its persuasive force on the generalization and on phrasing the criticism negatively and incomprehensibly. 304 304 PHILOLOGY AND CRITICISM 290 The absence of 26–50 and 76–100 is no less characteristic of ν than the absence of 51–75 (the passages whose absence is characteristic of ν are thus 26–100). But as Grünendahl uses the term, he means passages absent only in ν. We have retained his usage. 291 A scribe could only know which interpolations are characteristically absent in ν if he knew which interpolations exist in γ and S. But to know this, he must have access to γ and S along with all of ν. No scribe can be expected to do this. 292 If Grünendahl had simply argued that contamination must be proved in terms of manuscripts not lacking the same passages as the purest members of their group, the problem would have been obvious: the demand does not address the impure manuscripts’ contamination but their lack of contamination with passages with which the purest members are also not contami- nated. Hence the obscure formulation “preferentially demonstrate […] in terms of textual editions whose not being present is characteristic of the hyparchetype,” necessary to render the argument opaque. Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 122 (italics in source). 293 This is true in the trivial sense that every manuscript is contaminated with the passages it is contaminated with and not contaminated with the passages with which it is not contam- inated. Grünendahl’s entire genius consists in phrasing the requirement so impenetrably that no one realizes that it places an arbitrary and exorbitant restriction on the editor: that he demonstrate A’s contamination only in terms of passages absent from α and in terms of all the passages absent from α. Consequently, every manuscript, even the most contami- nated, can now be proven not contaminated provided we select our examples appropriately. 294 Ibid., 122–23. 295 The source of the confusion is Grünendahl’s belief that the Ñ manuscripts were assigned to the γ recension for containing all of these passages. Consequently, he thinks that by pointing out that they do not, in fact, contain all of the interpolations in question, he can controvert their assignment to this recension. 296 Grünendahl’s igure is incorrect. D14 features 139* in addition to 132*. It also features the irst line of 137*. His error is probably due to the fact that the note to verse 1.2.148 is com- plicated. Sukthankar notes that D4 (marg).9–11 and S insert the passage 138* after 148 (G7, after 139*); Dr and D14 after 150 and then he continues:  “K3.4.6 V1 B D (except D14) G7 ins. after 148 (D4.9–11, after 138*):  139* [etc.].” V. S. Sukthankar, ed., The Ad̄ iparvan for the First Time Critically Edited (Poona:  Bhandarkar Oriental Research Institute, 1933), 50 (note to 1.2.148). In other words, whereas D14 inserts 139*, it does so following an additional insertion 138* characteristic of the southern manuscripts, but absent from the K manuscripts. This undermines Grünendahl’s theory that “D14 […] could have, with justiication, been assigned to the K manuscripts.” Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 123. 297 Sukthankar, “Prolegomena,” lxxi. 298 Ibid. 299 Ibid., xx. 300 Edgerton conirms Sukthankar’s assessment. He notes that the manuscript “described as ‘D14’ by Sukthankar on Ā di. Text of Sabhā originally had 78 folios (of which about a dozen are lost), 16–18 lines, 60–65 akṣaras. [It was] completely collated, but collations have been discarded (except for an occasional citation in the Addenda), since it is fragmentary and goes closely with Dn1.” Edgerton, “Introduction,” xix. 301 Grünendahl also errs in claiming that as D14 lacks two of the irst 25 passages it belongs with the K or Ñ manuscripts. As we observed earlier, a manuscript’s contamination must be demonstrated in terms of the passages with which it is contaminated. In the pre- sent case, what is decisive is not whether D14 features all or only two or none of the irst 25 passages in Sukthankar’s list. Rather, what is decisive is how many of the 244 smaller interpolations in the irst two adhyāyas are inserted in D14. When the insertion 305 CONFUSIONS REGARDING CLASSIFICATION 305 is characteristic of the northern manuscripts, D14 frequently does not feature it. In this respect, it usually agrees with the best K manuscripts, though not always: sometimes it is the K manuscripts that feature the insertion. On the other hand, when the insertion is characteristic of the southern manuscripts, D14 often features it, against the practice of the K manuscripts. This is exactly what Sukthankar found. As his list is speciically a list of insertions illustrating K3–6’s contamination with γ, that is to say, with manuscripts of the northern recension, it is not surprising that D14, a composite of a Kaśmīrī and a southern source, should not feature these insertions. In parts unafected by contamination, D14 follows the purer Kaśmīrī tradition. In the others, it follows the southern. At most, Grünendahl’s list demonstrates that D14 is not contaminated with a northern source. As no one claims this—D14 was discarded as contaminated with a southern source—it is not clear what Grünendahl hopes to prove with this argument. Every manuscript can be shown not contaminated if we restrict our discussion to passages with which it has not been contaminated. 302 Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 123–24. 303 As noted, K6 is discarded from the critical apparatus after the second chapter. The inal 21 passages therefore cannot apply to it, and it is disingenuous to suggest that Sukthankar claimed otherwise. The circumstance that K2 features eight of the passages is evidence of its contamination. It may not be as highly contaminated as K3 and still less than K4, but an editor must still note with which passages it is contaminated. Why Grünendahl thinks the editor should adopt the most contaminated manuscript of its group as his standard is a mystery. It contradicts what he said earlier about an editor having to adopt as his standard the non-contaminated manuscripts of its group, although there he was trying to prove that K3 was not contaminated. Likewise, the circumstance that K3 contains 37 of the passages is a strong argument for contamination. We can turn this evidence around and claim that K3 does not contain 9 of the 46 passages, but this is a bizarre argument against its contamination. No one claims K3 contains all the central recension’s interpolations, just suicient numbers to assume contamination with a γ source. Likewise, the fact that D5  “manifests diverse similarities with […] the non-contaminated manuscripts” only means it is less contaminated than other D manuscripts; not it is an apograph of a K source. Many of these passages are characteristically northern insertions and hence absent from the southern recension. Should we now claim the southern manuscripts are apographs of a K source? 304 Ibid., 124 (all italics Grünendahl’s). 305 These are the insertions 221*, 228*, 245*, 281*, 305*, 354*, 372*, 405*, 416*, 417*, 438*, 487*, 490*, 523*, 536*, 564*, 692*, 694*, 824*, 1000* and 1035* from Sukthankar’s list. The ive longer passages are App. 1, nos. 63, 75, 80, 81 and 85. 306 Edgerton addresses the underlying fallacy in his introduction to the critical edition of the Sabhā parvan:  “Retention of original features never even tends, in the slightest degree, to prove secondary relationship with any other manuscript or recension which also retains them. Such relationship can only be proved, or made in any way likely, by demonstration of common departures from the original.” Edgerton, “Introduction,” xlviii (Edgerton’s italics). Grünendahl evidently overlooked the passage. 307 It is evident why Grünendahl thinks these passages should have been the “touchstone” for evaluating Ñ4’s place in the stemma. According to him, since these insertions distinguish K0.1 from K2–6 and K2–6 again from γ, they constitute the speciic diference between ν and γ. Thus, if we want to determine whether a manuscript belongs under ν or under γ, we must examine whether it contains only some or all or none of the passages. The argument is doubly erroneous: the passages in question are neither characteristic insertions in K nor a complete list of characteristic insertions in γ. If our aim was to demonstrate that Ñ4 was sim- ilar to K0.1 by virtue of not containing many of the insertions K0.1 do not contain, we would 306 306 PHILOLOGY AND CRITICISM draw up a list of all the passages absent in Ñ4 irrespective of the recension in which they feature. This is exactly what Sukthankar did: he notes that “our MS omits most of the short insertions given in the foot-notes of the constituted text. There are more than two thousand of such passages, the majority naturally from the Southern Recension,” and then he cites 47 examples “selected at random.” Sukthankar, “The Oldest Extant Ms. of the Ā diparvan,” 204. On the other hand, if our aim was to demonstrate that Ñ4 was similar to K0.1 by virtue of not containing many of the insertions from the γ recension K0.1 also do not con- tain, we would draw up a list of all the passages absent in Ñ4 though characteristic of the γ manuscripts, but in no case would we draw up a list of just these 21 passages! Why Grünendahl thinks “the choice in his list of the shorter additions not attested in Ñ4 must not therefore have been a random one, as Sukthankar here claims too” is a mystery. Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 124. Perhaps he suspects this because he thinks they dem- onstrate something they cannot, namely, the iliation of Ñ4 with Ś 1 and K0–3. 308 See ibid., 127: “As regards the ‘important additions’ in ν that are discussed next (Prolegomena, p.  lxx), Sukthankar can only ‘recall’ one, namely, the similarity with Kṣemendra’s Bhāratamañjarı̄ (App I, 121), already discussed earlier (see above, p.  112). He [Sukthankar] tries to dispel the fact that ν at all manifests ‘important additions,’ which he manifestly found a defect, by referring to the fact that the addition is found ‘only’ in Ś 1, K0.1(4) and hence ‘cannot be considered as characteristic for ν as a whole.’ Such reasoning reveals Sukthankar’s remarkable lexibility in interpreting his material: was he not concerned previously to show that K0.1 in particular constituted the ‘pure’ core of the K version and that all of them—but especially K1—are especially close to the representative of the ‘Ś ā radā version,’ Ś 1 (which, to all excess, also features this addition!)?” 309 Grünendahl could have spared himself the word “recall” in quotation marks (see preceding note): App. 1, no. 121 is the only longer insertion in ν. Does he expect Sukthankar to make up passages where none exist? As for his claim that Sukthankar “manifestly” found its presence in ν a defect, he is right in that the presence of an interpolation is a defect, albeit a relatively harmless one, since it is easy to identify and to correct. But it is not a defect in the sense he means it, namely, as a ground for Sukthankar to revise his classiication. Sukthankar neither had a reason to suppress other important additions in ν, since there weren’t any, nor a reason to feel that citing them would undermine his classiication, since it wouldn’t. Only Grünendahl thinks these insertions would afect the stemma, because he does not consider readings and, above all, shared corruptions, but interpolations when grouping manuscripts. 310 Why exactly Grünendahl thinks the fact that ν itself has “important additions” is an argu- ment against it is a mystery. Perhaps he thinks this argument strengthens the case for his proposed edition or perhaps he thinks the stemmatic method does not work and therefore all we can do is take the shortest manuscript, discard its interpolations and arrive at the archetype. In the preceding section, we saw that, for him, all an editor needs to do to create a critical edition is begin with the shortest text, identify the longest open line and then brush aside the dead ends. But how are we to identify interpolations if not for the stem- matic method? 311 The fact that App. 1, no. 121 is found only in Ś 1, K0, K1 and K4 is not an argument against Ś 1 and K0.1 as the best ν manuscripts. Grünendahl thinks it is but that is because—as we know by now—he classiies manuscripts into the meliores and deteriores based on speciic insertions. In contrast, an editor will distinguish between a manuscript’s readings, which can be good or bad, likely the reading of the original or a corruption, and its expansion, which is not necessarily a sign of poor quality. Grünendahl errs in disregarding this distinction. From his perspective, it seems that, because Sukthankar concedes Ś 1 and K0.1 also have one longer insertion (albeit one not characteristic of ν as a whole, since it is absent from the remaining K manuscripts), he expresses “remarkable lexibility in interpreting his material.” Actually, the circumstance that 307 CONFUSIONS REGARDING CLASSIFICATION 307 Ś 1 and K0.1 also contain some interpolations does not negate their good readings. Neither does it negate the fact that they are the purest representatives of their group. Indeed, they contain several other insertions, but Grünendahl likely focuses on App. 1, no. 121 because he thinks, as an “important addition,” it especially disproves Sukthankar’s claims. 312 Among longer passages App.  1, nos. 1, 12, 33 and 58. The shorter additions were listed previously. 313 Ibid., 125–26. 314 Sukthankar, “Prolegomena,” liv. 315 Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 126. 316 Ibid., 126–27. 317 If the reading in γ was the original and that in ν and S the corruption, we would have to explain how both could contain the same corruption. Their scribes could have coincidentally made the same error several times, but if they agree consistently in their reading, this suggests a common source. Their reading must be the reading of the archetype and the divergent reading in γ a corruption. 318 Grünendahl is evidently projecting his anxiety on Sukthankar. Since his aim is to argue for a critical edition based on the Nepā lī manuscripts, he thinks Sukthankar must be negating the claims of the Nepā lī recension, since this is what he would have done in his place. See the next note. 319 Contrary to what Grünendahl thinks, Sukthankar had no reason to insist on K’s superi- ority. He had access to all the Mahā bhā rata manuscripts, including Grünendahl’s Ñ4. If he thought it the better manuscript, he would have used it. As a matter of fact, he studied it extensively—more so than Grünendahl, who relies exclusively on Sukthankar’s informa- tion about it—and produced the irst and, until today, sole collations of the manuscript. Grünendahl’s argument depends not only on imputing his own motives to Sukthankar but also on the assumption that Sukthankar in 1931 knew that in the future German Indologists would have preferential access to the Nepā lī manuscripts through the Nepal-German Manuscript Preservation Project and wished to preempt them from making a contribution to the Mahā bhā rata. 320 If the reading attested in ν and S is the correct reading, then it originated with the archetype. If it originated with the archetype, it could have descended to some manuscripts of the central recension. In this case, some γ manuscripts will preserve the correct reading and we should not be surprised that ν and Ñ share the same reading. Contrary to what Grünendahl thinks, this does not mean ν and Ñ are related, since correct readings are not evidence of iliation. 321 Sukthankar overlooks to write “mostly” only once in the 21st example in his list (the 15th in Grünendahl’s). This is the sole case where Grünendahl can claim to have identiied an excep- tion. But as the fact that at least one Ñ manuscript features the correct reading is only further conirmation for the constituted text, its relevance is trivial. 322 Grünendahl claims in four additional cases to have found a discrepancy between what the manuscript contains and what the critical apparatus notes. But as he provides no evidence, we could not verify the truth of this claim. 323 Grünendahl does not think so, but this is because he misinterprets Sukthankar’s comment, “One notable feature of ν to which I must now draw attention is its frequent agreement with S against γ, especially in the matter of isolated and even unimportant readings.” Sukthankar, “Prolegomena,” liv. He thinks Sukthankar means to assert ν and S’s independence from γ, when, in fact, he asserts ν’s independence from S or, rather, infers from this independence that their reading must be the reading of the archetype. 324 That is, in those cases when they agree with ν. When they do not, they, of course, contain diferent readings, but this is not the aspect that interests Grünendahl. 325 Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 127. 326 Ibid. 308 308 PHILOLOGY AND CRITICISM 327 “Since our codex (Ś 1) is fragmentary, it must be considered a piece of singularly good fortune that there has been preserved at least one nearly complete Devanā garī manuscript of the Ā di, namely, India Oice no. 2137, that may, as will presently be shown, be used, without hesitation, to supplement the missing portions, since it undoubtedly is a moderately trustworthy, though comparatively late and slightly contaminated and incorrect transcript of a Ś ā radā exemplar.” Sukthankar, “Prolegomena,” xlvii. And see also his comments concerning the ν recension: “An unbiased comparative survey of the diferent versions leads one to the conclusion that the Ś ā radā (Kaśmīrī) version is certainly the best Northern version, and probably, taken as a whole, the best extant version of the Ā di, a conclusion not based on abstract considerations, but one that may be veriied inductively and pragmatically. As is natural, this version is, not by any means, entirely free from corruptions and interpolations. These must be carefully corrected and controlled with the help of the other versions, particularly of those of the rival recension.” Ibid., lvi. 328 Ibid., li. 329 2.144b: K0 cābhibho; K1 cābhitaḥ ; K2 V1 B Da Dn D4 D2–6.13 vā vibho; K4.6 D7–12 vā vibhoḥ ; D1 yā vibho; T1 G4.5 tu prabhuḥ ; T2 G6 ca prabhuṃ; G1–3 M2.4 ca prabhuḥ ; G7 ca prabhoḥ ; D14 saṃjayaḥ sarvatraḥ prabho. 2.177b:  V1 ślokānāṃ; K1 ślokā caivātra kathyate; K3 ślokā granthārthasanmataḥ ; K4 Da D2.13 ślokasaṃkhyātra śabdya; K6 ścaivātra bhāṣitāḥ ; B1.2 D6 ślokasaṃkhyātra kı̄rtya; B3 ślokasaṃkhyā ca kı̄rtya; B4 nāṃ cāpi śabdya (marg. cātra kı̄rtya); D3 Dn D1.3.4.7.9–12 ślokasaṃkhyātra kathya; D5 ˚nāṃ cātra gı̄ya; D3 ślokasaṃkhyātra gaṇya; T1 gryaṃ yasya śabdya; G1.4.5 nāṃ cātra śabdya; G2 gryaṃ yatra gaṇya; G3 gryā yatra yasya; ślokā yatra praśabdya. Cd has at least ślokāgraṃ. 98.13a: Ś 1 kanyasaṃvāde; K1 kamahaṃ vade; K2 labdhā sa vade; K3.4 Ñ2 B D mā gamaḥ kāmaṃ; Ñ1.3 kanyasa bhrūhi; S kanı̄yastāt naiveha. 98.18c: K0–2 samudre; D5 samūhe; S samṛddhe; K3.4 Ñ2 B D (except D5) buddhoḍupe pari (B5 vini) kṣipya. 330 In a ifth case (1.2.189c), the diicult reading is found in all K manuscripts except K4. All southern manuscripts except one support this reading, so Grünendahl’s objection is moot. In a inal case (1.10.6c), all K manuscripts except K1 and K4 have the reading, so the objection is once again moot. Grünendahl appears to have misread the critical apparatus, since it notes that all Ñ manuscripts feature the reading kāmaṃ māṃ. 331 Sukthankar, “Prolegomena,” lxxxvi and see also ibid.:  “With that end in view, we must examine as many manuscripts—and above all as many classes of manuscripts—as possible, and group them into families” (all italics Sukthankar’s). 332 Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 127. 333 Sukthankar, “Prolegomena,” lvi. 334 As we have seen, this is Grünendahl’s misconception. Sukthankar’s point is that ν’s reading must be the correct one since it explains the other variants. This reading is also correct in any other manuscripts that contain it. The idea that he seeks to claim “exclusivity” for ν or that, because other manuscripts conirm its reading, Sukthankar should not have constituted the text on “ν alone” is Grünendahl’s interpretation. 335 The critical apparatus includes only the three manuscripts Ñ1–3. For Ñ4, Grünendahl appears to have consulted Sukthankar’s collations in “The Oldest Extant Ms. of the Ā diparvan,” and, from the fact that he does not list a variant for 1.166.23, concluded that Ñ4 has the reading of the constituted text. 336 Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 127 (italics added). 337 Grünendahl does not title the sole edition he produced—Reinhold Grünendahl, ed., Viṣṇudharmāḥ : Precepts for the Worship of Viṣṇu Part 1: Adhyāyas 1–43 (Wiesbaden: Harrassowitz, 1983)—a critical edition, but it is also not a true Bédierist edition, that is, an edition of the best available manuscript. He elects a group of manuscripts and reprints their readings with random “improvements” from the manuscripts of the other groups, following no apparent logic. He claims, “No single manuscript, and none of the established groups, taken by itself, ofers a satisfactory text. Given the choice between following in efect one 309 CONFUSIONS REGARDING CLASSIFICATION 309 manuscript singled out on account of its reliability, age, or other criteria, or selection the ‘best’ reading of all available manuscripts, a modiication of the former approach seemed most suitable to the material,” but neither identiies a single manuscript nor consistently follows the reading of the better group. Thus, while the edition appears to be based on the manuscript N4, many instances exist (1.32b, 1.47a, 6.4b, 6.13b, 7.21ab, 17.9c etc.) where its reading is rejected from the constituted text. He writes, “Of the three groups distinguished above, the two consisting mainly of palm-leaf MSS deserve preference over the paper MSS of Group III. Of the former, Group I  clearly stands out for its reliability. Its numerous insertions (always deined with regard to the constituted text) do not justify the assumption that its general ‘correctness’ is merely the result of thoroughgoing revision,” but at 2.25c, 2.42a, 2.57c, 2.73b, 2.78ab, 3.56c, 4.3ab and 5.22b he rejects readings contained in three of the four manuscripts—N3, 6 and 9—of Group I (the sole exception is N4); and at 1.34b, 1.42d, 7.11cd, 7.16a, 8.9cd, 13.30ab, 16.3b and 16.21d, 17.1b readings contained in N3, 4 and 6 of Group I. Sometimes all manuscripts of his better Group I are against the con- stituted text’s reading (2.48c, 4.40c, 4.49b, 5.4cd, 5.15c, 5.25a). At 3.29ab, 4.15b, 4.54cd, 6.23cd, 10.3c, 13.6b and 17.5cd all N manuscripts of either group are against the constituted text. Neither is he consistent in the application of the principle, “the bulk of the readings of Group I, as well as its predominance over the lacunae of Group II, is supported by another branch of tradition, i.e. Group III (cf. above). Therefore I have considered their far-reaching agreement as ofering the irmest basis for the present edition,” because sometimes the D manuscripts, either with or without some of the manuscripts of his Group I, support a diferent reading than that of the constituted text—for example, at 1.22d, 1.47d, 1.50a, 1.58a (at 1.32b with two of his better manuscripts N4 and N6 and at 1.47a with his best manuscript N4). Indeed, without drawing up a stemma and localizing his manuscripts, he cannot show that their agreement is “far-reaching.” Occasionally, he prints the reading of DB and DL in the constituted text, for example, at 19.14c where he notes “all mss. except DB, DL” contain the variant kāsāram (constituted text:  kṛsāraṃ), even though otherwise he discounts their reading (1.64b, 2.31d, 2.49a, 2.54c, 2.58c, 2.72b, 3.34a, 3.45c, 4.22a, 4.36cd, 4.40a, 4.45c, 4.45d, 4.50b, 5.18b, etc.). He writes, “concerning its insertions, the evidence of Group III, corroborated by Group II and independent sources (citations, parallel versions) proves to be superior to that of Group I,” but he often treats as original passages that are clearly insertions in some exemplars, since they are marginal insertions in the others and omitted in either the best groups or sometimes a majority of manuscripts (for example, 1.61cd, 2.82cd–84ab, 7.29ef, 16.8ab and 19.23). 1.61cd is absent from N1, 2, 5, 7 and 8 and inserted in the margin in N2, so that the evidence of Group III is not corroborated by Group II. Again, 7.29ef and 16.8ab are absent from N1, 2, 5, 7 and 8, and inserted in the margin in N1, suggesting its secondary nature. Ibid., 25–26 (all italics added). 338 Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 128. 339 Ibid. 340 “So also I  consider the starting point of his [Sukthankar’s] inductive demonstration, the Schriftartprämisse, to put it mildly, highly questionable. ‘Abstract considerations,’ to which Sukthankar pretends to be little inclined, speak, in my view, unambiguously against it.” Ibid., 128. 341 “How would the emergence of a version have to be thought ‘abstractly’ under the Schriftartprämisse? I basically see two possible explanations: either the text was ixed in a deinite script directly from the oral transmission or it was transferred from another script into this one. Whichever explanation one accepts neither of these two processes need have occurred only once. The text could have been ixed in a script from the oral tradition repeat- edly and in completely diferent periods or it could also have been transferred from completely independent sources. I consider it unlikely that this transition—be it from the oral tradition or from another script—should have occurred only once and that all the manuscripts in a speciic script should therefore be traced back to this prototype. In the present case, it is after 310 310 PHILOLOGY AND CRITICISM all far more likely that the manuscript tradition of the Mahā bhā rata that has come down to us in a speciic script goes back to multiple, independent sources.” Ibid., 128–29. 342 “When the Schriftartprämisse lose their place in the tool-box of Indological textual criticism, the classiication of manuscripts must be placed on a new foundation. It appears logical to focus initially on the manuscripts of individual regions. In particular, the relationship of the Newari manuscripts to each other, their position in the manuscript tradition of Nepal as well as of the Mahā bhā rata as a whole deserve a closer investigation. The conditions for this are given by the work of the Nepal-German Manuscript Preservation Project.” Ibid., 130. 343 “In particular, Sukthankar’s classiication is placed in question by Ñ4: (1) it is shorter than the textus simplicior constituted by him [sic], without standing in a demonstrable relationship to it; (2) it fulills many of Sukthankar’s criteria for his constitution of the hyparchetype ν discussed above; (3) it does not show any traces of contamination; (4) contrary to Sukthankar’s claim it stands in a closer relationship to only one other Newari manuscript, namely, to Ñ3; (5) according to Sukthankar’s information, it manifests conformities with the southern recen- sion that, when they cannot be traced back to contamination, are possibly to be ‘traced back to an original connection via the lost archetype.’ ” Ibid., 129. 344 See Reinhold Grünendahl, “Post-philological Gestures—‘Deconstructing’ Textual Criticism,” Wiener Zeitschrift für die Kunde Südasiens 52–53, Text Genealogy, Textual Criticism and Editorial Technique (2009–10): 27, n. 55: “In my view, a stemma may contribute substantially to what I consider the ultimate goal, viz., Nachvollziehbarkeit of the editorial process, but it is not a sine qua non. If it was, Indologists would ind themselves without critical editions of the Sanskrit epics, where contamination precludes such pretensions (see, e.g., Sukthankar, op. cit. [n. 23], p. lxxxii and lxxxvi; for general aspects, cf. Timpanaro, op. cit., p. 24). The same holds for various other ields where serious reservations towards strict stemmatology have been voiced; see, e.g., J. Hanneder’s edition of Abhinavagupta’s Mā linīślokavā rttika 1.1–399 (Groningen 1998, p. 40–45), and the Groningen edition of the Skandapurā ṇa by R. Adriaensen et al. (Vol. 1. Groningen 1998, p. 39).” 345 In his edition of the Viṣṇudharma, Grünendahl eschews genealogical analysis for a classiication of the manuscripts into the meliores and the deteriores. On pages 20–21 of the edition, he draws up a table, listing whether manuscripts feature certain passages. If a manu- script does not feature a given passage he enters a “–” sign before it; if it does he enters a “+” sign. If it repeats the passage, he enters an “r” (“repetition”); if it features it in another place he enters an “o” (“order changed”). Finally he also uses three other annotations: “m” (“mar- ginal insertion”), “d” (“common defect”) and “≠” (“divergence”). He prefaces the table with the words: “the interrelationship of the manuscripts can best be determined on the basis of variants comprising more than one pā da. With few exceptions that give convincing proof of a (more or less) direct dependence of individual manuscripts, smaller units hardly ever aford assistance in forming a picture as clear as circumstances would allow. The following table records the major variants of Adhyā yas 1–43. The conclusions drawn from it have been ver- iied by also checking Adhyā yas 67 and 102–105 (some 420 verse altogether).” Based on this table, he arrives at the following tripartite classiication: (1) Group I, which has “N6 and N3 at its centre and Ñ4 and N9 as gradually absenting associates [?]”; (2) Group II, which “clusters around N5 and N2, which are very intimately linked”; and (3) Group III, which “consists of the Devanā garī-manuscripts.” Grünendahl, ed., Viṣṇudharmāḥ : Precepts for the Worship of Viṣṇu Part 1, 22–23. The problem with this approach is that whether manuscripts contain or do not contain speciic passages is no guide to their iliation. Grünendahl does not examine their readings. Indeed, he appears unaware of the distinction between a signiicant error, a trivial error and a coincident error. He also appears not to know the distinction between horizontal and vertical transmission or that additional verses could most likely be transmitted through comparison, and do not mean that the two manuscripts featuring those verses were either copied from each other or constitute a “version.” His reconstruction is not stemmatic, and does not pretend to restore an archetype, though it presumably aims at an earlier state of the 31 CONFUSIONS REGARDING CLASSIFICATION 311 tradition, since he notes, “the bulk of the readings of Group I, as well as its predominance over the lacunae of Group II, is supported by another branch of tradition, i.e. Group III (cf. above). Therefore I have considered their far-reaching agreement as ofering the irmest basis for the present edition” and “In case of divergency [sic] between the two, I have generally preferred the readings of Group I. However, concerning its insertions, the evidence of Group III, corroborated by Group II and independent sources (citations, parallel versions), proves to be superior to that of Group I.” Ibid., 26. Actually, his edition represents a completely new text. It aspires neither to idelity to the original work nor to one of its witnesses, but randomly combines readings from diferent sources (cf. his comment “As against that the variants of other manuscripts had to be regarded as the last resort in those few cases, when the two constitutive groups [sic] ofered no acceptable alternative”; ibid.). with only the most supericial avowal of the principle of critical reconstruction. Gyula Wojtilla, review of Viṣṇudharmāḥ . Precepts for the Worship of Viṣṇu, Part 1 Adhyāyas1–43, by Reinhold Grünendahl, Acta Orientalia Academiae Scientiarum Hungaricae 39, no. 2/3 (1985): 393 is therefore incorrect to call Grünendahl’s Viṣṇudharmā a “critical edition.” 346 At 1.14.5b, Ñ1.3 and Ñ4 read pitāmahasute, while Ñ2 reads prajāpatisute. Grünendahl would reprint the reading of Ñ1.3.4, corroborated by T1.2, G4–6 and M, but, in fact, Ñ2 alone preserves the correct reading, attested in K0–4, V1, B1–5, Da, Dn, D1–7 and G1–3. Grünendahl could not know this without conducting a recensio of the entire tradition. In fact, he could not even know that T1.2, G4–6 and M support Ñ1.3.4’s reading. Alternatively, he could print Ñ4’s readings throughout, but Ñ4 does not always preserve the original. At 1.16.3c, it reads abhūt; Ñ1 and Ñ2 have adho bhūmeḥ , while Ñ3 has antarbhūme. All other manuscripts except D5 (which reads bhūmau cāpi) have adho bhūmeḥ . 347 Leonardi, “Il testo come ipotesi,” 8. 348 Ibid., 11. 349 Ibid., 12. 350 Grünendahl evidently does not consider contamination a problem except when it afects other editors. In his edition of the Viṣṇudharma, he notes the “degree of contamination is consid- erable” and “the material does not lend itself to the clear-cut determination of a stemma,” but this does not prevent him from reconstructing a presumed original. Grünendahl, ed., Viṣṇudharmāḥ : Precepts for the Worship of Viṣṇu Part 1, 22 and 3. 351 Leonardi, “Il testo come ipotesi,” 9. 352 Ibid., 7. 353 Ibid., 8. 354 Ibid., 12. 355 See Dembowski, “The ‘French’ Tradition of Textual Philology,” 531: “While the Bedierist critique successfully dealt with the excesses of restorative and interventionist practices in tex- tual criticism—or, more speciically, in the establishment of the ‘critical text’—it also dimin- ished the role of editing in general. In stressing the role of the base manuscript, this critique often ignored the rest of the manuscript tradition. The great successes of the CFMA series in practice conirmed, so to speak, the validity of the one-manuscript method and made any emendation, except that of the most obvious errors, somehow illegitimate. While we are most unlikely ever to return to the practices of a ‘critical’—that is to say, composite and restorative—text, I believe that the ideal of the critical edition itself should be maintained and redeined.” 356 Curiously for someone who commits manifold errors, Grünendahl is a passionate spokes- person for textual criticism. He sees himself as defending philology against those with no understanding of the art. See Grünendahl, “Post-philological Gestures—‘Deconstructing’ Textual Criticism,” 17–28 especially ibid., 27, n. 55: “In my view, a stemma may contribute substantially to what I consider the ultimate goal, viz., Nachvollziehbarkeit of the editorial pro- cess.” Witzel likewise notes that “Grünendahl (2008: 11) rightly stresses that a critical edition has the general goal of the Nachvollziehbarkeit, that is, the replication by readers of the editorial 312 312 PHILOLOGY AND CRITICISM process,” even though his sole edition is a single-manuscript edition based on conjecture. Michael Witzel, “Textual Criticism in Indology and in European Philology during the 19th and 20th Centuries,” Electronic Journal of Vedic Studies 21, no. 3 (2014): 21. 357 Grünedahl’s prior work scarcely inspires conidence in his proposed edition (whose publica- tion we still await). The sole full-length review of Grünendahl’s edition of the Viṣṇudharmāḥ that appeared is critical. “Grünendahl indicates that instead all his energies have gone into the editing of the text; I wish that I could therefore give it a warmer welcome. However, it contains too many discrepancies to inspire conidence. The table of major variants (pp. 20– 21) lists 44 items on which he bases his view of the interrelations of the manuscripts; in a quarter of them the manuscripts listed are at variance with the critical apparatus, while a further two are wrongly numbered (for 4.49cd–51ab and 52ab read 4.48cd–50ab and 51ab). Although his division of the manuscripts into three main groups looks broadly correct, the necessary corrections to his table might modify the picture. His irst two groups contain the bulk of the Newā rī manuscripts, which comprise 10 out of the 15 manuscripts used (a tribute to the success of the Nepal German Manuscript Preservation Project in making accessible such material), while the third contains the Devanā garī manuscripts. N8  ‘stands between Group I and Group II’ but is assigned overall to Group II; however its alignment with Group I in *(33), *(34) and *(36) shown in the table is contradicted by the critical apparatus. These three are among the passages which he regards as inserts by Group I, on which in general he bases his text, despite its presenting a slightly longer text than Group II. However, on three occasions (1.61ed, 7.29ef and 25.31ef) he retains a line found in Group I but absent in Group II and B (the sole Bengali manuscript) where the reduction of a 3-line stanza to a regular one seems preferable both by sense and meter. He regards B as independent, though noting ‘a rather close relation between B and N2’ in subsequent adhyayas, but even in this part of the text the evidence for its alignment at least with Group II is greater than he gives, since *(20) does not occur in B and 7.29ef (cf. above) does, according to the critical appa- ratus but contrary to the table and p. 25. N1, which he recognizes as having a ‘particular ainity with N5’ in addition to its eclectic character, seems in fact to be quite possibly derived from that manuscript with extensive additions from elsewhere. On his own evidence, DL is so derivative from DB that it should in general have been discounted. Finally, he notes the existence of South Indian manuscripts but was unable to utilise them—a potentially serious lacuna. A dependable critical edition of this text would be a real asset. It is regrettable that so many indications of carelessness, if nothing more, call in question the reliability of the present work.” John L. Brockington, review of Viṣṇudharmāḥ : Precepts for the Worship of Viṣṇu, Part 1: Adhyāyas 1–43, by Reinhold Grünendahl, Journal of the Royal Asiatic Society of Great Britain and Ireland 2 (1984): 291–92. Shorter book notices are in J. W. de Jong, review of Viṣṇudharmāḥ . Precepts for the Worship of Viṣṇu. Part 1, Adhyāyas 1–43, by Reinhold Grünendahl, Indo-Iranian Journal 29, no. 1 (1986): 64–65; J. W. de Jong, review of Viṣṇudharmāḥ . Precepts for the Worship of Viṣṇu. Part 2, Adhyāyas 44–81. Part 3, Adhyāyas 82–105 (with a Pāda-Index of Adhyāyas 1–105), by Reinhold Grünendahl, Indo-Iranian Journal 36, no.  2 (1993):  138–39; Hartmut Bobzin, review of Viṣṇudharmāḥ : Precepts for the Worship of Viṣṇu, by Reinhold Grünendahl, Zeitschrift der Deutschen Morgenländischen Gesellschaft 135, no. 1 (1985): 203–4; and Hartmut Bobzin, review of Viṣṇudharmāḥ : Precepts for the Worship of Viṣṇu, by Reinhold Grünendahl, Zeitschrift der Deutschen Morgenländischen Gesellschaft 136, no. 3 (1986): 658–59. 358 For a paradigmatic example of this contrast, see Heinrich von Stietencron, “Editor’s Introduction,” in Angelika Malinar, Rājavidyā: Das königliche Wissen um Herrschaft und Verzicht (Wiesbaden:  Otto Harrassowitz, 1996), 6–7. Stietencron writes:  “The analytical thinking of Western scholars trained in historical and philological methodology stood in contrast to the traditional Indian commentators. The latter not only generously harmonized all the disjunctions in the text but, above all, attempted to recognise in particular passages of the text their own philosophical and theological concepts. This was done in order to secure for them- selves the divine authority of Kṛṣṇa. In this manner, several philosophical schools developed 31 CONFUSIONS REGARDING CLASSIFICATION 313 Gītā interpretations of their own—a spectrum that has been further expanded through politically motivated, modern interpretations since the beginning of the Indian indepen- dence struggle.” 359 Jewish scholars have long recognized the motivations underlying German scholars’ devel- opment of new “critical” methods of biblical scholarship. Writing in 1935, Benno Jacob asked: “Has anybody considered what share in the immense sufering brought recently on mankind and on the Jewish people in particular has to be accredited to the modern German- Protestant science of the Old Testament?” Benno Jacob, The Jewish Quarterly Review 26 (1935– 36):  189, cited in Alan T.  Levenson, The Making of the Modern Jewish Bible:  How Scholars in Germany, Israel, and America Transformed an Ancient Text (Lanham, MD: Rowman & Littleield, 2016), 66. Solomon Schechter is even more radical: “my real sufering began later in life, when I emigrated from Roumania to so-called civilized countries and found there what I might call the Higher anti-Semitism, which burns the soul though it leaves the body unhurt. The genesis of this Higher anti-Semitism is partly […] contemporaneous with the genesis of the so-called Higher criticism of the Bible. Wellhausen’s Prolegomena and History are teeming with aperçus full of venom against Judaism, and you cannot wonder that he was rewarded by one of the highest orders which the Prussian Government had to bestow. Afterwards Harnack entered the arena with his ‘Wesen des Christenthums,’ in which he showed not so much his hatred as his ignorance of Judaism. But this Higher anti-Semitism has now reached its climax when every discovery of recent years is called to bear witness against us and to accuse us of spir- itual larceny.” Solomon Schechter, “Higher Criticism—Higher Anti-Semitism,” in Seminary Address and Other Papers (Cincinnati, OH: Ark Publishing, 1915), 36–37 (Schechter’s italics). We are in full agreement with them as to the specious nature of historical criticism, which in Mahā bhā rata studies as in biblical studies was applied with the sole intent of delegitimizing traditional communities of the book. 360 Grünendahl’s ideas of classiication have their source neither in textual criticism nor in biology, because editors know contamination has the efect of homogenizing the tradition. In contrast, the minor errors introduced into the text each time it is copied are responsible for the diferen- tiation of manuscripts. Rather, these ideas appear to derive from nineteenth-century anthro- pology of race:  his attempted classiication relied on distinguishing the manuscripts into two groups—the pure and the impure. The pure manuscripts constituted a family among them- selves and no further test of belonging was required than to show that a manuscript did not possess the characteristics of the impure group. In contrast, the impure copies played no role in reconstructing the archetype, and hence could be discarded. In his view, the text’s transmission occurred like the handing down of a perfect copy, while its changes were always attributed to an external source. According to him, if not for contamination, the pure tradition would not undergo corruption: its text would descend unchanged from source to apograph and from this source to a further apograph in perpetuity. This was the real reason typological classiication could substitute for genealogical: for him, the better manuscripts so obviously constituted one “family” and the worse ones another, that he considered it unnecessary to prove their iliation. Indeed, the copy was no less pure than its source: it never developed any errors unless contami- nated with an impure source. Once contaminated, it ceased to produce pure copies and set forth only the impure tradition. We encounter a similar notion in a distant ield where genealogical relations were also of concern: ideas of racial purity, as found in the work of Hans F. K. Günther (1891–1968) and later formalized in the Nuremberg Race Laws of 1935. In a manner reminis- cent of Grünendahl’s idea of transmission, these laws held that as long as extrinsic characteristics did not supervene through the much-reviled Rassenschande (racial pollution), the individual would reveal his true origin. The absence of speciic characteristics such as a Jewish nose were con- sidered positive evidence for the individual’s descent, because in parts unafected by contamina- tion this absence permitted the phenotype—the blond Nordic individual—to shine through. Likewise Grünendahl views the absence of certain characteristics as evidence of an original ainity, because, unlike additions, which can be transmitted through contamination, the absence 314 314 PHILOLOGY AND CRITICISM of these characteristics can never be shared with impure specimens. The reference is to Hans F. K. Günther, Kleine Rassenkunde des deutschen Volkes, 3rd edn. (Munich: J. F. Lehmann, 1934). 361 Witzel, “Textual Criticism in Indology,” 34, n.  136. The complete passage reads:  “Note Grünendahl, 1993, on the branches of Mbh. MSS, going beyond the facile division into various script traditions.” Elsewhere he notes:  “This is reminiscent of the branches of Mahā bhā rata MSS, and perhaps due to the same reasons: use of divergent local scripts. See however Grünendahl, 1993, who goes beyond the division into various script traditions.” Ibid., 26, n. 88. 362 “But this still does not explain the general failure to try for a stemma if more than a lim- ited number of MSS were available. (An exception obviously is the—ultimately futile— Winternitz-Sukthankar undertaking of establishing a stemma for Mahā bhā rata MSS early in the 20th century.” Ibid., 55. 363 Ibid.,  17. The reference is to Grünendahl’s “Post-philological Gestures—‘Deconstructing’ Textual Criticism.” 364 Witzel, “Textual Criticism in Indology,” 33 (Witzel’s italics):  “the many original Bardic compositions underlying the Mahā bhā rata, as well as its crystallization, remained unstable and open to additions—especially in the southern tradition. This is quite diferent from saying that the text as a whole was luid from its Bardic beginnings down to Gupta times, and then onward to the testimony of the late medieval MSS. Instead, we clearly have an early crys- tallization and later, local recensions.” And for his defense of critical editions, ibid., 20–25 and 47–50. 365 Georg von Simson, Mahābhārata:  Die Große Erzählung von den Bhāratas (Berlin:  Verlag der Weltreligionen, 2011), 688. 366 “Reinhold Grünendahl, Zur Klassiizierung der Mahābhārata-Handschriften, in: Studien zur Indologie und Buddhismuskunde. Festgabe <. . .> für Heinz Bechert, Indica et Tibetica 22, Bonn 1993, S. 101–130.” Ibid. 367 Walter Slaje, “The Mokṣopā ya Project,” Annals of the Bhandarkar Oriental Research Institute 77, no. 1/4 (1996): 210, n. 7. 368 Johannes Bronkhorst, “Archetypes and Bottlenecks:  Relections on the Textual History of the Mahā bhā rata,” in Pūr vāparaprajñābhinandam: East and West, Past and Present. Indological and Other Essays in Honour of Klaus Karttunen, ed. Bertil Tikkanen and Albion M. Butters, Studia Orientalia 110 (Helsinki: Finnish Oriental Society, 2011), 47. 369 His comments, however, are collected here for convenience in the original: “GRÜNENDAHL und andere haben gezeigt, dass im Bereich der niederen Textkritik im MBh noch nicht das letzte Wort gesprochen ist. Die Erkenntnisse dieser Kritik an der ‘Ersten Kritischen Edition,’ denn man darf nicht vergessen, dass SUKTHANKAR sie ganz bewusst so genannt hat, sollen auch in diese Arbeit einliessen und ihrerseits, wo die Beleglage dies aufdrängt, erweitert werden.” Bigger, Balarāma im Mahābhārata, 15. “Für die Hs. Ñ4 vgl. SUKTHANKAR (1938) und GRÜNENDAHL (1993b).” Ibid., 115, n.  24. “Zur Sonderstellung von D5 vgl. GRÜNENDAHL (1993b), insbesondere S.  116–18.” Ibid., 115, 25. “Will man die Menge der in der Kritischen Edition verwendeten Hss. zu eigentlichen Hss.-Traditionen ordnen, muss man sich über Unterscheidungskriterien Gedanken machen. Dies taten bereits die Herausgeber der Kritischen Edition selbst, allerdings nicht im Hinblick auf die Untersuchung der einzelnen Traditionen, sondern um die Qualität der einzelnen Hss. festzulegen. Das Hauptkriterium ist hierbei dasjenige, das GRÜNENDAHL später die Schriftartenprämisse genannt hat. Sie beinhaltet die Theorie, dass sich Traditionen innerhalb eines bestimmten Schriftgebietes leichter ausbreiten als über die Grenzen dieses Schriftgebietes hinaus. Damit können Hss., die in der gleichen Schrift geschrieben sind, zu einer Gruppe zusammengefasst werden. Die Schriftartenprämisse ersetzt für die Herausgeber der Kritischen Edition das Stemma.” Ibid., 118. “GRÜNENDAHL hat meines Erachtens überzeugend gezeigt, dass die Schriftartenprämisse in dieser Form nicht haltbar ist. Sie lässt sich aber andererseits auch nicht völlig von der Hand weisen. Wohl muss man annehmen, wie GRÜNENDAHL betont, 315 CONFUSIONS REGARDING CLASSIFICATION 315 dass nicht ‘alle in einer bestimmten Schrift vorliegenden Hss. […] auf […] einen Prototyp zurückzuführen wären. […] Meiner Meinung nach lässt sich die Schriftartenprämisse nach wie vor als ein erster Wegweiser im Dschungel der Hss. verwenden, auch wenn die strenge Trennung zwischen den Schriftarten in der Folge einer diferenzierteren Sicht weichen muss.” Ibid., 119. “Ñ4 scheint aufgrund der Schriftartenprämisse nicht zu dieser Gruppe zu passen. Diese hat aber, wie ich in Kapitel 4.2.1 dargelegt habe, mit den Ausführungen GRÜNENDAHLS ihre Unanfechtbarkeit eingebüsst. Ñ4 könnte also mit Ś 1 und K0–3 verwandt sein und mit ihnen zusammen eine Rezension bilden.” Ibid., 121. “So far, we know nothing about the develop- ment of the MBh after the normative redaction; there may even have been later redactions cov- ering only parts of the tradition. To sort their MSS for the edition, SUKTHANKAR and his followers used what GRÜNENDAHL (1993b) called the premise of scripts (Schriftartenprämisse). GRÜNENDAHL showed that this idea, namely that all MSS written in a certain script belong together, is not tenable, but it may still be useful as a irst criterion.” Ibid., 164. 370 Oskar von Hinüber, “Der vernachlässigte Wortlaut. Die Problematik der Herausgabe buddhistischer Sanskrit-Texte, Zur Überlieferung, Kritik und Edition alter und neuerer Text,” in Kleine Schriften, Teil I, ed. Harry Falk and Walter Slaje (Wiesbaden: Harrassowitz, 2008), 435. 371 Reinhold Grünendahl, “Zur Textkritik des Nā rā yaṇīya,” in Nārāyaṇıȳ a-Studien, ed. Peter Schreiner, Purā ṇa Research Publications Tübingen 6 (Wiesbaden:  Harrassowitz, 1997), 30–74. 372 Von Hinüber, “Der vernachlässigte Wortlaut,” 435, n. 19 (the references in square brackets are Hinüber’s and refer to footnotes in his text). 373 Ibid., 436. 374 Ibid., 435. 375 Oskar von Hinüber, review of The Sanskrit Epics, by John Brockington, Wiener Zeitschrift für die Kunde Südasiens 46 (2002): 268. 376 James L.  Fitzgerald, “Mahā bhā rata,” in Brill’s Encyclopedia of Hinduism vol. 2, ed. Knut A. Jacobsen, Helene Basu, Angelika Malinar and Vasudha Narayanan (Leiden: E. J. Brill, 2010); Brill Online, 2015; http://referenceworks.brillonline.com/entries/brill-s-encyclopedia-of- hinduism/mahabharata-BEHCOM_2020040 (accessed September 23, 2015). 377 Ibid. 378 James L. Fitzgerald, “General Introduction:  The Translation Resumed,” in James L. Fitzgerald, trans., The Mahābhārata, vol. 7:11. The Book of the Women; 12. The Book of Peace, Part One (Chicago:  University of Chicago Press, 2004), xvi. This paragraph also occurs verbatim (with an inserted “lectiones diiciliores” in parentheses after “diicult readings”) in James L. Fitzgerald, “Making Yudhiṣṭhira the King: The Dialectics and Politics of Violence in the Mahābhārata,” Rocznik Orientalistyczny LIV, no. 1 (2001): 64, n. 17. 379 Sheldon Pollock, “Sanskrit Literary Culture from the Inside Out,” in Literary Cultures in History:  Reconstructions from South Asia, ed. Sheldon Pollock (Berkeley and Los Angeles and London: University of California Press, 2003), 110–11. 380 Sheldon Pollock, The Language of the Gods in the World of Men:  Sanskrit, Culture, and Power in Premodern India (Berkeley and Los Angeles: University of California Press, 2006), 229. 381 Pollock, “Sanskrit Literary Culture from the Inside Out,” 111, n. 155. 382 Pollock, The Language of the Gods in the World of Men, 229, n. 11. 383 See Sheldon Pollock, “Towards a Political Philology: D. D. Kosambi and Sanskrit,” Economic and Political Weekly, D.  D. Kosambi Centenary Volume (2008):  52–59; Sheldon Pollock, “Future Philology? The Fate of a Soft Science in a Hard World,” Critical Inquiry 35, no. 4, The Fate of Disciplines (2009): 931–61 and Sheldon Pollock, “Kritische Philologie,” trans. Brigitte Schöning, Geschichte der Germanistik 45/46 (2014):  5–12. In “Kritische Philologie,” the term is deined as follows: “Philology becomes critical when it grasps its own historicity, constructedness, and changeability; when it understands that it is not and cannot be a local form of knowledge that passes as universal under the mask of science, but must be part 316 316 PHILOLOGY AND CRITICISM of a global—and, by preference, a globally comparative—discipline, seeking global—and by preference, globally comparative—knowledge; when it realizes that understanding by what means and according to what criteria thinkers in past eras have grounded their truth claims must be part of our own understanding of what truth is, and a key dimension of our knowledge politics.” Pollock, “Kritische Philologie,” 5 (the English is from Pollock’s original English text uploaded at: www.columbia.edu/cu/mesaas/faculty/directory/pollock_pub/ Kritische%20Philologie%20(+%20English%20original).pdf, accessed February 26, 2017). See also ibid., 5–6 (Pollock’s italics): “In 2004 while still a professor at The University of Chicago I organized a workshop on ‘critical philology,’ inventing the term—as I thought I had done—as a kind of provocative tautology: if philology has from its origins been con- ceived of as a critical practice—in the dictionary sense of an objective analysis and evalua- tion of something in order to form a judgment—it had ceased to be critical in the relexive way I felt to be essential to its continued viability. It had abandoned its large and ambitious theoretical aspirations, indeed, its leadership among the human sciences, and had become a completely routinized, self-complacent, and—the most deadly sin of all, in America at least—boring. Its celebrated rigor had turned into rigor mortis. What was needed was a new philology that really was critical.” And for the related “critical classicism” see Sheldon Pollock, “Crisis in the Classics,” Social Research: An International Quarterly, special issue titled India’s World 78, no. 1 (2011): 21–48. 384 “If philology is the discipline of making sense of texts, and making sense of texts is making sense of life, what does making sense of texts actually consist of ? Answering this question would seem to call for some heavyweight philological theory, but what I want to ofer instead is rather lightweight autobiography: a relection on how I myself have come to reconcile what for me were seriously conlicting modes of interpretation. I was trained to a very hard histor- icism, but also, as a Sanskritist, heir to a brilliant tradition of reception with its own strong claims to knowledge. Over time, I  have also been tempered by a critical hermeneutics of understanding and a neopragmatist conception of truth. This autobiography has prompted me to conceive of a philological practice that orients itself simultaneously along three planes of a text’s existence: its moment of genesis; its reception over time; and its presence to my own subjectivity.” Sheldon Pollock, “Philology in Three Dimensions,” Postmedieval: A Journal of Medieval Cultural Studies 5, no. 4 (2014): 399. And see ibid., 409: “Learning to read in three dimensions, which is the autobiography of my own philology, is learning to practice a deli- cate balancing act that requires both training and untraining. The act is especially diicult for hardcore historicists such as myself, but real historicism requires not only consistency but history.” 385 Sheldon Pollock, “What Was Philology in Sanskrit?” in World Philology, ed. Sheldon Pollock, Benjamin A. Elman and Ku-ming Kevin Chang (Cambridge, MA: Harvard University Press, 2015), 136. 386 Not to mention historical self-awareness. For Adluri’s criticisms, see Vishwa Adluri, review of World Philology, ed. Sheldon Pollock, Benjamin A.  Elman and Ku-ming Kevin Chang, American Historical Review 121, no. 3 (2016): 908–10. 387 Pollock, “What Was Philology in Sanskrit?” 136. 388 Alf Hiltebeitel, “The Nārāyaṇı ̄ya and the Early Reading Communities of the Mahābhārata,” in Between the Empires: Society in India 300 BCE to 400 CE, ed. Patrick Olivelle (New York: Oxford University Press, 2006), 231–32. 389 John L. Brockington, The Sanskrit Epics (Leiden: E. J. Brill, 1998), 58 (italics added). 390 Besides Madeleine Biardeau, whose work is cited later, some of Hiltebeitel’s early studies— for example, the article “Two Kṛṣṇas on One Chariot:  Upaniṣadic Imagery and Epic Mythology,” History of Religions 24, no.  1 (1984):  1–26—are insightful. In contrast, where he seeks to assimilate himself to the German historical paradigm—quoting, for example, Grünendahl, Bigger, Bronkhorst and Witzel—his work goes astray. This attempted assimi- lation is in striking contrast to his criticisms of German Mahā bhā rata studies in “Kṛṣṇa in 317 CONFUSIONS REGARDING CLASSIFICATION 317 the Mahābhārata: A Bibliographical Essay,” Annals of the Bhandarkar Oriental Research Institute 60 (1979): 65–110. 391 Nearly three-quarters of a century ago, V. S. Sukthankar prophetically observed: “It would be a rather hazardous conjecture to suppose that such a thing might perchance happen also to the works of the critics of the Mahābhārata, for within less than half a century the lucubrations of these wiseacres have approached perilously near the limbo of oblivion, from which they are periodically snatched out by the industrious pedagogue and the curious antiquarian.” V. S. Sukthankar, On the Meaning of the Mahābhārata (Bombay: Asiatic Society, 1957), 29–30 (Sukthankar’s italics). 392 “The learned philologist of the present day feels a deal of hesitation in answering this question, which to the unsophisticated Indian would present no diiculty whatsoever. If questioned, the latter will no doubt promptly and conidently answer that the Mahābhārata is a divine work recounting the war-like deeds of his ancestors, the god-like heroes of a past age, the unrigh- teous Kauravas on the one hand and the righteous Pā ṇḍavas aided by Lord Ś rī Kṛṣṇa on the other,—of the Golden Age when gods used to mingle with men, when the people were much better of, much happier, than they are today. And the illiterate Indian is right, to a very large extent, as he far more often is than his ‘educated’ brother. For, the Mahābhārata, as the poem itself tells us, arises out of the following question of Janamejaya addressed to the great Ṛṣi Kṛṣṇa Dvaipā yana Vyā sa on the occasion of the snake sacriice (1.54.19): kathaṃ samabhavad bhedas teṣam ̄ akliṣtạ kāriṇam ̄ | tac ca yuddhaṃ kathaṃ vṛttam bhūtāntakaraṇaṃ mahat || ‘How arose the quarrel among those men of unblemished deeds? How occurred that great war which was the cause of the destruction of so many beings?’ ” Ibid., 32–33. 393 See Sukthankar, “Prolegomena,” lxv: “We accordingly ind that he [Nīlakanṭ̣ ha] occasionally mentions (in about 125 places) variant readings and additional passages found in diferent pro- vincial versions (most of which can be identiied among the readings of the manuscripts com- prising our critical apparatus), and cites (as a rule, without naming the source) the explanations given by other scholiasts—information, scanty though it is, yet of immense interest and value for the history of the received text.” In his article on the Mahā bhā rata commentators, Sukthankar emphasizes: “The study of these commentaries must be now taken up more seri- ously, not so much for the sake of the explanations contained in the commentaries—though even the glosses of a commentator like Devabodha are extremely important—as for the readings and pā ṭhā ntaras recorded in them; because, most of the commentaries are older— some very much older—than our manuscripts; and therefore the documentation of these readings by the commentators takes us back a stage further in our investigation of the history of the epic.” Vishnu S. Sukthankar, “Epic Studies V: Notes on Mahā bhā rata Commentators,” Annals of the Bhandarkar Oriental Research Institute 17, no. 2 (1935–36): 185–86. 394 Madeleine Biardeau, Hinduism:  The Anthropology of a Civilization, trans. Richard Nice (Delhi: Oxford University Press, 1989), 170, n. 1. 395 Biardeau’s irst work, Théorie de la connaissance et philosophie de la parole dans le brahmanisme classique (The Hague: Mouton, 1964), was the outcome of her studies with Brahmans in Pune. It is testimony to her assimilation of their thought systems and commentary. For her applica- tion of these insights to the Mahā bhā rata, see the volumes Madeleine Biardeau, Études de mythologie hindoue I: Cosmogonies purāṇiques (Paris: École française d’Extrême-Orient, 1981) and Madeleine Biardeau, Études de mythologie hindoue II: Bhakti et avatāra (Pondichéry: Publications de l’École Française d’Extrême Orient, 1994). 318 The prevalent system of classiication: Hans F. K. Günther’s identiication of individuals of the “Nordic” type—in clockwise order, from Mecklenburg-Strelitz, Niedersachsen and Munich (front-facing and proile). Source: Reproduced from Günther, Kleine Rassenkunde des deutschen Volkes, 3rd edn., 23. 319 CONCLUSION: TEXTUAL CRITICISM AND INDOLOGY There are no new or old philologies, only good ones, based on long study, and mediocre, hasty ones, not supported by adequate training. —Paolo Trovato, Everything You Always Wanted to Know about Lachmann’s Method This now completes the second half of our two-part argument that, considered as an aca- demic discipline, German Indology fails to meet even basic canons of objectivity, truth and method.1 In the irst half, we argued that although Indologists claim a superior critical con- sciousness and decry all other approaches as nonscientiic, their work shares little in common with textual criticism.2 Their work cannot claim the title of philology at all; it is rather a mixture of dogmatic assertion and Protestant anti-traditional, anticlerical sentiments masquerading as rigorous textual scholarship. Its true antecedents, we argued, lay not in the work of editors and textual critics such as Karl Lachmann, Hermann Kantorowicz, Paul Maas and Eduard Fraenkel, but rather in the work of eighteenth-century Protestant theologians Johann Salomo Semler, Georg Lorenz Baur and Ferdinand Christian Baur.3 As we analyzed in our irst book, The Nay Science, it is this combination of a noncrit- ical polemical method with Romantic fantasies of the Aryans that largely explains why these scholars to this day tend to reject stemmatic methods in favor of a priori theories of the Mahā bhā rata textual transmission.4 It also explains why scholars from Edward W. Hopkins to Georg von Simson and Andreas Bigger thought to reconstruct the earliest epic using circular methods such as identifying supposed Kṣatriya passages or themes in the epic or a range of equally circular criteria such as the presumed style and meter of the earliest epic.5 But the claim that Indologists as a rule had little understanding of textual criticism could, at this stage of the argument, only be advanced as a promissory note and not fulilled within the conines of the irst book. Hence the need for this book. The present volume now fully redeems that claim: by tracing how Bigger and Reinhold Grünendahl, though they constantly belabored the language of “textual criticism,” had only the barest understanding of the principles involved, we have exposed Indology’s claims to being a science and, above all, to being a rigorously textual science. Not only Bigger and Grünendahl but also, by extension, all of the scholars who cited their work proved incapable of discriminating between rigorously stemmatic arguments and those that only had the appearance of being critical. Yet even today, whenever Indologists look to make a case for their discipline, they still invoke the comparison with classical philology.6 Grünendahl, for instance, was not even clear about the fact that if the editorial process aims at the reconstruction of the best and 320 320 PHILOLOGY AND CRITICISM most archaic text, then the true basis of classiication of a manuscript must be its text, and yet he saw himself in the role of having to defend philology against its “post-philological” critics.7 Yet so far, Indologists have succeeded in delecting criticisms of their discipline by portraying the critics as little more than amateurs, outsiders to the discipline motivated by resentment and envy to bring down a science that exceeds their comprehension.8 But if we set aside this narrative of initiates into and adepts of a highly evolved canon of method, on one hand, and parvenu intellectuals, on the other, we gain a new perspec- tive on the history of the discipline.9 We begin to see that the problem is not the contrast between an intricate method and its isolated and insuiciently appreciated practitioners but that the idea of method has itself become a hurdle to the examination and audit of the discipline. What are these methods? What is their history? Do they appear fully formed, or must they constantly be subjected to calibration? Who authored these methods and why? Whose interests do they serve? Every time Indologists have been called upon to legitimate their discipline, they have invoked the comparison with classical philology, even though, as we have seen, as a rule they had only the haziest notions of philology. Michael Witzel, for instance, lets the story of Indology begin with Richard Bentley (1662–1742) and F. A. Wolf (1759– 1824), though the link with Indology is tenuous at best. He writes that “diferent from his contemporaries Rousseau (1712–1778) and Herder (1774–1803), Wolf deined classical philology as the ‘knowledge of human nature as exhibited in antiquity,’ that was to be studied in its writings, art, and other forms of ‘national thought,’ ”10 but does not explain the relevance to Indology.11 He writes that “his [F. A. Wolf ’s] method inspired the crit- ical approach of the 19th century in analyzing ancient texts. He is often regarded as the founder of modern philology,” but the implied argument, namely, that Indology belongs within the tradition of “modern philology” inaugurated by Wolf, only draws force from the tacit assumption of a contrast between “modern” modes of inquiry and traditional ones.12 Citing an unnamed author, he notes that “ ‘philology is a Kulturwissenschaft based on texts,’ the study of a civilization based on its texts,” but does not speciically clarify wherein the philological nature of this study might lie or how a philological preoccupa- tion with texts might difer from other kinds of preoccupations with texts.13 He writes that “Philological study comprehends both the investigation of the available written and oral texts of a civilization and employs a range of tools (Hilfswissenschaften) necessary for understanding these texts; these tools deal with the realia met with in the texts and they range from archaeology to writing systems, and from astronomy to zoology,”14 but this only tells us that he understands philology as the historical study of past epochs and does not clarify wherein the speciically philological nature of this study might lie.15 As with Grünendahl, the wide-ranging focus of Witzel’s article (within the space of 80 pages he discourses, in sequence, on: “Textual criticism”; “Renaissance and Classicism”; “A new stimulus:  comparative linguistics”; “Philology”; “Manuscripts”; “Stemmatic method”; “Lectio diicilior”; “Contamination and its neglect”; “Stemma and arche- type outside classical European texts”; “ ‘Discovery’ of India:  Sanskrit languages and the European languages”; “Commentaries”; “Epic texts”; “Other approaches to (non-) European texts”; “Extra-Classical use of textual criticism”; “Ethnology/Anthropology”; “Mutual inluences”; “Critique of the stemmatic method versus lack of critical editions 321 CONCLUSION 321 in Indology”; “Criticism”; “Archetype”; “Abundance of Indian MSS”; “Continuing use- fulness of the stemmatic method”; “Recent inds of 2000-year-old MSS”; “The way forward in Indian textual studies:  critical editions”; “Commentaries (again), variants and editing”; “Types of editions, their critique”; “Oral and written traditions”; “History of writing”; “The scribes”; “Commentary and other testimonials”; “Oral and musical traditions”; “Other necessary prerequisites for editing and interpreting texts”; “Computer editing and stemma generation”; “Methods”; “Precursors”; “ ‘New Stemmatics’ ”; and “Prospects”—all of the sections are general; not one of them would convince a spe- cialist) is meant to detract attention from the fact that he knows almost nothing of textual criticism. Thus, Witzel repeatedly confounds historical criticism with textual criticism, arguing that “the early 19th century saw the establishment of both the modern method of histor- ical and textual criticism and of the development of historical comparative linguistics,”16 even though historical criticism and textual criticism have quite distinct genealogies and frequently pursue opposing aims. He writes that “The stemmatic method was largely developed by Lachmann. He developed the principle of recensio, that is the study of the received MSS of a text [?], the establishment of their ‘family tree’ (stemma), followed by the edition of the text,”17 even though Lachmann never drew up a stemma and was not the irst to appreciate the merits of conducting a systematic recensio.18 Witzel claims that “available MSS usually fall into a number of families, whose ancestor MS may no longer exist—biologists would call this a ‘bottleneck’ event,”19 even though the constriction in the tradition is only apparent and not real, being a consequence of the fact that the sur- viving copies of the text are descended from a single source and not because there was a real, observable reduction in the number of copies circulating at a certain moment in time. We already saw in the introduction the importance of not confusing the arche- type in the sense of an oicial or especially signiicant text (sometimes taken for the sole exemplar in existence) with a Neo-Lachmannian archetype. As Trovato notes, the archetype in this second sense refers to a manuscript that, “ ‘lacking any “oicial status” or normativity’ (Timpanaro), is [...] ‘randomly’ identiied by philologists while classifying witnesses.”20 Witzel, however, thinks the constriction is real, for he writes:  “Curiously, the parallel development in Indian MSS around 1000 CE has not even been noticed by scholars.”21 Now it would be very curious indeed if there were to be a similar reduction in the number of copies of any given manuscript circulating in India at a certain moment in time (we might wonder: what is this strange event that compels the reduction of all man- uscript traditions—east or west—at a certain moment in their history to one single exem- plar?), but, fortunately, the constriction here, as there, is only apparent.22 Only someone unfamiliar with the principles of textual criticism would think that because all our copies are descended from a single source, this means there really must have been only one source, yet Witzel’s work is rife with examples of this error.23 In general, Witzel seems to associate “textual criticism” with the genealogical-recon- structive method, but appears not to have quite grasped the principle of how editors proceed from creating a list of variants to establishing some of those variants as errors. In his account of how editors reconstruct the reading of the archetype, he notes: “In other words: irst comes noticing a mistake, and perhaps an inkling of what might be 32 322 PHILOLOGY AND CRITICISM the solution, all of which is then tested using the various tools and Hilfswissenschaften that we always use in philology. In this fashion, we can prove our point, or at least, make it a probable one. More on this, later.”24 It is clear from this account that, in Witzel’s model of textual criticism, there is no space for “a true, continuous, and systematic recension,”25 much less for genealogical analysis of the relations of iliation between manuscripts.26 His procedure is not stemmatic at all, since it relies on the editor “noticing” a mistake, which he immediately proposes redressing by ofering a conjecture.27 For instance, noting that Paippalā da Saṃhitā 8.12.9 reads “idaṃ kuru cemāṃ surām” in Raghu Vira’s edition, but “idaṃ kodacemāṃ surā” in D. Bhattacharya’s edition and that “another Orissa variant (MS mā) has:  idaṃ koda-dacemā,” Witzel proposes that “this is restored, with minimal emendation to:  +udaṃkodañcemāṃ (udank̇ a/udañca/imām)̣ surām ‘o ladle, scoop up the brandy.’ ”28 As is clear from this example, his method is not actually stemmatic and he is not interested in deriving any one of the readings from the other. This makes his comment on the following page—“We can thus approach a solution by using the stemmatic method combined with paleography and other Hilfswissenschaften, and if we can show how the reading of a certain MS has come about (Textgeschichte)”29—even more puzzling. In fact, it can be shown of Witzel’s critical edition of the Kaṭha Ā ranỵ aka that it is not a critical edition at all.30 For the major part of the text only one manuscript was available; the critical apparatus is inadequate; the principles followed in the reconstruction are unclear; much of the text has been “restored according to parallel passages, or even freely so”;31 emendations are not clearly signaled;32 and, most important, the reconstruction is not stem- matic. He writes that “next to the Tübingen Ṛcaka (T) that mainly contains KaṭhĀ I and III, the following manuscripts were available for text constitution, though they only contain parts of the mantra section (KaṭhĀ I). Ms. Tübingen MaI 396, fols. 274a–284b (t1), a part of the Ṛcaka (T) just mentioned and unfortunately damaged in almost the same fashion, accented (but not in all Mantras)—Paris, National Library (Bibliothèque Nationale), Ms. dév. 230, p. 304–317 (P), written in modern Ś ā radā on paper, no accents, dated (17)80 C. E. (or perhaps 1880)—Paris, Ms. dév. 227b, fols. 10 sqq. (P2), written in younger Ś ā radā script, on rough ‘country paper,’ accent rests—edition of the Laugā ksị Gṛhyasū tra (L) by M. S. Kaul, vol. II, p. 52–91 (unfortunately without variant readings, but with Devapā la’s commentary on the Mantras),”33 but explains neither according to which principle he has combined the readings of his sources or how he has dealt with variants, if any.34 Perhaps he simply means he has used the secondary sources to ill in the lacunae in his primary source. Likewise, he describes the method of reconstruction as follows: “Text in brackets [] has been restored according to parallel passages, or even freely so. However, restoration fre- quently results with remarkable certainty, due to the ‘inevitability’ of Brā hmanạ style. For the irst lines of a page of the manuscript that have mostly been broken of, […] has always been inserted in order to indicate the amount of missing text.”35 In the absence of a stemma, however, the reader has no way of knowing what stage of the tradition Witzel is reconstructing.36 Without identifying a pattern of conjunctive and separative errors between the manuscript sources that, as we saw, is key to establishing iliation, Witzel has no evidence his sources are related. Strictly speaking, each is a codex unicus. There is no evidence the texts were ever handed down as a whole. Given the 32 CONCLUSION 323 sensitivity of Vedic texts to even slight changes of wording or accent, the logical solution would have been to produce a scrupulous diplomatic edition, preserving all the features of the text. Even if Witzel imagines himself as a paleographer dealing with a lost script rather than a textual critic, his reconstruction remains purely hypothetical.37 In view of the emerging consensus among textual critics that the reconstructive edition should respect the integrity of a single manuscript, the idea that the text can be restored “even freely so” is especially problematic.38 In essence, all Witzel has done is use two partial manuscripts (Paris, Bibliothèque Nationale, Ms. dév. 230 and 227b) and one edition of a related work (M. S. Kaul’s edition of the Laugā kṣi Gṛhyasū tra) to attempt to restore the text to a readable state. While this is a laudable intention and Witzel may feel conident in being able to do so in view of what he, borrowing an expression from Karl Hofmann, considers the “ ‘inevitability’ of Brā hmaṇa style,”39 the fact is that a reconstructive edition that combines the available manuscripts, merely on the principle that without superimposing the texts over each other it is impossible to obtain a continuous text, is not a critical edition. Witzel has no corroboration for the authenticity of the text that results. Neither is the resultant text an accurate representation of the text’s diachronic evolution. His edition is, at best, a com- posite edition and, at worst, an attempt to pass of as a critical edition a text that is merely the transcription of a single manuscript with a mix of conjectures carried out with only occasional recourse to the manuscripts.40 This vacillation between a strictly stemmatic procedure and a relapse into conjectural criticism is not accidental. As we saw, it was also a feature of Bigger’s and Grünendahl’s work. Both scholars insisted that they were interested in “textual criticism,” even though what they ultimately ended up producing were reconstructions based on subjective and circular theories of the originally oral nature of the Mahā bhā rata. Likewise, Witzel, even though he repeatedly emphasizes the need for critical editions,41 proves incapable of producing a true critical edition, that is to say, a mechanical reconstruction of the archetype based on the principles of a systematic recensio and a genealogical analysis of the relations of iliation between manuscripts. Thus, under the cover of calling for more critical editions, what he is arguing for is much greater scope for emendation,42 since, like Otto Böhtlingk—whose work Patrick Olivelle has severely criticized43—his real interest is in proposing bold conjectures.44 In contrast, Witzel’s statements about the Mahā bhā rata critical edition, especially as regards the impossibility of producing a critical edition due to the widespread presence of contamination,45 are simply untrue: he appears to have only the haziest understanding of the Mahā bhā rata tradition, as evinced by his claim that “the many original Bardic compositions underlying the Mahā bhā rata, as well as its crystallization [?], remained unstable and open to additions—especially in the southern tradition.” 46 But how does Witzel know this? The only way he could know of the existence of “original bardic compositions” that “underlay” the Mahā bhā rata is if he were alive at the time of those compositions; otherwise, the assumption of an oral epic, as we have seen, only leads to speculative theories of the epic. Likewise, his claim that “this is quite diferent from saying that the text as a whole was luid from its Bardic beginnings down to Gupta times, and then onward to the testimony of the late medieval MSS. Instead, we clearly have an 324 324 PHILOLOGY AND CRITICISM early crystallization and later, local recensions”47 revives aspects of Grünendahl’s notion of independent recensions of the Mahā bhā rata, without, however, being able to prove this theory any more than Grünendahl could. We ind it surprising that Witzel should feel compelled to lecture Indians on the need for critical scholarship when he himself has only the haziest understanding of textual criticism.48 Witzel’s statements about the Mahā bhā rata’s critical edition illustrate just how problematic it is to unthinkingly accept “critical” views about the Mahā bhā rata and, by extension, about Indian texts in general. It is one of the deining characteristics of German Indology that its practitioners insist that the production of critical editions is the unique distinction of their tradition, while at the same time attempting to weaken the deinition of a critical edition.49 In this contrast between what the Indologists say it is that they do and what they actually do we see the central contradiction at the heart of the discipline: on one hand, in order to be recognized as a legitimate discipline within the university canon, they were forced to constantly seek the comparison with classical philology, the discipline that had most successfully mastered the transition from an indef- inite literary enterprise to a discipline modeled on the natural sciences and their rigorous procedures;50 on the other, the depth of expertise available in the ield was always scant as compared with their colleagues in classical philology. Ever since the nineteenth century, when Indology irst constituted itself as a discipline at the university by evoking the parallel with classical philology, Indologists have made a proitable living by claiming that what they specialized in was producing critical editions of Sanskrit texts.51 Yet, as we saw in this book, these claims are not entirely accurate. Not one of the Indologists whose work we examined had produced a true critical edition,52 yet all of them claimed that they practiced textual criticism. Given the inlation in the use of the expression critical edition in Indology, it appears appropriate to institute some criteria for its use.53 We propose the following deinition:  “Only those editions should be permitted to call themselves critical as make use of the genealogical-reconstructive method (also known as the common-error method) to reconstruct the relations of iliation between manuscripts and that propose a reconstruction of the archetype of the tradition on the basis of an explicit stemma.” This reconstruction, moreover, must be mechanical in the sense that it must be apparent, from a glance at the apparatus, what stage of the tradition the editor is reconstructing at any given moment. Further, no edition should be permitted to call itself a critical edition unless it is based on a systematic recensio of a large number of manuscripts (this would a priori exclude such one-manuscript “critical” editions as Witzel’s edition of the Kaṭha Ā raṇyaka).54 We are aware that in many cases this ideal will not be attainable, but if this reduces the number of critical editions of Indian texts in circulation, so much the better. Notes 1 As clariied in the introduction, it bears repeating that German in “German Indology” does not refer to national origins: it merely speciies a certain style of doing Indian studies. We could also have just used “Indology,” but this would be to subsume other traditions of Indian studies, perhaps unfairly, under this term. For this reason, we have chosen to retain the epithet German, well aware that it may result in misunderstanding and misrepresentation of our views. 325 CONCLUSION 325 2 See Vishwa Adluri and Joydeep Bagchee, The Nay Science: A History of German Indology (New York: Oxford University Press, 2014), especially the introduction and chapters 3–5. 3 See ibid., 11–19 (“The Origins of the Historical-Critical Method in the Neo-Protestantism of the Eighteenth Century”). 4 If one truly wished to understand the arguments for identifying various historical strata on the basis of considerations such as “redactional history” or “redactional ideology,” one would have to begin with biblical critics, whose work represents the closest parallel to contempo- rary Mahā bhā rata studies. Via Heinrich Ewald’s student, Veda scholar Rudolf von Roth, their method then entered German Indology. The earliest scholar to adopt the methods and termi- nology of biblical criticism in Mahā bhā rata studies was Christian Lassen. Mahā bhā rata studies really only came into its own as a discipline with the uncle–nephew duo of Adolf Holtzmann Sr. and Adolf Holtzmann Jr., who fused the methods of biblical criticism with nascent ideas of an Indo-Germanic/Aryan heritage to create modern Mahā bhā rata studies. We should also not overlook the role of the newly discovered science of geology, which, through its extension to ethnology, anthropology and, ultimately, history, signiicantly inluenced Indian historical writing, as Guha has shown. See Sumit Guha, “Lower Strata, Older Races, and Aboriginal Peoples:  Racial Anthropology and Mythical History Past and Present,” The Journal of Asian Studies 57, no. 2 (1998): 423–41. Guha does not discuss the Mahā bhā rata, but one vital link in the chain of “the mythic history of clashing races” (ibid., 438) is clearly Lassen’s pseudo-ethno- logical research based on the Mahā bhā rata as discussed in Adluri and Bagchee, The Nay Science, 41–48 and 124. 5 The paradigmatic author for this style of “criticism” is Holtzmann Jr. Although Lassen had already argued for the existence of a shorter, more ancient Kṣatriya epic, Holtzmann Jr. irst formalized the “criteria” for recovering this older epic. The following passage is revelatory of his concerns: “Here now, in the irst place,” writes Holtzmann, “the thoroughly warlike worldview is to be highlighted which constitutes the genuine soul of the old portions of the epic. […] Instead of the elegiac softness, the resignation, being tired of life, of later Indian literature, the raw warrior-like air of the old Germanic North blows against us here. If we were ever to succeed in determining the oldest cultural phase of the Indian race accessible to research and to dissolve away almost by means of a chemical process all inluences of the Brahmanism that is already slowly developing […] we would ind conditions before us only a little diferent from those described by Tacitus as unique to the ancient Germans. But even in its contemporary ruined form the Mahābhārata often delivers us the best commen- tary on Germania.” Adolf Holtzmann Jr., Zur Geschichte und Kritik des Mahābhārata (Kiel: C. F. Haeseler, 1892), 45–46 (emphasis in original). Compare Bigger’s claim that verse 13.135.81 from the Viṣṇusahasranā mastotram of the Mahā bhā rata might go back to an earlier stage of the epic’s history, that is, to a time when the names Aniruddha and Pradyumna were still identiied with Kṛṣṇa’s son and grandson rather than Kṛṣṇa’s vyūhas or hypostases, as the two are addressed here with the titles apratiratha (“peerless in chariot warfare” or “one possessing a matchless body”) and amitavikrama (“of limitless power”). “These are,” he continues, “in my eyes, statements such as one might make about Kṣatriyas.” Andreas Bigger, Balarāma im Mahābhārata: Seine Darstellung im Rahmen des Textes und seiner Entwicklung (Wiesbaden: Otto Harrassowitz, 1998), 64. 6 Most recently, Franco, in his review of our irst book, when he claims:  “The nature and origin of ‘Indology’ were already clearly stated in A. W. Schlegel’s founding essay, ‘Über den gegenwärtigen Zustand der Indischen Philologie,’ which marks the beginning of ‘German Indology’ as an academic discipline: ‘If the study of Indian literature is to thrive, the principles of classical philology have to be thoroughly applied to it, and that with the most scientiic rigor.’ [...] This opinion was still widespread and taken for granted when I irst came to Germany in the early 1980s. It lasted as long as classical philology itself was able to maintain its prestige, until the repeated waves of neo-liberalism, secondary school and university reform, and the cultural turn in the humanities marginalised it, and with that Indology as well. In other words, 326 326 PHILOLOGY AND CRITICISM ‘German Indology’ is not, at its core, a ‘nay’ science; rather, ‘German Indologists’ wanted to accomplish for India what their fellow philologists had accomplished for Ancient Greece and Rome—and presumably, some of them still have this aim. Looking back at what they have accomplished over the last 200 years, they have not done such a poor job.” Eli Franco, review of The Nay Science: A History of German Indology, by Vishwa Adluri and Joydeep Bagchee, South Asia: Journal of South Asian Studies 39, no. 3 (2016): 697–98. 7 See Reinhold Grünendahl, “Post-philological Gestures—‘Deconstructing’ Textual Criticism,” Wiener Zeitschrift für die Kunde Südasiens 52–53, Text Genealogy, Textual Criticism and Editorial Technique (2009–10): 17: “ ‘Post-philological’ sounds like just another of those ‘sign-post’ neologisms devised to mark a supposedly radical break with something, in this case philology”; ibid., 21: “To be sure, none of these post-philological arguments are really to the point, but that is as close as the critical theorist gets to the real thing. The basic misunder- standing that usually prevents an adequate assessment is the notion that textual criticism is about extracting a certain meaning from a text, while textual critics tend to follow the rule that questions concerning the meaning of a given text had better be postponed until its wording has been established as accurately as its textual tradition allows” (Grünendahl’s italics); and ibid., 22: “I see this caricature of textual criticism in terms of mechanistic pedantry as an attempt to discredit its very foundation, namely, what is called ‘Akribie’ in German, from Greek akríbeia, originally the accuracy of a craftsman in adjusting two workpieces to one another, and as such a itting term for the textual critic’s ambition, typically psychologized by Cerquiglini as ‘hysteria of detail.’ ” But Grünendahl could have spared himself the ref- erence to the German Akribie (meticulousness): other languages have similar words, and it is not a good idea to insist on the speciically German nature of this quality when one has pro- duced a work as appallingly full of errors as his 1993 article. We are told that “Frank Kolb has recently pointed out the importance of Akribie as a prerequisite of Geisteswissenschaften in general [...],” but “the ‘Newspeak’ of German education politics shows diferent tendencies in its efort to market its latest invention, the mass-produced Generalist. It may be interesting to see how Akribie will fare in the frantic deconstruction of Geisteswissenschaften according to the political dictate of ‘international standards,’ compounded by economization and the impact of social sciences.” Ibid., 22, n. 36 (Grünendahl’s italics). Only, the Akribie Grünendahl claims is the hallmark of German scholarship manifested in his case rather as an Akribie in generating errors. 8 Grünendahl argues that the “very reality” of textual criticism “is called into question by little more than lavish use of inverted commas and words to that efect, such as ‘construct,’ imaginaire, etc.” and that “consequently, the critical theorist can hardly be expected to enter into a detailed discussion of the actual ‘philological handiwork of critical editions, etymology, historical grammar’ and the like,” but without having irst shown that he either understands the “philological handwork” of critical editions or knows anything of etymology or his- torical grammar. Ibid., 24. In fact, he does not, as his discussion of Sukthankar’s editorial choices shows. He describes “textual criticism” as “just […] a tool” and asserts that “it is the faculty of handling texts according to reasonable principles,” but then commits the error of claiming that, as two manuscripts are lacking the same interpolations, one of them must have “contaminated” the other. Ibid. He argues that “the formulaic performance of the accusatory epic [sc. of the ‘post-philological’ critic] is directed irst and foremost at a favourable audience; an exchange with practising textual critics is not intended, as far as I can see,” but does not clarify whether he includes himself among the “practising textual critics.” Ibid., 26. 9 This contrast is invoked not only in Grünendahl’s writings but also in those of Indologists such as Stephanie Jamison and Sheldon Pollock. The former writes, “Ancient India remains the province of the philologists—among whom I am proud to number myself. It can only be approached through its texts, and the philological methods employed to investigate these texts 327 CONCLUSION 327 may, to the outsider, appear to approach the status of a mystery religion, to be even more esoteric than the contents of the texts they study. As a philologist, a practitioner of this cult, I am certain that the intricacies of the method are necessary. But I also regret the result: that the knowledge we gain too often remains walled of from the rest of the scholarly community, that the process of gaining it is so consuming that we often lack the energy to communicate it to others—and that others will not invest their energy in attending to the unfamiliar details that must be grasped in order to understand the whole.” Stephanie W. Jamison, Sacriiced Wife/ Sacriicer’s Wife: Women, Ritual, and Hospitality in Ancient India (New York: Oxford University Press, 1996), 3 (Jamison’s italics). 10 Michael Witzel, “Textual Criticism in Indology and in European Philology during the 19th and 20th Centuries,” Electronic Journal of Vedic Studies 21, no. 3 (2014): 12. 11 Witzel is obviously at pains to claim a separate genealogy for Indology than the one most often proposed for it, namely, as arising out of German Romanticism and sharing in the latter’s nationalism, for he emphasizes: “Incidentally, this semi-political cultural movement had little to do, contrary to what is now often asserted, with the strong German interest for [sic] Sanskrit and India, which was just one of the many avenues in the Romantic search for early textual materials. Still this interest is now frequently depicted, though erroneously, as having been instrumental for German nationalism of the Romantic and later periods, usually by those who know little of the original German language texts of the period—especially so in the Anglophone world: this has then been parroted all over the Indian right wing (Hindutva-leaning) internet.” Ibid., 15 (all italics Witzel’s). But his attempt at grounding the scientiic character of Indology in philology might have been more successful had he shown how exactly the Indologists dem- onstrate a knowledge of philology. Certainly, matters are not helped by the fact that he thinks Grünendahl “go[es] beyond the facile division into various script traditions.” Ibid., 34, n. 136. Pollock likewise implies a connection between Wolf and Indology by juxtaposing Wolf ’s phi- lology with Schlegel’s. See Sheldon Pollock, “Introduction,” in World Philology, ed. Sheldon Pollock, Benjamin A. Elman and Ku-ming Kevin Chang (Cambridge, MA: Harvard University Press, 2015), 7. For an assessment of this claim, see Vishwa Adluri, review of World Philology, ed. Sheldon Pollock, Benjamin A. Elman and Ku-ming Kevin Chang, American Historical Review 121, no. 3 (2016): 908–10. 12 This is something that every Indologist, of course, assumes, because, as we analyzed in The Nay Science, Indology as a discipline lives of of the postulated contrast with traditional scholarship. Thus, every time Indologists were asked what Indology was, they responded with a clariica- tion of what it was not: it was not the tradition. But the fact Indologists felt the need to belabor the contrast with tradition, especially insofar as it was “theological,” shows that theology has only been transformed and not overcome within Indology. Hence the need to constantly run down the tradition. It is precisely as a theological enterprise, speciically as a radicalization of the Protestant separation of faith from reason, that Indology felt that it was in competition with the tradition. 13 Even traditional exegetes, when they clarify the meaning of the śruti or the smṛti, are working with texts. If all that Witzel means by “philology” is a thorough acquaintance with the texts, then we would have to accede that traditional scholarship is also “philological,” indeed, “phil- ological” to a degree the Indologist could never hope to be. This is, of course, something no Indologist is willing to concede, since the discipline, as we argued in The Nay Science, draws its legitimation from the contrast with the tradition. 14 Witzel, “Textual Criticism in Indology,” 17. And see also ibid., 67: “This is not the case with a recent (re-)translation of Manu, where neither the readily available (semi-)critical edition of J. Jolly nor the oldest available commentary of Bhā ravi have been used and where matters of realia (for example the system of weights) are treated with cavalier neglect.” 15 Witzel’s insistence on the recovery of the realia from the text as the real task of “philology” corresponds to Semler’s distinction between the sensus literalis and the realia and the insistence 328 328 PHILOLOGY AND CRITICISM that the former could only be reached through a rigorous historical-critical methodology. Witzel’s insistence that recovery of the realia makes up either the whole or the largest part of careful “philological” reading reveals that he is confused about the distinction between his- torical criticism and textual criticism. Even though he claims to be interested in philology, all that he means when he says that Indologists contributed to philology is that they set forth the valorization of historical context over other kinds of exegetic concerns that is characteristic of Protestant scriptural hermeneutics. 16 Ibid., 16; see also ibid., 16–17: “The ensuing combination of the early development of his- torical and textual criticism, building on Bentley’s and Wolf’s [sic] philology and Jones’ and Bopp’s comparative linguistics, was especially due to the circle of scholars that W. v. Humboldt had assembled at the new Berlin (later, Humboldt) University, whose liberal and Wissenschaft- oriented structure was to become the model of many other western Universities”; and ibid., 19:  “Instead, it had been accepted, at least since Lachmann early in the 19th century, that we need a irm basis in order to study our classical texts:  we need a text that is ideally the same, or rather comes as close as possible to the text the author had in mind. This means to employ the methods of historical and textual criticism, strictly adhering to the principle of establishing a family tree of manuscripts (stemma)”; and ibid., 39–40:  “Since Lachmann we have been aware that we need to establish, as pointed out above (§1.4.), a text that comes as close as possible to the author’s text. This is to be carried out by historical and textual criti- cism, with a Lachmannian stemma, for which see P. Maas and M. L. West.” And see also ibid., 10: “Prominent was an historical and critical approach bolstered by the newly developed MSS stemmatics and the new evidence from comparative historical linguistics”; ibid., 30:  “Vedic and other Sanskrit texts were studied by employing the historical and critical method, already familiar from Classical texts, though some allowance was made for the strict oral transmission of the Vedic texts [?]”; and ibid., 33–34: “The recently acquired means of textual and histor- ical criticism were thus consistently applied, just as one had done with Classical and medieval European texts. The second part of the 19th and the early 20th century was the heyday of the critical and historical method (which in turn was heavily inluenced, especially for the older texts, by the expanding knowledge of comparative Indo-European linguistics).” Does he mean the historical-critical method or textual criticism? 17 Ibid., 19. 18 See Sebastiano Timpanaro, The Genesis of Lachmann’s Method, ed. and trans. Glenn W. Most (Chicago: University of Chicago Press, 2005), 96. Giovanni Fiesoli, in La genesi del lachmannismo (Florence:  Edizioni del Galluzzo, 2000), provides deinitive proof that Lachmann never used the common-error method. Witzel cites P. L. Schmidt’s essay “Lachmann’s Method: On the History of a Misunderstanding,” in The Uses of Greek and Latin:  Historical Essays, ed. A.  C. Dionisotti, Anthony Grafton and Jill Kraye (London:  The Warburg Institute, University of London, 1988), 227–36, but not Fiesoli’s work. 19 Witzel, “Textual Criticism in Indology,” 20, and see also ibid., 18 for the claim:  “Earlier European scholars of the Renaissance and Baroque periods had already noticed that their MSS frequently were just bad copies made by medieval scribes, whose mistakes were in part due to the change from uncial to miniscule characters in the 9th century CE, a bottleneck neck [sic] event, as we would now call it following biological parlance.” 20 Paolo Trovato, Everything You Always Wanted to Know about Lachmann’s Method:  A Non-Standard Handbook of Genealogical Textual Criticism in the Age of Post-Structuralism, Cladistics, and Copy-Text (Padua: Libreriauniversitaria, 2014), 141. 21 Witzel, “Textual Criticism in Indology,” 18. 22 Bronkhorst makes the same error, though in his case the error is attributable to his reliance on Witzel: “In both of these examples it is possible to think in terms of a bottleneck: of all the manuscripts that existed, only one became the ancestor of all those that survived (or have been taken into consideration). However, neither in the case of the Vākyapadı̄ya nor the Mahābhāṣya 329 CONCLUSION 329 do we have reason to think that only a few manuscripts existed at the time of the archetype. Quite on the contrary, there may have been many manuscripts in existence during the crea- tion of the one manuscript that would become the archetype of all manuscripts extant today. Seen this way, there may have been no real bottleneck. Some texts may have gone through a real bottleneck in the sense that few manuscripts existed at certain points in their history. An example may be found in the Paippalāda Saṃhitā of the Atharva Veda. The manuscripts of this text, preserved in both Kashmir and Orissa, go back to one written archetype from around 800–1000 CE in Gujarat. This, at any rate, is the theory presented by Michael Witzel (1985), who ofers the following explanatory hypothesis: For many centuries Brahmins of the Atharva Veda were centered in Gujarat, whence some were invited from time to time by kings in other parts of India. They arrived with their texts (i.e. the version of the Paippalāda Saṃhitā current in Gujarat). Other traditions of that text either did not exist or were overshadowed by the origi- nally Gujarati tradition. In other words, it is possible that only a small number of manuscripts of the Paippalāda Saṃhitā existed in that period. Here we can speak of a real bottleneck.” Johannes Bronkhorst, “Archetypes and Bottlenecks: Relections on the Textual History of the Mahā bhā rata,” in Pūr vāparaprajñābhinandam: East and West, Past and Present. Indological and Other Essays in Honour of Klaus Karttunen, ed. Bertil Tikkanen and Albion M. Butters, Studia Orientalia 110 (Helsinki: Finnish Oriental Society, 2011), 41. 23 See Witzel, “Textual Criticism in Indology,” 57:  “Such cases are due to a bottleneck situa- tion, mentioned above (§1), where a new stream of tradition evolved, superseding all others,” and see also ibid., 60–61, for the reference to the “Nā garī/Grantha bottleneck,” apparently responsible for the reduction in the number of copies in circulation to one: “Given the Nā garī/ Grantha bottleneck and the inherent lack of understanding of older MSS, therefore more easily discarded, it should not surprise that a new successful ‘species’ of textual tradition took over around 1000 CE. I suspect that most of our well-known ‘classical’ texts go back to such archetypes, written after the emergence of the (sub)regional scripts.” The argument parallels Bigger’s “normative redaction” hypothesis. Witzel also appears inluenced by Grünendahl’s idea that a text “could also have been transferred from completely independent sources” (Reinhold Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” in Studien zur Indologie und Buddhismuskunde, ed. Reinhold Grünendahl, Jens Uwe-Hartmann and Petra Kiefer-Pülz, Indica et Tibetica 22 [Bonn: Indica et Tibetica Verlag, 1993], 128–29) when he writes: “Such imports naturally lead to contamination, but importantly also to the superimposition of, and eventual replacement by a distant tradition over one’s own local one. This was common when a new, highly regarded version of a text was copied and the ‘older’ version ceased to be copied. Indeed, we can, occasionally, establish that there was just one late classical or medieval arche- type, from which all surviving MSS of a text derive.” Witzel, “Textual Criticism in Indology,” 55–56 (Witzel’s italics). But there is a diference between saying that all surviving manuscripts of a text are descended from the same copy and saying that just one copy existed. Which is it that Witzel means? The same error occurs in Michael Witzel, “On the Archetype of Patañjali’s Mahā bhā ṣya,” Indo-Iranian Journal 29 (1986): 249–59. Witzel begins the article thus: “W. Rau recently indicated that all available MSS of the Mahā bhā ṣya go back to ONE single archetype from which all have been copied” (ibid., 249; Witzel’s capitalization). Archetype in the sense of a reconstructed source from which all extant manuscripts are presumed to descend is only ever used in the singular. “ONE single” is redundant unless Witzel means the number of copies in existence at the time was reduced to one. The fact that all the extant manuscripts share certain features indicating a common ancestor, however, does not justify us in assuming this manuscript was the sole one in existence at the time. If Witzel’s attribution is correct, Wilhelm Rau, Die vedischen Zitate im Vyākaraṇa-Mahābhāṣya, Mainzer Akademie der Wissenschaften und der Literatur, Abhandlungen der Geistes- und Sozialwissenschaftlichen Klasse, Jahrgang 1985, no. 4 (Wiesbaden: Steiner, 1985) commits the same error. 24 Witzel, “Textual Criticism in Indology,” 21. 30 330 PHILOLOGY AND CRITICISM 25 The phrase is Wolf ’s (!)  and is cited and translated in Timpanaro, The Genesis of Lachmann’s Method, 71–72. 26 Compare Trovato’s ininitely more sophisticated account:  “The most commonly adopted procedure for singling out these errors—often inaccurately described by computer-assisted editors—is the following. We begin by setting as our term of comparison an apparently com- plete manuscript, one that is traditionally regarded as authoritative, or a randomly chosen one for want of a better alternative. We give it a label, such as A, and proceed to compare the irst verse or, in the case of a prose text, the irst line of A with the corresponding text portion of B, the second verse or line of A with the second of B, and so on, accurately recording any substantial diferences. Once we have gone through all the diferences between A and B, we go on to compare A with C, D, etc., until we have examined all the witnesses (manuscripts or printed editions, thus called because they bear witness to the content of the original text). The technical terms used for the action of comparing and the manuscript used for the comparison are respectively to collate (French collationer, It. collazionare, from the Latin collatio [comparison]) and collation text (It. esemplare di collazione). All the diferences thus observed can be called, in a broad sense, readings or variants.” Trovato, Everything You Always Wanted to Know about Lachmann’s Method, 52–53. Note that this procedure yields in the irst instance only a list of variants and not yet of errors as compared with the collation text—which of these variants must be considered an error cannot be established without further analysis. 27 Witzel’s procedure runs into the same objection as we found of Grünendahl’s “one recen- sion” edition, namely, that the identiication of errors, like the selection of the best manu- script, presumes a Lachmannian analysis. Contini expresses it best: “An objective deinition, developed in the Neo-Lachmannian milieu, of the best manuscript as the one which has been so resistant to banalization as to ofer the highest percentage of lectiones singulares to be preserved presupposes a Lachmannian edition. Indeed, only a Lachmannian editor, such as Bédier was for quite a long time, can point out the best manuscript, or even merely a good one.” Gianfranco Contini, “Filologia (1977),” in Frammenti di ilologia romanza: Scritti di ecdotica e linguistica (1932–1989), vol. 1, ed. Giancarlo Breschi (Florence: Edizioni del Galluzo, 2007), 34 (italics in original). He is equally persuasive on the concept of a latent error: “Stemmata are an objective and mechanical tool, invented to sort out, in the irst instance, the quarrel between variants that are in themselves equally acceptable: […] variants that are pronounced erroneous sever to brand as erroneous those that have remained equally acceptable.” Witzel’s procedure may correct obvious errors, but it would miss an equally large or larger number of latent errors. But here also it runs into a serious objection, which Contini voices thus: “Indeed, what does purging the ‘obvious’ errors in a manuscript mean? What is more disputable than obvious- ness, and the limits of the obvious?” Gianfranco Contini, “La ‘Vita’ francese ‘di Sant’ Alessio’ e l’arte di pubblicare i testi antichi (1970),” in Frammenti di ilologia romanza: Scritti di ecdotica e linguistica (1932–1989), vol. 2, ed. Giancarlo Breschi (Florence:  Edizioni del Galluzo, 2007), 961 and 967 (all italics in original; all translations Trovato’s). Insofar as Witzel’s emendations are not clearly signaled and entered into the body of the text, his procedure—under the pre- text of restoring a historically more accurate version of the text—ends up creating a new, composite edition. Olivelle’s criticisms are relevant here:  see Patrick Olivelle, “Unfaithful Transmitters: Philological Criticism and Critical Editions of the Upaniṣads,” Journal of Indian Philosophy 26 (1998): 173–87. 28 Witzel, “Textual Criticism in Indology,” 22. 29 Ibid., 23. 30 Michael Witzel, Katḥ a Ar̄ aṇyaka: Critical Edition with a Translation into German and an Introduction (Cambridge, MA:  Department of Sanskrit and Indian Studies, Harvard University, 2004). This is a reprint of the partial print of his 1972 dissertation published as Das Katḥ a-Ar̄ aṇyaka, textkritische Edition mit Uebersetzung und Kommentar (Erlangen: Nepal Research Centre, 1972). 31 Witzel, Katḥ a Ar̄ aṇyaka, xxiv. 31 CONCLUSION 331 32 See ibid., xxiv (boldface in original): “The fragmentary character of the only extant manuscript containing the major part of the text required an exact representation of destroyed passages as well as of lines or Akṣaras, of which only the upper or lower part, or even less, has been preserved. In the edition, such passages (and uncertain readings in general) have been indi- cated by italics. […] The approximate number of broken of Akṣaras has been suggested by x, occasionally a missing vowel sign by ə, especially when the original accent was still discernible.” Emendations, especially when Witzel’s own and not supported by any manuscripts, appear to be only occasionally noted in the endnotes. This is perhaps at the heart of his disputation with Olivelle regarding the tenability of inserting emendations into the body of the text. 33 Ibid., xxi–xxii. 34 The expression “though they only contain parts of the mantra section (KaṭhĀ I)” is especially infelicitous, since it is susceptible to misinterpretation: it might mean that these manuscripts contain only parts of the mantra section or that they contain only parts of the mantra section; in the latter case, the manuscripts would obviously be far more useful, and this seems to be what is meant (though this is then contradicted by the observation that only one manuscript contains the “major part of the text”). 35 Ibid., xxiv–xxv. 36 The absence of a stemma in an edition that calls itself a critical edition is especially puzzling, given Witzel’s insistence elsewhere that a stemma is the deining characteristic of a critical edition. See Witzel, “Textual Criticism in Indology,” 49: “It is remarkable that over the past 200 years or so only about a dozen truly critical editions, with stemma, of Sanskrit texts have been prepared” (Witzel’s italics); ibid.: “Older editions, including the fairly recent Poona efort (T. N. Dharmadhikari, R. S. Shastri, N. P. Jain, S. S. Bahulkar, Vedic Texts: A Revision, [...]), are semi-critical, i.e. without stemma”; ibid., 51:  “They were by no means critical, as we now understand, that is with a stemma”; ibid., 54:  “The 19th century ‘Boehtlingk et  al.’ semi- critical type, worst perhaps Boehtlingk’s Bṛhadā raṇyaka and Chā ndogya Upaniṣad. They give some variant readings and may even briely discuss the origin of the MSS involved (‘from Ahmedabad’) but they do not even try to establish a stemma. Most of the editions replicate those done before Lachmann” (Witzel’s italics); ibid., 54, n. 245: “The RV editions of Aufrecht, M. Müller, C. G. Kashikar, etc. are not critical in the strict sense as they do not give a complete apparatus nor do they have a stemma”; and ibid., 58: “First of all, it shows that we could make much more progress in the tracing the history [sic] of a particular text if we only had: (1) Proper editions, preferably with stemma, or at least a record of all the variants of the MSS available to an editor” (Witzel’s italics). 37 “The Katḥ Ā has come down to us only in one rather lacuneous [sic] birchbark ms. This puts certain limits to any sort of investigation and it added some additional constraints to my task of translation. I had to supply, as is the case with the translation of most Mesopotamian and Hittite texts on clay tablets, much of the lost text portions for consistency and continuity. I even did not know [sic] exactly how my text began, and its end was even more uncertain. While this certainly is not typical for Vedic texts, it presents some additional problems. Taking all of the preceding into account, I irst had to establish a reliable edition, which is diicult with just one ms. but can be established by carefully comparing the parallel traditions, ritual, style, and the Vedic grammar of the period in question.” Michael Witzel, “How to Enter the Vedic Mind? Strategies in Translating a Brā hmaṇa Text,” in New Horizons of Indological Research: Felicitation Volume in Honour of Prof. P.  C. Muraleemadhavan, ed. Dharmaraj Adat (Delhi:  Kunjunni Raja Academy of Indological Research and New Bharatiya Book Corporation, 2013), 25. 38 “As early as 1974, at the memorable round table of the congress of Romanists held in Naples, supporters of Neo-Lachmannism (Segre) and Neo-Bédierism (Lecoy) reached not too distant conclusions, summarizable as the desirability that the critical text should not stray too far from the text of one of the manuscripts (except in loci that were admitted of certainty, according to Segre), whereas the critical apparatus should be the place for comparisons of readings and 32 332 PHILOLOGY AND CRITICISM reconstructive hypotheses.” Leonard Lino, “Il testo come ipotesi (critica del manoscritto- base),” Medioevo romanzo 35 (2011): 8. 39 Witzel, Katḥ a Ar̄ aṇyaka, xxv and see also Witzel, “How to Enter the Vedic Mind? Strategies in Translating a Brā hmaṇa Text,” 27: “One item that come [sic] to our rescue in following such arguments and in our actual understanding of them, and indeed the most important one that came to my rescue in restoring the partially damaged text, is the so-called Zwangsläuigkeit of Brāhmaṇa style. The expression coined by K. Hofmann signiies the ‘inevitability of Brāhmaṇa style.’ The argumentation in the texts almost always follows a certain ixed pattern.” 40 Along with Trovato (Everything You Always Wanted to Know about Lachmann’s Method, 251), we acknowledge the necessity of emendation or, as Trovato elegantly puts it, “the ineluctability of critical judgment.” But Trovato’s observation that “many conjectures born of iudicium—that is, the competence and intelligence—of classical and modern philologists (or, more appropri- ately: many working hypotheses) have been conirmed by early witnesses discovered at a later date or fresh readings of wholly or partly illegible documents with new technologies, which carry precisely the conjectured reading. Not unlike what happens in physics nowadays, the best working hypotheses are later proved true” (ibid.) applies only to traditions in which much more is known about the authors and of the scribes’ practice. In contrast, the history of Sanskrit critical editions reveals an extraordinary overconidence on the part of the Indologists in their ability to uncover “original” meanings—meanings that have not withstood the test of time. Olivelle speaks of “philological hubris.” Olivelle, “Unfaithful Transmitters,” 177 and 183. 41 See Witzel, “Textual Criticism in Indology,” 49: “Surprisingly, Indian (or South Asian texts in general) texts have hardly seen any critical editions in the strict Lachmannian sense. What goes under this name usually are editions that merely include a selection of variants. It is remark- able that over the past 200 years or so only about a dozen truly critical editions, with stemma, of Sanskrit texts have been prepared”; see also ibid., 25:  “Well, it [a critical edition of the Praśna Upaniṣad] has not yet been done or even tried! We do not have a critical edition of the Upaniṣads”; ibid., 26: “In addition, Ś aṅkara’s commentaries have not been edited critically either. (I have frequently impressed that fact on Indian visitors who asked me what to do in their future work—to no avail.)”; ibid., 50: “Studies based on the present uncritical editions can, at best, only be provisional and are, at worst, plainly wrong. […] It is nothing short of a scandal that, after some 200 years of Indology, instead of preparing reliable texts and translations, a lot of ink keeps being spilled in works based on inadequate materials”; ibid., 51: “As mentioned (above §1.6.), the commentaries themselves have to be critically edited irst before actually using them in a truly scholarly fashion”; and ibid., 51 (with a reference to an earlier work): “I have criticized this approach nearly two decades ago (Witzel 1997:  Introduction):  ‘With millions of Indian MSS in public and private libraries, only a fraction of the texts contained in them—nobody knows how many—have been actually published or used for editions. So far, we have only a handful of critically edited Sanskrit texts which are based on a stemma of the manuscripts used”; ibid., 53: “I have criticized the ‘traditional’ Indological approach to editing, described extensively above, more than two decades ago (1997)” (footnoted with a reference to the same work and the same passage); ibid., 57: “Yet, a text such as the Ṛgveda cannot be understood if one does not know something about cattle, the historical climate of the Panjab, pre-state tribal societies and their social systems, about its complex system of Indo-European and Indo-Iranian poetics, about oral composition, canon formation and the techniques of crit- ically editing Sanskrit texts” (italics added); and ibid., 62: “However, as mentioned, commentaries have irst to be critically edited before we can fully rely on their testimony.” 42 See ibid., 22: “Housman, however, was of the opinion that ‘the truth or falsehood of a MS reading can never be conirmed or corrected by a<n equally> decisive test […] [that] would be the production of the author’s autograph.’ While he is right in principle, we can approach a solution by using the stemmatic method if it is combined with paleography and other Hilfswissenschaften”; ibid., 51–52:  “The main aim of Olivelle’s recent criticism of Upaniṣad 3 CONCLUSION 333 editions, however, remains controversial. He asks whether it is legitimate to incorporate emendations in the body of the text. One would have thought that we have for long discussed and acted on this question, for some 150 years in fact. It is clear that we can certainly insert our emendations in the text if we can in fact show that they are justiied and if we properly list all MSS variants in the footnotes, in the critical apparatus. (This is, however, not consis- tently done even by current editors, leading to Olivelle’s complaint.)”; ibid., 22: “After these initial steps of editing a text, higher textual criticism (emendatio) comes in. Based on our knowl- edge of the grammar, style, parallel passages or typical expressions of the author concerned— repeating here the Alexandrian model—we can scrutinize the archetype MS and propose certain corrections to that text. Obviously, here we are in the realm of conjecture (emendatio). The better one is as a philologist, the more one ‘employs one’s thought,’ and the more one uses (by now also electronic) tools, the better the resulting text will be” (all italics Witzel’s). 43 Olivelle, “Unfaithful Transmitters,” 173–87. 44 Though, as we know, the life of a conjecture is lamentably short. Böhtlingk’s reputation as a philologist was established to a great extent on his conjectures in the Upaniṣads, though, as Olivelle has shown, most of them are now considered erroneous. Perhaps this is why Witzel, along with his “old friend” Bernhard Kölver, complains: “It was a nice two hundred years.” Witzel, “Textual Criticism in Indology,” 17, n. 39. 45 See ibid., 28: “It is well known that contamination is the rule in the (edition) of the Sanskrit epics, which makes a true critical edition impossible (apart from the problems of oral bardic transmission), as stressed by the Mbh. editor Sukthankar”; ibid., 35 (Witzel’s italics): “At the time, Sukthankar could not yet fully appreciate the truly oral nature of the originally Bardic text (as exempliied by M. Parry (1930–32) and A. Lord (1991), which indicates that we will never reach a true Mahā bhā rata archetype, just as little as an ‘original Homeric’ text for the Greek epics”; ibid., 40–41:  “The situation is therefore comparable to that of other Bardic texts (Mahā bhā rata, medieval Indian Bhakti texts, etc.). Obviously, a stemma with an arche- type cannot be produced for such texts”; and see also ibid., 55:  “An exception obviously is the—ultimately futile—Winternitz-Sukthankar undertaking of establishing a stemma for Mahā bhā rata MSS early in the 20th century.” 46 Ibid., 34 (all italics Witzel’s). 47 Ibid. (all italics Witzel’s). 48 See ibid., 39 (Witzel’s italics): “Contrarily, colonial dominance in India apart, western inlu- ence on the theory and practice of Indian philology has been rather small. After some initial adaptations, notably by R. G. Bhandarkar (1837–1925) and his successors, the developments of the 20th century are of a rather mixed nature. Leaving apart the ever-diminishing number of learned traditional Pandits, irmly moored in their religious or śā stric mindset, some university-based scholars took over just some western methods of textual study. However many, if not most, constantly mingle this approach with traditional attitudes: we ind mythical and legendary ‘data’ (and dates) interspersed with minute and cogent observations on language, grammar, texts or cultural background. All in all, a picture emerges of a mindset untouched by the critical attitudes of the Enlightenment. Instead, the prevailing Hindu inclusivism super- icially incorporates some ideas external to Indian culture, certain procedures, or preferably, the latest technology, into a preexisting traditional framework. It does not see a contradiction.” But what could be a bigger myth than the idea that we can reconstruct the tradition “from the head,” because we know that an original heroic Aryan civilization must have existed in India? 49 Grünendahl, “Post-philological Gestures—‘Deconstructing’ Textual Criticism,” 27 and 27, n. 55: “But no matter which criteria we apply to an edition in order to call it critical we shall soon come to the conclusion that textual criticism in Indology has only just taken of: there are still preciously few critical editions of Sanskrit texts around, not to mention of texts in other Indian languages,” a statement he footnotes with the words: “In my view, a stemma may con- tribute substantially to what I consider the ultimate goal, viz., Nachvollziehbarkeit of the editorial 34 334 PHILOLOGY AND CRITICISM process, but it is not a sine qua non. If it was, Indologists would ind themselves without crit- ical editions of the Sanskrit epics, where contamination precludes such pretensions. […] The same holds for various other ields where serious reservations towards strict stemmatology have been voiced; see, for example, J. Hanneder’s edition of Abhinavagupta’s Mā linīślokavā rttika 1.1–399 (Groningen 1998, p.  40–45), and the Groningen edition of the Skandapurā ṇa by R. Adriaensen et al. (Vol. 1. Groningen 1998, p. 39)”; Witzel, “Textual Criticism in Indology,” 21: “Grünendahl (2008: 11) rightly stresses that a critical edition has the general goal of the Nachvollziehbarkeit, that is, the replication by readers of the editorial process”; and see also ibid., 74: “Yet, as Maas would point out, having a stemma (or we may add, an early version of a Bardic text) is only the starting point for the processes of emendatio and producing a critical edition.” 50 See Helmut Müller-Sievers, “Reading Evidence: Textual Criticism as Science in the Nineteenth Century,” The Germanic Review: Literature, Culture, Theory 76, no. 2 (2001): 167: “Over the entirety of Schleiermacher’s enterprise hovers the concept of the hermeneutic circle: that we cannot understand the parts (for example, the manuscripts) without the whole (for example, the author, or the ‘work’), and vice versa. Lachmann maintained that because textual criticism precedes any interpretation, it is exempt from hermeneutic circularity—and this is why textual criticism, if applied with the utmost rigor, can aspire to the dignity of the natural sciences and their inductive-deductive procedures. Indeed, [Müller-Sievers continues] the attribute most often evoked in conjunction with Lachmann’s method is rigor or severity (Strenge). No ‘reprehensible clemency’ (sträliche Milde) should be allowed in securing all the evidence for a textual tradition. The texts to be reconstructed have to be rigorously documentary (strengurkundlich); they must be subjected to the strictest exam, and they follow no writing before them (Vorschrift). […] Apart from imitating the natural sciences, this severity also served to recommend philology in gen- eral, and textual criticism in particular, to the pedagogical institutions of the German state.” 51 “Before our text can be studied from a philosophical or religious studies [religionskundlichen] perspective, it must be reliably edited. One cannot unseat this [fact] even though this truism is gladly and repeatedly ignored in those circles [Does he mean Indians? Or Western students of Indian philosophy?] that would rather [spiritually] elevate themselves with translations of the Upaniṣad rather than taking the trouble to read and understand these texts philologically and critically [philologisch-kritisch] in the original.” Wilhelm Rau, “Bemerkungen zu Ś aṅkaras Bṛhadā raṇyakopaniṣadbhā ṣya,” Paideuma 7, Festgabe für Herman Lommel zur Vollendung seines 75. Lebensjahres am 7. Juli 1960 (1959–61): 299; “Philology […] is the fundament of our science. It is the foundation on which we must build. The texts are our best source of testimony about classical India […] in many ways the only window that we have on classical Indian society. […] [O]nly through the creation of edited texts […] can [we] begin to place these texts in their proper context.” Richard Lariviere, Protestants, Orientalists, and Brāhmaṇas:  Reconstructing Indian Social History (Amsterdam: Royal Netherlands Academy of Arts and Sciences, 1995), 16–17. 52 In truth, the Indologists’ conidence in their ability to recover presumed “original” meanings has depended entirely on their belief that they were capable of displacing themselves into the conditions and/or mindset of ancient Indians. This belief, the root delusion of the Indologist, is manifest not only in the writings of Hermann Oldenberg, as discussed in The Nay Science, but also in Witzel’s work. See, for instance, Witzel, “How to Enter the Vedic Mind? Strategies in Translating a Brā hmaṇa Text,” 25 (the title itself is revelatory): “The most intriguing and dii- cult part of this undertaking [producing an edition of the Kaṭha Ā raṇyaka for his dissertation] was to translate according to the original intent of the composers of the text. One has to enter the Vedic mind in order to be able to achieve this aim” (all italics in original); ibid., 30: “The real task, however, is how to enter the Vedic mind, the mind of those Brā hmins who composed poetry and prose texts such as the Katḥ a Ar̄ aṇyaka”; ibid., 31: “In this fashion, and due to the large number of texts available to us, we have a real possibility to ‘interview’ our Vedic ‘informants’ of three millennia ago and to enter their mind as well as any anthropologist can do”; and see 35 CONCLUSION 335 also the story narrated on the same page: “If it sounds unbelievable that we can actually enter the Vedic mind and argue from the inside, following the thought pattern of the Vedic authors, I invite the reader to try beginning [sic] of the Katạ Ar̄ aṇyaka [sic]. As the facsimile shows, it is rather fragmentary. In order to restore and to translate the text, I had to study similar phrases or occurrences of a few key words retained [sic] the fragment. The restoration was supported [sic] the Zwangsläuigkeit of Brāhmaṇa style: the initial, half peeled of sentence is more or less repeated by a later one. But how to be sure? Fortunately, the passage contains another clue the frequently met with concepts of ‘thought-speech-action’ (manas-vāc-karman), a collocation that is found not only in the Veda but also in the closely related Old Iranian texts (manah-vacas-śiaoʋna, Y 34.1–2). Therefore, I was completely sure that I had restored the text incorrectly [correctly? A Freudian slip?]. When I inally went to Tübingen University Library to check the original ms. again (I had worked from a microilm), I found that a portion of my original lacuna was cov- ered by a small, dislodged piece of birchbark that half overlapped with my text. When I lifted the dislodged fragment, I found the text I had restored. If we can write Vedic texts that well, we can also translate them.” The story is also repeated, with only slight changes of accent, in Witzel’s 2014 article. The last line is updated to read: “If we can write Vedic Sanskrit texts that well, we can also edit, translate and understand them.” Witzel, “Textual Criticism in Indology,” 75 (italics in original). 53 We cite here the text of an exchange between Witzel and Grünendahl on the Indology list. First, Witzel’s remarks: “This makes for an interesting discussion topic—without laming. In my opinion the answer to most of A.  Thrasher’s questions put must be airmative. To put it in a few theses:  1. There are only a handful of critical editions of Sanskrit texts—in the strict sense, that is with STEMMA of the MSS. (Mbh., Ram., Koelver’s Rajatarangini work, Lariviere’s Narada etc.). 2.  All other ‘critical’ editions—including most of the often excel- lent editions made in the 19th cent.—are only attempts at critical editions, based on the rather limited MSS materials available then, and with occasional notes on their interrelation. MSS variants are recorded well—though not always consistently and while leaving out many of the important (regional) peculiarities of spelling (see below). Unfortunately this applies to Vedic editions as well, with the additional drawback that the—frequently better—oral tradition has been used only three times in editing a text: twice in the late 19th cent. and once in 1967! See below for an appreciation. 3. The bulk of the editions do not even mention variant readings of their not/badly reported MSS sources, or if they do so, they do so rather inconsistently. Take any Anandasrama edition, for example. Thus, the reader never knows what he/she is dealing with in a particular instance. 4. The rest relects the ‘commercial’ editions A. Thrasher asked about: editions made by someone (not always mentioned by name) on the basis of unreported MS(S) in a manner to be ‘deciphered’ by the user after long exposure. Such editions (e.g. Nirnayayasagara eds.) do not have more value than any MS (which also shows the hand<s> of the scribe/reader in its many corrections or ‘improvements’ of the text). Thus, in addition to case (1), ALL others are useful for a critical edition but have to be treated with caution— namely, as what they represent. Cases (3) and (4) can/often do represent a local tradition and are valuable as such: one should treat such editions on the same level as a MS: they relect nothing more than the more or less educated guess at a ‘correct’ text—producing many a lectio facilior. […]As far as I  am aware, this kind of question is rarely raised even among ‘professional’ indologists (otherwise we would have critical editions of Sankara, all works of Kalidasa etc. etc. by now!). Actually, I regard this neglect as nothing but scandalous. How can one make valid statements on advaita if one does not even come close to the text Sankara may have written? (The same applies, a fortiori, to Vedic editions made without using oral tradition. A recitation of the Rgveda is, after all, a tape recording of c. 1000 B.C. and as such better than any MS). Of course, the only way to achieve such an edition is long—tedious many would say—but unfortunately it is the only one we have, short of direct inspiration by the ancient Rsis or authors. It involves not only burning the midnight oil for a very long time when trying 36 336 PHILOLOGY AND CRITICISM to igure out the stemma, but it also involves the critical use of palaeography (where is one for MSS of this millennium??) and of the knowledge of local pronunciation (again, not collected anywhere!) and local orthography inluenced by it. […] In short, I think most students and professors are ill prepared even to begin REALLY critical editions—should they indeed be interested in doing such work (which often is called the death-knell for a prospective PhD stu- dent in North America, also in my University.) Still, we have to train a whole new generation of students to begin a task that has long been achieved in Classical Greek, Latin, Celtic etc. studies. […] After all, texts are a large part, if not the largest, of the materials we are dealing with on a daily basis and we should be sure of WHAT we are dealing with.” Grünendahl’s response follows: “M. Witzel’s contribution highlights a few ‘critical’ aspects of editions: 1. What makes an edition CRITICAL? If a stemma of the MSS used is the decisive criterion, then 200 years of indology have not produced very much deserving that name. Contrary to Witzel’s opinion, the critical editions of Mahabharata and Ramayana have to be dropped from the list (while others, as e.g. Wilhelm Rau’s Vakyapadiya, may be added). What is usually taken in the BORI and Gaekwad editions resp. do [sic] be a stemma of MSS is in fact a ‘pedigree of VERSIONS.’ [As to the term ‘version’ and its implications I may refer to my article ‘Zur Klassiizierung von Mahabharata-Handschriften’ in Studien zur Indologie und Buddhismuskunde: Festgabe des Seminars fuer Indologie und Buddhismuskunde fuer Prof. Dr.  Heinz Bechert.’ Bonn 1993 (Indica et Tibetica; 22); Grünendahl’s insertion.] But why should a stemma be the decisive criterion after all. This is a concept taken over from classical philology, and in my opinion, it still stands to reason whether it can really be applied to Sanskrit texts, except perhaps in a few cases with a very limited and rather ‘hermetic’ manuscript tradition, like perhaps the Rajatarangini. By the way, B. Koelver did not publish a critical edition of the Rajatarangini, but rather a study of its manuscript tradition with a lot of free advice for future editors. The actual work of editing the text is still waiting to be done by someone undeterred by the sound of the death-knell. My own modest attempts in the ield have taught me that the relation even of a limited number of MSS, even from a limited geographical area—like e.g. Nepal, is very diicult to determine with any degree of certainty. Consequently, it takes a fair amount of simpliication to press the often delicate interrelations into a stemma. It may be asked what we can expect from such a stemma. 2. Personally, I very much appreciate any ‘attempt’ at a critical edition. In all probability, it is a step forward, especially if the text in question has not been edited before. I don’t see any fault in limiting your ‘manuscript materials,’ as long as you do not leave the choice entirely to external circumstances. In view of the limitations of human life, not to mention the necessities and absurdities of its academic derivate, choices have to be made. Even if the result is not the type of DEFINITE edition some of us may expect to achieve—Sukthankar, by the way, didn’t!—it will help, if only in re-examining the MSS used, should that turn out to be necessary, and perhaps in preparing a better edition on that basis. Textual criticism is indeed a long and sometimes tedious process. But what can be more impor- tant for our discipline than injecting fresh material? 3./4. The same applies to the other types of editions, although with considerable qualiications, as already pointed out by M.  Witzel. His appeal to overcome the ‘scandalous neglect’ of textual criticism has my whole-hearted support.” http://list.indology.info/pipermail/indology_list.indology.info/1995-June/002673. html and http://list.indology.info/pipermail/indology_list.indology.info/1995-June/002674. html (both accessed March 28, 2015; all italics and uppercase letters the respective authors’). 54 Shillingsburg’s deinition “the result of any emendation is a critical text” is unacceptably wide. It would have the consequence of including all medieval copyists’ editions under the deini- tion of “critical text,” which can hardly be his intent. If the term is to have any meaning, it must entail a distinction between a stemmatic, mechanical procedure and emendation either ope ingenii or ope codicum (that is, with only occasional recourse to the manuscripts). Indeed, this is how the term has been understood for much of the past century. Perhaps Shillingsburg simply means any text in which an editor has exercised critical judgment, as his suggestion 37 CONCLUSION 337 that “the editor has a third choice: to edit a text that does for the authors what he expected to have done for him [that is, by a publisher], but avoids the extraneous alterations imposed by a publisher in his normal but misguided undertaking of the production process” indicates. Peter L. Shillingsburg, Scholarly Editing in the Computer Age: Theory and Practice (Ann Arbor: University of Michigan Press, 1996), 56. But this applies just as much to medieval copyists and Humanist editors as to the Lachmannian or neo-Lachmannian editor. 38 39 EPILOGUE In the world also, it is seen that the knower of the meaning is extremely adored (with money etc.) instead of a person who is only a reader (waving his hands according to svara, etc.).1 —Sā yaṇa, Upodghāta to the Ṛgvedasaṃhitā In two books now, we have found that whenever Indologists say that their work is sec- ular, scientiic, philological, critical and the like, what they really mean is that they are interested in a separation of the realia from the meaning of the text.2 This separation, however, is anything but secular, since the emphasis on the realia over the text’s philo- sophical meaning is a characteristic feature of neo-Protestant theology, corresponding to the belief that only a reconstruction of the historical conditions (not only the events and personalities but also the moral and social codes) prevalent at the time the books of the Old and New Testaments were composed can permit readers to represent to themselves the respective authors’ state of mind at the time and hence, in an act of sympathetic understanding or Einfühlung, feel or experience what the author felt or experienced when writing the book in question. This task was all the more urgent in the case of those books that were held to be divinely inspired, for at stake was nothing less than gaining access to the true meaning of scripture, which had become contentious in the wake of challenges to the authority of the church. Yet, if the realia truly hold the key to understanding the text—indeed, if only they permit a correct understanding of the text, as the Indologists contend3—then, in reading the German Indologists’ work, we also must focus on the historical realia. If knowing about cattle is important to understanding the Vedic mind,4 how much more important is it to know about the Indologists’ religious and political commitments, especially as these translated into positions of authority, state-sponsored salaries and the ability to make self-authenticating statements? In our next book, we plan to do exactly this. Notes 1 Sā yaṇa, Upodghāta to the Ṛgvedasaṃhitā, cited and translated in Saraswati Bali, Sāyaṇa’s Upodghāta to the Taittirnˉya Saṁhita and the Ṛgveda Saṁhitā: Introduction, English Translation of the Text and Notes (Delhi: Pratibha Prakashan, 1999), 163. 2 See Andreas Bigger, “The Normative Redaction of the Mahā bhā rata:  Possibilities and Limitations of a Working Hypothesis,” in Stages and Transitions: Temporal and Historical Frameworks in Epic and Purāṇic Literature. Proceedings of the Second Dubrovnik International Conference on the Sanskrit Epics and Purāṇas, ed. Mary Brockington (Zagreb:  Croatian Academy of Sciences and Arts, 2002), 24: “If, on the other hand, we accept the hypothesis of a normative redaction of the MBh as a fact, we can go on to ask further questions. First of all, one has to ask if it is possible 340 340 PHILOLOGY AND CRITICISM to localise the normative redaction in what one might call the space–time continuum of Indian history. […] We have two sets of evidences to date the normative redaction. The irst comes from within the text; realia mentioned in certain passages permit us to assume a terminus post quem for the passage in question. Since all the passages contained in the normative redaction must have come into existence before or while the normative redaction was being compiled, they also provide a terminus post quem for the normative redaction as a whole.” The search for the Mahā bhā rata’s dates, however, is not advanced by arbitrarily removing passages, adding others and generally reshaping the text to correspond to one’s prejudices about its Textgeschichte (an a priori racist or anti-Semitic history), all the while arguing that this is what the “Brahmans” must have done. 3 See, for instance, Michael Witzel, “How to Enter the Vedic Mind? Strategies in Translating a Brā hmaṇa Text,” in New Horizons of Indological Research:  Felicitation Volume in Honour of Prof. P.  C. Muraleemadhavan, ed. Dharmaraj Adat (Delhi:  Kunjunni Raja Academy of Indological Research and New Bharatiya Book Corporation, 2013), 21: “Before we can even attempt a translation of Brāhmaṇa texts, there [are] a number of procedures that must be discussed and several obstacles that must be overcome. Most of them can be taken care of by our old hand- maiden, philology. It is well-known that to merely mention this word is already the kiss of death in some circles, including Harvard. In fact, one of my colleagues here once explained philology to me as ‘the study of a word.’ I rather prefer to deine it, as we did in s [sic] symposium some ive years ago:  as ‘kulturwissenschaft based on texts,’ or ‘the study of a civilization based on texts.’ ” 4 See ibid., 27: “Furthermore, it goes without saying that we have to know a large amount of the realia of the period [if we want to interpret the text], whether it is items of cattle herding, local climate, tribal society, or ancient and beliefs [sic]”; and see also ibid., 24: “For example, the great commentator Sā yaṇa (d. 1387 CE) was a citizen of the last great Hindu empire of Vijayanagara, a medieval Hindu kingdom of South India with a full blown caste system, Bhakti/ Tantric Hindu religion, a tipical [sic] climate dominated by monsoon, and an economy based on rice, agriculture, crafts, and trade. This is quite diferent from being a member of one of the small tribal, pastoral societies of the Enastern [sic] Panjab, without or with only an incipient caste system, a Pre-Hindu religion, a cold winter, no real monsoon, without cities, and with an economy based on cattle herding. While medieval commentaries can help us in understanding the ritual and some of the grammar, syntax, and the general background of the texts, they cannot be relied on for the exact interpretation of individual words, of Brāhmaṇa sentences, and even less for the meaning of the archaic mantra-s of the original meaning of the rituals, and of Vedic religion and myth in general.” The idea evidently has tremendous appeal for Witzel, for it is repeated verbatim in Michael Witzel, “Textual Criticism in Indology and in European Philology during the 19th and 20th Centuries,” Electronic Journal of Vedic Studies 21, no. 3 (2014): 32–33. A few pages later, he again emphasizes: “A text such as the Ṛgveda cannot be understood if one does not know something about cattle, the historical climate of the Panjab, pre-state tribal societies and their social systems, about its complex system of Indo-European and Indo-Iranian poetics, about oral composition, canon formation and the techniques of critically editing Sanskrit texts—in other words: by a study of a culture through its texts but making use of various Hilfswissenschaften.” Ibid., 57. 341 342 Martin Luther (1483–1516) Fourfold sense of scripture (quadriga) Catholic intellectual tradition and spiritual ministry Karl Lachmann (1793–1851) Sensus spiritualis (mysticus) Sensus litteralis sive historicus Gottfried Hermann (1772–1848) Moriz Haupt (1808–74) Seperation of the Traditional The institution of a new clergy realia from the text canonical authority Don Mabillon (1632–1707) J. S. Semler (1725–91) Bernard de Monfaucon C. F. Baur (1792–1860) (1655–1741) Recovery of the literal Hermannn meaning through the Rudolf von Roth (1821–95) Oldenberg (1854–1920) historical-critical method Freedom to define new The sense of rivalry with the tradition ideals of scholarship The wistful glance backward The documentary impulse at classical philology GERMAN INDOLOGY A genealogy of German Indology, illustrating how the discipline emerges in the wake of Luther’s theology and also takes up nebulous—and barely understood—impulses from other disciplines (especially classical philology) 34 Appendices 1. THE VOLUMES OF THE CRITICAL EDITION The Mahā bhā rata critical edition project had 11 editors, four of whom were also gen- eral editors. The 11 editors (in order of the parvans or books edited) were Vishnu Sitaram Sukthankar (Ā di and Ā raṇyaka parvans), Franklin Edgerton (Sabhā parvan), Raghu Vira (Virā ṭaparvan), Sushil Kumar De (Udyoga and Droṇa parvans), Shripad Krishna Belvalkar (Bhīṣma, Ś ā nti, Ā śramavā sika, Mausala, Mahā prasthā nika and Svargā rohaṇa parvans), Parashuram Lakshman Vaidya (Karṇaparvan), Ramachandra Narayan Dandekar (Ś alya and Anuśā sana parvans), Hari Damodar Velankar, Raghunath Damodar Karmarkar (Ā śvamedhikaparvan) and Vasudeva Gopal Paranjpe (Strīparvan). Vaidya also edited the Harivaṃśa and the Pratīka Index to the Mahā bhā rata. The four general editors, V. S. Sukthankar (1925–43), S. K. Belvalkar (1943–61), P. L. Vaidya (1961–6) and R. N. Dandekar (joint general editor, 1957–61), also oversaw the other volumes completed during their tenures. The volumes oicial bore the title The Mahābhārata for the First Time Critically Edited by Vishnu S. Sukthankar, followed by the parvan name (for example, The Ādiparvan) and the addition: Being the First Book of the Mahābhārata, The Great Epic of India, for the First Time Critically Edited by Vishnu S. Sukthankar (the ordinal number obviously changing depending on the book). The Ādiparvan appeared in six fascicules between 1927 and 1932, with the sev- enth fascicule and complete volume appearing in 1933. It has since been reprinted in two volumes making up volume 1, parts 1 and 2 of the critical edition. The Sabhāparvan was irst published in 1944 as volume 2 of the critical edition. The Āraṇyakaparvan appeared as the 11th fascicule in 1941 and the 12th fascicule in 1942. It has since been reprinted in two volumes making up volumes 3 and 4 of the critical edition. The Virāṭaparvan was irst published in 1936 as volume 5 of the critical edition. The Udyogaparvan appeared in two fascicules in 1937 and 1940 and has since been reprinted as volume 6 of the critical edition. The Bhīṣmaparvan appeared as the 15th fascicule in 1945 and the 16th fascicule in 1947. It has since been reprinted as volume 7, the largest single volume of the critical edition. The Droṇaparvan, the second-longest of the Mahābhārata’s books, appeared as the 25th fascicule in 1953, the 28th fascicule in 1955, the irst half of second volume as fascicule 29A in 1957 and the second half of second volume and introduction and pre- liminaries as fascicule 29B in 1958. It has since been reprinted in two volumes making up volumes 8 and 9 of the critical edition. The Karṇaparvan was published as the 20th fascicule in 1950 and the 27th fascicule in 1954. It has since been reprinted as volume 10 of the critical edition. The Śalyaparvan appeared as the 30th fascicule in 1956 and the 31st fascicule 1961. It has since been reprinted as volume 11 of the critical edition. The 34 344 PHILOLOGY AND CRITICISM Sauptika and the Strī parvans were irst published in 1948 and 1956 as parts 1 and 2 of volume 12 of the critical edition, respectively. The Śāntiparvan, the Mahābhārata’s lon- gest book, was published in several parts. Parts of its irst section, the Rājadharmaparvan, were published as the18th and 19th fascicules in 1949 and 1950, whereas the complete Rājadharmaparvan was published in 1961 as volume 13 of the critical edition. The second part to appear was the Āpaddharmaparvan and the Concordance (titled Part II), which appeared as the 21st fascicule in 1950 and the 26th fascicule in 1954. The com- plete Āpaddharma and Concordance was published in 1954 as volume 14 of the critical edition. The Sāntiparvan’s third part, the Mokṣadharmaparvan, was published in two parts: Part III: Mokṣadharma, A and Part III: Mokṣadharma, B. The irst part appeared as the 22nd fascicule in 1951, the 23rd fascicule in 1952 and the 24th fascicule in 1953. It was inally published as the complete Mokṣadharmaparvan, part A in 1954 as volume 15 of the critical edition. The irst part of the Mokṣadharmaparvan, part B appeared in the 24th fascicule in 1953; the remainder appeared in the 25th fascicule in 1954. The complete Mokṣadharmaparvan, part B was inally published in 1954 as volume 16 of the critical edition. Belvalkar’s introduction to the volumes was published as a separate volume in 1966 (as part 2 of volume 16 of the critical edition). The Anuśāsanaparvan was irst published in 1966 in two volumes making up volume 17, parts 1 and 2 of the critical edition. The Āśvamedhikaparvan appeared in two halves: as fascicule 32A in 1958 and as fascicule 32B in 1960. It now constitutes volume 18 of the critical edition. The Āśramavāsika, Mausala, Mahāprasthānika and Svargārohaṇa parvans appeared in 1959 as parts 1–4 of volume 19 of the critical edition. The Harivaṃśa was published in two parts: volume 1, containing the critical text, appeared in 1961, whereas the extensive appendices were published as volume 2 in 1971. Both parts bore the title The Mahābhārata for the First Time Critically Edited by Vishnu S. Sukthankar† (August 1925–January 1943), S. K. Belvalkar (April 1943–March 1961), P. L. Vaidya (since April 1961). The Harivaṃśa, Being the Khila or Supplement to the Mahābhārata, for the First Time Critically Edited by Parashuram Lakshman Vaidya. The Pratīka index to the critical edition, an index of all the pādas (verse-quarters) included in the critical edition of the Mahābhārata whether the constituted text or the critical apparatus (that is, the star and appendix passages), was completed in six volumes between 1967 and 1972 under P. L. Vaidya’s editorship. Volume 6 also included a Pratīka index of the Harivaṃśa, although restricted only to its constituted text. 345 2. EDITIONS BESIDES THE CRITICAL EDITION The principal editions of the Mahā bhā rata before the critical edition were the Bombay, the Calcutta, the Kumbhakonam and the Chitrashala Press (Pune) editions. A  critical edition of the southern recension, published concurrently, has replaced the Kumbhakonam edition as the principal representative of the southern tradition.1 The Bombay and Calcutta editions are of the vulgate. The Bombay edition addition- ally includes Nīlakaṇṭha’s commentary. It is in pothı̄ (horizontal) form and two editions are known: an older edition published by Ganpat Krishnaji in 1863 and a newer one published by Gopal Narayan in 1901.2 The Calcutta edition was the irst printed edition of the Mahā bhā rata and hence constitutes the editio princeps. The Chitrashala Press edition, also known as the Kinjawadekar edition after its editor, is widely used for the text of Nīlakaṇṭha’s commentary. Atmaram Khadilkar, ed. Mahābhārata with Nı̄lakaṇt ̣ha’s Commentary. Bombay: Ganpat Krishnaji’s Press, 1863. Vasudev Balacharya Ainapure, ed. The Mahābhārata with the Commentary Bhāvadı̄pa of Nı̄lakaṇt ̣ha. 6 vols. Bombay: Gopal Narayan & Company, 1901. The Bombay edition is supposedly based on Nīlakaṇṭha’s text, but Sukthankar, who exam- ined the text, notes that it contains many readings and lines not found in the Nīlakaṇṭha manuscripts. The text is printed in large font in the center (when the volume is held open) and Nīlakaṇṭha’s commentary is printed in smaller font on the upper and lower margins of the open pages. The edition is available as a scan online. The Mahábhárata, an Epic Poem Written by the Celebrated Veda Vyása Rishi. Vol. 1, edited by the Pandits attached to the Education Committee. Calcutta:  Education Committee’s Press, 1834. Vol. 2, edited by Nimachand Siromani and Nanda Gopála. Calcutta: Baptist Mission Press, 1836. Vol. 3, edited by Nimachand Siromani, Jaya Gopála Tirkalanka and Ráma Govinda. Calcutta: Baptist Mission Press, 1837. Vol. 4, edited by Nimáichandra Siromani, Rámagovinda  and Rámahari Nyáya Panchánan. Calcutta:  Baptist Mission Press, 1839. Sukthankar refers to this as the best edition of the vulgate. It is printed in Devanā garī characters. Volume 1 contains the Ā di, Sabha and Ā raṇyaka parvans (the latter is called Vanaparvan). Volume 2 contains the Virā ṭa, Udyoga, Bhīṣma and Droṇa parvans. Volume 3 contains the Karṇa, Ś alya, Sauptika, Strī and Ś ā nti parvans. Volume 5 contains the Anuśā sana, Ā śvamedhika, Ā śramavā sika, Mausala, Mahā prasthā nika and Svargā rohaṇa parvans as well 346 346 PHILOLOGY AND CRITICISM as the Harivaṃśa. An index is included as volume 5 but no publication details regarding this volume are forthcoming. The edition is available online. T. R.  Krishnacharya and T.  R. Vyasacharya, eds. Sriman Mahabharatam. A  New Edition Mainly Based on the South Indian Texts with Footnotes and Readings. 19  vols. Bombay:  Javaji Dadaji’s “Nirṇaya-Sā gar” Press, 1906–14. Although published at the Nirnaya Sagar Press in Bombay, this edition was published for T. R. Krishnacharya, the proprietor of the Madhva Vilas Book Depot, Kumbakonam, and is therefore known as the Kumbakonam (or Kumbhakonam) edition. Sukthankar calls it a ine representative of the composite Telugu edition but notes that it is now rendered superluous and obsolete by P. P. S. Sastri’s edition. He also notes, of its third book, the Vanaparvan, that it contains “a good number of speciic Northern passages unknown to the true Southern recension.”3 Scans of this edition are partially available online. Ramachandra Kinjawadekar, ed. Mahābhāratam with the Commentary of Nı̄lakaṇt ̣ha. 6 vols. Pune: Chitrashala Press, 1929–36. The French scholar Madeleine Biardeau (1922–2010) used this edition exclusively. Curiously, Sukthankar does not mention it in his survey of printed editions in the “Prolegomena.”4 Jean- Michel Péterfalvi’s abridged translation into French (with Biardeau’s commentary) is based on this edition. Pandit Ramanarayanadatta Sastri Pandey, ed. Śrı̄manmaharṣi Vedavyāsapraṇı ̄ta Mahābhārata. 7 vols. (including Harivaṃśa). Gorakhpur: Gita Press, no date. This edition, published by Gita Press of Gorakhpur, features a Hindi commentary (printed beneath each śloka). Gita Press volumes typically do not feature the date of publication, so no information is available on when this edition was irst published. The order of the volumes is as follows: volume 1: Ā di and Sabhā parvans, volume 2: Vana and Virā ṭa parvans, volume 3:  Udyoga and Bhīṣma parvans, volume 4:  Droṇa, Karṇa, Ś alya, Sauptika and Strī parvans, volume 5:  Ś ā ntiparvan, volume 6:  Anuśā sana, Ā śvamedhika, Ā śramavā sika, Mausala, Mahā prasthā nika and Svargā rohaṇa parvans. Volume 7 is published under the title Śrı̄manmaharṣi Vedavyāsapraṇı ̄ta Mahābhārata-Khilabhāg Harivaṃśa. Pandey does not mention the manuscripts the edition is based on. The text is supposedly of Nīlakaṇṭha’s edition but with additional verses from the southern tradition. Scans of all seven volumes are available online. P. P.  S. Sastri, ed. The Mahābhārata (Southern Recension) Critically Edited by P.  P. S.  Sastri. 18 vols. Madras: Ramaswamy Sastrulu & Sons, Madras, 1931–33. Sastri’s “southern recension” of the Mahā bhā rata is published in 18 volumes, but the count does not match the Mahā bhā rata’s 18 books. The Ā di, Ā raṇyaka, Droṇa and Anuśā sana parvans are all in two parts, the Ś ā ntiparvan in three parts. The Ś alya, Sauptika and Strī parvans are combined into one volume as are the Ā śvamedhika, Ā śramavā sika, Mausala, Mahā prasthā na and Svargā rohaṇa parvans. In spite of the name, it is not a true critical edition. Sukthankar describes it as a true lineal descendant of the Mahā bhā rata manuscripts of south India. The edition is no longer in print, but scans are available online. 347 3. ENGLISH TRANSLATIONS OF THE MAHĀBHĀRATA Scholars have made ive attempts to translate the entire Mahā bhā rata into English so far.5 Kisari Mohan Ganguli inished the irst complete translation (of the vulgate) in 1896. Manmatha Nath Dutt issued a revised version of this translation, purging Ganguli of many archaisms. Dutt’s main innovation was the addition of verse numbers. When completed, the Chicago edition is likely to replace the Ganguli edition as the standard, as it is based on the critical edition’s text and is the most rigorous in terms of speci- fying the text it is based on, contextualization of the edition and citation of chapter and verse numbers. This is the easiest edition to examine against the constituted text. The second translation (and the only other besides the Chicago edition to translate the crit- ical edition) is Bibek Debroy’s in 10 volumes. The Clay Sanskrit Library translation was abandoned due to inancial problems. The existing volumes are inconsistent as regards the discussion of their text and the editorial practices adopted. The edition translates the vulgate, to avoid duplicating the Chicago edition. Finally, several abridgments and adaptations of the Mahā bhā rata exist, but they are not listed here as they provide only a glimpse into this complex and philosophically rich work. P. C. Roy, ed. The Mahābhārata of Krishna-Dwaipayana Vyasa Translated into English Prose. 11 vols. Calcutta: Bharata Press, 1884–96. This is the oldest and at present one of only two complete translations of the Mahā bhā rata into English. The edition is referred to either as the Roy edition (after its publisher) or as the Ganguli edition (after its translator, K.  M. Ganguli). The translator does not mention the edition or manuscripts the translation is based on, but it is in all likelihood the editio princeps. The translation reads well and luently and Ganguli provides extensive notes on some of the more diicult terms or expressions (often based on Nīlakaṇṭha’s glosses). The edition was reprinted in 12 volumes by Munshiram Manoharlal, New Delhi in 1990 and is still available from this publisher. Manmatha Nath Dutt, trans. A Prose English Translation of the Mahabharata (Translated Literally from the Original Sanskrit Text). 8 vols. Calcutta: Elysium Press, 1895–1905. This translation cleaves very closely to Ganguli’s (see examples below), while attempting to update the latter’s English. Dutt has been criticized for his obvious dependence on Ganguli, but his purpose difered somewhat: whereas Ganguli’s translation is in running prose and does not list the verse numbers, Dutt breaks up the text into numbered units to facilitate comparison with the Sanskrit original (the reprint by Parimal Publishers, Delhi in 1988 348 348 PHILOLOGY AND CRITICISM further aided this goal by adding the Sanskrit text on facing pages). The edition features sum- maries of the adhyāyas, but lacks interpretive aids such as an introduction or Ganguli’s exten- sive notes. Dutt also provides no information on his text or how he tackled diicult passages (in contrast to Ganguli, who almost always explains that he had recourse to Nīlakaṇṭha or other works, etc.). J. A.  B.  van Buitenen, trans. The Mahābhārata. 3  vols. Chicago:  University of Chicago Press, 1973–78. Continued by:  James L.  Fitzgerald, trans. The Mahābhārata, vol. 7. Chicago: University of Chicago Press, 2004. This edition is also referred to as the Chicago edition after its publisher. Thus far, six parvans of the Mahā bhā rata (and the irst part of a seventh) have appeared: the Ā di, Sabhā, Ā raṇyaka, Virā ṭa and Udyoga parvans and the Strī and Ś ā nti parvans (up to the end of the Rā jadharmaparvan). The translation is of the critical edition text and has largely replaced the older Roy edition. Van Buitenen has been criticized for his “medievalizing” expressions, and theological nuances or literary references are sometimes lost in his work. The remaining volumes will appear from diferent translators: at the time of the writing of this book, Gary Tubb had joined the project as associate general editor, and parvans 14–18 were due to be the next volume to appear (with Frederick M. Smith as translator). The following is a ten- tative list of forthcoming volumes with their translators:  Bhīṣmaparvan (David Gitomer), Droṇaparvan (Gary Tubb), Karṇaparvan (Adam Bowles), Anuśā sanaparvan  (David Brick) and the remaining parts of the Ś ā ntiparvan (James L. Fitzgerald). Bibek Debroy, trans. The Mahābhārata. 10 vols. New Delhi: Penguin Books India, 2010–14. This edition, targeted at the nonacademic market, eschews the paraphernalia of a scholarly edition such as diacritics, references to the Sanskrit text and an apparatus. Names are spelled phonetically (for example, Souti instead of Sauti, Droupadi instead of Draupadī, etc.). The same introduction repeats in all 10 volumes. The edition presents an accessible rendition of the text, supplemented with a minimum of notes. Debroy makes his preference for the critical edition text clear. Apart from the unconventional spelling, Debroy is scrupulously accurate in the rendering of names (that is, he does not substitute one epithet for another). Maha Bhárata. 15 vols. Clay Sanskrit Library. New York: New York University Press and the JJC Foundation, 2005–9. No general editor is mentioned for this series, which was unfortunately abandoned after 15 volumes (of a projected 32) appeared in it. The text is based on the Kinjawadekar edition. The individual translators emended the text in the case of unclear readings, but no gen- eral principles were evolved (the emendations and the translators’ rationale for them are listed at the back of the individual volumes). The critical apparatus and notes are minimal. The unusual transliteration scheme, though intended to make the volumes easier for novice readers, adds a further layer of diiculty. The following volumes are currently available: Paul Wilmot, trans. Maha·bhárata Book II: The Great Hall. 2006. William Johnson, trans. Maha·bhárata Book III: The Forest, volume 4 of 4. 2005. Kathleen Garbutt, trans. Maha·bhárata Book IV: Viráta. 2007. 349 APPENDICES 349 Kathleen Garbutt, trans. Maha·bhárata Book V: Preparations for War, volume 1 of 2. 2008. Kathleen Garbutt, trans. Maha·bhárata Book V: Preparations for War, volume 2 of 2. 2008. Alex Cherniak, trans. Maha·bhárata Book VI: Bhishma, volume 1 of 2. 2008. Alex Cherniak, trans. Maha·bhárata Book VI: Bhishma, volume 2 of 2. 2009. Vaughan Pilikian, trans. Maha·bhárata Book VII: Drona, volume 1 of 2. 2006. Vaughan Pilikian, trans. Maha·bhárata Book VII: Drona, volume 2 of 2. 2009. Adam Bowles, trans. Maha·bhárata Book VIII: Karna, volume 1 of 2. 2007. Adam Bowles, trans. Maha·bhárata Book VIII: Karna, volume 2 of 2. 2008. Justin Meiland, trans. Maha·bhárata Book IX: Shalya, volume 1 of 2. 2005. Justin Meiland, trans. Maha·bhárata Book IX: Shalya, volume 2 of 2. 2007. Kate Crosby, trans. Maha·bhárata Books X & XI: Dead of the Night and The Women. 2009. Alex Wynne, trans. Maha·bhárata Book XII: Peace (Part 2: The Book of Liberation), volume 3 of 5. 2009. For comparison, excerpts of the irst few verses of the principal translations are reprinted here. Every efort was made to retain the spelling, diacritic and punctuation conventions of the original. Kisari Mohan Ganguli: Manmatha Nath Dutt: Om! Having bowed down to Narayana and Having saluted the Supreme Deity (Narayana), and the Nara, the most exalted male being, and also to highest of all male beings (Nara) and also the Goddess of the goddess Saraswati, must the word success be Learning (Saraswati), let us cry success! uttered. 1–2. One day when the great sages of hard Ugra-srava, the son of Lomaharshana, surnamed austerities, who had been present at the twelve Sauti, well-versed in the Puranas, bending with years’ sacriice of Kulapati Saunaka, were humility one day approached the great sages of comfortably sitting in the Naimisharanya, Rishi rigid vows, sitting as their ease who had attended Lomaharshana’s son, Ugrasrava, popularly the twelve years’ sacriice of Saunaka, surnamed known as Souti—well-read in the Puranas—came Kulapati, in the forest of Naimisha. [...] to them with all humility. [...] The Rishis replied: “The Purana, irst The Rishis replied: promulgated by the great Rishi Dwaipayana, 17. The Purana which was irst told by the and which after having been heard both by the illustrious sage, Dwaipayana, and which was gods and the Brahmarshis was highly esteemed, greatly esteemed by the celestials and Brahmarsis which is the most eminent narrative that exists when they heard it, diversiied both in diction and division, possessing 18. And which, being full of various dictions subtile meanings logically combined and and divisions, is (undoubtedly) the most eminent embellished from the Vedas, is a sacred work. narrative (amongst all narratives) that exist, Composed in elegant language, it includeth containing (as it does) subtle and logically the subjects of other books. It is elucidated by combined meanings, enriched with (the essences other Sastras, and comprehendeth the sense of of) the Vedas, is a sacred work. the four Vedas. We are desirous of hearing that 12–21. It is composed in beautiful language, history, also called Bharata, the holy composition and it includes all other works. It is explained of the wonderful Vyasa, which dispelleth the by all Sastras, and contains the sense of the four fear of evil, just as it was cheerfully recited by Vedas. (But Souti,) we desire to hear Bharata, the the Rishi Vaishampayana, under the direction sacred history that drives away all fear—the holy of Dwaipayana himself, at the snake-sacriice of composition of the great Vyasa—just as it was Raja Janamejaya?” beautifully narrated at the great snake-sacriice of Raja Janamejaya by Rishi Vaishampayana as directed by Krishna Dwaipayana himself. 350 350 PHILOLOGY AND CRITICISM J. A. B. van Buitenen: Bibek Debroy: The Bard shall intone the song of the Triumph after “Jaya” must be recited after having bowed in having bowed to Nara and Nārāyaṇa, supreme among obeisance before Narayana and also Nara, the men, and to the Goddess Sarasvatı̄. supreme human being, and also the goddess 1.1 The Bard Ugraśravas, the son of Sarasvati. Lomaharṣaṇa, singer of the ancient Lore, once The great sages, performers of diicult came to the Naimiṣa Forest, where the seers of austerities, were present at the twelve-year strict vows were sitting together at the Twelve- sacriice of Kulapati Shounaka and were year Session of family chieftain Ś aunaka. [...] comfortably seated in Naimisharanya. The seers said: Ugrashrava, the son of Lomaharshana and the 15 Tell us that ancient Lore that was related by son of a suta, learned in the Puranas, and also the eminent sage Dvaipā yana, which the Gods known as Souti, once approached them, bowing and brahmin seers honored when they heard it! in humility. [...] That divine language of the sublime Histories, The sages replied: “Tell us that ancient story in all the varieties of words and books, the that was told by the supreme sage Dvaipayana, sacred Account of the Bhā ratas, that language that which was worshipped by the gods and the of complex word and meaning, ruled by brahmarshis when they heard it—and that which reinement and reinforced by all sciences, which is full of wonderful words and divisions and is Vaiśaṃpā yana, at Dvaipā yana’s bidding, repeated the supreme of narratives, with subtle meanings truthfully to the satisfaction of King Janamejaya and logic, adorned with the essence of the Vedas. at the king’s sacriice. We wish to hear that Grand That sacred history of the Bharatas is beautiful Collection, now joined to the Collections of the in language and meaning, and includes all other Four Vedas, which Vyā sa the miracle-monger works. All the shastras add to it and that sacred compiled, replete with the Law and dispelling all composition of great Vyasa has been added danger of evil! to the four Vedas. We wish to hear that holy collection, that drives away fear of sin, just as it was recited at King Janamejaya’s sacriice by Vaishampayana.” 351 4. HOW TO USE THE CRITICAL APPARATUS The critical edition comprises two parts: the constituted text and the critical apparatus. The constituted text is running text and can be read like any other text. It is whole and complete within itself and represents the editors’ best surmise of the text of the arche- type (the source from which all present-day manuscripts examined for the edition  are descended). The critical apparatus consists of the variant readings (also called variae lectiones) contained in one or more manuscripts, details of other changes (transpositions of sections or passages, names of chapters and numbering) and sometimes details about the manuscripts’ physical features (missing folios, change of hand, ink markings, mar- ginal insertions and the like). The critical apparatus is primarily a record of the variants found in the witnesses, but it is also a valuable tool to control and reconstruct the editors’ decisions. Since the editor will have explained his principles at the outset in the introduc- tion (for example, when two manuscripts disagree, he will always prefer the reading of the irst), the reader can in most cases understand why the editor selected the reading he did. The Mahā bhā rata critical edition also features an expanded critical apparatus in the form of an appendix:  since the diferent Mahā bhā rata manuscripts vary not just in terms of single readings or lines but sometimes by entire chapters and/or narratives, Sukthankar moved all such longer passages to an appendix at the end of the volume—a practice all the later editors followed. In contrast, the shorter insertions were printed along with the variant readings below the running text. These passages are marked with asterisks (for this reason, they are often called “star passages”) and numbered sequen- tially. Thus, 1* is the irst insertion in the Ā diparvan (the count resets in every volume so one must specify the volume in question). Likewise, the longer appendix passages are numbered sequentially and typically cited as App. 1, no. 1 and so forth (as with the star passages, the count resets in every volume). The Mahā bhā rata critical edition has a negative apparatus. That means the critical apparatus only records the variants and not the manuscripts containing  the reading adopted for the constituted text. To know which manuscripts the reading of the consti- tuted text is based on one must know the manuscripts collated for that section of the text. Ideally, the editor will have provided this information in a table like Sukthankar’s table of the manuscripts collated for the critical edition of the Ā diparvan (see later). If he has not, the reader must glean this information from elsewhere. Merely reading the list of manuscripts forming the critical apparatus for the edition or the volume is insuicient, since many manuscripts were fragmentary and/or were added or dropped throughout the edition. Thus, the reader must check at the beginning of the chapter if the editor 352 352 PHILOLOGY AND CRITICISM has listed additional information about the manuscripts used for that section; occasion- ally, she may need to leaf back several pages to see if the editor noted a change to his manuscripts from some previous section onward. Once she has this information to hand, she can proceed: she consults the critical apparatus for the manuscripts for which variant readings are recorded, then eliminates these manuscripts from the list of manuscripts used for this section of the text. In this way, she arrives at the manuscripts that are unac- counted for in the critical apparatus. These must be the ones whose reading the editor adopted for the constituted text. Explanations of the various abbreviations and diacrit- ical marks used (for example, marg. for a marginal insertion) can be found by consulting the respective editors’ tables (conveniently compiled in one table in Appendix 13). For instance, lines 1.1.1 of the critical edition read:  lomaharṣaṇaputra ugraśravāḥ sūtaḥ paurāṇiko naimiṣāraṇye śaunakasya kulapater dvādaśavārṣike satre | | Consulting the critical apparatus, the reader inds that the editor recorded the following variants for this line: K3 subst. for this passage lomaharṣaṇaputrastu naimiṣāraṇyavāsinaḥ . S rom˚ (T2 G6 śrı̄rom˚). D10 ˚pada for ˚putra. K0 sauti-; K1.2.5 D7 sūta-; K4.6 V1 B (B4 as in K0) D (except D14; D13 missing) T1 sautiḥ . D4 (by corr.).12 S Nīlp naimiśā˚. K2 ˚raṇya(vā)nivāsinaḥ (for ˚raṇye). D14 śaunakakula˚. D14 ins tu brahmarṣe vartamāne after ˚rṣike. D10–12 S (except G1) ins. vartamāne (G2 prava˚), while D14 ins. tatra after satre. From the table of manuscripts collated for this section of the text, the reader knows that the editor had the following manuscripts before him: K0–6, V1, B1–4, Da, Dn, Dr, D1–12.14, T1.2, G1–7 and M1–4. To aid herself, the reader may draw up the following scheme: lomaharṣaṇaputra ugraśravāḥ sūtaḥ paurāṇiko naimiṣāraṇye śaunakasya kulapater dvādaśavārṣike satre | | K0 sauti- K1 sūta- K2 sūta- ˚raṇya(vā)nivāsinaḥ (for ˚raṇye) K3 Subst. for this passage the foll. line: lomaharṣaṇaputrastu naimiṣar̄ aṇyavāsinaḥ K4 sautiḥ K5 sūta- K6 sautiḥ V1 sautiḥ B1 sautiḥ B2 sautiḥ B3 sautiḥ B4 sauti- Da sautiḥ Dn sautiḥ Dr sautiḥ D1 sautiḥ D2 sautiḥ D3 sautiḥ 35 APPENDICES 353 D4 sautiḥ (by corr.) naimiśā˚ D5 sautiḥ D6 sautiḥ D7 sūta- D8 sautiḥ D9 sautiḥ D10 ˚pada (for ˚putra) sautiḥ ins. vartamāne after satre D11 sautiḥ ins. vartamāne after satre D12 sautiḥ naimiśā˚ ins. vartamāne after satre D14 śaunakakula˚ ins. tu brahmarṣe vartamāne after ˚rṣike ins. tatra after satre T1 rom˚ sautiḥ naimiśā˚ ins. vartamāne after satre T2 śrı̄rom˚ naimiśā˚ ins. vartamāne after satre G1 rom˚ naimiśā˚ G2 rom˚ naimiśā˚ ins. pravartamāne after satre G3 rom˚ naimiśā˚ ins. vartamāne after satre G4 rom˚ naimiśā˚ ins. vartamāne after satre G5 rom˚ naimiśā˚ ins. vartamāne after satre G6 śrı̄rom˚ naimiśā˚ ins. vartamāne after satre G7 rom˚ naimiśā˚ ins. vartamāne after satre M1 rom˚ naimiśā˚ ins. vartamāne after satre 354 354 PHILOLOGY AND CRITICISM M2 rom˚ naimiśā˚ ins. vartamāne after satre M3 rom˚ naimiśā˚ ins. vartamāne after satre M4 rom˚ naimiśā˚ ins. vartamāne after satre Nīlp. naimiśā˚ Eliminating manuscripts for which variants have been entered, the reader can see that no variants are entered for the K manuscripts except K3 and for V1, B and all the D manuscripts except D10. The editor’s reading of lomaharṣaṇaputra ugraśravāḥ must hence be based on these manuscripts (that is, K [except K3], V1, B and D  [except D10]). Proceeding in the same way, the reader will see that the editor’s choice of sūtaḥ paurāṇiko must be based on the manuscripts K1, K2, K5, D7 and all the S manuscripts with the exception of T1. By repeating these steps as often as is necessary, she can identify the reading of the manuscripts for which no variants are entered. In fact, she can recon- struct not only the reading of these manuscripts but also of those for which variants are entered. For instance, the reading of the M group must be:  romaharṣaṇaputra ugraśravāḥ sūtaḥ paurāṇiko naimiśāraṇye śaunakasya kulapater dvādaśavārṣike satre vartamāne. Merely by using the constituted text along with the critical apparatus, the reader can thus gain a complete understanding of the entire manuscript tradition. 35 5. HOW EDITORS RECONSTRUCTED THE READING OF THE ARCHETYPE The Mahā bhā rata problem, Sukthankar famously remarked, is a problem sui generis.6 The canons of classical philology, developed to deal with the reconstruction of the archetypes of texts that difered only in a handful of variants and whose copies are more or less faithful transcripts of their sources, are strained to the breaking point in the case of the Mahā bhā rata. The epic, as Sukthankar comments, is a text with “about a dozen, more or less independent, versions, whose extreme types difer, in extent, by about 13,000 stanzas or 26,000 lines; a work which, for centuries, must have been growing not only upwards and downwards, but also laterally, like the Nyagrodha tree, growing on all sides.”7 If he could nonetheless apply the genealogical-reconstructive method and recover an archetype, it is because he had perfectly penetrated its under- lying principle: in essence, he treated the expansions of the text as variant readings and its interpolations as errors. The correct “reading” thus was always the text as preserved in its most conservative, least ampliied witnesses,8 and the reading of this text (that is, the text chosen as likely the most original) was always reconstructed on the basis of the simple principle that two witnesses at a great distance from each other are likely to have little or no possibility of communication and that, if they agree with each other (and this agreement is such that it cannot be attributed to chance, that is, to polygenesis of innovations), then their reading is likely the original.9 Figure 70 ofers insight into how the editors, using the “simple rules and calculations of probability” Glenn Most speaks of, reconstructed the reading of the archetype.10  356 newgenrtpdf ARCHETYPE READING OF N & S READING OF , , and M lomahar a aputra ugra rav [crux] paur iko naimi ra ye aunakasya kulapater dv da av r N satre lomahar a aputra ugra rav sauti paur iko naimi ra ye aunakasya kulapater dv da av r ike satre lomahar a aputra** ugra rav sauti paur iko naimi ra ye aunakasya kulapater 1.1.1A* lomahar a aputra ugra rav s ta paur iko dv da av r ike*** satre naimi ra ye aunakasya kulapater dv da av r ike satre romahar a aputra ugra rav s ta paur iko naimi ra ye aunakasya kulapater dv da av r ike S satre vartam ne**** romahar a aputra ugra rav s ta paur iko naimi ra ye aunakasya kulapater dv da av r ike satre vartam ne M romahar a aputra ugra rav s ta paur iko naimi ra ye aunakasya kulapater dv da av r ike satre vartam ne K0 sauti-; K1.2.5 s ta-; K4.6 sauti K3 subst. for this passage the foll. line: lomahar a aputrastu naimi ra yav sina K2 º ra ya(v )niv sina (for º ra ye) ** D10 º p da for º putra * Mss. D7 s ta-; V1 B (B4 as in K0) D (except D14; D13 collated for missing) sauti this portion D4 (by corr.).12 N lp. naimi º of the text: D14 aunakakulaº *** K0 6 V1 D14 ins tu brahmar e vartam ne after º r ike B1 4 Da Dn D10 12 ins. vartam ne, while D14 ins. tatra after satre Dr D1 12.14 T2 G6 r romº T1.2 G1 7 T1 sauti **** M1 4 G2 pravaº (G1 as in N) Figure 70 Reconstructing the reading of 1.1.1A 357 6. HOW TO CITE THE MAHĀBHĀRATA The Mahā bhā rata follows two or three diferent types of division. The broadest is the articulation of the work into 18 parvans or books (24 in the southern recension). Concurrent with this division, there is a second division into 100 upaparvans or sub- parvans. Finally, the Mahā bhā rata also features adhyāyas or chapters. The chapters con- tain verses or ślokas. When citing a verse, it is customary to specify the book, the chapter and the verse number (for example, Mahā bhā rata 1.1.1, which represents the irst verse of the irst chapter of the irst book of the Mahā bhā rata or the irst verse of the entire epic). When citing a passage, it is customary to cite beginning and ending line num- bers, occasionally with lowercase letters indicating the pāda or verse-quarter (for example, Mahā bhā rata 1.1.1–15, Mahā bhā rata 1.1.1c–2d, etc.) The upaparvans play almost no role in citation, but as they form well-deined units they are often mentioned in discussion (for example, “in the Aṇukramaṇiparvan” or “in the Paulomaparvan”). By contrast, the adhyāyas, though used in citation, are rarely men- tioned in discussion, as the divisions do not correspond to any one theme or narrative. For instance, the Ā stīkaparvan includes 40 adhyāyas, but many of its narratives extend across several adhyāyas; their end also does not necessarily coincide with the end of an adhyāya. As the book, chapter and verse numbers remain constant across translations of the same edition, it is customary to cite only these igures even for translations (for example, Mahā bhā rata 1.1.1 rather than the page number when citing from Van Buitenen’s trans- lation of the Ā diparvan). Note that these igures are not constant across the diferent editions so that Mahā bhā rata 1.45.10 might refer to a diferent line in the critical edition than in the vulgate. For this reason, it is always important to specify at the outset of one’s work which edition of the Mahā bhā rata one is relying on. 358 359 7. THE EXTENT OF THE MAHĀ BHĀ RATA’S BOOKS The traditional extent of the Mahā bhā rata is 100,000 verses, a igure that is computed by adding the Harivaṃśa, the khila or appendix to the Mahā bhā rata, to its 18 books. This igure is only approximate, as the diferent manuscripts and printed editions difer greatly in extent. P. P. S. Sastri, who computed the extent of the diferent editions (and the four manuscripts used for his edition), lists the following igures: 1. Bombay edition: 84,829 verses; 2. Calcutta, Telugu and Kumbhakonam editions: 84,836 verses;11 3. Grantha edition: 95,379 verses;12 4. Palm-leaf manuscripts क, ख, ग: 95,824 verses;13 and 5. Palm-leaf manuscript अ: 95,286 verses.14 The Kumbhakonam edition gives its own extent as 96,645, thus most closely approxi- mating the 100,000-verse ideal.15 No igures are known for the vulgate, but it is possible that, in some manuscripts at least, it approaches the canonical 100,000-verse mark. The critical edition reduced the text’s length to 73,640 verses and 297 prose units arranged in 1,995 adhyāyas or chapters (plus an additional 6,073 verses in 118 chapters in the Harivaṃśa).16 The breakdown of these verses is as follows:17 Parvan Extent Number of adhyaˉyas Ā diparvan 7,196 verses and 156 prose units 225 Sabhā parvan 2,390 verses 72 Ā raṇyakaparvan 10,141 verses and 84 prose units 299 Virā ṭaparvan 1,834 verses 67 Udyogaparvan 6,063 verses 197 Bhīsṃaparvan 5,406 verses 117 Droṇaparvan 8,112 verses 173 Karṇaparvan 3,871 verses 69 Ś alyaparvan 3,293 verses 64 Sauptikaparvan 772 verses 18 Strīparvan 730 verses 27 Ś ā ntiparvan, within which: 12,890 verses and 57 prose units 353 Rā jadharmaparvan 4,512 verses 128 Ā paddharmaparvan 1,560 verses 39 Mokṣadharmaparvan 6,737 verses and 57 prose units 186 Anuśā sanaparvan 6,536 verses 154 360 360 PHILOLOGY AND CRITICISM Parvan Extent Number of adhyaˉyas Ā śvamedhikaparvan 2,862 verses 96 Ā śramavā sikaparvan 1,062 verses 47 Mausalaparvan 273 verses 9 Mahā prasthā nikaparvan 106 3 Svargā rohaṇaparvan 194 5 Harivaṃśa 6,073 verses 118 361 8. THE 18 PARVAN S AND 100 UPAPARVAN S OF THE MAHĀBHĀRATA Besides the division into 18 parvans, the Mahā bhā rata also features a second division, namely, into 100 upa or sub-parvans (sometimes also called antaḥ parvans).18 These 100 upaparvans are mentioned at two places in the Parvasaṃgrahaparvan of the Mahā bhā rata (the second book of the Ā diparvan): once in the list of the 100 books from 1.2.34a–69c and once in the summary of the books from 1.2.72a–234a.19 This division is clearly a northern feature, for the southern manuscripts rarely, if at all, name the sub-parvan in their colophons.20 The names listed in the second column therefore rarely correspond to the southern manuscripts.21 The adhyāya divisions given are those of the critical edition.22 Parvan Upaparvan1 Adhyā ya(s) Ā diparvan (1) Anukramaṇī 1 The Book of the Beginning The List of the 100 Books (2) Parvasaṃgraha 2 The Summary of the Books (3) Pauṣya 3 The Book of Pauṣya (4) Pauloma 4–12 The Book of Puloman (5) Ā stīka 13–53 The Book of Ā stīka (6) Ā divaṃśāvataraṇa2 54–61 The Descent of the First Generations (7) Saṃbhava 62–123 The Origins (8) Jatugṛhadāha 124–38 The Fire in the Lacquer House (9) Haiḍimba3 139–42 The Slaying of Hiḍimba (10) Bakavadha 143–52 The Slaying of Baka (11) Caitraratha 153–73 The Book of Citraratha (12) Svayaṃvara4 174–89 The Choice of a Bridegroom (13) Vaivāhika 190–92 The Wedding 362 362 PHILOLOGY AND CRITICISM Parvan Upaparvan Adhyā ya(s) (14) Vidurāgamana 193–98 The Arrival of Vidura (15) Rājyalambha 199 The Winning of a Kingdom (16) Arjunavanavāsa 200–10 Arjuna’s Sojourn in the Forest (17) Subhadrāharanạ 211–12 The Abduction of Subhadrā (18) Haraṇahārika 213 The Fetching of the Nuptial Gift (19) Khāṇdavadāha 214–25 The Burning of the Khāṇḍava Forest Sabhāparvan (20) Sabhā 1–11 The Book of the Assembly Hall The Assembly Hall (21) Mantra 12–17 The Council (22) Jarāsaṃdhavadha 18–22 The Slaying of Jarāsaṃdha (23) Digvijaya 23–29 The Conquest of the World (24) Rājasū yika 30–32 The Royal Consecration (25) Arghābhiharanạ 33–36 The Taking of the Guest Gift (26) Ś iśupālavadha 37–42 The Slaying of Ś iśupāla (27) Dyū ta 43–65 The Gambling Match (28) Anudyū ta 66–72 The Sequel to the Gambling Ā raṇyakaparvan (29) Ā raṇyaka 1–11 The Book of the Forest The Forest Teachings (30) Kirmīravadha 12 The Slaying of Kirmīra (31) Kairāta 13–42 The Battle of Arjuna and the Mountain Man (32) Indralokābhigamana 43–79 The Journey to the World of Indra (33) Tīrthayātrā 80–153 The Pilgrimage (34) Jaṭāsuravadha 154 The Slaying of Jaṭāsura (35) Yakṣayuddha 155–72 The War of the Yakṣas (36) Ā jagara 173–78 The Boa (37) Mārkaṇḍeyasmāsya 179–221 The Meeting with Mārkaṇḍeya 36 APPENDICES 363 Parvan Upaparvan Adhyā ya(s) (38) Draupadīsatyabhāmāsaṃvāda 222–24 The Dialogue of Draupadī and Satyabhāmā (39) Ghoṣayātrā 225–43 The Cattle Expedition (40) Mṛgasvapnabhaya 244 The Deer in the Dream (41) Vrīhidrauṇika 245–47 The Measure of Rice (42) Draupadīharanạ 248–82 The Abduction of Draupadī (43) Kuṇḍalāharaṇa 283–94 The Theft of the Earrings (44) Ā raṇeya 295–99 The Fire Drilling Woods Virāṭaparvan (45) Vairāṭa 1–12 The Book of Virāṭa The Book of Virāṭa (46) Kīcakavadha 13–23 The Slaying of Kīcaka (47) Gograhanạ 24–63 The Cattle Robbery (48) Vaivāhika 64–67 The Wedding of Abhimanyu and the Daughter of Virāṭa Udyogaparvan (49) Udyoga 1–21 The Book of the Efort The Book of the Efort (50) Saṃjayayāna 22–32 The Coming of Saṃjaya (51) Prajāgara 33–41 The Sleeplessness (52) Sānatsujāta 42–45 The Book of Sanatsujāta (53) Yānasandhi 46–69 The Suing for Peace (54) Bhagavadyāna 70–137 The Coming of Kṛṣṇa (55) Vivāda5 138–48 The Quarrel (56) Niryāṇa6 149–52 The Marching Out (60) Bhīṣmābhiṣecana7 153–55 The Wonderful Installation of Bhīṣma (58) Ulū kadū tagamana8 157–60 The Arrival of the Messenger Ulū ka (57) Rathātirathasaṃkhyā 170–69 The Warriors and the Greater Warriors (59) Ambopākhyāna 170–97 The Narrative of Ambā 364 364 PHILOLOGY AND CRITICISM Parvan Upaparvan Adhyā ya(s) Bhīsṃaparvan (61) Jambū khaṇḍavirmāṇa 1–11 The Book of Bhīṣma The Creation of the Continent of Jambū (62) Bhū mi 12–13 The Earth (63) Bhagavadgītā 14–40 The Bhagavadgītā (64) Bhīṣmavadha 41–117 The Slaying of Bhīṣma Droṇaparvan (65) Droṇab ̄ hiṣeka 1–15 The Book of Droṇa The Installation of Dronạ (66) Saṃśaptakavadha 16–31 The Slaughter of the Sworn Warriors (67) Abhimanyuvadha 32–51 The Slaying of Abhimanyu (68) Pratijñā 52–60 The Promise (69) Jayadrathavadha 61–121 The Slaying of Jayadratha (70) Ghaṭotkacavadha 122–54 The Slaying of Ghaṭotkaca (71) Droṇavadha 155–65 The Slaying of Dronạ (72) Nārāyaṇāstramokṣa 166–73 The Casting of the Nārāyaṇa Weapon Karṇaparvan (73) Karnạ 9 1–69 The Book of Karṇa The Book of Karnạ Ś alyaparvan (74) Ś alya10 1–16 ́ The Book of S alya The Book of Ś alya (75) Hradapraveśa 17–28 The Entering of the Lake (77) Sārasvata11 29–53 The River Sarasvatī (76) Gadāyuddha 54–64 The Battle of the Bludgeons Sauptikaparvan (78) Sauptika 1–9 The Book of the Sleeping Warriors The Massacre of the Sleeping Warriors (79) Aiṣīka 10–18 The Aiṣīka Weapon Strīparvan (80) Jalapradānika12 1–8 The Book of the Women The Ofering of the Water (81) Strī 9–25 The Women (82) Ś rāddha 26–27 The Funeral Oblation Ś āntiparvan (83) Ā bhiṣecanikaṃ13 The Book of the Peace The Royal Consecration (84) Cārvākanigrahaḥ 14 The Subduing of Carvāka (85) The Distribution of the Houses 365 APPENDICES 365 Parvan Upaparvan Adhyā ya(s) (86) Rājadharma 1–128 The Law of the King (87) Ā paddharma 129–67 The Law of Emergencies (88) Mokṣadharma 168–353 The Law of Salvation Anuśāsanaparvan (89) Ā nuśāsanika15 1–152 The Book of the Instructions The Instructions (90) (Bhīsma)svargārohanạ 153–54 The Ascent to Heaven Ā śvamedhikaparvan (91) Ā śvamedhika16 1–96 The Book of the Horse Sacriice The Horse Sacriice (92) Anugīta17 The Anugītā Ā śramavāsikaparvan (93) Ā śramavāsa 1–35 The Book of the Hermitage The Sojourn in the Hermitage (94) Putradarśana 36–44 The Encounter with the Sons (95) Nāradāgamana 45–47 The Arrival of Nārada Mausalaparvan (96) Mausala 1–9 The Book of the Clubs The Battle of the Clubs Mahāprasthānikaparvan (97) Mahāprasthānika 1–3 The Book of the Great Journey The Great Journey Svargārohaṇaparvan (98) Svargārohanạ 1–15 The Book of the Ascent to Heaven The Ascension to Heaven Harivaṃśa (99) Harivaṃśa The Appendix of the Genealogy of Hari (100) Bhaviṣyat The Book of the Future Notes: 1 The list of sub-parvan names follows the northern recension; the southern recension list difers. As a rule, the southern recension manuscripts do not give the sub-parvan names in the colophons; an over- view of the southern sub-parvans must rely on the Parvasaṃgraha lists of those manuscripts. 2 This sub-parvan is referred to as the Ādivaṃsá ̄ vataranạ parvan at 1.2.34, but this name does not occur in the summary of the Ādiparvan given in the Parvasaṃgrahaparvan from 1.2.72–94. It is likewise not mentioned in the colophons of any of the manuscripts collated for the critical edition. 3 Hiḍimbavadhaparvan in the critical edition. 4 Draupadīsvayaṃvaraparvan in the critical edition. 5 This name is not found in the critical edition, which calls this section the Karṇopanivā daparvan. 6 This sub-parvan is called the Abhiniryā ṇaparvan in the critical edition. 7 The order of the upaparvans hereafter difers in the critical edition. The Ambopā khyā naparvan is the last upaparvan in the critical edition of the Udyogaparvan; the Bhīṣmā bhiṣecanaparvan, number 60 in the Parvasaṃgraha count, moves up, and the order of the Rathā tirathasaṃkhyā and Ulū kadū tagamana parvans, numbers 57 and 58 in the Parvasaṃgraha count, is inverted. 8 This sub-parvan is called the Ulū kayā naparvan in the critical edition. 9 The critical edition refers to this upaparvan as the Karṇavadhaparvan throughout in its header (upper right and upper left corners), though on whose authority is not clear as the colophon at the end of chapter 69 refers to it as the Karṇaparvan. 36 366 PHILOLOGY AND CRITICISM 10 The critical edition refers to this upaparvan as the Ś alyavadhaparvan throughout in its header (upper right and upper left corners), perhaps on the authority of the colophon at the end of chapter 16, which refers to it as the Ś alyavadhaparvan. 11 The Parvasaṃgraha (1.2.59c–60a) does not list these books in their correct order:  the order has been inverted in this table. The colophon at the end of chapter  53 refers to the upaparvan as the Tīrthayatraparvan. 12 The colophon at the end of chapter  8 refers to this parvan as the Viṣokaparvan. This name is adopted in the critical edition, both because all the editors after Sukthankar agreed with him that very little weight was to be attached to the Parvasaṃgraha igures and titles and because all the Strīparvan manuscripts were unanimous in referring to the parvan as the Viṣokaparvan. See Paranjpe, “Introduction,” xxiii. Paranjpe refers to a tradition in Sanskrit drama of ofering characters water for washing their eyes when they shed tears and suggests that jalapradāna may have been “an idiomatic expression for consolation of grief, but when it fell out of use the term Viṣoka was substituted.” Ibid., xxiv. This explanation is not very convincing, since Paranjpe—in agreement with all the other editors of the Mahā bhā rata critical edition—considers the Parvasaṃgraha late; in fact, he rejects the view that the Parvasaṃgraha divisions ought to be binding on the grounds that “the Viṣoka […] [would have been] before the authors of adhys. 1 and 2 of the Ādiparvan.” Ibid., xxiii. If so, it is unlikely that the author(s) of the Parvasaṃgraha used archaic expression, whereas, by the time the Strīparvan was composed, this usage was forgotten and replaced by viṣoka instead. 13 No upaparvans are found corresponding to this name and the next two. The Rā jadharmaparvan is the irst sub-parvan in the Ś ā ntiparvan. However, there are three brief episodes that might be intended:  chapter  39, titled variously Cā rvakavadha, Cā rvakanigrahaḥ , Cā rvakacaritakathanaṃ, Cā rvakacaritaṃ, Cā rvakotpattiḥ and Cā rvakotpattikathanaṃ, and chapter  40, titled variously Yudhiṣṭhirā bhiṣecanaṃ, Yudhiṣṭhirā bhiṣekaḥ , Dharmaputrabhiṣekaḥ and Yudhiṣṭhirarā jyā bhiṣekaḥ , and chapter  44, titled variously Gṛhapravibhā gaḥ (cp. pravibhāgo gṛhāṇam ̄ ̣, 1.2.63c), Gṛhavibhā gā ḥ , Gṛhapraveśaḥ (or veśanaḥ ), Bhīmā dīnā ṃ Gṛhanirū paṇaṃ and Bhrā tṛṇā ṃ Gṛhapraveśaḥ . But note that the order of the irst two chapters is the inverse of what we would expect from the Parvasaṃgraha and, furthermore, that they are chapters rather than sub-parvans. 14 See preceding note. 15 The text does not support this name. The critical edition adopts the name Dā nadharmaparvan for the irst sub-parvan of this parvan on the authority of the manuscripts. The adhyāya igures, accord- ingly, apply to this name. 16 The critical edition does not support this name (which occurs in the Parvasaṃgraha). The crit- ical edition of the Āśvamedhikaparvan contains only one upaparvan and this is called the Aśvamedhaparvan. The editor does not clarify on whose authority he makes this decision. Neither the initial colophons nor the inal colophon refers to the upaparvan name. Thus, all 96 chapters of the Āśvamedhikaparvan are part of the same upaparvan. 17 The Āśvamedhikaparvan includes the poem called the Anugītā from chapters 16–30. The colophons from 16–30 support this name as the name of a separate upaparvan, though not unanimously (a colophon is missing or omitted in chapter  27). From 28–30, the name Anugītā competes with Brā hmaṇagītā as the name of the sub-parvan; from 31–35 it is more often Brā hmaṇagītā (occasion- ally, also Anugītā ). At chapter 36 it reverts to Anugītā (until 50), but the more important name for this section appears to be the adhyāya name Guruśiṣyasaṃvā da (at least until chapter 48 and then again at chapter 50). The colophon for chapter 50 reads guruśiṣyasaṃvādonugı̄tā ca [samāptā] (Ś1 and K1.3); śiṣyānugı̄tā [samāptā] (K2) and anugı̄tā [samāptā] (T, G and M2–4). Thus, if there was ever a dis- tinct sub-parvan called the Anugītā parvan, it would have been from adhyāyas 16–50; the extent of the Anugītā itself, however, is much less clear: it may be considered the portion from 16–26, since the adhyāya name for 27 reads brahmaṇagı̄tā. See also the editor’s comment that “Adhyā yas 16–50 contain the episode of the Anugītā (containing the Anugītā proper, the Brā hmaṇagītā and the Guru-Śiṣya- saṃvā da),” but without trying to distinguish between these three sections more closely. Karmarkar, “Introduction,” xxiv. The critical edition does not feature Anugītā as the name of an upaparvan, in contradiction to the Parvasaṃgraha list. 367 9. THE ARRANGEMENT OF THE PARVAN S IN THE SOUTHERN RECENSION As mentioned earlier, the division of the Mahā bhā rata into 18 parvans and 100 upaparvans applies only to the northern recension. The southern recension has its own division.23 While the southern manuscripts feature the list of 100 upaparvans in the Parvasaṃgraha (in surprising agreement with the northern manuscripts), the books themselves do not rely on this division. For this reason, the attribution of the upaparvans to the parvans in the preceding table is only approximate in the case of the southern recension. This problem is further compounded by the fact that there is little clarity about where the southern parvan divisions occur.24 The printed editions of the southern recension uniformly follow the northern division into 18 parvans and are hence unilluminating. In the following table, we reconstruct the southern recension division using the critical apparatus (these remarks are only approximate, as the regional recensions and even individual manuscripts within the recensions difer in their divisions).25 Because the editors of the critical edition make no efort to mark out the divisions in the southern recension, this list is only provisional. The situation is further complicated by the fact that often the editors list the southern parvan names as sub-parvan names, even though certain details (for example, the fact that the adhyāya count has reset, indicating that it is a case of a parvan rather than a sub-parvan) suggest that the southern manuscripts read a (major) parvan here. The critical edition does feature a parvan at the beginning of the Strīparvan called the Viṣokaparvan, but the editor considers this a sub-parvan of the Strīparvan, rather than an independent parvan in its own right. We did not ind any evi- dence of a diference in the adhyāya count of the southern manuscripts (all of them read the exact same adhyāya number as in the critical edition) that might suggest a break here. It is, of course, entirely possible that the editor of the parvan (H. D. Velankar) overlooked this fact, but only a look at the manuscripts will settle the question. In any case, it is clear that, contrary to A. C. Burnell’s views, there is no division char- acteristic of the southern recension as a whole.26 The Telugu and Grantha manuscripts read 23 or 24 parvans, but their evidence is not unanimous. Whether the manuscripts that do not read 24 parvans do so because of contamination with northern manuscripts or whether they do so because they preserve an earlier division is impossible to settle. The Malayā lam manuscripts explicitly do not read the Gā daparvan (as a major parvan) after the Ś alyaparvan. Their parvan count is closer to the northern igure. Possibly, the practice of writing the Mahā bhā rata in 24 parvans arose in the Grantha manuscripts (typically, the largest) and was introduced into the Malayā lam manuscripts thence, albeit only partially and unsuccessfully. Sukthankar’s statement hence requires revision—if 368 368 PHILOLOGY AND CRITICISM “the South has never completely assimilated the (Northern) division of the epic into the conventional eighteen parvans,”27 it is equally true that the South has never fully assim- ilated the “southern” division into 24 parvans either. There is every reason for preferring the northern division.28 By way of comparison, the overviews of parvans and adhyāyas in the southern editions (P. P. S. Sastri’s Madras edition and T. R. Krishnacharya’s older Kumbhakonam edition) are reproduced in Figures 71, 72 and 73–77. Parvan Adhyā ya(s) Changes in the southern recension Ā diparvan 1–12 This is the irst parvan in the southern count. 13–53 Some manuscripts begin a new parvan (the Ā stīkaparvan) here; the parvan count is now two. G3.6 and M insert a colophon after 13.4. The adhyāya number (at the end of chapter 13) is listed as 2 (G3.6 and M1–5) (M2 sup. lin. sec. m. 1). In the colophon (after chapter 13), G3 omits ādiparvaṇi (in ādiparvaṇi astı̄ke) and interpolates instead parvaṇi after āstı̄ke. M reads iti āstı̄ke. 54–225 The southern recension (that is, T2, G1–5 and M) begins a new parvan (the Saṃbhavaparvan) here; the parvan count is now three. The adhyāya number (at the end of chapter 54) is listed as 1. Sabhāparvan 1–72 This is the fourth parvan in the southern count. Ā raṇyakaparvan 1–299 This is the ifth parvan in the southern count. Virāṭaparvan 1–67 This is the sixth parvan in the southern count. Udyogaparvan 1–197 This is the seventh parvan in the southern count. Bhīsṃaparvan 1–117 This is the eighth parvan in the southern count. Droṇaparvan 1–173 This is the ninth parvan in the southern count. Karṇaparvan 1–69 This is the 10th parvan in the southern count. Ś alyaparvan 1–53.37 This is the 11th parvan in the southern count. 54.1–2 After the inal colophon, T 2.4, G1, G2.3 conclude with śalyaparvaṃ samāptaṃ. G2.3 add gadāparvan. S (except M1; T3 damaged) reads 1–2 after 9.53.37. 54–64 T and G begin a new parvan (the Gadāparvan) here; the parvan count is now 12. T (T3.4 missing) G read the adhyāya number (at the end of chapter 54) as 1; M2–4 as 57. T (T3.4 missing) G read the adhyāya number (at the end of chapter 64) as 11. After the colophon at the end of chapter 63, M1.2 4 add śalyaparvaṃ samāptaṃ. M reads chapter 64 as the irst adhyāya of the Sauptikaparvan. After the inal colophon, T1 concludes with gadāparva samāptaṃ. Hereafter, TG and M will be listed separately. 369 APPENDICES 369 Parvan Adhyā ya(s) Changes in the southern recension Sauptikaparvan 1–9 This is the 13th parvan in the TG count (12th in the M count). T G1–3 read the adhyāya number (at the end of chapter 1) as 1. T G1–3 read the adhyāya number (at the end of chapter 9) as 9. After the inal colophon, G2 adds sauptikaparvaṃ samāptam. In the colophon (at the end of chapter 1), M1 reads the (major) parvan as Ś alyaparvan; the adhyāya number is 67 (of Ś alyaparvan). M2–4 read the adhyāya number (at the end of chapter 1) as 2. In the colophon (at the end of chapter 9), M1 reads the (major) parvan as Ś alyaparvan; the adhyāya number is 75 (of Ś alyaparvan). M2–4 read the adhyāya number (at the end of chapter 9) as 9. M3.4 add samāpta after the sub- parvan (?) or adhyāya name, while M1 adds samāpta after śalya + sauptika. 10–18 T G1.2 begin a new parvan (the Aiṣikaparvan) here; the parvan count is now 14.1 T G1.2 read the adhyāya number (at the end of chapter 10) as 1. After the inal colophon, T2 concludes with aiṣı ̄kaparvaṃ samāptam. M (M4 om.) begins a new parvan (the Aiṣikaparvan) here; the parvan count is now 13. M (M4 om.) reads the adhyāya number (at the end of chapter 10) as 1. Strīparvan 1–27 This is the 15th parvan in the TG count (14th in the M count).2 Ś āntiparvan 1–167 TG and M refer to this parvan as the Rājadharmaparvan; the parvan count is now 16 (15 in M). The Ā paddharmaparvan, a sub-parvan of the Ś āntiparvan in the critical edition, is considered part of the Rājadharmaparvan in the southern manuscripts and the colophons consistently read rājadharma for the parvan name. 168–353 TG and M introduce a new parvan (the Mokṣadharmaparvan) here; the parvan count is now 17 (16 in M). T1 G1.3.6 read the adhyāya number (at the end of chapter 168) as 1; T2 G2 as 15. After the inal colophon, G3 concludes with mokṣadharmaḥ samāptaḥ . M1.5–7 read the adhyāya number (at the end of chapter 168) as 1. After the inal colophon, M1 concludes with śāntiparva samāptaṃ; M7 with mokṣadharma samāptaṃ. Anuśas̄ anaparvan 1–154 This is the 18th parvan in the TG count (17th in the M count). Ā śvamedhikaparvan 1–96 This is the 19th parvan in the TG count (18th in the M count). Ā śramavāsikaparvan 1–47 This is the 20th parvan in the TG count (19th in the M count). 370 370 PHILOLOGY AND CRITICISM Parvan Adhyā ya(s) Changes in the southern recension Mausalaparvan 1–9 This is the 21st parvan in the TG count (20th in the M count). Mahāprasthānikaparvan 1–3 This is the 22nd parvan in the TG count (21st in the M count). Svargārohaṇaparvan 1–15 This is the 23rd parvan in the TG count (22nd in the M count). Notes: 1 G3 is discarded as it does not contain the Aiṣīkaparvan. 2 Contrary to Burnell’s comments, we found no evidence that the Sauptikaparvan was divided into three parvans: the Sauptika, Aiṣīka and Viṣoka parvans. Lüders’s comment that the Viṣokaparvan is, rather, the irst part of the Strīparvan did not help either. The critical edition features a sub-parvan called the Viṣokaparvan from 11.1–8, but the count of adhyāyas does not reset thereafter in any of the southern manuscripts, as we would expect if the Viṣokaparvan was a major parvan rather than a sub-parvan. There are also none of the usual concluding statements such as viṣokaparva samāptam, etc. 371 APPENDICES 371 Figure  71 P. P.  S. Sastri’s scheme of adhyāyas from his critical edition of the southern recension Source: Sastri, ed., The Mahā bhā rata (Southern Recension), xviii. 372 372 PHILOLOGY AND CRITICISM Figure 72 P. P. S. Sastri’s scheme of ślokas from his critical edition of the southern recension Source: Sastri, ed., The Mahā bhā rata (Southern Recension), xix. 37 APPENDICES 373 Figure 73 Parvan divisions and ́loka counts of the Kumbhakonam recension Source: Krishnacharya and Vyasacharya, eds., Sriman Mahabharatam, 4. 374 374 PHILOLOGY AND CRITICISM Figure 74 Parvan divisions and ́loka counts of the Kumbhakonam recension Source: Krishnacharya and Vyasacharya, eds., Sriman Mahabharatam, 5. 375 APPENDICES 375 Figure 75 Parvan divisions and ́loka counts of the Kumbhakonam recension Source: Krishnacharya and Vyasacharya, eds., Sriman Mahabharatam, 6. 376 376 PHILOLOGY AND CRITICISM Figure 76 Parvan divisions and ́loka counts of the Kumbhakonam recension Source: Krishnacharya and Vyasacharya, eds., Sriman Mahabharatam, 7. 37 APPENDICES 377 Figure 77 Parvan divisions and ́loka counts of the Kumbhakonam recension Source: Krishnacharya and Vyasacharya, eds., Sriman Mahabharatam, 8. 378 379 10. OTHER NARRATIVE DIVISIONS In addition to the divisions listed earlier—parvan, upaparvan and adhyāya—the Mahā bhā rata also features several well-deined narrative units, usually known by their own name, which coincide with neither the upaparvan nor the adhyāya divisions. These units are known as upākhyānas or subtales,29 and they are crucial to the epic’s narrative texture as they provide glosses on the main narrative. Many of them are narrated in response to the ethical dilemmas of the main protagonists, thus giving a story-within-a- story weave to the text. Counting all the units called upākhyānas, either in the text or in the Parvasaṃgrahaparvan or in the colophons and/or the running heads for units in the critical edition, Alf Hiltebeitel arrives at the following list:30 1. Ś akuntalā -Upā khyā na 1.62–69 2. Yayā ti-Upā khyā na 1.70–80 3. Mahā bhiṣa-Upā khyā na 1.91 4. Aṇimā ṇḍavya-Upā khyā na 1.101 5. Vyuṣitā ṣva-Upā khyā na 1.112 6. Tapatī-Upā khyā na 1.160–63 7. Vasiṣṭha-Upā khyā na 1.164–68, 173 8. Aurva-Upā khyā na 1.169–72 9. Pañcendra-Upā khyā na 1.189 10. Sunda-Upasunda-Upā khyā na 1.201–4 11. Ś ā rṅgaka-Upā khyā na 1.220–25 12. Saubhavadha-Upā khyā na 3.15–23 13. Nala-Upā khyā na 3.50–78 14. Agastya-Upā khyā na 3.94–108 15. Ṛśyaśṛṅga-Upā khyā na 3.110–13 16. Kā rtavīrya/ Jā madagnya-Upā khyā na 3.115–17 17. Saukanya-Upā khyā na 3.122–25 18. Mā ndhā tṛ-Upā khyā na 3.126 19. Jantu-Upā khyā na 3.127–28 20. Ś yena-Kapotīya-Upā khyā na 3.130–31 21. Aṣṭā vakrīya-Upā khyā na 3.132–34 22. Yavakrīta-Upā khyā na 3.135–39 23. Vainya-Upā khyā na 3.183 24. Matsya-Upā khyā na 3.185 25. Maṇḍū ka-Upā khyā na 3.190 380 380 PHILOLOGY AND CRITICISM 26. Indradyumna-Upā khyā na 3.191 27. Dhundhumā ra-Upā khyā na 3.192–95 28. Pativratā -Upā khyā na 3.196–206 29. Mudgala-Upā khyā na 3.246–47 30. Rā ma-Upā khyā na 3.257–76 31. Sā vitrī-Upā khyā na 3.277–83 32. Ā raṇeyam upā khyā nam yatra dharmo ‘nvaśā t sutam 3.295–99 33. Indravijaya-Upā khyā na 5.9–18 34. Dambhodbhava-Upā khyā na 5.94 35. Ambā -Upā khyā na 5.170–93 36. Viśva-Upā khyā na 6.61–64 37. Tripura-Upā khyā na; Tripura-Vadha-Upā khyā na 8.24 38. [Karṇa-Ś alya-Saṃvā da] Haṃsa-Kā kīya-Upā khyā na 8.28 39. Indra-Namuci-Upā khyā na 9.42 40. Vṛddha-Kumā rī-Upā khyā na 9.51 41. Ṣ oḍaśarā j[ik]a-Upā khyā na 12.29 42. Nā rada-Pā rvata-Upā khyā na 12.30 43. Rā ma-Upā khyā na 12.48–49 44. Mucukunda-Upā khyā na 12.75 45. Uṣṭragrīva-Upā khyā na 12.113 46. Daṇḍa-Utpatti-kathana-(Upā khyā na) 12.122 47. Ṛṣabha-Gitā / Sumitra-Upā khyā na 12.125–26 48. Kapota-Upā khyā na 12.141–45 49. Kṛtaghna-Upā khyā na 12.162–67 50. Jā paka-Upā khyā na 12.189–93 51. Cirakā ri-Upā khyā na 12.258 52. Kuṇḍadhā ra-Upā khyā na 12.263 53. Nā rā yaṇīye Hayaśira-Upā khyā na 12.335 54. Uñchavṛtty-Upā khyā na 12.340–53 55. Sudarśana-Upā khyā na 13.2 56. Viśvā mitra-Upā khyā na 13.3–4 57. Bhaṅgā śvana-Upā khyā na 13.12 58. Upamanyu-Upā khyā na 13.14–18 59. Mataṅga-Upā khyā na 13.28–30 60. Vītahavya-Upā khyā na 13.31 61. Vipula-Upā khyā na 13.39–43 62. Cyavana-Upā khyā na 13.50–51 63. Nṛga-Upā khyā na 13.69 64. Nā chiketa-Upā khyā na 13.70 65. Kīṭa-Upā khyā na 13.118–20 66. Ut[t]aṅka-Upā khyā na 14.52–57 67. Nakula-Upā khyā na 14.92–96 381 APPENDICES 381 Several units are discussed in the edited volume Argument and Design.31 The Mahā bhā rata also employs several other terms for narrative units such as saṃvāda (dialogue), carita (adventure), kathā (story) and ākhyāna (story or legend), the latter often interchangeably with upākhyāna. No count or overview of these units exists, as far as we could discover. 382 38 11. SUKTHANKAR’S TABLE OF THE MANUSCRIPTS COLLATED FOR THE Ā DIPARVAN This table shows the manuscripts collated for diferent portions of the text.32 It is only approximate and especially does not note the many minor lacunae in Ś1. It should there- fore be supplemented with the table on the next page.  384 newgenrtpdf Adhyā ya & Śloka Northern Recension Manuscripts Southern Recension Manuscripts 1.1–204 K0–6 V1 B1–4 Da Dn Dr D1–12.14 T1.2 G1–7 M1–4 1.205–2.39 K0–6 V1 B1–4 Da Dn Dr D1–14 T1.2 G1–7 M1–4 2.40–191 K0–4.6 V1 B1–4 Da Dn Dr D1–14 T1.2 G1–7 M1–4 2.192–243 K0–4.6 V1 B1–4 Da Dn Dr D1–14 T1.2 G2–7 M1–4 3.1–44 K0–4 Ñ1.2 V1 B1–5 Da Dn D1–7 T1.2 G2–6 M1–5 3.45–13.45 K0–4 Ñ1.2 V1 B1–5 Da Dn D1–7 T1.2 G1–6 M1–5 14.1–26.9 K0–4 Ñ1–3 V1 B1–5 Da Dn D1–7 T1.2 G1–6 M1–5 26.10–43.13 Ś1 K0–4 Ñ1–3 V1 B1–5 Da Dn D1–7 T1.2 G1–6 M1–5 43.14–47.19 S1́ K0–4 Ñ1–3 V1 B1.3–5 Da Dn D1–7 T1.2 G1–6 M1–5 47.20–53.36 Ś1 K0–2.4 Ñ1–3 V1 B1.3–5 Da Dn D1–7 T1.2 G1–6 M1–5 54.1–4 Ś1 K0–2.4 Ñ1–3 V1 B1.3–6 Da Dn D1–5 T1.2 G1–6 M3.5–8 54.5–55.3a Ś1 K0–4 Ñ1–3 V1 B1.3–6 Da Dn D1–5 T1.2 G1–6 M3.5–8 55.3b–60.61b K0–4 Ñ1–3 V1 B1.3–6 Da Dn D1–5 T1.2 G1–6 M3.5–8 60.61c–61.84a Ś1 K0-4 Ñ1–3 V1 B1.3–6 Da Dn D1–5 T1.2 G1–6 M3.5–8 61.84b–62.2 K0–4 Ñ1–3 V1 B1.3–6 Da Dn D1–5 T1.2 G1–6 M3.5–8 62.3–68.19 K0–4 Ñ1–3 V1 B1.3–6 Da Dn D1.2.4.5 T1.2 G1–6 M3.5–8 68.20–74a S1́ K0–4 Ñ1–3 V1 B1.3–6 Da Dn D1.2.4.5 T1.2 G1-6 M3.5–8 68.74b–69.41c Ś1 K0–4 Ñ1–3 B1.3–6 Da Dn D1.2.4.5 T1.2 G1–6 M3.5–8 69.41d–51 Ś1 K0–2.4 Ñ1–3 B1.3–6 Da Dn D1.2.4.5 T1.2 G1–6 M3.5–8 70.1–71.17c Ś1 K0–2.4 Ñ1–3 B1.3–6 Da Dn D1–5 T1.2 G1–6 M3.5–8 71.17d–72.8b Ś1 K0–4 Ñ1–3 B1.3–6 Da Dn D1-5 T1.2 G1–6 M3.5–8 72.8c-22 Ś1 K0–2.4 Ñ1–3 B1.3–6 Da Dn D1–5 T1.2 G1–6 M3.5–8 72.23–74.4 K0–2.4 Ñ1–3 B1.3–6 Da Dn D1–5 T1.2 G1–6 M3.5–8 74.5–76.35 K0–4 Ñ1–3 B1.3–6 Da Dn D1–5 T1.2 G1–6 M3.5–8 77.1–78.20b K0–4 Ñ1–3 B1.3–6 Da Dn D1.2.4.5 T1.2 G1–6 M3.5–8 78.20c–90.88 Ś1 K0–4 Ñ1–3 B1.3–6 Da Dn D1.2.4.5 T1.2 G1–6 M3.5–8 90.89–92.13d Ś1 K0–4 Ñ1–3 B1.3.5.6 Da Dn D1.2.4.5 T1.2 G1–6 M3.5–8 92.13c–96.37b Ś1 K0–4 Ñ1–3 V1 B1.3.5.6 Da Dn D1.2.4.5 T1.2 G1–6 M3.5–8 96.37c–127.21a Ś1 K0–4 Ñ1–3 B1.3.5.6 Da Dn D1.2.4.5 T1.2 G1–6 M3.5–8 127.21b–181.40 Ś1 K0–4 Ñ1–3 V1 B1.3.5.6 Da Dn D1.2.4.5 T1.2 G1–6 M3.5–8 182.1–225.19 Ś1 K0–4 Ñ1–3 V1 B1.3.5.6 Da Dn D1.2.4.5 T1.3 G1–6 M3.5–8 385 12. EXTENT OF THE ŚĀ  RADĀ CODEX FOR THE Ā DIPARVAN The Śāradā codex contains the irst three parvans, but only fragments of the Ādi and the Āraṇyaka parvans are preserved. Sukthankar notes that the beginning of the Ādiparvan is “particularly fragmentary; a continuous text begins only from fol. 63 (our adhy. 82).”33 We have therefore examined the critical apparatus only up to adhyāya 82 (81 was complete), but it is possible that minor lacunae may also be revealed in the remaining chapters on a closer inspection. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 137 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 Key: White indicates that the Ś ā radā codex is not extant for these sections; light gray indicates that it is extant but contains lacunae (see notes); and darker shading indicates that it is fully extant. Notes: Adhyāya 26: Ś1 commences with 26.10; lacunae at 21b–35d and 46ab–27.13a. Adhyāya 27: Ś1 has lacunae at 1–13a and from 24d (from yaśasvini)–28.5b. Adhyāya 28: Ś1 has lacunae at 1–5b. Adhyāya 37: Ś1 has a lacuna at 15. Adhyāya 43: Ś1 has a lacuna at 11cd. Adhyāya 47: Ś1 has a lacuna from 1–1b (up to mantribhiḥ) and from 12b–48.4b. Adhyāya 48: Ś1 has lacunae at 1–4 b (up to nāmānīha) and from 15c–49.3d. Adhyāya 49: Ś1 has lacunae at 1–3d and from 14a–28b. Adhyāya 50: Ś1 1 has lacunae at 1.50.9 (from viniścitaṁ)–1.51.2b. Adhyāya 51: Ś1 has lacunae at 1–2b and from 1.51.9 (from pyanuga)–1.53.27a (up to va). 386 386 PHILOLOGY AND CRITICISM Adhyāya 53: Ś1 has a lacuna at 1–27a (resumes from va). Adhyāya 54: Ś1 has a lacuna at 1–18ab. Adhyāya 55: Ś1 has a lacuna at 3b–60.61b. Adhyāya 60: Ś1 has a lacuna at 62abc (a portion is missing). Adhyāya 61: Ś1 has a lacuna at 1a (from ca)–12a (up to ketumā) and is partly damaged from 40–43; it also has a small lacuna at 45a (from kāla)–45b (up to narādhipāḥ) and a longer one from 61.84b–68.19. Adhyāya 68: Ś1 has lacunae at 1–19 and from 35d–46 and 64 (from prasūto’yaṁ)–75c (up to tayorapatyaṁ). Adhyāya 69: Ś1 has lacunae at 13d–22b (up to ca) and from 38–47ab (the lacuna begins with the latter half of 678*, inserted after 37 in all N manuscripts). Adhyāya 70: Ś1 has a lacuna at 10b–20a. Adhyāya 71: Ś1 has lacunae at 16c (from pure)–25cd (up to buddhvā dā) and from 39 (from śiṣyaṁ)–45 and 57 (from brahmabhūtaḥ)–72.7a. Adhyāya 72: Ś1 has lacunae at 1–7b (up to bhārga-) and from 23a (from karma)–78.20ab. Adhyāya 78: Ś1 has a lacuna at 1–20ab (resuming from bhartā hi in 20c) and is partly damaged from 21–24; the irst ive letters of 35 and all of verse 37 are lost on a missing portion; there is also a lacuna from 38–79.9b (up to punaḥ). Adhyāya 79: Ś1 has lacunae at 1–9b and from 23–80.2. Adhyāya 80: Ś1 has lacunae at 1–2 and from 20b–27. 387 13. ABBREVIATIONS AND DIACRITICAL SIGNS USED IN THE CRITICAL EDITION This list of abbreviations and diacritical signs is based largely on Sukthankar’s list in the irst volume of the critical edition. Additional signs later editors introduced were added at the appropriate place without noting these insertions. Where Sukthankar’s informa- tion is outdated (for example, his note on Sastri’s edition, not completed at the time of the release of the Ā diparvan), his annotation has been updated without comment; where there is a conlict between the editors’ conventions, both variants are given. Thus, this list is a complete reference to all the abbreviations, diacritical signs and other markers used in the critical edition. add. = adding. addl. = additional. adhy. = adhyā ya(s). App. = Appendix. Arj. = Arjunamiśra. B. = Bombay edition of the Mbh. (Ś aka 1799). BG. = Bhagavadgītā . B., Bom. or Bomb. = Bombay (edition). C. = Calcutta edition of the Mbh. (editio princeps). C., Cal. or Calc. = Calcutta (edition). chap. = chapter(s). comm. = commentary/commentaries. cont. = continue(s). corr. = corrected, correction. Crit. App. = Critical Apparatus. Dev. = Devabodha. ed. = edition. ig. = igure(s). fol. = folio(s). foll. = following. fragm. = fragment(ary). hapl. = halographic(ally). illeg. = illegible. inf. lin. = infra lineam (below the line). ins. = inserts. int. lin. = inter lineas (between the lines). interp. = interpolate(s). introd. = introduction, introductory. K. = Kumbhakonam edition of the Mbh. 38 388 PHILOLOGY AND CRITICISM Kumbh. = Kumbhakonam (edition). L. = Line. M. = Madras edition of the Mbh. (of P. P. S. Sastri, 1931–36). Mad. = Madras (edition). m or marg. = marginal(ly). Mañj. = Bhā ratamañjarī (Kā vyamā lā 65). Manu = Manusmṛti (ed. N. S. P.). Mbh. = Mahā bhā rata. Nīl. = Nīlakaṇṭha. O.-J. = Old-Javanese. om. = omit(s), omitting. orig. = original(ly). p. = pā ṭhā ntara (added to the abbreviation of the name of a commentator, e.g. Arjp, Nīlp or to the symbol denoting a commentary). post. = posterior. pr. m. = propria manu (by own hand). prob. = probably. r. = repeat(s), repeated. Rā m. = Rā mā yaṇa (ed. N. S. P.) ref. = refer(ence). resp. = respective(ly). sec. m. = secunda manu (by another hand). st. = stanza(s). subst. = substitute(s). Suparn.̣ = Suparṇā dhyā ya (ed. Grube). suppl. = supplementary. sup. lin. = supra lineam (above the line). transp. = transpose(s), transposition. v. = verse. (var.) = (with variation). v. l. = varia(e) lectio(nes). ❀❀ (in the critical footnotes) enclose citations from commentators. [] besides their normal uses, enclose additions to MS readings. () besides their normal uses, enclose superluous letters, which should be omitted from MS readings. * (superior star) in the text indicates an emendation. ** (in the MS readings) indicate syllables lost through injury to MS.              printed below (or along the side margin of) any part of the constituted text indicates that the reading of it is less than certain. ʗ denotes a “Commentary,” preixed to the abbreviation of the name of a commentator; thus, ʗd = Commentary of Devabodha; ʗa.d.n.v = Commentaries of Arjunamiśra, Devabodha, Nīlakaṇṭha and Vā dirā ja. A “p” is added on to denote a “pā ṭha”; thus ʗnp = a pā ṭha mentioned in the Commentary of Nīlakaṇṭha. 389 14. ABBREVIATED CONCORDANCE OF THE PRINCIPAL EDITIONS OF THE MAHĀ BHĀ RATA The following table lists the number of adhyāyas (occasionally also the verses) found in each parvan of the critical edition followed by the corresponding ranges for that parvan in the Bombay, Calcutta and Madras editions. The concordance is not comprehensive (for this the reader must consult the respective volumes of the critical edition; the com- plete concordance runs into 400 pages) but provides a useful overview of the diferent editions’ relative lengths.34 For instance, the critical edition frequently has a lower adhyāya count than the other editions. The numbers, however, are only indicative, because what is decisive is, of course, the length of the individual adhyāyas rather than their number. In order to know which chapter in the critical edition, chapter 43 of book 1 of the Calcutta edition corresponds to, the reader must consult the complete concordance.35 This abbre- viated concordance adopts the following scheme: the irst row lists the irst line of the respective edition (in the case of the critical edition, the irst line of the constituted text); the second lists its last line. Critical edn. Bombay edn. Calcutta edn. Madras edn. Ā di1 1 1 1 225 234 218 Sabhā2 1.1ab 1.2ab 1ab 1.5 72.36 81.38cd, 39 2709 72.36 ½ Ā raṇyaka3 1 1 1 299 315 2694 Virāṭa5 1 1 1 67.38 72.40cdef 67.57’–596 Udyoga7 1 1 1 197 196 186 Bhīṣma 1.18 1.1 2 1.1 117.27–34 122.32–39 5849–5856 118.30–379 Droṇa10 Karṇa 1.1 1.111 1–312 1.1–2013 69.43cd14 96.61–6515 5043–504716 110.52–5917 Ś alya 1.1–8ab 1.1–8ab18 1–8ab19 1.1–8ab 64.39–4320 65.42–46 3667–3671 59.42–46 Sauptika 1.1abc 1.1ac21 1abc22 1.1abc 18.1–26 18.1–26 786–811 18.1–6 390 390 PHILOLOGY AND CRITICISM Critical edn. Bombay edn. Calcutta edn. Madras edn. Strī 1.1–6abcd 1.1–6 1–6 1.1–6 27.24cd23 27.30ef 828 27.31 ½cd Ś ānti24 Rājadharma 1.1–22 1.1–2225 1–2226 1.1–22 128.49ef 130.50cd 4778cd 120.51 Ā paddharma 129.1ab 131.1ab27 4779ab 121.1ab 167.24 173.26 6456 148.163 ½ Mokṣadharma 168.1–5 174.1–528 6457–6461 149.1–5 353.1–9 366[5].1–9 13935–13943 162.1–9 Anuśāsana 1.1–5abcd 1.1–529 1–530 1.1–5 154.32–34 169(68).35cdef–37 7794–7796 156.35 ½–37 ½ Ā śvamedhika 1.1–10 1.131 1–1032 1.1–10 96.12cd–1533 92.50cd–53 2897cd–2900 100.12cd–15 Ā śramavāsika 1.1–7cd 1.1–734 1–735 1.1–7 47.25–2736 39.25–27 1102–1104 42.45 ½–47 ½ Mausala 1.1–6ab 1.1–6ab37 1–6ab38 1.1–6ab 9.38cd 8.38ef 292cd 8.43cd Mahāprasthānika 1.1–11 1.1–1139 1–1140 1.1–11 3.28–36 3.30–38 101–109 3.30–38 Svargārohaṇa 1.1–2ab 1.1–2ab 1–2ab 1.1–2ab 5.5441 5.6842 211 5.5343 Notes: 1 Sukthankar does not give a concordance for the Calcutta edition. His editions for comparison are the editions of Ganapat Krishnaji (Ś aka 1799) and of P. P. S. Sastri (1931). 2 Edgerton’s Bombay edition probably refers to the edition of Ganapat Krishnaji, but he does not say so. 3 Sukthankar does not give a concordance for the Calcutta edition. His editions for comparison are the editions of Ganapat Krishnaji (Ś aka 1799) and of P. P. S. Sastri (1931). 4 269 in the Madras edition corresponds to 298 in the critical edition; 270 is transposed to the Virā ṭaparvan to be the irst chapter there. 5 Raghu Vira does not give a concordance for the Calcutta edition. His Bombay edition refers not to the edition of Ganapat Krishnaji but to the Gujarati Printing Press edition of 1915. 6 These last verses are printed as *1174 and *1176 of the critical edition; the last line of the Madras edition corresponding to the constituted text is 67.56 (corresponding to 67.38 of the critical edition). 7 S. K. De’s Bombay edition refers to the edition of Ganapat Krishnaji (cited with a date of Ś aka 1798). His Madras edition refers to the Sastri edition (1932). 8 This is the irst passage in the constituted text corresponding to the other editions; 1* is the irst passage in the critical edition as a whole corresponding to the other editions (1 in the Bombay and Calcutta editions; the passage is not in the Madras edition). 9 The Madras edition also includes another four verses (118.38–42) but these do not have equivalents in the critical edition. 10 S. K. De does not provide a concordance for the Droṇaparvan. 11 1.1 is the irst numbered line in the Bombay edition; this edition also includes an unnumbered line corresponding to 1* of the critical edition. 12 1–3 are the irst numbered lines in the Calcutta edition; this edition also includes an unnumbered line corresponding to 1* of the critical edition. 13 1.1–20 corresponds to 7* in the critical edition. 14 This is the last line in the constituted text, but there are also several star passages thereafter. 1219*, 1220*, 1221*, 1222*, 1223* have equivalents in the Bombay, Calcutta and Madras editions, while 1224*, 1225*, 1226*, 1228* are found in the Madras edition. 391 APPENDICES 391 15 96.61–65 corresponds to 1221*, 1222* and 1223* in the critical edition. 16 5043–5047 corresponds to 1221*, 1222* and 1223* in the critical edition. 17 110.52–59 corresponds to 1224*, 65–80 in the critical edition. 18 1.1 is the irst numbered line in the Bombay edition; this edition also includes an unnumbered line corresponding to 1* of the critical edition. 19 1–8ab are the irst numbered lines in the Calcutta edition; this edition also includes an unnumbered line corresponding to 1* of the critical edition. 20 This is the last line of the constituted text; the critical edition also includes an additional line there- after: 403*, relegated to the critical apparatus. 21 1.1abc is the irst numbered line in the Bombay edition; this edition also includes an unnumbered line corresponding to 1* of the critical edition. 22 1abc is the irst numbered line in the Calcutta edition; this edition also includes an unnumbered line corresponding to 1* of the critical edition. 23 This is the last line in the constituted text, but there are several star passages (81*, 82*, 83*) corresponding to the texts of the Bombay and Madras editions. 24 The igures for the Śā ntiparvan have been broken up into its constituent sections. 25 1.1 is the irst numbered line in the Bombay edition; this edition also includes an unnumbered line corresponding to 1* of the critical edition. 26 1 is the irst numbered line in the Calcutta edition; this edition also includes an unnumbered line corresponding to 1* of the critical edition. 27 131.1ab is the irst numbered line in the Bombay edition; this edition also includes an unnumbered line corresponding to 301* of the critical edition. 28 The Bombay edition also contains an introductory mantra corresponding to 455* of the critical edition. 29 1.1 is the irst numbered line in the Bombay edition; this edition also includes an unnumbered line corresponding to 1* of the critical edition. 30 1 is the irst numbered line in the Calcutta edition; this edition also includes an unnumbered line corresponding to 1* of the critical edition. 31 1.1 corresponds to 1* of the critical edition; the irst line that has an equivalent in the constituted text is 1.2–11 (corresponding to 1.1–10). 32 1 is the irst numbered line in the Calcutta edition; this edition also includes an unnumbered line corresponding to 1* of the critical edition. 33 This is the last line of the constituted text. There follow 198*, 199* and App.  1, no.  4.  This last passage corresponds to adhyāyas 101–24 of the Madras edition (the conclusion of the parvan in that edition). 34 1.1 is the irst numbered line in the Bombay edition; this edition also includes an unnumbered line corresponding to 1* of the critical edition. 35 1 is the irst numbered line in the Calcutta edition; this edition also includes an unnumbered line corresponding to 1* of the critical edition. 36 This is the last line in the constituted text, but the critical edition also includes a star passage (108*) corresponding to 1105 of the Calcutta edition (the last line in that edition). 37 1.1 is the irst numbered line in the Bombay edition; this edition also includes an unnumbered line corresponding to 1* of the critical edition. 38 1 is the irst numbered line in the Calcutta edition; this edition also includes an unnumbered line corresponding to 1* of the critical edition. 39 1.1 is the irst numbered line in the Bombay edition; this edition also includes an unnumbered line corresponding to 1* of the critical edition. 40 1 is the irst numbered line in the Calcutta edition; this edition also includes an unnumbered line corresponding to 1* of the critical edition. 41 This is the last line of the constituted text; there follow three star passages—60*, 61* and 62*. 42 5.68 corresponds to 60* of the critical edition; the last line with an equivalent in the constituted text is 5.67 (corresponding to 5.54 of the critical edition). 43 This is the last line of the Madras edition, corresponding to 52*, lines 4–5 of the critical edition (occurring after 5.43 of that edition). 392 39 15. STEMMATA FOR THE DIFFERENT PARVAN S OF THE MAHĀ BHĀ RATA A. C. Burnell produced the irst known stemma of the Mahā bhā rata in 1875. Burnell’s stemma was based on a comparison of the Nā garī and Grantha recensions but does not mention the manuscripts involved (most likely, from the Tanjore Palace Library). His stemma is reproduced here for historical interest (Figure 78). Sukthankar produced the irst stemma based on an extensive survey of the principal versions in existence (Figure  79). Its inluence on all subsequent editors (with the possible exception of Edgerton) is palpable. Edgerton did not draw up a stemma for the Sabhā parvan. In con- trast to Sukthankar, he felt that three traditions of the Mahā bhā rata existed, a western (designated by W), an eastern (designated by E) and a southern (designated by S). The western recension included the Ś and K manuscripts; E included the Ñ, V, B and D manuscripts; and S included the T, G and M manuscripts. Edgerton also thought VBD might be derived from a common secondary archetype that he called “E1” (Ñ, in contrast, would have descended from E itself). Sukthankar also produced a second stemma for the manuscripts of the Ā raṇyakaparvan, the only other book he edited. See Figure 80. Raghu Vira’s and S. K. De’s stemmata for the Virā ṭa and Udyoga parvans (reproduced here as igures 81 and 82) are essentially variations of Sukthankar’s stemma. Belvalkar’s stemma of the Bhīṣmaparvan manuscripts is reproduced in Figure  35 and is therefore not reproduced here. After the Bhīṣmaparvan, no editor drew up a stemma. The only stemma to appear thereafter is Vaidya’s stemma of the Harivaṃśa (Figure 83). Original collection of ballads. Eclectic text with episodes added to it. A) N. Recension (Nāgarı̄), B) S. Recension (Grantha), a revision of the longer an independent revision of the books, settled finally by longer books. Nı̄lakantha’s ·· Commentary. Figure 78 A. C. Burnell’s stemma Source: Reproduced from Burnell, On the Aindra School of Sanskrit Grammarians, 80. 394 394 PHILOLOGY AND CRITICISM Vy rata Ur-Mah bh rata N S rada K Nep l Maithil Bengali Devan gar Telugu Grantha Malay lam (other than K) Figure 79 V. S. Sukthankar’s “Pedigree of Ādiparvan versions” Source: Reproduced from Sukthankar, “Prolegomena,” xxx. Vy rata Ur-Mah bh rata N S ν γ σ rad K Bengali (Non-K) Telugu Grantha Malay lam ( ) (B) Devan gar (T) (G) (M) (D) Figure 80 V. S. Sukthankar’s stemma of the Āraṇyakaparvan manuscripts Source: Reproduced from Sukthankar, “Introduction,” xiii. 395 APPENDICES 395 Vy rata Ur-Mah bh rata N S γ σ rad & K Bengali Devan gar Telugu Grantha Malay lam (other than K) Figure 81 R. Vira’s stemma of the Virāṭaparvan manuscripts Source: Reproduced from Vira, “Introduction,” xvi. Vy rata Ur-Mah bh rata Northern (N) Southern (S) Recension Recension ν γ ε σ rad K Bengali Devan gar (D) Telugu Grantha Malay lam ( ) (B) (other than K) (T) (G) (M) Figure 82 S. K. De’s stemma of the Udyogaparvan manuscripts Source: Reproduced from De, “Introduction,” xxiv. 396 396 PHILOLOGY AND CRITICISM Ur Text of Hariva a N S NW NE C T G M D K V B M4 1 Dn Ds D6 Dn1.n2 Ds1.s2 Figure 83 P. L. Vaidya’s stemma of the Harivaṃśā manuscripts Source: Reproduced from Vaidya, “Introduction [to the Harivaṃśa],” xxiv. 397 16. COMMENTARIES ON THE MAHĀ BHĀ RATA Surprisingly little is known about the Mahā bhā rata commentators, with the exception of Nīlakaṇṭha.36 Sukthankar, in his discussion of Mahā bhā rata commentaries,37 lists 22 commentators, whose names are either known to us or whose works have survived,38 but we possess manuscripts of only about a handful of these. There is unanimous agreement that Devabodha is the earliest—and, by some measures, the most important—of the Mahā bhā rata commentators,39 though opinions diverge about the precedence of commentators thereafter.40 In his study of the Mahā bhā rata commentators, Sukthankar proposes the following sequence: Devabodha, Vimalabodha, Sarvajña-Nā rā yaṇa, Arjunamiśra and Nīlakaṇṭha Caturdhara. Devabodha’s dates are not known. Christopher Minkowski assigns a date of 1000 CE to him, which Basile Leclere revises to an upper limit of 1100 CE.41 Since Vimalabodha is dated to 1150–1300 CE and he cites Devabodha (once as Devabodha and once as Devasvā min), we have a lower threshold for his dates.42 This would suggest that Minkowski is correct in asserting that a continuous tradition of Mahā bhā rata com- mentary did not exist before the twelfth century.43 The earliest commentaries such as Devabodha’s Jñānadı̄pikā are little more than glosses of the most diicult words or phrases, and this tradition continues in works such as Vimalabodha’s Durghat ̣ar̄ thaprakāśinı̄ or Viṣamaślokı̄.44 The practice of writing expansive commentaries explaining either entire narratives or the symbolic meanings of certain passages (for example, as Nīlakaṇṭha does in his Bhāratabhavadı̄pa commentary) appears of much later origin (though already Arjunamiśra’s Arthadı̄pikā commentary is, in the words of Sukthankar, “a revised and enlarged edition of the Jñā nadīpikā ”).45 Philosophical Ailiations and Milieu Leclere in his recent study of Devabodha argues for identifying Devabodha the com- mentator with Devabodha the medieval poet and dramatist (sometimes also called Devabodhi).46 He describes Devabodha as a versatile author, possessing rich “poetical skills” and familiar with the “codes of dramatic literature.”47 In his view, Devabodha likely belonged to a “religious order of wandering ascetics” and was himself titled paramahaṃsaparivrājakācār ya, “Paramahaṃsa and Master of Wandering ascetics.”48 Leclere locates Devabodha within a Vaisn ̣ ava devotional tradition. “The Jñānadı̄pikā and the Satyavratarukmāṅgada betray the same devotion to Visn ̣ as the Supreme Being. At the very ̣ u 398 398 PHILOLOGY AND CRITICISM beginning of both works, the author invokes, in elaborate verses, the god under his cosmic aspects of Nā rā yanạ and Trivikrama, and even though he does not ignore other divini- ties such as Śiva and Ganẹ śa, he undoubtedly gives preeminence to Visn ̣ throughout his ̣ u texts.”49 He argues that Devabodha was a yogic adept, familiar with the Sā ṃkhya school and its fundamental text, the Sāmḳ hyakārikā (cited in his commentary on the Ā diparvan).50 Devabodha’s work is “intended to facilitate the access to a meaning which can help people purify and eventually save themselves. As Devabodha makes it clear elsewhere, the readers can have thanks to these explanations an insight of the Inner and Supreme Self: ‘In order to put before the eyes the internal man through the light of consciousness, the Lamp of Knowledge [Jñānadı̄pikā] is placed in the Assembly by Devabodha.’ ”51 Leclere’s study provides evidence not only for the Mahā bhā rata’s embedding in the Sanskrit poetic tradition but also for a tradition of soteriological commentary in the manner of Ānandavardhana. The epic itself appears, at least within a Vaisn ̣ ạ va milieu, to have been regarded as a mystical work revealing the glories of Hari. Nīlakaṇṭha’s much 52 later Bhāratabhavadı̄pa also emphasizes that its aim is to illuminate the “internal, hidden meanings” (antarnigū dḥ ārtha) contained in the text.53 Aim in Reading the Mahābhārata Valuable insight into the commentators’ exegetic concerns is provided by their commen- taries on the Mahā bhā rata’s maṅgaláloka, which Jahnavi Bidnur is currently translating. The following excerpts from Devabodha’s and Nīlakaṇṭha’s commentaries are reprinted with her permission. Here is the commentary on the Mahābhārata’s maṅgaláloka. Devabodha begins  thus: iti namaskṛtya—jayaḥ (word jaya) paurānị kānāṃ rūdḥ yā54 (by the custom of the purānı̣ k̄ as) bhāratam (the Bhārata) ucyate (is said to be); svapaksạ sthāpanayā (by establishing own paksạ ) parapaksạ pratyākhyānaṃ (overpowering the opposition) (is) jayaḥ (triumph, victory) iti (thus) (is the) yogaḥ (original meaning); iha ca (so here) dharmārthakāmamoksạ sthāpanayā (by establishing dharma, artha, kāma and moksạ ) adharmānarthākāmabandhānāṃ tatpratipaksạ n̄ ạ m ̄ ̣ (opposition of that—adharma, anartha, akāma and bandha) pratyākhyānam (overpowering) iti (this) yogaḥ (original meaning) api (also) pratı̄yamānaḥ (implicated) na (not) hātum arhati (worthy to be denied); asya (of this) śas̄ trasya (knowledge branch) preksạ vatpravṛttyanġ avisạ yasambandhaprayojanāni (part of the wise beginning the subject matter, connection and purpose) jayasaṃjñāta (from the term jaya) upalabhyante (have been obtained); ata eva (therefore) adhikāribhedāt (by the diference in seeker) prayojanabhedat ca (by the diference in purpose) śas̄ trabhedaḥ (the subject matter difers). I translate: The word jaya in the benedictory verse means bhārata by the paurāṇika custom, but its original meaning is victory or triumph. So, one meaning of the word jaya, that is, the original meaning here is: by establishing dharma, artha, kāma and mokṣa, overpowering the opposition of adharma, anartha, akāma and bandha. This meaning [that is, the original meaning of victory as the overcoming of adharma, etc.], which is implied by this word [jaya], is not worthy to be denied. As a part of the wise beginning, the sub- ject matter, connection and purpose can be understood (obtained) from the word jaya. Therefore, the subject matter difers according to the seeker and the purpose. 39 APPENDICES 399 Thus, according to Devabodha’s commentary, the text’s anubandhacatuṣt ̣aya is as follows: (1) sambandha (connection): the word jaya connects the text with the Bhārata and all the four subject matters dharmārthakāmamokṣa; (2) adhikārin (seeker of knowledge): seekers of dharma, artha, kāma and mokṣa; (3) viṣaya (subject matter): dharma, artha, kāma and mokṣa; (4) prayojana (purpose):  narrating the Bhārata, according to the paurāṇika custom; and, by establishing dharma, artha, kāma and mokṣa, overpowering the opposition of adharma, anartha, akāma and bandha. While writing a commentary on the Mahā bhā rata’s maṅgaláloka, Nīlakaṇṭha begins with: iha (here) khalu (indeed) bhagavān (the revered one) pārāśaryaḥ (the son of sage Parā śara, that is, Vyā sa) paramakār uṇiko (superior among the compassionate) mandamadhyamamatın̄ (to the dull-witted and the mediocre/to the ignorant) anugṛahı ̄tuṃ (to bless or to favor) caturdaśavidyāsthānrahasyāni (14 lores of knowledge) ekatra (at one place) pradidarśayiṣuḥ (desirous to reveal) mahābhāratākhyam itihāsaṃ (a lokavṛtta or his- tory named the Mahā bhā rata) praṇeṣyan (establishing) prāripstitasya (of the initiated) granthasya (text) niṣpratyūhaparipūraṇāya (for accomplishment without any obstacle) pracayagamanāya ca (and to achieve progress or growth) kṛtaṃ maṅgalaṃ (ofered bene- diction) śiṣyaśikṣāyai (for the admonishment of the disciple) ślokarūpeṇa (in the form of verse) nibadhnan (composing) arthāt (that is to say) tatra (there) prekṣāvat (wise or learned) pravṛttyaṅgam55 (part of the commencement) abhidheyādi (subject matter along with prayojana) darśayati (points out) nārāyaṇamiti. I translate: Here, indeed, Lord Vyā sa, the son of Parā ́ara, being superior among the compassionate, to bless/oblige the dull-witted and the mediocre, desirous to reveal all the 14 lores at one place, establishes the lokavṛtta (history) named the Mahā bhā rata. For accomplishment of the initiated text, without any obstacle and to achieve growth or pro- gress, he has ofered benediction, for the admonishment of the disciple, composing in the form of a verse, that is to say, wise or learned part of the commencement, (and) points out subject matter along with prayojana: nārāyaṇamiti. Thus, according to Nīlakaṇṭha’s commentary, the text’s anubandhacatuṣṭaya is as follows: (1) sambandha (connection):  the words caturdáavidyāsthānrahasyāni and mahābhāratākhyam itihāsaṃ reveal the connection. The Mahā bhā rata is the itihāsa and it connects itself with the domain of 14 knowledge branches; (2) adhikārin (seeker of knowl- edge): mandamadhyamamatı̄n (to the dull-witted and the mediocre) anugṛahītuṃ (to bless or to favor)—here, the commentary mentions who the learners of this śas̄ tra are. The words used implicate the common people who might not be well-versed in any knowledge branches; (3) viṣaya (subject matter):  caturdaśavidyāsthānrahasyāni (14 lores of knowledge) ekatra (at one place) pradidarśayiṣuḥ (desirous to reveal)—here, the commentary explains what the text’s subject matter is; (4) prayojana (purpose): the same words in the commen- tary also explain the text’s purpose. It is to ofer knowledge of all knowledge branches in one place—śiṣyaśikṣāyai (for the admonishment of the disciple) ślokarūpeṇa (in the form of verse) nibadhnan (composing) arthāt (that is to say) tatra (there) prekṣāvat (the wise or the learned) pravṛttyaṅgam (part of the commencement) abhidheyādi (the subject matter along with the prayojana) darśayati (points out). These words also point out the prayojana of the maṅgala as well as the text. 40 400 PHILOLOGY AND CRITICISM Extent of the Commentaries and Published Editions P. K.  Gode refers to a list of manuscripts of the commentaries of the Mahā bhā rata Sukthankar prepared.56 In its absence,57 the sole means to ascertain which manuscripts of the Mahā bhā rata commentaries exist, whether they are complete, what portion has been published and whether the commentaries themselves are complete was to consult the editors’ introductions to the various parvans of the critical edition.58 A study of the parvans revealed the following situation: 1. For the Ā diparvan, manuscripts exist of the commentaries of Arjunamiśra (two), Nīlakaṇṭha (three), Ratnagarbha (four) and Devabodha (two).59 Of these, only Devabodha’s commentary has been published separately,60 although Nīlakaṇṭha’s is available in the printed vulgate editions.61 Additionally, Sukthankar notes that fragments of Sarvajña-Nā rā yaṇa’s commentary on the Ā diparvan are preserved.62 2. For the Sabhā parvan, manuscripts exist of the commentaries of Devabodha (two) and Vā dirā ja (cited from the printed text in the Sastri edition; the latter is based on two manuscripts63). Edgerton cites Nīlakaṇṭha’s commentary from the printed edition, but manuscripts may exist. In addition, Edgerton mentions two partial commentaries only occasionally cited in the Addenda. The irst is an “incomplete collation of the text of Arjunamiśra’s commentary, collated at the Visvabharati, Santiniketan, from a Calcutta MS, Sā hitya Parishad Library no. 24”; the second is “a copy of a commentary by Vimalabodha on a few (about twenty) scattered verses in this book,”64 bearing the number 84 of 1869–79 (of the Government Collection deposited at the Bhandarkar Oriental Research Institute). He refers to this second commentary as “the Viṣamaślokī.”65 Like the irst (which is cited as Ca), this manu- script is also occasionally cited in the Addenda (as Cb) “when it seemed interesting.”66 Only Devabodha’s commentary on the Sabhā parvan has been published.67 3. For the Ā raṇyakaparvan, only one copy of Arjunamiśra’s commentary turned up, and it was in poor condition. It is unclear whether a tradition of commentary on this parvan existed or only that the search failed to turn up manuscripts of commen- taries on this parvan.68 4. Raghu Vira, the editor of the Virā ṭaparvan, does not cite manuscripts for any of the commentaries he references. He mentions the edition of eight Mahā bhā rata com- mentaries (the commentaries of Vā dirā ja, Vimalabodha, Arjunamiśra, Caturbhuja, Nīlakaṇṭha, Rā makṛṣṇa, Sarvajña-Nā rā yaṇa and another of unknown author- ship, the Viṣamapadavivaraṇa; of these Vā dirā ja’s was “fragmentary,” Vimalabodha’s “too meagre”) published by Mahadeva Shastri Gangadhara Bhatta Bakre of the Gujarat Printing Press, Bombay in 1915. He also notes that for the Virā ṭaparvan the number of commentaries is “not less than ifteen,” but mentions neither which these 15 are (one of them must be Devabodha) nor the manuscripts in which they may be found.69 5. For the Udyogaparvan, manuscripts exist of the commentaries of Devabodha (one), Arjunamiśra (cited from the printed edition), Sarvajña-Nā rā yaṇa (two) and Nīlakaṇṭha (two). De notes that additionally two other commentaries for the Udyoga 401 APPENDICES 401 were available: the Durghat ̣ārtha-prakāśinı̄ of Vimalabodha and the Laksạ ̄bharaṇa by Vā dirā ja, both in the second volume of the edition of Mahadeva Bhatta Bakre (1920). The former was not used as it is “too meagre and deals occasionally only with what it considers to be durghat ̣a,” while the latter was not used as it “does not represent the best South Indian version and comments generally on the T G version.”70 Śaṅkara also wrote a commentary on the Sanatsū jā tīya section of this parvan. Only Devabodha’s commentary has been published.71 6. For the Bhīṣmaparvan, manuscripts exist of the commentaries of Devabodha (one), Caturbhuja (one), Vā dirā ja (one),72 Arjunamiśra (two), Nīlakaṇṭha (two).73 Belvalkar also mentions three other commentaries—Vimalabodha’s Durghat ̣ārthaprakāśinı̄, Yajña (or Sarvajña) Nā rā yaṇa’s74 Bhāratatātparyasaṃgraha and Rā mā nuja’s Viṣamaślokat ̣ı ̄kā also called the Bhāvārthadı̄pikā, the latter two available in two manuscripts each—all of which, however, “are far too meager and have been drawn upon only occasionally in the Critical Notes.”75 He later revised his ascription of the commentary in the two manuscripts nos. 84 and 167 of the Deccan College Library (deposited at the Bhandarkar Oriental Research Institute) to Vimalabodha.76 Only Devabodha’s commentary has been published.77 7. For the Droṇaparvan, manuscripts exist of the commentaries of Arjunamiśra (two), Caturbhuja (one), Nīlakaṇṭha (two) and Vā dirā ja (one). The parvan editor S. K. De also mentions that “the commentary of Devabodha was available in a transcript from MS. No. 482/5797, belonging to the Scindhia Oriental Institute, Ujjain,”78 but Dandekar’s comments in his introduction to the Śalyaparvan render the ascrip- tion problematic.79 No published editions exist. 8. For the Karṇaparvan, manuscripts exist of the commentaries of Arjunamiśra (one), Caturbhuja (one), Nīlakaṇṭha (one) and Vā dirā ja (one). The parvan editor P. L.  Vaidya also cites “Vimalabodha’s commentary known as Viṣamaśloki or Durghaṭā rthaprakā śinī” and adds, “a complete MS. of this work is found in the Bombay Government Collection (deposited at the BORI), No. 84 of 1869–70,”80 but according to Belvalkar’s information the ascription is incorrect.81 No published editions exist. 9. For the Śalyaparvan, manuscripts exist of the commentaries of Arjunamiśra (two), Vā dirā ja (one) and Nīlakaṇṭha (four). No published editions exist. 10. For the Sauptikaparvan, manuscripts exist of the commentaries of Arjunamiśra (two), Nīlakaṇṭha (one) and Vā dirā ja (one). Velankar also refers to Devabodha’s commentary, available in “a transcript from the Scindia Oriental Institute, Ujjain, bearing No. 482/5797.” He adds, “This happens to be the only MS. of Devabodha available for this parvan.”82 Dandekar, however, questions the ascription. According to him, the commentary is actually Arjunamiśra’s.83 No published editions exist. 11. For the Strīparvan, manuscripts exist of the commentaries of Arjunamiśra (two), Nīlakaṇṭha (two) and Vā dirā ja (one). Paranjpe also mentions a “transcript of Devabodha made at the BORI from No. 482/5797 of the Scindia Oriental Institute, Ujjain,” which “was used for the present edition.”84 Dandekar, however, questions the ascription. According to him, the commentary is actually Arjunamiśra’s.85 No published editions exist. 402 402 PHILOLOGY AND CRITICISM 12. It appears to have been the practice to hand down the Sā́ ntiparvan only in its subsections (that is, the Rā jadharma, Ā paddharma and Mokṣadharma parvans) rather than as a whole. Accordingly, the manuscripts of the commentaries are also speciic to these subsections. For the Rā jadharma and Ā paddharmaparvan, manuscripts exist of the commentaries of Arjunamiśra (three; an additional two contain only the Rā jadharma, while a solitary copy contains only the Ā paddharma),86 Vimalabodha (one) and Vā dirā ja (one).87 For the Mokṣadharmaparvan, manuscripts exist of the commentaries of Arjunamiśra (three), Vimalabodha (one), Paramā nanda Bhaṭṭā cā rya (two),88 Vidyā sā gara (three)89 and Vā dirā ja (one). Manuscripts also exist of Nīlakaṇṭha’s commentary:  one (Dn1) contains all three subsections; a second (Dn2) contains only the irst; a third (Dn3) contains the second and third; and a fourth (Dn4) contains the irst and third. In addition to these four, which were used for the critical edition, another manuscript containing only the Rā jadharmaparvan exists.90 No published editions exist of either of the three sub-parvans. 13. For the Anuśā sanaparvan, manuscripts exist of the commentaries of Arjunamiśra (one), Nīlakaṇṭha (two). The ascription of a commentary to Vimalabodha on the basis of two manuscripts in the Bombay Government Collection (nos. 84 of 1869– 70 and 167 of 1887–91) is erroneous.91 In addition, an unidentiied commentary exists in two manuscripts. No published editions exist. 14. For the Ā śvamedhikaparvan, only a solitary manuscript of the commentary of Vā dirā ja was used. Karmarkar cites Nīlakaṇṭha’s commentary from the printed edition, but, according to Belvalkar’s information, manuscript no.  29c of the Bombay (now Maharashtra) Government Collection also contains Nīlakaṇṭha’s commentary on this parvan. No published editions exist. 15–18. For the Ā śramavā sika, Mausala, Mahā prasthā nika and Svargā rohaṇa parvans, manuscripts exist of the commentaries of Nīlakaṇṭha (two)92 and Vā dirā ja (one). No published editions exist. Summing up, we can say that, for the principal Mahā bhā rata commentators (Devabodha, Arjunamiśra, Vimalabodha, Vā dirā ja and Nīlakaṇṭha), we possess near complete commentaries. Devabodha’s commentary is missing for the Ā raṇyaka and Virā ṭa parvans93 and for the Droṇa through Svargā rohaṇa parvans.94 It is preserved for the Bhīṣmaparvan, but without the Bhagavadgītā .95 Arjunamiśra’s commentary is missing for the Ā śvamedhika through Svargā rohaṇa parvans and is only partially available for the Sabhā parvan.96 Vimalabodha’s commentary is missing for the Ā di through Bhīṣma parvans, for the Droṇa, Śalya and Sauptika parvans and for the Ā śvamedhika through Svargā rohaṇa parvans. Vā dirā ja’s commentary is the most complete and is missing only for the Ā di and Ā raṇyaka parvans (though it is possible that Sukthankar, who was the editor of both these parvans, was unaware of Sastri’s transcript). Nīlakaṇṭha’s commentary is also missing for the Āraṇyaka and Anuśā sna parvans; it is cited from the printed edition in the Sabhā parvan. From the fact that these commentators appear to have written continuous commentaries, it seems likely that at one time commentaries existed on all the parvans 403 APPENDICES 403 of the Mahā bhā rata.97 Additionally, fragmentary commentaries exist on one or more parvans or subsections of parvans by Ratnagarbha, Caturbhuja, Rā makṛṣṇa, Rā mā nuja, Śaṅkara, Vidyā sā gara, Paramā nanda Bhaṭṭā cā rya and two unidentiied commentators.98 Except for Nīlakaṇṭha’s commentary, which has probably been continuously in print in one or more editions since the nineteenth century, no complete edition exists of any Mahā bhā rata commentary:  the one that has been published most extensively is Devabodha’s (Ā di, Sabhā , Udyoga and Bhīṣma parvans); additionally, published editions exist for the Virā ṭa and Udyoga parvans for the commentaries of Vā dirā ja, Vimalabodha, Arjunamiśra, Caturbhuja, Nīlakaṇṭha, Rā makṛṣṇa, Sarvajña-Nā rā yaṇa and an uniden- tiied commentator. No translations exist. Finding Guide to the Commentaries Given the Mahābhārata commentators’ importance in reestablishing the meaning of the text after the German critics’ willful and destructive interventions, we include a inding guide to the manuscripts here. The next task must be to edit and publish these com- mentaries, and, thereafter, to translate them for a new readership. Only thus can we revive a centuries-old tradition of literate reception and literary production based on the Mahābhārata. 1. Ā diparvan Commentary Sig. Manuscripts Arjunamiśra Da1 Poona, Bombay Government Collection (deposited at the (Mahā)Bhāratārtha(pra) BORI), no. 30 of 1879–90 CE. Folios 416, with about 7–10 dı̄pika or lines to a page; size 15½” x 6¾”. Devanāgarī characters. Old Bhāratasaṃgrahadı̄pikā. Indian glossy paper. Text with commentary of Arjunamiśra; Version probably an ofshoot written neatly but extremely corrupt and unintelligible of the Bengali version. in places, on account of the scribe’s inability to read the exemplar correctly. Da2 Poona, Bombay Government Collection (deposited at the BORI), Viśrāmbāg I, no. 468. Folios 415, with about 10 lines to a page; size 15.7” x 6.6”. Devanāgarī characters; dated V. Saṃ. 1676 (ca. 1620 CE). Indian paper. Text with commentary of Arjunamiśra.1 Nīlakaṇṭha Dn1 MS belonging to Sardar M. V. Kibe of Indore. Folios 446, Bhar̄ atabhav̄ adıpa ̄ with about 8–10 lines to a page; size 18.2” x 7.3”. Devanaḡ arī The text used or prepared characters. Thick Indian paper. Text with commentary of by Nīlakaṇṭha is a smooth Nīlakaṇṭha. The MS is, on the whole, correct and very clearly and eclectic but inferior written. Dated: Īśvara saṃvatsara, mārgaśīrṣa śuddha 13, text, of an inclusive rather which cannot be identiied. than exclusive type, with an inconsiderable amount of Southern element. 40 404 PHILOLOGY AND CRITICISM Dn2 Mysore, Oriental Library, no. 1064. Folios 448, with about 22 lines to a page; size 15¼” x 6¼”. Devanāgarī characters; dated V. Saṃvat 1864 (ca. 1808 CE). Paper. Text with commentary of Nīlakaṇṭha. Dn3 Poona, Bombay Government Collection (deposited at the BORI), no. 234 of 1895–1902. Folios 683, with about nine lines to a page; size 15.2” x 7.2”. Devanāgarī characters. Thick Indian paper. Text with commentary of Nīlakaṇṭha. Bold and clear letters; generally correct; margins are almost clean. Ratnagarbha Dr1 Tanjore, Saraswathi Mahal Library, no. 1246. Folios The critical notes contain 448, with about 11 lines to a page; size 15” x 6½”. only specimen collations of Devanāgarī characters. Paper. Text with the commentary of this version, which is a blend Ratnagarbha. between the Northern and Dr2 Tanjore, Saraswathi Mahal Library, no. 1199. Folios Southern recensions. Like 306, with about 10–13 lines to a page; size 16” x 6½”. the Telugu manuscripts Devanāgarī characters. Paper. Text with the commentary of […] it is eclectic, following Ratnagarbha. now the Northern tradition, Dr3 Tanjore, Saraswathi Mahal Library, no. 1313. Folios 366, now the Southern. It with about 11–13 lines to a page; size 16” x 6¼”. Devanāgarī seems to be an attempt to characters; dated Ś aka 1623 (ca. 1701 CE). Paper. Text with combine the two recensions the commentary of Ratnagarbha. by superposition, like the Dr4 Tanjore, Saraswathi Mahal Library, no. 1339. Folios Kumbhakonam edition. 108, with about 11–22 lines to a page; size 16” x 6½”. Devanāgarī characters. Paper. Text with the commentary of Ratnagarbha. This fragment contains only about 90 adhyāyas of this edition. The number of lines on each folio luctuates with the amount of commentary each folio contains, and which of course, varies considerably. Devabodha Cd Baroda Central Library (Sanskrit Section), no. 11372. The version of Devabodha Contains the commentary on Ā di only and is written in ́ is of the Sāradā–K type. Devanāgarī characters of about the seventeenth century. The bulk of the MS is in a fair state of preservation, though in many places the text is extremely corrupt. In our MS the name of the commentary is given as Jñānadı̄pikā; but, according to Holtzmann (op. cit. vol. 3, p. 71), it is also known as Mahābhārata-tātparyatı̣ ̄kā. MSS. of this commentary are rare, and no complete copy has yet been found.2 2. Sabhāparvan Commentary Sig. Manuscripts Devabodha Cd Poona, Bombay Government Collection (deposited at the BORI), no. 167 of 1884–87. Devanāgarī; paper; no date. Sabhāparvan alone, with commentary of Devabodha. Complete: 115 folios, (1a and 115b blank; text ends on 114b); 11–13 lines, 40–46 akṣaras.3 405 APPENDICES 405 Poona, Bombay Government Collection (deposited at the BORI), no. 74 of 1902–7. Devanāgarī; paper; dated Saṃvat 1821 (ca. 1763 CE). Sabhāparvan alone, with Devabodha’s commentary. Complete: 115 folios (1a and 115b blank; text ends on 114b), average 10–11 lines, 53–55 akṣaras.4 Nīlakaṇṭha Cn Cited from the printed text of the Bombay edition, of which the volume containing Ā di and Sabhā is dated 1807. Vādirāja Cv Cited from the printed text of Sabhā in the Southern Recension, ed. P. P. S. Sastri, being volume 3 of his edition (Madras, 1932).5 3. Ā raṇyakaparvan Commentary Sig. Manuscripts Arjunamiśra None6 Government Sanskrit College Collection in Calcutta (no. 310) 4. Virāṭaparvan Commentary Sig. Manuscripts Arjunamiśra Ca Caturbhuja Cb Nīlakaṇṭha Cn Rāmakṛṣṇa Cr Sarvajña-Nārāyaṇa Cs Commentary known as Cv Viṣamapadavivaranạ 5. Udyogaparvan Commentary Sig. Manuscripts Devabodha Cd no. 3399/4814 in Haraprasad Shastri’s Descriptive Catalogue v, Undoubtedly the Purāṇa, pp. 151–52. It contains folios 3–119, with ive lines oldest and deserv[ing] to a page, size 11½” x 1¾”, in clear bold Devanāgarī writing of foremost in ink, using frequent pṛṣtḥ amātrās. The following folios are consideration. In missing: 14, 39, 53–54 (which, however, are replaced by a folio the nature of brief without number), 62–63, 81, 84, 86, 98, 100–109 and 111. running glosses There are several extra folios containing other matters.7 on diicult words and phrases; but the commentator’s testimony, in striking agreement with that of Ś 1 K, has been of considerable value in the constitution of the critical text. 406 406 PHILOLOGY AND CRITICISM Commentary of Ca Cited from the printed text of the Gujarati Press edition Arjunamiśra of Mahadeva Shastri Gangadhara Bhatta Bakre (Bombay, As a Bengali, he 1915).8 should represent the Bengali version, but his great admiration for Devabodha often makes his commentary a direct copy or ampliied paraphrase of that of his predecessor, while his many divergences from speciic Bengali readings often raise the suspicion that he consulted other sources as well. Commentary of Cs Poona, Bombay Government Collection (deposited at the Sarvajña-Nārāyanạ BORI), no. 168 of 1884–87. 309 folios, with about 11 lines The exact provenance to a page; size 15½” x 6¾”. Devanāgarī characters. Modern of this commentary paper. Text with Sarvajña-Nārāyaṇa’s commentary. Clear and is not known, but fairly correctly written, with very few marginal corrections. Sarvajña-Nārāyanạ The pages are written in strips, with the original text in the was undoubtedly a middle.9 Northerner; and his Poona, Bombay Government Collection (deposited at the commentary shows BORI), no. 33 of 1879–80 CE. 309 folios, with about 10 to the direct inluence 13 lines to a page; size 15½” x 6¾”. Devanāgarī characters. of Devabodha. […] Modern paper. Text with Sarvajña-Nārāyaṇa’s commentary. Strangely enough, the Neat and clear writing, generally correct, the margins almost text does not always clean.10 represent the readings of the commentary. […] This makes it diicult to determine his provenance and to assign, even tentatively, any deinite version to him. 407 APPENDICES 407 Ś aṅkara (only on the Cś Unknown11 Sanatsujāta) If he is identical with the great Ś aṅkara, his text is undoubtedly older than the oldest of the extant Mahābhārata MSS., but this does not necessarily prove its absolute authenticity; for, his testimony afects only the authenticity of the particular recension or version used by him, and nothing more. Nīlakaṇṭha Cn Poona, Bombay Government Collection (deposited at the Nīlakaṇṭha’s BORI), no. 29B of 1879–80 CE. 255 folios, with about 17 preference for a to 19 lines to a page (being written in three strips, with the smooth and eclectic original text in the middle, the number of lines varies); size text leads him not 11¾” x 7¾”. Devanāgarī characters. Fine modern paper only to indiscriminate with watermarks. Text with Nīlakaṇṭha’s commentary, but incorporation and the commentary begins here with the Sanatsujāta. The MS athetization, but also is modern but some of the parvans copy, at the end, the date to the avoidance of the original exemplars, one of the latest dates being Ś aka of the lectio diicilior 1680 (= ca. 1758 CE). Neatly written and fairly correct.12 or grammatical Poona, Bombay Government Collection (deposited at the irregularities with a BORI), no. 271 of Viśrāmbāg II. 142 folios, with 19 to 20 view to simpliication lines to a page; size 17½” x 8”. Devanāgarī characters; and normalization. dated Ś aka 1666 (= ca. 1744 CE). Indian paper. Text with In this way authentic Nīlakaṇṭha’s commentary. The writing is clear and fairly archaisms are correct; and the margins are clean. As it is written in three completely obliterated, strips, with the original text in the middle, the number of lines but unskillful varies from page to page. The following folios are missing: 1– conlation has often led 2, 24, 34, 42–44, 86, 96, 100–101, 107, 109, 111, 113–24, to fresh confusion and 129, 137–40.13 obscurity. 408 408 PHILOLOGY AND CRITICISM 6. Bhīṣmaparvan Commentary Sig. Manuscripts Version of Devabodha; Cd Baroda Oriental Institute MS no. 12982. Written on paper the commentary called in Devanāgarī characters. Folios 1–5 (folio 1a blank) carry Jñānadı̄pikā the text up to the Bhagavadgītā; and folios 1–9 (written sec. m., with fresh pagination; fol. 3b, 4a and 9a being left blank) carry the text from 6.41.4 to the end of the parvan, there being no commentary available for the Bhagavadgītā. The date of completion of the MS is October 26, 1766. Version of Caturbhuja; Cc Oriental Institute, Baroda; no. 13036. Old country paper, the commentary called size 11¾” x 4”, slightly frayed at the lower right-hand corner. Tātparyaprakāśikā (?) 42 folios, with 10 lines to a page and 48 letters to a line. Written in Devanāgarī characters; not very correct. It gives the commentary on the Bhagavadgītā only up to stanza 57 of chapter 2, and remarks—ataḥ paraṃ granthābhāvānna likhitāni. The commentary is again resumed at 6.41.3, and goes on to the end of the parvan (fol. 40a, line 1). After this point, there are, in the same handwriting, ive pages of what purports to be deva(ko) bodhasya gı̄tātı̣ ̄kā; but its authenticity is uncertain. The MS is undated, and looks not more than 150 or 200 years old. The date of Caturbhuja lies between 1275–1350 CE. Version of Vādirāja; Cv Govt. Oriental Library, Mysore; no. C1577. the commentary called An almost complete MS of Vādirāja’s commentary on the Lakṣālaṃkāra Mahābhārata is available in the Mysore Oriental Library, besides parts of the commentary on stray parvans in several libraries of South India. For this edition, we were able to use a Devanāgarī copy of the Mysore MS, bearing Library no. C1577 and dated Ś aka 1597 (1675 CE). The original MS was written on paper in Devanāgarī characters. Its total extent is folios 2–104, 151–445. The Bhīṣmaparvan covers folios 151–88, with eight lines to a page, and 44 letters to a line. Vādirāja alias Kavīndratīrtha was eighth in the succession-list of the High priests of the Mādhva sect, and died Ś aka 1261 (1339 CE). Devanāgarī Version Ca Poona, Bombay Government Collection (deposited at the of Arjunamiśra; the BORI), no. 482 of Viśrāmbāg I. Written in bold and legible commentary called Devanāgarī characters in country paper, of size 16” x 7”. 285 Bhāratārthadı̄pikā (or folios, with folios 114, 132, 133, 135 and 136 missing, and folios pradı̄pikā) 126, 130 and 137 partially torn. Gives about 11 lines to a page and about 60 letters to a line. Not very correct, but corrections made here and there by a later hand. Gives the text in the center and Arjunamiśra’s commentary above and below the text. For the Bhagavadgītā (which extends from fol. 34a–125b) the commentary given is the Subodhinı̄ of Ś rīdhara. Not dated, but seems about 300 years old. On the blank sides of the irst and the last folios, the MS is said to have belonged to “saṃsthān dṛaṃbal [?14].”15 409 APPENDICES 409 Baroda, Oriental Institute Library, no. 10801. The portion up to the Bhagavadgītā covers folios 1–33. Thereafter, the Bhagavadgītā (text and comm.) is altogether omitted. After the Bhagavadgītā, with a fresh pagination (1–160), the parvan is brought to an end. Devanāgarī characters. The commentary of Arjunamiśra is written above and below the text as usual. Size 13” x 5½”. In the same bundle, written on paper of the same size but probably in a diferent hand, is the text (without commentary) of the Ā śramavāsaparvan, which is dated V. Saṃvat 1753 (ca. 1697 CE). This may have been the approximate date for the Bhīṣmaparvan also. Both the MSS. are stated at the conclusion to have been the property of Bhaṭṭa Ś rī-Govindajita.16 Devanāgarī Version Cn Poona, Bombay Government Collection (deposited at the of Nīlakaṇṭha; the BORI), no. 483 of Viśrāmbāg I. Country paper, Devanāgarī commentary called characters. Size 16¼” x 6¾”. 336 folios, with 10–12 lines to the Bhāratabhāvadı̄pa (or a page and about 48 letters to a line. Clear and legible hand, dı̄pikā) fairly correct. A few corrections are noted in the margin (sec. As this commentary m.). Gives the text in the center and the commentary above is readily available in and below the text, as usual. Although the last colophon of the print, and is the latest commentary reads: iti śrı̄madarjunamiśrakṛtau bhāratārthadı̄pikāyāṃ of them, it was not bhı̄ṣmaparvavivaraṇaṃ samāptam, the commentary from beginning thought necessary to go to end is actually Nīlakaṇṭha’s Bhāratabhāvadı̄pa. Not dated.17 into any critical details Indore, from the private collection of Sardar M. V. Kibe. Not about the commentary numbered. Tough country paper, Size 16” x 6”. 338 folios, proper beyond the data without about 11 lines to a page, and 50 letters to a line. supplied in connection Devanāgarī characters; written in a neat and legible hand, with the MSS. Dn1 and fairly correct. Gives the text in the center and the commentary Dn2 already described of Nīlakaṇṭha above and below the text, as usual. No date is on pp. xxxv–xxxvi above. given.18 Incidentally, it is worth noting, however, that Ganapat Krishnaji’s edition of 1863 cannot be said faithfully to present the text underlying Nīlakanṭ̣ ha’s commentary in every detail; and not all the Cn MSS. agree amongst themselves. Commentary19 Sig. Manuscripts Vimalabodha’s C Poona, Bombay Govt. Collection (deposited at the BORI), Durghatạ ̄rthaprakāśinı̄?20 no. 84 of 1869–70; dated Ś aka 1724 (1802 CE). Yajña (or Sarvajña) C Transcribed from MSS. Nos. R2169 and R1625 belonging to Nā rā yaṇa’s Govt. Or. MSS. Lib., Madras. Bhāratatātparyasaṃgraha Rā mā nuja’s Viṣamaśl C As given in the India Oice Codices, nos. 470 and 1411. okatı̣ k̄ ā also called the Bhāvārthadı̄pikā 410 410 PHILOLOGY AND CRITICISM 7. Droṇaparvan Commentary Sig. Manuscripts Devabodha?21 Cd Scindia Oriental Institute, Ujjain, MS no. 482/5797. Written at Kāśī in Ś aka 1571 (= ca. 1649 CE). Appears to be the same MS as was used for the Sauptika and Strī parvans. The commentary is unaccompanied by the epic text. Arjunamiśra Ca Benares Sanskrit College, MS no. 2278. Incomplete. Caturbhujamiśra Cc Poona, Bombay Govt. Collection (now deposited at BORI), no. 495 of Viśrāmbāg (i). 339 folios. Indian paper. Size 16” x 6 4/5”. Devanāgarī characters, with 11 lines to a page. Bold legible writing. The MS, which is in good condition, contains the commentary of Caturbhuja Miśra. The text is in the middle, and the commentary runs above and below the text.22 Nīlakaṇṭha Cn Poona, Limaye Collection of the BORI, no. 377. 305 folios. Indian paper. Devanāgarī characters, with 11–13 lines to a page. Good legible writing. The MS contains the commentary of Nīlakaṇṭha and extends from the Droṇa to Rājadharma sub-parvan of the Ś ānti only.23 Bhor State Library, Bundle no. 1, unnumbered. 188 folios. Paper. Size 14½” x 7”. Devanāgarī characters, with 16 lines to a page. Dated Saṃvat 1846 (= ca. 1790 CE). Good and clear writing. The MS gives the commentary of Nīlakaṇṭha written above and below the text.24 Vādirāja Cv From a transcript from Madras (now with the BORI).25 8. Karṇaparvan Commentary Sig. Manuscripts Arjunamiśra Library of the Oriental Institute, Baroda, no. 12927. The Arjunamiśra’s text of this commentary is identical with that found in Da1. commentary on this The MS consists of 21 folios. It is dated Ś aka 1605, the irst parvan is less copious than of the dark half of Phālguṇa, corresponding to Saturday, that of Caturbhuja, but 3rd March, 1683 CE. The copyist’s name is Sukhadeva slightly more extensive Pandit.26 than that of Nīlakaṇṭha and Vādirāja. Caturbhuja Oriental Institute, Baroda, no. 12983. The MS consists of 24 folios. It is very corrupt at places, and was therefore used wherever possible. Nīlakaṇṭha Available in printed form in the Ganapat Krishnaji and other editions. MS Dn1 contains the text of this commentary, which generally agrees with the Vulgate. It is very scanty and does not help much for text constitution. Vādirāja’s commentary Unknown27 Lakṣālaṃkāra Vimalabodha’s? Bombay Government Collection (deposited at the BORI), commentary Viṣamaślokı̄ no. 84 of 1869–70. or Durghatạ ̄rthaprakāśinı̄28 41 APPENDICES 411 9. Ś alyaparvan Commentary Sig. Manuscripts Arjunamiśra’s Ca Baroda, Oriental Institute Library, no. 12931. Folios 1–8a. commentary Bhāratārth The MS, which is dated Ś aka 1736 (= 1814 CE), gives only the apradı̄pikā commentary without the text. It also contains Arjunamiśra’s (ca. 1450–1500 CE)29 commentary on the Sauptikaparvan and the Strīparvan Not of any signiicant (beginning from folio 8a). use for the constitution Ujjain, Scindia Oriental Institute Library, no. 482/5797. Folios of the text. 1b–9a. The MS, which is dated Ś aka 1571 (= 1649 CE) and which is said to have been written by one Nārāyaṇabhaṭṭa Mahokṣa, gives only the commentary without the text. It also contains the commentary on the Sauptikaparvan (beginning from folio 9a). The colophons in this MS are generally very brief. Curiously enough, however, there is, at the end of the commentary on Crit. Ed. adhy. 28, a colophon, which reads: iti śrı̄paramahaṃsācār yaśrı̄devabodhakṛtau bhāratadı̄pikāyāṃ śalyaparvavivaraṇaṃ samāptam. The ascription of this commentary, on the basis of this colophon, to Devabodha is obviously a mistake. This MS was carefully compared with the Baroda MS no. 12931 referred to earlier, and was found to be identical with it. Vādirāja’s Cv From P. P. S. Sastri’s transcript of the commentary of Vādirāja. commentary The commentary on the Ś alyaparvan consists of 67 adhyāyas Lakṣālaṃkāra (30 for the Ś alya and 37 for the Gadā). (ca. 1571 CE)30 This commentary is very scanty and of little use for the constitution of the text. Nīlakaṇṭha’s Cn Poona, Bombay (now Maharashtra) Government Collection commentary (deposited at the BORI), no. 240 of 1895–1902. 57 folios; Bhāratabhāvadı̄pa written on Indian paper, in Devanāgarī characters, with 10– (ca. 1680–1693 CE)31 12 lines to a page and about 40 letters to a line; size 15½” x 7”; undated; text with Nīlakaṇṭha’s commentary. Besides the Ś alyaparvan, the MS contains the Sauptika, the Aiṣīka and the Strī, the folios of each parvan numbered separately. The handwriting is uniform throughout. Though the MS contains the entire Ś alyaparvan, it is used here only for adhys. 1–28 of that parvan. It is also used as Dn1 for the critical edition of the Sauptika (and Aiṣīka) parvans.32 Poona, Bombay (now Maharashtra) Government Collection (deposited at the BORI), no. 506 of Viśrāmbāg I. 49 folios; written on country paper, in Devanāgarī characters, with about 12 lines to a page and about 50 letters to a line; size 15 1/3” x 6 7/8”. The handwriting is uniform and beautiful. The MS contains only adhys. 1–28 of the Ś alyaparvan. The commentary of Nīlakaṇṭha is given at the top and at the bottom of the folios.33 412 412 PHILOLOGY AND CRITICISM Poona, Bombay (now Maharashtra) Government Collection (deposited at the BORI), no. 507 of Viśrāmbāg I. 65 folios; written on country paper, in Devanāgarī characters, with about 12 lines to a page and about 50 letters to a line; size 16” x 6¾”. The handwriting is uniform and good. The commentary of Nīlakaṇṭha is given at the top and at the bottom of each page. The MS is undated and is in good condition. It contains only the last two sub-parvans of the Ś alyaparvan. Presumably it is the continuation of Dn2 used for adhys. 1–28 of that parvan and forms part of a complete codex of the Epic.34 Poona, Limaye Collection of the BORI, no. 375 and 376. Folios 1–49 for the adhys. 1–28 and folios 1–65 for adhys. 29–64; written on Indian paper, in good legible Devanāgarī characters, with about 12 lines to a page; size 15½” x 6 7/8”. Besides the complete Ś alyaparvan, the MS contains the Droṇa, the Sauptika, the Strī and the Rājadharma sub-parvan of the Ś ānti. It is used as Dn3 for adhys. 1–28 and as Dn2 for adhys. 29–63 of the present parvan. The commentary of Nīlakaṇṭha is given above and below the text. The MS is also used for the Critical Edition of the Droṇaparvan (as Dn1). 10. Sauptikaparvan Commentary Sig. Manuscripts Commentary of Cd Scindia Oriental Institute, Ujjain; no. 482/5797. This is the Devabodha?35 only MS of Devabodha available for this parvan. It is dated Ś aka 1571 (1649 CE), and the name of the scribe is given as “mahokṣopanāmra nārāyaṇabhatṭ ẹ na.” Commentary of Ca MS no. 12931, dated Ś aka 1736 (1814 CE), of the Oriental Arjunamiśra Institute Library, Baroda. It is not possible to know either the scribe or the place of this MS, as this is not mentioned in the colophon. Commentary of Cn None36 Nīlakaṇṭha Commentary of Cv Transcript of no. C1577 in the Mysore Oriental Library?37 Vādirāja 11. Strīparvan Commentary Sig. Manuscripts Commentary of Cd Scindia Oriental Institute, Ujjain; no. 482/5797. The original Devabodha?38 manuscript was written by Nārāyaṇabhaṭṭa Mahokṣa in Ś aka 1571. Commentary of Ca MS no. 12931, dated Ś aka 1736 (1814 CE), of the Oriental Arjunamiśra Institute Library, Baroda. Commentary of Cn Unknown39 Nīlakaṇṭha Commentary of Cv Unknown (possibly Sastri’s transcript?)40 Vādirāja 413 APPENDICES 413 12. Ś āntiparvan, Rājadharma and Ā paddharma parvans Commentary Sig. Manuscripts Arjunamiśra’s Ca Paris, Bibliothèque Nationale, no. 379. Written on country Bhāratārthadı̄pikā or paper in Ś āradā characters. Size 260 mm x 355 mm (10 2/ pradı̄pikā 5” x 14 1/5”). Gives about 25 lines to a page and about 30 According to letters to a line. The original portion of the MS writes the the editor, it is commentary in Ś āradā characters on the margins all around Arjunamiśra’s the text. The commentary is by Arjunamiśra, as is expressly practice, when dealing stated on the top margin of folio 127b (iti śrı̄arjunamiśrakṛtāyāṃ with a particularly mahā […] samāptaḥ ) and on the lower margin of the same page diicult passage from (natvā jagadguruṃ viṣṇuṃ tataḥ pitarameva ca | kriyate’rjunamiśreṇa the Mokṣadharma, to mokṣadharmapradı̄pikā ||). The commentary seems to have been give a modest preface written by a later copyist, who at times attempts to correct or a conclusion to his the originally written text (inter. lin. or marg.) so as to make commentary. it agree with the commentary.41 The copyist who wrote the commentary seems to have been a diferent person from the one who copied the smaller-sized folios, so that the MS before us is the handicraft of three diferent persons more or less separated from one another in time. The MS cannot be said to be very old.42 Poona, Bombay Government Collection of MSS. (deposited at the Bhandarkar Oriental Research Institute, being part of the old Vishram Bag Collection, I. 499). Written on country paper in clear and legible hand. Size 15.4” x 6.4”, with 10 or all lines to a page and about 48 letters to a line. The text is written in the center and the commentary of Arjunamiśra written above and below the text. Total number of folios 179. The MS gives the Rājadharma and the Ā paddharma subsections only, and is undated. The MS Da2, giving the Ā paddharma and numbered Vishram Bag I. 500, seems to be a continuation of the present MS.43 Poona, Bombay Government Collection of MSS. (deposited at the Bhandarkar Oriental Research Institute, being part of the old Vishram Bag Collection, I. 500). Written on country paper in good legible hand with the text in the center and the commentary of Arjunamiśra above and below the text. Folios 176; size 16” x 7”, with 10 lines to a page and 40–42 letters to a line. The MS contains only the Rājadharma and the Ā paddharma subsections, the Mokṣadharma subsection being absent. The MS is undated.44 Vimalabodha’s Cb An unnumbered Maithilī MS from the Rāja Library, Durghtạ ̄rthaprakāśinı̄ Darbhanga. 41 414 PHILOLOGY AND CRITICISM Nīlakaṇṭha’s Cn Bhor, State Library, Nos. 80, 81 and 82, the three numbers Bhāratabhāvadı̄pa or giving respectively the Rājadharma, the Ā paddharma and dı̄pikā45 the Mokṣadharma subsections of the Ś āntiparvan. Folios, (1680 CE)46 respectively 186, 60 and 450. Size 16½” x 6¾”. Written on country paper in Devanāgarī characters, with legible and careful hand. There are, on average, 12 lines to a page with 48 letters to a line. The text is written in the center and the comm. of Nīlakaṇṭha above and below the text. Dated Ś aka 1691 (ca. 1769 CE) (in the Rājadharma and Mokṣadharma subsections).47 (Only for the Rājadharmaparvan.) Poona, Vishalgad Collection of the B. O. R. Institute, no. 245. The MS was presented to the Institute by Vishalgad State. It is now numbered 245, and contains the Rājadharma only. Written on country paper in Devanāgarī characters in a good and legible hand, with 12–13 lines to a page and 40–50 letters to a line. Size 15.5” x 6.8”, giving text in the center and the comm. of Nīlakaṇṭha above and below the text. It is dated ca. 1782 CE.48 (Only for the Ā paddharmaparvan.) Poona, Bombay Government Collection (deposited at the Bhandarkar Oriental Research Institute) no. 241 of 1895– 1902. Written on country paper in Devanāgarī characters. Size 14” x 5.6” with 10 lines to a page and about 50 letters to a line. The text is written in the center and the comm. of Nīlakaṇṭha above and below the text. Legible but somewhat carelessly written; in tolerable condition. The MS contains the Ā paddharma as well as the Mokṣadharma, but the latter subsection is not used in the present edition. The MS gives no date either at the end of the Ā paddharma or of the Mokṣadharma.49 Vādirāja’s Lakṣālaṃkāra Cv As in a copy from P. P. S. Sastri’s library. Mokṣadharmaparvan Commentary Sig. Manuscripts Arjunamiśra’s Ca Paris, Bibliothèque Nationale, no. 379 (Ś 1).50 Bhāratārthadı̄pikā or pradı̄pikā Poona, Bombay Government Collection of MSS. (deposited at the Bhandarkar Oriental Research Institute, being no. 476 of the old Vishram Bag Collection I). The MS has 433 folios, size 15.5” x 6.5”, with 3 to 5 lines to a page and between 40 to 45 letters to a line. The MS contains only the Mokṣadharma subsection, the text being written in the center and the comm. of Arjunamiśra above and below the text. It is written on country paper in a neat, careful and legible hand. The MS is undated.51 415 APPENDICES 415 Poona, Bombay Government Collection of MSS. (deposited at the Bhandarkar Oriental Research Institute) no. 38 of 1879–80 CE. Written on country paper in neat and legible Devanāgarī characters. Size 15.5” x 6.7”. Total number of folios 433. The MS contains only the Mokṣadharma subsection of the Ś āntiparvan with the commentary of Arjunamiśra written above and below the text. The MS bears no date.52 Vimalabodha’s Cb As in Bombay Govt. Collection (deposited at the BORI), NO. Durghtạ ̄rthaprakāśinı̄ 167 of 1887–91. Undated. Nīlakaṇṭha’s Cn Bhor, State Library, no. 82 (Dn1).53 Bhāratabhāvadı̄pa or Poona, Bombay Government Collection of MSS. (deposited at dı̄pikā the Bhandarkar Oriental Research Institute), no. 29c of 1879–80 CE. Written on country paper in Devanāgarī characters. Size 11.6” x 8.1”, with 17–18 lines to a page and about 40 letters to a line. The text is written in the center and the comm. of Nīlakaṇṭha above and below the text. At the end of the Mokṣadharma we read: śake 1670 bahudhānyanāmasaṃvatsare āṣāḍaśuddhe ṣaṣtḥ ı̄ cantravāsare lekhakabhagavānbhatṭ ẹ na likhitoyaṃ pustakam || The MS is dated ca. 1758 CE.54 Paramānanda Cp As in MS no. 12899 from the Oriental Institute, Baroda. Dated Bhaṭṭācārya’s Ś aka 1594 (ca. 1672 CE). Mokṣadharmatı̣ ̄kā The commentary of Paramānanda Bhaṭṭācārya was available after 1360 CE55 only for the Mokṣadharma subsection, and that too in a rare MS kindly lent by the Oriental Institute of Baroda: no. 12899. The MS was written on old country paper, 10½” x 4½”. The date is Ś aka 1594 (ca. 1672 CE). Vidyāsāgara’s Cs Poona, Bombay Government Collection of MSS. (deposited Vyākhyānaratnāvali at the Bhandarkar Oriental Research Institute), no. 39 of (ca. 1350 CE)56 A 1879–80. Written on yellowish country paper in Devanāgarī characters with the text of the Mokṣadharma in the center, and the commentary of Vidyāsāgara on the same, named the Vyākhyānaratnāvali, above and below the text. 681 folios, with about 10 lines to a page and about 36 letters to a line. Undated.57 Tanjore, Saraswathi Mahal Library, no. 1141. Written on country paper in Devanāgarī characters. Size 16½” x 6½”; total number of folios 413, with text in the center and the commentary of Vidyāsāgara called the Vyākhyānaratnāvali above and below the text. Undated.58 Vādirāja’s Lakṣālaṃkāra Cv As in a copy from P. P. S. Sastri’s library. 416 416 PHILOLOGY AND CRITICISM 13. Anuśāsanaparvan Commentary Sig. Manuscripts Visạ maślokı̄ Ṭık̄ ā of Cb Poona, Bombay (now Maharashtra) Government Collection Vimalabodha?59 (deposited at the BORI), no. 84 of 1869–70. 88 folios; written (a.k.a. on paper, in Devanāgarī characters, with 14–15 lines to a Durghatạ ̄rthaprakāśinı̄) page and about 30 letters to a line. The handwriting is good (1150–1300 CE)60 and legible. The MS is dated Ś aka 1724 (= 1802 CE). Acknowledges his debt to Poona, Bombay (now Maharashtra) Government Collection Devabodha (also referred (deposited at the BORI), no. 167 of 1887–91. 67 folios; to as Devasvāmin) written on paper, in Devanāgarī characters, with about 12 lines to a page and about 45 letters to a line. The handwriting is bold and legible. The MS is undated. Bhāratārthapradı ̄pikā Ca MS no. 2539 in the collection of the Raghunath Temple of Arjunamiśra MSS. Library, Jammu. 4 folios; about 22 lines to a page and (a.k.a. about 44 letters to a line; written on paper in Devanāgarī Mahābhāratārtha(pra)dı̄pikā characters. The MS is dated Vikrama Saṃvat 1707 (=1651 or Bhāratasaṃgrahadı̄pikā) CE). (1450–1500 CE) Commentary closely follows the Viṣamaślokī of Vimalabodha; has also been deeply inluenced by the Bhāratārthapradı̄pikā of Devabodha Lakṣālaṃkāra of Vādirāja Cv MS no. C1577 in the Mysore Oriental Library.61 (1571 CE) Bhāratabhāvadı̄pa of Cn Poona, Vishalgad Collection of the BORI, no. 241. 361 Nīlakaṇṭha folios; written on Indian paper, in Devanāgarī characters, (1680–93 CE) with 14 lines to a page and about 32 letters to a line; size 11½” x 6 7/8”. The MS, which is dated Ś aka 1706 (= 1784 CE), is carefully written and in good condition.62 Poona, Bombay (now Maharashtra) Government Collection (deposited at the BORI), no. 29C of 1879–80. Written on paper, in clear and legible Devanāgarī characters, with 16–18 lines to a page and about 30–35 letters to a line for the text and 45–48 letters to a line for the commentary; size 12” x 8¼”. The dates vary between Saṃvat 1810 (= Ś aka 1675 = 1753 CE) and Saṃvat 1815 (= Ś aka 1680 = 1758 CE).63 Unidentiied Cu Poona, Bombay (now Maharashtra) Government Collection (deposited at the BORI), no. 61 of 1871–72. 330 folios, written on old Indian paper, in bold and clear Devanāgarī characters, with 12–13 lines to a page and about 45 letters to a line; size 12½” x 5½”. The MS is dated Saṃvat 1717 and Ś aka 1582 (= 1660–61 CE).64 Poona, BORI Mahābhārata Collection; not yet registered and numbered. 327 folios; written on paper, in bold and legible Devanāgarī characters, with 12–13 lines to a page and about 48 letters to a line for the text and about 52 letters to a line for the commentary. The MS is undated. 417 APPENDICES 417 14. Ā śvamedhikaparvan Commentary Sig. Manuscripts65 Commentary of Cn As printed in the vulgate edition. Nīlakaṇtha Commentary of Cv On the evidence of a transcript procured from Madras. Vādirāja 15. Ā śramavāsikparvan Commentary Sig. Manuscripts Commentary Cn Jammu, State Library; no. 3732. Written on paper in of Nīlakaṇtha Devanāgarī characters. Size 14½” x 8¾”, with 15 lines to a (Bhāratabhāvadı̄pa page. Each line contains about 35 letters. Total number of or –dı̄pikā) folios 28. The text is written in the center and the commentary, Bhāratabhāvadı ̄pa of Nīlakaṇṭha, above and below the text. It is undated. The MS contains the four concluding parvans.66 Commentary of Cv On the evidence of a transcript procured from Madras.67 Vādirāja 16. Mausalaparvan Commentary Sig. Manuscripts Commentary of Cn Indore, from the private collection of Sardar M. V. Kibe; Nīlakaṇtha unnumbered and undated. Written on paper in Devanāgarī characters. Size 16” x 6”, with about 11 lines to a page and 50 letters to a line. The MS is written in a neat and legible hand, and is fairly correct. It gives the text in the center and the commentary of Nīlakaṇṭha above and below the text, as usual. The MS seems to have formed part of a complete MS of the Mahābhārata, of which 11 parvans are now available, those wanting being Sabhā, Droṇa, Karṇa, Ś alya, Sauptika, Strī and Ś ānti.68 Jammu, State Library; no. 3975. Written on paper in Devanāgarī characters. Size 15¼” x 7”, with 12 lines to a page and 40–44 letters in a line. The text is written in the center and the commentary, Bhāratabhāvadīpa of Nīlakaṇṭha, above and below the text. Total number of folios 12. The MS begins from fol. 1b and ends with fol. 13b. It is undated.69 Commentary of Cv Mysore Oriental Library; no. C 1577. Dated Ś aka 1597 (ca. Vādirāja 1675 CE).70 418 418 PHILOLOGY AND CRITICISM 17. Mahāprasthānikaparvan Commentary Sig. Manuscripts Commentary of Cn Indore, from the private collection of Sardar M. V. Kibe; Nīlakaṇtha unnumbered and undated. Written on paper in Devanāgarī characters. Size 15.8” x 6”, with about 7–10 lines to a page and 50–52 letters in a line. The MS is written in a neat and legible hand, and is fairly correct. It gives the text in the center and the commentary of Nīlakaṇṭha above and below the text. The MS seems to have formed part of a complete MS of the Mahābhārata, of which 11 parvans are now available, those wanting being Sabhā, Droṇa, Karṇa, Ś alya, Sauptika, Strī and Ś ānti. Total number of folios 6. The MS for the Mahāprasthānika begins from fol. 1b and ends with fol. 6a.71 Jammu, State Library; no. 3976. Written on paper in Devanāgarī characters. Size 15¼” x 7”, with 12 lines to a page and 40–45 letters in a line. The text is written in the center and the commentary, Bhāratabhāvadı̄pa of Nīlakaṇṭha, above and below the text. Total number of folios 7. The MS begins from fol. 1b and ends with fol. 7b. It is undated.72 Commentary of Cv Mysore Oriental Library; no. C 1577. Dated Ś aka 1597 (ca. Vādirāja 1675 CE).73 18. Svargārohaṇaparvan Commentary Sig. Manuscripts Commentary of Cn Indore, from the private collection of Sardar M. V. Kibe; Nīlakaṇtha unnumbered and undated. Written on paper in Devanāgarī characters. Size 15.8” x 6”, with about 7–10 lines to a page and 50–52 letters in a line. Total number of folios 14. The Svargārohaṇaparvan in the MS begins from fol. 1b and ends with fol. 14. It gives the text in the center and the commentary of Nīlakaṇṭha above and below the text. For a detailed description of Dn1, see the description of the same MS in the Mausalaparvan.74 Jammu, State Library; no. 3975. Written on paper in Devanāgarī characters. Size 15¼” x 7”, with 11 lines to a page, and 42 letters to a line. The text is written in the center and the commentary, Bhāratabhāvadı̄pa of Nīlakaṇṭha, above and below the text. Total number of folios 14. The MS begins from fol. 1b and ends with fol. 14. It contains the Mausala and the Svargārohaṇa parvans.75 Commentary of Cv Mysore Oriental Library; no. C 1577. Dated Ś aka 1597 (ca. 1675 Vādirāja CE).76 Notes: 1 Sukthankar notes that this manuscript agrees “page for page, with Da1.” However, he also states that “[as] there are many small diferences between them; neither can be a direct copy of the other; they must go back to a more remote common source.” Of the two, he thinks Da1 is the older and less corrupt manuscript. Sukthankar, “Prolegomena,” xvi. 419 APPENDICES 419 2 Sukthankar does not mention his source for the commentary of Devabodha in the “Prolegomena,” but in his “Notes on Mahā bhā rata Commentators,” 197–98. These details have been sourced from there. Sukthankar also mentions “another MS of the commentary in the Asiatic Society of Bengal, which was also consulted by [him],” but does not provide further details of this manuscript. See ibid., 197, n. 3. 3 This is the manuscript D1 used for the critical edition of this parvan. Edgerton additionally notes of this manuscript that it is “written in a slovenly hand and not very correctly, but extensively corrected […] [The] [t]ext of the commentary is far more corrupt, indeed extremely corrupt, and has not been corrected. There is no regular correspondence between the readings of Devabodha and those of the text given in this MS, which is deinitely not that of Devabodha. In general it is a relatively good D version.” Edgerton, “Introduction,” xv (italics in original). 4 D2 of the critical edition of this parvan. Edgerton additionally notes of this manuscript that it is “very well and clearly written; [and] the text is generally quite correct. […] The text of the commentary, however, is in little better shape than D1, both being frightfully corrupt, though fortunately their corruptions are diferent and independent of each other; by combining the two one can usually make a reasonable guess as to what Devabodha wrote. As in the case of D1 the text found in the MS is deinitely not that of Devabodha.” Ibid. (Edgerton’s italics). 5 Edgerton’s citations from Cv are actually a mix of two manuscripts. He notes that “The commentary, titled Lakṣā laṃkā ra, is described on p. xi of Sastri’s introduction, and the text of the single MS which he then knew is printed on pages 597–665. Before the book was printed, however, Sastri secured another MS of this commentary, which often corrects corruptions in the irst MS; variants of this second MS are cited by Sastri at the end, on nine pages numbered separately 1–9. My citations often make silent use of these corrections.” Ibid., xxiv. 6 Sukthankar does not assign a siglum to Arjunamiśra’s commentary, as the copy in his possession was “found to be full of clerical errors, and was not exactly suitable for collation.” Sukthankar does not make quite clear whether the manuscript was full of errors or the “rough copy” provided to him by the principal of the College. Based on his reference to “clerical” rather than “scribal” errors, we assume it was the latter. Two further copies of the commentary in the Sanskrit College at Benares (MSS no. 2279 and no. ग 16 पु 34) were incomplete. However, Sukthankar did occasionally consult the copy in his possession and he notes that “the text used by Arjunamiśra is, as usual, of a superior type, and his commentary would repay careful study, if good manuscripts of the commentary could be obtained.” Sukthankar, “Introduction,” xi. 7 The editor notes of this manuscript that it is “an old and much damaged Devanā garī palm-leaf copy belonging to the Royal Asiatic Society of Bengal, [and that it] is fragmentary and incomplete. As the MS presents the commentary unaccompanied by the epic text actually commented upon, the variants regarding stray words and phrases only, and not with regard to longer passages or the entire text, could be found and collated. It is partly for this reason and partly because the MS is fragmentary, [that] no inference would be safe regarding uncommented passages or omissions by Devabodha.” De, “Introduction [to the Udyogaparvan],” xiii. Note that, according to De, the man- uscript numbered 3398/5755 in Haraprasad Shastri’s catalog “ascribed to Devabodha in its inal colophon […] is really Sarvajña-Nā rā yaṇa’s commentary.” Ibid., xiii, n. 3. 8 The editor notes that the genuineness of portions of the edition were veriied by comparing them with “two Bengali MSS belonging respectively to the Calcutta Sanskrit College (no.  311)  and the Royal Asiatic Society of Bengal (no. 3373/4035 B); […] it was found that the printed text, though not critically edited, may be taken as reliable on the whole.” De, “Introduction [to the Udyogaparvan],” xiv. 9 This is the manuscript Ds1 used in the critical edition of this parvan. 10 This is the manuscript Ds2 used in the critical edition of this parvan. 11 De does not mention his source for this commentary. It may have been the Gujarati Press edition (we have been unable to examine this edition and therefore cannot conirm whether Ś aṅkara’s com- mentary is part of it). 12 This is the manuscript Dn1 used in the critical edition of this parvan. 13 This is the manuscript Dn2 used in the critical edition of this parvan. 14 The relevant page of the critical edition is missing here and is replaced by a photocopy; our tran- scription of the name may not be correct, as the Devanā garī characters were blurred. 15 This is the manuscript Da1 used for the critical edition of this parvan. 16 This is the manuscript Da2 used for the critical edition of this parvan. 17 This is the manuscript Dn1 used for the critical edition of this parvan. 420 420 PHILOLOGY AND CRITICISM 18 This is the manuscript Dn2 used for the critical edition of this parvan. 19 These three manuscripts were “far too meager, and have been drawn upon only occasionally in the Critical Notes.” Belvalkar, “Introduction [to the Bhīṣmaparvan],” xii. In addition to these three, Belvalkar also notes that “commentaries passing under the name of Sṛṣṭidhara (MS. in the RASB of Calcutta) and of Vidyā sā gara (MS. once reported as being in the Trivandrum Palace Library)” exist. Belvalkar could not examine them, but thinks they are “presumably equally meagre.” Ibid. 20 The ascription is problematic. See the discussion of this manuscript in the preceding summary. 21 The ascription is problematic. See the discussion of this manuscript in the preceding summary. 22 This is the manuscript Dc1 used for the critical edition of this parvan. 23 This is the manuscript Dn1 used for the critical edition of this parvan (Dn3 in the critical edition of the Ś alyaparvan). 24 This is the manuscript Dn2 used for the critical edition of this parvan (Dn1 in the critical edition of the Karṇaparvan). 25 This must be P. P. S. Sastri’s transcript, used also by all the other editors. 26 The editor notes that the same commentary “abridged […] with a few additions and alterations here and there” is also included in the manuscript D5. Its text “seems to be substantially the same as found in [the] MSS Da1, D1 and D2.” He therefore considers these manuscripts to represent “Arjunamiśra’s version of the Karṇaparvan.” Vaidya, “Introduction  [to the Karṇaparvan],” xxi. 27 Vaidya does not mention his source for the commentary. His sole note on it is: “Vā dirā ja’s commen- tary called Lakṣā laṃkā ra on the Karṇaparvan is also very scanty. It follows the text of the parvan as found in the S recension.” Ibid. 28 The ascription is problematic. See the discussion of this manuscript in the preceding summary. 29 Dandekar cites P.  K. Gode’s article in Indian Culture 2, 141–46 as his source. Unfortunately, all attempts to trace this journal have failed. 30 Dandekar cites Gode, “Dr.  P.  P. S.  Sastri, and the Date of Vā dirā jatīrtha,” 203–10 as his source. Gode controverts the earlier dates ascribed to Vā dirā ja and suggests a date of 1571 CE for the writing of his commentary. 31 Dandekar cites Gode, “Nīlakaṇṭha Caturdhara, the Commentator of the Mahā bhā rata—His Genealogy and Descendants,” 146–61 as his source. 32 This is the manuscript Dn1 used for adhyāyas 1–28 only of the critical edition of this parvan. 33 This is the manuscript Dn2 used for adhyāyas 1–28 only of the critical edition of this parvan. 34 This is the manuscript Dn1 (not an error; Dandekar uses the same siglum, but note that in the second half of the parvan it refers to a diferent manuscript) used for adhyāyas 29–64 only of the critical edition of this parvan. 35 The ascription is problematic. See the discussion of this manuscript in the preceding summary. 36 The editor appears not to have used a manuscript at all for the commentary of Nīlakaṇtha, for he notes: “The commentary of Nīlakaṇṭha is available in printed editions, and its basis is the Vulgate text. Citations from it are merely for pāt ̣hāntaras, or where its readings difer from those of the Vulgate.” Velankar, “Introduction,” xxiii. 37 The editor does not provide the manuscript particulars, but notes that “the commentary of Vā dirā ja, called Lakṣab̄ haraṇa is available on the whole of the Mbh. in a manuscript copy which was used by P. P. S. Sastri for the Madras edition.” Ibid., xxiii. This must be the same transcript the editors of the other parvans used and therefore the manuscript in question must be no. C1577 in the Mysore Oriental Library. 38 The ascription is problematic. See the discussion of this manuscript in the preceding summary. 39 Paranjpe does not provide any details of his source. See Paranjpe, “Introduction,” xix. He could be using either Dn1 or Dn2 (Collection of the Raja of Bhor, no. 2152 with the com. of Nīlakaṇṭha; D. A. V. College, no. 59, with Nīlakaṇṭha’s commentary) or, if he is following the practice of the editor of the preceding parvan, he may also be relying on one of the printed editions. 40 See preceding note. The editor’s source may be the Sastri transcript. 41 Belvalkar expressly airms this conclusion in his discussion of the commentary. “Regarding the passages from Ca as given on the margins of the MS Ś1, it is to be noted that in several places Ca explains text-reading and star-passages which Ś1 does not give, thereby proving that Ca quotations as supplied on the margins of Ś1 have no reference to the actual text as written in the centre.” Belvalkar, 421 APPENDICES 421 The Śāntiparvan: Introduction, cxxxix and see ibid.: “It is thus clear that Ca extracts in Ś1 have no nec- essary relation to the text of Ś1.” 42 This is the manuscript Ś1 used for the critical edition of this parvan. 43 This is the manuscript Da1 used for the critical edition of this parvan. 44 This is the manuscript Da2 used for the critical edition of this parvan. 45 Belvalkar’s statements on the number of Nīlakaṇṭha commentaries used in the critical edition are scattered and confusing. He notes that “The commentary of Nīlakaṇṭha, named the Bhāratabhāvadıp̄ a or dı̄pikā, printed in the Vulgate edition of the Mahā bhā rata, was used, in the present edition of the Ś antiparvan, from MS Dn1 which was available for all the three sub-sections, as also from Dn2 which was available for the Rā jadharma, Dn3 which was available for the Ā paddharma, and Dn4 which was available for the Mokṣadharma” (Belvalkar, The Śāntiparvan: Introduction, clv), but this is incorrect, as a brief perusal of his account of these manuscripts shows. The correct state of afairs is as represented in the following table: Dn1 Rā. Ā p. Mo. Dn2 Rā. Dn3 Ā p. Mo. (not used) Dn4 Rā. (not used) Mo. Thus, the igures for the Rā jadharmaparvan are four manuscripts (two used, one not used, plus the additional D. A. V. College, Lahore manuscript, no. 2935); for the Āpaddharmaparvan two (both used); and for the Mokṣadharmaparvan three (two used, one not used). The confusion in Belvalkar’s statement is caused by the “available,” which suggests that these manuscripts were not available for certain parts, when what he means is “used” or “used only for.” 46 Belvalkar cites Gode, Studies in Indian Literary History, vol. 2, 467f. as his source. 47 This is the manuscript Dn1 used for the critical edition of this parvan. 48 This is the manuscript Dn2 used for the critical edition of this parvan. 49 This is the manuscript Dn3 used for the critical edition of this parvan. 50 For details of this manuscript, see the comments in the preceding section under the Rā jadharma and Ā paddharma parvans. 51 This is the manuscript Da3 used for the critical edition of this parvan. 52 This is the manuscript Da4 used for the critical edition of this parvan. 53 For details of this manuscript, see the comments in the preceding section under the Rā jadharma and Āpaddharma parvans. 54 This is the manuscript Dn4 used for the critical edition of this parvan. Belvalkar notes of this man- uscript: “Besides the Mokṣadharma sub-section used for the Crit. Ed., the MS [Dn4] contains the Rā jadharma sub-section (written by the same Bhagavā nabhaṭṭa in Ś aka 1678, Caitra; not used in the Ś ā ntiparvan edition), as well as the Dā nadharma sub-section of the Anuśā sanaparvan, copied in Ś aka 1675, Śrā vaṇa, by a scribe named Rā jeśvara (in the MS, the Dā nadharma is said to belong to the Ś ā ntiparvan itself); the Aśvamedhaparvan (undated); theĀ śramavā sikaparvan (written Ś aka 1677; Mā rgaśīrṣa); the Mausalaparvan (Ś a ka 1677; Mā rgaśīrṣa); the Mahā prasthā nikaparvan (Ś aka 1677; Pauṣa) and the Svargā rohaṇaparvan (Ś aka 1677, Mā rgaśīrṣa). The hand-writing of the con- cluding Books difers from that of the Rā ja- and Mokṣa- sub-sections, as well as from that of the Dā nadharma sub-section. It seems, therefore, to be a patch-work MS, not perhaps very reliable.” Belvalkar, The Śāntiparvan: Introduction, lxxx. 55 The date is Belvalkar’s. He calculates it on the basis of the fact that “on fol. 52 [of the manu- script], Paramā nanda quotes  – phalavyāpyātvamevāsya śāstrakṛdabhirnivāritam | brahmaṇyajñānanāśāya vṛttivyaptiriheṣyate (= pañcadaśı,̄ 7.90cd, 92ab) ||.” Ibid., clviii. 56 Belvalkar notes that “according to Dr. V. Raghavan, the commentator lived under Kā madeva, the Kadamba ruler of Goa, one of whose inscriptions bears the date Ś aka 1315 (cir. 1393 CE).” Ibid., clix. According to Belvalkar, this places Vidyā sagara around 1350 CE. The reference is to V. Raghavan’s article in The Annals of Oriental Research IV (part I, Sanskrit section) (1939–40): 1–5. We could not trace this source. 57 This is the manuscript Ds1 used for the critical edition of this parvan. 58 This is the manuscript Ds2 used for the critical edition of this parvan. 42 422 PHILOLOGY AND CRITICISM 59 The ascription is problematic. See the discussion of this manuscript in the preceding summary. 60 Dandekar cites Gode, “Date of Vimalabodha’s Commentary on the Mahā bhā rata,” 395–97 and Gode, “New Light on the Chronology of the Commentators of the Mahā bhā rata,” 107 as his source. 61 Dandekar does not describe the manuscript. He refers the reader to its description in the Descriptive Catalogue of the library of 1922 (the manuscript is described on p. 155). 62 Dn1 of the critical edition of this parvan. 63 Dn2 of the critical edition of this parvan. 64 This manuscript is erroneously identiied as Dn3 of the critical edition of this parvan, but, on further examination, it became clear that the commentary it contains is not Nīlakaṇṭha’s but an unnamed commentator’s. Nonetheless the text of the manuscript substantially agrees with that of the version of Nīlakaṇṭha. The editor does not mention whether the commentary bears a name. 65 The editor of this parvan, R.  D. Karmarkar, does not provide any further information on these manuscripts or on the commentaries/commentators themselves. 66 The editor notes that Nīlakaṇṭha’s commentary was used in “the present edition of the Āśramavā sikaparvan, from MS Dn1 which was available for the four concluding parvans,” but this must be an oversight. According to the detailed account of the manuscripts, the manuscript Dn1 (Poona, Bombay Govt. Collection [deposited at the BORI]; no. 233 of 1895–1902) is “without the commentary, but the text generally agrees with Cn.” The editor must mean Dn2, which does it his description of Cn (the information above therefore has been sourced from the entry for Dn2). See Belvalkar, “Introduction [to the Āśramavāsikaparvan],” xxvii, xvii and xviii. 67 Karmarkar does not provide further details of this manuscript or of the text contained therein. 68 The editor notes that “the commentary of Nīlakaṇṭha, named the Bhā ratabhā vadīpa or -dīpikā , printed in the Vulgate edition of the Mahā bhā rata, was used, in the present edition of the Mausalaparvan, from MS Dn1 which was available for the last four concluding Parvans.” Belvalkar, “Introduction [to the Mausalaparvan],” xxiii. In this case, the editor’s Dn1 (that is, the manuscript from the private collection of Sardar M. V. Kibe) does match the description, but, if it contained the last four parvans, why was it not used for the Ā śramavā sikaparvan? Or has Belvalkar wrongly listed the Bombay Govt. Collection manuscript (no. 233 of 1895–1902) as his Dn1 when he meant, instead, the Kibe manuscript? We could not resolve the contradiction. 69 Contrary to the information in the “Detailed Discussion of the MSS” (see preceding note), in the list of manuscripts constituting the critical apparatus (ibid., ix–xi), the editor notes that “Cn,” the “Commentary of Nīlakaṇṭha,” has been used “as in Dn” (ibid., xi). In that case, the Cn must refer not only to the commentary as contained in Dn1 but also in Dn2, which is fortunate, as this man- uscript does contain the commentary. Consequently, the table above also lists this manuscript (that is, Dn2). 70 The editor comments on this manuscript as follows: “Vā dirā ja’s Commentary on the Mahā bhā rata in an almost complete form is available in the Mysore Oriental Library; no. C 1577. Dated Ś ā ka 1597 (ca. 1675 CE). The late P. P. S. Sastri, the editor of the so-called ‘Southern Recension’ of the Epic, got together a complete copy of Vā dirā ja’s Commentary in modern hand, which, after the completion of that edition, was procured by the B. O. R. Institute for use in the Critical Edition. The commentary is known as the Lakṣā laṃkā ra on the strength of the commentator’s concluding stanza of the Introduction to the Ādiparvan.” Ibid., xxiv. Note that this is the same manuscript as used for Vā dirā ja’s commentary in the critical edition of the Ā pnuśā sanaparvan and, in all likelihood, also in the critical edition of the Ā śvamedhikaparvan (cf. Karmarkar’s comment: “on the evidence of a transcript procured from Madras”). 71 The editor notes that “the commentary of Nīlakaṇṭha, named the Bhā ratabhā vadīpa or -dīpikā , printed in the Vulgate edition of the Mahā bhā rata, was used, in the present edition of the Mahā prasthā nikaparvan, from MS Dn1 which was available for the four concluding Parvans.” Belvalkar, “Introduction [to the Mahāprasthānikaparvan],” xx. This is the same manuscript as the one used in the Mausalaparvan (as Dn1). 72 Contrary to the information in the “Detailed Discussion of the MSS” (see preceding note), in the list of manuscripts constituting the critical apparatus (ibid., ix–xi), the editor notes that “Cn,” the “Commentary of Nīlakaṇṭha,” has been used “as in Dn.” Ibid., xi. In that case, the Cn must refer not only to the commentary as contained in Dn1 but also in Dn2, which is fortunate, as this manu- script does contain the commentary. Consequently, the foregoing table also lists this manuscript (that is, Dn2). 423 APPENDICES 423 73 Belvalkar’s comments on this manuscript are almost identical with those in the introduction to the Mausalaparvan and are therefore not repeated here. See ibid., xx. 74 The editor notes that “the commentary of Nīlakaṇṭha, named the Bhā ratabhā vadīpa or -dīpikā , printed in the Vulgate edition of the Mahā bhā rata, was used, in the present edition of the Svargā rohaṇaparvan, from MS Dn1 which was available for the four concluding Parvans.” Belvalkar, “Introduction [to the Svargārohaṇaparvan],” xxv. 75 Contrary to the information in the “Detailed Discussion of the MSS” (see preceding note), in the list of manuscripts constituting the critical apparatus (ibid., ix–xi), the editor notes that “Cn,” the “Commentary of Nīlakaṇṭha,” has been used “as in Dn.” Ibid., xi. In that case, the Cn must refer not only to the commentary as contained in Dn1 but also in Dn2, which is fortunate, as this manuscript does indeed contain the commentary. Consequently, the foregoing table also lists this manuscript (that is, Dn2). 76 Belvalkar’s comments on this manuscript are almost identical with those in the introduction to the Mahā prasthā nikaparvan and are therefore not repeated here (see also n. 12). See ibid., xxv. 42 425 17. COMMENTARIES ON THE BHAGAVADGĪTĀ Commentaries on the Bhagavadgītā signiicantly exceed those on the Mahā bhā rata, in part because the text was considered—at least since Śaṅkarā cā rya’s eighth-century com- mentary, the Gı̄tābhāṣya—part of the prasthāna trayı̄ or the three canonical works (the other two are the Upaniṣads and the Brahmasū tras) every orthodox (āstı̄ka) philosopher in India had to comment on. In contrast to the Mahā bhā rata commentators, the names of the Gītā commentators are well known and the major commentaries have been consistently in print.99 For this reason, this appendix does not devote a fuller discussion to either the commentaries100 or the commentators’ dates and precedence.101 Only the commentaries consulted for the critical edition are cited.102 Editions and sources are also mentioned only where relevant.103 Commentary Sig. Manuscripts Ā nandavardhana’s comm. Cā As found in the Bombay Government Collection called Jñānakarmasamuccaya1 (deposited at the BORI), manuscript no. 179 of 1883–84. Bhāskara’s comm. called Cb As found in a fragmentary MS, readings from which Bhagavadāśayānusaraṇa2 T. R. Cintamani reported in his edition of Ck, Madras, 1941. Caturbhuja’s comm. called Cc As found in the MS no. 13036 of Baroda Or. Institute, Tātparyaprakāśikā (?)3 already mentioned. The Gītā comm. in this MS covers only the irst two chapters. Abhinavagupta’s Cg Srinagar, MS belonging to the Shri-Rāma-Trika-Ā shram, comm. called Fateh Kadal, which Brahmachari Lakshman Raina used Bhagavadgı̄tārthasaṃgraha4 for his edition of the Bhagavadgītā with Abhinavagupta’s commentary, Srinagar, 1933. The MS is written on very old Kaśmīrī paper in Ś āradā characters, and forms the basis of the printed text.5 Rājānaka Rāmakaṇṭha’s (or Ck Poona, Bombay Government Collection (deposited at Rāmakavi’s) comm. called the BORI), no. 424 of 1875–76. Written on Kaśmīrī Sarvatobhadra6 paper in Śāradā characters and bound in a book-form. 581 folios, with 14–16 lines to a page and 11–16 letters to a line. Size 7¼” x 4”. Written by at least two hands. Occasional marginal additions and corrections in Ś āradā characters. The text and the commentary are written continuously. If Laukika Saṃvat 85 in Ś 1 corresponds to 1739 CE, the present MS may go back to 1754 CE. The MS looks to be a couple of centuries old.7 426 426 PHILOLOGY AND CRITICISM Commentary Sig. Manuscripts Poona, Bombay Government Collection (deposited at the BORI), no. 423 of 1875–76. Written on Kaśmīrī paper in Ś āradā characters. The irst six chapters (folios 1–64) written in upright book form, and the remaining chapters, probably by the same hand, in oblong pothı̄ form, numbered 1–96. The MS is bound in leather. Size 12” x 7”. About 24–27 lines to a page and 20 letters to a line in the irst part; about 13–14 lines to a page and 48 letters to a line in the second part. Occasional marginal corrections. The text and the commentary are written continuously, yellow pigment being used to diferentiate the text. If the Laukika year is taken to represent the cycle following the one intended in Ś 1, the present MS would be (15+27 =) 42 years younger than Ś 1—that is, it would belong to 1781 CE.8 Vallabha’s comm. called Cl As in the Gujarati Press edition (1935) of the BG. with 11 Sattattvadı̄pikā (or Tattvadı̄pikā)9 commentaries. The subcommentary of Puruṣottama on Cl is at times cited. Madhva’s comm. called Cm As in the Gujarati Press edition. The subcommentary of Gı̄tābhāṣya10 Jayatīrtha on Cm is also occasionally cited. Nīlakaṇṭha’s comm.11 Cn Poona, Bombay Government Collection (deposited at the BORI), no. 483 of Viśrāmbāg I. Country paper, Devanāgarī characters. Size 16¼” x 6¾”. 336 folios, with 10–12 lines to a page and about 48 letters to a line. Clear and legible hand, fairly correct. A few corrections are noted in the margin (sec. m.). Gives the text in the center and the commentary above and below the text, as usual. Although the last colophon of the commentary reads: iti śrı̄madarjunamiśrakṛtau bhāratārthadı̄pikāyāṃ bhı̄ṣmaparvavivara ṇaṃ samāptam, the commentary from beginning to end is actually Nīlakaṇṭha’s Bhāratabhāvadı̄pa. Not dated.12 Indore, from the private collection of Sardar M. V. Kibe. Not numbered. Tough country paper. Size 16” x 6”. 338 folios, with about 11 lines to a page and 50 letters to a line. Devanāgarī characters; written in a neat and legible hand, fairly correct. Gives the text in the center and the commentary of Nīlakaṇṭha above and below the text, as usual. No date is given at the end. The MS, however, seems to have formed part of a complete MS of the epic of which 11 parvans are not available, those wanting being Sabhā, Droṇa, Karṇa, Ś alya, Sauptika, Strī and Ś ānti.13 Rāmānuja’s comm. called Cr As in the Gujarati Press edition. The subcommentary, Gı̄tābhāṣya14 Tāparyaprakāśikā of Veṅkaṭanātha, is occasionally cited. 427 APPENDICES 427 Commentary Sig. Manuscripts Ś aṅkara’s comm. called Cś As given in the Ā nandāśrama edition, Poona (1909), and Gı̄tābhāṣya15 other standard editions.16 Dhanapati’s subcommentary Cu As in the Nirṇaya-Sāgar Press ed. (1936) of the BG. with on Ś aṅkara’s Bhāṣya, called Eight Commentaries. Bhāṣyotkarṣadı̄pikā on Cś17 Vādirāja’s commentary18 Cv Govt. Oriental Library, Mysore; no. C1577. Notes: 1 The text of this commentary is available in Belvalkar’s edition titled Śrı̄mad-Bhagavad-Gı̄tā with the “Jñānakarmasamuccaya” Commentary of Ānanda[vardhana]. 2 Bhā skara’s commentary is available in D. Subhadropadhyaya’s edition of 1965. Only the irst nine adhyāyas of this commentary are preserved. 3 No printed editions of Caturbhuja’s commentary appear to exist. 4 Abhinavagupta’s commentary is available in Pansikar’s edition. The earlier edition  by  Lakshman Raina (a.k.a. Lakshman Joo) is the source of Belvalkar’s text. The text has been translated twice: by Arvind Sharma and Boris Marjanovic. 5 This is the manuscript Ś3 used for the critical edition of this parvan. Belvalkar notes that Raina also used two other manuscripts, but as he only noted nine variant readings for the whole poem, “the printed text can reasonably be taken to represent Raina’s best MS.” Belvalkar, “Introduction [to the Bhīṣmaparvan],” lvi–lvii. 6 Rā jā naka’s commentary is available in the editions  by S. N. Tadpatrikar, T. R. Chintamani and Madhusudan Kaul Shastri. 7 This is the manuscript Ś4 used for the critical edition of this parvan. 8 This is the manuscript Ś5 used for the critical edition of this parvan. 9 Vallabha’s commentary is available in the Sadhale edition. We could not determine whether it also exists in an independent edition. 10 This commentary is available in the Sadhale edition and, with a translation, in Subba Rao, trans., Bhagavadgı̄tābhāṣya by Madhvācār ya. 11 Nīlakaṇṭha’s commentary is available in both the Sadhale and Pansikar editions and also in the Kinjawadekar edition (see Appendix 2). 12 This is the manuscript Dn1 used for the critical edition of this parvan. 13 This is the manuscript Dn2 used for the critical edition of this parvan. 14 The text of the commentary is available in the Sadhale edition and in the edition by P. B. Annang aracariyar. At least two translations—by M. R. Sampatkumaran and Swāmī Ādidevānanda—exist. 15 Ś aṅkara’s commentary is probably the most widely published. Besides occurring in both the Sadhale and Pansikar editions, it is also available in The Works of Sri Sankaracharya, 20  vols.  Three English translations—by A. G. Warrier, Swāmī Gambhīrānanda and Alladi Mahadeva Sastri—exist. A criti- cally edited text by Dinkar Vishnu Gokhale also exists, but we could not examine this edition. 16 We could not properly identify this edition. 17 Dhanapati’s commentary is available in the Pansikar edition. We could not determine whether it also exists in an independent edition. 18 No printed editions of Vā dirā ja’s commentary appear to exist. 428 429 18. THE USE OF VENN DIAGRAMS TO DEPICT MANUSCRIPT RELATIONSHIPS Thus the unfacts, did we possess them, are too imprecisely few to warrant our certitude, the evidencegivers by legpoll too untrustworthily irreperible where his adjugers are semmingly freak threes but his judicandees plainly minus twos. —James Joyce, Finnegans Wake In the major portion of this book we focused on the German Mahā bhā rata critics because their work, as noted in the introduction, poses the greatest single obstacle to the correct reception, interpretation and use of the Mahā bhā rata critical edition. This book also focused on genealogical reconstructions, whether based on an explicit or an implicit stemma. But what about other methods? And what about the contributions of other scholars? We would be misunderstood if readers thought we were concerned solely with the genealogical-reconstructive method. Rather, our broader argument is that, even when scholars evolved alternatives to the stemmatic method, they did so precisely in order to vali- date the German Indologists’ of a heroic, bardic epic and its “Brahmanic” corruption. As an example, this concluding appendix examines John L. Brockington’s work. We specif- ically focus on his suggestion that, given the Mahā bhā rata’s luid, oral nature, the Venn diagram is better suited to depict the relationships between manuscripts than a stemma codicum. We irst reconstruct Brockington’s views of the Mahā bhā rata critical edition. We then trace those views back to their German sources. Finally, we demonstrate how, even though he does not grasp the concept of a Venn diagram, Brockington invokes the concept to defend German prejudices about the epic.104 Brockington’s views on the Mahā bhā rata critical edition are scattered through diverse writings, but his clearest statement yet occurs in a lecture delivered at the Collège de France, Paris, on the occasion of the conference “Enjeux de la philologie indienne: Traditions, éditions, traductions/transferts” (December 2016).105 As the complete recording of this lecture is available online, we address some of his criticisms of the edition here, especially the suggestion that a Venn diagram may replace the stemma codicum.106 We irst quote from the relevant portion of his presentation (all italics are ours): [Fifty] years after completion of the Mahā bhā rata critical edition is an appropriate time to review its success, its limitations and its impact on epic scholarship along with that of the Rā mā yanạ critical edition, completed in 1975. I have myself studied the Rā mā yanạ more fully and shall rely on that for much of what I present, but I hope to make my remarks relevant to both epics, and more generally to the themes of this conference. Although they are classiied 430 430 PHILOLOGY AND CRITICISM by diferent names in the Indian tradition, for convenience I shall use the term epics to refer to both texts together. The basic, much debated issue is whether and how far oral performance gave way to manuscript transmission. External evidence is extremely meager and so internal evidence—for- mulaic diction, repetition, duplication of episodes and the like—has to be the prime source of information about this process of transition. I start with external evidence, which is only relevant for the Mahā bhā rata. Pā nị ni’s Asṭ ạ dhyāyi lists the addition of the suix ka to the names Vā sudeva or Arjuna, with the meaning of “one who has bhakti towards them” and mentions both Mahā bhā rata and some other names, while Patañjali mentions Yudhisṭ̣ hira and Arjuna as older and younger brothers. None of this, though, gives us much clue about what sort of text the grammarians knew, but in [unclear word] date it must have been oral. […] The internal evidence has mainly been seen in terms of formulaic diction under the inluence of the work of Parry and Lord on Homer and on the south Slavic epics. Indeed, the name oral formulaic theory usually given to this approach privileges formulae, perhaps unduly so. It is important to remember that while oral poetry is characteristically formulaic, this does not automatically mean that all formulaic poetry is oral. But what it means for us is that further markers of orality are needed before any text can rea- sonably be regarded as oral. Berkeley Peabody, working on Hesiod, produced a set of criteria at ive levels: phoneme, formula, enjambment, theme and song, and suggested that positive indications were needed for each before a text could irmly be considered a direct product of an oral tradition. While we must be aware of the limitations of extrapolating from one language and culture to another, without allowing for the inevitable diferences, these tests ofered me a convenient framework to assemble the evidence for the Sanskrit epics since they cover a broader range of evidence than do other schemes and conirmed with varying degrees of certainty the oral origins of both epics. Another feature which seems more characteristic of oral poetry than written literature is ring composition. Its earliest signiicant application to either epic was Renate Söhnen’s monograph on speeches and dialogues from the Rā mā yanạ . She argues persuasively for frequent deliberate structuring of speeches by anaphora, refrains, parallelism and the like but especially by means of ring composition. An oral origin for both epics has major implications for the duration of their growth period, which then has implications for issues about design and deliberate literary construction across their whole text as well as for our assessment of the sociocultural and religious contexts in which they evolved. The older view of this process of growth as essentially random has long been discarded by scholars who take a diachronic approach, through retained by those who take a narrowly syn- chronic approach as a stick with which to beat their opponents—discarded in favor of a more nuanced view of it as a design modiied and adapted over successive periods. […] The publi- cation of the critical editions has generated debate, focused mainly on the Mahā bhā rata, about the validity of the criteria used for the selection of readings and even the value of these editions. Some scholars, notably Madeleine Biardeau, have rejected the critical edition, regarding it as in essence a iction, and prefer to use the vulgate text, in efect the text established by Nīlakanṭ̣ ha Caturdhara late in the seventeenth century when he wrote his Bhāratabhavadı̄pa commentary. Often overlooked, though not by Biardeau herself, is that Nīlakanṭ̣ ha explicitly compared several manuscripts and was well aware of regional divergences. His comments to this efect come prominently at the start of his commentary. His text is avowedly eclectic while aiming at being comprehensive. Appeal to or preference for the vulgate as more authentic is therefore misguided in my view. The critical editions are not only on the whole the best we have, despite their limitations that I  discuss next, but also permit further advances in our understanding of them. Both critical editions are based on the readings of just a selection of the vast number of manuscripts known. A  larger number was initially examined, but then 431 APPENDICES 431 reduced to a manageable number. The principles followed were irst enunciated by Sukthankar in his Prolegomena to the Ā diparvan of the Mahā bhā rata and were broadly followed by all editors for both epics. These principles comprise grouping the manuscripts mainly by script into versions, relying on agreement between these versions for accepting any passage into the text in default of attestation in all manuscripts used, and accepting the two axioms of the textus simplicior and the lectio diicilior. Sukthankar recognized the impossibility of constructing a stemma codicum on the classical model and relied on grouping the script versions into primarily the northern and southern recensions. Since these are too far apart to permit the establishment of an archetype solely on the basis of manuscript readings, his—and the other editors’—consti- tuted text is necessarily eclectic though based on careful judgment on the basis of the available evidence and is an approximation to a written redaction of the text that became normative. The same considerations apply mutatis mutandis to the Rā mā yanạ . The irst limitation that must be recog- nized is the long period of transmission of the texts: around two millennia on Hiltebeitel’s view  that the Mahā bhā rata was composed between the mid-second century BC and the year zero by a committee of “out-of-sorts Brahmins”; more like two and a half millennia according to various other scholars. That is to be set against the recent date of most manuscripts. The oldest dated manuscript used for the Mahā bhā rata critical edition is dated 1261 CE while for the Rā mā yanạ critical edition it is one dated 1020. However, for both epics, the majority of the manuscripts used come from the sixteenth to nineteenth centuries. Most are essentially com- plete, whereas many older manuscripts lack portions of their text. The editors’ understandable tendency despite their good intentions to choose complete texts may well have resulted in the use of younger manuscripts. Moreover, several scripts are not represented: Oriya, Kannada, and Nandinā gari. My own investigations into an Oriya script Rā mā yanạ manuscript and of another one in Nevā rī script shows that such further evidence leads to a reassessment of the concept of script-based versions in favor of one based more on regional ainities. My investi- gation of another one in Malayā lam script in the Trivandrum collection necessitates a reassess- ment of the manuscript M4, previously dismissed as contaminated. These two manuscripts share features which indicate that they represent an alternative and probably older because shorter Malayā lam recension. The view that the various scripts generated relatively isolated manuscript traditions with little contamination between them goes back as far as Lüders. It was reairmed by Sukthankar in his Prolegomena. I quote: “The supericial diference of scripts corresponds, as a matter of fact, to deep underlying textual diferences.” But he goes on to state that in practice matters are never as clear-cut. Moreover, most literate Indians nowadays know more than one script and there is no reason to think that such was not always the case. Besides a scribe might have listened to an oral presentation of either epic or perhaps the reading aloud of a text in an area or by an individual who was using a script he did not know and he could have incorporated some episodes that remained in his memory into a subsequent manuscript copying. Certainly there is ample evidence of scribes deliberately comparing manuscripts or at least checking one from another, in particular through variant readings found in the margins or interlinearly, sometimes even in a diferent script from the manuscript itself. The supposed barriers of scripts are in fact highly permeable. We should, I submit, be thinking more in terms of regional groupings that may cut across script diferences. The number of marginal or interlinear additions also shows that the texts were commonly enlarged in successive copyings. Accidental loss is always a pos- sibility but much less likely than accretion. This is of course what underlies the principle of the textus simplicior which governs Sukthankar’s preference, followed by the other Mahā bhā rata editors, for the Sā́ radā script manuscripts and other closely related ones, despite their relatively recent date. The Malayā lam manuscripts also tend to contain a shorter text. Perhaps a feature 432 432 PHILOLOGY AND CRITICISM well known in linguistics that in language change the periphery tends to be more conservative than the center also applies here. On the other hand, attestation in all manuscripts is not infallible proof of a passage being part of the original text. It could have entered the tradition early on and then spread to all extant manuscripts, assisted by this habit of comparing manuscripts. The tendency to include everything found in the exemplars available may well account for the duplication of episodes that we ind in both texts. The scribe found varying versions in two manuscripts before him and, rather than choose one and reject the other, he incorporated both in an attempt at com- prehensiveness at the expense of consistency. So the text even of the critical editions cannot be regarded as a completely coherent whole. A  well-known example is the dice game in the Sabhā parvan of the Mahā bhā rata and Mehendale has collected multiple examples of interpolations and internal contradictions. Such lateral transfer of passages or episodes, verses between manuscripts is not the only complicating factor. Scribes did not or could not always copy a single exemplar throughout quite apart from comparing with others. […] Whereas the premise that manuscripts written in the same script belong together has signiicant limitations, then, the broader diferentiation into northern and southern recensions has a greater measure of validity. But even here we must recognize that a signiicant number of manuscripts combine features of both. This is probably more often by conlation of the two at a relatively recent stage in the chain of transmission but perhaps sometimes by retention of older features. Again I cite as my example my research on the Trivandrum Malayā lam script manuscript. Although it is close to M4 of the ̣ ạ of the Rā mā yanạ , neither could have copied from the other and their closeness is Ayodhyā kā nd marked by common absence of material rather than additions, suggesting a fairly early date. Neither unfortunately is dated but may be of sixteenth century date. However, they both combine features of the present northern and southern recensions. Within the northern recension, they align particularly with two manuscripts D1 and [D]2 which are usually assigned to the western subrecension within the northern recension and to a lesser extent with a third V1. The evi- dence of these two manuscripts, especially when linked with that from another fragmentary manuscript from Trivandrum, shows that they represent a deinite alternative tradition current within Kerala difering markedly from that of M1–3 and that it has clear links with the so- called western recension. A  related point is that the supposed uniformity of the southern recension of the Rā mā yanạ has been overstated. All in all, the evidence of a large number of Rā mā yanạ manuscripts shows that the simple opposition between the northern and southern recensions, however useful it may once have been as a heuristic device, does not adequately relect the complexities of the chain of transmission involved and consequently the question of their relative value is a more complex issue than was recognized in the constitution of the crit- ical edition text. The same is no doubt the case for the Mahā bhā rata too. […] So far I have mainly been demonstrating limitations in the way that the critical editions of both epics classiied the manuscript evidence into recensions and script-based versions, relying on what is only a modiication of the stemma concept. Can we go beyond this in our understanding of manuscript relationships? I will put forward two possible ways of doing so. Thirty years ago, I suggested that the model of the Venn diagram could be borrowed from mathematics and logic as it had already been borrowed into linguistics. This uses circles or ellipses overlapping as necessary to represent a relationship between separate groups or sets of items. In the case of the manuscripts I was studying this means that a circle or ellipse representing M4 and the Trivandrum manu- script will partly overlap with others for the other Malayā lam manuscripts with D1 and [D]2 and V1, in this way showing to some extent at least the multiple allegiances involved. The drawbacks, as you will realize, are that this is a decidedly visual model rapidly becoming too complex to describe verbally and that even visually the patterns soon become highly intricate 43 APPENDICES 433 and less easy to interpret. The other is not just a model but even more a new way of establishing relationships between manuscripts developed for the Mahā bhā rata by Wendy Philipps- Rodriguez. This uses phylogenetic algorithms based on the principles of cladistics taken from biology to plot the degree of closeness between manuscripts, producing what she has called unrooted trees. The length of the branch of each manuscript show[s] the degree of divergence from the rest individually and collectively. This is a more objective way of identifying the actual relationships involved and was, for example, the means of identifying the change of alignment of the Sabhā manuscript D6 that I noted earlier. It is reassuring to ind that the Mahā bhā rata critical edition is located near the center closer to the northern manuscripts, relecting both the editorial preference for them and the general connectedness [correctness?] of the readings chosen. Variant readings and new patterns of omission or addition are found in virtually every manuscript examined. Many of course are trivial, but others are signiicant for building up a truer picture of the complex relationship between the various recensions and versions. There is therefore a continuing need for further examination of the manuscript evidence. However, both the approaches just noted rely heavily on the information found in the apparatus of the critical editions and this will inevitably be true of any other initiative. We can only go beyond the critical editions of the Mahā bhā rata and the Rā mā yanạ because of the irm basis that they have provided for further research.107 Before we examine Brockington’s suggestion that the Venn diagram is an alternative to the stemma codicum for representing manuscript relationships, let us irst con- sider the sources of his other theories. Brockington does not provide citations, but his work is clearly indebted to the German Mahā bhā rata critics. The assertion that the editors grouped “the manuscripts mainly by script into versions” recalls Reinhold Grünendahl’s claim in his 1993 article “Zur Klassiizierung von Mahā bhā rata- Handschriften.”108 The statement “the premise that manuscripts written in the same script belong together” paraphrases Grünendahl’s “Schriftartprämisse.” Indeed, this is how the term has been translated in English-language summaries of Grünendahl.109 The suggestion that “Sukthankar recognized the impossibility of constructing a stemma codicum on the classical model and relied on grouping the script versions into primarily the northern and southern recensions” reproduces Grünendahl’s claim “on the basis of this theoretical fundament, which I  in the following would like to characterize by the term Schriftartprämisse[;] Sukthankar then groups the manuscripts included for his edition of the Ā diparvan into ‘recensions’ and ‘versions’ in the Prolegomena published in the 7th fascicule (of 1933).”110 In contrast, the latter half of Brockington’s claim that the Mahā bhā rata editors “classiied the manuscript evidence into recensions and script-based versions, relying on what is only a modiication of the stemma concept” owes more to Andreas Bigger’s notion that “for the editors of the critical edition, the Schriftartenprämisse [sic] replaces the stemma.”111 The notion that the “closeness” between two manuscripts can be “marked by common absence of material rather than additions” recalls Grünendahl’s argument that “the proximity of D2.5 to the K group manifests precisely not in terms of shared textual additions that are present but in terms of textual additions that are not present.”112 Except that Brockington translates “gemeinsam” with “common” rather than “shared” and simpliies Grünendahl’s “nicht vorhandenen Textzusätzen” to “absence of material” the sentences are equivalent. Like Grünendahl, 43 434 PHILOLOGY AND CRITICISM who argues that the similarities between K2 and K1 are “a case of not present textual additions, which [...] permits us to once again infer a relationship between these two manuscripts that reaches far back,”113 Brockington thinks that the circumstance that the Trivandrum manuscript and the Ā yodhyakā ṇḍa manuscript M4 exhibit “common absence of material rather than additions” suggests “a fairly early date.” The circumstance that two manuscripts lack the same passages, however, does not prove them closely related, since they need not owe this absence to an ancestor more recent than the archetype or, indeed, the irst source of the tradition. Neither do we need to assume the “inluence” of one manuscript on the other to explain this absence, as we saw. Brockington erroneously thinks that if the Trivandrum manuscript and M4 lack the same passages this means they derive from a more recent common ancestor than the one they share with the manuscripts containing the passages. But this is false. If the passages are insertions, the Trivandrum manuscript and M4 could be more closely related with the manuscripts that contain them than with each other. The circumstance that they both lack the passages could be coincidental. Only an evaluation of their readings and, more speciically, of their shared errors of transcription will permit us to conclude their ili- ation. Brockington does not address this issue. Like Grünendahl, he thinks that the cir- cumstance that the two manuscripts lack the same passage justiies him in assuming “a deinite alternative tradition current within Kerala,” even though, as yet, he only has two groups of manuscripts—the interpolated and the non-interpolated—and has not shown that the latter constitute a family in the genealogical sense. Whereas the idea that the Bhandarkar editors classiied manuscripts by script is owed to Grünendahl, the idea that the critical edition reconstructs “a written redaction of the text that became normative” is borrowed from Bigger. Bigger calls the Mahā bhā rata crit- ical edition “a redaction that became normative” or “had a normative efect” (normativ gewirkt hat).114 He also declares: “What makes it [the archetype reconstructed in the crit- ical edition] so important is the fact that this redaction came to dominate the whole written tradition of the Sanskrit MBh; it became normative for this part of the transmis- sion.”115 Brockington’s source could be Oskar von Hinüber, who argues that “What has arisen [in the critical edition] is not the Ur-text […] but a completely new normalized recension.”116 But since von Hinüber himself draws on Bigger, it is likelier Brockington is quoting Bigger. The oral nature of Brockington’s lecture makes it diicult to provide exact references. Most likely, his source is Bigger.117 At any rate, we ind no evidence he developed the idea independently.118 Regardless of its source, the idea that the critical edition reconstructs a “normative redaction” is false. Like Bigger, Brockington inds it convenient to suggest that the “constituted text […] is an approximation to a written redaction of the text that became normative” because this redeinition lets him rehabilitate the hypothesis of a heroic Ḳatriya epic corrupted by Brahmans.119 Like the German critics he draws on, Brockington identiies the task of textual criticism with recovering the oral, bardic epic(s) he posits as the Mahābhārata’s (and the Rāmāyaṇa’s) source.120 His work consis- tently seeks to legitimate the hypothesis of a heroic epic.121 He objects that “both critical editions are based on the readings of just a selection of the vast number of manuscripts known,” but eliminatio codicum descriptorum is well understood in textual criticism. There is 435 APPENDICES 435 no reason to include all manuscripts known if the majority can be eliminated as codices descripti. Manuscripts that the editor discarded as inferior would only clutter the apparatus. Brockington’s objection would carry weight if he could show that the editors overlooked a signiicant manuscript, occurring high up in the text’s transmission. But he explicitly negates this possibility.122 He claims that “attestation in all manuscripts is not infallible proof of a passage being part of the original text. It could have entered the tradition early on and then spread to all extant manuscripts, assisted by this habit of comparing manuscripts,” but if all manuscripts agree in their reading, there is no debate between them. Editorial judgment is not needed.123 To print a diferent reading in the face of their unanimous testimony is to enter the realm of a subjective ars critica scarcely worth the name. Brockington argues that “the text even of the critical editions [of the Rāmāyaṇa and the Mahābhārata] cannot be regarded as a completely coherent whole,” but this is to misunderstand what a critical edition can achieve. The circumstance that the constituted text contains “multiple examples of interpolations and internal contradictions” is not an argument against it.124 Neither should it mislead us into thinking that we can restore a more original state of the text by purging the text of its perceived contradictions, as Brockington, via a reference to Mehendale’s work, next suggests.125 Like his predecessor, Brockington would only arrive at a modernized version of the text relecting contempo- rary prejudices.126 Having considered Brockington’s sources, let us consider his suggestion that the “Venn diagram” ofers an alternative to the stemma codicum. Before we look at his suggestion, let us briely review the concept of a Venn diagram. A Venn diagram, named after its inventor, John Venn, is used to express the logical relationship between classes or sets of objects.127 The diagram uses closed-plane igures to denote sets, while the region they enclose stands for the sets’ members or elements. Sets can be given either an intensional deinition (for example, A is the set of the letters of the alphabet) or an extensional deini- tion (in which case, the convention is to list its members within braces; thus: A = {a, b, c, d…}). Once the sets under discussion are deined, the Venn diagram can be used to pro- vide a graphical representation of all possible logical relations between them such as their intersection, union, diference and complement. Brockington introduces the suggestion as follows: Thirty years ago, I suggested that the model of the Venn diagram could be borrowed from mathematics and logic as it had already been borrowed into linguistics. This uses circles or ellipses overlapping as necessary to represent a relationship between separate groups or sets of items. In the case of the manuscripts I  was studying this means that a circle or ellipse representing M4 and the Trivandrum manuscript will partly overlap with others for the other Malayā lam manuscripts with D1 and [D]2 and V1, in this way showing to some extent at least the multiple allegiances involved.128 Since these comments are not very illuminative, let us look at his earlier reference: Perhaps I  should at this point remark that the Critical Edition’s system of assigning manuscripts to recensions or to script versions is nearly as limiting as the construction of a stemma codicum (which does have the merit of corresponding on the whole to the way 436 436 PHILOLOGY AND CRITICISM that manuscripts are propagated). The drawback is that it still involves an either/or situa- tion, where a particular manuscript is assigned to a particular recension or version, or else is dismissed as “contaminated” or a “Mischcodex,” descriptions which seem almost to have become terms of opprobrium. Tentatively, I would like to propose that a better model might be the Venn diagram (which has already been borrowed from mathematics into linguistics to considerable advantage; see the sketch overleaf). Thus V1, which I was last considering, might be represented by a circle or oval mainly overlapping with one representing the Ñ and B mss but also partly overlapping with one representing the NW and W recensions. To a cer- tain extent this is a reversion to the stemma codicum in that it attempts to depict the diferent inluences or originals which have formed a speciic manuscript, while avoiding the impos- sible task of constructing the entire chain of transmission involved.129 As an example, he presents the diagram shown in Figure 84. There are several problems with Brockington’s proposal. Let us irst consider his reasons for thinking a stemma codicum is unachievable for the Mahā bhā rata and then evaluate his misconceptions about the Venn diagram. NE ~ D4 NW + W N B D1.2 V1 M4 + 14052 MI-3 S Sketch of a Venn diagram of some manuscript relationships Figure 84 Brockington’s “Venn diagram” Source: Reproduced from Brockington, “Textual Studies in Vā lmīki’s Rāmāyaṇa,” 204. The same diagram occurs in the handout accompanying Brockington, “Regions and Recensions, Scripts and Manuscripts.” Brockington’s label is misleading since the diagram does not depict manuscript relationships, but only common members (and only partially so, since he bases it on prominent features like interpolations). Moreover, the graphic is not a Venn diagram. Brockington neither deines the sets nor populates them with members (a problem he now acknowledges). Although not a Venn diagram, in the following we refer to it as one and only gradually introduce corrections to bring it closer to a real Venn diagram. 437 APPENDICES 437 1. Manuscript Filiation: Brockington erroneously links manuscript iliation with “alignment” and, from the observation that the manuscripts’ “alignment” changes, concludes that a stemma codicum is unachievable.130 Both assertions are false. Agreement in truth is not evidence of iliation and so a change in the manuscripts’ agreements should not prevent us from identifying their true iliation, which will reveal itself in a small albeit informative number of signiicant errors. Brockington is misled because he focuses only on selected passages, above all, obvious features such as additions and omissions, which just because of their itinerant nature will mislead us as to the manuscripts’ ilia- tion.131 Moreover, the promiscuity of more recent and inferior manuscripts should not prevent us from itting the older and better ones into a stemma.132 Brockington goes astray because, rather than study the entire tradition or individual manuscripts within the context of the larger tradition, he focuses on late and inferior specimens, and uses their recidivism to claim that, as their iliation cannot be satisfactorily determined, the very concept of a stemma codicum must be abandoned.133 He evidently confuses the stemma codicum with the “real tree,” that is, “the unknowable, but certainly much larger and more complex ensemble of all copies that ever existed.”134 From the circumstance that a genealogical tree in this sense is unattainable, he falsely concludes that we should use Venn diagrams, instead, to represent “the multiple allegiances involved.”135 Not only does this misunderstand the stemma codicum’s function (to depict, as unambiguously as possible, a manuscript’s descent from others, either surviving or hypothesized as having existed).136 It also overlooks that a Venn diagram cannot replace the stemma as it contains no genea- logically relevant information. From Brockington’s diagram, we could establish neither the ancestors of the manuscripts listed nor the part of their shared text owed to a common ancestor. Indeed, there is no place to locate an ancestor: unlike the stemma, which, though two-dimen- sional, uses the illusion of a third plane to convey height and thus temporal succession, a Venn diagram remains trapped within the plane of the paper. Brockington depicts the northwestern and western traditions by one circle, the northeastern by another and the southern by a third. But this is surely false, since the circles do not intersect and the three groups of manuscripts do not contain completely diferent texts.137 If the diagram is to have any sense, the intersecting regions must represent readings shared between manuscripts and hence their inherited text. But which regions? In Brockington’s dia- gram, some manuscripts are completely unrelated to the others, which must mean they are not Mahābhārata manuscripts at all. Even discounting these outliers, between two intersecting regions, which one indicates the manuscripts’ inherited text? Both cannot be original. If M4 ∩ V1 ∩ Ñ ∩B represents their common text and M4 ∩ D1.2 ∩ V1 represents their common text, one of the agreements must be secondary. More precisely, neither one can be completely original. Both intersecting regions will include verses or readings that are shared because they are inherited unchanged from the archetype and others that are shared either because they are owed to a more recent common ancestor than the arche- type or because they were transmitted horizontally by contamination. Brockington must disambiguate between the part of M4 ∩ V1 ∩ Ñ ∩B owed to the archetype, that is, is a subset of θ (since not all of θ’s reading will be transmitted unchanged to its descendants) and the part owed to later sources; and likewise for M4 ∩ D1.2 ∩ V1. But since this is the 438 438 PHILOLOGY AND CRITICISM issue at stake, the Venn diagram cannot help him. Manuscripts such as M1–3 must also share in the “archetypal” inheritance. Brockington has not mapped the common text of all manuscripts, but only selective ainities based on prominent characteristics.138 2. The Venn Diagram: Even if we grant Brockington that “variant readings and new patterns of omission or addition are found in virtually every manuscript examined” and they permit us to “build […] up a truer picture of the complex relationship between the various recensions and versions,” the problems with his “Venn diagram” remain. For starters, what is his set? If M4, V1, etc. denote sets whose members are the lines or words contained in the respective manuscripts, how does he account for repetition of lines or words given that every element of a set must be unique and that the order of elements is irrelevant? He notes that the Mahā bhā rata contains many repetitions (he calls them “for- mulaic expressions”).139 If we deine M4 as the set of lines contained in (the manuscript) M4, Brockington must irst eliminate all repetitions from the manuscript. To understand the implications, consider the following account from The Sanskrit Epics: Pāda-length phrases comprising a personal name and an epithet are the commonest formulaic expressions; however, they are not used just at random but fulil a deinite narrative function. The other main types are introductions and conclusions to speeches, various verbal formulae expressing emotion or emphasis, certain descriptive and hyperbolic phrases, stock expressions for battle scenes, phrases of time, place and number, proverbs and similar expressions, and stereotyped similes. Such formulae occur mainly in the second and fourth pādas, except those used after the end of speeches, which for obvious reasons usually occur in the irst (or less commonly the third) pāda; the metrical pattern of the śloka means, however, that there are reg- ularly diferent sets of formulae for the odd pādas from those for the even pādas. It is also note- worthy how often a phrase or passage is repeated within a short space of its irst occurrence. In several instances this results from the exact wording of a message being repeated or some event already described being narrated to another character. By contrast, the use of refrains, which occurs mainly in speeches, has a deinite emphatic purpose, as has parallelism within the verse, whereas another type of repetitiveness typiies certain didactic passages.140 Brockington could not preserve any of the elements that, in his view, demonstrate the Mahā bhā rata’s “oral” origins. A line that repeats in the irst adhyāya and the sixth may be listed only once, even if he considers the second occurrence signiicant for the manuscripts’ “alignment.” He could partially circumvent the problem by deining the verse as the smallest unit of analysis. He can now retain these lines despite their rep- etition (though identical verses must still be eliminated). But what he gains in material for comparison (that is, the number of elements in the set) he loses in accuracy, since he can no longer consider partial matches. For instance, if M4 and V1 contain a verse with an identical irst pāda, he cannot consider this pāda common to both sets, since their members are the complete verses only. Inevitably, this eliminates the majority of gene- alogically informative features such as shared signiicant errors. A smaller semantic unit than the complete verse thus appears preferable. But consider what happens when we deine the members of the sets as the words contained in the manuscripts. Except for their irst occurrence, we must now eliminate all frequently occurring words, including particles, proper nouns and common verbs. After elimination, we could identify elements 439 APPENDICES 439 common to M4 and V1. But not every shared word is genealogically signiicant. Many shared words will be commonplace names or terms. Even if M4 and V1 deviate from the other manuscripts in that they share not only common names but also words not found in the other manuscripts, this may not matter. The common elements could occur in diferent parts. Unlike a sequence, the order of elements in a set is irrelevant. Thus, the set M4 = {a, b, c, d, e} is identical with the set V1 = {d, e, c, a, b}, which can be rewritten as {a, b, c, d, e}. From a genealogical perspective, however, it is the occurrence of c in the same position in V1 as in M4 that is signiicant. Set notation cannot preserve this distinction and so we would consider V1 identical with Ñ = {a, b, c, d, e}, even though its text is closer to M4’s than Ñ’s, which contains the same elements in a diferent order. Brockington had a clariication for his diagram at the recent Dubrovnik International Conference on the Sanskrit Epics and Purā ṇas (DICSEP)  8.141 In a session on Venn diagrams as a tool of analysis, he explained that when he proposed the Venn diagram, it was “obvious” that the members of the sets were the manuscripts themselves. Thus, the set M4 contains the manuscript M4, V1 contains the manuscript V1 and so on. We must confess, this possibility did not strike us. If this is all Brockington means, some of our criticisms are moot. But this explanation does not cohere with his diagram. Let us irst address a basic problem. If the set M4 contains the manuscript M4, V1 contains the manuscript V1 and so on, no circle should overlap. M4 contains only M4 and hence no elements in common with V1 (which contains only V1). Their intersection is the empty set: M4 ∩ V1 = { } or ∅. This also holds for the intersection of the remaining sets. Our diagram would look like Figure 85. Actually, the circles should all overlap, because the Venn diagram must indicate all possible logical combinations of the sets, even if the containing intersections are empty (in which case, they can be shaded to indicate that they have no members). Not only is NE D1.2 NW + W V1 D4 M4 + 14052 S B Figure 85 A “Venn diagram” depicting sets whose members are individual manuscripts 40 440 PHILOLOGY AND CRITICISM NE D1 D2 NW + W V1 14052 D4 M1 M4 S M2 M3 B Figure 86 The corrected “Venn diagram” this extremely diicult in the case of 10 sets; we also wish to remain with Brockington’s understanding of “Venn diagrams” as we explore its problems, while gradually pro- ceeding toward a real Venn diagram. But even as drawn, Brockington’s explanation does not quite it the diagram. The set D1.2 contains two manuscripts (D1 and D2). The set M4+14052 likewise contains two manuscripts (M4 and 14052), whereas the set M1– 3 contains three manuscripts (M1, M2 and M3). These three sets difer quantitatively from the remainder. Although this may not make a diference, since the intersection of D1.2 with any other set is also the empty set (and likewise for M4+14052 and M1–3), it does raise the question: why are these seven manuscripts grouped into sets? Do they have a common property? If not, and we interpret Brockington’s statement that each set contains only the manuscript it is named for strictly, our diagram should appear as shown in Figure 86. This diagram is more accurate, but it does not solve the problem. If the sets’ members are the individual manuscripts themselves, what about NW+W, NE and S, given that no manuscripts correspond to them? Either NW+W, NE and S are empty sets, in which case we can eliminate them, or we must deine them diferently. If they mean anything, NW+W must denote the set of all sets that contain manuscripts belonging to the northwestern and western groups, NE must denote the set of all sets that contain manuscripts belonging to the northeastern group, whereas S must denote the set of all sets that contain manuscripts belonging to the southern group. Our dia- gram will look like Figure 87. NW+W remains an empty set. But we can now show that whereas NE contains six members (NE = {Ñ, D4, V1, B, D1, D2}), S contains ive members (S = {M4, 14052, M1, M2, M3}). We can, moreover, make logical inferences, for example, that NE and S are disjoint sets (they have no elements in common). But what do we gain by deining 41 APPENDICES 441 NE NW + W D1 D2 D4 B V1 14052 M4 S M1 M2 M3 Figure 87 A “Venn diagram” of sets that contain manuscripts of the S and NE groups NE as the set of all sets that contain manuscripts belonging to the northeastern group and likewise S as the set of all sets that contain manuscripts belonging to the southern group? It is senseless to place each manuscript in a set that contains no other manuscript and then group those sets into other, larger sets. A more intuitive deinition is that NE is the set of all manuscripts belonging to the northeastern group, whereas S is the set of all manuscripts belonging to the southern group. We can update our diagram as in Figure 88. This diagram corresponds more to a standard Venn diagram and its uses. But it no longer maps the resemblances or shared text between manuscripts. Rather, it represents the grouping of manuscripts after their analysis into groups, that is to say, after we have determined that Ñ, V1, B, D4, D1 and D2 belong to the NE group and so on. Most important, it does not “show […] the multiple allegiances involved,” since every manu- script either belongs to the set NE or does not belong to it; and likewise for S. Brockington may object that, actually, his point was that certain manuscripts have features of the northeastern and southern traditions, and hence should be considered members of both. Similarly, D1 and D2 in his original diagram represent manuscripts occurring halfway between the northwestern and northeastern traditions, rather than, as we have it, to the northeastern group alone. He may therefore propose returning to a modiied version of his original diagram (Figure 89). 42 NE NW + W D1 D2 D4 B V1 14052 M4 S M1 M2 M3 Figure 88 A “Venn diagram” of sets containing manuscripts of the S and NE groups NE NW + W D1.2 D4 B V1 M4 + 14052 S M1 M2 M3 Figure 89 A modiied version of Brockington’s original “Venn diagram” 43 APPENDICES 443 NE NW + W D4 D1.2 B V1 S M1 M4 + 14052 M2 M3 Figure 90 The corrected “Venn diagram” This is a valid objection, but the correct way to represent this situation is as shown in Figure 90. This is a “true” Venn diagram in contrast to Brockington’s. It shows that the sets NW+W and NE and S and NE each have two common members, whereas NW+W and S have no common members. Expressed in set notation, we have: NW+W ∩ NE = {D1, D2}; S ∩ NE = {M4, 14052}; and NW+W ∩ S = { } or ∅. But contrary to Brockington’s claim, this diagram does not let us analyze the relations between manuscripts. It only lets us represent their relationships after analysis, that is, after we establish the universal set (the set of manuscripts to be distributed) and determine which manuscripts belong in which of our inite sets NW+W, NE and S. Although we can now see at a glance which manuscripts are common to the sets, its advantages are not immediately apparent. We can neither express our reasons for assigning a manuscripts to two sets, for instance, that M4 has features of the S and NE groups, nor can we express the extent to which it does so, for instance, that M4 is a composite of 40 percent of S and 60 percent of NE. In set theory, an element is either a member of a set or not a member of a set. Venn diagrams neither obey cardinality (their size does not indicate the number of members in the set) nor can they be used to express ainities between members (for instance, by placing certain members nearer others), whereas Brockington’s original diagram tried to show precisely this, for instance, by displacing one manuscript toward another or by making some circles larger than others. But this is not a valid use of the Venn diagram. 4 444 PHILOLOGY AND CRITICISM A Venn diagram is a schematic, that is, an abstract diagram. The size, shape and posi- tion of circles or the location of elements within those circles, beyond the mere fact of intersecting or containing common elements, is irrelevant to their relationships. The real problem, however, is neither the Venn diagram nor that it cannot show what Brockington wishes. Rather, Brockington himself has not thought through what he wishes to show. Is it the manuscripts’ belonging to the regional groups? In that case, the groups are the sets and the manuscripts their members. Or is it the text shared between manuscripts? In that case, the manuscripts are the sets and their readings (words, lines or verses) the members. Without clarity on this fundamental issue, Brockington cannot make headway. If we now return to his original diagram, we see that all his problems stem from a failure to conceptualize what is the class and what is the member. On one hand, the diagram shows manuscripts contained either in one of the three regional groups or spanning the region between two groups. On the other, it shows manuscripts overlapping with each other such that their intersection subtends the extent of shared text. These two alternatives, however, are mutually exclusive. We can use the Venn dia- gram either to demonstrate the manuscripts’ membership of groups or to demonstrate the readings’ membership of manuscripts, but not both. For instance, if we wanted to show that 40 percent of M4’s text is shared with V1, we could populate their intersec- tion with 40 percent of their members, speciically the elements actually common to both. We could either list the verse numbers or use some other kind of notation to indi- cate their members. But where should we place this intersection? If we placed it inside D2 NE NW + W D1 D4 V1 B 14052 S M4 M1 M2 M3 Figure 91 Expressing relations between two orders of sets 45 APPENDICES 445 S, M4 would be entirely an S manuscript, whereas Brockington seeks to show that it is a composite of the S and NE traditions. Alternatively, if we placed the intersection of M4 and V1 inside NE, M4 would be partly a member of S and partly not a member of S, which cannot be. Every element is either a member of a set or not a member of that set. Brockington must choose between the alternatives. Figure 91 clariies why we cannot simultaneously express the relation between the sets NW+W, S and NE and their members, the sets M4, V1, etc., and the sets M4, V1, etc. and their members, the lines or words they contain. This diagram returns to treating NW+W, S and NE’s members as sets that contain the manuscripts M4, V1, etc. rather than as the manuscripts M4, V1, etc. themselves. This is essential: unless M4, V1, etc. are themselves sets capable of containing members, we cannot express the relationship between their members in turn (for instance, whether they are common to the sets M4, V1, etc. or found in just one manuscript, etc.). Even though the problem should already be evident, let us proceed. The region 14052 ∩ V1 indicates the elements common to 14052 and V1. The diagram displaces the circle denoting S so that 14052 remains inside it (recall that 14052 can either be a member of S or not a member of S, but it cannot be partly inside and partly outside S). The dia- gram similarly displaces the circle denoting NE so that V1 remains inside it. This already contradicts Brockington’s aim, since he wanted to show that 14052 was only partially an S manuscript (it also features elements characteristic of the NE manuscripts). More important, V1 is now partly contained inside S and partly outside it, whereas 14052 is partly contained inside NE and partly outside it, despite our best efort not to contravene the requirement that every element be either an element of a set or not an element of the set. Neither outcome is permissible. V1 can be either a member of NE or a member of S or a member of both. If it is a member of both, it must be so to the same extent, that is to say, it must satisfy the conditions of being NE and being S equally. This also holds for 14052. In this case, 14052 and V1 are members common to S and NE and we would locate them in the intersection of S and NE (S ∩ NE = {14052, V1}). Brockington may object that this is exactly what he wished to show. All we need do is shrink the circles denoting S and NE so that 14052 and V1 it inside their intersection. But consider the outcome (Figure 92). Initially, this diagram appears to meet all Brockington’s requirements. It expresses the intersection of 14052 and V1, while preserving their status as elements of S and NE. It also apparently indicates their halfway position between S and NE. In truth, however, the diagram does not show that 14052 is partly an S manuscript and partly an NE man- uscript. Rather, it shows that 14052 is both an S manuscript and an NE manuscript. In fact, it is an S manuscript to the same extent that it is an NE manuscript. Every line or every member of 14052 is now a member of S and likewise a member of NE. Contrary to Brockington’s aim, which was to show that 14052 owed part of its text to S and part of it to NE, that is to say, that it was a composite of the two traditions, this diagram shows that it owes all of its text to S and all of it to NE. There is simply no way that Brockington can show what he hopes using the Venn diagram. For the sake of argument, assume that Brockington discards the irst option. He decides it is less important to show whether the manuscripts are members of one of the 46 446 PHILOLOGY AND CRITICISM D2 NW + W NE D1 D4 14052 B S M4 M1 M2 M3 V1 Figure 92 The source of the error three groups or of two or three. He only seeks to show the common elements of cer- tain manuscripts. At this point, all the problems we identiied earlier with deining their members resurface. What are the sets’ members—the verses, lines, pādas or words of the manuscripts? What do we do with repetitions in a manuscript? Do we preserve as much of its text as possible (in which case the elements must be as large as possible to elimi- nate repetition, albeit at the price of iner resemblances between manuscripts), or do we preserve as much nuance as possible? How useful is it to indicate crude percentages of shared text, if they neither represent genealogical relationships nor indicate the extent of inheritance from the archetype or, indeed, a more recent ancestor? To circumvent these problems Brockington must irst deine the universal set: the set of passages (or elements) to be distributed. He may specify that his examination is restricted to the starred and appendix passages. Moreover, to avoid the problem of repetition (whether of words or longer units such as pādas) he regards these passages in their entirety. He seeks to show that, of x insertions, whereas M4 contains y and V1 contains z, w of their insertions are common. This is a valid use of the Venn diagram. But it is limited precisely to a consid- eration of genealogically irrelevant aspects. It is analogous to cataloguing the number of shared frocks and ribbons the children in various families have after they have been iden- tiied as members of their respective families and their genealogical descent established. Although useful for certain kinds of investigations (for instance, if we wanted to estab- lish that the children in family A had the most toys, but the children in family B also owned 30 percent of those toys), it does not address the problem of iliation. It can only 47 APPENDICES 447 U = Interpolations listed in the critical apparatus 16 20 U = {1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 15 17 D4 13, 14, 15, 16, 17, 18, 19, 20} 5 7 4 6 19 1 D1.2 V1 8 B 2 10 9 18 M4 + 14052 12 13 11 3 14 Figure 93 Mapping interpolations using a “Venn diagram” be undertaken once we have assigned the children to families, so that, efectively, what Brockington is doing is using information contained in the Mahā bhā rata’s appendices to “regroup” the manuscripts by their interpolations. The only value of doing so is that it provides a visual representation of the appendix passages diferent manuscripts share, but it cannot tell us anything about their iliation. Figure 93 clariies what Brockington most probably envisions when he proposes that we use the “Venn diagram […] to repre- sent a relationship between separate groups or sets of items […] in this way showing to some extent at least the multiple allegiances involved.” This diagram comes closest to reconciling what Brockington envisions. But it does not accurately represent the distribution of interpolations in the Rā mā yaṇa (remember that although Brockington features the diagram in his assessment of the Mahā bhā rata critical edition, the sigla actually stand for Rā mā yaṇa manuscripts). The diagram is limited to 11 manuscripts, which it groups into seven sets, so that it features only seven circles. It cannot handle the full complexity of the tradition. Even if we limited its use to representing interpolations, it would quickly break down. Granted that Brockington recognizes that “this is a decidedly visual model rapidly becoming too complex to describe verbally and […] even visually the patterns soon become highly intricate and less easy to inter- pret.” But he does not realize that it cannot represent even the limited set of “patterns” he wishes. This is where the inal failure of Brockington’s attempt at creating a “Venn diagram” manifests. A Venn diagram must represent all possible logical combinations between sets. Their number is given as 2n where n represents the number of sets in the diagram. Thus, for seven sets, the required number of regions is 128, whereas for Brockington’s original proposal, which featured 10 sets, the number is 1,024. Even our corrected diagram is far from meeting this condition. D1.2 and M1–3 do not inter- sect. D1.2 and Ñ likewise do not intersect. There is no region where only M4+14052 and Ñ meet. There is also no region where only V1 and B meet. Despite every attempt to save his hypothesis, this inal diagram is also not a Venn diagram! Not only is it not a Venn diagram; it also does not depict all possible combinations of interpolations in manuscripts, but only those Brockington selectively identiies. Insofar as it represents 48 448 PHILOLOGY AND CRITICISM anything, the diagram is symbolic of his approach to textual criticism: presenting partial information, selectively identifying ainities and using those ainities to draw erroneous conclusions based on a lawed method. Let us draw some conclusions: 1. The Venn diagram is useless as a tool of genealogical analysis. 2. It is also useless as a tool of analysis of the commonalities between manuscripts. 3. It can be used to represent manuscript groups, but not to analyze them. We can use it only once we have undertaken an analysis. 4. The Venn diagram can be used to represent the distribution of interpolations, but this presumes a prior stemmatic analysis to identify the interpolations as interpolations. 5. The Venn diagram cannot represent all interpolations, but only a limited number in a small group of manuscripts at a time. We can thus use it to study a selection of the evidence, but only if we clarify, in advance, our aims. Any conclusions we draw must be contextualized vis-à-vis the complete manuscript tradition. Finally, there is a possibility that what Brockington contemplates is not a Venn diagram at all, but simply some kind of “graphical representation.” In our dialogue, he admitted as much, saying that he merely wished to represent, in a rough way, the manuscripts’ relationships, remembering, of course, that by “relationships” he means not genealogical relationships but common features such as interpolations. The gen- eral consensus that emerged was that his diagram was not a Venn diagram and many confusions arose from his erroneous attribution of the term to it. Provided he no longer calls his igures Venn diagrams, understands that they contain no genealogi- cally relevant information and acknowledges that they are drawn on the basis of the editors’ existing analysis of the manuscripts’ relations of iliation and presume their labor in identifying and cataloguing interpolations, we see no objection to their con- tinued use. Brockington recommends the Venn diagram as permitting an initial, albeit crude, sorting of manuscripts. But he overlooks the following: 1. Not all shared variants have the same value. Only some variants are genealogically signiicant, whereas the great majority relect minor orthographic changes and the like. 2. Not all shared passages are evidence of a common inheritance. Some passages will have been transmitted horizontally by contamination and hence will mislead us as to the manuscripts’ true iliation. 3. The kind of resemblance—placement, repetition, extent, minor deviations and changes to neighboring text—contains more genealogically relevant information than the Venn diagram can capture. Set theory requires that we treat verses or variants as identical members of their containing sets, which implies an unacceptable loss in precision. In his desire to “mathematize” the manuscripts’ relationships, Brockington needlessly discards valuable information. His proposal is both excessive and insuicient: excessive because it entails a drastic simpliication of the manuscripts’ contents beyond even that he undertakes when he reduces their relationships to quantitative igures; insuicient because it does not adequately conceptualize the sets’ elements. He neither deines his sets nor explains how we should treat partial matches or repeated verses. His proposal 49 APPENDICES 449 is thus inadequate to the task. A Venn diagram based on grouping manuscripts by their most visible diferences appears to be an easier alternative. It replaces careful study of the manuscripts’ readings, above all, of their shared errors of transcription, with a handful of prominent characteristics such as shared interpolations. But the move from a simple quantitative relationship (for example, manuscript X shares 30  percent of the text of manuscript Y) to its diagrammatic representation presents insurmountable obstacles. Do placement and semantic order count? What about accidental identities? If M4 and V1 share a word or sequence of words not found in the other manuscripts, albeit in diferent chapters or books, is this evidence of ainity? Or, rather, is it a coincidence, even though M4 ∩ V1 will now contain more members than the manuscripts’ respective intersections with the other manuscripts? Brockington justiies his suggestion that editors use Venn diagrams “to represent a relationship between separate groups or sets of items” on the grounds that it evades “the impossible task of constructing the entire chain of transmission involved,” but he has considered neither the nature of those relationships (intersection, union or com- plement, or is their nature, rather, one of iliation?) nor the nature of the elements involved (are they additions, omissions, words or verses?). He emphasizes that the “Trivandrum manuscript” frequently omits the same passages as M4, even though the Venn diagram is singularly unsuited to capture such information. No precedent exists for deining a set in terms of members it does not contain. We could, at most, iden- tify the set’s complement, but this would not contain all of the set’s members rather than comprising a mixture of elements and omissions. Brockington also overlooks that if the “entire chain of transmission” cannot be reconstructed, it is especially unlikely that the relationship of the later and worse manuscripts can be stemmatized. Rather than focus on the part of the transmission that can be reconstructed, his solution is to map—imperfectly and with severe limitations—the extent of indiscriminate borrowing between the deteriores. He then reduces this borrowing to quantitative relationships and recommends using overlapping circles “showing […] the multiple allegiances involved.” The method’s supposed advantages, above all, the ability to overlap diferent circles, indicating a manuscript’s composite nature, are bought at the price of being unable to represent its (true) descent. Brockington’s approach is thus contrary to everything editors have attempted for the past two centuries, when they sought to determine manuscript iliation based on a system- atic recensio; counter also to the principle that an editor must irst establish the “vertical” constants of the tradition before he can determine the extent of “horizontal” transmis- sion. These errors are understandable insofar as Brockington has never edited a work and never produced full collations of a manuscript. All his examples are drawn from the * passages the editors identiied. Besides the questionable nature of drawing conclusions about the Mahā bhā rata based on the Rā mā yaṇa, Brockington’s work exhibits, once again, the tendency among Indologists to ignore readings for interpolations. His statement that “the critical editions are […] on the whole the best we have” is baling because if the stemma does not hold neither does the reconstruction based on it, and Brockington does not think we can draw up stemmata for the Sanskrit epics. A  crit- ical edition difers from an edition based on the editor’s conjectures or the concord of 450 450 PHILOLOGY AND CRITICISM a majority of witnesses precisely in that it relies on a stemma for most of the sifting between variants. Yet Brockington argues that, Essentially the point at issue is that the Critical Edition’s system of assigning manuscripts to recensions or to script versions is almost as limiting as constructing a stemma codicum (which does have the merit of corresponding on the whole to the way that manuscripts are propagated). The drawback is that this still involves an either/or situation, where a particular manuscript is assigned to a particular recension or version, or else is casually dismissed as “contaminated” or a “mischcodex.” The reality is that script boundaries are by no means as rigid as this scheme of classiication would suggest and that the only efective way of tracing alignments is by plotting the shared features of speciic manuscripts.142 Although intended as his deinitive statement on the Rā mā yaṇa (and mutatis mutandis the Mahā bhā rata),143 it is less conclusive than he imagines. Consider the following problems: 1. The critical edition does not “assign manuscripts to recensions or script versions.” It undertakes a genealogical analysis of their relations of iliation. 2. The stemma does not correspond “on the whole to the way that manuscripts are propagated.” It is the best means yet found for representing manuscript iliation. Brockington’s proposed alternative—the Venn diagram—is unequal to the task. We await clariication of his statement that “a new way of establishing relationships between manuscripts [has been] developed for the Mahā bhā rata by Wendy Philipps-Rodriguez,” but at the recent DICSEP 8 he did not elaborate.144 3. The “either/or situation, where a particular manuscript is assigned to a particular recension or version, or else is casually dismissed as ‘contaminated’ or a ‘mischcodex’ ” does not arise from “the Critical Edition’s system of assigning manuscripts to recensions or to script versions” (which it does not do). It is a relic of the fact that we can either establish a manuscript’s descent with a degree of certainty or it resists stemmatization. Ultimately, it is a consequence of whether vertical or horizontal transmission domi- nated in the manuscript’s formation, and thus an unavoidable feature of manuscript transmission. Brockington’s proposed solution cannot eliminate the problem. The Venn diagram appears to cope better with this ambiguity, but only because it is less sensitive to iliation. To argue that the Venn diagram is preferable as its blurry nature corresponds to the blurring of manuscript boundaries is like suggesting we should look at indistinct objects through a frosted lens because we will then not notice their indef- inite boundaries. 4. The objection that “script boundaries are by no means as rigid as this scheme of classiication […] suggest[s]” relects a justiied anxiety that horizontal transmission played a greater role in an exemplar’s formation than vertical transmission. But we must irst establish this. The stemma is the best means for this.145 If all Brockington means by the statement is that manuscripts rejected as late and inferior nonetheless contain good readings, this principle is well known in textual criticism (recentiores non deteriores). But does it repay to study the manuscripts for a few potentially superior readings? Brockington, at any rate, appears to think not, since he has neither collated a manuscript completely nor proposed a single conjecture that betters the editors’ readings. 451 APPENDICES 451 5. Finally, although it is true that the “only efective way of tracing alignments is by plotting the shared features of speciic manuscripts,” this is to misunderstand the aim of textual criticism. Agreements between manuscripts are the easiest way to sort them, but ofer only a preliminary criterion. They do not reveal their genealogical relationships. If two or more manuscripts agree against the rest, which of the two groups preserves the correct reading? A  table of agreements such as the one Brockington proposes cannot eliminate what Paolo Trovato calls “the ineluctability of critical judgment.” “Plotting the shared features of speciic manuscripts” will most likely provide the same information as a simple observation.146 6. The proposal is also not new. Before Brockington, John Griith proposed a sim- ilar approach, albeit in greater detail.147 Martin L. West provides a concise summary of Griith’s procedure: A more elaborate way of using such a table has recently been advocated by J. G. Griith. It involves comparing manuscripts in respect of the whole numerical series of each one’s agreement with the others; no distinction is made between primary and secondary readings, though agreements that may represent coincidental innovation are excluded. Suppose manu- script A shows the following numbers of agreements with the others: B F G H J K L N O P R T U V Z 50 44 61 49 52 48 62 62 48 42 44 53 61 47 43 For each manuscript a similar series of numbers is found. The one whose series matches that of A most closely is then grouped most closely with A. Finally the manuscripts are all arranged in a “spectrum”: those with the most dissimilar patterns of agreement appear at opposite ends, with a continuous gradation from one end to the other, while certain clusters or “taxa” mark themselves of along the line.148 Griith’s tabulation of agreements is more systematic, but in essential respects it parallels Brockington’s approach. His diagram on page  121 can be considered a precursor of Brockington’s “Venn diagram” as it likewise aims to provide a graphic representation of manuscript agreement. But whereas Brockington uses circles and ellipses to represent his manuscripts and the points within the enclosed plane igures to represent their elements (their intersections representing the common elements), Griith adopts a more elegant solution: an array that lists manuscripts along the x and y axes and uses three kinds of hatching to indicate either shared readings or the extent of sharing.149 What his approach gains in being able to represent diferent kinds of agreement, it loses in not being able to represent multiple coincidences, which must be read of from the array by considering pairs of manuscripts at a time. Vice versa, Brockington’s approach cannot capture verse position and order and remains completely abstract since he never populates his sets or their intersections with members. Diferences in representation aside, the two approaches share several features. Brockington shares Griith’s emphasis on the “open” nature of the tradition,150 the suggestion that the editors’ selection of manuscripts was random, that many exemplars were ignored or not known, that attention should focus rather on identifying alignments 452 452 PHILOLOGY AND CRITICISM or ainities between manuscripts,151 and the idea that, beneath the confused mass of contaminated exemplars resistant to genealogical classiication, statistical compila- tion of similarities can reveal distinct patterns or groups of manuscripts.152 Griith also echoes Brockington’s preference for arguing from agreement and his aversion to evaluating readings when he writes:  “It is essential to remember that the opera- tion is conducted in terms of similarities and diferences between the several mss. Thus there is no question of forming provisional judgments about the rightness or wrongness of a reading which is open to doubt.”153 Like Brockington, he is skeptical of the stemmatic method and considers it especially inadequate for open traditions.154 Like Brockington, Griith thinks a statistical approach has an advantage over the latter in that it reduces manuscript iliation to quantitative terms, and hence is “free” of its subjective and potentially circular nature.155 Griith also expresses a similar distaste for complete collations.156 More important, he concurs with Brockington that a stochastic sampling of data can replace “a true, continuous, and systematic recension” (Wolf).157 Like Brockington, Griith does not diferentiate typological from genealogical classiication. Again, like Brockington, he argues that “the taxonomic process” can “disentangle the underlying relations of the interpolated mss. and pen- etrate to the pattern of their ‘genetic histories’ latent beneath the over-burden of interpolation.”158 Both scholars propose graphical schemes that allegedly overcome the stemma’s weakness in representing contaminated traditions. Griiths argues, “in really complicated cases an admirable ingenuity is often displayed in adorning the stemma with an elegant arabesque of dotted lines to show the presumed channels of contamination.”159 But they overlook that the data they wish to capture are less rele- vant for reconstructing the archetype than the data their schemes fail to capture. Even discounting the problems with Brockington’s “Venn diagram,” West’s reservations about Griith’s statistical approach are a serious objection: The trouble with this kind of analysis is that it is not clear what useful conclusions can be drawn from it. Two manuscripts may be grouped together just because they show no partic- ular tendency to agree with any manuscript more than any other, in other words because they are equally promiscuous, even if they have no special similarity with each other textually. In some cases it is evident that the taxa relect real ainity-groups, in others it is not. In any case we are given no guidance as to the distribution of ancient readings.160 Griith’s and Brockington’s work illustrates the limits of statistical approaches to man- uscript classiication.161 A  mere tabulation of agreements (in Brockington’s case, also “omissions”) between manuscripts cannot replace a genealogical analysis. Above all, it does not let us reconstruct earlier stages of the text, despite Brockington’s belief that removing certain kinds of repetitions (those he considers “Brahmanic”) and collecting others (those he considers remnants of “oral epic poetry”) lets him reconstruct a text older than the archetype. Criticisms of the Mahā bhā rata critical edition relect wider trends in textual crit- icism, where the neo-Lachmannian edition has fallen out of favor. In an age when few scholars have either the patience or the resources to create an edition based on 453 APPENDICES 453 an exhaustive survey of manuscript sources and a consideration of their readings, ad hoc criticisms based on partial investigations are appealing. Yet, as this inal section revealed, these criticisms are often based on an insuicient familiarity with the material, or an insuicient knowledge of the principles involved, or both. The charge that the “critical editions are based on the readings of just a selection of the vast number of manuscripts known,” for instance, is relevant only if we can show that it incorrectly discarded some higher-value manuscripts. Invoking “a new way of establishing relationships between manuscripts” likewise requires that we irst demon- strate we understand the principles involved.162 Before we raise unfounded and mis- leading objections to the critical editions, let us irst learn to appreciate the editors’ achievements, remembering that they were men of great probity and dedication163 who worked untiringly for an inconsequential remuneration.164 We should also require scholars who criticize the available edition(s) to provide alternatives or, at least, explain which text we should use.165 Brockington criticizes “appeal to or preference for the vulgate” as “misguided,” but simultaneously argues that “the text even of the critical editions cannot be regarded as a completely coherent whole.” In fact, neither the receptus nor the constitutio textus should have binding value for us. No less than Grünendahl, who has yet to produce his promised critical edition of the Nepā lī recension, Brockington must tell us what editorial approach he favors, what manuscripts he proposes editing and how he plans to go about it. Until someone proposes a better text, the objections to the existing editions stemming from higher criticism are against argument and good sense.166 As Brockington’s work attests, spec- ulative, a priori histories based on unprovable claims of “oral” transmission have led only to grievous misunderstandings of the manuscript evidence. A responsible textual criticism therefore must begin by rejecting the German scholars’ tendentious theory of an original epic and its “Brahmanic” corruption.167 Only then can we ponder what textual criticism is still called to do in relation to the Mahā bhā rata and what we, as readers and critics, can contribute. In the hope that a younger generation of scholars will take on this awesome responsibility, we now end this book.168 The inal word, how- ever, is reserved for Nietzsche: When we have words for something, we’ve already gone beyond it. In all speaking there is a grain of contempt. Language, so it seems, was invented only for what is mediocre, common, communicable. In language, speakers vulgarize themselves right away.169 Notes 1 This edition should not be confused with Sukthankar’s “S,” “the ultimate source from which all versions of the Southern recension are, directly or indirectly, derived.” V.  S. Sukthankar, “Prolegomena,” in The Ad̄ iparvan for the First Time Critically Edited, vol. 1  (Poona:  Bhandarkar Oriental Research Institute, 1933), xxx. Despite the name, Sastri’s edition is not a true critical edition. It does not undertake a genealogical reconstruction of the ancestor of the southern manuscripts, instead providing a composite text based on the southern sources. See ibid., cvi: “the editor is avowedly aiming only at reproducing the text of one manuscript, categorically renouncing the obligation of the textual critic to restore the text, as far as possible, to its original form”; “a true lineal descendant of the Mahā bhā rata manuscripts of South India” (all italics Sukthankar’s). 45 454 PHILOLOGY AND CRITICISM 2 The dates of these editions should be cited with caution. In many cases, the editions are no longer in circulation and we were compelled to rely on secondary sources. 3 V. S. Sukthankar, “Introduction,” in The Ar̄ aṇyakaparvan for the First Time Critically Edited, vol. 1 (Poona: Bhandarkar Oriental Research Institute, 1942), xxxiii. 4 But see the brief remark in his introduction to the Ā raṇyakaparvan that “The Bombay edition, like the Calcutta, is based on the Nīlakaṇṭha version but (like yet other editions of the Mahā bhā rata prepared in the same fashion, e.g., the Chitrashala Edition) does not represent the Nīlakaṇṭha tradition very faithfully.” Ibid., xxxiii (italics added). 5 A sixth attempt, the 18-volume “transcreated” edition (author’s term) of Purushottam Lal (The Mahābhārata of Vyāsa:  Transcreated from Sanskrit, 18  vols. [Calcutta:  Writers Workshop, 1968– 2008]), is not considered here as it does not claim to be a translation. The edition is not widely available (most recently only directly from the website of the Writer’s Workshop) and we could not examine the volumes. 6 Sukthankar, “Prolegomena,” lxxxvi. 7 Ibid., lxxvii. 8 The role expansion played in the text’s transmission is often exaggerated, as we have seen with Grünendahl. With the exception of the additional passages (which we can often easily identify), the Mahā bhā rata tradition is remarkably conservative. Doubtless, scribal errors were committed and their scope was magniied by frequent copying of the text (it has been estimated that about 400–500 years is the normal limit for a palm-leaf manuscript in the Indian climate), but a competent editor will often be able to identify the true reading (the one exception was the Virā ṭaparvan, a book that appears to have been shifting until a very late date). 9 Pasquali originally formulated this geographic criterion that innovations originate from the center and radiate outward to the periphery as follows:  “Just as in linguistics it is universally agreed today that earlier stages are preserved for a longer time in peripheral areas, and that hence the occurrence of the same phoneme, form, term or construction in two peripheral areas distant from one another guarantees their antiquity, so the agreement of codices written in areas far removed from the cultural center and from one another constitute an argument for the gen- uineness of a reading. Often texts that were much read, both in antiquity and in the Middle Ages, form a vulgate text which spreads, as fashions are wont to do, from the center towards the periphery, but do not always reach the periphery.” Giorgio Pasquali, Storia della tradizione e critica del testo (Florence: Le Monnier, 1952), xvi–xvii, cited and translated in Paolo Trovato, Everything You Always Wanted to Know about Lachmann’s Method: A Non-Standard Handbook of Genealogical Textual Criticism in the Age of Post-Structuralism, Cladistics, and Copy-Text (Padua: Libreriauniversitaria, 2014), 72–73. Sukthankar does not formalize the principle, but it is clearly present in his mind: “This version [the Malayā lam] has several striking agreements with Ś 1, a fact all the more impressive, because M, a Southern version, hails from the province at the opposite end of India from the province of Ś 1, a Northern version; for instance, Malayālam supports Ś 1 (against TG) in omitting the spurious parts of adhy. 128–129 of the Bombay edition.” Sukthankar, “Prolegomena,” lxxiv  (Sukthankar’s italics). 10 Glenn W. Most, “Editor’s Introduction,” in Sebastiano Timpanaro, The Genesis of Lachmann’s Method, ed. and trans. Glenn W. Most (Chicago: University of Chicago Press, 2005), 7–8. 11 Sastri refers to an edition of the Mahā bhā rata printed in Telugu characters and published, according to him, four times in Madras (1850, 1855, 1860 and 1929) as the “Telugu edition.” See P. P. S. Sastri, ed., The Mahābhārata (Southern Recension) Critically Edited by P. P. S. Sastri, vol. 1, Ā di Parva, part  1 (Madras:  Ramaswamy Sastrulu & Sons, Madras, 1931), v.  Sukthankar also referred to this edition in his “Prolegomena,” cv, but it was not examined for the critical edition. No further details are forthcoming of this edition. 12 This is the edition A.  Rangaswami Dikshitar printed in 1895 at the Veda Vyasa Press in Sarabhojirajapuram. Sukthankar refers to it as the Grantha edition of Sarfojirajapuram 1896. Sukthankar, “Prolegomena,” cv. No further details are forthcoming of this edition. 45 APPENDICES 455 13 क is Burnell’s catalog no.  11838, Grantha, Tanjore Palace Library. ख is Burnell’s catalog no. 11851, Grantha, Tanjore Palace Library. ग is Burnell’s catalog no. 11809, Telugu, Tanjore Palace Library. 14 Burnell’s catalog no. 11860, Grantha, Tanjore Palace Library. 15 Sukthankar’s source is T. R. Krishnacharya, ed., A Descriptive Contents of Sriman Mahabharatam (Kumbakonam: Madhva Vilas Book Depot, 1912), 4. 16 For the source, see John L.  Brockington, The Sanskrit Epics (Leiden:  E. J.  Brill, 1998), 60. Brockington’s igures were checked, where possible, against the igures the editors supplied in their respective volumes. 17 Debroy arrives at a slightly diferent breakdown. For the main text, he counts 73,784 verses in 1,995 chapters (6,073 verses in 118 chapters for the Harivaṃśa); thereafter his count of the verses in each adhyāya difers slightly from Brockington’s (though his count for the adhyāyas in each parvan is the same). In contrast to Brockington, Debroy also provides verse counts for the sub-units of each parvan (that is, the 100 upaparvans). See Bibek Debroy, The Mahābhārata, vol. 1 (New Delhi: Penguin Books India, 2010), xxiii–xxvi (the table repeats in all volumes). 18 The 18-parvan division is a northern feature; the southern manuscripts are inconsistent in the number of parvans they read. Curiously, their Parvasaṃgraha lists feature the northern division, leading the editors of the southern editions to override the southern manuscripts when creating their editions. See Sukthankar, “Prolegomena,” xxxiii. Thus, when reconstructing the southern division, we had to have recourse to the critical apparatus, as we had no access to southern manuscripts. 19 Both sections are called parvasaṃgrahaḥ (cf. Mahā bhā rata 1.2.33c and 71c), but the latter is more in the form of a narrative summary, and hence, the divisions are referred to less consistently. The list that follows takes the names from the irst section (1.2.34a–69c). The translations of the titles are Van Buitenen’s. 20 Sukthankar, “Prolegomena,” xxxiii. 21 Surprisingly, the Parvasaṃgraha list itself is extremely consistent in both the northern and southern recensions. Except for a handful of words, no part of it features a wavy line, suggesting that no efort was made to make the list conform to the southern contents. 22 As a matter of fact, the adhyāya counts cannot apply to the southern recension at all, since the southern manuscripts do not mention the names of the sub-parvans in their colophons. The list of upaparvans in these manuscripts is compiled on the basis of the Parvasaṃgraha list, but as this only gives the names of the 100 sub-parvans without mentioning the adhyāyas in each, we have no way of knowing how many adhyāyas a particular sub-parvan contains. See Sukthankar, “Prolegomena,” xxxiii, n.  4. The igures are thus only an approximate guide to where the southern manuscripts would have broken the parvan, had they done so. 23 Several sources report this as 24 parvans uniformly in the southern manuscripts, but this is incorrect. As a matter of fact, there is no division characteristic of the southern recen- sion as a whole, since the diferent manuscript groups do not agree (and sometimes there is disagreement even within the manuscripts of a group). Burnell, in his survey of the Mahā bhā rata manuscripts in the palace library at Tanjore, lists the following division as characteristic of the “southern recension”:  Ā diparvan (8 chapters), Ā stīkaparvan (40 chapters), Saṃbhavaparvan (200 chapters), Ś alyaparvan (56 chapters), Gadā parvan (11 chapters), Sauptikaparvan (9 chapters), Aiṣīkaparvan (11 chapters), Viśokaparvan (8 chapters), Rā jadharmaparvan (172 chapters), Mokṣadharmaparvan (191 chapters). A. C. Burnell, A Classiied Index to the Sanskrit MSS in the Palace at Tanjore (London: Trübner, 1880), 180. But, as Burnell did not possess a Malayā lam exemplar (see ibid., 181–83, for his list of manuscripts), this division cannot be extended to the southern recension as a whole. Likewise, Winternitz only had access to Grantha manuscripts. Even though he claimed to have seen and examined some “highly valuable […] Malayâlam MSS” in the collection of the Royal Asiatic Society, London, these are not included in his description. See Moriz 456 456 PHILOLOGY AND CRITICISM Winternitz, “On the South-Indian Recension of the Mahā bhā rata,” The Indian Antiquary 27 (1898): 68. The one Grantha manuscript he examined (no. 65 in the Whish collection) had two parvans—the Ā diparvan and the Saṃbhavaparvan—rather than the three Burnell reported. Ibid., 122. Winternitz appears not to have examined any manuscripts for the books after the Ā diparvan (no.  65 itself is incomplete and includes only the Ā diparvan) and his comments therefore are applicable only to the irst of the additional parvans found in the southern recension. Lüders, the next to consider the problem, wrote that Burnell’s distinction of the Sauptikaparvan into three parvans—the Sauptika, Aiṣīka and Viṣoka parvans—was not quite correct, since the Viṣokaparvan, in the printed editions of the northern recension, was not considered part of the Sauptikaparvan, but rather, of the Strīparvan. “Burnell ought, rather, to have said that, in south India, the Sauptikaparvan is broken up into two parvans—the Sauptikaparvan and the Aiṣīkaparva—and the Strīparvan likewise into two parvans—the Viṣokaparvan and the Strīparvan.” Heinrich Lüders, Über die Grantharecension des Mahābhārata, Abhandlungen der Königlichen Gesellschaft der Wissenschaften zu Göttingen, Phil.-Hist. Klasse, N. F. vol. 14, no. 6. (Berlin: Weidmannsche Buchhandlung, 1901), 66. Lüders only had access to a fragmentary manuscript of the Ś ā ntiparvan (Mackenzie 64, India Oice Library, London), but he nonetheless found that the manuscript included the Ā paddharmaparvan along with the Rā jadharmaparvan (both under the parvan title of Rā jadharmaparvan). Ibid., 68–69. 24 Even the critical edition is unsatisfactory in this regard. Sukthankar is scrupulous about noting the deviations from the northern division in the Ā diparvan (see his “Prolegomena” xxxiii and xxxiii, n.  2), but thereafter no editor, to our knowledge, appears to have paid much attention to this problem. Dandekar devotes much more attention to the fact that the Devanā garī manuscripts break the Ś alyaparvan after adhyāya 28, reading thereafter a parvan called the Gadā parvan (adhyāyas 29–64), and does not discuss the much more important fact that the southern manuscripts break the Ś alyaparvan into two independent parvans. See R.  N. Dandekar, “Introduction,” in The Śalyaparvan for the First Time Critically Edited (Poona: Bhandarkar Oriental Research Institute, 1961), ix. Velankar spends much more time discussing the Parvasaṃg raha igures and only mentions the fact that M1 lists the Sauptika sub-parvan as part of the Ś alyaparvan at the margins of his discussion. See H. D. Velankar, “Introduction,” in The Sauptikaparvan for the First Time Critically Edited (Poona:  Bhandarkar Oriental Research Institute, 1948), xxxii. Belvalkar, in contrast, does not seem to have attached any signiicance to the southern division at all: both in his introduction and in his critical apparatus, he does not mention the fact that the southern manuscripts consider the Rā jadharma and Ā paddharma parvans one parvan (the Rā jadharmaparvan rather than the Ś ā ntiparvan) and treat the Mokṣdharmaparvan as its own, independent parvan. 25 In creating this table, we were guided by Burnell’s comments, but it is entirely possible that we have overlooked other divisions. Sukthankar refers to the 24-parvan division as a feature of the southern recension as a whole, but clearly on Burnell’s and Winternitz’s authority. See Sukthankar, “Prolegomena,” xxxiii, n.  1. As a matter of fact, he could have had irsthand knowledge only of the divisions in the Ā diparvan, as the search for manuscript of the other parvans had probably only gotten under way in his time. Only a look at the colophons of all the chapters of the Mahā bhā rata will really settle the question. 26 As a matter of historical interest, Burnell’s views on the division of the Mahā bhā rata into 24 parvans in the southern tradition were irst published in the volume On the Aindra School of Sanskrit Grammarians (Mangalore:  Basel Mission Book & Tract Depository and London: Trübner & Company, 1875), 75–80. This section contains what must be the irst stemma of the Mahā bhā rata manuscripts ever drawn up. Burnell divides the Mahā bhā rata into two recensions—a northern (Nā garī) recension, which he claims was “a revision of the longer books, settled inally by Nīlakaṇṭha’s Commentary,” and a southern (Grantha) recen- sion, which he calls “an independent revision of the longer books.” Ibid., 80. Had Burnell 457 APPENDICES 457 had access to manuscripts from the Malayā lam area, he would have seen that the Grantha was not the whole of the southern tradition (and certainly not the best representative of it). He would also have hesitated to make the Devanā garī manuscripts the main bearers of the northern tradition, seeing as they are late and corrupt exemplars and as the northern tradition was established long before Nīlakaṇṭha’s commentary. It was left to Sukthankar, following the extensive search for manuscripts in all scripts from all parts of India, to draw up the irst true stemma of the Mahā bhā rata tradition. Nonetheless, Burnell’s stemma is of interest as the irst attempt to diagram the relation of the main Mahā bhā rata traditions to each other and because of the way it integrates the two traditions into the stemma: Burnell imagines that the tradition would have begun with an “original collection of ballads,” giving rise later to an “eclectic text with episodes added to it” before it inally broke up into the northern and southern traditions. Ibid. The historical inluence of his stemma on later editors (Sukthankar not excepted) is evident. Burnell’s stemma is reproduced in Appendix 15. 27 Sukthankar, “Prolegomena,” lxxiv (italicized in original). 28 Though not, of course, in the printed editions of the southern manuscripts, which cannot be true to their claim to print the southern recension and also follow a division into 18 parvans. 29 On the translation and deinition of upākhyāna, see Robert P. Goldman, “On the Upatva of Upā khyā nas:  Is the Uttarakā ṇḍa of the Rā mā yaṇa an Upā khyā nas of the Mahā bhā rata?” in Argument and Design: The Unity of the Mahābhārata, ed. Vishwa Adluri and Joydeep Bagchee (Leiden and Boston, MA: Brill, 2016), 69–82. 30 Alf Hiltebeitel, “Not without Subtales: Telling Laws and Truths in the Sanskrit Epics,” Journal of Indian Philosophy 33 (2005): 467–69. 31 Vishwa Adluri and Joydeep Bagchee, eds., Argument and Design:  The Unity of the Mahābhārata (Leiden and Boston, MA: Brill, 2016). Hiltebeitel’s 2005 article is reprinted as the irst of the contributions. 32 Sukthankar remarks: D13 added at 1.205; K5 discontinued from 2.10; G1 has lacuna from 2.192 to 3.11; K6 Dr D8–14 G7 discontinued, and Ñ1.2 B5 M5 added, from 3.1; Ñ3 added at 14.1; Ś 1 added at 26.10; B2 ends at 43.13; K1 has lacuna from 47.20 to 54.4; D6.7 M1.2.4 discontinued, and B6 M6–8 added, from 54.1; Ś 1 has lacuna from 55.3b to 60.61b and from 61.84b to 68.19; D3 (which transposes the Ś akuntalā and Yayā ti episodes) has lacuna from 62.3 to 69.51; V1 has lacuna from 68.74b to 92.13; K3 has lacuna from 69.41d to 71.17c, and from 72.8c to 74.4; Ś 1 has lacuna from 72.23 to 78.20b; D3 ends at 76.35; B4 ends at 90.88; V1 has lacuna from 96.37c to 127.21a; T2 ends as 181.40; T3 begins from 182.1. Sukthankar’s remarks concerning Ś 1 are partial and should therefore be supplemented with the next appendix. 33 Sukthankar, “Prolegomena,” x. 34 Brodbeck and Black present a concordance of the critical edition and the Ganguli/Roy edition in Simon Brodbeck and Brian Black, eds., Gender and Narrative in the Mahābhārata (Abingdon and New York: Routledge, 2007), 279–84. 35 Sukthankar also warns in his introduction to the Ā raṇyakaparvan that “there are some grave mistakes in the numbering of the stanzas of the Calcutta edition. The irst mis- take occurs after stanza 3095, where the next number suddenly jumps up to 4000 (when it should be 3100). Similarly after 8895, the following number jumps up to 9900 (when it should be 8900). For the third time, after 10,095, the following number is given as 11,000 (instead of 10,100). For the fourth time, after 11,095, the next number jumps up to 12,000, when it should be 11,100. At the next step, however, the number equally suddenly jumps back to 10,200! Thus at one place three successive numbers (which are supposed to increase by ive at each step) are 11,095, 12,000, 10,200! After adhy. 117, there appear to be no new mistakes, and from that point onwards the numbering may be taken to be continuous 458 458 PHILOLOGY AND CRITICISM and correct. But the result of these earlier miscalculations is that the extent of the parvan appears to be 17,478, when it should in reality be (according to Jacobi’s computation) 12,848, the Parvasaṃg raha igure being 11,664! Several numbers occur twice, and in giving references to the Calcutta edition, it is advisable to cite the adhyā ya number as well.” Sukthankar, “Introduction,” xxxiii. This is probably why Sukthankar does not cite igures for the Calcutta edition in his concordance. To our knowledge, no other editor has drawn attention to this problem. 36 For a discussion of Nīlakaṇṭha’s dates and background, see Christopher Minkowski, “Nīlakaṇṭha Caturdhara’s Mantrakāśı ̄khaṇḍa,” Journal of the American Oriental Society 122, no. 2 (2002): 329– 44 and Christopher Minkowski, “Nīlakaṇṭha’s Mahā bhā rata,” Seminar 608, The Enduring Epic (2010):  32–38. Most of what we know about the Mahā bhā rata commentators comes from a handful of studies:  Vishnu S.  Sukthankar, “Epic Studies V:  Notes on Mahā bhā rata Commentators,” Annals of the Bhandarkar Oriental Research Institute 17, no. 2 (1935–36): 185–202; Vishnu S.  Sukthankar, “Arjunamiśra,” in Dr.  Modi Memorial Volume:  Papers on Indo-Iranian and Other Subjects, ed. Dr. Modi Memorial Volume Editorial Board (Bombay: K. J. Dubash, 1930), 565–68 (reprinted in Sukthankar Memorial Volume, vol. 1, ed. P.  K. Gode [Bombay:  Karnatak Publishing House, 1945], 403–5) and Sukthankar, “Prolegomena” (the latter absorbing much of the material of the irst two, but also adding much new material). 37 Sukthankar, “Notes on Mahā bhā rata Commentators,” 185. Additionally, P. K. Gode in a series of articles in the Annals of the Bhandarkar Oriental Research Institute took up the question of the dates of speciic commentators (these are all cited in the bibliography and in this appendix where relevant). 38 They are, in no particular order:  Anantabhaṭṭa, Arjunamiśra, Ā nanda, Caturbhuja(miśra), Jagadīśa-cakravartin, Devabodha, Nīlakaṇṭha, Mahā nandapū rṇa, Yajña-Nā rā yaṇa, Ratnagarbha, Rā makiṃkara, Rā makṛṣṇa, Rā mā nuja, Lakṣmaṇa, Varada, Vā dirā ja, Vidyā sā gara, Vimalabodha, Ś aṅkarā cā rya, Ś rīnivā sa, Sarvajña-Nā rā yaṇa and Ṣ ṛṣṭidhara. Leclere inexplicably gives the number as “no less than twenty-one” even though Sukthankar’s numbering is unambiguous. Basile Leclere, “New Light on Devabodha, the Earliest Extant Commentator on Mahābhārata,” Asiatische Studien/Études Asiatique 70, no. 2 (2016): 490. 39 See Sukthankar, “Prolegomena,” lxx (Sukthankar’s italics):  “A commentary older and more important than the Arthadīpikā of Arjunamiśra, and one more neglected still, is the Jñā nadīpikā of Devabodha, cited here as Cd. Devabodha is certainly older than Vimalabodha, Arjunamiśra and Nīlakaṇṭha, all of whom cite him with great respect, and probably earlier than Sarvajña-Nā rā yaṇa and Vā dirā ja. He is, therefore, most likely, the earliest commentator of the Mahā bhā rata hitherto known, and, in my opinion, also the best.” 40 If there were commentators earlier than Devabodha, he does not mention them (at least in the parts of his commentary that have been preserved). In contrast, almost all of the commentators thereafter cite Devabodha; Arjunamiśra even mentions him with special reverence as the irst in a series that begins with Vyā sa and Vaiśaṃpā yana:  vedavyāsavaiśaṃpāyanadevabodhavim alabodhasarvajñanārāyaṇaśāṇḍilyamādhavapitṛbhyo namaḥ | śrı̄devabodhapādādimatamālokya yatnataḥ | kriyate’rjunamiśreṇa bhāratārthapradı ̄pikā ||. Cited in Sukthankar, “Notes on Mahā bhā rata Commentators,” 189. The fact that Arjunamiśra gives Devabodha irst place after Vyā sa and Vaiśaṃpā yana is indicative. 41 See Minkowski, “Nīlakaṇṭha’s Mahā bhā rata.” Minkowski does not clarify how he arrives at this date. Leclere’s relections are in “New Light on Devabodha,” 501–5. Leclere validates many of Sukthankar’s views about Devabodha except Sukthankar’s ascription to Kaśmīr on the basis of his text. According to him, it is more likely that Devabodha was “an intellectual from southern India who completed his studies and probably prepared his commentaries in the great centre of traditional learning that Banaras already was by that time. Having spent several decades of his life travelling in the North, he moved back to the South and composed a devotional play on a Vaiṣṇava hero at the demand of his patron the king Tribhuvanamalla, 459 APPENDICES 459 who probably belonged to the Hoysaḷa dynasty” (that is, in modern-day Karnataka). Ibid., 520. 42 The dates are Gode’s. See P.  K. Gode, “Notes on Indian Chronology XXXIV:  Date of Vimalabodha’s Commentary on the Mahā bhā rata Called the Viṣamaślokī—after 1150 CE,” Annals of the Bhandarkar Oriental Research Institute 17, no.  4 (1935–36):  395–99. Gode provides the upper date in P. K. Gode, “New Light on the Chronology of the Commentators of the Mahā bhā rata,” Annals of the Bhandarkar Oriental Research Institute 25, no. 1/3 (1944): 107. 43 Minkowski traces the origins of the Mahā bhā rata commentarial tradition back to the “vyasaghatṭ ạ literature,” that is, “compendia of thorny parts of the text.” Christopher Minkowski, “What Makes a Work ‘Traditional’? On the Success of Nīlakaṇṭha’s Mahābhārata Commentary,” in Boundaries, Dynamics and Constructions of Traditions in South Asia, ed. Federico Squarcini (Florence: Firenze University Press, 2005), 236. He also suggests that “as a phenom- enon, epic commentaries in Sanskrit appeared relatively late in the life of the epics, with extant commentaries datable not much earlier than the twelfth century.” Ibid. 44 In Sukthankar’s words, the Jñānadıp̄ ikā is “a concise ṭīkā ; that is, a running commentary, explaining, as a rule, only the diicult words and passages in the text. Occasionally it ofers explanations of constructional obscurities and grammatical diiculties, and gives the gist of passages; in the latter case, usually, under citation of entire verses (i.e. half ślokas) from the text.” Sukthankar, “Prolegomena,” lxx. But see Leclere, “New Light on Devabodha,” 520– 21: “Admittedly the philological approach does prevail in the Jñānadıp̄ ikā, which often explains a diicult word by its synonyms or its etymologies, but Devabodha at the same time could nonetheless have pursued a higher goal. For a devotee of Viṣṇu and an adept of Sā ṃkhya- Yoga, understanding the Great Epic correctly might have signiied approaching the truth of God, and it is not irrelevant that the commentary enlarges and deepens each time it tackles philosophical passages like the Sanatsujātıȳ a. As a matter of fact, the soteriological aspect of the enterprise is underlined by Devabodha himself at the end of such explanations.” 45 Ibid. 46 Leclere argues that the name Devabodhi arose due to the misinterpretation of a verse quoted in Ś rīdharadā sa. See ibid., 492–94. 47 Ibid., 494–95. 48 Ibid., 496–97. 49 Ibid., 498. In Leclere’s translation, the two inaugural benedictory stanzas (the Satyavratarukmāṅgada’s and the Jñānadı̄pikā’s) read as follows: As long as the moon and the sun exist, they assume the nature of seeds for provoking the reappearance of the Fortune of the gods in the ield of the triple world divided by the water of the River of the Immortals which springs from the foot [of Trivikrama]; they fall from every side, out of desire for reaching the earth, the ether and the heaven which are trembling: may the lines of dust coming from the lotus-foot of [Visṇ ̣u], the enemy of the Dā nava, give you happiness! It causes the hot and cool-rayed [celestial bodies] to turn away rapidly and conceals the sky- roads; refulgent, it pervades the earth, the ether and the heaven and makes them shine with its luminous nails it tears away the wall-like shores of the eastern and western regions from the world of mortals up to the [celestial] abode; may it protect you, the foot of Nā rā yaṇa which grants [serenity] to the moving and stationary beings frightened for long! 50 Ibid., 516–19. 51 Ibid., 520–21. 52 Ibid., 521–22, and see nn. 141 and 144. 53 See Minkowski, “What Makes a Work ‘Traditional’?” 238–41 for the full passage and for a broader discussion of Nīlakaṇṭha’s innovations in reading the text. Nīlakaṇṭha reads the epic 460 460 PHILOLOGY AND CRITICISM from a Vedā ntic perspective, as conirming the truth of advaita (non-dual) philosophy, but his devotion to Viṣṇu is evident. 54 A word’s meaning is obtained in two ways. The original or etymological meaning is called yoga. The meaning established by the customary usage is called rūḍhiḥ . Thus, Devabodha explains both meanings of the word jaya. 55 This compound can be split in two ways: (1) pravṛtteḥ aṅgam, which means part of the beginning; and (2) pravṛttiḥ aṅgaṃ yasya tat, meaning (the subject matter) of which pravṛtti is a part. 56 Gode, “New Light on the Chronology of the Commentators of the Mahā bhā rata,” 103. Note that this list is not the same as the list of Mahā bhā rata commentators and their dates (without reference to manuscripts) given in Sukthankar, “Notes on Mahā bhā rata Commentators,” 202 (the Appendix). 57 All attempts to trace this list have failed. It is not included in the two-volume edition of Sukthankar’s articles and works—Sukthankar Memorial Edition, 2  vols., ed. P.  K. Gode (Bombay:  Karnatak Publishing House, 1944–45). It is also not found in the digitized volumes of the Annals of the Bhandarkar Oriental Research Institute (available through JSTOR). It is possible that Gode was referring to a list for private and/or internal use, known only to staf and colleagues at the Bhandarkar Oriental Research Institute, as he does not mention a source. 58 This measure is imperfect, of course, since more manuscripts may have turned up in the mean- time or editions of more commentaries published. Our access to editions published in India especially has been limited. But we hope that this appendix will be a starting point for other scholars to ill in the missing data. 59 The number of manuscripts, strictly speaking, refers to the number used for the critical edition and not the total number of manuscripts available. But as this was one of the widest searches conducted for manuscripts of the commentaries and the editors did not discount any important source, it may be taken as representative of the latter igure (except in the case of Nīlakaṇṭha, of whose commentary likely many more copies exist than were used). 60 R. N. Dandekar, Commentary of Devabodha on the Ad̄ iparvan of the Mahābhārata (Poona: Bhandarkar Oriental Research Institute, 1941). 61 Hereafter, this statement will not be repeated. It is true, of course, of Nīlakaṇṭha’s commentary for all 18 parvans of the Mahā bhā rata. 62 “The Bombay Government MSS Collection (no. 180 of 1891–95) contains a fragment of his commentary on the Ā di, comprising merely the irst adhyā ya with the beginning of the second; while the Madras Government Collection (no. 2169) contains another fragment which lacks the beginning.” Sukthankar, “Notes on Mahā bhā rata Commentators,” 187. 63 Hereafter, this statement will not be repeated. Except for Edgerton, who used the printed text of the Sabhā parvan in the P. P. S. Sastri edition, all of the other parvan editors made use of a Devanā garī transcript of Vā dirā ja’s commentary prepared by Sastri and later acquired by the Bhandarkar Oriental Research Institute. In this case, the number of manuscripts is cited as one, since the editors could not have had access to a second manuscript that Sastri acquired later and whose readings are cited in the printed edition. However, a second manuscript must also exist for these parvans. 64 Franklin Edgerton, “Introduction,” in The Sabhāparvan for the First Time Critically Edited (Poona: Bhandarkar Oriental Research Institute, 1944), xxiv. 65 Note that this manuscript (and a second that Edgerton does not cite) later proved not to be copies of Vimalabodha’s commentary, as per Belvalkar’s information (discussed later). 66 Edgerton, “Introduction,” xxiv. 67 R. D.  Karmarkar, Commentary of Devabodha on the Sabhāparvan of the Mahābhārata (Poona: Bhandarkar Oriental Research Institute, 1949). 68 There appear not to have been many commentaries on this parvan; Sukthankar notes that Devabodha’s commentary on the Ā raṇyakaparvan “has unfortunately been lost, or at least has not been recovered so far.” Sukthankar, “Introduction,” xi. 461 APPENDICES 461 69 Raghu Vira, “Introduction,” in The Virātạ parvan for the First Time Critically Edited (Poona: Bhandarkar Oriental Research Institute, 1936), x. 70 S. K. De, “Introduction,” in The Udyogaparvan for the First Time Critically Edited (Poona: Bhandarkar Oriental Research Institute, 1940), xiii. 71 S. K. De, Commentary of Devabodha on the Udyogaparvan of the Mahābhārata (Bombay: Bharatiya Vidya Bhavan, 1944). For the Virā ṭa and Udyoga parvans, we also have the aforementioned edition of Mahadeva Shastri Gangadhara Bhatta Bakre (1915 and 1920). If we include this edition (now extremely rare), then published versions exist of eight Mahā bhā rata commen- taries for these two parvans. 72 More than one copy may exist, for Belvalkar remarks: “An almost complete MS of Vā dirā ja’s commentary on the Mahā bhā rata is available in the Mysore Oriental Library, besides parts of the commentary on stray parvans in several libraries of South India.” S.  K. Belvalkar, “Introduction,” in The Bhı̄ṣmaparvan for the First Time Critically Edited (Poona:  Bhandarkar Oriental Research Institute, 1947), lxix. 73 Manuscripts including commentaries of the Bhīṣmaparvan appear to have been especially numerous. The editor of this parvan notes:  “there exist, besides these 125 MSS of the text alone, nearly 40 other MSS giving the text of the parvan together with a Sanskrit commen- tary: namely, Arjunamiśra’s Bhāratārthadı̄pikā, some 10 MSS; Nīlakaṇṭha’s Bhāratavbhāvadı̄pa (or dı̄pikā), about 30 MSS; and two MSS of an anonymous commentary. Manuscripts containing commentaries alone on the Bhīṣmaparvan number more than 30, there being, besides the two commentators already mentioned, about ten others, including Devabodha, the oldest of them all.” Belvalkar, “Introduction [to the Bhīṣmaparvan],” ix. 74 The identity of this commentator is uncertain. Belvalkar calls him “Yajña (or Sarvajña) Nā rā yaṇa,” but Sukthankar in his list of Mahā bhā rata commentators cites the two separately. 75 Belvalkar, “Introduction [to the Bhīṣmaparvan],” xii. 76 See S. K. Belvalkar, The Śāntiparvan: Introduction (Poona: Bhandarkar Oriental Research Institute, 1966), cxxxix–cxli and cliv–clv: “This commentary is mentioned in the Critical Apparatus of the Sabhā , Virā ṭa, Udyoga and Bhīsm ̣ aparvans, on the authority of two paper MSS, viz. no. 84 of 1869–70 and no. 167 of 1887–91, both belong to the Deccan College, MS Library, now deposited in the B. O. R. Institute. The irst of these three MSS is 9˝ x 5½˝ in size, with 88 folios, giving 16 lines to a page and about 36 letters to a line. The second is 13½˝ x 6˝ in size, with 67 folios, and giving about 12 lines to a page and 48 letters to a line. […] There are, however, certain indications showing that what we have in these two MSS is not the original commen- tary of Vimalabodha, but only some select extracts from that commentary. […] Thus we have three diferent names to deal with:  Vimalabodha’s original Durghatạ r̄ thaprakāsí nı;̄ Trivikrama’s Visạ mapadyatātparyārthavivaranạ tı̣ k̄ ā (in MS no. 84 of 1869–70 only), and Durbodhapadabodhinı ̄ or bhañjikā according to the two MSS before us. […] The two MSS above described do not, there- fore, constitute the real commentary of Vimalabodha, but only a selection of diicult extracts from the same. This was proved by the fortunate discovery of a MS belonging to the Rā ja Library at Darbhanga, which contained an unnumbered MS written in Maithili characters and giving the original commentary of Vimalabodha on the Rā jadharma (which, according to the commentator, includes the Ā paddharma). A  Devanā garī transcript of the MS was obtained through the courtesy of Shri Jīvā nanda Ṭhakkara, the Head Pandit of that Library. […] For the Moksạ dharma, unfortunately, we have to rely on the extracts from Vimalabodha’s commentary as supplied by our two MSS, no.  84 of 1869–70 and no.  167 of 1887–91. As Vimalabodha quotes Bhojarā ja’s Vyavahāramañjari and is himself quoted by Vidyā sā gara (Cs), his date, according to Dr. P. K. Gode (Studies in Indian Literary History, vol. I, p. 422) lies between 1150 and 1300 CE. The commentator is therefore fairly old, and Arjunamiśra, although difering from him once in a whole, generally follows him.” 77 S. K.  Belvalkar, Commentary of Devabodha on the Bhı̄ṣmaparvan (Poona:  Bhandarkar Oriental Research Institute, 1947). 462 462 PHILOLOGY AND CRITICISM 78 De, “Introduction [to the Droṇaparvan],” xiii. 79 See Dandekar, “Introduction [to the Ś alyaparvan],” xl (italics added):  “Ujjain, Sanskrit Oriental Institute Library, No. 482/5797 […] The MS., which is dated Ś aka 1571 ( = 1649 CE) and which is said to have been written by one Nā rā yaṇabhaṭṭa Mahokṣa, gives only the commentary without the text […] there is, at the end of the commentary on Crit. Ed. adhy. 28, a colophon with reads: iti śrı̄paramahaṃsācār yaśrı̄devabodhakṛtau bhāratadı̄pikāyāṃ śalyaparvavivaraṇaṃ samāptam. The ascription of this commentary, on the basis of the colophon, to Devabodha is obviously a mistake.” He continues: “Colophons 3, 5, and 6 in the Baroda MS., as cited above, are fairly detailed and seem to put it past doubt that that MS. contains Arjunamiśra’s commentary on the Ś alyaparvan. The Ujjain MS., which, as mentioned above, is essentially identical with the Baroda MS., must, therefore, be also regarded as containing Arjunamiśra’s commentary. […] The ascription of the com- mentary to Devabodha may have been due to the occurrence of the name of Devabodha in the longer colophon in Arjunamiśra’s commentary.” Ibid., xli. And see ibid., xli, n. 3, where Dandekar speciically mentions:  “The same Ujjain MS.  seems to have been used—again erroneously—as embodying Devabodha’s comm. also for the Crit. Ed. of the Droṇaparvan (see: Introduction to that parvan, p. xiii). De had access to Arjunamiśra’s commentary in man- uscript no. 2278 of the Benares Sanskrit College, so it is unclear why he did not note the error, since his two manuscripts presumably contained the same text. 80 P. L.  Vaidya, “Introduction,” in The Karṇaparvan for the First Time Critically Edited (Poona: Bhandarkar Oriental Research Institute, 1954), xxi. It is not clear whether by “com- plete MS” Vaidya means complete for this parvan or complete for the entire Mahā bhā rata. 81 See Belvalkar, The Śāntiparvan: Introduction, cxxxix–cxli and cliv–clv, cited previously. Belvalkar does not speciically mention the Karṇaparvan, but it is the same manuscript. 82 Velankar, “Introduction,” xxiii. 83 See Dandekar, “Introduction [to the Ś alyaparvan],” xl–xli, cited previously. As with De, it is unclear why Velankar did not notice the discrepancy, since he had access to Arjunamiśra’s commentary. 84 Paranjpe, “Introduction,” xix. 85 See Dandekar, “Introduction [to the Ś alyaparvan],” xl–xli, cited previously. As in De’s and Velankar’s cases, it is unclear why Paranjpe did not notice the discrepancy, since he had access to Arjunamiśra’s commentary. 86 See Belvalkar, The Śāntiparvan: Introduction, cxxxvi, who mentions that he also used “for com- parison, copies procured from two other MSS in the Tanjore Sarasvathi Mahal Library, num- bered, Burnell 1214 (= no. 8652 in P. P. S. Sastri’s Descriptive Catalogue, vol. XV, no. 8652) and 1237 (= ibid., no. 8651), both undated and written on paper in Devanā garī characters.” Ibid. These manuscripts are not mentioned in Belvalkar’s “Chart of the Ś ā ntiparvan Critical Apparatus” on pp. xix–xx, which therefore needs to be supplemented with the information in the “Detailed Account of the Manuscripts.” “These Tanjore copies gave, without the text, the commentary of Arjunamiśra on the Rā jadharma sub-section only, while a third copy pro- cured from Mysore Oriental Research Institute Library, no. S. A. 125, folios 138, also written in Devanā garī, gives Arjunamiśra’s commentary on the Ā paddharma sub-section.” Ibid., cxxxvi–cxxxvii. Additionally, Belvalkar made use of the text of Arjunamiśra’s commentary on adhyāyas 12.260–62 (the Gokapilīya chapters) in Friedrich Weinreich’s dissertation of 1928 (commentary available in the manuscript numbered 330 and in the London manuscript). Since it is not known whether these two manuscripts ofer a relatively complete text of the commentary or only this section, they have not been added to the count above. Weinreich’s remaining manuscripts are enumerated in the critical edition of the Mokṣadharmaparvan; see S. K. Belvalkar, ed., The Śāntiparvan, Part III: Mokṣadharma, B, For the First Time Critically Edited (Poona: Bhandarkar Oriental Research Institute, 1954), 2194. 87 More than one copy may exist, for Belvalkar remarks:  “Vā dirā ja’s commentary on the Mahā bhā rata in an almost complete form is available in the Mysore Oriental Library, and stray 463 APPENDICES 463 portions of it in the several libraries of Southern India.” Belvalkar, The Śāntiparvan: Introduction, clx. He does not cite his source for this information. 88 In addition to the manuscript listed in the table later, Belvalkar notes that “there is also a MS of Paramā nanda’s Mokṣadharmatı̣ k̄ ā mentioned by Burnell. It is a palm-leaf MS in Grantha characters, dated Saṃvat 1690 (cir. 1633 CE), of which no further details are available in P. P. S. Sastri’s Descriptive Catalogue. This MS was not used for the present edition.” Ibid., clviii. 89 In addition to the two manuscripts listed in the table later, Belvalkar notes that “subsequently, a MS of the same comm. was procured from the Hyderabad State Library, which did not ofer much diference.” Ibid., clix. 90 See ibid., clv: “one more MS, Lahore, no. 2935 from D. A. V. College, Lalchand Library (now transferred to Sadhu Ashram, Hoshiarpur, East Panjab), giving the Rā jadharma alone, was collated but not used for the present ed. except in 12.79.41.” 91 See R.  N. Dandekar, “Introduction,” in The Anuśāsanaparvan for the First Time Critically Edited (Poona:  Bhandarkar Oriental Research Institute, 1966), xxxvi–xxxvii. Belvalkar’s informa- tion in The Śāntiparvan: Introduction, cxxxix–cxli and cliv–clv overturns Dandekar’s ascription. Dandekar’s edition must have appeared concurrently with Belvalkar’s introduction in 1966, for Belvalkar does not mention the Anuśā sanaparvan and if his introduction had appeared earlier, Dandekar would surely have corrected it. 92 Belvalkar used only one manuscript for each of these parvans, but, according to his information in the Ś ā ntiparvan, a second manuscript containing Nīlakaṇṭha’s commentary on these parvans also exists (no. 29c of the Maharashtra Government Collection). 93 Belvalkar notes in the introduction to the Bhīṣmaparvan that “a paper MS in Bengali characters of Devabodha’s comm. on the Ā raṇyakaparvan is mentioned by Rajendra Lala: Notices of MSS, Vol. v, no. 3009. The MS belonged to Pratapacandra Bhattacharya of Singura, Dist. Virabhuma” and that “no.  3010 is a Bengali MS of Devabodha on the Virā ṭaparvan, of the same owner.” Belvalkar, “Introduction [to the Bhīṣmaparvan],” lxii, n. 4 and lxiii, n. 1. If his information is correct, we would possess a near complete commentary of the earliest of the Mahā bhā rata commentators up to the end of the Bhīṣmaparvan. 94 Belvalkar notes that “of this old and venerable commentator, no MS of the commentary on the entire Mahā bhā rata is known to exist. The Critical Edition could use his commentary on the Ā diparvan, the Sabhā parvan, the Udyogaparvan and the Bhīṣmaparvan (without the Bhagavadgītā ). We still lack his commentary on the Ā raṇyakaparvan, the Virā ṭaparvan, the Bhagavadgītā , the Karṇaparvan, the irst eight chapters of the Strīparvan and all the remaining parvans from the Ś ā nti to the end.” Belvalkar, “Introduction [to the Bhīṣmaparvan],” lxii– lxiii. His count is obviously based on assigning the manuscript in the collection of the Scindia Oriental Institute, Ujjain (no. 482/5797) to the Devabodha commentaries. However, this com- mentary later turned out to be a commentary of Arjunamíra, with merely an incorrect colo- phon attributing it to Devabodha. 95 We know that this must have existed, because, according to Belvalkar, “there is a MS (Baroda Or. Inst. no. 13036) of Caturbhuja’s commentary on the Bhīṣmaparvan which seems to have preserved a portion of Devabodha’s commentary on BG. 2.11–16 and 4.17–31. But we cannot be certain of its authenticity. Sarvajña-Nā rā yaṇa [also] quotes Devabodha’s gloss on BG. 11.36.” Belvalkar, “Introduction [to the Bhīṣmaparvan],” lxiii. 96 A complete commentary must have existed at one time, for Belvalkar remarks:  “Complete MSS of Arjunamíra’s commentary on the entire Mahā bhā rata are no longer met with, although Rā jendralā la Mitra, in his Notices, vol. V, p.  viii, records having come across one such at the village of Guḍā p (District Burdwan). MSS of the commentary for parvans 14, 16–18, as also for occasional passages of the Rā jadharma and the Ā paddharma sections of the Ś ā ntiparvan, have not yet been found in any of the collections of MSS of which lists are available.” Belvalkar, “Introduction [to the Bhīṣmaparvan],” lxxi. 46 464 PHILOLOGY AND CRITICISM 97 This is not a conclusive argument, of course. Only a study of the colophons of the commen- taries or the commentaries themselves can help settle the question. Dandekar, for instance, notes that, in his commentary Durghatạ ̄rthaprakāśinı̄ on folio 58a of the manuscript in the Bombay (now Maharashtra) Government Collection (deposited at the BORI), no.  84 of 1869–70, Vimalabodha remarks:  śalyādiparvāṇi sugamānı̄ti rājadharme nirupyante. “So there is not available any commentary by Vimalabodha on the Ś alyaparvan as such.” Dandekar, “Introduction [to the Ś alyaparvan],” xxxix. But even this is not always a guide, since there is no necessity that the commentator speciied the books he commented. On the other hand, it is dangerous to conclude ex silentio from the fact that we do not possess a commentary for certain parvans of the Mahā bhā rata that the author of that commentary did not write one on them. It may simply be that the commentary for that section has not survived. In the case of Caturbhuja, for instance, Belvalkar remarks: “Complete MSS of Caturbhuja’s commentary on the entire Mahā bhā rata are no longer extant. MSS of his commentary on the Ā raṇyaka, Virā ṭa (published by the Gujarati Printing Press, Bombay, 1915), and Droṇa parvans have been reported.” Belvalkar, “Introduction [to the Bhīṣmaparvan],” lxvi. At least for the commentators, whose practice it appears to be to write continuous commentaries, then, we might be justiied in assuming that commentaries for the remaining parvans existed at one time. 98 With the exception of Ratnagarbha and Caturbhuja, who seemingly wrote complete com- mentaries, the remaining commentators appear to have only commented on speciic sections of the epic. A search for more extensive commentaries would probably be in vain. But see the preceding note. 99 The standard work, often referred to as the Sadhale edition, is Shastri Gajanana Shambhu Sadhale, ed., The Bhagavadgita with Eleven Commentaries, 2nd edn., 3  vols. (Bombay:  The “Gujarati” Printing Press, 1935; reprinted Delhi: Parimal Publishers, 2000), which includes the commentaries of Ś aṅkara, Ā nandagiri, Rā mā nuja, Vedā nta Deśika, Vallabha, Madhva, Jayatīrtha, Hanumat, Venkaṭanā tha, Nīlakaṇṭha and Yamuna. A  second multi-commen- tary edition is Wasudev Laxman Sastri Pansikar, ed., Śrı̄madbhagavadgı̄tā with the Commentaries of Śrı̄mat-Śaṅkarabhāṣya with An̄ andagiri; Nı̄lakaṇtḥ ı̄; Bhāṣyotkarṣadı̄pikā of Dhanapti; Śrı̄dharı̄; Gı ̄tārthasaṅgraha of Abhinavaguptāchār ya; and Gūḍhārthadı̄pikā of Madhusūdana with Gūḍhārthatattvā- loka, 2nd edn. (Bombay: Nirṇaya-Sā gar Press, 1936; reprinted Delhi: Munshiram Manoharlal, 1978). 100 See, however, W.  M. Callewaert and Shilanand Hemraj, Bhagavadgı̄tānuvāda:  A Study in Intercultural Translation (Ranchi: Satya Bharati Publication, 1983), 96–110, who count a total of 249 commentaries on the Gītā (either extant or known from a reference or cited in a man- uscript catalog). 101 Almost all of the literature on one or more of the major Vedā nta schools discusses the rela- tionship of the principal commentators to each other. See also T. G. Mainkar, A Comparative Study of the Commentaries on the Bhagavadgı̄tā, 2nd edn. (Delhi: Motilal Banarsidass, 1969). 102 No commentaries more recent than Nīlakaṇṭha were collated, perhaps because their value for establishing a critical text was minimal. 103 Editions of Ś aṅkara and Rā mā nuja are the most numerous by far, and therefore we have cited only a few standard editions. For those commentators whose editions are rare, we have striven to be more comprehensive. 104 Franco describes Brockington as “arguably the greatest living scholar on Indian epic litera- ture.” Eli Franco, review of The Nay Science: A History of German Indology, by Vishwa Adluri and Joydeep Bagchee, South Asia: Journal of South Asian Studies 39, no. 3 (2016): 696. Von Hinüber calls Brockington’s The Sanskrit Epics “the foundation for all further researches into ancient Indian epic and a secure and reliable guide for everyone who preoccupies himself in some way with these texts.” Oskar von Hinüber, review of The Sanskrit Epics, by John Brockington, Wiener Zeitschrift für die Kunde Südasiens 46 (2002): 269. 465 APPENDICES 465 105 International colloquium, “Enjeux de la philologie indienne: traditions, éditions, traductions– transferts,” December 5–7, 2016, Paris, Collège de France. For the conference theme, program and abstracts, see:  www.iran-inde.cnrs.fr/evenements-scientiiques/colloques- et- conferences- 2016/ enjeux- de- la- philologie- indienne- traditions- editions- traductions- transferts.html?date_debut=2017-03-28&lang=fr (accessed August 30, 2017). 106 John Brockington, “Regions and Recensions, Scripts and Manuscripts: The Textual History of the Rā mā yaṇa and Mahā bhā rata,” www.college-de-france.fr/site/jean-noel-robert/ symposium-2016-12-06-14h00.htm (accessed June 30, 2017). We thank Silvia D’Intino for the link. The website incorrectly lists the title as “Religions and Recensions, [etc.]” We thank Brockington for the hint. 107 Ibid. (all italics ours). Brockington refers to Renate Söhnen, Untersuchungen zur Komposition von Reden und Gesprächen im Rāmāyaṇa, 2  vols., Studien zur Indologie und Iranistik 6 (Reinbek:  Dr.  Inge Wezler, Verlag für orientalistische Fachpublikationen, 1979). For his review, see John Brockington, review of Untersuchungen zur Komposition von Reden und Gesprächen im Rāmāyaṇa, by Renate Söhnen, Journal of the Royal Asiatic Society of Great Britain and Ireland 114, no. 1 (1982): 65: “The irst volume begins with various general observations, such as the utility in an oral narrative of clear markers at the beginnings and endings of speeches. The author then distinguishes three types of dialogue, dramatic (infrequent, despite providing all four passages chosen for detailed study), narrative, and emotive. She takes as her basis the Bombay edition on the grounds that, unlike the Critical edition, it represents a relatively uni- form text. However, she lacks the courage of her convictions, regularly discussing the Critical edition’s readings, frequently—especially when it its her analysis—accepting its rejection of stanzas found in the Bombay text, and occasionally suggesting that certain passages found in both texts are not original. Her comments on such occasions are sensible but inconsistent with her own initial position. Nevertheless, Söhnen argues persuasively for frequent delib- erate structuring of speeches, especially for ‘Ringkomposition’, the organization of a speech symmetrically around a central passage (less commonly there is no central point as such but just an axis of symmetry). There is room for discussion about details of her analysis, and in particular the varied nature of what constitutes a ring, for the similarity may rest on theme, mode of expression, wording, or merely the attitude of the speaker, while the bal- ancing elements may difer considerably in length. As the author recognizes, the interpreta- tion therefore involves a substantial subjective element. More objective evidence of attention to structure in speeches is provided by the undoubted greater frequency of refrains than in the narrative parts, and also of chiasmus (to which she gives a more extended meaning than I would).” 108 Reinhold Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” in Studien zur Indologie und Buddhismuskunde, ed. Reinhold Grünendahl,  Jens Uwe-Hartmann and Petra Kiefer-Pülz, Indica et Tibetica 22 (Bonn:  Indica et Tibetica Verlag, 1993), 101–30. Alternately, Grünendahl may himself owe the error to Brockington, who asserts, “Apart from the Devanā garī manuscripts, the [Rā mā yaṇa] Critical Edition groups its manuscripts by the script employed but, as I stress in the previous paper, the numbers [sic] of manuscripts in each version varies considerably. The point is equally valid of the Critical Edition of the Mahā bhā rata.” John L.  Brockington, “The Text of the Rā mā yaṇa,” Indologica Taurinensia 15–16, Proceedings of the Seventh World Sanskrit Conference, Leiden, August 23–29, 1987 (1989–90): 80. The error repeats in John L. Brockington, “Textual Studies in Vā lmīki’s Rā mā yaṇa,” in Epic Threads: John Brockington on the Sanskrit Epics, ed. Greg Bailey and Mary Brockington (New Delhi: Oxford University Press, 2000), 195–206 (verbatim on page 197). (The article was irst published in 1986 in the Journal of the Asiatic Society of Calcutta.) We cannot determine who borrowed the thesis from whom, since neither scholar cites the other. Von Hinüber, review of The Sanskrit Epics, 268 notes the pointed omission: “Only very rarely has a title that deserved mention escaped the author’s notice, whose overview of scholarship 46 466 PHILOLOGY AND CRITICISM is current up until the book’s very publication date. This applies, for example, to the impor- tant contribution by R.  Grünendahl, [...] who with good reasons disputes the connection between the script and recension, which V.  S. Sukthankar had already questioned.” The reasons for this silence may have to do with Brockington’s negative review of Grünendahl’s dissertation. See John Brockington, review of Viṣṇudharmāḥ :  Precepts for the Worship of Viṣṇu Part 1: Adhyāyas 1–43, by Reinhold Grünendahl, Journal of the Royal Asiatic Society of Great Britain and Ireland 2 (1984): 291–92. 109 Bigger’s “English Summary” states:  “To sort their mss. for the edition, SUKTHANKAR and his followers used what GRÜNENDAHL (1993b) called the premise of scripts (Schriftartenprämisse). GRÜNENDAHL showed that this idea, namely that all mss. written in a certain script belong together, is not tenable, but it may still be useful as a irst cri- terion.” Andreas Bigger, Balarāma im Mahābhārata:  Seine Darstellung im Rahmen des Textes und seiner Entwicklung (Wiesbaden:  Otto Harrassowitz, 1998), 170. Mahadevan circumscribes Grünendahl’s notion thus: “the texts of a given script hang together” and “A critical edition (CE) of a work, by deinition, is an assemblage of its available texts, anchored on the twin maxims of textual scholarship: brevior lectio praeferenda est (the shortest text is to be preferred) and Schriftartenprämisse (manuscripts [MSS] in the same script belong together.).” T. P. Mahadevan, “The Ś akuntalā -Yayā ti ‘Transposition’, the Southern Recension of the Mahā bhā rata, and V. S. Sukthankar,” in The Churning of the Epics and Purāṇas at the 15th World Sanskrit Conference, ed. Simon Brodbeck, Alf Hiltebeitel and Adam Bowles (New Delhi: Dev Publishers, 2018), 60 and T. P. Mahadevan, “Three Rails of the Mahābhārata Text Tradition,” Journal of Vaishnava Studies 19, no. 2 (2011): 23. 110 Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 104. 111 Bigger, Balarāma im Mahābhārata, 118. 112 Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 117 (Grünendahl’s italics). 113 Ibid., 121 (italics in original). 114 Bigger, Balarāma im Mahābhārata, 15. 115 Andreas Bigger, “The Normative Redaction of the Mahābhārata: Possibilities and Limitations of a Working Hypothesis,” in Stages and Transitions: Temporal and Historical Frameworks in Epic and Purāṇic Literature. Proceedings of the Second Dubrovnik International Conference on the Sanskrit Epics and Purāṇas, ed. Mary Brockington (Zagreb: Croatian Academy of Sciences and Arts, 2002), 20 (italics added). 116 Oskar von Hinüber, “Der vernachlässigte Wortlaut. Die Problematik der Herausgabe buddhistischer Sanskrit- Texte, Zur Ü berlieferung, Kritik und Edition alter und neuerer Text,” in Kleine Schriften, Teil I, ed. Harry Falk and Walter Slaje (Wiesbaden: Harrassowitz, 2008), 436. 117 Brockington cites Andreas Bigger, “Balarā ma im Mahā bhā rata,” Asiatische Studien/ Études Asiatiques xlviii, no. 1 (1994): 1297–99 with the words: “These episodes are briely studied by Andreas Bigger (1994), as a foretaste of his thesis on the same subject,” but does not indicate whether he has read this latter work. See Brockington, The Sanskrit Epics, 288, n. 129. Most likely, it had not appeared at the time (both works appeared concurrently in 1998). There is no reference to the Mahā bhā rata critical edition as a “normative redaction” and the term redaction itself appears only sparingly (for example, on pages 21 and 156; on page  71 it is linked to von Simson’s work). Whenever Brockington discovered the idea that the critical edition reconstructs a normative redaction, it must be after 1998. 118 Ordinarily, this level of dependence would raise suspicions of plagiarism. In Brockington’s case, it is likelier he lost sight of when he was speaking on the German scholars’ behalf. We know of only one instance of demonstrable plagiarism, but here also it is likelier Brockington forgot he was repeating in translation something his German source already said. The passage occurs in John Brockington, “The Bhagavadgı̄tā: Text and Context,” in The Fruits of Our Desiring: An Inquiry into the Ethics of the Bhagavadgı ̄tā for Our Times, ed. Julius Lipner 467 APPENDICES 467 (Calgary: Bayeux, 1997), 32. “Also, in the visit of the Pā ṇdạ vas to the enemy, which follows the Bhagavadgı̄tā, we see a late piece of Brahmin apologetic, which equally with the second visit to Bhīṣma (6.103.40–84) must have been inserted for the moral justiication of the Pā ṇḍavas.” It is a literal translation from Georg von Simson, “Die Einschaltung der Bhagavadgı̄tā im Bhı ̄ṣmaparvan des Mahābhārata,” Indo-Iranian Journal 11, no. 3 (1968/69): 174: “Auch in dem auf die Bhagavadgı̄tā folgenden Besuch der Pā ṇḍavas im feindlichen Heer (6.41) sehen wir ein spätes Stück brahmanischer Apologetik, das ebenso wie der zweite Besuch bei Bhīṣma (6.103.40–84) zur moralischen Rechtfertigung der Pā ṇḍavas eingefügt wurde.” Brockington does not enclose the passage in quotation marks, and no citation is provided. 119 For Brockington’s views of the Mahā bhā rata’s “brā hman redactors,” see The Sanskrit Epics, 11, 19, 237 and 332. In John Brockington and Mary Brockington, Rāma the Steadfast: An Early Form of the Rāmāyaṇa (London: Penguin Books, 2006), the distinction between the Vedic-Brahmanic and the Epic-Heroic traditions is encapsulated thus:  “Transmission of the Rāmāyaṇa and Mahābhārata in their oral stages had been a diferent process from transmission of the Vedic texts that preceded them, which in traditional circles still retain their original oral form. Knowledge of the Vedic texts was restricted to the brā hmanic class, and the words themselves took on a ritual, even a magic, signiicance largely unrelated to their semantic meaning. They were handed down verbatim over the centuries without variation, and the reciter’s or hearer’s understanding of them—or lack of it—was immaterial. The case of the Rāmāyaṇa and the Mahābhārata has been completely the opposite. They were heroic tales, narratives conceived as entertainment. Both began, and both continued for several centuries, as the preserve of the warrior kṣatriya class, before being taken over by brā hman redactors.” Ibid., xxiv. And see also ibid., 364: “Around the irst to third centuries AD brā hman rather than kṣatriya values came to dominate the text. The details of the process are not entirely clear, but seem to have involved a change to a new set of redactors; probably at this time the text was irst committed to writing, although it continued to be present orally.” Most recently, Mislav Ježić has revived the hypothesis, distinguishing “brahmanical, hieratic poetry” from “aristocratic, heroic poetry.” Here is how he expresses their distinction and transformation: “Vedic poetry is brahmanical, hieratic poetry. Formulations about celestials, divinities, world and the powers inherent in it are not narratives, no stories, and are not to be taken literally, but are expressions of a hidden, invisible order, which can be understood as righteous if seen in its truth (ṛta), in which the sages have insight (dhī), and which they express in an appropriately covert, enigmatic manner (brahman, brahmodya). We may term Vedic mantras or brahmans, or formulations in them myths, but we should be aware of the fact that they are not narrations, not even elliptical ones, and that celestials praised in them are not anthropomorphic, have no clear genealogies, and are essentially invisible to those who have no insights into their nature. Epic poetry is aristo- cratic, heroic poetry at the base, as evident in the case of the Mahā bhā rata, which can repre- sent a form of archaic court poetry, which may attain a certain poetical reinement close to kā vya, as in the case of Rā mā yaṇa. It describes individual heroes who are human and mortal, or partly superhuman, but certainly visible, for whom their genealogy is an essential condi- tion of their nobility, whose feats can and should be narrated to glorify their superior virtues, strength and skills. They serve as a model for other members of the aristocracy to follow. This epic poetry, if it reaches out for universal hieratic mythic models, either to describe the rela- tionship of mortals and immortals, or to enhance the glory of heroes by comparing it with that of celestials, or even by elevating them to the level of immortal divinities, may assume a form of religious poetry. This tendency can be reinforced if epic compositions are recited at religious ceremonies, or if epic tradition starts being transmitted by brahmans or priests. The idea that some epic heroes are sons of celestials, or that some of them are even incarnations (avatā ra) of God, gave the Mahā bhā rata and Rā mā yaṇa, and even more their khilas, their religious signiicance. If hieratic topics, like the origin or creation of the world, its dissolution and renewal, the cosmic periods, and origin of celestials and supernatural persons, become 468 468 PHILOLOGY AND CRITICISM the material for epic poets and singers, they will elaborate these by means of their poetical technique, and give the immortals anthropomorphic features, ixed genealogy, and individual (instead of regular or cyclical) feats, which can be narrated and tend to be understood liter- ally, as mythological facts. They need not be decoded, solved as riddles, but may be taken at their face value, remembered and transmitted (smṛti), in the same way as the genealogies of heroes and (legendary or historical) kings. That is the literary genre of transmitting ancient stories, ancient lore, namely the Purā ṇas.” Mislav Ježić, “Vedic Myths and Epico-Purā ṇic Mythology—Transformation of the Sense of Vision into Narration,” paper presented at the Dubrovnik International Conference on the Sanskrit Epics and Purā ṇas (DICSEP) 8, IUC, Dubrovnik, Croatia, September 12, 2017. Insofar as they oppose Vedic literature to the Mahā bhā rata, all such attempts go back to Holtzmann Jr.’s claim that: “Alongside this epic literature ran the religious literature of the Brahmans, both completely independent of each other, only occasionally making use of each other. Epic and Veda are equally ancient; no bridge leads from the Veda to the Mahābhārata; ‘the world of the Veda is a world by itself ’ […] of course, the epic knows of the Veda, albeit, to conclude from the Mahābhārata, only most super- icially; but, on the whole, the two literatures, the Brahmanic and the heroic, are completely independent of each other.” Adolf Holtzmann Jr., Zur Geschichte und Kritik des Mahābhārata (Kiel:  C. F.  Haeseler, 1892), 61–62. Despite Ježić’s attempt to separate the Mahā bhā rata from the Veda, itihāsa is a Vedic genre. The Mahā bhā rata is thus no less “brahmanical” and “hieratic” than the Vedas. The very term “epic poetry” is a misnomer. It has no equivalent in the Indian tradition. Its use for the Mahā bhā rata originated with Lassen, who used it in the sense of “the narration [Epos] of the battle of the Kurus and Pândavas.” Only later, once Holtzmann Jr. identiied the Mahā bhā rata with the remains of a common “Indo-Germanic epic” that constituted the “epic inheritance” (episches Erbgut) of the “Indo-Germanic people,” did it enter scholarly discourse as a term for the entire Mahā bhā rata. Ibid., 42–43. For Lassen’s use of the term, see Christian Lassen, “Beiträge zur Kunde des Indischen Alterthums aus dem Mahâbhârata I: Allgemeines über das Mahâbhârata,” Zeitschrift für die Kunde des Morgenlandes 1 (1837): 61–86. 120 For Brockington’s comments on the Sanskrit epics’ oral beginnings, see John Brockington, “The Textualization of the Sanskrit Epics,” in Textualization of Oral Epics, ed. Lauri Honko (Berlin and New York: Mouton de Gruyter, 2000), 193–216. He declares: “It hardly needs airming that the Mahā bhā rata and the Rā mā yaṇa together represent the culmination of a lengthy tradition of oral poetry, transmitted through recitation by sūtas or bards.” Ibid., 193. The attached note merely states: “All references to the epic are to their Critical Editions, Sukthankar et al. 1933–66; Bhatt and Shah 1960–75. I use the term ‘* passages’ generically to designate all material excluded from the text of the Critical Edition in either the critical apparatus or Appendix I (unless context clearly indicates a more limited usage).” Elsewhere Brockington suggests: “In both the Sanskrit epics, the Rāmāyaṇa and the Mahābhārata, their framework stories present them as oral compositions; this suggests the importance of the roles not only of the bard or reciter but also of the audience whose presence is integral to this introductory frame.” Ibid., 193. Insofar as he refers to the epic’s narration at Ś aunaka’s sattra in the Naimiṣa Forest, the statement is false. Ugraśravas neither composes the Mahā bhā rata before his audience nor refers to an anterior “oral composition.” In this irst reference, the Mahā bhā rata is spoken of as a collection Vyā sa authored and taught his students (Mahā bhā rata 1.1.15–19 and 63–64). The transmission is irst Brahmanic and only then bardic if we understand this term to mean a narrator of “ancient Lore” (sūta paurāṇika) rather than the representative “of a lengthy tradition of oral poetry” Brockington imagines him as. Contrary to Brockington’s claim that “the Mahā bhā rata in its frame indicates very clearly its oral nature. […] Here we have the epic presented as if being narrated then and there, with reciters and audience actually within the text” (ibid., 202), the sages assembled in the Naimiṣa Forest ask to hear Vyā sa’s composition (dvaipāyanena yat proktaṁ purāṇaṁ paramarṣiṇā, 1.1.15). 469 APPENDICES 469 The bard responds that he will narrate Vyā sa’s “entire thought” (mataṁ kṛtsnaṁ, 1.1.23). The Mahā bhā rata’s account of its transmission is more complex than most scholars imagine. It is keenly attuned to issues of memory, philosophical insight and narrative time. See Vishwa Adluri, “The Perils of Textual Transmission:  Decapitation and Recapitulation,” Seminar 608, The Enduring Epic (2010): 48–54 and Vishwa Adluri, “Literary Violence and Literal Salvation: Ś aunaka Interprets the Mahābhārata,” Exemplar: The Journal of South Asian Studies 1, no. 2 (2012): 45–68. 121 Brockington’s The Sanskrit Epics begins with this distinction: “It is clear that the Mahābhārata and the Rāmāyaṇa represent the culmination of a lengthy tradition of oral poetry, trans- mitted through recitation by the sūtas or bards. The oral origins of the Sanskrit epics have, indeed, long been recognised.” Ibid., 3 (italics added). Thereafter, Brockington cites several German sources. And see ibid., 18–19 for similar claims: “No doubt eulogies of heroes and cycles of stories underlie the extant epics; traditionally the Mahābhārata had an earlier and perhaps more clearly epic version in the Bhārata. The irst stage in the establishment of an epic tradition would then have been the progressive clustering of ballads and other material around some central theme, whether the personal vicissitudes of a hero or events of particular importance. Eventually, these cycles may have been linked together into works with a more com- plex plot. The Mahābhārata and Rāmāyaṇa would on this view constitute the end-point of a similar evolutionary process which can be identiied, according to some, through a number of intermediate stages. The reciters and preservers of this heroic poetry would be the sūtas or bards, who were attached to the courts of chieftains and who recited in short songs, and on major festivals the glorious deeds of their lords.” Later he adds: “Although the two extant epics are based on such traditional bardic material, in the course of time more religious and didactic material was included in them, to such an extent that the Mahābhārata in particular has rather lost the character of an epic, since so much of its volume is didactic material. On the face of it, the Mahābhārata does give some indication of its own development. The very setting of the scene suggests the process of expansion in transmission by which a Bhārata lay of 24,000 verses grew to the Mahābhārata of 100,000. Elsewhere there occurs an instance not only of self-reference but also of its narration by brā hmans (brāhmaṇāḥ kathayiṣyanti mahābhāratam āhavam, 5.139.56ab). In the irst stage of growth the basic story must have been subject to expansion from within, then in a second stage mythologised and in a third stage entirely taken over by the brā hmanical tradition; inally, after the epic was committed to writing, the number of manuscripts needed for the purpose seems to have become virtually a library of Indian tradition, to which new material of all sorts could be added.” Ibid., 20– 21 (all italics ours). The passage quotes Van Buitenen nearly verbatim, who writes: “The original story was in the irst phase of complication expanded from within, in the second phase mythologized, in the third phrase brahminized. One might even discern a fourth phase, after the epic was irst written down, when this collection of manuscripts became, as it were, a library to which new books could be added.” J.  A. B.  van Buitenen, “The Mahā bhā rata: Introduction,” in J. A. B. van Buitenen, trans., The Mahābhārata, vol. 1: The Book of the Beginning (Chicago:  University of Chicago Press, 1973), xxiii. In contrast, the idea of a “Bhārata lay of 24,000 verses” is probably owed to Hopkins’s idea of “Bhā rata (Kuru) lays,” allegedly “combined into one, but with no evidence of an epic before 400 B.C.” Van Buitenen also mentions a “Bhārata of 24,000 couplets” and an “old Bhārata lay.” Ibid., xxiii and xxiv. Except that he changes “The Bhārata” to “a Bhārata lay” and replaces “couplets” with “verses,” the statement is an exact facsimile of Van Buitenen’s claim: “Thus The Bhārata of 24,000 couplets grew to The Mahābhārata of 100,000.” Ibid., xxiii. For the Hopkins quote, see Edward W. Hopkins, The Great Epic of India: Character and Origin of the Mahābhārata (New York: Charles Scribner’s Sons, 1901), 397. 122 Brockington, “The Text of the Rā mā yaṇa,” 79–80:  “These eleven manuscripts add well over a third to the number used for the Critical Edition of the Ayodhyākāṇḍa (and more than a 470 470 PHILOLOGY AND CRITICISM quarter even to those that were initially collated). It is not surprising, therefore, that together they add substantially to the body of variant readings recorded in the critical apparatus. This material is signiicant, even if of no great value in itself, for building up a truer picture of the com- plex relationship between the diferent recensions and versions” (italics added). 123 “What principle should we follow if not the manuscripts’ univocal testimony? If they are obviously corrupt, the editor may propose a conjecture, though this situation will be rare, since a problematic reading is unlikely to be conserved in all manuscripts” (Contini: difraction). Sukthankar therefore rightly preferred documented readings (Sukthankar, “Prolegomena,” xcii: “Emendation has played a very inconspicuous role in the preparation of the constituted text. Interpretation has in general been given preference over emendation. Even in the case of corrupt passages, the reading of some manuscript or other gives sense, though it may not be the original sense, not even a wholly satisfactory sense. Precipitate emendation is, however, to be deprecated; for experience has shown that but a small proportion of scholars’ corrections are really amendments”; Sukthankar’s italics). 124 The error relects a prevailing confusion about critical editions. Despite many excellent technical handbooks, many still think a critical edition is an edition that has been crit- icized. They are unaware that critical edition refers to an edition based on an exhaustive survey of the manuscripts that uses clearly enunciated principles and simple rules to recon- struct a deinite state of the text. The term is frequently applied to a Lachmannian or neo-Lachmannian edition and hence as a synonym for the genealogical-reconstructive edition. Brockington’s criticisms are irrelevant because Sukthankar anticipated the charge, and warned against expecting more from the edition than it could provide: “The reader will ind that the constituted text is by no means smooth. It contains fresh instances of loose and archaic linguistic forms and constructions, anacoluthons and lack of syntactical concord. There remain many contradictions and superluities. There is evident lack of inish in the hidden parts. These blemishes—if they be blemishes in epic poetry, which is dynamic poetry, with no necessary pretensions to niceties of style, in the narrower sense of the term—must have been inherent in the old poem. Where they are met with in the critical text, they are not speculative iction; they are documented by the manuscripts them- selves or at least are inferable from them with a high degree of probability.” Sukthankar, “Prolegomena,” ciii–civ. What Brockington seeks is not a critical edition, but an emended edition meeting his expectations of simplicity and consistency. From a critic’s perspective, this is the well-known tendency of banalization. 125 The reference could be to any of three works:  M. A.  Mehendale, “Interpolations in the Mahā bhā rata,” Annals of the Bhandarkar Oriental Research Institute 82, no.  1/4 (2001):  193– 212; M.  A. Mehendale, “The Critical Edition of the Mahā bhā rata:  Its Achievement and Limitations,” Annals of the Bhandarkar Oriental Research Institute 88 (2007):  1–16; and M.  A. Mehendale, “The Critical Edition of the Mahā bhā rata:  Its Constitution, Achievements, and Limitations,” in Text and Variations of the Mahābhārata: Contextual, Regional and Performative Traditions, ed. Kalyan Kumar Chakravarty (New Delhi: National Mission for Manuscripts, 2009), 3–23. But it does little to bolster Brockington’s claims. Brockington likely cites Mehendale because, like him, Mehendale also proceeds from the assumption that the Mahā bhā rata “was orally transmitted for many centuries.” He divides interpolations in the Mahā bhā rata into three categories:  “1. Those revealed by ms. evidence and set aside by the editors, 2. those that are not revealed by ms. evidence and part of the constituted text but are quite obvious due to contradictions in consecutive stanzas and 3. those revealed either by contradictions that are not so obvious because they are not found in consecutive passages but are found in the text as a whole, or by the criterion of intrinsic probability.” Mehendale, “Interpolations in the Mahā bhā rata,” 195–96. Insofar as Brockington refers to the consti- tuted text, he cannot mean the irst category. Insofar as his claim concerns the second cate- gory, the manuscript evidence is unambiguous and no Lachmannian or neo-Lachmannian 471 APPENDICES 471 editor would emend the text against the manuscripts’ uniform testimony. This leaves the third possibility—that Brockington, like Mehendale, thinks the critical edition does not pro- vide a facsimile of the “original” epic, because of its perceived inconsistencies. This suggests that Brockington’s work aims, rather, at a subjective Konjekturalkritik. Insofar as he thinks this continues the editors’ work he commits the same error as Mehendale when he suggests his work contributes to Mahā bhā rata textual criticism. See Mehendale, “Interpolations in the Mahā bhā rata,” 196:  “The task of removing the spurious matter that still remains in the constituted text and bringing it as close as possible to the older version of the Mbh. has been left by them [the Mahā bhā rata editors] to future students of the text.” Mehendale annotates the statement with a reference to Edgerton. “This task has been characterized as the appli- cation of ‘higher criticism’ to the text. See F.  Edgerton, Introduction to Sabhā -parvan, p. XXXIII.” Ibid., 196, n. 12. In Mehendale, “The Critical Edition of the Mahā bhā rata: Its Achievement and Limitations,” he notes: “The Editors of the critical edition have already pointed out in their respective introductions to the parvans edited by them instances of such contradictions, especially when they occur in passages close to each other. They have, how- ever, not said which of the two contradictory passages could be original and which spurious. That, according to them, would mean going beyond the scope of the work they had under- taken. In all such cases, taking decisions would be the task of later researchers by adopting ‘higher criticism.’ ” Ibid., 7–8. The attached note reads: “To try to make the text consistent on such points would be to enter the realm of ‘higher criticism.’ F. Edgerton, Introduction to Sabhā parvan, p. xxxiii.” Ibid., 7–8, n. 16. Mehendale misrepresents Edgerton both times. What Edgerton actually says is: “To try to make the text consistent on such points would be to enter the realm of ‘higher criticism.’ I agree with Sukthankar that our job is diferent.” Edgerton, “Introduction,” xxxiii (italics added). The reference is to Sukthankar, “Prolegomena,” lxxxvii. Sukthankar notes: “The above examples will show that the diaskeuasts did not always employ any great art—I may add, fortunately—in conlating two discrepant accounts of an incident, which is by no means an easy task. To resolve such anomalies, however, is beyond the scope of this edition, since the entire manuscript evidence unanimously supports the conlation, which is too old and deep-rooted to be treated by the ordinary principles of textual criticism. If we went about, at this stage of our work, athetizing such passages as were self-contradictory or as contradicted the data of some other part of the epic, there would not be much left of the Mahā bhā rata to edit in the end.” 126 A critical edition in the sense of an edition free of contradictions is neither achievable nor is it necessarily an improvement over the present text. Contrary to the expectation that a critical edition must provide a completely coherent text, it attempts, rather, to undo the emendations of scribes and editors. A  critical edition prefers diicult readings over easier ones (the diiciliores over faciliores) and generally works against the tendency to banalization. F. A. Wolf famously wrote: “when the witnesses require it, a true recension replaces attractive readings with less attractive ones. It takes of bandages and lays bare the sores. Finally, it cures not only manifest ills, as bad doctors do, but hidden ones too.” F. A. Wolf, Prolegomena to Homer, trans. with an introduction and notes by A. Grafton, G. W. Most and J. E. G. Zetzel (Princeton, NJ: Princeton University Press, 1985), 44. Sukthankar explicitly noted the advantages of the vulgate over the constituted text: “The place for resolving contradictions is interpretation, as, indeed, the work of glossators shows. The Vulgate text of the Mahā bhā rata is fairly readable and will appear in places, at irst sight, to be even ‘better’ than the critical text, because the former has been purged by the continuous emendations of scholars for centuries. A whole army of anonymous scholars and poets must have worked at the text to make it smooth and easy of comprehension and to increase its popularity and usefulness by adding to it interesting anecdotes, incorporating into it current and popular versions and explanations, bringing it in a line with the ethical, moral, religious and political ideas of essentially diferent ages.” Sukthankar, “Prolegomena,” ciii. 472 472 PHILOLOGY AND CRITICISM 127 This reference is essential. It comprises Venn’s main innovation over the earlier Euler dia- gram. Edwards notes: “When such pictorial representation [the Euler diagram] is extended to more complex sets of logical propositions, however, diiculties set in quickly. For example, a logician faced with a series of propositions might wish to ascertain whether they were mutually consistent, and Euler diagrams do not provide a graphical ‘algorithm’ for settling such a question. Venn’s own description of the impasse, in 1880, is conclusive: ‘… we cannot readily break up a complicated problem into successive steps which can be taken indepen- dently. We have, in fact, to solve the problem irst, by determining what are the actual mutual relations of the classes involved, and then to draw the circles representing this inal result; we cannot work step-by-step towards the conclusion by aid of our igures.’ Venn had a better idea.” A. W. F. Edwards, Cogwheels of the Mind: The Story of Venn Diagrams (Baltimore, MD, and London: Johns Hopkins University Press, 2004), 3–4. Venn himself explains his advance over Euler thus:  “Whereas the Eulerian plan endeavoured at once and directly to repre- sent propositions, or relations of class terms to one another, we shall ind it best to begin by representing only classes, and then proceed to modify these in some way to make them indi- cate what our propositions have to say. How, then, shall we represent all the subclasses which two or more class terms can produce? Bear in mind that what we have to indicate is the successive duplication of the number of subdivisions produced by the introduction of each successive term, and we shall see our way to a very important departure from the Eulerian conception. All that we have to do is to draw our igures, say circles, so that each successive one which we introduce shall intersect once, and once only, all the subdivisions already existing, and then we have what may be called a general framework indicating every possible combination producible by the given class terms.” John Venn, “On the Diagrammatic and Mechanical Representation of Propositions and Reasonings,” cited in Edwards, Cogwheels of the Mind, 4. There is no Venn diagram not based on a prior establishment or analysis of sets. This is the greatest problem with Brockington’s work: he does not conceptualize much less deine his sets. 128 Brockington, “Regions and Recensions, Scripts and Manuscripts.” Brockington does not provide a citation for the use of Venn diagrams in linguistics, but all the examples we exam- ined had two features in common: (1) they deined the members of the sets under consid- eration; (2)  they employed Venn diagrams to depict either the extent of shared features between two languages or to classify certain types of statements. Neither feature applies in Brockington’s case. 129 Brockington, “Textual Studies in Vā lmīki’s Rā mā yaṇa,” 203–4. The proposal was probably announced earlier at the 23rd International Congress for Asian and North African Studies in Hamburg, as the following abstract records: “Of the varied manuscripts examined, the most signiicant is ms. 14052 (unfortunately containing only half the Ayodhyā kā ṇḍa) for it is closely related to M4 and probably of similar age. Their closeness is marked by omissions— more exactly absence of material—rather than additions, indicating relative antiquity. However, omissions unique to each show that, despite their regular shared readings, neither could have been copied from the other, nor probably from an immediate ancestor. […] Such details suggest that the CE scheme of recensions and subrecensions is too limiting. I propose as a possible model the Venn diagram, which can better represent overlapping inluences and avoids a straight either/or classiication.” John Brockington, “Textual Studies in Vā lmīki’s Rā mā yaṇa,” Zeitschrift der Deutschen Morgenländischen Gesellschaft, Supplement IX, Proceedings of the XXXII International Congress for Asian and North African Studies, Hamburg, 25th–30th August 1986 (1992):  509–10. Brockington also refers to the Venn diagram in his latest published work:  “Elsewhere I  have proposed that a better model than either the Critical Edition’s system of script versions or a stemma codicum would be the Venn diagram, which with its overlapping circles or ovals can better indicate degrees of commonality between various manuscripts, as well as being capable of arrangement to a certain extent in a form reminiscent 473 APPENDICES 473 of geographical relationships.” John Brockington, “Is the Script Relevant? Further Evidence from a Nevā rī-Script Rā mā yaṇa Manuscript,” in Epic and Argument in Sanskrit Literary History. Essays in Honor of Robert P. Goldman, ed. Sheldon Pollock (Delhi: Manohar, 2010), 24. 130 Compare Brockington, “Regions and Recensions, Scripts and Manuscripts”:  “Within the northern recension, they align particularly with two manuscripts D1 and [D]2 which are usually assigned to the western subrecension within the northern recension and to a lesser extent with a third V1.” Compare also Bigger, Balarāma im Mahābhārata, 113 and 133, n. 88. For Grünendahl’s use of the term (sich anschließen), see Grünendahl, “Zur Klassiizierung von Mahā bhā rata-Handschriften,” 126: “Additionally, there are nine passages where, according to the list all the Ñ manuscripts feature the reading of γ, but, according to the critical appa- ratus, individual manuscripts align themselves with the reading of ν and the southern recension.” 131 The following is typical: “Within this same span of the irst thirty sargas, in addition to its unique omissions, the Trivandrum manuscript omits two other passages found in M4 but omitted by some other manuscripts; the second of these is interesting, since it shows the Trivandrum manuscript diverging from M4 on one of the few occasions when M4 goes with the S recension in these irst thirty sargas. After sarga 31 (its own sarga 34) M4 shows a greater tendency to side with the S recension, and so in general does the Trivandrum manuscript, but it still omits some S material occurring in M4 (such as 42.25cd + 1047*, and 1107*1) and includes some N material lacking in M4 (such as 1097*3–4 + 1098*1–2). Equally, though, it includes some S material absent from M4 (for example, 50.11 and 1200*), so it is not possible to simply say that it shows a greater tendency than M4 to retain the ainity with the N recen- sion. Indeed, there are two instances where it seems to give even clearer evidence than M4 of combining both N and S traditions. These are at the start of sarga 50, where it has both 1217* (read by many N mss and M4) and 1218* (read by S mss, including some Devanā garī mss), and at 54.16cd where it has both the text (representing the S version) and the N variant successively. Moreover, at 50.11, although its reading is basically that of the text (= S), it has a unique variant reading in the irst pā da which is similar to the N substitute (1197*1 pr.).” Brockington, “Textual Studies in Vā lmīki’s Rā mā yaṇa,” 202–3. 132 As West also notes: M. L. West, Textual Criticism and Editorial Technique (Stuttgart: B. G. Teubner, 1973), 14–15. 133 Brockington’s approach is peculiar. He writes: “In view of the large numbers of Rāmāyanạ manuscripts preserved in various collections, and in particular of their relative lateness (and therefore the impracticality of establishing a stemma codicum), the number of manuscripts used for the Critical Edition is really quite small. The largest number—41—was used for the Uttarakā ṇḍa and the smallest—29—was for the Ayodhyā kā ṇḍa and also for the Ā raṇya and Sundara kā ṇḍas, although admittedly a larger number was initially collated (e.g. 43 for Ayodhyā ). It is not surprising, therefore, that variant readings and new patterns of omission or addition can be found in virtually every new manuscript examined. I would argue that this material is signiicant if we are to build up a truer picture of the complex relationships between the diferent recensions and versions.” Brockington, “Textual Studies in Vā lmīki’s Rā mā yaṇa,” 197. Compare also John L.  Brockington, “A Malayā ḷam-Script Rā mā yaṇa Manuscript,” Indologica Taurinensia 21–22 (1995–96): 79: “However, the Critical Edition neces- sarily utilised a relatively small number of manuscripts (29 in the case of the Ayodhyā kā ṇḍa) and so variant readings and new patterns of omission or addition can be found in virtually every new manuscript examined.” The late manuscripts probably deserve classiication as inutiles. Even if they contain good readings, their lateness makes it unlikely they will afect the stemma. Brockington has not grasped the principle of eliminatio codicum descriptorum, which Trovato elucidates as follows: “In operative terms, if all the sources of a contaminated ms. or of a subfamily of contaminated mss. are preserved, there is no need to use the contam- inated ms. or subfamily, which can be eliminated as codices descripti, or rather inutiles.” The 47 474 PHILOLOGY AND CRITICISM sole exception to this rule is: “If, instead, a ms. that descends to some extent from a known exemplar contains high-quality variants not found in any known ms. or subfamily, we shall need to assume that ‘extra-stemmatic’ (Timpanaro) or rather ‘extra-archetypal’ contamina- tion has occurred.” Trovato, Everything You Always Wanted to Know about Lachmann’s Method, 134. But here also, rather than alter the stemmatic relationships of older family members, this is more likely to trigger a revision in our judgment about (the utility of) late and inferior specimens. Moreover, the circumstance that “new patterns of omission or addition can be found in virtually every new manuscript examined” does not negate the relationships already established, since they were not established using “patterns of omission or addition.” Possibly, what Brockington means is that some discarded manuscripts contained better readings due to their scribes’ access to good, old sources. But he does not argue from the lectio diicilior. He merely claims that several lower-value manuscripts exist. This is like saying we should not establish family trees for blue-bloods because we cannot establish the parentage of the bas- tard children of prostitutes. 134 Trovato, Everything You Always Wanted to Know about Lachmann’s Method, 80. 135 Brockington, “Regions and Recensions, Scripts and Manuscripts.” The reference is to Brockington, “Textual Studies in Vā lmīki’s Rā mā yaṇa,” speciically the claim that “to a certain extent this is a reversion to the stemma codicum in that it attempts to depict the diferent inluences or originals which have formed a speciic manuscript, while avoiding the impossible task of constructing the entire chain of transmission involved.” Ibid., 204. This is a novel interpretation. A stemma codicum would only “depict the diferent inluences or originals which have formed a speciic manuscript” if every manuscript had lines not only diverging from it but also converging on it, that is to say, if every manuscript was copied not from one exemplar, but several. Though not impossible, this is an unusual interpretation of the stemma, since genealogical reconstruction under these conditions is impossible. 136 If he wanted, an editor could draw dotted lines converging on manuscripts to depict the direction and extent of contamination (for example, by varying the lines’ thickness or aixing numbers). Belvalkar actually undertook this. But besides mapping the extent of contami- nation, his stemma did not provide new insights. It did not afect the relationships already established. From an editor’s perspective, showing “multiple allegiances” is less relevant than establishing the manuscripts’ descent as unambiguously as possible. The tradition may, of course, be so hopelessly contaminated that we cannot reconstruct an archetype. In this case another solution like editing the best manuscript must be pursued. But this is not Brockington’s contention. He does not recommend another editorial strategy. He defends the critical editions as “on the whole the best we have.” Brockington probably emphasizes the “multiple allegiances involved” because he has not collated the manuscripts himself and does not seek to provide an edition, critical or otherwise. 137 But see our proposed amendments to his igure later, for what we think Brockington is really trying to express. Here also the circles must intersect even though they have no common members. A Venn diagram must indicate all possible logical combinations of sets. 138 We did not ind a single example of complete collations in the articles we examined. To our knowledge, Brockington has never edited a manuscript. The closest he comes to providing collations is in Brockington, “A Malayā ḷam-Script Rā mā yaṇa Manuscript,” 77–91, but even this does not amount to a complete collation. Although he provides a detailed account of variants for the irst 20 sargas, he then focuses on “major features,” especially additions and omissions. The following passage is typical: “The Trivandrum manuscript agrees with D1.2 M4 in, for example, omitting—as well as 9.14ef, 18–19, 26, 43, 10.6cd, 21–25, 13.25, 18.19cd (not D2) and 40 (not D2)—20.23c–34d, App. I 12.35–39, 58, 80–81 and 24.16, in inserting 168*, 203* and 362* (also in B2.4), in reading 620*4 after 632*, in the sequence of stanzas at 25.4–14, and in very many common readings elsewhere. It is still more interesting to note 475 APPENDICES 475 the frequency with which it agrees with D1 or D1.2 only against M4. It agrees with D1 only but not M4 at 17.24b, 451*3, 452*5, the insertion of 469* (also in D5), 489*19–20 (but not 37–40), 508*1 and 570*2; it agrees with D1.2 but not M4 at 281*2, 295*3 (also in D5), 320*3, 14.18c, App. I 12.80, 623*2 and 658*1. These are certainly not enough to indicate a link between this manuscript and these two of the supposed Western recension manuscripts to the exclusion of M4, but it does suggest that D1.2 are aligned in some way with this alter- native Malayā ḷam recension of M4 and ms. 14052, which is still more obvious when we note the common readings, insertions and omissions of these four manuscripts. There are also traces of links with V1 either of all four manuscripts or of the two Malayā ḷam-script ones.” Ibid., 88–89. 139 See Brockington, The Sanskrit Epics, 103–11. He estimates that “the proportion [of formulaic expressions] for the Mahābhārata is at least as great as in the Rāmāyaṇa where around one-third of the ślokas contain signiicant formulaic material.” Ibid., 103. 140 Ibid., 105–6. 141 An earlier version of this section was presented at the DICSEP 8 conference. This version beneits from Brockington’s input on this paper. 142 Brockington, “Is the Script Relevant?” 18. 143 The attached note makes this explicit: “the point that script boundaries are less signiicant than assumed by the editors of the Critical Edition is one that is valid for the Mahābhārata also, as research being done in Cambridge under John Smith by a former student of mine is demonstrating by the rather diferent technique of applying cladistics (see Wendy J. Phillips- Rodriguez, ‘The Evolution of a Sanskrit Epic: Some Genetic Considerations about Scripts’, in The Evolution of Texts: Confronting Stemmatological and Genetical Methods, ed. C. Macé, P. Baret, A. Bozzi, L. Cignoni, Linguistica Computazionale 24–25 (IEPI, Pisa-Roma).” Ibid., 27–28, n. 4 (italics added). 144 We referred to the diagrams in Wendy Phillips-Rodriguez, C. J. Howe and H. F. Windram, “Chi-Squares and the Phenomenon of Change of Exemplar in the Dyū taparvan,” in Sanskrit Computational Linguistics, ed. Gérard Huet, Amba Kulkarni and Peter Scharf (Heidelberg: Springer, 2009), 383–86. 145 Reeve rightly notes that “scholars [who] believe that contamination and interpolation were so common that stemmatic method seldom or never works […] are biting of the hand that fed them, because unless one can actually see scribes contaminating and interpolating under one’s nose it is only by applying stemmatic method that one can detect contamination and interpolation.” Michael D. Reeve, “Stemmatic Method: ‘Qualcosa che non funziona?’ ” in The Role of the Book in Medieval Culture: Proceedings of the Oxford International Symposium, September 26–October 1, 1982, ed. P. Ganz, Bibliologia 3 (Turnhout: Brepols, 1986), 65. 146 This is also the opinion of no less an authority than West: “Of the value of other sorting methods, in particular statistical methods, I remain sceptical. A numerical table of signiicant agreements between every two manuscripts, as described on p. 38, will provide objective con- irmation of groupings suggested by casual inspection, and will indicate how clear-cut they are (e.g. how much more often GHIKL agree with each other than with other manuscripts); but simply collecting the evidence, without reducing it to igures, will probably have given a clear enough picture already. Indeed, where groups of three or more manuscripts are concerned it will have given a clearer one, for the information that A agrees with B 81 times, B with C 92 times, and A with C 79 times does not enable us to deduce that ABC all agree together 73 times, or even once.” West, Textual Criticism and Editorial Technique, 46. 147 John G.  Griith, “A Taxonomic Study of the Manuscript Tradition of Juvenal,” Museum Helveticum 25, no. 2 (1968): 101–38. 148 West, Textual Criticism and Editorial Technique, 46–47. 149 Griith provides the following, not entirely perspicuous explanation: “A dot is then placed in each square that corresponds to a coincidence of reading in each pair of mss. The agreement 476 476 PHILOLOGY AND CRITICISM of an isolated pair is recorded by a single dot in the appropriate square; three dots are needed for a trio of mss. in agreement, 6 for a quartet, and so on. Obviously an aberration or singu- larity on the part of an individual ms. has to be neglected, if only because there is nowhere to register it. (One might in such a case record the agreement of the other 15, but it is doubtful whether the extra labour would afect the result and isolated eccentricities are more likely to falsify the picture than to clarify it.) The only critical activity demanded of the recorder is that he should be alert to exclude from the count any non-signiicant similarities, such as lines lost by homoeoteleuton (such as I 87–88) or homoearchon (as at II 24–25). Spelling trivialities are also inadmissible, as in these details scribes may have unconsciously conformed to ingrained habit and so not reproduced faithfully what was before them. The recording process is not however completed by merely illing in the Squares wherever there is agreement between pairs of mss. and leaving it at that. Thus at V 116, PRA agree in spumat, as is shown by the three entries in distinctive hachuring (white bands running from ‘North-West’ to ‘South- East’) in diagram I below[.] But the agreement of the other 13 in the alternative fumat must be plotted in too, as has been done in the diagram by a rather diferent hachuring, with the white bands running from ‘South-West’ to ‘North-East.’ This second operation entails 78 recording acts, in addition to the three already made: in practice no distinction need be made between the agreement within a small and the agreement within a large group, which in dia- gram I has only been done for clarity of exposition. Obviously the least tedious distribution to record is a split of 8 mss. against 8, which requires only 28+28 or 56 recording acts in all. The maximum number occurs when 14 mss. agree against 2: this works out at 91+1 or 92 in all. Where mss. split into 3 groups, each group is treated separately, as at VI 322. There AFNPR agree in luctum, LO in frictum and the rest (BGHJKTUVZ) in fructum. This calls for 10+1+36 recording acts, or 47 in all, as is shown in diagram I. Reference to the diagram and its attendant key should dispel obscurity.” Griith, “A Taxonomic Study of the Manuscript Tradition of Juvenal,” 120–22. 150 “This enquiry sets out to see how far an unsophisticated but, I  hope, judicious use of taxonomical methods (4.22 f., below) may lead towards a rational sorting of those mss. which, being fully collated, represent for this purpose the ‘open’ or ‘horizontally interpolated’ ele- ment in the tradition.” Ibid., 102. 151 “I now turn to those other mss. which have been collated in suicient detail to make quan- titative comparison possible. There is an undoubted element of randomness inherent in the data, in as much as the choice of mss. for collation was made by diferent individuals long ago, and was, so far as can be seen, largely a haphazard one. Indeed much of the work of collation was undertaken before many of the possible candidates for scrutiny were known or even listed. It may well be that the choice was on the whole not a bad one. Whatever its merits or failings, it is only practical to make the most of the information at present available. Any improvement to the text of the author will thus be consequential, neither is this an efort to establish an order of merit of the witnesses being examined. This would be tantamount to reviving the meaningless concept of a ‘best ms.’ ” Ibid., 112–13. 152 “What I set out to do is to explore a promising method of achieving a meaningful calibration of the supericially disparate array of manuscript-characteristics in terms of ‘near-neighbour’ ainity, which is quite another matter. If an arrangement of clusters of mss. can be detected underneath the confusion of data before us, the path to the next stage of enquiry may be clearer than it would otherwise be. If the process is properly conducted, it should result in a scale of such near-neighbour relationships, with the most heavily interpolated documents coming together at one end of the scale and as far apart from the sincere witnesses as pos- sible. Clusters with less pronounced characteristics should appear in the middle section of the line, which can be thought of as similar to a spectrum-line in the ield of Physics or Chemistry.” Ibid., 113. 153 Ibid. 47 APPENDICES 477 154 “From what has just been said, it will be clear that one must cut oneself loose from the preconceptions and limitations of the conventional family-tree presentation, whose inade- quacy has become increasingly evident as the intricacies of such open traditions come to be better understood. A further advantage of the process is that the investigator is entirely igno- rant of the way the data he is analysing will work out until his counting is completed and he comes to the inal stage of the resolution of the igures he has arrived at for each section of the satires.” Ibid. 155 See ibid.: “In the conventional methods of evaluating mss. there is an inherent danger of circular argument, from which the taxonomic process is free.” 156 “I take the view that the return on the labour of further full-scale collation is unlikely to be justiied. The use of reliable sampling techniques to associate promising witnesses, as yet imperfectly collated or unknown, with well-established taxa is obviously a matter that will have to be faced at the next stage. It is obviously impracticable to re-calculate similarity- matrices de novo to take account of new material as it arises; fortunately other techniques (such as ‘split matching-scores’ and the like) may enable the classiication to be achieved with an acceptable economy of efort.” Ibid., 132. 157 “So long as one is dealing with compact arrays of data which show characteristics common to the whole group of specimens being studied, the accepted methods are generally ade- quate, in biological study as elsewhere. Such systems are now called ‘monothetic’ and their treatment and interpretation are mostly straightforward. They are obviously analogous to a ‘closed’ textual tradition, in which features such as lacunae of identical extent [are] common to a whole group of mss. They tell their own story. ‘Polythetic’ systems are another matter. Here characteristics are not universal to the group under investigation: thus there exist in nature birds without wings, mammals lacking red corpuscles and countless other ‘untidy’ phenomena to bedevil classiication. Concentration on single characteristics leads nowhere in such situations, and attention has to be directed to an aggregate of properties. Certainly, acute observation, aided by luck, may detect some striking features of resemblance, but discovery of this kind is at best fortuitous and at worst misleading, in that what has been observed may have come in only at a late stage in the evolutionary process, and so mean less than it might seem. Obviously, for the taxonomical approach to be valid, a substantial number of points of similarity between specimens or groups of specimens is essential. There seems general agreement that a minimum of 40 mutually unrelated elements should be taken for any single operation, but a igure of the order of 60–70 is safer. If therefore in a literary text a signiicant divergence occurs about once in every 4 or 5 lines, reliable results may be expected where stretches of not less than 300 lines have been treated; in most of what follows considerably longer blocks have been used.” Ibid., 114–15. 158 “One question has been deliberately left undiscussed. Has the taxonomic process efectively disentangled the underlying relations of the interpolated mss. and penetrated to the pattern of their ‘genetic histories’ latent beneath the over-burden of interpolation they now present to our observation? The degree of regularity of their behaviour as disclosed in this study makes this at least a tenable supposition. It has been put to me that the position should be stated more strongly: that it is incumbent on those who are sceptical of the method or of its results in Juvenal’s case to indicate the kind of explanation which in their view would cover both the supericial disarray of the data and the underlying regularity revealed by the analysis, certiied as this is by rigorous and accepted criteria of statistical probability.” Ibid., 134. 159 Ibid., 119. 160 West, Textual Criticism and Editorial Technique, 47. 161 Unlike Griith, who applies a statistical approach only to the manuscripts, Brockington also applies it to the constituted text of both epics itself. On one hand, he lists interpolations and “omissions” in manuscripts, and uses this to undermine the critical editions’ claims 478 478 PHILOLOGY AND CRITICISM to authority. On the other, he lists repeated words or verses in their constituted texts, and uses this to argue for the works’ oral origins. Pollock notes: “In the most recent discussion Brockington (1984) stubbornly reairms Jacobi’s view:  ‘Far from being a Vaiṣṇava epic, Vā lmīki’s Rā mā yaṇa is no religious epic at all. It is lamentable that misunderstanding of this point […] should still persist so long after Jacobi’s explicit declaration’ (p. 13). Brockington employs an elaborate ive-stage scheme of text evolution to demonstrate the developing con- ception of the hero (from one who is ‘thoroughly human’ to a god, pp. 218–25).” Sheldon I.  Pollock, “Introduction,” in The Rāmāyaṇa of Vālmı̄ki:  An Epic of Ancient India, ed. Robert P. Goldman, vol. III: Araṇyakāṇḍa, trans. Sheldon I. Pollock (Delhi: Motilal Banarsidass, 2007), 17, n. 21. For Pollock’s rejection of the thesis of a “ ‘heroic epic’ [transformed] according to a later theological program,” see ibid., 18–19. The reference is to John Brockington, Righteous Rāma: The Evolution of an Epic (New York: Oxford University Press, 1984). 162 Trovato’s judgment is apt and devastating: “More generally, as many of the considerations discussed above suggest, one gains the clear impression that almost all the debatable elements discussed so far, and possibly even the search for a computer based alternative, arise from inadequate knowledge of traditional philological theory and methods.” Trovato, Everything You Always Wanted to Know about Lachmann’s Method, 199. 163 For an insight into the sense of humility and service that motivated the Mahā bhā rata editors, see Belvalkar, “Introduction [to the Bhīṣmaparvan],” cxxxii (“The edition of the Bhīṣmaparvan, completed by me after several years’ labour, is now ofered, as a sort of vāṅmaya-śrāddha, to the revered memory of Ramakrishna Gopal Bhandarkar on the occasion of his twenty-second death-anniversary. May it be regarded as a tribute worthy of that great savant, and may it also be, in accordance with BG. 18.70, acceptable to Lord Kṛṣṇa as a proper iṣtị of the Jñānayayajña that is to redound to His Glory!”) and Sukthankar, “Introduction,” xxxviii (“If Mahā rṣi Kṛṣṇa Dvaipā yana Vyā sa tells us that he has cried himself hoarse urging people to follow the Path of Duty:  ūrdhvabāhurviraumyeṣa na ca kaś cicchṛṇoti mām | dharmādarthaś ca kāmaśca sa kimarthe na sevyate ||, his shouting with uplifted arms has not been entirely in vain. He has not failed in his mission. Across the reverber- ating corridors of Time, we his descendants can still hear dimly his clarion call to Duty. It is in response to that call and in a spirit of reverent homage to that sage of unfath- omable wisdom—that embodied Voice of the Collective Unconscious of the Indian people—[that]  we ofer this work, pledged to broadcast to mankind, in this hour of its need and its peril, the luminous message of the Maharṣi: na jātu kāmānna bhayānna lobhād; dharme tyajejjıv̄ itasyāpi hetoḥ | nityo dharmaḥ sukhaduḥ khe tvanitye; jı ̄vo nityo heturasya tvanityaḥ ||”) (all italics Sukthankar’s). See also V.  S. Sukthankar’s statement in “Bhandarkar Oriental Research Institute, Poona Silver Jubilee Celebrations 4th and 5th of January 1943. Roger Lumley, N.  C. Kelkar, J.  R. Gharpure, R.  N. Dandekar, C.  Kunhan Raja, V.  M. Apte, P.  V. Bapat, A.  P. Karmarkar, B.  R. Kulkarni, H.  D. Sankalia, C.  R. Sankaran, S.  R. Shende, Ludwik Sternbach,” Annals of the Bhandarkar Oriental Research Institute 24, no.  1/2 (1943): xlvii: “There is a danger that in our pseudo-scientiic mood, we may be tempted to discard this great book, thinking that we have out-grown it. That would be a capital blunder! That would in fact mean nothing but an indication of our will to commit suicide, national suicide, the signal of our national extinction. For never was truer word spoken than when the late German Indologist Hermann Oldenberg said that ‘in the Mahā bhā rata breathes the united soul of India, and the individual souls of her people.’ And why is that? Because the Mahā bhā rata is the national saga of India. It is, in other words, the content of our collective unconscious. And just for that reason it refuses to be discarded. We must therefore grasp this great book with both hands and face it squarely. Then we shall recognize that it is our past which has prolonged itself into the present. We are it: I mean the real WE! Shall we be guilty of strangling our own soul? NEVER!” (italics and capitalization Sukthankar’s). Penned on January 5, 1943, 16 days before he died, they are the last words Sukthankar wrote. 479 APPENDICES 479 164 P. L. Vaidya, the last general editor, recollects: “The late Dr. Sukthankar, the irst General Editor, was paid a salary of Rs. 500 per month for a period of nearly 18 years. After his death in 1943, the Institute found that it could pay to the General Editor not more than Rs. 250 a month. When, therefore, a scholar of the eminence of Dr. Belvalkar, who had held a post in the Indian Education Service, was appointed General Editor, we could not ofer him more than Rs. 250 per month as salary. In order to make this paltry sum look digniied, we started referring to his ‘salary’ as ‘honorarium of Rs. 3,000 per annum.’ Out of his 18 years of General Editorship, Dr. Belvalkar received this honorarium for 10 years only, and, although the cost of living was steadily rising all those years, the amount of his honorarium of Rs. 3,000 per annum actually happened to dwindle down to a zero. The Parvan-editors were paid at the rate of Rs. 5/-per printed page of demi quarto size, no doubt, but this rate, in my opinion as well as in actual experience, worked out to be not more than one rupee a day. The departmental staf was paid a salary every month, but, according to Government Audit Party visiting the Institute annually, it was far below the standard.” “The Completion of the Critical Edition of the Mahā bhā rata,” Annals of the Bhandarkar Oriental Research Institute 47, no. 1/4 (1966): iii–iv (italics in original). 165 This also applies to Grünendahl, who, after casting suspicion on Sukthankar in “Zur Klassiizierung von Mahā bhā rata-Handschriften,” states that “a more detailed evaluation of the Nepali manuscripts of the Mahā bhā rata in general and of the Ā diparvan manuscript Ñ4 in particular must be reserved for a work currently in preparation.” Ibid., 129–30. Twenty- ive years later this work is yet to appear (despite being told “the conditions for this [task] are provided by the work of the Nepal-German Manuscript Preservation Project”; ibid., 130). In fact, the sole collations we possess for this manuscript are Sukthankar’s, provided in 1939 in Vishnu S. Sukthankar, “Epic Studies VII: The Oldest Extant Ms. of the Ā diparvan,” Annals of the Bhandarkar Oriental Institute 19 (1938–39): 201–62. Grünendahl himself relies on them, though, like Brockington, he neglects readings for interpolations. Not one Mahā bhā rata critic has addressed himself to the task Sukthankar set when he wrote: “In the light of such variants, some readings of the Critical Edition will have to be reconsidered, and it may, in sporadic cases, be found necessary even to alter slightly the readings of the Ā diparvan in the Critical Edition. That work must however be left over now for further research by a future generation of scholars.” Ibid., 209. 166 To clarify a pervasive misunderstanding, it is not that we think oral transmission is completely excluded: we think scholars have not made the case for it. Their attempts to reinstate the oral hypothesis after the critical edition are circular and question-begging. The orality hypothesis frequently conceals real deicits in their knowledge of textual criticism, to say nothing of the fact that they can more easily make claims about a nonexistent text since these claims are, in a literal sense, unveriiable. As Sullivan observes: “As [Kevin] McGrath writes in his most recent volume (McGrath 2013: 10), ‘In this book I examine what I envision to be an earlier and thoroughly heroic status of the poem when it still existed in a preliterate form, examining the narrative as it concerns Kṛṣṇa as he exists in that hypothetically “earlier” telling.’ In fact, we do not have such a text, so McGrath is examining the text we have while omitting portions he chooses to ignore, namely, parts he regards as didactic, concerning ‘ediication’ (McGrath 2004: 5).” Bruce M. Sullivan, “An Overview of Mahā bhā rata Scholarship: A Perspective on the State of the Field,” Religion Compass 10, no. 7 (2016): 170. 167 See Vishwa Adluri and Joydeep Bagchee, The Nay Science:  A History of German Indology (New York: Oxford University Press), 153: “In spite of Fitzgerald’s and McGrath’s eforts, however, the Holtzmannian Ur-epic remained elusive. As far back as they pushed this ideal epic—and in his most recent iteration McGrath has advanced a date of the Bronze Age— they were still unable to ind evidence of it and for the simple reason that its origins lay nei- ther in the fourth century CE nor in the fourth century BCE nor even in the twelfth century BCE, but in the nineteenth century CE when the thesis was irst proposed by Lassen. The 480 480 PHILOLOGY AND CRITICISM idea of an original epic, as we have seen, was a speciically German notion answering to spe- ciically German needs. German scholars had deployed the idea in pursuit of an ideal of a heroic Aryan race. It permitted them to make polemical points against Roman Catholicism as being essentially alien to the spirit of the German people. It permitted them to identify deining traits of the ‘Indo-Germanic’ or ‘Aryan’ peoples—nobility, free-spiritedness, suspi- cion of priestly authority, a warrior culture, and so on—in contradistinction to the ‘Semitic’ tribes. It permitted them to undertake a sustained critique of non-Germanic cultures as well as to issue stern warnings of what might befall the German nation if it permitted itself to be overrun by these cultures. Beyond these ends, the theory of a heroic epic had no relevance to Mahā bhā rata criticism.” 168 Leonardi expresses this sense of responsibility well: “The concept of a critical edition [also] implies a responsibility to propose a text that, while meeting the requirements of science and elucidating the manuscript tradition, is not merely accessible to specialists, but recovers a work of the past for a contemporary public to read; one that does not reproduce a document, but interprets the tradition as a whole as a means to transfer its textual reality into something that is readable today.” Lino Leonardi, “Il testo come ipotesi (critica del manoscritto-base),” Medioevo romanzo 35 (2011): 6 (Trovato’s translation). Indology’s greatest law is that it lacks a similar sense of responsibility, to say nothing of readers to which it could feel responsibility. 169 Friedrich Nietzsche, Twilight of the Idols, trans. Richard Polt (Indianapolis and Cambridge: Hackett Publishing Company, 1997), 65–66 (italics in original). 481 GLOSSARY ad fontes. Latin for “[back] to the sources,” the term refers to the belief or the neces- sity that one must consult the original manuscripts rather than rely on medieval or Renaissance copies. Generally attributed to Erasmus of Rotterdam (1466– 1536):  “Sed in primis ad fontes ipsos properandum, id est graecos et antiquos” (“Above all, one must hasten to the sources themselves, that is, to the Greeks and ancients”; De ratione studii ac legendi interpretandique auctores, 1511). analytic and synthetic. The two main schools of Mahā bhā rata criticism. The ana- lytic school holds that the epic is a work of composite origins. Accordingly, the critic’s task is to distinguish its several sources or layers. In contrast, the synthetic school holds that the Mahā bhā rata is a unity, the product of a conscious creation. The critic’s task thus is to understand the text or the author’s message. Joseph Dahlmann (1861–1930) introduced the terms in his Genesis des Mahābhārata (1899). ancestor. The source from which one or more manuscripts, whether extant or nonex- tant, were copied; occasionally reserved for lost sources alone. anti-Brahmanism. The view that the Brahmans seized control of an original Kṣatriya epic and altered it for apologetic reasons. Anti-Judaic in intent (its origins lie in the German Protestant suspicion of priestly authority), anti-Brahmanism has been a central principle of German Mahā bhā rata criticism, since it “explained” why the epic, though attributed to a heroic warrior culture, appears as a work of theology, law and ritual. apograph. Literally meaning “written or copied from” (from Greek apo-, “from,” and graphō, “to write”), a copy of a work made from an older exemplar. Contrasted with the autograph, the work from the author’s hand (or someone he commissioned to write for him). apparatus criticus/critical apparatus. A list of variant readings and their manuscript sources placed beneath the constituted text, the critical apparatus provides the information required to evaluate the latter (for example, the manuscripts it is based on and the stage of the tradition it reconstructs). A critical apparatus may some- times also include other editorial annotations such as the manuscripts available and their state or peculiarities of the tradition. arbre réel/complet. French for “real tree,” the real or historical tree that depicts the rela- tionship of all the exemplars ever produced as it must have existed as opposed to the stemma codicum, which is reconstructed on the basis of the available information and is thus always a partial representation of the historical reality. 482 482 PHILOLOGY AND CRITICISM archetypus/archetype. The latest common ancestor of the manuscripts examined for an edition. An archetype may be extant or nonextant. In the latter case, it may be possible to reconstruct it using the available manuscripts. assimilation/harmonization. The incorporation of passages from a second source into one’s copy with the intent of achieving greater consistency, assimilation is responsible for the production of composite manuscripts. autograph. The authorized, original copy of a work, produced by the author himself or someone he commissioned to write. Occasionally multiple autographs may exist, if the author signiicantly revised his work in the time after the irst edition was produced. banalization. Replacement of a word by a more familiar or easier form. Bédierism. The view that editors should edit the text of the best manuscript (also called the bon manuscrit), correcting obvious errors, but otherwise not attempting to restore an earlier state of the text (for example, through genealogical reconstruction). Bédierism is directed against Lachmannism with its belief in restoration of the text of the archetype. The name comes from French critic Joseph Bédier (1864– 1938), who after producing a Lachmannian edition, turned against the method after he observed that most stemmata had two branches and suspected editors were attempting to increase the scope for subjective iudicium or subjective choice. Bipartite (also: biid) stemma. A stemma having two branches. Contrasted with tri- partite, quadripartite, pentapartite and so on. A bipartite stemma poses problems for the genealogical-reconstructive method, since no stemmatic justiication exists for choosing the reading of one branch over the other. canon. Any collection of authoritative books, usually applied to scripture. From the Greek kanon, meaning “measuring stick” and hence, by extension, a “rule” or a “standard.” closed tradition/recension. A tradition or recension free of contamination in which no manuscript is copied from more than one source. All variants therefore can be accounted for within that tradition or recension (that is, as either inherited from an ancestor or an innovation within the particular exemplar). codex (pl.: codices). In the ancient Near East and Greece, texts were written on long, continuous scrolls of paper or parchment. Around the late irst century CE, the codex, a bound book with cut pages, appeared and largely replaced the scroll by the fourth century CE. The distinction is irrelevant in the Indian context, as the use of scrolls died out in the third century CE and all our Mahā bhā rata manuscripts are codices (either in book format or the Indian long format known as pothı̄), written on palm leaves or birch-bark pages. codex descriptus (pl.:  codices descripti). A codex copied from an extant witness. As the codex contains additional errors over its source and ofers no new information for the reconstruction of the archetype, it should be discarded from the stemma. codex inutilis (pl.:  codices inutiles). A codex discarded as of lesser or negligible value for the reconstruction. Sebastiano Timpanaro introduced the term in his article “ ‘Recentiores’ e ‘deteriores,’ ‘codices descripti’ e ‘codices inutiles’ ” in 1985. He distinguished it from the descriptus as follows: “Thus, if a witness that is certainly a descriptus can contribute to the tradition with a variant acquired by contamination, 483 GLOSSARY 483 it can happen, on the other hand—and it does happen very often—the opposite case of a witness that is not necessarily a descriptus, which nonetheless does not contribute any ‘relevant’ reading and this is so not because it is not true or corrupt but simply because it does not carry any previously unattested reading.” codex unicus. The sole manuscript of a work in existence at a given moment in time. The archetype of a tradition reconstructed on the basis of extant manuscripts is the sole witness for that stage of the tradition, even though it may not have been the only exemplar in existence at that time. codicological/manuscript evidence. Codicology is the study of the physical aspects of the book (materials, paper, inks, techniques of binding, etc.). Codicological evidence, sometimes also called manuscript evidence, thus refers to the physical evidence. collation/collation text. Collation (from the Latin collatio, meaning “to compare”) refers to the comparison of witnesses.  Typically, one manuscript (usually an authoritative exemplar) is chosen as a reference manuscript (collation text) and the irst verse of the irst manuscript is compared with its irst verse and so on for the remaining verses. The procedure is then repeated for the remaining manuscripts. colophon. Information found at the end of a work, a book (for a work in several books), a chapter or a section thereof, the colophon provides information on the title or subject matter, the chapter number or book count and sometimes the scribe or date and place of production. The following is an example of a Mahā bhā rata colophon (from Mahā bhā rata 1.1): iti śrı̄mahābhārate ādiparvāṇi prathamo’dhyāyaḥ || 1 || samāptamanukramaṇı ̄parva ||. common-error method. The genealogical-reconstructive method is also called the common-error method because it is based on the insight that two manuscripts are related only if they have at least one signiicant error in common. Agreement in the correct reading or in the truth, in contrast, proves nothing for their relation- ship because the correct reading could have been passed down to them from any branch of the tradition, in contrast to the error, which is unique to the family. composite edition. An edition produced through the combination of readings from diferent sources. An edition may be accidentally composite, as is the case with manuscripts copied over time, or intentionally composite, as is the case with the Lachmannian edition, which is not based on one manuscript but takes the evi- dence of several representative exemplars into consideration. Nonetheless the Lachmannian edition is composite to a lesser degree than a historical manuscript, as it attempts to undo the errors and alterations of centuries and arrive at a text as close as possible to the author’s hand. concordance. An index of the initial words or phrases of all the verses of a work, with a reference permitting the identiication of the location of any verse in the work. The two main Mahā bhā rata concordances—Jacobi’s and Sørensen’s— only permit the tallying of the Bombay and Calcutta editions against each other (Sørensen’s also against the P. C. Roy translation). conlation/hybridization. The combination of readings from two diferent sources, creating a new reading diferent from either. Conlation is similar to assimilation, but applied at a more granular level. 48 484 PHILOLOGY AND CRITICISM conjunctive error/errore conjunctivo. A signiicant error shared between two manuscripts, which serves to establish that the two manuscripts are related. For instance, if both A and B contain an error that is such that either they or their respective ancestors could not have made it independently, then either A got it from B or B from A or both got it from a common source. constitutio textus/constituted text. The text the editor reprints as the (hypothetical) text of the archetype. The constituted text is printed in the main body of the edition, with variants the editor rejected appearing beneath. contamination. The inclusion of readings from memory from one source, while actually copying from a diferent source. Contamination is sometimes used inter- changeably with conlation, but the former is accidental, whereas the latter implies a conscious activity (that is, a hermeneutic or critical evaluation of the variants available to the scribe). corruption. A change in a manuscript’s reading due either to scribal error or to contamination. Corruptions are the only truly probative evidence for establishing iliation. Corruptions may be contrasted with emendations, which are intentional changes to the text’s reading. critical text. The reconstructed text based on a systematic evaluation of variants. The critical text may be based on the best manuscript or manuscripts or undertake a genealogical reconstruction of the archetype. critique génétique/genetic criticism. An approach less interested in the inished, static work than in the process of evolution (writing, drafting, editing) that precedes and leads to the inished work. Genetic criticism treats works less as inalized, bounded units and more as stages of concretion within a dynamic, creative movement. crux. From the Latin for “cross,” a crux arises when we have two equally valid alter- native readings, either of which could be the reading of the source. A crux typ- ically arises in a bipartite stemma, as it presents no stemmatic way of choosing between them. descendant. A manuscript copied from another. It may be possible to establish a manuscript’s descent using the genealogical-reconstructive method. deteriores. The inferior manuscripts, contrasted with the meliores. dittography. Greek for “written twice,” dittography refers to the accidental repetition of a letter, word or line. The opposite of haplography, in which a semantic unit is omitted through oversight. divinatio. A conjectural emendation, usually proposed when none of the available alternatives makes sense and a third reading both makes sense and explains the attested readings. Also sometimes used to designate the last stage of producing a critical edition, when following the recensio and examinatio, the editor attempts to restore damaged or unattainable parts of the text through emendation. ecdotics. Sometimes used as a synonym for textual criticism, ecdotics encompasses a wider ield that includes not just traditional critical editing of texts but all aspects concerned with the reading, production and transmission of texts (that is, a textual culture in the widest sense). 485 GLOSSARY 485 editio princeps (pl.:  editiones principes). Latin for “irst edition,” editio princeps refers to the irst printed edition of a text (previous exemplars all being handwritten manuscripts). The editio princeps of the Mahā bhā rata is the Calcutta edition the Education Committee/Baptist Mission’s Press published in four volumes between 1834 and 1839. Its text is based on Nīlakaṇṭha’s version. eliminatio. In the attempt to unify the manuscript branches, the elimination of variant readings (or sometimes of entire manuscripts) is the most important step. Eliminatio is of two types: eliminatio codicum descriptorum (elimination of derivative manuscripts) or eliminatio lectionum singularium (elimination of unique readings). eliminatio codicum descriptorum. This type of eliminatio removes entire manuscripts, the so-called codices descripti, from the stemma. If a manuscript derives from another extant manuscript, it contains no information useful for the reconstruction and hence the editor need not consider it further. eliminatio lectionum singularium. Since readings shared between independent families of manuscripts are more likely to preserve the reading of the archetype (unless there was simultaneous innovation or contamination between the families), readings unique to one branch of the manuscript tradition (or to one manuscript) can often be assumed to be corruptions introduced during copying. In reconstructing the archetype, the editor thus inevitably eliminates these readings in favor of the consensus of independent witnesses. emendatio ope ingenii and emendatio ope codicum. The two types of emendation practiced, emendation with the help of native wit and with the help of manuscripts, refer to the preference for subjective judgment and for the testimony of the manuscripts. Emendatio ope ingenii is now often disparaged. emendation. Conscious changes made to the text, either to correct something the scribe thought incorrect or could not construe or to improve upon the text. epigraphy. The study of inscriptions, including their writing systems, decipherment and historical classiication. errors, innovations or secondary readings. An error in textual criticism does not neces- sarily refer to a mistake. It simply means a non-original reading. Hence, sometimes the word innovation or the term secondary reading is preferred. See also: latent error. exegesis. From the Greek exēgeisthai, meaning “to explain, interpret” and, literally, “to lead out” (from ex- + hēgeisthai), exegesis is the practice of drawing out the meaning of the text through careful study. family. Two or more manuscripts related by descent. The grouping of manuscripts into families is the irst step in genealogical reconstruction. A  family is distin- guished from a group or manuscripts having similar features that are not related in a genealogical sense. iliation. The familial or genealogical relationship between manuscripts or entire branches of manuscripts, typically illustrated by means of a stemma codicum. genealogical analysis/reconstruction. The analysis of the relations of iliation between manuscripts or groups of manuscripts, typically carried out using the common-error method. genealogical-reconstructive method/edition. Also known as the common-error method and Lachmann’s method, the genealogical-reconstructive method is based 486 486 PHILOLOGY AND CRITICISM on the insight that two manuscripts must be related if they have one or more signif- icant errors in common. The presence of such errors can be used to establish the genealogical relationship between manuscripts, which can be expressed in the form of a stemma and be used to reconstruct the reading of their common ancestor (if lost). The genealogical-reconstructive edition, also known as the Lachmannian edition, is an edition based on the genealogical-reconstructive method. haplography. The accidental omission of a letter, word or line due to its similarity or identity with a preceding one. When the haplography involves a larger section of text, it is common to speak of a saut du même au même. hermeneutics. Hermeneutics, from the Greek hermeneuein, meaning “to express, to explain, to translate, to interpret,” signiies the art of interpretation. It designates the rules and techniques for elucidating the text. Hermeneutics is sometimes contrasted with exegesis and is understood as establishing or making sense of the bare meaning of a text. hiatus. Both Greek and Sanskrit avoid hiatus, a break in a sentence, through modiications to either the end of the word preceding the hiatus or the beginning of the word succeeding it (crasis or contraction of two vowels or diphthongs into a new vowel or diphthong, elision of the initial vowel of the following word, changes in word order and, in Sanskrit, other types of vocalic changes). Epic Sanskrit is lexible regarding hiatus, but later scribes appear to have strenuously avoided it. The critical edition restores many hiatuses. higher criticism/internal or inner criticism. Sometimes contrasted with lower criticism, which culminates in the reconstruction of the archetype, higher criticism attempts to restore still earlier forms of the text using specious criteria such as con- tent, ideology, assumed form and metrical and stylistic criteria and the like. historical criticism/historical-critical method. As a historical movement and as a method, historical criticism is the view that a text can only be understood out of its historical context. Practically, it expresses itself as a skepticism about texts’ unity and literary context and their received meaning. Historical criticism seeks to ascer- tain the realia, the historical and social conditions obtaining at the time of the text’s composition, which it considers the key to the text’s “true” (that is, literal) meaning. With its emphasis on the literal and historical dimension, historical criticism sets forth central principles of Protestant scriptural hermeneutics. horizontal and vertical transmission. Pasquali introduced the terms in 1952 to distin- guish a tradition in which readings are transmitted from a source to its descendant without deviation (vertical transmission) from one in which a witness reproduces the characteristics of a diferent branch of the stemma, suggesting that some of its readings are inherited from a manuscript family other than its own (horizontal trans- mission). A tradition may comprise a mixture of vertical and horizontal transmission. hyparchetype. A hyparchetype or a sub-archetype is a source in the stemma, either extant or nonextant, from which a family of manuscripts is descended. In the latter case, it may be possible to reconstruct it. Hyparchetypes may be represented as intermediary nodes in the stemma (occurring beneath the archetype, which occupies the stemma’s vertex) from which lines branch. 487 GLOSSARY 487 indicative error (see: signiicant error). interpolation. Verses or passages inserted into a manuscript, either from another source or possibly composed by the scribe, thus altering the original text. Interpolations can only be securely detected through the comparison of man- uscript exemplars. Lachmann’s method/Lachmannism/Lachmannian edition. Widely attributed to Lachmann, though he had several important predecessors and was inconsis- tent in applying the method, Lachmann’s method reconstructs the reading of the hypothetical archetype, using the common errors between manuscripts to establish their genealogical relationships or iliation. Lachmannism endeavors to reconstruct a presumed historical archetype and is opposed to Bédierism, which rejects this possibility. The Lachmannian edition is also known as the genealogical-reconstructive edition and is a composite edition based on the consensus of independent manuscripts or witnesses. lacuna (pl.: lacunae). Physical damage to a manuscript caused by careless handling, improper storage, weathering or vermin that has left a gap or a fragment missing in the text. latent error. A latent error is a reading whose non-original nature is not immediately apparent due to its unexceptionable character. A latent error can only be identiied using the stemmatic method. It is contrasted with the manifest error, which reveals itself as a mistake on reading. layers/strata. Associated more with biblical than with textual criticism, the term refers to materials of difering antiquity and origin found within the same text. Layers are typically identiied using internal criteria and stylistic or linguistic changes to the text in conjunction with what is known about the historical conditions at the time, but often face the charge of circularity. lectio brahmanica and lectio heroica. The Brahmanic reading and the heroic reading. Not terms in textual criticism, but useful to understand the Indologists’ work. For almost all Indologists, a good reading is identical with the lectio heroica (regardless of its attestation or state of preservation) and their philology consists of a single prin- ciple: lectio heroicior praeferenda est (the more heroic reading is preferable). By contrast, the lectio brahmanica must be discarded as by deinition a corruption (and between two competing versions, the lectio brahmanicior, the more Brahmanic version, must be discarded). lectio brevior/brevis potior. The principle, associated with J.  J. Wettstein (1693–1754), that when two versions of a text or a passage are found, the shorter one is prefer- able. The underlying idea is that ancient scribes are more likely to add details (by making implicit ideas explicit, by adding glosses or additional lines or, when faced with alternate versions, including both in order to preserve the putative original) than to omit them. Given this determination to transmit the text intact, texts tend to expand over time. The editor interested in restoring the original will thus regard the longer version with justiied skepticism. lectio diicilior probabilior/praeferenda est. The principle, irst articulated by Jean Le Clerc (1657–1736), that of two readings, the more diicult or the more obscure one is 48 488 PHILOLOGY AND CRITICISM likely the original. The principle is the inverse of the well-known phenomenon of banalization, that is, that a scribe was more likely to replace a more diicult reading, either consciously or unconsciously, with an easier one. lectio facilior/clarior (pl.: lectiones faciliores). The lectio facilior is the simpler reading (com- pare banalization) and is therefore most often rejected for the lectio diicilior in reconstructing the reading of the ancestor or archetype. lectio singularis (pl.: lectiones singulares). Literally, “singular reading.” The lectio singularis is a unique reading found in a single witness, that is to say, a reading lacking support from the other witnesses. It is therefore usually assumed to be the result of scribal error and hence can be eliminated. Sometimes, however, the lectio singularis can be the correct reading, especially if the manuscript containing it represents an inde- pendent branch of the tradition and all the other witnesses innovated. lower criticism. Sometimes encountered as a synonym for textual criticism, lower criti- cism is contrasted with higher criticism. meliores. The better manuscripts, the term referring to the quality of the text found in the manuscripts and not their physical condition. negative apparatus. An apparatus that only lists the manuscripts that contain readings difering from the constituted text and not the manuscripts on which the consti- tuted text is based. A negative apparatus is less useful than a positive apparatus. In the latter, the manuscripts on which the constituted text is based are typically listed to the left and followed by a single square bracket (]) and thereafter by the other manuscripts with their respective variants. The Mahā bhā rata critical edition has a negative apparatus. Neo-Lachmannism/trans-Lachmannism/Italian school. A perspective associated with Giorgio Pasquali (1885–1952), though others such as Cesare Segre, Michel Barbi, Gianfranco Contini and Silvio Avalle also made important contributions. Neo-Lachmannism refers to a Lachmannian approach tempered with investigations into the history of the text, that is to say, it attempts to undergird the abstract and mathematical nature of Lachmannian reconstruction with positive knowledge of the text, its transmission and the scribes’ practices. Sometimes also referred to as the Italian school in view of Italian scholars’ signiicant theoretical contributions. open tradition/recension. A tradition or recension in which some exemplars in the stemma have lines converging on them, so called because there has been contami- nation from exemplars outside of the tradition or recension modeled in the stemma. original. The text of the author’s hand or the text as authorized for publication by the author, the originals of ancient texts have not been preserved in all but a handful of cases. What we have in some cases are the archetypes, later exemplars from which all extant manuscripts are descended. Very rarely is the original identical with the archetype. paleography. The study of ancient writing, including its development from one script to another. Paleography sometimes overlaps with epigraphy, the study of ancient inscriptions, though it actually refers only to the writing system. palimpsest. From Greek palimpsēstos, meaning “scraped again,” a palimpsest is a manuscript that has been reused after erasing the original text. The original text 489 GLOSSARY 489 may sometimes be detectable, especially using infrared or ultraviolet photography, leading to the wider use of the term for something that has been altered and still bears traces of that alteration. polygenesis of errors. If an error is found in two branches of the tradition neither having gotten it from the other or from a common ancestor, it is common to speak of polygenesis or the development from more than one source of the error. Polygenesis is typically restricted to common orthographic confusions or other kinds of errors two scribes could have committed independently. recensio/systematic recension. From Latin recenseo, meaning “to examine, review, or survey.” A systematic recensio is a comprehensive survey of the available manuscripts of a work conducted prior to creating an edition. Occasionally, recensio is distin- guished from examinatio, even though it also entails an evaluation of manuscripts. recension. A recension strictly refers to an editorial revision of a text, but it is more widely applied to signiicantly diferent versions of a work. Recensions of a work may arise due to conscious changes to the source, but they may also be the product of the reinforcement of scribal corruption, emendation, interpolation and even conlation over centuries or millennia. recentiores. The newer manuscripts, typically applied to Latin and Greek manuscripts from the mid-twelfth century onward. redactor. From Latin redigere, meaning “to edit,” a redactor is someone who edits, arranges and otherwise revises existing materials into a new literary composition. saut du même au même. French for “leap from the same to the same,” saut du même au même refers to a copyist’s accidental skipping of a line. Its occurrence is facilitated by verbal similarities at the beginning or the end of the line (known as homoearchon and homoeoteleuton, respectively). Schichtanalyse/Textenschichtung. The analysis or separation of a work into its presumed constituent layers or texts. Favored by German Indologists, these terms ind no mention in handbooks of textual criticism. Schriftartprämisse. The mistaken view that the Mahā bhā rata manuscripts were classiied by script. scripts. The writing system employed for a work, the script is a system of marks or signs and therefore distinct from the text. separative error/errore separativo. A term introduced by Paul Maas (1880–1964), the separative error (German: Trennfehler) is a signiicant error that serves to establish that one manuscript is not the source of another. For instance, if A has a signii- cant error not found in B, then A cannot be the source of B (presuming the error is not such as could have been corrected by a scribe through conjecture), though B might still be the source of A (to exclude this latter possibility, we must identify a separative error of B versus A). signiicant error/errore signiicativo. The concept of a signiicant error (also known as the indicative error) refers to the fact that, in order to establish iliation using the common-error method, the errors under consideration cannot be simple spelling mistakes or orthographic confusions (that either scribe could commit indepen- dently) or be attributable to regional and idiomatic variations. In other words, the 490 490 PHILOLOGY AND CRITICISM errors must be such as to indicate that one exemplar is the source of the other’s error or that the manuscripts in question must have had a common source to which they all owe the error. simultaneous innovation. Also referred to as polygenesis of errors, simultaneous innovation refers to the circumstance that two manuscripts may feature the same corruption, especially if it is a trivial one, without one having gotten it from the other or both having gotten it from a common ancestor. source. The exemplar from which another manuscript was copied (compare German Vorlage). stemma codicum (pl.: stemmata codicum). The genealogical tree of manuscript relationships  established on the basis of the signiicant errors shared between manuscripts, a stemma codicum may be either implicit or explicit (that is, drawn up or represented as a graphic) and includes extant witnesses, as well as hypothetical ancestors. text transmission. The descent of the text from one or more exemplars, usually established through genealogical analysis of manuscripts in cases where external testimony is unavailable. textual criticism. The two-stage process of a systematic recensio followed by a gene- alogical analysis of manuscripts. Sometimes applied to the methods of biblical criticism, but should be restricted to the production of a critical text, either of the hypothetical archetype or of the best surviving witness, after a careful study of the available variants alone. textual history. Also history of the text, the practice of determining the text’s genesis and transmission and the circumstances surrounding its reproduction. Should not be confused with Textgeschichte in German Mahā bhā rata studies, which refers to an a priori history based on neither positive historical evidence nor stemmatic reconstruction. textus receptus. Latin for “received text.” The textus receptus of the New Testament is either the 1633 edition of Elzevir or the 1550 edition of Robert Stephanus. More generally, the term can be used to refer to the traditional text, as received or handed down within a particular community. usus auctoris/usus scribendi. The author’s or scribe’s habitual style, the distinction between the two is not hard and fast. varia lectio (pl.: variae lectiones). Literally, “variant reading.” An alternate wording of a passage or word in one manuscript in another manuscript of the same work. variantistica. The study of the author’s variants or versions of a text as an end in itself, an intellectual enterprise revealing much about the author, his work and the creation of texts, rather than as a problem to overcome in the attempt to recon- struct the archetype. variation place. It is sometimes helpful to distinguish a variation place, a passage or unit of the text in which a variant can occur, from the actual variants themselves. vetusti. The old manuscripts of a work. vetustissimi. The oldest manuscripts of a work. 491 GLOSSARY 491 vulgate/vulgata. The term for any successful edition of a text (from editio vulgata). Accordingly, there may be several vulgatae of a work. For the Bible, the vulgate is the version of the Latin Bible the medieval Church recognized as the “vetus et vulgata editio” (the “old and popular edition”). The vulgate version of the Mahā bhā rata is Nīlakaṇṭha’s late seventeenth-century edition. The attribution is Sukthankar’s. witness. Extant manuscripts of a work, so called because they witness or attest to that work. Sometimes also called surviving manuscripts. 492 493 ANNOTATED BIBLIOGRAPHY This bibliography includes the works cited in this book and additional sources that provide an overview of the main currents in Mahā bhā rata scholarship. It is not intended as a comprehensive overview of literature on the Mahā bhā rata. The background of Mahā bhā rata studies is covered in our book The Nay Science, while the bibliographies in Argument and Design ofer a better overview of literary and philosophical approaches. The focus in this bibliography is, rather, on Mahā bhā rata textual criticism, supplemented with information about the main debates in Mahā bhā rata schol- arship and suggestions for further study. Additionally, introductions and theoretical overviews of textual criticism are included, especially those from the Italian school, as the most germane for Mahā bhā rata textual criticism. The Mahā bhā rata Critical Edition The full titles of the Mahā bhā rata critical edition volumes are clariied in Appendix 1.  The volumes, however, are almost never cited like this. The shorter forms that have become scholarly convention and that are adopted in this book are listed in this section. Also included here is the edition’s predecessor, N. B. Utgikar’s critical edition of the Virā ṭaparvan. Belvalkar, S. K., ed. The As̄ ramavāsikaparvan for the First Time Critically Edited. Poona:  Bhandarkar Oriental Research Institute, 1959.1 ———. The Bhı̄ṣmaparvan for the First Time Critically Edited. Poona: Bhandarkar Oriental Research Institute, 1947. ———. The Mahāprasthānikaparvan for the First Time Critically Edited. Poona:  Bhandarkar Oriental Research Institute, 1959. ———. The Mausalaparvan for the First Time Critically Edited. Poona: Bhandarkar Oriental Research Institute, 1959. ———. The Śāntiparvan for the First Time Critically Edited. 4  vols. Poona:  Bhandarkar Oriental Research Institute, 1954–61. ———. The Śāntiparvan, Part III: Mokṣadharma, B, For the First Time Critically Edited. Poona: Bhandarkar Oriental Research Institute, 1954. ———. The Svargārohaṇaparvan for the First Time Critically Edited. Poona:  Bhandarkar Oriental Research Institute, 1959. Dandekar, R. N., ed. The Anuśāsanaparvan for the First Time Critically Edited. 2 vols. Poona: Bhandarkar Oriental Research Institute, 1966. ———. The Śalyaparvan for the First Time Critically Edited. Poona:  Bhandarkar Oriental Research Institute, 1961. De, S. K., ed. The Droṇaparvan for the First Time Critically Edited. 2 vols. Poona: Bhandarkar Oriental Research Institute, 1958. ———. The Udyogaparvan for the First Time Critically Edited. Poona: Bhandarkar Oriental Research Institute, 1940. Edgerton, Franklin, ed. The Sabhāparvan for the First Time Critically Edited. Poona:  Bhandarkar Oriental Research Institute, 1944. Karmarkar, R. D., ed. The As̄ ́vamedhikaparvan for the First Time Critically Edited. Poona: Bhandarkar Oriental Research Institute, 1960. 49 494 PHILOLOGY AND CRITICISM Paranjpe, V. G., ed. The Strı̄parvan for the First Time Critically Edited. Poona:  Bhandarkar Oriental Research Institute, 1956. Sukthankar, V. S., ed. The Ad̄ iparvan for the First Time Critically Edited. 2  vols. Poona:  Bhandarkar Oriental Research Institute, 1933. ———. The Ar̄ aṇyakaparvan for the First Time Critically Edited. 2 vols. Poona: Bhandarkar Oriental Research Institute, 1942. Utgikar, N. B., ed. The Virātạ parvan of the Mahābhārata Edited from Original Manuscripts as a Tentative Work with Critical and Explanatory Notes and an Introduction. Poona: Bhandarkar Oriental Research Institute, 1923. Vaidya, P. L., ed. The Harivaṃśa for the First Time Critically Edited. 2 vols. Poona: Bhandarkar Oriental Research Institute, 1969–71. ———. The Karṇaparvan for the First Time Critically Edited. Poona:  Bhandarkar Oriental Research Institute, 1954. ———. The Pratı̄ka-Index of the Mahābhārata. 6 vols. Poona: Bhandarkar Oriental Research Institute, 1967–72. Velankar, H. D., ed. The Sauptikaparvan for the First Time Critically Edited. Poona: Bhandarkar Oriental Research Institute, 1948. Vira, Raghu, ed. The Virātạ parvan for the First Time Critically Edited. Poona:  Bhandarkar Oriental Research Institute, 1936. Editors’ Introductions from the Mahā bhā rata Critical Edition The editorial introductions to the respective volumes of the critical edition remain the best source of information about it. The most important introductions from the perspective of textual criticism are Sukthankar’s “Prolegomena” (the introduction to the Ā diparvan), his “Introduction” (to the Ā raṇyakaparvan) and Edgerton’s “Introduction” (to the Sabhā parvan). Sukthankar’s “Foreword” (to the irst fascicule of the Ā diparvan) and the series of “Editorial Notes” (written for succeeding fascicules) ofer a glimpse into his evolving understanding of the Mahā bhā rata tradition. Most of their information was absorbed into the “Prolegomena,” which replaced them from the seventh and inal fascicule onward. Belvalkar, S. K. “Introduction.” In The As̄ ́ramavāsikaparvan for the First Time Critically Edited. i–xlviii. Poona: Bhandarkar Oriental Research Institute, 1959. ———. “Introduction.” In The Bhı̄ṣmaparvan for the First Time Critically Edited. ix–cxxxii. Poona: Bhandarkar Oriental Research Institute, 1947. ———. “Introduction.” In The Mahāprasthānikaparvan for the First Time Critically Edited. ix–xxvi. Poona: Bhandarkar Oriental Research Institute, 1959. ———. “Introduction.” In The Mausalaparvan for the First Time Critically Edited. ix–xxxv. Poona: Bhandarkar Oriental Research Institute, 1959. ———. “Introduction.” In The Svargārohaṇaparvan for the First Time Critically Edited. ix–xxxi. Poona: Bhandarkar Oriental Research Institute, 1959. ———. The Śāntiparvan: Introduction. Poona: Bhandarkar Oriental Research Institute, 1966.2 Dandekar, R. N. “Introduction.” In The Anuśāsanaparvan for the First Time Critically Edited. Vol. 1. ix–lxxxv. Poona: Bhandarkar Oriental Research Institute, 1966. ———. “Introduction.” In The Śalyaparvan for the First Time Critically Edited. ix–lvi. Poona: Bhandarkar Oriental Research Institute, 1961. De, S. K. “Introduction.” In The Droṇaparvan for the First Time Critically Edited. Vol. 1.  i–xxxii. Poona: Bhandarkar Oriental Research Institute, 1958. ———. “Introduction.” In The Udyogaparvan for the First Time Critically Edited. i–xlix. Poona: Bhandarkar Oriental Research Institute, 1940. Edgerton, Franklin. “Introduction.” In The Sabhāparvan for the First Time Critically Edited. ix–l. Poona: Bhandarkar Oriental Research Institute, 1944. 495 ANNOTATED BIBLIOGRAPHY 495 Karmarkar, R. D. “Introduction.” In The As̄ ́vamedhikaparvan for the First Time Critically Edited. ix–xlvi. Poona: Bhandarkar Oriental Research Institute, 1960. Paranjpe, V. G. “Introduction.” In The Strı̄parvan for the First Time Critically Edited. vii–xxiv. Poona: Bhandarkar Oriental Research Institute, 1956. Sukthankar, V. S. “Editorial Note.” In The Mahābhārata for the First Time Critically Edited, Ad̄ iparvan: Fascicule 2. i. Poona: Bhandarkar Oriental Research Institute, 1928. ———. “Editorial Note (2).” In The Mahābhārata for the First Time Critically Edited, Ad̄ iparvan: Fascicule 3. i–iii. Poona: Bhandarkar Oriental Research Institute, 1929. ———. “Editorial Note (3).” In The Mahābhārata for the First Time Critically Edited, Ad̄ iparvan: Fascicule 4. i–iv. Poona: Bhandarkar Oriental Research Institute, 1930. ———. “Editorial Note (4).” In The Mahābhārata for the First Time Critically Edited, Ad̄ iparvan: Fascicule 5. i–iii. Poona: Bhandarkar Oriental Research Institute, 1931. ———. “Foreword.” In The Mahābhārata for the First Time Critically Edited, Ad̄ iparvan: Fascicule 1. i–ix. Poona: Bhandarkar Oriental Research Institute, 1927. ———. “Introduction.” In The Ar̄ aṇyakaparvan for the First Time Critically Edited. Vol. 1.  i–xxxviii. Poona: Bhandarkar Oriental Research Institute, 1942. ———. “Prolegomena.” In The Ad̄ iparvan for the First Time Critically Edited. Vol. 1.  i–cx. Poona: Bhandarkar Oriental Research Institute, 1933. Utgikar, N. B. “Introduction.” In The Virātạ parvan of the Mahābhārata Edited from Original Manuscripts as a Tentative Work with Critical and Explanatory Notes and an Introduction. i–lii. Poona: Bhandarkar Oriental Research Institute, 1923. Vaidya, P. L. “Introduction.” In The Harivaṃśa for the First Time Critically Edited. Vol. 1.  ix–l. Poona: Bhandarkar Oriental Research Institute, 1969. ———. “Introduction.” In The Karṇaparvan for the First Time Critically Edited. ix–xlii. Poona: Bhandarkar Oriental Research Institute, 1954. Velankar, H. D. “Introduction.” In The Sauptikaparvan for the First Time Critically Edited. ix–xxxiv. Poona: Bhandarkar Oriental Research Institute, 1948. Vira, Raghu. “Introduction.” In The Virātạ parvan for the First Time Critically Edited. i–xxvii. Poona: Bhandarkar Oriental Research Institute, 1936. Reviews of the Mahā bhā rata Critical Edition No attempt has been made to collect all the reviews of the Mahā bhā rata critical edition, but the list that follows is fairly representative. The most important reviewer of the edition was Franklin Edgerton: in a series of reviews published in the Journal of the American Oriental Society, Edgerton expressed his keen understanding of the principles behind the edition. Lüders and Winternitz also wrote early and detailed reviews. Also included here are reviews and discussions of the tentative critical edition of the Virā ṭaparvan under N. B. Utgikar’s leadership and P. P. S. Sastri’s southern recension. Lévi’s 1929 review is notable for his suggestion that Sukthankar abandon the attempt to reconstruct the archetype in favor of reprinting the text of the vulgate with an apparatus of variants (see the introduction). Of the reviews we examined, the majority were positive. The few longer, negative articles are listed under “German Scholarship/Errors in Textual Criticism” in light of their frequent misunderstandings of the critical edition. Agrawala, V. S. Review of The Mahābhārata: Drona-Parvan, Parts I, 2, 3A, 3B, by S. K. De. Journal of the American Oriental Society 82, no. 2 (1962): 231–32. Barnett, L. D. Review of The Mahābhārata:  For the First Time Critically Edited by V.  S. Sukthankar, Fascicules 1 and 2. Bulletin of the School of Oriental Studies 5, no. 2 (1929): 379–80. ———. Review of The Virātạ parvan of the Mahābhārata, by Narayan Bapuji Utgikar; Shrimant Balasaheb Pant Pratinidhi. Journal of the Royal Asiatic Society of Great Britain and Ireland no. 2 (1924): 293–95. ———. Review of The Virātạ parvan of the Mahābhārata. Edited from original manuscripts by Narayan Bapuji Utgikar. Journal of the Royal Asiatic Society 56, no. 2 (1924): 293–95. 496 496 PHILOLOGY AND CRITICISM Belloni Filippi, F. “The Critical Edition of the Mahā bhā rata.” East and West 5, no. 4 (1955): 265–70. Belvalkar, S. K. “Publication of Fascicule 22 of the Critical Edition of the Mahā bhā rata.” Annals of the Bhandarkar Oriental Research Institute 32, no. 1/4 (1951): 340–43. Burrow, T. Review of The Ar̄ aṇyakaparvan (Part  2). For the First Time Critically Edited by Vishnu S. Sukthankar. Journal of the Royal Asiatic Society 78, no. 1–2 (1946): 111–12. ———. Review of The Mahābhārata. For the First Time Critically Edited by Vishnu S. Sukthankar and S. K. Belvalkar, vols. VII, X, XIII, XIV, XV. Journal of the Royal Asiatic Society 84, no. 3–4 (1952): 165. Dewhurst, R. P. Review of The Mahābhārata, Ad̄ iparvan, Fascicule 6, edited by Vishnu S. Sukthankar. Journal of the Royal Asiatic Society 65, no. 2 (1933): 443–44. ———. Review of The Mahābhārata. For the First Time Critically Edited by Vishnu S.  Sukthankar, Ā diparvan fascicules 1–4. Journal of the Royal Asiatic Society 63, no. 2 (1928): 465–66. ———. Review of The Mahābhārata. For the First Time Critically Edited by Vishnu S. Sukthankar, Fascicule 5. Journal of the Royal Asiatic Society 64, no. 3 (1932): 668–69. ———. Review of The Mahābhārata (Southern Recension). Vol. I: Ad̄ i Parva, Part I, by P. P. S. Śastri. Journal of the Royal Asiatic Society of Great Britain and Ireland 64, no. 2 (1932): 446–47. ———. Review of The Mahābhārata (Southern Recension). Vol. III: Sabhā Parvan by P. P. S. Śastri. Journal of the Royal Asiatic Society of Great Britain and Ireland 65, no. 1 (1933): 142. Dikshitar, V. R. R. Review of The Mahābhārata, Vol. I, Parts 1 & 2, Critically Edited by P. P. S. Sastri. Annals of the Bhandarkar Oriental Research Institute 13, no. 2 (1931–32): 192. Edgerton, Franklin. “A Critically Edited Text of Nala, 1–5.” Journal of the American Oriental Society 62, no. 3 (1942): 198–200. ———. Review of The Mahābhārata for the First Time Critically Edited, by V.  S. Sukthankar. Ā diparvan: Fascicules 2 and 3. Journal of the American Oriental Society 49 (1929): 282–84. ———. Review of The Mahābhārata for the First Time Critically Edited, by V.  S. Sukthankar. Ā diparvan: Fascicule 5. Journal of the American Oriental Society 52, no. 3 (1932): 252–54. ———. Review of The Mahābhārata for the First Time Critically Edited, by V.  S. Sukthankar. Ā diparvan: Fascicule 7. Journal of the American Oriental Society 56, no. 3 (1936): 360–62. ———. Review of The Mahābhārata for the First Time Critically Edited, by V. S. Sukthankar. Text of 1.1.1 to 1.2.233. Journal of the American Oriental Society 48 (1928): 186–90. Emeneau, M. B. Review of The Mahābhārata. Vol. 2. The Sabhāparvan by Vishnu S. Sukthankar; S. K. Belvalkar; Franklin Edgerton. Journal of the American Oriental Society 66, no. 3 (1946): 267–69. ———. Review of The Mahābhārata. Vol. 5. The Virātạ parvan by Vishnu S. Sukthankar; Raghu Vira; The Mahābhārata. Vol. 6. The Udyogaparvan by Vishnu S. Sukthankar; Sushil Kumar De. Journal of the American Oriental Society 62, no. 3 (1942): 205–6. Gurner, Walter. Review of The Mahābhārata. Vol. XV:  Ś ā ntiparvan by S.  K. Belvalkar; The Mahā bhā rata. Vols. VIII and IX:  Droṇaparvan by S.  K. De; The Mahā bhā rata. Vol. X: Karṇaparvan by P. L. Vaidya. Journal of the Royal Asiatic Society of Great Britain and Ireland 88, no. 1/2 (1956): 108–9. ———. Review of The Mahābhārata by V. S. Sukthankar and S. K. Belvalkar. Journal of the Royal Asiatic Society of Great Britain and Ireland 90, no. 3/4 (1958): 211–12. Hopkins, E.  Washburn. Review of The Mahābhārata for the First Time Critically Edited, by V.  S. Sukthankar. Ā diparvan: Fascicule 4. Journal of the American Oriental Society 51, no. 2 (1931): 179–80. Johnston, E. H. Review of The Mahābhārata. Vol. V.  The Virātạ parvan. Critically edited by Raghu Vira. Journal of the Royal Asiatic Society 70, no. 4 (1938): 588–89. ———. Review of The Mahābhārata. Vol. 1. The Ad̄ iparvan. Critically edited by Vishnu S. Sukthankar. Journal of the Royal Asiatic Society 68, no. 2 (1936): 317–22. Lévi, Sylvain. Review of The Mahābhārata. For the First Time Critically Edited, by Vishnu S. Sukthankar. Journal Asiatique 215 (1929): 345–48. ———. Review of The Mahābhārata. For the First Time Critically Edited, by Vishnu S.  Sukthankar. Journal Asiatique 225 (1929): 281–83. 497 ANNOTATED BIBLIOGRAPHY 497 Lüders, Heinrich. Review of The Mahābhārata for the First Time Critically Edited, by Vishnu S. Sukthankar. Deutsche Literaturzeitung 24 (1929): 1137–146. Pisani, Vittore. “A Note on Anuśā naparvan.” Annals of the Bhandarkar Oriental Research Institute 48/49, Golden Jubilee Volume 1917–67 (1968): 59–62. Thomas, E. J. Review of The Sabhaparvan, by Franklin Edgerton. Bulletin of the School of Oriental and African Studies 12, no. 2 (1948): 458–59. Weller, Hermann. “Zum zweiten Heft der neuen Mahā bhā rata Ausgabe.” Zeitschrift für Indologie und Iranistik 7 (1929): 91–95. ———. “Zum 7.  Heft der neuen Mahā bhā rata-Ausgabe.” Zeitschrift für Indologie und Iranistik 10 (1935–36): 116–20. Winternitz, Moriz. “The Critical Edition of the Mahā bhā rata: Ā diparvan.” Annals of the Bhandarkar Oriental Institute 15, nos. 3–4 (1934): 159–75. ———. Review of The Virātạ parvan of the Mahābhārata Edited from Original Manuscripts as a Tentative Work with Critical and Explanatory Notes and an Introduction, by N. B. Utgikar. Annals of the Bhandarkar Oriental Research Institute 5, no. 1 (1923–24): 19–30. Yarrow, Andrew H. Review of The Mahābhārata. Vol. 7. The Bhīṣmaparvan by Vishnu S. Sukthankar; S.  K. Belvalkar; Shripad Krishna Belvalkar. Journal of the American Oriental Society 70, no. 4 (1950): 317–20. Editions Other than the Critical Edition The editions listed here were described in Appendix 2 and therefore are not discussed here again. The critical edition will be the standard edition for the foreseeable future, but readers should not neglect the vulgate and other regional editions of the Mahā bhā rata, where available in print. In many cases, they remain the best guides to the archetypal text, exceeding in value most of the scholarly and “scientiic” literature composed about the epic. Ainapure, Vasudev Balacharya, ed. The Mahābhārata with the Commentary Bhāvadı̄pa of Nı̄lakaṇtḥ a. 6 vols. Bombay: Gopal Narayan & Company, 1901. Khadilkar, Atmaram, ed. Mahābhārata with Nı ̄lakaṇtḥ a’s Commentary. Bombay:  Ganpat Krishnaji’s Press, 1863. Kinjawadekar, Ramachandra, ed. Mahābhāratam with the Commentary of Nıl̄ akaṇtḥ a. 6  vols. Pune: Chitrashala Press, 1929–36. Krishnacharya, T. R., ed. A Descriptive Contents of Sriman Mahabharatam. Kumbakonam:  Madhva Vilas Book Depot, 1912. Krishnacharya, T. R., and T. R. Vyasacharya, eds. Sriman Mahabharatam. A New Edition Mainly Based on the South Indian Texts with Footnotes and Readings. 19 vols. Bombay: Javaji Dadaji’s “Nirṇaya- Sā gar” Press, 1906–14. The Mahábhárata, an Epic Poem Written by the Celebrated Veda Vyása Rishi. 4 vols. Calcutta: Education Committee’s Press/Baptist Mission Press, 1834–39. Pandit Ramanarayanadatta Sastri Pandey, ed. Śrı̄manmaharṣi Vedavyāsapraṇıt̄ a Mahābhārata. 7  vols. (including Harivaṃśa). Gorakhpur: Gita Press, no date. Sastri, P. P.  S., ed. The Mahābhārata (Southern Recension) Critically Edited by P.  P. S.  Sastri. Madras: Ramaswamy Sastrulu & Sons, Madras, 1931–33. Translations (Including Reviews) of the Critical Edition or the Vulgate With the exception of the abridged translation of the Mahā bhā rata into French, the translations listed here were described in Appendix 3 and therefore are not discussed here again. With the excep- tion of Biardeau’s and Péterfalvi’s editions, cited for their valuable commentaries, abridgements (such as John D. Smith’s) are not listed here. 498 498 PHILOLOGY AND CRITICISM Aklujkar, Ashok. Review of The Mahabharata. I. The Book of the Beginning, by J. A. B. van Buitenen. Paciic Afairs 48, no. 1 (1975): 133–34. Biardeau, Madeleine. Le Mahābhārata. Extraits traduits du sanscrit. Vol. I: Livres I à V. Vol. II: Livres VI à XVIII. Translated by J.-M. Péterfalvi. Introduction, commentaries, summaries and glossary by M. Biardeau. Paris: Flammarion, 1985–86. ———. Le Mahābhārata: Un récit fondateur du brahmanisme et son interpretation. 2 vols. Paris: Éditions du Seuil, 2002. Bisschop, Peter. Review of Mahābhārata. Book Two. The Great Hall by Paul Wilmot; Mahābhārata. Book Three. The Forest. Volume Four by William J. Johnson; Mahābhārata. Book Nine. Śalya. Volume One by Justin Meiland. Zeitschrift der Deutschen Morgenländischen Gesellschaft 159, no. 1 (2009): 239–41. Bodewitz, H. Review of Le Mahābhārata. Livres I à V. Livres VI à XVIII. Extraits traduits du sanscrit par Jean-Michel Péterfalvi. Commentaires, résumé et glossaire par Madeleine Biardeau, by J.-M. Péterfalvi, M. Biardeau. Anthropologica xxix (1987): 167–68. Bolle, Kees W. Review of The Mahābhārata, Book 1: The Book of the Beginning, by J. A. B. van Buitenen. Journal of the American Academy of Religion 43, no. 3 (1975): 617. Van Buitenen, J. A.  B., trans. The Mahābhārata. 3  vols. Chicago:  University of Chicago Press, 1973–78. Debroy, Bibek, trans. The Mahābhārata. 10 vols. New Delhi: Penguin Books India, 2010–14. Dutt, Manmatha Nath, trans. A Prose English Translation of the Mahabharata (Translated Literally from the Original Sanskrit Text). 8 vols. Calcutta: Elysium Press, 1895–1905. Fitzgerald, James L., ed. and trans. The Mahābhārata. Vol. 7:11. The Book of the Women, 12. The Book of Peace, Part One. Chicago: University of Chicago Press, 2004. Ganguli, Kisari Mohan, trans. The Mahābhārata of Krishna-Dwaipayana Vyasa Translated into English Prose. 11 vols. Calcutta: Bharata Press, 1884–96. Goldman, Robert. “India’s Great War.” Review of The Mahabharata. Book I. The Book of the Beginning, by J. A. B. van Buitenen. The Journal of Asian Studies 35, no. 3 (1976): 463–70. Hazlehurst, Leighton. Review of The Mahābhārata: Book 1, the Book of the Beginning, by J. A. B. van Buitenen. Ethnohistory 21, no. 2 (1974): 181–83. Hiltebeitel, Alf. “On Reading van Buitenen’s Vyā sa.” Review of The Mahābhārata. Vol. 1: The Book of the Beginning, by J. A. B. van Buitenen. History of Religions 14, no. 3 (1975): 230–32. ———. “On Reading Fitzgerald’s Vyā sa.” Review of The Mahābhārata. Vol. 7: 11, the Book of the Women; 12, the Book of Peace, Part One, by James L. Fitzgerald. Journal of the American Oriental Society 125, no. 2 (2005): 241–61. Hudson, Dennis. Review of The Mahābhārata, by J.  A. B.  van Buitenen. Journal of the American Academy of Religion 46, no. 3 (1978): 387–88. Lal, Purushottam, trans. The Mahābhārata of Vyāsa: Transcreated from Sanskrit. 18 vols. Calcutta: Writers Workshop, 1968–2008. Maha Bhárata. 15 vols. Clay Sanskrit Library. New York: New York University Press and the JJC Foundation, 2005–9. Mehta, Mahesh. Review of The Mahābhārata. VII: The Astika Parva in the Adi Parva, by P. Lal. Books Abroad 45, no. 1 (1971): 183. Olson, Carl. Review of The Mahābhārata. Book 11:  The Book of the Women; The Mahābhārata. Book 12: The Book of Peace, Part One, by James L. Fitzgerald. International Journal of Hindu Studies 10, no. 1 (2006): 109–10. Regier, Willis G. Review of The Mahābhārata, Volume 7, by James L. Fitzgerald. MLN 119, no. 5, Comparative Literature Issue (2004): 1103–06. Rocher, Ludo. Review of The Mahābhārata by J. A. B. van Buitenen. Journal of the American Oriental Society 98, no. 2 (1978): 193–94. Smith, John D. Review of The Mahābhārata: 11, The Book of the Women; 12, The Book of Peace, Part One, ed. and trans. James L. Fitzgerald. Bulletin of the School of Oriental and African Studies 68, no. 1 (2005): 131–33. 49 ANNOTATED BIBLIOGRAPHY 499 Von Simson, Georg. Review of Mahābhārata. Book Four. Virātạ , trans. Kathleen Garbutt; Mahābhārata. Book Seven. Droṇa. Volume One, ed. and trans. Vaughan Pilikian; Mahābhārata. Book Eight. Karṇa. Volume One, trans. Adam Bowles. Zeitschrift der Deutschen Morgenländischen Gesellschaft 159, no. 2 (2009): 481–89. Problems in Mahā bhā rata Textual Criticism Lüders and Winternitz were the earliest and most acute observers of the Mahā bhā rata tradi- tion (see the section “Background of the Mahā bhā rata Critical Edition; Biographic Sources” for additional writings). The main persons to address problems in Mahā bhā rata textual criti- cism thereafter were the editors of the critical edition. Sukthankar produced eight important studies titled “Epic Studies” (I–IV and VI–VII are listed here; V is listed under “Mahā bhā rata Commentators, Commentators’ Editions, and Chronological Surveys” and VIII under “Textual Traditions and Editions of Texts Other than the Mahā bhā rata”; all eight are reprinted in the irst volume of the Sukthankar Memorial Edition). Edgerton planned to write a series of articles titled “Epic Studies: Second Series” in homage to Sukthankar, but we could locate only the irst. Some of the most important articles thereafter concerned the Parvasaṃgraha igures, Sukthankar having already discussed the argument from them in 1933 and 1939. Also included here are Mahesh Mehta’s and M. A. Mehendale’s articles, as examples of the kind of higher critical ideas frequently projected onto the Mahā bhā rata critical edition. Belvalkar, S. K. “The Interpretation of the Parvasaṃgraha Figures.” Annals of the Bhandarkar Oriental Research Institute 27, no. 3/4 (1946): 303–9. ———. “The Mahā bhā rata Text-Transmission Problems: Problem No. 4.” Journal of the American Oriental Society 72, no. 1 (1952): 34–37. ———. “Some Interesting Problems in Mahā bhā rata Text-Transmission.” Annals of the Bhandarkar Oriental Research Institute 25, no. 1/3 (1944): 82–87. Bhattacharya, Ramkrishna. “The Mahā bhā rata 12.224–30: A Text-Critical Note.” Annals of the Bhandarkar Oriental Research Institute 83 (2002): 153–63. De, Sushil Kumar. “Some Lexicographical Notes on the Udyoga-Parvan of the Mahā bhā rata.” Bulletin of the Deccan College Research Institute 8, Dikshit Memorial Volume (1947): 1–33, 1–2. De, S. K., and E. H. Johnston. “A Passage from the Udyoga-Parvan.” Journal of the Royal Asiatic Society of Great Britain and Ireland 1 (1940): 69–73. Edgerton, Franklin. “Epic Studies: Second Series.” Bulletin of the Deccan College Research Institute 5, Sukthankar Memorial Volume (1943–44): 1–12. Kosambi, D. D. “The Parvasaṃgraha of the Mahā bhā rata.” Journal of the American Oriental Society 66, no. 2 (1946): 110–17. Kosambi, D. D., and E. D. Kulkarni. “Parvasaṃgraha Figures for the Bhīṣmaparvan of the Mahā bhā rata.” Journal of the American Oriental Society 71, no. 1 (1951): 21–25. Krishnamurti Sarma, B. N. “An Opening Verse of the Mahā bhā rata.” Annals of the Bhandarkar Oriental Research Institute 26, no. 3/4 (1945): 307–12. Kulkarni, E. D. “The Parvasaṃgraha Figures.” Journal of the American Oriental Society 66, no. 2 (1946): 118–45. Lüders, Heinrich. Über die Grantharecension des Mahābhārata. Abhandlungen der Königlichen Gesellschaft der Wissenschaften zu Göttingen, Phil.-Hist. Klasse, N.  F.  vol. 14, no.  6. Berlin: Weidmannsche Buchhandlung, 1901. Mehendale, M. A. “The Critical Edition of the Mahā bhā rata: Its Achievement and Limitations.” Annals of the Bhandarkar Oriental Research Institute 88 (2007): 1–16. ———. “The Critical Edition of the Mahā bhā rata:  Its Constitution, Achievements, and Limitations.” In Text and Variations of the Mahābhārata:  Contextual, Regional and Performative Traditions, edited by Kalyan Kumar Chakravarty, 3–23. New Delhi:  National Mission for Manuscripts, 2009. 50 500 PHILOLOGY AND CRITICISM ———. “Interpolations in the Mahā bhā rata.” Annals of the Bhandarkar Oriental Research Institute 82, no. 1/4 (2001): 193–212. Mehta, Mahesh M. “The Mahā bhā rata: A Study of the Critical Edition with Special Reference to the Suparṇā khyā na of the Ā diparvan: Part II.” Bhāratı ̄ya Vidyā 32 (1972): 3–72. ———. “The Problem of the Double Introduction to the Mahā bhā rata.” Journal of the American Oriental Society 93, no. 4 (1973): 547–50. Sukthankar, Vishnu S. “Epic Questions I: Does Indra Assume the Form of a Swan?” Bulletin of the Deccan College Research Institute 1, no. 1 (1939): 1–7. ———. “Epic Questions II: The Parvasaṃgraha Figures.” Annals of the Bhandarkar Oriental Institute 23, Silver Jubilee Number (1939): 549–58. ———. “Epic Studies I: Some Aspects of the Mahā bhā rata Canon.” Journal of the Bombay Branch of the Royal Asiatic Society, n.s., 4 (1928): 157–78. ———. “Epic Studies II: Further Text-Critical Notes.” Annals of the Bhandarkar Oriental Institute 11, no. 2 (1930): 165–91. ———. “Epic Studies III: Dr. Ruben on the Critical Edition of the Mahā bhā rata.” Annals of the Bhandarkar Oriental Institute 11, no. 3 (1930): 259–83. ———. “Epic Studies IV: More Text-Critical Notes.” Annals of the Bhandarkar Oriental Institute 16, no. 1 (1934–35): 90–113. ———. “Epic Studies VI: The Bhṛgus and the Bhā rata: A Text-Historical Study.” Annals of the Bhandarkar Oriental Institute 18, no. 1 (1936–37): 1–76. ———. “Epic Studies VII:  The Oldest Extant Ms. of the Ā diparvan.” Annals of the Bhandarkar Oriental Institute 19 (1938–39): 201–62. Winternitz, Moriz. “On the South-Indian Recension of the Mahā bhā rata.” The Indian Antiquary 27 (1898): 67–81, 92–104, 122–36. ———. “Der Sabhā parvan in der südindischen Rezension des Mahā bhā rata.” Wiener Zeitschrift für die Kunde des Morgenlandes 17 (1903): 70–75. Mahā bhā rata Commentators, Commentators’ Editions and Chronological Surveys The works listed here appear in the notes to Appendix 16 (“Commentaries on the Mahā bhā rata”). They are listed again here for convenience. Additionally listed are the most important articles and essays on the Mahā bhā rata commentators. The Mahā bhā rata commentators remain the least researched aspect of the Mahā bhā rata tradition. They are often passed over for a fetish history of origins or allusions to the Mahā bhā rata as a Wikipedia-like undertaking or a text so luid that it cannot be said that a single Mahā bhā rata ever existed. Newer studies of the commentators, how- ever, reveal a deinite text with established canons for its interpretation so that we must radically revise our notions of the work and its reception. Apte, V. M., and D. V. Garge. “Mahā bhā rata Citations in the Ś abara Bhā ṣya.” Bulletin of the Deccan College Research Institute 5, Sukthankar Memorial Volume (1943–44): 221–29. Bakre, Mahadeva Shastri, ed. Udyogaparvan. Bombay: Gujarati Printing Press, 1920. ———. Virātạ parvan. Bombay: Gujarati Printing Press, 1915. Belvalkar, S. K. Commentary of Devabodha on the Bhı̄ṣmaparvan. Poona: Bhandarkar Oriental Research Institute, 1947. Bhattacharya, Dinesh Chandra. “Vidyā sā gara’s Commentary on the Mahā bhā rata.” Annals of the Bhandarkar Oriental Research Institute 25, no. 1/3 (1944): 99–102. Dandekar, R. N. Commentary of Devabodha on the Ad̄ iparvan of the Mahābhārata. Poona: Bhandarkar Oriental Research Institute, 1941. ———. “Devabodhakṛtamahābhāratatātparyatı̣ ̄kā—ādiparva.” Annals of the Bhandarkar Oriental Research Institute 22, no. 1/2 (1941): 135–70. 501 ANNOTATED BIBLIOGRAPHY 501 Dandekar, R. N., and V. M. Bedekar. “Vā dirā ja’s Lakṣā laṃkā ra on the Viṣṇusahasranā mastotra.” Annals of the Bhandarkar Oriental Research Institute 46, no. 1/4 (1965): 81–98. De, S. K. Commentary of Devabodha on the Udyogaparvan of the Mahābhārata. Bombay: Bharatiya Vidya Bhavan, 1944. Gode, P. K. “Dr. P. P. S. Sastri, and the Date of Vā dirā jatīrtha.” Annals of the Bhandarkar Oriental Research Institute 17, no. 2 (1935–36): 203–10. ———. “New Light on the Chronology of the Commentators of the Mahā bhā rata.” Annals of the Bhandarkar Oriental Research Institute 25, no. 1/3 (1944): 103–8. ———. “Nīlakaṇṭha Caturdhara, the Commentator of the Mahā bhā rata—His Genealogy and Descendants.” Annals of the Bhandarkar Oriental Research Institute 23 (1942): 141–61. ———. “Notes on Indian Chronology XXXIV:  Date of Vimalabodha’s Commentary on the Mahā bhā rata Called the Viṣamaślokī—After 1150 AD.” Annals of the Bhandarkar Oriental Research Institute 17, no. 4 (1935–36): 395–99. ———. Studies in Indian Literary History. 2 vols. Bombay: Bharatiya Vidya Bhavan, 1953–54. Kane, P. V. “The Mahā bhā rata and Ancient Commentators.” Annals of the Bhandarkar Oriental Research Institute 19, no. 2 (1938): 161–72. Karmarkar, R. D. Commentary of Devabodha on the Sabhāparvan of the Mahābhārata. Poona: Bhandarkar Oriental Research Institute, 1949. Leclere, Basile. “New Light on Devabodha, the Earliest Extant Commentator on Mahā bhā rata.” Asiatische Studien/Études Asiatique 70, no. 2 (2016): 489–526. Minkowski, Christopher. “Nīlakaṇṭha Caturdhara’s Mantrakā śīkhaṇḍa.” Journal of the American Oriental Society 122, no. 2 (2002): 329–44. ———. “Nīlakaṇṭha’s Mahā bhā rata.” Seminar 608, The Enduring Epic (2010): 32–38. ———. “What Makes a Work ‘Traditional’? On the Success of Nīlakaṇṭha’s Mahābhārata Commentary.” In Boundaries, Dynamics and Constructions of Traditions in South Asia, edited by Federico Squarcini, 225–52. Florence: Firenze University Press, 2005. Misra, V. N. Mahābhārata-Tātparyaprakāśa of Śrı̄ Sadānanda Vyāsa. Varanasi: Ratna, 1992. ———. Śrimahābhāratatātparyanirṇaya of Śrı̄ An̄ andatı̄rtha (Madhvācār ya). Varanasi: Ratna, 1992. Narsimha Acharya, N. C. “The Ā ndhra Mahā bhā ratamu.” Annals of the Bhandarkar Oriental Research Institute 22, no. 1/2 (1941): 97–102. Raghavan, V. “Notes on Some Mahā bhā rata Commentaries.” In A Volume of Studies in Indology, Presented to Prof. P. V. Kane, edited by S. M. Katre and P. K. Gode, 351–55. Poona Oriental Series 75. Poona: Oriental Book Agency, 1941. Sukthankar, Vishnu S. “Arjunamiśra.” In Dr.  Modi Memorial Volume. Papers on Indo-Iranian and Other Subjects, edited by Dr.  Modi Memorial Volume Editorial Board, 565–68. Bombay: K. J. Dubash, 1930. ———. “Epic Studies V: Notes on Mahā bhā rata Commentators.” Annals of the Bhandarkar Oriental Research Institute 17, no. 2 (1935–36): 185–202. Commentators’ Editions of the Bhagavadgītā The works listed here appear in the notes to Appendix 17 (“Commentaries on the Bhagavadgītā ”). They are listed again here for convenience. Recent scholarship on developments in Bhagavadgītā criticism in the past 100 years is listed under “Overviews of Bhagavadgītā Scholarship.” Annangaracariyar, P. B., ed., Rāmānujagranthamāla. Kā ñcīpuram: Granthamā lā Oice, 1956. Belvalkar, S. K. Śrım ̄ ad-Bhagavad-Gıt̄ ā with the “Jñānakarmasamuccaya” Commentary of An̄ anda[vardhana], Edited from an unique Śārada MS, by S. K. Belvalkar, MA, PhD, with an Introduction Discussing the Problem of the “Kashmir Recension” and Two Appendices. Poona:  Bilvakuñja Publishing House, 1941. Chintamani, T. R., ed. Śrı̄madbhagavadgı̄tā, with Sarvatobhadra of Rājānaka Rāmakaṇtḥ a. Madras University Sanskrit Series no. 14. Madras: University of Madras, 1941. 502 502 PHILOLOGY AND CRITICISM Gambhīrā nanda, Swā mī, trans. Bhagavad Gı ̄tā: With the Commentary of Śaṅkarācār ya. Calcutta: Advaita Ashrama, 1995. Krishna Warrier, A. G., trans. Srı̄mad Bhagavad Gı̄tā Bhāṣya of Sri Saṃkarācār ya. Madras:  Sri Ramakrishna Math, 1983. Mainkar, T. G. A Comparative Study of the Commentaries on the Bhagavadgı̄tā. 2nd edn. Delhi: Motilal Banarsidass, 1969. Marjanovic, Boris, trans. Abhinavagupta’s Commentary on the Bhagavad Gita: Gı ̄tārtha-saṃgraha, Translated from Sanskrit with an Introduction and Notes. Varanasi: Indica Books, 2002; 2nd edn. 2004. Pansikar, Wasudev Laxman Sastri, ed. Śrı̄madbhagavadgı̄tā with the Commentaries of Śrı̄mat- Śaṅkarabhāṣya with An̄ andagiri; Nı̄lakaṇtḥ ı̄; Bhāṣyotkarṣadı ̄pikā of Dhanapti; Śrı̄dharı̄; Gı̄tārthasaṅgraha of Abhinavaguptāchār ya; and Gūḍhārthadı ̄pikā of Madhusūdana with Gūḍhārthatattvāloka. 2nd edn. Bombay: Nirṇaya-Sā gar Press, 1936. Reprint, Delhi: Munshiram Manoharlal, 1978. Raina, Lakshman (a.k.a. Lakshman Joo). Śrı̄madbhagavadgı̄tā with Commentary by Mahāmaheśvara Rājānaka Abhinavagupta. Srinagar: Kashmir Pratap Steam Press, 1933. Rao, Subba, trans. Bhagavadgı ̄tābhāṣya by Madhvācār ya. Madras: n.p., 1906. Sadhale, Shastri Gajanana Shambhu, ed. The Bhagavadgita with Eleven Commentaries. 2nd edn. 3 vols. Bombay: The “Gujarati” Printing Press, 1935. Reprint, Delhi: Parimal Publishers, 2000. Sampatkumaran, M. R., trans. The Gı ̄tābhāṣya of Rāmānuja. Madras:  Professor M.  Rangacharya Memorial Trust, 1969. Sastri, Alladi Mahadeva, trans. Gı ̄tā Bhāṣya of Śaṅkara. Madras: Samata Books, 1988. Sharma, Arvind, trans. Abhinavagupta: Gı ̄tārthasaṅgraha, Translated with an Introductory Study. Leiden: E. J. Brill, 1983. Shastri, Madhusudan Kaul, ed. The Bhagavadgı ̄tā with the Commentary Called Sarvatobhadra by Rājānaka Ramakaṇtḥ a. Kashmir Series of Texts and Studies 64. Srinagar:  Research and Publication Department, Srinagar, Jammu and Kashmir Government, 1943. Śrimadbhagavadgı̄tā. The Bhagavad-Gı̄tā: With the Commentary of Śankaracarya, Critically Edited by Dinkar Vishnu Gokhale. Poona Oriental Series 1. Poona: Oriental Book Agency, 1950. Subhadropadhyaya, D., ed. Bhagavadāśayānusaraṇābhidhānabhāṣyaṃ. Sarasvatī Bhavana Granthamā lā 94. Varanasi: Varanaseya Sanskrit Vishvavidyalaya, 1965. Swā mī Ā didevā nanda, trans. Śrı̄ Rāmānuja Gı̄tā Bhāṣya:  With Text and English. Madras:  Sri Ramakrishna Math, 1991. Tadpatrikar, Shriniwas Narayan, ed. Rājānakarāmakavikṛtasarvatobhadrākhyātı̣ k̄ āyutākāśmı̄rapātḥ ānusāriṇı ̄ śrı̄madbhagavadgı̄tā. Ā nandā sramasaṃskṛtagranthā valiḥ 112. Pune: Ā nandā śramamudraṇā laya, 1939. The Works of Sri Sankaracharya. 20 vols. Srirangam: Sri Vani Vilas Press, 1910. Introductions to Textual Criticism Besides the earlier manuals on textual criticism, this section lists some of the most useful current ones. West is invaluable for his conciseness and precision. Trovato combines a comprehensive over- view of the ield with a systematic approach that builds up to ever more complex problems. We relied extensively on both for this book. Most provides a good overview of the aims of textual crit- icism. Katre is the sole work on textual criticism of Indian texts. A work like Reynolds and Wilson, a standard reference work for the Greek and Latin classics, is badly needed for the Indian context. Bieler, Ludwig. The Grammarian’s Craft: An Introduction to Textual Criticism. 3rd edn. New York: The Catholic Classical Association, 1960. Havet, Louis. Manuel de critique verbale appliquée aux textes latins. Paris: Libraire Hachette, 1911. Joshi, Venkatesh Laxman, ed. Prauḍha Manoramā with Commentary Śabdaratna. Deccan College Monograph Series 31. Poona: Deccan College, 1966. Kantorowicz, Hermann. Einführung in die Textkritik: Systematische Darstellung der textkritischen Grundsätze für Philologen und Juristen. Leipzig: Dieterich, 1921. 503 ANNOTATED BIBLIOGRAPHY 503 Katre, S. M. Introduction to Indian Textual Criticism. With Appendix II by P.  K. Gode. Deccan College Handbook Series 5. Pune: Deccan College, 1954. Maas, Paul. Textkritik. 4th edn. Leipzig: B. G. Teubner, 1960. ———. Critica del testo. Translated by Nello Martinelli. Florence: Le Monnier, 1972. Most, Glenn W. “Editor’s Introduction.” In Sebastiano Timpanaro, The Genesis of Lachmann’s Method.  Edited and translated by Glenn W.  Most. 1–32. Chicago:  University of Chicago Press, 2005. Reynolds, Leighton Durham. Texts and Transmission: A Survey of the Latin Classics. Oxford: Clarendon Press, 1983. Reynolds, L. D., and N. G. Wilson. Scribes and Scholars: A Guide to the Transmission of Greek and Latin Literature. Oxford: Clarendon Press, 1991. Stählin, Otto. Editionstechnik, Completely Revised Second Edition. Leipzig: B. G. Teubner, 1914. Trovato, Paolo. Everything You Always Wanted to Know about Lachmann’s Method: A Non-Standard Handbook of Genealogical Textual Criticism in the Age of Post-Structuralism, Cladistics, and Copy-Text. Padua: Lib reriauniversitaria, 2014. West, M. L. Textual Criticism and Editorial Technique. Stuttgart: B. G. Teubner, 1973. Advanced Works in Textual Criticism The more advanced works here deal either with objections to the genealogical-reconstructive method or special applications of it. Also listed are works addressing the problem of contamina- tion. A few, such as Timpanaro, are concerned with the history of the method, but are also useful as a guide to the method itself. Gane and Soulen are Soulen are two useful lexica. Avalle, D’Arco Silvio. “Di alcuni rimedi contra la ‘contaminazione’. Saggio di applicazione alla tradizione manoscritta di Rigaut de Berbezilh.” In La letteratura medievale in lingua d’oc nella sua tradizione manoscritta. 159–78. Turin: Einaudi, 1961. Bornmann, Fritz. “Sui criteri di ‘recensio’ meccanica enunciati da Lachmann nel 1817.” Rivista di Letterature Moderne e Comparate 15 (1962): 45–53. de Contenson, Pierre Marie. “Principles, Methods and Problems of the Critical Edition of the Works of Saint Thomas as Presented in the ‘Leonine Edition’.” Tijdschrift voor Filosoie 36, no. 2 (1974): 342–64. Dearing, Vinton A. “Some Notes on Genealogical Methods in Textual Criticism.” Novum Testamentum 9, no. 4 (1967): 278–97. Eklund, Sten. “The Traditional or the Stemmatic Editorial Technique?” Eranos 104 (2006–7): 5–18. Froger, Jacques. La critique des textes et son automatisation. Paris: Dunod, 1968. Gane, Yorick Gomez. Dizionario della terminologia ilologica. Turin: Accademia University Press, 2013. Greetham, D. C.  “Criticizing the Text:  Textual Criticism.” In Textual Scholarship. An Introduction. 295–346. New York: Garland, 1994. Greg, W. W. The Calculus of Variants: An Essay on Textual Criticism. Oxford: Clarendon Press, 1927. Irigoin, Jean. “Quelques rélexions sur le concept d’archétype.” Revue d’histoire des textes 7 (1977): 235–45. ———. La tradition des textes grecs: Pour une critique historique. Paris: Les Belles Lettres, 2003. Lachmann, Karl. “Der Nibelungen Lied, herausg. von v. d. Hagen. Breslau 1816.” In Kleinere Schriften zur deutschen Philologie. Vol. 1. Edited by Karl Müllenhof. 81–114. Berlin: G. Reimer, 1876. Lejay, Paul. “Revue critique d’histoire et du littérature.” N. S. XXVI (1888): 281–83. Maas, Paul. “Leitfehler und stemmatische Typen.” Byzantinische Zeitschrift 37 (1937): 289–94. Marichal, Robert. “La critique des textes.” In L’histoire et ses méthodes, edited by Charles Samaran, 1246–366. Paris: Éditions Gallimard, 1961. Pasquali, Giorgio. Review of La critique des textes, by P. Collomp. Gnomon 8 (1932): 128–34. ———. Review of Textkritik, by Paul Maas. Gnomon 5, no. 8 (1929): 417–35. ———. Review of Textkritik, by Paul Maas. Gnomon 5, no. 9 (1929): 498–521. 504 504 PHILOLOGY AND CRITICISM Quentin, Henri. Essais de critique textuelle (Ecdotique). Paris: Picard, 1926. Van Reenen, Pieter, and Margot van Mulken, eds. Studies in Stemmatology. Amsterdam and Philadelphia, PA: John Benjamin Publishing Company, 1996. Van Reenen, Pieter, August den Hollander and Margot van Mulken, eds. Studies in Stemmatology II. Amsterdam and Philadelphia, PA: John Benjamin Publishing Company, 2004. Rizzo, Silvia. Review of Il lapsus freudiano. Psicanalisi e critica testuale, by Sebastiano Timpanaro. Rivista di ilologia e di istruzione classica (1977): 102–5. Schmidt, P. L. “Lachmann’s Method: On the History of a Misunderstanding.” In The Uses of Greek and Latin: Historical Essays, edited by A. C. Dionisotti, Anthony Grafton and Jill Kraye, 227–36. London: The Warburg Institute, University of London, 1988. Segre, Cesare. “Appunti sul problema delle contaminazioni nei testi in prosa.” In Studi i problemi di critica testuale. Convegno di Studi di Filologia Italiana nel Centenario della Commissione per i testi di lingua. 63–67. Bologna: Commissione per i testi di lingua, 1960. Shillingsburg, Peter L. Scholarly Editing in the Computer Age: Theory and Practice. Ann Arbor: University of Michigan Press, 1996. Soulen, Richard N., and R. Kendall Soulen. Handbook of Biblical Criticism. 4th edn. Louisville, KY: Westminster John Knox Press, 2011. Timpanaro, Sebastiano. The Genesis of Lachmann’s Method. Edited and translated by Glenn W. Most. Chicago: University of Chicago Press, 2005. ———. Per la storia della ilologia virgiliana antica. Rome: Salerno Editrice, 1984. Varvaro, Alberto. “Considerazioni sulla contaminazione, sulle varianti adiafore e sullo stemma codicum.” In Storia della lingua italiana e ilologia. Atti del vii Convegno asli Associazione per la Storia della lingua italiana, edited by Claudio Ciociola, 191–96. Florence: Cesati, 2010. Wolf, F. A. Prolegomena to Homer. Translated with an introduction and notes by A. Grafton, G. W. Most and J. E. G. Zetzel. Princeton, NJ: Princeton University Press, 1985. Problems in Textual Criticism/Computer-Aided Analysis Several works listed here are concerned with the overrepresentation of bipartite stemma in crit- ical editions (often called “Bédier’s paradox” after Joseph Bédier, who irst articulated the problem in 1928). Others concern probabilistic calculations of the type:  given x manuscripts, what are the stemmatic combinations possible, or, given x manuscripts at a certain point of time, what are the chances that a certain percentage survived and so on. Though not essential to understanding textual criticism, they sharpen the reader’s sense for its probabilistic nature. For obvious reasons, these problems often invite and sometimes demand computer-aided analysis. For understanding how problems can be solved using intuitive logic and a sense for the possible outcomes given certain premises, Reeve’s articles are unequalled. They should be required reading for everyone interested in learning to think abstractly about stemmata. Bédier, Joseph. “La tradition manuscrite du Lai de l’ombre. Rélexions sur l’art d’éditer les anciens textes.” Romania 54 (1928): 161–96, 321–58. Flight, Colin. “A Complete Theoretical Framework for Stemmatic Analysis.” Manuscripta 38, no. 2 (1994): 95–115. ———. “How Many Stemmata?” Manuscripta 34, no. 2 (1990): 122–28. ———. “Stemmatic Theory and the Analysis of Complicated Traditions.” Manuscripta 36, no. 1 (1992): 37–52. Grier, James. “Lachmann, Bédier and the Bipartite Stemma: Towards a Responsible Application of the Common-Error Method.” Revue d’Histoire des Textes 18 (1988): 263–78. Guidi, Vincenzo, and Paolo Trovato. “Sugli stemmi bipartiti. Decimazione, assimmetria e calcolo delle probabilitità.” Filologia italiana 1 (2004): 9–48. Hall, J. B. “Why Are the Stemmata of So Many Manuscript Traditions Bipartite?” Liverpool Classical Monthly 17 (1992): 31–32. 50 ANNOTATED BIBLIOGRAPHY 505 Hering, Wolfgang. “Zweispaltige Stemmata: Zur Theorie der textkritischen Methode.” Philologus 111 (1967): 170–85. Irigoin, Jean. “Stemmas biides et états de manuscrits.” Revue de Philologie, third series, 28 (1954): 211–17. Kleinlogel, Alexander. “Das Stemmaproblem.” Philologus 112 (1968): 63–82. Najock, D., and C. C. Heyde. “On the Number of Terminal Vertices in Certain Random Trees with an Application to Stemma Construction in Philology.” Journal of Applied Probability 19, no. 3 (1982): 675–80. Poole, Eric. “The Computer in Determining Stemmatic Relationships.” Computers and the Humanities 8, no. 4 (1974): 207–16. Reeve, Michael D. “Archetypes.” In Manuscripts and Methods: Essays on Editing and Transmission. 107– 17. Rome: Edizioni di Storia e Letteratura, 2011. ———. “Classical Scholarship in the Renaissance.” In Manuscripts and Methods: Essays on Editing and Transmission. 255–81. Rome: Edizioni di Storia e Letteratura, 2011. ———. “Cuius in Usum? Recent and Future Editing.” The Journal of Roman Studies 90 (2000): 196–206. ———. “Editing Classical Texts with a Computer: Hyginus’s Astronomica.” In Manuscripts and Methods: Essays on Editing and Transmission. 361–93. Rome: Edizioni di Storia e Letteratura, 2011. ———. “Eliminatio codicum descriptorum: A Methodological Problem.” In Editing Greek and Latin Texts, edited by J. N. Grant, 1–35. New York: AMS Press, 1989. ———. “Errori in Autograi.” In Manuscripts and Methods: Essays on Editing and Transmission. 3–23. Rome: Edizioni di Storia e Letteratura, 2011. ———. “Gruppenarbeit an Handschriften.” In Manuscripts and Methods:  Essays on Editing and Transmission. 315–21. Rome: Edizioni di Storia e Letteratura, 2011. ———. “A Man on a Horse.” In Manuscripts and Methods: Essays on Editing and Transmission. 211–19. Rome: Edizioni di Storia e Letteratura, 2011. ———. “Reconstructing Archetypes:  A New Proposal and an Old Fallacy.” In Hesperos:  Studies in Ancient Greek Poetry Presented to M. L. West on His Seventieth Birthday, edited by P. J. Finglass, C. Collard and N. J. Richardson, 326–40. Oxford: Oxford University Press, 2007. ———. “The Rediscovery of Classical Texts in the Renaissance.” In Manuscripts and Methods: Essays on Editing and Transmission. 229–54. Rome: Edizioni di Storia e Letteratura, 2011. ———. “Shared Innovations, Dichotomies, and Evolution.” In Manuscripts and Methods: Essays on Editing and Transmission. 55–103. Rome: Edizioni di Storia e Letteratura, 2011. ———. “Some Applications of Pasquali’s ‘Criterio Geograico’ to 15th-Century Latin Manuscripts.” In Manuscripts and Methods: Essays on Editing and Transmission. 221–28. Rome: Edizioni di Storia e Letteratura, 2011. ———. “Stemmatic Method: ‘Qualcosa che non funziona?’ ” In The Role of the Book in Medieval Culture: Proceedings of the Oxford International Symposium, September 26–October 1, 1982, edited by P. Ganz, 57–70. Bibliologia 3. Turnhout: Brepols, 1986. Salemans, Benedictus Johannes Paulus. “Building Stemmas with the Computer in a Claddistic, Non-Lachmannian Way:  The Case of Fourteen Versions of ‘Lanseloet van Denemerken.’ ” PhD Diss., Katholieke Universiteit Nijmegen, 2000. www.neder-l.nl/salemans/diss/salemans- diss-2000.pdf. Spencer, Matthew, and Christopher J. Howe. “Collating Texts Using Progressive Multiple Alignment.” Computers and the Humanities 38, no. 3 (2004): 253–70. ———. “Estimating Distances between Manuscripts Based on Copying Errors.” Literary and Linguistic Computing 16, no. 4 (2001): 467–84. Timpanaro, Sebastiano. “Ancora su stemmi bipartiti e contaminazione.” Maia 17 (1965): 392–99. ———. “‘Recentiores’ e ‘deteriores,’ ‘codices descripti’ e ‘codices inutiles.’ ” Filologia e critica 10 (1985): 164–92. 506 506 PHILOLOGY AND CRITICISM Tov, Emmanuel. “Computer-Assisted Tools for Textual Criticism.” In Tradition and Innovation in Biblical Interpretation: Studies Presented to Professor Eep Talstra on the Occasion of His Sixty-Fifth Birthday, edited by W. Th. van Peursen and J. W. Dyk, 245–60. Leiden: E. J. Brill, 2011. Trovato, Paolo. “Bédier’s Contribution to the Accomplishment of Stemmatic Method: An Italian Perspective.” Textual Cultures 9, no. 1, Texts, Contexts, Interpretation (2014): 160–76. ———. “La tradizione manoscritta del ‘Lai de l’ombre’:  Rilessioni sulle tecniche d’edizione primonovecentesche.” Romania 131 (2013): 338–80. Weitzman, Michael P. “The Analysis of Open Traditions.” Studies in Bibliography 38 (1985): 82–120. ———. “Computer Simulation of the Development of Manuscript Traditions.” Bulletin of the Association for Literary and Linguistic Computing 10, no. 2 (1982): 55–59. ———. “The Evolution of Manuscript Traditions.” Journal of the Royal Statistical Society 150, no. 4 (1987): 287–308. Whitehead, Frederick, and Cedric E. Pickford. “The Introduction to the Lai de l’ombre: Sixty Years Later.” Romania 94 (1973): 145–56. Windram, Heather F., Matthew Spencer and Christopher J. Howe. “Phylogenetic Analysis of Manuscript Traditions, and the Problem of Contamination.” Linguistica computazionale 24–25 (2004–5): 141–55. Theoretical Perspectives, Romance Philology and Italian Textual Criticism Beginning with Pasquali, Italian editors developed a rich and parallel tradition of textual criticism that, although imbibing the lessons of Lachmannism, characteristically pays greater attention to the history of the text, its reception and its variations. Somewhat simplistically, one might say that whereas German textual criticism is shaped by its Protestantism (the principle of sola scriptura, a desire to purify the receptus and the desire for a pure text unmediated by an interpretive tradition), Italian textual criticism is still a child of its Renaissance heritage. No one should approach textual criticism without studying the developments over a narrow Lachmannism from the South. For obvious reasons, this tradition is more relevant to the Mahā bhā rata, where Protestant prejudices presented a major obstacle to the work’s correct interpretation and appreciation. Pugliatti and Varvaro ofer good introductions to the distinctive features of this tradition. Dembowski and Altschul discuss the related French tradition. Dante ofers an introduction to Contini, perhaps the most important Italian theoretician of the past century. Avalle, D’Arco Silvio. Principî di critica testuale. Padua: Antenore, 1972. Barbi, Michele. La nuova ilologia e l’edizione dei nostri scrittori da Dante al Manzoni. Florence: Sansoni, 1938. Berdozzo, Fabio. “Platon, Wilamowitz und Giorgio Pasquali.” Hermes 133, no. 2 (2005): 245–55. Contini, Gianfranco. “Come lavorava l’Ariosto.” In Esercizi di lettura sopra autori contemporanei con un’appendice su testi non contemporanei. 2nd edn. 232–41. Turin: Einaudi, 1982. ———. “La critica testuale come studio di strutture (1971).” In Frammenti di ilologia romanza: Scritti di ecdotica e linguistica (1932–1989). Edited by Giancarlo Breschi. Vol. 1. 63–74. Florence: Edizioni del Galluzo, 2007. ———. “Filologia (1977).” In Frammenti di ilologia romanza: Scritti di ecdotica e linguistica (1932–1989). Edited by Giancarlo Breschi. Vol. 1. 3–62. Florence: Edizioni del Galluzo, 2007. ———. “Rapporti fra la ilologia (come critica testuale) e la linguistica romanza (1970).” In Frammenti di ilologia romanza:  Scritti di ecdotica e linguistica (1932–1989). Edited by Giancarlo Breschi. Vol. 1. 75–97. Florence: Edizioni del Galluzo, 2007. ———. “Ricordo di Joseph Bedier.” In Esercizi di lettura sopra autori contemporanei con un’appendice su testi non contemporanei. 358–72. Turin: Einaudi, 1974. ———. “La ‘Storia della tradizione e critica del testo’ di Giorgio Pasquali (1935).” In Frammenti di ilologia romanza: Scritti di ecdotica e linguistica (1932–1989). Edited by Giancarlo Breschi. Vol. 1. 99–112. Florence: Edizioni del Galluzo, 2007. 507 ANNOTATED BIBLIOGRAPHY 507 ———. “La ‘Vita’ francese ‘di Sant’Alessio’ e l’arte di pubblicare i testi antichi (1970).” In Frammenti di ilologia romanza: Scritti di ecdotica e linguistica (1932–1989). Edited by Giancarlo Breschi. Vol. 2. 957–85. Florence: Edizioni del Galluzo, 2007. Dembowski, Peter F. “The ‘French’ Tradition of Textual Philology and Its Relevance to the Editing of Medieval Texts.” Modern Philology 90, no. 4 (1993): 512–32. Durante, Erica “Sous la rature, la littérature. L’expérience de la philologie italienne au service de la littérature comparée.” TRANS—Révue de littérature générale et comparée 2 (2006). http://trans. revues.org/171 (accessed October 13, 2005). Fahy, Conor. “Old and New in Italian Textual Criticism.” In Voice, Text, Hypertext: Emerging Practices in Textual Studies, edited by Raimonda Modiano, Leroy F.  Searle and Peter L.  Shillingsburg, 401–11. Seattle and London: University of Washington Press, 2004. Fiesoli, Giovanni. La genesi del lachmannismo. Florence: Edizioni del Galluzzo, 2000. Flores, Enrico. Elementi critici di critica del testo ed epistemologia. Naples: Lofredo, 1998. Fraenkel, Eduard. Review of Storia dello spirito tedesco nelle memorie d’un contemporaneo, by Giorgio Pasquali. Gnomon 26, no. 5 (1954): 337–41. Isella, Dante. “Contini e la critica delle varianti.” Filologia e critica 15 (1991): 281–97. Kenney, E. J. “Recentiores non Deteriores.” Review of A. Persi, Flacci, Saturarum liber, ed. W. V. Clavsen. The Classical Review n.s., 7, no. 3/4 (1957): 226–28. Klingner, Friedrich. “Giorgio Pasquali.” Gnomon 25, no. 1 (1953): 60–62. Leonardi, Lino. “Il testo come ipotesi (critica del manoscritto-base).” Medioevo romanzo 35 (2011): 5–34. Lernout, Geert. “Genetic Criticism and Philology.” Text 14 (2002): 53–75. Montanari, Elio. La critica del testo secondo Paul Maas: testo e commento. Florence: SISMEL edizioni del Galluzzo, 2003. Nadia, Altschul. “The Genealogy of Scribal Versions:  A ‘Fourth Way’ in Medieval Editorial Theory.” Textual Cultures 1, no. 2 (2006): 114–36. Pasquali, Giorgio. Storia della tradizione e critica del testo. Florence: Le Monnier, 1952. ———. “Arte Allusiva.” In Stravaganze quarte e supreme, 11–20. Venice: Pozza, 1951. Perrotta, Gennaro. “Intelligenza di Giorgio Pasquali.” Quaderni Urbinati di Cultura Classica, n.s., 21, no. 3 (1985): 7–12. Pugliatti, Paola. “Textual Perspectives in Italy: From Pasquali’s Historicism to the Challenge of ‘Variantistica’ (And Beyond).” Text 11 (1998): 155–88. Renzi, Lorenzo and Alvise Andreose. Manuale di linguistica e ilologia romanza. Bologna: il Mulino, 2009. Seel, Otto. Review of Storia della tradizione e critica del testo, by Giorgio Pasquali. Gnomon 12, no. 1 (1936): 16–30. Segre, Cesare. “Critique des variantes et critique génétique.” Genesis 7 (1994): 29–46. Stussi, Alfredo. Introduzione agli studi di ilologia italiana. 4th edn. Bologna: il Mulino, 2011. Ullman, B. L. Review of Storia della tradizione e critica del testo, by Giorgio Pasquali. Classical Philology 32, no. 4 (1937): 371–73. Varvaro, Alberto. “Elogio della copia.” In Atti del XXI Congresso Internazionale di Linguistica e Filologia Romanza, vol. VI, edited by Giovanni Ruino, 785–96. Tübingen: Niemeyer, 1998. ———. “The ‘New Philology’ from an Italian Perspective.” Translated by Marcello Cherchi. Text 12 (1999): 49–58. Zaccarello, Michelangelo. “Metodo stemmatico ed ecdotica volgare italiana. Brevi considerazioni su alcuni recenti contributi metodologici.” Textual Cultures 4, no. 1 (2009): 55–71. Discussions of the Mahā bhā rata Critical Edition There are very few sensible discussions of the Mahā bhā rata critical edition. Sukthankar’s “Prolegomena” remains the best introduction, but it is daunting in its technical brilliance. Dunham’s shorter essay is an excellent summary. Some of the articles listed in this section are technically 508 508 PHILOLOGY AND CRITICISM incorrect, but the truly problematic works are listed later. Biardeau’s and Bedekar’s articles preserve an important exchange concerning the value of the vulgate and historical methods in the study of the epic. Phillips-Rodriguez and colleagues apply methods of computer analysis to the critical edition. Adarkar, Aditya. “Turning a Tradition into a Text: Critical Problems in Editing the Mahabharata.” In Theoretical Approaches to the Transmission and Edition of Oriental Manuscripts, Papers of the Symposium Organized in Istanbul, Turkey, 2001, edited by Judith Pfeifer and Manfred Kropp, 133–46. Beiruter Texte und Studien 111. Wurzburg: Ergon, 2007. Adluri, Vishwa. “Introduction.” In Ways and Reasons for Thinking about the Mahābhārata as a Whole,  edited by Vishwa Adluri, vii–xxxii. Pune:  Bhandarkar Oriental Research Institute, 2013. ———. “The Critical Edition and Its Critics: A Retrospective of Mahābhārata Scholarship.” Journal of Vaishnava Studies 19, no. 2 (2011): 1–21. Austin, Christopher. “Evaluating the Critical Edition of the Mahā bhā rata: Inferential Mileage and the Apparatus Materials.” Journal of Vaishnava Studies 19, no. 2 (2011): 71–86. Bagchee, Joydeep. “The Epic’s Singularization to Come: The Ś akuntalā and Yayā ti Upā khyā nas in the Southern Recension of the Mahā bhā rata.” In Argument and Design:  The Unity of the Mahābhārata, edited by Vishwa Adluri and Joydeep Bagchee, 83–126. Leiden and Boston. MA: Brill, 2016. Bedekar, V. M. “Principles of Mahā bhā rata Textual Criticism:  The Need for a Restatement.” Purāṇa 11, no. 2 (1969): 210–28. Bhate, Saroja. “Methodology of the Critical Edition of the Mahābhārata.” In Mahābhārata Now:  Narration, Aesthetics, Ethics, edited by Arindam Chakrabarti and Sibaji Bandyopadhyay, 29–36. New Delhi and Abingdon: Routledge, 2014. Biardeau, Madeleine. “Letter to the Editor—Dr. Madeleine Biardeau’s Letter in Reply to Sri V. M. Bedekar’s Article.” Purāṇa 12, no. 1 (1970): 180–81. ———. “Some More Considerations about Textual Criticism.” Purāṇa 10, no. 2 (1968): 115–23. Dunham, John. “Manuscripts Used in the Critical Edition of the Mahā bhā rata:  A Survey and Discussion.” In Essays in the Mahābhārata, edited by Arvind Sharma, 1–18. Leiden: Brill, 1991. Hiltebeitel, Alf. “On Sukthankar’s ‘S’ and Some Shortsighted Assessments and Uses of the Pune Critical Edition (CE).” Journal of Vaishnava Studies 19, no. 2 (2011): 87–126. Metzger, Bruce M. “Trends in the Textual Criticism of the Iliad, the Mahabharata, and the New Testament.” Journal of Biblical Literature 65, no. 4 (1946): 339–52. Phillips-Rodriguez, Wendy J. “The Evolution of a Sanskrit Epic: Some Genetic Considerations about Scripts.” In The Evolution of Texts: Confronting Stemmatological and Genetical Methods, edited by C. Macé, P. Baret, A. Bozzi and L. Cignoni, 175–90. Linguistica Computazionale 24–25. Pisa: Istituti editoriali e poligraici internazionali, 2006. Phillips-Rodriguez, Wendy, C. J. Howe and H. F. Windram. “Chi-Squares and the Phenomenon of Change of Exemplar in the Dyū taparvan.” In Sanskrit Computational Linguistics, edited by Gérard Huet, Amba Kulkarni and Peter Scharf, 380–90. Heidelberg: Springer, 2009. ———. “Some Considerations about Bifurcation in Diagrams Representing the Written Transmission of the Mahā bhā rata.” Wiener Zeitschrift für die Kunde Südasiens 52–53, Text Genealogy, Textual Criticism and Editorial Technique (2009–10): 29–43. Proudfoot, Ian. Ahiṃsā and a Mahābhārata Story:  The Development of the Story of Tulādhāra in the Mahābhārata in Connection with Non-violence, Cow Protection and Sacriice. Asian Studies Monographs, n.s., 9. Canberra: Faculty of Asian Studies, Australian National University, 1987. ———. “Interpreting Mahābhārata Episodes as Sources for the History of Ideas.” Annals of the Bhandarkar Oriental Research Institute 60, no. 1/4 (1979): 41–63. 509 ANNOTATED BIBLIOGRAPHY 509 Discussions of the Mahā bhā rata and the Mahā bhā rata Tradition Masses have been written on the Mahā bhā rata since the early nineteenth century. Most of these works, as Sukthankar predicted in 1942, are only of antiquarian interest now. The Nay Science is the best guide currently available to the history of Mahā bhā rata studies. Because of the great variation in literature, we have subdivided the category by themes. Overviews of Mahā bhā rata Scholarship The Nay Science covers the irst phase of “critical” Mahā bhā rata studies, and thus complements the present book. Sullivan contains a balanced appraisal of recent literature. Brodbeck has a good account of why Euro-American Mahābhārata studies developed as they did. Sukthankar is a classic work that defends the Mahā bhā rata’s integrity and traditional reception against the German critics’ specious theory of a Kṣatriya “Ur-epos.” Adarkar, Aditya. “The Mahā bhā rata and Its Universe: New Approaches to the All-Encompassing Epic.” History of Religions 47, no. 4 (2008): 304–19. Adluri, Vishwa, and Joydeep Bagchee. The Nay Science: A History of German Indology. New York: Oxford University Press, 2014. Brodbeck, Simon. “Analytical and Synthetic Approaches in Light of the Critical Edition of the Mahā bhā rata and the Harivaṃśa.” Journal of Vaishnava Studies 19, no. 2 (2011): 223–50. ———. Review of Strı̄: Women in Epic Mahābhārata, by Kevin McGrath. Indo-Iranian Journal 53, no. 1 (2010): 89–94. Sukthankar, Vishnu S. On the Meaning of the Mahābhārata. Bombay: Asiatic Society, 1957. Sullivan, Bruce M. “An Overview of Mahā bhā rata Scholarship: A Perspective on the State of the Field.” Religion Compass 10, no. 7 (2016): 165–75. Overviews of Bhagavadgītā Scholarship The Nay Science, “Paradigm Lost” and “Who’s Zoomin’ Who” provide a comprehensive overview of historical-critical research on the Bhagavadgītā , the main current in contemporary scholar- ship. The earlier Callewaert and Hemraj is still useful as a resource. Malinar encapsulates the German approach to the Gītā but is now superseded. Adluri, Vishwa. Review of The Bhagavadgı̄tā: Doctrines and Contexts, by Angelika Malinar. History of Religions 50, no. 1 (2010): 102–07. Adluri, Vishwa, and Joydeep Bagchee. The Nay Science: A History of German Indology. New York: Oxford University Press, 2014. ———. “Paradigm Lost: The Application of the Historical-Critical Method to the Bhagavadgītā .” International Journal of Hindu Studies 20, no. 2 (2016): 199–301. ———. “Who’s Zoomin’ Who:  Bhagavadgītā Recensions in India and Germany.” International Journal of Dharma Studies 4, no. 4 (2016): 1–41. Callewaert, W. M., and Shilanand Hemraj. Bhagavadgı ̄tānuvāda:  A Study in Intercultural Translation. Ranchi: Satya Bharati Publication, 1983. Malinar, Angelika. The Bhagavadgıt̄ ā:  Doctrines and Contexts. Cambridge:  Cambridge University Press, 2007. Philosophical Interpretations Biardeau is a rare example of a Western scholar who appreciated the Mahā bhā rata as a work of philosophy. The other works listed here are indebted to her insights. The edited volume Argument and Design represents new direction in Mahā bhā rata studies after The Nay Science. 510 510 PHILOLOGY AND CRITICISM Adluri, Vishwa. “The Divine Androgyne:  Crossing Gender and Breaking Hegemonies in the Ambā Upā khyā na of the Mahā bhā rata.” In Argument and Design:  The Unity of the Mahābhārata, edited by Vishwa Adluri and Joydeep Bagchee, 275–319. Leiden and Boston, MA: Brill, 2016. ———. “Ethics and Hermeneutics in the Mahā bhā rata.” Review of Disorienting Dharma:  The Aesthetics of Sufering in the Mahābhārata, by Emily Hudson. International Journal of Hindu Studies 20, no. 3 (2016): 385–92. ———. “Hindu Studies in a Christian, Secular Academy.” International Journal of Dharma Studies 5, no. 1 (2017). doi:10.1186/s40613-016-0037-5. ———. “Literary Violence and Literal Salvation:  Ś aunaka Interprets the Mahābhārata.” Exemplar: The Journal of South Asian Studies 1, no. 2 (2012): 45–68. ———. “The Perils of Textual Transmission: Decapitation and Recapitulation.” Seminar 608, The Enduring Epic (2010): 48–54. ———. “Philosophical Aspects of Bhakti in the Nā rā yaṇīya.” In The Churning of the Epics and Purāṇas at the Fifteenth World Sanskrit Conference, edited by Simon Brodbeck, Alf Hiltebeitel and Adam Bowles. 127–54. New Delhi: Dev Publishers, 2018. Adluri, Vishwa, and Joydeep Bagchee, eds. Argument and Design: The Unity of the Mahābhārata. Leiden and Boston, MA: Brill, 2016. Biardeau, Madeleine. Études de mythologie hindoue I:  Cosmogonies purāṇiques. Paris:  l’École française d’Extrême-Orient, 1981. ———. Études de mythologie hindoue II: Bhakti et avatāra. Pondichéry: Publications de l’École Française d’Extrême Orient, 1994 ———. Hinduism:  Anthropology of a Civilization. Translated by Richard Nice. Delhi:  Oxford University Press, 1989. ———. Théorie de la connaissance et philosophie de la parole dans le brahmanisme classique. The Hague: Mouton, 1964. ———. “The Two Sanskrit Epics Reconsidered.” In Studies in Hinduism: Vedism and Hinduism, edited by Gerhard Oberhammer, 73–119. Vienna: Austrian Academy of Sciences, 1991. ———. “Nara et Nā rā yaṇa.” Wiener Zeitschrift für die Kunde Südasiens 35 (1991): 75–108. Bowles, Adam. Dharma, Disorder and the Political in Ancient India:  The Aˉpaddharmaparvan of the Mahābhārata. Leiden and Boston, MA: Brill, 2007. ———. “Framing Bhīṣma’s Royal Instructions: The Mahābhārata and the Problem of Its ‘Design’.” In Parallels and Comparisons. Proceedings of the Fourth Dubrovnik International Conference on the Sanskrit Epics and Purāṇas, edited by Petteri Koskikallio, 121–35. Zagreb: Croatian Academy of Sciences and Arts, 2009. Dahlmann, Joseph. Das Mahâbhârata als Epos und Rechtsbuch. Ein Problem aus Altindiens Cultur- und Literaturgeschichte. Berlin: Felix L. Dames, 1895. ———. Die Genesis des Mahâbhârata. Berlin: Felix L. Dames, 1899. Hiltebeitel, Alf. “Two Kṛṣṇas on One Chariot: Upaniṣadic Imagery and Epic Mythology.” History of Religions 24, no. 1 (1984): 1–26. Sullivan, Bruce M. “The Epic’s Two Grandfathers, Bhīṣma and Vyā sa.” In Essays on the Mahābhārata, edited by Arvind Sharma, 204–11. Delhi: Motilal Banarsidass, 2007. ———. Kṛṣṇa Dvaipāyana Vyāsa and the Mahābhārata: A New Interpretation. Leiden: E. J. Brill, 1990. ———. “The Religious Authority of the Mahābhārata: Vyā sa and Brahmā in the Hindu Scriptural Tradition.” Journal of the American Academy of Religion 62, no. 2 (1994): 377–401. Oral Epics, Metrical and Statistical Analysis, Search for the Heroic Epic Metrical and statistical analyses characteristically aim to identify the oral, bardic epic assumed as the Mahā bhā rata’s “core.” Their usefulness is limited, since they run into the charge of circu- larity. Brockington’s The Sanskrit Epics presents a summary of the Sanskrit epics’ development based on German views of the past two centuries. 51 ANNOTATED BIBLIOGRAPHY 511 Brockington, John. The Sanskrit Epics. Leiden: E. J. Brill, 1998. ———. “The Spitzer Manuscript and the Mahā bhā rata.” In From Turfan to Ajanta:  Festschrift for Dieter Schlinglof on the Occasion of His Eightieth Birthday, edited by Eli Franco and Monika Zin, 75–87. Lumbini: Lumbini International Research Institute, 2010. ———. “The Textualization of the Sanskrit Epics.” In Textualization of Oral Epics, edited by Lauri Honko, 193–216. Berlin and New York: Mouton de Gruyter, 2000. Carroll Smith, Mary. “The Mahā bhā rata’s Core.” Journal of the American Oriental Society 95, no. 3 (1975): 479–82. Edgerton, Franklin. “The Epic Triṣṭubh and Its Hypermetric Varieties.” Journal of the American Oriental Society 59, no. 2 (1939): 159–74. Fitzgerald, James L. “The Rā ma Jā madagnya ‘Thread’ of the Mahā bhā rata: A New Survey of Rā ma-Jā madagnya in the Pune Text.” In Stages and Transitions: Temporal and Historical Frameworks in Epic and Purāṇic Literature. Proceedings of the Second Dubrovnik International Conference on the Sanskrit Epics and Purāṇas, edited by Mary Brockington, 89–132. Zagreb: Croatian Academy of Sciences and Arts, 2002. ———. “Toward a Database of the Non-Anuṣṭubh Verses in the Mahā bhā rata.” In Epics, Khilas, and Purāṇas: Continuities and Ruptures. Proceedings of the Third Dubrovnik International Conference on the Sanskrit Epics and Purāṇas, edited by Petteri Koskikallio, 137–48. Zagreb: Croatian Academy of Sciences and Arts, 2005. ———. “Triṣtụ bh Passages of the Mahābhārata.” In Epic Undertakings: Papers of the 12th World Sanskrit Conference, vol. 2, edited by Robert P. Goldman and Muneo Tokunaga, 95–117. Delhi: Motilal Banarsidass Publishers, 2009. Hopkins, Edward W. “The Bhā rata and the Great Bhā rata.” The American Journal of Philology 19, no. 1 (1898): 1–24. ———. The Great Epic of India: Character and Origin of the Mahābhārata. New York: Charles Scribner’s Sons, 1901. Van Nooten, Barend A. “Redundancy in Mahā bhā rata Verse Composition.” Journal of the American Oriental Society 89, no. 1 (1969): 50–58. Schlinglof, Dieter. “The Oldest Extant Parvan-List of the Mahā bhā rata.” Journal of the American Oriental Society 89 (1969): 334–38. Yardi, M. R. “The Multiple Authorship of the Mahā bhā rata: A Statistical Approach (Paper III).” Annals of the Bhandarkar Oriental Research Institute 62, no. 1/4 (1981): 49–65. ———. “The Multiple Authorship of the Mahā bhā rata:  A Statistical Approach (Paper IV).” Annals of the Bhandarkar Oriental Research Institute 63, no. 1/4 (1982): 181–93. ———. “The Multiple Authorship of the Mahā bhā rata: A Statistical Approach (Paper V). Annals of the Bhandarkar Oriental Research Institute 64, no. 1/4 (1983): 35–58. ———. “The Multiple Authorship of the Mahā bhā rata:  A Statistical Approach (Paper VI).” Annals of the Bhandarkar Oriental Research Institute 65, no. 1/4 (1984): 59–80. Histories and Historical Reconstructions The “histories” in this section should be used with caution. Many are speculative and prejudicial. Bühler and Kirste are a welcome exception. Coomaraswamy addresses the sense in which the Mahābhārata is a “history.” Bühler, Georg, and J.  Kirste. “Indian Studies. No. II. Contributions to the History of the Mahâbhârata.” Sitzungsberichte der Philosophisch-Historischen Classe der Kaiserlichen Akademie der Wissenschaften 127. Vienna: F. Tempsky, 1892. Van Buitenen, J. A. B. “The Mahābhārata: Introduction.” In J. A. B. van Buitenen, trans. The Mahābhārata. Vol. 1: The Book of the Beginning. xiii–xliv. Chicago: University of Chicago Press, 1973. Coomaraswamy, Ananda K. “Mahābhārata, Itihāsa.” Annals of the Bhandarkar Oriental Research Institute 18, no. 2 (1937): 211–12. 512 512 PHILOLOGY AND CRITICISM Fitzgerald, James L. “General Introduction: The Translation Resumed.” In James L. Fitzgerald, trans. The Mahābhārata. Vol. 7: 11. The Book of the Women; 12. The Book of Peace, Part One. xv–xxxi. Chicago: University of Chicago Press, 2004. ———. “History and Primordium in the Mahā bhā rata.” In Thinking, Recording, and Writing History in the Ancient World, edited by Kurt Raalaub, 41–60. Malden, MA, and Oxford: Wiley-Blackwell, 2014. ———. “India’s Fifth Veda: The Mahābhārata’s Presentation of Itself.” In Essays on the Mahābhārata, edited by Arvind Sharma, 150–70. Leiden: E. J. Brill, 1991. ———. “Introduction [to the Book of Peace].” In James L. Fitzgerald, trans. The Mahābhārata. Vol. 7: 11. The Book of the Women; 12. The Book of Peace, Part One. 79–164. Chicago: University of Chicago Press, 2004. ———. “Mahā bhā rata.” In Brill’s Encyclopedia of Hinduism, vol. 2, edited by Knut A. Jacobsen, Helene Basu, Angelika Malinar and Vasudha Narayanan, 72–94. Leiden:  E. J.  Brill, 2010. Brill Online, 2015; http://referenceworks.brillonline.com/entries/brill-s-encyclopedia-of- hinduism/mahabharata-BEHCOM_2020040 (accessed September 23, 2015). ———. “Mahābhārata.” In The Hindu World, edited by Sushil Mittal and Gene Thursby, 52–74. New York and London: Routledge, 2004. ———. “Making Yudhiṣṭhira the King:  The Dialectics and the Politics of Violence in the Mahābhārata.” Rocznik Orientalistyczny 54, no. 1 (2001): 63–92. ———. “The Great Epic of India as Religious Rhetoric: A Fresh Look at the Mahābhārata.” Journal of the American Academy of Religion 51, no. 4 (1983): 611–30. ———. “The Many Voices of the Mahā bhā rata.” Review of Rethinking the Mahābhārata: A Reader’s Guide to the Education of the Dharma King, by Alf Hiltebeitel. Journal of the American Oriental Society 123, no. 4 (2003): 803–18. ———. “No Contest between Memory and Invention: The Invention of the Pā ṇḍava Heroes of the Mahā bhā rata.” In Epic and History, edited by David Konstan and Kurt A. Raalaub, 103– 21. Malden, MA: Wiley–Blackwell, 2010. Hiltebeitel, Alf. “Introducing the Mahā bhā rata.” Religious Studies Review 41, no. 4 (2015): 153–74. Kirste, J.  “Zur Mahā bhā ratafrage.” Wiener Zeitschrift für die Kunde des Morgenlandes 14 (1900): 214–24. Mahadevan, T. P. “On the Southern Recension of the Mahābhārata, Brahman Migrations, and the Brā hmī Paleography.” Electronic Journal of Vedic Studies 15, no. 2 (2008): 43–147. ———. “The Ś akuntalā -Yayā ti ‘Transposition,’ the Southern Recension of the Mahā bhā rata, and V. S. Sukthankar.” In The Churning of the Epics and Purānạ s at the 15th World Sanskrit Conference, edited by Simon Brodbeck, Alf Hiltebeitel and Adam Bowles, 46–70. New Delhi:  Dev Publishers, 2018. ———. “The Southern Recension of the Mahābhārata, the Harivaṃśa, and Ā ḻvā r Vaiṣṇavism.” In Ways and Reasons for Thinking about the Mahābhārata as a Whole, edited by Vishwa Adluri, 87–136. Pune: Bhandarkar Oriental Research Institute, 2011. ———. “Three Rails of the Mahābhārata Text Tradition.” Journal of Vaishnava Studies 19, no. 2 (2011): 23–69. Indian History, Epigraphy and Manuscript Culture This section lists the literature we consulted in Indian epigraphy, paleography and writing materials. Baums, Stefan. “Gandhā ran Scrolls: Rediscovering an Ancient Manuscript Type.” In Manuscript Cultures: Mapping the Field, edited by Jörg Quenzer, Dmitry Bondarev and Jan-Ulrich Sobisch, 183–226. Berlin, Munich and Boston, MA: de Gruyter, 2014. Bhoi, Panchanan. “Scribe as Metaphor:  Patterns of Processing and Writing Palm Leaf Manuscripts.” Indian Anthropologist 40, no. 1 (2010): 71–92. ———. “The Writer and the Text: The Palmleaf Scribe as Chronicler.” Social Scientist 33, no. 5/ 6 (2005): 73–92. 513 ANNOTATED BIBLIOGRAPHY 513 Hoernle, A. F. Rudolf. “An Epigraphical Note on Palm-leaf, Paper and Birch-Bark.” Journal of the Asiatic Society of Bengal LXIX, II (1900): 93–134. Mahadevan, Iravatham. “From Orality to Literacy:  The Case of the Tamil Society.” Studies in History, n.s., 11, no. 2 (1995): 173–88. Perumal, P. “The Sanskrit Manuscripts in Tamil Nadu.” In Aspects of Manuscript Culture in South India, edited by Saraju Rath, 157–72. Leiden: Brill, 2012. Rath, Saraju. “The Evolution of Inscriptional Nā garī from Early 7th till 12th CE.” In Epigraphika Vostoka (Epigraphy of the Orient) 29 (2011): 187–201. ———. “The Oriya Script: Origin, Development and Sources.” Heritage India 5, no. 2 (2012): 54–61. ———. “Scripts of Ancient India—Siddhamātṛkā.” In Nyāya-Vasiṣtḥ a. Felicitation Volume of Prof. V. N. Jha, edited by Manabendu Banerjee et al., 717–29. Kolkata: Sanskrit Pustak Bhandar, 2006. ———. “Varieties of Grantha Script: The Date and Place of Origins of Manuscripts.” In Aspects of Manuscript Culture in South India, edited by Saraju Rath, 187–206. Leiden: Brill, 2012. Rath, Saraju, and Jan Houben. “Introduction:  Manuscript Culture and Its Impact in ‘India’: Contours and Parameters.” In Aspects of Manuscript Culture in South India, edited by Saraju Rath, 1–53. Leiden: Brill, 2012. Sircar, D. C. Indian Epigraphy. Delhi: Motilal Banarsidass, 1965. Textual Traditions and Editions of Texts Other than the Mahā bhā rata No text is as widely referenced within Indian literature as the Mahā bhā rata. A  list of relevant textual traditions would easily extend into the hundreds. The following list ofers an initial orienta- tion to the two most important traditions for the Mahā bhā rata: the Dharmaśā stras (especially the Mā nava-Dharmaśā stra) and the Rā mā yaṇa. Like Mahā bhā rata studies, Rā mā yaṇa studies also rest on foundation the German scholars provided. Once again, the central problem concerns the search for an oral epic devoid of its theistic references. Kane, Pandurang Vaman. History of Dharmaśas̄ tra: Ancient and Medieval Religious and Civil Law. 5 vols. Poona: Bhandarkar Oriental Research Institute, 1930–62. ———. “The Two Epics.” Annals of the Bhandarkar Oriental Research Institute 47, no. 1/4 (1966): 11–58. Olivelle, Patrick. “Structure and Composition of the Mānava Dharmaśāstra.” Journal of Indian Philosophy 30, no. 6 (2002): 535–74. ———, ed. Manu’s Code of Law:  A Critical Edition and Translation of the Mānava-Dharmaśāstra. New York: Oxford University Press, 2005. Pollock, Sheldon I. “Introduction.” In The Rāmāyaṇa of Vālmı̄ki: An Epic of Ancient India. Edited by Robert P. Goldman. Vol. III: Araṇyakāṇḍa. Introduction and translation by Sheldon I. Pollock. 3–84. Delhi: Motilal Banarsidass, 2007. ———. “Part I.  Introduction.” In The Rāmāyaṇa of Vālmı̄ki:  An Epic of Ancient India. Edited by Robert P. Goldman. Vol. II: Aȳ odhyakāṇḍa. Introduction and translation by Sheldon I. Pollock. 1–76. Delhi: Motilal Banarsidass, 2007. ———. “The Rāmāyaṇa Text and the Critical Edition.” In The Rāmāyaṇa of Vālmı̄ki:  An Epic of Ancient India. Vol. I:  Bālakāṇḍa. Introduction and translation by Robert P. Goldman. 82–93. Delhi: Motilal Banarsidass, 2007. Rocher, Ludo. “Sanskrit Literature.” In Scholarly Editing:  A Guide to Research, edited by David C. Greetham, 575–99. New York: Modern Language Association of America, 1995. Sukthankar, Vishnu S. “The Rā ma Episode (Rā mopā khyā na) and the Rā mā yaṇa.” In A Volume of Studies in Indology Presented to Prof. P. V. Kane, edited by S. M. Katre and P. K. Gode, 472–87. Poona: Oriental Book Agency, 1941. German Scholarship/ Errors in Textual Criticism The works listed in this section are the main ones criticized in this book. Their reasons for inclusion are transparent, but not every work criticized in this book is listed. Some it better in other sections 514 514 PHILOLOGY AND CRITICISM or had value nonetheless. Additionally, we list works here that we think deserve mention for their exaggerated claims on behalf of German philological prestige. Austin, Christopher R.  “Vedic Myth and Ritual in the Mahābhārata:  A Critical Study of the Mahāprasthānika- and Svargārohaṇa Parvans.” PhD diss., McMaster University, 2007. Bigger, Andreas. “Balarā ma im Mahā bhā rata.” Asiatische Studien/ Études Asiatiques xlviii, no.  1 (1994): 1297–99. ———. Balarāma im Mahābhārata:  Seine Darstellung im Rahmen des Textes und seiner Entwicklung. Wiesbaden: Otto Harrassowitz, 1998. ———. “The Normative Redaction of the Mahābhārata: Possibilities and Limitations of a Working Hypothesis.” In Stages and Transitions:  Temporal and Historical Frameworks in Epic and Purāṇic Literature. Proceedings of the Second Dubrovnik International Conference on the Sanskrit Epics and Purāṇas, edited by Mary Brockington, 21–33. Zagreb: Croatian Academy of Sciences and Arts, 2002. ———. “Wege und Umwege zum Himmel. Die Pilgerfahrten im Mahā bhā rata.” Journal Asiatique 289, no. 1 (2001): 147–66. Bronkhorst, Johannes. “Archetypes and Bottlenecks:  Relections on the Textual History of the Mahā bhā rata.” In Pūr vāparaprajñābhinandam: East and West, Past and Present. Indological and Other Essays in Honour of Klaus Karttunen, edited by Bertil Tikkanen and Albion M. Butters, 39–54. Studia Orientalia 110. Helsinki: Finnish Oriental Society, 2011. ———. Greater Magadha: Studies in the Culture of Early India. Leiden and Boston, MA: Brill, 2007. ———. How the Brahmins Won: From Alexander to the Guptas. Leiden: Brill, 2016. Goldstücker, Theodor. “Hindu Epic Poetry: The Mahâbhârata.” The Westminster Review, n.s., 33 (1868): 380–419. Grünendahl, Reinhold. “History in the Making:  On Sheldon Pollock’s ‘NS Indology’ and Vishwa Adluri’s ‘Pride and Prejudice.’” International Journal of Hindu Studies 16, no. 2 (2012): 189–257. ———. “Post-philological Gestures—‘Deconstructing’ Textual Criticism.” Wiener Zeitschrift für die Kunde Südasiens  52–53, Text Genealogy, Textual Criticism and Editorial Technique (2009–10): 17–28. ———. “Zur Klassiizierung von Mahā bhā rata-Handschriften.” In Studien zur Indologie und Buddhismuskunde, edited by Reinhold Grünendahl, Jens Uwe-Hartmann and Petra Kiefer-Pülz, 101–30. Indica et Tibetica 22. Bonn: Indica et Tibetica Verlag, 1993. ———. “Zur Textkritik des Nā rā yaṇīya.” In Nārāyaṇı ̄ya-Studien, edited by Peter Schreiner, 30–74. Purā ṇa Research Publications Tübingen 6. Wiesbaden: Harrassowitz, 1997. ———, ed. Viṣṇudharmāḥ :  Precepts for the Worship of Viṣṇu Part  1:  Adhyāyas 1–43. Wiesbaden: Harrassowitz, 1983. Hacker, Paul. “Zur Methode der geschichtlichen Erforschung der anonymen Sanskritliteratur des Hinduismus. Vortrag gehalten auf dem XV. Deutschen Orientalistentag Göttingen 1961.” Zeitschrift der Deutschen Morgenländischen Gesellschaft 111, no. 2 (1961): 483–92. Hanneder, Jürgen. “Introduction.” Wiener Zeitschrift für die Kunde Südasiens 52–53, Text Genealogy, Textual Criticism and Editorial Technique (2009–10): 5–16. ———. To Edit or Not to Edit:  On Textual Criticism of Sanskrit Works. Pune Indological Series 1. Delhi: Aditya Prakashan, 2017. Von Hinüber, Oskar. Review of The Sanskrit Epics, by John Brockington. Wiener Zeitschrift für die Kunde Südasiens 46 (2002): 268–69. ———. “Der vernachlässigte Wortlaut. Die Problematik der Herausgabe buddhistischer Sanskrit- Texte, Zur Überlieferung, Kritik und Edition alter und neuerer Text.” In Kleine Schriften, Teil I. Edited by Harry Falk and Walter Slaje. 431–50. Wiesbaden: Harrassowitz, 2008. Holtzmann Jr., Adolf. Zur Geschichte und Kritik des Mahābhārata. Kiel: C. F. Haeseler, 1892. ———. Das Mahābhārata nach der nordindischen Recension. Kiel: C. F. Haeseler, 1894. Holtzmann Sr., Adolf. Indische Sagen. Zweite verbesserte Aulage in zwei Bänden. Vol. 1. Stuttgart: Adolph Krabbe, 1854. 51 ANNOTATED BIBLIOGRAPHY 515 Lariviere, Richard. Protestants, Orientalists, and Brāhmaṇas:  Reconstructing Indian Social History. Amsterdam: Royal Netherlands Academy of Arts and Sciences, 1995. Lassen, Christian. “Beiträge zur Kunde des indischen Alterthums aus dem Mahâbhârata I: Allgemeines über das Mahābhārata.” Zeitschrift für die Kunde des Morgenlandes 1 (1837): 61–86. ———. Indische Alterthumskunde. 4 vols. Bonn and Leipzig: H. B. Koenig, 1847–61. ———. Indische Alterthumskunde. Vol. 1: Geographie und die älteste Geschichte. Bonn: H. B. Koenig, 1847. Mylius, Klaus. Geschichte der altindischen Literatur. Bern: Scherz, 1988. Oberlies, Thomas. “Arjunas Himmelreise und die Tīrthayā tra der Pā ṇḍavas:  Zur Struktur des Tīrthayā trā parvan des Mahā bhā rata.” Acta Orientalia 56 (1995): 106–24. ———. “Die Ratschläge des Sehers Nā rada: Ritual an und unter der Oberläche des Mahā bhā rata.” In New Methods in the Research of Epic/Neue Methoden der Epenforschung, edited by Hildegard L. C. Tristram, 125–41. Tübingen: Gunter Narr Verlag, 1998. ———. “(Un)ordnung im Mahābhārata: Rahmenerzählungen, Gesprächsebenen und Inhaltsangaben.” Studien zur Indologie und Iranistik 25 (2008): 73–102. Oldenberg, Hermann. Das Mahābhārata: Seine Entstehung, sein Inhalt, seine Form. Göttingen: Vandenhoek & Ruprecht, 1922. Pollock, Sheldon. The Language of the Gods in the World of Men: Sanskrit, Culture, and Power in Premodern India. Berkeley and Los Angeles: University of California Press, 2006. ———. “Sanskrit Literary Culture from the Inside Out.” In Literary Cultures in History: Reconstructions from South Asia, edited by Sheldon Pollock, 39–130. Berkeley and Los Angeles and London: University of California Press, 2003. Rau, Wilhelm. “Bemerkungen zu Ś aṅkaras Bṛhadā raṇyakopaniṣadbhā ṣya.” Paideuma 7, Festgabe für Herman Lommel zur Vollendung seines 75. Lebensjahres am 7. Juli 1960 (1959–61): 293–99. ———. Die vedischen Zitate im Vyākaraṇa-Mahābhāṣya. Mainzer Akademie der Wissenschaften und der Literatur, Abhandlungen der Geistes- und Sozialwissenschaftlichen Klasse, Jahrgang 1985, no. 4. Wiesbaden: Steiner, 1985. Ruben, Walter. “Schwierigkeiten der Textkritik des Mahā bhā rata.” Acta Orientalia 8 (1930): 240–56. ———. Studien zur Textgeschichte des Rāmāyaṇa. Stuttgart: W. Kohlhammer, 1936. Shalom, Naama. Re-ending the Mahābhārata: The Rejection of Dharma in the Sanskrit Epic. Albany: State University of New York Press, 2017. Von Simson, Georg. “Altindische epische Schlachtbeschreibung:  Untersuchungen zu Kompositionstechnik und Entstehungsgeschichte der Bücher VI bis IX des Mahā bhā rata.” Habilitationsschrift, University of Göttingen, 1974. ———. “Die Einschaltung der Bhagavadgı̄tā im Bhı̄ṣmaparvan des Mahābhārata.” Indo-Iranian Journal 11, no. 3 (1968/69): 159–74. ———. Mahābhārata: Die Große Erzählung von den Bhāratas. Berlin: Verlag der Weltreligionen, 2011. ———. “Text Layers in the Mahābhārata: Some Observations in Connection with Mahābhārata VII.131.” In The Mahābhārata Revisited. Papers Presented at the International Seminar on the Mahābhārata Organized by the Sahitya Akademi at New Delhi on February 17–20, 1987, edited by R. N. Dandekar, 37–60. New Delhi: Sahitya Akademi, 1990. Slaje, Walter. “The Mokṣopā ya Project.” Annals of the Bhandarkar Oriental Research Institute 77, no. 1/4 (1996): 209–21. Srinivasan, Srinivasa Ayya. Vācaspatimiśras Tattvakaumudı̄:  Ein Beitrag zur Textkritik bei kontaminierter Überlieferung. Alt- und Neu-Indische Studien 12. Hamburg: Cram, De Gruyter & Co., 1967. Von Stietencron, Heinrich. “Editor’s Introduction.” In Angelika Malinar, Rājavidyā: Das königliche Wissen um Herrschaft und Verzicht. 6–11. Wiesbaden: Otto Harrassowitz, 1996. Weller, Hermann. “Zur Textkritik des Mahā bhā rata.” In Festschrift Moriz Winternitz:  1863– 23. Dezember 1933, edited by Otto Stein and Wilhelm Gampert, 37–40. Leipzig:  Otto Harrassowitz, 1933. ———. “Who Were the Bhriguids.” Annals of the Bhandarkar Oriental Research Institute 18, no.  3 (1937): 296–302. 516 516 PHILOLOGY AND CRITICISM Winternitz, Moriz. “Genesis des Mahā bhā rata.” Wiener Zeitschrift für die Kunde des Morgenlandes 14 (1900): 51–77. ———. Geschichte der indischen Literattur. Vol. 1. 2nd edn. Leipzig: C. F. Amelangs, 1909. ———. Geschichte der indischen Litteratur. Vol. 3. Leipzig: C. F. Amelangs, 1920. ———. “The Mahabharata.” The Visva-Bharati Quarterly 1 (1924): 343–59. ———. “Notes on the Mahā bhā rata, with Special Reference to Dahlmann’s ‘Mahā bhā rata.’ ” Journal of the Royal Asiatic Society of Great Britain and Ireland (1897): 713–59. Witzel, Michael. “How to Enter the Vedic Mind? Strategies in Translating a Brā hmaṇa Text.” In New Horizons of Indological Research: Felicitation Volume in Honour of Prof. P. C. Muraleemadhavan, edited by Dharmaraj Adat, 21–33. Delhi: Kunjunni Raja Academy of Indological Research and New Bharatiya Book Corporation, 2013. ———. Katḥ a Ar̄ aṇyaka: Critical Edition with a Translation into German and an Introduction. Cambridge, MA: Department of Sanskrit and Indian Studies, Harvard University, 2004. ———. Das Katḥ a-Ar̄ aṇyaka, textkritische Edition mit Uebersetzung und Kommentar. Erlangen:  Nepal Research Centre, 1974. ———. “On the Archetype of Patañjali’s Mahā bhā ṣya.” Indo-Iranian Journal 29 (1986): 249–59. ———. “Textual Criticism in Indology and in European Philology during the 19th and 20th Centuries.” Electronic Journal of Vedic Studies 21, no. 3 (2014): 9–91. The Background of the Mahā bhā rata Critical Edition/Biographic Sources A history of the Mahā bhā rata critical edition has never been written. The present work is not this his- tory, even though we gleaned almost everything known about the edition for this book. This section ofers a list of the sources consulted. The Annals of the Bhandarkar Oriental Research Institute are the best source for the history of the edition. They are now fully archived on JSTOR. The list also includes a few additional sources on the history of the Bhandarkar Institute, its collections and important igures associated with it (even if not directly involved in the Mahā bhā rata critical edition itself). “150th Birth-Anniversary of Ramakrishna Gopal Bhandarkar and 70th Foundation-Day of the Bhandarkar Oriental Research Institute, Poona July 6, 1987.” Annals of the Bhandarkar Oriental Research Institute 69, no. 1/4 (1988): i–viii. Bedekar, V. M. “The Contribution of Dr.  S.  K. Belvalkar to Indology.” Annals of the Bhandarkar Oriental Research Institute 81, no. 1/4 (2000): 35–67. “Bhandarkar Oriental Research Institute, Poona Silver Jubilee Celebrations 4th and 5th of January 1943. Roger Lumley, N. C. Kelkar, J. R. Gharpure, R. N. Dandekar, C. Kunhan Raja, V. M. Apte, P. V. Bapat, A. P. Karmarkar, B. R. Kulkarni, H. D. Sankalia, C. R. Sankaran, S. R. Shende, Ludwik Sternbach.” Annals of the Bhandarkar Oriental Research Institute 24, no. 1/ 2 (1943): iii–xlvii. Bhandarkar, R. G. “Text of the Inaugural Address.” Annals of the Bhandarkar Oriental Research Institute 1, no. 1 (1918–19): 1–6. “The Completion of the Critical Edition of the Mahā bhā rata.” Annals of the Bhandarkar Oriental Research Institute 47, no. 1/4 (1966): i–xiv. Dandekar, R. N. “Geheimrat Professor Dr. Heinrich Lüders (1869–1943).” Annals of the Bhandarkar Oriental Research Institute 24, no. 3/4 (1943): 282–83. ———. “The Mahā bhā rata and Its Critical Edition.” In Exercises in Indology: Select Writings. Vol. 3. 292–310. Delhi: Ajanta Publications, 1981. ———. “Parashuram Lakshman Vaidya.” Annals of the Bhandarkar Oriental Research Institute 60, no. 1/4 (1979): 350–51. ———. “Professor P.  V. Kane.” Annals of the Bhandarkar Oriental Research Institute 53, no. 1/4 (1972): 311–22. 517 ANNOTATED BIBLIOGRAPHY 517 Deshpande, Madhav M. “Professor R. N. Dandekar March 17, 1909–December 11, 2001.” Indo- Iranian Journal 45, no. 2 (2002): 109–19. “Dr.  Shripad Krishna Belvalkar.” Annals of the Bhandarkar Oriental Research Institute 47, no. 1/4 (1966): 132–33. Edgerton, Franklin. “Tribute from the West.” Annals of the Bhandarkar Oriental Research Institute 24, no. 1/2 (1943): 136. “Editorial Notes.” Annals of the Bhandarkar Oriental Research Institute 1, no. 1 (1918–19): 81–85. “Editorial Notes: Presentation of the First Volume of the Critical Edition of the Mahā bhā rata to Shrimant Balasaheb Pant Pratinidhi, B. A., Ruler of Aundh, 6th July 1934.” Annals of the Bhandarkar Oriental Research Institute 15, no. 3/4 (1933–34): 250–58. Emeneau, M. B. “Franklin Edgerton.” Language 40, no. 2 (1964): 111–23. Hahn, Adelaide E. “Franklin Edgerton: Personal Reminiscences.” Journal of the American Oriental Society 85, no. 1 (1965): 3–8. Jacobi, Hermann. “Bericht über den Beschluß des Komitees für eine kritische Textausgabe des Mahabharata.” In Verhandlungen des XIII. Internationalen Orientalisten-Kongresses, Hamburg September 1902. 84–85. Leiden: E. J. Brill, 1904. Jacobi, Hermann, Heinrich Lüders and Moriz Winternitz. “Promemoria über den Plan einer kritischen Ausgabe des Mahābhārata.” Almanach der kaiserlichen Akademie der Wissenschaften 54 (1904): 268–78. Johnson, Donald Clay. “On the Origins of the Bhandarkar Oriental Research Institute Library, Part I:  The Initiation of the Searches for Sanskrit Manuscripts in Bombay.” Annals of the Bhandarkar Oriental Research Institute 69, no. 1/4 (1988): 111–24. ———. “On the Origins of the Bhandarkar Oriental Research Institute Library, Part II:  The Work of Ramakrishna Gopal Bhandarkar and Peter Peterson.” Annals of the Bhandarkar Oriental Research Institute 71, no. 1/4 (1990): 71–82. Katre, S. M. “In Memoriam:  Vishnu Sitaram Sukthankar 1887–1943.” Annals of the Bhandarkar Oriental Research Institute 24, no. 1/2 (1943): 123–35. ———. “Vishnu Sitaram Sukthankar and His Contribution to Indology.” In Sukthankar Memorial Edition. Vol. 2: Analecta. Edited by P. K. Gode. 463–505. Bombay: Karnatak Publishing House, 1945. Lad, Gouri P. “Sumitra Mangesh Katre: A Biographical Note.” Bulletin of the Deccan College Research Institute 51/52, Professor S. M. Katre Felicitation Volume (1991–92): 3–29. Laddu, S. D., and A. N. Gokhaley. “The Government Manuscripts Library at the Bhandarkar Oriental Research Institute.” Annals of the Bhandarkar Oriental Research Institute 77, no. 1/4 (1996): i–x. Lévi, Sylvain. “Tato Jayam Udīrayet.” Translated by L. G. Khare. Annals of the Bhandarkar Oriental Research Institute 1, no. 1 (1918–19): 13–20. Lüders, Heinrich. Druckprobe einer kritischen Ausgabe des Mahābhārata. Leipzig: Drugulin, 1908. Marathe, N. B. “Index to Writings on the Mahā bhā rata from the Annals (1919–1997).” Annals of the Bhandarkar Oriental Research Institute 79, no. 1/4 (1998): i, iii–xx. Mehendale, M. A. “Professor Dr. Sumitra Mangesh Katre.” Annals of the Bhandarkar Oriental Research Institute 79, no. 1/4 (1998): 301, 303–9. Morgenroth, Wolfgang. “Vishnu Sitaram Sukthankar as a Student in Berlin, 1911–1914.” Annals of the Bhandarkar Oriental Research Institute 58/59, Diamond Jubilee Volume (1977–78): 193–201. Paranjpe, V. G. “The Late Dr. Sir R. G. Bhandarkar. An Appreciation.” Annals of the Bhandarkar Oriental Research Institute 7, no. 1/2 (1926): 153–59. Pratinidhi, Balasaheb Pant, ed. A Prospectus of a New and Critical Edition of the Mahabharata. Poona, 1919. “Prof. Hari Damodar Velankar.” Annals of the Bhandarkar Oriental Research Institute 47, no. 1/4 (1966): 134–35. 518 518 PHILOLOGY AND CRITICISM Pusalkar, A. D. “Vishnu Sitaram Sukthankar (4 May 1887–21 January 1943).” Journal of the Bombay Branch of the Royal Asiatic Society 19 (1943): 89–92. Radhakrishnan, S. “Silver Jubilee Address (Monday, 4th January 1943).” Annals of the Bhandarkar Oriental Research Institute 24, no. 1/2 (1943): 1–8. Rajwade, V. K., and Harold H. Mann. “Report and Proceedings of the Bhandarkar Oriental Research Institute Poona for the Year 1919–20.” Annals of the Bhandarkar Oriental Research Institute 2, no. 1 (1920–21): 1–80. “Ramchandra Narayan Dandekar (1909–2001).” Indian Literature 45, no. 6 (2001): 152–54. Von Schroeder, Leopold. “Bericht über den Stand der vorbereitenden Arbeiten für eine kritische Ausgabe des Mahabharata.” In  Verhandlungen des XIII. Internationalen Orientalisten-Kongresses, Hamburg September 1902. 83–84. Leiden: E. J. Brill, 1904. Souvenir of the Completion of the Critical Edition of the Mahābhārata. Poona:  Bhandarkar Oriental Research Institute, 1966. “Speech by Shri Rajendra Prasad, President of the Indian Union.” Annals of the Bhandarkar Oriental Research Institute 35, no. 1/4 (1954): 311–12. “Statement Presented by the Chairman of the Executive Board of the Institute.” Annals of the Bhandarkar Oriental Research Institute 35, no. 1/4 (1954): 299, 301–4. “A Statement Presented by the General Editor at the Inauguration of the Mahā bhā rata Printing Department of the B. O. R. Institute Press.” Annals of the Bhandarkar Oriental Research Institute 25, no. 1/3 (1944): i–xii. “Statement Presented by the General Editor of the Mahā bhā rata.” Annals of the Bhandarkar Oriental Research Institute 35, no. 1/4 (1954): 305–10. Sukthankar Memorial Edition. 2  vols. Edited by P.  K. Gode. Bombay:  Karnatak Publishing House, 1944–45. Sukthankar, Vishnu S. “Appendix VII:  Critical Edition of the Mahā bhā rata.” Annals of the Bhandarkar Oriental Research Institute 24, no. 1/2 (1943): xliv–xlvii. ———. “In Memoriam: Professor Moritz Winternitz (1863–1937).” Annals of the Bhandarkar Oriental Research Institute 18, no. 3 (1937): 313–20. ———. “Lebenslauf.” In Sukthankar Memorial Edition. Vol. 2: Analecta. Edited by P. K. Gode. 81. Bombay: Karnatak Publishing House, 1945. ———. Prospectus of a Critical and Illustrated Edition of the Mahabharata:  India’s Great National Epic. Poona: Bhandarkar Oriental Research Institute, 1937. “Sushil Kumar De: Curriculum Vitae.” Bulletin of the Deccan College Research Institute 20, no. 1/4, Sushil Kumar De Felicitation Volume (1960): vii, ix–xv. Thomas, G. A. “Appendix:  Report of the Working Committee of the Bhandarkar Oriental Research Institute, Poona City, from 6th July 1915 to 10th September 1918.” Annals of the Bhandarkar Oriental Research Institute 1, no. 1 (1918–19): 1–59. Utgikar, N. B. “The Devanagari Recension of the Mahā bhā rata.” Annals of the Bhandarkar Oriental Research Institute 1, no. 2 (1919–20): 145–55. ———. “The Devanagari Recension of the Mahā bhā rata II.” Annals of the Bhandarkar Oriental Research Institute 2, no. 1 (1920–21): 73–77. Winternitz, Moriz. “The Critical Edition of the Mahā bhā rata.” Paper read in Section Va, 1 (Ancient India) of the XVIIth International Congress of Orientalists, held at Oxford, on August 28, 1928. Indologica Pragensia 1 (1929): 58–63. ———. “Die kritische Ausgabe des Mahā bhā rata.” Forschungen und Fortschritte 8 (1932): 427. ———. “On the Mahā bhā rata MSS in the Whish Collection of the Royal Asiatic Society.” Journal of the Royal Asiatic Society of Great Britain and Ireland (1898): 147–50. ———. “Promemoria über die Nothwendigkeit einer kritischen Ausgabe des Mahā bhā rata, insbesondere der südindischen Recension.” Almanach der Kaiserlichen Akademie der Wissenschaften 51 (1901): 206–10. 519 ANNOTATED BIBLIOGRAPHY 519 ———. “A Proposal for the Formation of a Sanskrit Epic Texts Society, to Be Laid before the Indian Section of the XII International Congress of Orientalists, Held at Rome, in October 1899.” XIIme Congrès International des Orientalistes, Bulletin no. 3 (1899): 46–49. Philology, Textuality and the Value of Textual Criticism The works listed here range from general discussions about the value of philology to specialized treatments of particular traditions. They are all instructive, and provide insight into the debates concerning the study of texts occurring today. Tanselle is a good if lengthy introduction. Canfora, Luciano, and Jean Burrell. “From the Quest for the Archetype to the History of Texts. A  Short Note on the French Critical Style.” Translated by Jean Burrell. Diogenes 47, no. 2 (1999): 28–29. Debray-Genette, Raymonde. “Génétique et poétique:  esquisse de méthode.” Littérature 28 (1977): 19–39. Greetham, David C. Theories of the Text. Oxford: Oxford University Press, 1999. ———, ed. Scholarly Editing:  A Guide to Research. New  York:  Modern Language Association of America, 1995. Hay, Louis. “Genetic Criticism: Origins and Perspectives.” In Genetic Criticism: Texts and Avant-textes, edited by Jed Deppman, Daniel Ferrer and Michael Groden, 17–27. Philadelphia: University of Pennsylvania Press, 2004. ———. “History or Genesis?” Yale French Studies 89 (1996): 191–207. ———. “Does ‘Text’ Exist?” Studies in Bibliography 41 (1988): 64–76. Kenney, E. J. The Classical Text: Aspects of Editing in the Age of the Printed Book. Berkeley, Los Angeles and London: University of California Press, 1974. Mathijsen, Marita. “The Concept of Authorisation.” Text 14 (2002): 77–90. McGann, Jerome. A Critique of Modern Textual Criticism. Chicago: University of Chicago Press, 1983. McGillivray, Murray. “Towards a Post-Critical Edition: Theory, Hypertext, and the Presentation of Middle English Works.” Text 7 (1994): 175–99. Müller-Sievers, Helmut. “Reading Evidence:  Textual Criticism as Science in the Nineteenth Century.” The Germanic Review 76, no. 2 (2001): 162–71. Nichols, Stephen G. “Introduction:  Philology in a Manuscript Culture.” Speculum 65, no. 1 (1990): 1–10. Richardson, Peter. “The Consolation of Philology.” Modern Philology 92, no. 1 (1994): 1–13. Storey, Wayne H. “Interpretative Mechanisms in the Textual Cultures of Scholarly Editing.” Textual Cultures 4, no. 1 (2009): 129–47. Tanselle, G. Thomas. “Textual Criticism at the Millennium.” Studies in Bibliography 54 (2001): 1–80. Tov, Emmanuel. “The Relevance of Textual Theories for the Praxis of Textual Criticism.” In A Teacher for All Generations: Essays in Honor of James C. Vander Kam, vol. 1, edited by Eric Farrel Mason, Samuel I. Thomas, Alison Schoield and Eugene Ulrich, 23–35. Leiden: E. J. Brill, 2012. Watkins, Calvert. “What Is Philology?” Comparative Literature Studies 27, no. 1 (1990): 21–25. Additional Sources This section contains sources that were cited in the footnotes but did not it into the previous sections. The majority of works concern art history and art criticism. Adluri, Vishwa. Review of World Philology, ed. Sheldon Pollock, Benjamin A. Elman and Ku-ming Kevin Chang. American Historical Review 121, no. 3 (2016): 908–10. Adluri, Vishwa, and Joydeep Bagchee. “Jews and Hindus in Indology.” Paper published on Academia.edu. www.academia.edu/30937643/Jews_and_Hindus_in_Indology (accessed April 14, 2017). 520 520 PHILOLOGY AND CRITICISM ———. “Theses on Indology.” Paper published on Academia.edu. www.academia.edu/30584042/ Theses_on_Indology (accessed October 13, 2017). Aner, Karl. Die Theologie der Lessingzeit. Halle a.d. Saale: G. Olms, 1929. Arrowsmith, William. “The Shame of the Graduate Schools: A Plea for a New American Scholar.” Arion: A Journal of Humanities and the Classics, third series, 2, no. 2/3 (1992–93): 159–76. Austin, Christopher R. “The Sārasvata Yātsattra in Mahā bhā rata 17 and 18.” International Journal of Hindu Studies 12, no. 3 (2008): 283–308. Bali, Saraswati. Sāyaṇa’s Upodghāta to the Taittirı̄ya Saṁhita and the Ṛgveda Saṁhitā: Introduction, English Translation of the Text and Notes. Delhi: Pratibha Prakashan, 1999. Bobzin, Hartmut. Review of Viṣṇudharmāḥ : Precepts for the Worship of Viṣṇu, by Reinhold Grünendahl. Zeitschrift der Deutschen Morgenländischen Gesellschaft 135, no. 1 (1985): 203–4. ———. Review of Viṣṇudharmāḥ : Precepts for the Worship of Viṣṇu, by Reinhold Grünendahl. Zeitschrift der Deutschen Morgenländischen Gesellschaft 136, no. 3 (1986): 658–59. Brockington, John L. “The Bhagavadgītā: Text and Context.” In The Fruits of Our Desiring: An Inquiry into the Ethics of the Bhagavadgītā for Our Times, edited by Julius Lipner, 28–47. Calgary: Bayeux, 1997. ———. “Is the Script Relevant? Further Evidence from a Nevā rī-Script Rā mā yaṇa Manuscript.” In Epic and Argument in Sanskrit Literary History. Essays in Honor of Robert P.  Goldman, edited by Sheldon Pollock, 17–29. Delhi: Manohar, 2010. ———. “A Malayā ḷam-Script Rā mā yaṇa Manuscript.” Indologica Taurinensia 21–22 (1995–96): 77–91. ———. “Regions and Recensions, Scripts and Manuscripts: The Textual History of the Rāmāyaṇa and Mahābhārata.” www.college-de-france.fr/site/jean-noel-robert/symposium-2016-12-06- 14h00.htm (accessed June 30, 2017). ———. Review of Untersuchungen zur Komposition von Reden und Gesprächen im Rāmāyaṇa, by Renate Söhnen. Journal of the Royal Asiatic Society of Great Britain and Ireland 114, no. 1 (1982): 65. ———. Review of Viṣṇudharmāḥ : Precepts for the Worship of Viṣṇu, Part 1: Adhyāyas 1–43, by Reinhold Grünendahl. Journal of the Royal Asiatic Society of Great Britain and Ireland 2 (1984): 291–92. ———. Review of Viṣṇudharmāḥ : Precepts for the Worship of Viṣṇu. Part 2: Adhyāyas 44–81, by Reinhold Grünendahl. Journal of the Royal Asiatic Society of Great Britain and Ireland 1 (1986): 126–27. ———. Righteous Rāma: The Evolution of an Epic. New York: Oxford University Press, 1984. ———. “The Text of the Rā mā yaṇa.” Indologica Taurinensia 15–16, Proceedings of the Seventh World Sanskrit Conference, Leiden, August 23–29, 1987 (1989–90): 79–90. ———. “Textual Studies in Vā lmīkī’s Rā mā yaṇa.” In Epic Threads: John Brockington on the Sanskrit Epics. Edited by Greg Bailey and Mary Brockington. 195–206. New Delhi: Oxford University Press, 2000. ———. “Textual Studies in Vā lmīkī’s Rā mā yaṇa.” Zeitschrift der Deutschen Morgenländischen Gesellschaft, Supplement IX, Proceedings of the XXXII International Congress for Asian and North African Studies, Hamburg, August 25–30, 1986 (1992): 509–10. Brockington, John, and Mary Brockington. Rāma the Steadfast:  An Early Form of the Rāmāyaṇa. London: Penguin Books, 2006. Brodbeck, Simon. The Mahābhārata Patriline:  Gender, Culture and the Royal Hereditary. Farnham: Ashgate, 2009. Brodbeck, Simon, and Brian Black, eds. Gender and Narrative in the Mahābhārata. Abingdon and New York: Routledge, 2007. Brullé, Pierre, and Marketa Theinhardt. “Painting Despite Everything: František Kupka on Creation in the Plastic Arts.” In Painting the Universe: František Kupka, Pioneer in Abstraction, edited by Dorothy Kosinski and Jaroslav Anděl, 151–77 . Ostildern-Ruit, Germany: Verlag Gerd Hatje, 1997. Von Büren, Veronika. “La place du manuscrit Ambr. L 99 sup. dans la transmission des Étymologies d’Isidore de Séville.” In Nuove ricerche su codici in scrittura latina dell’Ambrosiana, Atti del Convegno 521 ANNOTATED BIBLIOGRAPHY 521 Milano, 6–7 ottobre 2005, edited by Mirella Ferrari and Marco Navoni, 25–44. Milan: Vita e Pensiero, 2007. Burnell, A. C. A Classiied Index to the Sanskrit MSS in the Palace at Tanjore. London: Trübner, 1880. ———. On the Aindra School of Sanskrit Grammarians. Mangalore: Basel Mission Book & Tract Depository and London: Trübner & Company, 1875. Chambers, Mortimer. Review of Thucydidis Historiae, Vol. I: Libri I–II, by Carolum Hude and Otto Luschnat. Classical Philology 51, no. 1 (1956): 44–45. Copi, Irving M., Carl Cohen and Kenneth MacMahon. Introduction to Logic. 14th edn. London and New York: Routledge, 2016. Dawe, R. D. The Collation and Investigation of Manuscripts of Aeschylus. Cambridge:  Cambridge University Press, 1964. Dickerman, Leah. Inventing Abstraction, 1910–1925:  How a Radical Idea Changed Modern Art. New York: Museum of Modern Art, 2013. Doniger, Wendy. “How to Escape the Curse.” Review of The Mahabharata, translated by John Smith. London Review of Books 31, no. 19 (2009): 17–18. www.lrb.co.uk/v31/n19/wendy- doniger/ howto-escape-the-curse (accessed October 13, 2015). Edwards, A. W. F. Cogwheels of the Mind: The Story of Venn Diagrams. Baltimore, MD, and London: Johns Hopkins University Press, 2004. The Etymologies of Isidore of Seville. Translated by Stephen A. Barney, W. J. Lewis, J. A. Beach and Oliver Berghof. Cambridge: Cambridge University Press, 2006. Everdell, William R. The First Moderns: Proiles in the Origins of Twentieth-Century Thought. Chicago and London: University of Chicago Press, 1997. Fontaine, Jacques. Isidore de Séville:  genése et originalité de la culture hispanique au temps des Wisigoths. Turnhout: Brepols, 2000. ———. Tradition et actualité chez Isidore de Séville. London: Variorum, 1988. Foucault, Michel. The Order of Things: An Archaeology of the Human Sciences. London and New York: Routledge, 2002. Franco, Eli. Review of The Nay Science: A History of German Indology, by Vishwa Adluri and Joydeep Bagchee. South Asia: Journal of South Asian Studies 39, no. 3 (2016): 695–713. Gaugin, Paul. Gaugin’s Intimate Journals. Translated by Van Wyck Brooks. Mineola, NY:  Dover Publishing Company, 1997. Gehrts, Heino. Mahābhārata: Das Geschehen und seine Bedeutung. Bonn: Bouvier Verlag, 1975. Goldman, Robert P. “On the Upatva of Upā khyā nas: Is the Uttarakā ṇḍa of the Rā mā yaṇa an Upā khyā nas of the Mahā bhā rata?” In Argument and Design: The Unity of the Mahābhārata, edited by Vishwa Adluri and Joydeep Bagchee, 69–82. Leiden and Boston, MA: Brill, 2016. Griith, John G. “A Taxonomic Study of the Manuscript Tradition of Juvenal.” Museum Helveticum 25, no. 2 (1968): 101–38. Guha, Sumit. “Lower Strata, Older Races, and Aboriginal Peoples:  Racial Anthropology and Mythical History Past and Present.” The Journal of Asian Studies 57, no. 2 (1998): 423–41. Günther, Hans F. K. Kleine Rassenkunde des deutschen Volkes. 3rd edn. Munich: J. F. Lehmann, 1934. Gurd, Sean Alexander. Iphigenias at Aulis: Textual Multiplicity, Radical Philology. Ithaca, NY: Cornell University Press, 2005. Hall, J. B. “Claudian, De Bello Gildonico.” Review of Claudii Claudiani, De Bello Gildonico, ed. E. M. Olechowska. The Classical Review 33, no. 2 (1983): 203–7. Hellwig, Oliver. “Stratifying the Mahā bhā rata: The Textual Position of the Bhagavadgītā .” Indo- Iranian Journal 60, no. 2 (2017): 132–69. Hiltebeitel, Alf. “Kṛṣṇa in the Mahābhārata: A Bibliographical Essay.” Annals of the Bhandarkar Oriental Research Institute 60 (1979): 65–110. ———. “The Nārāyaṇı ̄ya and the Early Reading Communities of the Mahā bhā rata.” In Between the Empires: Society in India 300 BCE to 400 CE, edited by Patrick Olivelle, 227–55. New York: Oxford University Press, 2006. 52 522 PHILOLOGY AND CRITICISM ———. Nonviolence in the Mahābhārata: Śiva’s Summa on Ṛṣidharma and the Gleaners of Kurukṣetra. London and New York: Routledge, 2016. ———. “Not without Subtales: Telling Laws and Truths in the Sanskrit Epics.” Journal of Indian Philosophy 33 (2005): 455–511. ———. Rethinking the Mahābhārata:  A Reader’s Guide to the Education of the Dharma King. Chicago: University of Chicago Press, 2001. Hudson, Emily T. Disorienting Dharma: The Aesthetics of Sufering in the Mahā bhā rata. New York: Oxford University Press, 2013. Jacobi, Hermann. Mahābhārata: Inhaltsangabe, Index und Concordanz der Calcuttaer und Bombayer Ausgaben. Bonn: Friedrich Cohen, 1903. Jamison, Stephanie W. Sacriiced Wife/Sacriicer’s Wife: Women, Ritual, and Hospitality in Ancient India. New York: Oxford University Press, 1996. Jerusalem, Johann Friedrich Wilhelm. Zweyte Sammlung einiger Predigten. Braunschweig:  Rudolph Schroders Erben, 1752. Ježić, Mislav. “Vedic Myths and Epico-Purā ṇic Mythology—Transformation of the Sense of Vision into Narration.” Paper presented at the Dubrovnik International Conference on the Sanskrit Epics and Purā ṇas (DICSEP) 8, Dubrovnik, Croatia, September 12, 2017. de Jong, J.  W. Review of Viṣṇudharmāḥ . Precepts for the Worship of Viṣṇu. Part  1, Adhyāyas 1–43, by Reinhold Grünendahl. Indo-Iranian Journal 29, no. 1 (1986): 64–65. ———. Review of Viṣṇudharmāḥ . Precepts for the Worship of Viṣṇu. Part  2, Adhyāyas 44–81. Part  3, Adhyāyas 82–105 (with a Pāda-Index of Adhyāyas 1–105), by Reinhold Grünendahl. Indo-Iranian Journal 36, no. 2 (1993): 138–39. Kosambi, D. D. “The Autochthonous Element in the Mahā bhā rata.” Journal of the American Oriental Society 84, no. 1 (1964): 31–44. Kosegarten, Johann Gottfried Ludvig. Pantschatantrum, sive, Quinquepartitum de moribus exponens:  ex codicibus manuscriptis. Bonn: H. B. Koenig, 1848. Kupka, František. Abstraction, création, art non-iguratif, 1932–36. New York: Arno Press, 1968. ———. La Création dans les arts plastiques. Paris: Cercle d’art, 1989. Leighten, Patricia. The Liberation of Painting: Modernism and Anarchism in Avant-Guerre Paris. Chicago and London: University of Chicago, 2013. Levenson, Alan T. The Making of the Modern Jewish Bible: How Scholars in Germany, Israel, and America Transformed an Ancient Text. Lanham, MD: Rowman & Littleield, 2016. Lipsey, Roger. The Spiritual in Twentieth Century Art. Mineola, NY: Dover Publications, 2004. Meecham, Pam, and Julie Sheldon. Modern Art: A Critical Introduction. 2nd edn. Oxford: Routledge, 2005. Mládek, Meda, and Margit Rowell. “Chronology.” In František Kupka 1871–1957:  A Retrospective. New York: Solomon R. Guggenheim Foundation, 1975. Nietzsche, Friedrich. Anti-Education:  On the Future of Our Educational Institutions. Edited by Paul Reitter and Chad Wellmon. Translated by Damion Searls. New York: New York Review of Books, 2016. ———. The Gay Science. Edited by Bernard Williams. Cambridge:  Cambridge University Press, 2001. ———. “On Truth and Lying in a Non-moral Sense.” Translated by Ronald Speirs. In The Birth of Tragedy and Other Writings. Edited by Raymond Geuss and Ronald Speirs. 139–53. Cambridge: Cambridge University Press, 1999. ———. Twilight of the Idols. Translated by Richard Polt. Indianapolis and Cambridge: Hackett Publishing Company, 1997. Olivelle, Patrick. “Unfaithful Transmitters:  Philological Criticism and Critical Editions of the Upaniṣads.” Journal of Indian Philosophy 26 (1998): 173–87. O’Neil, Edward N. Review of Juvénal, les baladins et les rétiaires d’après le manuscrit d’Oxford, by Jean Colin. Classical Philology 51, no. 1 (1956): 42–44. 523 ANNOTATED BIBLIOGRAPHY 523 Pollock, Sheldon. “Crisis in the Classics.” Social Research: An International Quarterly, special issue titled India’s World 78, no. 1 (2011): 21–48. ———. “Future Philology? The Fate of a Soft Science in a Hard World.” Critical Inquiry 35, no. 4, The Fate of Disciplines (2009): 931–61. ———. “Introduction.” In World Philology, edited by Sheldon Pollock, Benjamin A. Elman and Ku- ming Kevin Chang, 1–24. Cambridge, MA: Harvard University Press, 2015. ———. “Kritische Philologie.” Translated by Brigitte Schöning. Geschichte der Germanistik 45/ 46 (2014):  5–12. English translation:  www.columbia.edu/cu/mesaas/faculty/directory/ pollock_pub/Kritische%20Philologie%20(+%20English%20original).pdf (accessed February 26, 2017). ———. Kritische Philologie:  Essays zu Literatur, Sprache und Macht in Indien und Europa. Edited by Christoph König. Translated by Brigitte Schöning. Göttingen: Wallstein Verlag, 2015. ———. “Philology & Disciplinarity.” Baraza:  Critical Collaboration on the Middle East, South Asia, and Africa, https://baraza.cdrs.columbia.edu/philology-disciplinarity/ (accessed February 26, 2017). ———. “Philology in Three Dimensions.” Postmedieval: A journal of Medieval Cultural Studies 5, no. 4 (2014): 398–413. ———. “Towards a Political Philology: D. D. Kosambi and Sanskrit.” Economic and Political Weekly, D. D. Kosambi Centenary Volume (2008): 52–59. ———. “What Was Philology in Sanskrit?” In World Philology, edited by Sheldon Pollock, Benjamin A. Elman and Ku-ming Kevin Chang, 114–36. Cambridge, MA:  Harvard University Press, 2015. Raeder, Siegfried. “The Exegetical and Hermeneutical Work of Martin Luther.” In Hebrew Bible/ Old Testament: The History of Its Interpretation, vol. 2: From the Renaissance to the Enlightenment, edited by Magne Saebø, 363–406. Göttingen: Vandenhoeck & Ruprecht, 2008. Ramanujan, A. K. “Repetition in the Mahā bhā rata.” In Essays on the Mahābhārata, edited by Arvind Sharma, 419–43. Delhi: Motilal Banarsidass, 2007. Reich, Tamar C. “Ends and Closures in the Mahābhārata.” International Journal of Hindu Studies 15, no. 1, The Mahābhārata—Perspectives on its Ends and Endings (2011): 9–53. Rowell, Margit. “František Kupka: A Metaphysics of Abstraction.” In František Kupka 1871–1957: A Retrospective. New York: Solomon R. Guggenheim Foundation, 1975. ———, ed. František Kupka 1871–1957:  A Retrospective. New  York:  Solomon R.  Guggenheim Foundation, 1975. Schechter, Solomon. “Higher Criticism—Higher Anti-Semitism.” In Seminary Address and Other Papers. 35–39. Cincinnati, OH: Ark Publishing, 1915. Schreiner, Peter. “Die Hymnen des Viṣṇupurā nạ: Materialien zur Textanalyse des Viṣṇupurā nạ.” Habilitationsschrift, University of Tübingen, 1980. Sheehan, Jonathan. The Enlightenment Bible: Translation, Scholarship, Culture. Princeton, NJ: Princeton University Press, 2005. Söhnen, Renate. Untersuchungen zur Komposition von Reden und Gesprächen im Rāmāyaṇa. 2  vols. Studien zur Indologie und Iranistik 6.  Reinbek:  Dr.  Inge Wezler, Verlag für orientalistische Fachpublikationen, 1979. Sørensen, Søren. An Index to the Names in the Mahabharata: With Short Explanations and a Concordance to the Bombay and Calcutta Editions and P. C. Roy’s Translation. 13 vols. London: Williams & Norgate, 1904–25. Teugels, Lieve. “Textual Criticism of Late Rabbinic Midrashim:  The Example of Aggadat Bereshit.” In Recent Developments in Textual Criticism:  New Testament, Other Early Christian and Jewish Literature, edited by Wim Weren and Dietrich-Alex Koch, 207–41. Assen:  Van Gorcum, 2003. Tieken, Herman. “The Mahābhārata after the Great Battle.” Wiener Zeitschrift für die Kunde Südasiens 48 (2004): 5–46. 524 524 PHILOLOGY AND CRITICISM Tuchman, Maurice. The Spiritual in Art: Abstract Painting 1890–1985. New York: Aberville Publishers, 1986. Vachtová, Ludmila. Frank Kupka, Pioneer of Abstract Art. Translated by Zdeněk Lederer, with an Introduction by J. P. Hodin. New York: McGraw-Hill, 1968. Wojtilla, Gyula. Review of Viṣṇudharmāḥ . Precepts for the Worship of Viṣṇu, Part 1 Adhyāyas1–43, by Reinhold Grünendahl. Acta Orientalia Academiae Scientiarum Hungaricae 39, no. 2/3 (1985): 392–93. Notes 1 The Ā śramavā sika, Mausala, Mahā prasthā nika and Svargā rohaṇa parvans appeared in one volume. 2 The Ś ā ntiparvan is the only book of the critical edition with a separate introduction. The Mahā bhā rata’s longest book was published in ive parts:  part I:  Rā jadharma; part II:  Ā paddharma and Concordance; part IIIA:  Mokṣadharma A  (Mahā bhā rata 12.168–317) and part IIIB:  Mokṣadharma B (Mahā bhā rata 12.318–53), along with the aforementioned introduction, which appeared as the last of the ive parts. 52 INDEX Page numbers in italics indicate tables, igures and illustrations. abstract art and concept of critical edition reconstruction of reading of 355, 356 xxiii–xxviii resilience, argument from, and redeinition adhyāyas 357, 455n22. See also structure of of archetype 61–62 Mahābhārata special status of 94n28 Adluri, Vishwa, and Joydeep Bagchee, The spread, argument from, and redeinition of Nay Science xv–xvi, 11, 27, 156n68, archetype 55–59 319, xxixn8 Aristotle 159n76, 275n14 analytic school 5, 53, 95n37, 112n139, Arjunamiśra 397, 400–403, 458nn39–40, 119, 126, 141, 143, 145n8, 146n9, 461n73, 462n79, 462n86, 463n96 148n16, 149n32 Aryanist/neo-Aryanist ideology xvii, 145n7, Ānandavardhana 398 319, 479–480n167, xxxn13 ancestor, concept of 12–13, 18 Austin, Christopher R. 111–13n139, anti-Brahmanism 70, 88–89, 145n6, 269 144–45n3, 276n26 anti-modernism xxiv, xxxn14 Āyodhyakāṇḍa manuscript 432 anti-Semitism xvii, 28n2, 140–41, 145n7, 157n70, 313n359 Bagchee, Joydeep, and Vishwa Adluri, The apographs Nay Science xv–xvi, 11, 27, 156n68, classiication issues 319, xxixn8 expertise, argument from 313n360 Barbi, Michele 14 extensive contamination, argument from bardic hypothesis 27, 93n25, 160n79 212, 230, 231, 233, 234–35, 240, 243, Baur, Ferdinand Christian 319 291n184, 298n250 Baur, George Lorenz 319 independent recensions, argument from Bédier, Joseph, and Bédierism 31n20, 33n30, 259, 301n277, 305n303 248, 268, 308–9n337 contamination, identifying source of Belvalkar, S. K. xvii, 89n2, 198–99, 288n162, 124–25, 126 401, 402, 461n72, 461n74, 461n76, deined 13 463nn86–88, 463nn93–97 apparatus criticus (critical apparatus) 13, Bengālī manuscripts 30n16, 80, 82, 187, 192, 17, 351–54 204, 205, 223 archetype Bentley, Richard 320, 328n16 Bigger’s assumption of 157n71 Bhagavadgītā 27, 40n88, 99n72, 100n81, coincidence of archetype and normative 289n168, 402, 425, 425–27, redaction 65–67, 66 463n94, 464n99 deined 13 Bhārgavas 102n98, 147n13 extant tradition, as source of irst branching Bhatta Bakre, Mahadeva Shastri Gangadhara of 96n47 401, 461n71 multiple meanings of 13, 18–20, 36n65, 67, Bhaṭṭācārya, Paramānanda 402, 403, 94n34, 97n60 463n88 normative redaction hypothesis, archetype Biardeau, Madeleine 42n92, 274, and original in 46, 46–49, 48, 92n14 317n395, 430 526 526 PHILOLOGY AND CRITICISM biblical criticism and theology, as background brevior lectio potior 171, 175 of German Indology 325n4, brevity, argument from 169–82, 170, 174, 327nn12–13 177–78, 179 Bidnur, Jahnavi 398 Brockington, John L. 142, 271, 273, 429, Bieler, Ludwig 31n20 464n104. See also Venn diagrams Bigger, Andreas Bronkhorst, Johannes 63, 65, 99nn75–76, archetype, assumption of 157n71 100n80, 269, 328–29n22 Brockington on Venn diagrams and 433, Burnell, A. C. 367, 455–56n23, 434, 466n109, 466n117 456–57nn25–26 on classiication and Grünendahl’s work 163–64, 167, 169, 275n13, 275n17, Catholicism 342, 479–80n167, xxx–xxxin15 277n28, 314–15n369 Caturbhuja 400, 401, 403, 463n95, on critical edition 157n72 464nn97–98 on iliation 275n17 Cicero, Ad Atticum 18 normative redaction hypothesis of xv, xvii, citation of Mahābhārata 357 21, 22, 28, 35n54, 45–46, 47–49, classical philology, spurious relationship of 101n88 (See also normative redaction German Indology to 319–24 hypothesis) classiication of Mahābhārata manuscripts oral source, on contamination from 132, 5–8, 163–274 133, 135–39, 140 brevity, argument from 169–80, 170, 174, on prehistory of normative redaction 121, 177–78, 179, 182 122, 143, 144 expertise, argument from 269–74 uncertainty, on contamination argued from extensive contamination, argument 126–29, 131 from 209–46 Witzel’s thesis compared 269 assimilation of K1 manuscript to Śāradā bipartite stemma 39n81, 125, 127, 167, 168 codex 241–46, 244–247 birth and death of manuscripts 24–26, 25 corruption, contamination as 210–15, Bornmann, Fritz 124 211, 212, 213, 214 bottleneck theories 18, 20, 36n65, 58, 65, 67, iliation, Grünendahl’s failure to 99n75, 321, 328n19, 328–29nn22–23 understand 233–35 Brahmā-Gaṇeśa interpolation 106n114, 201 Grünendahl’s understanding of Brahmanic tradition overriding Kṣatriya concept of classiication and 228–33, tradition 228–29, 231–33 anti-Brahmanism 70, 88–89, 145n6, 269 inlation (transmission of interpolations), Bhārgavas and 102n98, 147n13 contamination as 215–17, 216, 217 Brockington on Venn diagrams and non-contamination (absence 429, 432–33 of interpolations) as type of conservatism of Brahmanic contamination 217–30, 220, 222, 223, tradition 160n79 225–26, 227, 228, 295–96n218 as contamination 119–22, 126–31, 132, ‘purity’, Grünendahl’s deinition of 139, 140–41, 144, 150–51n40, 153n59 294n208 distinction of Vedic-Brahmanic and Sukthankar’s identiication of K group as Epic-Heroic traditions 467–68n118 separate class 235–41, 237–40 in normative redaction hypothesis 45, false premises, argument from 52, 54, 59–60, 62, 63, 66–67, 68, 169–70, 180–82 71, 72, 74–75, 87–88, 97–98n65, iliation, determining 164–67 100n78, 100n80 independent (regional) recensions, argument origins of theory of 155–56n65 from 247–69 textual critics’ response to critical edition confusion leading to 290n175 and 22, 24, 27, 28n2, 38n74, ‘criticism of the base manuscript’ 38n79, 40n88 approach 248, 266–69 527 INDEX 527 superiority of Ñ4 to K1, 248, 257–66, contamination 4–5, 119–44. See also under 261, 265 classiication of Mahābhārata superiority of Newārī to Kaśmīrī manuscripts manuscripts generally 248–56, 250, as Brahmanic takeover or redaction 252, 254, 256 of Kṣatriya tradition 119–22, by meliores and deteriores 310–11n345 126–31, 132, 139, 140–41, 144, misapprehension concerning classiication 150–51n40, 153n59 (Schriftartprämisse), argument from deining 119–20 182–209 extra-archetypal 146–47n10 content of text, classiication properly extra-stemmatic 120–22, 126, 127–30, 132, based on 192–96, 205, 278n50 143, 145–46n9, 152n54 geographic or script-based presentation, hyperarchetypal 120–22, 132, 143, assumptions regarding 191–92, 146–47n10 204–8, 207 intra-stemmatic 120, 131, 132, K siglum, misunderstanding of use 146–47n10, 149n33 of 203–4 normative redaction hypothesis and 121, Kaśmīrī version and K siglum, critique of 122, 126–27, 129, 131 Sukthankar’s use of 200–5 objective criteria for identifying source of presentation or organization of material 122–26, 123 185–92, 187–88, 192 from oral source 131–39, 133–35, 137, 138, reproduction of Sukthankar’s stemma 152n53, 155n64 and 184, 184–85 postulated antiquity and ideology, single manuscript as evidence of version arguments from 140–44, 142, 143 of text 196, 196–200, 198, 199 ‘purity’, Grünendahl’s deinition of racial classiication theories inluencing 294n208 313n360, 318 Sukthankar on 301n271 typological versus genealogical 163–64 uncertainty, argument from 126–31, witness texts, eliminating 167–69, 168, 127–30, 151n46 284–85n127 content of text, classiication properly based codex descriptus (codices descripti) 197, on 192–96, 205 284–85n127, 434–35, 473–74n133 Contini, Gianfranco 13, 50, 156n67, codex optimus 31n20 281n100, 330n27 codex unicus 19, 35n53, 47, 50, 60, 141, corruption, contamination as 210–15, 211, 248, 322 212, 214, 215 commentaries and commentators. critical apparatus (apparatus criticus) 13, 17, See also speciic commentators 351–52, 352–54 on Bhagavadgītā 425, 425–27 critical edition. See also Mahābhārata critical on Mahābhārata 397–403, 403–18, edition 458n38 abstract art and concept of xxiii–xxviii role of 12, 63, 99n75, 159–60n78 Bigger on 157n72 Sukthankar on study of 317n393 deined 324 variant readings, not concerned by 292n188 German Indologists on concept of common-error method 47, 324, 328n18 335–36n53 conjunctive errors 34n36, 55–56, 57, 67, 74 plurality of texts, as arrangement of xxviii, conservatism xxxin28 of Brahmanic tradition 160n79 reasons for creating 11–12 of geographic peripheries relative to as reconstruction of reading of archetype linguistic center 432, 454n9 355, 356 of Mahābhārata tradition as text versus working tool 15, 32n24 154–55n63, 454n8 critical philology of Pollock 272, constitutio textus (critically constituted text) 13 315–16nn383–84 528 528 PHILOLOGY AND CRITICISM critically constituted text (constitutio textus) 13 extra-archetypal contamination 146–47n10 crux 7, 14, 94n31, 129–30, 247, 267, 356 extra-stemmatic contamination 120–22, 126, 127–30, 133, 134, 143, Dandekar, R. N. 401, 456n24, 145–46n9, 152n54 462n79, 463n91 death of manuscripts 24–26, 25 false premises, argument from 169–70, Debroy, Bibek 347–48, 455n17 180–82 decimation rate 24 iliation Dembowski, Peter F. 165, 311n355 Bigger on 275n17 descendant manuscripts, concept of 96n56 Brockington’s Venn diagram proposal Devabodha 317n393, 397–403, and 435 458–9nn40–41, 459n44, 460n68, classiication of manuscripts and 163–67 461n73, 462n79, 463nn93–94, 463n95 Grünendahl’s failure to understand Devanāgarī manuscripts 233–35 Burnell’s stemma and 456–57n26 interpolations/absence of interpolations as classiication issues insuicient basis for 214, 216–17, 219, expertise, argument from 271 225–26, 230 extensive contamination, argument from normative redaction hypothesis and 209, 223, 224, 235 69, 102n92 independent recensions, argument from inal redaction (normative redaction) 248, 308n327, 310–11n345, 312n357 hypothesis of von Simson 19–24, 27, misapprehension concerning classiication 28, 33n34, 35n54, 36–37nn66–67, (Schriftartprämisse), argument from 184, 40n88. See also normative redaction 186–87, 193, 195, 200, 201–4, 205, hypothesis 206nn8–10, 281n103, 281–82n108, Fitzgerald, James L. xv, xvi, 21, 28–29n3, 285–86n136, 287–88nn160–62, 35n54, 36n64, 41–42nn91–92, 289nn166–68 112–14n139, 271–72 commentaries and 462n86 Flores, Enrico 55–60, 58, 62 in normative redaction hypothesis 80, Fontaine, Jacques 142 102–3n100, 105–6n112, 110n130 Foucault, Michel, The Order of Things, vii Telegu recension and 30n16 Fraenkel, Eduard 319 Venn diagrams and 465–66n108 Franco, Eli 325–26n6, 464n104 diaskeusis 92n20, 101n86, 114n143, 155n65 Frankensteinism 14–15, 29–30n12 Doniger, Wendy 14–15 Froger, Jacques 164, 166, 275n18 double agreement 280n84 Gauguin, Paul xxx–xxxin15 ecdotics xviii, xviii–xixn7 genealogical-reconstructive method 13, Edgerton, Franklin 89–90n2, 200–201, 268, 324, 355, 429, 470n124. See also 286nn140–42, 288–89nn164–65, stemma/stemmata 296n224, 300–301n262, 304n300, genealogy of German Indology 342 305n306, 400 genetic racism 313–14n360, 318 Edwards, A. W. F. 472n127 genetic relationships between manuscripts empty reference, argument from 67–75, 33n35, 217, 218, 289n165, 292n191, 72, 102n91 293n193, 452, 477n158 English translations of Mahābhārata genetics, compared to philology xxiii, 37n69, 347–50 276n24, 313–14n360 epic, concept of 153–54n62 geographic presentation of material, epistemic prejudices 145n7 classiication misapprehension based Euler diagram 472n127 on 191–92, 203–8, 207 Euripides, Iphigenia in Aulis xxvii geology, Indian historical writing inluenced expertise, argument from 269–74 by 325n4 529 INDEX 529 German Indology and German Indologists structure and arrangement 367 319–24, 339. See also speciic German textual critics’ reception of Mahābhārata Indologists critical edition and 30n16, 38n73 anti-Semitism of xvii, 28n2, 140, 145n7, Griith, John G. 451–52, 475–77nn149–58 157n70, 313n359 Grünendahl, Reinhold xv, 21 Aryanist/neo-Aryanist ideology of xvii, Brockington on Venn diagrams and 433, xxxn13, 145n7, 319, 479–80n167 434, 465–66n108, 466n109, 479n165 biblical criticism and theology inluencing classiication theories of 164, 169 (See 325n4, 327nn12–13 also classiication of Mahābhārata Brahmanic tradition overriding Kṣatriya manuscripts) tradition, belief in 22, 24, 27 (See critical edition, on concept of 335–36n53 also Brahmanic tradition overriding Mahābhārata manuscripts, lack of Kṣatriya tradition) experience with 289n167, 290n173 Brockington’s Venn diagrams and 433–35, normative redaction hypothesis and 86 466–67n118 sole edition (of Viṣṇudharmāḥ) produced classical philology, spurious relationship by 308–9n337, 310–11n345, 311n350, to 319–24 312n357 deined 324n1 on textual criticism 311–12n356, 326nn7–8, genealogy of 342 333–34n49 Mahābhārata critical edition, reception of Guidi, Vincenzo 24 20–28, 23, 25 (See also Mahābhārata Günther, Hans F. K. 313–14n360, 318 critical edition) Gupta redaction hypothesis of Fitzgerald xv, oral hypothesis, reliance on xv–xvi, 28, 21, 26, 35n54, 41–42n91, 113n139 28–29n3, 36–37n66, 41n90, 89–90n2, Gurd, Sean Alexander xxix–xxxn10, xxxin23 429–30 (See also oral hypothesis) ‘original’ meanings, conidence in ability to Hacker, Paul 94–95n35 recover 332n40, 334–35n52 Hahn, Michael 116–17nn147–49 Protestantism, inluence of 15, 32n23, 88, Hellwig, Oliver 100n81 140, 154–55n63, 156n69, 160n79, Herder, Johann Gottfried 320 313n359, 319, 327n12, 327–28n15, higher criticism, concept of 49–53, 51 339, 342 Hiltebeitel, Alf 273, 316–17n390, 379 realia and 339 Rethinking the Mahābhārata xvi recognition of 299–300n256 Hinüber, Oskar von 31–32n22, 46, responsibility, lack of sense of 480n168 270–71, 434 Gode, P. K. 400, 460n57 historical and textual criticism, confusion Goldstücker, Theodor 155n65 of 328n16 Grantha edition 359 historical-critical method 114–15n144, Grantha manuscripts 327–28nn15–16, 342 Burnell’s stemma and 393, 456–57n26 Hitler, Adolf xxxn13 classiication issues Holtzmann, Adolf, Jr. 27, 28n3, 40– expertise, argument from 271, 273 41n89, 88, 115n146, 153nn60–61, extensive contamination, argument from 325nn4–5, xxixn8 210, 211, 213, 214, 291n181, 293n192 Holtzmann, Adolf, Sr. 325n4 independent recensions, argument from Homer, Mahābhārata compared to 35n58, 303n285 40n88, 97n60, 270, 333n45, 430 misapprehension concerning classiication Hopkins, Edward W. 114n142, xxixn8 (Schriftartprämisse), argument from horizontal transmission 50, 80, 91n6, 183, 186, 187, 192, 193, 204, 206n7, 108n122, 119, 216, 255, 292n191, 282n109, 283–84n121, 289n170 296n225–26, 450 commentaries and 463n88 Howe, Christopher J. 39n82 Nāgarī/Grantha bottleneck 329n23 Hudson, Emily xxxi–xxxiin29 530 530 PHILOLOGY AND CRITICISM Human Genome Project 37n69 independent recensions, argument hyparchetypes from 248–49, 252, 257, 304–5n301, classiication issues and 308n327 brevity, argument from 280n86 misapprehension concerning extensive contamination, argument from classiication, argument from 193, 209, 212, 216, 218, 235, 238, 243, 200–204, 281–82n108, 284–85n127, 293–94n200 286n142, 287–88nn159–60 independent recension, argument contamination and 131 from 253, 254, 256, 257, 301n277, normative redaction hypothesis and 302nn283–84, 304n292, 310n343 106–7n115 misapprehension concerning classiication textual critics’ responses to Mahābhārata (Schriftartprämisse), argument from 185, critical edition and 16 191, 197, 282n110, 285n130, 285n133 Kosambi, D. D. 147n12 witnesses, eliminating 168 Kṣatriya tradition. See Brahmanic tradition contamination and 148n27 overriding Kṣatriya tradition deined 14, 29n10, 51 Kupka, František xxiii, xxvi–xxviii, in normative redaction hypothesis 14, xxixn1, xxixn7 29n10, 51, 63 Hindu Motif (Graduated Red) xxvii, xxxin27 hyperarchetypal contamination 120–22, 132, Le Premier Pas (The First Step) xxiii–xxvii, 143, 146–47n10 xxviii hyperarchetypal inference. See normative redaction hypothesis Lachmann, Karl 15, 33n30, 93n26, 122–23, 196–97, 248, 319, 328n16, 334n50 ideology, argument from 140–44, 142, 143 Lariviere, Richard 334n51 independent recensions, argument from. Lassen, Christian 88, 101n86, 114n143, See under classiication of Mahābhārata 155n65, 157n70, 325n4 manuscripts latent errors 23, 50, 93–94n27, 330n27 inlation (transmission of interpolations), Lears, T. J. Jackson xxxn14 contamination as 215–17, 215, 216, 217 Leclere, Basile 397–98, 459n44 intra-stemmatic contamination 120, 131, 132, lectio diicilior 27, 263, 264, 265, 320, 431, 146–47n10, 149n33 474n133 Irigoin, Jean 119, 296–97n225 lectio heroicior praeferenda est 139, 152–53n58 Isidore, Etymologies 141–43, 142, 158–59n76 Lejay, Paul 275–76n18 Leonardi, Lino 13, 15, 163, 248, 267–68, Jacob, Benno 313n359 281n100, 331–32n38, 480n168 James, William, The Varieties of Religious Lévi, Sylvain 16, 33n30 Experience xxxn14 Lipsey, Roger, Disks of Newton xxv–xxvi Jamison, Stephanie 326–27n9 loss, argument from 75–88 Janamejaya 28 lost original, eforts to reconstruct 93n25 Jerusalem, Johann 88 lower criticism 50, 52, 93n24, 94n34, 276n26 Joyce, James, Finnegans Wake 429 Lüders, Heinrich 183, 185, 192–93, 204, 209–10, 270, 281–82n108, 290n175 Kaṇikanīti xxxi–xxxiin29 Luther, Martin 32n23, 342 Kannada script 431 Kantorowicz, Hermann 18, 319 Maas, Paul xvii, 50–52, 51, 55, 60, 92n17, Kaśmīrī version 94n28, 165, 167, 319 classiication issues Maas, Philipp 36n65 brevity, argument from 173 Mahābhārata. See also structure of extensive contamination, argument from Mahābhārata 242, 293–94n200 abbreviated concordance of principle false premises, argument from 181 editions 389, 389–90 531 INDEX 531 average Indian’s understanding of 274, salaries of editors of 479n160 317n392 stemmata used to construct 13 (See also citation of 357 stemma/stemmata) conservative nature of tradition table of manuscripts collated for 384 154–55n63, 454n8 as text versus working tool 15, 32n24 editions prior to critical edition 345–46 textual critics’ reception of 20–28, 23, 25 English translations 347–50 variant readings, scribes and commentators northern and southern recensions not concerned by 292n188 150nn38–39, 361, 367–68, 368–70 volumes 343–44 number of verses 359, 359–60 witness texts, role of xxxi–xxxiin29 oldest dated manuscript of Ādiparvan written text, Mahābhārata viewed as xvi 39n83, 77–78, 82, 110n127, Mahābhārata vulgate 16, 471n126 110n129–30, 162 Mahadevan, T. P. xvi, 466n109 Mahābhārata critical edition xv–xvi, 1, 11–28 Maithili manuscripts 82, 223, 461n76 abbreviations and diacritical signs 387–88 Malayām manuscripts abstract art as means of understanding Brockington on Venn diagrams and concept of xxiii–xxviii 432, 435 archetype, concept of 13, 18–20 (See also Burnell’s stemma and 456–57n26 archetype) classiication issues 184, 192–93, 198–99, Brahmanic tradition overriding Kṣatriya 210–11, 257, 283–84n121, 289n166, tradition, German Indologists’ belief in 291n180 22, 24, 27 (See also Brahmanic tradition parvans, arrangement of 367 overriding Kṣatriya tradition) Marichal, Robert 45 contamination and 119–44 (See also McGrath, Kevin 479–80nn166–67 contamination) Meecham, Pam xxixn7 critical apparatus, deinition and use of 13, Mehendale, M. A. 470–71n125 17, 351–52, 352–54 mimesis and Mahābhārata xxvii–xxviii deined and described 12–16, 13, 17 Minkowski, Christopher 397, 459n43 editors of 343 Müller-Sievers, Helmut 334n50 inal redaction hypothesis of von Simson Mylius, Klaus 147n12 19–22, 24, 27, 28 irst editorial board 44 Nāgarī recension 204, 393, 456–57n26 Frankensteinism, charged with 14–15, Nāgarī/Grantha bottleneck 329n23 29–30n12 Nandināgari script 206n11, 288n163, 431 German scholars, recognition of narrative units (upākhyānas or subtales) of 299–300n256 Mahābhārata 379–81 Gupta redaction hypothesis of Fitzgerald xv, Nepālī manuscripts 21, 26, 35n54, 41–2n91, 113n139 classiication issues interpretation of 14–20 expertise, argument from 271 mimesis, central role of xxvii–xxviii extensive contamination, argument from normative redaction hypothesis 21, 28, 224, 233 35n54, 45–88 (See also normative false premises, argument from 278n50 redaction hypothesis) independent recensions, argument from oral hypothesis, German Indologists’ reliance 247, 248, 259, 267–68, 307nn318–19 on xvi, 28, 29–30n3, 36–37n66, 41n90 misapprehension concerning classiication (See also oral hypothesis) (Schriftartprämisse), argument from 190, plurality of texts, as arrangement of xxviii, 285n129, 289n167, 290n173 xxxin28 in normative redaction hypothesis 77, 79, reasons for creating 11–12 82–86, 110n129 reconstruction of reading of archetype oldest dated manuscript of Ādiparvan 355, 356 39n83, 82–86, 107n116, 110n127, 162 532 532 PHILOLOGY AND CRITICISM Newārī manuscripts 170, 206n3, 248, 252, failure of scholars to make case for 312n357 479–80nn166–67 Nibelungenlied xxvii, 123, 196 normative redaction hypothesis and Nietzsche, Friedrich xvii, xxxn11, 89–90n2 xxx–xxxin15, 75, 99n69, 105n106, 453 textual critics’ responses to Mahābhārata Nīlakaṇṭha 12, 16, 33n31, 45, 91n6, 200, 270, critical edition based on belief in xvii, 317n393, 397–403, 430, 456–57n26, 28, 28–29n3, 36–37n66, 41n90 461n73, 463n92 oral source, contamination from 131–40, non-contamination (absence of interpolations) 133–35, 137, 138, 152n53, as type of contamination 217–30, 220, 155n64 222, 223, 225, 227, 228, 295–96n218 organization or presentation of material, Nordic type, classiication of 313–14n360, 318 classiication misapprehension based normative redaction hypothesis 2–3, 45–89 on 185–91, 187–88, 191, 192 anti-Brahmanism of 88 Oriya script 431 archetype and original in 46, 46–49, 48, 92n14 Paramānanda. See Bhaṭṭācārya, Paramānanda as if logic of 88, 115–16n146 Paranjpe, V. G. 401 coincidence of archetype and normative Paris, Gaston 275–76n18 redaction in 65–67, 66 parvans 357, 359–60, 361, 361–65, 367–68, comprehensive revision of text, pointing 368–70, 371–77, 393, 393–96, to 92n18 455n18, 456–7n26. See also structure of contamination and 121, 122, 126–27, Mahābhārata 129, 131 Pasquali, Giorgio 31n19, 98–99n68, 454n9 deined and described 45–46 Storia della tradizione e critica del testo empty reference, argument from 67–75, xvii, 134–35 72, 102n91 Philipps-Rodriguez, Wendy 433 loating oral tradition and 101–2n88, philology 112n139, 138, 152n56 classical philology, spurious relationship of higher criticism and 49–53, 51 German Indology to 319–24 loss, argument from 75–88 critical philology of Pollock 272, prehistory of 121, 122, 143, 144 315–16nn383–84 resilience, argument from 54, 61, 61–67 deined by Witzel 340n3 space-time continuum of Indian history “post-philological,” as concept 320, 326n7 and 339–40n2 Pollock, Sheldon I. 272, 315–16nn383–84, spread, argument from 54–61, 58 478n161 textual critics’ reception of Mahābhārata popular narratives critical edition and 21, 28, 35n54 classiication and 155n65, 158n75 uniform redaction, based on xv, 45, 47, normative redaction hypothesis and 63, 89–90n2 76, 78, 79–80, 81, 82, 91n6, 105n105, Nuremberg Race Laws (1935) 313–14n360 108n122 textual critics’ response to critical edition Oldenberg, Hermann xxixn8 and 28, 38n74 oldest dated manuscript of Ādiparvan 39n83, Venn diagrams, Brockington on 435 82–86, 107n116, 110n127, 162 “post-philological,” as concept 320, 326n7 oral hypothesis postulated antiquity, argument from 140–44, bardic origins attributed to Mahābhārata 142, 143 and 27, 93n25, 160n79 presentation or organization of material, Brockington on Venn diagrams and 429–30, classiication misapprehension based 434, 468–69nn120–21 on 185–91, 187–88, 191, 192 distinction of Vedic-Brahmanic and Protestantism and Mahābhārata textual Epic-Heroic traditions 467–68n119 criticism 15, 88, 140, 154–55n63, 53 INDEX 533 156n69, 160n79, 313n359, 319, classiication issues 327n12, 327–28n15, 339, 342 brevity, argument from 169, 171, ‘purity’, Grünendahl’s deinition of 294n208 172, 175, 176, 178, 180, 278n48, 280n82 quadripartite stemma 125 extensive contamination, argument from Quellenkritik 27, 158n75 235, 236, 241, 243, 293–94n200 false premises, argument from 181, 182 rabbinic literature and textual criticism misapprehension concerning classiication 154–55n63 (Schriftartprämisse), argument from 186, racial classiication 313n360, 318 197, 199, 200, 201, 202, 285n127, Rāmakṛṣṇa (commentator) 400, 403 286n142, 287n155, 287n159 Rāmānuja 401, 403, 464n103 contamination and 151n43 Ramanujan, A. K. xxvii extent of, for Ādiparvan 385 Rāmāyaṇa and Rāmāyaṇa critical edition textual critics’ responses to Mahābhārata Brockington on Venn diagrams and 429, critical edition and 38n73, 39n83 431–33, 435, 447, 450, 465–66n108, Śāradā manuscripts 322 467–68n119, 468–69nn120–21, Brockington on Venn diagrams and 431 473–74n133, 475n139, 477–78n161 classiication issues classiication issues 230–31, 282–3n113 brevity, argument from 176 contamination and 153n61, expertise, argument from 270, 153–54n62 271–72, 273 criticism and 104–5n104 extensive contamination 210, 225, 228, Ratnagarbha 400, 403, 464n98 230, 235–36, 242–43, 291n180, Rau, Wilhelm 334n51 293–94n200, 297n235, 297n239, real trees 19, 23, 24, 63, 65, 66, 67, 98n67, 300n259 99n75, 437 independent recensions, argument from recentiores deteriores 137, 138 258, 308n327 recentiores non deteriores 134–35, 135, misapprehension concerning classiication 136, 450 (Schriftartprämisse), argument from 184, Reeve, Michael D. 55–59, 61–2, 111n132, 186, 187, 195–97, 199–204, 205, 145–46n9, 148–49n28, 168, 212, 206n12, 281n103, 282n108, 285n129, 278n49, 284n125, 475n145 286n136, 287n155, 287–88n160, regional recensions, argument from. See 288n162, 289n168, 289n170 under classiication of Mahābhārata contamination and 131 manuscripts normative redaction hypothesis and 63, 77, Reichenbach, Karl von xxxn13 91n6, 106–7n115, 110nn129–30 resilience, argument from 54, 61, 61–67 textual critics’ responses to Mahābhārata Rizzo, Silvia 279n62 critical edition and 38n73 Roman Catholicism 342, 479–80n167, Sarvajña-Nārāyaṇa 397, 400, 401, xxx–xxxin15 403, 463n95 Rousseau, Jean Jacques 320 Śaunaka 28 Rowell, Margit xxiv Schechter, Solomon 313n359 Ruben, Walter 282–83n113, 283–84n121, Schriftartprämisse. See under classiication of 290n173 Mahābhārata manuscripts script-based presentation of material, Sāṃkhya school 398 classiication misapprehension based Sāṃkhyakārikā 398 on 191–92, 204–9, 207 Śaṅkara 401, 403, 464n103 Sedulius Scottus 278n49 Śaṅkarācārya 425 Semler, Johann Salomo 319 Sannazaro, Arcadia 24 Sheehan, Jonathan xixn8 Śāradā codex Sheldon, Julie xxixn7 534 534 PHILOLOGY AND CRITICISM Shillingsburg, Peter L. 336–37n54 ‘style’, concept of 145n5 signiicant errors 7, 56, 82, 172, 209, 233, subjective reconstruction 37n67 276n25, 292n189, 302n280, 310n345, subtales (upākhyānas or narrative units) of 437, 439 Mahābhārata 379–81 single manuscript, as evidence of version of Sukthankar, V. S. xvii–xviii, 10. See also text 196, 196–99, 198, 199 Mahābhārata critical edition single manuscript, emerging consensus that abstract art, creation of critical edition reconstructive editions should respect compared to xxiii–xxviii integrity of 331–32n38 Calcutta edition numbering criticized by Slaje, Walter 269, 270 457–58n35 śodhapatra 11 commentaries, on study of 317n393 Söhnen, Renate 430, 465n107 on conservative nature of Mahābhārata Spencer, Matthew 39n82 tradition 154–55n63 spread, argument from 54–61, 58 on contamination 301n271 stemma/stemmata describing critical edition 13, 16, 18 abstract art, creation compared to Edgerton on death of 286n140 xxiii–xxviii educational background xxixn9 apparent narrowing of tradition indicated irst(?) attempt at stemma codicum of by 46, 46–49, 48 Ādiparvan versions xxii bipartite 39n81, 125, 127, 167, 168 German scholars, recognition of criticism of stemmatic method 93–94n27 299–300n256 deined xxixn3, 96n58 on higher criticism 53 Griith’s tabulation of agreements on lack of loss in the tradition 26, compared 451–52, 475–77nn149–58 109–10n124 Grünendahl’s attack on (See classiication of motivation and purpose of 478n163 Mahābhārata manuscripts) on northern and southern manuscripts Maas’s hypothetical stemma 50–52, 51 150nn38–39 for parvans 393, 393–96 plurality of texts, critical edition as quadripartite 125 arrangement of xxviii, xxxin28 real trees 19, 23, 24, 63, 65, 66, 67, 98n67, on readability of critical edition 32n24 99n75, 437 script and version, on concommitance recensio, stemma emerging from 185–86 between 288n161 shapes of 98n66 search for lost oral epic rejected by 95n37 stemma codicum, deined 13 on spurious passages 107–8n121 Sukthankar’s irst(?) attempt at stemma table of manuscripts collated for codicum of Ādiparvan versions xxii Mahābhārata 384 tripartite 30n13, 50, 125–26, 169 on vulgate versus critical edition 471n126 Venn diagrams versus 429, 448, 450 (See also synthetic school 145n8 Venn diagrams) Witzel on 331n36 Taine, Hippolyte 299n256 Stercutus 159n76 Telugu edition 346, 359, 454n11 Stietencron, Heinrich von 312–13n358 Telugu manuscripts stock phrases, use of 118 classiication issues structure of Mahābhārata 357 expertise, argument from 271, 273 adhyāyas 357, 455n22 extensive contamination, argument from narrative units (upākhyānas or 209–11, 291n181 subtales) 379–81 misapprehension concerning classiication number of verses 359, 359–60 (Schriftartprämisse), argument from parvans, arrangement of 357, 361, 361–65, 182, 183, 187, 193, 206n10, 367–68, 368–70, 371–77, 455n18, 288–89nn163–64 456–57n26 normative redaction hypothesis and upaparvans 357, 361, 361–65, 455n22 104n100, 107–8n121 53 INDEX 535 parvans in 367 Vaiśaṃpāyana xxvi, 28, 458n40 textual critics’ reception of Mahābhārata Vaisṇava devotional tradition 397, 398 critical edition and 14, 30n16, 38n73 van Buitenen, J. A. B. 145n6, 147n12 Teugels, Lieve 154–55n63 variant readings (variae lectiones) 12–13, 33n34 text-critical approach 27, 37n67, 53, 73, 76, Velankar, H. D. 367, 401 82, 94–95n35, 144–45n3, 272 Venn diagrams 429–53 text-historical approach 52, 94–95n35, 235 concept of 435, 436 textual criticism Euler diagram versus 472n127 classical philology, spurious relationship of iliation issues 437–38 German Indology to 319–24 German Indologists inluencing Brockington of German Indology (See German Indology on 433–35, 466–67n118 and German Indologists) graphical representation versus 447–48 historical and textual criticism, confusion Griith’s tabulation of agreements of 328n16 compared 451–52, 475–77nn149–58 Mahābhārata critical edition and (See ideological stance of Brockington and Mahābhārata critical edition) 477–78n161 single manuscript, emerging consensus that inadequacies of using 448–51 reconstructive editions should respect Mahābhārata critical edition, Brockington’s integrity of 331–32n38 views on 429–33, 449–50 textual unconscious xixn8 oral hypothesis and 429–31, 434, 453, textus simplicior 176, 242, 245, 280n86, 468–69nn120–21 310n343, 431 problems with Brockington’s presentation of theology and biblical criticism, as background 438–43, 440–43 of German Indology 325n4, stemmata versus 429, 448, 450 (See also 327nn12–13 stemma/stemmata) theosophy xxiv, xxxn11, xxxn13 true Venn diagrams, use of 443–48, Timpanaro, Sebastiano 15, 18, 23, 34–5n46, 444–46, 447 120, 123, 124, 126, 145–46n9, 146n10, vertical transmission 80, 216, 219, 230, 152n54, 196–97, 275n14, 276n27 310n345, 450 tripartite stemma 30n13, 50, 125–26, 169 Vidyāsāgara 402, 403 Trivandrum manuscript 431, 432, 434, 435, Vimalabodha 397, 400–403, 461n76, 464n97 449, 473n131 Vira, Raghu 200, 285–86n136, 289n166, Trovato, Paolo xvii, 11, 18–20, 23, 24, 34n36, 400 67, 93n25, 146–47n9, 147n10, 319, Viṣamapadavivaraṇa 400 330n26, 332n40, 451, 473–74n133, von Büren, Veronika 141–44, 142, 158n75 478n162 von Roth, Rudolf 325n4 Tuchman, Maurice xxv, xxvi, xxxn13, xxxin24 von Simson, Georg, inal redaction (normative redaction) hypothesis of 19–22, 24, Ugraśaravas xxvi, 28, 48–69n120 27, 28, 33n34, 35n54, 36–37nn66–67, uncertainty, argument from 126–31, 40n88 (See also normative redaction 127–30, 151n46 hypothesis) uniform redaction, concept of xv, 45, 47, vulgate text of Mahābhārata 16, 471n126 89–90n2 Vyāsa xxvi, 28, 40n88, 41n90, 92n20, upākhyānas (narrative units or subtales) 379–81 107–8n121, 180, 274, 458n40 upaparvans 357, 361, 361–65, 455n22. See also structure of Mahābhārata Wagener, Otto xxxn13 Urepos 12, 28n2, xxxn11 war narrative hypothesis 27 usus scribendi 27, 142, 265 Weitzman, Michael P. 39n81 West, Martin L. 131, 135, 136, 171, 186, Vādirāja 400–403, 460n63, 461n72, 281n106, 290n172, 451–52, 475n146 462–63n87 Textual Criticism and Editorial Technique xvii Vaidya, P. L. xvii, 288n163 Whitman, Walt xxxn14 536 536 PHILOLOGY AND CRITICISM Winternitz, Moriz 12, 29n5, 41n90, Witzel, Michael 269, 314nn361–62, 114–15nn144–45, 153–54n62, 314n364, 320–24, 335–36n53, 155–56n65, 158n75, 274n2 340nn3–4 witness texts xxxi–xxxiin29, 167–69, 168, Wolf, F. A. 31n21, 320, 328n16, 284–85n127 471n126