Bloomfield and the Sanskrit Origin of the Terms Exocentric and Endocentric

TABLE OF CONTENTS HISTORIOGRAPHY LINGUISTICA IX: 1/2 Editorial..............................................iii ARTICLES / AUFSATZE Jakob Ilornemann Bredsdorff: On the Causes of Linguistic Change (1821), transl., with an essay by Hcmiing Andersen (Copenhagen) .............I Robin N. Campbell (Sterling, Scotland) and Robert Grieve (Perth, Australia): Royal Investigations of the Origin of Language...................43 Genevieve Clerico (Reims): A Propos du Cesar Chesneau Bit Marsais et son role dans/'evolution de la Grammaire Generate (1928) de Gunvor Sahlin - 75 Martin L. Manchester (Somerville, Mass.): Philosophical Motives in Wilheini von Humboldt's Defense of the Inflectional Superiority Thesis..........107 Ingrid Tieken-Boon van Ostade (Leiden): Benjamin Martin the Linguist ... 121 REVIEW ARTICLE / RAPPORT CRITIQUE / FORSCHUNGSBERICHT Stanley S. Newman (Albuquerque, N.M.) Toward a History of American Lin- guistics ............................................135 REVIEWS / COMPTES RENDUS / BESPRECHUNGEN N. A. Kondrasov, htorija Hngvisticeskix ucenij (Moskva, 1979), reviewed by David L Olmsted (Davis, Calif.) ...........................145 Konrad Koerner, Hans-J. Niederehe, and R. H. Robins, eds., Studies in Medieval Linguistic Thought (Amsterdam, 1980), reviewed by Charlcnc McDcrmott (Albuquerque, N.M.)...................................152 Peter C. Rollins, Benjamin Lee Whorf (Ann Arbor, Mich., 1980), reviewed by Stephen O. Murray (Berkeley, Calif.)........................J 56 R. W. Hunt: Tlie History of Grammar in the Middle Ages: Collected papers, ed. by G. L, Bursill-Hall. (Amsterdam, 1980), reviewed by Sten Ebbescn (Copenhagen).......................................161 MISCELLANEA: DISCUSSIONS / DISKUSSIONEN - NOTES / NOTIZEN Robert A. Hall, Jr. (Ithaca, N.Y.): Karl Bartsch (1832-88)............165 Jan Noordegraaf (Amsterdam): The Port-Royal Grammar: A bibliographical note..............................................169 Pierre Swiggers (Leuven): 'Portraits of Linguists' anno 1927 ........... 175 Dominik Wujastyk (Oxford): Bloomfield and the Sanskrit Origin of the Terms 'exocentric' and *endocentric*.............................179 Frederick J. Newmeyer: Reply to Murray's Review.................185 Stephen O. Murray: The Reviewer responds. . ....................187 PUBLICATIONS RECEIVED / OUVRAGES REfUS / EINGEGANGENE SCHRIFTEN...........................................189 Wujastyk, Dominik: "Bloomfield and the Sanskrit Origin of the Terms 'ExoGen^c* and 'Endocentric'". In: Historiographia Linguistica, Volume IX, nol/2(1982). HISTORIOGRAPHY LINGUISTICA latemational Journal for the History of Linguistics Revue Internationale pour Fhistoire de la linguistique Internationale Zeitschrift fdr die Geschichte der Sprachwissenschaft EDITOR ASSOCIATE EDITOR E. F, KONRAD KOERNER (Ottawa) HANS-JOSEF NIEDEREHE (Trier) ADVISORY EDITORIAL BOARD Raimo Anttila (Los Angeles) Arthur L. Blumcnthal (Cambridge, Mass.) Ranko Bugarski (Belgrade) Lugenio Coseriu (Tubingen) Jacek Fisiak (Poznari) Dell Hyrnes (Philadelphia) Giulio C. Lepschy (Reading) Paolo Ramat (Pavia) Rosane Rocher (Philadelphia) Jean Stefanini (Marseille) Daniel J. Taylor (Appleton, Wis.) Karl D. Ustti (Princeton) Hans Arens (Bad Hcrsfeld) Herbert E. Brcklc (Uegcnsburg) G. L, Bursill-Hall (Vancouver, B. C.) Rudolf Engler (Bern) Tetsuro Hayashi (Fukuoka) Norman Kretzmann (Ithaca, NT.) J. Peter Maher (Chicago) R, H . Robins (London) Luigi Romeo (Boulder, Colo.) Barbara M. H. Strang t (Newcastle) Zsigmond Telegdi (Budapest) C. H. M. Versteegh (Nijmegen) V. A. Zvegincev (Moscow) AIM AND SCOPE This periodical is intended to serve the ever growing scholarly interest of lin- guists, psycholinguists, and philosophers of language of divergent persuasions in the history of linguistic thought. Central objectives of HL are the discussion of the epistemologtcal and method- ological foundations of a historiography of the discipline and the critical pre- sentation of particular areas or aspects of actual or potential research. HL is published in 3 issues per year of ca. 450 pages altogether. Each issue contains at least three major articles, one review article and/or a bibliography devoted to a particular topic, and many reviews, review notes and announce- ments of recent publications in the field. LC no. 73-88206 - ISS no. 0302 - 5160 © John Benjamins B, V. Publisher AmsteJdijk 44 - P O Box 52519 - The Netherlands 1007 HA AMSTERDAM Telephone: (020) 73 81 56 - Telex 15798 Wujastyk, Dominik: "Bloomfield and the Sanskrit Origin of the Terms 'Exocentric1 and 'Endocentric'". In: Historiographia Linguistica, Volume DC, nol/2 (1982). BLOOMFIELD AND THE SANSKRIT ORIGIN OF THE TERMS EXOCENTRIC* AND ENDOCENTRIC* DOMINIK WUJASTYK Brasenose College, Oxford As pointed out by Lyons (1968:231-32, 468), the terms 'exocentric1 and 'endocentric* were introduced into linguistic terminology by Leonard Bloom- field in Language (1933:194ff., 235-36). In fact, Bloomfield discusses these terms twice in his influential book. The first occasion is in his chapter on syntax (1933:194-97). There, the terms are presented as describing properties of open syntagmata, or phrases. John ran is exocentric because "the resultant phrase belongs to the form class of no immediate constituent"; poor John is endocentric because "the forms John and poor John have, on the whole, the same functions" (1933:194), Further- more, 'endocentric1 constructions are subdivided into "co-ordinative (or serial) and subordinative (or attributive)" (1933:195). Broadly, an endocentric con- struction is co*ordinative when "the resultant phrase belongs to the same form- class as two or more of the constituents", and subordinative when "the resul- tant phrase belongs to the same form-class as one of the constituents'* (1933: 195). Examples are boys and girls and poor John respectively. Finally, Bloom- field notes that these constructions may nest: "there can be several ranks of subordinative position" (1933:195). The explication of these terms is taken up a second time in the chapter on morphological types, in the discussion on compound constructions (1933: 233ff.)< Bloomfield introduces the Sanskrit classification of compounds into dvandva, tatpurusa, amredita1 and kamiadharaya. These he describes as endo- centric constructions. Next he mentions the bahuvrihi, the other main type of Sanskrit compound, and calls it exocentric (1933:235). After some explana- tion of these compound types, with examples, he briefly mentions two further types, the dvigu and the avyaylbhava (1933:237). Thus Bloomfield introduces and characterises the terms 'exocentric' and 180 MISCELLANEA 'endocentric* in two distinct contexts in Language: as types of phrase and as types of nominal compound. The description of the latter case is intimately bound up with his description of Sanskrit compound types. It would appear, then, that Bloomfield first worked out the idea of this distinction in dealing with phrases, then applied it to compound constructions and used the distinc- tion to highlight some features of Sanskrit compound classification. I hope in what follows to show that the opposite is true: Bloomfield found the distinc- tion being applied by the Sanskrit grammarians to their compounds, turned the device to his own use in classifying English compounds, and then extended the notion to apply to phrases too.2 it is a fact not requiring extensive justification that Bloomfield, whose uncle was a professor of Sanskrit,3 was himself very familiar with Sanskrit, in particular with the superb grammar of that language by Panini. His admira- tion for Panini was often expressed explicitly in Language (1933:11, 19, 63) and elsewhere (e.g., 1929:273-74). He called Panini's grammar "one of the greatest monuments of human intelligence" (1933:11), a remark which has often been repeated by western linguists who study Panini. Bloomfield wrote articles such as *'On Some Rules of Panini" (1927) and the review of Bruno Liebich's Konkordanz Panini-Candra (1929), which demonstrate his detailed knowledge of Panini's system, and of the history of traditional Sanskrit grammars in general. The first article is an excellent example of specialist investi- gation, while the second could be given to any student today as a useful introduction to the Indian grammatical tradition. Other writings of his, such as the famous "A Set of Postulates for the Science of Language" (1926) and "Menomini Morphophonemics" (1939) are depply influenced by Paninian methodology, as is plain to anyone familiar with Panini (e.g., Allen 1955:112). Bloomfield*s attraction to the scientific rigour of Panini's system has frequently been noted by those who knew him closely (e.g., Bloch 1949:90; Hall 1950: 121), and by those who know his work and Panini's system (e.g., Staal 1972: 264). Having established that the terms *exocentric* and 'endocentric* were part of Bloom field's apparatus for describing compounds, in particular Sanskrit compounds, and bearing in mind the influence which the Indian grammarians appear to have had on him, let us turn to the Paninian description of com- pounds, and its development by his earliest commentators.4 Panini divides compounds into four major groups: avyaylbhava, tatpurusa, bahuvrihi and dvandva. All other compound types belong to subsets of these groups. The purpose of the classification is to group together items which NOTES / NOTIZEN 181 will later be subject to the same syntactic, grammatical or morphophonemic operations. Therefore the groups are distinguished by a number of different criteria, some formal, some syntactic, some semantic and some enumerative. Thus, following Panini's sequence, first avyayibhavas are defined by enumera- tion, with semantic conditioning (P.2.1,5-21). Tatpurusas are defined by enumeration and by form, with syntactic as well as semantic conditioning (P.2.1.22-2.2.22). Bahuvrihis are defined in a very interesting way. First a rule declares that "what remains is a bahuvrlhr (P.2.2.23: s'eso bahuvrthih). A second rule then states that a bahuvrlhi is "two or more nouns [in composition] when the sense is that of another word" (P.2.2.24: anekam anyapadarthe). Further rules give special cases (P.2.2.25-28). The point of the first rule, P.2.2.23, is rule ordering. Suppose a compound were formed which satisfied the conditions for classifica- tion both.as an avyayibhava and as a bahuvrPiL By P.2.2.23 the compound would be classified as the former, because only those forms which "remain", that is remain after applying the rules for avyayibhavas and tatpurusas, are liable to be classified as bahuvrihis. The second rule, P.2.2.24, is the chief criterion for classifying bahuvrihis. In particular, the term anyapadartha which appears in this rule will be relevant to our discussion below. Finally, a dvandva is defined simply in a single rule by the purely semantic criterion that it should consist of "two or more nouns [in composition] when the sense 'and' is present" (P,2.2.29: carthe dvandvah). Although Panini obviously took great care in framing these definitions, even his earliest commentators have attempted to reduce the criteria to a set of simple, purely semantic conditions.5 Thus, on three occasions Patafijali (fl.140 BC) has said: In this connection, some compounds have as their head the sense of their first word, some have as their head the sense of their last word, some have as their head the sense of another word, some have as their head the sense of both words. The avyayibhava has the sense of the first word as its head. The tatpurusa has the sense of the last word as its head. The bahuvrlhi has the sense of another word as its head. The dvandva has the sense of both words as its head.6 In other places, when contrasting the bahuvrihi with the other types of compound, Patafljali lumps the others together as "those whose sense is one of their own words (svapadartha)" (Kielhorn 1880 1,404. See also pp.405-420), as opposed to the bahuvrihi, "whose sense is some other external word (anya* padartha)" (ibid.). J 182 MISCELLANEA Thus we see that Patafijaii, whose Mahabhasya is our earliest commentary on Panini, described compounds in terms of the locus of their head, and in particular drew a clear distinction between a compound whose head is a word outside the compound itself, and one whose head is a word within the com- pound. The terms used by him are anyapadartha and swpadartha, the first of which is actually introduced by Panini in P.2.2.24. Thus one may trace the fundamental distinction to Panini, although Patafijali generalized it into a full classificatory system for all compounds. Finally, it remains to be said that the English terms *exocentric* and 'endo- centric' are most apt translations of the Sanskrit anyapaddrthapradhdna and svapadarthapradhana. In view of Bloomfield's acquaintance with Panini and his followers it does not seem too much to see in the Sanskrit words and in Pataft- jali's compound theory the source of Bloomfield's own writing on the subject, and his resulting inceptive use of the terms 'exocentric' and 'endocentric'. This is endorsed when one sees that his subclassification of the endocentric construc- tion into co-ordinative and subordinative types corresponds exactly to Pataft- jalfs description of the dvandva and the tatpurusa, and that furthermore Bloom- field himself equates the terms 'subordinative' and tatpurusa (1933:235). If this thesis is correct, then Bloomfield took the description of the Sanskrit grammarians, which they apply only the categorization of nominal compounds, and generalized it to describe in addition the syntax of certain phrase construc- tions. In more modern works, however, the terms are used exclusively to des- cribe phrase constructions, and their origin in and application to compound classification is forgotten, or at least not mentioned (e.g., Lyons 1963:23Iff; Crystal 1980:131,136-37).7 Author's address: Dominik Wujastyk 28 Dale Close Oxford, OX1 ITU England NOTES 1) Contrary to Bloomfield's statement (1933:235), the term 'amredita' is not the name of a compound type for the Hindu grammarians. Pinini defines the term precisely as the NOTES / NOTIZEN 183 second of two repeated words (P.8.1.2: tasya dviruktasya param amreditam), 2) Panini treats nominal compounds and phrase constructions strictly as optional syn- tactic transformations of one another (P.2.L11: vibhdp). This may have suggested to Bloomfield the application of the classification of compounds to their transformationally equivalent phrase forms. 3) Maurice Bloomfield (1855-1928), professor of Sanskrit and comparative philology at Johns Hopkins University, published distinguished works on Vedic and other topics. 4) What follows is a mere thumbnail sketch of Panini's treatment. For a slightly fuller description in English see Roodbergen's introduction (1974). 5) These, though easier to grasp and apply, are seriously inadequate in several cases and cannot properly replace Panini*? treatment (Whitney 1893:187-91; Cardona 1976: 213-14). 6) Kielhorn 1880 1,378-79, repeated p.382, 392: "iha kai cit samS^ah purvapadartha- pradhanah kai cid uttarapaddrthapradhanah kai cid anyapaddrthapradhdnah kai cid ub- hayapaddrthapradhdnah / purvapadarthaprudhdno *vyaytbhdvah / uttarapaddrthapradhdnas tatpuruqah f anyapaddrthapradhdno bahuvrihih j ubkayapaddrthapradhdno dvandwth /". 7) This process has come full circle with the publication of Coulson's Sanskrit (1976) in which the term *exocentric* is employed to help the beginner in Sanskrit Ln his first steps towards grasping the character of the bahuvrlfii (Coulson 1976:117). REFERENCES Allen, W. Sidney. 1955. "Zero and Panini". Indian Linguistics 16.106-113. Bloch, Bernard. 1949. "Leonard Bloomfield". Language 25.87-94. Bloomfield, Leonard. 1926. MA Set of Postulates for the Science of Language". Language 2.153-64. (Repr. in Bloomfield 1970,128-38.) -----. 1927. "On Some Rules of Panini". JAOS 47,61-70. (Repr. in Bloom- field 1970.157-65.) -----. 1929. Review of Bruno Liebich> Konkordanz Pdnini-Candra (Breslau: M. & H. Marcus, 1928). Language 5.267-76. (Repr. in Bloomfield 1970.219- 226.) -----. 1933. Language. New York: Holt, Rinehart & Winston. -----. 1939. "Menomini Morphophonemics". Travaux du Cercle Linguisti- quedePrague 8.105-15.(Repr. in Bloomfield 1970.351-62.) -----. 1970. A Leonard Bloomfield Anthology. Ed. by Charles F. Hockett. Bloomington & London: Indiana Univ, Press. Cardona, George. 1976. Panini: A survey of research. The Hague: Mouton. Coulson, Michael. 1916. Sanskrit London: HodderA Stoughton. Crystal, David. 1980. A First Dictionary of Linguistics and Phonetics. London: A. Deutscji. 184 MISCELLANEA Hall, Robert A., Jr. 1950.4*ln Memoriam Leonard Bloomfield (April 1,1877 - April 18, 1949)". Lingua 2.117-23. (Repr. in Bloomfield 1970.547-53.) Kielhorn, Franz (1840-1908), ed. 1880-85. The Vyakarana-mahabhasya of PatanjalL 3 vols. Bombay: Government Central Book Depot. (References to third edition, by Kasinath V. Abhyankar, Poona: Bhandarkar Oriental Re- search Institute, 1962-72.) Lyons, John. 1968. Introduction to Theoretical Linguistics. Cambridge: Cam- bridge Univ. Press. (References to 1977 reprint.) Panini. Apadhydyl See 3ns'a Chandra Vasu 1891. Pataftjali. Mahabhdsya. See Kielhorn 1880-85. Roodbergen, Jouthe A. F. 1974. PatanjalVs Vydkarana-Mahdbhdsya, Bahuvrl- dvandvahnika (P. 2.2.23-2.2.3 8): Text, translation and notes. Poona: Univ. of Poona. Staal, Johan Frederik. 1972. A Reader on the Sanskrit Grammarians. Cambridge, Mass.: MIT Press. Vasu, Srisa Chandra (1861-1918?). 1891. The AstddhyayT of Panini Edited and Translated into English. Allahabad: The Panini Office. (References to second reprint, Delhi: Motilal Banarsidass, 1977.) Whitney, William Dwight (1827-94). 1893. "On Recent Studies in Hindu Gram- mar'* . A merican Journal of Philology 14.171-97. MANUSCRIPTS Manuscripts submitted for publication in Historiographia Linguist tea should be in English, French, or German, occasionally, contributions in other languages using Latin script will be considered as well. All correspondence concerning editorial matters - inquiries regarding suitable topics, reviews, formal presentation, etc. - as well as manuscripts and hooks tor review should be sent directly to: E. F. Konrad Koerner, Editor Historiographia Unguistica Department of Linguistics University of Ottawa OTTAWA, Ont. CANADA-KIN 6N5 or Hans Josef Niederehe. Assoc Id //istorii ygraph ia I. ingu is tu *a FB II: Romanistik Universttat Trier D-5500 TRIER W. GERMANY SUBSCRIPTION The prices for subscriptions to llistoriogrciphid Linguistica for one volume O issues of about 450 pages altogether) arc: □ Hfl. 140,- for institutions, □ A reduced rate of Hi]. 80,- postage and handling included, is offered to individual subscribers, provided that their prepaid order is placed directly with the publisher. FOR SUBSCRIPTION, and other business information, write to: John Benjamins B.V. Publisher Amsteldijk 44 - P 0 Box 52519 - The Netherlands 1 1007 HA AMSTERDAM Telephone: (020) 73 81 56 Telex 15798