1 The Three Recensions of the Versio Dionysii §1. The state of the question: The Versio Dionysii (that is, John Scottus Eriugena’s translation of the four treatises and 10 letters of Dionysius the pseudo-Areopagite) was lasted edited by Joseph Heinrich Floss in 1853, and published in the Patrologia Latina, volume 122. Within the last 150 years, Floss has been severely criticized for failing to recognize the manipulations to which the Versio was subjected during the Middle Ages1. The textual tradition of the Versio is, in the words of H. Dondaine, extremely ‘plastic’2. In 1886, L. Traube attributed this plasticity to the interfering hand of Anastasius Bibliothecarius3. Eriugena began his translation of the Dionysian corpus some time between 860 and 8624. Shortly after this, Anastasius Bibliothecarius translated the Dionysian scholia of John of Scythopolis and Maximus Confessor, and appended them to a copy of the Versio, which he sent to Charles the Bald, Eriugena’s patron, in 875. In the letter to Charles which accompanied the Versio Anastasii5, the papal librarian noted that, where he found descrepancies between the translation of Eriugena and the lemmata of the scholia, he noted them, adding a very few notes of his own6. Traube, believing that Hincmar of Reims had used the Versio of Eriugena for his citations of the Latin Dionysius in his Liber de praedestinatione, and realizing that these citations in Hincmar correspond to no extant MS of the Versio, concluded that Hincmar must have known the Versio before Anastasius had added his corrections. For this reason, Traube despaired of ever finding the ‘pure’ version of Eriugena. 1  2 However, as G. Théry later discovered, Traube’s point of departure—the citations of Dionysius in Hincmar’s treatise on predestination—was faulty. Since Traube published his notes on the manuscripts of the Versio, Théry has proven that the citations in Hincmar’s Liber de praedestinatione come from Hilduin’s translation rather than that of Eriugena7. Despite the defects of Traube’s ‘Anastasius theory’, Théry also believed he had discovered the intruding hand of Anastasius in certain Viennese MSS. Théry found that Vienna Nationalbibliothek 754 and 971 witnessed an interlinear commentary, the goal of which was the latinization of Greek terms which Eriugena either preserved or transliterated8. Following this ‘discovery’, Théry suggested that the work of the court philosopher could be disentangled from that of the papal librarian by comparing the MSS of the Versio to the Greek MS of the Dionysian corpus which we know Eriugena to have used for his translation—Paris BN, gr. 437 (=Pa); or more simply, to divide the MSS of the Versio into those which contain the scholia of John of Scythopolis and Maximus Confessor, and those which do not. The former should witness the contamination of Anastasius, while the latter should present the ‘pure’ work of Eriugena. Théry’s work, like Traube’s, suffers from certain shortcomings. First, his alleged Anastasian glosses are, in fact, extracts from Eriugena’s commentary on the first Dionysian hierarchy, the Expositiones in hierarchiam caelestem9. Second, Théry seems to have overlooked the fact that his two Viennese MSS, which, according to Théry, witness the glosses of Anastasius, are conspicuously lacking the scholia of John and Maximus. 2  3 Despite these shortcomings, in 1956, H. Dondaine pursued Théry’s suggestion— that the MSS of the Versio ought be divided into those which contain the scholia and those which do not, and that both groups ought be compared against Pa—and, indeed, discovered two separate recensions of Eriugena’s translation. The recension A, witnessed by the Paris MSS BN 1618 (XI) and 2612 (XII), has been corrected against a Greek MS other than Pa—the corrector has supplied many of the lacunae which characterize Eriugena’s Greek source10. The recension T, witnessed by the French MS, Troyes, Bibliothèque municipale 802 (IX/X) and the English MS, Oxford, St. John’s College 128 (IX/X), is free of such contamination. Dondaine naturally concluded that T should represent the ‘pure’ work of Eriugena11. However, further examination of these two recensions against one another revealed that the variants which separate the text of T from that of A represent an attempt at correcting the translation of Eriugena—again, the transliterations and transcriptions of Greek terms in A have been replaced in T by their Latin equivalents12 and Eriugena’s more complex vocabulary, often the result of his extreme literalism, has been simplified13. Finally, Dondaine realized that most of T’s corrections were suggested by Eriugena himself in his Expositiones14. Thus, Dondaine suggested that Eriugena himself was responsible for the corrections of T15, proposing the following timeline: “recension A—Expositiones—recension T”16. Although it far surpasses the work of his predecessors, Dondaine’s work suffers from its own shortcomings. First, his chronology fails to reflect the dynamic relationship between the recensions A and T. A is characterized not only by its contamination against a 3  4 Greek MS which does not share the lacunae of Pa, but also by an extensive interlinear apparatus, which reports, amongst other things, the variants of T. And while the ‘pure’ text of A—that is, the translation minus contamination and apparatus—does precede the text of T chronoligically, this ‘pure’ text of A exists in no MS of the Versio. The matter is complicated by the fact that, in certain instances, the variant of T has replaced the original reading of A. In most of these cases, the apparatus of A preserves the original reading. Nevertheless, in a very few cases, the original reading of A has been omitted altogether. It would be altogether impossible to discover this contamination of A by T were it not for the existence of a third recension of the Versio, which leads us to the second major shortcoming of Dondaine’s research—his meagre selection of MSS §2. A third recension of the Versio Dionysii: Of the more than 30 MSS which witness the Versio Dionysii, I have found 617, the significant variants of which define a third recension of Eriugena’s translation, M, quite distinct from both A and T. Variants separating M from A and T (De caelesti hierarchia II): PL 122 1039C7-7 aestimamus AT, aestimemus M 1039C12 oportet AT, oporteat M 1040A1 imaginamus AT, imaginemus M 1040A7 poeticis AT, facticiis M 1040B6-7 immaterialibus quoquomodo AT, materialibus quomodo M 1040B12 dissimilitudines AT, dissimulationes M 1040B13 simul iniuste AT, illegitime M 1040B15 immundas AT, immundanas M 1040C6 dum describuntur AT, describendo M 1040C11 sic AT, si M 1040D2 in AT, om. M 1040D4 possibiles AT, passibiles M 4  5 1041A3 diuina AT, et diuina M 1041A7 deformes AT, deiformes M 1041A7-8 descriptionis causas aetimauerit quis inhonestum AT, descriptione causas aestimauerit quis dehonestari M 1041B1 et colendam AT, colendam M 1041B1 diuinitatis AT, QEARXIAE M 1041B4-5 rationalitatem AT, rationabilem M 1041B9-10 castioribus manentibus AT, gloriosioribus quidem existentibus M 1041B10-11 quoquomodo AT, quomodo M 1041B11 diuina AT, diuini M 1041C3 et incomprehensam AT, incomprehensibilemque M 1041C4 quae ex quibus AT, ea ex quibus M 1041C7 hoc quidem (non esse) AT, om. M 1041C10 negationes AT, depulsiones M 1041C11 affirmationes AT, intentiones M 1042A2-3 per quidem enim praeciosiores sacras formationes AT, in quidem enim praeciosioribus sacris formationibus M 1042A6-7 candidum et igneum innocue respergentes AT, candide et ignee innocueque resplendentes M 1042A13 remanens quiescere AT, manens requiescere M 1042A14 sursum ferens animae AT, sursum uersus animam M 1042B8 materialibusque formare AT, materiis reformare M 1042B11 in irrationabilibus AT, irrationabilibus M 1042B12 omnis irrationabilitatis AT, omni irrationabilitate M 1042B13 in intellectualibus AT, intellectualibus M 1042C10 circumformamus AT, circumformemus M 1042D3 sufficientia et inconuersabilitate et a nulla affligitur uirtute AT, sufficienti inconuersibili et a nullo affligi ualente M 1043A1 immutabilem AT, incommutabilem M 1043A8 supereminentias AT, supereminens M 1043A9-11 nostram transitoriam et corporalem rationem et materialem et alienatum incorporalibus animis sensum excellentes AT, nostra transitoria et corporali ratione et materiali alienato incorporalibus animis sensu M 1043A11 dissonas AT, et dissonas M 1043A11-12 formare AT, reformare M 1043B11-12 incircumuolute AT, incircumuelate M 1043B12 inuisibiliter AT, intellectualiter M 1043C8 diuina AT, diuini M 1043C12 recte accepta AT, tractabilia M 1043C13-14 contemplationis AT, contemplatores M 1044A3 utique AT, aeque M 1044A10 extendere AT, extendi M 1044B1 aestimamus, quidque ab ipsa hierarchia prosunt hierarchiam sortientes AT, aestimemus, quidue ipsa hierarchia prosit 5  6 hierarchiam sortientibus M 1044B3 hierarchiae AT, hierarchicae M 1044B7 diuina AT, in deo et M M is genarally unaware of the corrections of T (De caelesti hierarchia II): PL 122 1040B5 reformare AM, formare T 1041B8 reformant AM, formant T 1041B9 reformationibus AM, formationibus T 1041C13 reformationes AM, formationes T 1042A3 et AM, om. T 1042A11 et AM, om. T 1042C1 immitem AM, immanem T 1043A15 resonantias AM, imagines T 1043B8 thearchicis AM, diuinis T 1043C13 agalmatum AM, imaginum T 1044A2 reformant AM, formant T Oddly, M does agree with T against A in certain, fewer instances (De caelesti hierarchia II): PL 122 1041C15 eos TM, eas A 1042A5 decora TM, decore A 1042A9 uisibilibus bonis TM, uisibilium bonorum A 1042B14 irascibile TM, irascibilem A 1043B15 in TM, om. A 1043C10 prophetae TM, ypophetae A 1043C15 negationibus TM, depulsionibus A Nevertheless, there are separative errors which distinguish all three recensions from one another (De caelesti hierarchia II): PL 122 1042D2 uelut impotentiam ueluti potentiam T potentiam M A 1043D1 resultationum A resonantiarum T imaginationum M 1044A4 uero A om. T quoque M 1044A13-14 angeliformes A angelicas T angelicarum M 6  7 The three recensions are, therefore, independent of one another, siblings with a common parent (something similar to A prior to the contamination against its Greek source and the addition of its apparatus). Further, the majority of variants in M cannot be attributed to scribal error. M has been thoroughly corrected, independently of T. But if Eriugena was responsible for the corrections of T, who was responsible for correcting M, and is this recension valuable for the textual criticism of the Versio Dionysii? The answer, once again, comes from Eriugena’s Expositiones. For, while the corrections of T are reflected in the commentary, those of M make up the actual lemmata of the text commented. And that the text presented by the lemmata is, in fact, the version Eriugena had before his eyes as he composed his commentary is confirmed by the commentary itself. For example, at 1042A14, Eriugena has translated the Greek dianistw~sa to\ a0nwfere\j th~j yuxh~j as purgans sursum ferens animae according to A and T. M, on the other hand, gives the less literal, slighly simpler, sursum uersus animam. His comment on this passage in the Expositiones shows that Eriugena was reading M, rather than A or T: ‘Pro eo quod transtulimus sursum uersus animam, in graeco scribitur ANWFERES THS YUXHS…’18 The same occurs at 1042C13. The Greek texts reads: w9j mh/te ta\ qei~a toi~j bebh/loij eu0xei/rwta ei]nai…. Both A and T give the overly literal recte accepta for eu0xei/rwta. M, however, gives tractibilia, and Eriugena confirms this reading in the same manner as above: ‘Et notandum quod in loco quo transtulimus tractabilia, in graeco scriptum est EUXEIRWTA…’19. We must conclude, then, that the corrections of M have a greater claim to authenticity than do those of T. 7  8 The recension M, therefore, has a double significance for the textual criticism of the Versio Dionysii. First, it represents the final version of his translation as Eriugena, himself, intended it. Second, because M is unaware of the changes made in T, it preserves the original readings of A where A has been contaminated by T, with no evidence of the intrusion in the apparatus of A. De caelesti hierarchia I A/T M 8.5-6 Kai\ ga\r ou0de\ au0th\ 1038C6-7 Etenim neque Etenim neque ipse usquam pw/pote th~j oi0kei/aj ipse usquam unquam radius unquam propria singulari e9nikh~j e9no/thtoj propria singulari unitate unitate deseritur… a0polei/petai… deseritur… De caelesti hierarchia II A/T M 14.1 a0napla/sai 1042B8 formare reformare 14.3 o9 qumo\j eoi~j me\n 1042B11 furor in furor irrationabilibus a0lo/goij irrationabilibus quidem quidem §3. W and the stemmatization of the three recensions: The MS evidence points to three recensions, the relationship of which is rather fluid. On the one hand, A appears to be the archetype of T, which latter has attempted to remove the literalism of its parent, using the suggestions supplied by the Expositiones. A Expositiones T On the other hand, the recension A is characterized by the variants of T, reported above the lines of the text. Where the variant of T and the original reading of A have been switched, or where the original reading of A has been omitted altogether, A is contaminated. From this point of view, A and T point to a hypothetical archetype, roughly 8  9 the equivalent of A, minus the Greek contamination and the contamination against T; or T, minus its corrections over A. W Greek MS (R) Expositiones A T M, like T, has a double relationship with A. First, it represents a correction of A— perhaps with the Expositiones as mediator. A Expositiones M Second, it witnesses W as a sibling of A, where the latter is contaminated by T. W T M A The question remains: which version should be edited? W can be reconstructed from A and M. But an edition which presents only the archetype would be omitting critical information: the corrections of Eriugena, himself, witnessed in M. An edition of M alone would fail to represent the work of Eriugena in its entirety: after all, the process of correction is as much a part of the translating process as is the final product. Further, 9  10 both of these options would misrepresent the state of the Latin Dionysius in the Middle Ages. For, while the recension T has less claim to authenticity than do both A and M, T was widely read up to, and through the 13th century. For, the Corpus Dionysiacum—the 13th century manual of Dionysius which contained the scholia of John and Maximus, and the commentaries of Eriugena, Hugh of St. Victor and John Sarrazin surrounding the text of De caelesti hierarchia—was based on T20 rather than M. Thus, it was T—or a close offspring—which was read by Albert the Great, Ulrich of Strassbourg and Thomas Aquinas. I suggest, then, that the critical edition of the Versio Dionysii should present three separate editions of three distinct versions: the first version presents W, reconstructed from A and M; the second version is that witnessed by M, indicating the corrections of Eriugena; and the third version is that represented by T, the version as it was later known to the 13th century schoolmen. I envision an edition similar to Édouard Jeauneau’s edition of the Periphyseon, which is likewise a text in transition throughout the Middle Ages. This edition would have two advantages. First, it would present the work of Eriugena as translator in two phases—his initial attempt at rendering the Areopagite, and his corrected version. This aspect of the edition will be invaluable for our understanding of Eriugena’s knowledge of Greek, his competence and development as a translator, and, generally, the Hellenism of the 9th century. Second, such an edition would represent a valuable tool for the study of the Latin Dionysius throughout the Middle Ages, and a reliable resource for disentangling the sources at the heart of the commentaries of Hugh and Albert as well as 10  11 the translations of Sarrazin and Grosseteste—both of whom used the Versio in composing their own translations. §4. The manuscripts21: A. The recension A Avranches, Bibliothèque Municipale, 47, XII (=Av) Berlin, Staatsbibliothek, Phillipps 1668, IX-X (=Bp) Bern, Burgerbibliothek, 19, IX-X (=Be) Paris, Bibliothèque Nationale, lat. 1618, XI (=A) Paris, Bibliothèque Nationale, lat. 2612, XII (=B) B. The recension T Basel, Universitätsbibliothek, O.III.5, XII (Ba1) Boulogne-sur-mer, Bibliothèque Municipale, 27, XII (=Bo) Cambridge, Trinity College, B.2.31 XII (=Ct) Città del Vaticano, Biblioteca Vaticana, Reg. lat. 67, XII (=Vr) Darmstadt, Hessische Landes- und Hochschulbibliothek, 30, XII (=E) Firenze, Biblioteca Medicea Laurenziana, Plut. 89 sup. 15, XI (=Fl) Heiligenkreuz, Bibliothek des Zisterzienserstift, 111, XII (=H) Lilienfeld, Stiftsbibliothek, 128, XIII (=Li) Lyon, Bibliothèque Municipale, 598, XII (=Ly) Montecassino, Biblioteca dell’Abbazia, 221, XI (=Mc) Oxford, Bodleian Library, Laud. Misc. 639, XII (=Ox) Oxford, St. John’s College, 128, X (=Os) Tours, Bibliothèque Municipale, 338, XIII (=To) Troyes, Bibliothèque Municipale, 802, IX-X (=T) C. The recension M Bamberg, Staatsbibliothek, Patr. 66, X (=Ba) München, Bayerische Staatsbibliothek, lat. 14137, XI (=Clm1) München, Bayerische Staatsbibliothek, lat. 17072, XII (=Clm2) Wien, Österreichische Nationalbibliothek, 754, XII-XIII (=Va1) Wien, Österreichische Nationalbibliothek, 971, XII-XIII (=Va2) Zwettl, Bibliothek des Zisterzienserstift, 236, XII (=Z) §5. The contents of A: 11  12 Apart from their textual variants, the MSS of this recension are distinguished by several features. The A MSS witness the Corpus Anastasii, that is, they report the scholia of John of Scythopolis and Maximus Confessor, and all five MSS contain the interlinear apparatus recording the variants of T, with the exception of the Avranches MS, which, from chapters 4 to 14 of the De caelesti hierarchia, abandons the interlinear glosses22. The apparatus of A has several sources beyond the corrections of T. Dondaine divided the apparatus into four types of glosses: 1) explanations of words, introduced by id est, 2) textual variants introduced by aliter or uel, 3) corrections usually introduced by pro or graecus and which introduce a reading from R, and 4) certain longer, doctrinal glosses23. The first two generally report the readings of T, but certain may be found rather in the Expositiones. Of the third, Dondaine proposed that those introduced by pro represent a correction of A against a Greek MS of the corpus other than Pa, while those introduced by gr. or graecus are less intelligent, and, of 50, only about 15 reveal the presence of another Greek MS24. However, a comparison of these glosses with the scholia of John and Maximus reveals the glossator’s source: not a Greek MS at all, but Anastasius’ translation of the scholia. For example, in A, 49r , where the glossator has suggested dicemus in place of diximus, the scholion says: Aliud commentum se pollicetur fore scripturum symbolicam uidelicet theologicam25. And on fol. 74v, where the glossator replaces diuinitus with diuine siue deo decenter, the scholion reads: Interdicens in deo ea quae de corporalibus motibus dicta sunt intelligi quonam modo decenter deo oporteat intelligi motum dei edisserit…26 12  13 What remain are the doctrinal glosses, which are relatively rare. These glosses are particularly dense in chapter four of the De diuinis nominibus, the famous treatise on evil. Generally, they represent points of clarification, whether philosophical or grammatical, guiding the reader through especially difficult passages. The glosses are merely the notes of an interested reader: A, 63r: Et enim si natura mali neque ex bono neque in existentibus [id est in his quae sunt]… A, 63r: neque ex bonis conuersum natura et semper mali existunt [scilicet ad ista inferiora et ea diligeret siquidem boni angeli ad bonum conuersi qui est forma illorum]… A, 63r: Si autem aliis [scilicet mali] quomodo corrumpunt A, 63r: Natura enim incorruptibilia [id est quae sunt ultra lunam] non corrumpunt [scilicet daemones] sed acceptiua corruptionis [id est quae sunt in hoc mundo], sequitur neque hoc omnino et omnino malum [scilicet est]. It is impossible to trace the provenance of these glosses. However, it is worth noting that this passage, so heavily glossed in A is quoted at length and glossed in the Periphyseon. Indeed, the A MSS contain an interesting gloss at the beginning of the De caelesti hierarchia, a mixture of etymologies and explanations of Greek words which come from the Periphyseon and the Expositiones. ANAGWGH. Sursum reductio uel contemplatio uel ascensio. IERWS. Templum. Iereus sacerdos. Ierarchia episcopatus uel summum sacerdotem. Duobus nominibus uocatur deus apud grecos quibus maxime iste intellectus multiplicationis uidelicet dei per omnia et recollectionis iterum omnium in deum insinuatur. KALWS et AGAQWS. Kalos enim dicitur, quia omnia ad se uocat ut unum in ipso sint. Agathos uero, quia ualde currit per omnia, dans eis essentiam, differentiam, proprietatem et uniuersaliter et singulariter. Agan enim apud grecos ualde significat. QEWS currens. KALWS. uero uoco. Hinc uerisimile datur quod et apud nos bonus dicatur, a uerbo greco quod est BWO, hoc est clamo. Bonus ergo dicitur deus, quia omnia ad se clamat27. 13  14 Sursum ductio for anagoge comes from the Expositiones II, 140-1: anagogicas, sursum scilicet animum ducentes. The etymology of qe/oj comes from Eriugena’s Periphyseon I, 442ff.: cum uero a uerbo QEW QEOS deducitur, currens recte intelligitur. Ipse enim in omnia currit et nullo modo stat sed omnia currendo implet, sicut scriptum est: “Velociter currit sermo eius”. The etymologies of bo/w and ka/lw also come from the Periphyseon II, 1737ff.: Nam et hoc nomen, quod est bonitas, non aliunde originem ducit nisi a uerbo graeco quod est BOW (hoc est clamo). BOW autem et KALW (id est clamo et uoco) unum sensum possident. Etenim qui uocat saepissime in clamorem erumpit. Deus ergo non inconuenienter bonus dicitur et bonitas, quia omnia de nihilo in essentiam uenire iltelligibili clamore clamat; ideoque graece dicitur deus KALOS (hoc est bonus), DIA TO PANTA KALEI EIS OUSIAN (hoc est: eo quod omnia uocat in essentiam). The greek words for temple and priest are not expressly given in either the Expositiones or the Periphyseon, but may be found in the famous 9th century greek glossary compiled by Martin of Laon in the manuscript Laon 44428. Thus, the compiler of A had at his disposal, not only a MS each of A (prior to the apparatus and contamination by T) and T, but also the Periphyseon, the Expositiones, the scholia and, perhaps, Martin of Laon’s Greek wordlist. §6. The contents of T. The T MSS are not always as consistent in their content as those of A. Most witness the Corpus Anastasii (the scholia of John and Maximus and the glosses of Anatasius). Nevertheless, some, notably, T, Ct and Vr, are conspicuously lacking the scholia. One could suspect that these MSS represent an early stage in the life of T, 14  15 perhaps prior to the work of Anastasius. And certain variants and marginalia support this view. For example, at 1041C5, Eriugena translates the Greek comparative, kuriw&tero&n, potentius. This reading is shared by A, T and M. However, many of the T MSS report the variant magis proprium—sometimes above the line, sometimes in the text itself (magis proprium potentius in Dt and Os, magis proprium et potentius in H). Nevertheless, Ct, T and Vr are free of this gloss—it is found neither in the text, nor above the line. Further, several of the A MSS report magis proprium in the apparatus. It would seem, then, that magis proprium was a later addition, perhaps initially above the line, and eventually added to the text itself without removing the word, potentius, it was meant to replace. The absence of this phenomenon in Ct, T and Vr would seem to give these MSS priority over the other MSS of T, perhaps confirming Théry’s suspicion that the MSS of the Versio may be divided into those which contain the scholia and those which do not. The presence of the variant in the apparatus of A would seem to indicate that the compiler of A must have used a later MSS from the T family. Several of the MSS of the T recension are also contaminated as against M. This contamination is obvious in Fl, H and Li. In the case of Ox, the contamination is more complex. Ox shows all the signs of belonging to the family T. However, there are certain odd textual variants which Ox shares only with the lemmata of the Expositiones. For example, at 1040C, Ox is missing the words in animorum caelestium formationibus. This lacuna is found only in the lemma of the Expositiones (II, 300). Again, at 1040C, Ox inserts the inquit of the lemma of this very passage in the Expositiones (II, 327). Thus, the 15  16 contamination of M in Ox comes not from a separate MS of the M recension, but from a MS of the Expositiones. Finally, more than half of the MSS of this family contain a curious piece immediately following the Versio—a mixture of several extracts from Eusebius’ Historia Ecclesiastica. There are nine extracts in all: the first from Polycrates’ letter to Victor, the bishop of Megala (preserved only in Eusebius=HE 3.31.3), the second from Clement of Alexandria’s Quis diues saluetur (42.1-42.15=HE, 3.23.6-3.23.19), and the third through ninth from Philo Judeaus’ De uita contemplatiua (21-23, 26-27, 28-29, 29-30, 34-35, 68 (paraphrase) and 78=HE, III, xvii, 7-8, 9, 10-11, 13, 16-17, 19, and 20). The first and second discuss the fate of John the Evangelist after his exile to Patmos and seem to be meant as an explanation of Dionysius’ tenth letter, addressed to the Evangelist, wherein he predicts the return of John from Patmos. The extracts from Philo discuss the use of the term therapeutes as an appelation for monks, a term used by Dionysius in addressing his first letter to Gaius the monk. The Philonian pieces are followed by an explanation of Philo’s discusion of the therapeutes: Meminit horum et Iosebius Pamphili. Quidam autem dicunt haec Philonem de sociis iudeis dicere. Alii de Nazareis iudeis, alii ex circumcisione fidelibus et credentibus in christum et custodientibus legem Moysi, alii de perfectis christianis. Talis autem erant heresis monachicam uiuentes uitam therapeyte merito nominati sunt. Non solum autem iosebius pamphili sed et philo iudeus sed et beatus Dionisius Ariopagita discipulus sancti pauli apostoli factus athenarum episcopus in eo qui est de ecclesiastica ierarchia monachos ait, antique et therapeutas nominat. Beati autem Dionisii Ariopagite meminit Lucas euangelista in actibus apostolorum et Dionisius episcopus corinthi uir antiquus et beatus Policarpus in epistola ad ecclesiam athenarum et Iosebius Pamphili in Ecclesiastica Historia29. 16  17 The presence of these extracts in these manuscripts of the Versio, together with their relationship to the Dionysian corpus are explained by the scholia of John of Scythopolis. In his scholion on the first letter, John of Scythopolis considers the identity of this mysterious monk, and explains why Dionysius refers to Gaius as therapeutes: Tou/touj de\ e1oike qauma/zein Fi/lwn o9 0Ioudai~oj e0n tw~| Peri\ qewrhtikou~ h2 9Iketw~n sunta/gmati, qerapeuta\j au0tou\j kalw~n, kai\ th\n di/aitan au0tw~n dihgou/menoj pro\ o9li/gou tou~ te/louj tou~ bibli/ou tou/tou. 0Ana/gnwqi ta\ tou~ au0tou~ Fi/lwnoj.30 The passages from Philo seem to be a continuation of John’s scholion, pursuing his directive to read what Philo has to say. In his scholion on the tenth letter, John of Scythopolis, using the date of the reign of Tiberius, the solar eclipse which occurred that year and which Dionysius claims to have witnessed, and assuming that the convert of Saint Paul would have been around 20 at the crucifixion of Christ, notes that Dionysius would have been around 90 when he wrote this letter. All these historical facts, which allow him to suggest such an age, are recounted by Clement of Alexandria in his treatise: Quis diues saluetur? Me/mnhtai de\… kai\ Klh/mhj o9 0Alecandreu\j, e0n tw~| lo/gw~| tw~| peri\: Ti/j o9 swzo/menoj plou/sioj;31 It seems, then, that the Eusebian extracts are simply glosses on the scholia of John of Scythopolis. And indeed these extracts are found in several of the more ancient greek manuscripts of the Corpus Areopagiticum. The question remains: how did they enter the MSS of Versio? The evidence points in different directions. First, Eriugena seems to have known the extracts. His dedicatory letter to Charles the Bald betrays an acquaintance with the section which concludes the extracts from Philo: 17  18 Fertur namque praefatus Dionysius fuisse discipulus atque adjutor Pauli apostoli, a quo Atheniensium constitutus est episcopus, cujus Lucas commemorat in Actibus apostolorum, et Dionysius, episcopus Corinthi, vir antiquus, beatus quoque Polycarpus in epistola ad ecclesiam Athenarum, Eusebius item Pamphili in Ecclesiastica Historia.32 Second, although Pa does not contain the extracts, its last folios are missing. Perhaps it would have been strange for Pa to contain these pieces so closely connected with the scholia and not the scholia themselves. Nevertheless, it is by no means impossible. For the greek manuscript that Robert Grosseteste had copied for the purposes of translating the Dionysian corpus, Oxford, Bodleian, Canonici, gr. 97, contains the Eusebian extracts, while its prototype, Paris Bibliothèque nationale, gr. 933, does not. Grosseteste collated his manuscript against Pa. Is it not possible that Pa is the source of these fragments in Bodl. Can. gr. 97?33 Third, some of the peculiar vocabulary of the Latin translation of the Eusebian extracts matches that of Eriugena in the Versio: eloquia for logi/a, and inuisibilibus for nohma&twn. Nevertheless, there remain some translations which are extremely difficult to understand if they were made by the translator of the Dionysian corpus: thus, the Latin extracts give formare in place of metasxei~n, a crucial word in the Dionysian system, and one which Eriugena generally translates participare. Further, the absence of these extracts in both A and M is suspicious. Were Eriugena their translator, one would expect to find them in the MSS which are closer to his original work. 18  19 If Eriugena is not the translator of these extracts, perhaps Anastasius was responsible for translating them when he translated the scholia. If he had a Greek MS of the Corpus Areopagiticum, it is not at all unlikely that they contained the extracts. Further, Anastasius’ letter to Charles the Bald gives a litany of historical persons who substantiate the apostolicity of the Areopagite which is rather similar to that of Eriugena (quoted above) and which may come from the last extract from Philo. Et quidem Lucas in Actibus apostolorum, Eusebiusque Pamphili, et Dionysius Corinthi antistes, de eo quaedam miranda tradunt: quae vero scripserit nisi me oblivio fallat, omnino non tradunt34. Nevertheless, if the MS T, which contains the extracts, represents the recension T prior to the work of Anastasius, then the Latin extracts themselves must likewise antedate the work of the papal librarian. All that can be said for certain, at this point, is that whoever was responsible for the extracts was also, most likely, responsible for the corrections of T. §7. The contents of M: All of the MSS of this family lack the scholia. This recension is clearly separate from the Corpus Anastasii, and therefore, lacks the letter of Anastasius to Charles the Bald (Inter caetera, 1025-1030), which precedes the Versio in almost every other MS Further, Clm1, Va1 and Va2 contain the verse Nobilibus quondam, once attributed to Eriugena35. These are the only two external features of the M manuscripts which distinguish them from those of the other recensions. §8. Conclusion: 19  20 As we can see, the Expositiones in caelestem hierarchiam is at the core of the textual tradition of the Versio Dionysii. Eriugena’s commentary on the first hierarchy was both the source of the corrections of the recension T and the impetus behind the corrections of M (whether it represents the philosophic reflection which pushed Eriugena to correct his translation or a justification of the corrections once already made). As a result, the distinction between the recensions drops off considerably in the other three treatises. The stemmatics hold up in the presence of the rest of the corpus, but the source of corrections, contamination, etc., is less clear. The edition of the Versio, therefore, ought be presented in two separate works, first, the edition of the De caelesti hierarchia, and afterwards, the edition of the De eccelsiastica hierarchia, the De diuinis nominibus, the De mystica theologia and, finally, the letters, in as many volumes as is necessary. For my part, I am currently preparing the edition of Eriugena’s translation of the De caelesti hierarchia. Timothy R. Budde Centre for Medieval Studies University of Toronto 1 Cf. Hyacinthe Dondaine, Le Corpus Dionysien de l’Université de Paris au XIIIe siècle (Rome, 1953), p. 37 and Phillipe Chevallier, Dionysiaca vol. 1 (Paris, 1937-49), p. CVII, n. 4. 2 Dondaine, Le corpus Dionysien, p. 37. 3 Cf. MGH, Poetae Latini III, 525-6. 4 The Versio was begun after Hincmar of Reims composed his Liber de praedestinatione in 860, and before Eriugena began his Periphyseon around 864. Between the Versio Dionysii and the Periphyseon are situated his Versio Maximi and De imagine. Allowing another two years to complete these further translations, the Versio Dionysii cannot have been completed long after 862. 5 PL 122, 1026-9; Anastasius’ letter is dated 23 March 875. 6 Cf. PL 122, 1028: Sane ubi a verbis interpretis scholia ipsa dissentire vidi, ut lector, quid de apposita dictione interpres senserit, quid scholion insinuet, indifficulter agnoscat, et verba interpretis scholio inserui, et, qualiter ea scholii compositor praetulerit, innui. Sed et, sicubi opportunum fore conspexi, ex me quoque, quoniam esse aliter non potuit, paucissima quaedam, et quae facilius ab intelligente agnosci poterant, interposui. 20  21 7 Cf. Gabriel Théry, Études Dionysiennes vol. 1. Hilduin traducteur de Denys (Paris, 1932), ch. VIII, and Gabriel Théry, “Catalogue des Manuscrits Dionysiens des Bibliothéques d’Autriche,” Archives d’histoire doctrinale et littéraire du Moyen Ages, 10/11 (1936), pp. 167-73. 8 Cf. Théry, “Catalogue des manuscrits,” p. 172. Théry gives the following examples from the De caelesti hierarchia: “… Scot Erigène Gloses d’Anastase MS 14137 (autogr. d’Othlon) Ch. I. symbolice id est significatiue fol. 5r anagogice id est contemplatiue fol. 5r ierarchias id est summa sacerdotia fol. 5r teleta/rxij teletarchis, id est princeps fol. 5r hostiarum qe/wsin theosin, id est deificationem fol. 5r Ch. II. archistrategos id est dux, principium fol. 6r exercitus; stratos exercitus; strategos, princeps theologia id est diuina disputatio fol. 6r qearxi/aj id est diuinitatis fol. 7r caracterizante id est imaginante fol. 7r theosophi id est diuini sapientes fol. 8v …” 9 For example, the gloss on archistretagos in Vienna Nationalbibliothek, 754, 6r (see the previous note), id est dux, principium exercitus; stratos exercitus; strategos, princeps, may be found at Expositiones II, 114-5: Archistrategos enim dicitur quasi ARXWN TOU STRATOU HGEMWN… 10 Le corpus dionysien…, p. 40. 11 : “Cette première constatation met en valeur la pureté du texte de T…,” Dondaine, Le corpus dionysien, p. 40. 12 Cf. Dondaine, Le corpus dionysien, p. 41; the MS A=Paris BN 1618 and the MS T=Troyes BM 802. “… PL 122: A T HC 4 (1046B3) fol. 9v teletarchicis fol. 160r perfectissimis B6 thearchia divinitas C12 thearchicam divinam HC 7 (1052 B12) fol. 12v teletarchia fol. 162r perfectionis principe (1053B11 13r a theodochis v ex Deum recipientibus HE 1 (1072B6) 23v qewsin fol. 172r deificationem HE 7 (1109 B5) 44v ypopheta 192v propheta (1109C6) ypophetice prophetice (1110C14) 45r teletarchica 193v perfectiva …” 21  22 13 Cf. Dondaine, Le corpus dionysien, pp. 42-4; the MS b=Pa. “… HC 4 (1046 B3): b en qeoprepei sebasmiothti (Dionysiaca, II, 8002). A in Dei predicta religiositate (f. 9v). T in divina religiositate (f. 160r). HC 4 (1047 A4): b pollakwj ekfantorikai (Dionysiaca, II, 8063). A innumerabiliter manifestatrices (f. 9v). T multipliciter manifestatrices (f. 160v). HC 4 (1047B7): b kata taj prepousaj qew… ekfantoriaj (Dionysiaca, II, 8094). A secundum predictas Deo… manifestationes (f. 10r). T secundum decentes Deum… manifestationes (f. 160v). HC 4 (1048B4): b ouk apophda thj up’ autou taxqeishj te kai ai9reqeishj A non resilit sub se et bellicali et libenti T non resilit a se ordinata et assumpta anqrwpoprepouj eutaciaj (Dionysiaca, II, 8094). humana bona ordinatione (f. 10v). humana ordinatione (f. 161v). …” 14 Cf. Dondaine, Le corpus dionysien, pp. 42-3, nn. 24-7. 15 “[I]l n’est plus guère douteux qu’il faille l’attribuer à Scot en personne”, Dondaine, Le corpus dionysien, p. 63. 16 Dondaine, Le corpus dionysien, p. 49. 17 Bamberg, Staatsbibliothek, Patr. 66, (X =Ba), Bayerische Staatsbibliothek, Clm. 14137 (XI, =Clm1) and Clm. 17072 (XII, =Clm2), Osterreichische Staatsbibliothek 754 (XII, =Va1) and 971 (XIII, =Va2), and Zwettl, Stiftsbibliothek 236 (XIII, =Z). 18 Exp. II, 668-9. 19 Exp. II, 1147-9. 20 The recension found in the 13th-century Corpus Dionysiacum is, in fact, an expansion of T, which Dondaine calls P; essentially T contaminated against another Greek MS of the Dionysian corpus; cf. Dondaine, Le corpus dionysien, pp. 77ff. 21 I.P. Sheldon-Williams’ list of mss. (“A Bibliography of the Works of Johannes Scottus Eriugena,” The Journal of Ecclesiastical History 10 (1959), pp. 216-7) is based on the work of Dondaine. Sheldon-Williams identifies “four principal stages in the development of Eriugena’s text: (a) the First Version; (b) the Second Version; (c) the Eleventh- Century Corpus ((a) + the additions and corrections of (b)); (d) the Thirteenth-Century Corpus, in which Eriugena is represented by (b) + the Expositiones + a considerable portion of the Periphyseon + the Dedicatory Epistle to Charles the Bald”, 216. How 22  23 Sheldon-Williams decided to divide the MSS is impossible to discern—his ‘Eleventh- Century Corpus’ is witnessed by 9th- and 10th-century MSS, and he also includes in his list at least one MS, Chartres BM, 131, which no longer exists, and another, Barcelona, Ripoll o, which, the archivist of the Arxiu de la Corona d’Aragó has assured me, never existed. In any case, Sheldon-Williams’ division is incorrect. He combines T mss. (e.g., Troyes 802) and M mss. (e.g., Munich 14137 and Vienna Natbibl. 754 and 971) in his second family, (b), and A, T and M mss. in his third (Berlin Phill. 1668, from A, Darmstadt 30, from T and Munich 17072, from M). The recent article by Ernesto Mainoldi (“Iohannes Scottus Eriugena” in La trasmissione dei testi latini del medioevo TE.TRA.2, (Firenze, 2005), pp. 186-264), based as it is on Sheldon-Williams’ article, suffers from the same problems. 22 Cf. Dondaine, Le corpus dionysien, p. 38, n. 10. 23 Cf. Dondaine, Le corpus dionysien, pp. 57ff. 24 Cf. Dondaine, Le corpus dionysien, pp. 58ff. 25 Cf. PG 4, 209C: Sumbolikh~? qeologi/a?. 1Allo su/ntagma u(pisxnei~tai gra/fein, o3per kalei~ Sumbolikh\n qeologi/an. 26 Cf. PG 4, 384A: )Alla\ kai\ kinh/seij qeou~. )Apagoreu/saj ta\ tw~n swmatikw~n kinh/sewn e)pi\ qeou~, o3pwj xrh\ qeoprepw~j noei~n ki/nhsin e)pi\ qeou~ safhni/zdi… 27 Paris BN 1618, 5r. 28 Fol.122r: ieron, sacrum templum, fol.122r: iereuj, sacerdos. 29 Oxford, St. John’s College 128, 219r. 30 PG4, 528A. 31 PG4, 573CD. 32 PL122, 1032A. 33 Cf. Ruth Barbour “A manuscript of ps.-Dionysius Areopagita copied for Robert Grosseteste,’ Bodleian Library Record 6 (1958), pp. 401-16. The presence of these extracts in Grosseteste’s manuscript led Barbour to suggest that Bodl. Can. gr. 97 was collated against yet another manuscript. She corroborated her claim by certain variants recorded in Bodl. Can. gr. 97 which she found neither in Pa nor in Paris BN gr. 933. However, she did not personally consult either of these manuscripts. For Pa, she used Chevallier’s Corpus Dionysiacum and for Paris BN gr. 933 relied upon a collegue at the BN. However, an examination of these two manuscripts shows that several of her unaccounted-for variants are, in fact, variants in the margins and lines of these two Paris manuscripts. For example, on fol. 117r, Bodl. Can. gr. 97 gives the reading a0nalo/goij together with the variant. Barbour notes that both Paris, BN gr. 933 and BN gr. 437 report a0nalo/goij. In fact, a0nalo/goij is corrected to a0nalo/gwj above the line on fol. 105v. On f. 182r, Bodl. Can. gr. 97 gives the reading proeilhfe/nai together with the variant proseilhfe/nai. According to Barbour, both Paris, BN gr. 933 and BN gr. 437 give the reading proeilhfe/nai. For the latter MS, Barbour cites Chevallier’s transcription (480.4). However, Chevallier has misread or mistranscribed the MS here, which, in fact, reads proseilhfe/nai on fol. 179r. 34 PL 122, 1028-9. 35 On this piece, cf. Théry, “Catalogue des manuscrits,” pp. 168-169, “D’après Traube, cette pièce aurait été composée par un napolitain peu après 878: «Id quod non pono, sed possum probare, probaboque alio loco, versus illi contra Romam scripti sunt paulo post a. 23  24 878 a grammatico aliquo Neapolitano» […] E.K. Rand, Supposed Autographa of John the Scot, dans University of California, Publications in classical Philology, 1920, t. V, p. 3, ne pense pas que l’érudit allemand ait maintenu cette manière de voir. H. Bett, Johannes Scotus Erigena: A Study in mediecal Philosophy, Cambridge, 1925, p. 12, conserve, par ailleurs, l’attribution de cette pièce à Scot Erigène; M. Esposito, The poems of Colmanus «Nepos Cracavist» and Dungalus «Praecipuus Scottorum», dans Journal of Theological Studies, t. XXXIII (1932), p. 118, no 7 pose à nouveau le problème de l’origine érigénienne de cette poésie, mais sans apporter de solution définitive.” 24