BAKER, AARON, Eunapius' "Nea Ekdosis" [Greek] and Photius , Greek, Roman and Byzantine Studies, 29:4 (1988:Winter) p.389 Eunapius' NEa "EKOOO"LS' and Photius Aaron Baker T HIS STUDY of the v(a €ICOO(TLS of Eunapius' History seeks plausible answers to two questions: what the via ("aOCTLS was, and how it came to be. 1 Our principal testimony for the vla €KaO(TLS is codex 77 ofPhotius' Bibliotheca, and it will be a primary concern here to determine the meaning of that testimony and and whether it is likely to be reliable. Fragment 41 of the History, much cited in the debate regarding £ICaO(T(£S,2 will be treated in an appendix below. In codex 77, the great ninth-century patriarch says this of the two editions in which he found the History (54a26-b3): l> ' l> \ '\. \ ,<, , ", uVO uE 7rpaYfLaTf.LaS T7JV aVT7JV 7rEPLEXOVCTas LCTTopLav CTvvEypa."aTO, 7rPWT7JV Kat OEVT£pav. Kat EV fLEV Tn 7rPWTV 7rOA>"~V KaT a TijS Ka8apas < n ~ X ~, , Q~ ",' \ \ l> ' ~ ~ \ n • Q~ Q ~ , 1\7JVLK7JV a7rOCTEfLvVVEL uELCTLuaLfLovLav, 7rOl\l\a TWV EVCTE,..WV ,..aCTLI\EWV 8 I • '1>\ n 'I> ' " \ , "5> • I", ' , Ka a7rTofLEvOS' EV uE TV UEVTEPI!-, "IV KaL VEav EKuOCTLV E7rLypa."EL, T7JV fLEV ~ ~ \ " Q \.,~ " \ ~ , Q' , '5> ~ < 7r01\1\7JV V,..PLV KaL aCTEI\YELaV, "IV KaT a T7JS EVCTE,..ELas ECTKEua~EV, V7rO- I 'I>' ~ 'n "'~ n " "'I> < Tf.fLVETaL, TO uE I\OL7rOV T7JS CTvyypa.,,7JS CTwfLa CTVVELpas VEav EKuOCTLV, WS "", ' , '" " ~ ~ \ ~. ~ ~, < "" , '" ~ E.,,"IfLEV, E7rL,),pa."EL, En 7rOl\l\a T"IS EKELCTE I\VCTCT"IS V7ro."aLvoVCTav. afL."oLV OE TaLS EKOOCTECTLV EV 7raAaLOLS EVETVXOfLEV /3L/3AtOLS, 1.0tWS EKaT£pav EV C" I , (' I I 'c'" , '" "~ A.. \ ETf.P~ Tf.VXEL KaL ETEP~ CTVVTf.TaYfLEV7JV· E~ WV aVTWV KaL T7JV uLa."opav , ~ t'" Q'''' • n , ''I>' ~~, n aval\E.,.afLEvoL E,),VWfLEV. CTvfL,...aLVEL ovv EV TV VEl!- EKuOCTEL 7rol\l\a TWV , '1>\ \ 'n < ~ \ , "'~ , ~8 XWPLWV uLa Tas YEYEV7JfLEvas TWV P7JTWV 7rEpLK07ras aCTa."ws EKKELCT aL, I", " ~ ",n .~~," ,~, ," \ KaLToL ."povnCTT7JS ECTn TOV CTa."ovs· al\l\ OT~ Tp07r~ I\EyELV OVK EXW, fL7J ~~ \ \ \ <, \~, '~50 I '50' Kal\ws KaTa Tas 7rEpLK07raS apfLoCTas TOVS l\oyOVS EV TV uEVTEPI!- EKuOCTEL \ ,... ,\' ,..", , TOV vovv I\VfLaLVETaL TWV ava')'LVWCTKOfLEVWV. He composed two works embracing the same history, a first and a second; and in the first he scatters much blasphemy against the pure 1 The vta E'"aoau is especially important in that those portions of Eunapius' History preserved in the tenth-century Excerpta de sententiis (produced at the command of Constantine VII Porphyrogenitus) are known to have been drawn from it (the Eunapian passages, together with a heading that attributes them to the vEa l"aoau, are on pages 71-103 of U. P. Boissevain's edition of these Excerpta [Berlin 1906]). The vEa E'"aOITlS is thus the likely source of those Eunapian passages preserved in the Ex­ cerpta de legationibus, another volume of historical extracts produced underConstan­ tine VII (for the Eunapian passages here, see C. de Boor's edition, II [Berlin 1903J 591-99). The tenth-century Suda drew its historical quotations almost exclusively from Constantinian excerpta; thus quotations from Eunapius in the Suda very likely go back to the vta E'"aoIT's as well (for the historical sources of the Suda see A. Adler, REIVA.l [1931J 675-717 s.v. "Suidas," at 700-06). 2 Fr.41 Muller=fr.41.1 Blockley=no. 39 in the Excerpta de sententiis. 389 BAKER, AARON, Eunapius' "Nea Ekdosis" [Greek] and Photius , Greek, Roman and Byzantine Studies, 29:4 (1988:Winter) p.389 390 EUNAPIUS' NEA EK.D.O~I~ AND PHOTIUS faith of us Christians, and glorifies pagan superstition, inasmuch as he attacks the pious emperors in many ways. But in the second, which he also entitles "new edition," he trims the great violence3 and wantonness that he scattered against piety, and after stringing together the remaining body of his composition, he entitles it "new edition," as we were saying, although it still contains much of the madness there [i.e., in the previous edition]. We came upon both editions in old books, arranged separately in one volume and anoth­ er.4 Having read from them, we recognized the difference. The result, then, is that in the new edition many passages are obscure because of the cuttings that have occurred in the text, although clarity is a concern of his. But, in what way I cannot tell, since he has not adjusted well the language in respect to the cuttings, in the second edition he ruins the sense of what is read. The first modem scholar who sought to define the character of the vEa rlCaO(T'~ was B. G. Niebuhr, who suggested that the History was abbreviated by an ignorant copyist to remove especially offensive anti-Christian passages. He gave as a parallel the expurgation of cer­ tain books after the Council of Trent. S C. de Boor, elaborating im­ plausibly on Niebuhr, argued that someone edited the History as part of a "Weltgeschichte in Einzeldarstellungen," and that the resulting clipped version was the vta rlCaO(T,~.6 J. C. Vollebregt, also following Niebuhr, ventured a simpler suggestion: the History was rearranged by someone other than Eunapius to present its historical data in a 3 T~V . . . 7rOAA~V iJ/3p'v .•• V7roT'p.Vf:Ta, is difficult; alternatively, it might be rendered "he cuts away most of the violence," which gives a different meaning. 4 a.p.t/>oiv ~t Tais fICMuf:u,v fV 7raAa,ois fVf:TtJXOp.f:V /3'/3Alo,s, l~lCl)s ElCaT'pav fV ET'PCP UVVTfTayp.'V7Jv, seems translatable in roughly two different ways. Photius' most recent editor, R. Henry, offered this interpretation (I [Paris 1959] 159f): "Nous avons trouve ces deux editions dans de vieux exemplaires; dans l'un, chacune des deux etait a part; dans l'autre, elles etaient combinees." A very different translation is given by A. Nogara, "Note sulla composizione e la struttura della Biblioteca di Fozio, Patriarca di Costantinopoli I," Aevum 49 (1975) 213-42, esp. 233 n.60: "separatamente era scritta in un volume una (edizione), in un altro l'altra." The same interpretation is given independently by R. Goulet, "Sur la chronologie de la vie et des oeuvres d'Eunape de Sardes," JHS 100 (1980) 60-72, esp. 68. Nogara's (and Goulet's) proposal is less ingenious than Henry's but perhaps more plausible; as Nogara points out, one manuscript containing two different editions of the same work was unusual. S See the introduction to his edition of Dexippus, Eunapius, et al. (CSHB [Bonn 1829]) xix. 6 "Die via llC~ou,s des Eunapios," RhM 47 (1892) 321ff. De Boor appealed to a scholium preserved at the beginning of the Eunapian portion of the Excerpta de sententUs (p.71 Boissevain). This scholium apparently came from the text of the via ;IC~OU'S used by the imperial excerptors and shows that this text was in a volume that contained also a text of Priscus of Panium; thus the theory of a number of works edited to form a world history. BAKER, AARON, Eunapius' "Nea Ekdosis" [Greek] and Photius , Greek, Roman and Byzantine Studies, 29:4 (1988:Winter) p.389 AARON BAKER 391 more regular order, with a severe pruning of redundant polemic.7 Like de Boor, however, Vollebregt was guilty of an unwarranted assump­ tion, for Photius stresses the removal of anti-Christian material, with­ out hinting at any re-ordering of contents. A contrary opinion, that Eunapius himself produced the vta (lCaOcrtS, has been asserted and, with some modifications, now prevails.8 The case for Eunapian authorship is made to rest on a number of con­ siderations. First, Eunapius appears to have published his History more than once. In his Vitae sophistarum he twice says that, God willing, he will describe certain events in the History.9 In all other cross-references from the Vitae to the History, he asserts that this or that has already been recounted in the latter work. 10 The natural inference is that Eunapius brought out the History in installments, one or more of which were available as of the publication of the Vitae (probably late in 399).11 Several scholars have therefore equated the old (lCaOcrtS with one installment or set of installments (extending to ca 378 or 395),12 and the vta (lCaoa'LS with a completed and revised form of the work, ending with 404.13 Secondly, Photius unambiguously 7 Symbola in novam Eunapii Vitarum edition em (Amsterdam 1929) Theses ix-x. 8 The following will be cited by author's name: W. R. CHALMERS, "The NEA 'EK~Ol:Il: of Eunapius' Histories," CQ N.S. 3 (1953) 165-70; F. PASCHOUD, Cinq etudes sur Zosime (Paris 1975) 171 if, and "Eunapiana," BHAC 1982183 (1985) 239- 303, at 284-92 (here, however, he proposes that the vfa f'lCooau was abbreviated by some other person or persons [290]); T. D. BARNES, The Sources of the Historia Augusta (Brussels 1978) 114-17 (he concedes that Eunapius "did not necessarily write both versions in the form described" [114]); A. B. BREEBAART, "Eunapius of Sardes and the Writing of History," Mnemosyne SER. IV 32 (1979) 360-75, esp. 361f; R. C. BLOCKLEY, The Fragmentary Classicising Historians of the Later Roman Empire I (Liverpool 1981) 2f; A. BALDINI, Ricerche sulla storia di Eunapio di Sardi (Bologna 1984) 75-117. 9 VS 7.3.4 Giangrande (476 Boissonade2 ); 8.2.3 (482). 10 6.3.8 (464), 6.11.7 (472), 7.1.5 (473), 7.3.4 (476), 7.3.7 (476), 7.3.8 (476) 7.4.10 (478), 7.6.5 (480), 9.1.3 (483), 10.1.1 (485), 10.7.13 (493), 21.1.4 (498). 11 See T. M. Banchich, "The Date of Eunapius' Vitae sophistarum," GRBS 25 (1984) 183-92. 12 How far Eunapius had taken the History when he brought out the VS is disputed. C. Muller argued that he had reached the death of Theodosius I in 395 (FHG IV [Paris 1851] 8); MUller is followed by Chalmers (165). Barnes initially maintained that Eunapius had reached the battle of Adrianople in 378 (114-17), but in response to criticism from Paschoud, who argues for 395 ("Quand parot la premiere edition de l'histoire d'Eunape?" BHAC 1977178 [1980] 149-62), Barnes has conceded that the History may have continued down to ca 383 (in Constantine and Eusebius [Cam­ bridge (Mass.) 1981] 403f n.5). Banchich provides additional arguments for the theory that the portion of the History available when Eunapius published the Vitae concluded ca 378 ("Eunapius and Jerome," GRBS 27 [1986] 319-24). 13 Chalmers, Paschoud ("Cinq etudes" 171if, "Eunapiana" 284-92), Barnes (114- 17), Baldini (75-117). Breebaart (362 and n.13) and Blockley (3) believe that Eunapi­ us produced the vfa f'lCooau but do not equate the two ;1C060'"£&~ seen by Photius with BAKER, AARON, Eunapius' "Nea Ekdosis" [Greek] and Photius , Greek, Roman and Byzantine Studies, 29:4 (1988:Winter) p.389 392 EUNAPIUS' NEA EK~OLrE AND PHOTIUS attributes the vEa (I(OOCTtS to Eunapius. I4 Finally, these contentions have been made: both editions must have had some independent value if they were preserved together;I5 the vEa ;l(oou"LS retains enough anti-Christian bile to render bowdlerizing by a Christian hack un­ likely; 16 the deficiencies of this edition are best explained by Eunapi­ us' haste, carelessness, mental decline, or death.I7 To determine whether this theory-that Eunapius produced the vEa (I(OOlTLS-Can be maintained, the statements that Photius makes in his comparison of both editions must be carefully examined. IS Having asserted at the beginning of codex 77 (53b.36-54a.4) that the vEa (1(00- lTLS" covered the years from 270 to 404, Photius now makes clear that both it and the first edition had the same temporal span (Mo 7fpaYlla- , \ , \ 't')' (' b " " ) TEL as T1JV aVT1JV 7f€PL€XOVlTas LlTTOpLaV • 7f€PL€XCJ> em race, compnse is unproblematic; Photius evidently liked to use it for the encompass­ ing of periods and events. I9 Whether I.lTTopla be translated "history," the installments that Eunapius published. Goulet, who believes (71) that "Nous n'avons donc aucune preuve decisive de l'existence d'une seconde 'edition' de la Chronique faite par Eunape lui-meme," also maintains that the question of install­ ments of the History has no bearing on that of editions. 14 Chalmers (168) notes this fact without giving it the weight it deserves. 15 So Chalmers 167. Of more than a little relevance to this argument is the question whether Photius ever saw the two fleaoCTnS bound in one volume (cj supra n.4). If in fact he found them together in one place, but bound separately, of course no per­ tinent inference could be drawn from their proximity. Chalmers actually says that the two editions were preserved together "often," an assertion justified by no reading of Photius. On the interpretation proposed by Henry, Photius saw both fleOOUfLS bound in one codex with apparently some matter intervening, and bound next to each other in another codex (that by TfVXOS Photius means 'codex' is urged, correctly I believe, by T. Birt, Das antike Buchwesen in seinem Verhtiltniss zur Litteratur [Berlin 1882] 26). Twice is not often, and of course one codex could have been copied (with the addition or elimination of intervening matter) from the other. If both editions were ever bound together (and the assertion that they were rests only on a doubtful reading of Photius' difficult Greek), the reason could simply be that they were known to be a first and a second. 16 So Blockley 3. 17 So Chalmers 170; Blockley 3. 18 What follows will show that I agree fully with Goulet's observation (69) that "on peut retenir qu'elle [the vEa E'IeOOULS] se caracterisait essentiellement par des coupures ayant laisse un texte parfois decousu et que son extension etait la meme que Ie premiere, du moins au jugement de Photius qui a examine assez attentivement Ie texte pour deoouvrir la disparition de certains passages." Unfortunately, as will be­ come clear, more than a bald assertion of the obvious is necessary. 19 Cj, for example, cod. 57 on the books of Appian's Roman History, 15b.26f, I ~,\ (h lei f R ) " \ J. ~ I 37f ~ \ , ~ TOVTWV TWV f'7l'Ta t e ngs 0 ome fpya Tf leaL '7I'P~f£S '7I'fPLfXf£; , a P.fV 0 '7I'PWTOS \ I I ~ I , 41 f • ~ \ '"" ~~ I \ \ \ T' I 'p I I\oyos '7I'fPLfXfL, TaVTa funv; , 0 uf f.,.f~71S '7I'fPLfXf£ TOV '7I'pOS TOVS ~aVVLTas wp.aLWV '7I'OAfP.OV; 16a.3f, b O£ TETapTos, fwfL TOV '7I'pOS KfATOtS '7I'fPLEXf£ 'Pwp.alwv '7I'OAfP.OV; 7-9, b O£ ¥~oop.os •.• f'll'fL TOV 'll'pOS TOV 'Avvl~av TOV KapX71oovLOV 'll'fp"Xf£ 'll'OAfP.OV; 24-31, fJl.'II'fp"XfTaL O£ TOts fP.4>vAloLS 'll'PWTOV JI.£V Tet 'll'fPL MapLov leaL l:fiAAav a.AA"AOLV BAKER, AARON, Eunapius' "Nea Ekdosis" [Greek] and Photius , Greek, Roman and Byzantine Studies, 29:4 (1988:Winter) p.389 AARON BAKER 393 "story," or "narrative" is of little importance here; both editions had one I.CTTopia in common, and if Photius knew that something had been added to the I.CTTopia of the vl.a ;1C.1jOCTL~, he could reasonably be ex­ pected to make clear, in the next sentence or so, what it was. Instead, he says that the v£a ;KaOCT'~ had suffered abridgment: anti-Christian matter had been removed and what remained was strung together. Moreover, this job of editing had been done so ineptly that the meaning of the work was often obscured. If Photius has given a cor­ rect account of what he saw, the simplest and most plausible explana­ tion is that the old ;KOOCTLS represented a complete version by Eunapi­ us, and the new a clumsy bowdlerizing of the same. The two EKaoCT£L~ should then pose a different problem from that of the installments which, on the evidence of the Vitae sophistarum, Eunapius seems to have produced. It must then be asked whether Eunapius was likely to have been the expurgator of the vl.a ;KOOCTL~. Against the opinion that he was, one may first adduce a not very compelling argument from silence: neither reference in the VS to a continuation of the History indicates that the work will otherwise be altered. More important, Eunapius was a fanatic pagan and, when he finished the History, an aging one.20 Also, whatever modems may think of his gifts, he was a stylist. That his hatred of Christianity abated is scarcely likely; even if it had, and if he thought himself obliged to moderate his tone, he would hardly have mutilated his life's work in the way Photius describes, even if his faculties were failing. Nothing, moreover, suggests that he was com­ pelled, or found it politic, to cut portions from his History. Further­ more, tampering by a Christian is no less likely because offensive , , ' " \ \ n ,.. \'1" \ K' ',,- t~ ~\ \ U7fO"fp-fJuaUTO'U, f7fUTa Ta 7ffP' oP-7ffJ'ou lCa, ov,,'ou TOU a,uapa ... f'f'f~fJS uf Ta 7Tfipt ' AVTri>VlOV lCat ' OICTciv,'ov Katuapa. 20 Eun. Hist. fr.87 MUller (=fr.72.l Blocldey=Exc. de sent. 79) is usually believed to provide a terminus post quem for the History at 414, because it begins ()TL {7ft llOVAXtplas TijS /3au'AtuuTJS (Pulcheria became Augusta in 414). But because of a possible chronological difficulty posed by this fragment in conjunction with fr.86 (=71.4=78), which can be dated to 404, Blockley suggests that {7ft llOvAXEplas is a mistake of the excerptor, and that the events described in fr.87 took place during the tenure of the empress Eudoxia, i.e., no later than 404: "The Ending of Eunapius' History," Antichthon 14 (1980) 170-76; see also Historians 15, II (1983) 149 n.176. If the correct terminus post quem should be 404-the earliest propounded by any scholar-then Eunapius, who was born in or near 348, could have been no younger than his late fifties when he finished his work. For the evidence bearing on the chronology of Eunapius' life, see Goulet 60-64, with criticism by Banchich, "On Goulet's Chronology of Eunapius' Life and Works," JHS 107 (1987) 164-67, and by A. E. Baker, Eunapius and Zosimus: Problems of Chronology and Composition (diss.Brown Univ. 1986) 1-18, 103-10. BAKER, AARON, Eunapius' "Nea Ekdosis" [Greek] and Photius , Greek, Roman and Byzantine Studies, 29:4 (1988:Winter) p.389 394 EUNAPIUS' NEA EKAO~Il: AND PHOTIUS matter remains. 21 In a history whose hero was Julian the Apostate, the elimination of all that gave offense might have meant the elimination of the work.22 That Photius ascribes to Eunapius the second (lCaOC7U and its expurgation is of course important, and Photius must not be lightly dismissed. If the new (lCaOC7'&S was entirely the result of a later hand or hands, Photius was guilty of an easily explained error. The heading of the manuscript (or manuscripts) he saw was probably th O l'k E' I 'r ~ ~ • I. \ A 'i: I ., ~ some lng I e vvaw&ov ~apu&avov &CTTOp&a 11 p.f.Ta ~f.~&WWOV, Vf.a f.lCuO- CT&S; and so he could have made a reasonable inference from that title.23 Thus the theory of Niebuhr should be retained: the vta (lCaOCT&S represented an expurgated version, produced by a Christian, who strove to render the work more tolerable to the pious. The arguments given here would collapse if Photius could be proved to have erred regarding the chronological limits or content of the EICOOCTf.&S. Two scholars, Baldini and Paschoud, who believe that the ElCooCTns differed substantially,24 have in fact charged Photius with error. Baldini (220) considers it unlikely that Photius read in its entirety every book he cites, and to support his skepticism he adduces codices 76 and 82. Codex 76, on the Jewish Antiquities of Josephus, after giving the chronological limits of the work, presents a summary confined to Book 20.223-57 (52b.18-53a.34). Codex 82, on Dexippus 21 In fact, among the extracts from Eunapius' History that survive in the Excerpta de sententiis and de legationibus, as well as in the Suda, criticisms of Christian beliefs and institutions (as opposed to attacks on important Christians, such as Constantius II and Theodosius I) are almost lacking. There is a sneer at monks in fr.55 (=4S.2= 53), and the pagan Fravitta is called ;'Yfp1.>U Tfjf p.~u SfOq,&AOVf lCal Sflaf p.fpl~Of (fr.60= 6S=Exc. de leg. no. 7), with an implicit gibe at Christianity, but little else can be found (see also fr.7S=6S=Exc. de leg. no. 72). Unfortunately, the paucity of explicitly anti-Christian passages proves nothing about the efficiency with which the expurgator performed his job; the excerptors working for Constantine VII, and the compilers of the Suda, as Christians, would probably have omitted anti-Christian matter when they could and when they recognized it as such. 22 Frr.8-27 (=15-28.7) remain from Eunapius' account of Julian (7a=14.1=4, al­ though about Julian, belongs to Eunapius' account of Constantius 11). Fr.S (=15=5), from the preface to Book 2, is especially important here, for in it Eunapius affirms that Julian is the principal subject of the History. 23 Photius introduces codex 77 with the statement &'ufyuwrrS." Evua1l'lov x,pOUtlCfjS tUToplaf Tfjf P.fT'O. LlE!t1l'1I'OU, uEaf flCMufCllf (53b.34f). The passages from Eunapius preserved in the Excerpta de sententiis are introduced by the heading flC Tfjf < IUToplaf Evua1l'lov l:apa&allov Tfjf P.fT'O. LlECt1l'1I'oll uEaf fKMufCllf (p. 71). Both notices were probably drawn from titles. That Photius made the inference in question may be supported by his language: he twice says that Eunapius f1l'typo.q,n the work uEau f'lCaOutU (54a. 31 f, 34f). 24 Baldini suggests that the first f'1C~OUtf of the History began with Augustus (20S-30) and concluded with Adrianople in 378 (100-17) and that the uEa llCaOUtf was to some extent a restructuring and re-ordering of the first edition (SO). Paschoud's latest judgment ("Eunapiana" 2S4-92) is that the two (ICMum differed in content and per­ haps in temporal span. BAKER, AARON, Eunapius' "Nea Ekdosis" [Greek] and Photius , Greek, Roman and Byzantine Studies, 29:4 (1988:Winter) p.389 AARON BAKER 395 of Athens, epitomizes only an initial portion of that author's Ta IUTa , A).Etavopov (64a.21-b.31). To these might be added codex 63, on Procopius of Caesarea, which summarizes his Wars no further than Book 2.19. Baldini also (218f n.45), though without much confidence, raises the possibility that the genitive after aVEl'v6Jcr87J at the beginning of codex 77 is partitive (53b.34f, aVf-yvwCT(JT/ Evva7l'iov XPOVLlcijS LCTTopias Tij!l fLETu' AitL7r7TOV, via!l EK06crEO)!I). He also draws from this sentence the correct inference that codex 77 refers principally to the new, not the old, €l(oocr&!I (220f). Paschoud (289) urges that one may take it practically for granted that Photius did not compare the (KOOcrE&!I page by page "comme ferait un bibliographe modeme." Moreover, Pa­ schoud asserts that the strangeness of the exemplars with which Pho­ tius dealt could easily have led him astray.25 Neither Baldini nor Paschoud makes a satisfactory case, for neither takes sufficient account of what is known of the methods Photius used to compose his Bibliotheca, and both ignore pertinent information provided by codex 77. To ascertain the temporal bounds of both editions, Photius had merely to check the first and last leaves of both, and from the information he gives regarding many works he cites, he very likely made for these the requisite inspection: a quick check of the beginning and end as well as a few other portions of the cited work to give such details as "the author, the title, the dedication, the number of parts in the work, the number of works in the volume, the table of contents, the chapter headings, and a few facts from the first page, the last page, or a page in the middle. "26 Codex 77 is one of many in the Bibliotheca whose arrangement is best explained by the theory of a summary check: it gives the author, the title, the edition, 25 Paschoud ("Eunapiana" 285) accepts Henry's interpretation of latoos (/CaTtpav €V h£p~ Tt:VXE' Kal Er£P~ (J'vvTETayp.'VTfv. His other arguments are based on a misinterpre­ tation of (J'vvdpoo, which, apparently following Henry (I 159), he renders as "ramener a l'unite" (284f); "on peut supposer, sans risquer trop de se tromper, un remaniement comportant des suppressions ainsi que des adaptations visant a masquer les lacunes et les sutures resultant de ces suppressions" (290). Paschoud complains because, on his reading of (J'VVEtpoo, Photius later contradicts himself with his statement about clumsy cuttings that obscured the meaning of the new (/Cao(J"s. To diminish the imagined contradiction, Paschoud posits these stages: abbreviation and restoration by Eunapius, then mutilation by some other hand or hands. In fact, (J'vvElpoo never means 'restore to unity', but simply 'string together', 'connect'. I have inspected every example of (J'VVELPoo available in the computerized TLG as of April 1988, and find none that requires or justifies Paschoud's (and Henry's) interpretation. Paschoud should have been guided by LSJ s. V., which correctly defines the word. 26 W. T. Treadgold, The Nature of the Bibliotheca of Photius (Washington 1980) 92; at 84-88 Treadgold describes codices of the Bibliotheca in which this procedure was probably followed (under the heading Class II). K. Ziegler, RE 20.1 (1941) 667-737, at 716, s. v. "Photios (13):" plausibly suggests that Photius used /3[0' appended to the beginnings or ends of cited books for some of his biographical details. BAKER, AARON, Eunapius' "Nea Ekdosis" [Greek] and Photius , Greek, Roman and Byzantine Studies, 29:4 (1988:Winter) p.389 396 EUNAPIUS' NEA EK~Ol:Il: AND PHOTIUS the number of books (fourteen) and chronological span, as well as noting the existence of two editions and how they were arranged in the manuscripts that Photius had seen (53b.34-54a.4, 54a.26-39). If he customarily inspected the beginning and end of a work, it is hardly over-bold to suggest that he did so for both editions, particularly as he seems to have been interested in how they did and did not differ.27 To determine, with a good chance of escaping error, how the old and new €ICSouns were dissimilar, Photius must have read extensively from both. It may be objected that his tendency to extract informa­ tion from beginning, middle, and end of a given work proves that when he did so, he did nothing else. This objection is ill-founded. In many codices of the Bibliotheca that show this tendency, he also gives indications that he did read to a considerable extent, perhaps in full, the work in question.28 Codex 77 is again typical. It begins with the assertion that Photius read the vEa (ICSOUlS 29 and offers a judgment of Eunapius as stylist that seems based on a close acquaintance with at least one edition (54a.12-25). Thus the quick checking of cited works appears in many cases to have been a quick re-checking. Elsewhere in the Bibliotheca, moreover, Photius admits his inabil­ ity or unwillingness to read certain works from beginning to end. In codex 35, on Philip of Side, Photius admits that he has seen only the first twenty-four books of that author's Christian History (7a.31-33). In codex 40, on Philostorgius, Photius makes clear that he had read the first six books of his Church History, but then found a volume containing the final six books (8b.24-27). In codex 41, on John Diacrinomenus, Photius says that he has been able to read five of the ten books of John's Church History (9a.15-17). In codex 97, on the enumeration of Olympic victors by Phlegon of Tralles, Photius im­ plies that he found the work too boring to finish (83b.35, 84a.37- 84b.2). In codex 176, on Theopompus of Chios, Photius reports the opinion of "ancient writers" that the sixth, seventh, ninth, twentieth, and thirtieth books of the Philippica had disappeared, and says that he has not seen them either; but he asserts that he has read the twelfth 27 Since the History ended with 404, the year when John Chrysostom was per­ manently deposed from the patriarchate of Constantinople (53b.39-54a.4), and since it might have recounted this event in some detail, Photius may have had a special interest in the final parts of both editions: he admired John, and described more works by Chrysostom in the Bibliotheca than by any other author (see codd. 172-74, 270, 274, 277; also 96 and 273 [on works whose subject was John), and 59 [summar­ izing the Synod of the Oak, which first deposed him in 403)). 28 See Treadgold (supra n.26) 84-88 on several codices of this sort. 29 aV~'Yvwu811 or aVE'Yvwu811 {If. are the formulae with which Photius introduces by far the greater number of codices in the Bibliotheca. BAKER, AARON, Eunapius' "Nea Ekdosis" [Greek] and Photius , Greek, Roman and Byzantine Studies, 29:4 (1988:Winter) p.389 AARON BAKER 397 book, which, according to Menophanes, was lost (120a.8-14). In codex 224, on Memnon's history of Heraclea Pontica, Photius says that he has not yet seen either the first eight books or those following book 16 (240a.9-11). Furthermore, Photius frequently admits or im­ plies that he was unable to find a given work at all; this practice also encourages belief in his candor.30 Such apparent candor, combined with the circumstantial detail and accuracy of so much of the Biblio­ theca,31 renders plausible (though it cannot prove) the inference that, unless he says otherwise, Photius did read each work he cites from start to finish. 32 At the least, unless he was very disingenuous, he read extensively in most of the writings mentioned by the Bibliotheca. Of course one must not ignore the incomplete summaries noted by Baldini. Photius may not have read the Jewish Antiquities, Ta JJ-ETa 'A>..ffavopov of Dexippus, or Procopius' Wars page for page. Incom­ plete summarization, however, might indicate instead an especial interest in one portion of a work (for instance in the account of the High Priests at AJ 20.10) or, on the contrary, a loss of interest in the work as a whole. Photius ends his partial epitome of Dexippus' work on the Successors with these words: Kat Ta l1>..>..a odfELCnv EV 7I"0>">"0'S', wS' , I 'A ~ " \ ~ I "" I"" (d 82 Kav TOVTOLS', ppLav~ KaTa TO 7I"I\ELCTTOV CTVJJ-."wva "Ypa."wv CO. , 64b.31f). In codex 92 the Bibliotheca epitomates all ten books of Arrian's history of the Successors. If, as seems likely, Dexippus did little more than imitate Arrian, and if Photius had already composed his summary of Arrian's work, he may have decided, after beginning, that a full summary of Dexippus was redundant. The strange genitive at the beginning of codex 77 also merits an explanation. As Baldini has suggested, it may be partitive, as may the genitive at the beginning of codex 76: aVE"Yv~CT8TJ CP>..a{jlov 'IwCT~7I"ov 30 See, e.g., codd. 14 (4b.19-21), 58 (17b.21f), 98 (84b.32-36, where Photius infers, rightly or wrongly, the existence of a tirst edition of Zosimus' New History, and admits that he has not seen it), 111 (89b.38-90a.l), 213 (171a.19-27), 268 (496b.38- 40 [in this codex, devoted to the orator Lycurgus, Photius admits that he has read none of his speeches]). 31 Photius' detail and accuracy are made apparent in these treatments of the Bibliotheca: Ziegler (supra n.26) 713-19; Nogara (who emphasizes, however, the casual, unsystematic nature of Photius' reading: 224-40); Treadgold (supra n.26) 37- 96. For a somewhat less favorable assessment, see N. G. Wilson, Scholars of Byzantium (London 1983) 93-111. Unlike Wilson (95), I see no reason to doubt Pho­ tius when he says that he has read speeches by Hyperides (cod. 266, 495b.2). 32 For the importance of Photian candor, see C. Mango, "The Availability of Books in the Byzantine Empire, A.D. 750-850," Byzantine Books and Bookmen (Washington 1975) 29-45, at 39 (reprinted in Mango, Byzantium and its Image [London 1984] no. vii). The regular practice of reading books from beginning to end is recommended by the eleventh-century writer Cecaumenus (Strategicon pp.212, 240 Litavrin), and so should not be assumed alien to Byzantine readers. BAKER, AARON, Eunapius' "Nea Ekdosis" [Greek] and Photius , Greek, Roman and Byzantine Studies, 29:4 (1988:Winter) p.389 398 EUNAPIUS' NEA EK~Ol:IE AND PHOTIUS 'IovlIa'·".qs apxaLoAoylas, (V AcSYOLS Ie' (S2b.18f). A partitive genitive might indicate (though it could not prove) partial reading. Yet codex 59, on the Synod of the Oak, also begins with such a genitive: 4V£- I (J I~ - I \ -,. I '1 I - X yvwu 7J UVVOuov T7JS 7fapavoJJ.ws "aTa TOV £v aYLOLS wavvov TOV pVUO- UTcSJJ.OV uvy"poT7J(J£lu7Js (17b.2S-27). The summary that follows (so far as can be judged; the acts of the Synod are lost) seems to be based on a complete reading. Indeed, given his interest in Chrysostom, Photius is unlikely to have read the record of the Synod superficially. These genitives are perhaps better interpreted as dependent on an unex­ pressed word for 'book' or 'manuscript' than as partitive. Although it is likely that Photius read the vla l"lIo(TLS (the principal subject of codex 77, as Baldini asserts) from beginning to end, it cannot be established how extensively Photius explored the first 1,,­ aOULS. But TO at- AOL7fcSV T.qS uvyypac/>.qs uwp.a uvv£Lpas does imply a close acquaintance with the uwp.a of the earlier edition; and (' ~v aVTwv "at T~V aLac/>opav avaA£fo.p.£VOL lyvwp.£v makes clear that Photius did read in both the old and new (ICMuns, whether or not he went through the first in its entirety. If ff ~v in this clause is taken, together with aVT(;w, to refer to ai fICacSUtLS, then these are possible renderings: "having read from them, we recognized the difference," or "from them we recog­ nized the difference, having read [them]." Instead of being taken with aVTwv, ff ~v might mean "from that," "from that fact."33 Unfor­ tunately, regardless of how these words are interpreted, they prove neither that the reading involved was partial nor that it was complete. Evidence from elsewhere in the Bibliotheca shows that Photius did check several divergent manuscripts of a given work, although how closely he compared them is beyond determining. For instance, in codex 112-13, on works falsely attributed to Clement of Rome, Pho­ tius says that some manuscripts of the pseudo-Clementine Recogni­ tions are prefaced by a letter supposedly from St Peter to St James the Greater, others by a letter purporting to be from Clement to James (90a.IS-21). Treadgold points out34 that while only manuscripts of the Pseudo-Clementine Homilies now contain the letter from Peter, the main text of the Homilies has an initial portion identical to the beginning of the Recognitions.35 Thus Treadgold conjectures that Photius, having read the Recognitions, saw some manuscripts of the Homilies; he noted the different letter, then read a short way, and, 33 As Baldini suggests, 222 n.S2. 34 Treadgold (supra n.26) 10f. 35 So far as can be determined from Rufinus' Latin translation of the latter work, which is otherwise lost (Recognitions pp.6-10.6 Rehm corresponds to Homilies 23- 26.16 Rehm2). BAKER, AARON, Eunapius' "Nea Ekdosis" [Greek] and Photius , Greek, Roman and Byzantine Studies, 29:4 (1988:Winter) p.389 AARON BAKER 399 assuming that he was dealing again with the Recognitions, went no further. Treadgold's hypothesis is plausible, but one fact should be noted: Photius states emphatically that after the varying introductory letter, all the manuscripts he saw were identical: EV 1TaO"L yap TO'S ~ ~). , ",'~ """).,, ' , ,....L,....I\LOLS a nUOJLEV ••• T7JV aVT7JV EVpOJLEV a1Tapal\l\aKTWS 1TpaYJLaTnav (90a.29-33). He may have been guilty of hyperbole, or he may indeed have seen, and examined with care (perhaps to the extent of reading them all cover to cover), manuscripts of the Recognitions that con­ tained a letter supposedly from Peter. Photius speaks with similar assurance and emphasis about the manuscripts he saw of the Stroma­ teis of the Alexandrian Clement: from Books 1 to 7, the manuscripts were uniform (€VLa'i.'oL); but in some, the eighth book was replaced by what is now usually known as the treatise Quis dives salvetur? (cod. 111, 89b.27-35). The most plausible conclusion is this: Photius read the vEa lKOOO"LS cover to cover, and read enough of the first lKaOO"LS to feel confident in asserting how the two editions differed and were alike. He may have read large portions of that lKaOO"LS, or he may have read it whole. Given his apparent conscientiousness, he may have carried out what Paschoud is certain he did not: a page-by-page comparison of the two editions. Nor, since Photius seems to have been a careful reader, is he likely to have been misled by the manuscripts of the lKaDO"ELS, however strange they might have been. Thus the vEa lKaOO"LS was probably what Photius says it was, and the burden of proof rests on those who would argue otherwise. Once the testimony of Photius concerning the vEa lKaOO"LS is taken seriously, an important conclusion follows. Since the vEa lKaOO"LS was probably in its entirety the work of a later hand, it may not be used in support of any theory regarding Eunapius' methods of composition. No evidence exists that Eunapius revised his work by either adding or subtracting material. He evidently brought out his History in install­ ments, but that he did so is a conclusion established without reference to the Bibliotheca. ApPENDIX: Fragment 41 of the History Fragment 41 of the History is often cited in the controversy over install­ ments or editions because of these words in its first sentence: , , • ~ ~ .I.~'~' ,~, .I.' \' " Ta P.EV ovv 7fpWTa T71S (Tv'Y'Ypa.,,71S, OVuEVOS OVuEV (Ta."ES ,M'YEtV EXOVTOS "8 " < O· If I 'E' I ~ , I~ o EV TE OVTEs OL VVVOL u7r'!l TE KELP.EVOL T71V VpW7r1Jv 7ra(7"av E7rEupap.ov , ,~ (J' "./, I , ~ \ ~ (J I " KaL TO .L.KV LKOV ETpL."av 'YEVOS, EK TWV 7fal\aLWV (TVvn Evn KaT a TOVS BAKER, AARON, Eunapius' "Nea Ekdosis" [Greek] and Photius , Greek, Roman and Byzantine Studies, 29:4 (1988:Winter) p.389 400 EUNAPIUS' NEA EK~Ol:IE AND PHOTIUS " ~ ,., , "" ~, ~~ 1 "i:/~ , fLICOTas AO"YLCTILOVS fP7JTaL, Ta uf flC TCI)V a7ta"Y"YfAAOlLfVCI)V uo~al>0VTL 7tpOS \' IJ. I TO alCpL,..fS • ••• Cardinal Mai, reading IJ-EV otv 7tpWTa TfjS CTv"Y"Ypa4>fjs, interpreted the initial phrase to mean "In priore Historiae editione." 36 This interpretation was accepted by Niebuhr (supra n.5: 75). C. MUller, however, understood the phrase to mean: "In priore Historiae parte" (FHG IV 30). More recently, Chalmers (168f) has revived Mai's interpretation without, however, adopting his conjecturallCaTa. Chalmers is followed by others. 37 It should be noted that these two interpretations would gain greatly in plausibility if lCaTa conjectured by Mai could be proved; but Boissevain, in his apparatus to the Excerpta de sententiis (84), is more cautious: "Ta IJ-EV otv 7tpWTa dedi, r~ I!-fV otv 7tpWTa legi addidique in re praesenti 'tres litterae priores prorsus incertae.' " Blockley (II 140 n.90), who accepts Boissevain's reading, raises this cogent objection against both Mai and Miiller, as well as their respective followers: on the normal interpretation of Ta IJ-fV and Ta af, the things joined by the two particles should closely parallel each other. Eunapius "simply says that he is incorporating two different versions of the origin and situation of the Huns, the old, inaccurate version ofthe written records (Ta IJ-EV otv 7tpWTa ••• (IC TWV 7taAaLwv) and the new, accurate oral reports (Ta aE (IC rwv Q.7ta"Y"YfAAOlJ-fVCI)V) ... with the expression ra 7tpwra he is making no reference to a first edition or part of his own History." Blockley himself, however, incurs a serious objection. Others who have dealt with this passage have understood rfjs CTv"Y"Ypa4>fjs to refer to Eunapius' History, if they are right, Blockley's interpretation collapses. His rendering (II 59: "The first accounts of the history of the Huns") fails to allay the objection, for ra 7tpwra rfjs CTv"Y"Ypa4>fjs cannot be plausibly rendered "first accounts. " One reasonable solution has been proposed by T. M. Banchich.38 It is well known that the excerptors who worked at the behest of Constantine VII often altered the beginning or end of an excerpted passage.39 Banchich suggests that this tendency may account for the first words of fragment 41; apparently dismissing TfjS CTv"Y"Ypa4>fjs as a hopeless corruption, he comes independently to the same conclusion as Blockley. 40 Preferable, however, would be an interpretation, otherwise similar to those of Blockley and Banchich, that could plausibly account for the text as it stands. CTv"Y"Ypacf>~ often means 'written account', and so it has been inter­ preted by practically all who have dealt with fragment 41. But a CTv"Y"Ypa4>~ can also be a subdivision of a written account. The second-century historian 36 A. Mai, Scriptorum veterum nova collectio e Vaticanis codicibus edita II (Rome 1827) 269. 37 Barnes, Sources 116f; Paschoud (supra n.12) 152f; Breebaart 362 (accepting lCaTa: n.14). 38 The Historical Fragments of Eunapius of Sardis (diss.SUNY Buffalo 1985) 39-47, 58-60. 39 See P. A. Brunt, "On Historical Fragments and Epitomes," CQ N.S. 30 (1980) 477-94, at 483ff; Banchich (supra n.38) 43-47, 59f. 40 Supra n.38: 44ff, 60, 69, 102. BAKER, AARON, Eunapius' "Nea Ekdosis" [Greek] and Photius , Greek, Roman and Byzantine Studies, 29:4 (1988:Winter) p.389 AARON BAKER 401 Appian regularly uses the word when referring to books and other divisions, both larger and smaller than books, of his Roman History.41 Pausanias, his contemporary, uses uvyypa4>~ in the same fashion in his Description of Greece. In Book 9, his Boeotica, he says that what he has heard about Hyettus the Argive and Olmus son of Sisyphus 7Tpou'uTaL Kat aUTa Tfi ' OpXop.£vlQ. O"vy­ ypatPii (9.24.3). And indeed both Olmus and Hyettus are mentioned again in that portion of the Boeotica which refers to Orchomenus (Olmus, here called Almus, 9.34.10; Hyettus, 9.36.6ff). Plainly, therefore, when Pausanias refers h· , 8' I h' I , I h now to IS AT LY O"vyypa4>'1I, now to IS MEyapLK'1I or MEyapLY O"vyypa4>'1I, e means the relevant sections of Book 1.42 Likewise, Koptv8la, ' Apyo,\ls, and 1:,LKvwvLa O"vyypa4>~ all refer to the pertinent divisions of Book 2.43 Pausanias could also use O"vyypa4>~ for his description of a single building (10.19.5, r \ ~ Il' ~, , • E\ \ 'll ' I'" " ., , al\aTWV uE T1]S ES T1]V I\l\aua E7TtUTpaTEtas EXEt IJ-EV Ttva IJ-v1]IJ-1]v /Cat 1] ES TO {jOVAEVT~P&OV ~p.I.V TO 'ATTtKOV uvyypa4>~, a reference to a digression at 1.3.5- 4.6). Yet O"vyypa4>~ does duty also for whole books of his Description.44 What may be another example of the usage here described is provided by the sixth­ century chronicler John Malalas, who speaks of the O"vyypa4>aL of Diodorus Siculus (CSHB p.54). Malalas might, however, have been unsure how many separate works Diodorus wrote. Nowhere else in his extant oeuvre does Eu­ nap ius employ uvyypa4>~ to mean a division (defined, apparently, by subject matter) of a literary work. Yet he does refer once to his History with the plural of the related word O"vyypap.p.a (VS 6.11. 7 Giangrande=472 Boissonade2: '\\' " , " ~ 8 \ ~ ~. I I ") al\l\a 7TEpL TOVTWV P.EV KaL £v TOLY Ka OI\LKOLY T1]Y &O"TopLay O"vyypap.p.aULV np'1lTaL • Nothing precludes his having used O"vyypa4>~ with the desired sense in his Hunnic passage. If the word was so used, the uvyypa4>~ in Question was probably that 41 Hisp. 1.3, avct'YlC71 P.OL crvV€u€'YIC€iv €Y T~V 'I/371PLIC~v crv'Y'Ypa4>~u €'YfU€TO; Hann. 1.2, aKpt/3fcrTaTa p.~v fV Tfj 'I/3TJptKfj crvY'Ypa4>fj a€a~ACIlTa,; Syr. 51.260, aAAa TO-a€ p.~v €VT€AOOY (V Til ilapOucil crvyypa'Yil AfCCIl (the ilapOLIC~ crvyypay~ was perhaps never written; see E. Gabba in the Teubner Appian, edd. P. Viereck, A. G. Roos, and Gabba [Leipzig 1962] vii n.1); BC 1.6.24, lin P.OL Tijy Al'Yv1TT{a~ crvyypa4>ijs TaO€ 1TPOTJyovp.Eva lCat TEAEvT~crovTa El~ EKE{VTJV avaYlCaiov ~v 1TpoavaypO-"'acr8at; 1.6.25 (the conflict between Antony and Octavian), aPx~ Kat Tij~ Al'Yv1TnaK7j~ crvY'Ypa4>7j~ f'crTat; 1.34.151, Kat ahov (the Social War) o,a TO-OE crvv~yayov fY T~VOE T~V crvnpa4>~v (context does not make clear whether this crvY'Ypa4>~ is his entire account of the Civil Wars, in five books, or Book 1 alone); 2.90.379, aAAa TO-OE P.fV ¥KacrTa lI1TCIl~ €YfUETO, aKpL/3fcrTEPov ~ 1TEpt Alyv1TTov crvnpa4>~ OLfCELcrL (the Egyptian crvnpa4>~ was appar­ ently in four books: see Gabba vii n.2); BC 2.92.385, ;;,~ P.OL KaTa T~U 'AcrLav~v crvyypa4>~v OEa~ACIlTaL (this apparently was that section of Appian's 'EAATJV'K~ crv'Yypa- 4>~ that dealt with Roman actions in Asia Minor: see Gabba vi-vii n.4). 42 'AT8t~ crvnpa4>~: 2.21.4, 3.11.1, 3.17.3, 4.28.3, 5.10.4, 7.7.7, 7.20.6, 9.6.5; ME)'apLK~ (Mf'YapL~) crvl"Ypa4>~: 2.19.8, 9.19.2. 43 Kop,v8{a: 5.18.8; , Ap-yOA{~: 4.2.4, 8.4.6; I.lKvCllv{a: 9.5.5. 44 AaKCIlvLK~, 3.3.5, 4.2.4; MEcrcr71v{a, 3.7.5, 3.15.10, 4.6.4, 8.31.1; (Y 'AxaLOv~, 6.3.8; ~ of Herodotus, it is unclear whether Pausanias had in mind the first book of Herodotus as a whole, or a Lydian portion distinct from, for instance, a Babylonian portion (Herodotus tells of Arion in 1.230. BAKER, AARON, Eunapius' "Nea Ekdosis" [Greek] and Photius , Greek, Roman and Byzantine Studies, 29:4 (1988:Winter) p.389 402 EUNAPIUS' NEA EKilO~I~ AND PHOTIUS portion of the History dealing with the Huns (almost certainly a digression rather than a full book). Eunapius can be correctly translated thus: The first elements of the account [i.e., the Hunnic account]-since no one could say clearly where the Huns were from nor where they were situated when they overran all of Europe and crushed the Scythian race-have been stated by composing from the Ancients on the principle of reasonable reckonings; the other elements [have been or will be stated] from oral reports, by forming an opinion with a view towards accuracy .... The passage so interpreted is (for Eunapius) plain in meaning and irrelevant to the controversy regarding editions and installments of the History. THE PENNSYLVANIA STATE UNIVERSITY January, 1989