Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Missing links in the development of scholia McNamee, Kathleen Greek, Roman and Byzantine Studies; Winter 1995; 36, 4; ProQuest pg. 399 Missing Links in the Development of Scholia Kathleen MeN amee S CHOLIA, understood as extensive commentaries on Greek or Latin literary texts, are typically compilations of earlier commentaries with long, scholarly pedigrees. For Greek authors, the best scholiastic sources date from the Alexandrian period. Scholia themselves appear in their most familiar form in manuscripts of the ninth century and later, when professional scribes copied them neatly into margins that had been left wide to accommodate written commentary above, below, or beside the subject text. The question whether they first assumed this form in the ninth century or earlier, however, is unresolved, despite much discussion in the last thirty years. Gunther Zuntz opted strongly for a ninth-century genesis. Nigel Wilson found reasons to push the first appearance, for some texts at least, back to late antiquity. 1 Meanwhile, however, it escaped attention that a good two dozen annotated papyrus codices of the fourth century and later display, for the first time, a correlation between large for­ mat and heavy, planned annotation (see the Table, 413 infra).2 Their marginalia are longer, more frequent, and more carefully written than was ever the case in book rolls or in codices of con­ ventional design, and their margins are distinctly wider than in I G. ZUNTZ, Die Aristophanes-Scholien der Papyri (Berlin 1975; hereafter 'Zuntz, Aristophanes-Scholien') Nachwort, with the argument enlarged at An Inquiry into the Transmission of the Plays of Euripides [Cambridge 1965: 'Zuntz, Inquiry') 272-75; N. G. WILSON, "A Chapter in the History of Annotation," CQ N.S. 17 (1967: 'Wilson, ·Chapter"') 244-56; with L. D. Reynolds, Scribes and Scholars] (Oxford 1991) 46; Scholars of Byzantium (Baltimore 1983: 'Scholars') 33. See also J. Zetzel, ·On the History of the Latin Scholia," HSCP79 (1975) 335-54. 2 Where possible, I refer to papyri by the numbers assigned them in the cata­ logues of Pack and Uebel: R. A. Pack, The Greek and Latin Literary Texts from Graeco-Roman Egypt2 (Ann Arbor 1965); F. Uebel, APF21 (1971) 167ff. 399 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 400 THE DEVELOPMENT OF SCHOLIA typical codices that lack marginalia. 3 More than three-quarters of all heavily annotated papyrus codices are in this group. By contrast, only about a third of all heavily annotated book rolls have wide margins. Book rolls with broad margins survive in good numbers, to be sure; but here a lavish margin was the mark of a deluxe production: it was not intended to receive copious marginalia, and normally it did not. o4 Parchment codices, whatever the provenance, have dimensions and layout that differ from those typical among papyrus codices. 5 Yet among annotated parchment codices also, a correlation between mar­ ginal breadth and the density of annotation seems to develop in the fourth century.6 Taken altogether, the evidence of the Greek and Latin codices of late antiquity suggests that it is time to revisit the question of how scholia developed. I should say at once that most annotations in the literary texts of the group were aimed at a scholastic, not a scholarly reader­ ship. The learned elements of ancient commentaries, detectable often enough in scholia, are missing here: we find predominant­ ly paraphrases, glosses, and elementary mythological informa­ tion.7 This would seem to direct the search for the 'invention' of 3 I consider a marginal note 'long' if it is at least eight words long; notes are 'frequent' if three or more have been added beside a passage of text extending for at least fifteen lines; a 5-cm. margin in a papyrus codex is 'wide'. These are the admittedly arbitrary standards I have used, with slight modification, in dealing with rolls: • Annotations in Book Rolls of Greek Literature, D to appear in the proceedings of a conference on the Ancient Book held at the University of Minnesota, September 1992. 4 W. Johnson, The Literary Papyrus Roll: Formats and Conventions. An Analysis of the Evidence from Oxyrhynchus (diss.Yale 1992). Exceptional cases are P.Oxy. V 841 (Pindar, Paeans, Pack2 1361) and P.Schub. 10 (Iamb. Eleg. Gr. I Hippon. 148b, Pack2 1840), papyrus rolls in which the margins, incolumnia, and interlineations were deliberately left broad so that annotations could be added. 5 E. G. Turner, The Typology of the Early Codex (Philadelphia 1976) 30-33. 6 Among the twenty-two annotated parchment codices listed in the catalogues of Pack and Uebel or published subsequently, six have margins noticeably broader than usual in parchment books (they measure 4 cm. or more); and eight have markedly long or frequent annotations. 7 In the Vienna fragment of Pind. Pyth. 2 (Pack2 1365), for example, sub­ stantive material that appears in the scholia and treats history, biography, and geography is altogether missing. The annotator restricts himself to myth: K. McNamee, ·School Notes," Proceedings of the XX International Congress of Papyrology (Copenhagen 1993) 177-84. Annotations in the Antinoe Theoc­ ritus (A. S. Hunt and J. Johnson, Two Theocritus Papyri [London 1930], Pack 2 1487), though abundant, are also intermittent and fall short of the rela- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. KATHLEEN McNAMEE 401 scholia away from late antiquity and back again, with Zuntz, to the ninth century, when marginal commentary becomes sud­ denly dense and often learned. Two exceptional papyri of ear­ lier centuries, however, demonstrate that the scantiness of the papyrus evidence and its haphazard preservation are decep­ tive. The first of these is the Oxyrhynchus Callimachus, P.Oxy. XX 2258, which even Zuntz acknowledged as a "missing link'" between books of ancient format and scholiastic manuscripts. 8 Its marginalia are uniquely long, varied, thorough, and learned. The scribe's objective, to judge from the suffocating quantity of notes he added, was to squeeze in as much material as the mar­ gins would hold. Most of the manuscript contains fragments of lost poems, so comparison is possible only for the meager scholia on Callimachus' Hymns. Yet the papyrus' comments on the Hymns outnumber scholiastic comments on the same lines by a factor of four to one, an atypical proportion. 9 Indeed, the notes of the papyrus (in the left column below) occasionally provide information that is more specific or more expansive than that offered by the scholia (on the right): 3.84 1l0V10V OOK~: ot Ko.n]pOl t01K[~ VEIlOIlEVOl' I o.1t(XYEMi:~OV't] ~ 1 yap o.AA:Tt [Amv . 1l0VlOV 'to K:a'ta 1l0va<; VEIlO­ IlEVOV. Oo.KO<; of: 'to 91)piov. 6.15 'tpi.<; 0' tnl KaUlxoPpT)'ti: ]e . ~ [ I 8EOllO ]q>[ 0 ]pia 0 [ I KaAAlXOPOV q>peap tKaA.t:~'to tv KaA]Aixopov q>[peap' r ?ita~o]a 'EAEuo~vl (EO'tl of: Kai. oi1ll~ ~T)'toUcra[ • . 'AUlK11<;). tively high standard of Theocritean scholia. The comments in the Berlin Aratus (P.BeroL inv. 5865, M. Maehler, APF 27 [1980] 19-32, Pack2 119) are a somewhat better match for scholia in coverage and content. The annotator was selective in what he recorded, however, and his notes would have served well in the classroom. The twenty comments that survive omit astronomical points covered in the scholia, except those dealing with the myths of constel­ lations. Even the scholia's occasional vague invocations of literary parallels to popular authors (e.g. w<; Eupm{oTJ<;. w<; EupmioTJ<; ~:~$ 1 OlOV 1vc.O~ 1'1. v . 'tov 1 ~EV'tOl lCa'tayy£A9& l'ta lCOAaC£o9m' d ~EV l'tOl 5uo i1~ap'to~ ~~[ap]'t1J~a'tC~v, on 1 [E]~pl'l.V K~:~~l'l.Vuoa\' ~OUAEl 'It ro~av £ [ 1 'ltpay~a'tEU£Oeal; £:yo1 ~£v 't[o] oov dlCEval .. \ . . ..... . .. .. . .. Eoo~al EIKroV ou KroA.UEl 5£ o£ 'tUX£tV ou(y)yvro~l1C; b VO~OC; 1 roc; ..... ( .. ."." I IOX:n~lcp. b., ,,0 .uv e~ 'B]cpuP1lC;· o/twv 'til> IOO\~lCP· 'EqnJpa 5£ " Kopw- 9(oc;)· OEA,WOv'tal. Pf. 384.22 (Sosibiau Nike): a "ot~ cixo ne1.p1\v1l~ fiyay(ev) 'A[Pyo]A1.1(a): a1to KOPlV­ eou 1(at 'Io9~[oJ\). " y[a.Jp nE\PI1Vll1(P"Vll Ev KOPlV9cp [ I] .... v . 1( .. ov ayrova 'P(llow) £1ta1(oA,ou~1([£]val 'ta. N[EJ~£a. de; a 1(at 0 [£JXlVl1(oc; yEypa1t'ta\ [ I ~ov NE~Ea1(OV aymva 5llA.ol. b.'tOle; 'I(o)9~lovl1(o"\~l[OU VlK'f1e; a1(OlJOOXJlV oi. (supp!. Lobel) I 1toppm Oi1(OUV'tEe; £1tt 'tC[> Klvuqn ~il ~ovov oi. Ev 'AA.E~av5pd~. KlV[U'I/ I . voc; 'tile; ap . ~mc; oplCmv ~v KapX'T1[5]ovlmv xropav. ton 5£ 1(at) 1tOA,tc;. [ 5\o'tE'P£a 5£ 5ux'to 5tc; vuciloa\ 'tOY I MEA,lIC£pl'tU 'ti9E'ta\ 'ta. I "Ioe~w. £1t1. 5£ 'til> 'APXE~OPCP 'ta. N£~Ea. b. 'til> MEA,[ \JICEp'tU. That the scribe included near-identical lemmata in each of the first two pairs of notes proves nothing in particular. One ele­ ment of each fair is written in a side margin, one at the bottom of the page. I the annotator got these notes from a single com­ mentary he might understandably have copied twice, for the sake of clarity, all or part of a single introductory tag in his source. Alternatively, his source may have provided him these four distinct notes. These he dutifully copied out in their en­ tirety, omitting nothing that he found under his nose, as scribes Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. KATHLEEN McNAMEE 409 often did~ven though this meant duplicating all or part of a lemma. (This automatic behavior has been characterized as a .. scribal act," as distinguished from "'creative acts," which in­ volve conscious thought on the part of the scribe.) 22 Significant in this papyrus, however, is not the repetition of lemmata, but certain duplicate information provided in the notes. In lines 65ff of the Coma, each of two surviving notes con­ veys roughly the same information-the time of year when the constellation is visible in the sky-but they do so in astronomi­ cally different terms. The Lock rises, according to the first com­ ment, before the winter solstice and it sets after the spring equinox. The second note, more precisely worded, extends the period at each end: it reports that the constellation rises at the fall equinox and sets at the summer solstice. In pairs of notes on the Victory of Sosibius we also find differently worded but essentially congruent statements about the relative timing of the Isthmian and the Nemean games (Pf. 384.22). The two notes on 384.23-24, further, each identify the location of the Kinyps River. Both are essentially in agreement, although one is more precise and ample. It reports that the Kinyps bounds the land of the Carthaginians and is also the name of a city. The second makes the simple statement that the Kinyps is a river of Libya. Degree of amplification also distinguishes the notes on Pf. 384.4 and 384.25. In each case one element of the pair of notes is a simple gloss, the other the gloss with further explanation. In each of the five cases, the likelihood that the annotator com­ piled his material from divergent sources is strong, given the repetitive nature of the comments. Wording varies here more than content. Ancient hypomnemata have their organizational deficiences, to be sure, but so much redundancy in so little space is not typical of their faults. Even in the absence of an obvious source-marker like liAAWC;, the substantive repetitions of information in these five pairs of notes must mean that they were compiled from at least two sources. Beyond papyri, evidence accumulates for the practice in late antiquity of preserving excerpts of multiple commentaries in the margins of the texts they explain. This is the import, as Alan Cameron has shown, of certain observations by the early sixth- 22 C. E. Murgia, quoted by Zetzel (supra n.t) n.32. The phenomenon is ob­ servable, for example, in the scholia in the Bembine Terence, where the second annotator took no steps to avoid repeating information entered by the first. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 410 THE DEVELOPMENT OF SCHOLIA century mathematician Eutocius on the nature of his own work.23 Eutocius' preface describes the accessibility of explana­ tory material accompanying his edition of Book 4 of the Conica of Apollonius of Perga. In Cameron's translation: "'Book IV, my dear friend Anthemius ... is particularly elegant and clear for readers. thanks to my edition with its commentary: the notes make up for anything left unsaid." 24 The "commentary" to which he refers are axoAuI. a word used apparently interchan­ geably by Eutocius with at 1tapaypacpai, the "notes." These notes appear. from context and from the term used. to have been written in margins beside the text. In putting together this work. moreover, Eutocius says that he worked from multiple sources, chose the best text where he had a choice. and put alternatives "outside" the text, in what he seems to describe as .. concatenated· scholia: £~{J)eEV tV 'tOLe; auv'tE'taYI!£VOle; axoAiotc;. £~{J)e£v is admittedly ambiguous. but it seems likely that it denotes the marginal area outside the space occupied by the main text. From beyond Egypt, furthermore. as we saw above, come examples of parchment texts prepared to receive. and receiving. large quantities of scholia compiled from multiple sources. The margins of the fifth-century Verona Vergil contain voluminous scholia by two different hands. both dated to the fifth century. The Bembine Terence of the fourth century has copious anno­ tations in two different sixth-century hands. The famous Vienna manuscript of Dioscorides. probably copied in Constantinople in the early sixth century, is another. Carefully labelled excerpts from Galen and Crateuas in the vast empty spaces at the bottom of its oversize pages come very close to the definition of scholia that governs this discussion: they are comprehensive. learned. and compiled from multiple sources. They do not surround the text in the manner of many later scholia. Yet they constitute another missing link in their history.25 Wilson found evidence for compilation in yet another source. the Latin scholia to Vergil's Eclogues by one Philargyrius. who may have 21 A. Cameron, -Isidore of Miletus and Hypatia: On the Editing of Mathematical Texts," GRRS 31 (1990) 103-27. 24 Eon 5£ Xap{EV Kal oa .., ::c t"'" tTl t'Tl Z ~ (') Z > ~ t'Tl t'Tl ~ ::: ~ Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 414 THE DEVELOPMENT OF SCHOLIA uation is correct (Scholars 36.): "When all the facts are taken into account, the balance of evidence entitles us to revert to the view that scholia began to be amalgamated from the earlier monographs and to take on their mediaeval shape during the late Roman empire. It is possible that the process had started by the end of the fifth century and that Procopius was taking note of it when he invented the catena." Indeed, if the evidence and date of Philargyrius are reliable, or if two of the legal codices in question are accurately dated to the turn of the fourth century, then the process may have begun even earlier, possibly as early as 400.30 The question of the origin of scholia comes to this: is it reasonable to imagine that some readers of late antiquity conceived of filling the broad margins of their books with ex­ tensive exegetic material compiled from more than one source? The evidence of the very heavily annotated Oxyrhynchus Callimachus in particular evokes an affirmative answer. Given the limitations of this evidence, however, one wishes all the more to know what scriptoria in contemporary Constantinople and Rome and Gaul (the possible provenance of the Verona Vergil) were producing. For the drastic changes in the form of books and in the quantity and form of their annotation in late antiquity suggests strongly that if we had a window on the past through which we might view scribes at work in major centers of culture in the fifth century, we would indeed find some engaged in transcribing full commentaries, and possibly compilations of commentaries, into wide margins: in a word, prod ucing scholia. 31 WAYNE STATE UNIVERSITY July, 1996 30 Legal papyri dated to the 4th /5U, C.: P.Ryl. III 476 (CLA 2.225, Pack 2 2282); P. Vindob.Lat. 110 ined. (Pack 2 2984, Seider 2.2.38); P.S.I. XIII 1348 (Pack 2 2982). 31 A version of this paper was first read at the 1995 Byzantine Studies Conference. I am grateful to members of that audience, particularly Alan Cameron, for their comments.