Text as Paratext: Pindar, Sappho, and Alexandrian Editions Enrico Emanuele Prodi W HAT LITTLE SURVIVES of the archaic Greek lyricists has come down to us as bare text, shorn of music, dance, location, ambience, occasion, ceremony.1 Our texts ultimately go back to Alexandria and the late third century B.C., when the scholars of the Museum compiled what were to become the canonical editions of those poets; and what those editions preserved and enabled to circulate anew throughout the Greek-speaking world were written words alone. But that from sung spectacle to written text, from body and voice to papyrus and ink, was not the only change of state to which lyric poetry was subjected between the archaic and the Hellenistic age. Another, equally momentous transforma- tion took place: individual compositions which were originally independent of, and unrelated to, one another became joined together in a fixed sequence as constituents of a larger unit, the book.2 Lyric was not the only kind of poetry that was affected by this 1 Fragments of Pindar are cited from Snell-Maehler, fragments of Sappho and Alcaeus from Voigt. All translations are my own. 2 G. O. Hutchinson, “Doing Things with Books,” Talking Books: Readings in Hellenistic and Roman Books of Poetry (Oxford 2008) 1–2, cf. 4–15. On ancient poetry books see also J. van Sickle, “The Book-Roll and Some Conventions of the Poetic Book,” Arethusa 13 (1980) 5–42. The interrelation between Pindaric song and the materiality of the book is now the subject of T. Phillips, Pindar’s Library: Performance Poetry and Material Texts (Oxford 2016), a volume I was regrettably unable to consult until rather late in the composition of the present article. ————— Greek, Roman, and Byzantine Studies 57 (2017) 547–582 2017 Enrico Emanuele Prodi 548 TEXT AS PARATEXT process. A case in point are epigrams, which about the same time were being gathered in purposeful, artistically arranged collections that became one of the main vehicles—or indeed the main vehicle—for their transmission.3 One recently published example is P.Vindob. G 40611 (late third century B.C.), a list of opening lines of epigrams which may have been preliminary to the compilation of such a collection.4 The Posidippus papyrus too, P.Mil.Vogl. 309 (also late third century B.C.), although seemingly representing the work of a single author,5 brings together epigrams that must have been written at different times and is arranged according to principles that resemble those operative in multi-author collections—and in the editions of the lyricists.6 Ptolemaic papyri also preserve a great deal of non-epigrammatic anthologies.7 Indeed, as Luigi Enrico Rossi 3 See A. Cameron, The Greek Anthology from Meleager to Planudes (Oxford 1991), esp. 3–12, 19–33; K. J. Gutzwiller, Poetic Garlands: Hellenistic Epigrams in Context (Berkeley 1998); L. Argentieri, “Epigramma e libro. Morfologia delle raccolte epigrammatiche premeleagree,” ZPE 121 (1998) 1–20. 4 P. J. Parsons, H. Maehler, and F. Maltomini, CPR XXXIII (2015), esp. pp.10–12, including a discussion of previously published similar lists. 5 Dissenting voices are H. Lloyd-Jones, “All by Posidippus?” in The Further Academic Papers of Sir Hugh Lloyd-Jones (Oxford 2005) 246–249, first published in D. Accorinti and P. Chuvin (eds.), Des géants à Dionysos: Mélanges de mythologie et poésie grecques offerts à Francis Vian (Alessandria 2003) 277–280; S. Schröder, “Skeptische Überlegungen zum Mailänder Epigrammpapyrus,” ZPE 148 (2004) 29–73; F. Ferrari, “Posidippo, il papiro di Milano e l’enig- ma del soros,” in J. Frösén et al. (eds.), Proceedings of the 24th Congress of Papyrology (Helsinki 2007) 331–339. 6 Posidippus papyrus as a poetry book: K. J. Gutzwiller, “A New Hel- lenistic Poetry Book: P.Mil.Vogl. VIII 309,” in B. Acosta-Hughes et al. (eds.), Labored in Papyrus Leaves: Perspectives on an Epigram Collection Attributed to Posidippus (Washington 2004) 84–93, and “The Literariness of the Milan Papyrus or, ‘What Difference a Book?’” in The New Posidippus: A Hellenistic Poetry Book (Oxford 2005) 287–319. Organization of Hellenistic collections of earlier lyric: A. Dale, “The Green Papyrus of Sappho and the Order of Poems in the Alexandrian Edition,” ZPE 196 (2015) 26–30. 7 Now collected and re-edited by F. Pordomingo, Antologías griegas de época helenística en papiro (Florence 2013). ————— Greek, Roman, and Byzantine Studies 57 (2017) 547–582 ENRICO EMANUELE PRODI 549 suggested, in the Hellenistic period “the anthology” (or, better, “the collection”) should be recognised as a genre in its own right.8 These collections ranged from informal, private com- pilations to works that were evidently meant to enter the book market and circulate, with a clear identity of their own, some- times under the compiler’s name.9 The canonical Alexandrian editions of the lyricists are evidently closer to the latter pole, but they have nonetheless a clearly distinct status. They were intended as authoritative, definitive, and—crucially—complete texts of the authors whose poetry they contained. They do not seem to have circulated under their editor’s name: of the nine ‘canonical’ lyricist, only one is explicitly associated with an editor in the ancient sources—Pindar, with Aristophanes of Byzantium—and in few sources at that.10 The organising principles of many of them seem to have been mechanical rather than artistic. Et cetera. The intriguing common ground of Hellenistic collection-making should not obscure the variety of the phenomenon. The present article focuses on one consequence of the com- bination of several lyric poems into one book: the role that the opening poem has to play in the economy of the book that it opens. We are all used to finding an introductory poem of some kind at the opening of a poetry book: perhaps a dedica- tion, a more or less overt self-presentation, a declaration of the contents of the book, or an anticipation of one or more key 8 Letteratura greca (Florence 1995) 635–636. 9 A case in point is Meleager’s Garland, which the introductory elegy calls a πάγκαρπον ἀοιδάν, 1.1 Gow-Page (Anth.Gr. 4.1.1). “The singular noun indicates that the Garland was conceived not as a mere anthology … but rather as an aesthetic whole, a poetry book”: M. S. Santirocco, “Horace’s Odes and the Ancient Poetry Book,” Arethusa 13 (1980) 48. 10 Vita Vaticana (I 7 Drachmann), P.Oxy. 2438 col. ii.35–36 (partly restored but certain). Dion. Hal. Comp. 22.17, 26.14 (II 102, 140 Usener-Rader- macher) does not explicitly attribute the edition to Aristophanes but shows that his name was the first that came to mind when thinking about Pindar as an edited text. ————— Greek, Roman, and Byzantine Studies 57 (2017) 547–582 550 TEXT AS PARATEXT themes within it, whether it be Meleager’s Garland or Baude- laire’s Fleurs du Mal, Petrarch’s Rerum Vulgarium Fragmenta—or Heaney’s Death of a Naturalist; something that leads the reader into the book at the same time as it tells, in some way, about the book; a preface to the text embedded within the text itself; simultaneously text and, in Gérard Genette’s terminology, “paratext.”11 Of course no archaic Greek lyricist really wrote anything of the kind; partly because, with all likelihood, they never envisaged a reading public being presented with a set of their poems in a fixed sequence at all.12 However, some books of ancient lyric opened with poems that can be seen to be in- vested with a similar paratextual function. After all, Hellenistic and Roman readers were used to seeing poetry books of their own time begin with such introductory poems, just as they were used to seeing longer poems begin with a proem fulfilling a similar function.13 It may have come naturally to expect pro- oemial overtones in the first poem in a book, and therefore to read them into it regardless of the poem’s original purpose. In the case of the archaic lyricists, any such prooemial over- tones arose from a later, non-authorial arrangement. Any in- 11 G. Genette, Paratexts: Thresholds of Interpretation (Cambridge 1997). In his definition, the paratext consists of “a certain number of verbal or other pro- ductions, such as an author’s name, a title, a preface, illustrations” which “surround [the text] and extend it, precisely in order to present it, in the usual sense of this verb but also in the strongest sense: to make present, to ensure the text’s presence in the world, its ‘reception’ and consumption in the form (nowadays, at least) of a book … the paratext is what enables a text to be- come a book and to be offered as such to its readers” (1).� 12 Compare how the arguable paratextual overtones in the first poem of Corinna’s ϝεροῖα (PMG 655) have been taken as evidence of a Hellenistic date: M. L. West, “Corinna,” CQ 20 (1970) 203–204; “Dating Corinna,” CQ 40 (1990) 553–554. The case of the ostensibly paratextual “seal elegy” that (almost) opens the corpus Theognideum—which accordingly R. Reitzen- stein, Epigramm und Skolion (Giessen 1893) 267, judged to be “das älteste nachweisbar vom Autor selbst edierte Buch”—is too different, and too com- plex, to be dealt with properly here. 13 Gutzwiller, Garlands 10; see also J. van Sickle, “Poetics of Opening and Closure in Meleager, Catullus, and Gallus,” CW 75 (1981) 65–75. ————— Greek, Roman, and Byzantine Studies 57 (2017) 547–582 ENRICO EMANUELE PRODI 551 troductory function was bestowed on the opening poem ex post facto by the editor who opted to place it before the others. When there are grounds for taking a poem’s position as a spe- cific choice on the editor’s part, such choice implies a statement that this particular poem is a fitting opening to the book—good to think about the book with, as it were. Thus an editor’s reasons for choosing a given poem as a preface for the respec- tive book invite exploration. So does the very act of extrapo- lating one poem to introduce the collection to which it belongs, repurposing and paratextualizing one of the author’s own poems retrospectively. This act is an implicit editorial state- ment both on the book and on the opening poem. Like all editorial activity, it both responds to the text and intervenes in it, seeking to orient subsequent responses to it. Readings of the book will be variously influenced by the suggestions that reso- nate from the opening poem; in turn, the opening poem will find itself charged with implications broader and more complex than it would have otherwise. There are two poets for whom we are well placed to attempt an inquiry along these lines: Sappho and Pindar. Olympian 1 and Sappho fr.1 are well studied in their own right, but a read- ing of these poems from the vantage point of the book has much to give.14 In the case of Olympian 1, we are aided by the complete survival of the Olympians as a book and of (most of) the rest of the Epinicians, of which the Olympians constituted the first book. We are not equally lucky with Sappho, but we have enough fragments of Book 1 to be able to form of an idea of its contents and of its internal organization. These two poems will therefore take up the first and last sections of the present article. Much less is known about the fragmentary books in the Pindaric corpus, but we have some information about the opening poems of some of them (the Hymns and, to a much lesser degree, the Dithyrambs and the Partheneia), and speculation 14 Phillips, Library 122–142, offers a reading of Ol. 1 in its editorial context which usefully complements the one given here. ————— Greek, Roman, and Byzantine Studies 57 (2017) 547–582 552 TEXT AS PARATEXT based on this information will be the subject of the middle sections.15 All of our enquiry will inevitably be speculative. We cannot read the mind of Pindar’s ancient readers or Sappho’s ancient editor. We do have one detailed ancient reading of Sappho fr.1 in Dionysius of Halicarnassus’ On the Arrangement of Words 23.12–17 (II 116–117 Usener-Radermacher), but its concerns are only style and sound, and thus it is of no great help to our undertaking. We likewise have a corpus of scholia to Pindar, of which some use will be made in the section on Olympian 1, but they are mostly silent on the topic that concerns us. Another difficulty is the very crux of the issue, namely the superposition in our material of several layers of agency—the poet’s, the editor’s, the reader’s—the combination of which produces the effect we are seeking to investigate. On the one hand, these layers are clearly distinct. The composition of the individual poems and their being gathered together into books are separate act performed by different people at very much different times; contrast the poet-editors of the Hellenistic and Roman era, who not only collect and arrange their poems themselves but often compose them with an eye to the resulting collection from the start.16 On the other hand, these layers of agency cannot be cleanly separated. An editor constructs hierarchies, adds emphases, shifts perspectives, builds (and severs) connections, but can do all this only to the extent that the text permits. Likewise, a reader can take up these hierar- chies, perspectives, meanings, etc.—in a way that is, of course, not necessarily subordinated to the author’s or the editor’s di- rections—only to the extent that they can be inferred from, or mapped onto, the authored-edited text. But reading can be complicated by countless factors: social 15 On the division of Pindar’s poetry into books and their respective or- ganization see S. Schröder, Geschichte und Theorie der Gattung Paian (Stuttgart/ Leipzig 1999) 136–149. 16 G. O. Hutchinson, “Propertius and the Unity of the Book,” JRS 74 (1984) 99. ————— Greek, Roman, and Byzantine Studies 57 (2017) 547–582 ENRICO EMANUELE PRODI 553 and cultural priming, preconceptions or existing knowledge about the author, guidance by a teacher or commentator… Nor was the reading of such editions the only kind of reception that is relevant to our subject: the editions compiled in third- century Alexandria were preceded by centuries of multifarious transmission, in both oral and written form, about which we know very little indeed, but which may have influenced the editorial choices that produced the editions we do know.17 The best one can offer, then, is judicious (hopefully) speculation on what the editor may have meant, what a reader may have understood. Pindar: the Olympians The significance of Olympian 1 being the first of the Olympians was already noted by the anonymous author of the Vita Vati- cana, or Thomana (I 7 D.):18 ὁ δὲ ἐπινίκιος οὗ ἡ ἀρχή· Ἄριστον µὲν ὕδωρ προτέτακται ὑπὸ Ἀριστοφάνους τοῦ συντάξαντος τὰ Πινδαρικὰ διὰ τὸ περιέχειν τοῦ ἀγῶνος ἐγκώµιον καὶ τὰ περὶ τοῦ Πέλοπος, ὃς πρῶτος ἐν Ἤλιδι ἠγωνίσατο. The victory ode that begins “Best is water” has been placed first by Aristophanes, who arranged Pindar’s works, because it con- tains a panegyric of the games and the tale of Pelops, who was the first to compete in Elis. Like so much in ancient Pindaric scholarship, this statement frames the matter in terms of causation rather than of effect; the focus is on the agent—Aristophanes of Byzantium—rather 17 Pindar before Alexandria: J. Irigoin, Histoire du texte de Pindare (Paris 1952) 8–28. Sappho before Alexandria: D. Yatromanolakis, Sappho in the Making: The Early Reception (Washington 2007). As concerns Pindar, there is no particular evidence that such traditions did influence the Alexandrian edition: Ol. 1 does not loom large in the pre-Alexandrian record at all, and neither does the first Hymn if one excepts the fictional and quite possibly post-Alexandrian story of Pindar’s encounter with Corinna (Plut. De glor. Ath. 4, 347F–348A). As for Sappho, see 572 ff. below. 18 Commentary in C. Daude, S. David, M. Fartzoff, and C. Muckensturm-Poulle, Scholies à Pindare I (Besançon 2013) 135–150. ————— Greek, Roman, and Byzantine Studies 57 (2017) 547–582 554 TEXT AS PARATEXT than the recipient—the reader. But in this case its assertion of agency is to the point, as a further element shows. As Jean Irigoin remarks, each book of the Epinicians is organised following a recognisable order which is consistent across the four books.19 The criterion is a hierarchy of the disciplines in which the respective victories were achieved, from the chariot race and other equestrian competitions (keles, apene) to contact sports (pankration, wrestling, boxing), pentathlon, and finally foot races (race in armour, dolichos, diaulos, stadion).20 The only surviving victory ode that celebrates a victory in a musical contest, Pythian 12, was placed at the very end of its book. True, the order is not impeccable. Sometimes similar disciplines are mixed up a little: the Olympians have a sequence boxing (7)/wrestling (8–9)/boxing (10–11), the Nemeans even pankration (2–3)/wrestling (4)/pankration (5)/wrestling (6). Special circumstances can also be accommodated: Pythian 3 interrupts a sequence of odes for chariot victories in order to keep the three odes for Hieron of Syracuse together.21 Lastly, at the end of both Nemeans and Isthmians there are a few poems, respec- tively three and at least two (the Isthmians after the eighth are fragmentary), that are not actually Nemean or Isthmian odes: for the the Pythia in Sicyon (Nem. 9), the Heraia in Argos (10), and the election of a town official in Tenedos (11), 22 for the 19 Histoire 43–44. 20 The single book of Bacchylides’ Epinicians was arranged in a similar way, with the bulk of the odes organised by a combined hierarchy of disci- plines and games (victories in the crown games first, orderd by discipline and then by games, then victories in local games), with the first pair of poems out of sequence: J. Irigoin, Bacchylide: Dithyrambes, épinicies, fragments (Paris 1993) xxiv–xxv. There is an interesting difference in the order of disciplines, however—foot races (6–7, 10) come before contact sports (11– 13)—and an ode for a pentathlete interrupts the sequence for runners (9). 21 Irigoin, Histoire 44. 22 Inscr. Nem. 9 (III 150 D.), cf. inscr. a Nem. 11 (III 184–185 D.) = Didymus fr.62 Braswell. Incidentally, the common notion (starting from Irigoin, Histoire 40–41) that these extra poems were added here because the Nemeans were the last book of the Epinicians is severely problematized by the ————— Greek, Roman, and Byzantine Studies 57 (2017) 547–582 ENRICO EMANUELE PRODI 555 Hellotia in Corinth (Isthm. frr.6a(i)–(l)), and the Oschophoria in Athens (fr.6c).23 That they should be grouped at the back of the book is understandable: the primary ordering criterion is the games in which the victory was achieved, so poems commem- orating victories in other games must be placed out of sequence regardless of the discipline. And among these extra odes the normal order is respected: the last three Nemeans celebrate a chariot victory, then a wrestling victory, and lastly a political event with no direct link to sports. The ordering criterion is sometimes circumvented but never fatally undermined. Olympian 1, therefore, stands out conspicuously. It celebrates a victory with the keles although three odes for victories in the chariot race come next. Following the normal order it should come fourth or even as late as sixth (depending on where one places the apene [Ol. 4–6] relative to the keles), certainly not first. Despite not acknowledging this fact explicitly, the author of the Vita Vaticana (or his source) did recognise the position of Olym- pian 1 as significant. He represents this position as the result of a positive choice on Aristophanes’ part, a choice he takes to have been made on account of a positive reason—or indeed two: the panegyric of the games and the aetiological tale of Pelops. For us too the contravention of the normal order that governs the body of the book highlights the significance of the choice. Evidently, Olympian 1 is there not because an overarch- ing ordering principle resulted in its being there, but because it was specifically meant to be there. This prompts the double question of cause and effect. What may have prompted Aristophanes’ choice to place Olympian 1 first, and what may Olympian 1 tell the reader when so placed? The matter has been excellently explored by Monica Negri.24 ___ fact that there were extra poems at the end of the Isthmians too. 23 M. Negri, “L’oschophorikon di POxy 2451 B fr. 17.6 e la sua posizione nell’edizione di Pindaro,” ZPE 138 (2002) 35–36; G. B. D’Alessio, “The Lost Isthmian Odes of Pindar,” in P. Agócs et al. (eds.), Reading the Victory Ode (Cambridge 2012) 28–57. 24 M. Negri, Pindaro ad Alessandria: Le edizioni e gli editori (Brescia 2004). ————— Greek, Roman, and Byzantine Studies 57 (2017) 547–582 556 TEXT AS PARATEXT A first answer lies in the ode’s celebrated opening (1–7): Ἄριστον µὲν ὕδωρ, ὁ δὲ χρυσὸς αἰθόµενον πῦρ ἅτε διαπρέπει νυκτὶ µεγάνορος ἔξοχα πλούτου· εἰ δ’ ἄεθλα γαρύεν ἔλδεαι, φίλον ἦτορ, µηκέτ’ ἀελίου σκόπει ἄλλο θαλπνότερον ἐν ἁµέρᾳ φαεννὸν ἄστρον ἐρήµας δι’ αἰθέρος, µηδ’ Ὀλυµπίας ἀγῶνα φέρτερον αὐδάσοµεν Best is water; gold like fire that blazes in the night outshines wealth that makes men great; but if you wish to voice the games, my heart, look no further for another star warmer than the sun shining in the day across the empty sky, nor let us call a contest greater than Olympia. With its emphatic exaltation of the Olympic games above all others, this proem proclaims the hierarchy of the four Pan- hellenic games just as it is embodied in the four books of Pindar’s victory odes, with the Olympians at the start.25 Pindar’s own words are used to justify the architecture of the collection and the position of the Olympians within it. By placing the ode here, Aristophanes of Byzantium foregrounds an alleged Pin- daric rationale for his own editorial activity. He thus appears as the faithful but artful executor of Pindar’s will, the editor who understands the requirements of his material and presents the evidence for his decision up front while thereby signalling his own scholarship and editorial craft. Aristophanes’ skill is thus brought to the fore at the same time as it ostensibly recedes into the background by deferring to the higher authority repre- sented by the poet. To an extent, this confirms the first claim made by the Vita Vaticana, that Olympian 1 contains the praise (ἐκγώµιον) of the games; but it also qualifies it. What we have in Olympian 1 is an ἐγκώµιον of a very specific kind. Other odes too exalt the 25 Negri, Pindaro 37–38, 125–126. ————— Greek, Roman, and Byzantine Studies 57 (2017) 547–582