Jump to ContentJump to Main Navigation
BacchylidesPolitics, Performance, Poetic Tradition$

David Fearn

Print publication date: 2007

Print ISBN-13: 9780199215508

Published to Oxford Scholarship Online: February 2010

DOI: 10.1093/acprof:oso/9780199215508.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2017. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber: null; date: 25 February 2017



Oxford University Press

Here I offer a reconstructed text for Bacchylides 13.155–67. This is an attempt, based on a personal inspection of the papyrus in the British Library, to shed some further light on the lines beyond what has already been achieved by Maehler, who had taken advantage of Barrett’s unpublished notes reassessing the first of two fragments correctly relocated here by Blass (see Blass 104, nn.; Kenyon 126). I offer a diplomatic and interpretative transcript, an image of the reconstruction, a translation, and a full textual and interpretative commentary.

Diplomatic Transcript

  • […]εγ’ ημιθεοιc
  • […]̣ϊι̣θεωνγ’ιορμα***
  • […]ρ̣ονε*** ημεγαλαιcινελπιcιν
  • […]οντεcυπερφ[…]λον
  • Χ̣
  • […]υ̣[̣̇]ν̣[…]cιππευταικυανώπιδαcεκ
  • [  space for±16 letters  ]νεαc
  • [……]ρ̣ο̣ιcεξεινθ̣[…]ματονπογι***
  • [̣]ᾶνταφοινιξει̣[̣̣]αμανδρ̣[
  • [̣]ναcκοντεcυπ̣[…]κιδαιc
  • ερειΨ[̣]π̣ύ̣[

Interpretative Transcript

  • [ὀξὺ μ]εγ’ ἡμιθέοις      155
  • [1F30;θει̂α]ν̣ ἰσοθέωθ δι’ ὁρμάν̣
  • < – >
  • [ἆ δύσφ]ρ̣ονες ὑπερφ[ία]λον
  • [θρόησαν α]ὐ̣χ̣[ὰ]ν̣


  • Т̣[ρὦε]ς ἱππευταὶ κυανώπιδας ἐκ-      160
  • [πέρσασιν Ἀσάσειν] χ[ο]ρ[ὸ]ν̣ [εἰλα]πίνας τ’ ἐν
  • [λαοφό]ροις ἕξειν θ[εόδ]ματον πόλιν̣
  • [μ]έλλον ὴρα πρότε̣[ρο]ν δι-
  • [ν]ᾶντα φοινιξειν [Σκ]αμανδρ[ον,      165
  • <)–>
  • [θ]νάισκοντες ὑπ[’ Aἰα]κίδαις
  • ἐρειψ[ι]π̣ύ̣[ρλοι***]

Figure 6 Image of Bacch. 13.155–69 on papyrus (BM inv. no. 733 col. 27), together with a drawing of supplementation for lines 155–67.

Image of papyrus by permission of the British Library (Pap. 733 folio 3 (right), order number 21374).

Translation (155–63)

…, a mightily keen, bitter contest for the heroes though the direct assault of those godlike men. Ah, what fools! High-spirited in their great hopes those Trojan horsemen arrogantly boasted that, once they had utterly destroyed the dark-prowed Argive ships, their god-built city would set up a chorus and would hold feasting in its streets in their honour.

Textual Notes

155 18 mm of space from projected left-hand margin for 4 letters.

156 19.5 mm of space for 5 letters.], vertical ink trace at two-thirds height (p.353) just below left-hand dot of following trema, 1 mm to the left, on very edge of papyrus, folded back and overhanging, so no chance of other ink remaining. This could be part of a right-hand vertical (η, ι, µ, ν, π), the far r.-h. end of a mid-height crossbar, i.e. c (pace Barrett), γ, or ψ, or very right-hand edge of ρ, φ, or ω. The trace is too low for a letter-tops crossbar, therefore ruling out τ, and if it were anything else we would have expected more ink to the lower right to be visible, therefore ruling out e.g. α.

cο̣θ, as Jebb and Kenyon rightly say, an omicron has been written over an erasure. δ'ιορµαν, it seems right that A 2 or A 3 inserted an apostrophe and put a deletion stroke through the ι, and the gap between ι and ο was also filled with another stroke. It is possible that A originally intended to write δια but felt some uncertainty (did he find his exemplar difficult to read here?) and left a space instead of writing α. 157 Overlap here with small fragment added by Blass. Projected 19 mm space for 5 letters. ]ρ, trace of bottom of descender. No ink at top (since papyrus breaks off), so it is impossible to say whether originally there was any ink above or to the right. Maehler, on this basis, suggested]ϊ as an alternative possibility, but this does not account for the extension below the bottom line. 158 Projected 15.5 mm space for 4 letters. It is difficult to see why Barrett could expect to be so accurate with tracing that he thought πνε(ι)] οντεs would use up the space better than πνεί] οντες. Without the iota, π, ν, and ε would have to be at their broadest to fill up the available space.

At the bottom of the first fragment incorporated by Blass, beneath τεc in the line above, there appear very faint remnants of ink between the lines, followed by the blotchy remains of one letter. There are faint spidery traces of two (or three?) letters, so faint and thin as to be hardly visible on a photograph of the papyrus; even under strong magnification on a view of the original, they seem very hard to interpret. They might be said to resemble CΜ in form, but if so, the letter forms do not correspond with any other supralinear additions elsewhere in the papyrus. Certainly, they do not seem compatible enough, in stroke-shape or thickness, to represent letters in the same hand as the following trace or the hands of the main scribe or those of the subsequent editors. It is unclear whether these initial traces are to be interpreted as original ink at all, and it seems that they are best regarded as accidental, since the ink traces only remain on the very top of the top surface of the papyrus; they are quite clearly different from the following trace. The final trace seems more blocky and thick in appearance than letters written in the hand of A, and this may suggest that it had been corrected, written over, or even erased, but perhaps just damaged. But it seems best interpreted as χ; though the surface is damaged, and some of the ink seems anomalous, on the left-hand side the two obliques characteristic (p.354) of χ can be made out. Indeed, it seems to correspond well with the form of the supralinear χ at col. 15(11).9 (Bacch. 5.164), by A 2 to correct κρη to χρη. There is no way of telling whether the letter beneath was cancelled or not.

159+160 Projected 24 mm space for 7 or 8 letters.̣]υ̣δ̣[̣]ν̣τ̣[̣̣̣], beneath the interlinear traces, at the break-off of the papyrus, there appear to be two traces level with letter-tops, separated by 4 mm. Between these there are two minute specks. These two traces correspond to Barrett’s ]α̣υ̣[. The first trace is a speck ±1 mm wide, and could be the remnants of either the apex of α or δ, or the tip of the upper left-hand oblique of any tall letter which falls away to the lower right. If it were ν or µ we would possibly expect there to be a further upper extension of the vertical beyond the join of the oblique. Therefore I propose υ, the trace corresponding to the uppermost tip of the left-hand part of the bowl. The two tiny specks before the next trace could be then interpreted as ink from the other side of the bowl of this υ. The second trace, which Barrett takes as υ̣, is slightly larger, about 1 mm lower in the line than the previous trace, preserved on a small piece of papyrus hanging off to the bottom. The trace appears to be angled more overtly in the direction of the lower right. I believe it to be perhaps a little low for υ, and therefore suggest δ̣. If this is the case, it could represent part of the small extension to the upper left often seen at the apex of this letter. Then, after a gap of 9 mm (space for 1 or 2 letters), there again appear letter-top ink traces, beneath the π and ε of the line above. The first is a horizontal hairline trace, below and to the left of which the top surface is missing, followed, after a gap of 1.5 mm, by a blob of ink, and then, after a 1 mm gap, a group of traces 3 mm long level with letter-tops. The first of these appears as a 1 mm long horizontal stroke; the rest are minute dots. This group of traces would be compatible with the cross-bar of τ. Also possible would be π, but often the l.-h. descender projects above the cross-bar, and there is no evidence of this here. Certainly Barrett’s τ̣ seems very likely, forming the start of Τ̣[ρω̑>ε] S, but ν̣ is very much more uncertain. The trace described above as a blob could certainly be interpreted as the top of a descender; in following Barrett and reading this as the r.-h. descender of ν, it is important that the trace 1.5 mm before this be taken into account. If we read this as part of the same ν, the spacing would mean that this would have to be part of the oblique of ν. Although some concern may be felt that the trace is only slightly lower than the top of the next trace, it does seem that there may be just enough space to allow for it indeed to represent part of the l.-h. part of an oblique of ν. After the τ there is enough space for three letters before the remnants of the c.

161 Projected 57 mm of space, enough for about 15 or 16 letters. (p.355) 162 Overlap with Blass’ second fragment, the top-left of which has surface missing. Projected 30 mm of space, enough for about 7 or 8 letters. ]χ̣[̣]ρ[̣]̣[, above ξ, r.-h. descender of ν, and ι, in εξειν in line below, three small traces of ink (the second 7.5 mm from the first, the third 5.5 mm from the second), all at low-line level. The first is slightly rounded and pointing toward lower left, compatible with bottom l.-h. tip of α or χ (not δ: no ink to the right necessary for horizontal stroke of δ). The second and third are mere specks, most plausibly bottom tips of verticals; the second is lowest of three and angled slightly down to the left, thus compatible with tip of descender of ρ. There is enough space between first and second to suggest a missing letter in between. The last trace is roughly 13 mm before start of π of πιναc, enough room for about 4 letters. Barrett offered π] ά̣ [λ]ι̣ν̣. I would suggest, however, that the trace which Barrett interprets as ι is too low in the line: this I take as the ρ. It is also possible for a small letter like ο or c, that would sit higher in the line, to fit in the gap between the second and third traces here; hence my ρ[ ̣ as opposed to Barrett’s ··. Furthermore, I interpret Barrett’s reading of the third trace as the base of the left-hand descender of ν. This fits the space slightly better than Barrett’s reading of it as the right-hand descender, and suits the projected space before] πιν. Furthermore, on this analysis the spacing for the χ, ρ, and ν matches that for the χ, ρ, and ι in χοροί in Bacch. 14.14, col. 35(29).26 163 ]ροιc, surface missing (17 mm space, enough for 5 letters, before first trace to projected l.-h. margin), then descender of ρ, then, after 2 mm gap, trace of bottom of a curved letter, between bottom and mid-height, then 2 mm gap, then lower half of a descender, almost certainly ι. Total space between ρ and ι 4.75 mm. This is exactly the same as for ραι in µίτραισιν in Bacch. 13.196, col. 34(28).12, so Desrousseaux (Rev. Phil. 1898) is incorrect if he is interpreted as saying that α is ruled οut on spacing (194: ‘… il n’y a pas la place d’un Α, mais plutôt d’un Ο.’). However, we would expect the base of α to show much lower than the trace here actually does; therefore certain that this is in fact ο not α. Irigoin (1993) reads ]ρειc, but again the trace seems too high to represent the bottom curve of ε. 164 τ̣ [̣̣]ν, after τ bottom half of a short descender, then space for two letters before ν. 165 ι [, after ι, speck of ink at break-off level with top of preceding letter. ρ[, tip of below-line descender, too far to left to be υ or φ. 166 Maehler’s text ought to read ύπ[’ Αἰα]κίδαις rather than ύπ[’ Αἰακἰδαις, since the end of the line is preserved. Omission of the square-bracket must be a misprint. 167 ψ[x0323;]π̣ύ̣ [, below κον in line above, remnants of a topstroke, compatible with π, followed by an acute accent, either side of which there is a speck of ink, therefore compatible with ύ, which, coupled with the compound-word stoke beneath ψ, would produce the supplement ἐρειψ[̣]π̣ ύ̣ [ργοις].

(p.356) Notes on Supplementation

Metre: Dactylo-Epitrite. In the discussion of individual lines below I give Maehler’s metrical analysis (Maehler I.1. 17) at the start, though there are difficulties when the length of the link anceps is in question.

Lines 155–6 are a self-contained problem. There is a stop after άρµὁν, and in the context of the antistrophe end, it seems that we require a two-line phrase summing up the action of the Trojans’ assault in 141–54. µ]εγ’ appears most plausible. (Possible alternatives might be verbs ending in -εγω; possibilities whose forms might be compatible with the metre and spacing might be λέγω, ἀλέγω, φλέγω, ὀρέγω, στέγω, and ψέγω. Something like ἥν ὀρέγ’ might be possible after χειρός in 154 (‘under the hand of Hektor, which he stretched out …’) but it seems extremely difficult to see what sense this would make when coupled with what follows. I also briefly considered ἔλ]εγ’, but couldn’t see where that might lead us either.) It seems that we require something in the neuter for µεγ’ to agree with, or else an adjective (of any gender) for µεγ’, if understood adverbially, to intensify. Maehler suggests that we need a noun with the sense ‘pain’ or ‘fright’, describing the negative effect on Greek morale of the Trojan upsurge. Barrett alternatively suggests [ὄναα]ρ in 156, turning the sense round to describe the positive effect on Trojan morale here (citing Il. 22.432 ff. as a parallel). The problem is that the most obvious supplement, Jebb’s πηôµα in 155, seems about a letter and a half too long for the available space. This has led to the positing of a simple verb for ‘there was’ at the start of 155, leaving the start of 156 for a noun (hence Barrett). The problem is also compounded because of uncertainty over the quantity of the first-syllable anceps in 156. Maehler in his notes suggests ἴυγµα for the start of 156, but it now seems that this final alpha is ruled out because of incompatibility with the initial trace (see textual note on 156 above). If, agreeing with Jebb, we might supply an epithet to modify ὁρµαν, reading the original ἰσοθέων instead of Tyrrell’s and Barrett’s alteration to ἰσοθέον (Jebb seems right that ‘these two verses speak of heroes pitted against heroes’, with ἡµιθέοις balancing ἰσοθέων in reference to Greeks and Trojans respectively), Jebb’s ཀξεΐαν seems too long for the space, as does Blass’ βαρείαν. I suggest ༰θείαν, my preference for which is discussed below. We are then left with the start of 155. None of the supplements so far offered seem to fill the space well enough. Barrett’s η̑̓ν <δέ> µεγ̉ is an obvious compromise over Schwartz’ offering, and perhaps seems a little short. If, as I believe, we still need a word to work in apposition to the previous phrase, I suggest ἰξύ, in the metaphorical sense of ‘grievous’ or ‘painful’, or, working closely in apposition to the previous phrase, meaning (p.357) something like ‘keenly contested’ by analogy with its usage with µαχη, (see further below) and intensified by µεγ’, used here adverbially. The resultant word-order is unusual, but cf. Il. 22.88 for emphatic use of µέγα following the adverb it intensifies (ανευθε δ e σε µέγα νώϊν | Άργείων παρα νηυσι κύνες ταχέες κατεδονται).

157 is the start of an epode, and the narrator describes the Trojans’ over-indulgent ambitions for success against the Greeks. It seems likely that a main verb expressing the Trojans’ confidence and hopes is required in 159, rather than another participle after the one we have already had in 158. It therefore seems unwise of Maehler to give preferential treatment in his text to Barrett’s supplements for 159 (he prints Barrett’s θΊέντες] α · [δα]ν). This main verb (‘the Trojans cried out/boasted/thought that …’) is to introduce an accusative-and-infinitive construction (or a mixture of a nominative and an accusative with separate infinitives); since the Trojans cannot be the subject of ἕξειν (on sense grounds the object would then have to be πάλιν, and it is difficult to see where this could lead us), πάλιν must at least be the internal subject of this verb. There are two courses of action here. Either (i) we take πάλιν as the internal subject of the whole clause, with the τ’ linking two separate infinitives to describe two actions the Trojans hope that the city will carry out, or (ii) we take the τ’ as making a divide between the first half of the clause (nominative & infinitive: ‘the Trojans hope that, once victorious, they will do x’) and the second (acc. & inf.: ‘and that the city will do y’). Jebb objects to a nominative participle in 161 (pp. 349–50: ‘A nomin., εκπέρσαντες, would … imply that they actually destroyed the ships.’), since this nominative could be understood to have its force outside of the Trojans’ projected hopes (we know that in fact the Trojans didn’t destroy the Greek ships). Jebb therefore supplies a dative εκπέρσασιν dependent upon ἕξειν, therefore anterior to their prospective celebrations. Given the validity of this objection, it seems that for option (ii) no nominative could be suitable to satisfy the conditions for the first half of the supposed construction. Barrett, siding with option (ii), offers the verb introducing the infinitive construction (ωισθεν) in 161 after his εκπέρσαντες, and this seems openly to fall foul of Jebb’s objection over the force of the nominative. Barrett offers νεΐσθαι παλιν in 162 (’The Trojans thought that they would come back home); as Maehler notes, he cites Il. 8.498–500 for the sense, but it seems to me that perhaps we do not need a verb to state the Trojans’ return; this would surely already be implicit in the city holding celebration for them. After some deliberation, I have decided therefore to side with option (i). This has the advantage of not forcing a change of subject midway through the infinitive construction, but it might be said on the other side that a dative participle (p.358) dependent on a noun that appears two lines later at the end of the clause is little more satisfactory. However, a comparison with Pindaric syntax on this issue suggests that a dative would be possible, and possibly even desirable: cf. Hummel (1993) §132: ‘On compte un certain nombre de tournures formulaires qui font partie des composantes conventionelles de l’épinicie où sont associés poète, éloge et destinaire’; she cites e.g. Pind. Nem. 4.73–4,Θεανδρίδαισι δ’ ἀεχιγυίων ἀέθλων κάρυξ ἐτοι̑µος ἔβαν, and Nem. 1.7, ἔέργµα-σιν νικαφཹροις ἐγκώµιον … µέλος. With such Pindaric parallels, though we obviously have to be careful in making analogies between the syntax of these different poets, the separation of the dative participle from έ'ξειν πཹλιν seems less of a concern. The translation I have offered tries to bring out this force for the dative (‘… in their honour’). The double infinitives linked by τ’ also seem quite satisfactory on this alternative hypothesis, given that they now have the same subject.

155 Metre: D ἠ̑ν δὲ µέγ’ (Schwartz) too long. ἠ̑ν δx1F72;æµ έγ’ (Barrett) perhaps a little too short. ἠ̑ε µέγ’ (Pfeiffer) too short. πηx0311;µα µέγ’ (Jebb) and δει̑µα µέγ’ (Jurenka) both too long. Although I am unable to find any parallels for ὀξὺ standing without a noun in close proximity for it to qualify, there are enough parallels for it modifying a noun meaning ‘pain’ that I believe it would be understood in the light of these connotations in this context. See, for instance, ὀξ#x1F7A;… ἅχος (H.H.Dem. 40), ἅχος ὀξ#x1F7A; (Il. 19.125). Furthermore, in apposition to the last clause describing the Trojan assault, it could be likened to its usage with µάχη: see ὀξει̑αν µάχαν, also at Bacch. 13.117, in the sense of ‘keenly contested’ (LSJ cite Hdt. 9.23). This appositional usage would also strengthen its position, qualifying the previous phrase, despite the fact that, being in the neuter, technically it has no noun to qualify. Kühner-Gerth II 1.285 cite among others Soph. Ant. 44, ἀπόρρητον πόλει for this appositional use of a neuter adjective or participle to pass judgement on the clause preceding. For the usage of µέγα adverbially with adjectives in positive form, see e.g. µέγ’ ἔξοχος (Il. 2.480), µέγα νήπιος (Il. 16.46) and µέγα πλούσιος (Hdt. 1.32 and 7.190); again, cf. ἄνευθε … µέγα (Il. 22.88). I believe that the force of both senses is at play here: hence my translation (‘…, a mightily keen, bitter contest …’).

156 Metre:—E—||| ϕέροντος (Jebb ap. Kenyon) ruled out by spacing. τεύχοντος (Desrousseaux) ruled out by spacing, as are ὀξεîαν (Jebb) and βαρεîαν (Blass). Barrett’s ὄνααρ fits the spacing, and also takes account of the minute ink trace at cut-off of papyrus. However, to go along with Barrett also requires the dangerous alteration (first offered by Tyrrell) of the uncorrected extant ἰσοθέων to ἰσοθέον to agree with his ὄνααρ. ἰθεîα]ν is paralleled by Bacchylides at 15.54, qualifying Δίκαν and with the sense (p.359) there of ‘straight’ i.e. ‘upright’, and has the added advantage over Jebb’s suggestion that it modifies ὁρμάν in a very logical fashion, in the sense of ‘direct’. Although I can find no parallels for the coupling of ἰθεîαν with ὁρμάν, and no adjectival usage of ἰθύς in Homer, I suggest that Bacchylides could be thought to have elaborated on Homeric phrases such as ἰθὴςμαχέσασθαι (Il. 17.168), and ἰθύς μεμαŵτος (Il. 8.118), which the scholia gloss with ᾄντικρυς βουλοµένου ὁρµᾶν; consider also Σ Il. 6.79a, where Homer’s πᾶσαν ἐπἰθύν (used as a noun in the sense ‘endeavour’) is glossed as εἰς πᾶσαν ὁρµήν, and similarly Σ Od. 4.434.1, where ἰθύν is glossed with ὁρµὴν as well as πρᾶξιν. The supplementation here is complicated somewhat by the metrical problem of the quantity of the initial anceps. As Jebb correctly notes, in all other corresponding places the initial anceps is long where it is preserved. A further question to ask is to what extent short anceps is theoretically ruled out here. Barrett (Hermes 84 (1956) 248–9), cited by West (1982a) 74, states for Pindaric practice that the first triad sets the precedent for long or short anceps for the rest of a given poem. He also suggests that ‘short ancipitia in Bacchylides tend to correspond: a place which admits ˘ tends to admit it in more instances than one (though not nor mally in all). Pindar, in whom ˘ is relatively only half as c om m on, has the same ten den cy to co r respondenc e, though it is in gen eral less stron gly marked’. Therefore, despite the fact that Bacch. 13 ‘shows an exceptionally fr ee tec hnique in both frequ ency and position of ˘’ (ibid. 250) it seems that a short anceps in this p lace would be betrayed b y a cor re spond in g instance elsewhere in the poem. Maehler therefore seems correct to print the initial position as long rather than anceps in his metrical schema. If this is the case then it is a further objection to Barrett’s ὄνααρ, as well as to ϕέροντοc, though Barrett obviously felt that considering the nature of the evidence a short syllable would not be impossible.

157 Metre:—Dxe| Of compatible supplements, Blass offers ᾆ δύσφρονες, Barrett χαλίϕρονες, Jurenka ᾆ π ά ρφρονες. The otherwise attractive τλάμ]ονες read by Kenyon, Desrousseaux, and Jebb is now unfortunately ruled out: the amount of ink at below-line level in the initial trace is unaccounted for by Barrett, and is incompatible with a r.-h. descender of e.g. µ. ἆ accords well with Homeric usage; cf. Jebb on Bacch. 3.10: ‘The exclamation is regularly found in expressions of pity or reproof, as in the Homeric ἆ δεîλ(Il. 11.441 etc.)’. Jebb objects to Blass’ δύσϕρονες, suggesting that it more frequently means ‘melancholy’ or ‘malevolent’ rather than ‘misguided’. Though it seems that this is very much a subjective matter, there seems no obvious and unambiguous parallel for ‘misguided’ which couldn’t also be interpreted as ‘malevolent’ or ‘hostile’. More often than not, in fact the word does seem to mean ‘hostile’ (cf. e.g. Aiskh. Supp. 511, (p.360) Ag. 547, 608). Jebb parallels Aiskhylos’ usage at Sept. 875 (with reference to Eteokles and Polyneikes), but this could refer to the brothers’ mutual hostility as well as, if not more than, to their misguided actions. However, I believe that this works in favour of δύσφρονες: in the context of the Trojan assault, it is clear that they are hostile as well. Furthermore, Bacchylides may also be offering a modification of the Iliadic δνσϕενής, ‘hostile’, here, with the sense of pity or reproof added by analogy with the Homeric usage of νήπιος and with the coupling with (see Jebb above). Once again, the supplementation here is complicated somewhat by the metrical problem of the quantity of the initial anceps. As with the previous line, the initial link anceps is long in all corresponding extant places (though in this case we are lacking information for the start of the first two of the poem’s seven epodes). Theoretically there appears the same indeterminacy as there was in the case of the previous line. Barrett’s offering of χαλίϕρονες, with initial short, again seems to indicate that he thought that an initial short anceps here could not be theoretically ruled out.

158 Metre:—D Though theoretically the length of the link syllable is difficult to determine, as Jebb states, ‘πνείοντες (Jurenka, Ludwich) is more likely than πνέοντες (Blass), because in all the corresponding verses (59, 92, 125, 191, 224) the first syllable is long’. Here Barrett’s πνε<ί>]οντες indicates that he considered that the initial link here ought to be long too.

159 Metre: xe—| All supplements previous to Maehler are outdated since Barrett’s analysis of the ink traces. I offer here a main verb for the utterance of the Trojans’ overweening ambitions. A main verb the length of ἔκλαγξαν would be slightly too long to fit the space. The word I adopt, θρόησαν, is paralleled at Bacch. 3.9. The short anceps here accords with the evidence of the majority of other places (short in five out of six of the extant corresponding places) and with Barrett’s analysis of this poem’s predilection for short anceps in the position ˘e(x) (art. cit. 250). Furthermore, the phrase τίνα θροεîς αὐδά ν; occurs relatively frequently in Euripides (Troades 12 39, Or. 1249, Hipp. 571), providing a parallel for A’s error, on my hypothesis (see below). On the possibility of this phrase or verb introducing indirect speech with an infinitive construction, though I can find no precise parallels, Bacch. 3.9 ff. is of interest. Although Maehler argues against taking 3.10–14 as direct speech after the verb (as Jebb), his arguments (he suggests that such appreciation of the victor would not square with the excited state of the hypothesized spectators) strike me as weak. See Carey, ‘Ethos and Pathos in Bacchylides’ 24: ‘Only modern printing methods can distinguish whether these lines 10–12 are a statement by the poet or a cry by the crowd at Delphi; it is difficult to see how the identity of the speaker could have been brought out in performance. The result is an indeterminacy in the authority for the (p.361) statement.’ If therefore this verb can be used to introduce oratio recta (at least as a possibility) in Bacchylides, it seems only a short step to using it to introduce oratio obliqua. What I believe has happened in the text of the papyrus is that A wrote αὐδὰν, and that A 2 offered the variant αὐχὰν by writing a χ over the δ (or else that the δ originally had cancellation stroke though it); this χ can still be seen above the traces of this line (a reconstruction of the line by tracing alerted me to this hypothesis). I believe αὐχὰν to be the better alternative, since it conveys more of an emotional charge to the Trojans’ arrogant utterances: boasts, not mere words. Supralinear correction with concomitant deletion is paralleled in most columns of the papyrus; examples of variant readings added without deletion of letter beneath are col. 19(15).7: A 3 offering [δ]ωκεν for [θ]ηκεν A in Bacch. 9.26, and col. 31(25).6: A 3 offering υφαυχαϲ for A in Bacch. 13.84.

161 Metre:—E|| Once we have a main verb in 159 to govern the infinitive construction, we no longer need Barrett’s ingenious ὤσθεν, so can return to Jebb’s Ἀργείων (or Desrousseaux’s Ἑλλάνων). The use of the rather grandiose ἐκπέρσασιν (Jebb) rather than the more descriptive ἐκφλέξασιν(Nairn) or ἐκκαύσαντες (Desrousseaux) captures perfectly the overweening ambitions of the Trojans. The initial long anceps, though not theoretically established beyond doubt, accords with all five extant (or securely supplemented, as with v. 62) corresponding places, where it is also long. Furthermore, if a compound of πέρθειν is the right, we may find its usage ironic in the context: normally it is used of the Greeks whose wish it is utterly to destroy Troy: cf. Il. 1.19, 9.20, etc., and indeed it is their main priority; see also Sim. fr. 11.13 W. For a parallel for the ironic usage, see Rhesos’ overbold wish to destroy Greece at Eur. Rhes. 471–2: ξὺν σοὶ στρα-τεύειν γη̑ν ἐπἈργείων θέλω | καὶ πα̑σαν ἐλθὼν Ἑλλάδἐκπέρσαι δορί.

162 Metre:—D×| All supplements (apart from εἰλα]πίναϲ, which seems unassailable) previous to Maehler are outdated since Barrett’s analysis of the ink traces. The initial long anceps, though not theoretically established beyond doubt, accords with the five extant (or securely supplemented, as with vv. 63 and 96) corresponding places, where it is long. If, as I believe, we need an infinitive explaining something that the Trojans hoped that their city would provide for them, in parallel to εἰλαπίνας … ἕξειν, and theirreturn is already implicit, an alternative interpretation of the ink could produce χορὸν rather than Barrett’s πάλιν. This is particularly useful, since Jebb (following Nairn) and Jurenka actually offered χορὸν here before Barrett interpreted the ink traces he found, and it frees up the start of the next line. Headlam (CR 14, (1900), 13), who suggested καὶ χο]ρ[ο]ι̑ς for the start of the next line, remarked ‘χοροί are the natural accompaniments of peace and joy, Ar. Pax 976, Eur. H. F. 755, Hes. Scut. 272–285 after Hom Σ (p.362) 491 sqq.’ Once we have χορὸν, a useful verb to go with it could be στάσειν, for ἵστηµι is used elsewhere for the establishment of a chorus or choruses: e.g. Hdt. 3.48; Σ Aiskhin. 1.10 (29 p. 15 Dilts); in particular Bacch. 11.112 (χοροὺς ἵσταν γυναικω̑ν: part of the festivities celebrating the return of the Proitidai). Musical celebration is also apposite (quite apart from the self-reflexive, or ‘projected’ usage in a myth within a poem performed chorally) by analogy with the celebration of Akhilleus and the rest of the Greeks in Iliad 22 after his killing of Hektor (22.391); cf. also Timoth. 791 PMG lines 196–201. For ἵστηµι of choruses see too Calame (1977) 88.

163 Metre: E×|| Now that the problems over the remnants of the start of the line have been resolved, i.e.]ροιϲ not]ραιϲ or]ρειϲ, and ‘dancing’ has been displaced to the previous line, Barrett’s λαοφόροις seems plausible. Headlam’s [καὶ χ]ορ[ο]ι̑ς would fit, going after Blass’ µετ’ εἰλα]πίνας τ’ ἐν (Headlam cites Pind. Ol. 2.28, 7.26, Pyth. 10.58 and Nem. 7.31 as parallels for ἐν καὶ), but this no longer squares with a reading of the previous line, and it is very unclear how to fit this with the rest to make plausible sense, especially considering the earlier τ’. Desrousseaux reads [εὐπό]ρ[ο]α̑ς, which would also fit and mean something similar to Barrett’s offering. However, despite Maehler’s qualms, a parallel which neither Barrett or Maehler appears to have spotted for λαοφόροις seems to me to alter the balance in its favour: the word λαοφόρον is used in Homer, only once, in Iliad 15, in the simile describing Aias’ leaping from ship to ship in warding off the rampaging Trojans (λαοφόρον καθ’ ὁδόν· at 15.682). Since Iliad 15 is a recognized source for much of the material in this poem’s myth, it seems highly likely that Bacchylides may have also been influenced by some of its unique diction. Alternatively, it might be possible that εἰλαπίνας should have read εἰλαπίναις, but was left uncorrected, and that therefore the word at the start of 163 could have been a compound adjective in the dative to agree. This highlights the problems of supplementation in these lines, but it would seem extremely rash to alter a preserved ending where the existing one can be reconstructed to fit correctly into a meaningful sentence. The problem with having a compound adjective at the start of 163 (not to mention the problems we might have deciding what it should be, considering Bacchylides’ innovatory coinages; in any case, this use of λαοφόροις would itself be innovatory considering that the only usage before Bacch. 13 (Il. 15.682) is as a compound adjective rather than a noun) is that without εἰλαπίνας as the accusative object of ἕξειν, the temptation would be to make πόλιν the object of ἕξειν, and it seems unlikely that anything could be made from this. It would destabilize the τ’, which might also have to be deleted, again a rash procedure.