Thursday, 21 December 2017

Gongyla: a difficult reconstruction

Last poem was addressed to Atthis, this one… to Gongyla? Depends on what version of the text you take. I have a long story behind this poem, which I have no time to tidy up and set into English, so this is another WIP-spoiler critical note poem, surrounded by what I wrote far later, on 3-4 April 2018. There are also some mysteries on the sources. The sum-up is that Edmonds clearly had only P.Oxy;. 1231 fr. 15, Campbell also seems to have had only that (and fr. 12 which will be addressed in another post), but in reality the fact he reads more suggests he used another scrap, possibly a P.Oxy. 2166(a) fragment, which I cannot find anything about online. Except P.Oxy. 2166(a) is from higher up in the column, so it's just Lobel-Page reading extra stuff on fr. 15 and Campbell following suit. A couple letters seem to have completely revolutionized the interpretation and restoration of the poem, but in fact, as will be seen below, the real problem is, in most cases, who the addressee is, and only in l. 7 do the extra letters alone matter a lot. What I present below are the vulgate version set to music by Greek songwriter Manos Khatzidakis (version 1), which only uses P.Oxy. 1231, and Campbell's version, which I initially rejected since I couldn't make sense of it. Edmonds has yet another version. The meter is Sapphic stanzas, which is rendered the usual way. Let's get to it!

Ν]ῦν, φ[ίλα, σὲ μὰν κ]έλομαι· σ[πόλαν σάν,
Γο]γγύλα̣, π[έφα]νθι λάβοισ̣α̣ μά[λαν
Γλα]κ̣τίνα[ν·] σ̣ὴ δηὖτε πόθος̣ τ[ε αὖτο]

Τὰν κάλαν· ἀ γὰρ κατάγω̤γις αὔ[τ]α [μ'
Ἐπτόαις' ἴδο[ι]σαν· ἔγω δὲ χ̣α̣ίρω·
Καὶ γὰρ αὔτα̣ δ̣η τ[όδ]ε μέμφ[εταί σοι

Τ]ς ἄρ̣αμα
Τοῦτο τὦ[πος
… ]

[Gon]gila, [or ti p]rego͜ [invero,͜ amata]:
La [la]ttea tu͜a veste indossata
[Ap]pari, mentre͜ attorno a te͜ adornata
Vola ’l disìo,

Bella; la veste stessa oltremodo
Sconvolge [me] vedente,͜ ed io ne godo:
La stessa de͜a [C]iprig[na] qu[es]to odo
Bia[smarti] io:

[Q]uella i’͜ am[o]…
Questa par[ola]…
I͜o [d]isì[o]…
[Ā]pprĕcōr [tē]: [lā]ctĕăm āpprĕhēndĕ
V[ēst’, ămātă Gōn]gy̆lă, [tēn]dĕqu’ ōs tē.
Tē quĭdēm dēsīdĕrĭūmqu[ĕ] cīrcūm-
vōlĭtăt [īpsŭm]

Pūlchrăm; īllă vēstĭs ĕn’ īpsă tūrbăt
[Mē] vĭdēntēm mūltŭm, ăt īpsă ga͞ude͞o:
Nāmquĕ d’ h[ōc]cĕ īpsă rĕprēndĭt īllă
[Tē] dĕă [C]ȳprĭ[s:

Ī]ll’ ăm’ īps[ă]…
Hōccĕ vē[rbŭm]…
Nūnc [v]ŏlō…

[I p]ray [you, darling]: out your d[ress now] take,
And, [mi]lky in it, your [app]earance make,
O [Gon]gyla, an[d] Wish [itself] please make
Flutter ’round you,

So glorious; th’ dress itself doth [me] distract
Indeed, but I with joy to that react:
The Cyprus-born herself – and that’s a fact -
Bla[mes] t[hi]s [of you:

T]hat I wis[h] for…
[T]his wo[rd] here…
[I] now wan[t]…
Νῦν] δ' ἔγ[ωγ' ἦ μὰν κ]έλομαι σ[' ἀείδην
Γο]γγύλα̣ν, [Ἄβα]νθι, λάβοισ̣α̣ν α[ὔταν
Πᾶ]κ̣τιν, [ς] σ̣ε δηὖτε πόθος̣ τ[ε αὖτο]

Τὰν κάλαν· ἀ γὰρ κατάγω̤γις αὔ[τ]α[ς σ'
Ἐπτόαις' ἴδο[ι]σαν· ἔγω δὲ χ̣α̣ίρω·
Καὶ γὰρ αὔτα̣ δ̣ηπ[οτ'] ἐμέμφ[ετ' ἄγνα

ς ἄρ̣αμα
Τοῦτο τὦ[πος

Abanti, or ti prego inver, suonando
L’arpa c’hai presa, Gongila cantando,
Fa‘ che disìo di lei vada volando
Intorno a te,

Bella; la veste stessa sua te un tempo
Vedente sconvolgeva; gioia io sento:
La stessa dea Ciprigna infatti un tempo
Biasmava me

Ché io disio
Questa parola
Āpprĕcōr tē prēndĕrĭ’ c’ īllăm ārpăm
Gōngy̆lam, ō Ăbānthĭ, mĭhī cănās tū;
Ēiŭ’ tē dēsīdĕrĭūmquĕ cīrcūm-
vōlĭtăt īpsŭm

Pūlchr’; ĕnīm tūrbābăt ĕ’ īpsă vēstĭs
Tē vĭdēntēm mūltŭm; ăt īpsă ga͞ude͞o:
Nāmquĕ rēprēndēbăt ĕ’ īpsă ōlĭm
Mē dĕă Cȳprĭs

Quodqu’ eg’ opto
Hocce verbum
Nunc volo…

I pray you now, who’ve taken that harp there,
Abanthis, sing of Gongyla, while there
Flies ’round you once again the wish for her,
So pretty you;

For once her dress indeed did thee distract
As you did see it; now with joy I act:
The Cyprus-born herself – and that’s a fact -
Blame me used to

For I wish
This word here
I now want…

Critical note

More wild variations in a text, eh? Yeah, and with even more mysteries. But let us first leave the spoiler with the old critical note, and then proceed with the history of the text of this poem, and what I did about it back in the days.

With that being said, let's look into the history of this poem. In 1914, the 10th volume of the Oxyrhynchus Papyri was published. It contained P.Oxy. 1231 fr. 15, the only (or almost - see below) source for this poem. Starting from that volume's transcription, Edmonds produced his own restoration, which was published in Edmonds' 1922 edition of Sappho. W-M had previously "recognized" this was addressed to Gongyla, and that was what got into Edmonds' text. As time passed, another addressee presented itself. In 1951, Lobel published volume 21 of the Oxyrhynchus Papyri, where P.Oxy. 2299 fr. 10(b) was:

In particular, the final line read ὄρχησθ' [ἐρό]εσσ' Ἄβανθις, "graceful Abanthis danced". And in l. 2 of our poem, a convenient lacuna had νθι right after it. Hence, the idea that Ἄβανθι was to be read there was born, and for some reason it just stuck: it was a possibility given in LP's apparatus criticus in 1955, it stayed like that in Voigt 1971, who seems to prefer the perfect imperative solution for which cfr. below, and Campbell adopted it in 1982, and so did Carson in 2002, and Powell in 2007. So let me restate this: Ἄβανθι stuck among translators of Sappho to English. However, another idea had been proposed by Wilamowitz, and was reported in Voigt: filling the lacuna with the perfect imperative πέφανθι. Greek translator Panagis Lekatsas (Παναγής Λεκατσάς) took it up in his Σαπφούς Άπαντα (All of Sappho), dated who knows when after 1937 (date of publication of Reinach which is among Lekatsas's sources), and presumably from that book the alternative reached Greek songwriter Manos Khatzidakis (Μάνος Χατζιδάκις), who set it to music and included it in his 1972 album Ο Μεγάλος Ερωτικός (O Megalos Erotikos, the big lover), where he set it to music in a way totally irrespectful of the meter, with lyrics pronounced halfway between modern and ancient pronunciation (as highlighted by the lyrics' spurious umlauts that divide some "ai"s and "oi"s but not all of them), lines divided according to the music and not the original meter, and even a wrong word division (γλακτὶν ἄν or with different diacritics perhaps), and possibly coupled (as it is on the link I gave) with the horrible translation by Odysseas Elytis (Οδυσσέας Ελύτης), which skips the second half of l. 5 and the first half of l. 6 entirely. Then came I to gaze upon this mess. But wait a sec: one more thing must be noted. The P.Oxy. 1231 fragments are numbered together (I presume) because they are believed to be parts of a single papyrus roll. I do not know why this is, but let us assume that is the case. How do you decide where each fragment was in the papyrus, when there is no physical joining between them? And above all, how do you decide that a single half-letter from the end of fr. 12 belongs in l. 1 of fr. 15, when fr. 12 hardly makes any sense because of mutilation, and the two fragments don't join at all? In any case, the joining stuck starting from Lobel and all the way to Powell and Carson.
Then came I to gaze upon this mess. I started with Greek Wikisource, and Bibliotheca Augustana, which, modulo critical notation, have exactly Campbell's text. Now I don't know you, but I am not used to seeing a person as the object of to sing, so my attempt to interpret that accusative Γογγύλαν ended up with «WTF does the harp (πᾶκτιν) of l. 3 have a name for? Since when do instruments have names?», and this, together with the unfindable name Ἄβανθι, made me reject that version, and go googling. I thus landed on some site with Xatzidakis's music version, and, though horrified at the lack of critical notation, the wrong line division, the misspellings, and the bad translation, I took up the perfect imperative, which solved the problem of the weird name and also eliminated the "name of the harp" by making it vocative. Two problems were left: what is a perfect imperative, and what is γλακτὶν in l. 3 (remember I had the wrong version from Xatzidakis, γλακτὶν άν or whatever it was included). So I searched for the sources, and finally landed on P.Oxy. vol. X, by Grenfell and Hunt, and found [γλα]κτίναν, which solved problem nr. 2. As for the perfect imperative, I knew that present and aorist imperative are distinguished by being of imperfective and perfective aspect respectively, which is a distinction of the tenses themselves, so I thought, what characterizes the perfect tense? Probably influenced by the way "present perfect" was explained to me in English classes, I answered, well, the present visibility of the action's result. Putting this together with imperative, I concluded the perfect imperative was a command with the implicit statement that the result of the execution of said command was known to the speaker. And this fit perfectly with having "that dress makes me distraught when I see it" in stanza 2, so case closed, translation time. I conjured up a completion that fit the GH text, et voilà, the text and translations of the "Vulgate version" above. Far later on, as I fought against "O Atthis!" (which I then called Ἀριγνώτα as it was titled on Greek Wikisource), I looked for its sources (that is P.Berol. 9722 fol. whatever number), and I landed instead on Campbell, read the translation of Gongyla for some reason, and went "Oh THAT's what Gongyla is doing! It's the object of ἀείδην in l. 1! That makes total sense!», so I conjured a new completion for the Campbell text, and translated it, relegating the translations in the Paracritical Note. And that gives us the text and translations of the Campbell version above. And I was over with this fragment.
Fast forward some 8 years, and the blog appeared. Go back to this fragment, and whoa whoa whoa, what the bleep are these differences of critical notation from GH to Campbell, and the magical letter conjurations and transfigurations that happen in Campbell? Stack Exchange suggested I check out Lobel-Page and Voigt, so I found the former on Amazon and it was a Christmas present for me, but Voigt is not buyable online, though it subsequently appeared on Scribd. Now Lobel-Page partially answered the question with "lectio nonnullis locis difficillima" (reading is very hard in some points), but that left open problems. Then, the Bodleian library published an image of some P.Oxy. 1231 fragments, and I analyzed and transcribed the Gongyla fragment. All of this is of course documented in dear old transcriptions post, where I reject the joining of fr. 12 on grounds of why would I join it that way (and besides my completions are both destroyed by a Γ/Ζ/Ξ/Π/Τ somewhere in the middle of l. 1), I made notes answering all the questions about the variations, and gave the following papyrus text:

] . . . [ . . . . . . . . . κ]έλομαι σ[
Γο]γγύλα̣ . [ . . . . ]νθι λάβοισ̣α̣ . . [
]κ̣τιν . [.] σ̣ε δηὖτε πόθος̣ . [
τὰν κάλαν· ἀ γὰρ κατάγω̤γις αὔ[τ]α[
ἐπτόαις' ἴδο[ι]σαν· ἔγω δὲ χ̣α̣ίρω·
καὶ γὰρ αὔτα̣ δ̣η . [. .]εμεμφ[
.] . ς ἄρ̣αμα
τοῦτο τὦ[πος

So now the time has come to play a fun little game called "Effects of the Addressee". The game is simple: let us choose who this poem is addressed to, and see what consequences this has on the text. We have two options: Abanthis or Gongyla.
  • If we choose Abanthis, this means the lacuna in l. 2 is to be filled as [Ἄβα]νθι; great, now we have two vocatives: Γογγύλα, Ἄβανθι; but there is only a singular σ(ε) in l. 1, so we cannot have two addressees unless we assume that sigma is not for a pronoun but some other word, and we find a way to stick ὔμμε (you pl. acc.) into l. 1, and also to deal with that vestige before the lacuna, which would have to be the A of Abanthi, but cannot because the options are ΒΓΗΙΚΜΝΠΡ; wait, did you say a nu is possible? What about Γογγύλαν? Great! Problem solved… um, wait, what's that accusative doing there? Putting it with an incomplete word turns a problem into another one, so we will have to suppose κ]έλομαι in l. 1 is followed by a verb that takes an object; what could this verb be? Well, what about πᾶ]κτιν in l. 3 and σ' [ἀείδην or σ'· [ἄειδε in l. 1? Great, but which do we choose? Well, the former does't allow reading λάβοισα̣ν without placing it with Γογγύλαν before, but then the harp would be taken by Gongyla, and Abanthis would be singing a cappella about Gongyla playing the harp? Not too convincing right? But if we want to place λάβοισα̣ν with the σ' we need to eliminate the high dot, which the imperative forbids; so we either take the infinitive route, or are at a loss as to how to complete l. 2; hence, we go for the infinitive, and read λάβοισα̣ν α[ύταν as the end of l. 2; we already said we want l. 3 to start with πᾶ]κτιν; what next though? The vestige suggests alpha or chi, but chi would make the iota in πᾶ]κτιν long, which means inmetricality; what is the next letter then? That minimal pseudo-trace would suggest sigma or nu; the sentence at hand already has an accusative, so ἆν would need to be genitive plural, referred to what? Gongyla and the harp? Why not have ἆ[ς], either as "whose" or as "while" (=ἔως)? And with that, from κέλομαι to πόθος is Campbell's text, plus the αὔταν completion; l. 1 was completed as it was starting from the Campbell text; having put my hand to the papyrus, I now propose Νῦν] δ' ἔγ[ωγ' ἦ μὰν κ]έλομαι as a better alternative, or any other way to fill the middle lacuna with intensifying particles (γὴ, μὰν, ἦ, γε or the likes); τε αὖτο in l. 3 is just auto, no other option presents itself; and stanza 1 is filled up; but we can go on: now that we have no reference to a garment in stanza 1, the κατάγωγις must be of Gongyla, hence l. 5 must end with αὔ[τ]α[ς; we are then left with the choice as to whether this garment made any woman beholding it distraught, or just Abanthis; the latter case requires an extra σ' at the end of l. 5; why did Campbell go that way? Maybe because why would only women be made distraught, and not men? If we add that σ', we keep things within the friend circle of Sappho, whereas if we don't, we would expect to open the distraughting effect to the world, but no, just women; mmh…; and with that, the mere choice of Ἄβανθι and some thought led us to Campbell's text up to l. 6, modulo the missing letter in l. 1 (from fr. 12 which we rejected) and the extra completions; l. 7 is a beast of its own kind, in a way, but the choice of Ἄβανθι means the accused one cannot be σοι, the addressee, and hence εμεμφ must give a past tense; from there, we can choose δήπ[οτ'] ἐμέμφ[ετ' ἄγνα, as in Campbell, or δὴ τ[όδ'] ἐμεμφ[όμαν τὰν, but at this point we just go with Campbell;
  • We now try choosing Gongyla as the addressee; this means Γογγύλα is vocative, and we expect a verb to go with κέλομαι; we now look at the lacuna we can no longer fill with Ἄβανθι: what do we do with it? We then have a choice: we either follow Edmonds with β[ρόδα]νθι, or Wilamowitz (and Xatzidakis) with π[έφα]νθι; in the former case, we are fed up, and just follow Edmonds entirely, instead of trying not to invent μάνδυν in l. 2 and to put instead a garment name in l. 1's lacuna; in the latter case, however, we have the verb in that lacuna, and the next thing we want to do is complete that . . [. Vestige one is ΜΝΗ, vestige 2 is «ΑΔ€ΘΛΜΟ, though the pointy nature of that blot would naturally suggest ΑΔΛ», so the likely combinations without weird clusters are ΜΑ, ΝΑ, ΗΔ, ΗΛ; of course, back in the days, I had no idea this was the case, so I simply completed the μα from Xatzidakis somehow; let's try to see if we can get there without that start; now, the eta options mean we have a horrible hiatus λάβοισα η, so we discard them; we then proceed to browse our dictionary for fitting terms with the first option, and find μάλαν meaning white; we have Xatzidakis with his [γλα]κτίναν (well, that is actually GH) in mind; so white, and then milky-white: a nice way to intensify the whiteness; but then where is the garment name? Well, we do have that sigma in l. 1, hitherto believed to be a pronoun; it so happens that στολή, or σπόλα in Aeolic, means "dress", so that is it: κ]έλομαι· σ[πόλαν σάν, will be our completion, and with the auto-τέαυτο we have the Vulgate version, save for still having to conjure up a completion for l. 1 before κ]έλομαι; having to include the σε in there, and starting from the GH text, we end up with the Vulgate version text above, though now the transcription contrasts it, suggesting Νῦν] δ' ἔγ[ω σε μὰν κ]έλομαι instead; but let us go on: in stanza 2, we can now refer to "that dress", or κατάγωγις αὔ[τ]α; then, once more, do we add an object "me" or not? Well, again, generalizing but not too much is weak, let's keep things between Sappho and Gongyla, and add [μ' at the end of l. 5; l. 7 is a beast of its own kind, because we still have the option between τ[όδ]ε μέμφ[εταί σοι and τ[όδ'] ἐμέμφ[ετ' ἄγνα, or even Campbell's text or Edmonds' text.
The last thing I want to discuss is what the perfect imperative actually is. I asked for usage examples here, and the answerer stated that «Perfect: the act is envisaged as being done once and for all, i.e. with permanent effect. “Fix it so that it stays fixed, got that?”». This more or less confirmed the idea I had from Idk where (probably a now-deleted comment here, I suspect) that it meant "Do this (and make it stay done)" or "Make this stay done", as the example from the Iliad I gave in the Quora question seems to hint at. So in this case, it would mean "appear (before me) once and for all (so that I can see you and be distraught by your lovely milky dress)".
Oh and finally, lots of googling and Wikipediaing got me to finding out about Wilamowitz's perfect imperative proposal, as illustrated here, but I could have found the answer in Voigt immediately. However, this way I am more certain that the perfect imperative actually got to the music version starting from Wilamowitz, and not some other proposer, since Lekatsas definitely got it from Wilamowitz, as his own apparatus criticus clearly states.

No comments:

Post a Comment