The tradition for this fragment is a horrible mess which I will detail in an addendum at the end (modulo updates) of the question by pasting the critical note I prepared this morning for it. Line 2 is especially problematic. This question focuses on safopoemas's (cfr. 1 especially, and also 2 and 3) text:
/έσπερε πάντα φέρων δσα φαΐνολις ¿σχίδασ' αδως φέρρες δϊν, φέπες αίγα,
φέρες (δ') α"πυ μάτερι παΐδα.
Noting that the linebreak may just be the document's margin, I proceed to emend the text to make it intelligible:
Ϝέσπερε πάντα φέρων ὄσα φαίνολις ἐσκέδασ' Αὔως
Φέρρες ὄϊν, φέρες αἶγα, φέρες ‹δ'› ἄπυ μάτερι παῖδα.
Given that the critical notes say nothing, and the introduction (cf. link 3 above) says they follow the 1937 French edition of Sappho curated by Reinach and reprinted in 1966, and that this is not a "fragmento no recopilado" (non compiled fragment), I would assume this is what Reinach had. So my question is:
Is there any justification for the double rho in the Φέρρες that opens l. 2? If so, which is it? If not, who (if not Reinach) proposed this, and why?
Last in the category1 we have Bergk 96 Edmonds 149 Lobel-Page 104(a) Campbell 104(a). OK, this is a huge mess. I already kind of knew that before I actually looked things up for this note, but it's way, waaaay, worse than I thought. For starters, it has 10 (yes TEN) sources, which I will try to reconstruct all below as far as Voigt allows me to, even in the oscillations of the different manuscripts for those sources. There is every sort of weird oscillation here.
- In Voigt's order, source one is Demetrius's treatise De Elocutione, where he wrote «χαριεντίζεται δέ ποτε (sc. ἡ Σαπφώ) καὶ ἐξ ἀναφορᾶς, ὡς ἐπὶ τοῦ Ἑσπέρου, "Ἕσπερε πάντα φέρεις", φησί, "φέρεις οἶνον φέρεις αἶγα φέρεις μάτερι παῖδα"· καὶ γὰρ ἐνταῦθα ἡ χάρις ἐστὶν ἐκ τῆς λέξεως τῆς "φέρεις" ἐπὶ τὸ αὐτὸ ἀναφερομέης», or in Campbell's traslation «Sometimes, also, Sappho makes charming use of repetition as in the description of the Evening Star: "Hesperus, you bring everything", she says, "you bring the wine, you bring the goat, you bring the child to the mother": here the charm lies in the repetition of the word "φέρεις" (bring), always with the same reference», that is with the same subject Hesperus. The translation of the quote is mine, for Campbell's text is amended. And here already we have a problem: what is the meter? The first part could be the first half of a hexameter with pentemimeral (is that the term?) caesura, but the rest oesn't fit the hexameter meter, it's like u–––u––uu––uu–u. Welp, let's leave the discussion for below, and just note that φέρεις is probably to be Aeolicised to φέρῃς, which Lobel himself corrected according to Voigt, though Lobel-Page and Campbell both keep φέρεις.
- Source two is a scholiast on Euripides' Orestes, who wrote, in Voigt's quoting cut, «ἑσπέρα [...] ἀπὸ τοῦ πάντας εἰσφέρειν [...] καὶ ἡ Σαπφὼ [...] τὸν ἕσπερο ἀστέρα εἶπε [...] ἐτυμολογοῦσα [...]· ἕσπερε πάτα φέρων ὄσα φαινολὶς ἐσκέδασ' αὔως» (ἑσπέρα, evening, [is called thus] from [its] εἰσφέρειν, bringing, everything […] and Sappho […] called Hesperus a star […] explaining the etymology […]: Hesperus, bringing everything which shining Dawn scattered). This source has its own oscillations, since a cod. B for this has φαινόλης, which is either a paenula (thick upper garment) or a doric form for the plural of φαινόλις, therefore we discard it and keep the other codices' form, with Aeolic barytonesis applied, to produce φαίνολις. Also, cod. M for this source has αὔων, but that is accusative AFAIK, and we don't want an accusative here, so we follow the other code. Said codex also has ἔσπειρε at the beginning, which could be construed as the imperfect of σπείρω, thus meaning "he/she/it sowed", which is out of place, so we discard it. We thus give this line the form Ἔσπερε πάντα φέρων ὄσα φαίνολις ἐσκέδασ' Αὔως. And now it's a hexameter.
- Source 3 is the Etymologicum Gen[uinum] A, a passage more or less the same as one from the Etymologicum Magnum, where we read «Σελεύκου· ἕσπερος· [...] ἀπὸ τοῦ ἔσω περᾶν πάντα τὰ ζῷα [...] καὶ Σαπφώ· Ἕσπερε πάντα φέρων ὄσα φενόλις ἐσκέδασ' Αὔως» (Of Seleucus: Hesperus: [called thus] from all the animals going inside […] and Sappho: Hesperus bringing everything which (shining?) Dawn scattered), where Seleucus may be the author of this particular entry (the Etymologica are essentially etymological dictionaries), and φενόλις is probably a corruption of our friend φαινόλις, obviously to be Aeolically accented. So the same form is given to l. 1 as in source 2.
- Source 4 is the Etymologicum Gen[uinum] B, the only other source with l. 2. It reads «Σαπφώ· Ἕσπερε πάντα φέρων ὄσα φαινόλης ἐσκέδασ' Αὔως φέρεις οἶον φέρεις οἶνον φέρεις αἶγα φέρεις ἄποιον μητέρι παῖδα», oh my God what happened to l. 2? The thing translates to «Sappho: Hesperus bringing everything which shining Dawn scattered you bring the lonely man you bring the wine you bring the goat you bring the absent child to the mother». Apart from amending φαινόλης the usual way, all we will do is fix φέρεις, again as usual, and to Aeolicise μητέρι to μάτερι.
- Source 5 is the Etymologicum Gud[ianum], where we read much the same as in source 3, except for having the line as Ἕσπερε πάντα φέρων ὄσα φαινολὶς ἐσκέδασ' Αὔως, where we amend φαινολὶς in terms of accent getting the same old form for the line as all sources save for 1.
- Source 6 is another passage of the Etymologicum Gud[ianum], again similar to source 3, and it reads «ἑσπέρα ἀπὸ τοῦ σαφῶς εἰς πέρας πάντα φέρειν, καὶ Σαπφώ· Ἕσπερε πάντα φέρειν ὄσα φαινολὶς ἐσκέδασ' Αὔως, where we amend φαινολὶς as usual, and don't know what to do with φέρειν: φέρων or φέρῃς? The former is less distant, so we'll go with that, getting the same old form for l. 1.
- Source 7 is the Etymologicum Gen[uninum] A, bearing the exact same passage as the Etymologicum Magnum (not the one above though), which reads «Σαπφὼ [...]· Ἕσπερε πάντα φέρων ὄσα φαινόλης ἐσκέδασ' Αὔως», simply "Sappho said" and the quote, and you know the drill by now, and nothing new comes from here. Note that only the E.G. and cod. V of the E.M. have φαινόλης, while other codices have φαινολὶς.
- Source 8 is Cramer's Anecdota Or., which reads «ἠώς· ἤτοι [...] ἢ ἀπὺ τοπυ διεξιέναι τοὺς ἀνθρώπους· Σαπφώ· ὅσα φαινώλης ἐσκέδασ' Αὔως» (ἠώς, Dawn: either […] or from men going out; Sappho: all which shining Dawn scattered», and you know the drill. Except the line is not complete, but whatever.
- Source 9 is again the Etymologicum Gud[ianum], which reads «Σαπφώ· Φέσπερε πάντα φέρω ὄσα φαινοτέρα ἐσκέδασ' ἡώς» (Sappho: Hesperus I bring everything which the rather bright Dawn scattered). That is some change. The leading phi is obviously spurious, and may indicate the presence of a digamma in the original. φέρω is an interesting variation, but it creates a hiatus which seems unlikely, especially since that would probably cause a shortening of the vowel which would be inmetrical, so we amend it as usual. φαινοτέρα is inmetrical, and we'd surely like to amend it as usual, but as far as this source is concerned, a more conservative thing is to elide it to φαινοτέρ'. The last word is obviously non-Aeolic, so we Aeolicise it as usual. And so we get Ϝέσπερε πάντα φέρω ὄσα φαινοτέρ' ἐσκέδασ' Αὔως.
- Source 10, finally, is another passage of the Etymologicum Gud[ianum], which reads, without an author name, «ὅσα φαινόλης ἐσκέδασ' ἄνθρωπος» (which shining man scattered), where the last word is obviously a spectacular corruption, as it is inmetrical and "shining man" is just nonsense. So after the usual fix to φαινόλης we're back to the usual point.
So l. 1 is pretty much undoubted, except for the outsider of source 9,which we promptly reject, and φέρῃς vs. φέρων. The tradition seems to indicate the latter, since only source 1 has the former, but the former would strengthen the anaphora Demetrius praises. This is probably why everybody outside Edmonds and safopoemas has the former. I beg to differ. I mean, an overwhelming majority of sources indicates the latter, so what is more likely, that the anaphora was only in l. 2, or that 8 sources out of 10 made the same mistake? However, I followed Bibliotheca Augustana for the translations, so I'll have to keep the former for this post, but the all-Sappho posts will get the latter.
Line 2, however, is a huge mess. We have two sources, and they give us «φέρεις οἶνον φέρεις αἶγα φέρεις μάτερι παῖδα» (I) and «φέρεις οἶον φέρεις οἶνον φέρεις αἶγα φέρεις ἄποιον μητέρι παῖδα». What do we make of all this?
- Bergk, who is btw the only one to keep the digamma in l. 1 (I won't either, because only one source has a hint to it), raises his hands and says «quæ frustra restituere conantur» (which in vain people try to reconstruct). – Lobel-Page, Campbell, and Voigt all do more or less the same: φέρεις ὄιν φέρεις αἶγα φέρεις ἄπυ μάτερι παῖδα, with Voigt having no cruces, treating this as a iambic meter + xx–uu–uu–uu–x (pherecratean with two dactyls inserted), and amending to φέρῃς all times, Lobel-Page wraps the whole thing up to and including ἄπυ in cruces, and Campbell putting the closing crux after the second φέρεις.
- Bibliotheca Augustana and Greek Wikisource and The Complete Poems of Sappho do the same, disregarding φέρῃς-φέρεις oscillations, with the first two breaking l. 2 into two lines with the breaking point after ὄιν.
- Edmonds writes φέρεις ὄιν, / φέρεις αἶγα, φέρεις ἀπὺ ϝὸν μάτερι παῖδα, suggesting perhaps ἀπὺ ϝϝὸν on the basis of some Homeric quotes where a third-person adjective/pronoun with no consonant at the start (like ἕο) lengthens the omicron in the preceding ἀπὸ. This is just turning a metrical problem into a worse one, since l. 2 (or ll. 2-3) is still not a hexameter, and the meter is even less recognizable than before. The double digamma is just weird. So screw Edmonds.
- Finally, safopoemas has the whole thing as «/έσπερε πάντα φέρων δσα φαΐνολις ¿σχίδασ' αδως φέρρες δϊν, φέπες αίγα, φέρες (δ') α"πυ μάτερι παΐδα.». Apart from observing φέρων and the slash that could be a digamma in l. 1, let's look at l. 2. It is now a hexameter, presumably reading «Φέρρες ὄιν, φέρες αἶγα, φέρες ‹δ'› ἄπυ μάτερι παῖδα». This makes it a past action, an imperfect, to be precise. While this is nice because we have a hexameter with the anaphora, the doubled rho in the first word is totally unjustified AFAIK, though I will ask Stack Exchange about it. I mean, why would it be double? And on the other hand, the absence of a critical note seems to suggest Reinach had that text, so what reason did he have for that? So I'll put it on standby for the all-Sappho posts untill Stack Exchange answers me.
In this post, however, I must follow the text I translated, and I dismissed this reading, both because of the unjustified double rho, and because I felt like this must be a present action, a habitual one. Now obviously I didn't know anything about the above mess, so I just reordered Bibliotheca Augustana a little, getting the hexameter Αἶγα φέρῃς τε σ' ὄιν τε, φέρῃς ‹δ'› ἄπυ μάτερι παῖδα, the δ' addition being kept because it sounded good. Maybe τ' would be better. Yeah, I'll change it, the translations are left unaffected.
I'm not the only one who tried this: in Voigt's words, «hexam[etrum] restit[uere] conatus est Bowra l.c.: αἶγα φέρῃς καὶ ὄιν τύ, φέρῃς τ' ἄπυ μάτερι παῖδα, sed anaphoram destruxit» (Bowra tried to reconstruct a hexameter in the quoted place: [reconstruction], but destroyed the anaphora). Now, while the anaphora is certainly weaker, I wouldn't say it's "destroyed". Not sure why τύ and not σύ. I'll keep my version, despite Bowra's translating much the same, because the anaphora is stronger.
Finally, some names to emendations:
- ὄιν instead of (οἶον φέρεις) οἶνον is "Fab. Benevolentius ap[ud] Ursin[um]", probably arguing that φέρεις οἶον was an accidental and slightly corrupted repetition of φέρεις οἶνον and that -ον again was an accidental and corrupted repetition of just οιν; also, sheeps and goats fit together and are brought back in the evening, wine is an intruder and how is it brought back in the evening?
- ἄπυ is "Bgk.", and seeing Bergk amend this surprises me: if you think it's a vain effort to try reconstructing this, why amend it? The recessive accentuation is probably justified since this is an adverb and not a preposition, which even in Aeolic is accented on the final syllable.
- The addition of δ' is by Sitzler, and the breaking found in B.A. and G.W. is by Wilamowitz.
Wow that took forever! Oh look, I had the τ' addition already! I wonder why…
Told you it was huge. Btw, bonus question: are Et Gen and Et Gud Etymologycum Genuinum and Gudianum respectively, or is this disabbreviation incorrect?