George Hinge


Dialect Colouring in Quotations
of Classical Greek Poetry



0. The testimony of the indirect transmission is often ignored altogether or at best considered inferior, when the linguistic form of classical poetry is being evaluated. The relative ignorance of the copyists and the large amount of corruptions in the fragmentary quotations are indeed an impediment to the full exploitation of this material. However, I shall try in the present article to throw some light on the performance of classical poetry starting from two different situations of indirect transmission: a quotation of Alcman in the grammarians Apollonius Dyscolus and Herodian and an apparent allusion to Stesichorus in the 4th book of Herodotus.

1. The papyrus fragments of Alcman all show the expected “Doric” α for original long ā, and the same holds true for most of the quotations as well. There are in the scattered fragments c. 200 instances of α for *ā. One is therefore tempted to exclude the non-Doric η altogether from the Hellenistic edition of Alcman, and in fact all modern editors carry out the α vocalism in all fragments. There is however a small amount of quotations that appear to have Attic-Ionic η:

30 P κεκλήγη, ἡ (bis), 64 P Εὐνομίης, ἀδελφή, Προμηθείας, 68 P μέμηνεν, 77 P βωτιανείρηι, 80.1 P Κίρκη, 89.5 P πορφυρῆς, 102 P ἀνάγκη, 106 P βροτήσια, 114 P ἡγεῖται, 115 P γυνή, 119 P δῆμος, 126 P ηὔλησε, 135 P ἡδυμέστατον

The question is whether we can discard these examples as secondary in the sense that they are due to the unreliable transmission of the fragments in question, as the editors do, or we have to admit their presence or at least the presence of some of them in the Hellenistic edition. All fragments with spurious η are quoted in late or even Byzantine authors, normally through intermediary sources, and since the scribes responsible for the transmission of most of these texts were not at all prepared for passages in another dialect, the quotations are notoriously unreliable.

Interference with the Koine cannot, however, be an explanation in fr. 77 P βωτιανείρηι and 64 P Εὐνομίης: After r and i, Attic and Koine have preserved the α vocalism, and only East Ionic shows the advanced η vocalism in this context. Fr. 64 P is in its linguistic form perfectly Ionic: Εὐνομίης τε καὶ Πειθοῦς ἀδελφὴ καὶ Προμηθείας θυγάτηρ (only είας for είης). Plutarch, the author responsible for this fragment, is normally not afraid of presenting Laconian verses in the vernacular (cf. fr. 41 P and 57 P), but the dactylic metre may of course have altered the transmission. Fr. 30 P is quoted in a dialect-neutral form: ἡ Μῶσα κεκλήγη/ ει ἡ λίγεια Σειρήν (except for the typical Μῶσα). Our informant Aristides has elsewhere a fragment of Alcman in a perfectly Doric form (fr. 127 P). Hephaestion, who quotes fr. 119 P with an η, keeps the Doric α in 27 P Καλλιόπα and 50(b) P ἅν (beside more specific dialectal features: σαλασσομέδοισ᾿, μασδῶν), and Strabo has η in 126 P ηὔλησε, but α in 16.5 P ἀκρᾶν, 55 P ἱμερτάν, περιρρύταν, and 98 P παιᾶνα.

2. There is at least one example, which is not easily dismissed without suggesting some codical conspiracy: Fr. 115 P is transmitted in two independent traditions, both of whom showing the Attic-Ionic η. The fragment is quoted not only by the 2nd cent. AD grammarian Apollonius Dyscolus (Adv. I 193):

... ἔστι δὲ καὶ ὑποδείγματα τῶν <διὰ> τοῦ τε τὸ πότε πότα, ὅτε ὅτα, ἑτέρωτε ἑτέρωτα, ἄλλοτε ἄλλοτα. διαφέρει δὲ πάλιν τὰ Αἰολικὰ τοῦ Δωρικοῦ, ᾗ τὸ μὲν Δωρικὸν τὸ τ εἰς κ μεταλαμβάνει, ὅτε τὸ πότε πόκα ἐστί, τὸ ἄλλοτε ἄλλοκα, ὅτε ὅκα, καὶ μετὰ περισσοῦ τοῦ κ, «ὅκκα δὴ γυνή».
“Other examples of derivations in τε are πότε πότα, ὅτε ὅτα, ἑτέρωτε ἑτέρωτα, ἄλλοτε ἄλλοτα. Again there is a difference between Aeolic and Doric, as far as Doric changes τ to κ: πότε is πόκα, ἄλλοτε is ἄλλοκα, and with an extra κ, ὅκκα δὴ γυνά.”

but also in the Byzantine lexicons Etymologicum genuinum and Etymologicum magnum:

«ὄκκα δὲ τύνη εἴην»· παρ᾽ Ἀλκμᾶνι. τὸ ὅτε ὄκα λέγει ἡ διάλεκτος, εἶτα διπλασιάσασα ὄκκα. Περὶ παθῶν.
“ὄκκα δὲ τύνη εἴην: in Alcman. Instead of ὅτε, the dialect has ὄκα and then with gemination, ὄκκα. On word forms.”

The transmitted form τύνη is a palaeographically trivial misinterpretation of γυνη. The lemma ends with the reference Περὶ παθῶν, i.e. it is excerpted from the 2nd cent. AD grammarian Herodian’s fragmentary work ”On word forms” (= Hdn. II 302 Lenz). As Aelius Herodianus happens to be the son of Apollonius Dyscolus, the matter is even more delicate. Both grammarians quote numerous Alcman fragments in their oeuvres. Thus, Apollonius presents 33 quotations and Herodian 24, each no more than one line or just half a line. This is by no means natural, as Alcman was not part of the average syllabus in Roman times.

In general, these two grammarians are faithful to the linguistic form of the fragments quoted. One possibility is therefore that they did in fact write a Doric γυνά, but that it has been corrupted in the transmission. That model cannot of course be ruled out a priori, but it is less economical than the assumption of a common source of the innovation (as I shall argue below, it is an innovation). If Herodian has borrowed the quotation from the work of his father, the form γυνή must go back to Apollonius’ own hand, as it is less likely that Herodian consulted a corrupted manuscript of his father. However, the excerpts of Herodian consist of four words, whereas Apollonius quotes only three words of the fragment. We must therefore assume either that Apollonius wrote in fact four words, but that the last word fell out in the transmission (a not very economical assumption, either), or that the common source was the Alcman text, which Herodian and Apollonius consulted.

Even though the testimony is very sparse, it seems probable that the Hellenistic text of Alcman did in fact have some instances of Attic-Ionic η. If that holds true, we have to ask ourselves how it did enter the text in the first place.

Some scholars ascribe the spurious η in the dactylic hexameters fr. 77 P Δύσπαρις Αἰνόπαρις κακὸν Ἑλλάδι βωτιανείρῃ and fr. 80 P καὶ ποτὲ Ὀδυσσῆος ταλασίφρονος ὦτά θ᾿ ἑταίρων Κίρκη ἐπαλείψασα to an epic colouring. Fr. 119 P ταῦτα μὲν ὡς ἂν ὁ δῆμος ἅπας is according to Hephaestion an ἑφθημιμερές, i.e. it goes like a hexameter until the 7th position. On the other hand, the Doric α is preserved in other dactylic hexameters, e.g. 26.1 P μελιγάρυες, 28 P ὠρανίαφι and 56.6 P Ἀργεϊφόντᾳ - all of these words are more epic than vernacular. The hexametric fragments show in general the severe Doric open ω, too, instead of the Attic-Ionic closed ου, e.g. 28 P Μῶσα, ὠρανίαφι and 80.1 P ὤ{τ}αθ᾿.

The epic vocabulary is not characterised with a non-Doric vocalism either: e.g. 1.5 P κορυστάν, 1.48 P παγόν, καναχάποδα, 1.62 P ἀμβροσίαν, 7.4 P ἀσανάτας, 59(a) P ἕκατι. Thus, neither metrics nor phraseology was a sufficient parameter for choosing an Ionic pronunciation. The passages, in which we find Attic-Ionic η, are not necessarily more epic than the passages, in which Doric α prevails, and the passages, which are in fact more epic than others, do not necessarily show η, but more often α.

3. To my opinion, the testimony of the archaic and classical verse inscriptions is decisive. They are in fact the only direct source to the phonetics of archaic poetry. The dactylic hexameters and pentameters were not marked with an Ionic η either in Laconia or in the other non-Ionic cities, and the older Attic verse inscriptions do not allow η after r, i, e as well. It is therefore improbable that Alcman and the other poets of the mainland marked dactylic metre and epic phraseology with an East Ionic pronunciation, as long as the West Ionic Athenians did not.

Nevertheless, we seem to have the Ionic η present in the transmitted Alcman text. The only solution is to accept that the fragments ascribed to Alcman in antiquity were not all performed and transmitted in the same way. One may distinguish a group of ”normal” and in Hellenistic times more popular fragments, like the famous Louvre Partheneion, and a group of more or less abnormal fragments, which are not transmitted but in quotations, probably because of their shape and content. Among the abnormal fragments, we have several dactylic hexameters, a metre totally absent from the Egyptian papyri. However, they might very well have been included in the six volumes Collected Works of the greater libraries.

In classical Athens, the poetry of Alcman appears to have been practically unknown except for a couple of popular songs like the famous fr. 26 P about the halcyon, in dactylic hexameters; it shows in fact ”Doric” α, but the linguistic form is otherwise neutral (the α is itself neutral in the classical dialect geography). I suppose that bound up as it was with the cultic context most of Alcman’s poetry was of such character that no one bothered to export it to other Greek cities in classical times. A small amount of poems, or parts of poems, was however more appealing. Those early exported verses were neutralised linguistically or even Ionicised. In post-classical times, antiquarians got interested in the more difficult poems, and the cult songs (such as the Partheneions) were recorded according to the epichoric performance.

The conjunction ὅκκα in fr. 115 P cannot be described as Ionic or linguistically neutral, it is true. Probably it has resisted the adjustment to the Ionic performance, because it could not be replaced by the Attic-Ionic ὅτε (or ὅταν) without altering the metre. But the very coexistence of ὅκκα and γυνή in the transmission demonstrates that we are dealing with two layers in the performance, just like the Ionic epic shows vernacular forms along with metrically unmatched archaisms and dialect features

4. Let us turn to another author: In the excursus on Scythian ethnography, Herodotus refers different myths about the origin of the Scythians. The Greek version starts with Heracles (Hist. 4.8): Ἡρακλέα ἐλαύνοντα τὰς Γηρυόνεω βοῦς ἀπικέσθαι ἐς γῆν ταύτην ἐοῦσαν ἐρήμην, ἥντινα νῦν Σκύθαι νέμονται ... “Having the oxen of Geryon with him, Heracles came into this country, which was then empty, but is now inhabited by the Scythians...” Whereas most manuscripts (A, B, C and T) have the proper Ionic Γηρυόνεω βοῦς, one manuscript D (hence R, S, V) has Γηρυόναο βόας. The editions of Hude (Oxford) and Legrand (Budé) print the regular form, but Rosén (Teubner) - and with him Medaglia (Mondadori) - prefers Γηρυόναο βόας, obviously the lectio difficilior, since it is not in accordance with the dialect of Ionic prose. He suspects an allusion to Pindar, who has the same two words in the same order in fr. 169 S.-M., v. 6.

That Herodotus did in fact know Pindar’s poetry is stated with the reference to Hist. 3.38.4, where the father of history is quoting him explicitly: οὕτω μέν νυν ταῦτα νενόμισται, καὶ ὀρθῶς μοι δοκέει Πίνδαρος ποιῆσαι νόμον πάντων βασιλέα φήσας εἶναι “Such are the customs, and Pindar says - rightly, I think - that Custom is the king of all”. With the words Νόμος ὁ πάντων βασιλεύς etc., Pindar introduces in fact the very same choral song, which speaks about the theft of the Geryonic oxen, fr. 169 S.-M. In the Herodotean variant Γηρυόναο, we would have an explicit testimony for an Ionic η in the text of Pindar not much later than the poet himself. The name is derived from γῆρυς, γηρύω, which is spelled regularly with α in Doric and Aeolic texts (e.g. Pind. O. 1.3, 2.88, Alcm. fr. 26.1 P, IG 5(1).315.4), and is itself testified with α: Stes. S13.4 P Γ]αρυόνα, S15 P Kol. II.14 Γαρ[υόνας and Ibyc. S176.18 P Γαρυόναν.

In the text of Pindar, Γηρυόνα is transmitted unanimously: Isthm. 1.13, fr. 81 and fr. 169.4 S.-M. Fragment 81 and 169 are quoted in Aristides and in the scholia to Aristides, and both sources have η. Isthm. 1.13 is transmitted in two manuscripts and in the papyrus commentary P.Oxy. 2451 col. i.32, and there, too, the lemma has Γηρυ[όνα. The Pindar text apparently had Γηρυόν- in all instances. Bernhard Forssman argues that Pindar has coloured the epic vocabulary with an Epic-Ionic η, because the name in question was known to him primarily through epic sources.

5. After the discovery of large fragments of Stesichorus’ monumental choral epic the Geryoneid, it has however become evident that the Stesichorean version of the myth about Geryon was considered the classic version in antiquity: Apollodorus, Bibl. 2.10, seems in his narrative to follow this poem in most details, and Pindar is depending on Stesichorus too in fr. 169; it is no coincidence that the vase paintings show an immense interest in the story in the period right after Stesichorus. It is therefore less obvious than Forssman thought, that Pindar should have marked the name of Geryon with an epic η, and less obvious, too, than Rosén thought, that Herodotus should quote Pindar on the same matter.

In the Pindar verse in question, we have not Γηρυόναο βόας as in Herodotus, but Γηρυόνα βόας. The papyrus P.Oxy. 2450 has only ] βόας, but the scholion to Aristides (1.1.3, 408 Dind.), who quotes the first six verses, has Γηρυόνα βόας that is also supported by the responsion. If Herodotus does allude to Pindar, why has he not recorded an exact quotation of Pindar? Of course, the historian may in fact have written *Γηρυόνα himself, but the transmission has corrected this unusual genitive to αο, which is more common to a copyist educated in the Homeric idiom.

Another possibility is however that Herodotus did not have Pindar in mind in the 4th book. He knows the initial words of the poem in question, but they may have become a favourite refrain already. In Gorgias 484b-c, Plato quotes the first five verses, and he refers explicitly to the same verses in Leg. 3, 690b-c, 4, 714e-5a and 10, 890a. Herodotus must not have known the exact wording of the continuation with Geryon. Though the maxime about King Nomos was fancied, Pindar’s song has not exploited the Geryon legend, the mythic exemplum, exhaustively or authoritatively. It would therefore be meaningless for Herodotus to quote this song of all and only very implicitly. It would make much more sense to allude to Stesichorus’ Geryoneid that has influenced the art and poetry of the following period.

In the extant fragments of the Geryoneid, which come mostly from P.Oxy. 2617, we have only Γ]αρυόνα (without βόας) in the genitive case. It is however no audacious conjecture, that *Γαρυόναο βόας was somewhere in the poem. The oxen are only mentioned once, in fr. S11, v. 27 περὶ βουσὶν ἐμαῖς, probably in the direct speech of Geryon, but they must have mentioned in numerous places in the more than 1300 verses long poem. For the genitive in -αο cf. 222 col. i.9 P Εἰλατίδαο.

6. We have at least the a-declension in both Stesichorus (S13.4 P Γ]αρυόνα) and Ibycus (S176.18 P Γαρυόναν), whereas the dactylic epic has a derivation with the suffix -εύς (Hes. Th. 287 Γηρυονῆα, 309 -ῆι, 982 -έα / -ῆα), a form, which is unusual outside the epic, cf. Plt. Euthd. 299c Γηρυόνην, Gorg. 484b Γηρυόνου, Leg. 795c Γηρυόνου, Aristoph. Ach. 1082 Γηρυόνῃ. In the passage in Herodotus, both the majority variant Γηρυόνεω and the lectio difficilior Γηρυόναο are a-stems; the first is the ordinary Ionic form, the latter a common poetic archaism (cf. Myc. -a-o). A little later in the same paragraph, we have the a-stem accusative Γηρυόνην in the same manuscript branch that has Γηρυόναο too, but Γηρυόνεα in the other manuscripts. A eu-stem accusative would be inconsistent with the a-stem genitive of both manuscript traditions, and Herodotus may therefore be alluding to Hesiod Th. 982-3 Γηρυονέα τὸν κτεῖνε βίη Ἡρακληείη βοῶν ἕνεκ᾿ εἰλιπόδων ἀμφιρρύτῳ εἰν Ἐρυθείη. On the other hand, the manuscripts have Γηρυόνεα, and Herodotus have numerous instances of paroxytonic -εα as a heteroclitic pre-consonantal variant of -ην.

Herodotus did know Hesiod (cf. Hist. 4.32, referring, as it seems, to Cat. fr. 150 M.-W.), but he nowhere mentions Stesichorus. However, since he lived for several years in Magna Graecia (as a citizen of the colony Thurii), it is by no means improbable that he was familiar with his poetry. Since the phrase Γηρυόναο βόας must be due to some poetic model, and the genitive form seems to be incompatible with the Ionic epic tradition, the most obvious proposal is Stesichorus.

If that holds true - and I must admit that the evidence is anything but decisive - the Ionic η of the Herodotean Γηρυόναο may be a testimony to the performance or at least to one performance of Stesichorus’ poetry in the 5th cent. BC - and taking into consideration that the paradosis of Pindar has chosen η in the very same word already in the papyri, we have a tiny piece of evidence there too for an Ionic performance of Stesichorus in classical times. In that case, there is a discrepancy between the 5th cent. performance with Γηρυον and the papyrus tradition; P.Oxy 2617, with the Geryoneid, is from the 1st cent. AD, but also the oldest papyrus P.Lille 73, from the 3rd cent. BC, has carried the α vocalism and other Doric features through consequently. That the Doric form of Stesichorus’ poetry may in fact be post-classical is also supported by the fact that Himera, the city, in which he wrote his poems, seems to have been Ionic-speaking in the 6th cent. BC.

8. In Archaic Greece, which was still basically an oral culture, the poetry was probably performed with the pronunciation that happened to be predominant in the city in question, as it is evident from the verse inscriptions. Thus, we have the η vocalism in the Ionic-speaking cities and α elsewhere. In classical times, and afterwards, certain performances were fixated for certain genres. The phenomenon of genre dialects is the consequence of developed literacy. The dactylic hexametric poetry decided on an Ionic surface, since Homeric Epic developed in an Ionic-speaking milieu, whereas the choral lyric decided on a more neutral Greek surface with the α vocalism that was predominant in most of Greece until Hellenistic times.

Some portions of Alcman slipped out of the Laconian context, and they were eventually performed with an Ionic pronunciation. Stesichorus was known by Herodotus and Pindar, it seems, in Ionic, but the Hellenistic papyri have recorded a Doric performance. In the meagre evidence of Apollonius’ and Herodian’s quotations of Alcman and Herodotus’ allusion to Stesichorus, we discern but a shadow of the original variety, which yielded eventually to the uniformity of mass-published books.