Portal de Periódicos da CAPES

The Dialect Basis of Choral Lyric and the History of Poetic Languages in Archaic Greece

2013; Taylor & Francis; Volume: 87; Issue: 1 Linguagem: Inglês

10.1080/00397679.2013.822726

1502-7805

Борис Маслов,

Ancient Near East History

AbstractAccording to a widely accepted view, the dialect used in Archaic Greek choral lyric is predominantly Doric and is strictly correlated with this genre. This article reconsiders the linguistic data as well as earlier attempts to dispute this scholarly consensus (which included interventions by Antoine Meillet, Carlo Pavese, and Natan Grinbaum), and puts forward an alternative theory of the nature and the origins of the distinctive idiom used in choral lyric. It argues that choral lyric preserves an ancient poetic language which predates the Homeric epic and in most likelihood arose in the (Proto-)Aeolic realm. This account, on the one hand, makes it possible to correlate the history of cult-embedded lyric with the history of narrative epic, which is generally held to include an Aeolic phase. On the other hand, it explains the progressive Ionization of the language of choral lyric by the rise of the Ionic poetic language during the 6th-5th c. BCE. AcknowledgementsI am grateful to Leslie Kurke, Anna Morpurgo, Nikolaos Papazarkadas and Viktor Zhivov for their advice at various stages of preparation of this article, as well as to the audiences of the Classics Department seminars at the University of Oslo and at the RGGU in Moscow for their insightful comments.Notes1. Alexander Veselovsky discusses cases of the use of foreign languages in cultic contexts, as well as in poetic insets in narrative texts (Veselovskii Citation2006, 205-6; cf. Meletinskii Citation1986, 33-34)2. Originally formulated in Meister Citation1921. According to a minority view, represented, e.g., by Geoffrey Horrocks (Citation1997, 214-217), Aeolic forms are to be explained by infiltration into a dominant Ionic tradition. For recent discussions of the evidence, see Haug Citation2002, esp. 151-162, Nagy Citation2011.3. To quote Gregory Nagy's formulation, “Homeric diction defaults to Aeolic forms when it has no metrically equivalent forms in Ionic” (Citation2011, 144). There are examples of the use of metrically equivalent variant forms, like Aeolic αἰ and Ionic εἰ ‘if’; on closer examination it turns out that in our text of Homer αἰ is preserved only in certain set combinations (with κεν [Aeolic equivalent of ἄν], γάρ, and as αἴθε), where it is also being replaced by εἰ (statistics in Nöthiger Citation1971, 5).4. To the same effect: “These regional traditions may be defined initially in terms of dialect. Broadly speaking, the Archaic Greek poetry known to us is composed in three dialects: (i) Ionic.. (ii) Lesbian (iii) … Doric” (West Citation1973, 181). In West's account, it is the “Doric” dialect of choral lyric that proves the existence of an independent regional Dorian poetic tradition. Ignoring extensive scholarly work on this subject, West refers to just one shared grammatical feature (fem. participle ending -οισα) to demonstrate the commonality of the dialect of his “Dorian poets” (Citation1973, 181, Citation1982, 46); in fact, this is a feature of the Lesbian branch of Aeolic. There is no evidence to support West's view of the “Dorian-Ionian” tradition as an inheritor of a “South Greek” tradition “differentiated from the Aeolic” (Citation1973, 183-4).5. Cf. examples from Cassandra's lament (she speaks in dialogue with the chorus) in Aeschylus's Agamemnon: ἀλκά (1103), τύχαν (1129: Weil emends to τέχναν), τύχᾳ (1136). Note also that Greek tragedians only used the non-Ionic form of the adjective δαρός, including in dialogue (Eur. IT 1339, [Aesch.] PV 648, 940).6. E.g.: Buck Citation1955, 15; Forssman Citation1966, xi: “Zur Sprache Pindars – und der übrigen griechischen Chorlyriker – haben mehrere Dialekte beigetragen. Besonders klar erkennbar sind bei Pindar die dorischen Elemente.” Verdier Citation1972, 9: “Le fond de la poésie de Pindare est composé par le dorien; cela n'a rien d'étonnant puisque c'est la langue de la lyrique chorale, par tradition.” To the same effect: Palmer Citation1980, 119-30, Cassio Citation2005, Colvin Citation2007, 54-55.Watkins (Citation1995, 59) describes the language of Alcman and Stesichorus as “a literary Doric dialect”, of Pindar and Bacchylides as “a mixed literary Doric and Aeolic,” and of choral passages in Attic drama as “literary Doric”.7. In fact, the evidence on the dialects of Himera and Rhegium, the home poleis of Stesichorus and Ibycus respectively, is far from certain (Nöthiger Citation1971, 1-3). A. C. Cassio has demonstrated that in the time of the poets the dialect spoken in these Euboian colonies was Ionic, and that they were Doricized in a later period (1999, 204-207). West views Eumelos as the earliest representative of the Dorian tradition, but his only fragment may well be a late forgery (D'Alessio Citation2009, 137-145).8. Gentili Citation1988, originally published in Italian in 1985, quotes an early article by Grinbaum (Citation1968) in extenso and concurs with him in claiming that the choral poetic tradition has its roots in “an ancient Mycenaean koine containing both Aeolic and Ionic forms” (60). (The notion of Mycenaean as a koine was advanced by Georgiev [e.g. Georgiev Citation1972] but is not shared by the majority of Mycenologists. The more widely accepted view is that Mycenaean writing, since it was used over a wide linguistic area, from Crete to Thebes, can document occasional dialect differences.) In his later work, not referred to by Gentili, Grinbaum posited an “Ionic” basis for the point of origin of choral language and an additional Aeolic substratum, as it became clear that Mycenaean could not be assumed to have a significant Aeolic component. The pattern of discussing only Grinbaum Citation1968 is observed in Palmer Citation1980, 129-30, Bowie Citation1981, 59, Trümpy Citation1986, 18-25, Hinge Citation2006, 340-1.9. García's summary of the negative conclusions of Grinbaum and Pavese: “… the classification of [a shared fund of ‘dialect features’], on linguistic principles, as ‘Doric’ is misleading, for in most cases it could be classified in other ways” (Citation1998, 137). Cf. the conclusion of Trümpy's monograph: “Wir sind wiederum zu einem negativen Resultat gelangt: Ursprünglich dorische Herkunft der Chorlyrik ist aus deren Sprache nicht zu erschliessen” (Citation1986, 126; emphasis in the original).10. See Peters Citation1986, Haug Citation2002, 39-69, Maslov Citation2011.11. Grinbaum eschews Risch's division of Greek dialects into a Northern and a Southern branch (Risch, based on the lack of assibilation, assigned Aeolic to the northern branch, although it patterns with Mycenaean in other respects), in favor of an older dialect division, recently resurrected in Finkelberg Citation1994, which places Ionic and Aeolic closely together.12. Notably, in his later work Grinbaum spoke of two strata of this koine – older Aeolic and younger Ionic, which calls to mind the constitution of Homer's dialect. See Grinbaum Citation2008, 106 (on the evolution of Grinbaum's theory cf. Kazansky Citation1997, 17 n. 36).13. David Goldstein (Citation2006) makes interesting observations on the distribution of some variant forms in Pindar based on phonetic environment. More work in this direction is clearly a desideratum.14. For a similar assessment of Grinbaum's theory, see Kazansky (Citation1997, 17 n. 36) who views the formation of choral lyric dialect as part of “the general process of Panhellenic cultural development”.15. Another scholar writing on Pindar's dialect in the early 1970s, Michel Casevitz (Citation1972) notes the apparent avoidance of Doric forms in Pindar. According to Casevitz, this pattern is due to Pindar's avoidance of forms that are used in his local dialect, Boeotian (a dialect commonly classified as Aeolic but sharing many features with Doric). For the argument that Pindar in general avoids characteristic dialectisms, cf. Christ Citation1891, 58-62.16. On apparent examples of Doric futures due to editorial choices in the accentuation in the text of Ibycus and Stesichorus, see Cassio Citation1999.17. Admittedly, Pindar in all likelihood did not mark the phonetic distinction between a closed (Ionic) and an open (Doric) [ō]; the choice of later transcribers or copyists is likely to be significant. See n. 28 for further discussion.18. Christ Citation1891, 58. Cf.: “Les traits qui caractérisent les parlers occidentaux, et en particulier le dorien, sont ou employés sans constance ou non employés […] Si la langue de la lyrique chorale n'est pas ionienne, elle n'est donc pas non plus dorienne et, sauf l'emploi sporadique de l'aoriste en -ξα [this is not probative, since the variation is also found in Homer: Risch Citation1955.73 – B. M.] on n'y trouve rien qui mérite le nom de dorien, pour peu qu'on attribue à ce nom une sens linguistique défini” (Meillet Citation1975.211). Even as he recognized that choral lyric was written by non-Dorians and even pointed to its Aeolic origins (see below), Meillet did not abandon the notion of a Doric “base” of the choral dialect (Citation1975.209). I believe this was due to Meillet's reluctance to posit archaic features, which can be deemed Proto-Aeolic (such as the 3rd p. pl. ending -οντι), as an intrinsic part of the choral idiom.19. Pavese, improbably, suggests that this is an independent Pindaric innovation (Citation1967, 174).20. Similarly, in the case of γλέπω (once in Alcman) = βλέπω and its derivate γλέφαρον = βλέφαρον we are dealing with a dissimilatory loss of the labial element, which is common across Greek dialects (cf. Trümpy Citation1986, 123), even though these two particular forms are attested only in Doric. Pindar only uses forms of βλέπω, but the deverbal noun and related adjectives appears only with γλέφ-, suggesting that those were perceived as elements of poetic diction.21. For a detailed discussion, see Nöthiger Citation1971, 96-104. The short infinitive ending -εν is a result of innovatory abstraction from the Common Greek infinitive ending -εεν (Sihler Citation1995, 608).22. Originally, there was a variation between short accusatives that preceded words beginning with a consonant and long accusatives that preceded words beginning with a vowel (Risch Citation1954, 33; Pavese Citation1972, 35-9).23. Pindar uses ἐν with the accusative in Pyth. 2.11, Pyth. 5.38, fr. 75.1. It is likely that this feature was a Common Greek archaism, preserved in conservative dialects as well as in poetic language (Nikitina and Maslov Citation2013, 132-133). On parallels for pre-verbal ἐν governing an accusative in Attic tragedy, see Troxler Citation1964, 114-5.24. The inscriptional attestation in Thessalian (IG IX.2.270 = SEG 40.473) cited by Grinbaum (Citation2007, 183) is not probative, as it comes from Kierion, whose dialect is known to have many (North)-West Greek elements (Thumb Citation1959, 51-52); the epsilon may also conceal a long vowel; further discussion in Lorenz Citation1976, 41.25. Morpurgo Davies Citation1964, 140 n.1; West Citation1978, 312; in favor of keeping it: Troxler Citation1964, 102.26. Pindar: O. 1.3, O. 3.25, P. 4.56, 115, P. 9.21, Pae. 9.36; μελίζεν in N. 11.18 is an emendation; Bacchylides: 16.18, 17.41 and 88, 19.25.27. Irigoin Citation1952, 28. In favor of emending short infinitives: M. Christ Citation1867, 628-9, W. Christ Citation1891, 35-38, Gerber Citation1982, 15-16.28. It is instructive to dwell on another case of significant variation for which we are dependent on the MSS evidence: the distribution of the forms of the infinitive of the verb “to be”. The most common form in our texts of choral lyric is ἔμμεν (Thess.), whereas Homer's preferred form, also normal in choral lyric, is a secondary development ἔμμεναι peculiar to Lesbian. (Ionic form ēnai, spelled εἶναι, is avoided by choral lyric, but not by Homer.) Crucially, the form with compensatory lengthening ēmen found in Doric and Boeotian is infrequent: it occurs once in Alcman (where it is the only form attested), once in Bacchylides, and in Pindar only as a variant reading (Nöthiger Citation1971, 39-43). The traditional explanation (a Doric basis, plus admixtures from Lesbian and Homeric poetry) cannot account for this distribution, because it leaves the dominance of ἔμμεν unexplained. Instead, we need to posit a Proto-Aeolic basis and an influence of the Lesbian poetic tradition, with minimal admixture from other dialects. In principle, the geminate consonants tended not to be written, and the spelling most likely used by Pindar (ΕΜΕΝ) could reflect both (Thess.) ἔμμεν and (Doric) ēmen. In Irigoin's framework, the predominance of ἔμμεν in our texts would need to be explained by their dependence on a Thessalian edition, which is highly improbable. Most likely, we can trust our textual tradition to preserve ἔμμεν as an archaic poeticism.29. ὅκα (Alcman, Stesichorus, Ibycus), ὅκκα, πόκα (Alcman), τόκα (Pindar). This apparent Doricism is acknowledged as a problem in Pavese Citation1967, 177, Citation1972, 95-6.30. The defixio of Pella shows the “Doric” -ka in the correlative hopoka … [toka] (lines 3-4); cf. the discussion in Brixhe Citation1999, 45.31. In O.1.82, we find the form τά (= interrogative τί), which is only attested in Boeotian (Trümpy Citation1986, 101). The word ἑστία is only attested in Attic (Casevitz Citation1972, 25), which is in all likelihood an accident of preservation. Note too the marginal appearance of the West Greek vocalism in Ἄρταμις, ἱαρός. Pindar uses ἱερός; one dubious exception is fr.338.6. ἱαρός, attested in Thessalian (in a compound), Boeotian, and West Greek, may conceivably go back to a proto-Aeolic source (there is uncertainty as to which form is primary, see Lejeune Citation1972, 239; cf. Bowie Citation1981, 95-6). Ἄρταμις occurs once in Stesichorus (POx.2619.18; Nöthiger Citation1971, 88), also attested in Thessaly (Trümpy Citation1986, 80). Cf. also the form γεγάκειν in O. 6.49 whose root seems to point toward West Greek (Buck Citation1955, 118).32. García Citation1998, 182-3: “Being a facet of the overall ritual performance, the special speech of choral lyric took its place in the conservative and guarded sphere of correct, efficacious behavior that is ritual activity. The language suited to the occasions for ceremonial dance and song was handed down the ages by specialists trained in the conservation of metrical figures, turns of phrase, strophic design.”33. The usage of other poets is also divergent, although the numbers are not significant: Ibycus only has Lesbian Aeolic -οισι (2 instance), whereas Simonides has more “Doric/Thessalian” -οντι than Ionic -ουσι (3 vs. 2). The counts are from Nöthiger Citation1971, 58-60.34. For a discussion of differences between dialect forms of this word in choral lyric (as they are preserved in direct vs. indirect manuscript tradition, as well as in papyri), see Verdier Citation1972, 21-52.35. Note, however that Alcman probably did not have the graphic means of differentiating between a closed and an open long vowel (in Ionic alphabet, the distinction between ω and ου; Risch Citation1954, 24). The spelling Μῶσα probably reflects a performance tradition in Sparta. This usage is confirmed by indirect evidence (see Verdier Citation1972, 24-5), yet we cannot rule out intervention by the Alexandrian editors. In light of Alcman's preference for -οισα in feminine participles, Risch deems his apparent avoidance of the form Μοῖσα “rätselhaft” (Citation1954, 37).36. Note that 5 of the 15 surviving epinikia of Bacchylides were commissioned by Cean victors. By contrast, in the case of Pindar, there are 5 Theban epinikia out of a total 46. This distribution may, however, be due to an accident of preservation (our principal Bacchylides papyrus may derive from an edition with a “Cean” bias). In Ep. 19.13-14, Bacchylides uniquely uses an Aeolic form of the feminine participle λαχοῖσαν in a metapoetic statement, apparently to give it a heavier weight.37. Forssman Citation1966, 98-100; and further on the evidence for Ionic [ē] in Pindar outside P. 4: 101-163. Another important indication is Pindar's preference for the Ionic form παιήων when referring to genre, suggesting a literary-historical motivation (1966, 151-2). Forssman discusses a few other motivated Ionic intrusions, such as σοφίη (possibly directed against the new Ionic “wisdom”) in O. 9.38 (156); the variation between ἁσυχ-/ἡσυχ-, the “Ionic” form being reserved for sociopolitical “calm” (54-5); similar Ionic coloring in εὐνομίη (127-9). Another interesting case is the word μουσικά “poetry sung to music”; Pindar employs it with Ionic vocalism, but having “corrected” the final eta to alpha. The result is an artificial formation that attests, at the same time, to the Ionian cultural prevalence and the persistence of an archaizing poetic idiom. As suggested by inscriptional evidence (discussed in Schachter Citation1986, 150-179), in Pindar's native Boeotian dialect the word would have a long [ō] spelled with omega, as it is spelled in our text of Alcman. In this case, there is perhaps a greater chance that what is at issue is a decision by later editors who might have been more comfortable with the spelling μουσικά rather than μωσικά.38. Contra the earlier view (e.g., Buck Citation1955, 149) that the lack of assibilation is due to the West Greek (Doric) impact on Thessalian. In terms of Finkelberg's presentation, the place of origin of the choral dialect may be defined as the Lesbian-Thessalian-Boeotian dialect area; she dates the existence of this dialect continuum, bordering on Ionic to the south and West Greek to the north, to the post-Mycenaean period, 1200-900 BCE (Citation1994, 25-26). The end-date is determined by different results of the first compensatory lengthening, which preceded the change [ā] > [ē] in Ionic (the latter can perhaps be dated as no earlier than 900, cf. Risch Citation1955, 65, Morpurgo Davies Citation1964, 161).39. Additional arguments in favor of an Aeolic basis of Pindar's dialect come from Forssman Citation1966, In particular, he shows that πονάω and φωνάω – Aeolic, not Doric, forms – underlie Pindar's usage that is conventionally described as “hyper-Doric” (72-4, 81; on these forms cf. Struck Citation1964). Furthermore, in his discussion of variation between the use of the same word with vs. without aspiration in Pindar by reference to either Doric inscriptional evidence or to the influence of Lesbian poets (Lesbian is a psilotic dialect), Forssman acknowledges (33-34) that inscriptional evidence from Doric dialects very often contradicts Pindar's usage. It appears that these forms are Aeolic and – in some cases – specifically Lesbian.40. Notably, verse inscriptions do not display distinctly Aeolic features (Mickey Citation1981, 49) or some common Greek archaisms (such non-assibilated ti-) that are found in choral lyric. The former points to a Lesbian impact on the high poetic language, which bypassed the language of verse inscriptions. As for the lack of assibilation in choral lyric, one possible explanation is that it is an archaism on its way to disappearance; the usage of choral lyric in this case is more conservative than that of verse inscriptions.41. Finkelberg Citation1994, 9. Cf. “probably, the same explanation would hold good also for Homeric τοί/ταί, which in the historic period was found only in West Greek and Boeotian” (29 n. 54).42. As Pavese observes, the commonality of northern Greek features in Homer and choral poets suggests that they are Aeolic, rather than Doric (Citation1967, 169, 178).43. Sappho's fr. 44 provides a crucial piece of evidence in favor of this reconstruction, as this fragment points both to an ancient Aeolic literary idiom and to a pre-Homeric epic tradition (West Citation1973, 191, Nagy Citation1974, 118-39, Bowie Citation1981, 32-40, Trümpy Citation1986, 15-16 with further bibliography).44. As an example of Ionic (and possibly Homeric) influence, cf. the use of the gen. sing. ending -αο (vs. -ᾱ), attested in Homer, as well as Boeotian and Thessalian inscriptions. Pindar only uses it in personal names, many of them with Homeric resonance (cf. Casevitz Citation1972, 24).45. Similar results of the second compensatory lengthening are attested in the Doric dialects of Thera and Cyrene, where they appear to represent recent, independent innovations (Risch Citation1954, 33-4); forms in -οισα also occur in Theocritus, whose dialect is a literary simulation of spoken Doric, possibly informed by the usage in Cyrene. Non-epic aeolicisms in Pindar and other choral lyric are treated in Verdier Citation1972, who concludes “qu'il y a une tradition d'emploi de certaines formes lesbiennes dans ce genre poétique” [= la lyrique chorale], but offers no explanation: “quant au problème de savoir pour quelles raisons existe cette tradition, il n'est pas de notre propos de l'expliquer” (105).46. See p. 13 for counts. The greater consistency in the use of feminine participle is probably due to the influence of personal names that have the same ending.47. Strictly speaking, these are not metrically equivalent, since the -οισι and -ουσι endings allow for an epenthetical -ν; in any case, the forms in -οντι are fully replaceable by Ionic forms.48. A similar development took place in the Hellenistic period, when epigrams written in elegiac couplets started allowing for all possible dialect forms – the precedent of the Homeric Kunstsprache was probably of central importance here as well.49. A fuller discussion of the Northern roots of the Archaic Greek poetic culture can be found in Maslov Forthc. An interesting hint is suggested by an anecdote about Simonides (preserved in Plutarch, De audiend. poet. 15c) who is said to have deemed Thessalians “too uncouth (ἀμαθέστεροι) to be deceived by him.” Van Groningen (Citation1948) suggested that “deceit” stands for the use of myth, but this is unlikely, since myth was probably always present, in some form, in choral cult lyric. In any case, this anecdote appears to suggest that Thessaly was considered to be a region with a conservative poetic culture. For Hermann Koller, the Thracian provenance of the Muses proves the Northern origin of the entire Greek poetic culture: “Die halbmythischen thrakischen Sänger Orpheus, Thamyris, Eumolpos und andere sind die großen Vertreter jenes örtlichen Musendienstes, der […] die Grundlage griechischer Musike und Bildung schuf” (Citation1963, 37).50. It is notable that κίθαρις “kithara” (the only form used in Homer; cf. Ionic κιθάρη) has Aeolic accentuation (pointed out in Nagy Citation1990, 90); the same may be true of the accentuation of ὕμνος (Durante Citation1976, 159). On the proverb μετὰ Λέσβιον ᾠδόν referring to the superiority of Lesbian kitharodes, see Maslov Citation2009, 11-14. Virtually unknown is the poetess Kleitagora, mentioned in an Attic skolion; interestingly, her area of origin is variously reported as Thessaly, Laconia, or Lesbos (Reitzenstein Citation1893, 29).51. Risch Citation1954, 37, Nöthiger Citation1971, 43: “Als Quelle kommt dann nur frühe lesbische Dichtung in Frage”; cf. Verdier Citation1972, 17, Cassio Citation1999, 203-4.52. Hinge Citation2006 assumes the existence of a common choral Dichtersprache, which arose from the same roots as Homer's, but argues that the epichoric character of our text of Alcman is due to our tradition being based on a transcription of Hellenistic performance, or just on a Spartan text (as proposed by Pavese Citation1967, 178). Cassio extends a similar reasoning to the Doric accentuation in the text of Ibycus and Stesichorus. “Io ho l'impressione che il trattamento accentuale dorico riservato dagli studiosi ellenistici ad Alcmane, Ibico e Stesicoro sia dovuto, più che a considerazioni basate su una pretesa maggiore ‘doricità’ della loro lingua, soprattutto al fatto che i luoghi in cui i poeti operarono avevano in età ellenistica forti tradizioni doriche…” (Cassio Citation1999, 203).53. Nöthiger Citation1971, 119-121: Alcman uses the definite article in about 50% of expected cases, whereas for other choral poets the figures are much lower: 2% for Stesichorus, 19% for Ibycus, 36% for Simonides, 15% for Bacchylides, 10% for Pindar (all percentages are based on rough counts by Nöthiger). This suggests that Alcman's idiom is closer to the actual spoken language whereas the usage of other poets is archaizing (and for this reason closer to the Homeric usage). Similarly, of all poets composing in choral genres Alcman is the most sparing in his use of non-augmented forms of past tenses (another archaic/Homeric feature): 15% (contrast: Stesichorus 23%, Ibycus 33%, Simonides 44%, Bacchylides 56%, Pindar 43%; figures in Nöthiger Citation1971, 122-123).54. Apart from the -οισα ending in feminine participles, other notable Aeolicisms in the text of Alcman include πεδά, adjectives ending in -ιος, the dative plural -εσσι (also in Homer), geminates in the name Φαέννα and κλεννός; the latter has been claimed by Risch (Citation1954, 22) to be an artificial quasi-Lesbian formation. These features struck ancient philologists; Apollonius Dyscolus describes Alcman as “incessantly Aeolicizing” (συνεχῶς αἰολίζων; de Pron.136bc; cited, along with other ancient testimonies, in Risch Citation1954, 28). Note too that the dative pl. ending -οισι(ν), -αισι(ν) is so prominent that it is difficult to explain by Homeric influence alone (Nöthiger Citation1971, 23); in this case, however, we may be dealing with an undocumented Laconian dialect feature (-οισι is attested in an early Argolic inscription [Buck Citation1955, 283]).