Pindar’s Verbal Art: An Ethnographic Study of Epinician Style

  Wells, James Bradley. 2010. Pindar's Verbal Art: An Enthnographic Study of Epinician Style. Hellenic Studies Series 40. Washington, DC: Center for Hellenic Studies.

2. Epinikion as Event

For Dell Hymes the notion of a speech community, one of the fundamental notions of the Ethnography of Speaking, addresses how members of a community conceive of language and emphasizes the ethnographic description of language use (1974:47–51). The previous chapter presented evidence to show that, from the point of view of the community of artists and audiences who participated in epinician performance, epinician language is spoken, not written. The principle of intersubjective objectivity urges, then, a philological method that is dialogically responsive to how Pindar’s language represents itself. As it turns out, the opening of Pindar’s Pythian 1 provides a specific point of departure for building our local knowledge of epinician art: there are certain sâmata ‘signs’ that singers follow when they compose and perform epinikion and that guide the audience’s reception and evaluation of works of epinician art. Pindar’s description of composition in terms of reception and the audience’s co-creation of the communicative exchange highlight the sharedness of those sâmata, so that the signs point to some set of social conventions recognized by singers and audience alike: “communicative activity involves all participants; meaning is constructed within their relationship” (Bonifazi 2004a:391). The remaining chapters of Pindar’s Verbal Art describe the social conventions that epinician composer and audience use in their relationship, beginning in this chapter with another fundamental notion of the Ethnography of Speaking, the speech event (Hymes 1974:52).

Recalling from the Introduction that a register is a style whose use is dedicated to a particular context of situation in accordance with a community’s language conventions, the purpose of this chapter is to describe how Pindar’s epinician texts record ethnographic evidence for the relationship between epinician style and the event of epinician performance. Bauman calls such evidence “the keying of performance” (1977:15–24; Foley 1995:11–17 and 2002:85–93). Entextualized records of performance include features, or keys, that frame the speech event of performance and, through such framing, forge the link between an artistic idiom and the event to which it is dedicated. This idiom-event link is characteristic of a linguistic register. Pindar’s epinikia evidence each performance key that Bauman illustrates: special code, figurative language, parallelism, special paralinguistic features, special formulae, appeal to tradition, and disclaimer of performance. Echoing Bauman, I hasten to stress the suggestive nature of his performance keys, which ultimately are to be discovered ethnographically.

Special Code

Jakobson (1960:353) and Hymes (1974:10, 13, and 59), who adapts Jakobson’s model of language to the Ethnography of Speaking, apply the term “code” to a communicative vehicle shared by participants in the typical speech events that make up a community’s speech economy. In this sense, a linguistic code is a language—say, ancient Greek. In terms of verbal art, a special linguistic code is a communicative vehicle reserved for use in performance events. The difference between a special code and a register is one of relationality: the special code is one performance key among others that calls attention to the existence of a register. The features of epinikion’s special code include a blend of varieties of ancient Greek, archaizing morphological and lexical features, and prosodic patterning. [1]

Figurative Language

Praise and blame poetics is another example of how figurative language has more than denotational meaning. [11] One word that Pindar uses to refer to epinikion is ainos ‘praise’; [12] cognate verbal forms are ainein ‘to praise’ [13] and epainein ‘to praise’. [14] Opposing this designation for the poetics of praise is the designation for blame poetics, psogos ‘fault-finding’. [15] Also indicative is psogeros ‘bitter-tongued’; Nagy describes the passage in which this epithet occurs, Pythian 2.55–56, as “a programmatic description of blame poetry…as the opposite of praise poetry, in the specific context of rejecting blame within a poem of praise” (Nagy 1990:24): “bitter-tongued [psogeros] Archilochus with heavy–worded hatred fattening himself…” (ψογερὸν Ἀρχίλοχον βαρυλόγοις ἔχθεσιν / πιαινόμενον). While the opposition between praise and blame is explicit here, Nagy shows that this opposition is implicit even if it is not expressed: the potential of praise is blame, and vice versa. [16] In other words, praise and blame poetics entails an implicated meaning that contrasts with the literal, strictly denotational meaning of ainos and semantically related language. Since composer and audience share the same special code in order to carry off the communicative exchange, such implicated meaning indicates how figurative language partitions the performance event as a special context in which that implicated meaning makes sense, keying performance.


Bauman defines parallelism as “the repetition, with systematic variation, of phonic, grammatical, semantic, or prosodic structures, the combination of invariant and variant elements in the construction of an utterance” (Bauman 1977:19). [17] Epinikion exhibits elaborate and mutually imbricated patterns of regularity: prosodic structure and the patterned repetition of words, sounds, or themes. [18] Such patterned regularities enable an audience to evaluate a performer’s communicative competence on the basis of how well or poorly she creates and sustains a pattern of expectancy, while also artfully manipulating it within conventionally established parameters (Bauman 1977:19). In all epinician songs, prosody involves at least three patterns of repetition: (1) the basic metrical structure of an individual song is either dactylo-epitrite or Aeolic; (2) there is patterned repetition among individual lines; (3) there is patterned repetition at the level of strophic structure. For further examples of parallelism, Carola Greengard has described in detail aspects of formal arrangement in epinikion that involve the “adaptation and derivation of the archaic structure of ring composition” (1980:15): [19] framing, chiasmus, and recurrent diction. Devices that shape frames are recurrent diction, tautometric responsion, “cases in which similar rather than identical words are involved,” compound words, and “repeated patterns of sound and/or rhythm” (Greengard 1980:19–22). Although Greengard does not identify them as such, these structural features are forms of parallelism: they key—or re-key for the modern audience—epinician performance.

Special Paralinguistic Features

Special Formulae

Special formulae are, “in effect, markers of specific genres, and insofar as these genres are conventionally performed in a community, the formulae may serve as keys to performance” (Bauman 1977:21). [21] Bundy gives the following formulaic elements of epinikion:

None of these are formulae in the sense of “special phrases,” as described in Foley’s discussion of performance keys in How to Read an Oral Poem: “the guslar’s (and Bishop Njegoš’s) ‘well-wrought tower’ or ‘shaggy brown horse’ or Homer’s ‘wine-dark sea’ or the Old English poet’s ‘foamy-necked ship’” (Foley 2002:90). If we consider verbal art as performance cross-culturally, then we should not expect that special formulae will always be phraseological words, like the Homeric epithet, or that they are “of equal importance in all oral poetries,” as Foley stipulates (Foley 2002:90). I will show in Chapters Three and Four that the ways of epinician speaking—stylistic patterns that constitute the epinician way of speaking—are discursive formulae that are conventional to epinikion. [
23] The Appendix to Pindar’s Verbal Art presents evidence suggesting that the use of these ways of epinician speaking are so highly patterned that deviations from such patterns can be artistically meaningful. When special formulae occur, even conceived in terms of the list of Bundy’s formulaic conventions of epinikion, they serve, along with other performance keys, to segment performance as a specially framed speech event. [24]

Appeal to Tradition

An appeal to tradition keys performance by referring to traditional conventions as a standard for the audience’s evaluation of the performer’s verbal artistic competence (Bauman 1977:21). “Either explicitly or implicitly,” as Foley describes this key to performance, “oral poets are constantly establishing and reestablishing the authority of their words…by reaffirming their ties to an ongoing way of speaking, to an expressive mode larger than any one individual” (Foley 2002:91). Pindar’s use of humnos ‘song of praise’ illustrates his explicit appeals to tradition. As Calame has argued, the example of Pindar’s Nemean 5, whose mythological narrative depicts the Muses, led by Apollo, performing during the wedding of Peleus and Thetis, shows that humnos “defines itself as a song in which gods and heroes are celebrated” (2001:75). Calame goes on to make the important observation that the diversity of ways in which Archaic Greek forms of song use the term humnos indicates that it is not considered a lyric genre until the Alexandrian period (2001:78). As a strategy in a song of praise, rather than a genre of verbal art, Pindar at times appeals to the tradition of humnos. Olympian 2, for example, begins with an invocation addressed to humnoi:

Ἀναξιφόρμιγγες ὕμνοι
τίνα θεόν, τίν᾽ ἥρωα, τίνα δ᾽ ἄνδρα κελαδήσομεν;

Olympian 2.1–2

Lyre-ruling humnoi,
of what god, of what hero, of what man shall we sing in praise?

This passage is an appeal to tradition, serving to locate Olympian 2 in relation to existing conventions for praise associated with the strategy of humnos. Rather than resorting to literal exegesis to interpret Pindar’s references to song and dance, we can more empirically interpret them as explicit appeals to tradition that key performance, as in the passage’s opening, where Pindar aligns his composition with the traditional humnos. Some examples of other explicit appeals to tradition include references to Homer, a past authority who demonstrates the power of persuasive speech to create an enduring record of action (Pythian 4.277, Nemean 7.21, and Isthmian 4.37); [
25] one reference to Arkhilokhos (Olympian 9.1) positively compares that poet’s art with epinician praise; [26] another, discussed above, specifically contrasts Pindar’s praise poetry with Arkhilokhos’ blame poetry (Pythian 2.55).

Disclaimer of Performance

A first-person voice often interrupts the course of Pindar’s songs, a phenomenon that is part of the motivation for the hypothesis that Pindar uses an “oral subterfuge” to represent his putatively written compositions as spontaneous speech. [28] These Abbruchsformeln ‘break-off formulas’ can be thought of as saying one thing and doing another, in the sense that Foley (2002:93) explains disclaimers of performance: “saying you can’t means asserting that you can and you will.” [29] Here is an example of such a break-off:

ἐμοὶ δ᾽ ἄπορα γαστρίμαργον μακάρων τιν᾽εἰπεῖν· ἀφίσταμαι.

Olympian 1.52

For me it is impossible to say that any of the blessed ones is gluttonous; I stay away from that.

In this passage Pindar displays his observance of the conditions for appropriate speech expressed by gnomic statements that occur earlier in the song:

Truly wondrous are many things, and, as it seems, mortals’
speech in excess of a true account,
stories crafted with ornate lies, are utterly deceptive.

One example of “stories crafted with ornate lies” is the story that Pelops’ ivory shoulder replaced the flesh and bone shoulder eaten by Demeter in another version of the Pelops story. By refusing to tell such a story (Olympian 1.52), Pindar observes the conditions for propriety communicated at Olympian 1.28–29. This is one sense in which Pindar says one thing and does another: by saying that he will not tell a deceptive story, he performs a display of his competence in the appropriate use of speech. In this respect, we can see the break-off at Olympian 1.52 as an example of Bundy’s siga-motif, a Pindaric formula that corresponds to Bauman’s characterization of the disclaimer of performance, which “serves as a moral gesture, to counterbalance the power of performance to focus heightened attention on the performer, and a key to performance itself” (1977:23).

Performance Keys Conclusion

A crucial implication of the dedicated use of epinician language in epinician performance, the consequence of the fact that epinician language is a register, is that epinikion entails culture-specific ways of meaning. The communicative means used in epinician performance key that performance event as such, and this keying sets up a special frame for the interpretation of messages, so that it “contrasts with at least one other frame, the literal” (Bauman 1977:9). [32] Epinician performance keys enable an audience, ancient or modern, to evaluate and interpret epinikion in terms of the art form’s idiomatic, connotative meaning. Given that epinikion is a speech event, as the epinician performance keys indicate, our recognition of the existence of a conventional event-idiom link supports a further conclusion: Pindar is an oral poet. The special code, figurative language, forms of parallelism, and special paralinguistic features unambiguously cue audiences, both ancient and modern, to treat epinikion as a speech event, not a writing/reading event. Special formulae, appeals to tradition, and disclaimers of performance regulate the reception of each epinician song according to song and performance traditions and according to participation rules for performance—especially the participation of the audience as critical and highly competent evaluators of epinikion. In the next two chapters I will describe the ways of epinician speaking that constitute the epinician way of speaking, paving the way toward an understanding of the strategies of epinician oral composition.


[ back ] 1. For special code as a performance key, see Bauman 1977:17, Foley 1995:11–12, 83–85, and 93, 1999:23–25, and 2002:85–86. For descriptions of Pindar’s language, see, among many others, Gildersleeve 1890:lxiii–cxv, Dornseiff 1921, Forssman 1966, Horváth 1976, Palmer 1980:123–127, and Hummel 1993.

[ back ] 2. On the blend of ancient Greek varieties in Pindar’s language as artificial and reserved for literary usage, see Gildersleeve 1890:lxxvi–lxxvii, Palmer 1980:123, Hummel 1993:415–416, and Horrocks 1997:18. On the relationship between choral song and Panhellenic institutions, see Nagy 1979:7. Cf. Bowra 1964:197 and Horrocks 1997:19.

[ back ] 3. But see Davies 1988 for another view.

[ back ] 4. On the α (a)/η (ê) alternation, see Gildersleeve 1890:lxxviii–lxxix, with examples, Gentili 1988:58–59, and Horrocks 1997:21.

[ back ] 5. Note that Horrocks 1997:20 voices the concern that the evidence for Aeolic features may result from the influence of interpolations inserted by Alexandrian scholars in the process of textual transmission.

[ back ] 6. On Aeolic morphology in Pindar’s songs, see Gildersleeve 1890:lxxxv and Palmer 1980:124 and 126.

[ back ] 7. Note further that Pindar does not use the attested Doric form Μῶσα (Môsa).

[ back ] 8. See also Gildersleeve 1890:lxxxv and Palmer 1980:124–125.

[ back ] 9. The point of departure on Pindar’s prosody is Maehler 1989:178–188. See also Gildersleeve 1890:lxiii–lxxvi, Bowra 1964:317–321, West 1982:60–76, Race 1986:11–13, and Nagy 1990:416–418 and 439–464. With few exceptions the choral poetry of Simonides, Bakkhulides, and Pindar possesses a strophic structure that West (1982:60) associates with a “Dorian tradition of composition.”

[ back ] 10. On figurative language as a performance key, see Bauman 1977:17–18 and Foley 1995:12, 64, 85–86, 93, and 2002:87–88. See Bowra 1964:219–236 and 239–277 on Pindar’s word choices and imagery; Steiner 1986 and Loscalzo 2003:125–160 on Pindar’s metaphors.

[ back ] 11. For the benchmark description of praise and blame poetics, see Nagy 1979:147–150 and 222–242. See also Kirkwood 1984, Hubbard 1985:72 and 80, Gentili 1988:107–114, and Nagy 1990:147–150, 196–200, 203–206, 424–433, and passim. See Bonifazi 2004c:292–294 on the implicit connotations of the deictic demonstrative adjective (e)keinos in the context of blame poetics.

[ back ] 12. Olympian 2.95, 6.12, 11.7; Nemean 1.6.

[ back ] 13. Olympian 4.14, 7.16, 9.14, 9.48, 10.100; Pythian 1.43, 3.13, 4.140, 9.95; Nemean 1.72, 3.29, 4.93, 7.63, 8.39; Isthmian 5.59, 7.32, 8.69.

[ back ] 14. Olympian 13.2; Pythian 2.67, 4.168, 4.189, 5.107, 10.69; Nemean 5.19, 11.17.

[ back ] 15. Pindar Nemean 7.61, which Nagy 1990:223 discusses.

[ back ] 16. Kirkwood 1984 explicitly rejects Nagy’s view that a poetics of praise implies a (rejected) poetics of blame in the context of epinikion. Steiner 2002 explores the link between poetics of blame and characterizations of consumption as gluttonous in Olympian 1, confirming Nagy’s position.

[ back ] 17. On parallelism as a performance key, see also Foley 1995:12–13, 86–87 and 2002:89–90. On parallelism generally, see Jakobson 1960:358, 1966, and Waugh 1980:64–65.

[ back ] 18. On lexical, phonic, and thematic repetition, see Stockert 1969, Schürch 1971, and Greengard 1980.

[ back ] 19. On ring composition in Pindar’s songs see Illig 1932, Hamilton 1974:8, 56–71, 112–113, and Gerber 1982:xi (on Olympian 1, specifically). I discuss ring composition at greater length in Chapter 5 below.

[ back ] 20. For an admirable attempt to overcome this gap in our record of epinician performance, see Mullen 1982.

[ back ] 21. On special formulae as a performance key, see also Foley 1995:13–14, 81, and 2002:90–91.

[ back ] 22. Cf. Kyriakou 1996:18 on “hush passages” in Pindar’s epinikia and Burnett 2005:68-69 and 86 on “[t]he melodramatic choral trick of self-imposed silence.”

[ back ] 23. The existence of discursive formulae in epinikion will corroborate the research of Martin (1984 and 1989), who has demonstrated that the formulaic quality of Archaic Greek poetry can be identified at the discursive and speech act level of communication in addition to that of prosodically defined phraseological word formulae.

[ back ] 24. On framing of performance events see Bauman 1977:9; Foley 1995:14–15 and 87 and 2002:91–92.

[ back ] 25. See Nisetich 1989:22–23 and passim on the ways in which Pindar’s treatment of Homer reflect more Pindar’s conception of his own poetry than of Homer’s poetry.

[ back ] 26. The scholia report that the opening lines of Olympian 9 refer to the tênella kallinike cry. Originally in a hymn to Herakles composed by Arkhilokhos, tênella kallinike was used as an impromptu victory song. For a recent discussion of the passage, with sources, see Thomas 2007:144–145.

[ back ] 27. Here I am drawing from Nagy’s distinction between song and poetry (1990:17–51).

[ back ] 28. See Bauman 1977:21–22 and Foley 1995:15, 87, and 2002:92–93 on the disclaimer of performance as a performance key. On Pindaric Abbruchsformeln, see Schadewaldt 1928:267–268, 286, 312, Hamilton 1974:16–17, Race 1990:41–57, Kyriakou 1996, and Mackie 2003:9–37.

[ back ] 29. Mackie (2003:6, 9–37) explains Pindaric break-offs as “a safeguard against κόρος [koros] ‘excess’.” See also Race 1980 on rhetorical functions of break-offs.

[ back ] 30. Here I follow Bowra’s edition of Pindar (1935). Snell and Maehler’s edition (1997) has the noun θαύματα (thaúmata) ‘wonders.’ Their critical apparatus does not acknowledge the readings of the major MSS., which have either the unusual adjective θαυματά (thaumatá) ‘wondrous’ or the singular noun θαῦμα (thauma) with the neuter plural article τά (ta). The Codex Vaticanus evidences the prior reading, and Bowra identifies this manuscript as the most reliable at this locus (1935:v and 2). Gildersleeve, who maintains that he made no emendations to his edition, also accepts the reading θαυματά (thaumatá) (1890:132). So does Fisker 1990:35.

[ back ] 31. Thus Mackie 2003:13 does not account for the fact that Pindar actually does tell the story. With respect to Pindar’s attribution of the de-selected story about Pelops to a jealous neighbor, Bulman 1992:13 writes that “Pindar demonstratively rejects φθόνος [phthonos], damning it in others and quelling it in himself.”

[ back ] 32. Closely resonant with Bauman is Nagy’s view that “performance frames composition, and we cannot fully grasp the role of composition without knowing about this frame” (1994–1995:14; cf. 1994–1995:19). On frame analysis, I draw from Goffman 1974. Cf. Steiner’s comments: “The breakdown of literal levels of meaning sets us on the road to rediscovery and rediscription, allowing poet and audience both to step back from a world of ordinary reference where words function as signs, to one of symbols, where words become significant in themselves. This symbolic language is notoriously dense, making words into a more substantial matter which does not merely represent, but expresses. Such opaque discourse replaces denotation with connotation, the hallmark of metaphoric speech” (1986:149–150).