Master of the Game: Competition and Performance in Greek Poetry

  Collins, Derek. 2004. Master of the Game: Competition and Performance in Greek Poetry. Hellenic Studies Series 7. Washington, DC: Center for Hellenic Studies.

18. Modes of Innovation

The evidence for rhapsodic performance as we have it suggests that there were at least three basic types of improvisational activity in which rhapsodes engaged. The first involves the “stitching” or “weaving” of song, the second involves the insertion of newly composed “Homeric” verses into a preexisting text, and the third involves capping with hexameter verses. We are often at pains to determine which of these types were employed at given performance venues, but we have enough evidence to provide some suggestive indications. Let us begin with some familiar passages and scholia with regard to the etymology and meaning of the word ῥαψῳδός as ‘he who stitches the song’. The locus classicus for this word, as well as for the description of the mechanics of rhapsodic performance, is Pindar’s Nemean 2.1–3 and the scholia on those lines. At the beginning of Nemean 2, Pindar claims that he will begin where the Homeridae begin (Pindar, Nemean 2.1–3):

Ὅθεν περ καὶ ῾Ομηρίδαι
ῥαπτῶν ἐπέων τὰ πόλλ᾽ ἀοιδοί
ἄρχονται, Διὸς ἐκ προοιμίου

From the very point where the Homeridae,
singers of stitched-together utterances, most often
begin, from a proem of Zeus

There are several other testimonia in the same scholia that say the poetry of Homer had been, at some unspecified time, scattered and divided into parts, so that to sing it rhapsodically meant to do something on the order of sewing the parts together to produce a whole (scholia to Pindar, Nemean 2.1c 30.5–8 Drachmann):

οἱ δέ φασι τῆς ῾Ομήρου ποιήσεως μὴ ὑφ᾽ ἓν συνηγμένης, σποράδην δὲ ἄλλως καὶ κατὰ υέρη διῃρημένης, ὁπότε ῥαψῳδοῖεν αὐτην, εἱρυῶ τινι καὶ ῥαφῇ παραπλήσιον ποιεῖν, εἰς ἓν αὐτὴν ἄγοντας.

Some say that, since the poetry of Homer had not been brought together in one collection, and since it was otherwise scattered and separated into parts, whenever they would sing it rhapsodically they would do something similar to sequencing or sewing, producing it into one thing.

However one chooses exactly to define the word here for part, μέρος, clearly this definition of rhapsôidos or rhapsôidein suggests that each part was a longer segment of narrative, but how long we cannot say. It is possible that each meros ought to be scaled on the order of what we are told in Plato’s Ion, where popular scenes from the Iliad or Odyssey are singled out for mention by Socrates—such as Nestor’s advice to Antilokhos from Iliad 23, Odysseus at the moment when he leaps upon his threshold to kill the suitors from Odyssey 22, or the scene when Achilles lunges at Hector in Iliad 22 (all featured at Ion 535b3–7), each of which might constitute a performable “part.” [

The scholia on Nemean 2.1–3 also include other descriptions of how rhapsodes perform, notably from Philochorus (scholia to Pindar, Nemean 2.1c 31.7–9 Drachmann = FGH 328 F 212):

Φιλόχορος δὲ ἀπὸ τοῦ συντιθέναι καὶ ῥάπτειν τὴν ᾠδὴν οὕτω φησὶν αὐτοὺς προσκεκλῆσθαι.

Philochorus says that they [=rhapsodes] were thus called on account of the putting together and stitching of the song.

In this passage Philochorus, who may simply have rationalized his explanation based upon Nemean 2.1–3, connects the idea of assembling (syntithêmi) a song with the verb rhaptein. More tantalizing is that in conjunction with this Philochorus then cites a fragment attributed, perhaps wrongly, to Hesiod (F 357 MW):

ἐν Δήλῳ τότε πρῶτον ἐγὼ καὶ Ὅμηρος ἀοιδοὶ
μέλπομεν, ἐν νεαροῖς ὕμνοις ῥάψαντες ἀοιδήν,
Φοῖβον ᾽Απόλλωνα χρυσάορον, ὃν τέκε Λητώ.

At that time, Homer and I, as singers, sang for the first time on Delos,
stitching together a song in new hymns
about Phoibos Apollo of the golden sword, whom Leto bore.

The second type of improvisational activity by rhapsodes is attested in one final example from the scholia to Nemean 2 (2.1c 9–18= FGH 568 F 5):

῾Ομηρίδας ἔλεγον τὸ μὲν ἀρχαῖον τοὺς ἀπὸ τοῦ ῾Ομήρου γένους, οἳ καὶ τὴν ποίησιν αὐτοῦ ἐκ διαδοχῆς ᾖδον· μετὰ δὲ ταῦτα καὶ οἱ ῥαψῳδοὶ οὐκέτι τὸ γένος εἰς Ὅμηρον ἀνάγοντες. ἐπιφανεῖς δὲ ἐγένοντο οἱ περὶ Κύναιθον, οὕς φασι πολλὰ τῶν ἐπῶν ποιήσαντας ἐμβαλεῖν εἰς τὴν ῾Ουήρου ποίησιν. ἦν δὲ ὁ Κύναιθος τὸ γένος Χῖος, ὃς καὶ τῶν ἐπιγραφομένων ῾Ομήρου ποιημάτων τὸν εἰς ᾽Απόλλωνα γεγραφὼς ὕμνον ἀνατέθεικεν αὐτῷ. οὖτος οὖν ὁ Κύναιθος πρῶτος ἐν Συρακούσαις ἐραψῴδησε τὰ ῾Ομήρου ἔπη κατὰ τὴν ξθ´ ᾽Ολυμπιάδα, ὡς ῾Ιππόστρατός φησιν.

In this rather long example, we learn both about the lineage of the Homeridae, who once claimed to have descended from Homer, and then about Cynaethus, who is said to have been the first person to sing the epics of Homer rhapsodically at Syracuse. [11] We also learn that Cynaethus composed his own utterances (ἔπη), which here most likely means individual verses, and then his rhapsodes put them into the poetry of Homer. We do not know whether Cynaethus composed his utterances extempore during a performance and passed them off as Homer’s, or whether he composed them prior to his performance. Either way, two points are important here: 1) the Homeric poems are envisioned by this commentator—that is, Hippostratus, or as has more recently been argued, the Aristarchean school of Homeric criticism from which Hippostratus might have derived his ideas [12] —as being relatively fixed; and 2) Cynaethus composed lines that he then inserted into Homer. This story represents a type of rhapsodic improvisation in which a rhapsode creates his own lines for performance and display against the background of a more stable body of Homeric narrative. What remains implicit in the description of Cynaethus is why (beyond some generic desire for notoriety) he composed epic verses and a hymn and passed them off as Homer’s. I will return to this point later, but the evidence for rhapsodic performance as it accumulates will suggest that Cynaethus created new material to compete with his rhapsodic opponents rather than with Homer.

The game continues with “Hesiod’s” first challenge verse, in which he says δεῖπνον ἔπειθ᾽ εἵλοντο βοῶν κρέα καὐχένας ἵππων “then they took as their meal the flesh of cattle, and the necks of horses—.” At this point, which is to say right after the bucolic diaeresis, the noun aukhên looks as if it is going to remain the object of the verb εἵλοντο, until “Homer” successfully enjambs the next line with a verb and participle in agreement with the noun, ἔκλυον ἱδρώοντας “—they unyoked (those necks) dripping with sweat,” and then fills out the rest of the line with a further comment, ἐπεὶ πολέμοιο κορέσθην “when they had tired of war.” This does not just take a meaningless line of verse and turn it into a meaningful one, as Heldmann had so flatly observed, but rather successfully converts the outlandish idea of eating horses—a barbaric practice, perhaps reminiscent of what Herodotus tells us about the Scythians (4.61)—into a more mundane one about relieving them from their burdens during wartime. Another such inversion of cultural content concerns the next challenge, when “Hesiod” offers, καὶ Φρύγες, οἳ πάντων ἀνδρῶν ἐπὶ νηυσὶν ἄριστοι “And Phrygians, who are best of all men at ships—,” which is absurd because the Phrygians were never known as a maritime people. Yet when “Homer” counters with, ἀνδράσι ληιστῆρσιν ἐπ᾽ ἀκτῆς δόρπον ἑλέσθαι “—among thieves to take their dinner on the shore,” he accomplishes two things. First, he locates the Phrygians among thieves. Although we know little about the Phrygians through the second century BCE, “Homer’s” response may hint at a Greek antagonism toward them, perhaps due to their participation in the army of Mardonius against Greece in the Persian Wars. [27] Even if this interpretation is not correct, however, the second point “Homer” clearly makes is to situate the Phrygians on shore to eat, negating their pretended maritime prowess.

From a formal point of view, these examples suggest that the game depends upon enjambement and upon the possibilities of meaning created by a verse that does not express a complete thought. Wherever the sense break occurs in the lead verse spoken by “Hesiod,” this break helps to structure what kind of word can be placed in the runover position at the beginning of “Homer’s” following line. Aristotle and Demetrius of Phalerum both discuss the ambiguities created in incomplete (verse) periods in their rhetorical treatises, and thus it is possible that Alcidamas is dramatizing a rhetorical problem. [28] But this is certainly not his only, or even his main, objective, not least because there are always numerous possibilities involved in restoring sense to a nonsensical verse. Alcidamas is clearly interested in the extemporaneous choices that a rhapsode must make in performance. Moreover, we are simply not able to recover from the texts themselves any metalinguistic signals, such as changes in intonation or emphasis, let alone any kind of gestural cues, that could have been used by one rhapsode to signal the next rhapsode as to exactly which feature of the lead verse he would need to focus on for his enjambement. We may take such clues for granted, however, in a medium like this where dramatic enactment (μίμησις) also constitutes part of the rhapsodic performance of Homeric poetry. [29] We may recall that the rhapsode Ion tells Socrates how he is able to move his audience to tears with a riveting performance, or inadvertently to laughter with a poor one (Plato, Ion 535b–e). [30]

Sometimes the fictional “Homer” in the Certamen must wait until he hears the words that occupy the whole adonic at verse-end before he can decide how to enjamb them. So for example at lines 119–120, “Hesiod sings that ὣς οἳ μὲν δαίνυντο πανήμεροι, οὐδὲν ἔχοντες “so they feasted all day long, having nothing—” at which point “Homer” should be confounded, yet he twists the idea to his own advantage by enjambing an adverb οἴκοθεν “having nothing—from home” ἀλλὰ παρεῖχεν ἄναξ ἀνδρῶν ᾽Αγαμέμνων “but Agamemnon lord of men supplied it.” On this occasion the enjambement is an adverb, at other times it may be a noun or a participle (e.g. Certamen 128, 138) coordinated with the end of the previous verse by its case. Participial enjambement is especially significant because we have already seen in Part I how it was used in dramatic stichomythia (both in tragedy and Old Comedy) and lament to continue a verse sequence. This structural correspondence offers us strong evidence that we are dealing with a technique of oral verse-making that is in fact not unique to any one genre. On the contrary, participial enjambement, which might have originated with the earliest composers-in-performance of Homeric poetry, at a later time was incorporated into written literature as a reflection of that earlier oral heritage. Its preservation in such apparently disparate verbal-gaming genres as stichomythia and rhapsodic performance not only attests a close correspondence between these genres, it also attests the continued effort of Greek authors to represent extemporaneity in literature.

In this example we see that the noun φώς ‘man’ is governed by a verb in the runover position, and lest we think this is a formula, consider this next verse in which the same noun in the same position is governed by a different verb (Iliad 20.345–46):

ἔγχος μὲν τόδε κεῖται ἐπὶ χθονός, οὐδέ τι φῶτα

This spear of mine lies on the ground, and I cannot at all the man

In the Certamen “Homer” also enjambs infinitives to limit and transform a leading verse from “Hesiod.” So for example at lines 131–32, “Hesiod’s” lead verse says αὐτὰρ ἐπεὶ σπεῖσάν τε καὶ ἔκπιον οἶδμα θαλάσσης “but when they poured libations and drank, the swell of the sea—,” which implies that the men drank the sea, until “Homer” enjambs it with the infinitival construction ποντοπορεῖν ‘to sail’ + μέλλω ‘to intend’, ποντοπο ρεῖν ἤμελλον ἐυσσέλμων ἐπὶ νηῶν “—they were minded to sail (the swell of the sea) on web-benched ships.” We may compare this to another example from the Iliad, which although not exactly the same, similarly enjambs an infinitive that governs a preceding noun (11.704–5):

ἐξέλετ᾽ ἄσπετα πολλά· τὰ δ᾽ ἄλλ᾽ ἐς δῆμον ἔδωκε
δαιτρεύειν. μή τίς οἱ ἀτεμβόμενος κίοι ἴσης.

[Neleus] rook a huge amount; but the rest he gave to the people
to distribute, so that no one would go away deprived of a just share.


[ back ] 1. Schmitt 1967:300–30 and Chantraine 1968–:s.v. ῥαψῳδός. Cf. Tarditi 1968:144, who argues that the basic activity of the ῥαψῳδός involves interweaving (intessere) individual material into that derived from epic tradition, while performers like the Homeridae stitch (cucire) together Homeric material. Such a distinction is too rigid in my view because it presupposes a clear sense of what was “Homeric” versus “individual” poetry, but this is not always so clear.

[ back ] 2. As a response to Fränkel 1925, Patzer 1952:322–23 argued that the stitch (Stich, i.e. a line of hexameter verse) was the basic unit of composition implied by rhaptein, but he nevertheless conflated (like the scholiasts) the metaphors of weaving and stitching found in the scholia to Pindar.

[ back ] 3. I do not agree with the view (e.g. in Else 1957:27, Taplin 1992:29–31, et al., on which see chapter 19 below), reflecting a wider assumption in scholarship, that the entire and Odyssey, from what we know as their beginning to end, were performed at the Panathenaia. I leave open the possibility that different “parts” of the type just described in Plato’s Ion could have been performed as the respective contributions of competing rhapsodes.

[ back ] 4. Cf. Odyssey 8.429, where the expression ἀοιδῆς ὕμνος implies that humnos is a subdivision of song. But for the view that humnos in this passage “conveys the idea of the totality of a given performance of a song (italics in original),” see Nagy 2002:70–4, quote at p. 70.

[ back ] 5. As Else 1957:30–1 and most recently Martin 2000:411–15 have argued, if Hesiod F357 MW can be taken to refer to the Homeric Hymn to Apollo, we may plausibly account for the Delian and Pythian division of that poem as the competitive contributions performed respectively by “Homer” and “Hesiod.” Accordingly, line 241 of the Homeric Hymn to Apollo (with the exception of προΐει for προχέει) is from Hesiod, Catalogue fr. 70.18 MW; cf. the AB scholia on Iliad 2.522. As for the Homeric poems themselves, especially the Iliad, Eustathius already believed that many stylistic features could be through Homer’s improvisation; see Van der Valk 1976:xxvi–xxvii with n1, and xxxix with n3.

[ back ] 6. The sewing metaphor is embraced by Nagy 1996:66.

[ back ] 7. Ritoók 1962:226n7.

[ back ] 8. As Nagy 1990b:22n22 suggests, following Cingano 1985, the context of Cleisthenes’ war with Argos makes it likely that the content of these epic performances involved material from the Theban cycle.

[ back ] 9. The expression ἐκ διαδοχῆς in this context could be taken to mean ‘in [genealogical] succession’—in other words, the Homeridae would sing and teach the art of singing Homeric poetry from one generation to the next. But there are several reasons not to take only that meaning here. First, διαδοχή is from διαδέχεσθαι ‘speak next/succeed’, and δέχεσθαι ‘receive’ is the technical term for poetic capping, as at Aristophanes, Wasps 1222, 1225 (cf. Dionysius Chalcus fr. 1.1 West). Even if we extract the generational sense out of the expression, I do not believe that Hippostratus would have been unaware of the double meaning. However, the most compelling evidence for taking the expression to mean ‘by relay’ is not that it is already used to mean ‘by relay’ in Demosthenes, Philippics 1.21 (ἐκ διαδοχῆς ἀλλήλοις). Rather, it is that the A scholiast on Iliad 1.603, as we saw earlier, who is commenting on the singing of the Muses (ἄειδον ἀμειβόμεναι ὀπὶ καλῇ, 604), uses the same expression when he writes αἵτινες καὶ αὐταὶ [the Muses] ᾽Απόλλωνος κιθαρίζοντος ἐκ διαδοχῆς καὶ παρὰ μέρος ἦδον “the Muses themselves while Apollo played the cithara were singing by relay and by turns.” Thus we do violence either to the meaning of this scholiast or to the one on Nemean 2 to translate ἐκ διαδοχῆς otherwise.

[ back ] 10. Cf. Martin 2000:419n58, who suggests that the expression ἀνατέθεικεν αὐτῷ may mean that Cynaethus “dedicated it [the hymn] to him (autôi = Apollo) (italics in original).

[ back ] 11. For more on Cynaethus, see West 1975.

[ back ] 12. Nagy 2000a:99n6.

[ back ] 13. Dunkel 1979:252–53.

[ back ] 14. Although not involving rhapsodes, Dunkel 1979:252–53, following Dornseiff 1944:135, points to the parallel between these modes of poetic competition and those represented in Aristophanes’ Frogs between “Aeschylus” and “Euripides:” general tests of σοφία (1420–65), recitation of passages (1126–87), and capping a couplet given by the opponent (ληκύθιον ἀπώλεσεν, 1208–45). As an additional mode in the Frogs, the judge has them recite a line simultaneously to weigh the “heaviness” of its imagery (1378–1403).

[ back ] 15. Richardson 1981:1–2.

[ back ] 16. From Hesiod’s Melampodia = Frag. 278 MW. Cf. the tradition of the rhapsodic performance (rhapsôidein) of Empedocles’ Purifications (31 A 1 DK).

[ back ] 17. All text citations from the Certamen are taken from Allen 1912. For general background to the Certamen, especially the issue of dating, see Richardson 1981, which is a response to West 1967. For an approach to the Certamen similar to my own, cf. Graziosi 2001.

[ back ] 18. See the discussion of Ritoók 1991:160 and the more detailed analysis of Alcidamas’ views in O’Sullivan 1992.

[ back ] 19. For example, cf. the ἀμφίβολοι γνῶμαι ‘ambiguous sentences’ at Certamen 170–71, where “Hesiod” asks: τῆς σοφίης δὲ τί τέκμαρ ἐπ᾽ ἀνθρώποισι πέφυκεν; “what is the mark of wisdom for men?” to which “Homer” replies: γιγνώσκειν τὰ παρόντ᾽ ὀρθῶς, καιρῶ δ᾽ ἅυ᾽ ἕπεσθαι “to perceive present affairs correctly, and to keep pace with the right moment.” The translation cannot do full justice to this exchange, which among other things reflects the skills demanded in the very improvisational game in which “Hesiod” and “Homer” are presently engaged. On the posing of riddles in the Certamen, see generally Ohlert 1886:19–25.

[ back ] 20. Ritoók 1962:228–29. However, Ritoók did follow Davison in believing that rhapsodes merely recited memorized verses at an event like the Panathenaia.

[ back ] 21. See Glei 1992:48–9, on Aristotle, Poetics 1448a11–14.

[ back ] 22. Athens: Hegemon fr. 1.9–10 Brandt (possibly Panathenaia), and perhaps Plato, Laws 834e, on which see Glei 1992:50n37; Eretria (340 BCE, Artemisia): IG XII 9.189.10 and 19. Cf. Delos (fourth century BCE): IG XI 2.120.48 (παρῳδός).

[ back ] 23. As Lelièvre 1954:79 writes, “The Rhapsodes with their resources of memorized poetry and their practice of running passages together and interpolating their own work were unquestionably in possession of much of the parodists’ equipment. Parody is essentially a play upon an original brought about by verbal alteration, distortion, or change of context and implies a certain mastery both of the original and of the technique necessary for altering it: intrinsic probability might therefore suggest the professional rather than the amateur as the originator of the form.”

[ back ] 24. Heldmann 1982:81. The original reads: “Die Aufgabe besteht darin, einen Vers, der möglichst absurd sein muß … durch einen anderen Vers so fortzusetzen, daß beide zusammen eine einigermaßen sinnvolle Einheit ergeben ….” Froleyk’s 1973:55 analysis offers no further insight.

[ back ] 25. Wilamowitz 1916:402. The expression ἐξ ὑποβολῆς, to be discussed below, is from Diogenes Laertius 1.57 = FGH 485 F 6 and refers to rhapsodes at the Panathenaia.

[ back ] 26. Janko 1981:15.

[ back ] 27. Herodotus 9.32.

[ back ] 28. For Demetrius: On Style 58–9. For Aristotle: Rhetoric 1409b8–12; cf. 1407b 12–18 on the problem created when periods do not contain sufficient connecting particles, famously illustrated by the first line of Heraclitus: τοῦ δὲ λόγου τοῦδ᾽ ἐόντος αἰεὶ ἀξύνετοι γίγνονται ἄνθρωποι “Of the logos, existing always, men are uncomprehending” or “Of the logos, existing as it does, men are always uncomprehending” (fr. 1 Marcovich = 22 B 1 DK, with Marcovich’s note ad αἰεί).

[ back ] 29. Herington 1985:12–13. Rhapsodes are frequently compared to actors at Plato, Ion 532d, 536a, and Republic 395a; Aristotle, Rhetoric 1403b22 and Poetics 1462a5–6; Alcidamas, On Sophists 14. On the comparison between sophists and oral poets in Alcidamas, see Ritoók 1991.

[ back ] 30. Ion (Plato, Ion 535e) comes right to the point: δεῖ γάρ με καὶ σφόδρ᾽ αὐτοῖς τὸν νοῦν προσέχειν· ὡς ἐὰν μὲν κλαίοντας αὐτοὺς καθίσω, αὐτὸς γελάσομαι ἀργύριον λαμβάνων, ἐὰν δὲ γελῶντας, αὐτὸς κλαύσομαι ἀργύριον ἀπολλύς “I must pay very close attention to them [the audience], since if I set them crying, I myself will laugh because of the money I get, but if I set them laughing, I myself will cry because of the money I lose.”

[ back ] 31. Note the usage of the verb rhapsôidein to describe “Homer” at Certamen 56. Cf. Plato, Republic 600d, in which both Homer and Hesiod are described as rhapsodes (rhapsο̂idein).

[ back ] 32. For a discussion of the terminology of improvisation, see Hammerstaedt 1996:1215.

[ back ] 33. On the runover position, see especially Edwards 1966:138. On Homeric enjambement in general, I mention only Basset 1926, Edwards 1966, Kirk 1976:146–82, Clark 1994, and Higbie 1990. The work on enjambement by Bakker 1990 and 1997:152–55, focusing as it does on cognitive units rather than the runover position in hexameter verse, is not relevant to the game in the Certamen.

[ back ] 34. On this example cf. Clark 1994:113–14.

[ back ] 35. Bassett 1926:122.

[ back ] 36. Cf. Martin 2000:410 again on the Homeric Hymn to Apollo. His discussion of expansion of Homeric formulae can be found at 1989:209–10 (splitting and replacement), 214–15 (elaboration), and 216–19 (telescoping).