10. The Rhapsode in Performance

In the previous chapters I examined the evolution of epic performance from the point of view of inspiration, authority and authorship, and the increasing adoption of scripted delivery. In this chapter I reconsider the diachrony of rhapsodic performance more narrowly through the meaning of ῥαψῳδός and the distinctive recompositional poetics of traditional rhapsodic delivery.

10.1 Understanding the Rhapsode

The greatest hindrance to a proper understanding of the Homeric rhapsode and his craft is perhaps the entrenched, and diachronically invalid, opposition between a ‘creative ἀοιδός’ and a ‘reproducing (uncreative) ῥαψῳδός.’ Scholars who resort to this polarity often seek to preserve compositional creativity up until the time of the alleged ‘monumental composer’—Homer—who, with his dictated texts, arrested the traditional practice of recomposition in performance. The treatment in Pelliccia (2003:97–98) is typical. Because he explores this opposition explicitly and at length, I will review his argument below (§10.1). A few brief comments will suffice for now to set the background for my own treatment. On page 97 he offers in two opposing columns a “crude” outline of the “ends of the spectrum of synchronic possibilities” for the concept of the rhapsode. The left (a)-column posits: (1a) an oral culture; [1] (2a) a creative rhapsode who (re-)composes in performance; (3a) a fluid, evolving Homeric tradition. The right (b)-column espouses: (1b) uncertainty regarding the question of an oral vs. a literate culture; (2b) an uncreative rhapsode performing memorized texts verbatim; (3b) a fixed text of the Iliad and the Odyssey. On page 98 he notes that many proponents of the right-hand (b)-column “are happy to posit the entire left-hand (a)-column as the pre-historic antecedent for the historical (b)-column situation; the divide between them is the living space of Homer himself, and the key event of this interstitial moment is the dictation of the two poems.”
But the dichotomy of creative singer versus uncreative rhapsode flies in the face of ancient descriptions of Homer and Hesiod as rhapsodes. [2] To the Greeks, Homer was a rhapsode who traveled the Greek world composing and performing before audiences for a living. [3] Graziosi (2002:47) objects that “[Homer] is never said to be a rhapsodos.” But since his mode of performance is rhapsodic (ῥαψῳδεῖν, ἐπιδεικνύναι … ῥαψῳδίαν) [4] and he travels from place to place (περιιόντας, περιέρχεσθαι) like an ordinary rhapsode, her objection is without force. [5] More apropos is her observation that the sources present Homer as composing his poetry first, and only then performing it rhapsodically. [6] This would depart radically from the model of recomposition in performance only if one assumed that by ποιέω the exhaustive word-for-word formulation of the text was meant, and that the ensuing performance was understood as Homer’s verbatim delivery of the memorized verses. This assumption arguably paints an odd picture of the great poet/performer sedulously memorizing his own words. Even when writing intervenes, I dispute that this is what the sources require us to believe of Homer. It is more plausible that they take for granted a measure of improvisatory elaboration in performance. [7] After all, even late-classical and Hellenistic rhapsodic performances arguably featured a degree of recomposition, however limited. Witness the circulation of Ptolemaic papyri with plus-verses and the variant readings preserved by the indirect paradosis. It would be odd to deny the great Homer at least that much freedom in his own delivery. Furthermore, one should consider how the Homeric Vita traditions would have likely depicted Homer’s activity if in fact they had believed him, as I claim, to be a rhapsode. By their time, the construction of an authorial source for the poetry we call Homeric had already taken place—a diachronic process explored at length earlier in this book. [8] If Homer was to be the archetypal rhapsode with unequaled mastery of the tradition—i.e. able to shape it definitively into recognizable ‘poems’ ascribable to him and to perform it with unsurpassed authority—the natural way to convey this would have been to posit a prior stage of composition followed by serial performance. Such a representation need not even rule out the delivery of ‘Homer’s poems’ by rival performers, as was said of Thestorides in the Vita Herodotea §16: having acquired from Homer the Phōkais and other epē in written form, he moved to Khios where he performed them as his own (τὰ ἔπεα ἐπιδεικνύμενος ὡς ἑωυτοῦ ἐόντα 215–216) [9] and won much praise. Identifying the real source of the poetry—Homer, not Thestorides—is clearly the focus of the anecdote, but this should not obscure the fact that the poem could be performed by anyone with access to it and the requisite skill. My point, in short, is that what Graziosi sees as an objection to Homer’s rhapsodic identity is precisely what we would expect if late sources had wanted to describe Homer as a rhapsode (however exceptional in skill). If “poietai … are often contrasted to those who perform their compositions” (Graziosi 2002:47), then surely we must judge significant that the Vitae should make Homer the subject both of ποιεῖν and ῥαψῳδεῖν. [10] Pindar confirms that the rhapsode may be subsumed under the ἀοιδός in Nemean 2 (dated tentatively to 485 by Snell and Maehler), which opens with a picture of the Homeridai: [11]
Ὅθεν περ καὶ Ὁμηρίδαι
ῥαπτῶν ἐπέων τὰ πόλλ’ ἀοιδοί
ἄρχονται, Διὸς ἐκ προοιμίου …
Here the nexus between ἀοιδοί and ῥαπτὰ ἔπεα seems entirely unproblematic, and Pindar even compares his own compositional technique to the frequent recourse the Homeridai have to proemial hymns to Zeus (Kurke 1991:42–43). No attention is drawn to a contrast between the creative poet Pindar and, insofar as rhapsodes, allegedly uncreative Homeridai. Nor is there a hint in this lofty opening of the contempt for rhapsodes we find in Xenophon’s writings.
Three things must be done to remove Pelliccia’s objections to the notion of a creative rhapsode. First, the words “creative” and “verbatim” must be set in a diachronic perspective; this is effectively the overarching goal of this book, and in this connection I have already addressed the heart of Pelliccia’s argument in an earlier chapter. [12] Second, the implications of Pelliccia’s evidence for verbatim performance in classical Athens must be clarified; this, I do in brief immediately below. Finally, as I do above (§9.2), the quotations from Xenophon regarding Athenian rhapsodes must be contextualized. [13] Regarding the matter of verbatim performance to which this scholar helpfully draws our attention, it is important to bear in mind that memory is a rather complex faculty, variously deployed in all sorts of intellectual activities. Therefore, a reference to the use of memory does not eo ipso entail ‘rote memorization’ with no tolerance for mouvance. [14] Whether such exists and, if so, in what degree, is the question that cannot be begged. One of the more interesting findings of the evidence Pelliccia assembles in 2003:111n30 is precisely the primacy that aurality (and hence orality) still held during the classical period, even where bookrolls were in view. [15] Phaidros 228a6–b1 illustrates this well. [16] After all, this dialog is at least in part an examination of the propriety of scripts and the extent to which they may adequately stand for the corresponding performance. Towards its end, the scope of the discussion broadens to consider the role of writing in the preservation and transmission of knowledge, [17] a concern that also informs the conversation in Xenophon’s Αpomnēmoneumata 4.2. Pelliccia makes much of the extensive training of citizen amateurs to deliver previously memorized poetic scripts. But, whatever its import for the availability to the average classical Athenian of a “concept of fixed, absolutely verbatim accuracy” (102), this training entails very little for the classical rhapsodic delivery of Homeric poetry. For what matters is the conscious expectation of performer and audience whether such accuracy was possible, desirable, and requisite in the delivery of Homeric epic and not in the amateur performance of the citizen chorus. One must surely weight the fact that ordinarily these citizens were amateurs of rather limited poetic skill, not professionals who could readily recompose in the manner of traditional rhapsodes; that they performed in groups, not solo, and had to coordinate carefully their dancing and singing, both melody and lyrics; and that they did not—and this is crucial—reperform the same pieces time after time. Thus, the poetry in view is not traditional and cannot illuminate the manner of composition and performance of traditional poetry. Neither the amateur performers nor their citizen audience would have had the same expectation of, or tolerance for, mouvance in these performances that they had for the rhapsodic performance of Homer. And insofar as the audience did not have any independent access to the scripts delivered by choruses other than their actual performance, there is no obvious sense in which these choral performances could have met an alleged expectation of verbatim reproduction. This is not to deny, of course, that an individual interested in acquiring a particular composition, perhaps for its private performance in the setting of the symposium, would not exert every effort to memorize it. [18] Pelliccia’s observations about amateur choric performance are interesting but largely pointless for their ostensible focus, the rhapsode, since they are specific to a different genre and emphatically regard performers of poetry that—all agree—are to be sharply distinguished from the authors of their poetic scripts. If one is not to beg the question, the same cannot be said of the rhapsode without further study. [19]

10.1.1 Etymology

Having briefly considered Pelliccia’s arguments, I will now seek to advance our understanding of the word ῥαψῳδός by approaching it from the point of view of etymology and usage in the extant texts. Scholars have long agreed that ῥαψῳδός is a composite noun of the τερψίμβροτος type. [20] The first element is verbal (perhaps in origin an abstract verbal noun in -τι-); the second, a noun. Because abstract verbal nouns often echo the aorist, the terpsimbrotos compound was closely tied to it in form: τερψι- to τέρψαι, and ῥαψ- to ῥάψαι. [21] Words of this type receive the accent on the first element, but ῥαψῳδός analogically follows the more numerous compounds whose second element conveys a verbal action: χρησμῳδός, ὑμνῳδός, θεσπιῳδός, κιθαρῳδός, etc. [22] The ancients were divided about its etymology, tracing it both to ῥάπτειν and ῥάβδος. [23] As the more transparent, the latter alternative was especially popular: the classical and Hellenistic rhapsode recited without an instrument, clad in gorgeous attire, and leaning on a staff (ῥάβδος). [24] The ῥάβδος was thought the equivalent of Hesiod’s σκῆπτρον (Theogony 30), itself an emblem of performative authority and the functional equivalent of the singer’s instrument. But there is no linguistic path from ῥάβδος to ῥαψῳδός, and although this erroneous etymology may inform the meaning of late instances of the term, it cannot explain its origin or early use.
Pindar is usually thought to vacillate and support both etymologies. I have already quoted Nemean 2.1–3, which speaks to ῥάπτειν. Isthmian 3/4.55–57 points to ῥάβδος:
ἀλλ’ Ὅμηρός τοι τετίμακεν δι’ ἀνθρώπων, ὃς αὐτοῦ
πᾶσαν ὀρθώσαις ἀρετὰν κατὰ ῥάβδον ἔφρασεν
θεσπεσίων ἐπέων λοιποῖς ἀθύρειν.
Patzer (1952:316n2), however, remarks that ῥάβδος θεσπεσίων ἐπέων in the first instance cannot be the staff of the rhapsode. It is far more likely that these words point to the preeminent acceptation of ῥάβδος in the Homeric poems, the ‘wand’ with divine or magic powers that Hermes, Kirke, and Athena wield. [25] Metaphorically, the expression would amount to ‘according to/by the enchanting power of his divine poetry’. [26] As Patzer further notes, “[e]ine Anspielung auf den ‘stabhaltenden’ Rhapsoden liegt freilich vor, aber pindarische Anspielungen zeugen noch nicht für ein gangbares und auch für den Dichter geltendes etymologisches Bewußtsein.” Whether or not Patzer is right, the popular connection between the rhapsode and the ῥάβδος, the emblem of performative authority, is undeniable, even if, as alleged, for Pindar the tie was not etymological.
ῥάπτω, then, is the only valid etymological explanation for the first element of ῥαψῳδός; ᾠδ-, with the appropriate masculine ending -ος, can be traced to ἀοιδή. The etymology of ῥάπτω is disputed. In the past, it was often assigned the root *wr̥p-, ‘to bend’, ‘to turn’. [27] But two arguments preclude this analysis. Mycenaean Greek, with its ra-pi-ti-ra2 (ῥάπτρια) and ra-pte (*ῥαπτήρ; cf. ῥάπτης), “ruine l’étymologie traditionnelle” (Chantraine 1999 s.v.). [28] There is also, as many have recognized, no clear semantic path from ‘to turn’ to ‘to sew/stitch’. [29] Unfortunately, the alternative *serp- is not semantically illuminating because it lacks parallels in other IE languages. [30] For ῥάβδος Patzer (1952:316) offers the gloss “die ‘biegsame’ Gerte.” But why should pliancy in a staff be desirable, much less its defining quality? The contrary can be easily argued. This gloss seems superficially plausible at Μ 297—the one Homeric use of the word that is clearly related to sewing—if ῥάβδοισι there means ‘wires’. [31] One might imagine the golden thread for sewing as a ‘pliant rod’ of sorts. But the manufacture of the shield is being described from the point of view of its final appearance. It is not easy to imagine a member of the audience visualizing the wires stretched like ‘pliant rods’ before the stitching. It is far more likely that ‘stitches’ were meant. [32] The scholiast’s ῥαβδοειδέσι ῥαφαῖς perhaps hints at the appearance of the stitches: if the artisan sewed the golden thread in and out of the hides along their perimeter, the dashed circle of exposed thread would have looked like a series of small rods.

10.1.2 Stitching the song: creative work?

If ῥάπτω is ‘to sew/stitch’, how are we to visualize the action of sewing? This question entails two related queries: does the metaphoric use of ‘sewing/stitching’ connote creative or uncreative work? What precisely in the practice of the rhapsode might have led him or his audience when the term was first coined to associate his performance with ‘sewing/stitching a song’? Related issues are where and when the term was first coined and used, how it gained currency, and why it became the preferred term for the rhapsode. Additionally, one should address the instances of ῥάπτω in the Iliad and the Odyssey and explain why its metaphoric use regularly refers to contriving evil or devising plots.
It is not hard to establish that early uses of ῥάπτω consistently point to creative work. This is doubtless the case in its metaphoric acceptation ‘to contrive’, ‘to plot’, where it takes the objects κακά (Σ 367 γ 118 π 423), φόνον (π 379), and θάνατόν τε μόρον τε (π 421–422). Creativity is explicitly asserted in the celebrated Hesiodic fragment that states, ‘I [Hesiod] and Homer sang Phoibos Apollo, stitching [our] song with new hymns’ (μέλπομεν ἐν νεαροῖς ὕμνοις ῥάψαντες ἀοιδὴν | Φοῖβον Ἀπόλλωνα, fr. 357 MW). However one construes the grammar and the relationship between ὕμνοι and ἀοιδή, [33] it is impossible to gainsay the adjective νεαροῖς. It is easy to find modern scholars who believe that the label ‘rhapsode’ was not coined for the reciter of memorized scripts at the twilight of the oral tradition of Homeric epic. If they are right, the rationale for the term must be sought in something other than its alleged contrast with the creative ἀοιδός, in whose hands (by common consent) traditional diction and thematic structure were still malleable. So, for example, Wilamowitz-Moellendorff (1927:175) wrote: “Der Rhapsode will Dichter sein; nichts anderes besagt der Name, aber niemand fragt, wieviel die Muse ihm von dem eingegeben hat, was er vorträgt, zumal der Stoff allmählich den Hörern kaum etwas ganz Neues war.” Fränkel (1925:4) translates ῥάπτω ἀοιδήν as “ich erfinde kunstreich ein Lied” and relates ῥάπτω to ἀρτύ(ν)ω. Schwartz (1940:9), a believer in the historicity of Homer, remarked: “[E]in Rhapsode des Namens muß tatsächlich gelebt haben und in einem Kreis … berühmt geworden sein zu einer Zeit, wo ein Teil der Epen schon bleibende Gestalt gewann, aber die Rhapsoden zu dem Vergangenen, das sie rezitierten, noch Neues hinzufügten.” Pagliaro (1951:44) comments that “la qualifica di ῥαψῳδός … nel senso di chi ‘cuce canti altrui’ diventa in sè poco qualificante e non necessaria: a chi sarebbe venuto in mente di darla al cantore, quando nessuno si sentiva impegnato a distinguere se i canti che costui recitava erano cuciti di epi propri od altrui?” Durante (1968b:279) so strongly insists on the originality of the rhapsode, that he even wrongly rejects the notion that the corresponding ‘stitching’ had anything to do with preexisting epic songs: “[D]ie Definition dreht sich um einen Punkt der epischen Dichtung, der—aus der Sicht der Späteren—wichtig erscheinen kann, der aber sicher eine ganz am Rande liegende Bedeutung einnehmen muß, wenn wir von Ort und Zeit ausgehen, als der Ausdruck geprägt worden ist.” With a more nuanced view, Patzer (1952:318) writes that “[die] Verwendung fester vorgegebener ‘Werkstücke’ nichts mit dichterischer Originalität oder gar Güte zu tun hat … sie [ist] so alt wie die epische Dichtung selbst und [war] mit deren ursprünglichem Stegreifvortrag gegeben … [E]ine atomistisch verstandene Originalität in der epischen Gattung [konnte] für ein frühgriechisches Publikum kein Maßstab für Güte sein.” Finally, Ford (1988:300) observes that “in a number of relatively early sources we find ῥαψῳδός, ῥαψῳδία, and ῥαψῳδεῖν used of performers who do not merely recite the poetry of others but also create their own”; and he adds that “[i]t is … unwarranted to attribute to the archaic age a clear and significant distinction between creative singers and mechanically imitative rhapsodes.” This is only a representative sample of those who variously trace their way to Wolf (1876:60): “Quamvis vero artis huius nomen videatur posterius esse Homero, ipsa ars et professio iam antiquissimis temporibus viguit … nullumque prope fuisse rhapsodum, quin idem probabilis poeta esset, manifesta historiae vestigia arguunt.”
The competing view that ‘rhapsode’ stricto sensu precludes or severely limits compositional creativity, although broadly represented, is characteristic of modern British scholarship and its American adherents. Even though Kirk (1962:312) allowed that the stitching metaphor probably referred “not to the joining of different poems but to the interlocking of phrase with phrase, verse with verse and theme with theme,” he nevertheless made a clear divide between “the ἀοιδός or true oral singer” (312) and the rhapsode or “reciter, the performer, the reproducer by rote” (313) who “polluted” the “unadulterated oral period” (318). [34] If the two Companions to Homer are at all representative of the consensus where they address themselves to the transmission of the Homeric text, we must note their agreement on this one point: “[R]hapsodes were primarily executants, performing works composed by others, rather than original composers” (Wace and Stubbings 1962:218); “I see no alternative to dictation … . The Homerids of Chios will have performed hereditary functions, and perhaps the poems … . A change in mode of performance is signalled by the switch from the Homeric lyre to the Homeric staff: Homer sang (at least, his bards do), rhapsodes recited” (Haslam 1997:81–82). These views mislead for their lack diachronic depth. They retroject classical and late-classical attitudes about rhapsodes that cry for historical contextualization and diachronic nuance. Sometimes we even meet with the claim that ‘rhapsode’ is a late coinage actually devised to draw a stark contrast between the scripted reciter of epic and the ‘true poet.’ Advocates of an essential difference between ‘rhapsode’ and ‘singer’ or ‘poet’ often espouse a flat synchronic reading of the Panathenaic rule. [35] Although their interpretation is therefore ahistorical, its fundamental features are in turn judged inherent to rhapsodic performance: epic performers qua rhapsodes recited a written script in sequence, each picking up the thread of the story precisely where the previous one left off. Rote memorization and verbatim reproduction are held to be the rhapsode’s defining traits. Proper diachronic analysis clarifies where these readings go astray and puts us on the right path to an understanding of rhapsodic performance. Armed with this understanding, we shall be in a position to grasp the essential nature of Homeric poetry and the role of performance in its genesis, diffusion, textualization, and preservation.

10.2 Understanding Rhapsodic Performance

Seeking to determine what precisely about his performance qualified the rhapsode for the label ‘song-stitcher’, Ford 1988 has put the accent where it belongs: the correct answer must concern something peculiar to him that was intelligible and readily perceptible to the audience. Otherwise, it is hard to see why the new coinage should have been widely received and adopted as apropos. Ford thinks that rhapsody was understood “irrespective of originality” as the “solo presentation, in public, of a poetic text without musical accompaniment” (303). I will deal with his argument presently. For now, mark his emphasis that this sense of the word “is based on an obvious aspect of the performance rather than a literary conception of genre” (305). [36] Since the earliest instances of metaphoric ῥάπτειν do not support the suggestion that ῥαψῳδός is an ironic or derisive designation (pace Schmid and Stählin 1929:1.157 [37] and Else 1957b:33), we must look for distinctive thematic or formal characteristics that would distinguish the rhapsode from the more generic ἀοιδός. [38] For Meyer (1918:332) it is his style of performance, the artful stitching together of individual recitations delivered piecemeal over several consecutive days. [39] For Fränkel (1925:4) the emphasis falls on “artful invention,” which evil scheming and outstanding oratory share. [40] For Pagliaro (1951:44), rhapsody refers to the technique peculiar to epic recitation, specifically, to individual performers reciting cooperatively their song (‘the rhapsody’). For Patzer (1952:323), it is the “‘monostichische’ Prinzip,” that is, the epic singer who composes “[eine] Reihung unzähliger aneinander schließender gleicher Glieder.” For Else (1957b:32–33), rhapsodes competed by reciting individual contributions not of their own making, which, “taken together, produced some kind of whole.” For Durante (1968b:281), they composed by advancing, intertwining, or plaiting the thread (οἴμη) of their narrative. [41] For Ritoók (1962:228), rhapsodes engaged in sequential delivery, joining their songs one to another before an audience not merely by following an established script but also by composing extemporaneously. All of these proposals have one or more elements of truth; some are even insightful. But, because they were not formulated with the diachronic development of epic performance in view, they often set in opposition mere matters of emphasis that shifted with time or else they set at odds what are in fact distinct stages of the history of Homeric performance. My aim below will be to take what is true, or plausibly true, in the opinions surveyed above, and place it in a diachronic framework that offers an authentic picture of the origins and development of the Homeric rhapsode and his craft.

10.2.1 Non-melodic recitation?

To start with Ford 1988, as I observed above, he argues that the defining qualities of ‘rhapsody’ were solo delivery and the absence of musical accompaniment—a mode of performance characteristic of stichic and elegiac verse in the classical period [42] —not its degree of originality or its specific association with the meter of epic. In support of his conclusion, he adduces that fourth-century instances of ῥαψῳδός and its derivatives (some of them referring to the fifth century) concerned the performance of poetry other than epic—in particular, poetry by Arkhilokhos. Ford claims that a study of the relevant Aristotelian loci confirms that the non-melodic recitation of stichic and elegiac verse is in view.
If there is an overriding objection to Ford’s argument, it is that he cannot, and does not even attempt to, explain why the metaphor of sewing or stitching should have been thought apposite to the mode of performance and poetic material peculiar to the rhapsode. What associative semantic path leads from a solo, non-melodic public recitation to ‘stitching’ a song? Unless he is willing to re-embrace Patzer’s view (which he seemingly rejects as insufficient at 301–302), [43] what might possibly suggest the metaphor of stitching to portray performance without instrumental (and vocal) melos? Since Ford does not tender any explanation for the coinage of the term ῥαψῳδός, his article only proves what is hardly in dispute: that by the mid fourth century ῥαψῳδέω was applied to the solo, non-melodic public performance of stichic and elegiac verse. But this state of affairs could just as easily be explained as the outcome of a diachronic evolution from ‘song’ (melic meters with musical accompaniment) to ‘poetry’ (stichic and elegiac meters, first with reduced melody, then in recitative [44] with or without an instrument, finally simply recited as ordinary, if emphatic, speech, and without instrumental accompaniment). [45] This diachronic scenario allows for the view that rhapsodic performance—i.e. the mode of performance peculiar to the professional for whom ῥαψῳδός was originally coined—was at first tied specifically to epic and made use of a κίθαρις or φόρμιγξ (as was true of the Homeric ἀοιδός).
I believe that ‘singer’ (ἀοιδός) and ‘rhapsode’ (ῥαψῳδός) were concurrent designations for the performer of epic. Homeric poetry, in a characteristic display of conservatism, does not depict its own performance as it actually was for most of Greek history. It even excludes features already typical of the formative archaic period. Instead, it offers a suitably refracted portrayal of what it must have been in origin or at least imaginatively recreates its melic roots with authentic diachronic insight. As many have noted, neither the Iliad nor the Odyssey makes explicit reference to the rhapsode or to the stitching of poetry. There are good reasons for this silence which I shall detail below. [46] But this no more rules out the application of the name ‘rhapsode’ to the early-archaic performer of the poems than the failure to mention festivals precludes the eighth- and seventh-century performance of Homeric poetry at various archaic poleis and Panhellenic festivals. As regards the particular mode of performance, ἀοιδός was the unmarked, ῥαψῳδός the marked, term. ἀοιδός was applicable (at least originally) to (professional) performers of all sung poetry, including lyric meters with choral dancing [47] —i.e. to all performers of ‘song,’ the broader field of marked speech out of which hexameter epic evolved. [48] ῥαψῳδός referred specifically to the performer of traditional epic. It reflected the audience’s perception of what was formally and thematically characteristic of this oral traditional medium of recomposition in performance. The formal criterion did not in origin rule out instrumental accompaniment, although I do believe that it regarded a particular type of music and choice of instrument. This is not to say that the rhapsode qua rhapsode at any one time necessarily performed with instrumental accompaniment any more than the Homeric singer qua singer. Both were, after all, alternative labels for one and the same performer. The point of my argument is that the profile of the rhapsode is no more and no less ‘melic’ than that of the Homeric singer. The title ‘rhapsode’ focuses attention on features of his craft that are largely indifferent to the use of an instrument and the presence of melody. The Homeric performer owes to his presence in the poems qua ἀοιδός his association to instrument and melos in the epic medium. His absence qua ῥαψῳδός means that references tying rhapsode, melos, and instrumental playing must be sought elsewhere. Whether as ‘singer’ or ‘rhapsode,’ with the passing of time the Homeric performer shed the instrument. Since his was the preeminent public performance during the classical period without any instrumental accompaniment, this differentia was readily seized then as the defining characteristic of the rhapsodic mode of performance. This explains the easy naivete with which the performance practices of the classical rhapsode were sometimes retrojected without warrant to the archaic period, when a uniform non-melodic practice cannot be assumed. Whether epic poetry was then sung or delivered with reduced melody or as recitative, with or without an instrument, or else merely declaimed as spoken verse, cannot be finally determined for any particular occasion and performer (e.g. Xenophanes, on whom see below, §10.2.1). It is likely that all modes of performance were used at one time or another, as the medium evolved first towards reduced melody, then to recitative, and, finally, to spoken declamation. If the practice, then, of the fourth-century epic performer—and hence the fourth-century use of ῥαψῳδός—does not necessitate a static view of the rhapsode as one who ab origine never accompanied his delivery with an instrument, to assume that the rhapsode qua rhapsode always declaimed without musical accompaniment is simply to beg the question that must be answered.
But even as Ford fails to explain the coinage of ‘rhapsode’ and its differentiae, he faces three further obstacles. Out of all early stichic and elegiac composers, why should Arkhilokhos alone have formed part of the repertory of rhapsodes? The prologue and ῥῆσις of tragedy, allegedly Thespis’ innovations, must have been solo, non-melodic, stichic set speeches. Why then was ῥαψῳδέω never used in connection with them or, for that matter, applied to the performance of the longer spoken passages of Attic drama? (To messenger speeches, for example.) Finally, Ford fails to explain the ready association of ἀείδω and its semantic family with ῥαψῳδός and its derivatives.
To start with this last point first: the usual reply is that ἀείδω often seems interchangeable with verbs of speaking, or that at least it is not always clear that it must mean ‘to sing [poetry to melody]’. Therefore, that ῥαψῳδός points to ἀείδω does not necessarily support the claim that rhapsodic performance ever involved a melic component. [49] But the very fact that we are not always able to draw a clear synchronic contrast between verbs of speaking and singing is itself diachronically significant. Just as the marked ῥαψῳδός stands in contrast to the unmarked ἀοιδός, so also ἀείδω and various verbs of speaking (εἰπεῖν, (κατα)λέγειν, ἐννέπειν, etc.) may see their underlying diachronic contrast activated by a given context, even though ordinarily they overlap as synchronically interchangeable within the system of Homeric diction. Whether ‘speaking’ is the marked, and ‘singing’ the unmarked, term when the contrast obtains or vice versa depends on the context. The reason for this superficially puzzling semantic structure is that, synchronically, Homeric diction puts on a par older and newer modes of performance (melodic, reduced melodic, and recitative). [50] Diachronic analysis in turn reveals them as successive stages. As Nagy (1990c:21) writes, “[s]elf-references in Archaic Greek poetry may be diachronically valid without being synchronically true.” [51] But, to return to ῥαψῳδός, if by the time of its coinage the rhapsode’s craft precluded even a reduced melodic delivery, why should the coinage not have resorted to an alternative like λόγος or ῥῆσις, rather than the potentially melodic ἀοιδή? And if ἀείδω had, in fact, already shed its necessary association with melodic singing when ῥαψῳδός was coined, what need was there for a new term, different from ἀοιδός, to designate a performer whose delivery was non-melodic? On the other hand, if ἀοιδός had not lost its intimate association with melos, why coin a potentially confusing term whose second element might connote melody (-ῳδός), only to purge its melic connotation with the choice of ῥάπτω for its first? There is something paradoxical, if not incoherent, about the logic. Furthermore, if Ford is right, μέλπω, with its strong semantic association with instrumental song and dance (some might even say, with choral performance), [52] seems an especially perverse choice for the Hesiodic fragment quoted above. Why not resort instead to the traditional and more neutral φαίνομεν or ἐντύομεν, governing ἀοιδήν apo koinou with ῥάψαντες? [53]
As is often the case, a learned Hellenistic poet helps to illuminate the diachronic evolution of the rhapsodic craft. [54] Kallimakhos’ Hymn to Zeus 78–79
[ὑδείομεν] … Φοίβου δὲ λύρης εὖ εἰδότας οἴμους·
‘ἐκ δὲ Διὸς βασιλῆες’

[we say that] those who know well the tracks of the lyre belong to Phoibos
‘but basileis are from Zeus’
reprises Hesiod’s Theogony 94–96:
ἐκ γάρ τοι Μουσέων καὶ ἑκηβόλου Ἀπόλλωνος
ἄνδρες ἀοιδοὶ ἔασιν ἐπὶ χθόνα καὶ κιθαρισταί,
ἐκ δὲ Διὸς βασιλῆες·

For from the Muses and far-shooting Apollo
there are men on the earth who sing and play the cithara,
but basileis are from Zeus.
At first, Kallimakhos’ reprise seems straightforward: cithara corresponds to lyre, and ἀοιδοί is glossed by εὖ εἰδότες οἴμους. But, as I make clear below (§, the word οἶμος, intimately related to οἴμη, is peculiar not to the unmarked ἀοιδός but to the marked rhapsode. In effect, Kallimakhos has applied to the rhapsode and his craft the Hesiodic reference to the epic ‘singer’ and his instrument [55] —a refraction of the melic origins of the epic genre. By implicitly substituting the marked, narrower term for the unmarked, broader one while keeping the reference to melic delivery, Kallimakhos confirms that ‘rhapsode’ is not a late term coined for a reproducing reciter of textually fixed Homeric poems, but a label that focuses attention upon the rhapsodic craft of recomposition of traditional epic themes and diction in performance.
A similar point may be made on the basis of Pindar’s Isthmian 3/4.55–57, which was already cited above (§10.1.1) and I shall have a further occasion to revisit below (§10.2.1). Note the verb ἀθύρειν at 57. As Wilamowitz-Moellendorff (1922:339n3) observed, it is an epexegetic infinitive. As to meaning, after asserting that “[f]erner ist ἀθύρειν immer nur ‘spielen’,” he glossed it thus: “Homer made the deeds of Ajax conspicuous to future men, to make it easy now to mention them occasionally.” But there is more to ἀθύρειν than the bland ‘to mention with ease’. Race (1997:167), like Slater (1969:15) before him (“sport with, take delight in”), enlarges the verb’s semantic field to embrace the meaning ‘to enjoy’ (“for future men to enjoy”). Thummer is nearer the truth with “spielend besingen” (1968–1969:74), although he misses the performative implications of ἀθύρειν, which are more pointed that a generic ‘playing’ or ‘playfully’, and in the end he reverts to Wilamowitz’s “spielerisch leicht.” Willcock (1995:80) is right to particularize λοιποῖς to the professional of poetry (“for later poets to make their entertainment from”), although he seems to have in view composition apart from performance and only non-epic poets. [56] I submit that Pindar is drawing on the language of epic performance, specifically, the performance of the Homeric hymns, which makes ἀθύρω and ἄθυρμα preferred terms for the joyful playing of the lyre and the singing and dancing that regularly accompany it. In the Hymn to Hermes 52 the lyre is called a ‘lovely plaything’ (ἐρατεινὸν ἄθυρμα) and the tortoise is called ἄθυρμα proleptically at 32 and 40. At 484–485 we further learn that, in the context of the ‘plentiful banquet, the lovely dance, and the glorious revel’, [57] the utterance of the lyre teaches every manner of thing delightful to the mind when she is played (ἀθυρομένη 485) with ease and tender intimacy. Homeric Hymn 19.15 portrays Pan ‘playing his song (μοῦσαν ἀθύρων) sweetly to the reed pipes’. His is not, as one might think, a solo instrumental performance. To his melodious playing, compared to a lament (θρῆνον 18) and a honey-voiced song (μελίγηρυν ἀοιδήν 18), the mountain nymphs dance and sing (μέλπονται 21), hymning the gods (ὑμνεῦσιν δὲ θεούς 27) and telling of Hermes alone above all, [58] of his love for Dryops’ daughter and of Pan, their child. Thus, in circular fashion, what started with the rhapsode’s invocation of the Muse to tell of Hermes’ child, Pan, becomes in performance one and the same with the hymnic performance of the nymphs, who sing and dance to the piping of Pan himself. This self-referentiality endows the rhapsodic performance of the epic hymn with the representational character of melic khoreia. In Pindar, the melic character of the epic performance entailed by ἀθύρειν is clear, and even the scholiast marked it: τὸ δὲ ἀθύρειν ἀντὶ τοῦ μέλπειν καὶ ὑμνεῖν. [59] It is not that in Isthmian 3/4.57 Pindar has only rhapsodic performance in view, but that with the language of rhapsodic hymnody he encompasses the performance of all future poets who might make Ajax their theme. That the epic medium should portray the performance of Homeric hymns as melic may be attributed to the conservative impulse that animates diachronic skewing. [60] That Pindar’s intertextual echo should embrace this traditional melic representation (note his reference to hymnody at 61, πυρσὸν ὕμνων) betrays a keen sensitivity to the diachronic development of rhapsodic performance.
I believe that Else (1965:59–60) is right to look to the rhapsode and his craft in his search for a model that might have inspired the dramatic prologue, which Themistios, on the authority of Aristotle, ascribed to Thespis. [61] Else’s proposal that the rhapsode and his performance help to motivate and explain significant features in the origin and growth of Attic drama agrees with my own convictions, independently arrived at and explored at length in this book. [62] Although I often differ with him in matters of detail, this essential insight is sound and productive. If so, it is all the more surprising that ῥαψῳδέω was never used in connection with the performance of the prologue. The retort that this verb was too closely tied to epic performance to suggest itself in the context of a new genre—however much epic and its professional performer might have inspired various aspects of it—undermines Ford’s argument that stricto sensu this verb must not be understood as the delivery of epic but as the solo, public performance of non-melodic verse. [63] Neither is the dramatic ῥῆσις the object of ῥαψῳδέω, [64] although, as Else (1965:54) observes, “Wilamowitz pointed out, and Aurelio Peretti later showed in detail, that the structure of the tragic rhēsis in the earliest plays of Aeschylus has nothing in common with choral lyric but goes back to Ionian models: epic, elegiac, and iambic poetry.” [65]
A third significant objection to Ford’s argument is that of non-epic poets Arkhilokhos alone was associated with rhapsodic performance during the classical period (i.e. before the reforms of Demetrios of Phaleron). This suggests that there was something unique about him or his poetry that led to his inclusion in the rhapsodic repertoire. With him alone does Homer share Herakleitos’ censure (DK 22 B42), and his is the only name added by Sokrates to Homer and Hesiod as potentially within the expert purview Ion the rhapsode. [66] Had the rhapsodes’ engagement with Arkhilokhos merely concerned their public, non-melodic recitation of his iambic, trochaic, or elegiac poetry, why should he alone have been singled out for this honor? The implications of Sokrates’ question to Ion are arguably confirmed by Klearkhos, who in the first of his two books On Riddles wrote that Simonides of Zakynthos used to rhapsodize the poetry of Arkhilokhos in the theaters, sitting on a stool. [67] Given the date of Klearkhos and the context of Athenaios’ passage, it is just possible—but hardly certain—that Simonides’ Arkhilokhean performances followed the innovations of Demetrios. [68] Lysanias’ assertion (in the first book of his On Iambic Poets) that Mnasion the rhapsode in his public performances used to deliver (ὑποκρίνεσθαι) [69] some of the iambs of Semonides [70] only proves that after the reforms of Demetrios and the theatricalization of the rhapsodes, their repertoire embraced material that in the archaic and high-classical periods would never have been thought within their traditional purview. [71] None of this helps to advance Ford’s argument, whose interest is not the broadening of the repertoire of rhapsodes but the claim that ‘to rhapsodize’ ab origine concerned a particular type of performance. The paradoxical entailment of his views is that we need not even think of ‘rhapsodizing’ as the purview of the ‘rhapsode’. Any solo performer of non-melodic iambic, elegiac, or trochaic poems could be said to ‘rhapsodize’ his material, even if he was not a ‘rhapsode’ in any traditional sense of the term. [72] The singing of Solon’s poetry at the Apatouria (Plato Timaios 21b, cited by Ford 1988:302), falls under the category of rhapsodic performance understood as ‘relay poetics’ (on which more below, §, that is, the cooperative-competitive stitching of a song by two or more rhapsodes in turn, each taking the thread of the poem precisely where the previous one had left off. The boys’ ‘rhapsodic competition’ (ἆθλα … ῥαψῳδίας Timaios 21b3–4) transparently reproduces the arrangements and dynamics operative at the Panathenaia, attributed to festival reforms instigated either by Solon or by Peisistratos (see below, § [73]
To return to Arkhilokhos, critics of Notopoulos 1966 may be right that Arkhilokhos did not compose epic poetry now lost to us, although, in light of his versatility, [74] we should not be surprised to learn that epic poetry too had been attributed to him. [75] The most compelling alternative to the existence of Arkhilokhean epic relies on the typological opposition between epic as praise, and iambic as blame, poetry. [76] This proposal views either genre as the Panhellenic endpoint of the centuries-long evolution of Greek musical traditions. The corresponding diachronic development would have resulted in thematic specializations of opposite sign—epic κλέος versus iambic ψόγος—and a musical stylization that crystallized respectively into the dactylic hexameter and the iambic trimeter, both stichic meters performed to reduced melody and recitative speech cadences. The contrasting thematic emphases served broadly complementary cultural functions: public approval and public censure. The complementary uses of epic praise and iambic invective and their shared Panhellenic scope might well account for their joint adoption into the repertoire of the preeminent Panhellenic performers, the rhapsodes. [77] The Homeric Margites may itself instance a tradition that, as to form, blended epic and iambic meters precisely to effect a humorous abuse poem of Panhellenic projection. [78] At any rate, granting ex hypothesi that it took place, one would have to consider the incorporation of iambos into the rhapsodic repertoire a subordinate outcome of the essential character of rhapsodic performance at an already advanced stage of Panhellenic development. Panhellenic iambic invective would have remained a secondary growth that must be understood against the backdrop of the defining rhapsodic mainstream, Panhellenic praise poetry. [79] As perhaps the oldest of all named iambic poets, [80] whatever his biographical reality, [81] Arkhilokhos doubtless embodied the broadest range of traditional iambic poetry. Therefore, his poems would have had the greatest potential for, and claim to, Panhellenic status. [82] This fact would explain why he alone should have been grouped with Homer and Hesiod as the canonical trio of the rhapsodic repertoire. Note that this alternative explanation for the engagement of rhapsodes with Arkhilokhos differs crucially from Ford’s proposal. It relies primarily not on formal arguments but on the complementary cultural uses of iambos and epic. The fact that both featured non-melodic, stichic verse types perhaps facilitated their joint adoption by one and the same professional. But these thematic considerations are rather more significant than the formal ones that Ford alleges as the differentia of ancient ‘rhapsody’.
Ultimately, we will never know for sure what singled Arkhilokhos out as a likely companion of Homer and Hesiod in the repertoire of rhapsodes, but given that there are plausible alternatives to Ford’s argument, it is not safe to make this assemblage a cornerstone of our interpretation of ῥαψῳδός and ῥαψῳδέω. Finally, even if we can trust Diogenes Laertios’ report that Xenophanes ‘rhapsodized his own poetry’ (αὐτὸς ἐρραψῴδει τὰ ἑαυτοῦ 9.18), we can only guess what form his performance took (reduced melody for his hexameters? speech for elegy and iambos?) and whether Diogenes’ choice of ἐρραψῴδει is anachronistic or not. By the third century AD, the meaning of ῥαψῳδέω had unquestionably devolved into the looser one that Ford wishes to retroject into the archaic period: solo, non-melodic public recitation of non-lyric meters. But would Xenophanes (or a contemporary of his) have used this verb for his performance? And, if so, for poetry in all the three meters attributed to him? Even if we grant, arguendo, the non-melodic character of Xenophanes’ recitation of iambic and elegiac poetry, must we assume the same of his performance of hexameter poetry? [83] Perhaps Ford’s conjectures are right, but since the alternatives are not implausible we simply cannot know. [84] Even the support sought in Solon’s recitation of the Salamis is dubious. Plutarch writes that Solon ἐν ᾠδῆ διεξῆλθε τὴν ἐλεγείαν, [85] while both Demosthenes and Polyainos resort to ᾖδε. [86] The poem itself features the celebrated juxtaposition κόσμον ἐπέων ᾠδὴν ἀντ’ ἀγορῆς θέμενος, ‘having composed a song for my public address, a lay of epē’. [87] Here Solon explicitly disowns ordinary public speaking. It is possible that his performance was ‘song’ and not ‘speech’ not only because of its meter but also because Solon sang it with the reduced melodic contours typical of rhapsodic delivery without a kitharis.
The case of Khairemon’s Centaur most likely proves the Aristotle could resort to ῥαψῳδία for its etymological sense of ‘stitched song’—here a ‘mixed song consisting of all meters stitched together’ (Poetics 1447b22). Ford must be right that, in its context, the Centaur is marshaled as an example of a mimesis of non-lyric “mixed meters.” [88] But it hardly follows from this that Athenaios was in error when he called it a ‘polymetric drama’ (δρᾶμα πολύμετρον Athenaios 13.608e). It could still serve as Aristotle’s illustration if the non-lyric portions of this alleged ‘drama’ were of the mixed metrical type. [89] That the Centaur, although a drama, should serve to demonstrate the lack of a proper name (cf. ἀνώνυμος at 1447b9) when a mimesis ‘only uses meters, and these … mixing them one with another’ might seem problematic, since ordinarily drama calls for melos and dancing rhythms. There are two possible explanations for the philosopher’s choice. Examples of metrical mixtures of the sort he had in mind were probably rare and hard to find, and only Khairemon, his contemporary, came readily to mind; his work must have contravened Aristotle’s feeling for proper dramatic form and, consequently—here at least, though not in the Rhetoric—he refused it the corresponding generic label. [90] And, as we learn from Rhetoric 1413b12–13 (see further below, §14.2), the λέξις of Khairemon’s oeuvre was γραφική and, therefore, he, as a poet, ἀναγνωστικός: this hints at Khairemon’s emphasis as a composer on the non-choral passages of the Centaur and at a comparatively thin mimesis of action. This would have led Aristotle in turn to focus on the mixture of non-lyric meters. The argument of the Rhetoric ad loc. actually suggests that Khairemon is brought forward as an ‘anagnostic’ dramatist, just as Likymnios was as an ‘anagnostic’ dithyrambist. [91] Someone may object that if as Athenaios reports the Centaur had really been a drama, the label ‘drama’ (or the narrower ‘comedy,’ ‘tragedy,’ or ‘satyr play,’ as the case might be) should have sufficed Aristotle. But this need not follow. Aristotle wants to illustrate the statement that a mimesis that only uses ‘meters’ has no proper name ‘until now’. His point is that poems may have (or poets may use) the same meter and yet be of a very different mimetic character. Labels that merely specify their meter misrepresent their character and cannot be considered adequate nomenclature (the mimēseis remain ‘anonymous’). The same is true of metrical mixtures: a name that identified the meters used would still fail to capture the nature of the mimesis. As regards nomenclature, the best that one could do with Khairemon on the basis of his Centaur would be to call him a poet (‘one would have to call him “poet”’). In view of its exceptional mixture of meters, it would be a misrepresentation to call it a drama and thereby suggest its similarity to ordinary tragedies, comedies, or satyr plays.

10.2.2 Ῥάπτω and Homeric artistry

It is time to look at the instances of ῥάπτω and its cognates in the Homeric poems. As I do, I wish to probe the following claims made by scholars in this connection: first, that the poems do not apply ῥάπτω to verbal artistry and therefore the name ‘rhapsode’ for the epic performer must be post-Homeric; [92] second, that ῥάπτω does not belong among the traditional Indo-European metaphors for poetic composition; and third, that, given its consistent application to plotting and scheming, this verb does not designate performance, only composition, and calls into question the alleged practice of recomposition in performance.
Starting with the last point first, my contention is that the Homeric use of ῥάπτω is in every way parallel to that of ὑφαίνω. Insofar as the latter is accepted as a well-established metaphor for composition, so also must the former. The poems (especially the Odyssey) use both metaphorically for the fabricating of plots and evil schemes. We may read signal instances of either verb metapoetically with reference to the poem in progress, and, hence, to the performance of its narrative. There are good reasons for the bard’s choice of the comparatively more traditional ἀοιδός and ἀείδω in depictions of performance within the poems. Not only because ῥαψῳδός, with its contraction -αοιδ- > -ῳδ- belongs to later stages of the tradition (and correspondingly appears in the Homeric hymns only), [93] but also because the rhetoric of inspiration would be poorly served by a professional label that focuses so squarely on the singer’s compositional technique, performance practice, and role in the construction and delivery of the poems. [94] There is no contradiction in archaic thought between divine inspiration and fully conscious human instrumentality. But where artistry is concerned—precisely the aspect emphasized by manufacturing metaphors (carpenter, stitcher, etc.)—the performer is careful to attribute his skill to the goddesses. Β 488–493 provides a classic example: [95]
          ἔσπετε νῦν μοι Μοῦσαι Ὀλύμπια δώματ’ ἔχουσαι·
485    ὑμεῖς γὰρ θεαί ἐστε πάρεστέ τε ἴστέ τε πάντα,
          ἡμεῖς δὲ κλέος οἶον ἀκούομεν οὐδέ τι ἴδμεν·
          οἵ τινες ἡγεμόνες Δαναῶν καὶ κοίρανοι ἦσαν·
          πληθὺν δ’ οὐκ ἂν ἐγὼ μυθήσομαι οὐδ’ ὀνομήνω,
          οὐδ’ εἴ μοι δέκα μὲν γλῶσσαι, δέκα δὲ στόματ’ εἶεν,
490    φωνὴ δ’ ἄρρηκτος, χάλκεον δέ μοι ἦτορ ἐνείη,
          εἰ μὴ Ὀλυμπιάδες Μοῦσαι Διὸς αἰγιόχοιο
          θυγατέρες μνησαίαθ’ ὅσοι ὑπὸ Ἴλιον ἦλθον·
          ἀρχοὺς αὖ νηῶν ἐρέω νῆάς τε προπάσας.
Β 484–493
          Tell me now, you Muses who have Olympian dwellings
485    —for you are goddesses and are present and know all things
          but we only hear the report and know nothing—
          who the leaders and chiefs of the Danaans were.
          But the host I could neither narrate nor name,
          not even if I had ten mouths and tongues,
490    an unbreakable voice and a heart of bronze,
          unless the Olympian Muses, daughters of aegis-bearing Zeus,
          should mention all those who came beneath Ilios.
          I will tell in turn of all the ships and their leaders.
Some scholars, like West (1999a:189–190), have condemned verses 491–492 (or 491–493) without any textual support on the presumption that a later interpolator must have misunderstood the meaning of πληθύς as the number of the ἡγεμόνες Δαναῶν καὶ κοίρανοι (Β 487); since the bard does go on to enumerate and name them—so goes their conjecture—the interpolator would have felt the need for the exception ‘unless the Olympian Muses’ etc. I find this hypothesis unnecessary. The αὖ of Β 493 marks a contrast between the (allegedly misunderstood) πληθύς and the ἀρχοί. It seems clear to me that the logic of the thought is as follows: 1) tell me, Muses, who the leaders of the Danaans were (Β 484 487); 2) I could not list them without you (implied), for you alone were eyewitnesses and can convey the verbal report (κλέος) that we, in turn, hear (as I declare it) (Β 485–486); 3) the host I could not tell, regardless of personal vigor and skill—so many were they (Β 488–490); 4) unless the Muses chose to mention them too (Β 491–492); 5) but I will tell the catalog of leaders, for the Muses, whose speech I relay, are enumerating them to me as I speak (Β 493). In other words, in 485–486 the singer acknowledges the necessary mediation of the Muses for the catalog of the leaders; ἐρέω at 493 implies that the Muses are answering his request in 484–487 for this catalog. The intervening 488–490 serve rhetorically to underline the immensity of the Greek contingent that sailed to Troy: the singer’s abilities—the strength of his voice, his courage—would be insufficient to enumerate them all. Hence, he shies away from the task (which is mentioned solely for rhetorical purposes) and only asks for the leaders. But an allowance must be made for the overriding will of the inspiring goddesses. Who is the singer to limit what a god can or cannot do? Or to limit what the god can or cannot require him to do? Whether we accept the text as uniformly transmitted or as emended by West and others, there remains a rhetorical contrast between the skill and abilities of the poet—which cannot surmount the barrier of time, compensate for his lack of autopsy, and is physically unable to detail in full the massive undertaking of the Trojan expedition—and the capacities of the Muses, who can supply his need and make him equal to the task if so they choose. Individual skill and inspiration must ultimately cooperate in public delivery. They are distinct and complementary, in mutual tension to a degree, although priority is piously assigned to the divine element.
Anther signal instance that makes the same point is Phemios’ remarkable plea: αὐτοδίδακτος δ’ εἰμί, θεὸς δέ μοι ἐν φρεσὶν οἴμας | παντοίας ἐνέφυσεν (χ 347–348). To underline his divine commission, the singer speaks in the same breath of being ‘self-taught’ and of the god having caused narratives of every sort to grow in his mind. [96] αὐτο- in αὐτοδίδακτος does not oppose self to god, but emphasizes the divine source of Phemios’ song at the expense of his apprenticeship to another master: ‘I did not learn my songs from another man, but from god’. [97] The statement implicitly acknowledges that, ordinarily, exceptional skill was coordinately attributed to divine favor and to training under a superior master of the trade. What is commonly complementary can be set in tension or even opposition in order to serve extraordinary rhetorical ends.
Given that the authority of his performance was rhetorically premised on a totalizing view of divine inspiration, it would have been counterproductive for the ἀοιδός to highlight his artistry through the image of sewing or stitching. It is hardly a coincidence that, although the ἀοιδός is famously listed as a δημιοεργός (ρ 383–385), no manufacturing metaphors are applied to the singer of epic by the medium of epic poetry—not just stitching, but also weaving, carpentry, and others are excluded. Diachronic skewing, which by itself would have privileged the relatively more traditional ‘singer’ and ‘singing,’ is further reinforced by the underlying rhetorical emphasis on the divine source and the divine authority of the song. It is telling that in two exceptional passages that apply verbs of manual labor and manufacturing to the composition and performance of song the gods and the Sirens are the respective subjects. [98] Where gods and daimones are involved, the singer feels free to highlight artistic accomplishment. [99] Otherwise, in the Homeric poems we meet with a discreetly succinct ἐπισταμένως (λ 368); with a δεδαώς to which θεῶν ἒξ is explicitly appended (ρ 518–519); or with oblique rhapsodic terminology like κόσμος [100] (but not κοσμέω). [101]
Nevertheless, that the metaphoric use of ῥάπτειν embraced both composition and performance is clear from the later use of ῥάβδος for poetic delivery, a noun (as we have seen) commonly linked to the verb ῥάπτειν by popular etymology. This is arguably the point of Pindar’s Isthmian 3/4 56–57, [Ὅμηρος] κατὰ ῥάβδον ἔφρασεν | θεσπεσίων ἐπέων, where φράζω refers to a telling that bears the authority of the ῥάβδος. [102] The same point is made by the fragment in the scholia 1d to Nemean 2 ascribed to Kallimakhos (cf. fr. 26 Pfeiffer): καὶ τὸν ἐπὶ ῥάβδῳ μῦθον ὑφαινόμενον | ἠνεκὲς ἀείδω δεδεγμένος. [103] It seems clear that in the Hellenistic period ῥάβδος was primarily connected with the rhapsode in performance—what role could the staff play in his private ruminations, away from the public limelight? Kallimakhos portrays his own composition under the guise of rhapsodic performance as ‘singing what is being woven with the authority of the ῥάβδος’.
The converse of the rhetorical emphasis on the divine origin of the epic song is a focus on the deceptiveness of self-interested telling. To this, Homeric diction readily applies craft metaphors. Eumaios provides a good example at ξ 131–132, when he asserts that even Odysseus would ‘fashion a false story’ (ἔπος παρατεκτήναιο) in exchange for a cloak or tunic to wear. [104] This accounts for the negative metaphoric use of ῥάπτειν for plotting and scheming. [105] Following Scheid and Svenbro (1996:111–130), Nagy (1996c:64n23) concedes that of metaphors of weaving and sewing for songmaking only traces survive—with focus on content rather than form—and that the attested Homeric diction is in the process of phasing out the application of such metaphors to performance. This concession may be premature, for there is an alternative interpretation of the evidence that is preferable and does not entail the view that in Homer artisanal metaphors for song-making are vestigial. [106] The rhetorical emphasis on the divine source and authority of the song, with the corresponding de-emphasis of the performer’s individual artistry, sufficiently accounts for the failure of the medium to recognize explicitly the currency of stitching or weaving as metaphors for the performance of epic. Fränkel (1925:3–4) had already anticipated Scheid and Svenbro’s argument. β 236, which he took as proof that ῥάπτω signifies invention, not performance, actually shows how design and execution—in metapoetic terms, composition and performance—can be viewed as two sides of one and the same happening: ‘But truly I do not at all grudge the proud suitors their doing violent deeds in the evil-sewing of their mind’. [107] The example from Herodotos 6.1 only proves that these are distinguishable elements that can, but need not, be chronologically coextensive or have one and the same subject. π 379 in turn does not preclude the performance of a scheme: it only acknowledges that schemes often fail in their execution. The sheer destruction that the suitors were sewing involved the careful keeping of watches, detailed at π 365–370, which were sedulously carried out ‘day by day’ (ἤματα π 365). Once again, ἐράπτομεν does not denote mere planning but the execution of the suitors’ machinations. In γ 118–119 Nestor summarizes the Greek campaign against Troy with the words, εἰνάετες γάρ σφιν κακὰ ῥάπτομεν ἀμφιέποντες | παντοίοισι δόλοισι, μόγις δ’ ἐτέλεσσε Κρονίων. The fulfillment of their war effort was long in coming, but this does not relegate the activity of κακὰ ῥάπτειν to arm-chair scheming. If there was any doubt, ἀμφιέποντες, ‘assailing them on every side’, makes clear that their fighting πόνος was comprehended in the statement, ‘for nine years we kept stitching evil against them with all manner of [martial] stratagems’. [108] The dative παντοίοισι δόλοισι makes explicit the ‘material’ out of which the final product, κακά, is stitched together. ‘Stitching’ does not fail to bear fruit at the planning stage but in execution. This hardly prevents its metaphoric application to songmaking that is composed in performance, for it is possible for a rhapsode to perform an utterance that, notionally, does not possess the performative authority to be effective (cf. the epic notion of ψεῦδος explored above, §8.2). [109]
Like ὑφαίνω, [110] ῥάπτω and its derivatives are also used in their literal sense in regard to: the manufacture of Sarpedon’s shield; [111] Laertes’ ‘patched tunic’ and ‘stitched greaves’; [112] a well-sewn (ἐϋρραφής) leather bag (β 354 = β 380); the seams (ῥαφαί) of the straps of Laertes’ old shield (χ 186); and winter clothing made of goat skins. [113] Penelope’s stratagem (δόλος β 93) to keep the suitors at bay—her promise to marry one of them only after she had completed a shroud (φᾶρος β 97) for Laertes’ burial, a web she would weave by day and unravel by night—provides a transitional instance of ὑφαίνω, both literal and metaphorical in that by weaving she was acting out her guileful scheming. Although the bard is never said to ‘weave’ his song, Penelope’s weaving here, like Helen’s at Γ 125–128, [114] may be read metapoetically to refer to the narrative in action. This accounts for the otherwise unexpected statement that she would not have her weaving perish in vain (β 98). Metapoetically, it restates in concrete narrative terms that Penelope’s epic κλέος will never perish. [115] The metaphoric uses of ῥάπτω are of a similar kind. [116] I have already dealt with γ 118 and π 379. [117] There are two other instances, of which one features an exchange between Zeus and Hera after Akhilleus has been roused to fight the Trojans: ‘You did it again, after all (ἔπρηξας καὶ ἔπειτα Σ 357) … you roused fast-footed Akhilleus’, he says; to which Hera replies, ‘How should I, who claim to be the best of goddesses … not have stitched evil for the Trojans in my anger?’ [118] The emphasis here is on accomplishment in deed (ἔπρηξας 357, τελέσσαι 362), and it is the perspective of a successful performance that must inform our reading of κακὰ ῥάψαι as combining inextricably both devising and doing. The last instance is π 421–423. Having learned from the herald Medon of ‘the destruction of her son in the halls’, [119] Penelope confronts Antinoos: ‘Why are you stitching death and doom for Telemakhos … ? It is impious to stitch evil against each other’. [120] The expression ‘destruction in the halls’ refers to the exchanges that took place during the assembly of suitors before the gates of Odysseus’ house (π 343–344). [121] These included the narrative of what they had unsuccessfully attempted (what Antinoos referred to at π 379 as φόνον αἰπὺν ἐράπτομεν οὐδ’ ἐκίχημεν), as well as what they were now intending to do (π 383–384). Hence, ὄλεθρος embraces the performance of past, and the planning of future, actions. That it cannot be strictly limited to inventio without actio is clear from the startling use of the present ἀποκτείνεις (π 432): ‘You are now devouring his household without repayment, you are wooing his wife, and you are murdering his son’ (π 431–432). As Hoekstra rightly notes, “ἀποκτείνεις [is] not so much conative as emphatic” (Heubeck and Hoekstra 1989:284). Penelope takes a comprehensive view of Antinoos’ doings and encompasses his deeds and counsels as one emphatic imperfective action: ‘you are killing Odysseus’ son!’ [122]
In sum, there is no great difference in the treatment Homeric diction gives to ῥάπτω and ὑφαίνω as metaphors of composition and performance. Both are used in their literal and metaphoric senses. Both are strongly associated with the construction and execution of plans and plots. Whether the corresponding guile is seen in a positive or a negative light only depends on the particular agent and on the narrative’s focalization. Since ὑφαίνω is unquestionably a metaphor for composition of Indo-European pedigree, whether ῥάπτω is a south-Aegean loanword or not, one must assume on the basis of this parity of treatment a comparable antiquity to its own metaphors for verbal (and, specifically, poetic) composition. An important finding from my survey above is that there are good rhetorical reasons why neither the Iliad nor the Odyssey applies artisanal metaphors to the bard’s singing, reasons that do not entail a disjunction between composition and performance. Very much to the contrary, my analysis demonstrates that, even in the limited sense in which ‘stitching’ is applied to individual authorship, one cannot and should not disjoin evil-scheming (composition) and evil-doing (performance). The application of ὑφαίνω to verbal artistry is explicit in Γ 212 (μύθους καὶ μήδεα πᾶσιν ὕφαινον); as it is to μῆτις in the sense of spoken counsels at Η 324 δ 678 739 etc., and to δόλος at Ζ 187 ε 356 etc. Weaving suggestively accompanied by singing takes place at ε 61–62 and κ 221 227 254. Although there are no instances of ῥάπτειν governing μῦθον vel sim., my analysis suggests that this is merely accidental and not the consequence of some inherent semantic property of the verb. Indeed, in δ 675–676 Penelope is said to have learned the μῦθοι that the suitors were ‘building deep in their hearts’ (βυσσοδόμευον δ 676, referring to their counsels against Telemakhos before Odysseus’ palace). But for the meter, nothing suggests that ἔρραπτον could not have stood for this word from the building trade. Such are indeed the counsels that Penelope will later describe as κακὰ ῥάπτειν.
There is another reason why Homeric poetry was likely to avoid the explicit application of ῥάπτειν to composition and performance. I observed above that the metaphor of stitching is specifically tied to the recomposition of traditional subjects in a traditional epic medium. And yet, from the internal point of view of the epic narrative, the recounted events are recent and their account is news to the hearers and the object of inquiry (e.g. κ 14–16). The best known expression of this perspective is α 351–352. Phemios sings of the woeful return of the Akhaians (α 325–327):
τοῖσι δ’ ἀοιδὸς ἄειδε περικλυτός, οἱ δὲ σιωπῇ
εἵατ’ ἀκούοντες· ὁ δ’ Ἀχαιῶν νόστον ἄειδε
λυγρόν, ὃν ἐκ Τροίης ἐπετείλατο Παλλὰς Ἀθήνη.

For them sang the famous singer, and they in silence
sat and listened; he sang of the return of the Akhaians,
the woeful return from Troy which Pallas Athena laid upon them.
Penelope hears the song and bursts into tears (α 337–341):
          Φήμιε, πολλὰ γὰρ ἄλλα βροτῶν θελκτήρια οἶδας
          ἔργ’ ἀνδρῶν τε θεῶν τε, τά τε κλείουσιν ἀοιδοί·
          τῶν ἕν γέ σφιν ἄειδε παρήμενος, οἱ δὲ σιωπῇ
340    οἶνον πινόντων· ταύτης δ’ ἀποπαύε’ ἀοιδῆς
          λυγρῆς …

          Phemios, since you know many other subjects spellbinding to mortals,
          deeds of men and gods that singers celebrate in song,
          sing for them one of these as you sit here, and let them in silence
340    drink their wine; but cease from this woeful song …
To this Telemakhos replies (α 346–352):
          μῆτερ ἐμή, τί τ’ ἄρα φθονέεις ἐρίηρον ἀοιδὸν
          τέρπειν ὅππῃ οἱ νόος ὄρνυται; οὔ νύ τ’ ἀοιδοὶ
          αἴτιοι, ἀλλά ποθι Ζεὺς αἴτιος, ὅς τε δίδωσιν
          ἀνδράσιν ἀλφηστῇσιν ὅπως ἐθέλῃσιν ἑκάστῳ.
350    τούτῳ δ’ οὐ νέμεσις Δαναῶν κακὸν οἶτον ἀείδειν·
          τὴν γὰρ ἀοιδὴν μᾶλλον ἐπικλείουσ’ ἄνθρωποι,
          ἥ τις ἀκουόντεσσι νεωτάτη ἀμφιπέληται.

          My mother, why do you begrudge the trusty singer
          to delight as his mind is stirred? It is not the singers
          that are to blame but Zeus, perhaps, who gives
          to men who eat grain, to each one as he will.
350    There is no cause for anger that this man here sings the evil doom of the Danaans;
          for men applaud more whichever song
          comes the newest to their ears.
The search for news is the motivation for Telemakhos’ journey, first to Nestor and then to Menelaos (α 281–286). It would have been rhetorically ill-advised for the bard to portray Phemios as song-stitching in performance a well-known and well-loved story. [123] Homeric poetry, then, adopts the rhetorical conceit, made effectual by the authoritative mimesis of performance, that it narrates history in progress. Not only are Odysseus’ woes the type of novel song that people love; Helen, too, in Ζ 357–358 adopts the perspective of present actors who, in a moment of quiet reflection, ponder how they will be regarded by future generations. [124] This conceit rules out an explicit self-reference to the performance of Homeric poetry as traditional in theme and diction. There are hints here and there that, in point of fact, it was traditional: Demodokos in the far-flung court of Alkinoos (ζ 8 204–205) has heard of the Trojan War (θ 489–491), and so have the Sirens (μ 189–190). [125] The metapoetic appreciation that marks the reception of traditional poetry allows us to take even Helen’s words in Ζ 357–358, and many a reference to κλέος, as affirmations of traditionality (González 2015). But the performance rhetoric of the profoundly traditional Homeric stories is that they are ever new in the telling.

10.2.3 Rhapsodic sequencing and relay poetics Kallimakhos and the rhapsodes
The fragment in the scholia 1d to Pindar Nemean 2, ascribed to Kallimakhos, opens a window into fundamental features of the rhapsode’s craft. It nicely combines weaving (ὑφαίνω), rhapsody (ῥάβδος), singing (ἀείδειν), and the sequencing of ‘relay poetics’ (δεδεγμένος; cf. Ι 191): καὶ τὸν ἐπὶ ῥάβδῳ μῦθον ὑφαινόμενον | ἠνεκὲς ἀείδω δεδεγμένος. [126] These lines were long ago recognized by Wilamowitz-Moellendorff 1911 as verses 5 and 8 of P.Ryl. 1.13 (MP3 203; cf. fr. 26 Pfeiffer = fr. 30 Massimilla). Before attempting a translation, one must decide how the scholiast’s suppression of the two intervening lines affected their meaning. Although certainty is impossible, one may reasonably assume that the excerpt preserves the substance of Kallimakhos’ point. Indeed, apparently taking πλαγκτύν closely with μῦθον, Wilamowitz offered the following translation: “[D]ie Geschichte, welche zum Stabe gedichtet die Leute der Vorzeit einzeln überliefert haben, erzähle ich im Zusammenhang ihnen nach” (Wilamowitz-Moellendorff 1911:472). But πλαγκτύς, ‘wandering’, perhaps referred not to ‘scattered’ transmission but to a transmitted meaning that had wandered away from the authoritative tradition, from the telling sanctioned by the rhapsodic staff. By receiving the story and singing it in continuity with the authentic tradition, Kallimakhos laid claim to its authority. Thus, on the whole I am in agreement with D’Alessio (1996:409n92): “Callimaco rielabora il canto che ha ricevuto dalla tradizione precedente (v.5), che doveva presentarsi discorde (v.7, secondo l’interpretazione di Wilamowitz). Con il termine ἠνεκές il poeta sottolinea in modo esplicito la specifica tecnica di connessione del materiale narrativo … , che reinterpreta il criterio di continuità contenutistica implicito nel διηνεκές di fr. 1.3.” Only, I understand ἠνεκές, ‘without a break’, not primarily as referring to a continuity of narrative technique but to the unbroken paradosis of the authentic tradition. [127] Hence, I offer the following translation of the excerpt: ‘And taking up the story upon the staff woven in uninterrupted sequence I sing’ (construing ἠνεκές apo koinou). [128]
Together, ἠνεκές and δεδεγμένος express the notion of performative sequence that is at the heart of rhapsodic poetics. As a performer of traditional poetry, there are three complementary ways in which the rhapsode might be said to sing in uninterrupted sequence. First, he situates his song with studied precision along the οἴμη, i.e. the traditional thematic sequence of the narrative plot. Pointers like τῶν ἀμόθεν (α 10) or ἔνθεν (θ 500) serve to locate the starting point of his performance. Second, he stands in a long succession of traditional singers from whom he has received the song and to whom he, in turn, bequeaths it. This succession embodies the authority of the paradosis and gives his performance a diachronic depth which the chain of inspiration in Plato’s Iōn, with the Muse as the fountainhead, celebrates. [129] And third, in the competitive setting rhapsodes take turns singing their song. Such performance in relays became central to the competitive poetics of the Panathenaia, one rhapsode taking up the song right where the preceding one had left off. I call this competitive mode relay or hypoleptic singing. In Kallimakhos’ fragment, the participles δεδεγμένος and ὑφαινόμενον suggest that the last acceptation of sequence is primarily in view and serves as a dramatic synchronic reenactment of the second: the singer himself is the one who stands in continuity with previous performers as he takes up a song in progress. (δι)ηνεκές portrays his performance as an authentic actualization of the rhapsodic tradition. [130] Its point is not that the song never ceases but to validate the performer as a legitimate link in the long chain that ties together past and present through the imperishable memory of κλέος. The sequence remains unbroken so long as the narrative is reperformed yearly within the cycle of the festival calendar. This is the continuity of repetition, re-actualization, and reperformance, a continuity of like elements (here, discrete performances) arranged sequentially to make a whole. [131] Taking his place within the authentic paradosis, Kallimakhos lays claim to its authority.
The semantics of (δι)ηνεκές can also capture the thematic continuity and original comprehensiveness of the epic cycle. [132] It is in this light that we must read the famous opening to Kallimakhos’ Aitia, with its ἓν ἄεισμα διηνεκές (fr. 1.3). Whatever its immediate application to his own poetry, these words make rhapsodic poetics the indispensable reference for the Alexandrian’s poetic program (ἄεισμα denoting the product of the song as readily as the manner of its singing). [133]
Rhapsodic poetics regards each new performance as a segment of the notional thematic whole, whatever its particular beginning or end. Its Panhellenic tendency exhibits a marked propensity to embrace within its scope all that there is to say about its subjects. [134] This tendency toward expansion is signaled by the use Homeric poetry makes of διηνεκές and διηνεκέως to qualify the manner of telling. In η 241 Odysseus tells Arete that it would be grievous to narrate his woes fully (διηνεκέως ἀγορεῦσαι), i.e. one after another, continuously, without leaving anything out. It is well known that Odysseus is later compared to an ἀοιδός who possesses ‘shapeliness of utterance’ (μορφὴ ἐπέων λ 367), [135] a judgment that encourages us to consider his narrative representative of good rhapsodic practice. The length of Odysseus’ tale, [136] which doubtless serves to emphasize his long-suffering, also reflects the tendency towards comprehensiveness characteristic of rhapsodic poetics. By way of contrast, the negative use of διηνεκέως at μ 56 and δ 836 shows that, irrespective of the specific point from which the rhapsode takes up his telling, he always keeps the end notionally in view. [137] δέγμενος/δεδεγμένος
As I mentioned above, a second characteristic element of rhapsodic relay poetics underlined by Kallimakhos’ fragment is that of succession: one performer would take up his singing right where a previous one had left off. [138] From the point of view of the paradosis, one might restate this sequential singing as one performance taking up the song where a previous one had left off. The verb for ‘taking up the baton’ is δέχομαι. It involves waiting for the previous rhapsode to finish his singing and taking up the thread of the song at the very thematic juncture that marks the transfer. Such joints, syntactic and thematic, are also in play during solo epic performance as the singer strives to join one verse to another, especially in the technically more taxing cases of enjambment. This passing of the baton, expressive of the inherently agonistic nature of Greek performance, could be gesturally enacted by handing on a staff to the next singer. Each symposiast was handed the myrtle-branch when it was his turn to perform the skolia. [139] Rhapsodic serial performance fostered, and was in turn fostered by, a ‘poetics of seam-stitching’ already inherent in the genre of hexameter epic, with its long arrays of hexameter στίχοι. This mode of delivery is the focus of the various accounts of the so-called Panathenaic rule, which I review below (§ First, I wish to illustrate the use of δέχομαι to denote relay performance. A clear example in the context of the symposium is the passage in Aristophanes’ Wasps that starts with verse 1216. After listing the names of his fellow symposiasts, Bdelykleon tells Philokleon: ‘In the company of these men see to it that you take up (δέξει) the skolia well’ (1222). [140] Bdelykleon adds: ‘Suppose I am Kleon and I start singing the Harmodios song, and you are to take it up’ (δέξει δὲ σύ 1225). And at 1243–1244: ‘After him, Aiskhines, the son of Sellos, a learned man skilled in the art of the Muses, will take it up (δέξεται) and he will sing’. Eupolis’ fr. 395 K-A provides another instance from a symposiastic context: ‘Having taken it up (δεξάμενος) to the right, Sokrates stole the ladling-cup while he sang something of Stesikhoros to the lyre’. [141]
To move on to epic, one should mark the role of the σκῆπτρον in public assembly as a functional parallel to the rhapsodic staff in relay singing. [142] The σκῆπτρον identified the speaker that had the floor, attesting not only to his right to speak but also to the authority of his utterance. [143] Hence the remarkable use Akhilleus makes of it during his oath (Α 233–239 245–246). [144] Three passages illustrate the use of δέχομαι to denote relay singing. [145] The first is the fragment by Kallimakhos that introduced us to this notion. According to regular Homeric use, the perfect participle δεδεγμένος, besides the semantically predictable ‘having received’, and therefore ‘taking up’, somewhat surprisingly, simply means ‘awaiting’ or ‘expecting’. [146] Therefore, to ears attuned to the Homeric usage, besides positioning Kallimakhos rhetorically as the true heir of the authentic, i.e. the rhapsodic, tradition, there is in δεδεγμένος a distinct echo of our second passage, Ι 191, where Patroklos waits for the end of Akhilleus’ performance. In effect, there is a hint of Kallimakhos’ waiting to take up the song of the authentic tradition, a participant in hypoleptic delivery, as if that tradition had never vanished from the sphere of performance which is the life-blood of traditional poetry. The Alexandrian’s reception of earlier poetry is not presented as the fruit of library research. It is as tight and immediate as the passing of the baton from rhapsode to rhapsode in relay singing. [147]
Ι 191, my second illustration, depicts Akhilleus whiling away his time in epic performance. [148] This by itself is remarkable enough, but even more extraordinary is the manner of his performance. Here the hero engages in what amounts, on the lips of the performing rhapsode, to a magnificently self-referential metapoetic representation of hypoleptic rhapsodizing (Ι 186–191):
          τὸν δ’ εὗρον φρένα τερπόμενον φόρμιγγι λιγείῃ
          καλῇ δαιδαλέῃ, ἐπὶ δ’ ἀργύρεον ζυγὸν ἦεν,
          τὴν ἄρετ’ ἐξ ἐνάρων πόλιν Ἠετίωνος ὀλέσσας·
          τῇ ὅ γε θυμὸν ἔτερπεν, ἄειδε δ’ ἄρα κλέα ἀνδρῶν.
190    Πάτροκλος δέ οἱ οἶος ἐναντίος ἧστο σιωπῇ,
          δέγμενος [149] Αἰακίδην ὁπότε λήξειεν ἀείδων.

          Him they found, delighting his mind with the clear-sounding phorminx,
          beautiful, ornamented, and upon it was a silver bridge;
          this he won from the spoils after he destroyed the city of Eëtion.
          He was delighting his heart with it and was singing the glorious deeds of warriors,
190    and Patroklos alone sat facing him, in silence,
          waiting for when the son of Aiakos would leave off singing.
This description, which inspired the choice of cover for this book, raises the question of what Patroklos was waiting for. Was he simply waiting for Akhilleus to end his song so that, leaving the phorminx behind, both should move on to other things? Or is he singled out as ‘alone opposite him’ (οἱ οἶος ἐναντίος), waiting in silence, because he intended next to take the phorminx and continue the singing? The latter, I believe, is the point of the passage. The κιθαρῳδός (and κιθαριστής—it is usually impossible to distinguish them in the artistic record) could perform while sitting. [150] Patroklos is forcefully marked out as alone on a level with his leader, his attention fully engaged in his performance while he waits for him to end. As the scholia ad loc. long ago noted, ‘[Patroklos] is not alone in the hut, but he alone sat waiting for the son of Aiakos opposite him; for also Automedon was in the hut’ (190a Erbse). [151] It is no coincidence that this pair, the thematic heart of the poem in course, should have been selected to enact an inset of hypoleptic performance. The word ἐναντίος, which can be read as a simple marker of position, inevitably retains a measure of the antagonistic connotations that make it common in the context of hostile engagement. Its deliberate choice lends Patroklos’ waiting the competitive undertone that is the hallmark of even the most cooperative kind of relay performance. [152] An audience sensitive to Patroklos’ role in the poem and attuned to the hypoleptic performance practiced by competing rhapsodes would have no trouble understanding why he alone is singled out and why he is said to wait specifically for when Akhilleus would leave off singing. The audience would readily detect the metapoetic meaning. Without this motivation, the narrative’s focus on the waiting remains puzzling. [153] It is doubtless true, as some have suggested, that we may see in Patroklos the narrative equivalent of the rhapsodic audience. [154] This much can be derived from Ι 190. But there is no reason why the text could not have made clear that he was merely listening to the singing: why the otherwise odd focus on ‘waiting for when Akhilleus would leave off singing’? [155] This suggests that, the very moment Akhilleus should leave off singing, Patroklos was going to do something specific, the identity of which would have been readily inferable by the audience. This, I submit, is to take up the lyre and continue the song right where Akhilleus had left off. It never comes to pass, of course, since the embassy arrives to interrupt the relay performance. [156] Perceau (2005:71) objects that “in epic only solos are mentioned, never relay singing by several singers.” [157] She is right, of course, if she means singing by characters identified as professional ἀοιδοί. But this is too simplistic an approach to the intricate metapoetic texture and mimetic pragmatics of Homeric epic. I need only cite here the remarks by Tarditi (1968:140) concerning Andromakhe’s, Hekabe’s, and Helen’s laments for Hektor at Ω 720–776: “Ciascuna delle tre donne esprime il suo dolore e rievoca un aspetto dell’umanità o dell’eroismo di Ettore, ciascuna ‘cuce’ il suo canto in quella, potremmo chiamarla, rapsodia trenodica che viene cosí composta.” [158]
The last example comes from the Hymn to Hermes 476–477 and complements Ι 186–191, for it presents Hermes and Apollo as archetypes for hypoleptic performers. Hermes had struck up his song (γηρύετ’ ἀμβολάδην 426, in the manner of a hymnic prelude) and ‘ratified the gods’ (κραίνων 427), [159] starting with Gaia (427) and detailing their individual shares (μοῖραν 428): ‘First among the gods he honored Mnēmosynē in song, the mother of the Muses, for she received the son of Maia as her lot; and the immortal gods the splendid son of Zeus honored according to their seniority and how each was born, relating everything in good order, playing the lyre under his arm’ (429–433). As often in the Homeric hymns, this amounts to a transparent self-referential performance that overlaps thematically with the Hymn to Hermes and fuses with the rhapsodic delivery of this hymn currently in progress before the audience. At the marvelous sound, Apollo desires to learn the τέχνη of rhapsodic hymnic performance (song and instrumental playing, possibly dancing too). [160] Then comes the enactment of hypolepsis (475–477):
ἀλλ’ ἐπεὶ οὖν τοι θυμὸς ἐπιθύει κιθαρίζειν,
μέλπεο καὶ κιθάριζε καὶ ἀγλαΐας ἀλέγυνε
δέγμενος ἐξ ἐμέθεν· …

But since your heart is crazy about playing the lyre,
sing and play it and heed splendid festivity
taking it [161] up from me; …
The key word is δέγμενος, and in context its object must be all that Hermes’ τέχνη encompasses—preeminently the lyre, the focus of Apollo’s wonder, but also the ‘divine song’ (442) that the lyre accompanies and perhaps even dancing. Apollo is portrayed as permanently entering upon the art of Hermes (464–465), whose succession is enacted by an archetypal hypoleptic transfer in the course of a rhapsodic hymnic performance. Adopting the manner of a commercial exchange of divine τιμαί, the ceded privileges become the permanent property of the god who receives them. [162] Cooperation and contest
The relay poetics of hypoleptic performance comes in two flavors: one is predominantly cooperative, the other oppositive. But even the cooperative kind retains the antagonistic quality of competitive rivalry. This distinction is easily illustrated with reference to the singing of skolia at the symposium. Mure (1854–1859:3.101) writes that each guest whose turn to perform had come was “expected at once to carry on the strain, whether in the way of continuation or repartee.” When the connection between one skolion and the next was “supplied by innuendos or ambiguous allusions … to the character or circumstances of the individual performers … the principle of contrast would often be preferable, in point of effect, to that of conformity” (3.102–103). The cooperative character of relay poetics is at the heart of the Panhellenic process that drove the textual fixation of the Homeric poems. This cooperation was both diachronic—effective between successive rhapsodic performances, the later ones responding to the earlier—and synchronic—one rhapsode at a competitive event artistically stitching his performance in deliberate sequence to that of the previous rhapsode. Traditional diction, including formulaic language, is doubtless the instrument and product of this cooperation, but not the sole one. The traditional thematic sequence too guided the singing as a path [163] marked out by earlier singers and worn further by the ongoing performance. [164] This cooperative poetics is also in view in the scholion 1d to Pindar’s Nemean 2:
οἱ δέ φασι τῆς Ὁμήρου ποιήσεως μὴ ὑφ’ ἓν συνηγμένης, σποράδην δὲ ἄλλως καὶ κατὰ μέρη διῃρημένης, ὁπότε ῥαψῳδοῖεν αὐτὴν, εἱρμῷ τινι καὶ ῥαφῇ παραπλήσιον ποιεῖν, εἰς ἓν αὐτὴν ἄγοντας. οὕτω καὶ ὁ Πίνδαρος ἐκδέδεκται. οἱ δὲ, ὅτι κατὰ μέρος πρότερον τῆς ποιήσεως διαδεδομένης τῶν ἀγωνιστῶν ἕκαστος ὅ τι βούλοιτο μέρος ᾖδε, τοῦ δὲ ἄθλου τοῖς νικῶσιν ἀρνὸς ἀποδεδειγμένου προσαγορευθῆναι τότε μὲν ἀρνῳδούς, αὖθις δὲ ἑκατέρας τῆς ποιήσεως εἰσενεχθείσης τοὺς ἀγωνιστὰς οἷον ἀκουμένους πρὸς ἄλληλα τὰ μέρη καὶ τὴν σύμπασαν ποίησιν ἐπιόντας, ῥαψῳδοὺς προσαγορευθῆναι.
Some say that, when the poetry of Homer had not been collected as a single whole and was otherwise scattered and divided into sections, whenever they performed it rhapsodically, they would make something approaching a [coherent plot] sequence [165] or [narrative] stitch-work by bringing this poetry together into one. [166] This is also Pindar’s understanding. Others say that, since in former times [Homer’s] poetry had been handed down by sections and each of the contestants would sing whichever section he wanted to, a lamb being the designated prize for the victors, the contestants were then called arnōidoi; but in turn [they say that], after both poems were introduced, the contestants, ‘mending,’ so to speak, the parts one to another and going through [167] the entire poetry, were in turn called rhapsodes. [168]
Of especial significance here is the claim that, by singing Homeric poetry rhapsodically, performers refashioned it into a certain thematic sequence (εἱρμός τις) or stitch-work (ῥαφή). This reshaping is imagined as a cooperative endeavor that results in the artistic unity exhibited by the poems. The scholiast further underlines in the alternative account the collaborative nature of hypoleptic performance by his use of ἀκεῖσθαι, ‘to heal’ or ‘cure’, as a metaphor for ‘mending’ and ‘repairing’. Whereas formerly the poetry had been transmitted, and was correspondingly performed, piecemeal, with an overriding competitive focus on individual prize-winning and a consequent ‘willful’ injury to the notional integrity of the poems, once the outward circumstances of performance—the introduction of the poems—enforced the ‘correct’ sequence, the stitching of the contestants was reconceived as reparative, i.e. as the integrative recomposition of the erstwhile scattered members. Hence, the performers were no longer called arnōidoi, after the prize that was emblematic of their rivalry, but ῥαψῳδοί, after the corrective stitching enforced by the mandatory adherence to the notional canonical sequence. However anachronistic, this elucidation of the terms ἀρνῳδός and ῥαψῳδός (on which see Durbec 2003) illuminates the competitive and cooperative dimensions of rhapsodic performance. In particular, it tellingly acknowledges that, even when such performance was unconstrained and the rhapsode sang whichever parts he wanted, still the practice of his craft effected a narrative whole that exhibited a particular thematic sequence and resembled a kind of stitch-work. In other words, the aesthetics of rhapsodic performance may have moved its emphasis from competition to cooperation, but thematic stitching was, and remained, at the heart of the rhapsode’s craft.
Competitive performance in the context of the Greek ἀγών is dramatically illustrated by the Certamen of Homer and Hesiod. [169] The primary source for the contest has been persuasively identified as the Mouseion of Alkidamas, [170] the fourth-century sophist who was featured above (§9.5) as a signal witness to the increasing dependence of rhapsodes on scripts for their public delivery. Here, I need only recall that Alkidamas made the rhapsode and his craft a chief focus of analysis as he sought to illustrate the importance of improvisation (αὐτοσχεδιασμός) to accomplished oratorical delivery. [171] This suggests that he patterned his imagined ἀγών between Homer and Hesiod after actual rhapsodic dynamics at the Athenian Panathenaia. The Certamen features three kinds of competitive challenges: [172] first, as befits two renowned sages, a test of wisdom (75–101, 140–175); [173] second, what interests me now, a series of ‘capping’ challenges in which Hesiod’s deliberately perverse openings try to hinder Homer from successfully taking up the song from his adversary (107–137); [174] third, a solo reperformance by each contestant of a favorite passage from his own poems (180–204). In the second test, the artistry of stitching is dramatically on display as Hesiod does his best to prevent Homer from sewing a sensible seam. Here enjambment, the most challenging technique of hexameter stitching, takes center stage. This is not to suggest that rhapsodes at the Panathenaia must have exchanged one-liners in the manner of the Certamen, but that Alkidamas extracted the essence of their craft and displayed the overriding technical point at stake when they competed before the public. In all probability, not only would a competing rhapsode have to prove his skill at recomposing Homeric epic, thus demonstrating the ordinary stitching competence entailed by long stretches of solo performance; but the antagonistic challenge to pick up the delivery right where the previous rhapsode had left off would have provided a dramatic opportunity for the first rhapsode to devise a finish that would test to the utmost the ability of the succeeding one to stitch a felicitous seam.
Let us see how Hesiod in the Certamen stages the hypoleptic test of Homer’s skill. Before he turns to the heart of the challenge, Hesiod provides what is, in effect, an anti-invocation of the Muse (97–101): [175]
          Μοῦσ’ ἄγε μοι τά τ’ ἐόντα τά τ’ ἐσσόμενα πρό τ’ ἐόντα
          τῶν μὲν μηδὲν ἄειδε, σὺ δ’ ἄλλης μνῆσαι ἀοιδῆς.
ὁ δὲ Ὅμηρος βουλόμενος ἀκολούθως τὸ ἄπορον λῦσαι φησίν·
          οὐδέ ποτ’ ἀμφὶ Διὸς τύμβῳ καναχήποδες ἵπποι
          ἅρματα συντρίψουσιν ἐρίζοντες περὶ νίκης.

          Come now, Muse, to me what is, shall be, and was before—
          of these, sing nothing; but you [Homer?] give heed to the rest of the song.
Homer, wishing to solve the conundrum in a manner that was consistent [with Hesiod’s invocation], said:
          Never shall horses with sounding hoofs around the tomb of Zeus
          shatter chariots as they vie for victory.
Note the explicit focus on sequence: Homer wishes his solution to ‘follow’ (ἀκολούθως 99) Hesiod’s challenge. [176] Note also, incidentally, conclusive proof that ἄλλης in the common hymnic transition καὶ ἄλλης μνήσομ’ ἀοιδῆς (e.g. Hymn to Demeter 495) does not mean ‘another song’ but ‘the rest of the song’: [177] Hesiod is not challenging Homer to think of another song, but to take up the song where he left off and build the following lines upon his anti-invocation.
Seeing that Homer has met the challenge well, Hesiod turns to the ambivalent propositions. I will only quote the first as a representative illustration of the nature of Hesiod’s challenges: [178]
⟨Ἡ.⟩ δεῖπνον ἔπειθ’ εἵλοντο βοῶν κρέα καὐχένας ἵππων
⟨Ὁ.⟩ ἔκλυον ἱδρώοντας, ἐπεὶ πολέμοιο κορέσθην.

Hesiod: Then they dined on beef and the horses’ necks
Homer: sweating, they unyoked after they had their fill of fighting.
We are told that ‘speaking many lines Hesiod required Homer to answer each one in turn congruously’ (καθ’ ἕνα ἕκαστον συμφώνως ἀποκρίνασθαι 104). The verb ἀποκρίνασθαι makes clear that Homer’s line is a continuation of Hesiod’s; the adverb συμφώνως, that the sequence should be meaningful. Hesiod’s challenge hinges on the apparent syntactic parallelism between ‘horses’ necks’ and ‘beef’. But Homer uses enjambment to carry the day: ‘necks’ is now taken by ‘they unyoked’, and the added participle ‘sweating’, which qualifies ‘necks’, serves to strengthen the sequence he has forged. The Panathenaic rule
The drama of competitive hypoleptic performance must have thrilled the audience with its successes and failures. This explains the so-called Panathenaic succession rules, variously reported, among others, by [Plato’s] Hipparkhos and Diogenes Laertios. [179] First, the report in the Hipparkhos 228b5–c1:
Ἱππάρχῳ … , ὃς ἄλλα τε πολλὰ καὶ καλὰ ἔργα σοφίας ἀπεδείξατο, καὶ τὰ Ὁμήρου ἔπη πρῶτος ἐκόμισεν εἰς τὴν γῆν ταυτηνί, καὶ ἠνάγκασε τοὺς ῥαψῳδοὺς Παναθηναίοις ἐξ ὑπολήψεως ἐφεξῆς αὐτὰ διιέναι, ὥσπερ νῦν ἔτι οἵδε ποιοῦσιν.
[I mean] Hipparkhos … who displayed many fine accomplishments of learning, above all, he was first to bring the poetry of Homer to this land here and forced the rhapsodes at the Panathenaia to go through it in sequence by relay, just as they still do now.
The corresponding account in Diogenes Laertios 1.57 runs as follows:
τά τε Ὁμήρου ἐξ ὑποβολῆς γέγραφε ῥαψῳδεῖσθαι, οἷον ὅπου ὁ πρῶτος ἔληξεν, ἐκεῖθεν ἄρχεσθαι τὸν ἐχόμενον.
He has also enacted a law that rhapsodes are to perform the poetry of Homer reactively, that is to say, that the following is to start from that point where the first left off.
In the Hipparkhos, the key terms are ‘by relay’ (ἐξ ὑπολήψεως) and ‘in sequence’ (ἐφεξῆς). The former is unproblematic: each rhapsode takes up (ὑπολαμβάνω) where the former left off. [180] This expression corresponds closely to ἐκ διαδοχῆς, which the scholia D to Α 604 uses to gloss ἀμειβόμεναι: ἐξ ἀμοιβῆς καὶ διαδοχῆς ᾄδουσαι. Like ἐξ ὑπολήψεως, ἐκ διαδοχῆς could be understood not only synchronically but also diachronically: ‘Of old they called Homeridai those in the line of Homer who also sang his poetry in [genealogical] succession’ (i.e. as his successors). [181] What ‘in sequence’ (ἐφεξῆς) refers to is not so clear: either to the rhapsodes (one rhapsode after another) or to the subject matter (one episode after another). Collins (2004:193n4) thinks that it refers to “the sequence of performance by rhapsodes, i.e. one after another, rather than to the sequence of poetic material.” [182] But unless one can plausibly suggest alternative modes of singing that the specification ‘in sequence’ is supposed to rule out, what could out-of-sequence rhapsodes be? Is the point simply to rule out the unpredictable sequence that obtains for the singing of skolia at a symposium? Collins’s view depends on the notion that rhapsodes were enrolled to compete in a particular order agreed to in advance, from which they were not to depart once before the public. That performers participated in some pre-arranged order seems beyond question, but that there was a felt need to rule out a departure from this order seems improbable. [183] Far more likely, especially in view of the scholia to Nemean 2 reviewed above (§, both ἐφεξῆς and ἐξ ὑπολήψεως regard the sequencing of the subject matter: the former requires that episodes follow one another in agreement with an established traditional order; [184] the latter, that there be no thematic gap between two consecutive performances. The Panathenaic rule ascribes what must have been the late-classical and Hellenistic practice to an archaic πρῶτος εὑρετής, either Hipparkhos or Solon. Whether such a rule ever existed and was enforced is beside the point. What matters is that the thematic sequencing fundamental to rhapsodic practice and responsible for the artistic unity of the Homeric poems is re-imagined as the outcome of performance standards that force rhapsodes to recompose (in the sense of ‘ordering as an organic whole’) an otherwise scattered and disorderly body of Homeric material.
The expression in Diogenes Laertios that I have translated ‘reactively’, ἐξ ὑποβολῆς, raises difficult interpretive problems. [185] Nevertheless, there is general agreement that it is interchangeable in meaning with ἐξ ὑπολήψεως, however one justifies this conclusion. [186] Diogenes’ text is clear on this point, for it appends an explication preceded by the word ‘namely’ (οἷον). [187] The best analysis of the expression is in Pagliaro (1951:33–37), with whom Nagy (2002:19–21) overlaps. ὑποβάλλω in Τ 80 is given the Homeric acceptation ‘to interrupt’, and the derivative ὑποβλήδην in Α 292 is accordingly translated ‘interruptingly’. But, as Pagliaro (1951:35–36) astutely observes, strictly speaking, Akhilleus does not interrupt Agamemnon at Α 292 “perchè, difatti, Agamennone ha potuto terminare di parlare, dicendo tutto quello che aveva da dire.” Only if one extends the ordinary meaning of ‘to interrupt’ so as to embrace a heated follow-up to an otherwise semantically complete statement can one call Akhilleus’ reply ‘an interruption.’ Such a stylized ‘interruption’ would fit just as well under the interpretation championed by Pagliaro, which I will now develop with minor modifications. [188]
Pagliaro emphasizes that one cannot understand ὑποβάλλω at Τ 80 and ὑποβλήδην at Α 292 apart from the use of ἀμβλήδην at Χ 476, ἀμβολάδην in the Hymn to Hermes 426, and ἀναβάλλομαι at α 155 θ 266 ρ 262. In all probability, ἀναβάλλομαι with the infinitive ἀείδειν (and a fortiori its absolute use) derives from its use with direct objects like ‘tune’, ‘voice’, ‘song’ vel sim. (cf. μέλος ἀμβάλευ in Theokritos 10.22). Instances of it suggest that it was a technical expression from the domain of performance with currency among instrumentalists and singers, including rhapsodes. [189] Because ‘striking up the song’ was a terme d’art for the prelude, in time (how soon we cannot guess) it came to be used absolutely for ‘to start a performance’; and, from this, it was only a small distance to the complementary infinitive ἀναβάλλομαι ἀείδειν, ‘to start singing’. [190] Because performers often employed the verb as the semantic equivalent of ‘to begin’ when the action that was beginning concerned singing or playing an instrument, the adverbial derivatives ἀμβολάδην and ἀμβλήδην were also taken with a verb of performance (e.g. sing or play) to mean ‘to start singing or playing’. It is often possible to read a given instance with both the broader and the stricter senses in view. [191] This is the case at Χ 476: ‘But when Andromakhe revived and her spirit gathered into her breast, she started wailing (ἀμβλήδην γοόωσα) and spoke among the Trojan women’. I have translated the verse as if ἀμβλήδην γοόωσα stood for ἀνεβάλλετο γοήμεναι; but it is also possible to read it as ‘lifting up [her voice] in wailing’, and even as a not-so-subtle metapoetic reference to the overture of a song of lamentation. Similarly, one could render κιθαρίζων | γηρύετ’ ἀμβολάδην in the Hymn to Hermes 425–426 simply as ‘he started playing the lyre and singing [to it]’; [192] or, in this case, given the emphasis on technical virtuosity, more likely, ‘playing the lyre he sang (in the manner of) a prelude’. [193]
To come now to Diogenes Laertios 1.57: here the sense of ὑποβολή can only be understood against the corresponding ἀναβολή, ἀναβάλλομαι, and ἀμβλήδην. It is very hard to stumble upon the correct meaning from an etymological analysis that equates ὑποβάλλω with βάλλειν τι ὑπό τινα or βάλλειν ἑαυτὸν ὑπό τι, i.e. ‘to throw down [before someone]’ or ‘to throw oneself down under’. The former gloss metaphorically amounts to an interruption if one envisions disrupting a speaker by throwing words in his path; but, as I have argued following Pagliaro, no interruption is in view either at Α 292 or at Τ 80. [194] The latter gloss follows Della Seta (1910:335), who contended that, etymologically, the operative notion is “un movimento da farsi da una seconda persona per andare a prendere il peso del canto abbandonato dalla prima.” But this rationale seems far-fetched. The meaning becomes clear, however, if we assume that the coinage of the idiom was technical in origin and took place among professional performers. Indeed, one need only suppose that it took the form of a semantic proportion: just as ἀναβάλλομαι is ‘to cast up one’s melody or voice as one commences the song’, so ὑποβάλλομαι would mean ‘to cast one’s voice in reaction to’ words just uttered. [195] Or, reducing the former to its looser acceptation, ‘to start singing’, one might render the latter as ‘to resume singing’, with a strong emphasis on the fact that the continuing ‘song’ (i.e. what is performed next) picks up on, and reacts to, the words just uttered. [196] This is the reason for my translation of ἐξ ὑποβολῆς … ῥαψῳδεῖσθαι as ‘rhapsodes are to perform reactively’. That ὑπό may, and often does, carry this sense was conclusively demonstrated by Else (1959:85–100): “It appears, then, that one important function of ὑπό in early Greek … was to express psychological nearness or exposure, the experience of being within immediate range of a person or object … that was felt as having power to menace or protect. It would follow that a verb ‘compounded with ὑπό’ … will tend to denote a reaction or response to such a situation, if the meaning of the simple verb lends itself to such an idea” (his emphasis). [197] On my theory, archaic and classical audiences should not have found the meaning of ὑββάλλειν or ὑποβλήδην any less transparent than that of ἀναβάλλομαι or ἀμβλήδην. All were similarly technical in origin, all from the ambit of the rhapsode and his craft.
The tight consequence of thought entailed by reactive performance readily shades into immediacy of thematic or temporal sequence. [198] This explains why ἐξ ὑποβολῆς could be used as a synonym of ἐξ ὑπολήψεως and grasped unproblematically as a reference to hypoleptic performance. [199] It is not difficult to understand how the original rhapsodic meaning of ὑποβάλλω could give rise later to the dramatic ὑποβολεύς, ‘prompter’. This represents a technical development that, as I chronicle in this book, happened repeatedly: the adoption into the sphere of dramatic performance of an element of the rhapsodic craft, in this case, a modification of the all-important practice of hypoleptic delivery. [200] When the primarily reactive character of hypoboletic performance turns proactive, then ὑποβάλλω can refer to cuing and ἐξ ὑποβολῆς can be understood as ‘by cue’. Collins (2004:195) shows that the dynamics of hypolepsis point backward and forward, i.e. they are inherently reactive and proactive. Song sequencing and seam-stitching necessarily regard how the performer who leaves off sets up the thread that his successor must pick up. This form of cuing is wholly internal to the performance. I do not think, however, that we can dismiss out of hand the possibility raised by Boyd 1994 that at the Panathenaia the transition from one rhapsode to the next was at some point supervised by an official enforcer. [201] There is simply too much we do not know to be dogmatic about this, and it seems reasonable to imagine that there was a need for the external enforcement of competitive fairness. [202] But, whatever the truth of Boyd’s proposals, we must bear in mind that external rules would only have reflected (and perhaps enhanced) the intrinsic character and potential of rhapsodic hypoleptic performance and relay poetics. [203] That ὑποβολή in the sense examined here had its cradle among performers is supported by two epigraphical items. [204] The first is a second-century BC inscription from Khios, CIG 2214 (=SIG3 959). The second, CIG 3088 (=SIG3 960 n. 1), comes from Teos and is similarly dated. Comparing their competitive categories led Boeckh ad CIG 3088 to propose that the term ὑποβολή in CIG 3088.a1 and a4 stood for what the Khian text calls ῥαψῳδία. This would be strong evidence that ὑποβολή was still understood and used as a rhapsodic terme d’art.
The technical origin of the acceptations ‘reactive performance’ for ὑποβολή, ‘to perform reactively’ for ὑποβάλλω (i.e. ‘to perform in reaction to an earlier performance’), and ‘reactively’ for ὑποβλήδην (in a performance context) explains why rhapsodes used these words to project the dynamics of their own performance on the delivery of back-to-back speeches by Homeric characters. Martin 1989 has convincingly demonstrated that the speeches of Homeric characters are more than first-person thematic content to be quoted by the rhapsode in performance. From the viewpoint of pragmatics, as a matter of archaic and classical mimesis, they coalesce so decisively with the rhapsodic delivery in progress that the epic speakers in turn become performers in their own right. [205] This facilitated the application of specialized rhapsodic vocabulary to characterize them. As Martin (2000:410) observes: “The Glaukos and Diomedes exchange, and especially the speeches of Achilles (to Agamemnon, to the embassy, to Lykaon, to Aeneas, and to Hector) all show this consistent use by a second speaker of the first speaker’s phrases and topics, but with the added feature that the person who speaks second outdoes the first. … Furthermore, the paired speeches are carefully matched in structure.” I can illustrate the truth of this statement with reference to the first of two relevant Homeric passages, the confrontation between Akhilleus and Agamemnon in the first book of the Iliad. Indeed, the point of ὑποβλήδην at Α 292 is to mark the ensuing retort by Akhilleus as a supremely pointed answer to Agamemnon’s immediately preceding reproaches. The technical adverb underlines the thematic connection as singularly tight. As Kirk (1985:82) writes, Agamemnon “harps obsessively on Akhilleus’ domineering behavior.” Note his fourfold repetition at Α 287–289:
ἀλλ’ ὅδ’ ἀνὴρ ἐθέλει περὶ πάντων ἔμμεναι ἄλλων,
πάντων μὲν κρατέειν ἐθέλει, πάντεσσι δ’ ἀνάσσειν,
πᾶσι δὲ σημαίνειν, ἅ τιν’ οὐ πείσεσθαι ὀΐω.

But this fellow here wants to be above all others,
wants to rule all, lord it over all,
to boss all around—wherein I think someone will not obey him.
Akhilleus’ retort reacts closely to the formulation of Agamemnon’s contemptuous charges (Α 293–296):
          ἦ γάρ κεν δειλός τε καὶ οὐτιδανὸς καλεοίμην
          εἰ δὴ σοὶ πᾶν ἔργον ὑπείξομαι ὅττί κεν εἴπῃς·
295    ἄλλοισιν δὴ ταῦτ’ ἐπιτέλλεο, μὴ γὰρ ἔμοιγε
          σήμαιν’· οὐ γὰρ ἔγωγ’ ἔτι σοὶ πείσεσθαι ὀΐω.

          Well might I be called a coward and a nobody
          if I yield to you in everything you say.
295    Enjoin these things on others, but do not boss me around;
          for I, at least, am minded no longer to obey you.
Agamemnon has said, in effect: ‘who does he think he is, bossing everybody around?’ With his rhetorical question at Α 290–291, he further implies that, because Akhilleus owes to the gods his stature as a fighter, his greatness gives him no leave to insult others. [206] To paraphrase Akhilleus’ retort: ‘You will be right that I am a nobody if I yield to you and you still get to boss me around. I am the one who will not obey you!’ As Monro (1958:255) observes, “Achilles echoes l. 289, mockingly.” Akhilleus’ ‘reactive’ performance contains strong thematic and verbal echoes that closely mirror the words of the performer (Agamemnon) who has just left off speaking. This amply justifies the narrator’s recourse to ὑποβλήδην to qualify the manner of Akhilleus’ answer.
A similar analysis explains ὑββάλλειν at Τ 80. Agamemnon is not indirectly referring to himself, as is often assumed, but rather, as Pagliaro (1951:36–37) argued, to the approval of the Akhaians that immediately followed the end of Akhilleus’ address (Τ 74–75). Irritated by their obvious (and, to Agamemnon, disruptive) approbation of Akhilleus and their implicit reproach of him (cf. Τ 85–86), Agamemnon admonishes them for their breach of protocol. He remarks that one should ‘listen well’ (καλὸν ἀκούειν Τ 79) to him who stands up to address the assembly, and adds that it is not proper to answer him reactively, i.e. to offer his speech what amounts to tactlessly consequent criticism or praise. [207] This is irksome even to one who is skilled. If you react to a speaker with an uproar—he adds—how could anyone hear or speak? [208] The speaker is then hampered, clear-voiced though he be. [209] In this case, the Akhaians’ vehement approval is the ‘reaction’ that ὑποβάλλω refers to. The close connection is not wrought, as before, by a detailed response with verbal echoes, but by the audience’s immediate, unequivocal, and definitive embrace of Akhilleus’ address. What might well have been given verbal expression through one speaker who voiced his approval and the ensuing assent of the rest—see Η 399–404 and Ι 31–51—is here contracted to a two-line, third-person remark by the narrator.
It is important to emphasize that the Panathenaic rule—whatever its status as an actual stipulation—only formalizes what was already at the heart of rhapsodic poetics: the formal and thematic sequencing of which traditional diction and traditional plot were the outcome. Precisely because relay poetics is of the very essence of rhapsodic performance, one would not expect the various accounts of the rule to agree on its author or the time of its introduction. If traditional sequencing is fundamental to the rhapsode’s craft, the gradual fixation of the language and plot of epic narrative is the natural outcome of the rhapsodes’ cooperative performances, both in their diachronic and synchronic dimensions. That relay performance could be responsible for poems of recognized artistry is strongly suggested by the Hesiodic fragment 357 MW, analyzed at length below (§10.2.6). One plausible reading of it would be that Hesiod and Homer cooperated in stitching their respective hymns, the Delian and Pythian sections, as one song, namely, the extant Hymn to Apollo. [210] οἴμη and οἶμος
The rhapsodic emphasis on traditional thematic sequencing had another terme d’art: οἴμη, which seems to appear in a complementary distribution with the apparently related οἶμος. Bader 1990 offers a thorough and meticulous study of their etymology. She traces their origin, and that of ὑμήν and ὕμνος, to the IE root *sh2- ‘to link’, ‘to tie’, ‘to bind’. Osthoff (1898:161) had related οἴμη to Sanskrit sā́man- ‘song’ and old Norse seiđr ‘sorcery’. To these, Benveniste (1954a:39–40) had added Hittite ishamai- ‘song’. [211] Bader (1990:36) reconstructs *séh2-mn̥- > sā́man-, *sh2-°mi- > ishamai-, one from the e-grade, the other from the zero-grade of the root ‘to tie’. For seiđr and οἴμη, she posits e- and o-grade extensions to the zero-grade of the root: *sh2-ei- > *sai- > sei- and *sh2-oi- + -mā > οἴμη (1990:36–38). These elegant and persuasive etymologies allow her to relate these terms to the clearly relevant *ἱμᾱ́ < *sh2-i-mā (from the zero-grade of the root enlargement) [212] and Αἰσίοδος < *sh2-ei-ti-—the latter, according to the Etymologicum Magnum 452.37 (col. 1293 Gaisford), an Aiolic name for Hesiod (cf. Bader 1990:37 and n. 64).
οἶμος presents an added complication. There is evidence for both a psilotic οἶμος and an aspirated οἷμος. [213] Its single occurrence at Λ 24 apparently refers to the ‘stripes’ or ‘bands’ of a breastplate, [214] an acceptation for which there may be further evidence in the Mycenaean e-re-pa-te-jo-pi o-mo-pi (KN Se 891), if in fact o-mo-pi is οἴμοφι. [215] In Hesiod Works and Days 290 it clearly stands for ‘way’ or ‘road’ and it parallels ὁδός at 288. This justifies the persistent gloss ‘way’ vel sim., even in contexts where ‘song’ is in view and its relation to οἴμη is most immediate. So, for example, in the Hymn to Hermes 451, where Apollo says of the Muses that, among other things, they care about ‘the splendid path of song’, ἀγλαὸς οἶμος ἀοιδῆς. [216] Although I have printed the conventional translation ‘path’, it is in fact more likely that it should be translated as would οἴμη, ‘band’, ‘cord’, or ‘thread of song’. The variant ὕμνος ἀοιδῆς (cf. θ 429) [217] —about which Allen and Sikes (1904:336) observed, “it is doubtful if [it] should not be preferred”—puts οἶμος back in the context of the ancient technology of sewing and weaving. A review of ancient usage convinced Osthoff (1898:163) that the οἶμος of song and the οἷμος of path should be ascribed to different roots. Sommer (1905:29) made a proposal that has been well received by those who share Osthoff’s conviction: *oi-s-mo-s = *οἶhμος, a form attested by Lithuanian eismė̃ ‘motion’. [218] Bader (1990:36) restates Sommer’s form as *h1oi-(s)mo- (*h1ei- > εἶμι), and it is to be preferred to Osthoff’s *ϝοῖ-μο-ς (related to ἵεμαι ‘to desire’, ‘to pursue’), [219] against which Pagliaro (1951:26) raised valid objections on semantic grounds.
The two outcomes of *sh2-oi-m-, a first-declension οἴμη and a second-declension οἶμος, illustrate a lexical variation well attested in other cases. Compare: αὐχμός and αὐχμή, δεσμός and δέσμη, θάλαμος and θαλάμη, κάλαμος and καλάμη, γόνος and γονή, etc. [220] Pagliaro (1951:28) speculates that “il tema in -o rappresenta una funzione individuante nei confronti del tema in più generico e collettivo.” [221] Hence, οἴμη would be “legame, traccia” whereas οἶμος would be “striscia” (28). This suggestion is helpful, although I am not convinced that the distinction is attested where οἶμος is specifically used in connection with song, and the alternation -o/-ā, where it entails a semantic difference, can be motivated by a range of reasons. [222] That οἶμος had a longer afterlife than οἴμη may well be related to the influence of οἷμος ~ ὁδός (note its occasional feminine construal). After οἷμος had fallen together with οἶμος and by and large become its homophone, the semantic restriction of οἴμη to the technical sphere of rhapsodic performance would have been keenly felt. [223] Unless their intention was to draw narrowly on a background of traditional epic singing, authors would resort to the now unmarked οἶμος more readily than to the marked οἴμη. Later poets who use οἴμη are borrowing language from the rhapsode and his craft and making reference to traditional epic song. This is the case, for example, of Anakreontea 34 West, which speaks of cicadas as singers, an image straight out of the metaphoric world of epic. [224] Other late instances of οἴμη as ‘song’ are: Kallimakhos’ Hymn to Delos 9; Apollonios Rhodios Argonautika 4.150; Manetho Apotelesmatika 6.5 and 6.509; [225] [Lykophron] Alexandra 11; [226] Oppian Halieutika 3.3 and 28; and Anthologia Graeca 4.1.17 (οἴνης emended to οἴμης). [227]
The Iliad instances οἴμη three times, the first two in Odyssey 8. In the former, Demodokos sings about a quarrel (νεῖκος θ 75) between Akhilleus and Odysseus: Μοῦσ’ ἄρ’ ἀοιδὸν ἀνῆκεν ἀειδέμεναι κλέα ἀνδρῶν, | οἴμης, τῆς τότ’ ἄρα κλέος οὐρανὸν εὐρὺν ἵκανε (θ 73–74). Here οἴμης is an ablatival genitive of origin, ‘[starting] from that point in the song-thread whose kléos had then reached broad heaven’. [228] This passage sets οἴμη in close connection with the singing of κλέα ἀνδρῶν, as if the song were a cord tying the relevant subject matter along a traditional narrative sequence. In the second, Odysseus speaks in high praise of Demodokos and ἀοιδοί generally, who τιμῆς ἔμμοροί εἰσι καὶ αἰδοῦς, οὕνεκ’ ἄρα σφέας | οἴμας Μοῦσ’ ἐδίδαξε (θ 479–480). In the third, Phemios pleads for his life with the argument that he is ‘self-taught’: θεὸς δέ μοι ἐν φρεσὶν οἴμας | παντοίας ἐνέφυσεν (φ 347–348). These last two make the plural οἶμαι the sum total of traditional epic poetry and rhapsodic singing. Pagliaro (1951:29) aptly calls the οἴμη “[il] nesso di eventi che egli vuole sviluppare nella sua narrazione.” Because of its traditional character, once the rhapsode has chosen the epic οἴμη and his specific starting point along this song-thread, the sequence it entails constitutes “un dato obiettivo nella esibizione del cantore” (29). The synchronic limitation imposed by what Pagliaro calls “[la] congruenza dell’argomento” is the diachronic outcome of rhapsodic sequencing. The performative continuity is especially obvious in the relation between προοίμιον and οἴμη or οἶμος, the former being the ‘starting end’ of the rhapsodic performance. [229]
We see, then, that οἴμη and οἶμος, which must be understood etymologically as metaphoric connecting devices (i.e. ‘cord’ or ‘band of song’), were widely used in ancient Greece as metaphors for songmaking and the singing of traditional song. I suggest that these metaphors were so readily applied to Greek epic song because they underlined the tight performance sequencing and narrative connection that is at the heart of rhapsodic singing. They were also the notional ground for the rhapsode’s freedom to shift his performance forward and backward along the thematic thread. This freedom was gradually constrained to the point of eventual elimination, as the dynamic of accretion characteristic of Homeric poetry forced the explicit inclusion within the performance of the very stretches of narrative erstwhile bracketed (and only notionally incorporated) by these performative shifts. [230] The binding of verse to verse, speech to speech, episode to episode, and performance to performance (or rhapsode to rhapsode) is precisely captured by the image of a ‘cord’ or ‘[leather] band of song’. We find across the family of IE languages a similar metaphoric recourse to the technology of sewing in order to describe characteristic qualities of songmaking or singing or both. The old Norse seiđr was used for both ‘cord’ and ‘sorcery’, a trope that focused attention on the power of song to draw desired objects to the sorcerer (e.g. fish or a person); [231] or, as Heide (2006:164) has recently argued, it signaled that the mental images the sorcerer sent as his emissaries through seiđr ritual “could be regarded as something spun: a thread or rope.” In Hittite, ishima(n)- meant ‘string, line, cord, strap’; whereas the cognate ishamai- designated ‘song, melody’. Commenting on this fact, Puhvel (1984–:2.395) writes that “the semantic tie-in would be ‘rhapsodic’ in the literal sense,” and for his meaning he directs us to Hesiod fr. 357 MW ὕμνοις ῥάψαντες ἀοιδήν. In Greek traditional poetry, the semantic rationale for οἴμη and οἶμος is the sequencing, performative and thematic, that was the defining differentia of rhapsodic delivery: “l’οἴμη, in quanto è legame di contenuto nella recitazione di ciascuno e di tutti i cantori partecipanti, costituisce l’argomento e quindi il componimento ispirato a un motivo” (Pagliaro 1951:29). [232] ὕμνος
The conceptual correlative of οἴμη is ὕμνος. [233] Bader (1990:34–35) has shown that it too may be derived etymologically from the root ‘to tie, bind’. Furthermore, it is closely related to ὑμήν, which is attested both with a short and with a long υ: ῠ̔μήν, with the sense ‘thin skin, membrane’ (Apollonios Rhodios Argonautika 4.1648), and ῡ̔μήν, with the meaning ‘god of wedding’ (Theokritos 18.58). ῠ̔μήν follows readily from *sh2- with the affix -u-: ῠ̔μήν < *sh2-u-; ῡ̔μήν, on the other hand, resulted from the metathesis of the laryngeal and the affix -u-: *sh2-u- > *suh2-. [234] Before the suffix *-mn̥- in its long e-grade, [235] the loss of the laryngeal resulted in the compensatory lengthening of the -u-: *suh2-mn̥- > *sūmn̥-. -μην was used not only for agent nouns like ποιμήν but also for instruments like ὑμήν. [236] Like ῠ̔μήν ‘membrane’, ὕ̆μνος, with its short -ῠ-, followed from *sh2-u- by simple laryngeal loss without metathesis. Bader (1990:26) argues that this was in fact the older treatment (cf. 35 §10.3), and that metathesis arose as a ‘reaction’ against this loss. [237] Hence, besides the alternation of preconsonantal *sū- (e.g. sūtor) with prevocalic *sŭw- > *sŭ- (κασσύω, suō) [238] we also find preconsonantal *sŭ- and, for *sh2-i-, preconsonantal *sĭ̇-: to the contrastive pair ῡ̔μήν and ὕ̆μνος, add ῑ̔μάω and ῐ̔μάσθλη (and other derivatives of ἱμάς). ὕμνος is a thematic form built on the zero grade of the suffix < *sŭmn̥-os. Durante (1976:155–166) related ὕμνος to the Sanskrit sŭmná-, ‘benevolence, favor, song offered in tribute’. [239]
Etymologically, then, the notion of ‘hymn’ is built on a metaphor taken from the art of sewing, one that reifies the performance of epic song as a ‘membrane’ or, alternatively, as a product manufactured by tying or sewing. The etymology makes the formal and thematic sequence of traditional epic poetry metaphorically emphatic. The original generic meaning of ‘hymn’ is the unitary product of a rhapsodic performance, i.e. an integral whole of song and singing, the material delivered at a given performance and the performance itself viewed in its totality. [240] Nagy (2002:71) is therefore right to stress that ἀοιδῆς ὕμνον at θ 429 (cf. the variant at Hymn to Hermes 451) “conveys the idea of the totality of a given performance of song” (his emphasis). [241] Only the generic acceptation in evidence at θ 429 makes sense of the inscription with which Hesiod dedicated his tripod to the Muses after his victory over Homer in the Certamen: Ἡσίοδος Μούσαις Ἑλικωνίσι τόνδ’ ἀνέθηκεν | ὕμνῳ νικήσας ἐν Χαλκίδι θεῖον Ὅμηρον (213–214). As Martin (2000:414) remarks, “[these] lines may seem odd, because nothing in the narrated contest between the two poets resembles a ‘hymn’ in the narrower sense; the word [ὕμνῳ] seems to be used in this event as synonymous with epic verse” (his emphasis). I would slightly emend this observation to say that ὕμνος is synonymous with a performance of epic song. Evelyn-White (1914:587) translates ὕμνῳ νικήσας, “after he had conquered … in a contest of song”; West (2003b:341) writes, “having defeated in song”—they are doubtless right to avoid the (for us semantically much narrower) ‘with a hymn’ (vel sim.). [242]
Glossed as a ‘membrane-song’, the Hesiodic fragment that portrays Hesiod and Homer as rhapsodes stitching a song to Apollo ‘with new hymnoi’ gains the technical precision of professional suturing (see below, §10.2.6). On the relationship of ‘sewing’ to ‘tying’, Bader (1990:33) writes that “‘coudre’ n’est qu’un emploi spécialisé de ‘lier’,” and for this reason verbs derived from *sh2- with the former meaning are based on relatively late forms. [243] It is worth underlining here a fact recognized by many: irrespective of its original etymology, the combination of ὑφαίνω and ὕμνος in Bakkhylides 5.9–10 SM (ὑφάνας ὕμνον) [244] attests to a popular etymology that derived ‘hymn’ from ‘weaving’. Although popular etymologies are irrelevant to the meaning of a word at the time of its coining, they can greatly inform its later uses. I believe this was indeed the case with the word ‘hymn.’ It might be well to close this section with Bader’s own remarks on the use of technical terms to describe the art of singing and song:
L’emploi de formes d’une racine “lier” pour un art poétique qui comprend un chant est conforme à la conception générale de la poésie … comme comportant des techniques; celles-ci supposent un “agencement” (cf. *h2er-, §12.1), que peuvent préciser des métaphores de techniques mettant en oeuvre des agencements particuliers … . [La couture] est une forme particulière du “liage” … . Pour l’art poétique, c’est probablement un faux problème que d’hésiter … entre les sens “lier” et “coudre” de la racine [*seh2-], puisque la couture fut d’abord un liage, et qu’en outre les formes en *seh2-(i)- du vocabulaire de l’art poétique s’appliquent essentiellement au chant. [245]

10.2.4 Earliest attestations of ῥαψῳδός

The earliest attestations of the word ῥαψῳδός all come from the late fifth century BC. They are: Herodotos 5.67, SGDI 5786, and Sophokles’ Oidipous tyrannos 391. [246] From the former two there is very little that can be inferred about the essential character of rhapsodic performance. Herodotos 5.67 relates the intervention of Kleisthenes, the tyrant of Sikyon, against the rhapsodes’ performance of ‘Homeric poetry’ (τῶν Ὁμηρείων ἐπέων εἵνεκα) because it exalted Argos and the Argives, against whom he had then gone to war. [247] Herodotos’ report confirms the rhapsodes’ association with Homeric poetry. Ὁμήρεια ἔπεα perhaps designates the Homeric poems; or conceivably, as Cingano 1985 argued, [248] epic from the Theban cycle which at that time many still considered of Homeric authorship. [249] From the passage we also learn that rhapsodes took part in festival contests (ἀγωνίζεσθαι). [250] We cannot be certain that ῥαψῳδοί was the label applied to these performers during the rule of Kleisthenes in the late seventh and early sixth centuries. But I see no reason to prefer the view that the coinage of the term postdated these events and that the historian has resorted to an anachronism deliberately or in ignorance. The fourth-century BC (?) tripod inscription from Dodona tells us nothing about the manner of performance. [251] We only learn the name of the rhapsode, Terpsikles, and (presumably) the pride he takes in his trade—a fact, incidentally, that refutes those who believe that ‘rhapsode’ was coined as a term of shame (unless one accepts the convoluted conjecture that by Terpsikles’ time, ‘rhapsode’ had shed its negative connotations). [252]
We are left with Sophokles’ Oidipous tyrannos 391, which calls the Sphinx ἡ ῥαψῳδὸς … κύων. Why characterize the Sphinx as a ‘rhapsode’? For an answer, Jebb (1893:29) is typical: “ῥαψῳδὸς, chanting her riddle (in hexameter verse), as the public reciters chanted epic poems.” [253] It is worth considering that passages which mention the riddle also speak of the Sphinx in ways suitable to the unmarked ‘singer’. Euripides Phoinissai 48 calls her a ‘wise maiden’, a pun on the σοφία of the performer and the devious cunning of the female. She is said to perform μοῦσαι (50) whose meaning had to be learned by him who would rid Thebes of her; at 807 in the same play she is said to have ‘most discordant songs’ (ἀμουσοτάταισι ὠιδαῖς); [254] and in 1028, to have snatched youth away with her ‘lyreless Muse’ (ἄλυρον ἀμφὶ μοῦσαν), a manner of performance that well suits the classical rhapsode; finally, 1506–1507 speak of her μέλος and call her an ἀοιδός. Euripides Ēlektra 471 adds that her catch ‘involved song’ (ἀοίδιμον ἄγραν). [255] Sophokles Oidipous tyrannos 36 also calls her a ‘hard singer’ (σκληρᾶς ἀοιδοῦ); at 130 she is the Sphinx ‘of intricate song’ (ποικιλῳδός); and at 1199–1200 she is the ‘oracle-chanting maiden’ (παρθένον χρησμῳδόν).
What then are we to make of her characterization as a rhapsode? We can safely dismiss some proposals, for example, that she is so called because she engaged Oidipous in an oracular contest. [256] Hesiod fr. 278 MW attests to a contest of seers (specifically, Kalkhas and Mopsos in the Melampodia), but nothing in the transmitted multiforms suggests that Oidipous filled the place of a rival chanter of oracles who dared the Sphinx to unravel the meaning of a countering puzzle. [257] Ritoók (1962:229) suggests that the stitching of song by succeeding parties “wäre geeignet, um die Anwendung des Wortes auf die Sphinx zu erklären; die Sphinx hatte ja Rätsel gestellt, an die man die Antwort anzuknüpfen hatte, genau so wie in dem Agon.” But competitive song-stitching by consecutive rhapsodes, although important, is but one aspect of the notion of performative sequence at the heart of the rhapsode’s craft. It is neither a sine qua non—for solo performance of the proper kind may still be characteristically rhapsodic—nor is it plausible to argue that the Sphinx and Oidipous engage in sequential ‘verbal tying’ (Ritoók uses “anknüpfen” and “verknüpfen”) as rhapsodes would do at a festival contest (note Ritoók’s words, “genau so wie in dem Agon”). [258]
Can the notion of ‘rhapsode’ propounded in this chapter explain Sophokles’ designation of the Sphinx as a ‘rhapsode’? [259] I believe the answer is yes. In short, the Sphinx is readily called ‘rhapsode’ because: 1) her repertoire, formally and thematically, fits comfortably under the umbrella of traditional epic; 2) her manner of performance is characteristically that of the solo rhapsode; and 3) the source of her poetry and the terminology applied to her performance are peculiarly rhapsodic.
The riddle’s five hexameter verses are variously transmitted by Athenaios 10.456b (vol. 2, p. 491 §83 Kaibel); by the hypothesis to Sophokles Oidipous tyrannos (Dain and Mazon 1955–1960:2.70–71); twice by the scholia to Euripides Phoinissai, first in the hypothesis, then (with some changes and Asklepiades as the source) [260] in connection with verse 50 (Schwartz 1887–1891:1.243–244 and 256); the Anthologia Palatina 14.64; the scholia to Lykophron 7 Kinkel; and Tzetzes to Lykophron 7 (Scheer 1881–1908:2.11). Lloyd-Jones (1978:60) and Mastronarde (1988:6–7) offer convenient texts with apparatus. All of these are arguably formulaic multiforms of one another that bear the hallmarks of their oral composition and transmission. That Euripides included a different version in his own Oidipous (fr. 540a Kannicht) [261] sharply raised the question of his own source and the source of the more widely reported version. Lloyd-Jones (1963:447) had approved Robert’s conjecture that the better-known version (let us call it the ‘vulgate’) [262] probably came from the Thēbais or the Oidipodeia: [263] “[T]his version would not have been likely to be quoted by so many writers unless it had even better authority than that of Euripides” (Lloyd-Jones 1978:60). By 1978, his conviction had been undermined by Lesky 1966 [1928], unknown to him in 1963. Lesky’s argument combines a hypothetical reconstruction of the Oidipodeia that condemns the riddle as a later addition with the observation that, in at least two matters of diction, the vulgate proves to be post-archaic and cannot therefore be ascribed to the epic cycle.
Lesky’s further asseveration (1966:319) that, as a seven-thousand-line poem, the Thebaid had no room for the riddle displays his overconfidence. No matter. If the riddle belonged in the traditional Theban cycle at all, it merely needs a home there: if not the Thebaid, the Oidipodeia will do. Another question is whether it is an ‘original’ or ‘secondary’ element. Here we need not follow Lesky. Indeed, how much certainty one may have about the specific shape of archaic Theban epic can be judged from Mastronarde’s comment:
Most previous attempts to deal with the poorly documented background to the Thebaid myth … have been bedevilled by the desire to reconstruct specific literary artifacts of the archaic period … . Even if such reconstructions inspired more confidence than they do, they would give a misleading impression of the ‘sources’ of Thebaid myth. For … the literary distillations which survived in written form represented only a part of a larger oral tradition, a tradition that varied according to time, locale, and artistic temperament of the bards. … We have the scantiest possible fragments of the epic Oedipodeia, while those of the Thebaid are so few and uninformative that it can be debated whether they represent one or two poems. … Equally, our scant knowledge of local cults and of the stories attached to them … must create caution if not despair. [264]
In the face of such uncertainty, Lesky asserts that, although the Sphinx was doubtless featured in the Oidipodeia, it was no riddling monster defeated by Oidipous’ cleverness, rather, one he vanquished by sheer force (1966:320). [265] But what is the proof that such was the Sphinx of Theban cyclical epic? Lesky refers to the “certain impression” that the Sphinx in the much debated scholion of Peisandros [266] “did not pose riddles, but like other mythic pests merely picked its victims arbitrarily” (320). This comment seems at first incompatible with the scholion’s explicit language that “[Oidipous] then married his mother after he solved the riddle.” [267] Evidently, Lesky deems this statement one of those small interpolations that can be readily removed from the coherent subject matter of the scholion. [268]
But there is no narrative inconcinnity, much less contradiction, between the Sphinx’s use of the riddle and the scholion’s statement that she ‘seized up small and great and devoured them’. Even her arbitrarily picking out young victims [269] of high and low estate is compatible with her performing the riddle. [270] It was to an audience of male Thebans—which we have no compelling reason to believe excluded the young [271] —that she repeatedly tendered the riddle. Upon their failure to solve it, she might well proceed to snatch whomever within her reach she fancied. This is precisely what the scholion to Euripides Phoinissai 45 tells us. Difficulties arise only if we presume that all encounters with the Sphinx were one-on-one affairs—a presumption positively refuted by the vases. As Gantz (1993:496) observes, “youths rather than grown men will scarcely have been sent out to try to solve [the riddle].” But if we assume the concurrence of the riddle and youth-snatching, what Gantz pronounces scarcely probable would seem to follow from the literary references to young victims and the graphic emphasis on naked and unbearded preys (although Gantz 1993:496 admits that occasionally there are beards). [272] The consideration of a broader and inclusive male audience disposes of this specious difficulty.
Many have thought that the account by Asklepiades preserved by the Euripidean scholiast was an attempt to bridge a bifurcated mythic tradition that featured, on the one hand, a youth-snatcher and, on the other, a riddle-posing Sphinx. Vase depictions were marshaled to support this view: some showed the Sphinx pursuing or attacking youths; others, assemblies of men (seated and standing) around a Sphinx on a base (usually a column), in all probability listening to her performance and debating the meaning of her riddle and its answer. But the literary sources as transmitted—scholars’ suspicions of interpolations notwithstanding—fail to support any such bifurcation, and I believe that the iconographic evidence can readily be interpreted in harmony with the complex of Theban myth as we know it from our written sources. It is true that the earliest surviving items show males in flight, not figures seated or standing around the Sphinx. This seems to be a well-beloved decorative motif for drinking cups. One should not, however, infer ex silentio the non-existence of the riddle, for some vases that postdate the time when the riddle is attested still only show aggression and flight. A good example is the red-figure cup from ca. 510/500 listed as Cat. 23 by Moret 1984:168 (vol. 2 pl. 14). [273] Vase painting follows its own narrative conventions and one must be wary of making a simplistic correspondence between its iconography and the shape of mythic narrative. Consider, for example, Moret’s observation that, although our written sources portray a predatory Sphinx who kills and devours her quarry (see Moret 1984:10n2 for testimonia), “les images ne mettent pas en évidence ses instincts carnassiers. Tout au plus s’apprête-t-elle à bondir sur le dernier des fuyards (Cat. 4 et 5). Le plus souvent, elle est immobile (Cat. 2 et 3). La passivité de la Sphinx est le trait dominant, et le plus déroutant, de cette imagerie” (10). Neither does black-figure painting ever show the winged beast flying. But we are hardly to conclude from this evidence either that she could not fly or that in early myth she merely scared and did not attack her victims. [274]
A similar bifurcation on an evolutionary scheme is the proposal that in earlier versions of the myth Oidipous did not defeat the Sphinx by answering her riddle but by sheer force. Mastronarde (1994:20) notes the possibility merely on the grounds of the scholia to Euripides Phoinissai 26, which reports that the hero killed the Sphinx and the Teumessian fox, the latter clearly a feat of strength. The reasoning seems hardly safe or necessary. [275] Better the cautious statement that follows: “Yet it is unnecessary and unsafe to assume a subtle (or careless) reference to physical slaying in the phrase σῶμα φονεύσας at Phoen. 1507 … , and the illustrations on vases which have been thought to document the version without the riddle are both later in date than the illustrations of the riddle and open to other interpretations” (20). As it happens, two non-Attic vases attest to the riddle motif only some thirty years after our earliest vases pictured the fleeing victims (the C Painter Siana cups). Moret (1984:64–65) lists them as Cat. 100 and Cat. 101. They are an amphora in the British Museum (B 122) dated to ca. 540 BC and an amphora in Stuttgart (65/15) from ca. 530. The chronological gap between a hypothetical early form of the myth without the riddle and one that featured it is so small as to make the notion implausible and prejudiced. [276]
Katz 2006 is a tour de force that brilliantly demolishes such preconceptions. Devoted to the argument that “the story of the Sphinx and her riddle, if not also her name, goes back to Proto-Indo-European” (176), he offers compelling reasons to believe that ‘riddling’ (on whose definition see his 165n27) is of the Greek Sphinx’s essence. About Edmunds’s view that riddling is a secondary accretion in the story of Oidipous, Katz observes that there are plenty of Hellenists skeptical of Edmunds “and I expect that most Indo-Europeanists would be as well” (166n28). [277] To my mind conclusive is his reading of Hesiod Theogony 326–332, which follows Gotō 1995 in explaining ἐλεφαίρετο at 330 as “bamboozle (vel sim.)”: [278] “That it is, according to Hesiod, the Nemean Lion who does the bamboozling, rather than his sister the half-lioness Sphinx herself, is obviously a matter that needs further discussion (see, briefly, Gotō 1995: 366 and 368 n. 10). Still, it is surely not difficult to imagine either that both siblings have a share in riddling or that a trait of the sister has here been applied to the brother.” [279]
To move to Lesky’s linguistic arguments, they can be reduced to two: that ἀλλάσσω in the vulgate version of the riddle betrays late diction, for its first attested occurrence is as recent as Theognis; and that the use of ἀφαυρός with τάχος (the choice of some manuscripts over the more common μένος) deviates from epic usage. The force of Lesky’s arguments may be judged from his own ascription of the riddle to Antimakhos’ Thebaid. If its alleged source is an author steeped in the archaic epic tradition, any perceived divergence from it must be small indeed. [280] But, in point of fact, this is an idle ascription, devoid of probative value: for, if the late diction Lesky alleges finds thus its convenient ‘motivation’ in Antimakhos’ innovative use of epic diction, so also the riddle’s traditional language will follow readily from Antimakhos’ conventional reuse of Homeric poetry. [281] While this scenario is a priori possible, without any specific instances—which are not forthcoming—of ἀφαυρός or ἀλλάσσω in the fragments of the Colophonian that will vindicate his connection to the discontinuous usage charged by Lesky against the vulgate language of the riddle, Antimakhos amounts to little more than a rhetorical filler in Lesky’s argument. But what to say about these two points of epic diction? It is true that ἀφαυρός, whose etymology remains unsolved, [282] is almost always used in the Homeric poems to qualify diminished corporal strength. The LfgE s.v. glosses it as ‘schwach’: “[W]ohl das eigentliche (u[nd] nicht weiter spezifizierte) Adj. für körperliche Schwäche; außer Μ 458 (von der Wirkung eines geschleuderten Felsbrockens) stets unmittelbar von körperlicher Leistungsfähigkeit.” The five epic instances bear this out. [283] But it is not against these five, but against the ‘exception’ of Μ 458 that we must judge the riddle’s use of ἀφαυρότερον with τάχος. There, Hektor lifts up a heavy stone and hurls it against the gate of the Argive wall: ‘so Hektor lifted a stone and carried it straight to the boards … . He came and stood very close by, and leaning his weight into the cast he struck the gate in the middle, having planted his feet well apart so that his projectile may not be too weak’. [284] For my purposes, what matters in Hektor’s action is that he is hurling an object [285] and that the force of its impact is directly proportional to the speed imparted. This semantic development is not to be wondered at, since speed is a natural correlate of strength of limb. One would hardly be wrong to translate the last sentence of the passage, ‘so that his projectile may not be too slow’. Indeed, the text goes on to note that ‘the boards were severed [one from another] this way and that under the onrush (ὑπὸ ῥιπῆς) of the stone’ (Μ 461–462). ῥιπή is always used of fast and impetuous motion. With this in mind, we might re-read υ 110: twelve women worked at the mills grinding barley and wheat; the rest already slept, because they had finished their work, ‘but she alone was not yet ceasing [her work], but was slowest (ἀφαυροτάτη) [of them all]’. We may, of course, translate ‘feeblest’ rather than ‘slowest’, but it is clear that the focus is not on her bodily strength but on her slowness: she is taking much longer than her peers to finish her work. They have already gone to bed while she keeps on working. Once again, physical capacity is secondary to what it immediately translates into, i.e. speed. It bears emphasizing that these two make up one third of all the archaic epic instances of the adjective. Hence, for one in every three attested Homeric and Hesiodic occurrences, the gloss ‘slow’ for ἀφαυρός is at least as natural as, if not more pertinent than, ‘weak’. One may therefore question Lesky’s opinion that “man [kann] τάχος ἀφαυρόν nach dem sonstigen epischen Gebrauch des Wortes nicht ganz leicht sagen” (1966:319). Homeric usage may well have motivated Aratos’ choice of ἀφαυρός in Phainomena 227 to express speed (οὐδὲν ἀφαυρότερον τροχάει). The truth is that the degree of discomfort involved in applying ἀφαυρός to inanimate nouns (speed, strength of a bridge, intensity of light, etc.) remains substantially the same from the archaic through the Hellenistic period and beyond.
What about Lesky’s other objection, his assertion that ἀλλάσσω in the sense of ‘to change’ betrays post-archaic epic diction? Once again, my reply is that he is only partially right. It is true that ἀλλάσσω does not appear in the Homeric poems, and that, ceteris paribus, ἀμείβω might be considered a ‘more traditional’ choice (although it would be unmetrical in the same sedes and would call for a significant rewording of the riddle). All the same, Lesky has overstated his case, for ἀλλάσσω with the preverb ἐπί does occur in Homeric poetry once, at Ν 359, in a simile of hotly debated meaning:
τοὶ δ’ ἔριδος κρατερῆς καὶ ὁμοιΐου πτολέμοιο
πεῖραρ ἐπαλλάξαντες ἐπ’ ἀμφοτέροισι τάνυσσαν
ἄρρηκτόν τ’ ἄλυτόν τε, τὸ πολλῶν γούνατ’ ἔλυσεν.
Ν 358–360
The end of strong strife and leveling war
alternating it over both [camps] they tensed,
unbreakable, impossible to undo, which undid the knees of many.
This passage was already a crux in ancient times and the attempts to solve it are numerous. [286] Space prevents me from entering into a detailed exegesis here. I will only make the interpretive choices necessary for my present argument, which focuses on the word ἐπαλλάσσω. Homeric parallels and context show that the subject τοί must be Zeus and Poseidon (the plural can stand for the expected dual τώ). [287] This identification is confirmed by ἀμφοτέροισι; had τοί stood for the warring camps (Krates’ reading), this would privilege the alternative ἀλλήλοισι, apparently reported as a variant by Aristarkhos. [288] πεῖραρ is used metaphorically for the instrument that applies ‘strife’ and ‘war’ to both camps [289] and it is effectively a synonym of τέλος (Υ 101; cf. Β 121–122 Γ 291 Π 630). That it is conceived as a rope that effects conflict is clear from the adjectives ἄρρηκτον and ἄλυτον at Ν 360. The main verb τάνυσσαν conveys the tension of war and strife as it is wielded by the rival gods. It suggests the tensing of a bowstring (Δ 124) and is eminently suitable for a rope pulled in opposite directions. τείνω and τανύω are even used directly with ἔρις (Ξ 389 Π 662), πτόλεμος (Μ 436 Ο 413), μάχη (Μ 436 Λ 336 Ο 413), and πόνος (Ρ 400–401).
To come to the verb ἐπαλλάσσω, it arguably presupposes the availability of ἀλλάσσω. Thus, Lesky’s contention necessarily contracts to a single point: that ἀλλάσσω in the meaning used by the riddle is foreign to epic. [290] But the ἀλλάσσω that underlies the Homeric participle invalidates Lesky’s point and refutes his judgment that the riddle cannot therefore have an archaic-epic pedigree. This conclusion would seem to follow, unless of course ἐπαλλάξαντες in Ν 359 so differs in meaning from the riddle’s ἀλλάσσει that the latter breaks radically new semantic ground. I do not believe that the evidence supports the requisite degree of semantic discontinuity. The easiest way for me to make this point is to draw attention to the following semantic proportion. Heyne (1802–1822:6.438) translates ἐπαλλάξαντες as “alternando, alternis” and draws attention to Pindar’s use of ἐπαμείβεσθαι in Pythian 4.226. Heyne obviously construes ἐπαλλάξαντες intransitively. I do not believe that this is necessary. But neither voice is semantically determinative (see below), and my argument applies to the transitive and intransitive constructions just as well. Assuming, with Heyne, an intransitive construction, the semantic proportion is: ἐπαμείβομαι is to ἐπαλλάσσω as ἀμείβομαι is to X. The solution X is the intransitive ἀλλάσσω. For a transitive construction, the corresponding proportion is: ἐπαμείβω::ἐπαλλάσσω ~ ἀμείβω::X. X is now the transitive ἀλλάσσω. This proportion is significant because: 1) no one questions the archaic-epic pedigree of ἐπαλλάσσω in Ν 359; and 2) Lesky admits that, in contrast to ἀλλάσσει, there would have been no reason to impugn a hypothetical ἀμείβει as post-archaic. [291]
Now follow a few observations in support of the semantic proportion. The context decisively bears Heyne’s interpretation. [292] In view is the shifting battle fortune, as Zeus and Poseidon engage in a metaphorical tug of war. [293] This understanding is embraced, among others, by Köppen 1804–1823:4.48–50, [294] Bothe 1832–1834:2.178, Doederlein 1863–1864:2.16, [295] Ameis and Hentze 1905:25, [296] van der Valk 1963:99, [297] Michel 1971:52, [298] Heubeck 1972:142, [299] and Janko 1992:92. [300] The attentive reader of the preceding footnotes will have realized that scholars sometimes take the verb as transitive, sometimes as intransitive. This is not to be wondered at: the English ‘to alternate’ illustrates the reason well. Like ἐπαλλάσσω, it is as readily transitive (‘they stretched the rope of war over both camps, alternating it’) as it is intransitive (‘alternating, they stretched the rope of war over both camps’, i.e. ‘they stretched by turns’). The transitive meaning is clearly presupposed by those who think that the metaphor describes the tying or knotting of a rope. LSJ s.v. ἐπαλλάσσω I documents the transitive use; Ⅱ, the intransitive. Aristotle nicely illustrates the ready shift from transitive to intransitive: referring to animals who have interlocking teeth, he features the transitive construction in Historia animalium 501a18–19 (καρχαρόδοντα γάρ ἐστιν ὅσα ἐπαλλάττει τοὺς ὀδόντας τοὺς ὀξεῖς) and the intransitive in De partibus animalium 661b18–19 (τὰ δ’ ὀξεῖς καὶ ἐπαλλάττοντας [ὀδόντας ἔχει]). ἀλλάσσω allows for similar transitive and intransitive syntax. [301] The ἐπί of ἐπαλλάσσω, if intransitive, is best understood as referring to the succession of divine turn-taking. [302] If transitive, to the alternating target of the gods’ tug of war, now the Akhaians, now the Trojans; or else it should be taken closely with the ἐπ’ ἀμφοτέροισι that follows: the πεῖραρ (or τέλος) of war and strife devolves upon both camps, now the Akhaians, now the Trojans. [303] I think that word order commends the transitive reading: πεῖραρ immediately precedes ἐπαλλάξαντες and should be taken apo koinou with this participle and τάνυσσαν. πεῖραρ lends itself to a triple verbal pun. In the metaphoric tug of war, either god gains the upper hand and is overcome by turns—now Zeus prevailing over Poseidon, now Poseidon over Zeus—according as one or the other succeeds in pulling the ‘end’ (πεῖραρ) of the rope towards himself. The oppositive pulling explains the tensing (τάνυσσαν) of the rope (πεῖραρ in the sense of σειρή). As the ‘end’ of the rope moves, now in the direction of Zeus, now of Poseidon, the gods ‘repeatedly change’ (ἐπαλλάξαντες) the ‘end’ or ‘outcome’ (πεῖραρ as τέλος) of the strife and battle (ἔριδος κρατερῆς καὶ ὁμοιΐου πτολέμοιο), a change that can be readily envisioned as an exchange or alternation: victory is recurrently exchanged for (or alternates with) defeat and vice versa. The (ex)change affects ‘both camps’ (ἐπ’ ἀμφοτέροισι), that is, the divine rivals play out their tug of war ‘over’, and hence ‘upon’ or ‘against’, both military contingents, and hence ἐπί, with Cunliffe (1963:143 s.v. Ⅱ.1.a), should be understood as ‘on, upon’ “[i]n pregnant sense” (cf. Δ 178 Ε 384). [304] Insofar as πεῖραρ stands for a ‘rope’, it is natural to find the inescapable effect of the divine struggle upon the human agents restated metaphorically as an ‘unbreakable’ (ἄρρηκτον) rope, whose ends have been tied on either camp with knots that cannot be loosed (cf. θ 274–275). [305] The passage offers one final pun: the πεῖραρ that could not be loosed in turn ‘loosed the knees of many’. For my present argument, the main point in all of this is that ἐπαλλάξαντες construed with πεῖραρ as ‘outcome’ (a sense that Cunliffe 1963 s.v. 2.c renders ‘[t]he coming to pass’) readily translates as ‘changing the outcome’ (or ‘exchanging the outcome [of victory for defeat and vice versa]’). In the riddle of the Sphinx, ἀλλάσσω conveys the (ex)change of one external appearance (φυή) for another, [306] and it is therefore in semantic continuity with the Homeric ἐπαλλάξαντες.
Lesky’s linguistic arguments merely prove that, in the form preserved by the vulgate version, the riddle exhibits some features that, relatively speaking, are more recent than the corresponding Homeric ones. But just as the riddle displays formulaic language whose antiquity is neither the subject of skepticism nor without parallels in the Homeric poems, so also are there features in the Iliad and the Odyssey that some have claimed as interpolations precisely because they seem more recent than the scholar’s preferred termini ante quos for the poems’ composition. The truth is that the riddle’s formulaic multiformity proves its oral transmission, just as its traditional diction strongly suggests its oral-traditional composition. Traditional epic poetry, the Theban cycle included, continued to be recomposed in performance with some measure of variability into the classical period. Even if, arguendo, we accept Lesky’s view that ἀλλάσσω had not entered the poetic language with the meaning and in the construction exampled by the riddle until the time of Theognis—thus dismissing the implications of ἐπαλλάσσω at Ν 359—this hardly means that the riddle cannot derive from the Theban epic cycle. The Theognidea as a poetic corpus spanned the archaic and early classical periods. It too was traditional poetry, not epic but sympotic, whose growth must have had significant chronological overlap with the development of the Theban epic cycle, which rhapsodes continued to recompose in performance during much of the same period. Therefore, one cannot ascribe any one linguistic or lexical use by ‘Theognis’ to a particular period—certainly not to a period that is ex hypothesi later than a supposed dating of the Oidipodeia. One need only remember that attempts to fix in time a historical ‘Theognis’ resulted in seventh- and sixth-century dates as far apart from each other as a hundred years. Lesky’s objections only have force under the premises 1) that Theognis marks a terminus post quem for the use of ἀλλάσσω that the riddle exhibits; 2) that the text of the Oidipodeia was definitively fixed (composed in writing?) sometime in the archaic period and was not subject to recomposition that might have resorted to late-archaic or early-classical diction; and 3) that Theognis 21 can be securely ascribed to a time that postdates the conjectural archaic date by which the Oidipodeia had been textually fixed. These premises are at worst fallacious, at best hardly necessary, and in any case not in harmony with the diachronic perspective of epic composition and performance advanced in this book. They do not comport with the culture of performance, epic and sympotic, of archaic and early classical Greece.
As it happens, there is conclusive pictorial evidence for the text of the riddle that dates to ca. 520/10. The Vatican red-figure kylix (16541) from about 470 BC had long been the oldest extant (fragmentary) textual witness to the riddle; [307] on it, between the Sphinx and a seated Oidipous, we read a retrograde ΑΙΤΡΙ. This is presumably κ]αὶ τρί[πον … , i.e. the beginning of the second verse of the riddle. [308] Moret (1984:40) reported the existence of a new witness some fifty years older than the Vatican kylix, which has since been published by Kreuzer (1992:86–88). [309] It is a fragmentary Attic black-figure hydria from Basel (coll. Cahn 855) dated to 520/10 (Moret:1984 vol. 2 pl. 23). Moret (1984:40) is surely right to underline that “à cette date (520/510 avant J.-C.), la version courante n’a pu être que celle de l’épopée.” The photographs in pl. 23 are not easy to read. There is some disagreement whether the letters make sense. Immerwahr in his Corpus of Attic Vase Inscriptions no. 2072 writes that “they seem to be nonsense, but see below”—with a reference to Vollkommer (1991:60): “Auf dem genannten Fragment ist das Raten [der Bürger] besonders eindringlich wiedergegeben durch die Worte und Sätze, die über das ganze Bild hinweg verstreut sind.” Moret prints the readings τετράπουν οὗ and καὶ τρ[, [310] which are in broad agreement with the vulgate version of the riddle.
Perhaps the late sources that quote the vulgate version took it from a tragedy no longer extant that incorporated this excerpt from Theban cyclic hexameter poetry. The scholiast to Euripides Phoinissai 50 explicitly cites Asklepiades’ Tragōidoumena as his source for it. Lloyd-Jones (1978:60) writes that “this does not prove that the riddle was taken from tragedy, but it creates the presumption that it was.” If so, Aiskhylos’ Oidipous seems an obvious candidate. [311] Unlike the strongly innovative Euripides, the more conservative Aiskhylos is likely to have included a version substantially faithful to the rhapsodic recitation of the Oidipodeia in his own time. [312] If he is responsible for minor touches of linguistic updating, these cannot have been of any great significance: the failure of the later sources to ascribe the riddle to an author argues in favor of its ultimate traditional origin. In all probability, the original source was regarded as too obvious to call for specification. At the same time, given the tragedians’ numerous treatments of Theban myth, one cannot rule out that Asklepiades may have excerpted from Theban epic the eminently relevant language of the riddle, even if no tragedy actually incorporated the verses.
In light of these considerations, I am not impressed by Lesky’s arguments against the ascription of the vulgate version of the riddle to the Theban cycle. Neither seems West (2003a:41)—hardly a proponent of a performance-driven textual fixation of Greek archaic epic—who says of the vulgate version that “[it] is quoted by various sources which go back to Asclepiades of Tragilus … . There is a good chance that he took it from the Oedipodea.” [313] For my argument it is inconsequential whether the various details associated with the Sphinx and relevant to her characterization as a rhapsode belonged as one coherent narrative in any particular poem. What matters is that they are all traditional features of the Theban myth complex that formed the subject matter of the Theban epic cycle. These features need not even be entirely consistent with each other, much less need they harmonize with later tellings of the Oidipous saga. Discordant details are in the nature of traditional multiforms. Hence, for example, it is not important whether the people of Thebes met ‘daily’ to discuss the riddle’s meaning, as Asklepiades wrote, or just ‘often’, as we read in Apollodoros. More important is the agreement of these two sources—in all likelihood part of the traditional telling—that the Sphinx performed her riddle to an audience of Theban citizens often, like a rhapsode of rather limited repertory frequently reperforming the traditional poetry of his specialty (and exacting suitable ‘compensation’ for it).
There is, therefore, good reason to think that the riddle in its vulgate version goes back to well-known Theban cyclic epic. In my analysis of the Homeric Hymn 19 to Pan, I remarked on an important mimetic principle characteristic of archaic epic: it exhibits a poetics of performative self-referentiality that blends in complex ways the performance of a character within the narrative with the rhapsodic performance in progress. I also drew attention to Martin 1989, who has shown in detail how this mimetic principle, fundamental to rhapsodic poetics, decisively shapes the delivery of consecutive speeches in the Homeric poems. I suggest that a related kind of pragmatic blending accounts for Sophokles’ designation of the Sphinx as a rhapsode. After all, as a singer she is marked by the performance a popular ‘excerpt’ of a celebrated body of epic. She is to her internal audience what the rhapsode is to the real-world audience. In short, she is the narrative’s rhapsode. The traditional epic nature of her poetry, which bore the discernible hallmarks of an oral multiform, and the stability of her repertoire add to this portrayal. The riddle’s form was the rhapsode’s epic hexameter; its riddling subject matter, a recognized subclass of archaic epic, closely allied to oracular and gnomic poetry. [314] As Bollack also noted, her itinerancy is typically rhapsodic: the very line that designates her as the ‘she-rhapsode’ also alludes to her coming (‘when the she-rhapsode was here, the hound’, Sophokles Oidipous tyrannos 391). Euripides fr. 540a.12 features the word ὕμνον just two lines after Euripides’ version of the Sphinx’s riddle. In all probability, the dramatist has adopted rhapsodic terminology to typify her performance. Apollodoros Bibliothēkē 3.5.8 §52 tells us that, like any other rhapsode, she had learned her riddle from the Muses (μαθοῦσα δὲ αἴνιγμα παρὰ Μουσῶν). And, finally, her recurring performance before an assembled audience of Theban citizens marked her out as an infamous representative of the preeminent archaic performer, the rhapsode. Apollodoros writes: ‘after learning her riddle from the Muses, she would sit on Mount Phikion and would tender it to the Thebans … [who] coming together often (πολλάκις) would search what it was that was being told [by the Sphinx]; and when they would not find it out, after snatching one [of them] she would devour him’ (Bibliothēkē 3.5.8 §§52–53). The chronological extent of her recurring performances was dilated. Although Apollodoros notes that the citizens assembled to inquire into the meaning of the riddle, we are to imagine this meeting as an audience to the Sphinx’s performance, discussing its meaning in her presence. The vases fully bear this out, regularly representing the Thebans gathered around a Sphinx who typically sits on a column. Asklepiades confirms Apollodoros: the Thebans would gather daily in assembly on account of the woeful riddle of the Sphinx. And when they would not understand it, she would seize whomever of the citizens she wanted. [315]

10.2.5 The differentia of the rhapsodic craft

Having examined the essential character of rhapsodic performance, studied the relevant Homeric passages, and surveyed the earliest attested instances of ῥαψῳδός, it is time to answer those who seek to divorce the rhapsodes’ manner of performance from their poetic technique. These critics will accept relay singing as the defining differentia of rhapsodic delivery because the procedure was plain for everyone to see. But they reject the proposal that the solo stitching of traditional epic themes with traditional epic diction was also of the essence of the rhapsodic craft and set its practitioner qua ‘rhapsode’ apart from other kinds of ἀοιδοί. Yet this severance of composition from performance does violence to the fundamental insight that must guide a diachronic understanding of Homeric poetry: that the oral bard recomposed his song in performance. This implies that the manner of performance reflects the internal dynamics of composition and vice versa.
It should be clear from my analysis thus far that I view hypoleptic performance as entirely compatible with the poetics entailed by epic song-stitching in solo performance. In fact, I believe that the former was preceded by, and founded upon, the latter, for the principle of sequencing inheres in the composition of epic hexameter poetry and does not depend for its realization on a succession of performers at a festival. It is important, however, to add that rhapsodic compositional sequencing comes to fruition in the practice of relay performance. These two are complementary and intimately related facets of the classical rhapsode and his craft, and as such must be allowed to coexist as defining features. But the compositional sequencing practiced by a single rhapsode in performance holds diachronic priority over relays of several rhapsodes. And unless we are prepared to follow Else (1957b:33) in his judgment that ῥαψῳδός was a late label that issued from the practice of uncreative relay performance at the Panathenaia, we must accept the view that the performer himself or his audience could have coined, and in actual fact did coin, the professional label to make explicit reference to verse- and theme-stitching as was practiced by a solo epic performer.
Why should scholars prejudicially assume that audiences were not conscious of the traditional character of the diction and subject matter of Homeric poetry, and that they could not have tied this quality specifically to the performance of epic and its occasion? [316] Even children intuitively understand the traditional quality of bed-side story telling, as their insistence on a fastidious repetition of well-beloved story lines shows. The claim that the traditional character of the themes and the diction of epic was readily perceptible in their contrast with the themes and formal features of iambic, trochaic, elegiac, and lyric poetry does not entail an anachronistic focus of the epic performer and his audience on what was innovative over against traditional material in his composition. These other types of poetry adopted characteristic dictions peculiar to each and struck their own distinctive balance between tradition and innovation. [317] We should not marvel if this consciously realized perception received a precise articulation via artisanal metaphors. As soon as there were recognizable Iliadic and Odyssean traditions, with determinable thematic outlines and characteristic formulaic diction, there would have been identifiable episodes that a rhapsode in performance would string together in a sequence that in time would develop into the fixed plot lines (οἶμαι) attested by our written versions. ‘Stitching’ would focus attention on the industry and artistry of the resulting song, made of traditional episodes (the material stitched). All at once, the partial episodes as well as the whole song would have been traditional and also borne the stamp of the performer’s recomposition. There would have been no anachronistic focus on his ‘creativity,’ only on the artistry of the stitching—the felicity of the transitions, the compelling narrative of the stitched sequence, the skill of the diction used to render the individual episodes, the characterization of gods and heroes, the effect of their speeches, the adequacy and impact of the similes, etc.
It would not have been impossible to descend to the level of individual verse stitching. The traditional quality of Homeric epic must have been obvious to anyone who heard the rhapsode give expression to its long-established formal conventions. Scholars who dismiss the predictable formal quality of the hexameter as too abstract for an audience to realize in performance betray a surprising failure of imagination. I find no reason to doubt that listeners could and did perceive as recurring features its reduced pitch accentual melody, its rhythms, caesuras, cadences, and the overall effect of its bridges, cola, and the consistent arrangement of words by their metrical shapes. [318] This quality, cumulatively reinforced by each new line, would create the impression of a long array of near identical units of utterance, each perceptible as an individual whole, yet cleverly stitched by the rhapsode to the preceding and following verse with various semantic and morpho-syntactic strategies of periodicity and enjambment. [319] Even verses that are not self-standing semantic units enjoy the completeness of their formal structures. It is arguably the interplay between this prosodic wholeness and the rather common syntactic and semantic overruns [320] that would have impressed the audience the most, with an overall effect that the metaphor of stitching parts into a whole compellingly captures. This is the very effect Kirk (1985:34) notes: “The progressive form of enjambment, in particular, encourages the building up of longer sentences in a linear mode. Sometimes one can almost hear the components of a sentence [321] being cumulated one upon another in an accretive technique that can be prolonged or curtailed by the singer at will.” There is no need of the written page to feel the impact of the stitching technique. At first, Homeric epic was sung to the four-string kitharis or phorminx. [322] Even if we believe, with West (1981:121–122), that the notes traced the rise and fall of pitch accentual patterns and potentially each verse had its own melodic line, the reduced range of the melodic possibilities vis-à-vis those of lyric (which used the seven string lyre) [323] together with the overriding effect of the hexameter’s recurring rhythms, punctuated by the verse-end pause before picking up again, will have conveyed in clear perceptible outlines the traditional architecture of Homeric performance: the rhapsode performed his song by stitching smaller units one onto another—prosodic, formulaic, thematic, etc.—artfully expanding and abridging his material according to the need of the moment. These considerations should meet Ford’s desideratum that rhapsody refer to “concrete aspects of the performance or occasion” (1988:302). [324]

10.2.6 Stitching or weaving?

We have seen that etymology supports the view that the rhapsode is an ἀοιδός who ‘stitches’ or ‘sews together’ his song. But here I must deal with a final objection: some scholars have used ‘stitch’ and ‘weave’ as if they were interchangeable terms in practice. The equation, however, distorts the essential character of rhapsodic delivery, which should not be likened to the composition and performance of genres that do not share a comparable traditionality of form and subject matter. This misleading equivalence can only be put to rest once we acquaint ourselves with the technique of sewing in the Greek context. For as Durante (1968b:281n51) has shown, the verb ῥάπτω takes as direct object its product, not its means. [325] Therefore, the metaphor of song-stitching per se does not spell out in full the process by which the genesis of the song is envisioned. Our interpretation of the Hesiodic fragment 357 MW, that most important witness to the craft of the rhapsode, will depend crucially on the process entailed by ancient sewing and stitch-work. The scholiast informs us that, according to the late-classical Atthidographer Philokhoros, rhapsodes received their name from their practice of ‘composing [326] and stitching the song’ (ἀπὸ τοῦ συντιθέναι καὶ ῥάπτειν τὴν ᾠδὴν). In support, the following is allegedly quoted from Hesiod: [327]
ἐν Δήλῳ τότε πρῶτον ἐγὼ καὶ Ὅμηρος ἀοιδοὶ
μέλπομεν, ἐν νεαροῖς ὕμνοις ῥάψαντες ἀοιδὴν,
Φοῖβον Ἀπόλλωνα χρυσάορον, ὃν τέκε Λητώ. [328]
Marx (1925:399) construed the syntax as μέλπομεν ἐν νεαροῖς ὕμνοις … | Φοῖβον Ἀπόλλωνα and rendered it, “[i]n Delos haben damals zuerst ich und Homeros, die Dichter, besungen in neuen Hymnen, nachdem wir das Lied ersonnen hatten, den Phoibos Apollon.” But this translation depends on Fränkel’s unincisive “ersinnen” and entails the implausible view that ‘Hesiod’ is drawing attention to a compositional phase before the performance, at which the song(s) would have been stitched. There is a much better explanation for the aorist participle (see below). Clearly the focus of the fragment is on the performance in Delos, not on any preliminaries.
A further weakness of Marx’s reading is that his punctuation—“μέλπομεν ἐν νεαροῖς ὕμνοις, ῥάψαντες ἀοιδὴν, | …”—about which he is rather emphatic, [329] is not true to archaic epic diction: whenever archaic epic specifies the manner of performance designated by μέλπειν, μέλπεσθαι, or εὐμολπέω, it invariably resorts either to a simple adverb or, more often, to an accompanying participial phrase. To prove the point, I list here all the passages included by the LfgE s.vv.: ἐρατὴν δὲ διὰ στόμα ὄσσαν ἱεῖσαι | μέλπονται (Hesiod Theogony 65–66); θεαὶ δ’ ἐξῆρχον ἀοιδῆς | Μοῦσαι Πιερίδες, λιγὺ μελπομένῃς ἐικυῖαι (Hesiod Shield 205–206); εὐμόλπει μετὰ χερσὶν ἔχων λιγύφωνον ἑταίρην (Hymn to Hermes 478); νύμφαι ὀρεστιάδες λιγύμολποι | φοιτῶσαι πυκνὰ ποσσὶν ἐπὶ κρήνῃ μελανύδρῳ | μέλπονται (Hymn to Pan 19–21); μετὰ δέ σφιν ἐμέλπετο θεῖος ἀοιδὸς | φορμίζων (δ 17–18). So strong is this compositional habit, that μέλπεσθαι attracts the near-obligatory participle even when it only describes the performance in a rather indirect way; in the following example, its authority and excellence, as well as the corresponding social standing of the performer, are implied by the high regard in which he is held: μετὰ δέ σφιν ἐμέλπετο θεῖος ἀοιδός, | Δημόδοκος, λαοῖσι τετιμένος (ν 27–28). Α 472–474 exhibits a cognate construction: the unqualified participial form μέλποντες is coordinate with another participle, ἀείδοντες, which specifies it: οἳ δὲ πανημέριοι μολπῇ θεὸν ἱλάσκοντο | καλὸν ἀείδοντες παιήονα κοῦροι Ἀχαιῶν | μέλποντες ἑκάεργον. [330] μέλπεσθαι is used without qualification in Π 182 and Hymn to Hermes 476. The adjectives in Hymn to Apollo 197–198 [331] pertain to Artemis’ physique; if at all, only οὔτ’ αἰσχρή might hint at the quality of her performance. And in Η 241 [332] δηΐῳ … Ἄρηϊ cannot denote instrument or manner [333] but must be a locatival dative (‘I know how to sing-and-dance standing my ground [lit. ‘in stationary combat’] in destructive battle’) [334] or a dative of interest (‘I know how to sing-and-dance for hostile Ares in stationary combat’). [335]
In light of these parallels, it is perhaps unsurprising that the text is punctuated μέλπομεν, ἐν νεαροῖς ὕμνοις ῥάψαντες ἀοιδὴν, | … by Drachmann in his Pindaric scholia and by all the editors of Hesiod’s fragments without exception: Marckscheffel, Goettling, Kinkel, Sittl, Rzach, Evelyn-White, Merkelbach and West, and Most. [336] Indeed, the Hesiodic fragment follows the uniform usage reviewed above, that is, the verb μέλπομεν (here, in the rarer active voice) is qualified by the participial clause ἐν νεαροῖς ὕμνοις ῥάψαντες ἀοιδὴν. Although mistaken in his attempt to construe ἐν νεαροῖς ὕμνοις immediately with μέλπομεν, Marx cited helpful parallels for use of ἐν in adverbial clauses of manner or instrument. For example, Pindar Olympian 5.19 refers to ‘calling’ to the sound of (‘accompanied by’) Lydian pipes (Λυδίοις ἀπύων ἐν αὐλοῖς); in Isthmian 5.27 brave warriors are celebrated with lyres and the many-toned sounds of pipes (κλέονται δ’ ἔν τε φορμίγγεσσιν ἐν αὐλῶν τε παμφώνοις ὁμοκλαῖς); in Nemean 3.79 Pindar sends a drink to be sung to the accompaniment of the Aiolian breaths of pipes (πόμ’ ἀοίδιμον Αἰολίσσιν ἐν πνοαῖσιν αὐλῶν); and Sophokles Philoktētēs 1393–1394 envisions arguments that are instrumental to persuasion (τί δῆτ’ ἂν ἡμεῖς δρῷμεν, εἰ σέ γ’ ἐν λόγοις | πείσειν δυνησόμεσθα μηδὲν ὧν λέγω;). [337] We must now ask ourselves how ‘we sang Phoibos Apollo’ relates to the participial clause ‘stitching the song with/in new hymns’. [338]
A few comments about the aorist participle ῥάψαντες are in order. We know that participles detail aspect, not time. [339] As Kühner-Gerth Ⅱ.1.197 §389.E notes: “Während [das Partizip des Präsens] die Handlung in ihrer Entwickelung darstellt, bezeichnet [das Partizip des Aorists] dem Momentanbegriffe des Aorists entsprechend dieselbe bloss als Handlung gleichsam begebenheitlich.” “Momentan” need not be restricted to duration of time (‘instantaneous’). It may also refer to the verbal action seen complexively: “The complexive aorist is used to survey at a glance the course of a past action from beginning to end … . This is often called the ‘concentrative’ aorist, because it concentrates the entire course of an action to a single point” (Smyth §1927; cf. §1872.c.4). Given the imperfect μέλπομεν, nothing in the syntax would prevent Marx’s view that ῥάψαντες had preceded the action of the main verb. [340] But it is well known that, when the participle merely restates the main action under a particular aspect, it can be coincident with it. [341] I submit that this is precisely the case here: ‘we sang by stitching’ (‘in that we stitched’). In such cases, both the main verb and the participle are usually in the aorist. Although less common, an aorist participle with a future, present, or imperfect main verb are also sometimes found. [342]
But we still need a rationale for the choice of the aorist, since the present participle ῥάπτοντες would have suited the meter just as well. I submit that there are two equally valid alternatives. The first looks to the pragmatic contrast between imperfect and aorist as regards ‘focus function’: [343] by using the aorist, the fragment draws attention to the act of stitching, viewed complexively, and to its product, the song. The syntax effectively underlines the ‘stitching,’ i.e. the manner in which the action in progress, ‘singing,’ comes to pass. It is the concentrative aspect of the aorist that makes possible the pragmatic focus on the stitching. [344] ῥάπτοντες would have placed ‘stitching’ pragmatically on a level with ‘singing.’ The syntax that corresponds to this first alternative parallels that of ἐπαλλάξαντες in Ν 359, which I studied above. A second explanation is that the fragment portrays Hesiod and Homer as rhapsodes engaged in competitive relay poetics, with Hesiod taking up his singing right where Homer left off his. [345] Since ex hypothesi there would have been only one ‘seam’ at which the performances of Homer and Hesiod were joined—each discrete performance being one of the ‘new hymns’ stitched together—the stitching would have taken place only once in the course of singing. Hence the aorist.
Whichever our preference, I do not agree with Marx’s interpretation, which makes the stitching precede the singing. Neither do the many scholars who, whether instinctively or deliberately, have rendered the aorist as coincident with the imperfect, i.e. ‘stitching’ (not ‘after stitching’ or ‘having stitched’). [346] Rather, the νεαροὶ ὕμνοι are the raw material stitched in performance. This follows from the fact, noted above, that ῥάπτω takes as direct object the product of the sewing/stitching. If ‘new hymns’ had stood for the song stitched, it should have been in the accusative, in apposition to ἀοιδήν; or ‘song’ should have been in the genitive and construed with ‘new hymns’ in the accusative: νεαροὺς ὕμνους ῥάψαντες ἀοιδῆς vel sim., with a properly metrical complement (cf. θ 429). Most (2007:355 fr. 297) is surely right to translate “stitching together our song with new hymns.” [347]
The alternative construction that still keeps to the manufacturing metaphor but takes ἐν νεαροῖς ὕμνοις closely with ἀοιδήν is not readily intelligible. Stitching a part into a larger whole is possible, but what kind of ‘song’ could be stitched in/into new hymns? The resulting construction, moreover, effectively contravenes the usage rule that the product of the stitching should be the direct object of ῥάπτω. At any rate, this unattractive translation would still make the case that ῥάπτω denotes the combining of smaller units of performance into larger ones (in this case, the ‘song’ into ‘new hymns’). Evelyn-White (1914:281 “Doubtful Fragments” no. 3) adopts this alternative and makes it marginally acceptable to the English ear by translating ἀοιδήν, the concrete song of performance (‘our song’), merely as ‘song’ in the abstract (i.e. ‘the activity of singing’ and, by implication, its product in the abstract). ‘Stitching song in new hymns’ would be equivalent to ‘stitch-singing new hymns’, where the awkward ‘stitch-singing’ stands for ‘singing in a stitching mode’ (however understood). [348] There is yet another way to take ἐν νεαροῖς ὕμνοις, namely, as the performative context of the activity of stitching. Anakreon illustrates this alternative when he opposes Scythian drinking ‘with clatter and shouting over wine (παρ’ οἴνῳ)’ to drinking moderately ‘in [=accompanied by] beautiful hymns (καλοῖς | … ἐν ὕμνοις)’. [349] The parallel is not perfect, because Anakreon’s ‘drinking’ is only concomitant with the singing of hymns; in the Hesiodic fragment, ‘stitching our song’ is integral to ‘in new hymns’. This reading of the adverbial ἐν-clause does not construe it closely with the manufacturing metaphor of stitching. For this reason, although possible, I think it is less likely and compelling (note that the governing verb is ῥάπτω, not ῥαψῳδέω). It leaves undefined the particulars of song-stitching as a τέχνη: one merely learns that Hesiod and Homer stitched their (joint or respective) songs—with no further specification of manner—in the course of their new (individual or joint) performances. [350]
Now, if one adopts the view that the aorist regards ‘focus function’ and that ἐν … ὕμνοις does not specify the performative context of the rhapsodes’ singing, [351] how is their stitching with ‘new hymns’ to be envisioned? Given that the archaic ‘hymn’ was a unit of performance, one must imagine ‘Hesiod’ and ‘Homer’ to have joined smaller discrete performances into individual larger wholes in the course of their respective solo singing. By implication, the circumstances of their delivery would have had to provide notional ‘intermissions’ as occasions for seams. Something of the kind was true of Demodokos’ song in θ 87–92. [352] Alternatively, the point of the fragment may be to represent the larger song as stitched from smaller sections distinguished in turn by metabatic performance shifts—each section being an individual ὕμνος. [353] Or else the ‘hymns’ must refer to (the subject matter of) previous performances, and applying the adjective ‘new’ (νεαροί) to them proleptically—stricto sensu only true of the resulting song—would be a rhetorical figure not unlike a transferred epithet. [354]
In sum, the direct object of ῥάπτω, i.e. the product of the stitching, is the ‘song’ (ἀοιδήν). The material used, ‘new hymns’ (ἐν νεαροῖς ὕμνοις). To this Hesiodic fragment we should add the oblique testimony of Pindar about ῥαπτὰ ἔπη (Nemean 2.2), which commends equating ἀοιδή with ἔπη. The rhapsode, then, stitches the song of his epic performance out of smaller compositional and performance units, the ὕμνοι. There is no support here for Durante’s claim that what the rhapsode stitches into a song is the οἴμη (1968b:281), although οἴμη is arguably a rhapsodic term of great importance for our understanding of the rhapsode and his craft. In his analysis of ῥάπτω, Durante departs from the notion of ‘stitching’ and favors a meaning largely indistinguishable from ‘weaving’: “[D]ie antike ἱματιουργική unterscheidet sich nicht deutlich von der ὑφαντική” (1968b:280). His point is that in the ancient world, where rolls of fabric were not readily available for tailoring, sewing and stitching would have been marginal compared to weaving. But Cleland et al. 2007 s.v. “sewing” show that Durante overstates his case. While sewing was only one option for joining fabric, it was hardly rare: “The open side of Greek folded inner garments often seems to have been sewn … . Other garments (kandys, chitōn, cheiridōtos) were completely sewn, including the sleeves, but this manner of construction never became characteristic” (167). [355] Greek dress de-emphasized stitching and embroidery in favor of draping and folds; pins and brooches would have been more desirable than seams. But this does not mean that a metaphor based on stitching units one to another would have been hard for the average individual to grasp. [356]
At any rate, it is broadly acknowledged that skin and leather were materials commonly in use that did call for the stitching together of existing pieces. [357] The Linear B ra-pte, ‘sewer’ or ‘stitcher’, [358] perhaps in connection with armor, and ra-pi-ti-ra2, “sewing women,” [359] show that the practice of stitching was old enough to suggest the metaphor ‘song-stitcher’ as early as the late Bronze Age. Hesiod Works and Days 544 provides a well-known example of the manufacturing of winter clothing by stitching together skins of firstborn kids. Durante betrays a strange prejudice when he argues that the rhapsode’s craft should be connected with weaving and not stitching because the material most commonly stitched was leather and not fabric. Even if weaving became the dominant metaphor for poetic composition, this does not require us to reduce all craft metaphors to it. [360] Barber (1991:275) remarks with some plausibility that ῥάπτειν must be an old Protogreek loanword from a non PIE language parallel to the IE κασσύω. [361] Her analysis demonstrates that the existence of doublets in the Greek textile vocabulary is common, one of IE descent and the other borrowed from south Aegean substrate languages. [362]
To support his equating ῥάπτειν with ‘weaving’ Durante resorts to the rare words φορμορραφέω [363] and σχοινορραφέω. The former is a celebrated coinage by Demosthenes which Aiskhines ridicules as ‘incredible’ (ἀπίθανα). Demosthenes is alleged to have said: ἀμπελουργοῦσί τινες τὴν πόλιν, ἀνατετεμνήκασί τινες τὰ κλήματα τὰ τοῦ δήμου, [ὑποτέτμηται τὰ νεῦρα τῶν πραγμάτων,] φορμορραφούμεθα, ἐπὶ τὰ στενά τινες ὥσπερ τὰς βελόνας διείρουσι (Against Ktēsiphōn 166). [364] The Greek text is uncertain at points, but the reading φορμορραφούμεθα is not. Its meaning, however, is not obvious. It is possible that here the action implied by ῥάπτειν is the plaiting of rush mats. Hesykhios glossed it thus: ὡς φόρμοι κατα(ρ)ραπτόμεθα. [365] But how are we to understand the metaphor of plaiting the Athenian people as rush mats (or their state, if τὰ πράγματα is preferred)? Right after Demosthenes’ denunciation that the sinews had been cut, which evokes a body politic held together by ligaments and tendons, we might expect φορμορραφέω to depict an inferior way of political association that hampers the democratic state. Are we to think of each citizen as a plaited rush fiber? It is not clear why the resulting mat should be a natural metaphor for a defective organization (perhaps this adds to the ridicule). But in so close a proximity to needles, might this ῥάπτειν not refer rather to ordinary stitching? The metaphor of ‘patchwork’ seems immediately intelligible as inferior to ‘body politic’ and motivates the natural transition to needles. The patches might be individual plaited mats which, once stitched together, would make a rather weak overall structure. Stitching, in fact, may accompany plaiting in the making of rush mats. For example, in Tudor England rushes were plaited together into neat coils which were then sewn into mats. [366] The common acceptations of φορμός are ‘basket’ and ‘mat’, both of plaited reed, rush, or another suitable plant. [367] The scholiast to Aristophanes Wealth 542 shows that the suitable verb for plaiting is πλέκω, not ῥάπτω: [368] φορμὸς πᾶν πλεκτόν. ἐνταῦθα δὲ τὸ ψιάθιον; and: φορμὸς πᾶν πλέγμα, εἴτε ψιάθιον εἴτε ἄλλο τι. [369] These two activities and the corresponding stages in the manufacturing process should not be conflated. The only instances that bring together ῥάπτω and φορμός are those related to Demosthenes’ ‘incredible’ words. The TGL 8.1015 s.v. φορμοῤῥαφέω suggests the meaning “In tegete insuo. Vel potius, Ut tegetem trajicio acu.” If so, ῥάπτειν here would involve the usual needle work of stitching and not weaving or plaiting—only, not ordinary fabric but mats are involved. [370] For φορμός, the TGL 8.1016 collects statements from Hesykhios, Eustathios, the Suidas, Theophrastos, the Etymologicum Magnum, etc. All of them use πλέκω or derivatives like πλέγμα. [371]
The other example adduced by Durante (1968b:280) is σχοινορραφέω, from the scholia D to Κ 262, which to ‘ῥινοῦ ποιητήν’ notes: ἐκ δέρματος γεγονυῖαν, ἐσκυτορραφημένην. This, according to van Thiel’s edition, [372] is the reading in ZYQX (his sigla). QX also offer ἐσκηνορραφημένην, while Laskaris prints ἐσχοινοῤῥαφημένην, the hapax legomenon that concerns us. [373] The passage describes a boar’s-tusk helmet whose inner structure is leather; the teeth were sewn to it and one to another. [374] The text itself does not mention sewing. For my purposes, the relevant lines are Κ 261–263: … ἀμφὶ δέ οἱ κυνέην κεφαλῆφιν ἔθηκε | ῥινοῦ ποιητήν· πολέσιν δ’ ἔντοσθεν ἱμᾶσιν | ἐντέτατο στερεῶς. [375] The scholiast’s gloss ἐσκυτορραφημένην is reasonable: one must assume that the manufacturing involved the sewing of leather; [376] neither is ‘sewn like a tent’ wide of the mark, for the conical leather structure that supported the teeth was not unlike a small tent. But what could Laskaris’s hapax mean? If it means “stitched with cords,” as the LSJ s.v. notes, it is an odd attempt to explain the contested meaning of πολέσιν δ’ ἔντοσθεν ἱμᾶσιν | ἐντέτατο στερεῶς. A connection with weaving is marginally possible if σχοῖνος is used of anything twisted or plaited (cf. LSJ s.v. Ⅱ.1), in this case, twisted or plaited thongs. [377] This seems to be Lorimer’s view. She translates, “inside, it was made of thongs tightly strained,” and claims that the sense of ἐντέτατο is borne out by Ε 727–728, which she renders: “[Hera’s chariot] was made of … plaited strips.” [378] Cunliffe 1963 s.v. ἐντείνω seems to agree: “to strain tight (with interplaited straps).”
But I wonder whence these scholars have derived the notion that the making of the helmet involves plaiting at all. How plaited thongs should stiffen a leather helmet is a mystery, and as a substrate it would constitute a singularly poor structure onto which to sew the teeth. No wonder Reichel (1901:102) had already refuted these mistaken notions: “Wir nehmen demnach an, dass die Riemen dazu dienten, die Zähne mit der Haube als Nähte dicht zu verbinden.” [379] Borchhardt (1972:79) suggests that the helmet might consist of layers of broad leather strips sewn together and protected at the seams by the thongs. He does not, however, discount Lorimer’s “plaited” (at 80) on the basis of Hera’s chariot. [380] But ἐντείνω only means ‘to stretch tight’ and does not prejudge the structure of what is strained. Whereas an intransitive δίφρος … ἐντέταται [381] is readily intelligible as ‘the chariot floor stretches tight’ [382] (the floor stretches within a frame), ‘the helmet stretched tight within with many thongs’ is not: there is no frame to support a hypothetical substrate of tautly strained thongs. [383] In sum, this is a very uncertain foundation on which to build one’s understanding of ῥάπτω. Other professional terms ending in -ρράφος unproblematically entail ordinary sewing and stitching: σκηνορράφος (‘tentmaker’), κοσκινορράφος (‘sewer of leather sieves’), [384] ἡνιορράφος (‘saddler’), ἱστιορράφος (‘sailpatcher’), etc. As Durante observes, the preeminent classical professional of sewing and stitching was the shoemaker. [385] Among other names, he was called νευρορράφος by classical authors; post-classically, the rarer labels ὑποδηματορράφος and σκυτορράφος were also used. [386]


[ back ] 1. He places “oral” in scare-quotes.
[ back ] 2. Plato Laws 658b: εἰκός που τὸν μέν τινα ἐπιδεικνύναι, καθάπερ Ὅμηρος, ῥαψῳδίαν, ἄλλον δὲ κιθαρῳδίαν, κτλ. Plato Republic 600d: Ὅμηρον δ’ ἄρα οἱ ἐπ’ ἐκείνου … ἢ Ἡσίοδον ῥαψῳδεῖν ἂν περιιόντας εἴων … ; Certamen 55–56: ποιήσαντα γὰρ τὸν Μαργίτην Ὅμηρον περιέρχεσθαι κατὰ πόλιν ῥαψῳδοῦντα. Plato calls Phemios a rhapsode (Iōn 533c).
[ back ] 3. Cf. Wilamowitz-Moellendorff 1916:396–439, esp. 398; and Meyer 1918:335.
[ back ] 4. See footnote 2 immediately above for the references.
[ back ] 5. She notes that “[Homer and Hesiod] are assimilated to rhapsodes only when they are said to perform activities which are typical … of the rhapsodes proper” (Graziosi 2002:33). But the relevant point is surely that they are portrayed and said to perform as rhapsodes. The heart of Graziosi’s objection is that Homer is not only said to ‘rhapsodize’ (ῥαψῳδεῖν) but also to ‘compose’ (ποιεῖν) his poems. This, she believes, sets him apart from ‘true’ rhapsodes who are never said to compose (even if they arguably do). Graziosi does not consider that if the diachronic development of epic performance is as I argue here, precisely because Homer embodies the authority of Homeric epic one should expect biographical reports to separate his composing from his performing only to bring them together as the two stages of his professional activity. See my discussion immediately below.
[ back ] 6. E.g. Certamen 55–56, quoted in footnote 2 immediately above. For a discussion of the passages in the Vita traditions that ascribe composition to Homer, see Nagy 2010a:33–47.
[ back ] 7. Note, for example, the phrasing of Vita Herodotea §15. After leaving Kyme, Homer arrived in Phokaia, where he ‘earned his living in the same manner [as before], performing epē (ἔπεα ἐνδεικνύμενος 193) sitting in the public lounges’. At this point Thestorides, a teacher and would-be rival performer, ‘having learned of Homer’s verse-making’ (τοῦ Ὁμήρου τὴν ποίησιν 196), made him a proposal. Cf. Nagy 2010a:37–39, in particular his remark at 39 that the unscripted Homer is contrasted with the scripted Thestorides. Homer’s transcript becomes Thestorides’ script. That the performing of Homer is glossed by ‘verse-making’ is telling. For ποίησις as carminum factio see the TGL 6.1305: this is arguably its meaning at Vita Herodotea 68, 94, 218, 346, 372, and 375 (372 is of special significance because it summarizes Homer’s professional activity); it may also be similarly understood at 112 and 118. The instances at 145 and especially 209 perhaps regard the product, although verse-making remains in view: Thestorides does not merely want to appropriate the verses of Homer but to supplant him with his own activity as a rival performer. At 336, 349, and 521 we meet other marginal occurrences that straddle the making and the product. Notice ποιήσας at 350, which West (2003b:383) translates ‘writing’, perhaps because ‘composing that Odysseus entrusted his household to Mentor’ struck him as odd and unfamiliar. But preferring ποιέω to γράφω, the text emphasizes plot-construction over word-for-word composition, as one should expect of oral traditional poetry. The only instances of ποίησις for oeuvre are at 352, 380, 399, and 513. On occasion, Ps.-Herodotos writes that Homer performed τὰ ἔπεα—once τὰ ἔπεα τὰ πεποιημένα 143—as if he had in fact recited preexisting verses verbatim. This semantic slippage is motivated by the archaic meaning of ἔπεα, ‘utterances’, which does not concern detailed wording. Hence, ‘he performed the epē he had composed’ could, but need not, mean that Homer’s recitation slavishly followed a mental script. Cf., for example, Υ 204 θ 91 ρ 519. This brief survey confirms that ποίησις may indeed stand metonymically for its product, i.e. ποίημα and ποιήματα, but that even then the verbal notion—the ‘making’ in ‘verse-making’—often remains in view. An intuition of this fact underlies Aphthonios’ calling the Iliad a ποίησις while applying ποίημα to ‘the fashioning of Akhilleus’ armor’ (Progymnasmata 2.1 in Patillon 2008:113; cf. TGL 6.1305). As the verse-making of a poet, ποίησις readily stands for his ouvre, his life’s work, or for a representative collection of his ποιήματα. Cf. Chantraine 1933:287, Lledó Íñigo 1961, and Durante 1968b:267. For the use of ποίημα and ποίησις in Hellenistic (and later) criticism, see Ardizzoni 1953, Greenberg 1961, Ardizzoni 1962, and Häussler 1970.
[ back ] 8. See above, chapters 7–8.
[ back ] 9. A similar expression is used of Homer’s own performance at 112 and 143 (cf. 193).
[ back ] 10. Plato’s choice of Phemios, the celebrated Odyssean ἀοιδός, as the archetypal rhapsode (Iōn 533c1) further undermines the alleged incompatibility of ‘singer’ with ‘rhapsode’. There is no compelling reason to dismiss this portrayal of Phemios as tendentious or anomalous. The survey in chapter 1 of Graziosi 2002, which leads her to infer “substantial differences” (32) between rhapsodes and singers, lacks sufficient diachronic nuance. It neglects the use of ῥάπτω in the Homeric epics as “dangerous” evidence (23), inconclusive and possibly misleading; it infers too much ex silentio from the failure of Homeric poetry to mention rhapsodic composition (24); and it opposes the competitive setting of early instances of ῥαψῳδός to the Odyssey’s portrayal of Phemios’ and Demodokos’ performances (31–32). Her conviction that “the meaning of the term rhapsode must, in the first place, be sought in the fifth-century texts where it first appears” (22) precludes an adequate diachronic analysis and determines from the start the outcome of her survey. One is therefore not surprised to read that “it is attractive to suppose that these terms [rhaptō and rhapsōidos] indicated a new activity, to be contrasted, perhaps, with that of singers” (24). This is what she calls her “commonsense assumption” (32). But her discussion includes its own corrective, for she acknowledges that, in contesting the claims of the sophists, Plato “is relying on a current view of Homer” as a rhapsodic performer. Had this not been an ordinary and broadly accepted view of him, Sokrates’ remark at Republic 600d5–e2 would have lacked all force.
[ back ] 11. Nemean 2.1–3.
[ back ] 12. See above, §8.3.2.
[ back ] 13. Pelliccia (2003:112) acknowledges, as I do, that we glimpse here “a contemporary debate about education.” But, unlike him, I do not think that the passages from Xenophon serve in any significant way to emphasize, even implicitly, the different status of rhapsode and poet. What one might imagine Plato hypothetically doing with Euripides is not to the point. What matters is the traditional authority, cultural and educational, of the rhapsode, an authority that Euripides lacks. Furthermore, the claim that in the Platonic dialog that bears his name Phaidros performs “rhapsode-style” whereas Sokrates performs “as an aoidos” (113) is arguable. Since the text does not portray the action in those express terms, it can be obviously so only to those who have already decided how the alleged contrast of rhapsode with poet must be understood, articulated, and applied. The actual opposition in view in the Phaidros is the one between scripted and unscripted performance, a matter of vigorous debate in classical Athens. As I show in this book, this debate transcends the craft of the rhapsode. Where it overlaps with the dichotomy between creation and reproduction, it does so partially, not absolutely but in relative terms, as a matter of diachronic development. Scripted delivery by sophists and professional orators, not “professional memorizers” (whatever that may be), is what ἀπομνημονεύσειν in Phaidros 228a2 refers to (cf. Plato Greater Hippias 285e8).
[ back ] 14. For the meaning of mouvance, see Nagy 1996c:7–38.
[ back ] 15. Note also that κεκτῆσθαι is used not only of commercial purchases but also of acquiring mastery of a subject. This is the case, for example, in Plato Laws 829c8.
[ back ] 16. εὖ οἶδα ὅτι Λυσίου λόγον ἀκούων ἐκεῖνος οὐ μόνον ἅπαξ ἤκουσεν, ἀλλὰ πολλάκις ἐπαναλαμβάνων ἐκέλευέν οἱ λέγειν, ὁ δὲ ἐπείθετο προθύμως (‘Well I know that, hearing Lysias’ speech, he [=Phaidros] not only heard it once but, taking it up again often, told him [=Lysias] to deliver it, and [Lysias] eagerly complied’).
[ back ] 17. Cf. Yunis 2011:225–226.
[ back ] 18. To this speak Theophrastos Kharaktēres 27.7, cited by Pelliccia 2003:111n31, and the ᾄδουσι of Plato Iōn 534d7.
[ back ] 19. Pelliccia (2003:102) recognizes this difficulty, and the rest of his article seeks to justify extending his analytical schema to the rhapsode’s relationship to Homer.
[ back ] 20. E.g. Debrunner 1917:39 §79; Knecht 1946; and Patzer 1952:317, 320–321.
[ back ] 21. From *ῥαπτι̯- (Schmitt 1967:300–301 §609). Two further examples are: Στησίχορος and στῆσαι; φθεισίμβροτος and φθεῖσαι.
[ back ] 22. Cf. Patzer 1952:317, Schwyzer GG I.437–438, Bader 1972:205–206, and Probert 2003:108 §212.
[ back ] 23. ῥάβ-δος < ῥαπ- (cf. ῥαπίς). Among the witnesses to ῥάβδος, see Dionysios Thrax Tekhnē grammatikē §5 (Grammatici graeci I.1.8 Uhlig): εἴρηται δὲ ῥαψῳδία †οἱονεὶ ῥαβδῳδία τις οὖσα, ἀπὸ τοῦ δαφνίνῃ ῥάβδῳ περιερχομένους ᾄδειν τὰ Ὁμήρου ποιήματα; and the scholia vetera to Pindar Nemean 2.1d: τοὺς ῥαψῳδοὺς οἱ μὲν ῥαβδῳδοὺς ἐτυμολογοῦσι διὰ τὸ μετὰ ῥάβδου δηλονότι τὰ Ὁμήρου ἔπη διεξιέναι. Cf. scholia to Plato Iōn 530a5 and the near identical Photios Lexicon 3.317 no. 68 Theodoridis (=Suidas 4.287 Ρ no. 71 Adler). The scholia to Sophokles Oidipous tyrannos 391 supports ῥάπτειν: ἡ Σφὶγξ ἡ ῥάπτουσα τὰς ᾠδάς; so also the Etymologicum Gudianum s.v. ῥαψῳδία: λόγων συναφή. Cf., further, Ritoók 1962:225n1 and Ford 1988:300n4.
[ back ] 24. See Shapiro 1998:96 fig. 21 for a depiction on a red-figure neck amphora from ca. 490 BC, now in the British Museum (Vase E270).
[ back ] 25. Ω 343 ε 47 κ 238 293 319 389 ν 429 π 172 456 ω 2.
[ back ] 26. So also Wilamowitz-Moellendorff 1922:339n3.
[ back ] 27. Patzer 1952:316n1 (“‘krümmen’, ‘drehen’, ‘wenden’”); LIV2 690, doubtfully (“?[gr. ῥάπτω”; “[s]emantisch und formal unklar, könnte auch Anlaut *sr° haben”); Frisk 1973–79 s.v.; Watkins 2000:99 s.v. “wer-” ⅷ.
[ back ] 28. Cf. Lejeune 1972:119 §112 and n. 4, and Durante 1968b:278 and n. 43. Full bibliographical references are given in the DMic Ⅱ.221–223 s.vv.
[ back ] 29. Chantraine 1999: “l’étymologie traditionnelle … ne convenait guère pour le sens” (967); Lejeune 1972: “Les données mycéniennes … amènent à écarter la (médiocre) étymologie traditionnelle” (119); LIV2 690: “Semantisch und formal unklar.” Other words like ῥαμφή, ῥέμβομαι, or ῥάμφος can be related to *wr̥p- with greater plausibility.
[ back ] 30. For the possibility that it is a loanword, see below, §10.2.6.
[ back ] 31. Borchhardt apud Hainsworth 1993 ad loc.; also LfgE s.v. ῥάβδος B.2.
[ back ] 32. So Murray in his Loeb translation and Willcock apud Hainsworth 1993 ad loc. Lattimore translates ‘staples’.
[ back ] 33. I analyze this fragment at length below, §10.2.6. Graziosi (2002:33) remarks that “ἐν νεαροῖς ὕμνοις … may have been a necessary qualification: Homer and Hesiod did not perform in the manner of ordinary rhapsodes, their songs were new.” This special pleading reflects her conviction that rhapsodes did not explicitly admit to innovation and that, consequently, no one could have ascribed it to them (34). Rather than an illustration of rhapsodic poetics, for her this fragment espouses an untrue (and idealized) revision of it. But given Telemakhos’ celebrated recognition of the novelty of Phemios’ song at α 351–352, only a precommitment to wedge apart singer and rhapsode could turn the Hesiodic fragment on its head and make it an example of what rhapsodes did not do. I find implausible the notion that it is an aetiology, fashioned by rhapsodes about the origins of their craft, which does not however feature rhapsodic recitation, but a manner of performance that diverges from it at the crucial point of innovation. On Graziosi’s assumption, it is hard to see how the alleged aetiology could be rhetorically effective and useful to rhapsodes. Her reading moreover implies that neither Homer nor Hesiod had ever performed rhapsodically until they met in Delos to sing Apollo! (34).
[ back ] 34. For Kirk’s views on the rhapsodic “expansion or pollution” of Homer, see his index under ‘rhapsode’ (Kirk 1962:421).
[ back ] 35. See below, §
[ back ] 36. “‘[S]titching’ could signify a clearly audible difference between poetry that is sung and poetry that is not” (Ford 1988:306).
[ back ] 37. “Der Name, von den Alten verschieden abgeleitet, ist wohl von Haus aus als Spottname gemeint und nach Analogie von κιθαρῳδός, αὐλῳδός ironisch gebildet, da die Rhapsoden mit dem ᾄδειν überhaupt nichts zu tun hatten.”
[ back ] 38. I am not restricting ‘formal’ (in ‘formal characteristics’) to a literary conception of genre. In an oral performance culture, ‘genre’ conceptualized in terms of ‘performance occasion’ also exhibits ‘formal’ traits.
[ back ] 39. Meyer (1918:333) saw a clear example of the corresponding ‘seams’ in the ending to the Iliad that joined it to the Aithiopis.
[ back ] 40. Schmitt (1967:301 §609) is satisfied with the equation ῥαψῳδός = “Gesänge nähend” that he finds in Fränkel. He calls Patzer’s requirement for greater specificity “self-imposed” and conflates ‘stitching’ (ῥάπτω), ‘plaiting’ (πλέκω), and ‘weaving’ (ὑφαίνω). He translates ἐν νεαροῖς ὕμνοις ῥάψαντες ἀοιδήν as “in neuen Hymnen ein Lied webend”; ῥαπτῶν ἐπέων … ἀοιδοί as “die Sänger geflochtener Worte.” For him apparently ‘weben’, ‘flechten’, and ‘nähen’ are equivalent. Durante (1968b:280–281) makes a similar attempt to explain ῥάπτω not as ‘stitching’ or ‘sewing’ but as ‘weaving’ or ‘plaiting’ (see below, §10.2.6).
[ back ] 41. “Der Rhapsode ist der, der sein dichterisches Werk hervorbringt, indem er den Faden der Erzählung weiterführt, verwickelt oder verflicht.”
[ back ] 42. This assertion must be nuanced. The ordinary context for elegiac poetry was the symposium, where it was “sung with a piper (male or female) providing the accompaniment” (West 1992:25). Ford (1988:303 and n. 25) recognizes this fact, and it is therefore best to understand his restriction in negative terms, as denoting non-melic poetry or “poetry that can be spoken” (even if at times it was not). It is of course impossible to play and to sing to the auloi all at once, so that elegy sung to the pipe could not be a strict “solo” performance. One must also qualify the application of Ford’s statement to early iambos, at least to the iambos of Arkhilokhos (cf. Dale 1963:48, West 1992:40, and Rotstein 2010:230–232). For archaic iambos as a melic genre, delivered in melic, semi-melic, or non-melic performance, see Bartol 1992:70–71. In drama, aulodic recitative was the mode of performance for certain anapaests, iambic trimeters, dactylic hexameters, and iambic and trochaic tetrameters (West 1992:40 and Pickard-Cambridge 1988:156–167). For epic, see below.
[ back ] 43. Ford 1988:301–302 reviews previously suggested etymologies. Although he does not explicitly own his rejection of Patzer 1952—he uses the passive “Patzer’s view has … been criticized”—the logic of his argument is that Patzer’s treatment fails and the need remains for a new approach.
[ back ] 44. There is a degree of imprecision in the application of ‘recitative’ to ancient Greek music and poetry. The OED s.v. defines it as “a style of musical declamation intermediate between singing and ordinary speech.” ‘Declamation’, in turn, is the “uttering of a speech, etc. with studied intonation and gesture.” But “intermediate” lacks specificity, and it is sometimes hard to tell the “intonation” of ordinary speech, especially when emphatic, from the “intonation” associated with ‘to intone’ (=“to utter in musical tones; to sing”). Add to this semantic muddle: that ‘to chant’ and ‘to recite in monotone’ may be synonyms of ‘to intone’; that, among other acceptations, ‘to recite’, which recalls ‘recitative’, may denote ‘to intone’; and that any of these forms of delivery intermediate between singing and ordinary speech may have instrumental accompaniment. It is obvious, then, that by itself ‘recitative’ cannot situate a performance in relation to music and ordinary speech. It is easy to illustrate the confusion that ensues. As far as I can tell, West 1992 uses ‘recitative’ without defining it, as if its meaning were transparent. So, for example, he explains παρακαταλογή as “a technique of reciting verse with instrumental accompaniment” (40). This suggests ‘ordinary speech’ with an instrument in the background. But a few sentences later, West specifies that “the verses were recited in a more stylized manner” and that “[p]erhaps ‘chanted’ would be an appropriate term” (40). It is not clear what “stylized” or “chant” imply (monotone?), and we are little helped by the footnote on the same page that informs us that this is a “type of delivery … intermediate between ordinary speech and song.” The problem here is that παρακαταλογή arguably represents an intermediate stage between music and the rather cramped melody that some associate with recitative. At any rate, it is certainly not ordinary speech with an instrumental background. Hence, Nagy (1990c:46n140) is right to warn us that “[i]t may be misleading to some that West 1982.77 [West 1982] uses recitative to translate parakatalogē.” Nagy describes this mode of delivery as intermediate “between sung and spoken,” and, by writing thus, he makes clear that for him ‘recitative’ is stylized speech that cannot be considered between “sung and spoken.” Nagy’s own adjective for such intermediate forms is “reduced melodic” (1990c:46). For him, ordinarily, dactylic hexameter, elegiac distich, and iambic trimeter were recitative: although they still possessed prescribed pitch patterns, we would not recognize these as melody, not even as reduced melody (1990c:19–20). To West’s triad of melody, recitative, and ordinary speech, Nagy opposes the graded scale of (full) melody, reduced melody, recitative, monotone, and ordinary speech (but, nota bene, his ‘recitation’ may involve anything between reduced melody and ordinary speech). The limitations of West’s terminology are evident in his survey of the music of tragedy. Point (d) in West 1992:351 speaks of passages “recited … with instrumental accompaniment.” From the adjoined footnote, which mentions παρακαταλογή, we learn that what Nagy calls “reduced melody” is in view. The imprecision of ‘recitative’ is a reflection of its history. See, for example, the Oxford Companion to Music s.v.
[ back ] 45. Cf. Nagy 1990c:17–51.
[ back ] 46. See §10.2.2.
[ back ] 47. Cf. Appendix n. 5 below.
[ back ] 48. Cf. Nagy 1990c:33.
[ back ] 49. Ford 1988:305 and n. 35. The statement is common: e.g. Wilamowitz-Moellendorff 1916:341. Renehan (1976:91–92) and West (1981:113–114) cite numerous instances in which it is not always clear what synchronic contrast obtains between a verb of ‘speaking’ and a verb of ‘singing.’ Cf. Prauscello (2006:97–103), who takes issue with Renehan.
[ back ] 50. Just to cite one such synchronic leveling: the Iliad opens with an address to the Muse to ‘sing’ (ἄειδε); the Odyssey asks her to ‘tell’ (ἔννεπε).
[ back ] 51. See Nagy 1990c:21 for his comments about ἀείδω and ἐννέπω.
[ back ] 52. Cf. LfgE s.v. This choice is diachronically illuminating, as I hope to show in a future work.
[ back ] 53. Instead of μέλπομεν ἐν νεαροῖς ὕμνοις ῥάψαντες ἀοιδὴν, it might have been, for example, ἀοιδὴν | φαίνομεν ἐν νεαροῖς ὕμνοις ῥάψαντες ἀοιδοί (cf. θ 499) or ἐντύομεν νεαροὺς ὕμνους ῥάψαντες ἀοιδὴν (cf. μ 183 with Ε 720). I doubt that one must conclude from the distribution of ἐντύνω and ἐντύω in epic that only the former could govern ὕμνους. See LfgE s.v. ἐντύ(ν)ω.
[ back ] 54. For other examples of diachronic insight gained from Hellenistic poetry see González 2000 and 2010b.
[ back ] 55. I have no doubt that in its contrast of basileis to ἄνδρες ἀοιδοὶ … καὶ κιθαρισταί, the Theogony included its own performers among the latter. Just as βασιλῆες denotes the class of rulers comprehensively and generically, so also does ‘singers … and citharists’ potentially embrace all singers of archaic poetry (chiefly, the epic singer). ἀοιδοί and κιθαρισταί coordinately qualify the noun ἄνδρες. They are not two distinct classes of performers but a twofold description of the generic mortal performer. ἔασιν ἐπὶ χθόνα articulates the contrast between the divine and earthly spheres (the Muses and Apollo vs. mortal performers; Zeus vs. the basileis). It would be odd to think that to an utterly generic ἄνδρες ἀοιδοί a narrow class of instrumentalists, the κιθαρισταί, had been appended (καὶ κιθαρισταί might also be read epexegetically); or to conclude that ‘singer-citharists’ comprised a narrow class of performers that excluded the newly inducted ‘Hesiod.’ Despite its numerous ancient and modern proponents—e.g. Pausanias 9.30.3, who even thought that Hesiod had been barred from competing at Delphi because he had not learned to play the kitharis (10.7.3)—the σκῆπτρον of Theogony 30 was not a rhapsodic ῥάβδος that precluded a kitharis or phorminx (cf. Hymn to Hermes 210, 425–426, and 529), but an emblem of performative authority. My interpretation makes ‘Hesiod’ neither less of a rhapsode at Theogony 30 (cf. ἀείδειν 34) nor more of a singer at Theogony 95. The view that Theogony 94–96 includes epic singers even qua rhapsodes is supported by Homeric Hymn 25.
[ back ] 56. Cf. Farnell 1930–1932:1.254: “[Homer] hath told the story in the measured roll of his immortal verse, for other poets to disport themselves therewith.” Farnell (1930–1932:2.352) approves Wilamowitz’s reading of κατὰ ῥάβδον as a metaphor from weaving. But the fragment of Kallimakhos (fr. 26 Pfeiffer) on which it is based (see further below, § uses the preposition ἐπί, τὸν ἐπὶ ῥάβδῳ μῦθον ὑφαινόμενον. The acceptation of ῥάβδος instanced by Pollux 7.53 (‘streak’ or ‘stripe’) does not readily lend itself to construal with ἐπί (one would expect a simple dative or ἐν + dat.). Neither does ἐπὶ ῥάβδῳ seem an instrumental metaphor (one would expect ἐν + dat., cf. Pindar Olympian 5.19; ὑπό + gen., cf. Herodotos 1.17; or ἐπί + gen., cf. Bachmann 1928:1.425.9–10). We are simply left with the expression ἡ ἐπὶ ῥάβδῳ ᾠδή as a gloss for ῥαψῳδία or ῥαβδῳδία (Hilgard 1965:28.26 and 315.31). Pausanias 9.30.3 says of Hesiod that he ἐπὶ ῥάβδου δάφνης ᾖδε. Editors up to Bekker printed ἐπὶ ῥάβδῳ with codex Vindobonensis b, but this is a minority reading (Hitzig and Bluemner 1896–1910:3.1.361).
[ back ] 57. εἰς δαῖτα θάλειαν | καὶ χορὸν ἱμερόεντα καὶ ἐς φιλοκυδέα κῶμον (480–481). The connection between the lyre as an ἄθυρμα and the κῶμος resurfaces in Pythian 5.22–23. Cf. Bakkhylides Epi-grammata 1.3.
[ back ] 58. With West (2003b:200), I read οἶον for οἷον at 28: Hermes is not just one thematic choice among many exempli gratia. He is the preeminent theme of Pan’s cortège.
[ back ] 59. Scholia vetera to Isthmian 4.63e Drachmann.
[ back ] 60. On the notion of ‘diachronic skewing’ see Nagy 2003:39–48. He defines it as the perspective displayed by “the [poetic] medium [where it] refers to itself in terms of earlier stages of its own existence” (39). This skewing concerns “self-references in Archaic Greek poetry [that are] diachronically valid without being synchronically true” (Nagy 1990c:21). For example, whereas in the Odyssey the setting for the performance of epic is the evening’s feast, the earliest historical evidence points to festivals as the performance occasion.
[ back ] 61. Oration 26.316d: καὶ οὐ προσέχομεν Ἀριστοτέλει ὅτι τὸ μὲν πρῶτον ὁ χορὸς εἰσιὼν ᾖδεν εἰς τοὺς θεούς, Θέσπις δὲ πρόλογόν τε καὶ ῥῆσιν ἐξεῦρεν … ;
[ back ] 62. A clear statement is found in Else 1957b. He credits J. W. Donaldson for the idea (39n6) but notes that his article “did not spring from D[onaldson]’s suggestion.” See, for example, Donaldson 1860:61.
[ back ] 63. “Surely it is no accident that the nome, a solo performance and itself ultimately derived from epic song, is the one lyric genre with a formally marked prooimion. … [W]e cannot exclude the possibility that [in the prologue Thespis] appeared in character as the hero of the piece, introducing himself and reminding the audience of the background of his story” (Else 1965:60, my emphasis). Note that I do not endorse the view that the nome “derives from epic song.” The undeniable relationship between ‘nome’ and ‘epic’ is far more complex.
[ back ] 64. Cf. Philo De migratione Abrahami §111 CW; and Lucian Symposium §17.
[ back ] 65. With regard to the use of ἀείδω + ῥῆσιν at Aristophanes Clouds 1371, Herington (1985:225n15) writes that “a tragic rhesis may well on occasion have been intoned in the rhapsodic manner rather than delivered as ordinary speech.” The emendation of ᾖσ’ to ἦγ’ by Dover 1968a ad loc. prompted Renehan 1976:88–92. See, in particular, 90–91 and the treatment by Prauscello 2006:86–104.
[ back ] 66. Plato Iōn 531a1–2 (cf. 532a4–6).
[ back ] 67. Athenaios 14.620c–d.
[ back ] 68. Cf. Murray 1996:105 and Bartol 1992:67. Athenaios does not state unequivocally that Simonides performed Arkhilokhos in Athens, but the assumption seems reasonable (Burckhardt 1963:187).
[ back ] 69. With Bartol (1992:69–70), I do not believe that ὑποκρίνεσθαι here implies an out-and-out histrionic performance (so also Lennartz 2010:295; but cf. Bartol 1992:67–68 and nn17–19; Rotstein 2010:266 is neither incisive nor illuminating). Its use for ‘to perform’ stands in continuity with what I argue below must have been the original meaning of ὑποκριτής as a label for the actor: neither ‘expounder’ nor ‘answerer’, as the long debate thus far has had it, but simply ‘performer’. See pp. 647ff.
[ back ] 70. The text prints Σιμωνίδου, but West (1981:125) is surely right to correct it to Σημωνίδου. See further below, §11.1 n. 6.
[ back ] 71. The Suidas 2.403 Ε no. 2898 (Adler) makes Lysanias a teacher of Eratosthenes. Hence, he was roughly contemporaneous with Klearkhos. I deal further with Demetrios’ reforms below, §11.2.2. Modifications to the repertory and practices of the rhapsodes, with a growing emphasis on the histrionic potential of their subject matter, were already afoot. These changes must have been prompted by Homer’s rivalry with Attic drama for first place in the affections of the Athenians. Faced with the growing popularity of actors, rhapsodes must have been tempted to adopt stage manners that seemed to have gained their rivals the estimation of the people. The many speeches of the Iliad and the Odyssey gave ample occasion for character acting. Demetrios of Phaleron did not create these changes; he merely formalized and hastened them.
[ back ] 72. Apparent support for this claim dissipates on inspection. The statement about Stesikhoros in Athenaios 14.620c adds nothing to the argument since he was not a rhapsode. It is best to regard it as refracting a historical fact: the diachronic trajectory from song to poetry for the genres of epic (including gnomic), iambo-trochaic, and elegiac poetry. That Khamaileon must have thought Stesikhoros an exponent of the practice (adopted by some Hellenistic rhapsodes) of setting epic poetry to music follows not only from the comment Athenaios reports out of his work On Stēsikhoros, but is also consistent with [Plutarch] De musica 1132b8–c1 and 1133e7–f6. Cf. Nagy 1990c:27 and Russo 1999. A rhapsode’s recitation of Empedokles’ Purifications, even if not traditional in the sense of Homer and Hesiod—it certainly could not have involved oral recomposition—cannot be considered a radical departure from regular practice, since it was, after all, a hexameter poem (cf. Aristotle Poetics 1447b18). Little can be extracted from vague references to ‘other poets’ when Homer and Hesiod are mentioned (e.g. Plato Iōn 530b8–9 and 532a5; or Isokrates Panathēnaikos 18 and 33): they could well be hexameter poets such as Phokylides, the various putative authors of cyclic epic, etc. (cf. Notopoulos 1966:314).
[ back ] 73. Mark the use of ᾔσαμεν at Timaios 21b7. The movement from ‘many poems of many poets’ to ‘the poetry of Solon’—i.e. from the general to the specific—suggests that ἐλέχθη was the unmarked, ᾔσαμεν the marked description of the mode of performance. In other words, the text tells us that at least Solon’s poems were not delivered as ordinary speech.
[ back ] 74. Scholars admit as ‘securely Arkhilokhean’ iambic, trochaic, elegiac, and various lyric meters; for subjects, we have invective, consolatory poetry, epigrams, military themes, and many others. It is not unreasonable to imagine an over-zealous attempt to make him a master of all genres and meters—an attempt, therefore, that would also tie him to hexametric poetry.
[ back ] 75. Ford (1988:302n19) calls Notopoulos’s explanation “not convincing.” But other scholars are not so quick to dismiss it (cf., for example, Bartol 1992:65n3). Cf. also below, §12.3 n. 44.
[ back ] 76. Cf. Nagy 1976.
[ back ] 77. Cf. Lavigne 2005:12–57. Strictly speaking, epichoric traditions of praise poetry concern αἶνος/ἔπαινος rather than epic κλέος. Some of the ways in which this αἶνος converges on, and diverges from, epic κλέος are explored in Nagy 1976:194–199, Nagy 1986, and Nagy 1990c:146–198. Arkhilokhean iambos, its local origins notwithstanding, arguably had a Panhellenic projection. Cf. Nagy 1999a:250 §11.
[ back ] 78. Cf. Nagy 1999a:259 §9 and Lavigne 2005:48–57, both of whom draw attention to Aristotle Poetics 1448b24ff. On the Margites see Langerbeck 1958, Forderer 1960, Bossi 1986, Lennartz 2010:462–472, and Gostoli 2007.
[ back ] 79. Cf. Nagy 1999a:253–254 §2.
[ back ] 80. Only Semonides’ floruit was thought to rival Arkhilokhos’. Compare Hubbard 1994 and Lennartz 2010:473–474.
[ back ] 81. Cf. West 1974:27, Nagy 1999a:243–252, and Carey 1986.
[ back ] 82. This potential itself would have spurred the adoption of Arkhilokhean iambos by rhapsodes. Rhapsodic performance, in turn, would have been instrumental to actualizing its Panhellenic promise.
[ back ] 83. One cannot put much stock in Athenaios 14.632d, which numbers Xenophanes (with Solon, Theognis, and Phokylides) among those who did not bring singing (μελῳδία) to their poems. The passage blames Homer’s thoughtless metrical imperfections on his having composed his entire poetry for singing (μεμελοποιηκέναι). If Xenophanes did not exhibit comparable flaws, the logic of the argument would necessarily lead to the conclusion that melos played no role in the composition of his verses. Furthermore, if we embrace Athenaios’ testimony about Xenophanes, why should we not also accept that Homeric poetry was composed for singing?
[ back ] 84. It is also possible that by τὰ ἑαυτοῦ Diogenes only meant Xenophanes’ hexameter poetry. Before ἀλλὰ καὶ αὐτὸς ἐρραψῴδει τὰ ἑαυτοῦ he wrote: γέγραφε δὲ ἐν ἔπεσι καὶ ἐλεγείας καὶ ἰάμβους καθ’ Ἡσιόδου καὶ Ὁμήρου, ἐπικόπτων αὐτῶν τὰ περὶ θεῶν εἰρημένα. Since τὰ περὶ θεῶν εἰρημένα undeniably refers to ἔπη, it is not unreasonable to take τὰ ἑαυτοῦ as τὰ ἑαυτοῦ [ἔπη]. In other words, τὰ ἑαυτοῦ is not to be understood exhaustively as πάντα τὰ ἑαυτοῦ: ‘he rhapsodized his own’ means that he rhapsodized some of his poetry. Diogenes would then be drawing attention to a glaring inconsistency: however much Xenophanes criticized Homer and Hesiod, yet he himself too (ἀλλὰ καὶ αὐτός) embraced the composition and public delivery of epic verse as an itinerant rhapsode.
[ back ] 85. Plutarch Solōn 8.2. A few lines later we read, τότε δ’ ᾀσθέντος αὐτοῦ (8.3).
[ back ] 86. Demosthenes 19.252: ἐλεγεῖα ποιήσας ᾖδε. Polyainos Stratēgēmata 1.20.1: ἐλεγεῖα ᾖδε.
[ back ] 87. Or: ‘having composed a song instead of a speech, a lay of epē’. For κόσμος in the sense of ‘(epic) lay’ see above, §7.2.
[ back ] 88. The text, as often with Aristotle, is difficult to establish, and its meaning dense. I read it thus: ἡ δὲ [τέχνη] μόνον τοῖς λόγοις ψιλοῖς ἢ τοῖς μέτροις καὶ τούτοις εἴτε μιγνῦσα μετ’ ἀλλήλων εἴθ’ ἑνί τινι γένει χρωμένη τῶν μέτρων ἀνώνυμος τυγχάνει οὖσα μέχρι τοῦ νῦν (Poetics 1447a28–b2). The lines that follow illustrate this statement. The larger context shows that τέχνη stands for μίμησις: in 1447a13–16 ἐποποιία and ἡ τῆς τραγῳδίας ποίησις appeared not as products but as activities (i.e. as the making of epos and of tragic drama); with them were listed ἡ διθυραμβοποιητικὴ [τέχνη] and other arts, which were all said to be mimēseis.
[ back ] 89. Since trochaics, iambics, and anapaests were regularly mixed in drama, we must assume that Khairemon’s mixture went well beyond the ordinary canons.
[ back ] 90. Since Athenaios called the Centaur a ‘drama’, it must have included some mimesis of action. According to Rhetoric III.12, Khairemon’s plays also featured speeches. Because these were comparable to those of a λογογράφος in their stylistic precision, in the setting of public debate they struck a ‘meager’ pose. The further qualification ‘polymetric’ shows that the Centaur did not readily fit the canons of dramatic composition. Cf. Else 1957a:59 and Collard 1970:25–27. As noted above, Aristotle’s μικτὴ ῥαψῳδία is not intended as a generic classification, but it is used in its etymological sense of ‘a song stitched out of many meters’. Strictly speaking, Aristotle’s reluctance to classify the Centaur as a drama is not shown by his calling it a μικτὴ ῥαψῳδία—which, insofar as used in its etymological sense, is compatible with an overall dramatic form—but by his claim that no existing label captured its peculiar mimetic character. If naming it a drama had been unproblematic, the Centaur could not have served to illustrate the deficiency of existing mimetic nomenclature. Poetics 1460a1–2 states that Khairemon mixed iambic trimeters, trochaic tetrameters, and hexameters. I cannot agree with Ford (1988:304) that “[t]he implication [of this passage] is that he combined only these” (my emphasis). The text does not necessitate this restriction: its sole entailment is that Khairemon combined at least these.
[ back ] 91. Cf. Cope 1877:3.146–147 and Else 1957a:58. The text of Rhetoric 1413b8–14 is as follows: ἔστι δὲ λέξις γραφικὴ μὲν ἡ ἀκριβεστάτη, ἀγωνιστικὴ δὲ ἡ ὑποκριτικωτάτη. ταύτης δὲ δύο εἴδη· ἡ μὲν γὰρ ἠθικὴ ἡ δὲ παθητική. διὸ καὶ οἱ ὑποκριταὶ τὰ τοιαῦτα τῶν δραμάτων διώκουσι, καὶ οἱ ποιηταὶ τοὺς τοιούτους. βαστάζονται δὲ οἱ ἀναγνωστικοί, οἷον Χαιρήμων (ἀκριβὴς γὰρ ὥσπερ λογογράφος), καὶ Λικύμνιος τῶν διθυραμβοποιῶν. τὰ τοιαῦτα [δράματα] must be the ἠθικὰ καὶ παθητικὰ δράματα, both subspecies of ἀγωνιστικὰ δράματα. What remains is to illustrate the other side of the classification, γραφικὰ δράματα, which should be ἀκριβῆ (1413b9). This expectation is met when Khairemon is mentioned (ἀκριβὴς γὰρ ὥσπερ λογογράφος 1413b13). But instead of calling him γραφικός, Aristotle focuses on the character of the γραφικὴ λέξις—that its natural outworking is reading (τὸ γὰρ ἔργον αὐτῆς ἀνάγνωσις 1414a17–18; see below, §14.2). Hence, he labels him ἀναγνωστικός. Nothing in the context suggests, much less requires, that we think of Khairemon as an exponent of a genre other than drama. Cf. Sifakis 2002:157n23 and Innes 2007:159–160.
[ back ] 92. For the evolutionary model, ‘post-Homeric’ is an ill-defined terminus. We might think of it roughly as ‘the time when the text had largely achieved a substantial degree of fixation.’
[ back ] 93. Homeric Hymns 1.18, 2.494, 3.20, 6.2, 30.18, 32.2, and 32.19. (Also Hesiod Theogony 48?)
[ back ] 94. Cf. Ford 1992:35–39.
[ back ] 95. I have already considered this passage above, §7.2, where it is quoted (and provided with a slightly different translation) in connection with the view of inspiration it entails and its implications for our understanding of the epic medium.
[ back ] 96. More on this passage above at §7.2.3.
[ back ] 97. Cf. Ford 1992:34.
[ back ] 98. λιγυρὴν δ’ ἔντυνον ἀοιδήν [sc. Σειρῆνες] (μ 183; cf. Homeric Hymn 20); τεύξουσι δ’ ἐπιχθονίοισιν ἀοιδὴν | ἀθάνατοι (ω 197–198).
[ back ] 99. E.g. Hymn to Hermes 483, addressed to Apollo. Because the playing of the lyre is immediately in view, the focus on τέχνη and σοφίη is not problematic. This section of the hymn centers precisely on imparting art and knowledge (cf. 25). See also the simile at ψ 159–162. Without a context it is hard to judge the implications of Hesiod fr. 306 MW. Cf. Ξ 53–54.
[ back ] 100. θ 492 and the ubiquitous κατὰ κόσμον. For more on κόσμος as a rhapsodic term, see above, §7.2.
[ back ] 101. Used for ‘to marshal [troops]’, not for ‘to arrange [the song]’ (cf. Homeric Hymn 7.59).
[ back ] 102. Rumpel 1883:475 s.v. φράζω: “Act. indico, dico, eloquor, c. acc.” Also Slater 1969:535 s.v. a: “declare, expound.” Wilamowitz-Moellendorff (1922:339n3) translates, “Homer hat die Taten des Aias den Späteren deutlich gemacht” and “Homer … [hat] von Aias erzählt.” Thummer (1968–1969:2.74) wonders whether to construe κατὰ ῥάβδον with ὀρθώσαις or ἔφρασεν. Its significance as an emblem of authority commends that we take it apo koinou with both. See above, §10.1.1.
[ back ] 103. See above, §10.2.1 n. 56. More on this fragment below, §, in connection with relay poetics.
[ back ] 104. παρα- conveys the contrast of the hypothetically false story with a notional standard of truthfulness.
[ back ] 105. Planning is at times presented in a positive light, e.g. at Κ 19–20.
[ back ] 106. Nagy (1996c:64n22) is right to reject Scheid and Svenbro’s claim that there were no such metaphors in existence until the later era for which archaic choral lyric is extant. He is also right to draw attention to the Homeric use of ὕμνος and οἴμη as counter arguments (66n23).
[ back ] 107. ἀλλ’ ἦ τοι μνηστῆρας ἀγήνορας οὔ τι μεγαίρω | ἕρδειν ἔργα βίαια κακορραφίῃσι νόοιο· (β 235–236). The same observation applies to Ο 16. After Hera’s celebrated deception, Zeus wakes up from his sleep on mount Ida and sees Hektor lying on the plain. Realizing what has happened, he addresses Hera: ‘Hera, impossible woman! Forsooth, the trick you contrived with evil intent (κακότεχνος … σὸς δόλος) has stayed (ἔπαυσε) goodly Hektor from the fight and has routed (ἐφόβησε) his host. I know not but that you may in turn be first to reap the fruits of your grievous evil-stitching’ (Ο 14–17). Hera’s κακορραφίη, which stands for her κακότεχνος δόλος, is so far from mere inactive devising that it is personalized and said to have stopped Hektor and routed the Trojan armies.
[ back ] 108. κακὰ ῥάπτομεν ἀμφιέποντες is not merely “we kept busy plotting” (so Ameis and Hentze 1920:73). ἕπω means “to direct one’s energies at someone or something” (so LfgE s.v. B), and ‘to be busy at’ would not be an inadequate rendering if it did not suggest, as I think it does, planning apart from action. A survey of its Homeric use establishes that energetic action with bodily motion is the norm. As the LfgE notes: “it is noteworthy that many instances imply manu[facturing] act[ivity] (cf. deriv. ὅπλον) and that the subj. is always a pers[on] w[ith] the exc[eption] of χεῖρες … and πῦρ/ἀυτμή … , of animals 1× indir[ectly] in sim[ile].” With the preverb ἀμφι- (a simplex is instanced only at Ζ 321) it means “attend to, beset, iter[ative] and insistent.” One class of instances (LfgE s.v. 1a) refers to butchering and preparing meat; the second (1b) is the one that concerns us, “beset an enemy” (i.e. surround and harass, assail on all sides), which Λ 474 γ 118 and Hesiod Theogony 696 illustrate; and a third (1c), like the second in meaning, with πῦρ as subject (Π 124 Σ 348 θ 437). It should be clear that ἀμφιέποντες at γ 118 means ‘assailing them on all sides’, and that κακὰ ῥάπτομεν refers to the execution of their military campaign and stratagems on the battlefield. (S. West’s “probably ‘carefully, intently’, rather than ‘pressing them hard’” apud Heubeck et al. 1988:167 seems an unwarranted accommodation to the prejudice that ῥάπτειν must mean plotting apart from action.)
[ back ] 109. The four Homeric instances of ἄπτερος belong here, if ἀ- is privative (not copulative) and the implication of the statement ὣς ἄρ’ ἐφώνησεν, τῇ δ’ ἄπτερος ἔπλετο μῦθος is that the corresponding female failed fully to understand the implications of the utterance spoken. If not with certainty, this may be readily argued for ρ 57 τ 29 φ 386; χ 398 can be similarly interpreted if one gives emphasis to the manner in which Telemakhos addresses Eurykleia (ἥ τε γυναικῶν | δμῳάων σκοπός ἐσσι κατὰ μέγαρ’ ἡμετεράων χ 395–396)—little could she anticipate how Odysseus presently intended to apply her oversight of the servant girls! I cannot go into this famous crux here (for a good review see Russo et al. 1992:22–24), except to remark: 1) that the word was understood in both its privative and copulative senses (see LSJ ad loc.); 2) that ἀπτερέως in Hesiod fr. 204.84 cannot be determinative, for although it may be best understood as copulative (‘with wings’, i.e. ‘swiftly’) if the supplied ἐπί̣θ̣ο̣ν̣[το is right, ‘silently’ (=without demurring) is also possible (as is ‘without realizing the implications of what they were doing’!—although, admittedly, such a meaning here can only follow indirectly from the more common application of ἄπτερος to μῦθος). There is, at any rate, no guarantee that the Hesiodic fragment and Homeric epic agree in their use of the opening ἀ-. I find Russo’s defense of the equation ἄπτερος = πτερόεις attractive, but it is not without difficulties.
[ back ] 110. Γ 125–128 Ζ 456 Χ 440 etc.
[ back ] 111. ἔντοσθεν δὲ βοείας ῥάψε θαμειὰς | χρυσείῃς ῥάβδοισι διηνεκέσιν περὶ κύκλον (Μ 296–297).
[ back ] 112. ῥυπόωντα δὲ ἕστο χιτῶνα, | ῥαπτὸν ἀεικέλιον, περὶ δὲ κνήμῃσι βοείας | κνημῖδας ῥαπτὰς δέδετο (ω 227–229).
[ back ] 113. πρωτογόνων δ’ ἐρίφων, ὁπότ’ ἂν κρύος ὥριον ἔλθῃ, | δέρματα συρράπτειν νεύρῳ βοός (Hesiod Works and Days 544).
[ back ] 114. On Helen’s weaving see Nagy 1996c:64n23.
[ back ] 115. μή μοι μεταμώνια νήματ’ ὄληται (=τ 143 ω 133), to be compared with κλέος οὔ ποτ’ ὀλεῖται (Β 325 Η 91 ω 196). For Penelope’s κλέος, see ω 196–198. It is striking that scholars should expend their energies on the meaning and etymology of μεταμώνια (β 98), which simply adds emphasis to ὄληται, and fail to consider how Penelope’s rationale makes sense. The very few exceptions are superficial and generally unilluminating. Translating “so that the threads are not wasted uselessly,” Dawe (1993:99) observes that this must be her reply to “why not finish your weaving after your marriage?” Censuring her words as “vague,” he poses the follow-up question, “‘Why should laying your weaving aside for a little while mean that all your work to day … is wasted?’ We are given no answer, because we are given no question, and Homer hurries us on.” The truth is that Penelope’s words only make sense metapoetically. There is no compelling reason why she could not weave on after remarrying, unless we assume without evidence that her incorporation into a new household would preclude her continuing to care for her previous father-in-law (in which case, her having finished the shroud would hardly be of help). Even if this had been so, why express anxiety about the spoiling of the ‘threads’ (νήματα, standing for the ὕφασμα and her labor at the loom) rather than about her failure to provide for Laertes’ burial? Why the suitors should think Penelope’s rationale valid is another question. Presumably as an expression of piety they considered it right that it should take precedence over her remarriage. At any rate, the point of her weaving, of course, is to stave off her marrying one of the suitors. Were they not to wait for the completion of the φᾶρος, then her weaving—the ruse—would certainly be in vain, and the κλέος of her cunning and marital fidelity would indeed perish. Although harmonizing the three tellings of the ruse (in Books 2, 19, and 24) presents some difficulties, the threefold repetition itself, the prominence the ruse is given at τ 137–158, and the fact that at ω 128–150 the completion of the φᾶρος marks the arrival of Odysseus leave us in no doubt that Penelope’s guileful weaving defines her heroic identity and functions as a narrative metaphor of first importance (for the timing, see Heubeck et al. 1988:137). Cf. N. Austin 1975:127, Slatkin 1996:234–236, and Clayton 2004:21–52.
[ back ] 116. Besides ὑφαίνω, LfgE s.v. gives the related terms ἀρτύ(ν)ω, μήδομαι, μηχανάω/-ομαι, μενοινάω, μερμηρίζω, τεκταίνομαι, φράζομαι, φυτεύω, and (later) κασσύω.
[ back ] 117. See above, §10.2.2.
[ back ] 118. πῶς δὴ ἔγωγ’ … | οὐκ ὄφελον Τρώεσσι κοτεσσαμένη κακὰ ῥάψαι; (Σ 364–367).
[ back ] 119. πεύθετο γὰρ οὗ παιδὸς ἐνὶ μεγάροισιν ὄλεθρον (π 411).
[ back ] 120. τίη δὲ σὺ Τηλεμάχῳ θάνατόν τε μόρον τε | ῥάπτεις … | … οὐδ’ ὁσίη κακὰ ῥάπτειν ἀλλήλοισιν (π 421–423).
[ back ] 121. Cf. δ 675–678.
[ back ] 122. Porzig (1942:119–120) observes that τεύχειν (“das eigentliche und allgemeine Verbum für ‘handwerklich verfertigen’” in epic) may be used with an abstract noun as a periphrasis for the verbal action that corresponds to it. This is the case, for example, at υ 11, whose context precludes ‘to plan’ as the meaning of τεύξειεν: θάνατον … τεύξειεν is not ‘should prepare death’ but ‘should kill’. Instances like υ 11 unequivocally collapse onto actio the inventio that τεύχειν connotes at α 277 β 196 κ 18 etc.
[ back ] 123. Ford (1988:301) should not insist on novelty as a metaphoric connotation of ῥάπτειν. His glosses “ingenious machination” and “contrivance” are acceptable, and one may add that the schemes recounted advance the plot. But there is nothing inherently novel in them (i.e. new or unusual). ῥάπτειν and other semantically related terms neither imply new thematic material nor unprecedented or anomalous narrative developments.
[ back ] 124. Cf. Β 119 θ 579–580 ω 199–201.
[ back ] 125. It is true that the Sirens claim to know ‘all that happens on the fruitful earth’ (μ 191), but it seems likely from their addressing Odysseus as πολύαινε (μ 184) that they are not claiming omniscience but a comprehensive awareness of all that is ‘news’ and worthy of song. As has often been remarked, they are like the Muses in this regard (Β 485).
[ back ] 126. To fit the hexameter, Bentley corrected δεδεγμένος to δειδεγμένος with no change of sense.
[ back ] 127. The review of διηνεκές and its usage in van Tress 2004:31–38 is helpful, although I cannot fully endorse it.
[ back ] 128. Translating more freely: ‘Taking up the narrative that carries the authority of the rhapsodic tradition woven in unbroken continuity with it I sing’. Or, assuming μῦθον τὸν ἐπὶ ῥάβδῳ ὑφαινόμενον as the corresponding prose word order: ‘I take up and sing in unbroken continuity the story that is/was woven upon [=with?] the staff’. This seems the order assumed by Cameron (1995:363), who construes μῦθον with “heard” (for δειδεγμένος?): “I 〈heard〉 the story woven with the aid of a staff and tell it from beginning to end.” Asper (2004:89–90 fr. 29) implausibly assumes that ὑφαινόμενον is a middle: “und ihn, der am Stab Märchen webt … in steter Folge singe ich, wie ich empfangen habe.” The rendering of Ernesti (1761:497) is “Et carmen super virgam contextum perpetuo canto, comiter accipiens.”
[ back ] 129. What one can learn from the Iōn about the rhapsode and his craft is explored above, chapter 8 passim. Pelliccia (2003:106–108) disputes the legitimacy of placing ‘poet’ and ‘rhapsode’ in diachronic continuity and, a fortiori, my claim that diachronic analysis reveals the two as overlapping categories. But he fails to take the measure of Plato’s late and tendentious perspective, even though at 113 he dismissively anticipates and attempts to forestall this very objection to his analysis. In an author as late as Plato, what is significant is not what separates Homer the poet from Ion the rhapsode, but the ways in which a continuity of function is hinted at. See further above, §8.3.2.
[ back ] 130. Given the focus of ῥάπτω on the recomposition of traditional material, Kallimakhos’ self-conscious reappropriation of the authentic rhapsodic tradition is better designated by ὑφαίνω. For a somewhat different reading of this fragment that also stresses the programmatic importance of rhapsodic poetics for the Aitia, see Durbec 2003.
[ back ] 131. διηνεκές has this meaning in Strabo 3.1.3, where ὄρος … διηνεκές refers to an uninterrupted mountain range; in Μ 297, where ῥάβδοισι διηνεκέσιν denotes the continuous sequence of gold stitches along the circular perimeter of Sarpedon’s shield (on the meaning of ῥάβδος here, see further above, §10.1.2); and in IG Ⅱ/Ⅲ2 1666.b60, where ἐπε]ργάζεσθαι δὲ κατὰ τὸν στοῖχον ἕκαστον (λίθον) διανεκῆ refers to placing stones in a continuous line, one after the other. Cf. DGE s.v. A.I.3.
[ back ] 132. Here I am using ‘epic cycle’ broadly to refer, in the case of the Trojan-War myth, to all thematically related poems, including the Homeric. Cf. the statement of Proklos apud Photios concerning the ancients’ appreciation of the ἀκολουθία τῶν ἐν αὐτῷ [sc. τῷ ἐπικῷ κύκλῳ] πραγμάτων (PEG I.6 T22 Bernabé = EGF 1 T1 Davies).
[ back ] 133. For an informative and assertive reading of this fragment, see Cameron 1995:339–361. He rightly emphasizes that Aristotle’s views of dramatic unity lie behind Kallimakhos’ words (343–344). From the perspective of rhapsodic poetics, however, Cameron’s attempt to separate temporal from thematic continuity is fallacious (344–345), whatever the understanding of the neoteric Roman poets who glossed ἄεισμα διηνεκές as perpetuum carmen (cf. Horace Carmina 1.7.6 and Ovid Metamorphoses 1.4). This is not the place to compare at length the poems of the cycle with the Iliad and the Odyssey in regard to their poetics and aesthetics. I will only say that rhapsodic poetics embraced all manifestations of traditional oral epic, including the Homeric poems. But the differential manner in which Athenian festival regulations and audience expectations focused upon the Iliad and the Odyssey—in particular, the eventual thematic restriction of Panathenaic performance to these two, to the exclusion of cyclic epic, and, consequently, their comparatively more thoroughgoing Panhellenic shaping (cf. Nagy 2001 and Burgess 2001:14)—in their case led the rhapsodic tendency towards comprehensiveness to develop in a correspondingly differential manner. In contrast to the relatively less Panhellenic Trojan-War cyclic poetry, which preserved a longer temporal scope for its plot sequence, the Iliad and the Odyssey developed their own characteristic comprehensive monumentality by way of forward- and backward-looking references (cf. Burgess 2001:199n34). Aristotle (and the Alexandrians) viewed poems as self-contained products, as unitary collections of verses of specific authorship and with established texts, not as reflecting the potential for thematic expansion and contraction open to traditional performers according to the needs of the moment. The exclusion of the two Homeric poems from the cycle—like the contrast between cyclic poets and Homer—betrays a point of view characteristic of the late classical and Hellenistic periods. One must remember that Aristotle Sophistical Refutations 171a7–11 (PEG I.2 T8 Bernabé = EGF 13–14 T2 Davies) mentions κύκλος in connection with ἡ Ὁμήρου ποίησις. Insofar as a web of traditional myths connected with the Trojan War existed with a recognizable thematic coherence, this nexus of oral tradition could have been described as a ‘Trojan-War cycle’ (itself perhaps part of an even larger mythological cycle). This would have been true irrespective of the particular instantiations of such a thematic κύκλος. If Burgess (2001:143–146) is right to infer that the Kypria and the Little Iliad narrated the complete story of the Trojan War (doubtless with different emphases), the contrast between the comprehensiveness peculiar to such cyclic poems and the comprehensiveness characteristic of the Iliad and the Odyssey would have been pronounced. (Burgess’s central insights about the epic cycle do not depend on his assumption that particular instantiations of the Trojan-War cycle recognizable to the ancients as Proklos’ Kypria, Sack of Troy, Little Iliad, etc. were textually fixed at a time earlier than the textual fixation of the Homeric poems. For this view see, for example, Burgess 2001:12 and 198n28.) On all these matters, see Nagy 1990c:72–73, Scaife 1995, Holmberg 1998, and Burgess 2001:12–33.
[ back ] 134. The well-known strategy of verbal ornamentation characteristic of Akhilleus and the Homeric narrator is a feature of this tendency. Martin (1989:196) has called this phenomenon the “expansion aesthetic” (cf. 205–230). This aesthetic markedly shapes pairs of consecutive Homeric speeches: “He who talks bigger talks better” (Martin 2000:410).
[ back ] 135. μῦθον δ’ ὡς ὅτ’ ἀοιδὸς ἐπισταμένως κατέλεξας (λ 368). Eumaios praises him too as a skillful singer with a long tale (ρ 514–521).
[ back ] 136. λ 328–330, cf. Β 488; ρ 515–517.
[ back ] 137. S. West ad δ 836 observes that the “normal meaning [of διηνεκέως] is ‘continuously, from beginning to end’, but here it must be something like ‘positively, explicitly’; a familiar formula … has been carelessly used” (Heubeck et al. 1988:244). Not so: whether Odysseus is living or dead makes for a notional narrative end-point. One need only remember that, in the judgment of many, both ancient and modern, Penelope’s recognition that her husband had ideed returned was (or should have been) the proper τέλος of the Odyssey (see Russo et al. 1992:342–345 ad ψ 296).
[ back ] 138. In various publications, Nagy has championed the study of relay performance and relay poetics, which he has variously called “the esthetics of rhapsodic sequencing” (Nagy 1996c:71) or, more simply, “relay mnemonics” (Nagy 2002:123). Nagy (2002:36–69) provides an incisive exploration of this matter through Plato’s writings in the context of the classical Panathenaic festival.
[ back ] 139. Hesykhios Τ no. 796: τὴν ἐπιδεξιάν (-δέξια Reitz.)· περιέφερον ἐν τοῖς συμποσίοις ἐπὶ δεξιὰ τὸ πάλαι κιθάραν, εἶτα μυῤῥίνην, πρὸς ἣν ᾖδον. Cf. the scholia to Aristophanes Wasps 1222a (p. 192–193 Koster); the scholia to Plato Gorgias 451e; Athenaios 15.693f–694c (vol. 3, p. 535–536 §49 Kaibel); and Plutarch Quaestiones convivales 1.1.5 (615a–c). Most of these can be found in Campbell 1993:270–279.
[ back ] 140. Campbell 1993:271n2 writes: “Lit. ‘receiving’; he was expected to continue or to cap the line.” Cf. the DGE s.v. δέχομαι A.Ⅲ.4: “recibir y seguir, recoger, continuar cantos conviviales.”
[ back ] 141. δεξάμενος δὲ Σωκράτης τὴν ἐπιδέξι’ ⟨ἄιδων⟩ | Στησιχόρου πρὸς τὴν λύραν οἰνοχόην ἔκλεψεν. To δεξάμενος, the editors note: “δέχεσθαι cantu excipere.”
[ back ] 142. We do not know that a staff was passed on from bard to bard to signal whose turn it was to take up the song. But it is indisputable that the ῥάβδος was thought a badge of authoritative singing (cf. Hesiod Theogony 30 and Pindar’s κατὰ ῥάβδον at Isthmian 3/4 56–57) and that it was closely associated with performance (see above, §10.2.2).
[ back ] 143. Β 278–279 Γ 216–219 Σ 503–505 β 37–38 (cf. Β 185–186).
[ back ] 144. Cf. Κ 321 328 β 80.
[ back ] 145. For the acceptation ‘to take up’, what the LfgE s.v. I.2 glosses as “jem.m etw. abnehmen, etw. (von jem. in Funktion, Auftrag) übernehmen, um es zu bearbeiten, aufzubewahren, weiterzugeben usw.,” see Β 186 Ε 227 (=Ρ 480) Ρ 208 391 α 121 (~ο 282 π 40) ν 72 ο 132 φ 82.
[ back ] 146. So Leaf 1900–1902: “expecting, not ‘having received’” (1.477, his emphasis); and Ameis and Hentze 1906: “erwartend” (50), both commenting on Λ 124. The Homeric loci are Δ 107 Θ 296 Κ 62 Ο 745 Ψ 273 (cf. β 186 and LfgE s.v. B.I.5–6 and B.Ⅱ.5). See Debrunner 1956–1957, esp. 79–80, who resorts to the notion of “intensive Perfekta” (on which cf. K-G Ⅱ.1.148–150 §384.4).
[ back ] 147. A corresponding translation might be: ‘Waiting [for my turn to take up the song] I sing the story upon the staff woven in unbroken sequence’.
[ back ] 148. See Nagy 1996c:71–73.
[ back ] 149. Scholars are divided whether this is a perfect (Leaf 1900–1902:1.109 ad Β 794), a present (Hainsworth 1993:88, Schwyzer GG I.678 §V.1a.α6), or an aorist participle (Frisk 1973–1979 and Beekes 2010 s.v. δέχομαι). Cf. Chantraine 1999:256 (with further bibliography).
[ back ] 150. Images of performers seemingly processing or even standing while playing a string instrument of the family of the lyre (cf. West 1992:50) are not rare. But it is not difficult to find some that depict them in the sitting position. Cf. Paquette (1984:84–185, especially 99): “On joue de la cithare debout, plus rarement assis” (C22, C45, C47, and C49 are offered as examples of the latter). In the concert hall one might expect the performer to mount a platform and stand while playing; in the more intimate setting of Akhilleus’ tent, the sitting position is more suitable and relaxed (Ι 193–194). Cf. West 1992:64 and Zschätzsch 2002:29–44, 182–184.
[ back ] 151. Cf. 190b Erbse: “He means either that he alone was sitting or that he alone was opposite him as compared with the rest of the Myrmidons.”
[ back ] 152. Metapoetically speaking, as Akhilleus’ alter ego Patroklos almost steals the show at the expense of his leader when he dons Akhilleus’ armor and joins the fray. Akhilleus in turn proves jealous for his own honor when he considers the consequences of Patroklos’ involvement: ‘If the loud-thundering husband of Hera grants you to win glory, do not yourself apart from me be eager to fight against the war-loving Trojans—you will make me more dishonored’ (Π 87–90; cf. Π 269–274). Patroklos’ central role in the plot as Akhilleus’ stand-in represents a narrative hypolepsis of sorts. What Patroklos is ready to do at the end of Akhilleus’ performance but does not do because the arrival of the embassy interrupts the singing, he actually performs later at the level of the narrative. Then he takes not his master’s phorminx but his armor, and re-activates the κλέα ἀνδρῶν not in narrated song but in narrated deeds.
[ back ] 153. Perceau (2005:71) fails to grapple with the otherwise puzzling character of this important detail. Her citing of θ 87 is not apropos, for the Phaiakians are not portrayed as waiting for the end of the song. The text simply tells us that ‘whenever the divine singer would stop singing’ Odysseus would dry his tears and draw the cloak off his head. The optative here is iterative, and the ὅτε clause equivalent to a past general condition (‘as often as he would’ ~ ‘if ever he would’). Not so at Ι 191, where the waiting focuses on an anticipated one-time event; that is, λήξειεν is not iterative, but merely denotes time subsequent to that of the principal verb in secondary sequence (this is what Chantraine GH Ⅱ.260 §382 calls “valeur éventuelle,” to compare with “optatifs de répétition”). Perceau does not appreciate the puzzle because she has convinced herself that Ι 191 merely underlines the attentive manner of the singing entailed by σιωπῇ—only, we learn now that the respectful hearing does not extend beyond the end of the song. If she is right, the verse conveys a trite and predictable fact, and we are left without a rationale for the focus on Patroklos or for the curious manner of his singling out. Elsewhere, when waiting for a given happening is mentioned (δέχεσθαι + ὁπ(π)ότε), the context makes clear what specific actions were expected to follow right upon the happening. For example, in order to warn the Trojans of the marching of the Akhaians, Iris impersonates Polites, the Trojan sentinel who sat on the tomb of Aisyetes ‘waiting for when the Akhaians would sally forth from their ships’ (Β 794); it is clear that he was then to rush back and report the fact, which Iris does for him (cf. Η 415 Σ 524 υ 385–386). Not so here, unless we assume that Patroklos relieves Akhilleus in his singing, an inference commended both by the context and by the nature of epic performance in the culture of archaic and classical Greece.
[ back ] 154. For the interchangeability of audience and epic characters as a characteristic Homeric device, cf. Nagy 1996c:72n37 with bibliography.
[ back ] 155. Syntactically, Αἰακίδην has been extracted from the ὁπότε clause, where it would ordinarily sit, and placed into the main clause. Semantically, the object of δέγμενος is not ‘the son of Aiakos’, but ‘when the son of Aiakos would leave off singing’. Leaf 1900–1902:1.385 ad loc. writes: “Αἰακίδην is taken proleptically from the rel. clause.”
[ back ] 156. This is generally recognized by scholars. So, for example, Monro 1958: “δέγμενος, ‘waiting,’ apparently to take up the song: so the Muses sang ἀμειβόμεναι (1.603)” (344); Ameis and Hentze 1930:94 note: “δέγμενος Αἰακίδην gewärtig des Aiakiden, wohl um ihn im Gesange abzulösen: vgl. Α 604”; and Hainsworth 1993: “Patroklos is simply listening to Akhilleus, perhaps with the implication that he would take up the song at the point where Akhilleus left off” (88).
[ back ] 157. She refers here to Hainsworth 1993: “There are many ways of performing heroic song, including the employment of two singers, but the only one described in Homer is that of the solo singer” (88).
[ back ] 158. I cannot, however, agree with Tarditi’s rejection of the inherently competitive character of rhapsodic singing, or with his notion that the close sequencing entailed by οἴμη was incompatible with the festival ἀγών (1968, 141–142). Tarditi bases his views on an excessively rigid notion of rhapsodic sequence that would not survive the addition of a hymnic proem to the epic performance. Whether hymns always preceded an individual rhapsode’s performance, even when he followed another in relay singing, cannot be assumed without argument, as Tarditi does. But even if this were so, I dispute the notion that the hymn would necessarily be felt by performer and audience as an interruption that spoiled the narrative and performative sequence. One must reckon here with the techniques of expansion and contraction, which would allow the rhapsode to create a seam that met the pragmatic demands of his epic delivery. The peculiar genre and function of the hymn should also be taken into account in judging what sort of ‘break,’ if any, it might have effected in hypoleptic singing. One must also remember that the so-called Homeric hymns were preserved as their own distinct corpus, and not as openings to manuscripts of the Homeric poems. In a future work I hope to explore the implications of this fact for our understanding of these hymns in the context of the Panathenaia.
[ back ] 159. Lit., ‘to render true by an authoritative speech-act’. On the meaning of this verb in this passage, see further above, §7.2.1.
[ back ] 160. Amazed at the θέσπιν ἀοιδήν (442) and θαυμασίην … νεήφατον ὄσσαν (443)—presumably the combination of Hermes’ singing and the ‘voice’ of the lyre—Apollo asks: τίς τέχνη, τίς μοῦσα ἀμηχανέων μελεδώνων, | τίς τρίβος; (447–448). At first, dancing does not seem to be in view. Hermes ‘stands’ by Apollo when he performs (στῆ 424). But Apollo identifies himself as a follower of the Olympian Muses, who care about ‘dancing and the splendid path of song, and lively molpē and the lovely bray of auloi’ (τῇσι χοροί τε μέλουσι καὶ ἀγλαὸς οἶμος ἀοιδῆς | καὶ μολπὴ τεθαλυῖα καὶ ἱμερόεις βρόμος αὐλῶν 451–452). Not only does μολπή ordinarily involve a combination of singing, instrumental playing, and dancing, but χοροί explicitly refers to dancing (whether by the performer himself or by others to his music or both). On οἶμος, for which I have used here the traditional rendering ‘path’, see further below, §
[ back ] 161. It is an error narrowly to supply κίθαριν as the sole direct object of δέγμενος. The verb does not have an explicit object, and the context commends the verbal notion κιθαρίζειν or the abstract τέχνη as that which Apollo so earnestly desires. To be sure, there is wonder at the instrument Hermes has devised (443), but the focus of Apollo’s words are the entire art of κιθαρῳδία—playing the instrument and singing to it. He mentions Hermes’ θέσπις ἀοιδή (442); the lyre’s ὄσσα (443), described in 444 as something learned; the τέχνη (447), μοῦσα (447), and τρίβος (448) of Hermes’ performance; his κιθαρίζειν (455) and μήδεα (456). Some of these may readily include the lyre as a cunning artifact, but they are deliberately broader in scope and embrace playing and singing. Hermes in turn mentions his τέχνη (465) and κιθαρίζειν (475). (The ἔργα at 440 are probably Hermes’ ‘accomplishments’, not just the lyre, and δῶρον at 442 could be the lyre or, just as well, the skill necessary to devise it and play it well.) This is the context from which a direct object would have to be supplied to δέγμενος, if we must. If so, only the broader τέχνη of κιθαρῳδία will do. But I believe that the reason δέγμενος does not take an explicit direct object and the context does not readily supply one is because relay performance is in view, an acceptation that deploys δέχομαι absolutely (i.e. without an object). What Apollo is to receive and take up is the performance, the singing of the song. This entails, but is not limited to, receiving the lyre, a gesture that is otherwise commended by the archetypal setting. The case of 496 is different. There ὤρεξε and ἐδέξατο have no explicit direct object either, but the preceding context readily supplies the ἑταίρη (478) (cf. μιν 480, if one accepts Ilgen’s emendation [Ilgen 1796:41 and 461]; αὐτήν 482 and 486; and ταύτην 490) as the object held out by Hermes and received by Apollo (the κίθαρις of 499).
[ back ] 162. This hypoleptic transfer, central to the meaning of the hymn, would be contextually reinforced by the reference in 454 to “the passing to the right at young men’s feasts” (so West 2003b:149)—i.e. by the reference to the hypoleptic performance practice at the symposium—if in fact this is the correct translation of this line.
[ back ] 163. Cf. οἶμος ἀοιδῆς in the Hymn to Hermes 451.
[ back ] 164. Cf. τρίβος in the Hymn to Hermes 448.
[ back ] 165. Cf. DGE s.v. εἱρμός 3.
[ back ] 166. Cf. DGE s.v. εἷς I.2g.
[ back ] 167. For ἔπειμι in the sense of ‘to traverse’, see TGL 3.1460–1461 and LSJ s.v. Ⅲ. Relevant loci are: Lucian Hermotimos §1.10 MacLeod (ἐπιὼν τῇ μνήμῃ ἕκαστα); Aristophanes Frogs 897 (ἔπιτε δαΐαν ὁδόν); Plutarch Periklēs 17.2 (τοὺς ἐν Ἑλλησπόντῳ καὶ Θρᾴκῃ μέχρι Βυζαντίου τόπους ἐπῄεσαν); Heliodoros Aithiopika 2.6.2 (ἐπειρᾶτό τι τῶν ἐγγεγραμμένων ἐπιέναι); Strabo–28 ([οἱ ποταμοὶ] οἱ πολλήν τε καὶ μαλακόγειον χώραν ἐπιόντες); δ 411; and perhaps Herodotos 5.74. Cf. LSJ s.v. ἐπέρχομαι Ⅲ.1–2 and Herodotos 1.30.2.
[ back ] 168. For a somewhat different translation, see Nagy 1996c:67–68. Eustathios in Ad Iliadem offers a helpful reading of this scholion (van der Valk 1971–1987:1.10). By ἑκατέρας τῆς ποιήσεως, he understands the complete poems: ποίησις μὲν γὰρ ἡ ὅλη βίβλος (1.10–11). The scholion’s second interpretation assumes the rhapsodes’ dependence on notionally integral poems. Eustathios writes: ἢ καὶ ἄλλως, διότι κατὰ μέρος, φασί, τῆς ποιήσεως διαδεδομένης τὴν σύμπασαν ποίησιν ἐπιόντες οἱ ᾄδοντες καὶ τὰ ἐξ ἑκατέρας Ὁμηρικῆς βίβλου συρράπτοντες, ὡς ἐβούλοντο, ῥαψῳδοὶ ἐντεῦθεν προσηγορεύθησαν. The scholiast obviously believes that a notional sequence was established with the introduction of the poems, a sequence that stitching rhapsodes were to observe. Unless Eustathios has badly misunderstood the scholion at this one point, in writing ὡς ἐβούλοντο he cannot have had complete freedom of choice in view; only what was consistent with the poems as introduced, namely, the rhapsode’s choice of a starting point for his performance. Going through Homer’s poetry in performance, rhapsodes would move from one section to the next as if ‘mending’ the poetic fabric by respecting and reenacting in song the established connections (or ‘seams’) between sections. The metaphor of ‘mending’ served to contrast this performance practice with the one that relied on a notionally rent (κατὰ μέρος) poetic fabric. καὶ τὴν σύμπασαν ποίησιν ἐπιόντας need not imply comprehensive performances of the entire poems. The scholion describes the general practice of rhapsodes, not the specific practice at a given festival (even if one accepts its connection to Argive traditions). To clarify the distinction I am drawing here, it is helpful to consider a parallel: reading through virtually all the Bible over time as practiced in the liturgy of the Roman Catholic Church. This practice aims not at reading the entire biblical text at once but substantially to cover the canonical scriptures within a five-year cycle. The tendency towards comprehensiveness of scope characteristic of Homeric rhapsodic sequencing, the rhapsodic techniques of thematic expansion and contraction, and the traditional referentiality that marks the epic medium—all these rendered any one rhapsodic performance notionally homologous with performing the entire tradition. Because the rhapsode evoked the entire sequence by his adherence to whichever part of the traditional plot concerned his performance, there was as little need as there was little opportunity to perform the entire Iliad and Odyssey at any one festival.
[ back ] 169. Cf. Collins (2004:185–191). Else (1957b:30) had insisted that the label ῥαψῳδός should come from “a concrete occasion on which the ‘sewing’ would be overt, unmistakable. Such an occasion might be provided by the rhapsodes’ contests.” Else was anticipated by Pagliaro (1951:43–45), whose own important treatment exhibited greater diachronic insight. Taking up Else’s suggestion, Ritoók (1962:228–229) drew attention to the Certamen. He also conjectured on the basis of Pindar Nemean 2 that hypoleptic performance was only characteristic of the Homeridai (230). There is truth in Else’s and Ritook’s proposals, but they are wrong to narrow the characteristic craft of the rhapsode to competitive hypolepsis.
[ back ] 170. For the initial suggestion, see Nietzsche 1870 and 1873. Relevant scholarship, not all in support of Nietzsche’s hypothesis: Abramowicz 1938, Vogt 1959, Hess 1960 with Vogt 1961, Vogt 1962, West 1967a, Richardson 1981, Heldmann 1982, Avezzù 1982:84–87, Kawasaki 1985, O’Sullivan 1992:63–105, and Mariß 2002:21–24. For the controversy about the relevant papyrological finds, see the bibliography cited by O’Sullivan 1992:63–64.
[ back ] 171. See, for example, Alkidamas On the Sophists §14. More on Alkidamas above, §9.5. Richardson (1981:5) writes: “It has often been noted that the theme of improvisation in the Certamen forms a strong link with Alcidamas’ On Sophists … . Alcidamas seems to have been interested in the story of the contest … because it gave prominence to the value of improvisation … .”
[ back ] 172. Cf. Collins 2004:184.
[ back ] 173. I cite the line numbers of Allen’s OCT edition. Wilamowitz-Moellendorff 1929:34–47 and Avezzù 1982:38–51 should also be consulted for the text. West 2003b:318–353 provides a convenient translation.
[ back ] 174. Collins 2004:177 calls this “the ‘epic’ part of the Certamen.”
[ back ] 175. ‘[Hesiod] turned to asking conumdrums’ (ἐπὶ τὴν τῶν ἀπόρων ὥρμησεν ἐπερώτησιν 95) suggests that what follows deliberately frames the ‘ambivalent propositions’ (τὰς ἀμφιβόλους γνώμας 102–103). But whereas the preceding challenges had involved Hesiod’s querying Homer, the anti-invocation cannot be considered a ‘question’ in the ordinary sense of the word, even if arguably it seeks to put in question Homer’s skill. Instead, it is in character very much like the ambivalent one-liners, except that it does not call for enjambment—only for a sequence of thought that validates and builds upon what seems unmanageable nonsense.
[ back ] 176. The Flinders Petrie papyrus no. 25 (see Allen’s edition, p. 225, lines 41–44) merely picks up the language of 95 and is less illuminating.
[ back ] 177. Long ago argued by Koller 1956:177. Mirroring the hymnic use of ἄλλος, ὁ δ’ ἄλλος λόγος in [Demosthenes] 61 §2 also means ‘the rest of the speech’.
[ back ] 178. For further analysis, cf. Collins 2004:186–191.
[ back ] 179. For these and other testimonia, see Davison 1955a. Shapiro 1998 offers a fairly traditional reconstruction of the historical context for the Panathenaic rule and its impact on sixth-century BC rhapsodic performance. He pays particular attention to the iconographic evidence from vases.
[ back ] 180. Cf. LSJ s.v. I.3.
[ back ] 181. Ὁμηρίδας ἔλεγον τὸ μὲν ἀρχαῖον τοὺς ἀπὸ τοῦ Ὁμήρου γένους, οἳ καὶ τὴν ποίησιν αὐτοῦ ἐκ διαδοχῆς ᾖδον (scholion 1c to Nemean 2). Although one cannot be dogmatic, on balance the diachronic paradosis seems in view here, pace Collins 2004:183n9.
[ back ] 182. So also Boyd 1994:115. Nagy (2002:10) seemingly applies ἐφεξῆς to what he translates as the “poetic utterances of Homer,” for he quotes as parallels Plato Timaios 23d3–4 and 24a1–2, which involve only one narrator (cf. 15n23).
[ back ] 183. If the point was to prevent something like the unpredictable passing of the baton at a symposium, ἐφεξῆς was hardly the word to use. Only against an established pattern can the otherwise ambiguous ‘in sequence’ suffice. To the demand, ‘in sequence’, the obvious question is, ‘in what sequence’? At a symposium, the arrangement of couches produced a natural sequence against which ἑξῆς would make sense.
[ back ] 184. As Boyd (1994:115) observes, neither account of the rule requires a written text to attest to the traditional sequence. An oral telling of sufficient stability, at least in regard to the order of its episodes, would suffice. The degree to which the prevailing sequence of a received plot might descend to narrative minutiae would have heightened with time.
[ back ] 185. Cf. Nitzsch 1828:28–40; Boeckh 1858–1874:4.385–396; Hermann 1827–1877:5.300–311; Boeckh’s commentary to CIG 3088 (vol. 2.675–678); and Hermann 1827–1877:7.65–87. The footnote in Boeckh 1858–1874:4.385 offers a short history of the controversy up to 1874. See also Wilamowitz-Moellendorff 1884:265–266, who assumes that Diogenes Laertios (or his source) used ὑποβολή in error.
[ back ] 186. Cf., for example, Davison 1955a:11.
[ back ] 187. TGL 5.1830 s.v. οἷον: “Quinetiam redditur ‘id est’ … . Sic saepe grammatici.” Cf. LSJ s.v. οἷος 5.2.e. This obvious point is not usually observed by modern translations. Davison (1955a:11) correctly renders the clause, “viz. that the succeeding (reciter) should begin where the first left off.”
[ back ] 188. It begs the question to speculate that Agamemnon had more to say but Akhilleus prevents him from speaking on. Nothing in the context demands this view. Its sole basis would be the alleged meaning of ὑποβλήδην one is seeking to establish!
[ back ] 189. I do not restrict ‘prelude’ to ‘instrumental prelude’. Both singing and instrumental playing are wedded in the epic ἀναβολή. Because Pagliaro (1951:31) accepts the narrow sense of prelude, i.e. an instrumental opening that precedes the song, he must reject the connection of ἀναβολή with ἀναβάλλομαι ἀείδειν. No such rejection is necessary, once we consider the diachronic implications of the use of κιθαρῳδία to portray epic performance in the Odyssey. For instances beyond Homer of ἀναβάλλομαι as a technical word for performance, see Theokritos 6.20: ἀνεβάλλετο καὶ τάδ’ ἄειδεν (abs.); Pindar Nemean 7.77: εἴρειν στεφάνους ἐλαφρόν, ἀναβάλεο· (abs.); Aristophanes Peace 1269: αὐτοῦ παρ’ ἐμὲ στὰν πρότερον ἀναβαλοῦ ’νθαδί (sc. ᾄδειν). See further Nonnos Dionysiaka 24.242 (with ἀείδειν) and 19.102 (with ἁρμονίην); and Philostratos Imagines 1.29. For the corresponding noun ἀναβολή, see Pindar Pythian 1.4–5 and the other loci in the DGE s.v. A.Ⅳ.
[ back ] 190. Egan 2007 claims that, in poetic contexts, “ἀμβολάδην means ‘by way of resumption or sequel, after an interval, subsequently’ and ἀναβάλλομαι ‘[to] resume, [to] begin again after an interval’” (55). Since the function of epic proems was to furnish epic performance with a ritually sanctioned beginning, it would hardly be surprising if Homeric (and other epic) narrative should sometimes explicitly feature proems and proemial language. In fact, this book amply demonstrates that in their epic poetry rhapsodes often drew attention to their own performance medium. This chapter surveys many striking signs of this metapoetic tendency. Not the least of them are instances in which technical rhapsodic terms shaped epic diction. I submit that these include uses of ἀμβολάδην and ἀναβάλλομαι in performance contexts, terms that I believe to be proemial in origin (i.e. they regard proems in musical performance, with ‘music’ broadly understood). Egan, however, does not think their primordial poetic usage proemial: for him they denote, on the one hand, ‘resumption’; on the other, ‘proem’ as a specialized application of the common meaning ‘delay’, viz. ‘proem [that postpones the subsequent phase]’. Although the relationship between these two senses is unclear at first, eventually Egan also subsumes the acceptation ‘to resume’ under the notion of “delayed or postponed performance” (61). It is plain, then, that for him ‘delay’ is the fons et origo whence the various acceptations issue. Not so. The semantics of ἀναβάλλομαι is to be elucidated otherwise. Both meanings, ‘to perform a proem’ and ‘to delay’, derive from the local one ‘to cast up’ by semantic extension: the former, from ‘to cast up [the voice in song]’ vel sim.; the latter, from ‘to cast [further] along’ (or ‘up’, with ‘up’ notionally associated with ‘later’ for reasons we can only conjecture; cf. ὄπι, the o-grade of ἔπι, in ὄπισθεν and ὀψέ, on which see Morpurgo Davies 1983). In my view, then, ἀναβάλλομαι ‘to delay’ and ἀναβάλλομαι ‘to strike up’ (=‘to begin to sing/play’) are independent polysemes, and neither can be reduced to the other. Egan’s explanation of the proemial use of this word family—namely, that it comes from viewing the proem as the postponement of the rest of the performance—is unconvincing. Nothing in the ancient Greek song culture would encourage us to accept that a word that meant ex hypothesi ‘to postpone’ was embraced as characteristically proemial diction. It is implausible that either rhapsodes or their audiences should have likened the marked performative beginning of rhapsodic performance to a delaying utterance, much less characteristically identified it as such; nor, by extension, to view the ensuing performance as a “delayed utterance” (56; I do not quarrel with Egan’s “fresh start,” even though ‘fresh’ remains too vague to help his argument). This contention alone is so far-fetched that it makes Egan’s revisionary thesis inadmissible. Furthermore, to prove the alternative ‘to resume after an interval’, Egan needs to do more than showcase (what I claim is) proemial language featuring at the beginning (and as the beginning) of a performance that follows some other performance; he must also establish that the notion of ‘resumption’ or ‘sequel’, which requires an underlying continuity, obtains. Otherwise, his claim is unfalsifiably trivial: for, if any performance will do, identifying some earlier one that the present allegedly resumes is a trivial task. Such a light burden can hardly be considered probative. The onset of his argument is not reassuring. Egan cites a Pindaric scholion that resorts to the polyseme ‘to delay’ as one of several alternative interpretations (hardly the most compelling) of ἀμβολάδαν in Nemean 10.33: ‘he won the Panathenaia twice, not in succession but anaboladēn, i.e. with some delay intervening’. His other opening move amounts to a misreading of the scholion to Χ 476, which makes a thematic, not a performative, point—a point, moreover, entirely compatible with the performative sense of ἀμβλήδην (‘in proemial fashion’ or ‘with an overture’, see below). This term, the scholiast suggests, points to the fact that Andromakhe starts her mourning not with the current circumstances (Hektor’s death and the dragging of his body) but with her own birth in the house of Eëtion (προοίμια was sometimes used generally for ‘beginnings’; cf. LSJ s.v. I.2). I do not believe that the scholiast is right, apropos, or compelling. Like Egan, he has failed to take the measure of the performative setting and to see at play the metapoetic tendency that is the signature of just such Homeric contexts. All the same, on its own terms the scholiast’s point is clear and intelligible. Egan might argue that ἀμβλήδην picks up on the sequence ‘first Hekabe and then Andromakhe’ (57), but he will not find any support for this in the language of the scholion (the required performance sequence can hardly be extracted from ‘not from the present circumstances’). But if Egan’s opening examples fail to make his case, the notion that ἀναβάλλεσθαι means ‘to resume [singing] after an interval’ (strictly speaking still possible for Χ 476) cannot be sustained for θ 266. For even if one could prove that Demodokos’ Song of Ares and Aphrodite and his earlier performances share an overall notional unity that justifies rendering ἀνεβάλλετο … ἀείδειν at θ 266 as “he ma[de] a new start to his singing” (57, with emphasis on ‘new’, for I do not quarrel with ‘he started singing’), the argument stumbles on the language of θ 90: ὅτ’ ἂψ ἄρχοιτο καὶ ὀτρύνειαν ἀείδειν refers to an extended series of performances that led the rhapsode repeatedly to cease his singing, and his audience repeatedly to urge him to continue—with Odysseus having the time and opportunity during each break to dry his tears, to pull off the cloak, and to pour libations to the gods. It is clear that, but for the first time, this earlier sequence of performances offered ample occasion for Egan’s notion of resuming one’s singing after an interval, but ἀναβάλλεσθαι or its equivalent fails to materialize to make the point—unless Egan intends ‘new’ in “new start” to mark Demodokos’ singing about Aphrodite’s adultery not as a ‘sequel’ but as a thematically unconnected song. If so, Egan should exclude ‘sequel’ and ‘to resume’ from his definition or else must water it down to a bare ‘do again what has been done before in some form’ (whether the subject is one and the same and the actions stand in continuity or discontinuity). I doubt that such a weak sense is true or even useful. (Comparable refuting arguments can be made about the other loci discussed by Egan, including his bizarre interpretation of Arkhimedes’ Problema bovinum 37–38.)
[ back ] 191. The stricter sense would be ‘[to sing or play] in the manner of a prelude’ (=‘to sing or play a prelude’).
[ back ] 192. Or ‘playing the lyre he started singing’.
[ back ] 193. Or ‘he sang and played a prelude on the lyre’.
[ back ] 194. Here Agamemnon does not interrupt Akhilleus at all. Nor do the Akhaians, whose expressions of joy (Τ 74) follow what is obviously the end of Akhilleus’ address.
[ back ] 195. That in Τ 80 we find not the middle but the active ὑββάλλειν is no object, for the potential interchangeability of middle and intransitive active verbal forms is well documented in Homeric diction and specifically exampled for συμβάλλ-ω/-ομαι (cf. Π 565 Φ 578 φ 15 and Μ 377 Υ 335 etc.; see, further, LfgE s.v. βάλλω Ⅲ.1e and 2e). Chantraine writes: “Il arrive parfois, comme nous venons de le supposer pour οὐτάμεναι que les formes actives soient constituées parallèlement à un thème généralement moyen. De la racine de βάλλω il existe un aoriste βλῆτο ‘il a été blessé’ qui est très bien attesté … , βλήμενος … . Mais il existe deux formes à désinences actives qui présentent aussi un sens intransitif ‘se rencontrer, attaquer’: ξυμβλήμεναι … et ξυμβλήτην ἀλλήλοιιν … l’emploi absolu de formes actives au sens intransitif peut être ancien” (GH I.380). If ever the verb ὑποβάλλ- (not the noun ὑποβολή or the adverb ὑποβλήδην) was later borrowed in the Homeric sense, we should expect that its sole surviving occurrence at Τ 80 in (what I argue is merely by accident and indifferently) an intransitive active form (i.e. but for the meter, it might have been a middle) would commend its reappearing in the active. Cf. τ 584 with B 436.
[ back ] 196. Cf. Monro 1897: “ἀνεβάλλετο καλὸν ἀείδειν = ‘began the song’; so also ὑποβλήδην ‘taking up,’ ‘interrupting’” (395).
[ back ] 197. Although I do not agree with his conclusion on ὑποκρίνομαι or embrace all of his analysis (see below, Appendix), the general point is sound and incontrovertible.
[ back ] 198. This immediacy is key to Curtius’s attempt to motivate the label ὑποκριτής for the tragic actor (Curtius 1866). Although I do not think that his particular application of sequential immediacy to drama is tenable (see below, Appendix) or that the sense of ὑπό was the same in ὑποβάλλω and ὑποκρίνομαι, I do believe that the implication of sequence in the ὑπό of ὑποβολή, projected onto ὑποκριτής, would have encouraged the survival of this agent noun in connection with rhapsodic performance.
[ back ] 199. With these considerations in view, it is clear that Nagy (2002:21) is right to focus on “picking up the train of thought” as essential to its meaning, pace Schmidt in the LfgE s.v. ὑποβλήδην B.1.
[ back ] 200. Cf. Nagy 2002:19.
[ back ] 201. Collins (2004:195) and Nagy (2002:15n23) both disagree. Such supervision may have been more flexible than Boyd would allow for and need not have involved clocked performances (cf. Boyd 1994:115n16).
[ back ] 202. Just to mention one thing that is yet to receive significant thought: how did the Panathenaic officials make sure that the first rhapsode did not have an advantage over the succeeding ones, who unlike him were forced to take up the plot at a particular point and had to face the consequent challenge of forming a suitable narrative seam? To my knowledge, only A. Boeckh has mentioned this difficulty, touching upon it in passing during his discussion of ὑποβολή in CIG Ⅱ.676 ad no. 3088.
[ back ] 203. Collins (2004:195) points to living oral traditions, which show how one singer can offer oral cues to another with ease and virtuosity. Thus are links created—of subject matter, diction, etc.—which contribute to tighter dynamics of sequencing. Collins has in mind primarily a combination of proactive and reactive performance practices. In the case of rhapsodic delivery, ‘reaction’ is chiefly in view, although given the inherently competitive character of such delivery—and the competitive character of the speeches by epic speakers (see immediately below)—we should not discount that, at least in part, one rhapsode may have shaped his performance with a view to constraining the reaction of his successor. By extension, that too would fall under the scope of ἐξ ὑποβολῆς.
[ back ] 204. Cf. Pagliaro 1951:34 and Collins 2004:196.
[ back ] 205. See especially his chapter “Heroes as Performers” (Martin 1989:89–145).
[ back ] 206. This sentiment is ironic, because Athena has, in fact, commanded Akhilleus not to strike Agamemnon with the sword and only to revile him (Α 211).
[ back ] 207. In this case it is Agamemnon, not Akhilleus, who finds the Akhaians’ response tactless. The complementary ‘hearing well’ and ‘not reacting’ in the manner deprecated stand in continuity as elements of proper etiquette. An improper reaction might involve, but is hardly limited to, interrupting the speaker. It is doubtful that heroic assemblies live up to the standard articulated here by Agamemnon. But this, of course, is not the point of his self-serving reprimand.
[ back ] 208. As noted above, Akhilleus’ speech reaches an obvious endpoint at Τ 73 and, stricto sensu, the Akhaians’ rejoicing does not interrupt it. Agamemnon’s rebuke addresses in general terms proper behavior at an assembly (εἰς ἀγορήν Τ 45), where participants are to speak and to listen by turns. The rowdy reaction of the audience (ἐν πολλῷ ὁμάδῳ Τ 81) might disrupt the carefully calibrated pragmatics on which rhetorical persuasion often depends, and hence threaten its social objectives. Such unfavorable circumstances might inhibit proper hearing and speaking and strain even a knowledgeable (ἐπισταμένῳ περ ἐόντι Τ 80) and clear-voiced (λιγύς περ ἐὼν Τ 82) speaker. One should recall that λιγύς and λιγέως are regularly used to qualify metonymically the eloquent and rhetorically effective speaker (e.g. at Γ 214). Hence, λιγύς is not merely a matter of tone or volume (so, correctly, LfgE s.v. B.1b). Cf. Martin 1989:117 and λιγύφθογγος, the epithet of heralds.
[ back ] 209. For an alternative interpretation, cf. Nagy 2002:21.
[ back ] 210. This old idea is explored at length by Martin 2000. See further below, §10.2.6 n. 345.
[ back ] 211. Cf. Lazzeroni 1967:53–55.
[ back ] 212. From *ἱμά derives the secondary formation in -ντ- ἱμάς ‘leather strap’ (Schwyzer GG I.526n5; cf. Beekes 2010:1.590 for the quantity of the anlauting iota). Durante (1976:176 §76) had already connected οἴμη and ἱμάς.
[ back ] 213. So Lentz 1867–1870:1.546.16–17 (τὸ οἷμος οἱμῶ δασύνεται). Cf. φροίμιον (LSJ s.v. προοίμιον) and ἄοιμος· ἄπορος (Hesykhios Α no. 5665).
[ back ] 214. See Hainsworth 1993:218–219.
[ back ] 215. Cf. DMic Ⅱ.24 s.v. “o-mo-pi.”
[ back ] 216. For a novel exploration of this motif, see Giannisi 2006. Conceptualizing the performance of song as walking along a path explains the ready use of μεταβαίνω for ‘to move along’ in the performance (θ 492; Homeric Hymns 5.293, 9.9, 18.11). Similar spatial notions may underlie the transitive use of ἐπιβαίνω in Hesiod Works and Days 659 for ‘to initiate’ as a performer (‘where [the Muses] first caused me to step [mount? embark?] on clear-sounding song’; cf. Hymn to Hermes 464–465). But the parallel (intransitive) use with abstract nouns like ὁσίη (Hymn to Hermes 173) and ἀναιδείη (χ 424) makes the precise nature of the metaphor elusive. See Ford 1992:41–43.
[ back ] 217. On ὕμνος, see further below, §
[ back ] 218. For the transposition of the -h- see Lejeune 1972:138 §133.
[ back ] 219. Osthoff 1898:171. Beekes (2010:2.1058 s.v. οἶμος) suggests that the hiatus δέκα οἶμοι in Λ 24 (τοῦ δ’ ἤτοι δέκα οἶμοι …) “requires [οἶμος with] initial ϝ-.” But this is no more necessary here than in Β 87 (ἠΰτε ἔθνεα εἶσι …), where the hiatus ἔθνεα εἶσι hardly justifies positing initial ϝ- for εἶσι.
[ back ] 220. All these and a few more in Osthoff 1898:163. Whether such forms differ at all one from another in meaning and to what degree is not a priori predictable.
[ back ] 221. Cf. Gagnepain 1959:57–104, with the pair οἶμος/οἴμη at 85–86.
[ back ] 222. Cf. Wackernagel 1920–1924:2.12–15.
[ back ] 223. The confusion is well chronicled by Osthoff 1898:163–164.
[ back ] 224. Γ 151–152; Hesiod Works and Days 582–584, Shield 393–396.
[ back ] 225. Ameis et al. 1862:59 and 69 (the pagination noted starts anew with Aratos’ Phainomena).
[ back ] 226. οἶμαι of αἰνίγματα, ‘lays of dark oracular statements’ that were delivered of old in epic verse, following conventions of recomposition in performance. With mantic and oracular poetry, epic shares its cultural beginnings (a matter explored at length above in chapter 8). Like the scholiast, Osthoff 1898:162 (and, with him, many others in recent times) thought that οἶμαι here referred to “‘gänge, pfade, weisen’ der dunklen reden,” adducing τρίβος (11) and κελεύθῳ (12) in the same sentence. I do not deny that οἴμας plays on οἵμους; hence διοίχνει … οἴμας (10–11). But these are δυσφάτους … οἴμας, i.e. ‘hard to utter’ (cf. 713). And τυλίσσων, which more immediately governs οἴμας, evokes the notion of unraveling a thread: “déroule les énigmes” (Hurst 2008:2). One must also recall that the manufacturing of thread—an obvious metaphor for epic composition—involved twisting between one’s fingers fibers from the mass of carded wool. For a cognate use of τυλίσσω as ‘to roll up’, see the scholia to ζ 53 and Eustathios Ad Odysseam ad ν 107 (Stallbaum 1825–1826:2.42.30–31). There might also be a hint in this verse of Theseus’ escaping from the labyrinth by winding up the thread of Ariadne (so Müller 1811:1.6 and Mooney 1921:2n11; contra, Holzinger 1895:166). τυλίσσω stands for convolvo more often than for evolvo.
[ back ] 227. But οἴμη is used for ‘way’ in Cougny 1890:486 epigr. 120.16 (on Plotinos) and Quintus of Smyrna 7.320, 9.508.
[ back ] 228. The genitive is certainly not one of inverse attraction. Although it is possible to classify it as partitive, with Pagliaro (1951:28 and n. 2), I believe that the rhapsodic usage attested by the syntax of Ι 97 α 10 θ 500 (cf. Φ 142) decisively points to a genitive of origin. Cf. Schwyzer GG Ⅱ.92 §4e.I.1c and Nagy 1996c:63n19.
[ back ] 229. Cf. Pagliaro 1951:29 and Nagy 1996c:63. Cf. Nagy 2009b:241 2§109, 330–335 2§§311–323; and Nagy 2010a:79–102.
[ back ] 230. See §1.2 above on the characteristic tendency of Homeric epic towards ‘monumentality,’ that is, towards an explicit, no longer merely notional, comprehensiveness in performance that progressively limited the ability of rhapsodes to shift their focus from one episode to another not immediately contiguous along the plot line.
[ back ] 231. Heide 2006: “In Icelandic seiđr tradition, from recent times, attraction dominates and most of the sources have the fixed expression seiđa til sín ‘attract by seiđr’ … . In some of the sources, it is as if the victim is pulled by an invisible rope” (164).
[ back ] 232. Compare Pagliaro’s attempt (1951:29) to put this notion in a diachronic perspective by taking into account that it must have referred to creative composition: “L’uso del termine οἴμη nell’Odissea … fanno pensare che l’innovazione, cioè l’applicazione di questa nozione di ‘legame, traccia’ alla composizione, sia probabilmente dovuta al fatto che la novità del poetare epico è costituta dal ‘legame’ contenutistico che informa l’esibizione del cantore, del ‘racconto’, nei confronti della poesia lirica divisa in membri … .”
[ back ] 233. The bibliography on the Greek hymn is voluminous and this is not the place for a review. I will only mention, among recent surveys, Cassio and Cerri 1991, Lattke 1991, Burkert 1994, Devlin 1995, and Furley and Bremer 2001; and the important treatments in Koller 1956, Durante 1976:155–166, Nagy 2000a and 2002:70–98, Vamvouri Ruffy 2004, and Nagy 2009b:189–248 2§§8–122.
[ back ] 234. Bader (1990:34) deals with Sanskrit syū́man- ‘band, thong, bridle’, which is regularly mentioned in connection with ὑμήν.
[ back ] 235. *-mn̥- is known in long e- and o-grades, -μην and -μων, as well as in the zero grade -μα. Cf. Chantraine 1933:170–174.
[ back ] 236. Chantraine 1933:174 §133.
[ back ] 237. Cf. Bader 1990:11 §3.5 and 16 §4.4.
[ back ] 238. Cf. Bader 1990:25.
[ back ] 239. Frisk 1973–1979 s.v. ὕμνος was skeptical. Cf. Mayrhofer 1953–1980:3.485 (with references). Beekes 2010:2.1531–1532 surveys the proposed etymologies for ὕμνος without committing himself to any of them.
[ back ] 240. This totality is notional insofar as hymnic poetics allows the rhapsode the freedom to shift to another point along the narrative thread. Such shifts include the possible transition from a ‘hymn’ understood in the narrower sense of prooimion (e.g. a shorter Homeric hymn) to a subsequent performance of epic. Thus, in practice, to speak from the perspective of a cultural outsider, by shifting the rhapsode may ‘bracket’ an intervening narrative stretch.
[ back ] 241. Pace Collins (2004:181n4), ἀοιδῆς ὕμνος need not imply “that humnos is a subdivision of song.” There is no compelling reason to read the genitive thus. One would hardly construe in this manner the English ‘a song’s performance.’
[ back ] 242. Cf. the remarks in Vine 1999:575–576. It is not clear whether of old Certamen 213–214 was thought compatible with Hesiod Works and Days 656–658. The Works and Days does not mention Homer, and modern scholars have often identified the ὕμνῳ of Works and Days 657 as the Theogony—a work whose formal and thematic canons more readily fall under the narrower notion of ‘hymn.’
[ back ] 243. Cf. Bader 1989:23.
[ back ] 244. See Bakkhylides 19.6–9 SM and Pindar fr. 179 M. Cf. Nünlist 1998:112–116.
[ back ] 245. “The use of forms of a root ‘to bind’ for an art of poetry that includes singing is in line with the general concept of poetry … as entailing technical skills; these assume an ‘arrangement’ (cf. *h2er-, §12.1), which metaphors of crafts effecting particular arrangements can specify … . [Sewing] is a particular form of ‘bonding’ … . For the art of poetry, to hesitate … between the senses ‘to bind’ and ‘to sew’ of the root [*seh2-] is probably only a false problem, since sewing was from the start a [type of] bonding, and, furthermore, the forms in *seh2-(i)- of the vocabulary of the art of poetry apply mainly to singing” (Bader 1990:38–39).
[ back ] 246. To these, one may add Herakleitos’ censure of Homer (DK 22 B42), if the attractive suggestion that ῥαπίζεσθαι puns on ῥάβδος is true. Cf. Nagy 1989:38 and Graziosi 2002:29.
[ back ] 247. Κλεισθένης γὰρ Ἀργείοισι πολεμήσας τοῦτο μὲν ῥαψῳδοὺς ἔπαυσε ἐν Σικυῶνι ἀγωνίζεσθαι τῶν Ὁμηρείων ἐπέων εἵνεκα, ὅτι Ἀργεῖοί τε καὶ Ἄργος τὰ πολλὰ πάντα ὑμνέαται.
[ back ] 248. See above, §2.2 n. 24.
[ back ] 249. See, for example, PEG I.22 F1 Bernabé = EGF 22 F1 Davies. On the authorship, cf. PEG I.21 T5 Bernabé and EGF 21 de Homero auctore Davies.
[ back ] 250. Cf. Nagy 1990c:22.
[ back ] 251. Τερψικλῆς τῶι Δὶ Ναίωι ῥαψωιδὸς ἀνέθηκε (in the Ionic alphabet). Usually dated after Kirchhoff to the mid fifth-century BC, this is only possible if the inscription is by an Ionian artisan. If of local manufacture, I believe rather that it ought to be downdated to sometime in the first half of the fourth century BC. See below, §12.3.1, item 9.
[ back ] 252. Cf. Tarditi 1968:144 and Graziosi 2002:25–27.
[ back ] 253. Cf. Bollack (1990:245), who after noting this fact remarks that the mobility associated with the rhapsodic profession may also have suggested the label.
[ back ] 254. As we shall see, the point of the adjective is not to ascribe their inspiration to a source other than the Muses, but to draw attention to the absence of merriment that ordinarily accompanies all that the Muses oversee.
[ back ] 255. Another semantic pun.
[ back ] 256. Cf. Ritoók 1962:229n17. For contests of riddles, see Schultz 1909–1912:2.73–81 and the article by the same in RE 1 A (zweite Reihe), coll. 62–125 s.v. “Rätsel,” esp. 70–72.
[ back ] 257. An elegiac version of a possible answer has been transmitted by the scholia to Euripides Phoinissai 50 (Schwartz 1887–1891:1.257), but it is clearly of late composition and cannot rival the antiquity of the hexametric account of the riddle, even if the latter is judged post-archaic.
[ back ] 258. Tarditi’s interesting but ultimately unconvincing article is at least right on this point: bearing in mind that, when he uses ‘to sew, stitch’, he means ‘to perform sequentially after another’, he is surely right to aver that “la Sfinge non cuciva il suo canto in nessun contesto” (1968:142). Of course, as my argument in this chapter shows, if taken absolutely, the words “in nessun contesto” are wrong.
[ back ] 259. According to Graziosi (2002:24), the scholiast to Oidipous tyrannos 391 complains that calling the Sphinx a ‘rhapsode’ is anachronistic. But the scholion, whose authority at any rate is limited by its chronological horizon, is not so definitive. In Papageorgius’s edition its text reads: τὸ δὲ ὄνομα τοῦ ῥαψῳδοῦ καθ’ Ὅμηρον ἢ μεθ’ Ὅμηρον ἦν· ⟨ἀνεχρόνισεν οὖν ὁ Σοφοκλῆς⟩. The bracketed sentence (with ἀνεχρόνησεν corrected by Dindorf to ἀνεχρόνισεν) appears only in G (Laurentianus 2725), which also omits ἢ μεθ’ Ὅμηρον. Other than in G, the meaning seems to be: ‘The word “rhapsode” [=‘calling her a rhapsode’] was according to (dated to the time of) Homer or postdated Homer’. This observation agrees with the Suidas 4.287 Ρ no. 70: ‘The word “rhapsode” [v.l. “to rhapsodize”] is Homeric, or it may have been post-Homeric’. This is hardly a ringing indictment for anachronism. It is not clear whether the statement is meant absolutely (‘the word rhapsode’) or is to be limited to the Sphinx (‘calling the Sphinx a rhapsode’). In other words, the scholiast may be (tentatively) ascribing to Homer an archaic epic that called the Sphinx a ‘rhapsode’ and whose usage Sophokles is deemed to follow. The reading of G is puzzling: ‘The word “rhapsode” was according to Homer [i.e. “followed Homer”]. Therefore Sophokles was anachronistic’. Does this condemn as anachronistic Sophokles’ (alleged) resort to Homeric usage? Hence Schneidewin’s plausible conjecture οὐ καθ’ Ὅμηρον (apud Papageorgius).
[ back ] 260. This is Asklepiades of Tragilos, whose fragments are in FGH 12 (for the Sphinx, see F7, vol. 1a p. 169).
[ back ] 261. Snell et al. 1971–2004:5.1.571–574 nos. 540, 540a, and 540b. For a translation, see Collard and Cropp 2008:13.
[ back ] 262. I do not call it yet the ‘traditional’ version, since this is the matter at stake I am seeking to establish: may this vulgate form of the riddle be considered ‘traditional’ in the same sense that (other?) poetry from the epic cycle is considered ‘traditional’?
[ back ] 263. Robert 1915:1.56–57: “Da ist nun in der Tat doch das weitaus wahrscheinlichste, daß die fünf Hexameter aus einem Epos stammen, mag dies nun die Oidipodie, die Thebais oder eins der anderen epischen Lieder vom Oidipus gewesen sein, deren es gewiß noch viele gegeben hat.”
[ back ] 264. Mastronarde 1994:17–18.
[ back ] 265. A different kind of challenge is raised by Edmunds 1984 in a folklore study of the Sphinx’s role in the Oidipous myth. He argues that her presence is an over-determining addition to conventional folklore patterns of patricide and mother-incest. Why this is of little moment for the Greek Oidipous myth is best stated by Mastronarde 1994: “We know of no Greek version in which [the Sphinx] was absent.” Edmunds’s approach is open to arbitrary discretion (cf. Bremmer 1988:46–47). He notes, for example, that, of the two main categories of folktales cognate with the Oidipous legend, one does not feature a parricide: the hero returns to his native land, defeats an invading army, and then marries his widowed mother (1984:158). Why then should we not consider parricide a secondary element? And yet Edmunds assigns the Oidipous legend to the “divine kingship succession narrative” story-pattern, an illustration of which (from the Near-Eastern city of Dunnu) features “parricide and incest … in successive generations.” There is confusion, moreover, in Edmunds’s use of the terms ‘primary,’ ‘secondary,’ etc. Sometimes a ‘primary’ motif is one that is ‘integral’ to the plot; sometimes, it is its ‘high point.’ But what about a motif that no known version lacks? Does this fact make it ‘primary’? If so, eo ipso the riddling Sphinx would be ‘primary.’ That the identity of the divine sponsor of the Sphinx depends on the version or does not exist or goes unmentioned hardly entails the secondary status of the Sphinx. As Mastronarde (1994:19) writes, “[i]n myth some monsters appear as punishment for misdeeds, others are simply present in a locale as a chaotic, destructive force which must be conquered by a god or hero so that human civilization can be founded or maintained” (cf. Gantz 1993:495).
[ back ] 266. See PEG I.17–19 Argumentum? Bernabé = FGH 16 F10, vol. 1a pp. 181–182, with commentary in 495–496.
[ back ] 267. εἶτα ἔγημε τὴν μητέρα λύσας τὸ αἴνιγμα.
[ back ] 268. Here he follows Bethe. On this, cf. Moret 1984:80–81 with bibliography.
[ back ] 269. It is clear from the iconographic and literary records that she had a distinct preference for youthful quarry. Cf. Gantz 1993:497.
[ back ] 270. Cf. Moret 1984:10n6.
[ back ] 271. It is not difficult to find vases that feature audiences of beardless and bearded figures. See, for example, Moret 1984:1.34 fig. 4 (Cat. 33), ca. 470/60 BC, or vol. 2 pl. 18 (Cat. 31), ca. 480. Moret 1984:33 writes: “Sur plusieurs documents (Cat. 28–30 et 32), les interlocuteurs de la Sphinx sont des adolescents.” But “[e]n général, les classes d’âge sont mêlées” (34).
[ back ] 272. The two oldest items in Moret’s catalog (1984:165 nos. 1 and 2), a pair of Siana cups by the C Painter from ca. 570/60, feature divergent choices: on the first, eight beardless youths flee the Sphinx; on the second, five bearded men. Cf. Moret 1984:10.
[ back ] 273. Beazley Archive Vase no. 200716.
[ back ] 274. Moret (1984:11) speaks of the “Sphinx’s lack of aggressiveness” as a “fundamental particularity of the figurative tradition.” See his discussion at 12 about the subordination of the narrative to the need for compositional symmetry and regularity.
[ back ] 275. See the apposite remarks by Moret (1984:79–80), who is surely right to question on iconographic and literary grounds the broadly received dogma that the physical combat between Oidipous and the Sphinx was more primitive and must have preceded the invention of the riddle (cf. Bethe 1891:19–20).
[ back ] 276. Kock 1961 is no better guide for this matter than Lesky. Early in his article he commits himself to the notion that, in contrast with the cunning figure of later myth, “Oidipus in the Iliad [is] a figure of physical strength” (11); and that “the Sphinx did enter the Oidipus saga as a creature of brute force and only later became the poser of riddles” (10–11, his emphasis). But what does he have as evidence for this? Other than the weight of recognized authorities (Robert, Bethe, Lesky, and Wehrli, all cited at 10n23), he only notes that “we find proof of [the secondary nature of the riddles] in the appearance of the Sphinx” (10). Apparently, “a monster with the body of a lion is a figure of strength and force and hardly compatible with intellectual prowess” (10). No thought is given to the possibility that, whatever its earlier Phoenician background, this “monster” might have been adopted by Theban epic from the start as a figure of force and cunning. Or that, even if ex hypothesi cunning was not hers in the very early stages of Theban traditions of epic, by the sixth century it was a well established trait and played a role in the Oidipodeia. Kock’s prejudice seems to be that archaic epic could not have told the story so and that we must wait until the very late archaic period for a bard to come up with the notion. But why? Kock thought Pindar the terminus ante quem for the riddle and owned that “about the Sphinx’s riddle I follow Lesky” (1961:21). But we have seen above that one vase depiction dates the motif of the riddle to no later than ca. 540 BC; and we now know of one vase from ca. 520 with fragments of the riddle’s text in agreement with the vulgate version, which probably comes from the Oidipodeia (see below, §10.2.4). Given his views, Kock seems oddly unfazed by his admission that “the riddle as motif is as old as the hills” (21).
[ back ] 277. Of course the pairing of Sphinx and Oidipous might postdate the inherently riddling character of the monster, although Katz offers further reasons for its age in the same footnote.
[ back ] 278. Katz 2006:177–178, esp. 178n52, with the quotation at 177.
[ back ] 279. The scholiast ad loc. had already written of ἐλεφαίρετο: ἀντὶ τοῦ ἔβλαπτεν. Ὅμηρος δέ ποτε καὶ ⟨ἀντὶ τοῦ⟩ παρελογίζετο. Translators, however, often prefer the stronger and fatal ‘to destroy’ vel sim. But cf. scholia to Ψ 388 and Maehler 1982–1997:151–152 ad Bakkhylides 9.6–9. Modern scholars have hardly improved on the Hesiodic scholiast. West 1966 passes over the verb in silence. The LfgE s.v. ἐλεφαίρω offers the sole gloss “heimtückisch, durch Täuschung Schaden zufügen.” But, faced with Theogony 330, it notes “spez. Nuance unklar” and suggests δήλημα at Hymn to Apollo 364 as if it were a parallel.
[ back ] 280. Lesky himself admits that “Homerische Diktion ist wohl künstlich erstrebt—besonders auffällig der Einsatz eines homerischen Flickens in v. 2–3 = δ 417f.—, nicht aber erreicht” (1966:322).
[ back ] 281. For the combination of tradition and innovation peculiar to Antimakhos’ poetry, see Lombardi 1993 and Matthews 1996.
[ back ] 282. Cf. Beekes 2010:1.176 s.v. His references to “Fur.” are to Furnée 1972.
[ back ] 283. Η 237, said of a ‘feeble child’ (παιδὸς ἀφαυροῦ) who is paired with a woman without fighting experience; Η 457, of a god who, compared to Poseidon, is ‘much weaker in hand and strength’ (πολλὸν ἀφαυρότερος χεῖράς τε μένος τε); Ο 11, of Ajax, ‘not the weakest (ἀφαυρότατος) of the Akhaians’; υ 110, of a woman at a mill who was ‘feeblest of all [grinders]’ (ἀφαυροτάτη δὲ τέτυκτο); Hesiod Works and Days 586, of men during the heat of summer, when they are ‘weakest (ἀφαυρότατοι) because Sirius parches their head and knees’.
[ back ] 284. ὣς Ἕκτωρ ἰθὺς σανίδων φέρε λαᾶν ἀείρας | … στῆ δὲ μάλ’ ἐγγὺς ἰών, καὶ ἐρεισάμενος βάλε μέσσας | εὖ διαβάς, ἵνα μή οἱ ἀφαυρότερον βέλος εἴη (Μ 453 457–458).
[ back ] 285. LfgE s.v. βέλος B: “missile, projectile … a generic term for anything cast or propelled with intent to damage: of arrows, spears or stones … . The epith[ets]: ὠκύ [qualifies βέλεα in the sense] of spears (Ξ 407+) as well as arrows but would not be inappropriate with stones, especially sling-stones (see Ν 599f. 716).”
[ back ] 286. See Onians 1951:310–342; Michel 1971; Heubeck 1972 with older bibliography at 139n14; and Nothdurft 1978.
[ back ] 287. Cf. Janko 1992:92.
[ back ] 288. Scholia AT to Ν 359a–b Erbse. Cf. Michel 1971:55.
[ back ] 289. As Michel (1971:55) observes, there is only one πεῖραρ. Hence, the notion that two ropes or the two ends of a rope are being woven or tied together cannot be in play. According to the scholiast ad loc., the latter was Aristarkhos’ interpretation, on which see van der Valk 1963:97–99.
[ back ] 290. Of course, Lesky does not put it so. By pronouncing the use “foreign to epic” (“der Gebrauch … ist dem Epos fremd”) and adding that one would instead expect ἀμείβειν, he makes clear that he objects to the word itself, not just its meaning (Lesky 1966:322).
[ back ] 291. I have set up two proportions—one for the transitive, another for the intransitive, construction—because Greek usually opposes the transitive ἀμείβω to the intransitive ἀμείβομαι (cf. DGE s.v. ἀμείβω), whereas it regularly deploys the active ἀλλάσσω for both. There are exceptions, however, and ἀμείβω is used intransitively at Ψ 712.
[ back ] 292. Leaf (1900–1902:2.29) remarks that “[t]he general sense of the passage would be better given if we could translate ἐπαλλάξαντες alternately. The use of ἀλλάσσειν makes this possible, but we should require the pres. part. in place of the aor.” Bergren (1975:48–49) agrees: “[T]he (metrically equivalent) present rather than the aorist participle would be required. The action of ἐπαλλάξαντες, whatever it means, is not continuous.” But van der Valk (1963:2.99) long ago corrected this syntactic misunderstanding: “Leaf (ad loc.) says that the participle of the aorist cannot be used in this meaning. However, this participle may convey the idea of concomitant circumstances, cf. Schwyzer, GG Ⅱ, 388.” I address this important syntactic point at length below, §10.2.6, in connection with Hesiod fr. 357 MW.
[ back ] 293. Heyne 1802–1822:6.438: “At h. l. Iupiter et Neptunus ἐτάνυσσαν πεῖραρ, utrinque extendunt, arreptis funis oris, nec coeunt et iunguntur, sed recedunt et disiungunt se; ut intento fune alter deficientibus viribus cedat alteri et est ἐπαλλάξαντες, alternando, alternis, dum modo hic funem intendit, modo validius ille.” Cf. scholia A to Ν 359a Erbse; for the game (under σκαπέρδα), see Pollux Onomastikon 9 §116 (vol. 2, p. 179 Bethe) and Hesykhios Σ no. 854.
[ back ] 294. “Ich fasse es so: wechselnd mit dem Ausgang des Kampfs, theilten sie beiden Völkern den Sieg zu, impenderunt utrisque, d. i. ἐπαλλάξαντες ἔδωκαν νίκην ἀμφοτέροις. … Also ist ἐπαλλάττειν πεῖραρ πολέμου nichts mehr als permutare victoriam; und τανύειν πείρατα πολέμου ἐπί τινι, gleichbedeutend mit διδόναι νίκην τινί” (48–49).
[ back ] 295. Interpreting τοὶ δὲ as the Akhaians and the Trojans, he translates: “Utrique alternantes vincebant … victoria in utramque partem varie inclinabat.”
[ back ] 296. “[D]as Leitseil des Kampfes … spannten sie wechselnd über beide Parteien, … in verderblicher Wirkung bald für die eine, bald für die andere Partei.”
[ back ] 297. “Now ἐπαλλάττω also has the meaning of ‘to alternate’. This meaning especially suits a situation in which a battle remains undecided. Therefore, we must render ‘Both gods stretched the rope of war alternating it’, i.e. both parties were alternately victorious.”
[ back ] 298. He translates: “Des heftigen Streites und allen gemeinsamen Krieges | Seil spannten sie wechselweise über beide Heere.”
[ back ] 299. Tentatively: “mit wechselndem Erfolg (ἐπαλλάξαντες)?”
[ back ] 300. “Alternately, they pulled taut the rope of violent strife and equal war over both sides … , i.e. the two gods made the armies fight a fierce and indecisive battle.”
[ back ] 301. DGE s.v., which presents under I the transitive ‘change, modify, alter, vary’; under Ⅱ the transitive ‘exchange’ and ‘repay’; under Ⅲ the intransitive ‘alternate, take turns’; and under Ⅳ the intransitive ‘change, alter’.
[ back ] 302. LSJ s.v. ἐπί G.4 speaks of “[a]ccumulation of one thing over or besides another.”
[ back ] 303. This ‘sympathetic’ construction, in which the preverb and the preposition motivate each other, must be responsible for the variant to Ζ 230 δ’ ἐπ’ ἀλλήλοις ἐπαμ. apud cod. E4 of Porphyry Quaestiones Homericae 1.96.24 (Schrader), which Erbse reports in his entry of the LfgE s.v. ἀμείβω Ⅲ.2.a. Schrader himself does not report it, although, following Erbse, I have given the corresponding locus in Schrader’s edition for ease of reference. On cod. E4 see Erbse 1969–1988:1.xx–xxi. Since I have not inspected the manuscript, I cannot tell whether the reading is reported as an unmetrical variant to the Iliadic text or whether the δ’ should be suppressed and one should print the resulting asyndetic reading τεύχε’ ἐπ’ ἀλλήλοις. For the elision, cf. τεύχε’ ἐπ’ αὐτῷ (Δ 504 etc.). For asyndeton in Homeric poetry, see Hermann 1806:98–100 ad Hymn to Aphrodite 177; Thiersch 1826:579–580 §312.33; Nägelsbach 1834:266–280 (“XⅣ. Beitrag zur Lehre vom homerischen Asyndeton”); Seymour 1895:18–19 §§l–n; and Chantraine GH Ⅱ.351.
[ back ] 304. I believe that the choice of ἐπί with ἀμφοτέροισι is further motivated by the preverb of ἐπαλλάσσω. See n. 303 immediately above.
[ back ] 305. Either contingent suffers loss as it is pulled together with its divine patron in the direction of his prevailing opponent.
[ back ] 306. The Homeric meaning is ‘form’ or ‘figure’ (see Α 115, with Leaf’s note ad loc.; cf. Β 58 Γ 208 etc.).
[ back ] 307. Moret 1984:40 and 175 no. 87 (vol. 2 pl. 51.1) = ARV2 451.1.
[ back ] 308. Or κ]αὶ τρί[πουν if we assume the Attic form. This cup is often thought to reflect Aiskhylos’ satyr-play Sphinx. Cf. Simon 1981:28–31. To weaken its alleged connection with the epic version of the riddle, Lesky 1966:322 suggested that ΑΙΤΡΙ might stand for καὶ τρία. He added that the space after the second iota supported this reading over against καὶ τρίπον, because it could only accommodate one more letter. (So also Hoffmann 1997:80–82, apparently following Hartwig 1893:664 without argument.) But Lesky’s rationale is invalid: whereas the missing kappa of καὶ falls on a line of fracture and we may legitimately assume it was lost, there was never anything after the iota of ΤΡΙ. The inscription reads ]ΑΙΤΡΙ, not ]ΑΙΤΡΙ[. It might seem odd to us that vase painters should depict speech fragments, but this is a well-established fact which Beazley (1927:348–349) noted in connection with this very cup: “Such snatches of song, it is well known, are not uncommon on archaic red-figured vases. … [S]ometimes the snatch ceases in the middle of the word. … So in the Vatican cup with Oedipus and the Sphinx, all that Douris writes of the riddle is [κ]αὶ τρι … .”
[ back ] 309. Beazley Archive Vase No. 43112. See also Vollkommer 1991:60 and n. 92.
[ back ] 310. Without autopsy, a definitive confirmation of Moret is impossible. But judging from a careful inspection of the photographs, he seems to be right. On pl. 23.1 I tentatively read a retrograde and horizontal ΕΠΕΙΔΑΝ ΓΕΡΑΣ̣. This, Moret reports there as ἐπειδὰν γῆρας. The epsilon graphemes (for ε and η), of admittedly odd shape and defectively drawn like digammas, correspond to Immerwahr 1990:xxⅱ S6. Remarking on this shape in connection with nonsense inscriptions, Immerwahr notes that “although these forms [S6–7] may sometimes represent faulty epsilons, they are often used with consistency and in company with well-written epsilons, and I think it likely that they are to be read as digammas” (1990:140–141). Pl. 23.2 does, in fact, seem to offer regularly shaped epsilons. Nevertheless, I am satisfied that the odd S6’s are epsilons too and that the readings reported by Moret are at least substantially right and make good sense. (I cannot say to what degree what appear as odd shapes might be due to faint traces invisible on the photographs.) Also on pl. 23.1 I tentatively read a retrograde, horizontal ΕΤΡΑΠ which shifts direction at the Π and continues downward with ΟΥΝ. If this is Τ]ΕΤΡΑΠΟΥΝ, it agrees with Moret’s report of τετράπουν οὗ. But I cannot confirm his οὗ. Two letters follow ΠΟΥΝ; they seem an eta and a upsilon (ὑ?). The vase painter might have intended ΗΟΥ = οὗ. There is, in fact, a precedent for ὑ- ~ οὑ- in the Neandros black-figure cup, Boston 61.1073 dated to ca. 540, which spells ΗΥΤΟΣ for οὗτος (see Threatte 1980:260 and Immerwahr 1990:49 no. 229, 162, and pl. 14.64). At any rate, these readings seem substantially in harmony with the vulgate version of the riddle. The Attic form τετράπουν appears as an (unmetrical) variant in some of the multiforms, e.g. in Athenaios’ version (corrected by Kaibel to τετράπον). As is well known, the source of the Attic -ου- remains a puzzle; it should have been -ο-, like the Homeric, or -ω- if from a lengthened grade (cf. Sihler 1995:117–118 §116.1 and Rix 1976:95 §105, 126 §137).
[ back ] 311. Snell et al. 1971–2004:3.287–288. Cf. Nauck2 no. 173.
[ back ] 312. Tragedy can occasionally incorporate hexameter poetry as necessary (cf. West 1982:98, 128–132). This is precisely what happens, for example, in Euripides’ Oidipous fr. 540a.7–10 Kannicht (Snell et al. 1971–2004:5.1.573).
[ back ] 313. A change of mind since West 1978a:293.
[ back ] 314. The scholion to Euripides Phoinissai 50 calls her a χρησμολόγος, and her riddle, χρησμοί. Cf. Nagy 1990c:147–150, 328–329 on αἶνος.
[ back ] 315. Ἀσκληπιάδης δὲ λέγει τοὺς Θηβαίους εἰς ἐκκλησίαν καθ’ ἑκάστην ἀθροίζεσθαι διὰ τὸ δυσαίνιγμα τῆς Σφιγγός … . ὁπότε δὲ μὴ συνίοιεν, ἁρπάζειν αὐτὴν ὅντινα ἂν βούλοιτο τῶν πολιτῶν (scholion to Euripides Phoinissai 45).
[ back ] 316. So, for example, Ritoók 1962: “Es fragt sich auch, ob das Publikum im 7. Jahrhundert einen so grossen Unterschied in Hinsicht des Stiles und besonders der stichischen Art fühlte, und ob es gerade diesen Unterschied für etwas so wesentliches hielt, wie Patzer es glaubt” (226).
[ back ] 317. I am anxious to disown an unproblematic view of ‘lyric’ as poetry that stands in more or less straightforward synchronic contrast to ‘epic.’ I believe that epic (formally and thematically) derived diachronically, by a narrowing and stylization of its generic and performance conventions, from the larger whole of geometric and early-archaic Greek song traditions. This hypothesis, which I believe has comparative, anthropological, historical, and methodological support superior to its alternatives, means that archaic lyric in all of its manifold variety, though often thought to follow epic and to reflect epic themes and diction to various degrees, actually represents a parallel development to archaic traditions of epic performance. I do not, of course, a priori rule out the possibility of synchronic interactions between them at various points of the diachronic continuum.
[ back ] 318. Cf. Kirk 1985:17–30 and 1966a, with bibliography. The Latin grammarian Diomedes reflects the ancients’ controlling awareness of sequence in epic (and, incidentally, of epic’s connection to oracular language) when he writes in his Ars grammatica (GL 1.484 Keil): “epos autem appellatur, ut Graecis placet, παρὰ τὸ ἕπεσθαι ἐν αὐτῷ τὰ ἑξῆς μέρη τοῖς πρώτοις. praecipue vero hexameter versus epos dicitur, quoniam quidem hoc versu verba responsi in mutuam, ut sic dixerim, consequentiam primus deus vates conprehendit, unde postea abusive verbum et solutae orationis ipsa scriptura consequens ab aliis epos dictum” (‘epos moreover receives its name, as the Greeks think, “from the succeeding parts in it following the former.” It is particularly the hexameter verse, however, that is called epos, since indeed the god first as oracle-giver expressed in this verse the words of his response in mutual (con)sequence (so to say), whence later by an abuse of terminology a word even of prose was called epos by others because it “follows” by the very writing’). The popular etymology ἔπος < ἕπεσθαι attested by this statement speaks forcefully for the view that notions of sequence are intrinsic to epic poetry.
[ back ] 319. Cf. Kirk 1985:30–37 and 1966b.
[ back ] 320. Kirk 1985: “[W]hole-sentence verses are now in a minority of one in six or more” (31).
[ back ] 321. For Kirk, sentence is the supra-linear unit of meaning. His previous section is titled “From verse to sentence.”
[ back ] 322. Or at least it is the Panhellenic stylization of local poetries that were so sung. Cf. West 1981:125–126 and West 1992:52–53.
[ back ] 323. West 1986:43–44 notes that the hexameter hymn to Asklepios from Epidauros (West 1992:279 no. 11), although written for seven strings (the fashion since the seventh century BC), featured the same melody for every line. After proposing that it may be a traditional one, if slightly modernized, inherited from the archaic period, West adds: “[T]he eighth-century aoidos sang epic poetry on four notes, the four notes to which his phorminx-strings were tuned, and … he followed the contours given by the word accents. That would mean a different melodic configuration for each verse, though it would be compatible with a broad pattern repeating from line to line” (45).
[ back ] 324. Cf. Notopoulos 1964:58–59.
[ back ] 325. “Als Objekt von ῥάπτειν ist … das Ergebnis der Arbeit und nicht der bearbeitete Stoff anzusehen.”
[ back ] 326. In its etymological sense ‘com + ponere’ = ‘to put together’, ‘to construct’.
[ back ] 327. It is not clear whether the quotation too is taken from Philokhoros. The fragment is classified among the “dubia” by Merkelbach and West (fr. 357). The reference to ‘Homer’ dates it to the late sixth or early fifth century BC.
[ back ] 328. Because the correct translation of this fragment depends crucially on the analysis below, I forgo a translation here which would seem to beg the question.
[ back ] 329. “Wie durch die Interpunktion deutlich gemacht ist, gehört μέλπομεν ἐν νεαροῖς ὕμνοις zusammen” (Marx 1925:399).
[ back ] 330. Because μέλποντες ἑκάεργον reprises the main clause μολπῇ θεὸν ἱλάσκοντο, it functions effectively as μέλποντο ἑκάεργον. This happens to be the reading of Etymologicum Magnum 657.4–5, col. 1858 Gaisford (cf. scholia A ad Α 474a), which is only marginally metrical (cf. West 1982:38 under ‘d’; the verse-initial sequence ‘-οντο CV-’, here μέλποντο *ϝε-, is not exampled in Homer or Hesiod). The sixth Pindaric paean, fr. 52(f).15–18, cited as a parallel by the LfgE s.v. μέλπω B.1, exhibits similar syntax. It assigns to the main verb κροτέο[ντι the description of the performance and adds to it the participle μελπ̣[ό]μεναι qualified simply by θαμινά: τόθι Λατοΐδαν | θαμινὰ Δελφῶν κόραι | χθονὸς ὀμφαλὸν παρὰ σκιάεντα μελπ̣[ό]μεναι | ποδὶ κροτέο[ντι γᾶν θο]ῷ.
[ back ] 331. τῇσι μὲν οὔτ’ αἰσχρὴ μεταμέλπεται οὔτ’ ἐλάχεια, | ἀλλὰ μάλα μεγάλη τε ἰδεῖν καὶ εἶδος ἀγητὴ | Ἄρτεμις ἰοχέαιρα (Hymn to Apollo 197–199).
[ back ] 332. οἶδα δ’ ἐνὶ σταδίῃ δηΐῳ μέλπεσθαι Ἄρηϊ.
[ back ] 333. As might be the case with Α 521 Ν 684 Π 79 (so Chantraine GH Ⅱ.78 §106 under “Remarques I”).
[ back ] 334. Perhaps an ironic oxymoron, as Mader suggests in the LfgE s.v. Ἄρης B.2eβ.dd (col. 1261, lines 48–52), where the reference to Π 610 should read Π 617. See generally the LfgE s.v. Ἄρης B.2e for the god as a metonymy of ‘fight, battle’. This locative would parallel the use of ὑσμῖνι in μέμασαν δ’ ὑσμῖνι μάχεσθαι (Β 863), for which see Ameis and Hentze 1877:158 and Chantraine GH Ⅱ.78 §106 (under “Remarques I”). Cf. Θ 56.
[ back ] 335. So Leaf 1900–1902:1.316 (“to dance the war-dance to Ares”; “Hector means, ‘I can dance the war-dance not only in mimicry at a feast of Ares, but in grim reality on the battle-field’”) and Kirk 1990:268 (“dance for Ares”). Cf. Chantraine GH Ⅱ.73 §95.
[ back ] 336. Marckscheffel 1840:376 fr. 221; Goettling 1878:343 fr. 227; Kinkel 1877:174 fr. 233; Sittl 1889:625 fr. 180; Rzach 1908:221 fr. 265; Evelyn-White 1914:280 fragmenta dubia no. 3; Merkelbach and West 1967:176 fr. 357; Most 2007:354 fr. 297. It is so received by the LfgE s.v. μέλπω B.2.
[ back ] 337. All these in Marx 1925:399.
[ back ] 338. In accordance with my aspectual analysis immediately below, I have translated the aorist with an English present participle and forgo the conventional past participle ‘having stitched’.
[ back ] 339. K-G Ⅱ.1.182 §389.3: “[sie] bezeichnen ebenfalls nicht die Zeitstufe, sondern nur die Beschaffenheit des durch sie ausgedrückten Verbalbegriffes.” Cf. Smyth §1872.
[ back ] 340. Cf. K-G Ⅱ.1.199 §389.E, Anm. 8.
[ back ] 341. Stahl 1907: “In besonderen Fällen jedoch kann das Partizipium des Aoristes auch bei Gleichzeitigkeit eintreten, und zwar zunächst … wenn die durch das Partizipium und das übergeordnete Verbum bezeichneten Erscheinungen in der Weise als Tatsachen an sich hingestellt werden, daß sie vollständig zusammenfallen, die eine in der anderen oder durch die andere besteht, in welchem Falle das Partizipium im Deutschen durch ‘dadurch daß, damit daß, darin daß’ wiedergegeben werden kann” (212–213).
[ back ] 342. See Stahl 1907:213–214.
[ back ] 343. Sicking and Stork 1996:103. ‘Focus function’ concerns the pragmatics of discourse organization: “‘[F]ocus function’ [is] the part of the unit involved that, from a viewpoint of information, is the most prominent in the sense of being its ‘nucleus’, or the part ‘to which the speaker especially draws the hearer’s attention’” (Sicking and Stork 1996:75n121). Sicking surveys “postposed aorist participles” in Herodotean narrative at 51–52. They describe further the action of the main verb or give additional details connected with it. From the viewpoint of Herodotean discourse pragmatics, Sicking shows that this information is not presented as an element of the main sequence of events. These participles are akin to narrative ‘footnotes,’ although the information they convey is hardly nugatory. In the Hesiodic fragment, ῥάψαντες could be similarly read as an ‘aside’: “It was then that Homer and I first sang on Delos—we stitched our song with new hymns—[we sang] Phoibos Apollo … .” This ‘aside’ is of great interest, for it adds the specific manner of the singing, viewed complexively. Sicking summarizes his investigation thus: “Aorist indicative verb forms and participles 1) are to be assigned focus function (or: are the ‘nucleus’) in the clause they are part of, and 2) are the predicate of a self-contained statement” (103). Although I believe that Sicking has correctly identified the pragmatic implications of the aorist stem, I also share Wakker’s concerns about excessive oversimplification (Wakker 1998). In particular, I too believe that the pragmatic value of the aorist vis-à-vis the present stem can be understood as a consequence of the semantic opposition of the ‘complexive’ to the ‘non-complexive’ aspect. Wakker prefers ‘completed’ and ‘not completed’ (361–362 and n. 9), where completion may be understood in temporal or discursive terms. I would rather steer clear of labels that seem to privilege time.
[ back ] 344. As noted, the aorist, as the unmarked verbal stem, does not per se imply anything about the temporal extent of the action. It only contemplates it in its ‘bare form.’ It is sometimes called ‘punctual,’ and although the ‘point’ in question is often considered ‘a point in time’ the aorist does not inherently refer to time at all. In this case, inasmuch as the participle is the unmarked aorist, the syntax does not pronounce on whether the ‘stitching’ goes on while the song is in progress or else happens before, or at a given point during, the singing. Had it been a present participle, we would know that the stitching had some overlap with the μέλπομεν; but only the context could settle whether it was a parallel, unconnected action or one upon which the main verb depended. The aorist allows the participial action to receive pragmatic emphasis without prejudging whether it is actually durative or non-durative. This is what Ruipérez (1991:81 §132) calls “aoristo neutro,” indifferent to duration. Note his examples ad loc. of “neutral aorists” that belong to non-transformative durative semantemes.
[ back ] 345. See Martin 2000 for the suggestion that fr. 357 MW has the Hymn to Apollo in view, stitched together by Homer and Hesiod in competitive performance (the Delian section being Homer’s and the Pythian, Hesiod’s). Martin’s proposal is not new, as his review of previous scholarship shows. Among earlier contributors is Else (1957b:30–31), who taking up Crusius (1895:717–719) had written the following: “Here ἀοιδήν denotes the common content of the whole, viz. Apollo (Ἀπόλλωνα in apposition with ἀοιδήν), which is the joint product of the two bards’ efforts while the ὕμνοι are their individual contributions: ‘We sang, “stitching” the song [namely Apollo, etc.] in (the medium of, out of) new hymns.’”
[ back ] 346. Most 2007: “In Delos then for the first time Homer and I, bards, sang, stitching together our song with new hymns” (355); Evelyn-White 1914: “Then first in Delos did I and Homer, singers both, raise our strain—stitching song in new hymns” (281); Nagy 1996c: “Then it was, in Delos, that Homer and I, singers [aoidoi], for the first time sang, in new hymns, sewing together [rháptō] the song [aoidé]” (73); Powell 1997: “In Delos, Homer and I, singers of oral song, sang of Phoebus Apollo … stitching together oral song in fresh hymns” (31–32); Taplin 1992: “Then first in Delos Homer and I, the singers, stitching together our song in novel hymns” (41n57). On the hypothesis of respective solo performances, the point of the fragment’s πρῶτον would not be that Hesiod and Homer had never sung or stitched before. Rather, that they performed together “then for the first time” (never before had they sung for the same audience); or that they performed (together) in Delos “then for the first time” (never before had they sung, or sung together, at the Delian festival). If, on the other hand, we prefer the hypothesis of hypolepsis, πρῶτον would make clear that never before had Hesiod and Homer engaged each other in competitive relay poetics (so Crusius 1895:717).
[ back ] 347. Cf. Lucian Toxaris §10: πολλοὺς καὶ ἀξιοπίστους μάρτυρας τοὺς ποιητὰς παρεχόμενοι τὴν Ἀχιλλέως καὶ Πατρόκλου φιλίαν … ἐν καλλίστοις ἔπεσι καὶ μέτροις ῥαψῳδοῦντας. Here, the classification of ἐν καλλίστοις ἔπεσι καὶ μέτροις as an adverbial clause of ‘means’ shades readily into one of ‘manner.’ The choice of one over the other depends on how closely the writer is adhering to the manufacturing metaphor. With ῥαψῳδέω, ‘manner’ might be thought preferable; with ῥάπτω, ‘means.’
[ back ] 348. Evelyn-White’s translation sounds acceptable to the English ear because of expressions like ‘blowing glass into bottles’, where ‘into’ is not understood locally (‘putting glass into existing bottles by blowing’) but as capturing the manufacturing process through which the glass is turned into bottles (cf. OED s.v. “into” A.I.6a). His “stitching song in new hymns,” even as it evokes manner (cf. OED s.v. “in” I.12c)—and in so doing impermissibly disregards epic diction—effectively amounts to ‘stitching song into new hymns’. This inverts the intended terms of the manufacture, making ‘song’ the raw material and ‘hymns’ the product.
[ back ] 349. ἄγε δηὖτε μηκέτ’ οὕτω | πατάγωι τε κἀλαλητῶι | Σκυθικὴν πόσιν παρ’ οἴνωι | μελετῶμεν, ἀλλὰ καλοῖς | ὑποπίνοντες ἐν ὕμνοις (PMG 356b).
[ back ] 350. Besides Anakreon’s fragment and Herodotos 4.35 (where the adverbial clause qualifies ἐπονομάζω), I have not found any other early examples of ἐν ὕμοις or ἐν ὕμνῳ (with or without the interposition of the article and/or an adjective). Cf. Aristophanes Birds 906. The plural ἐν ὕμοις is very common with Christian writers but, other than the three cases above, wholly absent until the second century BC. From Philodemos’ reference to Ἐμπε⟨δοκλῆς ἐν τ⟩οῖς ὕμνοις in De pietate (DK 31 A33) one cannot infer that the expression ἐν τοῖς ὕμνοις goes back to Empedokles’ time. The fact that early instances of the adverbial expression should be so rare even with verbs of speech, performance, and the like, commends the interpretation that construes it more closely with the manufacturing metaphor entailed by ῥάπτω.
[ back ] 351. That is, assuming that ‘Hesiod’ and ‘Homer’ are not making just one seam to a joint hypoleptic performance and that the ὕμνοι are instrumental to the stitching.
[ back ] 352. It would be idle to speculate beyond the little that can be inferred from the fragment itself about the corresponding circumstances of performance.
[ back ] 353. For ‘metabasis’ as ‘performance shift’, see above, §7.2 and § n. 240. The Hymn to Apollo is precisely one such ‘larger song’ with a thematic bipartition and performative markers—openings (1, 179), closings (165–176, 545–546), and a transition (177–178)—that suggest its analysis into two smaller ὕμνοι. Just as one may speculate that the Hesiodic fragment portrays Homer and Hesiod as stitching together these two ‘hymns’ into the one resulting song, we may also consider this Homeric hymn a model of the kind of larger ‘new composition’ a solo performer might claim to have stitched together from smaller ὕμνοι. I do not think the singular number of the noun ἀοιδήν in the fragment an obstacle to this view: it could indeed mean ‘we sang (to) Apollo, stitching new hymns together into one joint song’; but the singular may stand, as often, for the (naturally singular) abstract verbal activity of ‘singing’ (cf. LfgE s.v. B.2: “Gesang als Tätigkeit, auch Resultat der Tätigkeit”); or it may adopt the familiar use of singular for plural hardly unique to ancient Greek. Cf., for example, the sentence ‘soldiers fought with a brave heart,’ which hardly implies the numerical restriction of ‘heart’ to one.
[ back ] 354. ‘A song of new sewn hymns’ ~ ‘a new song of sewn hymns.’
[ back ] 355. The sleeves could be buttoned rather than sewn. Cf. Cleland et al. 2007 s.v. “cheiridōtos.” And s.v. “seams”: “Essential to modern dress … seams were not similarly important in Greek and Roman dress, although many garments were sewn in some way.”
[ back ] 356. Cleland et al. 2007 s.v. “stitching.”
[ back ] 357. Cf. Durante 1968b:280–281, Cleland et al. 2007 s.v. “stitching,” and Marcar 2005:37, only to mention three examples.
[ back ] 358. DMic Ⅱ.221–223 s.v.: “con el significado de ‘talabartero’ o ‘guarnicionero’.” Cf. Lindgren 1973:132–133.
[ back ] 359. Lindgren 1973:134. Cf. “costurera” in DMic Ⅱ.221 s.v.
[ back ] 360. Cf. Barber 1975:294.
[ back ] 361. Frisk (1973–1979 s.v.) is skeptical about the derivation *κατ-sju- (cf. Latin suere), but I do not find Lagercrantz’s and Kretschmer’s objections persuasive (references apud Frisk). Cf. Schwyzer GG I.321, Bader 1990:33, and Chantraine 2009:1313 s.v. κασσύω. LIV2 545 states the IE root as *syewH-.
[ back ] 362. Barber 1991:277–278.
[ back ] 363. Cf. φορμορραφίς.
[ back ] 364. ‘Some are pruning the city like a vine, some have cut off the twigs of the demos, [the sinews of the state have been severed,] we are being stitched together like a mat, some are drawing us through narrow straits like needles’.
[ back ] 365. Hesykhios Φ no. 782.
[ back ] 366. “The plaiting itself is carried out with the aid of [a] wall beam. A bundle of rushes is tied to the beam and the plaiter works and walks backwards as the plait grows. As the coil increases and will not keep taut it is wound round the beam and the plaiting continues. … The coils for the body of a mat or carpet are mainly of 9 ply, while the borders are worked of a wider plait. The finished coils are handed over to other women for sewing up. This is done on a special table with a toothed edge to prevent slipping. The sewing must be done with great care. There must be even stitches and no puckering or the mat will not lie flat.” This passage, reportedly taken from Doris Stephens’s Memories of Sherington, is currently found at http://www.mkheritage.co.uk/shhs/rush.htm.
[ back ] 367. In Herodotos 3.98, Theophrastos Peri phytōn historia 2.6.11, and Aristophanes Wealth 542 they are made of plaited plant fibers. This seems to be the case also in Theokritos 21.13. In Pausanias 10.29.8 it must refer to coarsely woven fabric, a meaning that is marginally possible in Theokritos too.
[ back ] 368. Cf. Nünlist 1998:110–112, 114 §3.84.
[ back ] 369. Cf. the scholia to Aiskhines Against Ktēsiphōn 379 Dilts. The scholiast struggled to make sense of the expression and read it as ‘some are plotting against us’: ἀντὶ τοῦ συρράπτουσί τινες καθ’ ἡμῶν πράγματα, ἀπὸ τοῦ φορμοῦ. φορμὸς γὰρ λέγεται τὸ ψιαθῶδες πλέγμα, ὡς καὶ ὁ Θουκυδίδης (4, 48) φορμηδὸν λέγει. Note the word πλέγμα; in view of the manufacturing process for plaited rush mats (see above), its relation to συρράπτουσι is unproblematic and does not bear out Durante’s claim.
[ back ] 370. Construing the ‘stitching’ of φορμοῤῥαφέω with εἰς τὰ στενά, the TGL s.v. also offers “[i]n tegetem storeamve conjectum insuere. … Potest enim aliquis in tegete vel storea insui ut calescat, vel sudet.”
[ back ] 371. Hesykhios Φ no. 781: ἀγγεῖόν τι πλεκτὸν ψιάθοις ὡς κόφινος; Suidas 4.752 Φ no. 608.3: φορμὸς οὖν πλέγμα, ὡς κόφινος; Theophrastos Peri phytōn historia 2.6.11: καὶ πλέκουσιν ἐξ αὐτοῦ τάς τε σπυρίδας καὶ τοὺς φορμούς; Etymologicum Magnum 798.57–59 (col. 2233 Gaisford): φορμός ἐστι πλέγμα τινὸς μεγάλου εἴδους, ἐν ᾧ ἐκοιμῶντο οἱ πένητες, ἢ κατετίθεντο τὰ ὄσπρια; etc.
[ back ] 372. Available at http://kups.ub.uni-koeln.de/1810.
[ back ] 373. Laskaris 1517, three pages after the obverse of sheet νiiii (the last explicitly numbered before the relevant page).
[ back ] 374. Somewhat varying descriptions of its manufacture are found at Reichel 1901:101–105; Lorimer 1950:212–213; and Borchhardt 1972:18–28, 79–80. For bibliography, see Buchholz and Wiesner 1977 E 224–225.
[ back ] 375. ‘About [Odysseus’] head [Meriones] placed a helmet made of hide; it was stretched firm within with many thongs’.
[ back ] 376. From archaeology we know that the boar teeth were sewn to the hide and to each other, but since the text neglects these details one does not expect the scholiast to address them.
[ back ] 377. Presumably, the scholiast’s gloss would mean ‘woven with plaited thongs’; i.e. σχοῖνος would stand for the twisted thongs and ῥάπτω, as Durante contends, for ‘to weave’ or ‘to plait’.
[ back ] 378. Lorimer 1950:213; cf. 326. Borchhardt (1972:79n352) adds three other secondary sources who believe that the helmet itself was made “aus geflochtenen Lederriemen.”
[ back ] 379. “Abgesehen davon, dass es methodischer ist, wenn der Dichter erst die Hauptsache, die Lederhaube, darauf deren Zuthaten, Zähne, Nähte derselben und Filzfutter erwähnt, sind die bildlichen Beispiele, welche die Genannten zur Stütze ihrer Ansicht heranziehen, theilweise missverstanden” (Reichel 1901:102n1).
[ back ] 380. Although a majority of scholars believes that the platform of Hera’s chariot consisted of plaited strips, there are exceptions: “[D]oes this mean that the floor is made out of straps under tension, or that the front and sides are so constituted? Critics differ … ; the former is surely impracticable, since the leather would stretch and a foot find its way through somehow” (Kirk 1990:132).
[ back ] 381. For the intransitive value of the perfect and its affinity with the middle voice, see e.g. Sicking and Stork 1996:130–137. For intransitive middles generally see Schwyzer GG Ⅱ.230. Examples of the middle intransitive (plu)perfect of τείνω are Sophokles Antigonē 600 (τέτατο, ‘for the light just now spread over the last roots in the house of Oidipous’), Philoktētēs 857 (ἐκτέταται, ‘the man lies stretched out’), and Aias 1402 (ἐκτέταται, ‘already much time has elapsed’).
[ back ] 382. The ‘gold and silver thongs’ would be a dative of material (i.e. ‘the body made of gold and silver thongs stretched out tight’), semantically very close to means and manner. Cf. K-G Ⅱ.1.438 §425.10 and Smyth §1508c. An English example would be, ‘the tapestry stretches out with beautiful embroidery,’ where sense makes clear that the ‘embroidery’ is not strictly instrumental to the stretching out.
[ back ] 383. The point is that ‘thongs’ cannot easily be the (woven or plaited) substrate that ‘stretches out’ (in the intransitive sense). They are more plausibly the means by which the helmet is somehow stretched tight. I suggest that it was by the very process of sewing the boar tusks onto the hide that the tight stretching was effected. That ‘thongs’ (ἱμάντες) are used for sewing is clear from χ 186, where ῥαφαὶ … ἱμάντων are not, as correctly observed by Fernández-Galiano in Russo et al. (1992:253, with the concurrence of the LfgE s.v. ἱμάς 2d), ‘the seams of the straps’ (vel sim.) but ‘the seams/stitches [made with] thongs’ (i.e. ἱμάντων is a genitive of material; cf. Smyth §1323).
[ back ] 384. Cf. Harrison 1904:253–254.
[ back ] 385. ἀφωρισάμεθα καὶ τὴν τρήσει καὶ ῥαφῇ χρωμένην σύνθεσιν, ἧς ἡ πλείστη σκυτοτομική (Plato Statesman 280c3–5).
[ back ] 386. Ordinarily, besides the occasional νευρορράφος, classical authors used σκυτοτόμος and σκυτεύς.