Homeric problems and bibliographical challenges, Part 1: On the performances of rhapsodes at the festival of the Panathenaia

2018.11.22 | By Gregory Nagy

§0. No one who claims expertise in the study of Homer will ever have the last word on Homer. But those who study Homer can still hope to come up with a cumulative formulation of their own understanding of Homeric poetry, and such a formulation, published at a given time, could be considered their own last word—or, more accurately, their latest word. In my case, the closest thing to such a formulation, at least so far, has been a set of twin books entitled Homer the Classic (2009|2008) and Homer the Preclassic (2010|2009). In those two books, I figured out various solutions to various Homeric problems. And wherever such solutions did not depend on secondary bibliography, that is, on published work of experts who had confronted the same problem but whose solutions were unsuccessful, at least in my opinion, the approach I took in those two books was to avoid entering into detailed debates with such experts and to concentrate instead on trying to consolidate my various solutions by integrating them into what I have just now called here, at the start, a cumulative formulation. In taking such an approach, however, I was faced with bibliographical challenges: for example, what about situations where evidence was lacking, and where I had only a partial solution to a given Homeric problem? In such situations, I wanted to cite secondary bibliography, yes, but without losing sight of the partial solution that I had already figured out for the given problem. That is what I mean when I speak of “bibliographical challenges”: how do I keep track of a partial solution without getting bogged down in secondary bibliography? In this posting, as in other postings to come, I concentrate on such challenges. And I choose as my first example here some of the problems I encountered in figuring out the historical circumstances of Homeric performances by professional reciters called rhapsōidoi ‘rhapsodes’ at the seasonally recurring festival of the Panathenaia in Athens.

Amphora with red-figure painting of a rhapsode (490–480 BCE). British Museum 1843,1103.34.
Amphora with red-figure painting of a rhapsode (490–480 BCE). British Museum 1843,1103.34. Image via the British Museum.

§1. In the twin books I already mentioned, Homer the Classic and Homer the Preclassic, hereafter abbreviated as HC and HPC, I examined at length the details reported in ancient sources about rhapsodic performances of Homeric poetry at the Panathenaia. According to these sources, there was a customary regulation, the conventional term for which nowadays is the Panathenaic Rule, which required the rhapsodes to perform the Homeric Iliad and Odyssey in their proper narrative sequence. The clearest attestation of this regulation comes from the fourth century BCE: the relevant text is Lycurgus Against Leokrates 102, as analyzed in HPC I§§43–45. As I argued there about this “rule,” such continuity in narration made it necessary for the rhapsodes to perform the Iliad and Odyssey in relay, which involved cooperation as well as competition. By way of this argument, I offered a solution to one of the biggest of all Homeric problems, which boils down to this basic question: how to explain the narrative unity of the Iliad and Odyssey? In terms of my solution, the so-called “unitarian” approach to Homeric poetry could be explained in terms of an ongoing need, over a lengthy period of evolution in the medium of epic, to sequence the narrative by way of cooperation in the act of narration—that is, in the performance of the narrative.

§2. But this solution was only partial. What made it incomplete was the fact that we have no evidence about the duration of Homeric performance by rhapsodes in the historical context of the Panathenaia. And here is where I addressed the bibliographical challenge of tracking the many different attempts at reconstructing such a historical context. In doing so, I restricted myself in HPC 1§46n65 to citing just one secondary bibliographical source, an article by Jonathan Burgess, published in 2004. I had two reasons for such a narrow citation. First, this article by Burgess provided what was to my mind the broadest available bibliographical survey of attempted explanations of the historical context. And, second, the explanation offered by Burgess himself was for me the most persuasive of them all, even though I did not fully agree with it.

§3. In a general book on Homer published later, in 2015, Burgess recapitulates his explanation at p. 73:

A notable opportunity for performance of epic was at the Panathenaic festival at Athens, held every fourth year. Some testimony from the fourth century BCE reports regulation for sequential performance of ‘Homer’ at this event. […] Homerists are naturally eager to suppose that all of the Iliad and Odyssey were thus performed, with the regulation perhaps based on authentic texts of them; this is optimistic; it is questionable whether there was enough time to perform both epics […].

I agree with Burgess—to the extent that there has been as yet no plausible reconstruction of the events at the Panathenaia that would allow for a complete performance of the entire Iliad and Odyssey as we know these two epics.

§4. I also agree with Burgess (2004, especially pp. 17–18) when he argues that the ultimate solution for understanding the regulation of sequencing in rhapsodic performance can be found by way of broadening our understanding of the process of sequencing itself in oral poetics. In this regard, Burgess (p. 17n68) cites the relevant arguments of Cook 1999 (especially p. 159n29). I too cite these arguments of Cook in HC (1§119n106).

§5. I go further in HPC (especially at I §§167, 188, 231) by tracing the principle of sequencing in the performance of Homeric poetry at the Panathenaia back to an earlier pattern of sequencing at the festival of the Panionia in the late eighth and early seventh centuries BCE, as analyzed by Douglas Frame (2009 chapter 11).

§6. Still further, in chapter 4 of HPC, I argue that that the so-called epic Cycle, which was also part of the epic repertoire in earlier phases of the Panathenaia, shows a less continuous kind of narratological sequencing than does the epic of the Iliad and Odyssey. I have more to say about the Cycle in Nagy 2015.12.24 and, in what follows, I include some relevant aspects of the argumentation presented there.

§7. With regard to the epic Cycle, I stress here its eventual obsolescence in the epic repertoire of the Panathenaia—though this obsolescence may have to be downdated even further down than what I formulated in HPC E§32. A case in point is the image of a rhapsode as painted on an amphora housed in the British Museum, 1843,1103.34, featured on the cover of this posting. This painting is generally dated to somewhere between 490 and 480 BCE (Burgess 2004:2). (I note here in passing an observation by Shapiro 1993:95–97: he thinks that this figure is not a rhapsode but an aulode, accompanied by the aulos-player represented on the other side of the vase.)

§8. The problem remains: is there any way to imagine the performance of the Iliad and Odyssey in their entirety at the Panathenaia by the time we reach the fifth and the fourth centuries BCE? If we accept the arguments we read in the article of Burgess (2004), then it is more likely that only selections of the entire narrative sequence were performed by that time. But such selections, I would maintain, would have to operate on the same principle of narratological sequencing that had led to the evolution of an Iliad and Odyssey into epics containing over 15,000 and 12,000 lines respectively. And such sequencing was a matter of notional totality.

§9. I say it best, I think, in what I epitomize here from a general argument I once presented in another context (Nagy 2006 §§37–40):

{§37…} In the archaic era of the Panathenaia, that is, in the sixth century BCE, the idea of the epic Cycle was simply the idea of epic as a comprehensive totality: the term ‘Cycle’ or kuklos was sustained by metaphors of artistic comprehensiveness.

{§38.} In the classical era of the Panathenaia, however, that is, in the fifth and fourth centuries BCE, newer ideas of comprehensiveness had replaced the older idea. These newer ideas were now being determined by the artistic measure of tragedy. Aristotle says explicitly that only the Homeric Iliad and Odyssey are comparable to tragedy because only these epics show a comprehensive and unified structure, unlike the epics of the Cycle (Poetics 1459a37–b16). In Plato as well, the standards of tragedy are evident in descriptions of Homer as a proto-tragedian in his own right. For Plato and Aristotle, the Homeric Iliad and Odyssey measured up to the standards of tragedy, whereas the epics of the Cycle did not.

{§39.} Thus the criteria of epic comprehensiveness vary from age to age—from the archaic notion of the epic Cycle to the classical notion of Homer the tragedian. What remains an invariable, however, is the basic institutional context in which the very idea of epic comprehensiveness took shape: that context is the festival. In the case of epic as performed in Athens, that context remained the festival of the Panathenaia. In its archaic phase, to repeat, the Panathenaia featured the epic Cycle, including the repertoire of what we know as the Homeric Iliad and Odyssey. In its classical phase, this same festival of the Panathenaia featured only the Homeric Iliad and Odyssey, excluding the repertoire of the epic Cycle. Even the term ‘Cycle’ was no longer appropriate, since the epic Cycle no longer embodied the notion of epic as a comprehensive totality.

{§40.} A typological comparandum for the notion of epic as a comprehensive totality is the case of heroic epics and dramas at festivals in latter-day India: the measure of totality in the performing of these epics and dramas is determined by the ideologies of the festivals that serve as the historical contexts for such performances. Impartial observers of actual performances of epics at festivals in latter-day India have found that there are various different ways of imagining and realizing a notional totality for these epics (Flueckiger 1996:133–134). Here is a most telling formulation: “although scholars have spent considerable energy recording epic stories ‘from beginning to end’, counting the number of hours and pages required to do so, this is not how the epic is received by indigenous audiences” (Flueckiger p. 134).



Burgess, J. S. 2004. “Performance and the Epic Cycle.” Classical Journal 100:1–23.

Burgess, J. S. 2015. Homer. London and New York.

Cook, E. 1999. “‘Active’ and ‘Passive’ Heroics in the Odyssey.Classical World 93:149–167.

Flueckiger, J. B. 1996. Gender and Genre in the Folklore of Middle India. Ithaca, NY.

Frame, D. 2009. Hippota Nestor. Hellenic Studies 34. Cambridge, MA, and Washington, DC. http://nrs.harvard.edu/urn-3:hul.ebook:CHS_Frame.Hippota_Nestor.2009.

Nagy, G. 2005. “The Epic Hero.” A Companion to Ancient Epic (ed. J. F. Foley) 71–89. Malden and Oxford. For an expanded version, see Nagy 2006.

Nagy, G. 2006. “The Epic Hero.” Expanded version of Nagy 2005. http://nrs.harvard.edu/urn-3:hlnc.essay:Nagy.The_Epic_Hero.2005.

Nagy, G. 2009|2008. Homer the Classic. Printed | Online version. Hellenic Studies 36. Cambridge, MA, and Washington, DC. http://nrs.harvard.edu/urn-3:hul.ebook:CHS_Nagy.Homer_the_Classic.2008.

Nagy, G. 2010|2009. Homer the Preclassic. Printed | Online version. Berkeley and Los Angeles. http://nrs.harvard.edu/urn-3:hul.ebook:CHS_Nagy.Homer_the_Preclassic.2009.

Nagy, G. 2015.12.24. “Pindar’s Homer is not ‘our’ Homer.” Classical Inquiries. http://classical-inquiries.chs.harvard.edu/pindars-homer-is-not-our-homer/.

Shapiro, H. A. 1993. “Hipparchos and the Rhapsodes.” Cultural Poetics in Archaic Greece (ed. C. Dougherty and L. Kurke) 92–107. Cambridge.