Genre and Occasion

[[This article was first published in 1994 in Mètis: Anthropologie des mondes grecs anciens 9–10:11–25. In this online version, the original pagination will be indicated within braces (“{” and “}”). For example, “{11|12}” indicates where p. 11 of the original article ends and p. 12 begins. These indications will be useful to readers who need to look up references made in previous scholarship to the original printed version of this article.]]

In a recent work, Pindar’s Homer: The Lyric Possession of an Epic Past, I explored the relationship of epic and lyric as “genres,” reaching the conclusion that lyric is too broad a category to qualify as a “genre” in ancient Greece, while epic is too imprecise a term to apply to Homeric poetry. [1] A more fundamental problem, I argued, was the definition of poetry in an archaic social context where the technology of writing was involved in neither the composition nor the performance of any given poem or song. Within such a context, I will now argue, definitions of genre have to be correlated with questions of occasion. [2] And the occasion is captured, long-term, in a process known to the ancient Greeks as mīmēsis. [3]

My invocation of the two factors of composition and performance implies a derivation of ancient Greek poetry from oral poetry, as defined through the comparative fieldwork criteria developed by Milman Parry and Albert Lord, who was my mentor at Harvard University. [4] From Lord’s {11|12} empirical study of living oral traditions, especially the case of South Slavic epic, it becomes clear that composition and performance are aspects of the same process in oral poetry. The English noun song, along with the verb sing, expresses admirably the coexistence of composition and performance as a continuum. Further, the performative connotation of song, which is absent in the word poetry, makes it more useful to apply the word song rather than poetry to archaic Greek traditions, which do not explicitly distinguish song from poetry. [5] Clearly, it is the resonance of performance that led Albert Lord to describe the medium of the South Slavic guslar—and of Homer—as song rather than poetry. And it is for the same reason that the same word figures so prominently in his pathfinding 1960 book, The Singer of Tales.

In order to achieve a more accurate taxonomy of the earliest phases of the Greek songmaking tradition, the two factors of composition and performance must be consistently kept in mind. Only then can we arrive at a basis for considering the utility of a concept such as genre—and of a related concept, occasion.

In addressing these two factors of composition and performance I propose to bring into play a crucial work that has taken them both into account. I am referring to the 1989 book of Richard P. Martin, entitled The Language of Heroes. [6] Martin has pioneered an explicit connection between Lord’s empirical observations about performance in living oral traditions and J. L. Austin’s theories about the performative uses of language, as articulated in his book, How to Do Things with Words. [7]

As Martin demonstrates in Language of Heroes, Austin’s formulation of the performative, where you do something when you say something, meshes with Lord’s formulation of performance as the key to bringing the words of a song to life. To use Austin’s wording, song is a speech act, as Martin shows in detail with reference to Homeric poetry. Ironically, Austin himself resisted the idea that poetry could count as a speech act, and we can see clearly the reason for his reluctance: for Austin, poetry is a matter of writing, not speaking. For Austin, the dimension of oral tradition is utterly removed from his own conceptualization of poetry.

Martin’s book demonstrates not only the self-definition of Homeric song {12|13} as a speech act. It shows also that this medium is capable of demonstrating the function of song as “quoted” within its overall frame of song. That is to say, Homeric song dramatizes, as it were, the performative aspects of songs that it quotes. Ironically, the performative aspects of Homeric song itself are shaded over while the performative aspects of the songs contained by it are highlighted, including laments, acts of praise and blame, proverbs, and so on. To put it another way, Homeric song specifies the occasion of songs that it represents, or even presents, while it leaves vague any potential occasion for its own performance.

I have used the word occasion here in referring to the context of speech acts “quoted” by Homeric song. In fact, I am ready to define occasion as the context of a speech act. [8] Further, I define genre as a set of rules that produce a speech act. In offering this definition of genre, I follow Tzvetan Todorov in his Genres du Discours, Chapter 2. [9] For Todorov, genres are “principles of dynamic production” of discourse in society. [10] Here I propose to build on this most useful formulation in two ways:

1. I hope to tighten up the notion of speech act, correlating it with the specific interweaving of myth and ritual in traditional societies and dissociating it from purely philosophical considerations that center on individual judgments concerning when is a speech act a speech act. For purposes of this presentation, a speech act is a speech act only when it fits the criteria of the community in which it is being used. To determine the validity or invalidity of a speech act is to observe its dynamics within the community in question.
2. The genre, the set of rules that generate a given speech act, can equate itself with the occasion, the context of this speech act. To this extent, the occasion is the genre. [11] For example, a song of lament can equate itself with the process of grieving for the dead. [12] Moreover, if the occasion is destabilized or even lost, the genre can compensate for it, even recreate it. [13]

{13|14} An extreme illustration is the following assessment, offered by Gordon Williams, of Hellenistic poetics:

What I have described as the loss or destabilization of the occasion of the speech act can be linked with the eventual or episodic breakdown of myth and ritual as correlative self-expressions of traditional society. As I argued at length in Pindar’s Homer, ritual frames myth in traditional societies: myth is performed, and the performance is ritual. To put it another way, performance frames composition, and we cannot fully grasp the role of composition without knowing about this frame. [15] If the frame is lost, then the occasion has to be re-created by the genre. [16]

So I propose to reformulate Todorov’s convincing formulation of genre, shifting from “principles of dynamic production” [17] to “principles of dynamic re-production” of discourse in society. The shift from the wording production to reproduction corresponds to what I propose is a historical specialization of the semantics of the Greek word mīmēsis.

In the earliest attested Greek usage, as I argue extensively in Pindar’s Homer, mīmēsis (which I henceforth write in Roman letters) had meant primarily ‘re-enact’ (or, let us say, ‘re-produce’) and only secondarily ‘imitate’. [18] In my newer work on mimesis, I offer the following formula:

{14|15} When it is your turn, your moment to re-enact something in this forward movement of mimesis, you become the ultimate model in that very moment. As a way of understanding occasion, I propose to equate it with the moment of mimesis.

In later Greek usage, however, that is, by the middle of the fifth century before our era, it becomes clear that the primary meaning of mimesis as ‘reenactment’ was becoming lost or at least destabilized, and that the secondary meaning was encroaching on the primary meaning. [20] This semantic specialization of mimesis is paralleled by that of Latin imitor, imitārī. Other Latin words containing the same root im-/aem- retain semantic nuances of the older primary meaning of ‘re-enact’ inherent in imitor, imitārī. Striking examples are aemulus in the sense of ‘understudy,’ as it were (Cicero, pro Murena 61), and imāgō, which means not only ‘image’ generally but also, specifically, the ‘death-mask’ of one’s ancestor (Cicero, pro Milone 33). [21]

It is this semantic specialization of mimesis from ‘re-enactment’ to ‘imitation,’ I submit, that helps explain the discomfort of some contemporary literary critics with mechanical applications of genre-taxonomy. It also explains the desire of such earlier critics as Benedetto Croce to define any great poem as sui generis. [22] The concept of an “original” work that defies genres recovers, albeit indirectly, the concept of an archetype, of an absolute model. Such an absolute model is also presupposed by the world view of myth and ritual. In resisting the classifications of genre, critics like Croce hope to define literary works that transcend imitations of models. To use the terms that I have just introduced, such transcendent works are supposed to achieve genuine enactment, not just re-enactment.

My interpretation of mimesis as an authoritative ‘re-enactment, impersonation’ is supported by the celebrated description of mimesis in the Poetics of Aristotle as the mental process of identifying the representing “this,” as in the ritual of acting the drama, with the represented “that,” as in the myth that is being acted out by the drama: in Greek this mental process can be expressed by way of the equation houtos ekeinos ‘so this is that!’ (1448 b 17). [23] The same equation, as restated in Aristotle Rhetoric 1.1371 a 21, makes it clear that the media of representation that Aristotle has in mind {15|16} are not just the visual arts but also the verbal arts, primarily the art of songmaking and poetry as performed in drama. So long as the represented “that” remains absolute—that is, absolutized by the myth—the representing “this” remains a re-enacting “this.” [24] So long as “this” imitates an absolute “that,” it re-enacts as it imitates; the re-enactment remains primary, and the imitation remains secondary. [25] Once you start imitating something that is no longer absolute, however, you can no longer re-enact the absolute: then you can only make a copy, and your model may be also just a copy. I have just described here the general mentality induced by the destabilization of the conceptual world of mimesis. [26]

Given that mimesis was a traditional function of the khoros, a song-and-dance ensemble or “chorus,” let us turn to the Homeric Hymn to Apollo, which describes a chorus of Delian Maidens performing at a festival on the island of Delos. [27] In my earlier work, I argued that the Delian Maidens represent an idealization of choral lyric. [28] We may compare Plato’s picture of the ultimate divine khoros, comprised of the Olympian gods themselves, serving as model for all human choruses (Phaedrus 247a, θείου χοροῦ). I also argued that these Delian Maidens are presented in the Hymn as archetypes meant to be re-enacted in the local ritual context of real choral performances at Delos—in which context any real chorus-members would be equated, for the ritual moment, with the archetypal Maidens. [29] Such a reenactment of a model would be mimesis in the primary sense that I have just outlined. [30]

The Delian Maidens show the way for others to re-enact them by demonstrating their own power to re-enact all other peoples, in all their varieties. These Maidens are models of mimesis by practicing mimesis: they can repeat everyone’s voice, mīmeîsthai (Hymn to Apollo 163), and {16|17} everyone who hears the repetition will think that it is his or her own voice (163-164). [31] We may compare the traditional image of the sacrificing god, as recently studied by Kimberley Patton in a wide variety of different cultures: when gods take the stance of sacrificing, they act as models, authoritatively showing the way for others to sacrifice by being the first to do so themselves. [32]

Similarly in Alcman’s Partheneion, I argue that archetypal figures, including the primary archetypal figures named Hagesikhora and Agido, are models being acted out by real chorus-members in performances held on a seasonally-recurring basis. [33] Even their names—Hagesikhora and Agido—designate models, either divine or royal. [34] We may reconstruct a similar principle at work in the earliest stages of Athenian State theater: the real chorus-members of a tragedy would be re-enacting an archetypal ensemble that is interacting with archetypal figures of the heroic world, figures acted by actors playing roles differentiated out of the ranks of the chorus. [35]

Of course there are many roles to be played out in performance traditions, whether they are overtly dramatic or otherwise. Still, it is justifiable to consider drama, with all its ritual background, as a primary form of mimesis. Moreover, it may well be the ultimate status of drama as State Theater in Athens—and as the near-equivalent of that concept in other city-states as well—that conferred upon the word mimesis its ultimate importance and seriousness. The word’s prehistory, to be sure, suggests that mimesis once had a less important and less serious tone, since it is after all {17|18} derived from mīmos, the meaning of which never really went far beyond the relatively lowly meaning that corresponds to our own notion of ‘mime’ (Aristotle Poetics 1447 b 10). [36] The eventual importance, however, of the derivative of mīmos, that is, mimesis, is quite clear even in its earliest attested usage, in the Homeric Hymn to Apollo, where we have seen the stately Delian Maidens themselves being described as engaging in the activity of making mimesis, that is, mīmeîsthai (Hymn to Apollo 163). [37]

All this is not to lose sight of something that I have noted from the start, that the word mimesis became destabilized later on in the history of ancient Greek society, and in fact we start seeing traces of a weakening in authority as early as the second half of the fifth century, as for example in some passages of Herodotus. [38] For the moment, however, I concentrate on the surviving authority of the word, and I suggest that any eventual diminution in its authority may be simply a symptom of an eventual diminution in the authority of Athenian State Theater itself. Plato’s negative treatment of mimesis as a concept can be interpreted as a sign of the surviving power and prestige that marked the poetics of State Theater even so late as in Plato’s era, the fourth century. [39] Another sign is the attitude of another fourth century figure like Isocrates, who throughout his own extensive corpus of written work uses the word mimesis in a positive light, without implications of disapproval: in his eyes, mimesis seems to be a matter of utmost importance and seriousness (e.g. Euagoras 75, Antidosis 3). [40]

Earlier, I made the claim that genre can compensate for the occasion. But now, applying the semantics of mimesis, I extend the argument by claiming that genre can even absolutize the occasion. A striking example is the “epinician moment” as dramatized in the epinician songs (“victory odes”) of Pindar. [41] Seth Schein has analyzed Pindar’s Pythian 6 as an illustration of that moment, [42] and he quotes in this context the remarks of Hans-Georg Gadamer, who has this to say about the element of the occasional in the epinician songs of Pindar: {18|19}

I suggest, then, that any given Pindaric composition defies the realization of all the signs of occasionality that it gives out about itself. This defiance is not the result of any failure to adhere to the given occasion of real performance. Rather, it is a mark of success in retaining aspects of occasionality that extend through a diachronic spectrum. If we think of occasion as a performative frame, even a ritual frame—the frame that makes a speech act a speech act—then what we see in a Pindaric composition is an absolutized occasion. Moreover, this occasion is absolutized by deriving from the diachrony of countless previous occasions. In other words, a Pindaric composition refers to itself as an absolute occasion that cannot be duplicated by any single actual occasion. Only an open-ended series of actual occasions, occurring in a continuum of time, could provide all the features of an absolutized occasion.

Where a medium refers to itself over time, the referent, which in this case is the medium itself, may not keep up with its own development. That is to say, the medium in real life may change over time, but the medium as referent may stay relatively the same. I have elsewhere referred to this phenomenon as diachronic skewing, in that “self-references in archaic Greek poetry may be diachronically valid without being synchronically true.” [44] I cite, as a premier example, the case of the self-accompanying singer Demodokos in Odyssey viii. [45] It goes without saying, however, that the medium may refer to itself not only in terms of earlier stages of its own existence: it can also overload, as it were, references to its own occasion, so that all the given self-references could not possibly fit any one time and any one place of performance. I will presently cite a salient example, the case of Pindaric references to an ensemble of revelers, or kômos, who are celebrating in song the moment of victory.

Before we take up this example, however, I propose to deepen my {19|20} argument on the question of genre and occasion. Up to now I have been claiming, in the most general terms, that genre can compensate for the destabilization of occasion, and that such a compensation corresponds to the specialization of mimesis to mean mere ‘imitation.’ When you compose in a given genre, according to the surface logic of this argument, you may be simply imitating your predecessors.

But now let us explore beneath the surface, turning to the insights of Wolfgang Rösler on the social contexts of archaic Greek songmaking. [46] He objects to what he calls the “immanentistic” [47] attitudes of those who interpret references to the self in archaic Greek songs as an exercise in creating a Rollencharakter, and who view the poetic “I” as a function of poetic conventions. [48] Rösler tries to navigate between the Scylla of the “autobiographical I” and this Charybdis of a “fictional I,” positing what seems to be a “historical I” as a compromise.

I propose to build on Rösler’s formulation by positing, as middle ground between “autobiographical I” and “fictional I,” not a “historical I” but a “reenacting I.” [49] For me, the major problem with the “immanentistic” solution is the concept of fictional in the construct of a “fictional I.” A “fictional I” is a matter of mimesis as mere imitation, to be contrasted with my proposed construct of a “re-enacting I.” A term more apt than “fictional I” might be “generic I,” provided we are allowed to understand genre as a formal device for recapturing the authoritative occasion. [50]

Pursuing the elusive concept of a “generic I,” with the goal of reaching a more precise concept of a “re-enacting I,” I am ready to accept Rösler’s dictum that the identity of Alcaeus as a poet is a function of his social group, his hetaireía, which he seems to be addressing in his poetry: “ohne Hetairie kein Lyriker Alkaios.” [51] I only need to add that the hetaireía is diachronic—and so too, for that matter, is the persona of Alcaeus.

With these adjustments, I can follow Rösler in arguing that there is no “fiction” per se on the occasion of, for example, an interchange {20|21} between Alcaeus and his comrades. I would add, though, that there is indeed reenactment. There are model situations to be acted out in the song. There is, in a word, a ritual frame, a performative frame.

Once we distance ourselves from the idea of a “fictional” dimension in the performance of archaic Greek song, we acquire a ready counterargument to the idea, as we see it take shape in Plato’s Laws (Book III 658a–659c, 669b–670b), that Athenian State Theater is an exercise in theatrokratía, in that Theater supposedly appropriates real genres from real occasions and makes them make-believe. [52] Let us take the example of the “genre” of the thrênos or lamentation. As Nicole Loraux points out, the condemnation of mimesis as a representative of theater in general and of tragedy in particular is specifically correlated in Republic Book III (395d-e) with the condemnation of imitating women’s behavior, especially when it comes to lamentation. [53] For Plato, a lament must implicitly be a lament in “real life,” where real living persons mourn for a real dead person in a song that marks a real occasion, while a thrênos in tragedy is supposedly an imitation, a fiction, where make-believe persons mourn for a make-believe dead person in a song that merely imitates a real occasion.

As we now see from a deeper reading of mimesis as re-enactment, however, the songs of lamentation in State Theater are really archetypal, as it were, with relation to the “real-life” laments of “real-life” people. Far from being an imitation of “real-life” genres, the dramatized thrênos of Athenian State Theater is intended as a model. [54] There is an authority inherent in mimesis, and this authority confers an absolute status upon the person or thing to be represented.

The theatricality of the song of lament or thrênos in Greek tragedy can be compared with the intense dramatization, even theatricalization, evident in Pindar’s victory songs. In this case again, the song as performance {21|22} presents its own occasion not as a mere instance of the genre but as a model of the genre, even an archetype. Here I return to the point that I made earlier, that the typical Pindaric victory song can overload, as it were, references to its own occasion, so that all the given self-references could not possibly fit any one time and any one place of performance. And I come finally to the promised example of Pindaric references to the kômos.

There is a convention, commonly found in the victory songs of Pindar, which allows the khoros or chorus, an ensemble of performers who ostensibly sing and dance such a song, to be described as a kômos, an ensemble of revelers (e.g. Nemean 3.1–12). In the victory songs of Pindar (and of Bacchylides), it is in fact regularly the word kômos—and not khoros—that designates the joyous ensemble that celebrates the victory. This fact has led one scholar to argue that the kômos was a performing ensemble that was exterior to the performance of the victory ode. [55] Moreover, since all aspects of group performance are attributed by the victory songs themselves to the kômos, this scholar infers that these songs of Pindar must be a matter of solo performance. [56] This view of Pindar as solo performer of his epinician poetry is shared by another scholar who argues that the “I” of Pindar’s victory songs cannot be the chorus—let alone the kômos—and that therefore this “I” must be Pindar himself, performing solo. [57] Others have counterargued that the “I” of Pindar’s victory songs can indeed refer to the chorus, so that these songs must be a matter of group performance. [58] The methods used by both sides in this debate, as yet another scholar points out, are remarkably similar: “They focus on problematic passages, that is, on suggestive references to solo or choral performance, and read them as stage direction.” [59]

{22|23} As an alternative solution, this other scholar proposes what she calls a “slightly different” method: “Instead of understanding references to solo or multiple voices as literal indications of performance conditions, I will examine them as examples of a rhetoric of performance, indications of how the poet locates himself in the celebration, and as expressions of Pindar’s totalizing claim for his poetic discourse.” [60]

I can accept most of this formulation, but I disagree with the emphasis on Pindar’s own personal poetic program as the source for the rhetoric of performance. Yes, Pindar may superimpose his own “I” on the activities of the chorus, making them act out his role as a master-poet, but we must remember that this role is primarily a function of the victory song itself. The epinician program imposes the solo role [61] of an authoritative xenos or guest-stranger who is obliged by the rules of xenia or guest-friendship to praise a counter-xenos, his host:

The chorus of the victory song can re-enact such a solo role, in which case the “I” is the laudator as guest who owes praise to the laudandus as host, but it can also re-enact a group role, in which case the “I” plays the part of an exuberant ensemble of spontaneously celebrating youths, a kômos. [63] In songs other than victory odes, such as paeans, it is generally agreed that the {23|24} “I” of Pindar’s compositions regularly stands for the whole chorus that is celebrating. [64] Thus we cannot really claim that the “I” of Pindaric poetry is a totalizing Pindar. All we can say is that the “I” of victory songs composed by Pindar—or of those composed by Simonides, Bacchylides, and others as well—is an exponent of epinician poetry. [65]

Granted, the kômos of historical times seems too “primitive” an institution to be equated synchronically with the khoros of a Pindaric epinician. And yet, the chorus of Pindar’s victory songs plays the role of a kômos—or, better, it re-enacts a kômos. If the kômos, as we see it represented in such songs as Alcaeus F 374 V, is indeed “primitive,” then all the better for purposes of epinician mimesis, since Pindar’s own poetry represents the proto-epinician as a kômos improvised once upon a time by a prototypical poet, Archilochus (Pindar Olympian 9.4). [66] In other words, epinician represents the kômos as its own prototype, to be re-enacted in the here and now of the victory song’s mimesis.

In my earlier work, I had analyzed Pindar Nemean 7.75–76 as an example of the stylized exuberance of a kômos representing the epinician moment:

And I translated as follows:

Homeric poetry, as I argued, already shows traces of this convention, as in Odyssey xiv 462–467:

And I translated as follows:

{24|25} The spirit of merriment in a kômos is lost on most translators of Pindar, as in this rendition of Pindar Nemean 7.75–76:

In sum, it seems to me unnecessary to argue that the internal references to a kômos in Pindar’s victory songs reflect a reality that is external to the performance of these songs. Even in and of itself, the kômos has great mimetic potential: in historical times, according to the definition of a historian who has studied the institution closely, the kômos was a “ritual drunken riot at the end of the symposion, performed in public with the intention of demonstrating the power and lawlessness of the drinking group.” [70] It goes without saying that a Pindaric choral performance must be a far cry from a drunken riot, even it be a ritual drunken riot. Still, Pindar’s victory ode may ritually express the exuberance of its own celebration by representing itself as a spontaneous kômos. The Pindaric chorus can reenact the role of the kômos, and the diachronic skewing inherent in this reenactment can only reinforce the sense that something primordial is happening. The occasion produced by the genre of the victory song thus becomes prototypical, even absolutized. Such is the power of mimesis.