Chapter 1. Performance, Speech-Act, and Utterance

{1} Does it really matter whether or not Homer's Iliad is a piece of oral poetry? In the final analysis, no. Even if the 15,693 hexameters printed in T. W. Allen's Oxford Classical Text happen to represent the exact transcription of an actual performance by one "singer of tales" from the eighth century B.C., we still do not have an oral Iliad, because the poem has, somehow, become a text; and that has made all the difference. To put it another way, our Iliad is no longer an action, as it must have been if it was ever an oral composition-in-performance. Instead, it is an artifact. [1]

To concede that our Iliad is a text, however, does not excuse us from making the effort of imaginative reconstruction to interpret the poem as closely as possible in its own context. Athenian drama, after all, was never intended to be read simply as isolated texts, and few scholars today would dare study it without some attempt at understanding the circumstances of dramatic performance. The fresh emphasis on a sociocultural reading of tragedy and comedy, in the work of such critics as Helene Foley, Simon Goldhill, Nicole Loraux, C. P. Segal, J.-P. Vernant, and Froma Zeitlin, stands in disconcerting contrast with the most recent tendencies among professional readers of Homer. Whereas tragedy has been cut loose from the bonds of New Criticism, and approached more cautiously in all its strangeness, Homer has become for some a haven safe from critical storms. A new {1|2} reaction has set in against the work of Milman Parry and other exponents of an "oral" Homeric poetry—or, we should say, against a certain portion of this work, for many of Parry's insights are ignored by the new critique. The oralists' concern with technique has earned them the label "Formalists," and their emphasis on the traditional nature of Homeric craft has prompted the charge that they neglect the individual genius of the poet. [2] Of course, such criticisms were leveled at Parry from the outset, not surprisingly given the climate of Anglo-American literary study at the time. More puzzling is the resurgence today of this reactionary criticism, half a century after Parry's seminal work. It is disturbing that young philologists such as David Shive find it necessary to attack the alleged flaws in Parry's first publications, and to defend the "creativity" of Homer, while failing to reexamine the very idea of what creativity in an oral tradition might mean. [3]

This wave has been building; in 1978 David Bynum could note a "palpable ennui" among scholars first attracted to the Parry view, "as the practice of formula-counting has become more common, lost its first blush of novelty, and for the most part failed to deliver the innovations in the substantive understanding of oral traditions which were expected of it from the first." [4] The reaction has been aided to some extent by the honest appraisals of Homeric tradition produced by philologists who followed the Parry direction. One turning point came as it was gradually recognized that "oral poetry" and "formulaic poetry" were not convertible terms, and that the "orality" of our poems must remain an open question. In one of his last articles, Adam Parry subjected his father's work to a critical reappraisal. He concluded that although the style of Homer "shows many features of a style originally created for oral composition," the oral composition of the two epics "probably cannot now be proved." [5] From another perspective, the apparent uniqueness of the Iliad, at least among the European epic traditions, has been noted by British scholars generally sympathetic to Parry's work. J. B. Hainsworth remarks that "the {2|3} greater architecture of the poem appears to be unlike typical oral poetry. It is more like drama, and therefore more amenable to the canons of orthodox criticism." [6]

Yet should we practice orthodox criticism simply because the poetry will permit it? The temper of some contemporary Homeric study answers "yes." Thus, even Bernard Fenik, whose earlier studies of typical scenes did much to expand an oralist perspective, writes in his 1986 book that "the artistry of the Greek epics draws them back from that alien strangeness where formalist studies have isolated them. The Iliad and the Odyssey belong instead within literature's conventional ambit and they respond, with certain adjustments, to familiar and demanding criticism." [7] He speaks of forcing attention "back to the poetry itself" and contends that side-by-side comparison with written poetry shows Homer excelling because the poet successfully employs traditional compositional devices, such as prolepsis and juxtaposition, which "belong to the art of good storytelling." [8] I do not have time to trace the connections between such a view and the antique arguments as to whether Homer is as good as Vergil. [9] Ironically, the phrase "alien strangeness," which in Fenik's view denotes a critical wasteland, describes for some Hellenists an important facet of Greek poetry.

Another strategy for rescuing Homer from the oralists has a more contemporary ring: we are told by the critic Martin Mueller that "whether audiences read the Iliad or listen to it, they must construe and respond to the meaning of the words, and this act of making sense may justly be called 'reading.'" [10] Again, the "literariness" of Homer is vindicated, for whatever reason. I do not claim that these "readers" of Homer obtain erroneous results from their method: both Mueller and Fenik have produced thoughtful, graceful essays, comparable to the acme of New Criticism (Whitman's 1958 book Homer and the Heroic Tradition). But I do regret the dismissal of the Parryan perspective because it opens the door for atomistic explications in the nineteenth-century Analyst mode, and because it does not do justice {3|4} at all to the wealth of insight gained from the post-Parry work in so-called oral literature. [11] Perhaps too narrow a focus on the definition and description of "oral literature" has produced ennui. The term itself perpetuates an unhelpful stance, as Michael Herzfeld notes: "Even the recognition of folk texts as Oral literature' . . . merely projected an elegant oxymoron: by defining textuality in terms of 'literature,' a purely verbocentric conception, it left arbitration in the control of 'high culture.'" [12] Inevitably, the text-centered nature of academic study shifts the emphasis from "oral" to "literature," from performance to script. In what follows I intend to redress the balance.

Only within the past few decades have social anthropologists, folklorists, linguists, sociologists, and a few literary critics begun to detect the crucial importance of performance in the study of verbal behavior. One of the earliest and most influential books in the field was Erving Goffman's study of personal interaction routines, published in 1959, the year before Lord's Singer of Tales. Goffman borrowed the concepts of actor and role from dramaturgy and game theory in order to show how everyday communication, and the more stylized communication of art and "performances" in a strict sense, share essential features. To use Goffman's definition, both types of communicative "performance" represent "the activity of a given participant on a given occasion which serves to influence in any way any of the other participants." [13]

This approach, which sees verbal art as part of a spectrum of human communicative performance, has led to significant research into discourse strategies. We have learned that orally produced "texts," artistic or not, establish cohesion by a number of means undeveloped in written texts: they involve the audience through direct quotation and increased use of deictic pronouns and present-tense verbs, or they ease comprehension by reduced sentence complexity. At the same time, written communication can be seen as often elaborating "strategies associated with speaking, in order to create involvement." [14] Such findings regarding everyday communication surely have relevance for the Homerist's judgments concerning "orality" in the {4|5} poems. With the notion of performance as their guide, a number of linguists, reacting against the abstract syntactic descriptions of Noam Chomsky, have begun to investigate a grammar of context, which aims at defining the role of pragmatics in determining morphology, syntax, and semantics. [15] In this perspective, discourse analysis must precede grammatical study: the sociolinguist William Labov, who perfected the study of language use during "speech events," puts the case trenchantly: "The student of his own intuitions, producing both data and theory in a language abstracted from every social context, is the ultimate lame." [16] Nor has the artfulness of "ordinary" discourse remained solely of interest to sociolinguists; performance has become a key concept for philosophers of speech-act theory, whose insights will prove helpful when we come to examine "winged words" in Homer. [17]

Along with a new appreciation for the precious individuality of each verbal performance, a second variable has gained attention in this research—the audience. Whether it determines the "success" of a speech-act, or affects a salesperson's pronunciation of the phrase "fourth floor," the presence of a particular audience in the speech situation is crucial. Oral communication must know and face its audience. Written texts, while not responsible to one specific audience, are nevertheless defined by their readers, as even the most text-centered literary critics will acknowledge. The recent attempts to "recontextualize" literature must adopt a view similar to the oralist's: "There is no universal listener. There are only individual listeners, real or fictional, but all time-bound." [18]

Where we can observe the audience during performance, as in living oral poetic traditions, its large role contrasts enormously with that of the reader. In some African communities, as Daniel Biebuyck notes,

the performances of the epics are highly complex events which must be viewed as total social and artistic phenomena. Besides the actual bard and his aides (eventually including the apprentices), there is a diverse and sometimes large, actively participating audience. There is a constant {5|6}interplay among these three categories of participants. The actual presentation of the epic narrative is enhanced with musical performance (one or more musical instruments, eventually of different type); appropriate costumes and adornments; singing, chanting, praising, dialoguing; dancing, gesticulation, handclapping; dramatic reenact-ments; and gift-exchanges. [19]

At times, an audience can inspire the performing bard to digress and expand, as when some African poets fill their compositions with talk of themselves and their families, teachers, proverbial wisdom, and the audience itself. [20] Audiences can force a poet into an agonistic stance either against other poets (as happened to the modern Cretan bard Barba-Pantzelyo) or against the audience itself, as in the case of a Romanian epic singer who accused his backup musician of falling asleep, in such a way that the audience knew its own attention was being criticized. [21] A Philippine storyteller (who faces a wall, not his listeners) might be encouraged by his audience to keep on in one vein, or to get rid of characters that bore them, speed up narration, or tell another story with a specific set of protagonists. Yet the performer is still in control, guided by his audience. [22] A Lao audience, sitting close to a mohlam performance duo, will shout approval, and even imagine themselves individually as the "thou" addressed by one performer to the other, answering the poetic lines being traded on "stage. " In turn, a performer might address certain sad verses in this genre of lyric drama to a member of the audience who she knows is leaving the region. [23] The reports about nineteenth-century performances of the Central Asian Manas epic, among the Kirghiz, stress the vital bard-audience bond: "He sought to discover a theme that would suit their mood, while they in turn incited him to feats of virtuosity with their applause. At such peaks of fervor, when the bard and his listeners were as one, the patron would rise, peel off a costly robe, and throw {6|7} it to the triumphant performer." [24] Nothing prevents us from attributing some or all such features of audience behavior to the context of Homeric performance. If rhapsodic performance, as described in Plato's Ion, is at all traditional, Homer might well have resembled the African scepter-carrying epic singer; the stories about Homer's life might preserve memories of a time when poems could be objects in gift-exchange. [25] Even the rhapsodic habit of "explaining" Homer appears to be more traditional when we examine other epic performance: Dennis Tedlock, in his studies of the Mayan epic Popol Vuh, as also of Zuni Indian poetry, has shown how poet and audience interact during performance and thereby actually interpret the poem in tandem. The text becomes simply the flexible springboard from which the performer continually takes off and to which he or she returns—it has no rigid fixity, any more than any other actual oral poem, even though its content is allegedly sacred myth. [26]

One of the most sustained and accessible studies arising from the new performance-centered approach to verbal art is by Elizabeth Fine, a folklorist. It surveys the work done since the 19605 and contains her own elegant demonstration, using fieldwork with Southwestern storytellers, of how meaning emerges only through performance. Time and again the observer of performances can note that timing, gesture, voice inflection, tempo, proximity to the audience, the past relation of a particular performer with his or her audience, the setting, the season, the time of day—are factors that determine the meaning of the actual words spoken by a performer as much if not more so than the literal meaning of the words themselves. This is to say that it is the performance, not the text, which counts. [27] {7|8}

The publications of Milman Parry and Albert Lord record their respect for the individual performance, a knowledge gained from intensive fieldwork as yet unequaled by students of Homeric poetry. Richard Bauman, one of the leading exponents of the contemporary performance approach, recognizes that Lord's Singer of Tales "opens a range of productive questions . . . the constitution of the individual repertoire, the acquisition of performance skills, the individual's performance career, differences across performances of what a singer considers 'the same song' and so on." [28] Yet, perhaps because many of their studies were read by scholars who knew only texts, and were less interested in the mode of performing these "texts," Parry and Lord's work was first mined by Homerists and medievalists primarily as evidence for certain verse-making techniques, notably the existence of the formula. [29] This circuitous route, however, leads us eventually back to the Homeric performance itself. To put it briefly, on the one hand, it is impossible not to believe that the Iliad comes from a long tradition of oral performance—to this date Parry's demonstration has remained, in its essentials, undisturbed. [30] On the other hand, given what we now know about actual oral verbal art, we can more fully appreciate the width of the chasm separating readers of Homer from watchers and hearers of an archaic performance. Readers of the poems possess only the husk of a performance, this collection of hexameters, the program, only, of the event. [31]

There is an alternative strategy in contemporary Homeric criticism that welcomes comparative study of performances to recover the feel of the Greek epic tradition. Although I am in favor of such an approach {8|9}, as I have shown thus far, it clearly involves scholars in the riskier enterprise of finding increasingly "exotic" analogues, which only deepens the resistance of the "literary" Homerists. Balinese shadow theater, for example, has recently been offered, by Agathe Thornton, as bearing some resemblance to Homer's work. Like the Homeric epics, it employs a Kunstsprache—the Balinese have in fact a separate language for the gods on stage. Shadow theater requires remarkable stamina on the part of its lone performer, the dalang, and I can imagine a really good Homer sweating profusely as he performed. Of course, the potential for imperfect understanding, accidental likeness, and misapplied comparison increases as one moves farther from traditional narrative verse into an exploration of other living oral genres. [32] Shadow theater really makes a better analogue for such Greek traditions as Old Comedy and Karagheozis performance than it does for the Iliad. But even were this not so, the value of any such analogues lies in their suggestive power; they never constitute proof concerning any points about the orality of Homer or the original mode of performance. As living oral genres die off, we need to be even more clinically scrupulous in our interpretations. It appears there are good living traditions that are generically akin to the Homeric—the Mongol and Kirghiz, largely neglected by non-Soviet scholars and demanding investigation. [33] Meanwhile, on home ground the first step for re-imagining Homeric performance is the Alexandrian scholarly principle—to explain Homer out of Homer. Internal analysis must precede any external comparison.

Where can we turn, within Homer, to find performance? Not, I believe, to the poets. Most of what can be said about the significance of Demodokos and Phemios, the so-called bards, and Odysseus, bard manqué, has been said by now, some of it thoughtfully. [34] This path has been a useful dead end; it has been worth it if only because we have learned to reread the Odyssey as a narratological labyrinth. I {9|10} cannot see that it has had any impact on Iliad studies. [35] Ultimately, the evidence is too thin for us to draw conclusions about Homer from his depictions of bards.

If we start with the idea that Homer was an oral poet, it seems to me essential that we should delve more deeply and concentrate not on poets in the texts but on orality itself, to look at the very notion of speech within the poems to discover the parameters of this very basic sort of performance. Then we can extend the notion of performance, or rather, recapture what Greeks considered to be a "performance, " and compare it with our own notions. The task is ethnographic; the society to be observed happens to be extant only in the remnants of its poetic production. Yet some reconstruction can be attempted. To my knowledge, this has not been done yet; I find the task all the more compelling precisely because workers in the other fields I have mentioned now seem agreed in stressing the importance of performance as the distinguishing feature of all speech events. We know what Homer says about the power of memory and of art: Odysseus is an emblem for their dual potency. But what does this poetry say about its very stuff, words themselves? And can this tell us something about the poetry?

Performance and Speech Taxonomy

We should begin with words for speech itself. Again, Lord provides valuable hints from field experience, when he relates that some modern singers claim to repeat a composition "word for word" like an earlier song, yet are shown to have made wide changes by the transcripts of their performances. In fact, as Lord and Parry found, the idea of a single "word" made no sense to their informants, who regularly used the same term to mean an utterance of any length. [36]

We can press further this insight about terminology, for metalanguage—talk concerning talk—is highly language-specific. That is, the spectrum of speech, like the spectrum of colors, can be described in various ways by different languages. Irish, for example, denotes with one adjective—glas—the shades of color that English distinguishes as green, gray, and blue-green. As regards speech notions, {10|11} English speakers have at their disposal an estimated 4,800 verbs and phrases describing the act of speaking (along with its enactment, motivation, and valuation), but they have only a relative handful of nouns and general terms for speech-acts. [37] By contrast, the villagers of Chamula, a Tzotzil-speaking Mayan community in southern Mexico, subdivide into dozens of categories the spectrum of speech. The Chamulas classify all speech into either ordinary talk; words that come from the heart (e.g. political talk, angry speech, courting language); or "pure words." The third category comprises so-called recent words and ancient words—and ancient words, in turn, encompass many kinds of talk, ranging from "ancient true narratives of the First Creation" to "prayers for evil people, Protestants, witches, murderers, and thieves." Gary Gossen has demonstrated that these native categories are implicit in the everyday life of the village, although it requires the skill of the ethnographer to uncover the total system through which the terms obtain their meanings by structural oppositions. [38] It is significant for our purposes that the Chamulas cannot even talk about speech without indicating the speech-genre to which a given piece of discourse belongs. This means that it would make no sense to investigate, for example, Chamula storytelling, without first finding out what kind of "speech" word the Chamulas use to describe it. Recent studies in the ethnography of speaking confirm that this is the norm for many cultures. [39] The western Apache, for instance, have a system of speaking/genre terms which is similarly highly specific: the ethnographer is able to define a term such as "wise words," goyaayo yaiti, as "a distinctive speech genre associated with adult men and women who have gained a reputation for balanced thinking, critical acumen, and extensive cultural knowledge. " [40] The term is inexplicable without knowledge of the kind of performances it refers to and the status of the performers. In more abstract terms, the study of "oral literature," linguistics, and anthropology cannot operate as separate disciplines when it comes to understanding tradi-tional {11|12} folk cultures such as these. The same must apply to study of archaic Greece.

The ethnographer of speaking who attempts to reconstruct Greek talk about words, then, will not be surprised to find a folk taxonomy of speech that is askew from the standpoint of our own notions. The difficulty lies in recapturing the semantics of words for speech when we have no native informants and only poetic texts. Homerists have a model for overcoming part of this problem: I refer to the brilliant work of Leonard Muellner, which explains the problematic semantics of the speech-act verb eukhomai—boast/pray—by analyzing its formulas in the text of Homer. [41] I find Muellner's method useful for analyzing the two terms that demand attention when we turn to words for "speech" in Homer—namely muthos and epos. In hopes of recapturing the intricacies of the oral poetic world behind Homeric verse, I have investigated these two words in their context and can now redefine the words as follows: muthos is, in Homer, a speech-act indicating authority, performed at length, usually in public, with a focus on full attention to every detail. I redefine epos, on the other hand, as an utterance, ideally short, accompanying a physical act, and focusing on message, as perceived by the addressee, rather than on performance as enacted by the speaker. In short, I believe the analysis of speech terms within Homer offers us an immediate entryway into notions of performance, through those speeches in the poems which are called muthoi.

In what follows, I shall explain how I arrive at this reconstruction of notions regarding speech in archaic Greek. The dichotomy of speech-performance and utterance can be used, along the way, to answer such questions as what kind of speech-act the epic is, and whether "winged words" is just a convenient filler or a meaningful phrase. In Chapters 2 and 3, as I examine the poetics and rhetoric of the major types of Iliadic performances, it will be seen that the word muthos comprises a range of speech-genres similar to that of Chamula "words for heated hearts": political talk, angry speech, and affectionate recollection. Heroes can be distinguished as performers by their ability in these genres. Chapters 4 and 5 will focus on one heroic performer, Achilles, and my conclusion on another—the poet Homer.

Before beginning with the semantic distinctions between these two {12|13}words, however, it is worth pointing out why these in particular should be selected for investigation in the first place, out of all the Homeric words for voice and speech. A major reason is that they are the two most important terms designating speech of some sort which, as it happens, also become names for genres—one poetic, one applicable to many narratives—in post-Homeric times. Determining the semantics of these two words can illuminate for us the meaning of "epic" and "myth. " Furthermore, their semantics diverge so widely and rapidly after Homer that one must wonder whether signs of the difference in meaning can already be discovered in Homeric poetry. As Détienne has shown, the meanings "tale, fiction, lie" for the word muthos appear to be as ancient as Hesiodic poetry. On the other hand, it seems that the word epos develops the sense "poetic utterance" and "hexameter verse" equally early. It then becomes marginalized in Greek, to be replaced by logos in the sense of a "single word. " [42] Oddly enough, this semantic divergence, attested not much later than Homer's time, is absent at first sight from Homer, in whose poetry both muthos and epos seem to function as synonyms meaning "speech, word." The synonymity has been widely accepted. Cunliffe's Homeric lexicon accords the two terms nearly identical definitions. Both words, the lexicon records, mean "something said, an utterance, a word; speech, discourse, words; speech as distinguished from action," and "the sense of the word colored by the context." The last category is a catchall lexicographic net whereby both words can be translated as command, counsel, injunction, and so forth— with no difference in meaning. [43] And yet, given what we now know from Milman Parry about the economy of Homeric diction, and about the specificity of Homeric vocabulary, from the work of Benveniste especially, such massive synonymity in the case of two words denoting an essential human act seems suspicious. [44]

{13|14}We must rewrite the dictionaries, by looking afresh at the exact contexts, associations, and disjunctions in which these words play a part. When we do pay attention to context, synonymity recedes. Or rather, it emerges that epos resembles muthos only in a small fraction of cases, where another overriding concern bleaches out the sharp contrast between these two terms that can be found in the majority of their occurrences. In the singular number, when they occur inside or out of formulas, epos and muthos occupy opposite ends of the speech spectrum. Their semantic opposition can be sketched out in terms of four pairings. I hesitate to call them polarities since, in some instances, the pairings approach complementarities; as we shall see, in one major group, formulas using epos and epea actually fill a lexical gap left by formulas employing the word muthos. All in all, however, it is most helpful to view the two words as each expressing a completely different focus on the phenomenon of speech.

Because this is a book about the Iliad, my examples will be drawn in the main from that poem. This is not to say that the system I am about to sketch does not apply in the Odyssey. I have verified it there as well. Yet a general working principle, which I borrow from Bolling, is that Homeric discourse grammar should consist of three segments: a description for each poem, and a third comparing the two. [45] What follows then represents my contribution toward a grammar of the Iliad.

The Speaker and the Spoken Word

For the first pairing of contrasts, I must turn to the well-known model of the speech event as refined by Roman Jakobson. [46] Communication involves a speaker, an addressee, and a context; a message to be spoken; a code to carry the message; and contact between the parties. If we apply this system to the words for "speech" in Homer, muthos can be viewed as the kind of speech that focuses on the speaker. This connection between the viewpoint of the person talking and the talk labeled muthos is so close and consistent that those few scholars who have tried to trace differences between muthos and epos have invariably defined muthos as some form of thought. Thus Ebeling {14|15} glosses the word "sermo intimus"; Hofmann derives the senses of "fable" and "opinion" from an original meaning "cogitatum"; Four-nier followed along similar lines, giving the definition "pensée qui s'exprime, le langage, l'avis, langage intérieur"; and even Chantraine seems to feel this way about the term: "Suite de paroles qui ont un sens, propos, discours; associé à έπος qui désigne le mot, la parole, la forme." [47]

It is certain that, in the language of the Iliad, muthos is associated with words for thinking. For example, Paris in the assembly of the Trojans alleges that Antenor knows how to think of another and better proposal than the muthos he has just made (that Helen should be returned):

"You know how to think of another muthos better than this one."


Earlier in the poem, when Antenor recalls during the teikhoskopia the speech styles of the Achaean heroes who came to Troy, he associates the word muthos with well-made plans (mēdea):

"But when they wove speeches (muthoi) and plans for all, Then, you know, Menelaos discoursed in running fashion, Speaking little, but very clearly, since he is not much with words

( polumuthos )

nor one to cast words about. And, indeed, he was younger. But when indeed Odysseus much with wiles (polumētis) arose he'd stand, he'd look down fixing his eyes on the ground. ..."


The same passage shows a clear correspondence between the adjectives polumuthos and polumētis, "with much clever intelligence." "We recall that mētis, in turn, bears a close relation to mēdea in Greek. [48] Elsewhere, muthos is correlated with words for counsel (boulē) and intellect (noēma). Adjectives such as "painless" modify muthos but {15|16} not epos, and denote speech that is meant to have an active role in resolving a crisis, as when Polydamas addresses the Trojans: "And the painless word pleased Hektor" (muthos apēmōn 12.80, 13.748).

Close as is the connection between muthos and "intent," however, the word always refers to actual speech accompanying a speaker's thought. Thus, one can never justify translating the word as "thought." [49] This problem, by contrast, never arises with the word epos. Unlike muthos, this word has a clear Indo-European derivation, which connects it with the root seen in Greek ossa and opa and in Latin vox. [50] The root refers to voice, and this original sense survives in epos. A muthos focuses on what the speaker says and how he or she says it, but epos consistently applies to what the addressee hears. We can see the root meaning in a number of places in the Iliad, as when Hektor does not "fail to recognize the epos" of the goddess Iris (2.807), and Andromakhe says she wants to be out of hearing of Hektor's death: "May the epos be away from my ear" (22.454). [51]

Given the etymology of epos, we can see that a consistent image underlies Aeneas' words to Achilles at 20.203-204:

"We know each other's genealogy, we know the parents, from hearing the famous epea of mortal men. "

Literally, the adjective proklut' means "heard before." Only in the context of oral tradition can this word come to mean "famous, " as happens also with the noun kleos, "glory." Gregory Nagy has explicated Aeneas' speech in Book 20 showing that epea here in fact refers to poetic utterances, in the form of traditional narratives about Aeneas. [52] I would underline in this passage the significant distinction between "telling" blameful things (oneidea, line 202)—an act described with the verb muthēsasthai—and hearing utterances (epea). Once again, muthos is associated with the speaker's action in giving a message, whereas epos refers to the transmission of the message, the end-product of the speech process.

The notion of product (as opposed to action) seems to be embedded {16|17} in the word epos. First, it is an inanimate neuter noun (as opposed to the animate noun muthos). [53] In the few places where speech is described by means of both words, it appears that the term epos refers to the smallest elements of connected discourse, to single words or emergent sounds. Antenor's description of Odysseus' rhetoric mentions voice (opa) and words (epea) in the same breath, and vividly compares the latter with winter snowflakes (3.221—22). The image is that of a powerful, silent natural phenomenon itself composed of single powerless parts. The description of the speech style of Thersites (a foil for Odysseus) concentrates on his inability to organize the discrete small units of his talk, the epea:

"Thersites alone, of unmeasured speech, still brawled, he who knew many disordered epea in his mind, rashly, not according to good order."


During one of the rare moments when fighters discuss speech for more than a few lines, we see again that muthoi are the large units, epea the small. Toward the end of the challenging speech to Achilles, in which Aeneas refers to "famous words" that have acquainted the warriors with one another's deeds, he calls for a fight, to put an end to childish talk. The contrasting mention of "speech" versus "deeds" is a frequent Homeric topos. Only at 20.246-250, however, is the first part of the contrast further subdivided.

"We both have insults to speak formally (muthēsasthai), many of them, nor could a ship with a hundred benches bear the load. The tongue of men is pliant, in it there are many muthoi of all types, of epea there is much share-land here and there. Whatever epos you say you hear in return."

"Insults" (oneidea) are, in this image, the weightier form of speech: like goods on a merchant vessel they are tangible, substantial. Note that the verb Aeneas uses for "speak (insults)" is derived from the word muthos. As we shall see later, the verb is markedly more restricted than "say" (eipein, from the root in epos): it always means {17|18} "speak in detail and at length." Furthermore, the genre comprising threats and insults, (oneidea), is one of the few speech types introduced by the word muthos. [54] In short, the image of "heavy" speech that Aeneas applies to "insults" should be extended to muthoi as a whole. [55] Contrast this with the images of the "pasturage" of epea: unlike muthoi, which can be characterized individually (pantoioi, "of all types"), these travel about here and there in a wide field—the image accords with that describing Thersites' speech. [56] Moreover, epea are like small objects that can be batted to and fro, part of a general system of exchange. [57] Aeneas' gnomic statement, "whatever word you say, you hear," points once again to the notion of transmission of speech, a notion that goes with epea (instead of muthoi). [58] Finally, this line (250) embodies the idea that an epos is the least unit of speech: the smallest piece of praise or blame conies back to its speaker in turn. More evidence for this view comes from verses in which epos (and never muthos) is modified by the indefinite pronoun, as in Hera's reported address at 19.121: [59]

Ζέυ' πάτερ άργικέραυνε, έπος τί τοι εν φρεσΐ θήσω
"Zeus father with the flashing bolt, I will place some word in your mind."

The physical quality of speech, (epos), as well as the function it fulfills in reciprocal social relations, is well expressed by two common groups of formulas: those describing hand movements coordinated with an epos; and those naming expressive social acts, such as chiding or weeping, accompanied by epea. The first class of formulas is employed by Homer mainly in type-scenes of welcoming or consolation, {18|19} but "hand and word" descriptions can also occur whenever one speaker establishes contact with a listener for an emotional private conversation, as when Athena persuades Ares to leave battle (5.30): "Taking his hand she spoke to rushing Ares with words (epeessi)." [60] The focus is on speech as a social bonding mechanism, the equivalent of a handshake, an affirmation, like that between helping divinities (Athena and Poseidon) and Achilles (21.286): "Taking hand in hand they pledged faith by means of words (epeessin)." The parallel between verbal and physical gesture is highlighted particularly in the following formulas:

ἔν τ' ἄρα oἱ φῡ χειρί, ἔπος τ'ἔφατ' ἔκ τ'ὀνόμαζεν χειρί τέ μιν κατέρεξεν έπος τ' ἔφατ' ἔκ τ'ὀνόμαζεν

She put her hand on him and spoke a word and called. She stroked him with her hand and spoke a word and called.

These introduce motherly, comforting language—Thetis to Achilles (1.361, 19.7, 24.127), Dione to Aphrodite (5.372), Hekabe to Hektor (6.253)—or words between intimates, like Hektor and his wife (6.406, 485). [61]

The phrase "spoke a word and called" ends, literally, "called the name. " It has long been a puzzle, since not every one of the forty-three occurrences of this formula in Homer is followed by an explicit vocative. Couch saw a more general consistency in that this half-line forms a "prelude to the words of a god or mortal, who is the superior of the person addressed, whether through recognized rank, mistaken identity, or the moral force of circumstances at the time that the speech is introduced." [62] I cannot solve the problems concerning the meaning of onomaze, "called the name," but by putting this half-line formula in the larger context of the physical connotations which I observe in the word epos, I can suggest that the coordinated hand and speech gestures are what is important in the poetic employment of this half-line, rather than any usefulness for introducing vocatives. {19|20}

This becomes clear when one compares lines in which a mention of the "hand" appears before the phrase "spoke a word and named" with lines that have a different introductory phrase. The lines without "hand" phrases (quoted earlier) seem to be employed randomly. They can introduce a speech full of taunts and insult, such as Helen's retort to the disguised Aphrodite (3.398-99):

She was amazed, and spoke a word and called the name, "Strange one, why do you want to deceive me?"

Or, they introduce simple questions (14.297), a berating speech (15.552), a speech of concession (21.356), an invitation to pray (24.286). In short, the "hand-and-speech" lines represent a consistent poetic expression drawing the audience's attention to a genre of speech, consolation. Even though the lexical form of the "hand" phrase may vary, the underlying semantics of the line are preserved. Gesture and utterance, the latter imagined to be just as physical as the former, claim notice. The lines that share only the second segment seem to represent an innovation and a reuse of an original full-line formula without a sense of the line's original poetic purpose. [63]

In the second class of formulas that draw attention to the physical nature of the word, epos appears in the dative case to indicate the means by which a communicative act occurs. [64] The act itself— reviling, mourning, praying, urging, answering—is described by Homer with a verb-phrase. The dative phrase containing the term epea then describes the tone or mood of the words. Positive or negative overtones can be created by the choice of an adjective, and words can be "cutting" (4.6), "childish" (20.200, 211, 431), "angry" (4.241), and "insulting" (16.628, 2.277, 1.519), or, less often, they can be "pleasant" (12.267), "soft" (1.582, 6.337), and "mild" (2.164, 180). It is true that some of the speeches introduced by these epeessi formulas function to describe speech events, such as blaming, which can also be introduced as muthoi. But it should be pointed out that most often this class of formulas is employed only for a narrative description, not to introduce an actual "speech" in the poem. This may explain why {20|21} the epos formulas so often occur with modifying adjectives (unlike lines containing muthos): the performance of a speech action is not the focus of the word epos. The elided, unnarrated words of the heroes are considered, in epos descriptions, only as verbal accompaniment to a more important act. Since we do not "hear" the words, the poet must color the narrative by using a descriptive adjective for the kind of talk that his characters used.

We have noted that epea in the system of Homeric diction represent the means of conducting social life; they participate in an economy of exchange. Unlike muthoi, which are the full, exaggerated speech-acts of heroes, "utterances" simply receive passing notice by the poet, in their role as concomitants to physical gestures. It is not that these words are unimportant; they matter in a different way. One of the first occurrences of an epeessi formula in the poem can illustrate the phenomenon. Hephaistos attempts to patch up the quarrel between Zeus and Hera by code-switching, as it were. Instead of aggravating the confrontation between two powerful performers, in which Hera has just challenged the muthos of her husband (1.565), Hephaistos advises Hera to use soft language (1.582-83):

"But approach him with soft epea and right away, the Olympian one will be propitious to us."

As prayer and supplication work for mortals, so these soft words will supposedly help Hera, making the angry divinity propitious (hi-laos). [65] Words alone do not complete the reconciliation, though. We do not hear Hera speak to Zeus. Instead, Hephaistos' advice to Hera is joined to his gesture of offering her a cup (585) and to his comic autobiographical story (586-94). The ensemble of word and action carries the scene to its harmonious end.

Much remains to be said about the role of this social-poetic metaphor in Homeric speech depictions. It can be shown, for example, that the "supplication" of the godlike Achilles by the embassy depends on this concept of speech as reciprocal exchange, as the realm of epea. Nestor instructs the Achaeans to propitiate Achilles with "mild words and pleasing gifts," significantly equating the two {21|22} (9.113): δώροισίν τ' αγανοΐσι/ν 'έπεσαι τε μειλιχίοισι. But Achilles operates within a different system, it appears. Rejecting the exchange value of "pleasant words," the supreme hero constructs the most powerful muthos in the Iliad, as I shall show in Chapter 5.

The Authoritative Speech-Act

The second major distinction between muthos and epos pertains to the sorts of discourse labeled by these terms within the poem. I have found that all but 12 of the 167 occurrences of the stem muth- (noun and verb forms) in the Iliad can be categorized as marking proposals and commands or threats and boasts. As we shall see, these four types of discourse constitute, essentially, just two types of speech-act. More abstractly, muthos in the Iliad is always the speech of one in power, or of someone, for example a boasting warrior who is laying claim to power over his opponents. The word muthos implies authority and power; epos implies nothing about these values. Not surprisingly, Book 1 of the poem is permeated with muthoi. Agamemnon commands Khryses to depart and the heralds to fetch Briseis by using a "powerful speech" (krateron d' epi muthon etelle—1.25, 1.326). Athena's warning to Achilles is referred to by the narrator as a muthos (1.221). Achilles, in his colloquy with Thetis, views both of Agamemnon's acts, against Khryses and himself, as authoritative misuse of muthoi:

But he sent him off badly, and ordered a hard muthos." (1.379)

Then anger took hold of Atreus' son, and quickly standing, he threatened with a muthos which indeed has been fulfilled.


Here and throughout the poem the authority underlying muthoi is acknowledged in the audience's response to a given speech; the addressee is most often persuaded. At times the success of a muthos is evident only from the behavior of the person listening—submission (willing or not), indignation at being commanded, fear. Often, however, the result is instantaneous and explicit in formulaic language. The poet emphasizes the moment of persuasion thereby: [66]

{22|23} Thus he spoke, and the old man feared and obeyed the muthos. (1.33, Khryses = 24.571, Priam)

Nor did he disobey the muthos of Athena. (1.220-21, Achilles)

Be quiet; obey my muthos. (4.412, Diomedes)

Those who are equals, socially or as "performers" of deeds, can challenge one another's muthoi. As we have seen, Paris does so with Antenor's proposal in the Trojan assembly (7.358). When the lines of power are clearly drawn, the speaker who uses a muthos prevails. Thus, Odysseus speaks to his social equals in the testing scene of Book 2 and uses epea. The poet says:

Whatever king and outstanding man he met, he stood beside and held back with mild words (epeessin). (2.188-89)

To Thersites, however, and to the men of the dēmos, he uses a muthos:

But whatever man of the people he saw or found shouting he would drive on with the scepter and berate with a muthos: "Strange one, sit still and listen to the word of others (αίΐδη muthon) who are stronger, while you are unwarlike and strengthless."


The Iliad is largely about situations in which power is in dispute, up for grabs. The anthropologist Michael Herzfeld has recently drawn attention to the role of expressive rhetoric, accomplished by verbal and nonverbal means, in a similar cultural context—the disputes of contemporary Cretan mountain villagers. [67] It is precisely this sort of expressive use of language in dispute settings that Homer characterizes as muthos. Like the "poetics of manhood" that Herzfeld explicates, this rhetorical skill at self-presentation can be learned by Homeric heroes; indeed, it must be acquired, and Homer shows how the education takes place. Diomedes, in the Iliad, becomes the model of the young Greek male initiated into forceful speaking, a learning of muthoi. As we might expect, Nestor, the veteran performer and {23|24} orator, whose authoritative speech was always obeyed by heroes of the past (e.g. 1.273), teaches the skill of words.

It might help to compare this sort of initiation in words with the experience of the young singer of epic songs, as Lord describes him. [68] Like the poetic performer, the young orator Diomedes starts out slowly, by listening to his elders and repeating their phrases. He throws back at Agamemnon some of the older man's harsh words from an earlier encounter, warning him that he will fight in the speaking-place "as is fit" (9.33, hē themis estin—Agamemnon's phrase at 2.73), in a clear allusion to Agamemnon's remark that Diomedes is a better speaker than fighter (4.400). We get the impression that Diomedes, like the young singer, again, needs time to compose his reply to Agamemnon; for, when attacked in Book 4, he made no direct answer (4.411-18), but at the start of Book 9 he has a ready supply of words. Yet the phrases he employs are a curious blend of rhetorical tacks taken by Thersites and Achilles (significantly, two other marginalized characters, one by status, one by choice). Agamemnon has only authority, not strength, Diomedes asserts; he can go home if he likes, while Diomedes and his companion win the war. Besides the upstart rhetoric of this speech, there is a lot of repetition, as if Homer were characterizing the inexperience of Diomedes through his style. [69] His line-initial repetition of alkēn, "strength" (9.34, 39) might seem forcefully expressive; but the repetition of the longer phrase "unwarlike and strengthless, (9.35, 41) which, in turn, encloses a triple occurrence of "he gave" (dōke—9.37—39) strikes the listener as a clumsy attempt at sounding forceful. And so it seems to affect Nestor, as well. Perhaps Nestor is prompted by the hyperbole of Diomedes' assertion (after all, nowhere in the testing scene did Agamemnon actually call Diomedes "unwarlike and strengthless"— those were Odysseus' words to the low-status fighters in Book 2.201—6). Perhaps it is Homer's purpose to show us Nestor allied to Agamemnon. At any event, the old warrior is conspicuously not a part of the general acclamation of Diomedes' muthos:

So he spoke, and all the Achaeans' sons clamored assent, in awe at the muthos of horse-taming Diomedes. But standing up, the horseman Nestor spoke to them: "Son of Tydeus, you are especially strong in war, {24|25} and in counsel you were best of all your peers, No one will find fault with your muthos, as many as are Achaeans, nor speak against it. But you have not reached the utmost point of muthoi."


Nestor notes the difference in their ages, saying he could be Diomedes' father, and approves the substance of his remarks (9.57—59). Then he speaks his own mind, as explicit instruction for the younger hero in how one reaches the "perfection of speeches" (telos muthōn, line 56):

"But, come, I who claim to be more honored than you, I will say out all and go through it all, and no one can devalue my muthos, not even Agamemnon who rules."


He proceeds to deliver a simple proposal in elegant form, full of polyptoton, asyndetic gnomic statements (line 63), and, most important, full praise for his audience—Agamemnon. As he predicted, all approve Nestor's muthon (9.79). The stage is set for the next contest of counsels (9.74-75), in which Nestor again convinces his audience, thereby causing the embassy to be sent (9.94-113).

Having witnessed such expert teaching, Diomedes grows in rhetorical ability through the rest of the poem. By Book 14, we see that he has learned how to construct an impregnable speech, by taking preemptive action against possible objections. As does Nestor, Diomedes prefaces his speech with disarming recollection of the past, his own ancestry in this case, and thus rejects possible claims that he is too young to speak well, while at the same time raising his status as a warrior by blood. As a result, his muthos is assured of success:

"Therefore, you cannot devalue my muthos by saying I am strengthless and have a bad ancestry, whatever I say. "


To indicate that Diomedes' proposal is a convincing performance the poet concludes the speech with the same line that followed Nestor's exemplary muthos in Book 9:

Thus he spoke and then they listened and obeyed. (14.133 = 9.79) {25|26}

We shall return in Chapter 3 to trace the development of Diomedes as heroic speaker within the Iliad. For now, we can note that the heroic imperative as crystallized in Phoinix's words to Achilles demands that a hero learn to "be a doer of deeds" and a speaker—not of words, but of muthoi, "authoritative speech-acts": μύθων τε ῥητήρ' ἔμεναι πρηκτῆρά τε ἔργων (9.443). Both deeds and such words can be enacted; Diomedes shows us a Homeric hero striving for this ideal.

Synonymity and "Winged Words"

Thus far we have seen that epos can be distinguished from muthos through the former's close relation to voice and hearing and the latter's consistent association with powerful, status-related speech. Yet it could be objected at this point that epos also means "command" or "proposal" in more than a few contexts. Does this mean that the systematic distinctions shown thus far are a mirage caused by the formulaic language of the poem? A closer look at the employment of epos shows that this is not the case; the system of Homeric speech terms is instead even more complex (and yet economical) than at first suspected. This becomes clear if we examine contexts in which epos seems to be synonymous with muthos and compare the results with patterns of co-occurrence, either of the two words in the same description of a speech or of the two words in association with mutual modifiers.

At first it seems there is a striking difference in the use of the word epos to mean command: that it, unlike muthos, refers to the secondary transmission of an original command by someone else. The focus appears to be on the message itself (as we have seen with this word in other contexts). Thus, Achilles' demand that Thetis supplicate Zeus is referred to by his mother as an epos: Looking forward to the moment when she will transmit the message, she promises: [70]

"I will go back to snowy Olympus to tell Zeus, who delights in thunder, this epos if he may obey." (1.419-20)

The same phenomenon seems to occur when Achilles tells Athena that he will "preserve" her command (epos eirussasthai) to fight Agamemnon {26|27} with words, instead of sword. Her directive, as the poet twice makes clear (1.195, 208), comes from Hera, who sent her; therefore, we might say that the command is not Athena's self-willed act, and does not require extended performance to show her own authority. It is not a muthos, it would seem—and yet, a few lines later, it is called precisely that by the poet (1.221). Once more, we might want a similar interpretation at first to explain why Zeus' orders to Hera are epea when she transmits them (15.156); but this too becomes overingenious and simply does not work. In this example, the original command was designated epea (15.48). In short, it looks as though the two terms which we thought distinct, epea and muthos, can denote exactly the same sort of speech-acts. Although the word epos is used more often in actual speeches (47.2 percent) whereas muthos more often occurs in the narrator's framing of speech (38.2 percent in speeches), again there is no rigid demarcation when it comes to commands: this meaning can occur when either a character or the poet uses either word.

Two solutions can explain the overlap in meaning, I believe, without forcing us to abandon an otherwise systematic distinction between the two terms for speech. First, it will be seen that other factors, especially the speech situation in which the "command" occurs, can determine whether epos or muthos names the speech-act described. The latter is reserved for public performances. A second, more general explanation works by organizing these two words for speech in a hierarchy, in which muthos has restricted range but greater semantic weight. But before outlining these factors, Ι should point out the sort of pressure in the system of Homeric diction which leads to some leveling out of the observable distinction between muthos and epos in some contexts.

To this point, I have concentrated on the role of speaking and the ways in which it is described in the Iliad. Yet we must remember that the heroic ideal of speaking and fighting virtuosity is always being propounded in the poem. "Word and deed" becomes a merismus, expressing an ideal totality by reference to the extremes which shape it. [71] This is the positive rhetorical strategy concerning "words." At the same time, however, the way in which one talks about speech can itself become part of the rhetorical repertoire of the warrior, as well as a poetic topos. Because the ideal is to speak and talk well, any hero (or the poet) can characterize another man as deficient in either, thus {27|28} using the rhetoric of speech description negatively. For our project. both strategies of description tend to reduce any difference between muthos and epos in favor of the overriding contrast with the larger category of "action. " [72]

In an expression denoting instantaneous action, the poet can use muthos:

αὐτ,ίκ' ἔπειθ' ἄμα μῦθος ἔην, τετέλεστο δέ ἔργον (19.242)

Then, as soon as the muthos, the deed was done.

The word epos however, will do just as well:

οὔ πω πᾶν εἴρητο ἔπος ὅτ' ἄρ' ἤλυθον αὐτοί (10.540)

Not yet was the whole epos said when they themselves came.

Still, there are differences in the acts referred to by these two terms: reading context, we see that Odysseus at 19.216-37 makes an authoritative proposal, called by him muthoi (19.220). This the poet refers to at 19.242. But Nestor (10.533-39) merely voices his suspicion that the best of the Achaeans are returning from their night mission: his speech is labeled epos — as one could predict, given the associations of this word with reported speech. A similar explanation will show the differences beneath the surface likenesses when the two different nouns share an adjective, giving the appearance that they are formulaically interchangeable words. When Hera tells Athena that they are dangerously close to reneging on their promise that Menelaos will take Troy, she refers to the original speech-act as muthos:

ἦ ῥ' ἅλιον τὸν μῦθον ὑπέστημεν Μενελάῳ


"Vain, indeed, is the muthos we promised Menelaos."

Achilles also mentions a past speech-act when he uses the same adjective halion, "in vain," and we might expect him to label it muthos, but the line concludes:

ἦ ῥ' ἅλιον ἔπος ἔκβαλον ἤματι κείνῳ (18.324)

"A vain epos I tossed forth that day." {28|29}

But here the nature of the original speech-act gives us a clue to explain why it can be called epos in retrospect: as a boast, it is attracted into the language of actual boast descriptions, which often contain the formula εὐχόμενος ἔπος ηὔδα. [73] Once again, then, external factors—formulaic pressure, poetic themes in the larger discourse, or accidents of the system—can mask the inherent semantic distinction between muthos and epos.

And yet, even after we exclude and explain such apparent synonymity, there remain passages in which the two different terms that we are studying co-occur. For example, we might look at Agamemnon's words during the troop-rousing episode (4.337-48). The poet introduces his speech to Menestheus and Odysseus as "winged words" (337), and Odysseus responds to the speech by labeling it with a formulaic line (350): Ἀτρεῖδη, ποῖόν σε ἔπος φύγεν ἕρκος ὀδόντων ("Atreus' son, what sort of word has escaped your teeth's fence?"). Yet at 357, in reference to Agamemnon's apology the poet says, "He took back the muthos." It seems either that the poet is simply manipulating formulas without regard for how they correspond (and so is not stopped by any semantic differences between the two terms, if they exist); or that the two terms simply are synonyms. [74]

Rather than being a flaw in the system, this co-occurrence is the key to Homeric usage of speech terms, and can help us understand the seemingly arbitrary deployment of lines such as "she spoke winged words." For the origin of this usage, in which epos resembles muthos, can be explained if we return to Prague School linguistics for a moment, in particular to the notion of "marked" versus "unmarked" members of an opposition. The "marked" member of a pair carries greater semantic weight, but can be used across a narrower range of situations, whereas the unmarked member—the more colorless member of the opposition—can be used to denote a broader range, even that range covered by the marked member: it is the more general term.

Turning to the words for speech, we can now say that muthos is {19|30} the marked member of the pair, and epos the unmarked. [75] This means that a muthos can always be referred to periphrastically as epea, "utterances," since the latter word, singular and plural, has primary reference to anything uttered or heard; to "words" in the most general, unmarked sense. One can never simply substitute the semantically restricted term muthos—meaning authoritative speech-act, or "performance"—for the ordinary term epos, however. Thus, Odysseus can refer to Agamemnon's words, in the scene just mentioned, as an epos—the formula stresses the physical reality of the single word or utterance—and the poet can use epea to introduce them, even though the speech also partakes in the more restricted term and can be tagged with it if the poet or speaker wants to emphasize the speech's power or importance. The reverse never happens: that is, in Homer, a speech explicitly said to be an epos, and not also represented as epea (the plural), is never called a muthos. The irreversibility of singular and plural is good evidence that the words are not synonyms: epea can co-occur to refer to a muthos, but muthoi in the plural is never correlated with the singular form epos, to describe a speech. Furthermore, the same distinction applies to the use of verbs formed from these two roots. There exists a formula "he spoke a muthos (muthon eeipe)" but there is never a collocation of the type "he authoritatively spoke an epos," even though we can imagine a metrically possible phrase (*epos muthēsato).

With this distinction in mind, I can elucidate perhaps the most famous and least understood phrase in Homer, "he/she spoke winged words. " Now we have seen that epea, on the one hand, can be a periphrastic expression for muthos (but not vice versa). On the other hand, as we saw earlier, epos and epea have a reference not shared by muthos, to speech as utterance, as thing heard and transmitted, as an item of exchange that is at the same time a physical object, like a weapon. Thus, the formula epea pteroenta can fit two functions. This is what makes it such a useful, economical poetic device. On a purely imagistic level, the phrase evokes the swiftness, irrevocability, directness, and reciprocal nature of speech—the adjective affirms the basic notion in the noun. On the narrative level, as a synonymous phrase for muthos, this formulaic line can be used to introduce highly "marked" speeches. That is, the possibility lies open that there is real {30|31} meaning to Homer's description of just sixty-one of the poem's approximately six hundred speeches as "winged words." [76]

Milman Parry has been the most forceful advocate of the view that the line means nothing in relation to its context. His article on the phrase was meant to counter the view of Calhoun, who thought that the Homeric line introduced speeches of high emotion: "Pteroenta evidently does not mean 'winged' or 'swift' in the general sense that all words fly with the speed of sound, but is intended to say that the particular words which follow were spoken quickly, or with animation, or some symptom of emotion." Parry was certainly correct to challenge this rather vague formulation by Calhoun (after all, as one critic pointed out, all speeches in Homer are emotionally charged), but his analysis stopped short because he did not have a framework for describing common elements in the "winged words" speeches. [77]

I now suggest that there are two common elements in these speeches. First, it has not been noted that they share a pragmatic speech situation: winged words are spoken by one person to one other (rarely to two), and the addressee is in close contact with the speaker, usually as a comrade-in-arms. Even more consistently, the deployment of these speeches by Homer enables us to describe all examples of "winged words" in terms of one category of "speech-act": in the terminology adopted by Searle and others, the sixty-one speeches thus introduced are all "directives." That is, every speech called "winged words" is meant to make the listener do something. [78] Certainly not all utterances in Homer function this way; yet the Iliad always pays attention to the motivation and effect of heroic speeches.

That the "winged words" all constitute one specific speech-act class has not been noticed, because typologists have previously used a haphazard mixture of descriptive criteria to analyze Homeric speeches, relying on such elements as scene, emotional content, or use of certain phrases. [79] And, on the surface, the "winged word" speeches do {31|32} not offer immediate similarities. Some contain direct imperatives, some indirect (third person); others are only hints, statements of need, questions, or even straightforward narratives seeming not to demand response at all. It is only by abstracting, by analyzing function rather than form, that we discover the underlying similarity of "winged words." For, as speech-act theory can tell us, precisely such diverse surface structures characterize the class of "directives." In the appropriate context, diverse utterances—"Enter the battle"; "We are losing, I'm afraid"; "Could you join with us?"; and "I heard of a battle like this once" (see Phoinix's speech in Book 9)—these represent the same directive speech-act. [80] Once again, we must remember that not every "directive" in the Iliad is tagged epea pteroenta. But the converse is true.

There is a range of "directive" expression. For example, Achilles' winged words upon recognizing Athena in Book 1 (201—5) take the form of an interrogative: "Why have you come, child of aegis-bearing Zeus? Is it to see the hubris of Agamemnon, son of Atreus?" Here the question functions as a directive: Athena should perceive things the way Achilles does. Her reply confirms that Athena fully realizes what Achilles commands her to acknowledge: "Sometime," she says, "there will be three times the glorious gifts because of this hubris." In other words, the goddess successfully reads the illocutionary force of Achilles' words, rather than replying to them as if they constituted an actual request for information. [81] Agamemnon's scornful speech to Odysseus and Menestheus in the troop-rousing scene (4.337-40) employs the interrogative form similarly: "Why do you stand off cowering and wait for the rest?" The directive force is amplified by the next words: "It is fitting for you two to stand with the front-lines and confront burning battle" (341-42). This statement is just as "directive" in function although it, too, lacks the linguistic form of an imperative.

Because form and function can vary in the employment of "directives," there is opportunity for the individual speaker to "perform," to exercise expressive creativity. Later in this book, I shall draw out the analogy between this everyday linguistic phenomenon and Homer's {32|33} own "formulaic" art. For now, we can observe that "character" in the poem (as in life) arises from our perception of a speaker's selectivity and sensitivity in matching linguistic expression to internal motivation. Some people always make their directives into imperatives. Some have more tact. Achilles, whose expressive repertoire we shall examine later, is good at hinting: to his mother, he speaks "winged words" (19.20-27) to say he will arm. These end with what is almost an afterthought: "But I fear terribly that flies, meanwhile, might breed worms down in the bronze-cut wounds, might defile the corpse—his life being destroyed, the flesh might all go rotten" (19.23-27). The description is worthy of poetic narrative at its best. It convinces both audiences—that of the poem (persuaded of the height of Achilles' emotion) and that in the poem, Thetis, who replies to this extended hint /directive by infusing Patroklos' corpse with nectar and ambrosia.

Four passages make clear the tone implied by the mention of epea pteroenta: the death of Patroklos, the encounter of Hektor and Achilles in Book 20, the death of Lykaon, and the encounter of Priam and Achilles in Book 24. We have seen that, out of the hundreds of passages where speech appears, "winged words" in the Iliad highlight only "directive" speeches between those sharing a social bond. They are language appropriate to an "in-group." As the poem nears its end, enemies exchange "winged words." It must be noted that the four passages in which this occurs are not casual encounters, but rather highly charged events important to the outcome of the plot and, furthermore, that they are given lengthy, elaborate ornamentation by the poet. Much of the powerful effect in these scenes comes from their inclusion in the conventional pattern of fighters addressing comrades-in-arms with "winged words." For here, the fighters are paradoxically bonded by their very determination to kill one another. What seems like a violation of formulaic conventions is actually a creative extension of the usual meaning of the phrase. [82] The speeches of Apollo and Hektor resemble one another as warnings to Patroklos that Troy will not be taken by him or Achilles. Apollo's speech, furthermore, is an explicit directive: "Fall back, god-born Patroklos!" What is the force of Hektor's words? They are introduced as a boast (ἐπευχόμενος, 16.829). They actually contain an embedded directive, {33|34} in the form of Hektor's rendition of what he imagined Achilles to have commanded: "Do not return . . . until you pierce the bloody khiton of man-slaying Hektor." But their directive force, I believe, comes from Hektor's description a few lines before this "quotation": "Here the vultures will eat you." Although this is cast as a prediction, the sentence conveys what Hektor wants to happen to the corpse of his enemy. To say "the vultures will eat you" is to direct Patroklos to die. Confirmation of this reading comes in the other death scene in which the killer speaks "winged words," at 21.121-35. Like Hektor, Achilles seeks to stun his victim and vent his rage with a vivid description of the enemy's defilement. But, whereas Hektor condenses this narration to a half-line (16.836), Achilles expands it to six (21.122-27), Furthermore, Achilles begins this full version of the motif with an explicit directive in the form of an imperative: "There now, lie with the fishes." Only then does he revert to future verbs, in Hektor's manner predicting that Lykaon will be unmourned, his corpse whirled in the river, and his flesh eaten by fish. [83]

The brief encounter between Achilles and Hektor in Book 20 resembles these two death scenes in tone. The directive force of Achilles' words at 20.449-54 is that of a warning: he will finish off Hektor on their next meeting (452). From the three scenes we can begin to appreciate the effect that Homer achieves by using the "winged words" formula in such encounters. It is not unlikely that the regular employment of the formula, usually introducing friendly, same-side directives between just two speakers, carried with it an implicit tone of voice and volume. Such speeches could properly be performed by a poet in a stage whisper, or with an intimate low modulation. If this same auditory effect was produced by Homer as he imitated Achilles' or Hektor's voice in the speeches we have just noted, the audience, recalling the more common employment, would hear the intimate voice of violence.

Finally, with both possibilities for the enactment of "winged words" now in mind—the intimate and the violent—we can recover something of the tone of Achilles' words to Priam as the poem ends (24.518-51). They are a mixture of reproach, pity, and encouragement. Priam is a "wretch" and foolish, with an "iron heart," to have dared approach the Achaean camp alone. Achilles directs him gently, however, to sit, mourn, and ultimately endure (line 549, anskheo {34|35} forming a ring-composition with line 518, anskheo). The triumph of this conclusion is tonal: enemies can exchange "winged words" without a killing.

Now that I have traced the functional meaning of this speech description, what of the metaphor itself? Are words winged "because they fly through the air rapidly, like birds," as the most recent commentary suggests? Or are they "feathered," "fledged," like arrows going to the mark? [84] Perhaps there is a mixing of metaphors, as Gladstone imagined: "It is not the mere feather, but the wing which is described. It is not a random, but a carrying force. The word is a weapon, and bears its mission through the air." [85] There is no guarantee that the metaphor underlying epea pteroenta actually was perceived in one specific way by the audience; it could be a dead metaphor, completely unrelated to the function of "winged words" as a speech-introducing phrase. If we wish to attempt a consistent interpretation, however, Gladstone's solution might be reconsidered. For the phrase, as we have seen, cues the listener to a directive (thus the "arrow" is an apt image); at the same time, it focuses attention on the physical, enduring nature of words, as epea. These words are like birds, then, but not in their free flight. Rather, epea pteroenta resemble the insistent hovering motion of a bird beating its wings. This image fits the formulaic evidence, as well, since the adjective puknos and its by-form pukinos are the only words associated with both epos and pteroeis when these are examined separately. When applied to speech, the adjective carries the sense of language that is dense with meaning and filled with urgency. Priam's formal bid to call a truce (7.375), the advice Patroklos is to give Achilles (11.788), Zeus' directive for Thetis to transmit to Achilles (24.75), and the word which Andromakhe hoped to hear from her husband (24.744) are the four instances of a pukinon epos in the Iliad. The last-mentioned makes clear the special quality of this sort of speech:

"For you did not die on a bed and stretch out to me your hands, nor say to me any close-set word (pukinon epos), one I could recall, shedding a tear for days and nights." {35|36}

The word she wanted from Hektor would have been enduring through time, unassailable in the way of well-constructed, solid, or dense-packed objects in the poem that are called pukinon. The adjective and related forms modify arms and armor, beds, troop formations, house construction, branches, and clouds. But mental products can also be thus qualified: a tightly constructed plan (2.55), Odyssean wiles (3.202), an ambush (4.392), a trick (6.187). [86] If epos generally is speech as product, as I have argued, then this particularized form of speech is the paradigm for the best kind of epos, speech that has become a lasting possession for its hearer, the "last word" to remain with an intimate (Achilles, Andromakhe) or to put an end to strife (the truce in Book 7, the ransom in Book 24).

The adjective obtains this sense of "unassailable" from the basic reference to density; the same root meaning in the adjective emerges in adverbial use, but this time with reference to a series of rapid movements, a density in temporal terms which we might call in the language of physics "frequency." Homeric similes associate the temporal with the physical sense of "density," as when Hektor's rapid striking of his enemies' heads (pukna karēath') is compared to the action of wind piling up wave and cloud (11.305-9). I suggest that the same combination of "density" and "frequency" occurs in the phrase ptera pukna, so that it describes both the close-packed construction of the wing and the resultant rapid wing-beating of the bird: this is a prime example of "interaction" in poetic imagery. [87]

It is the aural quality resulting from the flight of birds that is the primary association in the phrase "winged words." This is to say that the one poetic phrase is built on the image of the other, of "thick-and-fast wings" (ptera pukna). Such a close association between physical density or frequency and aural effect is found elsewhere in Homeric diction where emotional, forceful speech is being described. For the adverb pukna can also be applied to the sound of the lamenting voice:

τοῖσι δὲ Πηλεῖδης ἁδινοῦ ἐξῆρχε γόοιο . . . πυκνὰ μάλα στενάχων (18.316-18)

{36|37} For them, the son of Peleus began the resonant dirge . . . groaning thickly (pukna).

The enduring physical quality of the epos which we have found accords with its meaning "utterance"—the physical result of speech. But in some contexts—as for instance, that of intimate directives— the intensity of the speaker's purpose imparts to the otherwise ordinary epea a tautness, power, and movement that makes them whir and beat, like the motion of a wing. The powerful language thus produced makes its impression because it continuously reaches the hearer's consciousness, like a wave of sound. We could apply the image to Homeric poetry itself, a medium of enduring motion.

Public and Private Speech

The association of muthos with power, planning, and self-presentation goes naturally with the third set of oppositions between this word and epos. I have found that muthos always implies "public speech," and involves a performance before an audience. Such speech can be called "pleasing" (headota) to all (9.173). Groups of hearers can approve it (epainēsantes) or be astounded by it (agassamenoi) as is the embassy at Achilles' refusal speech (2.335, 9.710). It is significant that none of these audience-reaction phrases occurs with epea. Furthermore, as we might expect, muthoi are the object of dispute, which takes place in public, as the poet implies when characterizing Thoas in Book 15. Along with his excellence at the javelin, this hero excels in speech: "Few of the Achaeans would beat him at speaking, whenever they used to engage in strife concerning muthoi" (15.283—84). We can note that the word for a style of speech connected with muthoi here is also used precisely for a place, the public arena of speech.

There is good formulaic evidence as well for the association between muthoi and agoreuein, "speaking in public." In the poet's language, one can be said to begin speech-acts in an assembly muthōn ērkhe (2.433, 5.420), or to begin to speak, ērkh' agoreuein (1.571, 7.347). In the Odyssey, I might add, similar phrases function as formulaic complements: ērkh' agoreuein at line-end (2.15, 16.345) is interchangeable with ērkheto muthōn (1.367, 15.166).

A study of the word epos shows a complete contrast: unlike muthos, it is associated with private and reciprocal speech, such as {37|38} that between husbands and wives, companions, or kin. When Hera says that she is going to patch up marital difficulties between her relatives Okeanos and Tethys, she calls her speech epea:

"If persuading them with words in the dear heart (epeessi paraipepithousa) I might manage to unite them in affection I would be called dear and respected always by them." (14.208-10)

To carry out Hera's project, Aphrodite grants her parphasis, the sort of speech characterized as "soft," "gentle," and "sweet." [88] It is personal appeal, not authoritative performance, that generates speech denoted as epea. The ideal "utterance," the enduring pukinon epos, is set in a context of intimate relationship: it is the language Andromakhe expected from her husband, and that which Patroklos is instructed to provide for Achilles: "But tell him well a pukinon epos and instruct and give him directions" (11.788-89; cf. 24.744). This kind of discourse is language one can personally "keep"—or fail to keep, as is the case when Patroklos forgets Achilles' personal warning (16.686-87): "If he had guarded the epos of the son of Peleus, he would have fled the evil fate of black death." In this function, it is worth noting, epea are often spoken by women, a convention that appears to be canonical in the deployment of the phrase epos t'ephat' ek t'onomaze, as we observed earlier.

The private nature of the epos explains the use of this word (rather than muthos) in those scenes where the poet privileges us with seeing the communication between heroes and divinities. Diomedes labels his talk with Athena in this way: "I recognize you, goddess, daughter of aegis-bearing Zeus. Therefore I will candidly say a word (epos) and will not conceal it" (5.816-17). Achilles refers to Athena's private advice (1.216) and his own prayer to Zeus (16.236) with the word epos. Prayer, which epitomizes private communication, is in fact never designated muthos in the Iliad. A further indication of the personal nature of epos comes from Homer's adjective usage: possessives frequently accompany the word epos: Achilles speaks of "your word" (1.216), as does Aphrodite to Hera (14.212); Achilles, praying to Zeus, and Hera, using a similar formula to Hypnos, say "You have {38|39} heard my word before" (emon epos eklues, 14.234, 16.236). Zeus refers to his own command with this phrase (8.8). By contrast, muthos is rarely referred to by a speaker as his or her own or as that of his or her addressee: it is "impersonal" in the sense that it is public discourse (although it is certainly personal expressive performance).

To talk of one's own "speech" as opposed to "word" would be equally aberrant in normal American conversation. I have been working with a distinction between epos as "utterance" and muthos as "speech-act." Were we to attempt a definition in terms of familiar English vocabulary, "word" or "talk" might best translate epos, and muthos could be paralleled by "speech" in the sense of "making a speech." The marked character of the latter can be appreciated if we examine patterns of co-occurrence in English: one cannot "make a word" or even "make a talk." The distinction between the terms in English depends on a number of contrasting features including occasion, tone, audience, and length of discourse: "speech" implies an audience of more than one, a formal routine (for instance, speakers take turns, without engaging in cross-talk) and elaborated use of language occupying a significant amount of time. Precisely these features fit the deployment of the term muthos in the Iliad.

First, we see this in the simple fact of poetic mimesis: speeches called muthos are almost always quoted in full by the poet, whereas those designated epos or epea, if quoted at all, occupy only a few lines. When muthoi are not represented by the poet, we are still given to understand that the discourses were lengthy, as in the description of the stories exchanged between Makhaon and Nestor (11.642-43):

When those two drinking put off the parching thirst, with muthoi they found pleasure narrating to one another.

Not only does the imperfect tense here reinforce our perception of lengthy storytelling: the narrative itself shows us this pair of heroes still drinking as Book 14 opens. [89] Notice that a translation for muthos as "story" accords with our earlier definition of the term—an authoritative speech-act performed in detail. The only unusual feature distinguishing the "stories" in Nestor's tent is their nonpublic performance. {39|40}The characteristics of muthos speaking emerge with particular clarity when we look at the poet's use of the denominative verb mutheomai, "to make a muthos." When this word for speech occurs, the accompanying discourse has a formal nature, often religious or legal; full detail is laid out for the audience, or is expected by the interlocutor in the poem; at times, a character comments on the formal qualities of the discourse labelled with this verb. Thus, when Ajax performs a muthos before Achilles during the embassy (9.623-42), Achilles approves the format of the carefully made speech, although he seems little swayed by its content:

"God-born Ajax, son of Telamon, commander of troops, You appear to have said formally (muthesasthai) all to me according to your heart (thumos). (9.644-45)

It is worth noting that Achilles believes that Ajax has made a full disclosure of his views.

Three times the verb appears in the infinitive at line-end, (muthēsasthai), with another verb meaning "command." The passages share certain rhetorical features. Kalkhas (1.74-83) answers Achilles by interpreting the hero's previous speech as a specific kind of request: "You bid me to make a muthos about the divine wrath of Apollo" (muthēsasthai/mēnin). Reassured by Achilles, Kalkhas proceeds to make a formal declaration of the god's will; Achilles' own precise formulation of the problem seems to have elicited this response (see 1.65-67, raising three possible religious delicts; and 1.85: a call for the theopropion which Kalkhas knows). The formality of Kalkhas' declaratory speech is enhanced by the priamel at lines 93-96, as also the double prin construction (97-98) and asyndetic legal phrasing (99). At 7.284-86 a different situation elicits this highly marked verb "to make a muthos. "The duel of Hektor and Ajax would have continued, if not for the intervention of the heralds, one of whom, the Trojan Idaios, "said a muthos" (eipe te muthon, 277). Ajax replies, "Bid Hektor to speak these things (muthēsasthai). For he himself challenged all the best to battle. Let him take the lead" (284-86). Hektor then rephrases what had been suggested (290-93): "Let us cease from battle and strife. "

Why does the poet give us such a roundabout description to end the duel which, itself, has struck critics as inessential? Because Homeric poetry so keenly attends to socially correct forms of speech. Hektor's {40|41} words are more than a simple directive; using the viewpoint of speech-act theory, we can state that lines 290-93 represent an actual "performative utterance"—language that brings about a desired end by virtue of the very pronouncement. As elsewhere, this language is formal: compare the standard examples of such speech, "I thee wed," or "I find the defendant guilty." [90] Furthermore, there are "felicity conditions" for performative utterances (only judges, for instance, can speak certain pronouncements); the poem respects just such parameters. Therefore, to be linguistically correct, Hektor, who began the duel with the challenge, must "pronounce" its end. Ajax, with a good sense of propriety that marks the most successful Homeric warriors, defers to his opponent. Formulaic repetition here (e.g. 7.282 = 293 on obeying night) is not merely poetic filling: it draws our attention to the sociolinguistic value of formulas; Idaios has an exact analogue in the religious or legal official who instructs the groom to say "I thee wed." [91] The herald's religious role is involved with learning of ritual formulas; he too is necessarily a "poetic" performer, even apart from Homeric poetry's depiction of his role.

The third time we see the infinitive used to fill out a command is at 11.201. Zeus, intervening in the battle, commands Iris to announce his order (muthon enispes, 186). An important detail here is the correlation of this phrase, "narrate/report the word," with the verb "to speak a muthos." [92] This confirms the association of the word with the act of telling in detail. In this instance, Zeus' command indeed is marked by a series of detailed instructions to guide Hektor's progress. [93]

The set of oppositions we have been examining includes features that relate to the setting of performance and to the performance itself. If muthos is used of public speech, and this usually means detailed {41|42} pronouncements, does the private nature of epos discourse correspond with brevity? Hints of this have come from our analysis of the utterance as "compact" or "close-set." I now suggest that the later Greek use of the word epos to mean "poetic line" is in fact implicit in the distribution of this term within the Iliad. [94] Let me point out that many of the discourses labeled with this term in the poem not only contrast with muthoi by being much briefer, but they often center around a gnomic one- or two-line utterance. Iris tells Poseidon that she will bring his declaration of status (muthos, 15.202) back to Zeus, but asks whether he will not reconsider, the speech having been truculent and harsh. "Pliable are the minds of fine people. You know how the Erinyes always follow elders," she warns (15.203-4). Poseidon replies, "Iris, you have in proportion (kata moiran) spoken this word (epos)." That he refers to the gnomic utterance, not to the description of what Iris intends, becomes clear from his next line, which is a capping gnomic verse: "It is also fine (esthlon, cf. her use of the word) when a messenger knows suitable things (aisima)."

With this mention of gnomic verse, I come to the final polarity between muthos and epos. The latter word is unmarked, as I have shown; this applies also to its use as a term designating types of discourse: epos can refer to any sort of speech. On the other hand, I shall show that the term muthos is the name that the poet gives to actual genres of discourse which are also poetic genres, and which we find embedded in the speeches of the Iliad. The evidence of formula once more gives us the clue to this usage. Line-final muthōi is regularly correlated with the word ēnipape (2.245, 3.427, 5.650, etc.), and associated with the verb erethize (5.419-20). The verbs mentioned are regularly employed to signal the poetry of blame, appropriate to neikos scenes. [95] Although this points us only toward one embedded genre of discourse, a fuller examination can show that muthos actually designates the conventional ways of speaking found in two other types as well, and lets us view Homer's heroes as poetic performers in their own right, as stylists. To these larger issues I turn now.


[ back ] 1. "An oral poem is not composed for but in performance" Lord 1960:4. "Dans le texte, le discours homérique se trouve en quelque sorte 'dénaturé'" Svenbro 1976:14. All translations from the Iliad in this book are mine, unless otherwise noted.

[ back ] 2. Lynn-George 1982 has a salutary critique of such reactionary criticism.

[ back ] 3. On the controversy over Parry's demonstration, the best short summary can be found in Latacz 1979: 1-17. Shive 1987: 139 intends "to help cure Homer of blindness and to put a pen in his hand. " A bridge between the old criticisms of Parry and the new reactionary strain is provided by the work of Goold 1977.

[ back ] 4. Bynum 1978:5; see also 3-13.

[ back ] 5. A. Parry 1972:I.

[ back ] 6. Hainsworth 1970:40. On the "uniqueness" of Homer, see also Griffin 1977. There is some confusion, in these arguments, between uniqueness of style and of subject matter or treatment, and I am not convinced that the former has been proved.

[ back ] 7. Fenik 1986:171.

[ back ] 8. Fenik 1986:151;xi-xiv.

[ back ] 9. Clarke 1981:116-21 gives a good introduction to this seventeenth-century debate and its later manifestations.

[ back ] 10. Mueller 1984:14.

[ back ] 11. On Analyst criticism, see Latacz 1979 and Clarke 1981:156-82. It is not a coincidence that reaction to Parry has paralleled the rise of Neo-analysis, on which see now M. Clark 1986.

[ back ] 12. Herzfeld 1985b: 202.

[ back ] 13. Goffman

[ back ] 14. Tannen 1982:18-19.

[ back ] 15. The number of such studies is now quite large. For an introduction, see van Dijk 1976:27.

[ back ] 16. Cited by Pickering 1980:5. On the development of his performance-centered fieldwork, see Labov 1972:49-69.

[ back ] 17. On this issue, see Searle 1976.

[ back ] 18. Ong 1986:148-49.

[ back ] 19. Biebuyck 1978: 351. Reichl 1985:614-43 observes a similar context for Uzbek and Karakalpak epic performances. Okpewho 1979:52 points out that only a full-length color film could accurately recreate a contemporary African epic performance. Recognizing this role of the audience, Renoir 1986:105-10 stresses the need for readers of ancient and medieval texts to re-imagine the original milieu.

[ back ] 20. Biebuyck 1978:352.

[ back ] 21. On the Cretan, see Notopoulos 1952:239-40; Ghil 1986:607-35 discusses the relation of the Romanian bard to his audience.

[ back ] 22. Wrigglesworth 1977:104.

[ back ] 23. Compton 1979:13, 122-29.

[ back ] 24. Hatto 1980:307, citing V. V. Radlov.

[ back ] 25. On one such story, see Burkert 1972. Herington 1985:13, discussing the links between rhapsodic performance as in the Ion and Homeric composition, says: "Homeric poetry . . . seems to have been designed from the first to be acted."

[ back ] 26. Tedlock 1980 shows that parts of the performance can indeed be "fixed," without benefit of written tradition, by stress, pitch, and pause. But this is not the same as saying that an entire text is immutably fixed and canonized as some one person's authoritative version. The latter approach has been tried by Homerists attempting to account for the gap between postulated oral composition and attested written transmission: see Mueller 1984:60-61, who cites Kirk 1976; 1978.

[ back ] 27. See Fine 1984. On performance as more important than text, see also Hrdličkovâ 1976:171-90. An entire oral epic performance is recorded and analyzed by Slyomovics 1987, the fullest such examination to date. A book could be written on the roots of the performance-centered approach; I have been selective. Fine 1984:32-37 recognizes the concept in the work of Kenneth Burke, Gregory Bateson, Victor Turner, Clifford Geertz, and Erving Goffman, as also in the work of folklorists since the 1960s. Bauman 1986tb:112-15 cites Walter Benjamin, Roman Jakobson, and Mikhail Bakhtin as models for criticism that "recontextualizes" literature. Pratt (1977) 68 nicely compares Labov's sociolinguistics with V. Shklovskij's poetics. Howell 1986:79 traces a concern with performance to B. K. Malinowski in the 1920s.

[ back ] 28. Bauman 1986b:78.

[ back ] 29. On this topic, a red herring, see Chapter 4.

[ back ] 30. Shive 1987:10-20 in his recent attack claims that the economy and extension demonstrated by M. Parry for the noun-epithet system does not apply to the other formulaic phrases in the poems. Yet Parry never asserted that it did, and this restraint on his part has long been acknowledged, by Hainsworth, Hoekstra, and others. The more careful study by Paraskevaides 1984, not cited by Shive, demonstrates that even Homer's use of most synonyms shows clear marks of formulaic economy and extension.

[ back ] 31. A few Homerists have recognized the import of comparative studies. Hainsworth 1970:29 saw that "the oral poem, properly speaking, is knowable only through its performances."

[ back ] 32. On the shadow theater, see Thornton 1984: n17, 25-32. Boon 1984:158 also notices the importance of the Balinese evidence. On perils of wider analogies, see Lord 1975 and Foley 1985:67-70.

[ back ] 33. Reichl 1985:613 lists some. On the lasting value of Serbo-Croatian for comparison, see Auty 1980:196-97.

[ back ] 34. 34Schadewaldt 1965 is the best attempt. Maehler 1963:9-34 does not press the evidence too far; see also Skafte-Jensen 1980:116-20. Macleod 1983:3 offers a useful reminder that we can learn not just from Phemios and Demodokos: "When Odysseus relates his adventures truly to the Phaeacians, or falsely to Eumaeus, when Helen, Menelaos, and Nestor recall their experiences at Troy or afterwards, they are to all intents and purposes poets."

[ back ] 35. See Todorov 1977:53-65 on the Odyssey as a poet's self-reflexive epic.

[ back ] 36. Lord 1960:25 and in general 99-123.

[ back ] 37. Ballmer and Brennenstuhl 1981:5, 33-67.

[ back ] 38. See Gossen 1974:vii, 78-83 and esp. 247-49; als° Gossen 1978:81-115 for a summary of his holistic approach to speech taxonomy.

[ back ] 39. The collection of essays edited by Bauman and Sherzer 1974 contains dozens of examples for this point, some of which I shall refer to later in this chapter.

[ back ] 40. Basso 1976:99. On the central role of American Indian descriptionist studies in the development of the ethnography of speaking as a discipline, see Saville-Troike 1982:5-11.

[ back ] 41. Muellner 1976.

[ back ] 42. On the semantics of muthos see Détienne 1986:47-51, and the review by Nagy 1982. For epos, see Koller 1972; but note Ford 1981:137-52 who argues that the "poetic" meaning of epos occurs not before the sixth century. Bynum 1976:47-54 traces the use of epos meaning "epic" to Aristotle. On the later history of muthos, see Bompaire 1977. I suggest that Aristotle's use of the word to mean "plot" is directly in line with a Homeric meaning "detailed, authoritative speech-act."

[ back ] 43. 43Cunliffe 1924:152-53 and 274; Ebeling 1885:464 defines epos as "verbum, perpetuitas verborum, atque res quae narrantur" and contrasts muthos as "sermo, Ansprache, quo quae sentimus aperimus. " For a similar definition of epos, see now Beck 1987.

[ back ] 44. For excellent illustrations of the specificity of Homeric vocabulary, see Benveniste 1969.

[ back ] 45. See Boiling 1946:343.

[ back ] 46. See Jakobson 1960. On the roots of his model, see Fine 1984:32 and on further refinements to the model, Dirven 1982:2.

[ back ] 47. Ebeling 1885:1122-24; Hofmann 1922:28-33; Fournier 1946:215-16; Chantraine 1968-80:718. Frisk 1960-70:2:264 defines muthos as "Wort, Rede, Gespräch, Überlegung, Erzählung, Sage, Märchen, Mythus," in an unhelpful collection of attested meanings. Some would go so far as to connect muthos (which has no known etymology) with words meaning "thought" in other Indo-European languages. See Hofmann 1922:47.

[ back ] 48. See Détienne and Vernant 1974:222, 231.

[ back ] 49. The demonstration by Russo and Simon 1968 that thought is often dramatized as internal speech in Homer can explain the tendency to translate the word this way.

[ back ] 50. and perhaps even in the word for "human," anthrōpos. See Pisani 1981.

[ back ] 51. Compare 22.451, describing how her thoughts are darkened by the sound of Hektor's voice: (opos ekluon). The vocal quality of epos also underlies the phrase at 17.695: amphasiē epeō.

[ back ] 52. Nagy 1979:265-75 and on the semantics of kleos, 16.

[ back ] 53. 53On the determinate, material nature of epos, see Fournier 1946:211-12 and Beck 1987. I cannot agree with the latter that the use of the two terms in one line is simply hendiadys.

[ back ] 54. Cf. the formula at 21.393 and 21.471, oneideion . . . muthon.

[ back ] 55. The Limba of Africa employ a similar metaphor of "heavy words" for important performative speech: see Finnegan 1969:550. For the image of the loaded ship, see Od. 3.312.

[ back ] 56. On the word nomos and ancient interpretation, see Ebeling 1885 s.v.; Hofmann 1922:5 wants the metaphor to refer to a flock of "winged words." I translate the word as "share-land" in order to capture the root meaning, relating to distribution (cf. the related verb nemō). If this meaning is overt in the noun, we could explain the image as connected with that of reciprocal movement in the next line.

[ back ] 57. As in Rotinese, in which the verb "talk" (fada), belongs to the same semantic sphere as words for "exchange, "barter": Fox 1974:78.

[ back ] 58. The same notion underlies Nestor's words to Agamemnon at 9.100: "You especially must say the epos and hear it back again."

[ back ] 59. See further instances at 1.108, 1.543, 3-83, 10.540, 24.767.

[ back ] 60. An expanded version of this formula occurs at 15.126-27. See also 14.137-38.

[ back ] 61. See also 18.384, 423 (Thetis with Kharis and Hephaistos).

[ back ] 62. Couch 1937:140. D'Avino 1969 sums up earlier attempts and concludes that an originally "durative" sense of "naming fully" is retained in some instances and transformed in others.

[ back ] 63. On this phenomenon of full-formula segmentation, see Muellner 1976:21-24.

[ back ] 64. One indication that this function is most important for epos rather than muthos is statistical: the former term appears most often in the dative in Homer; the latter rarely takes that case, occurring usually injust two formulas. See 1.565 and 2.245 for the types.

[ back ] 65. The propitiation of Apollo not long before this scene uses similar language (i. 100), as does the embassy to Achilles (9.639), the only mortal to whom this diction is applied.

[ back ] 66. In addition to these examples, cf. 1.273, I-565, 2.156-66, 20.295, 23.157; Helenus' prophetic advice to Hektor (7.43-53) transmits a directive of Apollo, obeyed by Athena (oud' apithēse). The force of Helenus' muthos comes from its representation of divine voice—op' akausa thēon.

[ back ] 67. See Herzfeld 1985 and 1985b.

[ back ] 68. Lord 1960:23-24.

[ back ] 69. On the characterization of heroes by style, see Chapters 3 and 4.

[ back ] 70. For similar uses of epos to refer to transmitted messages, see 1.652, 11.652, 17.701, 24.92.

[ back ] 71. A similar ideology has been traced in the Poema de Mio Cid: see Read Malcom 1983:2-21.

[ back ] 72. Barck 1976 examines the ramifications of this Homeric topos. An example of che contrast word versus deed (epos versus ergon) to mean "in all ways" is Thetis' supplication at 1.504; similar expressions with epos: 1.395, 5.879, 11.70.3, 15.106, 15.234.

[ back ] 73. On the formula, see Muellner 1976:127. Note that similar phrasing occurs at 14.44-45 describing Hektor's boast/threat.

[ back ] 74. The same might be concluded after we read Zeus' declaration at 8.8, "Let no male or female divinity cut through my word (epos)"—a speech referred to as muthos (29) by the poet. Here, however, it can be argued that the poet's label refers to the whole threatening performance of Zeus, while Zeus' term denotes just his (personal) command.

[ back ] 75. On marked versus unmarked, see Ducrot and Todorov 1972:148.

[ back ] 76. See Vivante 1975:2-8 on this image-evoking utterance in its purest form. To recognize the aptness of the metaphor, however, is not to specify the function of the speech introductions in which it occurs, and I do not agree with Vivante's impressionistic conclusion that the phrase refers to "sudden" words at points of reunion, recognition, danger, and perception.

[ back ] 77. Calhoun 1935:226. M. Parry 1971:414-18. Combellack 1950, a good summary of the controversy, includes J. A. Scott's comment to Calhoun on emotionality; Combellack himself saw no particular quality shared by "winged word" speeches.

[ back ] 78. On directives, see Searle 1976:11.

[ back ] 79. See Fingerle 1939 for the fullest description.

[ back ] 80. See Ervin-Tripp 1976:127—41 for an analysis of the types of directive. On "indirect" directives, see Searle 1979:36-48. Voloshinov 1930 was one of the first to draw attention to the linguistic importance of contextual implications (as opposed to verbalization).

[ back ] 81. On locutionary and illocutionary force, see Bach and Harnish 1979:4-8.

[ back ] 82. On the poet's formulaic artfulness in describing the death of Patroklos, see Lowenstam 1981:106-18.

[ back ] 83. On the thematic importance of these threats, see Segal 1971.

[ back ] 84. Birds: Kirk 1985:74; arrows: Latacz 1968;27-32, following Durante 1958.5-8. These are the two commonest solutions: see D'Avino 1981:89, who favors "wings," but sees a reference to the divine origin of sacral speech, epea. Hainsworth 1960:204n1 doubts that either image applies.

[ back ] 85. Gladstone 1874:844. Vivante 1970:5 seems to want a similar mixture.

[ back ] 86. Cf. Cunliffe 1924 s.v. On the phrase's associations, see also Lynn-George 1988:232-33.

[ back ] 87. For the Greek phrase, see 11.454, peri ptera pukna balontes (of scavenger birds). A comparison with 9.588, thalamos puk' eballeto (of a chamber under frequent assault) shows that the word pukna in 11.454 can be either adjectival or adverbial. In 23.879 it is clearly the former. On "interaction," see Silk 1974.

[ back ] 88. On persuasion and malakos speech, see 6.337 (Paris describing Helen's words); for the association of pareipon and epos, see also 6.54-62 (Agamemnon to Menelaos).

[ back ] 89. The same implication can be seen in the nearly exact line describing Odysseus' storytelling session with Penelope, Od. 23.300-301.

[ back ] 90. On these utterances, see Austin 1962.

[ back ] 91. We should recall that the herald's primary role is religious in Homeric poetry, and that the term and the institution are cognate with that of the Indic "ritual singer." On the religious function of Homeric heralds, see Mondi 1978.

[ back ] 92. For these meanings in enispes see Risch 1985.

[ back ] 93. 93See further on this speech the analysis of Zeus' style in Chapter 2. All other examples of the verb mutheomai are in similar contexts of formal and detailed speech: 7.76 (Hektor's declaration of rules for duel); 2.488 and 3.235 (reference to exact naming and enumerating of heroes); 6.376 (request for exact directions); 8.40 and 22.184 (reference to formal threats); 21.462 (of a formal declaration); 23.305 (of Nestor's detailed advice); 17.200 and 17.442 (Zeus' formal and detailed prophecies); 1.291, 20.246, and 20.433 (reference to lengthy abuse speeches, on which see Chapter 2).

[ back ] 94. On the later development of the term epos see Koller 1972 and Ford 1981.

[ back ] 95. Note here that the phrase kertomiois epeessi, "with cutting words," is also associated with the act. This shows the plural epea in its function as a periphrasis for muthos, as in the "winged words" phrase discussed earlier. On this blame genre, see Nagy 1979:222-42.