5. Language about Achilles: Linguistic Frame Theory and the Formula in Homeric Poetics

Charles Stocking

The Formula: Some Theoretical Considerations

Few if any scholars today can deny that the formula is an essential feature of Homeric verse. G. S. Kirk says in the preface to his Iliad commentary:
The whole question of the formular, conventional or traditional component in the Homeric language is extremely important for the exact appreciation of any particular passage, and of course of the whole poem. Something of a reaction is detectable at present from the extreme claims and inconclusive statistics that proliferated after the Milman Parry revolution, but it remains true, nevertheless, that the deployment of a partly fixed phraseology is a fundamental aspect of Homer’s style and technique—one that shaped his view of life almost. One can well ignore Homer’s “use of phrases” as an ordinary poet’s “use of words.”
Kirk 1985:xxiii
Kirk rightly asserts that to understand Homer, one must understand his use of the formula. And yet, Kirk also seems to hint that formula studies are in a state of aporia with the “extreme claims and inconclusive statistics that proliferated after the Milman Parry revolution.” [1] One may wonder to what extent one can have an “exact appreciation of any particular passage” without an exact understanding of what the formula is and what it does. One of the major difficulties in understanding the formula and its function, I think, can be attributed to the implicit assumption of many Homeric scholars that the formula is somehow unique to Homer, or at least to his genre, oral poetry. Some recent and not-so-recent studies in the field of linguistics, however, would indicate that Homer’s use of formulae may not be so different from an everyday speaker’s use of natural language. At a 1976 conference on statistical linguistics, C. J. Fillmore says about natural language, “an enormously large amount of natural language is formulaic, automatic, and rehearsed rather than propositional, creative, or freely generated” (Fillmore 1976:9). His comments suggest that something similar to the Homeric formula exists as an aspect of language in general, with as much significance as one attributes to the phenomenon in Homeric poetry. Thus a linguistic analysis of the formula, with the aim of gaining some new insights as to how the formula operates in Homer and to what end, seems worthwhile. Fillmore proposed a theory to account for the enormously large amount of formulaic natural language, known today as frame theory semantics. This essay will apply frame theory to the Homeric formula. As we will see, not only is the theory consistent with the way the formula operates within Homer, but I believe it will make for a useful tool in the interpretation of Homeric poetry, as I hope to demonstrate with an analysis of a speech by Achilles at XVI 200–209.
The groundwork for a linguistic frame theory analysis of the formula was first laid out in 1974, at the Ann Arbor Conference in Michigan, by Paul Kiparsky, prior to Fillmore’s own work on formulaic language. At the conference, Kiparsky equated formulae in Homeric poetry with the bound expressions of ordinary language. Bound expressions are of two types, fixed and relatively bound, which correspond nicely to Hainsworth’s (1968) fixed and flexible formulae. Bound utterances are considered to have at least one of three properties: non-compositional semantics, arbitrarily limited distribution, and frozen syntax. In terms of a generative grammar, fixed phrases are ready-made surface structures that can be considered as entries in a lexicon, whereas flexible phrases are defined by co-occurrence, insofar as there is a restriction between lexical items. This is the principle of non-compositional semantics, which would call for Homeric formulae not to be entered as individual words, but as formulae, when fixed, or with reference to their co-occurring words when flexible. Cunliffe’s lexicon of the Homeric dialect does reflect this principle with regard to word co-occurrences, though fixed formulae themselves have yet to be entered separately. [2] An example of arbitrarily limited distribution in English would be the phrase livelong day, which has no corresponding *livelong night. An example in Homer of arbitrarily limited distribution is θυμὸς ἀγήνωρ, where ἀγήνωρ is limited only to the term θυμός, in the Iliad, and only once used predicatively, even though as an adjective it could theoretically be applied to any number of nouns. [3] The equivalent property of limited distribution in flexible expressions in English would be kick the (blank) , which may be filled in with more than one word, but limited only to a few. An example of flexible limited distribution in Homer can be found in the phrase (blank) ἔμπεσε θυμῷ, which may be filled in with χόλος or δέος. An example of frozen syntax in English would be foregone conclusion, the syntax of which would not allow a phrase such as *foregone though the conclusion may be. The formula θυμὸς ἀγήνωρ also demonstrates relatively frozen syntax because it appears almost always in the nominative case at line end, and only once in the dative at Iliad XXIV 42, ἀγήνορι θυμῷ. Kiparsky’s analysis of the formula as bound expression has two major conclusions. First, Parry’s insight about the formula as a means for improvisation in metrical verse may be reconsidered as a special utilization of the formula, not its cause. [4] And ultimately Kiparsky concludes that “the language of oral literature does not differ qualitatively from ordinary language.” [5]
Thus, Kiparsky provides us with the first step in our attempt to rethink the formula in terms of language universals. Nevertheless, his account provides us with only a partial understanding of the formula. We are able to classify and identify formulae as bound expressions, based on formulae having any one of the three properties listed above. But when Kiparsky discounts ease of versification as the primary cause of the formula, we are left with the question “What motivates the formula?” [6] Following in Kiparsky’s footsteps, I suggest we look toward the study of bound expressions and their motivation in the field of linguistics for an answer.
Kiefer provides a general framework that we may use for our “formulaic question,” through a pragmatic-semantic approach to bound expressions. Kiefer understands bound expressions to be “stereotypical utterances, which are automatically evoked by certain speech situations” (Kiefer 1996:575). To better define and elucidate this notion, he employs Fillmore’s frame theory semantics. The notion of the frame was first introduced in the field of artificial intelligence by Minsky, whose original working definition is perhaps the easiest to understand:
When one encounters a new situation (or makes a substantial change in one’s view of the present problem) one selects from memory a structure called frame. This is a remembered framework to be adapted to fit reality by changing details as necessary. A frame is a data structure for representing a stereotyped situation, like being in a certain kind of living room or going to a child’s birthday party. Attached to each frame are several kinds of information
Minsky 1975:212
Fillmore applies this notion of frame to lexical semantics, arguing that in place of semantic markers, [7] frames should mark all lexical items. He states that “if a language has a lexical item, a part of our understanding of a text containing it is an understanding of the culture or world in which the classifications the word implies are sensible.” [8]
According to Kiefer, every general frame has any number of specialized sub-frames, which are defined by a set of subevents, known as a script. The script is also a notion that began in cognitive science, an outgrowth of artificial-intelligence studies. A script is a socially stabilized order of events that is derived from a person’s experiences (see Shank and Abelson 1977). Kiefer provides the following example of a frame and script:
Frame: Commercial Transaction
      Sub-frame: Commercial Transaction at the flea market
           script: (A is the buyer, B the seller)
           e1) A stops in front of the stand.
           e2) B makes an offer.
           e3)      (i) A is not interested in the offer.
                      (ii) A is interested in the offer and asks for the price
           e4) B answers A’s question.
           e5) A thinks the price is too high and offers a lower price.
           e6) B makes a counteroffer
           e7) A thinks the price is still too high and offers a price slightly lower than B’s last price.
           e8) B agrees
           e9) A pays
           e10) B hands over the merchandise
           e11) A thanks and says goodbye
           e12) B thanks and says goodbye.
Kiefer notes that each subevent may (and some must) be accompanied by one or more bound utterances. More often than not, he argues, bound utterances associated with particular subevents of a frame are semantically transparent, although they often can become shortened and opaque, and can be semantically demotivated. [9] Kiefer also interprets degrees of boundedness, not according to grammatical features but according to the events in which bound expressions occur. There are three possibilities for the boundedness of an expression in relation to an event.
1) Bound Expression (Ui) ← ↑ Event (i)
Here there is a 1:1 correspondence between bound expression and sub-event of a given script. In this case, Ui always points to E(i) and hence to the respective frame of E(i). And E(i) always prompts Ui. For example the bound expression “Paper or plastic?” has a one-to-one correspondence with a particular event at a supermarket, where the bag-boy asks which type of bag the customer would like his groceries in. Any use of that phrase in a different event would somehow be making a reference to the event at the supermarket.
2) Ui ← ↑ {E(i)}
In this case the utterance is associated with a large number of events. And the larger the set {E(i)} the less bound the phrase, and the more restricted the set of events, the more bound the phrase. An example of this relation can be seen in the relatively bound phrase of hip-hop slang, “being down X” which has several manifestations related to different events. The phrase can appear with any subject and in any tense, but when it is “down with that” or “down for that,” it is in a situation where the speaker expresses his consent to participate in an activity. Whereas when the phrase is “down with him, her, you, etc.” the speaker is making claim to those with whom he relates and associates himself. Hence the less bound a phrase the more events there are to which it may refer.
3) E(i) {Ui}
Here the event cannot be recovered by the phrase, simply because there are free expressions in a script as well as bound expressions. Ultimately, we may come to understand bound expressions as those utterances that facilitate the fulfillment of a sequence of events according to an individual’s cognitive frames.
Frame theory semantics and its application to bound expressions as presented by Kiefer will provide a useful model for understanding the function and motivation of the formula in Homeric poetry. What we see from Kiefer’s study is that the central motivation of bound expressions is their ability to reference particular frames. A speaker will use a bound expression to place a person into a particular frame, and the person hearing the bound expression will know what frame he is to be thinking in and with. The same holds true, I believe, for the Homeric formula. In fact, application of this theory unifies several observations and approaches to Homeric language into a single cohesive model. [10]
Let us begin with the notion of the frame as it relates to Homeric poetry. Our definition of a frame, to restate, is a data structure for representing a stereotyped situation. This linguistic notion can be found in the term theme as it is used in the scholarship of oral poetry. Lord has defined theme as “the groups of ideas regularly used in telling a tale in the formulaic style of traditional song” (1960:68). Gregory Nagy, in Homeric Questions (1996), provides a working definition following Lord (1960) of theme as “a basic unit of content.” At the Ann Arbor Conference, Nagy also stated that “the theme is the key to all other levels of fixity in oral poetry—including both the formulaic and the metrical levels” (Nagy 1974b:274). Calvert Watkins suggests something along the same lines when he writes that formulae are “set phrases which are vehicles of themes. The totality of themes may be thought of as the culture of the given society” (Watkins 1977:9). In Watkins’s quote, we see the formula working in the precise same manner as a bound expression, as a vehicle of the theme or frame. Watkins also indicates why themes are significant: because the totality of themes of a given society are equivalent to its culture. Yet perhaps, with an understanding of frames as a cognitive phenomenon, we may correct this equation slightly. Themes would not constitute culture for the archaeologist, who is looking at a society’s material properties. And so it may be more appropriate to say that the totality of themes may equal the experience of a given culture. Thus, a frame may be understood not only as an individual’s data structure for an event, but a cultural data structure. [11] These various treatments of theme all stress its significance for poetic composition as a basic thought structure for the poet to work with. That is to say, a stereotyped situation must be represented in the mind of the oral poet for him to perform the poem. We also see in these notions of theme how the formula is in the service of theme. It places the audience within the experience of a culture.
Because a frame is as much a part of culture as it is a part of individuals within that culture, we can have a linguistic understanding of John Miles Foley’s appropriation of receptionalism (Foley 1985:39–60). Receptionalist theory is a theory of literature in which the role of the “reader” is considered essential to the meaning of a “text,” and thus a major factor in its production as well. [12] Foley applies this model to oral literature, where reader becomes audience. As a result he must modify the theory slightly: “We may say that all members of the audience interpret the text according to a shared body of knowledge that is their inheritance” (Foley 1985:45, emphasis added). We may understand this inheritance of the audience to be its cultural inheritance. And later Foley uses an analogy from E. H. Gombrich’s Art and Illusion (1960) to say of the embodiment of this inheritance that “these stereotypes comprise not simply a convenient method of representation but a set of cognitive categories. If the interpretive contract signed by painter and viewer alike is adhered to, both participants are perceiving according to these categories” (Foley 1985:50, emphasis added). Foley draws this conclusion based on the fact that, according to Gombrich, there is an agreement that a pictorial representation of reality is “a transposition, not a copy” (Foley 1985:49). What matters for a person viewing a painting, then, is the internal logic of a painter’s work and how well that logic may be understood by the viewer. It is because painting is meant to portray a virtual reality that one must employ cognitive categories for representation. And yet even in everyday speech, as Kiefer and Fillmore argue, these cognitive categories are employed. It is not a special feature of art alone, but a basic fact of communication. Frame theory thus accounts for the cognitive categories shared by poet and audience, and it also accounts for the cultural aspect of these cognitive categories. Thus, in the utilization of bound expressions, the poet’s aim is to cause his audience to employ particular cognitive categories, which we may better understand as frames. Foley, I think, is exactly correct in his understanding of how oral poetry works. However, his model may be difficult to follow because of the twofold appropriation of theories foreign to an oral medium, and thus may lead to unnecessary complications in working out the details. Frame theory does not change Foley’s model, but rather streamlines it, and is perhaps a more appropriate conception insofar as it functions within the same sphere as oral poetry, that is, the sphere of verbal communication.
Thus we see how the linguistic frame and the oral-poetic theme are conceptual equivalents. As Kiefer demonstrated, frames rely on scripts. I turn to Thomas Walsh’s work on χόλος to demonstrate how an oral-poetic theme functions like a frame in its utilization of scripts. Walsh states, “This word χόλος is the technical term in the oral poet’s vocabulary for the story of the withdrawn hero, who refuses to fight, seeks refuge with a companion, and returns only too late” (Walsh 2005:192). Thus, Walsh argues that the term χόλος does more than make reference to notions of anger, it is conceived of as representing an entire set of subevents. In other words, the term χόλος contains within it a potential script. [13] Fillmore explains that frames and scripts are useful not only for the interpretation of groups of words; often the interpretation of abstract nouns requires a particular script. Fillmore offers the example of charity: “We have a schema in which one person gives something to another under the following conditions. The giver is under no obligation to make the gift, the receiver is put under no obligation to the giver as a result of receiving the gift, and the giver believes that his gift will benefit the receiver” (Fillmore 1976:17). The theory thus accounts not only for formulae, but for culturally charged individual words as well. [14]
Whereas Walsh’s work demonstrates an implicit application of scripts, Elizabeth Minchin makes a formal application, and she also realizes the cultural implications of the cognitive script:
I suggest that these so-called themes have been laid down, as scripts, in the memories of aspiring singers long before their apprenticeship. Homer, along with other singers and all members of their audience, would have acquired them in the normal course of living, either through his experience of life in the real world, or through listening to the stories of others. These units are not particular to the repertoire of the singer; they are part of everyone’s repertoire.
Minchin 2001:39
Minchin points out that the scripts the poet employs will be just as familiar to the audience as to himself. The aspects of Homeric poetry to which she applies scripts are type-scenes, lists and catalogues, descriptive segments, similes, and invocations. In other words, her treatment of the cognitive script deals with the narrative aspects in Homeric poetry. The script by itself, after all, is only conducive to the interpretation of narrative, since the subject matter is the sequencing of events.
Yet scripts are equally applicable to the speeches of Homeric poetry, through the function of the formula as a means of referencing a given script and frame. The notion of formulae as reference is also addressed by Foley, in what he calls the metonymy and traditional referentiality of the formula: “Traditional referentiality, then, entails the invoking of a context that is enormously larger and more echoic than the text or work itself,” and he also states, “Such a process of generating meaning I call metonymy, designating a mode of signification wherein the part stands for the whole” (Foley 1985:7). Foley accurately describes the function of bound expressions as they relate to frames, without any reference to frame theory semantics. And yet he deals only with the interaction between poet and audience, and the same may also be said of Minchin. Is it the audience alone that hears the echoic signification of formulae? Frame theory semantics, as a theory of communication, tells us that is not the case, for bound expressions do more than make reference to frames, they help facilitate action. [15]
This relation between speech and action is specifically dealt with by Richard Martin in his taxonomy of speech terminology in Homer, distinguishing ἔπος from μῦθος, the latter term characterizing what may be understood as speech-acts. [16] Martin argues that the individual characters are performers in their own right, through their individual speech acts, and that the whole of the Iliad is a speech act of memory. Martin’s work on speech-acts has important ramifications for the application of frame theory to Homeric poetry, [17] as it suggests that the general signification of formulae will be the same for both the characters within Homeric poetry and the audience hearing the performance. For the audience, the speech-acts of Homer, as the realization of the poetic performance, function as a bond between the generalized subevents of a frame/theme and the surface formulae, which occur within the speech-acts. That is to say, the speech-acts mark major points in the narrative of the Homeric poems, and formulae thus tell the audience what the relation and significance of those speech-acts to the narrative is, by referencing the frame/theme. And within the poem, that same referentiality of formulae to their respective frames is also occurring, but with the primary motive of facilitating some immediate action. [18] As we saw earlier, Minchin observes that the poet’s repertoire is the audience’s repertoire as well. We may add that such a repertoire also belongs to the characters within the poem.
Thus, I believe frame theory is the model that accounts for the formula most accurately. Not only is the theory native to the medium of oral poetry, but it systematizes and codifies the work of many scholars into a productive model. We may schematize frame theory in relation to oral poetry:
Cognitive Understanding                       Poetic Product
      1) Frame/Theme
           2) Sub-Frame:
                3) Script
                     E1                          4) ←Speech act → {formulae X/Y/Z …}
                     E2                                                   "
                     E3                                                   "
                     E4                                                   "
                     …                                                    …
The Formula Sets X, Y, Z … may have correspondence in a 1:1 ratio to events in a theme, or they may correspond to multiple events but still only in application to one theme, or to multiple events in various themes. As Kiefer suggested, the number of events to which a formula may apply depends on that formula’s degree of boundedness, and Kiparsky’s three criteria can be used to determine the degree of boundedness. It is important to note that this is not a descriptive model of the formula, but a cognitive model. A frame is a generalized culturally bound notion of an event and others like it. As such, frame theory is able to give an account of both the composition and reception of oral poetry. From Nagy, Watkins, Lord, and others, we see that theme motivates formula; according to Foley, every occurrence of a formula also invokes all other occurrences for an audience. Thus the audience may have a model, like that of the poet, so that for the audience, we may say that formula motivates theme.
A version of frame theory has been applied to Homeric studies before; D. Gary Miller (1982 and 1985) employed cognitive science to formulate a new model of oral composition. Miller primarily relies on the notion of a schema, which is more or less equivalent to what Kiefer identifies as a frame. Miller gives a grid, which is very much like the model provided above.
Subschema                                    / / \ \
                                                 / / / / / \ \ \ \ \
However, what Miller fails to achieve is the precise relation of formulae to scripts and frames. The last level of his chart, the level of formulae, is the most nondescript. Kiefer’s work on the bound utterance allows us to begin where Miller left off. Miller’s model, though deeply involved with cognitive science, is still only a model of composition, and the formula has no role greater than as result of compositional need, for Miller seems to understand the formula only in terms of the cognitive processes of composition, rather than in terms of their linguistic manifestation in speech. [19] He argues that the primary function of the formula is ease of versification (Miller 1982:42), and for this reason he objects to Kiparsky’s theory. If we understand the formula by the three properties of bound expressions mentioned earlier, then the formula cannot simply be an aid in versification, because what formulae signify is different from what individual words signify. In other words, formulae have special semantic features, which are what cause the poet to utilize them in verse (more so than individual words). Miller claims that “the matter of meaning is ‘automatically’ incorporated into formulae by virtue of the fact that any phrase (bound or otherwise) necessarily has a semantic representation” (Miller 1982:40). While this is true, Fillmore shows that there is a difference between the semantic representation of a bound utterance and a regular phrase. Compare for example, the two bound phrases cited by Fillmore:
“He was on land briefly this afternoon.”
“He was on the ground briefly this afternoon.”
Fillmore explains that “on land” is part of a frame whose other member is “at sea,” and “on the ground” is part of a frame whose other member is “in the air” (Fillmore 1976:15). And so we see that bound utterances are specifically related to cultural frames, so that they signify events, facilitate actions, and add cultural value more easily than the equivalent propositional arguments with individual words. In a sense, ease of versification does motivate the use of formulae, but it is a secondary motivation, compared with the cultural-semantic motivation. [20]
The speech of Achilles at Iliad XVI 200–209, I believe, provides a unique opportunity for demonstrating how the proposed linguistic frame theory model of the formula works, and that its unique status as a conveyor of cultural-semantics allows it to be a useful tool in the interpretation of Homeric poetry. Like any piece of speech or narrative in the Homeric epics, Achilles’ speech contains both formulaic and non-formulaic elements. However, the formulaic elements in Achilles speech appear markedly so, and would indicate that Achilles is aware of both formulaic language and the implications of that language according to frame-theory semantics. [21]

Linguistic Frame Theory Applied: A Language about Achilles

Μυρμιδόνες μή τίς μοι ἀπειλάων λελαθέσθω,
ἃς ἐπὶ νηυσὶ θοῇσιν ἀπειλεῖτε Τρώεσσι
πάνθ’ ὑπὸ μηνιθμόν, καί μ’ ᾐτιάασθε ἕκαστος·
σχέτλιε Πηλέος υἱὲ χόλῳ ἄρα σ’ ἔτρεφε μήτηρ,
νηλεές, ὃς παρὰ νηυσὶν ἔχεις ἀέκοντας ἑταίρους·
οἴκαδέ περ σὺν νηυσὶ νεώμεθα ποντοπόροισιν
αὖτις, ἐπεί ῥά τοι ὧδε κακὸς χόλος ἔμπεσε θυμῷ.
ταῦτά μ’ ἀγειρόμενοι θάμ’ ἐβάζετε· νῦν δὲ πέφανται
φυλόπιδος μέγα ἔργον, ἕης τὸ πρίν γ’ ἐράασθε.
ἔνθά τις ἄλκιμον ἦτορ ἔχων Τρώεσσι μαχέσθω.
“Myrmidons, let no one forget the threats directed at me, you, which you made against the Trojans during the whole period of my μὴνις, and each one of you would accuse me, ‘Stubborn one, son of Peleus, your mother reared you on χόλος, hard one, you, who hold your companions by the ships though they are unwilling. Let’s just go back home again with the sea-faring ships, since evil χόλος has fallen upon your heart.’ These things would you say to me thick and fast, gathering together. But now the great work of war shows itself, which you desired before. So let each man fight against the Trojans with a brave heart!”
Iliad XVI 200–209
What makes this speech particularly relevant for a frame theory model of the formula is the fictive quotation Achilles delivers at lines 203–206. We know that this is not an attempt at verbatim quotation because of its generalizing quality. The catalogue of Myrmidons has just been completed and Achilles is giving the rallying cry to his men, allowing them to go to war. Part of his strategy is to recall their previous threats and claims. Mentioning previous threats, as Achilles does, is not an unheard-of rhetorical move in the speeches of the Iliad. At VII 96, Nestor rebukes the Achaeans and addresses them as ἀπειλητῆρες. At XIII 219–220, Poseidon, in the form of Thoas, asks Idomeneus, Ἰδομενεῦ Κρητῶν βουληφόρε ποῦ τοι ἀπειλαὶ / οἴχονται, τὰς Τρωσὶν ἀπείλεον υἷες Ἀχαιῶν; (“Idomeneus, council bearer of the Cretans, where do the threats go, which the sons of Achaeans threatened against the Trojans?”), and the same formula of XIII 219 is used at XX 83, where Achilles addresses Aeneas. When Akamas has killed Promachus, he boasts at XIV 479, Ἀργεῖοι ἰόμωροι ἀπειλάων ἀκόρητοι (“Argive Bowmen, untiring in threats”). There are two speech-acts in which the mention of threats or boasts appear, either in rebukes or directives calling for action; the two can but do not have to coincide. Thus Achilles is keeping with traditionally scripted behavior in mentioning threats. However, the threats themselves intended for the Trojans appear to be directed at Achilles, because the Myrmidons are not actually fighting, hence the dative at XVI 200. Coupled with those threats is the complaint quoted at XVI 203–206. The very fact that the verb indicating the quote has ἕκαστος as the subject means that the quotation itself is different in nature from other Iliad quotations, such as we find when Odysseus quotes Peleus at IX 254–258. [22] Nestor also uses the same quote technique as Odysseus, when speaking to Patroclus, at XI 786–789. The purpose of these quotations appears to be to add vividness in re-activating the memory of the addressee, or at least that is the rhetorical strategy, for both Odysseus and Nestor, after delivering the quote, say, ὣς ἐπέτελλ’ ὁ γέρων, σὺ δὲ λήθεαι (“Thus the old man enjoined upon you, but you forget.”). Both quotations appear in directives, where the addressee is not behaving in accordance with the speaker’s desires—desires that the speaker demonstrates are rooted in the addressee’s own upbringing vis-à-vis quoting the father. Achilles’ quotation would seem to have a similar function, to vivify the memory of the Myrmidons’ threat. However, what Achilles quotes is not the threat itself, but the attendant complaint to Achilles. In addition to this odd choice of quotation, the quotation itself cannot have the semblance of an actual quotation if it is to be what ‘each man’ has said.
Unreal quotation is characteristic of one particular speaker in the Iliad, Achilles’ antagonist, Hector, and yet to a different end. At VI 459–461: καί ποτέ τις εἴπῃσιν ἰδὼν κατὰ δάκρυ χέουσαν·/ Ἕκτορος ἥδε γυνὴ ὃς ἀριστεύεσκε μάχεσθαι / Τρώων ἱπποδάμων ὅτε Ἴλιον ἀμφεμάχοντο (“And someone seeing you pouring forth tears may say, ‘This is the wife of Hector, who was best at fighting among the horse-taming Trojans, when they fought around Troy.”). And at XXII 106–108: μή ποτέ τις εἴπῃσι κακώτερος ἄλλος ἐμεῖο·/ Ἕκτωρ ἧφι βίηφι πιθήσας ὤλεσε λαόν. / ὣς ἐρέουσιν (“Lest a man worse than me should say, ‘Hector, persuaded by his own might, destroyed his people.’ Thus they will speak.”). Since the situations Hector imagines are those of either praise (Book 6) or blame (Book 22) both of which are executed verbally, giving a quotation demonstrates those potential genres of discourse being realized. [23] And yet, unlike Hector’s speech, which expresses a potential, Achilles’ quotation expresses something in the past.
Achilles’ quotation is similar in tense and aspect to the formula used in narrative, τις εἴπεσκεν, of which Hector’s quotations are the future-tense version. However, the formula τις εἴπεσκεν is often used with a partitive genitive, “of Achaeans and Trojans.” Hence the subject, τις, is a collective pronoun, whereas ἕκαστος is distributive, so that the speech-event, which Achilles “quotes,” is distributed among all the Myrmidons. In addition, the verb introducing the unreal quotation, ᾐτιάασθε, is in the imperfect tense. The distributive subject, coupled with the imperfect verb, indicates that Achilles conceives of the situation as an iterated speech-event, in much the same way as one would conceive of a generic speech act in relation to its script. [24] Because Achilles’ quotation is not simply repeating what has been said, nor projecting what will be, it must be re-presenting what has been said, as a projection of speech into the past. Achilles’ quotation is a poetic mimesis. As the scholia ad loc. reads, ἀπέστροφε τὸν λόγον ἀπὸ τοῦ διηγηματικοῦ ἐπὶ τὸ μιμητικόν. From a stylistic analysis of the quotation, we will see that Achilles’ mimesis is filled with conventional formulaic language, part of which is involved with the cultural poetic frame of χόλος, with the result that Achilles is positioning this speech-event within that frame. It is by virtue of the speech’s attachment to this frame that it is able to be conceived of as an iterated speech-event in the first place.
σχέτλιε / Πηλέος υἱὲ / χόλῳ / ἄρα / σ’ ἔτρεφε / μήτηρ,
νηλεές, / ὃς / παρὰ νηυσὶν / ἔχεις / ἀέκοντας / ἑταίρους·
οἴκαδέ περ σὺν νηυσὶ / νεώμεθα / ποντοπόροισιν
αὖτις, / ἐπεί / ῥά / τοι / ὧδε / κακὸς / χόλος ἔμπεσε θυμῷ. [25]
Iliad XVI 203–206


This word is one that contains within it an implicit scripted series of events, which is referenced every time the term is used. σχέτλιος occurs in the first foot of a line twelve of the thirteen times it appears. The word is defined as ‘obstinate’ (coming from *sche-thlios; see Chantraine 1968–1980, s.v.) and appears only in speech. [26] There are three basic ways in which it is used. First, in conjunction with a relative clause, σχέτλιος ὅς, as part of a characterization or name-calling based on a particular action, where the relative clause contains the action that is the driving force of the speech about the person. In II 112, Agamemnon gives a false directive to go home, calling Zeus σχέτλιος, ὃς πρὶν μέν μοι ὑπέσχετο καὶ κατένευσεν [27] Agamemnon mentions Zeus, presenting the following logic to his men: “Zeus promised that we’d win, but we’re going to lose, so go home.” The same exact lines are used at IX 19, where this time Agamemnon is not secretly trying to rally the troops but is sincere about his directive. In fact, II 111–119 are exactly identical to XIX 18–25, the first half of Agamemnon’s speech. σχέτλιος is used the same way at V 403, when Dione tries to console Aphrodite and speaks of Diomedes as σχέτλιος ὀβριμοεργὸς ὃς οὐκ ὄθετ’ αἴσυλα ῥέζων. And it is used again in a similar manner at VIII 361, although without the relative clause, when Athena complains to Zeus, σχέτλιος, αἰὲν ἀλιτρός, ἐμῶν μενέων ἀπερωεύς. These uses occur toward the middle of a speech.
The second way in which σχέτλιος is used is in direct abuse, rather than for the sake of introducing some other main action. This occurs when Helen has enraged Aphrodite, and Aphrodite responds, μή μ’ ἔρεθε σχετλίη, μὴ χωσαμένη σε μεθείω (III 414). And in Book 10, Diomedes rebukes Nestor for having woken him up: σχέτλιός ἐσσι γεραιέ· σὺ μὲν πόνου οὔ ποτε λήγεις (X 164). Lastly, Apollo rebukes the gods for not rescuing Hector’s body from Achilles: σχέτλιοί ἐστε θεοί, δηλήμονες· οὔ νύ ποθ’ ὑμῖν / Ἕκτωρ μηρί’ ἔκηε βοῶν αἰγῶν τε τελείων; (XXIV 33–34).
The third use may be classified under metri causa, although it serves stylistic ends as well. In our first group, σχέτλιος introduced an action or characterization essential to the speech, in the second group, σχέτλιος is purely a term of rebuke. In our third type of usage, we find σχέτλιος wedged in between two complete thoughts, where σχέτλιος is structurally and semantically unnecessary to the preceding or following lines. Ajax speaks about Achilles, saying, Ἀχιλλεὺς / ἄγριον ἐν στήθεσσι θέτο μεγαλήτορα θυμὸν / σχέτλιος, οὐδὲ μετατρέπεται φιλότητος ἑταίρων (Iliad IX 628–630). Here we see a complete thought phrase before σχέτλιος and another directly after. The same occurs when Glaukos rebukes Hector: πῶς κε σὺ χείρονα φῶτα σαώσειας μεθ’ ὅμιλον / σχέτλι’, ἐπεὶ Σαρπηδόν’ ἅμα ξεῖνον καὶ ἑταῖρον (XVII 149–151). From a propositional perspective, removal of σχέτλιος would only cause the meter to suffer. But the enjambed name-calling after a statement seems to add an emotive effect. [28] Achilles says of Patroclus, ἦ μάλα δὴ τέθνηκε Μενοιτίου ἄλκιμος υἱὸς / σχέτλιος· ἦ τ’ ἐκέλευον ἀπωσάμενον δήϊον πῦρ / ἂψ ἐπὶ νῆας ἴμεν, μηδ’ Ἕκτορι ἶφι μάχεσθαι (XVIII 12–14). Priam also addresses Hector, Πηλεΐωνι δαμείς, ἐπεὶ ἦ πολὺ φέρτερός ἐστι / σχέτλιος· αἴθε θεοῖσι φίλος τοσσόνδε γένοιτο / ὅσσον ἐμοί (XXII 40–42). And Hecuba addresses Hector right after Priam: τείχεος ἐντὸς ἐών, μὴ δὲ πρόμος ἵστασο τούτῳ / σχέτλιος· εἴ περ γάρ σε κατακτάνῃ, οὔ σ’ ἔτ’ ἔγωγε / κλαύσομαι ἐν λεχέεσσι φίλον θάλος, ὃν τέκον αὐτή (XXII 85–88).
At the most mechanical level, σχέτλιος is merely a means of starting a new line. And yet, the form of address is not void of meaning. It is not socially relevant nomenclature, since mortals apply it to immortals (Agamemnon to Zeus), immortals to mortals (Dione to Diomedes), parent to child (Priam and Hecuba) and comrade to comrade (Ajax, Glaucon, Achilles). Rather, I would suggest that σχέτλιος is a term applied in a particular script. It may be used simply as the first foot of a line for emotive effect, or it may be syntactically connected to whatever thought is being expressed. But it is always used in the same scripted series of events: when someone delivers a directive/rebuke (as a result of not carrying out a directive) because the person in question is behaving contrary to expectation, and this leads to frustration on the speaker’s behalf. Achilles’ use of the term in the quotation is consistent with its other uses. However, in the case of Book 16, it is the first word of the quotation, which occurs only once else (Book 24). By positioning σχέτλιος as the first word of a mimetic quote, Achilles takes what is a common element of rebuke-directives and uses it as a typological marker for such speeches. In other words, he is establishing conventional modes of speech through the process of mimesis even as the epic progresses.
Achilles represents this particular speech type, a σχέτλιε speech, in connection with speeches directed at himself, through the use of the vocative, Πηλέος υἱέ. David Shive considers the string σχέτλιε Πηλέος υἱέ as a single, unique formulaic unit of address (Shive 1987:116). We have already seen that σχέτλιε often occupies the first foot of the hexameter line. Πηλέος υἱέ beginning the second foot occurs as a formula five times in the Iliad, three in speech, one in narrative, and one in direct address. Shive mentions three of the examples to demonstrate Homer’s favoring a creative rather than economical approach to formulae ending at the feminine penthemimeral caesura: XVIII 18, ὤ μοι Πηλέος υἱὲ δαΐφρονος ἦ μάλα λυγρῆς, where ὤ μοι adds emotion; XX 2, ἀμφὶ σὲ Πηλέος υἱὲ μάχης ἀκόρητον Ἀχαιοί, where he argues the prepositional phrase is unnecessary; and XXII 8, τίπτέ με Πηλέος υἱὲ ποσὶν ταχέεσσι διώκεις, where τίπτε breaks regular word order (although on what grounds he does not make clear). Shive’s notes on these three passages are part of his attack on Parry’s idea of Homer’s “traditional” language. Shive’s mistake is the need to consider σχέτλιε Πηλέος υἱέ as a single formulaic unit. Who is to say that the formula Πηλέος υἱέ did not have that very purpose of freeing up the first foot and the rest of the line following the caesura after it, as a boundary marker of sorts? It is significant that Shive notices the first hemistich of XVI 203 as a unique formula; it does seem to have formulaic status. What he failed to notice was the context of this unique formulaic unit. Achilles has constructed the first hemistich of XVI 203 from a scripted lexical item, the significance of which we have seen above, in conjunction with a formulaic form of address for Achilles. In connecting the lexical item and its implied script with the name of Achilles, Achilles connects the generic speech type with speeches addressed to himself, thus suggesting that the speeches addressed to Achilles are part of a certain generic script. [29] That script is part of the χόλος frame.


Like the initial address, the first phrase of Achilles’ mimetic quotation lacks any degree of boundedness and is comprised of individual lexical items: χόλῳ ἄρα σ’ ἔτρεφε μήτηρ. Thematically, Achilles’ quote is a speech about χόλος, given that it is used twice in the course of four lines. In fact, the majority of occurrences of the noun are in the two metrical slots that it occupies in this speech, after the feminine penthemimeral caesura and the hephthemimeral caesura, the two most common caesuras in the Iliad. A word such as χόλος has a large amount of metrical dexterity; it could appear in almost any spot, and does. However, Tom Walsh has demonstrated the poetic importance of χόλος. The fact that occurrences of χόλος cluster around the caesuras would indicate a confluence of meter and content in the poet’s composition of his lines.
The form is unique following the feminine penthemimeral caesura. The only forms found are χόλος and χόλον, both of which are metrically equivalent to χόλῳ, short-long, if followed by a word beginning with a consonant. In this metrical slot, there are only two fixed phrases using χόλος or χόλον: χόλος λάβειν before a bucolic diaeresis at I 387 and VI 166, and χόλον θυμαλγέα πέσσει at IV 513, with variant πέσσων at IX 565. The other seven occurrences of noun forms of χολ- in this spot are individual lexical items, usually filling the rest of the fourth foot with particles such as δέ τε or δέ μιν, or verbs. Achilles’ construction following the first hemistich is not unusual as far as formulae and metrics are concerned. But nowhere is it used as a dative of instrument. χόλῳ occupies a primary spot in the line, after a caesura, and is also the first element of the phrase, and so it stresses the innovation of the statement—that he was weaned on bile, as opposed to milk (cf. scholia A: ὑπερβολικῶς οὐ γάλακτι, ἀλλὰ χολῇ). This exceptional form of the noun anticipates how we are to render the verb. Forms of the verb occupy the fifth foot following the bucolic diaeresis only once. Hector speaks about Paris at VI 282: μέγα γάρ μιν Ὀλύμπιος ἔτρεφε πῆμα (“For the Olympian has nursed him as a great pain”). The same grouping of "pain" and "nursing" occurs at XXII 421; this time Priam speaks about Achilles: Πηλεύς, ὅς μιν ἔτικτε καὶ ἔτρεφε πῆμα γενέσθαι. In this second example we have a Hainsworth-type formula, a “repeated word group” (Hainsworth 1968:35), or a relatively bound phrase. This flexible phrase, which shares the same metrical placement of the verb, is also similar in sense to Achilles’ phrase at XVI 203 and may have had some influence on the construction of that phrase. We see from these other two examples that "rearing" can have negative connotations, best expressed by Thetis’ words to Achilles at I 414: ὤ μοι τέκνον ἐμόν, τί νύ σ’ ἔτρεφον αἰνὰ τεκοῦσα; (“My dear child, why did I raise you, bearing you bitterly?”). The last word of the first line of the quotation is a metrically fixed lexical item in the Iliad, occurring 42 of 63 times in the final foot, usually in the formula πότνια μήτηρ, τέκε μήτηρ, γείνατο μήτηρ, or otherwise.
And so the elements of this first phrase are attested elsewhere in the Iliad. What is significant is that all three major elements, χόλος, τρέφω, and μήτηρ, are associated with Achilles. The majority of the occurrences of χόλος in noun form pertain to Achilles or are used by him. χόλος relates to Achilles 25 of the 49 times that it appears, just over 50 percent, while the other uses are distributed among the gods and some heroes. Cunliffe understands χόλος here to stand for χολή, ‘bile’, but this does not preclude character association with Achilles. And Thomas Walsh has demonstrated that the term χόλος indicates more than anger, but rather the anger associated with a particular series of events. To restate, Walsh defines χόλος as “the technical term in the oral poet’s vocabulary for the story of the withdrawn hero, who refuses to fight, seeks refuge with a companion, and returns only too late” (Walsh 2005:192). And he defines three locations or contexts for this type of anger: 1) contest and challenge (168–171); 2) dispute and quarrel, or νεῖκος (171–175); and 3) the death of a φίλος (175–182). Achilles uses that very same poetically charged word here in a nonconventional manner as far as the Iliad is concerned, in connection with his own mother and his raising. In addition, the whole line lacks any degree of boundedness. In this way, Achilles gives a representation of the reworking of a generic poetic frame specifically in light of himself. It appears that Achilles is representing the Myrmidons as seeing Achilles fundamentally linked to this poetic script and frame. If Achilles is being reared on a particularly poetic and culturally significant form of anger, we may ask whose anger. Without a doubt it would be the χόλος of Thetis, the context of which would be Walsh’s site three, the death of a φίλος, namely her own son. [30] It is as though the poetic frame of χόλος has been genetically passed on to Achilles, so that Thetis’ χόλος becomes Achilles’ χόλος and informs the basic plot of the Iliad. [31] Thus, in the first line of Achilles’ mimesis, the beginning address, σχέτλιε Πηλέος υἱέ, provides information about the general form of what is said to him. They are directive rebukes. The phrase that follows is unique from a formulaic perspective, but again with elements common in speaking about Achilles, which still reference a particular frame. In his commentary on this quotation, Janko (1992:345) mentions Achilles’ sensitivity to Patroclus’ charge of savageness, which includes discussion of his mother at XVI 33–35:
νηλεές, οὐκ ἄρα σοί γε πατὴρ ἦν ἱππότα Πηλεύς,
οὐδὲ Θέτις μήτηρ γλαυκὴ δέ σε τίκτε θάλασσα
πέτραι τ’ ἠλίβατοι, ὅτι τοι νόος ἐστὶν ἀπηνής
“Hard one, your father was not the horseman Peleus, nor Thetis your mother, but the gray sea bore you, and the steep rocks, because your mind is unbending.”
Janko mentions the accusation of savageness only as the point of comparison, but there is at least one linguistic connection as well.


At XVI 204, Achilles uses the term of address νηλεές. Popularly, the term is an epithet of χαλκός, occurring eleven times as such in the Iliad. Most likely based on its use as an epithet, it also applies to ὕπνος, ἦμαρ, ἦτορ, and θυμός. Seldom does the word occur as a term of address by itself, except in Patroclus’ speech (above) and Ajax’s at IX 632 with hyphaeresis, νηλής. Both these occurrences are in the first foot of the line, and the only other occurrence in the first foot is in Phoenix’s speech in IX 497: νηλεές ἦτορ ἔχειν. And so we have the less regular usage and metrical occurrence of νηλεές occurring in other speeches about Achilles and also by speeches from actual Myrmidons, and Ajax as well. The use of νηλεές may be accounted for compositionally because of its earlier occurrence in Patroclus’ speech in Book 16 and so would be fresh in the mouth of the poet and in the ears of the potential listeners, a result of D. Miller’s “memory chunks” phenomenon. But it seems beyond chance that the other two metrical fits are also in speeches directed at Achilles. In light of its use here, and by Ajax, Phoenix, and Patroclus, [32] we may begin to create a formula set for the χόλος frame, although this is an extension of the proper frame for νηλεές as an epithet of bronze in battle scenes. [33]

XVI 204

The rest of XVI 204 does not have language that is markedly connected to Achilles, but does contain conventional elements. The first is the formula παρὰ νηυσί(ν). It occurs 36 of 49 times before the feminine penthemimeral caesura. There are several other formulae, which include παρὰ νηυσί(ν). One line of particular interest from a compositional view is II 391–393, Agamemnon’s threat,
ὃν δέ κ’ ἐγὼν ἀπάνευθε μάχης ἐθέλοντα νοήσω
μιμνάζειν παρὰ νηυσὶ κορωνίσιν, οὔ οἱ ἔπειτα
ἄρκιον ἐσσεῖται φυγέειν κύνας ἠδ’ οἰωνούς.
“Whomever I see having gone far from battle, remaining beside the curved ships, for him it will not be possible to escape the dogs and birds.”
The phrase μιμνάζειν παρὰ νηυσὶ κορωνίσιν occurs nowhere else, although μιμνάζειν παρὰ νηυσὶ occurs at X 549, about Nestor, and παρὰ νηυσὶ κορωνίσιν occurs at II 297, IX 609, XIX 44, XX 1, XXII 508, XXIV 115, and XXIV 136. Here is a case of a complex formula, composed of two separate formulae but acting as a single constituent. [34] The phrase παρὰ νηυσί(ν) is the kernel of a series of formulae, where the base form is used at XVI 204 as part of a constructed speech. παρὰ νηυσί(ν) also occurs a total of 32 times in speech. Use in speech more than narrative is significant, because in narrative the formula could simply be locating a particular act, and have a purely descriptive function. But within speech, such as in Agamemnon’s threat, παρὰ νηυσί(ν) often seems to have an implied antonymy, meaning not ‘on the battle field’ and not ‘at the walls of Troy’: ἀπάνευθε μάχης. Thus the phrase παρὰ νηυσί(ν), when used in speech, implies a script much like Fillmore’s “on land” vs. “at sea” example. And far from mere description, this small prepositional phrase contains within it the heroic ethic of action. We see this in Thetis’ wish that Achilles remain by the ships at I 416, where she speaks of Achilles as αἶσα μίνυνθα, and in Hector’s boast at VIII 180–183:
ἀλλ’ ὅτε κεν δὴ νηυσὶν ἔπι γλαφυρῇσι γένωμαι,
μνημοσύνη τις ἔπειτα πυρὸς δηΐοιο γενέσθω,
ὡς πυρὶ νῆας ἐνιπρήσω, κτείνω δὲ καὶ αὐτοὺς
Ἀργείους παρὰ νηυσὶν ἀτυζομένους ὑπὸ καπνοῦ.
“But when I am at the hollow ships, let there be some remembrance of the destructive fire, when I burn the ships with fire, and I will kill the Argives beside the ships panic stricken because of the smoke.”
The implied script of παρὰ νηυσίν explains why Achilles’ companions are presented as ἀέκοντας ἑταίρους; they want to fight.
Both ἀέκοντας and ἑταίρους frequent their respective metrical slots in XVI 204, but never occur together as such. The oblique cases of the adjective ἀέκων primarily occur in the bound phrase referring to horses, τὼ δ’ οὐκ ἀέκοντε πετέσθην, eight of the thirteen times that it occurs in the metrical slot of XVI 204. ἑταῖρος as an individual lexical item here does not so clearly derive from a bound expression. The various forms of ἑταῖρος (excluding the four-syllable dative plural) occur at line end 98 of the 121 times that the noun occurs in the Iliad. Nor is the noun attached to any particular formula. There are some formulae in which forms of ἐσθλός, ἔθνος, πίστος are attached to the noun to complete a line after the bucolic diaeresis, but these are few compared to the number of different lines that occur with ἑταιρ- at line end, most of which lack a bucolic diaeresis. The status of ἑταῖρος and its almost invariable position would indicate that the metrical fixity of individual lexical items in a line is not always derived from formulae, as above. Rather, it can have independent status and perhaps be a major keystone in the process of composition. [35] XVI 204 is the only line of Achilles’ mimetic speech that does not contain a bucolic diaeresis, and yet what appears to be a stylistic irregularity is in keeping with the conventions of the Iliad.

XVI 205

The conventional character of Achilles’ mimetic speech comes through most clearly at XVI 205: οἴκαδέ περ σὺν νηυσὶ νεώμεθα ποντοπόροισιν. This line presents overlapping bound formulae centered around the verb. νεώμεθα occurs at II 236, III 283, VII 335, XIV 505, and XXII 392, all in the metrical position of XVI 205, most likely because it allows for a bucolic diaeresis. At II 235–236, Thersites rebukes the Achaeans: ὦ πέπονες κάκ’ ἐλέγχε’ Ἀχαιΐδες οὐκέτ’ Ἀχαιοὶ / οἴκαδέ περ σὺν νηυσὶ νεώμεθα, τόνδε δ’ ἐῶμεν (“Oh weaklings, evil reproachable Achaean women, no longer Achaean men, let’s go back home with the ships”). Similarly, Agamemnon at III 283, praying to Zeus before the duel of Menelaos and Paris, says ἡμεῖς δ’ ἐν νήεσσι νεώμεθα ποντοπόροισιν (“Let us go back in the sea-faring ships).” Achilles does not seem to be referring to either of these speeches, in the way that he may be making reference with νηλεές to speeches about himself. Thersites is most likely using conventional language, especially since Nestor at VII 96 reproaches the Achaeans with the exact same phrase, Ἀχαιΐδες οὐκέτ’ Ἀχαιοί. [36] Also at VII 335, Nestor says οἴκαδ’ ἄγῃ ὅτ’ ἂν αὖτε νεώμεθα πατρίδα γαῖαν, which contains elements of XVI 205. In fact all the elements of these two formulae occur in a number of combinations with each other. XVI 205 compresses those combinations into a single line. Returning home is also a common theme in the speeches of the Iliad as a threat—Agamemnon in Book 2, Thersites in Book 2—and also as a positive sign of release from fighting—Agamemnon in Book 3, Nestor Book 7, the latter most likely the inspiration for the former.

XVI 206

αὖτις, ἐπεί ῥά τοι ὧδε κακὸς χόλος ἔμπεσε θυμῷ.
The last line of Achilles’ quotation has in part the same influence as νηλεές. αὖτις occurs occasionally in the first foot, although it also frequents the second and third feet, which would indicate that its positional appearance is variable and not fixed. What is interesting in this particular line is the fact that αὖτις is enjambed (which occurs nowhere else that it begins a line). Milman Parry discusses three ways in which one verse may relate to a following verse (M. Parry 1971:253). First, a sentence may end with the line, and a new sentence may begin the next line. The second type is when a thought is complete by verse end, but the sentence continues the next verse by adding free ideas through new word groups. This he refers to as unperiodic enjambment, a term taken from Dionysus of Halicarnassus. Finally, a verse can end without the thought completed, in which case enjambment is necessary. Carolyn Higbie adds two more types of enjambment to the list, and she also reclassifies Parry’s “necessary” as clausal and suggests necessary enjambment to be those instances when the verb is enjambed, since that is the crucial element in a clause (Higbie 1990:49). Her fourth type is violent enjambment, where there is close connection between the elements that verse end divides. The use of αὖτις by Achilles can be classified as this second, unperiodic, or adding type, given that the previous line could easily make a sentence on its own. Furthermore, αὖτις adds nothing semantically to the previous line, since returning home implies going back. At least here, αὖτις seems to have a purely functional status as line filler. The same appears true for the rest of the first half-line up until χόλος. αὖτις, ἐπεί has a strong phonetic similarity to the common line beginning αὐτὰρ ἔπειτα. Furthermore, the only cases in the quotation of the metrical phenomena of hiatus and correption occur at the first half of this line with τοι ὧδε. It is significant that there has not been hiatus or correption up until this point. Parry gives an account of hiatus as the desire to express a thought being stronger than the need for metrical regularity (M. Parry 1971:191–196). We may exclude from Parry’s generalization any such phrases that originally included a digamma, although that is not the case here. Thus far Achilles has been demonstrating conventional language, what “everyone” says, and so there has been no need to express a thought that has not already been manifested in formula. This first half-line does not seem to express any thought significant enough to require hiatus either. And yet, αὖτις, ἐπεί ῥά seems like a formulaic echoing already. The strain in the conventional language appears to be a result of Achilles’ efforts to get to the main concern of this line, the fixed formula χόλος ἔμπεσε θυμῷ.
All three elements χόλος, ἔμπεσε, θυμῷ occur in their metrical slots as individual lexical items, but it is their status as bound phrase that is most important at XVI 206. The entire phrase is used by Phoenix in Book 9: βάλλεαι, οὐδέ τι πάμπαν ἀμύνειν νηυσὶ θοῇσι / πῦρ ἐθέλεις ἀΐδηλον, ἐπεὶ χόλος ἔμπεσε θυμῷ (435–436). It may be the case that Achilles is referring to Phoenix’s speech at this point. The formula occurs in a line that Achilles himself heard. However, the formula occurs in various forms throughout the Iliad. At XI 155, we have a Hainsworthian word-expectancy formula with the same mention of fire that is in Phoenix’s speech: ὡς δ’ ὅτε πῦρ ἀΐδηλον ἐν ἀξύλῳ ἐμπέσῃ ὕλῃ. More important are the occurrences of the entire formula in Book 14. First, at XIV 206–207, Hera requests that Aphrodite give her φιλότητα καὶ ἵμερον (198) under the pretext that she is going to resolve the dispute between Oceanu and Tethys: ἤδη γὰρ δηρὸν χρόνον ἀλλήλων ἀπέχονται / εὐνῆς καὶ φιλότητος, ἐπεὶ χόλος ἔμπεσε θυμῷ. And that portion of her speech is repeated when she speaks to Zeus at XIV 300. So the formula broadly applies to all situations where strife is intended to be resolved. [37] While we do not know if the formula is Achilles-specific, it does apply to him and the situation, and Achilles is most likely aware of both these facts. Walsh identifies this formula as “a narrative marker, to indicate the exact moment at which the present situation came to be” (Walsh 2005:133). Thus the two occurrences of χόλος in this mimesis indicate the transfer of this poetically charged anger from mother to son, from past to present. Based on the information presented thus far, the conventional character of Achilles’ mimesis is clear. I would suggest even further that Achilles means to employ conventional language about himself. This is evident from the fact that many of the conventional elements, though not all, come from three major characters and their speeches to Achilles: Patroclus, Phoenix, and Ajax. The use of νηλεές suggests most strongly that the mimesis presents more than just formulaic language; it presents formulaic language about Achilles, because as a term of address, it is not the normal usage in the Iliad, except as it applies to Achilles. The Achilles-specific conventions are combined with mere conventional formulaic elements, such as line 205, and there are other aspects that are part of both categories, such as the term σχέτλιος and the last formula, χόλος ἔμπεσε θυμῷ. Through his employment of conventional elements in conjunction with his two mentions of χόλος, he achieves two ends. In the first place, we have a mimesis of a generic speech-act, a χόλος speech (which Achilles seems to identify by the term σχέτλιε). Ostensibly, this is a mimesis of the Myrmidons, but this analysis reveals that the language of Achilles’ quotation has more similarities with the speeches involved with the χόλος frame than with any specifically Myrmidonic language. The conventional language in Achilles’ mimesis reflects a phenomenon of formulae clustering as result of thematic motives, otherwise known as frames. And the fact of mimesis presented within the poem gives us the added bonus of character awareness in such an operation. The second achievement of this mimesis is the intimate connection of this generic speech-act with Achilles’ own character, as the address σχέτλιε Πηλέος υἱέ indicates, as well as the phrase χόλῳ ἄρα σ’ ἔτρεφε μήτηρ. In presenting the mimesis of an iterated speech-event, of which Achilles is the subject, and the frame of which is χόλος, we see Achilles become a part of the traditional frame. Still, we may wonder and even doubt whether there is any difference between the conventional character of Achilles’ mimesis and the conventional character of any other speech in the Iliad. One might consider it to be simply a good imitation of speech, without having to draw conclusions about Achilles’ special knowledge of how epic language works. This might very well be the case, were it not for the fact that the conventional character of Achilles’ mimetic quotation is offset by Achilles’ own non-conventional and self-reflexive language in the rest of the passage.

The Language of Achilles, Revisited

The very first line of Achilles’ speech, XVI 200, gives us a typological introduction to the activity and quality of this particular speech. As I noted earlier, it is common in speeches to begin by mentioning previous boasts. In fact, there is one other occurrence of ἀπειλάων in its position here: Ἀργεῖοι ἰόμωροι ἀπειλάων ἀκόρητοι (XIV 479). But the conventionality of Achilles’ first line stops there. Chantraine notes that nowhere else in the Iliad and only once elsewhere in the Odyssey, xvi 303, is there a third-person aorist imperative prohibition (Chantraine 1953:231). Achilles is thus in one respect “generating grammar.” The aorist imperative prohibition is necessary in achieving the generic subject, μή τις … ἕκαστος, which will be the object of his mimesis. The verb that takes this unprecedented form is λανθάνω. Although the form of the verb is unattested elsewhere, forms of λανθάνω do occur in the last foot and a half of the hexameter line, such as λανθάνοντο (IV 127), and ἐκλελαθέσθαι (VI 285). The verb is especially apropos to the mimesis that follows. λανθάνω is a rhetorical device regularly employed when calling someone to action. And we see the rhetorical use of the verb as a result of what appears to be a deep-structure connection between action and forgetting. [38] For forgetting may also be the pointed effect of blame speech acts, as Hector claims at XXII 281–282: ἀλλά τις ἀρτιεπὴς καὶ ἐπίκλοπος ἔπλεο μύθων, / ὄφρά σ’ ὑποδείσας μένεος ἀλκῆς τε λάθωμαι. And forgetfulness is the effect that death ultimately brings, as the description of the dead Cebriones indicates at XVI 775–776. While forgetting and fighting seem to be fundamentally linked within the action of the poem, forgetting and speech are also fundamentally linked, as we see with Nestor’s speech from XI 786–789 (above). [39] The work of Miller and Minchin has shown the importance of memory in frame theory and how this may apply to composition, through the abstract frame and cognitive script (Miller 1985:354–371; Minchin 2001:11–31). In light of the linguistic application of frame theory, I think we can account for the two classes of objects that λανθάνω takes in Homer. [40] Kiefer demonstrates that there is a correspondence between speech-acts and events in a script. Thus, by metonymy, forgetting a particular speech, as Achilles does (Book 9) and as Patroclus does (Book 11), entails the forgetting of an entire script, which, in a heroic context, would indicate proper paths of behavior. The cognitive frame as cultural tradition not only helps in the telling of stories, but in the culture depicted within the narrative, particular scripts may have a prescriptive function. Thus, when Cebriones is described as having forgotten his horsemanship, his death marks his disengagement from the culture of which he was a part, and the scripted action in which he participated as a part of that culture. When, in Book 11, Nestor makes that accusation of forgetting to Patroclus, he asserts that Patroclus has gone through the same process of disengagement. The verb λανθάνω, however, appears nowhere else in the Iliad with ἀπειλή or its derivatives, though the effect of other speeches may indicate precisely that semantic relationship. Still, the threat itself is something different from strength, or from horsemanship. To warn against forgetting one’s threat may be the equivalent of the bound phrase “put your money where your mouth is.” But an ἀπειλή is also a genre of discourse itself (Martin 1989:65–77), and as such there are two implications for Achilles’ warning not to forget one’s threat. First, Achilles is repositioning his men in the proper frame and event to which threat speech-acts belong, battle. And yet to warn against forgetting one’s boasts is also to warn against forgetting one’s power as a speaker, a power that Achilles then demonstrates. [41]


The cultural-poetic awareness with which Achilles might be operating suggests itself most strongly in the first hemistich of XVI 202, πάνθ’ ὑπὸ μηνιθμόν. The phrase is noted by Leaf as yet another “linguistic peculiarity” of the passage, since ὑπό + acc. with a temporal sense is used only at XXII 102 in narrative. Scholia A sees this as one possibility, the other being that μετὰ πάσης μήνιδος καὶ χόλου was meant, although this appears an entirely unheard of use of ὑπό. Eustathius uses the phrase as a temporal referent to situate the narrative, indicating that the rage was sixteen days long. What we may infer from Eustathius’ observation is that Achilles’ mention of the rage is what ends the time of his rage. The word in question, μηνιθμόν (from the verb μηνίω and cognate with μῆνις), [42] occurs only in Book 16: once at lines 61–62, where Achilles responds to Patroclus: ἀσπερχὲς κεχολῶσθαι ἐνὶ φρεσίν· ἤτοι ἔφην γε / οὐ πρὶν μηνιθμὸν καταπαυσέμεν “([it is not possible] to have unending χόλος in the mind, although I said before, that I would not stop the μηνιθμόν”). And again at XVI 281–282; this time the narrator speaks: ἐλπόμενοι παρὰ ναῦφι ποδώκεα Πηλεΐωνα / μηνιθμὸν μὲν ἀπορρῖψαι, φιλότητα δ’ ἑλέσθαι (“[The Trojans] expecting that the son of Peleus had left his anger beside the ship, and taken up friendship).” Leonard Muellner cites these examples in Book 16 as the “polar opposite of Achilles’ withdrawal” (Muellner 1996:135) in Book 1. Achilles’ use at XVI 61 seems to be an overt reference to his response to Ajax at IX 650–655, but in language that echoes IX 517–518, where Phoenix says, οὐκ ἂν ἔγωγέ σε μῆνιν ἀπορρίψαντα κελοίμην / Ἀργείοισιν ἀμυνέμεναι χατέουσί περ ἔμπης. [43] While μῆνις is not exclusive to Achilles, [44] μηνιθμόν and its associations with the μῆνις of Book 9 are specifically Achillean. The thematic character and connection of μηνιθμόν and μῆνις to Achilles are what allow for it to be an adverbial time marker. The infix -mo- of μηνιθμόν, according to Chantraine (1968:137), has an IE origin and is used in nouns indicating action derived from a verb, in this case μηνίω. Because the action form of the noun μηνιθμόν only applies to Achilles, the action has certain entailments, namely the withdrawal of Achilles in Book 1, mentioned by Muellner. The relation of μῆνις to action indicates that it functions as a frame, much like χόλος. Nagy was the first to suggest that μῆνις was the name of a theme in oral poetry (Nagy 1979:72–73). Muellner takes up Nagy’s suggestion with an in-depth analysis to understand the narrative “action” that the theme implies; as he says, μῆνις “is not separate from the action it entails.” [45] μῆνις through its reference to narrative action functions in the precise same manner as a frame. This narrative action to which μῆνις refers is linguistically included in the form μηνιθμόν and refers specifically to Achilles’ own action, although it is without a possessive, thus indicating the strong, and in this case exclusive, connection between cultural poetic frames and the figure of Achilles. At XVI 202, Achilles is not simply referring to his previous claim in Book 9, nor is this a formulaic or grammatically regular use of μῆνις/μηνιθμόν, but it appears that Achilles has the same formal understanding of his rage as does the poet. [46]

XVI 207

After Achilles has delivered his mimetic speech, he concludes with ταῦτά μ’ ἀγειρόμενοι θάμ’ ἐβάζετε. Both the form and position of ἐβάζετε are entirely irregular. Baz- appears as βάζεις twice and βάζειν once, and only at verse end. To even talk about the unprecedented mimetic speech act, Achilles must use unprecedented language. The adverb θάμα also reveals Achilles’ implicit accusation of ‘each man’s’ conventional speech. The adverb itself is related to the adjective θαμέες which Cunliffe glosses as 1) 'standing close together'; 2) 'flying, falling coming thick and fast'; 3) 'occuring at short intervals over a space'. The adjective always refers to some concrete object: pyres, arrows, people, etc. The adverb appears only once elsewhere in the Iliad, at XV 470, and even there it is in reference to arrows: ὄφρ’ ἀνέχοιτο θαμὰ θρῴσκοντας ὀϊστούς. The uses here are in the same metrical slot, but Achilles’ use of the adverb to refer to speech, rather than some physical object, is entirely innovative in our text. The metaphorical extension of this adverb is consistent with Cunliffe’s third definition, 'occurring at short intervals', since he imagines an iterated speech event. By depicting the Myrmidons’ language as an iterated speech event, and one with markedly conventional language, it appears that Achilles is making an argument for their conventional-formulaic language, and by extension, their conventional thought (given that language relates to cultural frames). We may ask: why doesn’t Achilles just come out and say, “you speak so conventionally about me”? Is he straining for words that the Homeric world cannot provide? [47] Perhaps this is the case. But the lack of abstract language provides an added bonus. Because we are dealing with speech, the demonstration is far more powerful than an abstract claim. Not only does Achilles make the argument, but he backs it up with his own verbal prowess.

XVI 208–209

As the next few lines show, however, the agonistic poetics implicit within the passage are not the major purpose of the speech. Our attention is focused away from the speech back to the present with the phrase νῦν δὲ πέφανται, which follows a bucolic diaeresis. From this point on, the speech is more or less conventional in its formulae, coinciding with the conventionality of the speech act, a call to battle. Still, the formulaic oddities are not entirely eliminated. πέφανται occurs three other times after a bucolic diaeresis, all in speeches calling troops to battle. At II 122 Agamemnon calls them to battle by feigning a desire to return, and at V 531 and XV 563 Ajax rallies the troops: αἰδομένων δ’ ἀνδρῶν πλέονες σόοι ἠὲ πέφανται. Fain- + μέγα ἔργον in the next line is a relatively bound formula. The phrase appears at XI 734 (ἀλλά σφι προπάροιθε φάνη μέγα ἔργον Ἄρηος), and at XII 416 (τείχεος ἔντοσθεν, μέγα δέ σφισι φαίνετο ἔργον). Both other uses have a narrative context: in the former, it is Nestor narrating his adventure, in the latter, it is simply the narrator. Thus, there are formulaic and metrically regular lexical elements up until the first hemistich of XVI 208.
After the hemistich, linguistic irregularities arise again. As Achilles mentions his men’s speech, it appears as though the conventional language fails. As Janko’s commentary tells us, ἕης is an artificial form for ἧς, by analogy with ὅου (from *ὅο) and the possessive ἕος / ὅς. The only motivation for such an analogy would be metrical fit. Furthermore, ἐράασθε appears to be an improvised thematic form of the older athematic ἔραμαι. Here it seems that Achilles (or the poet, as the distinction blurs in this passage) analogizes from other words ending the verse, such as ἀγοράασθε (II 337), ἠγοράασθε (VIII 230), μητιάασθε (XXII 174), and εἰσοράασθε (XXIII 495), which are four of the five thematic uncontracted second-person plural mediopassive endings in the entire Iliad, outside of this speech. The diectasis represents the older form, when the contacted theme vowel becomes opaque. Such a form would be in mind given that the suffix shows up just a few lines above, in ᾐτιάασθε at XVI 202. Finally, the last line of Achilles’ speech, ἔνθά τις ἄλκιμον ἦτορ ἔχων Τρώεσσι μαχέσθω, is comprised of entirely formualic elements. As would be expected, a common event necessarily has certain attendant bound expressions. This is the call to battle, upon which everything else in Achilles’ speech is predicated.
And so we see the conventional character of Achilles’ mimetic speech, which includes part of a formula set for the χόλος frame, and on the whole Achilles conscientiously refers to that frame, as the source for the Myrmidon’s conventional language. And this is offset by his own nonconventional language within the speech. Yet typologically the speech is entirely conventional, as the last line, the culmination of the speech, indicates. Why then is the mimetic speech present? To characterize Achilles? It most definitely does show him to be on a verbal par with the poet(s) of the Iliad. And yet characterization would not explain why the mimesis appears where it does. To answer this question, we need to think in terms of the narrative. In Book 16, Patroclus essentially presents us with the Second Embassy. Achilles abandoned the Argives; Agamemnon, inspired by Nestor, sends an embassy to bring him back; Achilles wholly rejects them; Nestor inspires Patroclus to get Achilles to join; Achilles agrees to send Patroclus in his place: a partial acceptance. And as we have seen, an overwhelming majority of the formulae involved in this speech point us back toward Book 9, but are fundamentally linked to the traditional, cultural poetic χόλος frame. Because of the metonymic connection these formulae have to the larger frame, by way of speech-acts, the speeches themselves, when employing traditional formulae, help in performing traditional scripts. The speeches themselves become a means of acting in accordance with tradition. This, I think, is what Adam Parry was attempting to explain in his famous article “The Language of Achilles,” when he discusses “the common language.” He says, “The unity of experience is thus made manifest to us by a common language. Men say the same things about the same things, and so the world to them, from its most concrete to its most metaphysical parts, is one.” And later he adds, “For the language of society is the way society makes things seem” (A. Parry 1989:4–5). Martin correctly demonstrates that there is not necessarily a 1:1 ratio between signifier and signified (Martin 1989:152). That connection is only made possible in speech through the cultural frame and the bound expressions or formulae, which are mechanisms for cultural frames. And so characters participate in the younger Parry’s notion of a “unified world,” what we may call the Iliad’s representation of Tradition, through their employment of frames. Achilles, on the other hand, imitates that very participation. For Achilles, the mimesis of his men’s conventional speech-behavior allows him to go through the motions, to take part in their traditionality, to accede to it, while still keeping himself at one degree of separation.
But why is he so recalcitrant in simply giving over to the traditional script, which has become so intimately connected with him through the speeches of various characters? Unlike any other heroic figure who might participate in the same script, Achilles’ return determines his death, as he explains to Odysseus at IX 412–413. This choice is particular to the character of Achilles and not part of the generic χόλος frame. Achilles’ separation, as it is presented through the eyes of various characters within this frame, has particular significance for him, because his separation keeps him alive. From an audience perspective, the figure of Achilles raises the stakes in the χόλος frame. But from the perspective of Achilles and other characters, the application of this generic frame downplays his unique status and situation. Characters such as Ajax, Odysseus, Phoenix, and Apollo view Achilles in a traditional and generic cultural poetic frame, and while the frame applies to Achilles, it is not who he is. [48] This other side, the Achilles who exists outside the generic cultural poetic frame, is only brought to attention by Hera in her response to Apollo’s attack on Achilles (XXIV 56–63). In Achilles’ defense, she mentions two “facts”: that Achilles, unlike Hector, is part immortal, and that Apollo himself was present at the wedding of Thetis and Peleus. Laura Slatkin has demonstrated the true significance of Thetis’ marriage to a mortal (Slatkin 1991). She brings to bear the story of Thetis in Pindar Isthmian 8.29–38, which tells how Zeus and Poseidon wanted to marry Thetis, but Themis reported that the son of Thetis would be stronger than his father. With this knowledge, Zeus exclaims:
… ‘ἀλλὰ τὰ μέν
παύσατε· βροτέων δὲ λεχέων τυχοῖσα
υἱὸν εἰσιδέτω θανόντ’ ἐν πολέμῳ,
χεῖρας Ἄρεΐ <τ’> ἐν-
αλίγκιον στεροπαῖσί τ’ ἀκμὰν ποδῶν.’
“But stop these things, let her have a mortal marriage bed and watch her son die in war, with Ares-like hands, and feet like lightning flashes.”
Pindar Isthmian 8.35a–37
Thus, it is in his connection to Thetis that Achilles has a heightened status, outside of the traditional χόλος frame and script. Nevertheless, he must accede to the culturally scripted series of events, as he states in XVI 60–62:
… οὐδ’ ἄρα πως ἦν
ἀσπερχὲς κεχολῶσθαι ἐνὶ φρεσίν· ἤτοι ἔφην γε
οὐ πρὶν μηνιθμὸν καταπαυσέμεν …
Indeed, it was not at all possible to have raged endlessly in my heart, although I said I would not stop my anger
In the course of two lines he refers to the two most poetically charged words of the Iliad, both of which exist as cultural scripts, in reference to his own activity and the impossibility of maintaining separation. Examination of the “embassy speeches,” which I consider to be more than simply those of Book 9, will make clear Achilles’ gradual accession to this traditional script.
Walsh, I think, is correct to see χόλος as the genre of the Iliad. Recalling the script that this word entails one last time, it is “the story of the withdrawn hero, who refuses to fight, seeks refuge with a companion, and returns only too late” (Walsh 2005:192). In this respect, the embassy is not a necessary feature of the story, and therefore is a means by which to expand the story. I consider an embassy speech to be any speech that employs the same formulae (some of which appear in Achilles’ mimesis, although there are more parallelisms in the Book 9 speeches), intended for the same effect, to have Achilles behave in accordance with the Achaean society’s expectations. [49] As I see it, there are in fact three sets of embassy speeches in the Iliad. The first set is in Book 9, the second set is Nestor’s speech to Patroclus and Patroclus’ subsequent speech to Achilles, and the third set is Apollo’s speech to the gods, which results in Thetis carrying Zeus’ orders to Achilles. [50] And we can see that Achilles’ responses to each speech and set of speeches is progressively shorter and more accommodating. In Book 16, the middle set of embassy speeches, we see him halfway between total refusal to accept and complete acquiescence, and the mimetic speech reflects this position through his quasi-participation in the traditional scripts of the Achaean society. By Book 24, he must give in entirely, though reluctantly, to the traditional cultural frames of the Iliad, and none other than his own mother, the reason for his unique existence, is the one to deliver this news. In Achilles’ response to Thetis, there is no sign of their intimate connection, only two impersonal lines:
τῇδ’ εἴη· ὃς ἄποινα φέροι καὶ νεκρὸν ἄγοιτο,
εἰ δὴ πρόφρονι θυμῷ Ὀλύμπιος αὐτὸς ἀνώγει.
“So be it, may he who brings the ransom, take away the corpse, if the Olympian himself, by his own urging commands.”
Iliad XXIV 139–140
It is only with Achilles’ acceptance of the traditionally scripted behavior that the Iliad is able to come to its conclusion. [51]
Formulaic studies have ceased to occupy Homeric scholars’ central interest, but I maintain that the formula is just as significant to Homeric poetry as Milman Parry had originally argued, only on different grounds. As we have seen, Kiparsky established that the formula is a linguistic phenomenon akin to bound expressions. Understanding the formula in this light requires a complete hermeneutic methodology, which we find in frame theory semantics, as it has been defined by Fillmore and refined by Kiefer. Frame theory is able to combine the work of a number of Homeric scholars into a single cohesive model, which can produce fruitful results for the reading and interpretation of Homeric language. Achilles’ mimetic quotation and the prevalence of χόλος within that quotation provide excellent evidence for frame theory as operative within Homeric poetry. What we see at work within the Iliad is a particular traditional cultural poetic frame becoming intimately connected with the figure of Achilles, so much so that he becomes a generic element in the frame. Thus a frame theory model of the formula allows us to see a “language about Achilles” developing in the Iliad.


Austin, J. L. 1962. How to Do Things With Words. Oxford.
Bakker, E. 1997. Poetry in Speech. Ithaca.
Chantraine, P. 1953. Grammaire homerique. Paris.
———. 1968. La formation des noms en grec ancien. Paris.
———. 1968–1980. Dictionnaire étymologique de la langue greque. Paris.
Collins, Derek. 1998. Immortal Armor: The Concept of Alke in Archaic Greek Poetry. Lanham, MD.
Cunliffe, R. A. 1963 [1924]. Lexicon of the Homeric Dialect. Norman, OK.
Devine, Andrew, and Stephens, Laurence. 2000. Discontinuous Syntax: Hyperbaton in Greek . New York.
Dickson, K. 1995. Nestor: Poetic Memory in Greek Epic . New York.
Dougherty, Carol, and Kurke, Leslie. 1998. “Introduction.” Cultural Poetics in Archaic Greece: Cult, Performance, Politics (eds. Carol Dougherty and Leslie Kurke) 1–14. Oxford.
Fillmore, C. J. 1976. “The Need for a Frame Semantics within Linguistics.” Statistical Methods in Linguistics (ed. Hans Karlgren) 5–29. Stockholm.
Foley, J. M. 1985. Oral Formulaic Theory and Research. New York.
———. 1991. Immanent Art. Bloomington, IN.
Gombrich, E. 1960. Art and Illusion . New York.
Hainsworth, J. B. 1968. The Flexibility of the Homeric Formula. Oxford.
Higbie, C. 1990. Measure and Music. Oxford.
Hymes, D. 1962. “The Ethnography of Speaking.” Anthropology and Human Behavior (eds. Thomas Gladwin and William C. Sturtevant) 13–53. Washington.
Iser, W. 1974. The Implied Reader . Baltimore.
Janko, R. 1992. The Iliad: A Commentary. Vol. 4. Cambridge.
Kahane, A. 1994. The Interpretation of Order . Oxford.
Kiefer, F. 1996. “Bound Utterances.” Language Sciences 18.1–2:575–597.
Kiparsky, P. 1974. “Oral Poetry: Some Linguistic and Typological Considerations.” Oral Literature and the Formula (eds. Benjamin A. Stolz and Richard S. Shannon III) 73–106. Ann Arbor.
Kirk, G. S. 1985. The Iliad: A Commentary. Vol. 1. Cambridge.
Lord, A. B. 2000. The Singer of Tales. Cambridge, MA.
Martin, R. P. 1989. The Language of Heroes. Ithaca.
Miller, D. G. 1982. Homer and the Ionian Epic Tradition. Innsbruck.
———. 1985. “Towards a New Model of Formulaic Composition.” Comparative Research on Oral Traditions (ed. J. M. Foley) 351–393. Columbus, OH.
Minchin, E. 2001. Homer and the Resources of Memory . New York.
Morris, Ian, and Powell, Barry, eds. 1997. A New Companion to Homer . Leiden.
Muellner, L. 1996. The Anger of Achilles . Ithaca.
Nagler, M. 1974. Spontaneity and Tradition . Berkeley.
Nagy, G. 1974a. Comparative Studies in Greek and Indic Meter. Cambridge, MA.
———. 1974b. “Formula and Meter.” Oral Literature and the Formula (eds. Benjamin A. Stolz and Richard S. Shannon III) 239–260. Ann Arbor.
———. 1979. The Best of the Achaeans . Baltimore.
Parry, A. 1989. The Language of Achilles and other Papers. Ed. H. Lloyd-Jones. Oxford.
Parry, M. 1971. The Making of Homeric Verse: The Collected Papers of Milman Parry. Ed. A. Parry. Oxford.
Shank, Roger, and Abelson, Robert. 1977. Scripts, Plans, Goals and Understanding: An Inquiry into Human Knowledge Structures. Hillsdale, NJ.
Shive, D. 1987. Naming Achilles. Oxford.
Slatkin, L. 1991. The Power of Thetis. Berkeley.
Todorov, Tzvetan. 1981. Introduction to Poetics. Trans. Richard Howard. Minneapolis.
Visser, E. 1988. “Formulae or Single Words? Towards a new Theory of Homeric Verse Making.” Wurzburger Jahrbucher fur die Alterumswissenschaft 14:21–37.
Walsh, T. R. 2005. Feuding Words and Fighting Words: Anger and the Homeric Poems. Lanham, MD.
Watkins, C. 1974. “Response to Kiparsky.” Oral Literature and the Formula (eds. Benjamin A. Stolz and Richard S. Shannon III) 107–111. Ann Arbor.
———. 1977. “Apropos Menis.” Bulletin de la Societe Linguistique de Paris 72:187–209.
———. 1995. How to Kill a Dragon: Aspects of Indo-European Poetics. Oxford.
Wilson, Donna F. 1999. “Symbolic Violence in Iliad Book 9.” Classical World 93.2:131–147.
———. 2002. Ransom, Revenge, and Heroic Identity in the Iliad. Cambridge.


[ back ] 1. I would like to thank Greg Nagy, Paul Kiparsky, Thomas Walsh, and Richard Martin for all their help and input on this project. Parry had originally defined the formula as “a group of words regularly employed under the same metrical conditions to express a given essential idea” (M. Parry 1971:13). The popularity of the formula in Homeric studies may be attributed to the central role the formula played in Milman Parry’s oral formulaic theory (for the history of influence and development of which see Foley 1985). Parry’s definition has undergone many changes, but no consensus has ever been reached by the scholarly world.
[ back ] 2. We see this principle in effect with dictionaries of slang. For example in an internet Rap dictionary (www.Rapdict.org) you can find entries such as around the way (adj.), defined as “From the neighbourhood. From around the way—Beastie Boys (No sleep till Brooklyn [1986]).” And being [down] with something (v.): “Favouring something, thinking the same way. Howie Tee, are you down with me?—The Real Roxanne and Howie Tee ((Bang zoom) let’s gogo).”
[ back ] 3. This example of arbitrarily limited distribution weakens Visser’s nuclear-semantics approach to the formula (1988). While it most definitely seems the case that there are functionally semantic units in Homeric phrases, and prosodic elements as well, it does not seem right to argue that those semantic units must be individual words. In the case of θυμῷ ἀγήνορι, Visser would have to argue that θυμός is the semantically functional unit and ἀγήνορι the prosodic, as he does with the Noun-Epithet formulae. Or he could argue the other way as well. However, there is a case (at Iliad XXIV 42) where this formula is used in a violation of Hermann’s Bridge. The usage is a clear example of a two-word phrase that is a semantically functional unit, so much so that it is able to violate one of Homer’s own prosodic tendencies.
[ back ] 4. Kiparsky 1974:88. As we have seen in the case of the formula θυμὸς ἀγήνωρ, the difference in prosodic shape corresponds to a difference in syntax, and, because of its relation to semantics, syntax outweighs prosody for formulaic variation. Nagy 1974a comes to a similar conclusion about the secondary significance of meter in Homeric poetry, but from a diachronic perspective.
[ back ] 5. Ibid. Bakker 1997 comes to a similar conclusion, though his is based more on stylistic features of ordinary language than on the formal linguistic features that Kiparsky addresses. Fillmore’s suggestions would indicate that there may not even be a quantitative difference between oral literature and natural language with respect to the formula.
[ back ] 6. Kiparsky’s account of syntactic motivation explains only why variations of particular formulae are used, not why formulae are used instead of propositional phrases.
[ back ] 7. According to semantic markers theory, originally outlined by Katz and Fodor 1964, lexical meaning is comprised of a series of atomic features, such as +Male, +Adult, and -Animate.
[ back ] 8. Fillmore 1976:27. While Kiefer does not entirely subscribe to Fillmore’s argument, he does believe frames can be productively applied to the lexical treatment of bound utterances.
[ back ] 9. This in many ways responds to Watkins’s objection to Kiparsky’s argument at the Ann Arbor conference that most formulae in Homer are not idioms such as “addled eggs” (Watkins 1974:108–109). Idioms are one subfield of bound expressions, and they share in the three properties of those types of expressions.
[ back ] 10. Both Russo and Edwards in Morris and Powell 1997 hint at what was an impending moratorium in the study of Homeric poetics at the level of formulae. I hope to bring the formula back into view and demonstrate in what ways it is relevant to the more recent studies that have moved past the “formulaic question.”
[ back ] 11. Watkins’s suggestion concerning the oral-poetic theme as constituting culture and its relation to frame theory, I believe, provides a mechanics of “cultural poetics” as it is defined by Leslie Kurke and Carol Dougherty. As they explain, “In a sense, this concept of art as necessarily and profoundly interactive with its social ‘frame’ inverts (and thereby compliments) Geertz’s theoretical model. Geertz teaches us to read ritual as art; the New Historicist would have us see art as ritual,” first noting that “archaic texts are often scripts for ritual performance” (Dougherty and Kurke 1998:6). It is not a coincidence that in defining cultural poetics, Kurke and Dougherty employ both terms, “script” and “frame,” metaphorically. The metaphors become substantive within frame theory proper. The social “frame,” I would suggest, is fundamentally related to the more technical cognitive frame, such that the praxis of art and ritual stem from the same cognitive source. Throughout this essay I will employ the term “cultural poetic frame” as an extension of Kurke and Dougherty’s original concept in light of the research in cognitive science and its relations to both art and ritual.
[ back ] 12. For receptionalist theory proper, see Iser 1974.
[ back ] 13. Walsh also argues that χόλος is the genre of the Iliad, as νόστος is the genre of the Odyssey. I will be working with the theme of χόλος and Walsh’s suggestion about the Iliad’s genre later in this essay. It is important to note that the events of a χόλος theme are implied, and are not predetermined. As Walsh himself argues, χόλος, unlike κότος, has no immediate telos, though it is thought to be short-lived. χόλος has a potentiality, which makes it worthy of an Iliad-sized epic (Walsh 2005:197–201).
[ back ] 14. As in the flea-market transaction, the χόλος script is a generalized sequence of events and not the instantiation of any particular sequence. Nagler outlined a similar concept of an arranged, but abstract, order of events, which he called a motif: “The pattern … identified by these key notions, or elements, is not meant to represent a rigid type-scene, in the original sense of the term; it is a kind of ideal configuration intermediate in abstraction between a much more general impulse (a branch of the ‘family,’ or subset of the Gestalt) and any particular exemplar” (Nagler 1974:68). He distinguishes his understanding of motif as different from the type-scenes that Fenik 1968 and Arend 1933 describe. Nagler himself also understood a problem with the Gestalt or sphota as being too “mystical.” But he prefers to be over-general rather than to “maintain that an Iliad was created with a finite number of well-defined formula systems undergoing a limited number of predictable transformations” (14). The notion of frame avoids the potentially infelicitous rigidity of such a formulaic system, which Nagler too was attempting to avoid. But the frame also avoids the imprecision of the “mysticism” involved with ideas of a Gestalt. To return to Nagy’s definition of theme, theme is a basic unit of content, and I would add that, as frame, it is the most basic culturally meaningful unit. Anything more generalized than the cultural frame would place oral poetry outside of poetry’s functional sphere of influence. The notion of frame reflects the basic understanding that oral poetry is culturally anchored.
[ back ] 15. When the person bagging your groceries at the supermarket asks “Paper or plastic?” he is doing more than helping you to situate yourself within a grocery store, he is attempting to carry out the action of bagging your groceries.
[ back ] 16. Martin 1989, chap. 1. Speech-acts are considered to be those statements that simultaneously communicate something and perform an action, beyond the act of communication. For speech-act theory proper see Austin 1962.
[ back ] 17. We see traces of the beginnings of a theory of frames and scripts in J. L. Austin’s work on speech-acts. When discussing the possible infelicities of speech-acts, he states, “There must exist an accepted conventional procedure having a certain conventional effect, that procedure to include the uttering of certain words by certain persons in certain circumstances” (Austin 1962:14, emphasis added). The flea-market transaction presented by Kiefer is basically an example of the very phenomenon that Austin describes. That the procedure is accepted, according to Austin, indicates its cognitive aspects, and that it is conventional indicates its cultural aspects.
[ back ] 18. I realize that this last suggestion may appear to be overreaching. And yet it follows when one considers the nature of semantic meaning applied to works of fiction. If I am watching a movie about a fast food restaurant, and a character working at the restaurant asks “Do you want fries with that?”, I will understand the phrase in the same manner as the character placing his order. But that bound expression will have an effect on the character, whereas it will not affect me. And perhaps fries will have a larger significance in the movie, which I will be able to recognize, but the character won’t. Nevertheless, basic semantic meaning and the frame to which “Do you want fries with that?” refers will be the same for both the character and the viewer. In narratological terms, we may say that formulae have a strong focalizing effect. What formulae focalize are the virtual experiences of a given culture (see Watkins 1977:9 and Foley 1985:45).
[ back ] 19. Miller 1985:360. Miller understands “formula” to cover any number of “conventions of usage,” which include freezes, collocations, etc. One might wonder how productive the term formula is under these circumstances. Clearly there would not be any singularly understood function to a formula in this respect. My project is an attempt to understand the formula in relation to speech. And the formula as bound expression seems most useful in that area.
[ back ] 20. Is a phrase such as “on land” for the sake of ease in speaking about the location of a person who had been, but no longer is “at sea”? It seems first and foremost to be the culturally established way to talk about such a situation, and only because of that fact is it more readily available for speakers.
[ back ] 21. In other words, Achilles’ awareness of the Homeric language indicates not merely that frame theory semantics is applicable to Homeric language (any passage would do for such a demonstration), but that the Homeric poet is aware that Homeric language works as such. Achilles’ awareness points to the poet’s own awareness, as would be logically consistent with the nature of frame theory semantics.
[ back ] 22. Odysseus’ quotation also has occurrences of elision absent from Achilles’ quotation (IX 255): δώσουσ᾽ αἴ κ’ ἐθέλωσι, σὺ δὲ μεγαλήτορα θυμὸν. Including what is an actual linguistic phenomenon of Greek speech makes Odysseus’ quote seem all the more plausible. The stylistic elements of Achilles’ speech will be discussed shortly.
[ back ] 23. Agamemnon gives a similar boast quotation at IV 178–182.
[ back ] 24. Tzvetan Todorov speaks of just such a phenomenon; he discusses three types of narratives in relation to the phenomenon of frequency: singulative narrative, repetitive narrative, and iterative discourse. Iterative discourse, he says, is a single discourse which "evokes a plurality of events” (Todorov 1981:31). Such is the discourse involved in scripts. Todorov comments on the role of iterative discourse in classical literature, but nevertheless undermines its significance.
[ back ] 25. The notation is an adaptation of Richard Martin’s notation for his analysis of Achilles’ speech in Book 9 (Martin 1989:166–171). Martin had identified two types of formulae, syntagmatic and paradigmatic; syntagmatic formulae consist of metrically fixed and flexible formulae, and paradigmatic formulae are those individual words that frequent a given metrical slot. According to frame theory, there is a need to distinguish between fixed and flexible formulae because flexible formulae may apply to more than one frame, whereas that is not the case for fixed expressions. Thus flexible formulae have a single underline, fixed formulae a double underline [bold], and metrically frequent individual words a dotted underline [italic].
[ back ] 26. As far as I can tell, this reconstruction is slightly odd, because according to Grassman’s law, deaspiration usually occurs with the first of two aspirated consonants. While there are exceptions, they are based on particular consonant clusters, and C + l is not one of them. Still, deaspiration of the second element may be a result of the aspiration in the first element being more semantically and pragmatically significant, as in the case of the aorist passive imperative ending, *-θηθι > θητι, where the aorist passive element is the more salient feature. This may also hold true for σχέτλιος, where retention of the initial aspirate makes it more semantically transparent than would be *σκέθλιος.
[ back ] 27. Agamemnon’s speech presents us with a figura etymologica, between σχέτλιε and ὑπέσχετο.
[ back ] 28. Devine and Stephens, in their discussion of hyperbaton, cite Russian, which can classify sentences as emotive or non-emotive, based on word-order and prosody. “It is natural to associate the high frequency of hyperbaton in the orators with a more emotive style that makes greater use of pragmatically marked word order, and the lower frequency in the historians with a preference for a more detached and analytical style” (Devine and Stephens 2000:59). The position of σχέτλιος does not change within the verse line, however. We see in this third group that it is an interjection and is not syntactically connected to either the preceding or following phrases. Thus, we may say that in these instances of σχέτλιος there is only a pragmatic, and hence emotive, function.
[ back ] 29. It should be noted that all occurrences of this vocative are after Book 16, again making it impossible to suggest that Achilles is simply repeating speech.
[ back ] 30. For Thetis is forced to marry a mortal and bear a mortal child. For the potential anger and power of Thetis, see Slatkin 1991. Most interestingly, Thetis’ χόλος, alluded to only here, is the result of the νεῖκος between Zeus and Poseidon, site 2 of χόλος! The progression of themes from νεῖκος to death of a φίλος in the Thetis story exactly mirrors the progression of νεῖκος between Achilles and Agamemnon to the death of a φίλος with Patroclus. The confrontation between Zeus and Poseidon in Book 15 also has the same formulaic language as Achilles’ language in Book 16 concerning his dispute with Agamemnon.
[ back ] 31. Achilles’ matrilineal connection and its tragic implications are emphasized most in Hera’s response to Apollo in Book 24.
[ back ] 32. Walsh (2002, chap. 10) discusses the three embassy speeches of Book 9 in light of the χόλος. Patroclus’ speech in Book 16 is also very much involved with the χόλος theme, as he says at line 30. Many of the formulaic and lexical and formulaic elements of Achilles’ quote are common to those speeches. σχέτλιος, for example, is also in Ajax’s speech in Book 9.
[ back ] 33. The use of the lexical item in Achilles’ quote demonstrates how Homeric language can be creative through its employment of traditional frames.
[ back ] 34. This is a common feature of Homeric formulae, which we can also see in a formula such as κραδίη καὶ θυμὸς ἀγήνωρ, composed of κραδίη καὶ θυμός, which is used in emotive situations as patients, and θυμὸς ἀγήνωρ, which is used as an agent. Both meanings are combined in Ajax’s use in his speech to Achilles at IX 635, for he is discussing what action (agentive aspect) one would take who is bereaved of someone near and dear (patient aspect).
[ back ] 35. A good example of single words having thematic and compositional significance is the use of ἄνδρα in the Odyssey, on which see Kahane 1994. It is also a good example of Richard Martin’s paradigmatic formulae (1989).
[ back ] 36. It is interesting that Nestor, the best speaker, and Thersites, (II 216), would use the same formula. This suggests that it is not the content of Thersites’ language that is the problem. As Martin remarks, “Thersites’ style deserves no respect because he does not have the heroic martial performance record needed to back up his words” (Martin 1989:111).
[ back ] 37. Furthermore, at IV 32 Zeus speaks to Hera with the formula ἀσπερχὲς μενεαίνεις, which Apollo uses toward Achilles at XXII 10. The situation in Books 4 and 22 are similar, since both deal with the sacking of Troy, and as a result there are similar frames involved and hence similar formulae are showing up.
[ back ] 38. Collins 1998 traces this relationship between action and forgetting in light of ἀλκή.
[ back ] 39. For Nestor and poetic memory, see Dickson 1995.
[ back ] 40. To my knowledge, a full analysis of the verb λανθάνω, and its subjects and objects, has not been done, though the results I think could tell us very much about the nature of Homeric poetry and its thematic issues.
[ back ] 41. This possible division between event/action and speech shows up later (XVI 626–631).
[ back ] 42. The exact etymology for μῆνις is not clear. Schwyzer 1931 had originally proposed *mnanis, a derivative from the root *mna-. Chantraine considers the etymology impossible to ascertain.
[ back ] 43. Watkins 1977 cites this example of μῆνις as the only example of the word spoken by a mortal. The uses of its cognate in Book 16 he considers as intentional avoidance of the actual word.
[ back ] 44. Following a long debate about to whom μῆνις may refer, Muellner says that “only gods and heroes have μῆνις” (Muellner 1996:127n69).
[ back ] 45. Muellner 1996:8. Many of Muellner’s observations on μῆνις work within the same sphere as χόλος, and there might not be a clear-cut distinction between the two even to the Homeric poet. In Achilles’ response to Patroclus (XVI 61), he uses variant forms of both χόλος and μῆνις.
[ back ] 46. Achilles’ use of μηνιθμόν demonstrates the confluence of primary and secondary focalizers within the Iliad, a phenomenon that Richard Martin recognized in his analysis of Achilles’ speech in Book 9: “The similarity [between Achilles and poet] arises because Homer, when he constructs Achilles by means of language, employs all his poetic resources and stretches the limits of his formulaic art to make the hero as large a figure as possible. In short, the monumental poem demands a monumental hero; the language of epic, pressed to provide speech for such a man, becomes the ‘language of Achilles’” (Martin 1989:223). The crossover in language from poet to character, which is all the more plausible when formulae are considered cultural mechanisms of communication, will necessarily conflate the two worlds, actual and represented, in performance.
[ back ] 47. This is Adam Parry’s argument in his famous essay “The Language of Achilles” (1956), reprinted in A. Parry 1989:1–7.
[ back ] 48. Although the poetic mimesis says as much when it claims, “your mother reared you on χόλος.”
[ back ] 49. Donna Wilson employs Bourdieu’s term “symbolic violence” in discussing the embassy of Book 9, and she explains,“The embassy speeches recast Agamemnon’s ἄποινα typologically, bringing them into exchanges conventional among φίλοι” (Wilson 1999:143).
[ back ] 50. There are a great number of formulaic parallels between the Ajax and Apollo speeches, one of which is the term σχέτλιος, also seen in Achilles’ mimetic speech.
[ back ] 51. Donna Wilson suggests that the mention of ἄποινα here evokes his initial rejection of that offered initially by Agamemnon (Wilson 2002:128). Achilles’ indifference thus marks a defeat of his larger purpose. As Wilson says, “What Achilleus had wanted, that is, to identify with his divine self and to attain universal dominance, has been unequivocally denied him” (130).