Homer’s Versicolored Fabric: The Evocative Power of Ancient Greek Epic Word-Making

Citation: Bonifazi, Anna. 2012. Homer's Versicolored Fabric: The Evocative Power of Ancient Greek Epic Word-making. Hellenic Studies Series 50. Washington, DC: Center for Hellenic Studies. http://nrs.harvard.edu/urn-3:hul.ebook:CHS_Bonifazi.Homers_Versicolored_Fabric.2012.

Chapter 4. Visual and Narrative Functions of αὐ-Discourse Markers

An epigram assigned to Pollianus (first or second century AD) in the Greek Anthology begins as follows:

Τοὺς κυκλίους τούτους τοὺς “αὐτὰρ ἔπειτα” λέγοντας
μισῶ, λωποδύτας ἀλλοτρίων ἐπέων.

Greek Anthology 11.130.1–2

These cyclic poets who say “αὐτὰρ ἔπειτα,”
I hate them; they are thieves of the poetic utterances of others.

These lines allude to Hellenistic “unimaginative imitators of Homer.” [1] Apart from the explicitly polemical intentions of the epigram, I underscore that the metonym chosen by the author in order to identify the style of some contemporary epic composers is “to say αὐτὰρ ἔπειτα.” The effectiveness of such a metonym was presumably high for the “there and then” receivers of that epigram, as it is still now for ancient Greek scholars. αὐτὰρ ἔπειτα is a very familiar phrase that is—and is felt as—typical of epic discourse, not only because it occurs frequently, but also because after Homer it basically disappears. [2]

The point is that this phrase is usually translated as “then” or “after that” and it is supposed to convey a temporal gap, for which the sole ἔπειτα could, in principle, suffice. If so, what is αὐτάρ for? What does αὐτάρ mean? This chapter is an investigation of the Homeric uses of αὖ, αὖτε, and αὐτάρ. Further αὐ-adverbs will be considered in the next chapter (αὖτις, αὐτίκα, αὔτως and αὐτοῦ (+ αὖθι/αὐτόθι). The αὐ- component is, of course, the basic reason for choosing these words. Surprisingly, the scholarly literature has not devoted much attention to {185|186} these adverbs/particles (Denniston even omits αὖ in his list of particles), [3] and, generally speaking, a unifying account of αὖ, αὖτις, αὐτός, αὐτίκα, and αὔτως encounters resistance. The meanings of these words are quite various; not only do they differ from word to word, but they also differ considerably for any single word. [4] This variety of meaning has so far presented some fundamental barriers to a unifying account. One barrier has been the distinction between grammatical categories; for instance, adverbs are distinct from particles. The former are mostly polysyllabic, semantically “stable,” and non-optional words; the latter are mostly monosyllabic, semantically elusive, and optional words. This distinction creates, as we shall see, several problems; while it might have a didactic convenience, it nonetheless complicates the understanding of the words. For example, adverbs that do not seem to behave as words that modify a constituent of a sentence are difficult to classify—some of them are called “sentence adverbials”—and yet it as presumed that they cannot be treated as particles. To pick another instance of grammatical barriers, why should an adverb share something with a pronoun? Yet, I will claim that αὐ- works as a crucial lexical element that unifies the words under discussion beyond their grammatical category. The upcoming pragmatic reading of αὖ, αὖτε, and αὐτάρ will, in fact, show that a number of functions lend support to a link with αὐτός. This link could, in turn, support the etymological thesis according to which αὐτός comes from αὐ-; my primary aim, however, is to study the effectiveness of using all these words and only secondarily to offer a contribution to IE studies concerning the origin of Greek αὐτός. What is common to all of these αὐ-terms is not what they say but what they do; in other words, they help producers and receivers of the linguistic communication to process the discourse unit in which they occur. Of course, they do so according to the specific grammatical contribution of each. For instance, while αὖτις may express the repetition of a certain action, the pronoun αὐτός may express repetition in recalling a certain object; while αὖ may convey the recognition of an already known event, αὐτός may convey the recognition of an already known hero. This is also what links the first three chapters of this monograph to the last two. Odysseus is pronominalized as αὐτός in local contexts, that is, as a prominent hero who is singled out in several different situations; and in the global context of the poem, he is the prominent leader isolated from the companions and whose real identity will be thematic throughout the entire poem. This is possible because of the pragmatic and cognitive work of αὐτός, which is also what characterizes αὐτός with respect to other third-person pronouns. The pragmatic and {186|187} cognitive work of αὐτός can be much better understood in light of the comparable work that underlies the Homeric uses of αὖ, αὖτε, and αὐτάρ.

The guiding purpose of this chapter is to enrich our account of the meanings of words, and this goal rests on the radically pragmatic view of propositions presented in the Introduction. Instead of assuming that the zero level of communicative units coincides with propositions, I will assume that the zero level of communicative units coincides with the utterances that situate each proposition and give them a certain meaning in a certain context. From this perspective, the use of words is assumed to embrace more possibilities: some words may deal with the content of the proposition, while others may deal with what the proposition is going to do. [5] This view permits the inclusion of different options, all of which make some sense, within the meanings of the same lexical item. This perspective differs from a more classical dichotomy between words that have an independent meaning and words that have no meaning per se.

A further challenge concerns the pragmatic and literary reading of the words at issue in different poetic genres—that is, epic, lyric, iambic and elegiac poetry. The heuristic reasons why I decided not to confine the analysis of the {187|188} uses of αὖ, αὖτε, and αὐτάρ to Homer rest on some pragmatic properties shared by epic and melic genres. One is the “center of gravitation” represented by the “I, here and now” of the oral performance. In linguistics, this is called zero-point of utterance. [7] While this property is widely recognized for melic texts, which abound in this deictic anchorage, [8] studies in epic poetry tend to pay less attention to this aspect, partly because the linguistic traces of this anchorage are quite infrequent. The prevailing modality for the narration of epic events seems to be impersonal, and the superordinate need to reach non-specific audiences in non-specific occasions harmonizes with this. Still, many scholars underscore the presence and the influence of the performing “I” at different levels. [9] Beyond the distinctive characteristics of epic and melic performances as far as metrical patterns, musical accompaniment, melodic structures, length, number of the performers, occasions, and public are concerned, both genres reaffirm traditional values and acknowledge a collective memory through an orally centered event. [10] Another shared pragmatic property can be summarized by the principle that Collins (1991) calls “kairotic play.” The idea of “play” points up the conventions (or rules) that are governing the understanding of communication; [11] kairos is intended to be the metaphor of a specific segment of time, recognizable and repeatable. [12] “Kairotic play” is “the willing suspension of disbelief in the difference between the action of portrayal (mimesis) and the portrayed action (mimema).” [13] With respect to epic and melic performances, {188|189} both performers and listeners have to suspend their disbelief in the difference between uttering words that describe actions and the actions themselves. This is true for those who participate in the narration of Hector’s death, as well as for those who participate in the praise of Theron of Acragas. This property is pragmatic because it is an extralinguistic (or contextual) pre-condition for the production and the reception of compositions. The “kairotic play” principle sums up the psychological assumptions, the cognitive activities, and the poetic effects underlying both epic and melic performances.

Discourse markers

Research into, and theory about, discourse markers started about thirty years ago; in recent years, such studies have constituted a remarkably productive field in pragmatics and in discourse analysis. [14] Several discourse markers have been analyzed within corpora from different live languages; [15] however, the present section will focus primarily on English examples. Here, I have selected three ways of summarizing what discourse markers are:

One of the most prominent functions of discourse markers is to signal the kind of relations a speaker perceives between different parts of the discourse.

Lenk 1997, HoP online

Discourse markers are items external to propositional content which are useful in locating the utterance in an interpersonal and interactive dimension, in connecting and structuring phrasal, inter-phrasal and extra-phrasal elements in discourse, and in marking some ongoing cognitive processes and attitudes.

Bazzanella 2006:456 {189|190}

… discourse markers constitute an extremely efficient, economical and felicitous means of facilitating the smooth production and interpretation of discourse.

Cohen 2007:xxii

Some crucial concepts are conveyed by these formulations: the meta-textual level of what the speaker/writer intends by speaking/writing; the subjective and the interpersonal components of relating parts of discourse; the marking of cognitive processes; the fact that discourse markers help not only to interpret but also to produce discourse.

Despite the notable variety of approaches and of terminologies that is acknowledged to pertain to studies of discourse markers, [16] several scholars agree on some general characteristics that allow a lexical item to be identified as a discourse marker. Here, I present those that I find most relevant to the topics of this chapter:

They embrace different grammatical categories: conjunctions (“but,” “like”), adverbs (“then,” “anyway,” “furthermore,” “well,” “finally”), verbal forms (“let’s say,” “you know,” “I mean”), various lexical phrases (“for example,” “before I forget,” “on the other hand,” “what else?”), and interjections (“oh,” “oops”). [17]

They are related to the utterance situation. [18] For example, an utterance such as “y’know, the students don’t get it” would be unacceptable if “know” would be used in the past tense (“??y’knew, the students don’t get it”). [19] {190|191}

They are syntactically detachable (or loosely attached) and prosodically independent. [20] In “well, I would not say that,” the discourse marker “well” occurs outside the syntactic structure “I would not say that.” Moreover, the same “well,” if uttered aloud, has its own intonational contour (in written language, this is normally rendered by a comma following the discourse marker). Thus, the prosodic independence is associated with intonational markedness.

Very often they occur in sentence initial position. [21] This reflects their role in connecting adjacent parts of discourse.

They typically do not add anything to the propositional content of the sentence. [22] In “moreover, the men were armed,” the conceptual content expressed by the proposition “the men were armed” is not affected or “touched” by the preceding adverb “moreover.” [23]

They are multifunctional. [24] For example, “okay” as a discourse marker can have two distinct pragmatic functions: one is “approval of an idea,” as in “okay, but how can you demonstrate that?”; another is “agreement with a proposed activity,” as in “okay, see you then.” [25] A more general aspect of multifunctionality concerns their global vs. local scope. [26] For instance, “to begin with” may be used to mark an entire section of an argument as a whole (global scope), whereas “but” may mark a frustration of expectation with respect to the immediately preceding words (local scope), as in “John is a Republican, but he is honest.” [27] Discourse markers can also serve multiple functions simultaneously: “oh” in dialogs may mark the speaker’s introduction of a new referent and {191|192} let the interlocutor know that (s)he lacks some information at the same time. [28]

Since discourse markers have one or more pragmatic functions, their lexical meaning is often difficult to retain in a translation or paraphrase. For example, while a discourse marker like “what else?” is easily to be paraphrased as “Is there something else that I could/should add? What?,” the initial “well” in the example mentioned above (“well, I would not say that”) cannot be paraphrased by maintaining any of the lexical meanings of “well,” such as “in a good way,” or “admirably,” or “healthy,” or “substantially.” [29]

They tend to occur more in spoken than in written language, [30] because spoken language allows more fragmentation and flexibility in the organization of discourse. Also, spoken language tends to include more explicit references to the interactional—or interpersonal—level of discourse. However, discourse markers are not an oral phenomenon per se.

Especially in spoken language they may occur in clusters or chains; [31] that is, more than one discourse marker may be uttered in a row, to reinforce an overall communicative intention (as in “Well, I mean, it depends”).

Discourse marker research rests on a basic notion: language contains more than one component; or, more precisely, language works at more than one level. Halliday and Hasan (1976) identified three levels, namely the ideational, the textual, and the interpersonal. The ideational refers to the ideas or concepts expressed (“language … being about something”); [32] the textual refers to the linguistic devices adopted to obtain a “text,” which are mainly concerned with cohesion; the interpersonal refers to different interactional aspects regarding speakers and interlocutors. Several scholars thereafter have re-defined {192|193} language levels according to its metalinguistic functions. [33] For the purposes of this chapter, I single out Kroon’s tripartition, which is the basic theoretical framework used for the analysis of several discourse markers in Latin; in order not to contribute to a confusing plethora of labels, I will make use of the same terms in my own analysis. Kroon isolates a representational, a presentational, and an interactional level of language. Language permits representing “some real or imaginary world outside the language itself.” The relationships between the constituent concern, in this case, states of affairs (actions, events, facts) or concepts and entities from a clause-internal point of view. [34] Language is also used to present—in a communicatively effective way—the ideas and the concepts that make up the represented world. The relationships between the constituent units concern the organization of discourse. [35] Finally, language allows speakers to interact with their interlocutors by marking the conversational structure and by situating the communicative events in terms of the attitudes, evaluations, and emotions of the discourse participants. The relationships between the constituent units concern interactional aspects of different types of exchange. [36]

To Kroon’s articulation of these macro levels of language, which correspond to three macro functions of language—the representational, the presentational, and the interactional—I would like to add some general remarks. First, the presentational and interactional functions rely on utterances as discourse acts. Second, some discourse markers are lexically indistinguishable; still, they may work either at the representational or at the presentational and/or interactional level. Third, presentational and interactional meanings are poles on a continuum. Let us comment on each of these points.

What is common to the way of organizing discourse and to the way of interacting with discourse participants is that the speaker/writer uses some words to cue the listener/reader about how to process the prior and the following discourse act(s). I foreground here the definition of discourse acts that I {193|194} have borrowed from Hannay and Kroon (2005:121): “distinct steps which the language producer executes as a result of strategic planning in order to realize her communicative intention.” From a pragmatic standpoint, every sentence is an utterance presupposing certain communicative intentions and enacting certain strategies to convey such intentions—which may be more or less explicitly stated; every utterance can be interpreted as a discourse act. If we take this into account, the meaning and the function of several connectives, for example, become clearer. Let us observe the use of “because” in the following two cases:

(1) I am going to the library because I have a book to return.

(2) Are you going to the library? Because I would have a book to return.

In (1) “because” connects two facts, namely “going to the library” and “having a book to return.” The conjunction “because” works at the representational level: the receiver of such a message identifies the return of books as the cause of going to the library. Conversely, “because” in (2) does not connect two facts; the “you” involved, who may or may not go to the library, and the speaking “I” that has to return a book are not representationally connected facts. “Because” here connects two discourse acts, namely, the question “are you going to the library?” and the statement “I would have a book to return”: “because” makes explicit why the previous question has been asked. “Because” helps the “you” interlocutor to process the relation between the previous and the following utterances. Furthermore, the whole of the two discourse acts pragmatically constitutes a “move,” to use Roulet’s terms. Such a move consists in indirectly suggesting to the interlocutor that (s)he could actually do a favor for the speaking “I” and return the book on behalf of her/him. This suggestive force is something that classical semantic analyses are unable to account for. [37] If we train ourselves to see that behind any sentence there are one or more discourse acts, we can better understand the connections between shorter or longer parts of discourse. Why? Because we can see connections at the level of facts and connections at the level of acts; we can see multiple uses of connectives. By contrast, if we limit our {194|195} study to the connections at the level of facts, we cannot explain some further uses of connectives. A larger spectrum of the uses of “because” is what allows us to illustrate how communication works in both (1) and (2), whereas a solely propositional definition of “because” prevents us from illustrating the use of “because” in (2). In addition to presentational adverbs and conjunctions we might also consider sentence adverbials, which quintessentially relate to the interactional level of discourse. The underlying view of acts presupposing intentions and strategies also holds for them. Let us observe the use of “really” in the following examples:

(3) They are not really my aunt and uncle.

(4) They are not my aunt and uncle. B: Really?

If we assume that “really” exclusively means “in actual fact” (cf. (3)), we are unable to comprehend the pragmatic sense of “really” in B’s comment to A’s statement (cf. (4)). The interactional value of the little exchange in (4) rests on B’s surprise or incredulity at hearing A’s statement, rather than on B’s questioning whether that is really the case or not. In other words, B’s utterance of “Really?” is a discourse act by means of which B expresses his/her reaction and attitude toward what A has just said; the intention is to show surprise or incredulity; the strategy is to politely use an interrogative form to express that. [38]

The preceding discussion shows that discourse markers work at the presentational and at the interactional level of discourse. However, in certain languages some lexical items working as discourse markers are lexically indistinguishable from adverbs and conjunctions working at the propositional level. I have deliberately chosen “because” and “really” in the instances analyzed above because they are such lexical items, and I have shown the difference between two possible uses. As we will see, several ancient Greek adverbs seem to behave in the same way. What allows us to identify when a certain lexical item is working as a propositional adverb and, conversely, when it is working as a discourse marker? In light of the preceding characterizations of discourse markers, I offer some “tips” that may help to disambiguate the functions (in English, for now): when a lexical item occurs in initial sentence position; when it is propositionally irrelevant; when it is syntactically optional; when it is prosodically independent; when its ultimate meaning is semantically opaque and the paraphrase cannot include the (original) lexical meaning; in all of these cases, we may have a discourse marker. When the lexical meaning works perfectly {195|196} for both functions, sentence position may make the difference. For example, in “I have lost my wallet again,” the adverb “again” works at the propositional level, whereas sentence initial “again” in “Again, the political implications are crucial” indicates a presentational feature connecting the upcoming utterance with a previous argument that was meant to be organizationally and hierarchically similar.

Finally, the dividing line between presentational and interactional functions is important in conceptual terms, but the reality of the uses of discourse markers in different languages (ancient Greek included) shows that often the two levels overlap. The difference may be seen, rather, in terms of degree, or in terms of higher vs. lower components. For instance, the discourse marker “for example” can be placed toward the presentational pole of a continuum, whereas discourse the marker “you know” can be placed toward the interactional pole of the same continuum. An initial “well” signaling, for instance, an utterance that is “dispreferred in terms of contents, structure or face” [39] can be placed somewhere in the middle. [40]

The Homeric adverbs that this chapter is focused upon work as discourse markers mainly concerned with the presentational functions of discourse. [41] The primary interest is to elucidate their contributions to signaling the discourse structure. Their interactional functions appear to be effective when internal characters utter them. However, it is possible to record presentational uses of discourse markers by characters as well as interactional uses by the primary speaking ‘I’. In this latter case, the exchange with the audience is not marked by any references to the listening “you” but by expressions of the speaker’s {196|197} attitudes, which have a cognitive and/or an emotional value, toward what is said. [42]

Presentational discourse markers fit particularly well the metaphor of discourse markers as road signs: [43] they signal where discourse comes from and where it is going. Because of the overarching structure of the Homeric poems as narratives, presentational signposts primarily help the cognitive processing of narrative transitions, of narrative suspensions and resumptions, [44] and the tracking of meta-narrative discourse acts; [45] within the same narrative section or sequence, they may mark topic shifts, the zooming in on single scenes (or on single characters), or genre-relevant structural elements (for example, the different entries of a catalogue).

Apart from ancient Greek, whenever narratives are intended to be orally performed, discourse markers help the processing of what is told, as they have to serve as what different devices usually do in written code: we might think of chapter-divisions, of paragraphs, of indentation, of quotation marks, of footnotes, and of other paralinguistic written tools, such as changes of fonts. Discourse markers used in oral narratives can potentially mark all of this, and even more. I will now present some research results from the analyses of very different narrative texts as they are orally marked, and in different languages (American English, German, Neapolitan dialect, and Serbo-Croatian).

Norrick 2001 is a work devoted to “well” and “but” marking narrative structures and storytelling procedures in spoken narratives. The author shows that such usages differ from more familiar conversational usages, and correspond to specialized functions because of the special discourse type in which they occur. In the example below, “well” signals the beginning of a new episode:

So my dad crated-up these two
beautifully little matched puppies
and shipped them from Missouri to Florida,
to this man.
And reminded him, {197|198}
“be sure you chain the puppies up.”
Well about, seven eight months later,
here those two dogs return to Missouri. [46]

The teller “keys well on the expected organization of the narrative in progress” (2001:854; italics in the text). Other narrative usages of “well” include the introduction of a summary coda and the return to the main theme of a story. As for “but”, in addition to the mentioned functions of “well”, it may mark “the resolution” of the narrative, it may “bridge the teller’s digression,” or it may mark the transition from “the orientation section to the main complicating action,” as in the following case:

And they were
I think because of the disgrace of it all,
they were going to move to Spain
where they’d built themself
this grand house in Spain.
But apparently
you uh they spent uh when it was finished
they went out there to spend a winter there … [47]

Norrick’s conclusion is that “discourse markers in oral narratives are elements dependent upon expectations about story structures and conventions and they bracket appropriate units in accordance with the organizational conventions of this genre” (2001:866).

Following the lead of Norrick, I have explored examples of German aber, which resemble the usages of “but” in the written version of German fairy tales by the Brothers Grimm. Here, I report one of these occurrences, from “Snow White:”

Als die Zwerge hörten, was geschehen war, sprachen sie: “Die alte Krämerfrau war niemand als die gottlose Königin: hüte dich und lass keinen Menschen herein, wenn wir nicht bei dir sind.” Das böse Weib aber, als es nach Haus gekommen war, ging vor den Spiegel und fragte: “Spieglein, Spieglein an der Wand, wer ist die schönste im ganzen Land?”

W. and J. Grimm 1987, I:273–274 [“Sneewittchen”] {198|199}

When the Dwarfs heard what had happened, they said: “The old storekeeper was none but the wicked Queen: beware, and don’t let any people in if we aren’t with you.” But the evil woman, when she had come back home, went in front of the mirror and asked: “Little mirror, little mirror on the wall, who is the fairest in the whole land?”

“But” (aber) in this case has the function of signaling the switch of setting from the Dwarfs’ house to the Queen’s palace, which possibly implies simultaneity and certainly does not imply anything adversative. Interestingly, the editors of the German text indent at that point. [48]

The recording and the transcription of a fairy tale orally transmitted in Neapolitan dialect, titled “Sarceniello” [49] (this name literally means “little bunch of wood”), provides us with the occurrence of a multifunctional discourse marker that is often ignored in the Italian translation—namely, the particle mò. Let us consider its presentational (and interactional) function in the following narrative transition:

[While Sarceniello, a crazy and lazy boy, goes to gather wood, he sees three fairies sleeping under the sun and decides to intertwine some foliage to screen them from the sun.]

Mentre ca isso faceva legne, se svegliarono ‘e Ffate. – Ma chi l’ha fatta sta bella capanna? Embè, chi ce ha fatto sta capanna, le sia cuncesso d’ave’ tutto chello che cerca –. E accussì affataieno a stu giovane. Mò, chisto steva facenno ‘a sàrcena, e, faticanno comme primma vota, ‘a facette bella grossa, ca nun ce ‘a faceva a ghizà.

While he was gathering wood, the fairies awoke.—Wow, who has made this nice shelter? Well, the one who made this shelter for us, let us grant him to have all he seeks—. And so, they cast a spell for this boy. So, this one was fastening a bunch of wood, and, while he was working hard on this like the first time, he made up a very big one, so that he was not even able to raise it.

While other discourse markers such as ma (“wow”) and embè (“well”) are spoken by the witches, mò is spoken by the storyteller and is forward-oriented: it shifts {199|200} attention (back) to the boy in order to introduce the next narrative section, in which the positive effect of the spell is narrated (the written edition I consulted indents, at this point). I see in the use of this mò an interactional function as well—that is, the listener is warned about what ensues from the witches’ decision, which is going to be something exciting; in fact, its introduction represents the fulfillment of the listener’s narrative expectation.

Finally, Elmer 2009 is a work devoted to some Serbo-Croatian expressions occurring in the transcriptions of nine South-Slavic oral epic songs recorded in June 1935. These expressions appeal to the listener and mark specific moments of the story, such as the introduction of a new character, the transition to a different temporal or spatial setting, or the focusing on a certain detail: “[t]he shifts in focus signaled by appeals to the audience often coincide with boundaries between distinct units of action” (Elmer 2009:52). An expression that is relevant to the Homeric analysis that will follow is kad evo ti:

sjede Marko pod žutu naranču
sjede Marko i opočinuo
kad evo ti careva telala
traži telal Kraljevića Marka

from Marko Kraljević i Nina od Koštuna (“Marko Kraljević and Nina of Koštun”). Milman Parry Collection, PN 6693

Marko sat beneath the yellow-green orange tree,
Marko sat and rested.
When—here you are!—the imperial messenger,
the messenger seeks out Kraljević Marko. [50]

kad evo ti literally means “when here to you” and it matches the characteristics of discourse markers that I have discussed: syntactically independent, sentence initial, related to the utterance situation (cf. the proximal deictic evo “here” and the “you” mark ti), exterior to the propositional content. The narrative function is twofold: on a local level, it shifts the attention of the receiver from Marko to the imperial messenger and it spotlights the messenger’s action; on a global level, it recalls another messenger who is ultimately responsible of Marko’s adventures at the beginning of the story. The two messengers become signposts of the articulation of the story and of the “thematic poles of foreign and domestic” (Elmer 2009:55). In addition to these presentational functions, kad evo {200|201} ti has an interactional value as well: Elmer underscores that the appeal to the listeners is an underlying feature characterizing all the expressions analyzed. [51]

Ancient Greek particles working as discourse markers

Some scholarship in discourse studies and in pragmatics uses “particles” instead of “markers,” since in many current languages the metalinguistic functions fulfilled by the broadly defined class of discourse markers actually pertain to particles—that is, single and often monosyllabic words that are grammatically mixed. [52] I do not adopt the label “discourse particles” (unlike Kroon, for example, in her study of Latin nam, enim, etc.) for two reasons. First, an ideological issue: “particle” is an ill-defined term that is neutral with respect to any possible language function and to any meaningful definition. It is simply a collective noun for all the indeclinable and short words that are not otherwise classifiable. This fits the very long tradition of studies in classical philology that privileges propositional content over speaker’s intentions, and immanent grammatical definitions over contextual significance. [53] The second reason is as psychological as practical. “Ancient Greek particles” almost automatically reminds ancient Greek scholars of Denniston’s work, The Greek Particles, whose first publication dates back to 1934 (the second edition published in 1954 is, however, the best known). Even though the author includes, especially in his Introduction, a number of insightful comments on the essence of such words—which I will recapitulate below—his approach differs considerably from that adopted by several recent works—this one included. This implies no dishonor to Denniston’s work; rather, I stress that the development of several fields of linguistics throughout the last century has stimulated us, at present, to fill in {201|202} some gaps in prior linguistic knowledge. For example, Denniston’s list of particles shows some lacunae: some are not included; his groupings are not always practical and useful; and, most problematically of all, some other lexical expressions that might work as particles are excluded. In addition to these gaps, I would add another, which has a didactic importance. In the last several years, especially in personal communication with several American and European colleagues, I have observed that teaching particles is still commonly regarded as a difficult task. In addition, the accounts of particles given in handbooks and grammars is considered to be unsatisfying or insufficient for teaching purposes. Chapters on particles (if there are any) tend to be last—in order of importance, too, I would add. This subordination of particles is in line with the phonology-morphology-syntax-semantics progression of most grammars. While such a progression is indisputably useful, it also entails that the less linguistic features have an immanent semantic value, the less they tend to be introduced and described. Unfortunately, these same features are also more difficult to grasp in their communicative value, and they are correspondingly difficult to teach and to translate—which is all the more reason to include them in contemporary grammars. After all, from the perspective of language-in-use, features such as discourse markers may constitute a powerful guide in the processing and understanding of textual units, rather than “remaining” optional or ornamental elements. For all of these reasons, I prefer to drop the term “particle’” in the description of the theoretical framework I am proposing and to adopt “discourse markers” instead; while it is also a broad term, it does indicate something about the function of these words or groups of words, and, most of all, it calls to mind a field of linguistic research that is more helpful than others for this topic. Finally, “discourse markers” permits the consideration of the non-propositional functions of some adverbs; “particles” is simply meant to exclude adverbs, no more than that.

Schenkeveld’s notable survey of views about particula (in Latin) and moria (in Greek) throughout the centuries shows that the modern use of the term “particles” and a modern theory about them started in the sixteenth century, mainly due to the influential works of Devarius (a Greek living in Italy who published Tractatus de Graecae linguae particulis in 1588) and of Sanctius (a Spanish scholar who published Minerva seu de causis linguae latinae in 1587). Some of their observations about the uses of particles present striking precedents for contemporary discourse studies. Devarius, in particular, makes some significant points: particles work if put in combination with other words and if they have vis (“force”) and efficacitas (“effectiveness”); they are “signs of grammatical mood, of order, and have the value of continuation, transition, interrogation, doubting, reasoning, or another quality of referring”; the group of particles {202|203} includes nouns and verbal forms that have lost their lexical meaning; finally, dropping them in reading would seriously compromise the understanding of the sentences (Devarius invites the reader to understand a passage from Aristotle’s Politics where all particles have been omitted). [54]

In the Preface to the first edition of The Greek Particles, Denniston defines particles negatively as “intangible and elusive words,” but he also speaks of their pragmatic nature, if indirectly: twice he pinpoints their “force,” and in remarking upon the difficulty of translating them, he mentions alternative renderings such as “italics” and “exclamation marks” as “convenient devices.” [55] The latter point is further developed in the Introduction (xxxix): “Often they cannot be appropriately translated into a modern language, and their effect must be suggested by inflexions of the voice in speaking, or by italics, exclamation marks, or inverted commas in writing.” I shall return to the analogies between particles, discourse markers, punctuation marks, and prosodic intonation later on. Denniston’s general definition of the particle (xxxvii) lends support to the presentational as well as the interactional level of interpretation: “I will define it as a word expressing a mode of thought, considered either in isolation or in relation to another thought, or a mood of emotion.” Insightful and relatively detailed descriptions particularly concern the particles’ presentational usefulness: their connective functions establish “a relationship between separate ideas. Relationships may be established in different ways. (a) The second idea is linked to the first by a connecting particle, which may do no more than connect, but may also give a logical turn (adversative, causal, or inferential) to the connexion” (xxxix–lx); “[some particles] pick up the thread of a thought” (xli); “Coherence of thought is adequately secured by the presence of a backward-pointing particle. The reader or listener, when he has reached a certain point, meets a particle which looks back to the road he has traversed, and beckons him on in a certain direction. But greater coherence is attained {203|204} if in addition a forward-pointing particle warns him in advance what path he will soon have to travel, the connexion being expressed reciprocally, from rear to van and from van to rear. It is characteristic of the Greek love of orderliness and lucidity that this double method of connexion is already present in Homer” (xli–xlii).

In the last twenty years, several scholars have started publishing works on particular ancient Greek particles or on multiple particles in particular authors, from a pragmatically oriented perspective. [56] Here, I mention some of these studies, along with some general outcomes. Levinsohn 1987 is a study of textual connections in the New Testament (specifically, in the Acts of the Apostles); Levinsohn argues, for example, that δέ begins a new developmental unit of the story, whereas καί marks single elements within these units. Sicking and Van Ophuijsen 1993 offers an analysis of the function of different particles in Lysias and in Plato. They underscore that different particles have “more than one alleged ‘sense’” (Sicking and Van Ophuijsen 1993:45); for example, different aspects of discontinuity and of continuity conveyed by the use of δέ are illustrated; as far as ἀλλά in Lysias is concerned, its function seems not to be the marking of an adversative relation, but rather the “replacing” of “one proposition or other notion by another one which is nearer to the heart of the speaker’s concern” (Sicking and Van Ophuijsen 1993:50). Wakker inserts in her work on conditional clauses a section about particles and their functions in combination with εἰ (Wakker 1994:303–364). [57] ἄρα, δή, που, and τοι are classified as modal particles (“by means of which a speaker may signal his own attitude towards the proposition he presents,” Wakker 1994:343), whereas γε, περ, and καί work as scope particles (“which mark the scope of (part of) the utterance …; they specify to whom/to what/in which case etc. the utterance applies,” 1994:307). [58] H. Dik (1995:31–51) discusses the pragmatic role of some particles—in particular καὶ δὴ καί and μέν … δέ—in terms of Focus constructions (within the Functional {204|205} Grammar’s framework), her corpus being Herodotean prose. Oréal 1997 synthesizes the pragmatic values of ἄρα, δή, οὖν, τοίνυν, καίτοι, μήν, and ἀλλά, as they occur in Demosthenes’ Speeches in terms of what the speaker signals about the ongoing utterance at the presentational and at the interactional level. Durán López 2000 proposes two criteria for the definition of the ambiguous term “particle,” one of which is the relationship with discourse functions (she refers to Jakobsonian phatic, conative, metalinguistic, and referential functions). The remarkable volume New Approaches to Greek Particles (Rijksbaron 1997), which represents a first attempt to update the accounts of several particles in several ancient Greek works—mainly prose texts—includes six articles that explicitly adopt pragmatic views (namely those by Basset, Jacquinod, Rijksbaron, Sicking, Slings and Wakker). [59] Finally, Van Ophuijsen and Stork 1999 devotes particular attention to particles and sentence adverbials in book VII of Herodotus’ Histories. As far as Homeric particles are concerned, Ruijgh 1971 (Autour de “ΤΕ épique”) constitutes a wide repertory of syntactical and semantic facts about many more particles than τε; however, the underlying approach is not pragmatic per se. Conversely, Bakker’s observations on different Homeric particles are consistently informed by a discourse-oriented perspective. The author (2005:135n51) indirectly points out an originally interactional value of τε that had been proposed by Bloch 1955 (τε derives from a second-person pronoun with the adverbial force “as you know”). Elsewhere, Bakker underscores that adverbial τε is presumably signaling the force of a particular conceptual link in a particular context (2002:77). Bakker 1988 focuses on περ as a scalar particle; [60] Bakker 1993a centers on δέ and its “discontinuous force.” Homeric ἄρα is interactionally characterized as a mark of visual “nearness,” “signaling the interpretive involvement of a speaker in a situation.” [61] About αὐτάρ, Bakker notes that its original meaning is “perhaps more specifically adversative than that of δέ”; however, since they both mark “switches and other transitions,” they are “equivalent,” and “serve as metrical alternatives for one another” (Bakker 1997b:96).

Ancient Greek particles share with discourse markers the following characteristics in terms of uses and functions (here, I refer to the numbered list above in the section “Discourse Markers”): they embrace different grammatical categories (conjunctions such as ἀλλά, adverbs such as δήπου, interjections such as οἴ, verbal forms such as ἄγε, φέρε, ἴθι, ἰδέ, and ἰδοῦ); they are {205|206} related to the utterance situation (as with νυν and τοι); they are syntactically detachable; they typically do not add anything to the propositional content of the sentence; [62] they are multifunctional (as with δέ); [63] since they primarily have one or more pragmatic functions, when paraphrased or translated their lexical or original meaning often cannot be kept (as with, e.g., καίτοι); they may occur in clusters or chains (as with ἀλλ’ ἤ and many other coalescences, such as μέντοι or γοῦν). I would further point out that the syntactical detachability of particles matches their occurrence in paratactic units of discourse, whereas conjunctions that subordinate (thus being hypotactically relevant) are obviously not detachable; in that case, their function is representational and their meaning is propositional. [64] As for prosodic independence or intonational markedness, ancient Greek particles seem to be all but prosodically independent, especially in the form of clitics; however, their occurrence may contribute to the prosodic prominence of the adjacent words. For example, γε may correspond to a special effect in the tone of voice (LSJ says that γε can be rendered also by an “emphasis in pronunciation”), which presumably involves the words that γε is close to. The same holds for the contrastive focus function of words uttered with prosodic prominence in many current languages. [65] Of course, the lack of oral recordings prevents us from establishing whether sentence initial particles or groups of particles were read aloud with specific intonational contours. Nevertheless, it does not seem unreasonable that clusters of particles at the beginning of answers in Plato’s dialogues, for example, which are often followed by a comma in written code, might have corresponded to a specific intonational contour. [66] This topic leads me to clarify a couple of {206|207} related points concerning the sentence initial position criterion for the identification of particles working as discourse markers. Fraser 2001 has included the so-called second position in ancient Greek clauses within a model that “integrates prosody and structure.” The second position—usually occupied by “prosodically light words”—can be defined by “the intonation group to which the words belong: either as enclitic on P1 [i.e. first position], or (following them in collocations) as preposed at the start of the basic clause.” [67] This means that in addition to sentence initial particles such as ἀτάρ and ἀλλά, particles occupying the so-called second position (after Wackernagel’s law [68] ) such as δέ and γάρ, and, even more to the point, enclitic particles such as που and τε, can be understood to contribute to the prosodic prominence of the entire initial group of words, whatever the semantic and/or pragmatic function of this prominence may be. [69] Therefore, particles occurring in second position can be accounted for as sentence initial discourse markers.

This raises a further relevant issue: does sentence initial position occur only after full periods, or does it correspond to the beginning of any independent clause? If the latter, we are dealing with a remarkable amount of sentence initial particles in the Iliad and in the Odyssey, as parataxis is a very conspicuous feature of Homeric language. Yet, our modern editions of the Homeric poems do not fully show that parataxis because of a seemingly secondary, but actually crucial, aspect of written code—namely, punctuation. And punctuation is a modern fact, which was not originally part of the text. [70] My claim is that the variety of written boundaries we find before δέ—for instance, full periods, semicolons, and commas—might be the result of arbitrary choices regarding syntactic coordination. Of course, the issue is complicated by the fact that punctuation marks—exactly like particles and discourse markers—are, in principle, multifunctional; therefore, the metatextual meaning of each punctuation mark may be driven sometimes by syntax (e.g. commas for hierarchically equal entries in lists), sometimes by prosody (e.g. question marks at the end of questions), and sometimes by the arbitrary interpretation of the presentational and of the interactional structure of discourse (e.g. the explanatory uses of colons, {207|208} semicolons, and commas, or the interactional purposes of suspension points). [71] Hannay and Kroon (2005) have pointed out the relevance of punctuation units in terms of discourse units. Their argument sheds light on a significant correlation: “discourse acts are preferably realized by intonation units for the spoken language and punctuation units for the written language” (2005:107). My own suggestion about Homeric particles in independent clauses is that they may mark the discourse acts underlying the clauses, and signal the latter’s relation to previous as well as to subsequent acts. Discourse acts do not necessarily have to coincide with entire clauses; they may relate to a part of the clause (such as, for example, a vocative) or they may embrace more than one clause. To transfer this formulation into practical terms: in order to grasp the discourse structure of a text, relying on particles works much better than relying on punctuation, because they are part of the original text (whereas punctuation is not) and because their frequency and their variety help to process the upcoming units in a potentially much more precise way, regardless of syntax. A final characteristic of discourse markers might be helpful with respect to the debate about whether particles are an oral feature or not. Discourse markers tend to occur more in spoken than in written language; however, they are not an oral phenomenon per se. Likewise, my view is that ancient Greek particles are not an oral phenomenon per se. Still, they may occur more frequently in texts that were supposed to be orally marked (such as, for instance, Plato’s dialogues) or orally/aurally received (such as tragedies and comedies). [72] Most of all, particles have a strong metatextual potential in terms of presentation and interaction, which in principle affects both oral and written communication.

Having argued these claims, I yet insist on the notion of discourse markers beyond the consideration of particles, for if we admit that ancient Greek texts (Homer included) were using discourse markers, the range of lexical expressions having this function widens. Particles are not the only linguistic features working as discourse markers. We can add the following: the so-called sentence adverbials, such as πρῶτον, δεύτερον, ἔπειτα (especially ἔπειτα δέ)—with a presentational force; exclamatory expressions containing interjections or not, such as οἴμοι τῶν κακῶν, εὖ γε, and νὴ Δία—with an interactional force; and various lexical phrases containing particles or not, such as ἄγε δή, ἔστιν ὅτε, ἑνὶ λόγῳ, and τὰ ἄλλα/τἄλλα. This larger view of lexical items that give {208|209} “instructions for the hearer how and where to integrate an upcoming unit of speech within the current discourse” (Kroon 1995:46) permits us to overcome a major difficulty concerning Greek particle—namely, the fact that they may work at the presentational/interactional, as well as at the representational level. For instance, καί can certainly be said to belong to the category of particles (Denniston includes it in his catalogue); however, it is not easy to distinguish when καί connects facts and when it connects acts. As we saw, the same holds for English “because.” Therefore, I propose not to group words lexically, by entries, but rather to group them by function and, specifically, by discourse function. In the following section, I have chosen to examine some adverbs that work as discourse markers; of this group, only αὐτάρ is regarded by Denniston as a particle.

αὐ-discourse markers

αὖ, ignored by Chantraine (1942–1953, II) and by Denniston (1954), is labeled in LSJ as an “adverb for a repeated action” and is translated as “again, anew, afresh, once more”; the five basic subsets identified refer to the following nuances of meaning: “again, once more” especially after numerals; “again” in questions (“expressing impatience”); “on the other hand, in turn,” analogous to δέ; “on the contrary;” “backwards.” Monro (1891:308) suggests that the original value of αὖ was “backwards”; [73] notwithstanding, he says, “in Homer this sense is only found in the form αὖτις.” Kühner and Gerth (1955, II:278) hypothesize that the possibly original spatial meaning “retro” early turned into the temporal one “rursu,” and only afterwards did the meaning “contra” emerge. Moreover, an evolution from the adverbial to the conjunctional function is posited. According to Schwyzer (1939–1971, II:560), αὖ never reached “die Stufe ‘aber’,” as Latin autem and Oscan aut did. Conversely, in Smyth (1980:637) αὖ is labeled an “adversative particle,” with the specification “often used with personal pronouns.” Humbert 1972:385 translates it “d’un autre côté,” which conveys returning movement and repetition. LfgrE (I:1521–1529) distinguishes between αὖ particle, with a slightly adversative or even a continuative function, [74] and αὖ adverb, mostly having the temporal meaning “again.” Its recurrence close to demonstratives is recorded as well (1527). Ruijgh (1971:690) points out the frequent combination with δέ (δ’ αὖ). {209|210}

αὖτε is mostly considered as a synonym of αὖ. LSJ underscores its usage “to mark sequence or transition” and reminds that δηὖτε (crasis for δή + αὖτε) occurs frequently in lyric. LfgrE (I:1582–1590) shows the same distinction as for αὖ—namely, between αὖτε as an adversative particle and αὖτε as an adverb meaning “again.” Nonetheless, usages of αὖτε that are also non-adversative—that is, continuative—are included in the former category (I:1589), while usages not implying repetition are included in the latter (1584). Ruijgh (1971:690) sees in the temporal suffix –τε (cf. τότε, ἄλλο-τε) an originally temporal mark. He also records that in Homer αὖτε occurs more than αὖ and αὖτις, and after Homer it basically disappears.

On the whole, these accounts of αὖ and its derivatives focus on the representational level of discourse. When they do refer to possible contextual implications, they list unrelated elements, such as “expressing impatience” for αὖ and “mark of sequence or transition” for αὖτε. Because of the completely different meanings of αὔτως—which I will focus on in chapter 5—some scholars have even argued for two different etymologies. These accounts do not explain why the same adverb can have such different meanings nor do they illumine the ground of these different meanings. I argue that the difficulty of classifying some occurrences, as well as the meaninglessness assigned to the Iliadic αὐτάρ mentioned above, rest on a partial view of the communicative effectiveness of discourse. This would not be an obstacle if the uses of αὖ and its derivatives were mainly representational. But that is not the case. αὖ and its derivatives {210|211} are multifunctional by nature. They have no independent meaning: their raison d’être corresponds to the communicative goals of the sentence (or clause) containing them. Thus, they must be analyzed within other (or new) conceptual categories. There is not an a priori translation that fits their multiple uses; there can be only a functional sense of what they are for.

Intriguing indirect indications of the discourse functions of what I call αὐ-discourse markers come from etymological hypotheses by Indo-Europeanists and from recent analyses of Latin autem. Let us start with the etymological contributions, which are remarkably various. Pokorny (1959:72–74) distinguishes between an IE *au- preverb, “away from” (cf. αὐχάττειν, Lat. aufero, aufugio), [79] from whence αὔτως meaning “in vain” is also said to come, and an IE *au-, u- third-person pronoun, “that one, the other one,” which is to be connected to Old Indic u, u-tā, and αὖ, αὖτε. Conversely, Chantraine (1999:137) interprets both basic meanings “away from” and “on the other hand” as being included in αὖ, which is to be considered either as a mark of separation (see Lat. aufero) or as a mark of opposition (see Lat. aut, autem, and Oscan auti). The debate about the etymology and the original functions of αὖ became disputatious through two contrasting studies, namely, Klein 1988 and Dunkel 1997. According to Klein 1988, the adverb comes from the Indic particle *au, u, whose oldest usage is in coreferential structures—see the prototypical ὁ/τὸ δ’ αὖ—and it works as discourse continuative. [80] Conversely, Dunkel (1997:170) hypothesizes the prior existence of three functionally distinct particles, i.e. a conjunctive one (IE * h₂o), a deictic one (IE *u), and a local one (IE *awo, “downward”). He argues that the coreferential usage is an Indic innovation and that Gr. αὖ comes from conjunctive *h _₂ o, originally a local adverb meaning “next to, up against” (1997:161 and 164). On the whole, αὖ seems to fluctuate etymologically in conveying separation/opposition and conjunction/continuation. To complicate the frame—or to make it more fascinating—some relevant work comes from Hittite, as well. Melchert (1994:134) states that Hittite had a preverb ḫu (from *-h _₂ o/u) whose meaning is “to, toward”; the implied movement seems to be “I”-oriented (for example, Hitt. eḫu means “come!”). Hittite had also a preverb aw– whose meaning is “away [from me]” (cf. IE *au-, mentioned by Pokorny in connection with Lat. aufero). [81] Thus, ancient Greek αὖ might be related to both a {211|212} movement toward the speaking “I” and a movement away from the speaking “I.” [82] Finally, αὐτάρ is, for the most part, interpreted as the blend of αὖτε + ἄρ. [83] Ruijgh (1971:716) instead suggests that we think of αὖτις + the evidential particle ἄρα. Katz (2007) argues that αὐτάρ may be read as αὖ + ταρ, the latter corresponding to Cuneiform Luwian particle tar. [84]

Kroon’s view about Lat. autem is summarized as follows: “[it] is a presentational particle which marks the discrete status of a piece of information with regard to its verbal or non-verbal context. … It can be characterized as a ‘highlighting’ or ‘focusing’ device … , or as a marker of the organization of the text” (Kroon 1995:226; italics in the text). Thus, the basic meaning is “distinctiveness”; the basic function is at the presentational level of discourse, either as a “marker of focus” [85] in local usages, or as a “marker of thematic discontinuity” [86] in “global” usages (Kroon 1995:270). In the former case, it can mark a “parallel” {212|213} or an “absolute” focus; parallel focus implies that the adverb signals an attention shift “from constituent A to constituent B”—for instance, in changes of addressee (Kroon 1995:229, 233 and 238). Autem in this parallel focus function often occurs in combination with a personal pronoun; in these cases, it is “apt to cluster with referents whose identity is salient, surprising or ambiguous in the particular context or situation” (Kroon 1995:235). Kroon also comments valuably on the “compatibility of strong deixis” with autem (Kroon 1995:237) and on the mixture of conjunction and disjunction implied in the usage et … et autem. [87] Orlandini 1999 analyzes the same Latin adverb as a connector that works at two discourse levels, i.e. not only at the presentational (cf. Kroon 1995) but also at the interactional. On the presentational level, autem operates with respect to the organization of the text; as such, it can present a proposition by marking the transition effect—“séparé du contexte précédent par une forte pause”—or it can narrow the scope of the topic and have a limitative value—“quant à”; it can also indicate a semantic opposition—“mais,” “alors que,” “tandis que” (see Orlandini 1999:150–157). On the interactional level, autem operates with respect to the performance of the text and can have a certain “force argumentative”; [88] as such, in dialogs it can underscore “la persistence d’une attitude de l’interlocuteur qui indigne le locuteur” (especially in questions; see Orlandini 1999:157–158); out of dialogs, it can introduce a parenthetical clause as an explicit speaker’s insertion, which creates “dissociation énonciative” (Orlandini 1999:158–159).

Recently, specific attention has been devoted to αὖ and αὖτε by Revuelta Puigdollers. The author examines the occurrences of these particles in a wide range of texts (both poetry and prose); he determines that they work “as boundary-markers that highlight the transition between different discourse units” (2009:106). Their contribution to topic management, in particular, is detectable through αὖ and αὖτε signaling the introduction of a second or further topic, the resumption of a given topic, and—in fewer cases—the introduction of an entirely new topic (2009:86–98). The ideas of topic change and topic chain concord with the main presentational functions I will illustrate in the next sections. Moreover, that these ideas underlie the uses of αὖ and αὖτε {213|214} well beyond Homer (in, for example, Herodotus, Euripides, Plato, Lysias, and Xenophon) is certainly a most relevant finding; it also shows that these particles are definitely not just optional elements of texts. Revuelta does not consider any possible interactional value of αὖ and αὖτε, nor does he include αὐτάρ in his analysis. He does, however, crucially disprove that αὖ and αὖτε express repetition, a claim often found in dictionaries (2009:83). As I will show, the principle of repetition is all but irrelevant to the pragmatics of αὖ and αὖτε; while αὖ and its derivatives do show repetition, that repetition pertains to acts rather than facts.

My own proposal about the functions and the meanings of αὖ, αὖτε, and αὐτάρ exclusively concerns the Homeric uses (though I will discuss some melic uses at the end of this chapter). My perspective of study encompasses more metatextual functions than those related to the handling of topics. In order to convey the horizons of such a perspective, I first resume some key concepts concerning the production and the reception of Homeric poetry—that is, involvement, visualization, and recognition.

In regard to the force of poetic utterances and the self-involvement of the performer, two cognitive processes are central; in Homeric literature, they have been called “focalization” and “re-enactment.” “Focalization” is borrowed from the narratological analysis of De Jong (1987), according to which the Homeric narrator is also a focalizer. While narrating consists in the “verbal presentation of the story,” focalizing consists in the “perceptional, emotional and intellectual presentation of the fabula” (De Jong 1987:32). Thus, focalization represents a subjective and engaging part of narrative, linguistically conveyed by particles (e.g. ἄρα or ἦ) and by verbs of perception. This does not mean that the only possible subjective point of view is that of the performer; in fact, Homeric narration also includes “embedded focalization”—that is, perceptional presentations by characters, typically occurring in reported speeches. [89] In any case, “the time-honoured dogma of Homeric objectivity” (De Jong 1987:221) is discredited. “Re-enactment” denotes the reactivation, the recreation, of the epic past “in the here and now of the performance.” [90] The epic singer anchors every mythical event to the present, as it is assumed to re-occur in his mind’s eye (as well as the audience’s). In Bakker’s view, augmented aorist verbs, constructions with μέλλειν, and particles such as ἄρα and δή are linguistic traces of that recreation; {214|215} they all convey temporal and visual nearness. [91] These two concepts—resting on different but complementary perspectives: the narratological and the pragmatic—converge upon the idea of cognitive involvement and of emotional engagement by the performer toward what is performed; such processes strongly condition both the production and the reception of Homeric texts.

Another crucial concept is visualization, a quintessential cognitive psychological means to engage the audience of the epic performance. A couple of arguments from cognitive psychology serve to ground this point. First, as Collins 1991 has shown, visual imagery in narratives and descriptions reflects the same cognitive processes as in immediate visual perceptions. [92] Second, as Rubin 1995 has shown, imagery is among the most powerful tools for recalling a song in oral traditions. [93] Some recent scholarship examines Homeric strategies of visualizing in narration from the cognitive point of view. Among the different kinds of memory required of the epic singer—memory for typical scenes, visual memory, spatial memory, and auditory memory—the visual and the spatial play a prominent role according to Minchin (2001:30 and 157). Within the singer’s cognitive map, the material is both thematically and spatially connected through mnemonic prompts that are associative, semantic, phonological, and/or visual (Minchin 2001:85–90). [94] Bakker 1997b argues for a close interdependence between the thematic and visual organization of Homeric discourse: “… all narration is description” (Bakker 1997b:57). The thematic organization of discourse harmonizes with the visual procedure of scanning the different scenes of the epic narrative. [95] The narrative articulations that Bakker analyzes as constituents of what he calls “syntax of movement” (close-up units, addition units, framings [96] ) reflect the oral techniques that are used to effectively introduce the content and the visualizations—about persons, objects, and events—that the performer intends {215|216} to share with his listeners. [97] To sum up, the thematic organization of discourse is informed by visual procedures. Moreover, mental imagery and actual perception, which, according to the cognitive sciences, share fundamental properties, tend, in Homer, to be conveyed in the same way. The aim of the performer is to create immediacy—that is, to nullify the temporal and spatial distance between the events of the past and the narration that lets those events take place again. [98] Among the linguistic markers of visualization as a prominent feature, αὖ and its derivatives seem to play a significant role. [99]

If visualization is a particular way of involvement, recognition is, in turn, a particular way of visualization. By recognition I mean a specific act implied by certain verbs and particles; it is a sudden as well as profound visual acknowledgement of something/somebody already known that is now seen under a new light. It is not only cognition “again,” but it is also new cognition. [100] The usual contexts for these acts are epiphanies and occasions of a god’s presence, or even noun-epithet formulas that work as “minirituals,” as when they introduce the epiphany of a hero on the stage. [101] A central verb that conveys recognition in this sense is νοέω. Such a verb denotes a cognitive activity that is mental and visual at the same time. [102] Recognizing through vision is essential to the interpretation of signs as well. [103] The link between recognition, the hieratic sphere, and the “material” effects of remembering is outlined by Bakker in his exploration of the full recognition of god’s presence during the performance of the Homeric Hymn III to Apollo. [104] Finally, in an article entitled “Symbolic Action in {216|217} the Homeric Hymns: The Theme of Recognition,” García argues that the narrative pattern of many hymnic or epic scenes within the Homeric Hymns includes motifs such as “deity in disguise, recognition by a mortal (or failure therein), and reward (or punishment).” [105] The recognition motif, in particular, may include, in turn, such submotifs as “failure to recognize,” “recognition” (in which includes the following elements: “mortal notes beauty, radiance or miracle of god,” “χαῖρε” and “mortal wonders or guesses which god”), and, finally, “the god acknowledges recognition.” [106] With respect to the topics of the current chapter, I underscore that the illustrative passages cited by García include occurrences of verba videndi, of νοέω or γιγνώσκω, and, sometimes, of αὐτίκα as well. [107] In sum, recognition represents a most relevant mix of cognitive and emotional activities, such as seeing and considering “again,” suddenly sensing something new about an already known subject, realizing somebody’s real identity, and acknowledging in the hic et nunc of the performance the truth of something being reenacted. This conceptual mix is what characterizes the majority of the uses of αὐ-discourse markers as well.

The following pragmatic and literary analysis of αὖ, αὖτε, and αυτάρ pays particular attention to the different visual aspects of the transitions being marked. A further basic focus is upon the fundamental dialectic of discourse connectivity, which is the interface between continuity and discontinuity. As I will show, continuity involves the control of the performer, as a pre-condition of his work, over the coherence and sequence of single utterances and between entire clusters of utterances. Discontinuity, conversely, concerns topic switches, setting switches, narrative switches, zoom switches, and emotional loads. Finally, as I noted in my Introduction, while my presentation will privilege {217|218} instances that show one particular function (in line with my overall argumentative purposes), the reader is invited to remember that multi-meaning uses of an αὐ-discourse marker in a single Homeric passage are regular, and are characteristic of poetry, which both induces specific inferences and leaves space for a wider range of inferences.

Presentational functions of αὖ, αὖτε, and αὐτάρ

“On the other side” and “then” shifts (signaling a “parallel focus”)

The main presentational functions of αὖ, αὖτε, and αὐτάρ deal with vision. All three primarily mark a shift from what is “on the one side” to what is “on the other side.” Within the visual framework that is “primed” [108] at a certain moment of narration, they grammaticalize the shift of focus to “another” item (often one or more individuals) or “another” event appearing/happening. Here is an occurrence of αὖ illustrating this basic use:

κόψ’ ὄπιθεν κώληπα τυχών, ὑπέλυσε δὲ γυῖα,
κὰδ δ’ ἔβαλ’ ἐξοπίσω· ἐπὶ δὲ στήθεσσιν Ὀδυσσεὺς
κάππεσε· λαοὶ δ’ αὖ θηεῦντό τε θάμβησάν τε.

Iliad 23.726–728; the primary speaking ‘I’ to the audience

He [Odysseus] caught him [Ajax] behind the hollow of the knee; he unloosed his joints;
he threw him down, backwards. On his chest, Odysseus
fell as well. [On the other side] the people watched and were impressed.

Odysseus and Ajax are wrestling at the funeral games in honor of Patroclus. Once the description of one of Odysseus’ moves is complete, the primary speaking ‘I’ narrates the reaction “on the other side” of the visual field. [109] {218|219}

What is “on the other side” does not have to be located in a specific place; the presentational relevance is simply the selection of one side within a more comprehensive visual framework. For example, immediately after Zeus has sent thunder and lightening favorable to the Trojans and terrifying to the Achaeans in Iliad 17 (595–596), the primary speaking ‘I’ begins to describe which Achaeans were fleeing. Πρῶτος (597) is an adverbial adjective that works as a discourse marker; the presentational meaning is clear: the Achaean warrior Peneleus, about to be attacked by Polydamas, is the first item of the narrative description. Then, two more warriors, Leitus and Hector, are visualized:

Λήϊτον αὖθ’ Ἕκτωρ σχεδὸν οὔτασε χεῖρ’ ἐπὶ καρπῷ
υἱὸν Ἀλεκτρυόνος μεγαθύμου, παῦσε δὲ χάρμης·

Iliad 17.601–602; the primary speaking ‘I’ to the audience

Then, Hector wounded Leitus in the hand, on the wrist,
the son of high-spirited Alectryon. And then he rested from his eagerness for combat.

The narrator is not interested in placing the second two warriors here or there; he simply visualizes them on another side of the primed framework (i.e. the overall scene of the Achaeans fleeing from the Trojans).

These first two passages show simultaneity with respect to the actions previously described: spectators wonder while seeing Odysseus falling; Hector pursues Leitus while Peneleos is attacked by Polydamas. Before commenting upon this, let us consider a different case, in which αὖτε presumably implies a slight posteriority:

… θέουσα δὲ Ἶρις ἐπέστη
βηλῷ ἔπι λιθέῳ· τοὶ δ’ ὡς ἴδον ὀφθαλμοῖσι
πάντες ἀνήϊξαν, κάλεόν τέ μιν εἰς ἓ ἕκαστος·
ἣ δ’ αὖθ’ ἕζεσθαι μὲν ἀνήνατο, εἶπε δὲ μῦθον·

Iliad 23.201–204; the primary speaking ‘I’ to the audience {219|220}

… Iris ran and stood
on the doorstep made of stone. As they [the winds] saw her with their eyes,
they all darted up and each of them called her toward himself.
But she [shift to the other side] said no, she didn’t want to be seated; she spoke …

Iris is first described as standing on the doorsill; then, the winds are said to call her (on one side of the frame). Finally, the mind’s eye shifts again to Iris’ side, to tell about her refusal and about her initiation of speech (204). The shift implies a slight temporal gap between the winds’ call and Iris’ refusal. My point is that sometimes the transition to what is “on the other side” includes posteriority: something happens on the one side; immediately after, on the other side (αὖ, αὖτε, or αὐτάρ), something else happens; but sometimes (more often, actually), the transition implies simultaneity: something happens on the one side; meanwhile (αὖ, αὖτε, or αὐτάρ), something else happens on the other side. Simultaneity in this kind of visual discontinuity is often expressed by the pronoun ὁ, μέν + (δ’) αὖ(τε). [110]

The question of the Homeric narrator’s ability—or inability—to narrate simultaneous events has been contentious ever since Zielinski’s (1899–1901) arguments, and has been extensively discussed, in particular by Patzer (1990), Rengakos (1995), and Nünlist (1998), as well as, very recently, by Scodel (2008). Patzer, Rengakos, and Nünlist present different examples that argue against a rigid succession of events in the larger narration. My own research on the presentational meanings of αὐ-discourse markers definitely supports an idea of narrative flexibility: αὐ-discourse markers may mark shifts involving temporal gaps as well as shifts involving simultaneity. [111] Further evidence of that claim will follow in this section. Translations often plays a role in misleading inter- {220|221} pretations of the meaning of αὐ-discourse markers. Let us consider, for example English “then,” which can work either propositionally or presentationally:

(a) He made a number of telephone-calls and fed the cat. Then he turned on the television.

(b) So much about pollution. Then there is our drought problem.

Kroon 1995:13 (italics in the text)

The difference between “then” in (a) and “then” in (b) is that in (a) it expresses a temporal gap in the described state of affairs, whereas in (b) it expresses performative succession (“a first point is about pollution; another one—“then”—is about drought”). [112] Performative succession is quintessentially presentational: it signals what comes next in the speaker’s organization of discourse. Accordingly, Lattimore’s translation “then” for αὖθ’ at Iliad 17.601 (above, p. 219) is correct, to the extent that it has a presentational meaning (which is why I added a comma after “then”); otherwise, one might interpret the word to mean that Hector was chasing Leitus once Peneleus was attacked by Polydamas, which is not the case. I will return to presentational “then” below.

αὐ-discourse markers can convey atemporal discreteness in location as well, as in this description of the two entrances of the Ithacan cave of the Nymphs:

… δύω δέ τέ οἱ θύραι εἰσίν,
αἱ μὲν πρὸς βορέαο καταιβαταὶ ἀνθρώποισιν,
αἱ δ’ αὖ πρὸς νότου εἰσὶ θεώτεραι· …

Odyssey 13.109–111; the primary speaking ‘I’ to the audience

… It [the cave] has two doorways,
one is toward the North wind, and is accessible by humans,
the other [on the other side] is toward the South wind, and it is for the gods …

The speaker uttering αὖ, αὖτε, or αὐτάρ helps the recipient to track items that are shifted in space in possibly different directions: on the side, on the {221|222} opposite side, upward, and backward. [113] If concerned with individuals, the spatial shift is not specific; it is generically “on the other side” of the current visual field. Such kinds of visual shifts might well underlie uses of αὖ, αὖτε, and αὐτάρ in “you”–“I” or “I”–“you” shifts, which all refer to actions that are to be accomplished simultaneously. [114] Especially interesting in this regard is αὐτάρ in Iliad 1.282, which is placed amongst “cases difficult to classify” in the Lexikon des frühgriechischen Epos, and which appears to be “meaningless” (“sinnlos”). [115]

Ἀτρεΐδη σὺ δὲ παῦε τεὸν μένος· αὐτὰρ ἔγωγε
λίσσομ’ Ἀχιλλῆϊ μεθέμεν χόλον …

Iliad 1.282–283; Nestor to Agamemnon

You, son of Atreus: make your fury stop; I [look at me, now] am the one
who is begging you to give over your anger against Achilles …

In the former sentence, Nestor’s request has the addressee (“you”) as the agent; the latter sentence, however, specifies the content of the request. By means of αὐτὰρ, Nestor suggests to Agamemnon that he face the speaking “I” (Nestor himself) as the subject of the current performative act (note the explicit performative verb λίσσομαι). The strong deictic mark ἐγώ and the particle γε reinforcing ἐγώ fulfill the visual and presentational function of αὐτάρ. Just as Agamemnon is invited to face Nestor, so, too, the recipients of the performance (including we readers) are encouraged to shift their attention from Agamemnon to Nestor. Thus, far from being “useless,” αὐτὰρ does have a communicative function. The point is that this function simply cannot be seen if the meanings of αὐτάρ can be only either adversative or continuative.

Quite often, occurrences of αὖ, αὖτε, and αὐτάρ that mark the shift to “the other side” and that include a slight posteriority are those exploited in turn-taking formulas and at the end of a speech. In familiar formulaic sentences, {222|223} such as τὴν δ’ αὖ Τηλέμαχος πεπνυμένος ἀντίον ηὔδα, [116] τὸν δ’ αὖτε προσέειπεν ’Αλέξανδρος θεοειδής, [117] or αὐτὰρ ὃ μειλιχίοισι προσηύδα ποιμένα λαῶν, [118] the communicative usage of αὐ-discourse markers is not to indicate opaquely “in turn” but, rather, to signal the shift from the previous speaking “I” to the next one “on the other side,” where visualization aids disambiguation. In line with this, once one character’s speech is over, transition formulas such as ὣς φάτο followed by αὐτὰρ + ὁ pronouns signal the shift of attention to someone other than the just preceding speaker, in order to tell about either the reaction or what happened afterwards, “on the other side.” The following instance includes αὖτε:

Ὣς ἔφαθ’, Ἕκτωρ δ’ αὖτε χάρη μέγα μῦθον ἀκούσας

Iliad 7.54; the primary speaking ‘I’ to the audience

So he [Helenus] spoke. Hector [on the other side] was happy to hear such a great speech [119]

Before proceeding with our analysis, it is important to clarify the relationship between δέ and αὖ or αὖτε (αὐτάρ is never accompanied by δέ, as I will discuss below) at the level of discourse functions. The “discontinuous force” (in Bakker’s words) of the multifunctional particle δέ relies on the marking of a new or of a different discourse act. When it appears with αὖ or αὖτε, δέ may signal that a new discourse act is going to be performed, while αὖ or αὖτε may specify the shift underlying this new different act. However, δέ may also substitute for αὖ or αὖτε; expressing “on the other side” is, indeed, one of the numerous functions of δέ. αὖ or αὖτε alone, without δέ, makes sense, as well: they are discourse markers specialized in marking what is “on the other side” and, thus, necessarily imply a new discourse act. [120] The point here is that more particles occurring side by side, {223|224} or even chains of particles (which is very typical of discourse markers, as well), are not to be seen as the mechanical addition of more lexical meanings, but as a flexible way to express the same ultimate pragmatic meaning. Instances of νῦν αὖ are exemplary, for they are, as we will see, pragmatically equivalent to νῦν δ’αὖ.

Everything that I have so far introduced permits me to make two points that are preliminary to the argument that follows. First, the item that is visualized “on the other side” presupposes that there is “one side” (often made verbally explicit by μέν). The idea of parallel focus is grounded upon the awareness (shared by the performer and by the audience) that there is an overarching visual frame—or a narrative structure, as we will see soon—within which to locate the upcoming chunk of information. In this way, the flow of discourse is always well organized. Performers and audiences always know from where discourse comes and to where it goes. I define this “bear-back”-cue to the frame of reference as performative continuity. αὐ-discourse markers express not only visual shifts and visual discontinuities, but also continuity. Continuity pertains to the performative level, as it is the primary speaking ‘I’ who has the control over all the threads of narration; he is the one who articulates narration and guides his recipients in connecting each part of the discourse to the next. [121]

The second point is that αὐ-discourse markers enable visual memory to go back to something that is already known (or that has been already visualized sometime before). This is primarily a cognitive aspect of anaphoricity. The presentational needs of storytelling also include “bearing-back” to a thread of discourse that had been followed in preceding sections, or “going back” to characters recently mentioned, in order to add something new. I argue that this is the basic process that led Klein to conclude that the Homeric instances of αὖ following third-person pronouns have an “anaphoric function” or are “continuative.” [122] The visual counterpart of this cognitive activity may range from what is visualized to happen “in the same way” or “along the same pattern” as the one introduced before. [123] As I hope to show, this activity of “going back” is also the basis for the expression of recognition, which I will introduce later on.

In passages dramatizing the speaker’s acknowledgment of the repetition of something already known or of something similar to what has just been said, αὐ-discourse markers may be translated as “by the same procedure” or “analogously.” [124] This happens, for example, when character X is said to say {224|225} something to Y and, “analogously,” character A is said to say something to B. [125] This sequence also holds for longer segments of text, as I will explain later.

When different kinds of lists and genealogies are performed, [126] αὐ-discourse markers are often employed to signal what I call “parallel repetition,” by which the upcoming discourse act is going to introduce a new and parallel entry. This idea is consistent with the concept of parallel focus introduced by Kroon about Lat. autem; only, it underscores that what is parallel is the presentation, the performative act of focusing on a state of affairs.

γείνατο δ’ Ἀντιφάτην καὶ Μάντιον, υἷε κραταιώ.
Ἀντιφάτης μὲν τίκτεν Ὀϊκλῆα μεγάθυμον,
αὐτὰρ Ὀϊκλείης λαοσσόον Ἀμφιάρηον,
ὃν περὶ κῆρι φίλει Ζεύς τ’ αἰγίοχος καὶ Ἀπόλλων
παντοίην φιλότητ’· οὐδ’ ἵκετο γήραος οὐδόν,
ἀλλ’ ὄλετ’ ἐν Θήβῃσι γυναίων εἵνεκα δώρων.
τοῦ υἱεῖς ἐγένοντ’ Ἀλκμάων Ἀμφίλοχός τε.
Μάντιος αὖ τέκετο Πολυφείδεά τε Κλεῖτόν τε·
ἀλλ’ ἦ τοι Κλεῖτον χρυσόθρονος ἥρπασεν Ἠὼς
κάλλεος εἵνεκα οἷο, ἵν’ ἀθανάτοισι μετείη·
αὐτὰρ ὑπέρθυμον Πολυφείδεα μάντιν Ἀπόλλων
θῆκε βροτῶν ὄχ’ ἄριστον, ἐπεὶ θάνεν Ἀμφιάρηος·

Odyssey 15.242–253; the primary speaking ‘I’ to the audience {225|226}

He [Melampus] had Mantius and Antiphates, powerful children.
Antiphates [on the one side] had great-hearted Oecles.
Then, Oecles had Amphiaraus, who drives armies,
whom aegis-holding Zeus loved in his heart and Apollo too,
with every favor. And he never reached the doorstep of old age,
but perished in Thebes, because of the presents given to his wife.
His sons were Alcmaeon and Amphilochus.
Mantius [on the other side] had Polypheides and Kleitus.
However, Dawn of the golden throne took hold of Kleitus
because of his beauty, to make him join the immortals.
Then, Apollo made high-hearted Polypheides a prophet,
far the best among the mortals since Amphiaraus died.

I have chosen this example, complicated though it is, as an exemplary instance of the work of αὐ-discourse markers in the organization of discourse. [127] αὖ and αὐτάρ work as signposts for specific hierarchical levels within the speaker’s list-making, so that the succession of generations in different family trees—on the one side, as well as on the other—appears to be clear. αὖ (249) locates what is going to be said about Mantius at the same hierarchical level as what is said about his brother Antiphates (’Αντιφάτης μέν, 243). Conversely, the two uses of αὐτάρ (244 and 252) mark a sub-topic of the genealogy (what happened to Oecles and what happened to Polypheides respectively). The ultimate communicative goal of these separate entries is to point out two important prophets, Amphiaraus and Polypheides, which is essential to the introduction of Teoclimenus (he is, in fact, the guiding reason for the performance of the genealogy).

Another instance of signaling an item that in visual memory follows a previous one in a parallel sequence occurs in the famous teichoskopia (in Iliad 3). The first Achaean hero about whom Priam asks Helen is Agamemnon. “Then,” he asks about another hero (Odysseus), and “then, again” he asks about Ajax.

191 Δεύτερον αὖτ’ ’Οδυσῆα ἰδὼν ἐρέειν’ ὁ γεραιός
. . . . . . . . . . . . . . . . . . . . . . . . . . .
225 Τὸ τρίτον αὖτ’ Αἴαντα ἰδὼν ἐρέειν’ ὁ γεραιός

Iliad 3.191 and 225; the primary speaking ‘I’ to the audience {226|227}

191 Second, having seen Odysseus, the old man asked:
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
225 Third, having seen Ajax, the old man asked:

Here, αὖτε + numerals marks the sequence of hierarchically equal segments of discourse, each of which is interposed with several hierarchically subordinate lines. As in the case of genealogies, the order of the entries—I submit—is primarily the order of the sequence to be memorized by the performer. [128] Numerals are used as presentational marks in many Indo-European languages, not only in ancient Greek (think of the presentational function of “first,” “second,” and so on in argumentative texts in English, where they are, futher, followed by a comma in written language and are intonationally marked when spoken aloud).

As a last example of listing, I cite the process of arming:

κνημῖδας μὲν πρῶτα περὶ κνήμῃσιν ἔθηκε
καλὰς, ἀργυρέοισιν ἐπισφυρίοις ἀραρυίας·
δεύτερον αὖ θώρηκα περὶ στήθεσσιν ἔδυνε

Iliad 3.330–332; 11.17–19; 16.131–133; 19.369–371; the primary speaking ‘I’ to the audience

First, he [Alexander/Agamemnon/Patroclus/Achilles] put the greaves around the shins,
beautiful, joined by silver ankle-clasps.
Second, he put the corselet about the chest [129]

In these cases, propositional and performative order fully coincide. The presentational mark is used to describe a representational procedure. [130]

Discreteness in parallel patterns is also clear whenever the speaker introduces “another” item hierarchically equal to the previous one: in those cases, ἄλλος regularly co-occurs next to αὐ-discourse markers. {227|228}

παντοίην ὄπ’ ἰεῖσαι ἀθέσφατον· ἄλλοτε μὲν γὰρ
φθέγγονθ’ ὥς τε θεοῖσι συνιέμεν, ἄλλοτε δ’ αὖτε
ταύρου ἐριβρύχεω μένος ἀσχέτου ὄσσαν ἀγαύρου,
ἄλλοτε δ’ αὖτε λέοντος ἀναιδέα θυμὸν ἔχοντος,
ἄλλοτε δ’ αὖ σκυλάκεσσιν ἐοικότα, θαύματ’ ἀκοῦσαι,
ἄλλοτε δ’ αὖ ῥοίζεσχ’, ὑπὸ δ’ ἤχεεν οὔρεα μακρά.

Hesiod Theogony 830–835; the primary speaking ‘I’ to the audience

They [Typhoeus’ voices] sent out a manifold, unutterable sound:
sometimes they were letting out a sound that gods could understand; sometimes,
the voice of a loud-bellowing bull, ungovernable in its strength, proud;
sometimes the sound of a lion with an reckless spirit;
sometimes sounds similar to those of puppies, wonders to hear;
sometimes he was hissing, and high mountains were resounding with it.

Each entry concerns the different sounds coming from the hundred heads of Typhoeus. The shifts between parallel focuses are temporal, because of ἄλλοτε; however, I submit that the recipients unavoidably process the visual shift from head to head as well—with the mind’s eyes. [131] Significantly, ἄλλος + αὐ-discourse markers may convey simultaneity as well, as the following famous simile shows:

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . ἐν δέ τε πολλὰ
κύματα παφλάζοντα πολυφλοίσβοιο θαλάσσης
κυρτὰ φαληριόωντα, πρὸ μέν τ’ ἄλλ’, αὐτὰρ ἐπ’ ἄλλα·
ὣς Τρῶες πρὸ μὲν ἄλλοι ἀρηρότες, αὐτὰρ ἐπ’ ἄλλοι,
χαλκῷ μαρμαίροντες ἅμ’ ἡγεμόνεσσιν ἕποντο.

Iliad 13.797–801; the primary speaking ‘I’ to the audience

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . the numerous
foaming waves of the loud-roaring sea
curved and gleaming, some of them ahead and the rest behind, {228|229}
so the Trojans, solidly packed, some of them ahead and the rest behind,
glittering in bronze armor, were following their leaders.

The visual component of the switch between what is ahead and what is simultaneously behind is most evident, and αὐτάρ is the grammaticalization of this.

To sum up, the idea of parallel repetition may be applied to the description of single details belonging to the same frame (e.g. waves ahead and waves behind), or to similar actions performed by the same character (e.g. asking about hero X, asking about hero Y), or to parallel events that are temporally sequential (e.g. A was the father of B, B was the father of C). The speaker thus uses αὐ-discourse markers to cue to the recipient that a “parallel” discourse act is going to come. This has important reflexes at the narrative level, as parallel repetition can embrace both shorter and longer narrative sections. Let us consider first a short narrative sequence:

421 ἀλλ’ ἄγ’ ὁ μὲν πεδίονδ’ ἐπὶ βοῦν ἴτω, …
. . . . . . . . . . . . . . . . . . . . . . .
423 εἷς δ’ ἐπὶ Τηλεμάχου μεγαθύμου νῆα μέλαιναν
πάντας ἰὼν ἑτάρους ἀγέτω, …
εἷς δ’ αὖ χρυσοχόον Λαέρκεα δεῦρο κελέσθω
ἐλθεῖν, …

Odyssey 3.421–426; Nestor to the servants

421 Come on, let one man go to the plain for a cow …
. . . . . . . . . . . . . . . . . . . . . . . . . . . . .
423 Then, let one go to the black ship of high-hearted Telemachus,
and bring all the companions back, …
Another one, then, let him invite Laerkes, who is a goldsmith,
to come here …

In giving instructions to the servants for the sacrifice to Athena before Telemachus’ departure, Nestor orders three different individuals to go simultaneously in three different directions, as if three different paths were visualized. [132] {229|230} As I have noted, the discourse marker “then” is not meant to render any posteriority but only succession in the speaker’s exposition. [133]

What happens when parallel repetition is applied to a different series of actions told in longer narrative units? Simultaneity and visual shifts structuring narrative are particularly evident in the following text, where the primary speaking ‘I’ tells what happens once the duel between Hector and Ajax stops (Iliad 7.306–345):

306 τὼ δὲ διακρινθέντε ὃ μὲν μετὰ λαὸν Ἀχαιῶν
ἤϊ’, ὃ δ’ ἐς Τρώων ὅμαδον κίε· τοὶ δὲ χάρησαν,
. . . . . . . . . . . . . . . . . . . . . . . . . . .
311 Αἴαντ’ αὖθ’ ἑτέρωθεν ἐϋκνήμιδες Ἀχαιοὶ
εἰς Ἀγαμέμνονα δῖον ἄγον …
. . . . . . . . . . . . . . . . .
344 Ὣς ἔφαθ’, οἳ δ’ ἄρα πάντες ἐπῄνησαν βασιλῆες.
Τρώων αὖτ’ ἀγορὴ γένετ’ Ἰλίου ἐν πόλει ἄκρῃ.

Iliad VII 306–345; the primary speaking ‘I’ to the audience

306 Once separated, one [Ajax] went among the Achaeans
while the other [Hector] reached the crowd of the Trojans. They rejoiced …
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
311 As for Ajax, on the other side, the well-greaved Achaeans
took him to noble Agamemnon …
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
344 So he [Nestor] spoke. All the kings approved.
Meanwhile, on the other side, the assembly of the Trojans was taking place on the summit in the city of Ilium.

The narrative flow simultaneously follows what happens on the respective sides; αὖτε at 311 and 345 sign-posts that simultaneity.

The presentational uses of αὐ-discourse markers for longer narrative sequences are, according to my analysis, the following: they may follow, as a sub-thread of discourse, what happens “on the other side” of the larger narrative frame as a sub-thread; they may re-join the main thread of discourse after an interruption (by expanding relative clauses or by similes, for instance); finally, {230|231} they may start a new thread of discourse or an entirely new narrative section. αὖ, αὖτε, and αὐτάρ are all employed in these three functions. Visual discontinuities underlie the presentational marking in all instances. However, it is possible to notice a kind of “specialization” of αὐτάρ in introducing new settings.

The first use—that is, introducing a sub-thread of discourse—is exemplified by a formulaic expression that is almost fixed in the first part, and includes a name and an epithet in the second part (most of the time, Athena is involved):

ἔνθ’ αὖτ’ ἄλλ’ ἐνόησ’ Ἑλένη Διὸς ἐκγεγαυῖα·

Odyssey 4.219; the primary speaking ‘I’ to the audience

Meanwhile, Helen, descended from Zeus, was making other plans. [134]

The use of ἔνθα at the very beginning of the sentence has a presentational function, as well (to which I will return when I discuss zooming-in techniques). The semantic element ἄλλος also occurs. By means of these expressions, different narrative threads (or, different visual shots) are set for resumed performance by the primary speaking ‘I’, while retaining their connection to the prior frame of reference. In the passage cited from Odyssey 4, the previous frame includes Menelaus’ invitation to the people present to eat and the start of the meal itself; then, a separate narrative thread follows what Helen—who is part of that people—does (line 219, which introduces her idea of putting a φάρμακον in the wine).

A clear instance of αὐτάρ used to shift to another thread of discourse distinct from the previous one is the very beginning of Odyssey 14, where the shift is also visual:

τώ γ’ ὣς βουλεύσαντε διέτμαγεν· ἡ μὲν ἔπειτα
ἐς Λακεδαίμονα δῖαν ἔβη μετὰ παῖδ’ Ὀδυσῆος.
Αὐτὰρ ὁ ἐκ λιμένος προσέβη τρηχεῖαν ἀταρπὸν
χῶρον ἀν’ ὑλήεντα δι’ ἄκριας, …

Odyssey 13.439–440 + 14.1–2; the primary speaking ‘I’ to the audience

The two [Athena and Odysseus] consulted and separated. She [Athena]
went to divine Lacedaemon, to follow Odysseus’ son.
As for the other [Odysseus], from the arbor he climbed up a rough path
in a wooded area, through the crests … {231|232}

One can easily visualize the two paths followed by the two individuals and, at the same time, one can process the first line of the new book as a narrative “move” that drops the thread “Athena” and follows the thread “Odysseus”—what he does, what happens to him—at least for a few sentences. [135]

ἔνθεν δ’ αὖ νήσοισιν ἐπιπροέηκε θοῇσιν,
ὁρμαίνων, ἤ κεν θάνατον φύγοι ἦ κεν ἁλοίη.
τὼ δ’ αὖτ’ ἐν κλισίῃ Ὀδυσεὺς καὶ δῖος ὑφορβὸς
δορπείτην …

Odyssey 15.299–302; the primary speaking ‘I’ to the audience [136]

Then, he [Telemachus] with the ships moved toward the islands,
pondering whether he would escape death, or be caught.
Meanwhile, the two in the shelter, Odysseus and the divine swineherd,
were having their supper …

In this example, the shift between two different situations—corresponding to two different visual settings—is presentationally committed to αὖτε, which divides but which is also a sign of continuity. In written code, the separation between different narrative threads is rendered by indentation, which is shown in more than one written edition of the Odyssey. [137]

A further use I mentioned is to re-join the main thread of discourse after a digression, an expansion, a simile, reported direct speech, or any kind of interruption (all of which correspond to different discourse acts). The meta-narrative meaning is “so, I was saying.” In this way, the upcoming topic is shown to be hierarchically (and typologically) equal to the one that was temporarily left behind. [138] In these cases, performative continuity appears to be almost explicit.

αἱ δ’ ἄλλαι μὰψ αὖραι ἐπιπνείουσι θάλασσαν·
αἳ δή τοι πίπτουσαι ἐς ἠεροειδέα πόντον,
πῆμα μέγα θνητοῖσι, κακῇ θυίουσιν ἀέλλῃ·
ἄλλοτε δ’ ἄλλαι ἄεισι διασκιδνᾶσί τε νῆας
ναύτας τε φθείρουσι· κακοῦ δ’ οὐ γίνεται ἀλκὴ {232|233}
ἀνδράσιν, οἳ κείνῃσι συνάντωνται κατὰ πόντον.
αἱ δ’ αὖ καὶ κατὰ γαῖαν ἀπείριτον ἀνθεμόεσσαν
ἔργ’ ἐρατὰ φθείρουσι χαμαιγενέων ἀνθρώπων
πιμπλεῖσαι κόνιός τε καὶ ἀργαλέου κολοσυρτοῦ.

Hesiod Theogony 872–880; the primary speaking ‘I’ to the audience

The other winds blow on the sea randomly:
after falling toward the cloud-covered dangerous sea
they rage with disastrous storms, a big calamity for mortals.
Sometimes other winds blow and scatter ships
and wreck seamen; there is no defense against this evil
for men who run into those ones on the dangerous sea.
Then, the other [winds that blow] on the limitless and flowery earth
destroy the lovely works of human beings, who are earth-born,
and fill them with dust and noisy, irritating tumult

In this passage from Hesiodic poetry, the winds that blow on the sea represent the former thread of discourse (872). A few more discourse acts describe their destructive effects (δή τοι, 873; δ’, 875; δ’ 876). Then, winds blowing on earth take up the latter thread of discourse (878). [139]

Another passage shows, with particular clarity, the meta-narrative intention to re-join a previous thread, along with the visual cue to go back with the mind’s eye to a previously mentioned subject:

306 Εὔμαι’, ἦ μάλα θαῦμα κύων ὅδε κεῖτ’ ἐνὶ κόπρῳ.
. . . . . . . . . . . . . . . . . . . . . . . . . . . .
311 τὸν δ’ ἀπαμειβόμενος προσέφης, Εὔμαιε συβῶτα·
καὶ λίην ἀνδρός γε κύων ὅδε τῆλε θανόντος.
. . . . . . . . . . . . . . . . . . . . . . . . . . . .
324 ὣς εἰπὼν εἰσῆλθε δόμους ἐῢ ναιετάοντας,
βῆ δ’ ἰθὺς μεγάροιο μετὰ μνηστῆρας ἀγαυούς.
Ἄργον δ’ αὖ κατὰ μοῖρ’ ἔλαβεν μέλανος θανάτοιο

Odyssey 17.306–326; Odysseus; Eumaeus; the primary speaking ‘I’ {233|234}

306 Eumaeus, what a spectacle! This dog lies on the dung-heap.
. . . . . . . . . . . . . . . . . . . . . . . . . . . .
311 Then, you, swineherd Eumaeus, replied to him:
“By all means, this is the dog of a man who died far away.”
. . . . . . . . . . . . . . . . . . . . . . . . . . . .
324 So he [Eumaeus] spoke, and went to the finely inhabited house.
He directly reached the proud suitors in the hall.
As for Argos, the destiny of black death seized him.

Re-joining the topic “Argos” and going back to the visualization of the dog are one and the same thing.

Finally, there is the marking of transitions to entirely new threads of discourse or to new narrative sections. αὐτάρ is quite frequently employed to convey this, even though αὖ and αὖτε are used as well. [140]

ἔνθ’ ἦ τοι εἷος μὲν ἐγὼ καὶ δῖος Ὀδυσσεὺς
οὔτε ποτ’ εἰν ἀγορῇ δίχ’ ἐβάζομεν οὔτ’ ἐνὶ βουλῇ,
ἀλλ’ ἕνα θυμὸν ἔχοντε νόῳ καὶ ἐπίφρονι βουλῇ
φραζόμεθ’ Ἀργείοισιν ὅπως ὄχ’ ἄριστα γένοιτο.
αὐτὰρ ἐπεὶ Πριάμοιο πόλιν διεπέρσαμεν αἰπήν,
[βῆμεν δ’ ἐν νήεσσι, θεὸς δ’ ἐκέδασσεν Ἀχαιούς,]
καὶ τότε δὴ Ζεὺς λυγρὸν ἐνὶ φρεσὶ μήδετο νόστον
Ἀργείοισ’ …

Odyssey 3.126–133; Nestor to Telemachus

Well, I tell you: for a while I and divine Odysseus
in assembly never expressed two different things, nor in council,
but having just one and the same mind in planning and in consulting
we showed to the Argives how the best events by far would happen.
So, after we sacked the steep-built citadel of Priam,
[we were going away in our ships, but a god scattered the Achaeans.]
At that moment, indeed, Zeus in his mind was meditating a mournful nostos
for the Argives … {234|235}

In these lines, Nestor first clarifies to Telemachus his own relationship with Odysseus; then, he starts telling about the mournful homecoming that the Argives had, once the Trojan war was over. αὐτάρ not only opens a new narrative section, but it also marks the beginning of a nostos tale. [141]

The next example shows that αὐτάρ may be used to mark narrative transitions, without implying any change of place:

πῦρ δὲ Μενοιτιάδης δαῖεν μέγα ἰσόθεος φώς.
αὐτὰρ επεὶ κατὰ πῦρ ἐκάη καὶ φλὸξ ἐμαράνθη …

Iliad 9.211–212; the primary speaking ‘I’ to the audience

And Menoetius’ son greatly kindled the fire, a man equal to a god.
So, when the fire burned out, and the blaze gradually died …

This description of the meal Patroclus is preparing for Odysseus, Ajax, and Phoenix—a sacrificial one [142] —does not include any spatial movement. Patroclus kindles the fire; then, he puts the meat on the embers of that same fire. Yet, the two images, as well as their respective temporal shifts, do belong to two different shots. As in a movie, a shot darkens and fades out, and a new one, departing from the same visual frame, fades in. What happens in the meantime is left unsaid, for the time-span of the fire’s blazing is, in fact, irrelevant. The new narrative section starts once the final flames are extinguished and the embers are ready for the barbecue to follow. In this case, the visual discontinuity arises from different “shots,” rather than from different spatial settings.

αὐ-discourse markers opening a book of one of the epics or an entire hymn might well reflect the same presentational intention—that is, to mark an entirely new narrative section. [143]

I draw a connection between the “specialization” of αὐτάρ in marking relatively clear-cut transitions and its prosodic status. Unlike αὖ and αὖτε, αὐτάρ is never accompanied by δέ. It is as if αὐτάρ is “strong” enough. If we observe the metrical position of αὐτάρ in its 838 instances, the vast majority occur either at the start of the line or after the bucolic caesura. Moreover, as a discourse marker, it always appears in sentence-initial position (the modern punctuation mark preceding each αὐτάρ may differ, but this is an arbitrary element, {235|236} after all [144] ). This suggests that a performative pause could have preceded the utterance of αὐτάρ, in line with the performative break characterizing the shift from a previous narrative section to the next one. In written code, this might be rendered by an indentation before each αὐτάρ. [145]

“In particular” singling out (zooming in)

Another presentational function of αὐ-discourse markers is the singling out of one individual (less frequently a group of individuals) from a more or less distinct plurality of characters. The primary speaking ‘I’ gives visual and thematic prominence to that individual. [146] What is common to this meta-narrative strategy and to the previous one (“on the other side” and “then”-shifts) is the idea of picking up one thread of narration and unfolding it. The common visual counterpart is a focusing on a partial section of the visual field. [147]

A first passage instantiates the singling out of a specific group of people. The speaker is Achilles, who is about to kill Lycaon:

νῦν δ’ οὐκ ἔσθ’ ὅς τις θάνατον φύγῃ ὅν κε θεός γε
Ἰλίου προπάροιθεν ἐμῇς ἐν χερσὶ βάλῃσι
καὶ πάντων Τρώων, περὶ δ’ αὖ Πριάμοιό γε παίδων.

Iliad 21.103–105; Achilles to Lycaon

Now there is not anyone who can escape death, whoever the god
might hurl into my hands before Ilium;
not one of all the Trojans and especially of Priam’s sons. [148] {236|237}

Another presentational αὖ spoken by the primary speaking ‘I’ and concerning a select group of people occurs at the beginning of the Catalogue of Ships:

488 πληθὺν δ’ οὐκ ἂν ἐγὼ μυθήσομαι οὐδ’ ὀνομήνω,
. . . . . . . . . . . . . . . . . . . . . . . . . . . .
493 ἀρχοὺς αὖ νηῶν ἐρέω νῆάς τε προπάσας. [149]

Iliad 2.488–493; the primary speaking ‘I’ to the audience

488 About the multitude of them [the chief men of the Danaeans], I could not give a speech, nor could I name them,
. . . . . . . . . . . . . . . . . . . . . . . . . . . .
493 I will single out the chiefs of the ships, and the ships in their multitude.

αὖ here occurs along with one of the very few explicit “I” marks referring to the primary narrator. I will return later to this significant fact.

On the whole, the Catalogue of Ships is an invaluable source of αὐ-discourse markers that zoom in on single individuals. Out of a total of forty-five entries concerning ships + leaders, twenty-three include instances of αὖ, αὖτε, or αὐτάρ that single out leaders (the remaining leaders are introduced by δέ [150] or lack any discourse marker). Here, I cite just one example:

Οἳ δ’ ἐκ Δουλιχίοιο Ἐχινάων θ’ ἱεράων
νήσων αἳ ναίουσι πέρην ἁλὸς Ἤλιδος ἄντα
τῶν αὖθ’ ἡγεμόνευε Μέγης ἀτάλαντος Ἄρηϊ
Φυλεΐδης …

Iliad 2.625–628; the primary speaking ‘I’ to the audience

Then, those who, coming from Dulichium and from the sacred Echinae
islands, which are situated on the other side of the sea, and face Elis,
of those [zoom in on him now] Meges was the chief, a man equal to Ares,
Phyleus’ son … {237|238}

The singling out of leaders or people higher in rank reminds us of one of the major uses of the intensifier αὐτός. I see a close relationship between αὐ-discourse markers and αὐτός in Homer, as far as isolating and centering are concerned. [151]

Further cases show the zooming-in technique clearly—namely whenever ἔνθ’ αὖ occurs: [152]

… οὐδ’ ἄρ’ Ἀχαιοὶ
ἀλκῆς ἐξελάθοντο, μένος δ’ ἰθὺς φέρον αὐτῶν.
ἔνθ’ αὖ Μηριόνης Τρώων ἕλεν ἄνδρα κορυστὴν
Λαόγονον …

Iliad 16.601–604; the primary speaking ‘I’ to the audience

… and the Achaeans
did not utterly forget the defence, but they drove their strength straight.
There he goes: Meriones caught a helmeted hero of the Trojans,
Laogonus …

My translation “There he goes” tries to render the pragmatic and visual force of ἔνθ’ αὖ, [153] which seems to give the performer and the audience a few instants to zoom in upon a specific hero. [154] I would underscore that the adverb ἔνθα can work as discourse marker on its own: I interpret its use as a way to zoom in on a specific point of the story—a mental place, a reconstructed spot along the oimē—to which the upcoming narration is attached. This important performative function of ἔνθα without any αὐ-discourse markers can be detected at Odyssey 1.11: once the proem is over, the primary speaking ‘I’ has to pick up the {238|239} starting-point point of the story (ἁμόθεν, 10): ἔνθ’ ἄλλοι μὲν πάντες, ὅσοι φύγον αἰπὺν ὄλεθρον, / οἴκοι ἔσαν, “So, all the remaining ones, who escaped utter death, were at home.” [155]

αὐτάρ singling out a prominent individual occurs very frequently, usually followed by the name of a hero or of a god (though that latter is much less frequent), for instance, αὐτὰρ Ἀχιλλεύς or αὐτὰρ Ἀθήνη. [156] The typical metrical position is after the bucolic caesura:

… ἀμφὶ δὲ κοῦροι
πρωθῆβαι ἵσταντο, δαήμονες ὀρχηθμοῖο,
πέπληγον δὲ χορὸν θεῖον ποσίν. αὐτὰρ Ὀδυσσεὺς
μαρμαρυγὰς θηεῖτο ποδῶν, θαύμαζε δὲ θυμῷ.

Odyssey 8.262–265; the primary speaking ‘I’ to the audience

… around him [Demodocus], young men
in their first youth stood, skilled in dancing,
and beat the divine dance with their feet. As for Odysseus,
he watched the twinkling of their feet, and was amazed in his heart.

The recipients are invited to zoom in on the main hero (who is “on the other side” of the dancers, as well) and to focus on his reaction and on his mental state (note that the prior focus is upon a plurality of individuals). [157] Equally frequent {239|240} is the nexus αὐτάρ + ὁ, which makes a pair with τὸν δ’ αὖτε; the pragmatic function is the same.

Finally, I also include in this group the strikingly numerous instances of αὐ-discourse markers zooming in on the speaking “I” (thus, almost always relating to characters). More precisely, they occur together with what I call an “I” mark—that is, any mark of the “I” deixis (Ich-Deixis): a singular first-person pronoun (ἐγώ, ἐμοί, ἐμέ), or a plural first-person pronoun (ἡμεῖς), [158] or an “I” deixis adverb of place (δεῦρο), [159] or an “I” deixis demonstrative (ὅδε), [160] or an “I” deixis adverb of time (νῦν). All of them mark the so-called zero-point of the utterance. [161]

The next passages include first-person pronouns and νῦν. Let us start with first-person pronouns:

ἀλλ’ αἰεὶ χαλεπὸς περὶ πάντων εἰς μνηστήρων
δμωσὶν Ὀδυσσῆος, περὶ δ’ αὖτ’ ἐμοί …

Odyssey 17.388–389; Eumaeus to Antinous

And you are always hostile beyond all the suitors,
with Odysseus’ servants; with me in particular …

αὖτε here serves to single out and isolate one individual—the speaking “I.” The very frequent phrase αὐτὰρ ἐγώ in the Iliad and in the Odyssey calls the attention of the internal audience to the speaking “I”; likewise, it invites the external audience to shift from what is previously mentioned to the character who is currently speaking; that character is the next topic of discourse and the subject to be visualized at that moment. [162] The closure of several Homeric Hymns shows that αὐ-discourse markers definitively work as both visual and narrative signposts:

αὐτὰρ ἐγὼ καὶ σεῖο καὶ ἄλλης μνήσομ’ ἀοιδῆς.

Homeric Hymn II to Demeter 495; the speaking ‘I’ to the goddess [163] {240|241}

As for me, I will remember you, as well as the rest of the song. [164]

The speaking “I” shifts from the final greetings and blessings addressed to the goddess to what comes next in the performance. He calls the attention of everyone to himself (αὐτὰρ ἐγώ means “zoom in on me now, look at me now”) because he represents the pivot subject of both the transition to the remaining part of the song and the performative continuation through the singing “I”—which is conveyed by all αὐ-discourse markers, as we have seen.

The forty-six occurrences of νῦν + au-discourse markers in the two poems (νῦν αὖ, νῦν δ’ αὖ, αὖ νῦν, [165] νῦν αὖτε, νῦν αὖτ’, and αὐτὰρ νῦν [166] ) tell us something pragmatically relevant. Indeed, as Bakker has already observed, the adverb νῦν in Homer can be followed not only by the indicative present tense but also by indicative aorist and by indicative future forms: “the deictic adverb nūn modifies not the event but its utterance.” [167] Different moods and tenses do actually accompany νῦν + αὐ-discourse markers. [168] Meanings may vary, as well. [169] However, there is, I submit, a common pragmatic feature characterizing all of them. Here, I select three illustrative instances of νῦν αὖτε:

ἐξ αὖ νῦν ἔφυγες θάνατον κύον· ἦ τέ τοι ἄγχι
ἦλθε κακόν· νῦν αὖτέ σ’ ἐρύσατο Φοῖβος Ἀπόλλων

Iliad 11.362–363; Diomedes to Hector = 20.449–450; Achilles to Hector

In this very moment you escaped death, dog. Well, actually the evil
came near you. But Phoebus Apollo saved you. {241|242}
νῦν δὲ δὴ ἐγγύθι μοι θάνατος κακός, …
… ἦ γάρ ῥα πάλαι τό γε φίλτερον ἦεν
Ζηνί τε καὶ Διὸς υἷι ἑκηβόλῳ, οἵ με πάρος γε
πρόφρονες εἰρύατο· νῦν αὖτέ με μοῖρα κιχάνει.

Iliad 22.300–303; Hector to Achilles

Now, see, wicked death is close to me, …
… For, well, for a long time it [having an escape] was pleasing
to Zeus and Zeus’ son who shoots far; previously, they
were glad to defend me. But now doom is finding me. [170]
οὗτος μὲν δὴ ἄεθλος ἀάατος ἐκτετέλεσται·
νῦν αὖτε σκοπὸν ἄλλον, ὃν οὔ πώ τις βάλεν ἀνήρ,
εἴσομαι, …

Odyssey 22.5–7; Odysseus to the suitors

This prize-contest has reached its accomplishment without trickery.
And now I see the remaining target, which no man has ever
hit …

Tenses vary: indicative aorist, present, and future appear respectively. Further, the contextual implications are not always the same: while in the first two examples a frustration of expectations is pointed out (evil came near, nevertheless Apollo saved Hector; Zeus and Apollo were formerly defending Hector, nevertheless death has reached him [171] ), in the third an allusion to a seemingly parallel repetition is made (Odysseus says he will shoot again, even though the upcoming strike is going to have a quite different outcome). Yet, in all three passages αὖτε marks the assertion of the utterance: it expresses the speaker’s acknowledgment of an event he senses at the moment of the utterance, despite the span of time involved. The pragmatic function of αὖτε, then, is to reinforce νῦν (which supports Bakker’s analysis). αὖτε, along with νῦν, zooms in on the instant in which a certain statement about a certain event is set forth. At the same time, αὖτε signals the speaker’s awareness and re-cognition of a reality he is experiencing in that moment. As such, it marks the speaker’s attitude towards {242|243} what is said. These latter cases constitute, in fact, an intermediate stage between the presentational and the interactional functions of αὐ-discourse markers.

Interactional functions of αὖ, αὖτε, and αὐτάρ

Re-cognition and emotional discontinuity

Bazzanella (2004:64, quoting Magri 1999) affirms that the emotions in recent scholarship have gained a full “citizenship in the conceptual system of mental states, representations and reasons.” Emotions are grounded on several components, which Elster (1999:56) lists as follows: “bodily arousal, physiological expressions, cognitive antecedents, intentional objects, valence (pleasure-pain), and action tendencies.” The diversity of levels at which emotions can be expressed—linguistic, paralinguistic, and extralinguistic—along with the complexity of their description should not prevent scholars of literary texts from excluding a priori an investigation into the use of words that imply an emotional effect, whether overt or covert. Some examples of linguistic devices employed to convey emotions are: pitch (a phonetic device), exclamation marks (graphic), prosody (phonological), diminutives (morphological), “admire verbs,” such as “enjoy” or “hate,” (lexical), marked word order (syntactical), variations in register (sociolinguistic), the use of dispositio (textual), and discourse markers and interjections (pragmatic). [172] Except for lexical evidence [173] (with such familiar terms as μένος, ἀχνύμενος, νηλέϊ θυμῷ, θάμβησε, τέρποντο) and, to some extent, for syntactical evidence (word order), all of the devices mentioned above are quite difficult to critically ascertain within the Homeric poems. One major feature that deals with emotions and that we have already noted is the encoding in Homeric language of non-verbal behavior, which includes such expressions of emotion as bursting into tears, going slack (at the knees), and so on (see Lateiner 1995). However, this approach does not wholly clarify what feelings a particular utterance is meant to convey when unambiguous lexical devices are missing and, most of all, when words are uttered by the performer and not by characters. [174] Sometimes, only context and cotext help us to guess {243|244} the speaker’s emotional attitude toward a certain fact. Let us think, for example, of every time a character reacts to his/her interlocutor’s words by saying “what sort of word escaped the barrier of your teeth,” ending either with an exclamation mark or a question mark. [175] The point I would like to underscore here regards just one part of the big picture concerning Homeric linguistic devices used to convey emotions: Homeric discourse markers, which are as much pragmatic devices as interjections are, may help to infer the emotions that are conveyed with the utterance. This is not a new idea, as it is commonly acknowledged that particles may indicate emotional moods (for example, Denniston 1996:xxxvii–xxxviii; and see above in this chapter). Many scholars would agree that utterances such as ἀμφὶ δ’ ἄρα σφι λιάζετο κῦμα θαλάσσης “And about them [Iris and Thetis] the sea current withdrew” (Iliad 24.96), [176] notably enriched by the particle ἄρα, signal the speaker’s emotional involvement in telling of the amazing detail. [177] Nevertheless, it may be the case that many other particles, whenever they work as discourse markers, give specific hints to the recipients on how to track the emotional involvement of the speaking “I.” My contribution in this direction concerns αὐ-discourse markers. The term “emotional discontinuity” indicates that the emotional charge of some utterances—signaled by αὐ-discourse markers—represents a discontinuity with respect to the narrative register preceding those utterances. This may occur in the switch from third-person narration to reported direct speeches, as well as within the third-person narrative itself.

Let us start with αὐ-discourse markers occurring in interrogative acts. αὖ, αὖτε, and αὐτάρ in those cases mark the speaker’s feelings, while wondering at what is happening before his or her own eyes. {244|245}

τίπτ’ αὖτ’, ὦ δύστηνε, λιπὼν φάος ἠελίοιο
ἤλυθες, ὄφρα ἴδῃ νέκυας καὶ ἀτερπέα χῶρον;

Odyssey 11.93–94; Teiresias’ soul to Odysseus

Oh my, why, o wrecked man, have you left the light of the sun
and come here, to see the dead and an unpleasant land?

Teiresias expresses wonder, surprise, and maybe even some disappointment, as he recognizes Laertes’ son (see 92) as a live individual visiting the land of the dead.

A sense of impatience and irritation seems to be conveyed by αὐτάρ in the following question:

οὐκ ἀγαπᾷς, ὃ ἕκηλος ὑπερφιάλοισι μεθ’ ἡμῖν
δαίνυσαι οὐδέ τι δαιτὸς ἀμέρδεαι, αὐτὰρ ἀκούεις
μύθων ἡμετέρων καὶ ῥήσιος; οὐδέ τις ἄλλος
ἡμετέρων μύθων ξεῖνος καὶ πτωχὸς ἀκούει.

Odyssey 21.289–292; Antinous to the beggar

Don’t you like that you, untroubled among us, who are heedless,
eat and are not deprived of any portion, are even listening to
our speeches and conversations? No other guest and no other beggar
listens to our speeches.

Antinous makes his indignation toward the beggar quite explicit: he is even allowed to listen to their conversations—is that not enough? I have selected these two citations as representative instances of αὐτάρ conveying emotional excitement, whether for good or ill. [178] Kroon’s analysis of Latin autem occurring in questions [179] helps us interpret the interactional aspects of a discourse marker whose primary function is to convey distinctiveness. Kroon disagrees with the common assumption in handbooks and grammars of two separate autems, one working as a connective/adversative adverb, and the other conveying specific emotions, such as indignation, surprise, admiration, and fear. Kroon acknowledges the emotional charge of several interrogative acts, including autem, but she encompasses that charge within the overarching pragmatic meaning of {245|246} autem—that is, to mark prominence. More specifically, Kroon speaks of prominence channeled through either counter-expectancy [180] or “emphatic identification,” which enhances the “emotional force” of autem. [181] I report here one of Kroon’s examples from Latin:

TH. Ius dicis
EP. Me decet
TH. Iam tu autem nobis praeturam geris?

Plautus Epidicus 25; Thesprio and Epidicus

TH. You speak like a judge
EP. The proper thing for me!
TH. What? Do you already hold the praetorship?

Thesprio’s surprise and wonder (and perhaps his scorn) arises from his recognition of a judgemental attitude in his companion. In this specific instance, “emphatic identification” matches counter-expectancy, as Thesprio presumably did not expect the haughty response of Epidicus. I submit that in the Homeric poems interactional values of αὐ-discourse markers concern either pieces of information that are contrary to the listener’s expectation or pieces of information with which the speaker emphatically identifies, or both. The emotional force of discourse markers basically rests on the speaker’s involvement, as he or she acknowledges facts to be wondered at. [182] The νῦν αὖτε passages that I have commented upon above are tinged with these interactional and emotional cues. Now, I am going to show first some instances of emotional discontinuity, as it relates to counter-expectancy. In Iliad VIII, Nestor is terrified by Zeus’ lightning while battling against Hector, and his horses are terrified as well. Therefore, he suggests to Diomedes:

Τυδεΐδη ἄγε δ’ αὖτε φόβον δ’ ἔχε μώνυχας ἵππους.
ἦ οὐ γιγνώσκεις ὅ τοι ἐκ Διὸς οὐχ ἕπετ’ ἀλκή;

Iliad 8.139–140; Nestor to Diomedes {246|247}

Son of Tydeus, come on [now I know], steer the solid-hoofed horses to flight.
Don’t you realize that Zeus’ defense doesn’t accompany you?

Nestor is panic-stricken, but his fear leads him to sense what is going on, to acknowledge what is happening, and to express that what he sees has to be done in the very moment. How can we dismiss the emotional implication in Nestor’s recognition of the necessary action? To recall the already cited Odyssean passage on Argos’ death (Ἄργον δ’ αὖ κατὰ μοῖρ’ ἔλαβεν μέλανος θανάτοιο “As for Argos, the destiny of black death seized him,” Odyssey 18.326), how can we dismiss the emotional implication in the primary speaker’s masterful visualization of the dog’s death upon seeing his master? As I have observed before, the reality and complexity of literary language is that it contains potentially overlapping meanings and functions; thus, the same αὖ can convey a narrative and visual shift “back to the dog,” as well as the contrary-to-expectation announcement of Argos’ death. A further component is the speaker’s emotional involvement—I call it empathy—with the dog. Analogously, I highlight the polyvalent effectiveness of αὖ placed near to one of the instances of κεῖνος by which Telemachus refers to Odysseus in Odyssey 3: “for we know of all the remaining ones who were battling in Troy; but of that one, instead, Cronus’ son has made the death something impossible to know about (κείνου δ’ αὖ καὶ ὄλεθρον ἀπευθέα θῆκε Κρονίων),” Odyssey 3.86–88. [183] αὖ here performs many functions: it singles Odysseus out from the indistinct plurality of “the others who battled in Troy”; it makes prominent the information that “even Odysseus’ death is unknown to us,” because it is contrary to Telemachus’ and to his listeners’ expectations; finally, it conveys the speaker’s empathy toward the subject involved. [184]

Some instances of αὖ mark relevant pieces of information that are contrary to the addressee’s expectations. For example, one of the suitors mistreats Eumaeus by asking him where he is going with the bow; he asserts: τάχ’ αὖ σ’ ἐφ’ ὕεσσι κύνες ταχέες κατέδονται / οἶον ἀπ’ ἀνθρώπων, “Quite soon the fast dogs will devour you beside the pigs, abandoned by human beings” (Odyssey 21.363–364). αὖ does not organize the narrative sequences, nor does it zoom in on a specific subject; rather, it has an interactional function. Specifically, it marks the content of the utterance as something the speaker supposedly knows and as something the addressee probably does not expect. Other utterances express, {247|248} conversely, the speaker’s surprise at events that he or she recognizes as unexpected. For example, when Odysseus asks Circe whether he could fight against Scylla, Circe’s first reaction is as follows: “Hard man, [that really surprises me] you care for belligerent actions and for toil, and you don’t make way for the immortal gods?” (σχέτλιε, καὶ δὴ αὖ τοι πολεμήϊα ἔργα μέμηλε / καὶ πόνος, οὐδὲ θεοῖσιν ὑπείξεαι ἀθανάτοισιν;, Odyssey 12.116–117). αὖ—here reinforced by καὶ δή—conveys Circe’s acknowledgment of, and amazement at, Odysseus’ stubborn fighting spirit. [185]

As for what Kroon calls “emphatic identification,” I cite a Hesiodic passage first and two Odyssean passages to follow that convey not only the speaker’s excitement in recognizing a certain reality, but that also channel the speaker’s empathy towards the subjects involved:

εἰ γάρ τις καὶ πένθος ἔχων νεοκηδέι θυμῷ
ἄζηται κραδίην ἀκαχήμενος, αὐτὰρ ἀοιδὸς
Μουσάων θεράπων κλεῖα προτέρων ἀνθρώπων
ὑμνήσει μάκαράς τε θεοὺς οἳ Ὄλυμπον ἔχουσιν,
αἶψ’ ὅ γε δυσφροσυνέων ἐπιλήθεται οὐδέ τι κηδέων
μέμνηται·

Hesiod Theogony 98–103; the primary speaking ‘I’ to the audience

For if one with sorrow inside, where recent cares are felt,
sighs and has grief in his heart, there he is, the singer:
servant of the Muses, he will sing the glories of prior people,
and the blessed gods that possess Olympus.
Immediately he [the man with grief] forgets the anxieties and
does not remember any of his cares.

“Blessed is the one whom the Muses love” (95–96): the singer has been just introduced. The communicative function of αὐτάρ (99) is to zoom in on this already mentioned figure by emphatically identifying him: the ἀοιδός is the one who makes human beings forget their sorrows by singing epic deeds and theogonies. [186] Emphasis comes from the emotional charge of such an identification, and also—in a subtle self-referential fashion—from empathy, from emotional nearness to the ἀοιδός. {248|249}

Recognizing the core of a person is what αὖ, αὖτε, and αὐτάρ convey in some polysemic utterances concerning the true identity of Odysseus, as well. The night after her first meeting with Odysseus, Penelope, in her wretched state, complains that the gods have sent her evil dreams (ὀνείρατα … κακά, Odyssey 20.87):

τῇδε γὰρ αὖ μοι νυκτὶ παρέδραθεν εἴκελος αὐτῷ,
τοῖος ἐὼν, οἷος ᾖεν ἅμα στρατῷ· αὐτὰρ ἐμὸν κῆρ
χαῖρ’, ἐπεὶ οὐκ ἐφάμην ὄναρ ἔμμεναι, ἀλλ’ ὕπαρ ἤδη

Odyssey 20.88–90; Penelope to herself

For tonight, really, someone resembling the true one has slept beside me,
as he was when he joined the army. And so, my heart
rejoiced, as I did not say that it was a dream, but the reality, indeed.

αὖ and αὐτάρ can be seen as the subjective marks of Penelope’s attitude towards what she describes. She “really” felt that someone like the real Odysseus (αὐτῷ, 88) was sleeping not far from her. The omniscient audience knows, of course, that this is quite true (Odysseus was laying in the forecourt, 20.1). However, the two discourse makers become polysemic, as we think of two possible layers of communication (see chapter 2): Penelope the unaware wife had only the impression that Odysseus was sleeping not far from her, whereas Penelope the aware character re-cognized what she was experiencing, she did realize that Odysseus was near. Of course, the primary speaking ‘I’ leaves the ultimate interpretation open. Analogously, αὖτε spoken by the eagle in Penelope’s dream (19.536–553) is twofold. The eagle asserts: “I was an eagle, a bird of omen for you, beforehand; now it’s your husband, I have arrived” (ἐγὼ δέ τοι αἰετὸς ὄρνις / ἦα πάρος, νῦν αὖτε τεὸς πόσις εἰλήλουθα, Odyssey 19.548–549). αὖτε works both as a marker of counter-expectancy (“but now”) and as a marker of emphatic identification (“now really”). The almost counterfactual assertion by the eagle “I am your husband” is balanced by the true fact that in the moment at which Penelope speaks these words in front of Odysseus, Odysseus “now really” has come home. The eagle empathizes with the beggar in front of the woman. Finally, I cite an instance from book fifteen. Once Eumaeus’ tale is over (Odyssey 15.390–484), Odysseus lets the swineherd know how impressed he was by the tale and concludes that Eumaeus, after much suffering, got a good life (ἀγαθὸν βίον, 491); he continues:

… αὐτὰρ ἐγώ γε
πολλὰ βροτῶν ἐπὶ ἄστε’ ἀλώμενος ἐνθάδ’ ἱκάνω.

Odyssey 15.491–492; the beggar to Eumaeus {249|250}

… As for myself,
I come to this place after much wandering through cities of mortals.

Who amongst the there-and-then listeners, as well as amongst here-and-now readers, would not take line 492 as a signature identifying Odysseus? I submit that the whole clause is one of those “unnecessary” sentences that are inserted at a certain point of the performance for exclusively playful reasons. [187] In other words, I am arguing that the primary speaking ‘I’ deliberately has Odysseus speak words that enhance layering. Through these words, the beggar empathizes with Eumaeus, who had many troubles, and through the same words Odysseus also quite explicitly reveals to his loyal ally that he has also came to the same place as Eumaeus—that he has, indeed, arrived. And the keyword for this layering is αὐτάρ, along with ἐγώ γε: αὐτὰρ ἐγώ γε presentationally marks the visual and narrative switch from “you”-Eumaeus to “I”-Odysseus; it also zooms in on Odysseus. At the interactional level, it emphatically identifies the subject of the sentence. It is the signpost for the re-cognition of the identity of Odysseus’ self. The variation in emotional intensity that underlies the whole clause is due to the empathy between the speaker of one layer (the beggar) and the speaker of the other layer (the master).

Some general remarks, in conclusion. More of the passages that I have commented upon in this latter section come from the Odyssey than the Iliad. Is this a sign that the interactional functions of αὐ-discourse markers are a subsequent development of the presentational functions? The latter abound in the Iliad; are they, then, more archaic? This is difficult to determine; the larger employment of interactional functions in the Odyssey may simply be due to different stylistic choices, such as a deeper introspection in rendering the personality of characters. It is, however, unquestionable that interactional functions occur much more in the characters’ utterances than in those of the primary speaking ‘I’. A last note concerns empathy. I draw a connection between empathy in identification through αὐ-discourse markers and empathy in identification through αὐτός. In chapter 3, I have shown that the intensifier αὐτός can be used as an empathic anaphora. In the current chapter, I have analyzed cases in which empathy constitutes a significant part of the emotional load. Also, both αὐ-discourse markers and αὐτός are frequently associated with “I” marks. I would propose that these “technical” devices to organize reference in the text may signal subjectivity, as well. As such, they give the recipients access to the mental states of the speaking “I” (whoever he or she is). Finally, if the speaking “I” is the primary speaking ‘I’—that is, the performer—the access concerns not {250|251} only his empathic attitudes but also the narrative strategies that are controlled by him.

In lyric, elegiac, and iambic poetry

My analysis now turns to lyric, elegiac, and iambic poetry. The interest in the use of αὐ-discourse markers in this quantitatively minor, but qualitatively significant, body of early Greek poetry stems from some basic questions: Are the pragmatic functions similar to those of early epic? What is the balance between presentational and interactional functions? Is it possible to draw genre-specific connections between their occurrences and specific stylistic features? Are there functions of αὐ-discourse markers that pertain exclusively to lyric poetry?

As far as frequency and distribution are concerned, [188] iambic poetry rarely displays αὐ-discourse markers. I have recorded just three instances of αὖτε: Ananius fr. 5.6 (αὖθ’), Hipponax fr. 122 (δηὖτε, commented upon below) and Archilochus fr. 88 (δηὖτε, also commented upon below).

First of all, some visual functions seem to be confirmed. Pindar, for example, employs αὖ not only to mark a parallel focus, but also to express spatial discreteness, even if that space is metaphoric:

… αἵ γε μὲν ἀνδρῶν
πόλλ’ ἄνω, τὰ δ’ αὖ κάτω
ψεύδη μεταμώνια τάμνοισαι κυλίνδοντ’ ἐλπίδες

Pindar Olympian 12.5–6a; the primary speaking ‘I’ to the audience

… Men’s expectations
several times roll up, they roll down
as they carve idle untruths.

Another interesting instance of αὖ signaling what is spatially “on the other side” appears in a fragment of Stesichorus:

ἔνθεν μὲν Λοκρ[οὶ
ἱζάνον αἰχματαὶ [
τέκνα φιλα[ἐρί-
ηρες Ἀχαιοὶ [
καὶ ὑπερθύμοι [ {251|252}
θ’ ἱαρὰν Βοιωτίδ[α ν]αίον [
χθόνα πυροφόρ[ον. ]
ἔνθεν δ’ αὖ Δρύοπ[ές] τε κα[ὶ
λ̣οι μενεχάρμα[ι

Stesichorus fr. 45, col. ii lines 1 to 9; the primary speaking ‘I’?

On one side there were the Locrians
spearmen made to sit …
dear sons …
faithful Achaeans …
and proud …
they were inhabiting sacred Boeotia …
wheat-bearing land.
On the other side there were the Dryopians and …
committed to battle …

Barrett (1972:117–118) hypothesizes that the two counter-posed groups are about to fight not against each other, but against the Calydonian boar. [189] What interests me is that the position taken up by the two coalitions is singled out by means of ἔνθεν μέν and ἔνθεν δ’ αὖ. The identification of a parallel focus matches a zooming-in effect. [190]

Sappho fr. 44 includes a visual mark conveyed by αὖ, which is reinforced by the adverb χωρίς:

… ἐ̣π̣[έ]βαινε δὲ παῖς ὄχλος
γυναίκων τ’ ἄμα παρθενίκα[ν] τ..[..] σφύρων,
χῶρις δ’ αὖ Περάμοιο θυγ[α]τρεσ[
ἴππ[οις] δ’ ἄνδρες ὔπαγον ὐπ’ ἀρ̣[ματ

Sappho fr. 44.14–17; the primary speaking ‘I’ to the audience

… the whole crowd
of women and maidens was getting on board
Then, apart from them, Priam’s daughters …
men brought horses under the yoke of the chariots … {252|253}

αὖ clearly singles out Priam’s daughters from the rest of the women.

Further significant examples of αὐ-discourse markers at the presentational level demonstrate parallel repetition—that is, the presentation of hierarchically equal entries. [191] A presentational αὐτάρ occurs in a very famous fragment of Alcaeus:

κάββαλλε τὸν χείμων’, ἐπὶ μὲν τίθεις
πῦρ ἐν δὲ κέρναις οἶνον ἀφειδέως
μέλιχρον, αὐτὰρ ἀμφὶ κόρσαι
μόλθακον ἀμφι<> γνόφαλλον

Alcaeus fr. 338.5–8; the primary speaking ‘I’ to the audience

Defy the storm, keep burning the fire,
put wine unsparingly in the pottery,
honey-colored. And around your forehead
a soft … cushion …

The presentational value of αὐτάρ here consists in expressing the simultaneity of the actions being suggested, as well as in implying a visual shift from what presumably surrounds the “you” being addressed (the fire and the wine) to the “you” that is a physical person. [192]

Theognis’ elegies include some interesting presentational instances of αὐτάρ and αὖ. In the following passage, αὐτάρ seemingly marks a parallel focus, though it more essentially effects a singling-out effect:

Ζεὺς μὲν τῆσδε πόληος ὑπειρέχοι αἰθέρι ναίων
αἰεὶ δεξιτερὴν χεῖρ’ ἐπ’ ἀπημοσύνηι,
ἄλλοι τ’ ἀθάνατοι μάκαρες θεοί· αὐτὰρ Ἀπόλλων
ὀρθώσαι γλῶσσαν καὶ νόον ἡμέτερον. {253|254}
φόρμιγξ δ’ αὖ φθέγγοιθ’ ἱερὸν μέλος ἠδὲ καὶ αὐλός·
ἡμεῖς δὲ σπονδὰς θεοῖσιν ἀρεσσάμενοι
πίνωμεν …

Theognis 1.756–762; the primary speaking ‘I’ to the audience

May Zeus hold his right hand over this city, he who dwells in the sky,
so that it may be safe from harm,
and (so do) the rest of the blessed immortal gods. As for Apollo,
may he straighten our tongue and our mind.
And so, may the lyre and the reed pipe make the sacred song resound;
and may we drink, after satisfying the gods with libations …

Apollo is singled out from the previously mentioned gods; and so, too, the ritual safeguards that follow. A new and a different communicative intention underlies the utterance starting with αὐτὰρ Ἀπόλλων “as for Apollo”—namely, to stress the exclusive relationship linking the god and the “we,” as they are invited to start playing and singing. [193] Then, αὖ marks a series of further discourse acts expanding the list of the desirable sacred activities (singing a sacred song accompanied by the lyre and by the reed; offering libations; drinking). Analogously, in another passage αὐτάρ signals a topic switch and spotlights the musical activity that is going to involve the speaking “I”:

Ἀλλὰ λόγον μὲν τοῦτον ἐάσομεν, αὐτὰρ ἐμοὶ σύ
αὔλει, καὶ Μουσῶν μνησόμεθ’ ἀμφότεροι.

Theognis 1.1054–1055; the primary speaking ‘I’ to a “you”

But let’s drop this discourse. So, play
the reed pipe for me, and let’s both remember the Muses.

Finally, a striking presentational αὖτε is recorded in the quotation of the two incipits of Stesichorus’ Palinodes:

… διτταὶ γάρ εἰσι πα-
λινωιδ<ίαι δια>λλάττουσαι, καὶ ἔ-
στιν ἡ μὲν ἀρχή· δεῦρ’ αὖ-
τε θεὰ φιλόμολπε, τῆς δέ· {254|255}
χρυσόπτερε παρθένε, ὡς
ἀνέγραψε Χαμαιλέων· …

Stesichorus fr. 16.7–12 (= P.Oxy. 2506, fr. 26, col. i; Stesichorus’ words underlined)

… For two are the palinodes,
and they differ from each other;
the beginning of one of them is:
“Here again, o goddess, lover of danced songs”;
the other one [starts]:
“O maiden of golden wings,”
as Chamaeleon recorded …

In accord with my analysis, the meaning “again” of αὖτε need not be intended at the propositional level. αὖτε might be a presentational adverb instead; indeed, it occurs in first sentence position, anticipating the vocative. Its pragmatic function might be the marking of the very beginning of the song. In fact, δεῦρο conveys the relevance of a specific hic, which is the hic of the performance that is going to start. [194]

As for the interactional functions of αὖ, αὖτε, and αὐτάρ, I first cite a question including δηὖτε:

Ἐρξίη, πῆι δηὖτ’ ἄνολβος ἁθροΐζεται στρατός;

Archilochus fr. 88; a speaking “I” to Erxies

So, Erxies, wherefore is the unfortunate army gathering, then?

On this translation of the isolated fragment (a trochaic tetrameter), δηὖτε seems to express the recurrence of the narrated event: the army is assembling “again,” “one more time.” Such an interpretation presupposes an exclusively propositional meaning of δηὖτε. A discourse-oriented analysis produces an alternative interpretation: not differently from the Homeric occurrences of αὖ, αὖτε, and αὐτάρ in questions, the function of δηὖτε can here be interactional, conveying wonder, or impatience, or even disappointment by the speaking “I” in asking about the army. A presentational meaning might also be suggested: the speaker resumes a previous topic: “So, to what end is the helpless army {255|256} assembling, then?” The lack of surrounding lines, however, makes this reading wholly conjectural.

I would interpret a further instance of αὖ as interactional in the following fragment of Simonides, which, in fact, resembles a Homeric phrase I have already analyzed: νῦν αὖ θάνατος καὶ μοῖρα κιχάνει (Iliad 18.478 = 672 = 22.436):

ὁ δ’ αὖ θάνατος κίχε καὶ τὸν φυγόμαχον.

Simonides fr. 19; a speaking “I” to the audience (?)

Death reaches even the one who escaped battle. [195]

The preceding lacuna prevents us from establishing which cohesive relationship αὖ might have underscored with respect to the previous topic. Nevertheless, the comparison with the Iliad enhances an interactional interpretation: αὖ might signal the sorrowful acknowledgment of a well-known (or experienced) phenomenon, which is the ineluctability of death; it might mark the recognition of that truth “once again.”

What the speaking “I” perceives and re-perceives is consistently marked by αὖτε, or, more frequently, δηὖτε, in a series of incipits, most of which have Eros as the topic. [196] Here, I cite some of them consecutively.

Ἔρος αὖτέ με κυανέοισιν ὑπὸ
βλεφάροις τακέρ’ ὄμμασι δερκόμενος
κηλήμασι παντοδαποῖς ἐς ἄπει-
ρα δίκτυα Κύπριδος ἐσβάλλει·

Ibycus fr. 6.1–4; the primary speaking ‘I’ to the audience

Once again, Love who gazes at me meltingly
with his eyes under the dark eyelids,
with all sorts of invites pushes me towards
the limitless nets of Cypris.
Ἔρως με δηὖτε Κύπριδος Ϝέκατι
γλυκὺς κατείβων καρδίαν ἰαίνει.

Alcman fr. 59a; the primary speaking ‘I’ to the audience

Again, Eros, on account of Cypris,
sweet, floods and melts the heart. {256|257}
Ἔρος δηὖτέ μ’ ὀ λυσιμέλης δόνει,
γλυκύπικρον ἀμάχανον ὄρπετον

Sappho fr. 130; the primary speaking ‘I’ to the audience

Once again, Eros who makes limbs loose shakes me,
bitter-sweet, ungovernable, the creeping one.
ἐρέω τε δηὖτε κοὐκ ἐρέω
καὶ μαίνομαι κοὐ μαίνομαι.

Anacreon fr. 83; the primary speaking ‘I’ to the audience

Once again, I am in love, and I am not in love,
and I am getting mad, and I am not getting mad.

Carson (1986:118–119) sees in δηὖτε two combined significations: namely, “a lively perception in the present moment,” conveyed by δή, and “a pattern of repeated actions,” conveyed by αὖτε. Aloni (1997:217) identifies δηὖτε as a thematic element of archaic lyric (“un vero elemento tematico della lirica arcaica”) whose function is to recall the recurrence of a single event within a tradition of songs, by virtue of the repeatability of a certain reality (“destinato a inserire il singolo evento attuale all’interno di un panorama tradizionale, dove la realtà si ripete identica”). Nagy (1996b:100–101) cites Carson and underscores that each time a certain event is re-enacted, the “I” perceives it anew, as each performance is a different experience. All of this is most relevant to the present investigation. δηὖτε, in fact, signals the perceived repetition of the whole, including the re-enacted event (for example, Eros pervading the singing “I”) and the utterances that re-enact it. δηὖτε does not simply refer to states of affairs (for instance, the first line of Anacreon fr. 83 does not simply say “I am in love for a second time, and I am not in love for a second time”), but it is also an interactional mark: it expresses the strong emotions qualifying the recognition of an event that is going to be re-experienced (“Here it is, it’s happening once again: I am in love and I am not in love at the same time”). Furthermore, δηὖτε is a presentational mark that refers to the performative level of repetition: the utterances that are going to be performed in the present moment repeat previous utterances of previous performances devoted to the same topic. This is a self-reference at the level of discourse, a meta-discursive function of literary significance: the speaking “I” deliberately binds the hic et nunc of the song to a chain of similar songs. The principle of cognitive anaphora in this case exceeds the borders of one text; it is an anaphoric reference to other realizations of other texts sharing the topic and/or the mode of discourse. {257|258}

Hipponax seems to be aware of all of these implications of lyric δηὖτε and seems to make allusive fun of them in the following fragment:

Μητροτίμωι δηὖτέ με χρὴ τῶι σκότωι δικάζεσθαι

Hipponax fr. 122; the primary speaking ‘I’ to the audience

Once again, I have to take Metrotimus, the obscure, to court.

Here, the prepositional content “once again” concords, in a masterfully irreverent way, with the speaker’s recognition of an already experienced event. The action described is codified as something that may repeatedly occur, but it has nothing to do with ritual or psychological events. It involves the speaking “I,” but it does not seem to provide any τέρψις. The speaker just has to take somebody to court, and the iambic choice is to communicate that by parodying lyric δηὖτε at the propositional, interactional, and presentational level.

Conclusion: distinctiveness and discontinuity within performative continuity

αὖ, αὖτε, and αὐτάρ have been chosen for study because their functions and meanings have often been underestimated (αὖ and αὖτε are absent from Denniston’s survey of Greek particles). [197] Further, their canonical descriptions show a mix of indications that indirectly invites to a closer investigation of their uses. On the whole, however, my analysis is intended to exemplify a new approach to the study of particles and of sentence adverbials. Indeed, I am proposing to abandon the idea of confining some words to the ill-defined label “particles” in favor of the idea of discourse markers, as illuminated by contemporary research in pragmatic and in discourse analysis. Even though findings may vary according to different scholarly traditions and standpoints, the underlying “philosophy” is helpful: discourse markers are positively described as single words and lexical phrases whose relevance is to signal how to process the immediately upcoming utterances or series of utterances. Their functions may range from the presentational to the interactional. At the presentational level, discourse markers typically signal transitions between sections; at the interactional level, they typically mark the speaker’s attitude toward what is said (either in front of listeners or in exchanges with other interlocutors). As their nature is sequential, [198] they tend to {258|259} have no independent meaning. Rather, their meaning is to be understood within the communicative intention underlying the utterances that contain them. From this perspective, ancient Greek particles turn out to be helpful guides in the comprehension of texts, rather than ornamental parts of discourse. Their optional syntactic standing is opposed to their crucial pragmatic standing.

A major point concerns the sentence initial position that is shared by discourse markers and by ancient Greek particles or sentence adverbials working as discourse markers. In this respect, I underscore that the effectiveness of αὖ, αὖτε, and αὐτάρ as segment transition markers is due to their position bridging the end of the previous clause and the beginning of the next. [199] They affect the listeners’ processing in that they provide what Redeker (2006:349) calls “attentional cues”: they cue the listeners (and readers, as well) to the kind of transition, so that the receivers may guess what to expect from the next discourse unit. The notable study by De Jong and Nünlist (2004) about the narrator’s standpoints in the Homeric epic telling does not focus on specific words used to convey different strategies of visualization; yet the authors point out that transitions from one scene to the next are usually “smooth” (2004:73). I claim that such smoothness is often the result of the use of discourse markers signaling different visual switches, which I call visual discontinuities. A related major point of my analysis is that αὖ, αὖτε, and αὐτάρ primarily—and perhaps also originally—represent visual cues that, in fact, also match narrative cues. My results fully confirm Bakker’s claim that “narration is description.” The marking of visual and narrative discontinuities perfectly harmonizes with Kroon’s definition of the basic meaning of Latin autem (1995:270) as “distinctiveness.” αὖ, αὖτε, and αὐτάρ very often help in visualizing the next subject or the next setting as a distinct one. From time to time, they interactionally add a sense of re-cognition or of a new cognition—by their speaker—of a special moment of the story, or they convey the speaker’s wonder toward a current event. In both cases, the visual component seems to be crucial.

An inseparable element characterizing the uses of αὖ, αὖτε, and αὐτάρ overall is anaphoricity. αὖ, αὖτε, and αὐτάρ are essentially anaphoric marks, and this does not conflict with the sense of distinctiveness. In a previous section of the current chapter, I have introduced two complementary anaphoric aspects of these markers. The first is the implicit “backward” reference to the same visualizing “I” that sees in his mind’s eye what is “on the one side” and what is “on the other side,” while the second is the backward reference to what has already been introduced in the discourse (i.e. “the same I have already mentioned {259|260} before”). The latter corresponds to the most common meaning of linguistic anaphora, whereas the former has a more performative significance. What I find striking is that both aspects converge once we think of the picking-up of threads of discourse. Whenever αὖ, αὖτε, and αὐτάρ are used, an overarching structure (the ongoing discourse and the ongoing performance) is recalled, and single threads (or paths) within this structure are followed one by one, each step corresponding either to a sequence of utterances—an entire narrative section—or to a single utterance. Distinctiveness presupposes anaphoricity. Discontinuity presupposes continuity. This is typical of discourse markers, whose primary function is to signal both where discourse comes from (backward orienting) and where discourse is going to (forward orienting). Picking up the different threads of discourse is a crucial task for each performer, not only because his telling must be coherent, but because the performer first has to recall what memory lets him produce, discourse act by discourse act. To reiterate: each presentational use of “then” does not primarily mark what is next in the narrated actions, but what is next in the performer’s mind.

In my view, the different etymological explanations and the different hypotheses about the original meaning of αὖ point to the same ambivalence. The “continuative” and the “separating” meanings—whether or not they stem from the same or from two different IE particles—are not contradictory. They both reflect the narrative need to anchor discourse to the speaking (and visualizing) “I,” and to discretely identify topics and individuals about whom something new is going to be introduced. And this is also what basically characterizes the Homeric uses of αὐτός. Anaphoricity, narrative centrality, and visual relevance pertain to αὐτός as well. The recognition of someone’s identity and the emotional nearness toward the αὐτός-subject are a notable part of its significance, as well. In chapter 3, I have introduced König’s notion of center and periphery relating to intensifiers. Now, I would like to make a comparison concerning αὖ, αὖτε, and αὐτάρ. Just as each utterance of αὐτός may suggest the presence of a periphery—either thematic or visual—so each utterance of αὖ, αὖτε, and αὐτάρ may likewise imply a peripheral side. The latter consists in the larger discourse frame within which a single thread of the story or a single character is focused, which has a corresponding visual counterpart: the piecemeal visualization of separate items (or separate scenes) is part of a coherent continuum guaranteed by the speaking “I.” He is the visual zero-point that produces the field of vision and creates the “syntax of movement.” [200] {260|261}

As a final point, I would like to briefly comment on the lyric, elegiac, and iambic instances of αὖ, αὖτε, and αὐτάρ. The psychological and cognitive aspects shared by epic and melic poetry—involvement, visualization, and recognition—derive from the idea of an oral center of gravitation and from the principle of kairotic play: the strategies enhancing them work only if performers and listeners suspend their disbelief and perceive the truth of what is uttered by the “I, here and now,” who reenacts the past and celebrates special moments of life. From this point of view, singing heroes’ glory, singing love, singing god’s presence, singing human victories activates the same cognitive activities; each is a pragmatically cognate event. In light of that, the presentational and the interactional values of αὖ, αὖτε, and αὐτάρ in melic poetry reflect cognate communicative intentions—namely, to anchor the discourse presentation to the hic et nunc of the performance and to convey the experiential and emotional effects of what is introduced. As for distinctions of use according to different genres, αὐ-discourse markers are almost entirely absent in iambic poetry, except for some parodic use. Lyric (especially Pindaric lyric) and elegiac poetry seem to confirm, generally speaking, the main pragmatic functions illustrated in epic. A notable distinctive feature, however, is the relatively high frequency and the peculiarity of δηὖτε (δή + αὖτε). The “once again” recurrent element, especially in erotic poetry, refers to both the events that are part of the propositional content and the performance that makes the participants experience those events. As such, δηὖτε works at both the propositional and the presentational/interactional level. Within Homeric diction, only αὖτε, according to my research, sometimes works exclusively as a propositional adverb; this happens when it occurs in subordinated clauses. [201] Apart from these cases, the majority of the occurrences of αὖτε and all the occurrences of αὖ and αὐτάρ can be read in presentational and/or interactional terms. That δηὖτε in main clauses can also have a propositional meaning anticipates my discussion in the following chapter, which will be devoted to the borderline between propositional and presentational/interactional uses of adverbs, such as αὐτίκα and αὔτως. The purpose of making such a distinction is to better understand a set of words to which a seemingly irreconcilable variety of meanings have been attributed. {261|}

Footnotes

[ back ] 1. Fantuzzi and Hunter 2004:249. The contrast alluded to is between the almost mechanistic repeatability of formulaic epic expressions and the “not stealable” style of Callimachus.

[ back ] 2. Ruijgh 1971:716.

[ back ] 3. A notable exception, however, is the recent study by Revuelta Puigdollers (2009).

[ back ] 4. Let us think of αὔτως, for example.

[ back ] 5. Relevance theorists say that some words encode concepts while some other words encode procedures; see Blakemore 1987:144 and Bezuidenhout 2004:2–3.

[ back ] 6. The English example is taken from the Oxford American Dictionary and Thesaurus.

[ back ] 7. See Lyons 1977, II:638 and 682, as well as Introduction, n32.

[ back ] 8. The meaning of such indications is quite complex. On the complexity of the lyric “I,” see, among others, Tsagarakis 1977:149–153; Gentili 1984:4–5; D’Alessio 1994; Calame 1997:436–438; Segal 1998:73–78; Nagy 2004b:27–28.

[ back ] 9. De Jong 1987 points out non-neutral ways to present characters or events by the speaker as focalizer. Frontisi-Ducrot 1995 addresses the hidden attention of the speaker to the public. Bakker (1997c, 1999, 2001, and 2005) argues that the aorist tense and some particles reflect a strong anchorage of the narration to the present of the performance.

[ back ] 10. “Every oral world view has had the severe limitations of particularism, since it is centered in a communal space no broader across than a speaker’s voice could carry. This center, governed by human agreement, is bonded by vocal presence and maintained generation after generation by oral mimesis and diegesis. … The center of the oral world existed wherever an authoritative voice was raised and accepted as such by any grouping of underlings” (Collins 1991:14).

[ back ] 11. Collins 1991:XXIII. The author relates his work to Bateson’s seminal article “A Theory of Play and Fantasy” (Bateson 1972), in which the evolution of play (as something removed from reality) is said to have been important in the evolution of communication. More recently, Bara (1999) traces the notion of “game” in linguistics (whose starting point is Wittgenstein’s Philosophische Untersuchungen, originally published in 1953) in order to explain all communicative acts as “behavioral and conversational game.”

[ back ] 12. “repeatable units of time … are the time of ritual, drama, narrative, song, and dance”; “kairos … is a durational totality all of whose temporal points are knowable and retrievable” (Collins 1991:77 and 78, respectively).

[ back ] 13. Collins 1991:77. The author borrows “willing suspension of disbelief” from Coleridge (1907, II:6).

[ back ] 14. The works that have been taken into account for this chapter include: van Djik 1979; Zwicky 1985; Schiffrin 1987 and 2006; Brinton 1996; Lenk 1997 and 1998; Jucker and Ziv 1998; Rouchota 1998; Byron and Heeman 1998; Schourup 1999; Fraser 1999; Bazzanella 1995, 2006 and 2008; Blakemore 2002; Bezuidenhout 2004; Fleischman and Yaguello 2004; Redeker 2006; Fischer 2006; Cohen 2007.

[ back ] 15. Beside English, the range of languages includes Chinese (e.g. Luke 1990), Japanese (e.g. Suzuki 1998), Spanish (e.g. Pons Bordería 1998), French (e.g. Hölker 1991), Italian (e.g. Bazzanella 1995), Finnish (e.g. Hakulinen 1998), and Hebrew (e.g. Maschler 1998).

[ back ] 16. The label “discourse markers” is, unfortunately, not the only one used in the literature; for example, we also find “discourse operators” (e.g. Redeker 1991), “pragmatic markers” (e.g. Brinton 1996), “discourse particles” (e.g. Fischer 2000), “pragmatic particles” (e.g. Östman 1981), and “pragmatic connectives” (e.g. Van Dijk 1979). The different labels reflect different terminological traditions and different theoretical standpoints. Two volumes that collect multiple views on discourse markers are Jucker and Ziv 1998 and Fischer 2006.

[ back ] 17. “We could speak of discourse markers’ ‘transversness’ with regard to other grammatical categories; the class of discourse markers is not grounded on fixed morphological and syntactical features but on their contextual use in a given text (both written and spoken)” (Bazzanella 2006:451). On the “multicategoriality” of discourse markers, see also Schiffrin 1987:40; Schourup 1999:234; Fraser 1999:450.

[ back ] 18. Cf. Östman 1981:6; Hölker 1991:79; Bazzanella 2006:449.

[ back ] 19. This example recalls an analogous instance presented by Bazzanella 2008:223. The double question mark preceding the sentence is a conventional sign indicating pragmatic unacceptability. Schiffrin (1987:322–326 and 2006:335–337) cites the indexicality of discourse markers, which rests on pointing to prior or upcoming sections of discourse, in involving the participants in communication, and in expressing the speaker’s stance. See also Aijmer 2002:14–16. On the deictic meaning of interjections, see Wilkins 1992.

[ back ] 20. See Brinton 1996:34; Zwicky 1985:303; Schourup 1999:232; Cohen 2007:v. On the relevance of “comma intonation” relating to conjunctions, see Sweetser 1990:82–84.

[ back ] 21. On discourse markers’ initiality, see Schourup 1999:233, Brinton 1996:33; van Dijk 1979:449.

[ back ] 22. See, for example, Lenk 1997; Bezuidenhout 2004:1–2; Bazzanella 2006:449.

[ back ] 23. Propositional content is here considered as a synonym for conceptual content and for representational information; see below, concerning levels of discourse. Exponents of the Relevance Theory (that is, a sub-field of pragmatics particularly devoted to cognitive inputs and outputs of pragmatic competence) claim that some discourse markers actually do give some conceptual contribution to the sentence. This seems to be the case, for example, of many English -ly adverbs, such as “frankly” or “seriously.” See Wilson and Sperber 1993:16–19 and Schourup 1999:246–247. For a Relevance Theory-based critical view of discourse markers as a unifying category, cf. Blakemore 2002.

[ back ] 24. See, among others, Aijmer et al. 2006:103; Walrod 2006:2; Brinton 1996:35 and 39.

[ back ] 25. The definition of the functions is by Schiffrin (2006:337); examples are mine.

[ back ] 26. See Traugott and Dasher 2002:154: the local scope of discourse markers is “between contiguous utterances,” whereas the global scope refers to the marking of episodes or the imposition of a hierarchical structure; see also Lenk 1997 and Georgakopoulou and Goutsos 1998:889.

[ back ] 27. The latter example occurs in Blakemore (2002:100).

[ back ] 28. “[‘oh’ is used] to propose that its producer has undergone some kind of change in his or her current state of knowledge, information, orientation or awareness” (Heritage 1984:299). Schiffrin devotes an entire chapter to the discourse functions of “oh” (1987:73–101).

[ back ] 29. “Since discourse markers are a functional, instead of a lexical category, they cannot be translated based on the meaning of the word”; such is the basic view of Mariano (2002:3) in a work devoted to the Italian translations of discourse markers in Harry Potter and the Sorcerer’s Stone. Lenk 1997 writes of emptiness of lexical meaning. Aijmer et al. 2006:103 write of “elusiveness of the meaning of pragmatic markers.” In certain cases, the lexical meaning may even fully contrast with the pragmatic sense: for example, Östman 1981 has shown that “you know” is used when the hearing “you,” in fact, is not aware of what the speaker is saying (it is just a rhetorical way to create a common ground).

[ back ] 30. See, among others, Brinton 1996:33; Schourup 1999:234

[ back ] 31. See Luke 1990:199; Zufferey and Popescu-Belis 2004:4; Bazzanella 1995:231.

[ back ] 32. Halliday and Hasan 1976:26; italics in the text.

[ back ] 33. Here I just mention some discourse-oriented divisions: Schiffrin’s model includes five “discourse planes” (“discourse domains” in Schiffrin 2006)—namely, “information state” (“what speaker and hearer know”), “participation framework” (“social side of speaker and hearer: identities, alignments, relationships”); “act structure,” “exchange structure” (in dialogs), and “idea structure” (i.e. propositions and topic/comment alternation); Sweetser’s tripartition is “content,” “epistemic,” and “speech act-domain” (1990), and Redeker distinguishes “content,” “discourse purpose,” and “sequential relations to the context space” (1991).

[ back ] 34. Kroon 1995:69.

[ back ] 35. Kroon 1995:73–75.

[ back ] 36. Kroon 1995:89–90. A theoretically important point by Kroon (1995:108–115)—inspired by Roulet et al. 1985—concerns the notion of “dialogical” features potentially present in monologues and of “monological” features potentially present in dialogues. In other terms, the idea of communicative exchange should also embrace cases of implicit or fictitious exchange—which are not infrequent in literary texts; for Kroon’s notion of “diaphony,” see ch. 3, n8.

[ back ] 37. On the presentational level of discourse regarding discourse units, such as acts and moves, see Kroon (1995:75–89). The original articulation of conversations into exchanges, moves, and acts is in Roulet et al. 1985. For pragmatic analyses of the uses of “because,” see, among others, Sweetser 1990:77; Schiffrin 1987:191–227. For the distinction between semantic and pragmatic connectives, see, in particular, van Dijk 1979. One may note that in (1) as well as in (2) “because” could, in fact, be dropped in favor of a full stop; (1) and (2) would still be understandable. The function of “because” is simply to make the procedure in the connection explicit. However, while in (1) no extralinguistic addition would be needed if “because” were dropped—as the logical connection of the two facts does not depend on the context of the utterance—in (2), conversely, a specific facial gesture or tone of voice would have to accompany one of the two acts, otherwise the interlocutor might not make sense of the adjacent clauses.

[ back ] 38. A close investigation of the semantic and pragmatic uses of “really” along with a survey of related literature is Haeyeon 2004. The author concludes that free-standing “really” used as a discourse marker in conversations “functions as a newsmark that indicates the need for specification or for more information in order for the information to be confirmed” (2004:880).

[ back ] 39. Cohen 2007:xiv.

[ back ] 40. It might also be the case that the same discourse marker has a particular presentational function in one context and a different function on the interactional level in other contexts. For instance, “oh!” starting a sentence such as “I don’t think I am going to spend the holiday with my husband” has an interactional function that differs from the presentational function of “oh!” starting a separate sentence in an e-mail concerning a research project, “oh, it occurs to me that we should record the timings as well.”

[ back ] 41. In the following analysis, I will retain Kroon’s term “presentational.” In Bonifazi (2008a) and in Bonifazi (2009a), I use the term “procedural,” after Rouchota 1998 and Bezuidenhout 2004. However, “procedural” in research on discourse markers pertains specifically to a Relevance Theory-based view that is not shared by all scholars. “Procedural meaning” (originally introduced by Blakemore 1987 as “constraints on the results of the pragmatic inferences involved in the recovery of implicit content”—Blakemore 2002:4; italics in text) corresponds to the pragmatic load characterizing not only the use of discourse markers but also of deictic expressions such as “yesterday” and that of third-person pronouns such as “he” (see Carston 2002:160–164). “Procedural” is seen as opposed to “conceptual”: some words convey concepts, while some others convey procedures. Since “procedural” in Relevance Theory can be applied to both presentational and interactional discourse markers—as the original sense is general: “helping to process the proposition”—and since I do not adopt Blakemore’s and Carston’s view of discourse markers, I prefer to keep “presentational” and “interactional.”

[ back ] 42. “A separate type of relationship on the interactional level of discourse is concerned with the way in which the transmitted content is … evaluated in terms of the non-verbal interactional context in which the text is integrated. … This interactional or communicative situation includes the intentions, beliefs, attitudes, emotions and knowledge (both general and situation-bound) of the discourse participants” (Kroon 1995:94–95).

[ back ] 43. See Wierzbicka 1986:524 and Mosegaard 1998:199.

[ back ] 44. Slings’ analysis (1997) of “push” and “pop” particles in ancient Greek is an example that shows ways of marking discourse units above the level of single clauses or sentences.

[ back ] 45. Redeker (2006:344) distinguishes between “paratactic transitions” (“between segments that follow each other at the same level”) and “hypotactic transitions” (“involving interruption or suspension of an incomplete unit with parenthetical material”).

[ back ] 46. Norrick 2001:854. The excerpt is the transcript of a story elicited during the first session of a graduate seminar at a large midwestern American University. Each line of the transcript contains a single intonation unit.

[ back ] 47. Norrick 2001:859–860. The terminology concerning parts of stories is adopted from Labov 1972.

[ back ] 48. The German text of the Brothers Grimm is after the 1987 Reclam edition; however, I have not reproduced the indentation.

[ back ] 49. The collection of fairy tales from the Neapolitan area by De Simone 1994 is the outcome of a systematic anthropological and linguistic inquiry into the currently surviving forms of oral storytelling. The tale at issue had been recorded in 1991 in Montesano Scalo (province of Salerno) from a 68-year-old peasant. The quoted passage is from p. 28 of the first volume (though I have not reproduced the original indentation).

[ back ] 50. The text of this song was dictated; the reported lines are 191–194. Translation is by Elmer (2009:54).

[ back ] 51. D. F. Elmer and I are preparing a pragmatic and performative commentary on an epic song from the Milman Parry Collection, which is intended to capture the interlacing of melodic discontinuities, intonation discontinuities, and verbal conjunctions that mark salient moments in the story.

[ back ] 52. In the history of discourse studies, some relatively early scholarship has delved into discourse markers by studying German particles, whence the terminological division “scope particles” (roughly overlapping with presentational particles) vs, “modal particles” (roughly overlapping with interactional particles). This is the framework adopted, for instance, by Wakker in her investigation of εἰ occurring along with other particles (on which see below).

[ back ] 53. “The term particle was used loosely in ancient rhetorical and grammatical theory to denote a variety of different linguistic elements: morphemes, words, parts of speech and cola” (Hilton 1997/1998:198). Sicking and Van Ophuijsen (1993:6) comment: “In plainer words we are dealing with a class of vocables which are not covered by the categories of traditional grammar, which in many cases cannot be translated, and whose omission does not leave the text incomprehensible or affect the truth conditions of the statements contained in it”. For a summary account of ancient views on particles, see Schenkeveld 1988 and Sicking and Van Ophuijsen 1993:4–6; also Sluiter 1997 on ancient “parapleromatic lucubrations;” see also Durán López 2000:50–51.

[ back ] 54. Schenkeveld 1988:88–89. The quotation is Schenkeveld’s translation of Devarius’ Praefatio to the Tractatus, p. 2. Later, J. Locke in An Essay Concerning Human Understanding (1691) claimed the subjective dimension of marking discourse through the use of particles (in English): “[he] who would show what significancy and force they have, must take … pains, enter into their own thoughts, and observe nicely the several postures of his mind in discoursing.” The concern about how to paraphrase particles (in German) is present in G. W. Leibniz’s thought: “For a proper explanation of the particles it is not sufficient to make an abstract explication … ; but we must proceed to a paraphrase which may be substituted in its place, as the definition may be put in the place of the thing defined. When we have striven to seek and to determine these suitable paraphrases in all the particles so far as they are susceptible of them, we shall have regulated their significations.” Both quotations are from Schourup 1999:256.

[ back ] 55. All the quotations are from Denniston 1996:vi. The latter originally runs: “Sometimes, indeed, the particle itself is left untranslated, and its force has to be gathered from the sentence as a whole. I have made free use of italics and exclamation marks, clumsy, but convenient, devices.”

[ back ] 56. A helpful online bibliography—periodically updated—about different studies in ancient Greek particles is included in “A Bibliography of Ancient Greek Linguistics” <vkc.library.uu.nl/vkc/antiquity/knowledgeportal/Wiki/A%20Bibliography%20of%20Ancient%20Greek%20Linguistics.aspx>. As for particles in Latin, a landmark study is Kroon 1995, which analyzes several discourse particles in several authors within a consistent theoretical framework.

[ back ] 57. I quote her definition of particles in a functional perspective: “a means of placing the unit they have in their scope into a wider perspective, which may be the interactional situation of which the text forms part or the surrounding context (and its implications). As such, particles may, for instance, enable the speaker to anticipate or respond to the (supposed) knowledge, presuppositions, expectations or questions of the addressee, or they may mark the coherence between different text units” (Wakker 1994:304–305). Already J. J. Fränkel (1947:201) was defining a particle as “a word without a meaning which … is able to discharge emotion, to make known intentions, to connect related clauses.”

[ back ] 58. Further works by Wakker on particles are Wakker 1996 and 1997.

[ back ] 59. I would add De Jong’s article (De Jong 1997) as it ultimately highlights an important presentational function of γάρ (which coincides with a specific narratological function).

[ back ] 60. “… scalarity may be used when one is dealing with ‘at least’ and ‘at most’ meanings” (Bakker 1988:29).

[ back ] 61. Bakker 1993c:15–25; 1997c:17–20; 2005:97–98.

[ back ] 62. See Kroon 1995:35 (“particles are words that do not contribute to the propositional content of a sentence or utterance”), and Durán López 2000:46.

[ back ] 63. “Particles are notoriously polyvalent” (Kroon 1995:96); see, also, Durán López 2000:68–70; 74–75.

[ back ] 64. Kroon is clear about this point (1995:103): “presentational relations obtain as a rule between independent acts (or clusters of acts), while within-act relations are always representational.” Among causal connectives, for example, several IE languages have different lexical entries distinguishing between presentational and representational functions: Latin has quoniam working at the presentational level and quia working at the representational (see Kroon 1995:105–106). As for English, “for” works only at the presentational level, while because works in both senses (see Davies 1979). For similar distinctions in French, see Anscombre 1984; in German, see Dunbar 1985. As for ancient Greek, Rijksbaron 1976 compares representational and subordinating ὡς vs. presentational and syntactically more independent ἐπεί. On the pragmatics of ἐπεί, see the recent work by Muchnová (2003). About the potentially primary function of εἰ and ἐπεί as particles, Denniston already states: “ἐπεί, εἰ and so forth, must themselves be regarded as particles” (1996:xl). Homeric αὖτε occurring in subordinated clauses has a propositional meaning; see, for example, Iliad XIX 151.

[ back ] 65. On the crucial relationship between information structure and prosody, see Gundel and Fretheim’s entry “Information Structure” in HoP online.

[ back ] 66. Consider, for example, the reading aloud of a sentence such as, καὶ μὲν δή, ὦ Σώκρατες, ἀτεχνῶς γέ μοι δοκεῖς … (Plato Cratylus 396d).

[ back ] 67. Fraser 2001:140. One of the advantages of this innovative reading of second position is that “the resulting initial prominence contributes to the typically contrastive nature of Greek co-ordinated clause linking.” (141).

[ back ] 68. Wackernagel 1892.

[ back ] 69. For example, marking topicalization or marking focalization, marking semantic or pragmatic contrasts.

[ back ] 70. For a short history of punctuation marks, see Truss 2003:72–79. In his recent edition of Plato’s Ion, Rijksbaron (2007:68–72) summarizes the main punctuation marks used in Byzantine texts, which were supposed to guide reading aloud.

[ back ] 71. Truss (2003:70) distinguishes between two major functions of punctuation, namely “to illuminate the grammar of a sentence” and “to point up—rather in the manner of musical notation—such literary qualities as rhythm, direction, pitch, tone and flow.” On the procedural vs. conceptual uses of punctuation marks, see also Borochovsky-Bar Aba 2003.

[ back ] 72. Pace Duhoux 1997, which finds particles occurring more frequently in non-dialogic parts of texts. For criticisms of Duhoux’s criteria and discussion of the distinction between orality and dialogicality, see George (2009).

[ back ] 73. Before Monro, Bäumlein (1861:44) asserted that the original meaning of αὖ, αὖτε, and αὖτις was “back” (“zurück”).

[ back ] 74. “In der Regel fungiert αὖ als Partikel mit schwach adversativer Bedeutung, die Entsprechung oder Ergänzung anzeigt (Verschiedenheit im Gleichen: korrelative Opposition) nicht selten in abgeschwächter Bedeutung einfach weiterführend (auch)” (1522; italics in the text).

[ back ] 75. Denniston (1996:55) already remarks: “αὐτὰρ ἐπεί often marks the successive stages of a narrative.”

[ back ] 76. Ἀτρεΐδη σὺ δὲ παῦε τεὸν μένος· αὐτὰρ ἔγωγε / λίσσομ’ Ἀχιλλῆϊ μεθέμεν χόλον …, which Lattimore (1951) translates as “Son of Atreus, give up your anger; even I entreat you to give over your bitterness against Achilleus” (Nestor is talking to Agamemnon).

[ back ] 77. I will return to an analysis of this passage in a later section.

[ back ] 78. On αὐτάρ, see especially Ruijgh 1957:29–55.

[ back ] 79. See Baltic au– and Old Slavonic. u- “weg, ab”, to be linked to Old Indic áva (Frisk 1960–1972, I:183).

[ back ] 80. “… to carry forward or continue the description, narrative etc.” (Klein 1988:281). Interestingly enough, Klein links the usage of αὖ to textual cohesion: he calls it “a discourse organizational element” (Klein 1988:271). The Rigvedic particle u reveals parallel usages in Klein’s view. Klein and Condon 1993 confirms this thesis by means of the comparative study of Gothic –(u)h.

[ back ] 81. Even though Puhvel (1984:245) does not see the connection as probable, Öttinger does not exclude that Greek αὖ in its isolating functions could have drawn from Hittite awan (personal communication).

[ back ] 82. I am indebted to Craig Melchert, to Norbert Öttinger, and to Georges-Jean Pinault for personal communications about the etymology of αὖ, which I benefited from during the Arbeitstagung der Indogermanischen Gesellschaft 2007, held in Marburg. Unfortunately, at the time of the writing of these pages I did not have access to Dunkel’s manuscript on IE particles.

[ back ] 83. See Boisacq 1907–1916, I:99; Schwyzer 1939–1971, II:559; Denniston 1996:55.

[ back ] 84. My primary interest in this chapter is to show the meaningfulness of the αὐ- component of αὐτάρ, rather than to delve into the question of whether what follows is to be interpreted as either τ’ ἄρ or ταρ. Nevertheless, I stress that some of Katz’ arguments are compatible with my synchronic pragmatic analysis. First, the Luwian particle is supposed originally to have a spatial connotation (due to the locatival etymology), which could match the use of αὐτάρ close to verbs of motion (Katz 2007:75). This might harmonize with what I am going to introduce in the next section, namely, αὐτάρ’s pragmatic function of visually “locating” people and events within the threads of the epic narration. Second, tar seems to be involved in the usages of αὐτάρ, but not in those of ἀτάρ (Katz 2007:72–73). This gives me the opportunity to outline briefly here a few differences between the Homeric uses of αὐτάρ and those of ἀτάρ. Katz (2007:73) suggests: “ἀτάρ has in general a somewhat different narrative function, namely to cast a new light or give a second perspective on a scene that is in the process of being described.” In line with Kroon’s analysis of Latin at (1995:333–370) and by considering the etymology of ἀτάρ as IE *át + ἄρ (see Dunkel 1988), I find that most of the Homeric occurrences of ἀτάρ work mainly at the interactional level of discourse; the higher degree of adversativity that is said to belong to ἀτάρ—in comparison to αὐτάρ—may come from a basic pragmatic value that expresses what Kroon calls a “dissonant reaction” (1995:349). Despite the Homeric instances that permit us to infer a somewhat interchangeable use of αὐτάρ and ἀτάρ (many scholars have claimed the metrical convenience of such an alternation, e.g. LfgrE I:1476; Humbert 1972:384), ἀτάρ has its own range of pragmatic functions, which are not shared by αὐτάρ. In particular, ἀτάρ typically signals a frustration of expectations; the flow of discourse is interrupted by the introduction of a comment, or an exclamation, or a self-correction (on the latter function of ἀτάρ, see Durán López (2000:60–61). ἀτάρ has a much less visual power of evocation, and it is never used as a presentational mark of entirely new discourse units.

[ back ] 85. The focusing device consists in singling out “constituents which count as salient in the immediate context or situation” (Kroon 1995:227).

[ back ] 86. Thematic discontinuity in Kroon’s view (1995:248) is “each interruption of a thematic strand (whether it concerns a referential discontinuity or a discontinuity of one of the other possible thematic strands).”

[ back ] 87. “The expression et … et autem is used to emphasize on the one hand the coordination of corresponding units of information, and on the other the individuality and mutual discreteness of the coordinated elements. This remarkable mixture of (formal) conjunction and (semantic or pragmatic) disjunction is … a recurrent feature of autem, which distinguishes it from other, comparable particles” (Kroon 1995:231; italics in the text). An example given by Kroon (1995:230) clarifies this: Facile istuc quidemst, si et ILLA volt et ILLE autem cupit “This (i.e., her leaving) is easy enough, if SHE wishes it, and HE, moreover, is crazy for it.” Plautus Miles gloriosus 1149; capital and italics in the text).

[ back ] 88. Orlandini’s theoretical model is Ducrot 1984b. See also Anscombre-Ducrot 1983, whose central thesis is that the (pragmatic) effects of language are implied by its “argumentation.”

[ back ] 89. See De Jong 1987:223. The adoption of one character’s point of view is discussed also in S. D. Richardson (1990:54–57) and in Rabel (1997:22–23; 80–87; 179–186).

[ back ] 90. Bakker 1993b:10. Already in Havelock (1963:160), minstrels are said to re-enact “the doings and sayings of heroes,” and to be emotionally identified with what they recite; this is what Plato calls mimesis. Nagy (1990a:373) asserts: “mīmēsis can designate not only the reenacting of the myth but also the present reenacting of previous reenactments.” On the notion of the “reenacting ‘I,’” see Nagy 2004b:27–28.

[ back ] 91. Bakker 2005:97–100 and 146.

[ back ] 92. Visual perception is not instantaneous but successive; objects are perceived part by part, through saccades (i.e. eye movements) alternating with fixations, where the focal field rests (see Collins 1991:96–98; 114–117). Hence, as Collins states (1991:98), “the consecutiveness of speech accords with the consecutiveness of visual perception.” From the cognitive point of view, perceptual and imaginal visuality follow the same procedure: “imaging is just as active a process as perceiving” (Collins 1991:91); see also Kosslyn 1995:268.

[ back ] 93. See Rubin 1995:304–307. Imagery includes, in Rubin’s view, both object imagery and spatial imagery.

[ back ] 94. Lists and catalogues basically constitute a performance within the performance (Minchin 2001:77 and 92; see also Létoublon 1998:176); the difference between them and the main narration is to be understood in terms of the particular visual/imaginative attention required of the listeners (and of the speaker, as well) during the enumeration of the entries.

[ back ] 95. See chapters 4 and 5 in Bakker 1997b; Bakker 1993a and Bakker 1997c.

[ back ] 96. “I define it [framing] as the demarcation of a frame limiting one’s field of vision for the next moments or speech units, the area within which addition of detail can meaningfully take place” (Bakker 1997b:89).

[ back ] 97. Let us take a plain sample from Bakker’s analysis. Iliad 4.459 can be divided into two information and intonation units: τόν ῥ’ ἔβαλε πρῶτος “He [Antilochus] first hit him [Echepolus],” and κόρυθος φάλον ἱπποδασείης “on the crest of his horse-haired helmet.” This represents an easy-to-visualize sequence articulated in two steps: the frame that is visualized and a detail that is added as a “second shot” (translation and quotation from Bakker 1997b:90).

[ back ] 98. More on immediacy in ch. 5, pp. 275 and 282.

[ back ] 99. Further relevant work from cognitive psychology is presented in Bonifazi 2008a.

[ back ] 100. On the difference between seeing and noticing, realizing, and recognizing in Homer, see Lesher 1981:10–12.

[ back ] 101. See Bakker 1997b:162–165; 162 for the quote.

[ back ] 102. Snell (1931:77) has already stated “Das νοεῖν steht … in naher Beziehung zum Organ des Auges.” See also Lesher 1981:8–9. Nagy 1983 connects the Homeric uses of νοέω, νόος, and νόησις to a specific cognitive signal of consciousness, that of “noticing of signs and recognition of what they mean” (Nagy 1983:38; italics in the text). An example cited by Nagy (1983:37) is Odyssey 8.94 and 533 Ἀλκίνοος δέ μιν οἶος ἐπεφράσατ’ ἠδ’ ἐνόησεν “Alki-noos was the only one who observed and noticed him [Odysseus]” (Nagy’s translation). See also ch. 3, n90.

[ back ] 103. As far as the episode of Odysseus’ scar is concerned, Scodel 2002 shows the relevance of flashbulb memory: a particularly intense visual memory that prompts the recognition and the interpretation of events; it is a sign of personal engagement as it involves “direct personal experience of the events narrated” (Scodel 2002:116).

[ back ] 104. Lexical root νο- refers to a “special awareness of the beyond, of the metaphysical” (Bakker 2002:78). Thanks to the poet’s νόος, the performance of the permits both him and his listeners to see and to accomplish Apollo’s presence (see Hymn ΙΙΙ to Apollo 186 ὥς τε νόημα). νοῆσαι, whose meaning is both “accomplishment and seeing with special clarity,” therefore becomes a means of “the very enactment of epic events” (Bakker 2002:79).

[ back ] 105. García 2002:14. In the Hymns, the uttered words are seen as corresponding to cultic actions and reenacted events: “The Homeric Hymns are not works of theology or cult history: they present the god to a group of celebrants” (García 2002:21).

[ back ] 106. García 2002:15.

[ back ] 107. See Hymn VII to Dionysus 15–16 κυβερνήτης δὲ νοήσας / αὐτίκα οἷς ἑτάροισιν ἐκέκλετο φώνησέν τε; Hymn V to Aphrodite 185–186: Αὐτίκα σ’ ὡς τὰ πρῶτα θεὰ ἴδον ὀφθαλμοῖσιν / ἔγνων ὡς θεὸς ἦσθα; Hymn IV to Hermes 87 τὸν δὲ γέρων ἐνόησε δέμων ἀνθοῦσαν ἀλωὴν; 208 παῖδα δ’ ἔδοξα φέριστε, σαφὲς δ’ οὐκ οἶδα, νοῆσαι; 213–214 οἰωνὸν δ’ ἐνόει τανυσίπτερον, αὐτίκα δ’ ἔγνω / φηλητὴν γεγαῶτα Διὸς παῖδα Κρονίωνος; Hymn II to Demeter 94–95 οὐδέ τις ἀνδρῶν / εἰσορόων γίγνωσκε βαθυζώνων τε γυναικῶν. We could also add Hymn III to Apollo 400–403, cited by García in his argument to retain the νοῆσαι of the codices, rather than Bolkestein’s conjecture βοῆσαι (García 2002:18–20). The text refers to Apollo’s epiphany in the guise of a dolphin: ἐν πόντῳ δ’ ἐπόρουσε δέμας δελφῖνι ἐοικὼς / νηῒ θοῇ, καὶ κεῖτο πέλωρ μέγα τε δεινόν τε· / τῶν δ’ ὅς τις κατὰ θυμὸν ἐπιφράσσαιτο νοῆσαι / πάντοσ’ ἀνασσείασκε, τίνασσε δὲ νήϊα δοῦρα. Further passages from the Hymns including αὐτίκα will be discussed in ch. 5.

[ back ] 108. “Priming” and “focusing” are two crucial concepts used by Emmott (1997) to distinguish between an overall context provided by the text and the reader’s reference to the set of primed participants in that context (“priming is like characters standing on a stage in front of us, whereas focusing is like someone drawing our attention towards one of these characters,” p. 217).

[ back ] 109. See also Iliad 23.881. Other instances illustrating this function of αὐ-discourse markers are Iliad 5.418 (Dione is “on one side,” while Athena and Hera are “on the other side,” αἳ δ’ αὖτ’); Iliad 7.268 (Hector acts “on one side,” while Ajax reacts “on the other side,” δεύτερος αὖτ’ Αἴας); Iliad 11.130 (Agamemnon attacks Pisander and Hippolochus, while the two of them, on the other side—τὼ δ’ αὖτ’—beg Agamemnon from the chariot); Iliad 13.178 (Teucer strikes Imbrius, while Imbrius, on the other side, falls down, ὁ δ’ αὖτ’). See also Hesiod Shield 282–285. “On the other side” is verbally reinforced by the adverb ἑτέρωθεν following αὖθ’ in some passages, such as Iliad 7.311, 8.55, 11.56, 14.388, 15.501, 16.755, 18.243, 20.3, 22.79.

[ back ] 110. For example, Hector and Teucer’s duel at Iliad 8.323–329: while Teucer (ὃ μέν, 323) is resting the arrow on the nerve between neck and chest, Hector strikes him (τὸν δ’ αὖ κορυθαίολος Ἕκτωρ, 324) right there with a boulder. See, too, Iliad 8.487 (the Trojans dislike the dark of night—Τρωσὶν μέν—while the Achaeans are happy about that—αὐτὰρ ’Αχαιοῖς); 9.1 (the Trojans held their night watches—οἳ μὲν Τρῶες—whereas the Achaeans are held by fear—αὐτὰρ ’Αχαιούς); Odyssey 1.110–111 (some servants of the suitors—οἱ μὲν—mix wine and water in the crater, while others—οἱ δ’αὖτε—clean the tables); Odyssey 23.243 (in order to permit Odysseus tell his tales to Penelope, Athena prolongs Night—νύκτα μέν—and at the same time detains Dawn—’Ηῶ δ’ αὖτε). For other uses of simultaneity in parallel situations, see Iliad 4.7–11 (Hera and Athena are watching over Menelaus from a distance; Aphrodite is close to him—τῷ δ’ αὖτε … ’Αφροδίτη); Iliad 10.419–421 (while the Trojans keep their guard, their allies sleep and let them battle—αὖτε … ἐπίκουροι / εὕδουσι).

[ back ] 111. Thus, I agree fully with Scodel, who emphasizes that “there is no single correct approach to Homer’s handling of time” (2008:107).

[ back ] 112. Schiffrin (1987:254) argues: “as then can mark temporal succession between events, so too can it mark temporal succession between the ideas and topics of talk” (italics in the text). Halliday and Hasan (1976:239) provide a similar distinction about the use of “next”: (a) “Next he inserted the key into the lock.” vs. (b) “Next, he was incapable of inserting the key into the lock.” In the latter sentence “the time sequence is in the speaker’s organization of his discourse.”

[ back ] 113. As in such instances as αὐτὰρ ὕπερθε(ν), αὐτὰρ ὄπισθε(ν) at: Iliad 2.218; 5.724; 7.101; 12.398 and 446; 13.682; 24.797; Odyssey 14.476; 24.230; Ηymn V to Aphrodite 158 (αὐτὰρ ὕπερθε); Iliad 17.718 and 746; 18.554; 23.763; 24.325; Odyssey 14.393; 19.436 (αὐτὰρ ὄπισθε[ν]).

[ back ] 114. Iliad 6.229 (πολλοὶ μὲν γὰρ ἐμοὶ, 227; πολλοὶ δ’ αὖ σοὶ, 229); 10.108 and 292; 11.367; 16.87; 21.399; 24.595 and 732; Odyssey 3.382; 13.303 (including αὖ); Homeric Hymn IV to Hermes 302–303 (ἑυρήσω … / … σὺ δ’ αὖθ’ …). Kroon (1995:233–235) reports Latin autem to be used in changes of addressee, including “I” vs. “you.” At Iliad 11.785, σοὶ δ’ αὖθ’ marks the shift from Peleus addressing Achilles to Menoetius addressing “you” Patroclus, i.e. the interlocutor of Nestor, who is giving one of his speeches.

[ back ] 115. So Bennekom and Beck in LfgrE s.v. αὐτάρ, 1578 “die Antithese σὺ / αὐτὰρ ἐγώ scheint sinnlos und ist am ehesten durch die Emotionalität der Rede zu entschuldigen.” I already quoted this above, p. 210.

[ back ] 116. This formula occurs nineteen times over the first four books of the Odyssey, and twenty-five more times from book xv on.

[ back ] 117. Out of 209 total occurrences of αὖτε in the Odyssey and in the Iliad, 99 (47.36%) belong to the formula τὸν/τὴν δ’ αὖτε προσέειπε + epithet and name. For a comprehensive examination of the diction of speech-introductory formulas (initial as well as reply formulas), see Beck 2005:32–43; 79–80; 129–134; 193–194; 278–285.

[ back ] 118. See Iliad 6.214; also 21.435 and 22.7.

[ back ] 119. Iliad 3.76 is very similar to this. The Odyssey includes frequent occurrences of ὣς + different forms of φημί, followed by a pronoun + αὐ-discourse markers: Odyssey 2.103 (ἡμῖν δ’ αὖτ’); 9.256; 10.466; 12.28; 24.138; see also Homeric Hymn IV to Hermes 304 (with αὖτε); Odyssey 4.375, 394, 481, 538, 548; 9.360, 522; 10.67, 261, 270, 336, 382, 406, 422, 438, 475, 496; 11.79, 138, 163, 204, 435, 462, 477, 504; 12.111; 19.551; 21.366 (with αὐτάρ).

[ back ] 120. See, for example, Odyssey 4.210–211: Menelaus praises the blessings that Nestor has received from Zeus, namely, a happy old age for himself (αὐτὸν μὲν λιπαρῶς γηρασκέμεν, 210) and [on the other side] wisdom and prowess for his sons (υἱέας αὖ πινυτούς τε καὶ ἔγχεσιν εἶναι ἀρίστους, 211).

[ back ] 121. “Homer’s narrative runs smoothly from one scene to another, but he [the poet] takes great care at the articulation points to make clear to his listeners exactly what he is doing—what is ending, and what is beginning” (Edwards 2002:39).

[ back ] 122. Klein 1988:281.

[ back ] 123. Several uses of αὖτις and of αὔτως seem to confirm this; see ch. 5.

[ back ] 124. See LSJ’s comment on αὖ (s.v.) “adverb used for repeated actions.”

[ back ] 125. See Iliad 2.188–189 (Odysseus’ reported speech to one of the leaders: Ὅν τινα μὲν βασιλῆα καὶ ἔξοχον ἄνδρα κιχείη / τὸν δ’ ἀγανοῖς ἐπέεσσιν ἐρητύσασκε παραστάς· “whatever king or prominent man he came upon, he would stay beside him and restrain him with soothing words”), which is followed by 198–199 (his reported speech to one of the soldiers: Ὃν δ’ αὖ δήμου τ’ ἄνδρα ἴδοι βοόωντά τ’ ἐφεύροι, / τὸν σκήπτρῳ ἐλάσασκεν ὁμοκλήσασκέ τε μύθῳ· “then, again, the man of the people whom he encountered and saw shouting, he would strike him with the staff and call out with speech”); 4.232–233 (what Agamemnon says to the warriors who are eager to fight: καί ῥ’ οὓς μὲν σπεύδοντας ἴδοι Δαναῶν ταχυπώλων, / τοὺς μάλα θαρσύνεσκε παριστάμενος ἐπέεσσιν “and those of the Danaans of the fast steeds, whom he saw hastening, he would stay beside them and encourage them with words”), which is followed by 240–241 (what he says to the reluctant warriors: Οὕς τινας αὖ μεθιέντας ἴδοι στυγεροῦ πολέμοιο, / τοὺς μάλα νεικείεσκε χολωτοῖσιν ἐπέεσσιν “and those whom he saw giving up the hateful war, he would reprove them much with wrathful words”); 17.414 (the suggestion of someone among the Achaeans: ὧδε δέ τις εἴπεσκεν Ἀχαιῶν χαλκοχιτώνων “one of the bronze-clad Achaeans was repeating words like these”), followed by 420 (the suggestion of someone among the Trojans: Ὣς δέ τις αὖ Τρώων μεγαθύμων αὐδήσασκεν “likewise, [on the other side] one of the high-hearted Trojans was speaking aloud). A similar pattern of analogy a procedure can be found at Iliad 19.107; 21.190–191; 23.315–316. See also Homeric Hymn V to Aphrodite 218 (Just as Zeus abducted Ganymedes, [in a similar way] Dawn abducted Tithonus), and Hesiod Works and Days 295.

[ back ] 126. This sub-genre of performance occurs much more in the Iliad than in the Odyssey.

[ back ] 127. Simpler instances can be found at Iliad 13.451; 20.215, 219, 231, 236; Odyssey 16.119. Hesiod’s Theogony offers some examples as well: cf. 139, 147, 214, 237, 270.

[ back ] 128. It can be argued that this is the function of αὐ-discourse markers in various kinds of lists; see Iliad 2.406–407; 10.283; 23.841; Odyssey 8.129–130; Homeric Hymn III to Apollo 159. Thus, the representational aspect of the temporal succession of the events is not a necessary one; it may or may not be implied.

[ back ] 129. See, also, Hesiod Shield 124 δεύτερον αὖ θώρηκα περὶ στήθεσσιν ἔδυνε.

[ back ] 130. On the Homeric techniques used to describe heroes arming themselves, see, in particular, Létoublon 1998. Parallel repetition involving αὐ-discourse markers occurs also in Iliadic lists concerning a sequence of warriors attacked by one individual (e.g. Iliad 6.184 and 186) or sequences of actions regarding other ritual procedures, such as preparing a sacrifice or meal (Iliad 1.458, 464, 467; 2.402, 421, 427, 430; 9.212; 23.750).

[ back ] 131. For similar constructions in Hesiod, see Works and Days 245; fr. 33a, 15. Further Homeric constructions: Iliad 18.159 and 472; 22.171; 24.10 and 511; Odyssey 4.102; 8.174; 11.303; 16.209; also Hymn XIX to Pan 10.

[ back ] 132. Different but parallel courses can also be marked by αὐ-discourse markers in argumentative texts, where the meaning is “on the other hand.” Such uses are infrequent in Homer, but see Iliad 4.415 (τούτῳ μὲν γὰρ κῦδος “to him glory, on the one hand”) and 417 (τούτῳ δ’ αὖ μέγα πένθος “to him great grief, on the other hand”); 16.84 (ὡς ἄν μοι τιμὴν μεγάλην καὶ κῦδος ἄρηαι “so that you carry off great honor and glory for me”) and 87–88 (εἰ δέ κεν αὖ τοι / δώῃ κῦδος ἀρέσθαι ἐρίγδουπος πόσις Ἥρης “if the resounding husband of Hera should grant you glory to be carried off”), where the visual shift from “I” to “you” plays a role as well. Representative instances, which also include εἰ-clauses, are: Iliad 1.340; 5.224 and 232; 9.135 and 277; Odyssey 5.221; 16.105; 18.371 and 376). εἰ, indeed, introduces disjunctive situations. The preceding prompts me to draw a connection between these Homeric instances of αὖ and the use of Latin aut … aut introducing alternative and exclusive situations.

[ back ] 133. For Latin autem similarly used in parallel constructions, see Kroon 1995:229–233.

[ back ] 134. See also Iliad 23.140 and 193; Odyssey 2.382 and 393; 4.795; 6.112; 16.409; 18.187; 23.344. The slight variation at the beginning of the line is ἡ δ’ instead of ἔνθ’.

[ back ] 135. Equally clear instances are, in my view, Iliad 3.383; 8.489 (with αὖτε); Odyssey 13.125; 21.22.

[ back ] 136. 15.301 is repeated at 16.1, where it starts a new book.

[ back ] 137. On the paragraph divisions of Homeric texts in general, see, in particular, Edwards 2002:47–52.

[ back ] 138. My analysis of this point is indebted to Revuelta (2009), whose work on αὖ and αὖτε as topicalizing devices in Homer, in Herodotus, in Thucydides, and in Xenophon has been illuminating.

[ back ] 139. See also Iliad 11.122 (after a simile at lines 113–121, the narrator re-joins the list of warriors killed by Agamemnon); 18.602; Odyssey 11.385 (after some sentence directly addressed to Alcinous at lines 378–384, Odysseus re-joins his narration of the trip to the land of the dead by using αὐτὰρ ἐπεί. Instances of αὖτε and αὐτάρ used to re-join the narration after a direct speech is over—signaled by the phrase ὣς ἔφατ’ and the like—are listed in n119.

[ back ] 140. For αὖ, see, e.g., Iliad 2.681 (a new section of the Catalogue is marked by νῦν αὖ) and Odyssey 24.546 (the last three lines of the poem represent the beginning of a new moment of narration, even if very brief). As for αὖτε, see, e.g., Iliad 3.121 (Iris going to Helen); 17.319 (the Trojans withdrawing from battle); Odyssey 4.20 (Telemachus and Peisistratus appearing at Menelaus’ palace in Sparta); 22.495 (Eurycleia goes to call the faithful maidens).

[ back ] 141. See Bonifazi 2009c. Another significant instance of αὐτάρ marking the beginning of a different narrative section is Iliad 18.609; Hephaestus has finished the shield and starts making the rest of the armor.

[ back ] 142. Nagy (1999:56).

[ back ] 143. See Odyssey 16.1 and Homeric Hymn XXXI to Helius 1 (αὖτε); Iliad 15.1, Odyssey 11.1, xii 1, 14.1, 19.1, 20.1, 22.1 (αὐτάρ).

[ back ] 144. On the potential mismatch between prosodic and syntactic functions of punctuation, which do not help processing the discourse function of particles, see above, pp. 207–208.

[ back ] 145. As I already noted in a previous section, Bakker hypothesizes a fundamental equivalence between αὐτάρ and δέ (1997b:96). I prefer to note that the equivalence of the two is only partial: δέ may always substitute for αὐτάρ, whereas αὐτάρ can substitute for δέ only in transitions to relatively longer narrative sections; in other words, αὐτάρ cannot mark adjacent and small narrative steps, while δέ conversely does.

[ back ] 146. This phenomenon only very rarely concerns objects. One instance, however, should be mentioned, as it clarifies a discourse marker function that is otherwise difficult to explain: at Iliad 20.481–482, the primary speaking ‘I’ narrates Achilles’ blow to Deucalion’s neck and severing of his head. Immediately afterward, a bloody detail is added by means of αὗτε, which works as zooming-in marker: μυελὸς αὖτε / σφονδυλίων ἔκπαλθ’ “[and there] the marrow spurted from the neckbone,” 482–483.

[ back ] 147. Sometimes, the two functions merge, as in Iliad 3.200, where Helen catches up with the list of heroes and, at the same time, zooms in on Odysseus: οὗτος δ’ αὖ Λαερτιάδης πολύμητις Ὀδυσσεύς “This one, then [zoom in on him], is Odysseus of much cunning.”

[ back ] 148. In the last book of the Iliad, after Andromache, Hecuba is singled out amongst the women performing the ritual lament over Hector’s corpse by means of αὖτε: ἐπὶ δὲ στενάχοντο γυναῖκες. / τῇσιν δ’ αὖθ’ Ἑκάβη ἁδινοῦ ἐξῆρχε γόοιο “and the women were groaning about this. Among them, Hecuba started her sobbing lament” (Iliad 24.746–747).

[ back ] 149. ἀρχοὶ δ’ αὖ occurs also at Odyssey 4.496.

[ back ] 150. I repeat that multifunctional δέ may mark singling out, as well (alone or with other particles). See, for example, Iliad 17.262: Τρῶες δὲ προὔτυψαν ἀολλέες· ἦρχε δ’ ἄρ’ Ἕκτωρ “The Trojans struck forward, all together; Hector was leading them.”

[ back ] 151. For example, I claim that αὐτάρ works like αὐτός at Odyssey 17.67 (it foregrounds Telemachus against the suitors); 18.343 (Odysseus against the maidens); 23.50 (Odysseus against the corpses of the suitors) and 24.225 (Laertes against the servants). See Homeric Hymn III to Apollo 179–181: ὦ ἄνα, καὶ Λυκίην καὶ Μῃονίην ἐρατεινὴν / καὶ Μίλητον ἔχεις ἔναλον πόλιν ἱμερόεσσαν, / αὐτὸς δ’ αὖ Δήλοιο περικλύστου μέγ’ ἀνάσσεις “Lord, you possess even Lycia and lovely Meonia and Miletus, attractive city on the sea. And you in person are the one who greatly reigns over Delos, washed around by the sea.” In ch. 3 (pp. 138–139), I have accounted for αὐτός with Menelaus as the referent—in the Catalogue of Ships (Iliad 2.588)—by the same idea.

[ back ] 152. See Iliad 5.1 (spotlight on Diomedes) and 471 (on Sarpedon); 12.182 (on Polypoetes); 16.477 (Sarpedon) and 603 (on Meriones). On the pragmatic similarity of ἔνθα and αὐτοῦ, see ch. 5.

[ back ] 153. ἔνθ’ αὖτε can occur as well. See Odyssey 22.281–283: some suitors attack Telemachus and Eumaeus. Then, on the other side (and in turn) Odysseus’ allies (τοὶ δ’ αὖτ’ ἀμφ’ ’Οδυσῆα, 281) attack as well. In particular, Odysseus kills Eurydamas (ἔνθ’ αὖτ’ Εὐρυδάμαντα βάλε πτολίπορθος ’Οδυσσεύς “At that point, Odysseus sacker of cities struck Eurydamas,” 283).

[ back ] 154. On the reasons why De Jong and Nünlist do not adopt the term “zooming in” (2004:67n6), while I keep it, see Bonifazi 2008a:54n55.

[ back ] 155. The quasi-technical use of ἔνθ- in the performative singling out of a specific point of the story is quite explicit in Odyssey 8, when the primary narrator shifts from the overall frame of Demodocus weaving his song (φαῖνε δ’ ἀοιδήν, Odyssey 8.499) to the telling of the specific point in the story from which the singer was starting: ἔνθεν ἑλών, ὡς οἱ μὲν ἐϋσσέλμων ἐπὶ νηῶν / βάντες ἀπέπλειον, which Lattimore (1967) translates “beginning from where the Argives boarded their well-benched / ships, and sailed away,” 500–501). Another significant case in which ἔνθα along with αὐ-discourse markers signals the shift and the zoom in—at the same time—to a different subject is the formulaic ἔνθ’ αὖτ’ ἄλλ’ ἐνόησε … (whose variant is ἡ δ’ αὖτ’ ἄλλ’ ἐνόησε) “At that point, different plans were in the mind of ….” See above, p. 231.

[ back ] 156. See Iliad 1.348; 9.628 and 663; 11.664 and 762; 16.124 and 220; 19.15, 188, 268; 20.75, 283, 423, 441; 21.520; 23.128 and 257; 24.3, 511, 675 (αὐτὰρ Ἀχιλλεύς); Iliad 5.844; Odyssey 6.2; 16.454; 17.360; 18.69; 24.367 (αὐτὰρ Ἀθήνη); Iliad 1.430; 11.321; Odyssey 1.57; 2.182; 5.370; 7.81; 8.83, 264, 367, 521; 13.28 and 367; 14.30 and 171; 16.119 and 177; 18.66, 100, 394; 19.209, 388, 479; 22.292, 493; 23.67, 181; 24.220 and 392 (αὐτὰρ Ὀδυσσεύς). Kahane (1994:119–120) says that proper-name nominative phrases ending the line are a “typical heroic feature,” a “typifying mark of role or character.”

[ back ] 157. A significant use of αὐτάρ zooming in on Odysseus occurs during the reported song of Demodocus at Odyssey 8: after framing the overall scene of the Achaeans coming out of the wood-horse and sacking Troy, the singer focuses on Odysseus reaching Deiphobus’ home (ἄλλον δ’ ἄλλῃ ἄειδε πόλιν κεραϊζέμεν αἰπήν, / αὐτὰρ Ὀδυσσῆα προτὶ δώματα Δηϊφόβοιο / βήμεναι …, Odyssey 8.516–518).

[ back ] 158. See Odyssey 2.205; 11.6; 12.148; 24.484.

[ back ] 159. See Odyssey 13.303 and 16.233.

[ back ] 160. See Homeric Hymn VII to Dionysus 27.

[ back ] 161. See p. 188 with n7. On the whole, αὐ-discourse markers and “I” marks co-occur in 22% of the total instances of αὐ-discourse markers.

[ back ] 162. In the Apologoi, αὐτὰρ ἐγώ often singles out “I”-Odysseus from the companions, as a thematized polarization (Odyssey 7.252; 9.173; 10.49, 475 and 480).

[ back ] 163. = Hymn III to Apollo 545 = Hymn IV to Hermes 579 = Hymn VI to Aphrodite 21 = Hymn X to Aphrodite 6 = Hymn XIX to Pan 49; Hymn XXVIII to Athena 18 = Hymn XXX to Earth, Mother of All 19. The slight variant αὐτὰρ ἐγὼν ὑμέων τε καὶ ἄλλης μνήσομ’ ἀοιδῆς occurs at Hymn XXV to the Muses and Apollo 7 = Hymn XXVII to Artemis 22 = Hymn XXIX to Hestia 13 = Hymn XXXIII to the Dioscuri 19.

[ back ] 164. In the translation, I follow Bakker’s interpretation (2005:144), which is in line with the proemial reading of the Homeric Hymns (see Nagy 1990a:353 and 359).

[ back ] 165. Only at Iliad 11.362 = 20.449, and 21.399.

[ back ] 166. Only at Odyssey 1.200 and 16.142.

[ back ] 167. Bakker 2005:170 (italics in the text). See also Bakker (2008:7): “the future, or rather the utterance of the future verb … is the beginning of a state of affairs that leads into the future” (italics in the text).

[ back ] 168. I have counted one subjunctive present verb, five imperative present verbs, fifteen indicative present verbs, two indicative perfect verbs, thirteen aorist verbs, six future verbs, and four verbal ellipses. In ch. 5, I will point to a similar phenomenon involving αὐτίκα.

[ back ] 169. For example, νῦν αὖ at Iliad 9.700 is the equivalent of “even,” conveying surprise (“now you [Agamemnon] even have driven him [Achilles] much deeper into his pride”); conversely, at Odyssey 4.817 νῦν αὖ marks a remarkable analogy (“first I lost a husband … ; and now my beloved son is gone”).

[ back ] 170. See also Iliad 17.478 and 672; 22.436.

[ back ] 171. I add that αὖτε in the second passage wither-joins the main thread of discourse after the thought about Zeus and Apollo; as such, it stresses the repetition of the concept “death is near me now.” A further—and intertextual—re-joining function of αὖτε concerns Patroclus’ prediction at the latter’s point of death: “but now already death and powerful destiny are standing beside you [Hector]” ἀλλά τοι ἤδη / ἄγχι παρέστηκεν θάνατος καὶ μοῖρα κραταιὴ, Iliad 16.852–853.

[ back ] 172. See Bazzanella 2004:64–65.

[ back ] 173. A landmark work on the lexical and conceptual range of the “emotional repertoire” of ancient Greek is Konstan 2006.

[ back ] 174. Even though it carries us beyond the scope of this chapter, I would briefly note that indirect sources may compensate for such a lack. To cite just a single example of the strong emotional involvement of Homeric rhapsodes, Socrates and Ion share the following thoughts at Plato Ion 535b–c (tr. Lamb 1952): “[Socr.] Stop now and tell me, Ion, without reserve what I may choose to ask you: when you give a good recitation and specially thrill your audience, either with the lay of Odysseus leaping forth on to the threshold, revealing himself to the suitors and pouring out the arrows before his feet, or of Achilles dashing at Hector, or some part of the sad story of Andromache or of Hecuba, or of Priam, are you then in your senses, or are you carried out of yourself, and does your soul in an ecstasy suppose herself to be among the scenes you are describing, whether they be in Ithaca, or in Troy, or as the poems may chance to place them? [Ion] How vivid to me, Socrates, is this part of your proof! For I will tell you without reserve: when I relate a tale of woe, my eyes are filled with tears, and when it is of fear or awe, my hair stands on end with terror, and my heart leaps.”

[ back ] 175. ποῖόν σε ἔπος φύγεν ἕρκος ὀδόντων. See Odyssey 1.64 (Zeus reacting to Athena inquiring about Zeus’ hostility towards Odysseus); 3.230 (Athena reacting to Telemachus’ incredulity about Odysseus’ return); 5.22 (Zeus reacting to Athena’s complaints); 19.492 (Eurycleia reacting to Odysseus’ threat if she should reveal the truth); 21.168 (Antinous reacting to Leodes’ ominous speech); 23.70 (Eurycleia reacting to an incredulous Penelope).

[ back ] 176. The scene relates to Iris and Thetis ascending from the sea floor to Olympus, so that Zeus can urge Thetis to convince Achilles to give Hector’s corpse to Priam.

[ back ] 177. The primary use of ἄρα for Denniston is “expressing a lively feeling of interest” (1996:33). On the relevance of ἄρα in Homeric performance according to Bakker, see above, p. 205.

[ back ] 178. The remaining questions that include αὐ-discourse markers occur at Iliad 1.540; 7.24; Odyssey 23.264 (αὖ); Iliad 1.202; 2.225; 7.448; 14.364; 18.6; 20.16; 21.394; Odyssey 6.119; 13.200; 10.281; 20.33 (αὖτε); Iliad 1.133; 15.134 (αὐτάρ). Autenrieth suggests that αὖτε in questions denotes “impatient tone”; Alpers (LfgrE I:1583) says it overemphasizes excitement (“affektiv übersteigert”) and may connote indignant communication.

[ back ] 179. See Kroon 1995:238–246.

[ back ] 180. The term is used by Givón (1984–1990, II:700). On counter-presuppositional focus, see S. C. Dik 1989. Kroon compares counter-expectancy with contrastive focus.

[ back ] 181. Kroon 1995:240. The passage to follow and the translation are quoted after Kroon (1995:238).

[ back ] 182. An extra-Homeric example of αὖ working in this way is Euripides Orestes 132–133, where Electra sees—and recognizes—the entering chorus of women, and says: αἵδ’ αὖ πάρεισι τοῖς ἐμοῖς θρηνήμασι / φίλαι ξυνωιδοί· “Here they are! Dear singers, whose song joins my laments”

[ back ] 183. The passage has been mentioned in ch. 1, n122.

[ back ] 184. Likewise, I see emotional charge and empathy in a further occurrence of κεῖνος δ’ αὖ—namely, when Odysseus speaks about Nausicaa’s imagined husband (Odyssey 6.158); the choice of such a demonstrative is significant in terms of the potential overlap between “that man” and Odysseus himself as κεῖνος. At Odyssey 7.313–315, Alcinous explicitly wishes that Odysseus might become his son-in-law.

[ back ] 185. Shorey (1928) had already pointed out several “pathetic” uses of αὖ, “sentimental” or “comic” or showing impatience.

[ back ] 186. See Nagy 1999:96.

[ back ] 187. On this point, see ch. 3, pp. 164–167 and 179.

[ back ] 188. A table summarizing the frequency and the distribution of αὖ, αὖτε, and αὐτάρ appears at the end of this chapter. The TLG editions from which the texts come are listed in ch. 1, n148.

[ back ] 189. “… it could be that the warriors are assuming hostile positions not against one another, but against the boar itself. In this case the fragment would belong not to the struggle with the Curetes but to the actual hunt” (Barrett 1972:118).

[ back ] 190. ἔνθεν need not to be read as an ad oculos deictic mark; it could simply work as an am Phantasma location of the overall frame in the discourse memory of the performer.

[ back ] 191. See, too, Mimnermus fr. 2.11–15 πολλὰ γὰρ ἐν θυμῶι κακὰ γίνεται· ἄλλοτε οἶκος / τρυχοῦται, πενίης δ’ ἔργ’ ὀδυνηρὰ πέλει· / ἄλλος δ’ αὖ παίδων ἐπιδεύεται, ὧν τε μάλιστα / ἱμείρων κατὰ γῆς ἔρχεται εἰς Ἀΐδην· / ἄλλος νοῦσον ἔχει θυμοφθόρον· “Many events are evil for one’s spirit. Sometimes the house is devastated, and the works of poverty are painful. Again, another lacks sons, and while desiring them most of all he goes to Hades, beneath the earth. Another has a spirit-destroying illness”; and, see also Pindar Pythian 4.142–145, where αὖ occurs within a little genealogical list: μία βοῦς Κρηθεῖ τε μάτηρ / καὶ θρασυμήδεϊ Σαλμωνεῖ· τρίταισιν δ’ ἐν γοναῖς / ἄμμες αὖ κείνων φυτευθέντες σθένος ἀελίου χρύσεον / λεύσσομεν “One and the same heifer was the mother of Creteus and of Salmoneus of stout thoughts. From those we in turn were procreated, in the third generation, and stare at the golden strength of the sun.” αὖ occurs significantly between the first-person plural mark (representing the new entry) and the social deictic pronoun κείνων (representing the previous entry).

[ back ] 192. Horace’s imitation of these lines (Odes 1.9. 5–8) stops at the pouring of wine.

[ back ] 193. For ὀρθόω connoting musical performances, see Pindar Olympian 3.3; Isthmian 1.46; Isthmian 3/4 56.

[ back ] 194. See Stesichorus fr. 63 δεῦρ’ ἄγε, Καλλιόπεια λίγεια “Come here, clear-sounding Calliope,” which is an incipit. Sappho fr. 127 is another incipit: δεῦρο δηὖτε Μοῖσαι χρύσιον λίποισαι “Here, once again, the Muses, while leaving the golden …” See also fr. 128: δεῦτέ νυν ἄβραι Χάριτες καλλίκομοί τε Μοῖσαι “Now once again, tender Graces and Muses of the beautiful hairs …” Further comments about the function of δή + αὖτε in lyric are below.

[ back ] 195. Cf. Horace Odes 3.2.14 mors et fugacem persequitur virum.

[ back ] 196. Sappho fr. 1.15, 16, 18; 22.11; 83.4; 127.1; 130.1; Archilochus fr. 88; Alcman fr. 59a.1; Anacreon fr. 4.1; 11a.6; 11b.1; 13.1; 26.1; 31.1; 49b; 55.1; 56.1; 67; 68.1; 83.1; Hipponax fr. 122.

[ back ] 197. Although their quantitative presence is not so small. See Table 3 at the end of this chapter.

[ back ] 198. Schiffrin’s (1987:31) early definition of discourse markers is “sequentially dependent elements which bracket units of talk.”

[ back ] 199. For a discussion of clitics and particles in the so-called second position prosodically reinforcing the first group of words in a sentence, see above, p. 207.

[ back ] 200. This point supports Bakker’s claim that “epic’s deictic orientation … is centered on the now of the performance,” as well as the idea of “syntax of movement” (see Bakker 2005:167–176, with 175 for the quotation, and Bakker 1997b:54–85 respectively).

[ back ] 201. See above, n64.

About

Fellowships

The Library

Programs