This book is a study of the full range of techniques for presenting speech in the Homeric epics, something that has been assumed not to exist in Homeric poetry. A range of speech presentation strategies can indeed be found in Homeric poetry, and they are well worth our attention. If there is such a thing, this fact alone changes our understanding of Homeric narrative in a basic and significant way. Moreover, individual pieces of this overall approach to presenting speech do not fully make sense if they are studied in isolation from other parts of the system. The main conclusion of this book is that each speech presentation technique has a stable set of functions and effects, and that these individual techniques add up to a unified, consistent speech presentation spectrum that underlies the entirety of both poems at all narrative levels. Different kinds of narrators use varied but overlapping subsets of this spectrum; a given speech presentation technique does not have one function or effect for a particular narrator or type of speech or narrative context, and a different function in a different context. Such a unified speech presentation spectrum gives a new kind of unity to the poems, which we must take into account in order to understand their power and effect as narratives.
This study draws mainly on two bodies of theory for its approach to speech presentation, namely narratology and pragmatics. Narratology, a branch of literary theory concerned with how stories are constructed and told, has developed widely used terminology for describing speech presentation. However, there are important aspects of Homeric speech presentation that narratology cannot explain. For example, whether a given speech is part of a conversation, and where it falls in a conversational exchange, affects how the speech is presented. So does what the speech is trying to do (or, in other words, the speech act type). Directives are presented in a much greater variety of ways than questions or emotional exclamations like vaunts or laments, but the speech presentation typology developed by narratology does not address this aspect of Homeric speech presentation. To complement narratological approaches to speech presentation, we must turn to ideas developed by subfields of linguistics. In addition to speech act theory, we need pragmatics, and in particular, the notion of the "move" of a speech, which relates various features of an utterance to its position within a conversation. Finally, expressivity, or the features of an utterance that mark it as the speech of an individual with feelings about what he or she has said, can explain why some speeches are presented with direct quotation and others, apparently identical in content and function, are presented differently. This study draws on consensus ideas of these different concepts developed within the individual disciplines from which each has emerged, and uses them in new combinations to shed light both on Homeric speech presentation and on each other.
This approach to speech presentation has several benefits. It provides a more well-rounded view of individual speech presentation techniques than we get from studying each one in isolation. Some techniques that have been almost completely neglected, like speech mention, can be understood for the first time as positive contributors to the narrative texture of the poems and not simply fallback mechanisms for times when the "default" presentation technique of direct quotation is not needed. Conversely, we gain a much richer appreciation of what direct quotation brings to the poems if we see it not simply as the default option for presenting speech, but against the larger background of what the other ways of presenting speech contribute to the narrative texture of the poems. Moreover, if both narratology and linguistics are combined into one set of interpretive tools for studying speech presentation, Homeric speech presentation takes on a qualitatively different and more compelling force as a way of understanding the poems. Not only does speech presentation make more sense with this interdisciplinary approach than if either narratology or pragmatics is used alone, but individual speeches themselves yield new insights when we look at them as human speech that can be understood along the same lines as the recorded speech of contemporary speakers that pragmatics uses for its inquiries. If the Homeric epics present speech partly according to pragmatic features that characterize "real" spoken speech, the speech in the poems is implicitly brought closer to the speech patterns of non-fictional humans. Just as direct speech has pushed other modes of speech presentation to the periphery of studies about Homeric speech presentation, so, too, the main narrators of the poems have generally been privileged over the characters as narrators in analyses of Homeric narrative. A speech presentation spectrum brings together into one cohesive unity not only individual modes of speech presentation, but also different narrators and different narrative levels. As with different ways of presenting speech, different narrators play complementary roles as presenters of speech that, together, form the Homeric poems.
These conclusions are important for two related reasons. First, this unified speech presentation spectrum entails a new kind of unity not only for each poem individually, but for the Iliad and the Odyssey together. Different kinds of speech acts, different kinds of narrators, and both Homeric epics draw on the same speech presentation spectrum to create a huge range of different effects. That the same set of building blocks creates all these different effects suggests a powerful unity both within and between the Homeric epics. The best way to explain this from a theoretical perspective is the implied author, defined by one scholar as "not the narrator, but rather the principle that invented the narrator, along with everything else in the narrative, that stacked the cards in this particular way." The implied author, although a perfectly valid narratological construct, is generally ignored in narratological work focusing on Homeric poetry. In fact, the main narrators and their narrative practices complement those of the characters, a fundamental aspect of the narrative texture of the poems. I attribute this to the implied author, the voiceless arranger of the voices found in the text across all narrative levels. Although scholars have studied some of the differences between the main narrator and the characters as narrators, this work has generally not gone on to consider the overall effect of the combination of this type of main narrator with this type (or types) of character narrator. The overall effect of the combination is, in fact, the poems themselves. The notion of the implied author appears here only in my Introduction and Conclusion, because the importance of the implied author is the cumulative result of all the chapters rather than the conclusion of any one chapter individually. The unity of the poems that I connect with the implied author is nonetheless one of the book's key conclusions.
The presence of an implied author points to a final benefit of the approach in this book, which is not a conclusion so much as a way of looking at the poems that I hope will contribute to the recent intellectual climate of seeing the Homeric epics both as orally based poetry and as fundamentally human and accessible forms of behavior. Both fictional storytelling and conversational interchange in the Homeric poems can be understood in ways that are similar to those applied to modern fictional narrative—namely the speech presentation spectrum and the implied author—and to the conversations of "real" people who are directly available for linguistic study. An overblown sense of Homeric epic as unique and inexplicable in literary history can stand in the way of understanding the poems. This has been particularly troublesome in relation to the tradition of oral poetry that underlies the Homeric epics. Despite important work showing the range and capacity of oral poetic traditions around the world, some scholars' minds have simply boggled at the idea that the Homeric epics could themselves be oral. Accordingly, some readers have abandoned the orality of the epics because they are undeniably subtle and complex fictional narratives. This is a pity, because a persuasive body of work shows that sophisticated aesthetics can exist in poetry that relies heavily on formulaic language. Similarly, theoretical frameworks for understanding speech presentation that have been developed for modern fictional narrative can profitably be applied to Homeric poetry to a greater extent than has been generally done, resulting in the insights that (for example) the poems do use unified speech presentational systems. In a broader sense, this both changes our idea of how Homeric poetry fits into the overall "storyline" of Western literature, and makes the poems themselves seem more intellectually accessible, insofar as we can use a modern theoretical construct to make sense of the product of an ancient and in some ways very alien oral poetic tradition.
In an effort to make this study as widely accessible as possible, I have intentionally avoided detailed discussions of theoretical background for its own sake. This may displease devotees of the various disciplines on which I draw, who might prefer a more extensive and exhaustive treatment of the theoretical framework both of relevant concepts developed by their discipline and of current debates going on within the discipline. I hope that this study will contribute to the individual disciplines from which it draws its terminology, insofar as it juxtaposes theoretical perspectives in new combinations, thus offering fresh views of each of the individual areas of study. But this book is really aimed at the broader audience of people who are interested from whatever point of view in the fundamental human pastimes of storytelling and presenting speech within the stories we tell, as we see those activities at play in the Homeric epics. I hope that this book is in the best sense a humanist study—in other words, a study of (one aspect of) what it means to be human.
Homeric Speech and Narratology
The most influential broadly based narratological works on Homeric poetry have included speech presentation among the topics they cover, but these studies have not led to the kind of fruitful and transformative insights about speech that they have produced about other features of Homeric narrative. This is because these scholars have implicitly or explicitly dismissed the idea that the range of techniques for presenting speech in Homeric poetry exists as a cohesive, unified system, or that individual parts of the system besides direct speech are meaningful in their own right. This is particularly clear in discussions of indirect speech used by the main narrator. One study categorizes indirect speech as the exception to the default option of direct speech: when indirect speech is used, it presents speeches that for various reasons are peripheral to the main storyline or where the exact wording is not necessary. Richardson makes important and valid observations about indirect speech, namely that the narrator uses indirect speech primarily to report commands, whereas characters regularly use indirect speech to report what other characters have said.
These observations are certainly true as far as they go, but they lack the wider context that would make them useful in explaining and understanding Homeric speech presentation more broadly. For example, Richardson does not discuss when the exact wording is or is not necessary when presenting speeches, and he does not mention that not only indirect speech, but also direct speech, is used primarily for presenting directives. So, directives are common in both direct and non-direct speech (a collective term for all forms of speech presentation other than direct quotation). This suggests that non-direct speech, far from differing fundamentally from direct speech, most frequently presents the same types of speech as direct speech does. What is the difference, we should ask, between directives that the main narrator quotes directly and apparently identical ones that are presented with indirect speech? Moreover, the reader is left wondering why characters and narrators use indirect speech for different kinds of speech reporting, or what the effect of this might be for the poems overall. Thus, although narratology has advanced our understanding of other aspects of Homeric narrative, it has not described a speech presentation spectrum for the Homeric poems, and accordingly has left important aspects of Homeric narrative unexplained.
Terminology for Speech Presentation Techniques
Scholars studying how speech is presented use two overlapping perspectives for classifying the different possibilities. Narratologists generally focus on the degree and nature of the resemblance between a presentation of speech and the speech that is being presented; linguists are more interested in the specific language features of different techniques of presentation that distinguish one from another, such as how deictic words like "you" or "tomorrow" are handled. Both of these approaches contribute to the definitions of speech presentational terms used in this study.
Most scholars of narratology who study speech presentation have related the different techniques to each other in one of two ways. Some have presented a set of discrete possibilities, each of which is clearly defined but at the same time contains a good deal of variety. An example of such a system is Genette's three categories of narratized speech, transposed speech in indirect style, and character speech. Others have favored a scalar or spectrum approach, either by defining a larger number of categories of speech presentation that are not as clearly distinguished from one another or by questioning the usefulness of the whole idea of "categories" in analyzing speech presentation strategies. Fludernik, who critiques the process of naming and categorizing methods of presenting speech and thought more thoroughly than any other person who has written on the topic, points out that although the scalar approach she favors has the advantage of describing most accurately the range of presentations that are actually found in narrative texts, this approach can become bogged down in its own subtleties and vitiate the categories it comprises.
Sternberg notes that many of the features that supposedly characterize individual modes of speech presentation in fact do not, and that direct, free indirect, and indirect discourse all can do most of the things that the other modes do, and/or things that they are not supposed to be able to do. He is in essence following a spectrum approach to these terms, although he does not say so. Perhaps as a result, he does not sufficiently acknowledge that although it is indeed an overstatement to call it a rule that (for example) indirect speech does not admit imperatives, it is perfectly true to say that this and similar statements remain accurate and useful as general tendencies. These categories do not lose their legitimacy as categories because their boundaries are fluid.
All of these scholars are striving to develop systems for describing speech presentation that can explain as wide a range of texts as possible. For Homeric poetry, which contains a small number of different methods of presenting speech that can be fairly clearly distinguished one from another and that display relatively little variation within individual categories, there is no need to worry about describing a wide range of subtly different approaches to speech presentation. Accordingly, this study is based on a few discrete categories of speech presentation, whose definitions are given below.
Traditionally, the notion of "mimesis" has been used as a criterion for organizing different modes of speech presentation. Since Plato first used this as a term of narrative analysis in the Republic, it has been used to refer to so many different aspects of narrative that it brings up as many problems as it solves as a term of analysis. Recent discussions of direct speech have proposed that what it provides is not the imitation of the "actual words" of the supposed speaker, even in non-fictional situations where there is such a thing as "actual words." No one interpretation has taken the place of "imitation of the original." In one view, what is important about direct speech is that "a voice other than the narrator's appears to take over"; other explanations of the effect of direct speech include offering the appearance rather than the reality of an original utterance, or requiring the audience to participate more directly as interpreters of the narrative.
A different way of distinguishing speech presentation modes from one another concerns the perspective of the deictic words in the presented speech, such as pronouns and temporal words. The term "deixis" covers the various words and grammatical structures (such as verb tenses, personal pronouns, and demonstratives) that take their meaning from the specific time and place of the utterance in which they occur. For example, the pronoun "you" has no intrinsic meaning; it has only a relational meaning based on some "I" that appears in or is implied by a particular utterance. Similarly, "here" or "tomorrow" do not designate any stable place or time. They refer to a place or time relative to the spatio-temporal perspective of the sentence in which they occur. In direct speech, the verb of speaking and the speech are deictically separate: the deictic words in the reported speech present the perspective of the quoted speaker. So, in the sentence, "Mary said, 'I left my hat on the bus yesterday,'" her speech refers to person and time in terms that are oriented toward herself, not toward the reporter who says, "Mary said …" In indirect speech, deixis in the reported speech is oriented to the reporting speaker. Mary's speech about her hat would then look something like, "Mary said that she had left her hat on the bus yesterday," or "Mary said that she had left her hat on the bus last Friday." Here Mary and the time of the incident are referred to from the perspective of whoever is telling us about Mary and her hat. In free indirect speech, some deictic features are oriented toward the speaker of the presented speech and some toward the voice that presents Mary and her speech.
The speech presentation spectrum used here is essentially the one that is used by the linguistically oriented critics Leech and Short. Their spectrum contains five methods of speech presentation: speech mention, indirect speech, free indirect speech, direct speech, and free direct speech. The most concise option for presenting a given speech act is "speech mention," which tells the audience that an act of speaking took place without giving any indication of the words that the speaker used. This approach treats speech as a narrative event. In Homeric poetry, speech mentions generally take the form of speaking verbs without direct objects. In "indirect speech," the text goes some way toward presenting the form and/or the content of an utterance.
A great deal of recent scholarship has focused on the variously named phenomenon that I am calling "free indirect speech." Free indirect speech, unlike indirect speech, lacks an introductory verb of speaking. In traditional accounts of free indirect speech, it combines the perspectives of indirect speech (by shifting verb tenses and personal pronouns to the presenting speaker's deictic orientation) and direct speech (by presenting other features, such as demonstratives, from the deictic perspective of the speaker of the utterance). Thus, free indirect speech can include the voice of both a presenter of the speech and the speaker of that speech. Most scholarship claims that this technique originated with modern fiction, and it is closely associated with defining features of modern fiction such as stream of consciousness.
The following quotation from Jane Austen's Emma
No invitation came amiss to her. Her Bath habits made evening-parties perfectly natural to her, and Maple Grove had given her a taste for dinners. She was a little shocked at the want of two drawing rooms, at the poor attempt at rout-cakes, and there being no ice in the Highbury card parties. Mrs. Bates, Mrs. Perry, Mrs. Goddard and others, were a good deal behind hand in knowledge of the world, but she would soon shew them how every thing ought to be arranged. In the course of the spring she must return their civilities by one very superior party—in which her card tables should be set out with their separate candles and unbroken packs in the true style—and more waiters engaged for the evening than their own establishment could furnish, to carry round the refreshments at exactly the proper hour, and in the proper order.
This passage uses the tense and pronoun shifting of indirect speech (here "she" and "was" at the end of the second line rather than Mrs. Elton's own "I am"); it retains the quoted speaker's perspective for deictics such as "now," as in direct speech; and it uses some expressive and stylistic features not permissible in indirect speech, such as vocatives, exclamations, and word choice (here the clue that the passage is FID rather than the narrator making fun of Mrs. Elton is the italicized she).
It has recently been argued that a number of premodern literatures do contain free indirect speech. One persuasive reading is that in fact, both of these arguments are at least somewhat accurate: free indirect speech is used to present speech quite regularly in premodern texts (including Homeric epic), but its use for the presentation of thought becomes widespread only in the nineteenth century. At all events, this technique has received essentially no attention in relation to Homer. Though free indirect speech is not one of the most prominent speech presentation strategies in the Iliad and Odyssey, it is by no means absent: both the main narrator and the characters use it, generally as a continuation of a speech that begins as unambiguously indirect speech. Free indirect speech in Homer functions, for the most part, just as theoretical treatments of it would lead us to expect. We recognize instances of free indirect speech mainly because they follow instances of indirect speech with which they are associated. And, although free indirect speech frequently entails emotional effects (empathy, irony) when used to present thought, it generally has either no particular emotional impact or an ironic effect when used to present speech. Probably scholars have not pointed out free indirect speech in Homer because it lacks the explicitly expressive elements found when it is used in modern fiction. In fact, other premodern texts besides Homeric epic contain free indirect speech that lacks such expressive markers, and these expressive elements are not a requirement for free indirect speech.
Free indirect speech in Homeric epic is much shorter than the example from Emma, but like the Austen quotation, it usually follows indirect speech, and it presents ambiguous information that might belong either to the presenting narrator or to the speaker being presented. For instance, in Iliad 9, when Odysseus tells Achilles that Agamemnon will swear an oath that he never slept with Briseis, he includes a relative clause in free indirect speech.
??? ????? ????? ???????
?? ???? ??? ????? ?????????? ??? ???????
? ????? ?????, ????, ? ?? ?????? ? ?? ????????.
He will swear a great oath
that he never entered into her bed and never lay with her
as is natural for human people, between men and women.
Verse 275 presents the oath with indirect statement, and the relative clause in 276 might be either part of the oath or Odysseus' aside to Achilles. Although this instance of free indirect speech lacks features like irony or clear signals of speaker focalization, ambiguity about whether it belongs to Agamemnon's oath or to Odysseus' presentation of the oath marks it as free indirect statement.
"Direct speech" is the only speech presentation strategy in Homer that has already been widely studied. Studies of direct speech in Homer have generally assumed that its defining characteristic in comparison to indirect speech—and therefore, the reason it predominates so heavily over indirect speech—is its faithfulness to a putative "original" speech and its vividness; a cogent critique of faithfulness and reproducibility suggests that direct speech is distinguished by its potential to reproduce whatever can be reproduced about an "original" utterance, although the reproducible elements fall short of the entirety of the reported speech event. That is to say, we can imagine that we are hearing the character's own words (if not a complete and exact replica of a speech event) after a verse like ??? ?? ????????????? ??????? ????? ???? ???????? ("Then in answer again spoke Achilleus of the swift feet"; 12 instances in the Iliad), whereas we have no such expectation for an expression like ?????? ?? ????? ?????? / ????????? ("[He] told his companion, Patroklos, / to sacrifice to the gods," Iliad).
Linguistics and Speech: Speech Act and Move
Various linguistic attributes of individual speeches strongly affect which of the speech presentation techniques just described is likely to present the speech. These features include what kind(s) of speech act are depicted in the speech; how the speech functions within a conversational exchange; and the subjective aspects of the speech, or its expressivity. Indirect speech in Homeric poetry, as I have already mentioned, is supposedly associated with directives, but in fact, directives are characteristic of Homeric speech in general. This basic misunderstanding shows that Homeric speech presentation cannot be properly understood without taking into account the nature of the speech being presented.
"Speech act type" is a way of classifying speeches that grows out of the work of Austin. His central insights—that speech not only states facts, but does things, or states feelings; that utterances cannot be understood without a context; and that utterances can be fruitfully classified and studied based on these features—have formed the basis of speech act theory. The typology of speech acts that Austin created, however, has not won general acceptance. There is no single criterion—or group of related criteria—that is consistently used as the basis for the categories in his typology. Rather, he seems to take a quite impressionistic approach, both in what distinguishes one category from another and in what justifies the existence of something as a category at all. The five families of speech acts Austin proposes at the end of How to Do Things with Words do not elicit unambivalent agreement even from Austin himself. Almost immediately, commentators began to overhaul, rework, and criticize Austin's categories. Even today, there is no particular "speech act typology" that is generally considered to be the consensus approach. Instead, different versions of speech act typologies proliferate, and scholars working on aspects of speech act theory tend to produce their own typology of speech acts as part of their inquiries.
Different speech act typologies include different speech act types, and are based on different criteria for classifying the individual speech acts. The speech act typology used in this book categorizes different types mainly according to what they are about—facts, emotions, and/or actions—and secondarily according to the orientation of the speech act toward the speaker, toward the addressee, or (sometimes) toward a third party. Speaker and addressee orientation will also play a central role in defining various subtypes of the large categories. The speech act types that I use in this study, which will be discussed in more detail in the next section, are directives (speech about action), assertives (speaker-oriented speech about fact), questions (addressee-oriented speech about fact), and emotives (speech about feelings).
How a given speech is presented in Homeric poetry relates not only to what kind of speech it is (its speech act type), but also to its role within an interactional exchange. In an unjustly ignored but important point, Bassett asserted many years ago that the kinds of speech that appear in non-direct forms are those which are "outside of the dialogue" . Homeric speeches that form part of an exchange (a conversation) differ in both content and presentation from those which appear singly. For example, the kinds of speech that are most frequently presented with non-direct speech are the same kinds that tend to appear singly rather than in conversational sequences when they are directly reported. This includes not just orders, but similar speech act subtypes like oaths, prayers, and so forth. Hence, we need to know where a speech falls in (or outside of) a conversational structure in order to understand speech presentation.
A further development of speech act theory adds information about how a particular speech works within an interactive structure. A "move" is essentially a speech act in a conversational context: "speech act" defines a particular utterance as a directive, assertive, and so forth in terms of particular linguistic and grammatical features of the utterance, whereas "move" concentrates on how a particular utterance operates in its context. Kroon defines a move as "the minimal free unit of discourse that is able to enter into an exchange structure. . . . A move usually consists of a central act (which is the most important act in view of the speaker's intentions and goals) and one or more subsidiary acts, which also cohere thematically with the central act."
The same basic categories apply to moves as to speech acts (a move can be a directive, assertive, question, or emotive), but the interactive perspective of move terminology entails a second dimension. Moves are classified both by what they are trying to do and by where they are in the interactional structure of the exchange in which they occur. So, a move can be initiating, reactive, or problematic, depending on whether it begins a new topic or theme (initiating), responds satisfactorily to a topic begun by a previous move (reactive), or somehow objects to or refuses to go along with the previous move (problematic). Problematic moves are both reactive and initiating at the same time. Most often, one initiating and one reactive move form an exchange, but from time to time a reactive move itself elicits a reaction, or two different speakers react to the same initiating move, particularly in a conversation that involves more than two speakers.
The following exchange between Iris and Achilles illustrates most of the permutations of what types of moves there are, how a move overlaps with an individual speech, and how individual moves interact to form an exchange. When Iris goes to Achilles in Iliad 18 and tells him to defend Patroclus, this is an initiating directive. Rather than immediately go along with this directive, Achilles asks not one but two questions about it. These are problematic moves: they are reactive insofar as they respond to the directive, but they are also initiating because they invite a response from Iris. Iris answers both questions in reactive assertive moves, and then repeats the directive a second time after answering Achilles' second question. This directive does not constitute a new move, but a continuation of her initial directive move at. This exchange illustrates several possibilities for how move and individual speech overlap. The first speech, Iris' directive, contains one move. It begins and ends with repetitions of one directive, while the middle section consists of subsidiary assertive acts that are intended to persuade Achilles to follow the directive. Achilles' question consists of a single question with no subsidiary acts. Iris' final speech at contains an assertive in answer to a previous question (reactive) and a directive. Here the directive is not a new move because she has already given this directive once before, but other speeches commonly introduce an initiating directive move after a reactive move. In contrast to direct speeches like this one, which contains both a reactive assertive and an initiating directive, non-direct speeches usually contain just a single move.
Move terminology offers one way of describing the conversational dimensions of speech. Expressivity provides another. As we will see, the move of a particular speech in Homeric epic and the expressive features it contains, as well as its speech act type, are relevant for understanding how the speech is presented. Expressivity offers a useful tool to describe in a quantitative manner what direct speech conveys that non-direct speech usually does not, and more importantly, what effect this has in a narrative. Expressivity is a somewhat slippery catch-all term covering the features of an utterance that make it the speech of a particular person with feelings about what he says. What distinguishes linguistically oriented discussions of expressive features from what a narratologist might say about (for example) focalization is primarily their focus on understanding the vehicles for conveying emotions and judgments rather than the specific emotions or judgments conveyed. Moreover, expressive elements may convey nothing more than that a particular speaker is the speaker (such as first-person forms) without implying any additional feeling on his part. Besides first- and second-person forms, expressive elements also include vocatives, exclamations like ? ???, and language that contains evaluations, emotions, and reasoning by the character speaking. As we will see, the interchange of conversation itself has an expressive value in Homeric poetry. Systematically bringing this idea to bear on Homeric epic has several benefits. We can see in a new way just how much of direct quotation in Homeric epic consists of expressive features rather than propositional content, and by extension, how central that expressive quality is to the poems. Non-direct modes of speech presentation have expressive qualities, too, which we are more likely to notice if expressivity is identified as one of the dimensions of speech presentation. These forms of expressivity make a positive contribution to the shape and effect of Homeric narrative that complements the more vivid and noticeable expressivity of direct quotation.
Defining a Speech Act Typology: Speech Act Types and Subtypes
As with terms for modes of speech presentation derived from narratology, I have chosen a system of speech act terminology that uses the fewest and clearest available terms that are nonetheless up to the task of describing speech in the Homeric poems: too much terminology is cumbersome to use and puts off nonspecialists, whereas too little or insufficiently specific descriptive language leads to analysis that is too general to be useful, or that leaves out important features of what is being studied. Questions and assertives are both about facts. Whereas an assertive is speaker-oriented ("The cat is on the mat"), questions seek out a position about some fact from the addressee ("Is the cat on the mat?"). An emotive speech act such as "I wish the cat were on the mat!" presupposes a fact and gives the speaker's feelings about it. Directive speech acts are aimed at getting some action accomplished. The directive "Put the cat on the mat" presumes certain facts, such as the cat not (yet) being on the mat. Commissives, where the speaker commits himself to a future action, hardly ever appear as the main speech act in Homeric epic, and accordingly are not included as part of this taxonomy. Instead, promises function as assertives, either to provide inducements to comply with a directive to the addressee, which is the main act, or in some more diffuse way to provide a guarantee for what the character presenting the promise is trying to achieve with his own speech. In Homeric speech, promises are best understood not as committing the speaker to a particular course of action, since that is rarely the main point of the utterances in which they occur, but instead as one of a variety of assertions that characters make to each other in order to produce compliance with a directive.
The category "directive" contains a number of subtypes depending on how obligatory the directive is and whether the directive advances the interests of the speaker, the addressee, or both. The speaker may give the addressee an option not to obey: noncompliance is essentially not available for an order, but is possible in the case of a request or a plea. The proposed action may benefit the addressee as well as the speaker, as in a suggestion like "Let's X," the subtype to which battlefield exhortations belong. An invitation makes an optional directive in which the speaker has a moderate interest. Supplication, a plea conducted in a particular way, benefits primarily the speaker, who generally tries to persuade the addressee to go along with his speech with emotional inducements of various kinds rather than by asserting his own power—usually the speaker has little or no power relative to the addressee—or by trying to align their interests. Directives may have negative consequences attached, generally within the speaker's control (a threat) or not (a warning). Instructions provide a series of directives for accomplishing a particular end in which the speaker does not have a strong interest. Instructions take the general form "If you want to accomplish X, do A, B, C," but generally speaking, the speaker of instructions is involved because of knowledge about X rather than interest in getting it done. Permission, a reactive directive, falls outside such a scheme, as does prayer, a kind of specialized or exaggerated plea in which a mortal issues a directive to a god.
Messages in Homeric poetry generally convey directives by means of an intermediary (a messenger), who has a moderate interest in the directive in addition to the more lively interest of the originating speaker. These are essentially two-stage directives. First, the originator of the message gives an order to the messenger to deliver a particular message to a third party. The messenger satisfies this first directive by setting out on a journey to the intended recipient of the message. Second, the messenger delivers the message, usually itself a directive, to the recipient. The messenger has an interest in seeing that the recipient acts on the message, and messengers regularly urge recipients to comply with a message even where their own emotions might seem to align their interests with the recipient rather than the originator of the message. When a speech is not quoted directly, sometimes it is impossible to tell from the context what kind of directive is depicted (an unspecified subtype). Indeed, characters who report directives often do not distinguish among different subtypes, presenting directives simply with a form of ?????? and an infinitive. All of these directive subtypes either use directive sentence types (usually an imperative, infinitive, or hortatory subjunctive), or they are not quoted and leave unclear what the subtype is. Directives that are directly quoted but whose content does not clearly convey that the speech act is a directive are implicit directives. The context identifies these speech acts as directives, but the speech act itself does not make this explicit. Although implicit speech acts are very common in most languages, especially for directives, they are quite rare in Homeric poetry. Indeed, the main narrator often points out explicitly speeches where the speaker's intentions are significantly at variance with what he actually says. This implies that the audience was not accustomed to speech that appeared to say one thing but actually meant something else.
The most common subtypes of assertives, statement and reply, are distinguished by whether they are oriented toward the speaker (statement) or the addressee (reply). Other kinds of assertives bring in various kinds of third-party authority for the statement being made. Oaths guarantee a statement that the speaker makes about himself, generally by invoking a god. Oaths can be about a state of affairs either past, present, or future. Clearly an oath referring to the past or present is an assertive; because Agamemnon's oath that he did not sleep with Briseis and oaths about future actions are presented the same way, all Homeric oaths are classed as assertives. Prophesy, a related kind of assertive that is less speaker-oriented, makes a statement about the future that is about something other than the speaker and is guaranteed by a god or some kind of supernatural intervention. Song is classified as a kind of assertive because it clearly involves a commitment to the accuracy of the speech act, since Homeric characters refer to poets in terms of their knowledge of what they sing about or praise them for singing as though they had been personally present at the events in their songs. It seems to be mainly addressee-oriented, insofar as characters can ask a poet to sing some particular song and poets are not described as singing when no one is present to hear them. At the same time, song differs from other speech act subtypes, because although an audience is necessary and the audience is often depicted responding to a song, the listeners are not so much addressees as an audience. This is a different kind of interaction than most conversation, even though interaction between a speaker and a listener takes place.
Questions have few variations that depend on the speaker or addressee orientation. Rather, questions vary primarily in relation to particular social contexts, where a small group of speech act types are used in different ways depending on the situation. These variations are important and interesting, but they do not translate into speech act subtypes. Questions can become more speaker-oriented than the prototypical addressee orientation in two ways. First, questions regularly appear in directive rather than interrogative sentence types, where the speaker orientation of the directive form heightens the speaker orientation of the question. A smaller subset of questions poses a more speaker-oriented question subtype by asking not for information unknown to the speaker, where the focus is on the addressee's knowledge of that information, but for information that the speaker already knows. Here, the goal of the question is not finding out the requested information, but exercising power by asking a question to which the speaker already knows the answer.
Emotives, as Risselada notes, are a grab-bag category. Emotive subtypes are mainly speaker-oriented, but challenges aim at producing fear in the addressee. Encouragement, conversely, seeks to create a positive frame of mind in an addressee. Breaking down emotives broadly into those that express positive emotions and those that express negative ones, we find in the former category vaunts (satisfaction that an enemy is dead), greetings (pleasure in the arrival of someone), and farewells (good wishes to a departing guest). Negative emotions motivate laments (sorrow on behalf of a dead person). Wishes, which express dissatisfaction with the current state of affairs insofar as they express the speaker's desire for a different one, also belong in this category. Emotives about sorrow express dissatisfaction with a current state of affairs without expressing a clear preference for something else instead; they are weaker and less ritualized than laments. Rebukes have elements of both an emotive and a directive speech act: the speaker expresses dissatisfaction with the current state of affairs and (at least implicitly) a directive to the addressee(s) to do something different. These two elements occur in different proportions in different rebukes. Some instances primarily express dissatisfaction with the current state of affairs and make a directive to change it only by implication rather than explicitly stating any particular action or result that the speaker desires. Given the diffuse nature of the directive component of such rebukes, and the metadirective quality of speech in general, it seems most appropriate to view these as emotives. Other rebukes rather perfunctorily refer to the speaker's dissatisfaction as an inducement to go along with a much more fully developed directive, and these make most sense as directives.
One subtype of emotive speech act consists of speeches that the speaker makes to himself. The same kind of language that introduces speech also introduces these monologues or soliloquies, most commonly the formula ??????? ?? ??? ???? ???? ?? ?????????? ????? ("And troubled, he spoke then to his own great-hearted spirit"). The ???? + (?)???? combination is one of the most common ways to introduce direct quotation, so it is clear that these speeches are presented as though they are direct quotations, too. Whether this is literally the case has been widely debated. For my purposes, what is important about these is that they are presented as if they were speech, not whether they are "actually" speech or thought. All of them in some way convey the speaker's emotions—hence the frequent appearance of ???????, "troubled," in the introductory verses and of emotional exclamations at the beginning of the speech—so they are classified as a subtype of emotive speeches.
For the most part, assertives, questions, and directives have typical sentence patterns that correspond to the speech act type. However, sentence type and speech act type do not always coincide. Questions in Homeric poetry are most often presented with interrogative sentence types, but they are also presented not only with directive sentences types (as noted earlier), but also with assertive sentences: "I want to know who broke that glass." Conversely, an assertive like "It's cold in here" can be a statement of information (someone has asked, "Is it cold in here, or is it just me?") or an implicit directive (to close the open window). Where the sentence type and the speech act type of a given speech act differ, the "speech act type" entered in my database is the sentence type, and the "speech act subtype" is the actual function of the utterance in its context. Usually this is simply a matter of using a nontypical sentence type (as in the common case of a question presented as a directive), but some of these speech acts are implicit.
Database Content and Design
The data that underlie the majority of this book are collected in a FileMaker database that I constructed and then revised several times over a multi-year period. In hindsight, there are some features that I would have designed differently, and in spite of many iterations of careful editing and standardizing, I am sure that mistakes and inconsistencies remain. The database contains information about each presentation of speech in the Iliad or Odyssey. The FileMaker format allows searches that not only tally a single feature, such as the number of speeches presented with direct quotation, but also collate multiple features of speech presentation, such as directives in indirect speech presented by the main narrator of the Iliad, or directives presented by characters except for those presented by Odysseus. For each speech presentation, the database includes the citation (work, book, starting and ending verse numbers); the length of the speech in verses based on the number of verses in which some part of the speech is presented; the narrative level at which the speech is presented (main narrator, character narrator, ambiguous between the two, third level of character narration); the names of the speaker and addressee of the speech, as well as their genders; the Greek word(s) of speaking that introduce or present the speech and the verse number(s) in which the words appear; for non-direct speech modes, any subordinate clauses depending on the verb of speaking; and the speech act type(s), subtype(s), and move type(s) of the speech. Speeches embedded within character speech collect most of this information a second time for the characteristics of the speech within which the speech is embedded.
Not all of these fields are equally important for my analyses, which focus mainly on three features of speech presentation: what level of narrator is reporting the speech; what kind of speech presentation is used; and various facets of the speech act being presented. Because this book focuses on how speech acts build on each other to create conversations, and what effect conversational exchange itself has on speech presentation, only the main speech act of each speech is tabulated. Some speeches have two main speech acts within them, but subsidiary acts (such as statements that explain why a directive should be obeyed) are not counted separately except under specific circumstances, which are explained as they arise.
Many directives presented by characters present the action that the enclosing speech is ordering the addressee to do. Thus, there is a sense in which directives are overrepresented in my database, since I have counted a speech like Odyssey 21.350–353 as a directive twice.
???? ??? ????? ????? ?? ?? ????? ???? ??????,
????? ?? ???????? ??, ??? ??????????? ??????
????? ?????????· ????? ?? ???????? ???????
????, ??????? ?? ????· ??? ??? ?????? ??????? ????.
Go therefore back into the house, and take up your own work,
the loom and the distaff, and order your handmaidens
to ply their work also. The men shall have the bow in their keeping,
all men, but I most of all. For mine is the power in this household.
The direct quotation overall presents the directive that Telemachus gives to Penelope, which in this case has two components. One is an action (attending to her own work), and the other is a speech (giving a directive to the maids). So, Telemachus' directive presents a second speech that is also Penelope's desired action, a further directive. I have tallied this speech overall as a directive from Telemachus to Penelope, but within that, I have tallied as a separate directive the directive by Penelope that Telemachus presents with indirect speech. The direct quotation is a directive presented by the main narrator; the underlined indirect speech presents a different directive, presented by a character to his addressee. These two directives have different properties as speech presentation—they occur at different narrative levels, are presented with different forms of speech presentation, and present two different speeches—so they are counted separately.
My main interest is in what Homeric characters say and how the audiences of the poems gain access to those speeches. The speech-related phenomena that I did not include in my data have been left out because in various ways they do not provide access to what people in the poems are saying: the references omitted are either not about speech, or are not about presentation. The largest category of arguably speech-related references that I have not included, from a numerical standpoint, is presentation of thought. Thought presentation and speech presentation differ substantially, particularly in premodern literature. Accordingly, thought presentation does not appear in this study, although clearly there is fascinating work to be done comparing speech and thought presentation in Homeric poetry. Some Greek words depict events that might or might not be speech. For instance, I have not counted objects of the verb ??????? as speech presentation unless the context requires that understanding, since ??????? depicts a wide range of actions, many of which are nonverbal. The boundary between presenting speech and presenting action is regularly a hard one to draw, and this verb gives a particularly clear example of that.
A short presentation of a speech that occurs at greater length elsewhere does not present the speech to the audience; rather, in order to help the audience follow the train of events or to position a speech within a conversational sequence, such cross-references point to a speech presentation that either has already happened or is about to happen. Cross-references are counted if they go beyond conventional references like ?? ????????, a phrase that regularly appears in character speech to position a particular speech within a conversational exchange and is normally not counted as speech presentation. For instance, an especially detailed presentation of what someone has already said, or one that characterizes a speech as a different kind of speech act than it appears to be from the main presentation, attempts to re-present the speech as something different from its original appearance, and so such instances are included. References to one's own speech, such as ??????? ?????, are not presenting the speech so much as characterizing it as a particular sort of speech. Finally, I have not counted references to talking about speech in general terms, such as how someone talks, because these do not present the content of a specific speech.
The first four chapters treat the four speech presentation techniques that make up the Homeric speech presentation spectrum. Each chapter begins with an overview of the basic properties and effects of a particular speech presentation technique, which includes how often and in what contexts that technique appears. These overviews analyze individual speech presentation techniques according to what kinds of speech acts they present, what kinds of moves they contain, and how they are distributed among various narrative levels. Typical examples illustrate these overall patterns. Each chapter ends with an in-depth analysis of examples of the technique that draw on common properties of that technique to create meaningful narrative effects connected to a key theme or story development. Thus, each chapter combines a broad overview of aggregate data patterns with detailed analysis of individual examples. This combination strives to present the enormous body of data that underlies this book in a way that is simultaneously broad, detailed, and manageable in scope. In these first four chapters, the Iliad and Odyssey for the most part are not distinguished from one another.
Chapter 1 focuses on direct quotation, the most common way of presenting speech in the Homeric poems. Chapter 2 focuses on free indirect speech, establishing that it does exist in Homeric poetry and showing what kinds of functions it has. These functions are consistent with the way free indirect speech has been shown to work in other premodern texts. Chapter 3, on indirect speech, shows that the characters use indirect speech for a wider range of functions than the main narrators do, but that all of these functions have underlying similarities related to the fundamental properties of indirect speech. In other words, focusing on indirect speech at the level of the main narrators, as previous studies have done, misses the full range of functions and effects that indirect speech has in the Homeric poems. Chapter 4 explores the effects and functions of speech mention. Speech mention differs from indirect speech mainly because it is more vague about the content of the speech being presented. This vagueness often helps to tell the story in a particular way, rather than simply being a fallback for speech that is not important to the narrative.
As we will see, each technique has a consistent set of effects across various narrative levels and speech act categories. Different speech act types and different narrators draw on different subsets of these effects that relate to the specific features of those speech acts or narrators. Characters have their own consistent approaches to speech presentation, but the focus of previous studies about Homeric speech on the main narrators as the main, only, and/or normative presenters of speech has obscured this fact. I argue that characters do have a coherent approach to speech presentation, one that overlaps with rather than differing entirely from what the main narrators do. Indeed, the speech presentation techniques themselves are remarkably stable and consistent. Each technique has a unique set of attributes and effects that makes a positive contribution to the overall narrative texture of the Homeric poems.
Chapters 5 (Odyssey) and 6 (Iliad) show how these different techniques work together in each of the Homeric poems. The first section of each chapter describes the speech presentation spectrum found in that poem, including the breakdown of different speech presentation techniques, different speech act types, and different move types. The characters in each poem show the same basic approach to speech presentation as the main narrator of that poem does. Differences in speech presentation in the two poems result from drawing on different parts of a stable set of functions and effects for a given speech presentation technique, not on using the same technique in completely different ways. Both chapters end with an extended analysis of speech presentation in a key aspect of the story of that particular poem. Chapter 5 discusses speech presentation for song in the Odyssey; Chapter 6 shows how fundamental speech presentation is for telling the story of Patroclus.
Finally, the Conclusion looks at the characters and the main narrators, at the Iliad and Odyssey, as a spectrum of a different kind: What kinds of "alternative" Homeric narratives can we see if we compare the varied norms of speech presentation for these different groups and narrators? That is to say, what if the entire Iliad were presented in the way that characters in the Iliad present their narratives? What kind of narrative would that be? Surely such a narrative falls within the scope of the Homeric poems overall, since the characters in the Iliad present half of the poem. Speech presentation in the Homeric epics shows a wide range of possibilities that draws on a stable set of attributes for individual speech presentation strategies across both poems and all narrative levels. These possibilities, deployed both with variation and with a fundamental underlying sameness, then imply a chooser. This figure is not necessarily a literal human composer, but the implied author, the voiceless force that creates the full range of phenomena in a given narrative. A narrative force beyond the main narrator arranges the Homeric epics so that they focus so strongly and expressively on the characters, so that the main narrator's speech presentation techniques complement those of the characters to achieve this result, and so that a stable and consistent set of speech presentation techniques underlie an Iliad where words often present isolation, conflict, and sorrow, and an Odyssey that depicts both the deceptive and the connective possibilities of conversation, and of narrative itself.