Consciousness and Cognition
Consciousness is the crucial interface between the conscious organism and its environment, the place where information from the environment is dealt with as a basis for thought and action as well as the place where internally generated experience becomes effective—the locus of remembering, imagining, and feeling.
—Wallace Chafe, Discourse, Consciousness, and Time
In the preceding chapters I have argued that a binary distinction between orality and literacy does not quite capture the way in which speech and written texts may coexist and interact in cultures other than our own. Whereas written texts are medially different from speech in any culture, the conception underlying their production and use may in some cultures be so close to the conception of spoken language that a simple distinction between “oral” and “literate” becomes inappropriate. To study such texts, we need to ask what it means for language to be spoken. The search for an answer to this question where the Homeric text is concerned requires concepts and methods that are not to be found in classical philology, a discipline that has historically been concerned with the study of language as text. Below and in the chapters that follow I shall draw on such recent work in discourse analysis, the empirical study of speech, as seems relevant and useful for our purpose. Discourse analysis will provide a basis for discussing those features of our Homeric text that go unnoticed and unaccounted for in the textual perspective.
But this proposed change in perspective will not only lead us to look outside the philological discipline, but also require us to rethink some of what lies within. This takes us into the realm of the style of Homer and other archaic Greek poetry and will prompt us to reexamine such central concepts as parataxis, adding style, and ring composition. We shall deal in some detail with the phenomena denoted by these terms, paying attention to the question whether they do indeed reveal pertinent properties of the object described—Homeric discourse—or are more indicative of the literate perspective of the stylistician. Our discussion of Homeric discourse in terms of the style of spoken language, then, will be concerned throughout not only with medium—spoken vs. written language—but also with the conception underlying each medium—oral vs. literate—and in particular with avoiding literate concepts in the study of the medium of speech.1
Periodic and Unperiodic Style
If the classical ideal of sentence structure, the rhetorical period, is taken as a norm, then Homer, who was held in antiquity to be the standard for many rhetorical virtues, falls short of it. Many ancient analysts of discourse, followed by the students of Homer in modern times, have observed that the style of Greek epic is different from the later, classical style characterized by a balanced and complex syntactic design that we call periodic structure. In fact, Homeric style is decidedly unperiodic, as Parry noted in a discussion of Homeric verse structure and sentence structure inspired by the analysis of these phenomena by the Greek critic Dionysius of Halicarnassus (first century b.c.e.): “The period … [is a sentence] in which there is a planned balance of the thought. The unperiodic sentence is one which lacks this balance and in which, to cite [Dionysius], ‘the clauses are not made like one another in form or sound, and are not enslaved to a strict sequence, but are noble, brilliant, and free.’ That is, the ideas are added on to one another, in what Aristotle calls the running style.”2 Parry is referring to Aristotle’s well-known concept of léxis eiroménē ‘strung-on way of speaking’, that is opposed, as an archaic way of expression, to the lexis katestramménē ‘turneddown way of speaking’, the periodic style that comes to a natural end:3
λέγω δὲ εἰρομένην ἣ οὐδὲν ἔχει τέλος καθ’ αὑτήν, ἂν μὴ τὸ πρᾶγμα τὸ λεγόμενον τελειωθῇ. ἔστι δὲ ἀηδὴς διὰ τὸ ἄπειρον· τὸ γὰρ τέλος πάντες βούλονται καθορᾶν· διόπερ ἐπὶ τοῖς καμπτῆρσιν ἐκπνέουσι καὶ ἐκλύονται· προορῶντες γὰρ τὸ πέρας οὐ κάμνουσι πρότερον. ἡ μὲν οὖν εἰρομένη [τῆς λέξεώς] ἐστιν ἥδε, κατεστραμμένη δὲ ἡ ἐν περιόδοις· λέγω δὲ περίοδον λέξιν ἔχουσαν ἀρχήν καὶ τελευτὴν αὐτὴν καθ’ αὐτὴν καὶ μέγεθος εὐσύνοπτον. ἡδεῖα δ’ ἡ τοίαυτη καὶ εὐμαθής, ἡδεῖα μὲν διὰ τὸ ἐναντίως ἔχειν τῷ ἀπεράντῳ, καὶ ὅτι ἀεί τι οἴεται ἔχειν ὁ ἀκροατὴς τῷ ἀεὶ πεπεράνθαι τι αὑτῷ. τὸ δὲ μηδὲν προνοεῖν εἶναι μηδὲ ἀνύειν ἀηδές. (Rh. 1409a29–1409b4)
I call “strung-on style” a type of discourse that has no inherent end point other than the completion of the discourse topic in question. This style is unpleasant by its unbounded nature, for all want to see an end point. This is why it is [only] at the end of the course that [athletes] are out of breath and exhausted: having the end point in view they do not tire beforehand. Such is the strung-on style; the “turned-down style,” on the other hand, is the periodic style. I call “period” an utterance with an inherent beginning and end as well as a length that can be beheld at a single glance. Such a type of discourse is not only pleasant, but also easy to learn. It is pleasant by the fact that it is the opposite of what is unbounded, and because the listener at every moment has the idea of having hold of something, by the fact that every moment is bounded in itself. For having no anticipation of what is to come or not completing anything is unpleasant.
Aristotle seems to take as periods not the long balanced sentences that the modern handbooks cite from the oratory of Isocrates or Cicero, but rather the constitutive elements that make up such elaborate linguistic structures.4 Immediately after the passage quoted above, Aristotle claims that a period is easier to understand because meter makes it easier to remember than unperiodic discourse, an observation to which we will return in Chapter 6. A second difference between Aristotle’s account and the conception of periodic style that we find in the modern handbooks of Greek grammar is that the former is hearer-oriented, whereas in the modern accounts the emphasis is on the hierarchical relations in a sentence as marked by the syntax of subordination. In fact, Aristotle seems to deal with a different medium: his concern is with spoken, not written discourse.
Central in Aristotle’s account of periodic style is hearers’ anticipation (pronoeîn ‘to have a sense as to what will come next’). Any present moment in the comprehension of a discourse, according to Aristotle, should be connected in expected ways to the next present moment, and it should be clearly defined with respect to a given end point. The teleology of periodic discourse, moreover, implies a “beginning” that is just as clearly marked as the “end,” yielding a conception of periodic style as “bounded”: it gives the listener a sense of “having hold of something,” a mental disposition that is not only conducive to the comprehension of a discourse, but also adds to the listener’s pleasure.5
In Aristotle’s account, the strung-on style of speaking is all that the periodic style is not: it is unpleasant because of its unboundedness. Without beginning, middle, and end, it does not provide the listener with a sense of being somewhere, of knowing where the discourse will lead and from which point it started. In other words, the lexis eiroménē is not viewed as a phenomenon in its own right; it is negatively defined with respect to the positive qualities of the periodic style. Aristotle may differ from us in his emphasis on the listener and in his attention to the spoken medium, but he is a clear precursor of modern stylistic practice in using a given conception of language as a norm to which some discourses conform and others do not.
It was not until Parry that Aristotle’s stylistic perspective came to be seen in the terms that are familiar to us today. Parry turned Aristotle’s well-known stylistic distinction between periodic and unperiodic style into a basis on which Homeric adding style, as he called it, could be considered an oral style and opposed to the literate style of later authors. Within Parry’s perspective and the conception of orality discussed in the first two chapters of this book, the adding style is what distinguishes Homer’s oral poetry from other literature. It is a necessary consequence of the fact that Homeric poetry is produced under circumstances that are quite different from those applying to written composition: “Oral versemaking by its speed must be chiefly carried on in an adding style. The Singer has not time for the nice balances and contrasts of unhurried thought: he must order his words in such a way that they leave him much freedom to end the sentence or draw it out as the story and the needs of the verse demand.”6
Just as in the case of the formula, the other major characteristic of the oral style, Parry puts strong emphasis on production as the criterion for distinguishing between oral and nonoral discourse and as the explanation of the distinctive features of the adding style. The hurried thought underlying oral verse-making renders the oral adding style a necessity, insofar as anything more complex is in practice unattainable under the specific circumstances in which oral verse is produced. In this way, Parry’s perspective is still (however implicitly) the primacy of writing and written language as a norm to which oral language does not yet conform, even though his work did much to establish oral poetry as a legitimate form of literature with its own poetics.
The strung-on style with its unbounded nature has indeed been upgraded from vice or necessary evil to stylistic virtue by some scholars working in the wake of Parry’s discoveries, the idea being that there is a need for an oral poetics opposed to and coexisting with the traditional literate view of language and texts.7 Such work, however, does little more than reinforce the binary and unproductive opposition between orality and literacy that I argued against in the first two chapters. It still views the unperiodic strung-on style from the point of view of periodic style; what is forbidden or reprehensible in the latter is simply permitted or even desirable in the former. We need an account of unperiodic style in its own right, without the bias of periodic style or its post-Aristotelian form, the literate conception of language. Such an investigation does not deal with orality or oral style but with something that lies outside the realm of style: ordinary speech. In this chapter and the next two I offer such an analysis, arguing that Homeric unperiodic discourse, if studied from the appropriate angle, involves just as much boundedness and controlled anticipation as Aristotle attributes to periodic style. The difference lies in the means used to that end, means that belong to the spoken, not the written medium.
Apposition and Parataxis
Parry’s conception of Homeric discourse in terms of adding style has been refined by later authors, leading to a proliferation of terms for what remains essentially the same observation. Kirk, for example, speaks of “cumulation,” the crucial property of a style in which “each new piece of information, as the story proceeds, can be envisaged as being heaped upon its predecessor”8 From a different angle, using syntactic rather than stylistic terminology, Antoine Meillet, Pierre Chantraine, and others have singled out the “appositional construction” as the crucial property of Homeric syntax.9 By a syntactic principle inherited from Indo-European, according to Meillet, phrases or even single words in Homer tend to have considerable syntactic autonomy, being loosely attached to each other by appositional relationships and having the semantic autonomy of independent sentences. As an example of appositional syntax in the adding style consider the following piece of Iliadic battle narrative:10
Πρῶτος δ’ ’Aντίλοχος Τρώων ἕλεν ἄνδρα κορυστὴν
ἐσθλὸν ἐνὶ προμάχοισι, Θαλυσιάδην ’Eχέπωλον·
τόν ῥ’ ἔβαλε πρῶτος κόρυθος φάλον ἱπποδασείης,
ἐν δὲ μετώπῳ πῆξε, πέρησε δ’ ἄρ’ ὀστέον εἴσω
αἰχμὴ χαλκείη· τὸν δὲ σκότος ὄσσε κάλυψεν,
ἤριπε δ’, ὡς ὅτε πύργος, ἐνὶ κρατερῇ ὑσμίνῃ.
Antilokhos was the first to kill a chief man of the Trojans,
valiant among the champions, Thalusias’ son, Ekhepolos.
Throwing first, he struck him, on the horn of the horse-haired helmet,
and the bronze spearpoint fixed in his forehead and drove inward
through the bone; and a mist of darkness clouded both eyes
and he fell as a tower falls in the strong encounter.
To the first mention of Antilokhos’s victim (ándra korustḕn ‘chief man’, 457) are added two pieces of information, which according to Meillet’s analysis have to be seen as loosely added appositional phrases (esthlòn ení promákhoisi ‘valiant among the champions’ and Thalusiàdēn Ekhépdlōn ‘Thalusias’ son Ekhepolos’, 458) that acquire a large amount of syntactic autonomy (they are equivalent to separate clauses: “he was… ; his name was… Furthermore, the recapitulative phrase tόn rh’ ébale prṑtos ‘Throwing first, he struck him’, 459 is followed by a phrase (kóruthos phàlon hippodaseíēs ‘on the horn of the horse-haired helmet’, 459) that can be seen as an added piece of information, specifying the verb ébale: ‘he struck him, [namely] on the horn of the horse-haired helmet’. The passage as a whole, finally, is characterized throughout by a paratactic structure, in that the clauses are added to one another, without there being the kind of complex syntactic interrelationships that come with hypotactic structure and periodic composition.11
Now adding style, appositional syntax, and parataxis are themselves apt characterizations of the passage cited, though I will be replacing them in the next two chapters with other terminology, but the way in which these terms are used may still to a certain extent be indicative of the perspective of the user. We have already seen that Parry’s use of the term “adding style” is entailed by his view that Homeric oral poetry is an art form composed under specific circumstances. Meillet’s and Chantraine’s appositional construction, on the other hand, is a concept from historical syntax: Homeric style is viewed, from a diachronic point of view, as archaic, conforming to inherited Indo-European patterns to a greater extent than do later texts from the Greek corpus. The closely related notion of parataxis frequently and typically conveys qualities such as primitive or crude. Paratactic discourse, it has been thought, fails to express certain logical relations and has to put ideas of unequal importance and syntactic status simply beside each other, in the absence of any grammatical means to effect syntactic subordination. Criticism of such a “prelogical” way of expression was frequent in the preoral era before Parry and continues to be heard occasionally. Thus Eduard Norden calls lexis eiroménē the prime characteristic of the language of “children and primitives,”12 and Kirk writes of “the unsophisticated tendency to state logically subordinate ideas as separate, grammatically coordinate propositions,” a very common characterization of “archaic style” in classical philology.13
The three terms all reflect an assumption that style is a set of properties assigned to what the researcher perceives and conceives of as a text, a text that is different in a number of ways from other (i.e., later) texts. Homer may be said to use an adding, appositional, paratactic, even an oral style, but these qualifications ultimately derive from a comparison with other texts that lack certain properties; even the notion of style may itself imply a textual perspective, in that texts other than the Iliad serve as a basis for the characterization of Homeric style as archaic or oral. This comparison is understandable, given the necessarily textual orientation of classical philology, but is it necessary? Must the analysis of the appositional, paratactic passage just cited be a matter of comparing supposedly primitive archaic texts with sophisticated classical texts? And is style, considered as a textual property, at all an appropriate concept for defining the passage’s nature and design?
Those questions might best be answered by considering a very different example. The following text is a transcript drawn from a linguistic investigation of spoken narrative under the supervision of Wallace Chafe and known as the Pear Film project, in which subjects were shown a short film and were asked afterward to verbalize what they had seen and experienced. The fragment has many of the features that we noticed in the passage from the Iliad, but its style does not seem to fit the historical and textual paradigms just discussed:14
And he rides off, [1.55] with the basket, [.3] in front of his handlebars, [.8] balanced,.. and he hits some bumps, and a few pears spill out. [.85] He goes down the road, [.75] and he passes a girl on a bicycle, [.35] and in passing her, his hat comes off, [.55] and he turns his head, and he looks back, [.5] then his bicycle hits a rock in the road, and he falls over, [.2] spills his pears, . . hurts his leg.
This text displays the appositional syntax and adding style attributed to Homer: one piece of information is heaped on another in small, relatively autonomous units (e.g., “with the basket | in front of his handlebars | balanced” or “and he falls over | spills his pears | hurts his leg”); further-more, the passage is paratactic throughout, with almost each new clause that contains a verb being linked to what precedes by the conjunction “and.” This parataxis may seem to be a textual feature, because we also find it in the Homeric text, but the Pear Film passage is, of course, a transcript of a spoken narrative. Its parataxis may be better understood if we study it not as a feature of the transcript as such, but of the speech serving as its model. Such a perspective quickly reveals that speech, our own everyday language, is pervasively paratactic too—the feature appears to be an inherent property of spoken discourse, naturally resulting from its production, and essential in some ways to its comprehension. If this is the case, parataxis can hardly be an archaism or a mark of primitive language. To question such a conclusion, we would have to say that we all “speak in an oral style.” But the very strangeness of that phrase brings us back again to the central issue: the concept of an oral style involves applying literate standards to what was once a spoken reality. What if we were to reverse that approach, applying what we know of speech to our received Homeric text? My brief discussion of the Pear Film passage would seem to allow such an experiment. It is suggestive, first of all, that a collection of spoken narratives should contain speech that in certain respects resembles so closely the language of Homer. Notice also that we have not listened to the original Pear Film narrative or to a tape of it. We have examined a transcript, using comparisons with other examples of spoken narrative as a way of understanding the style of the transcript. I would suggest, then, that we use a similar approach to the transcriptions that have come down to us as the Homeric poems.
Consciousness and Cognition
Perhaps the most striking difference between speech and written text—modern written text—is that the former is a dynamic process evolving in time, whereas the latter seems to be better characterized as a finished and hence static product. To call a text static may seem strange at a time when reader-response criticism and reception aesthetics have left their mark on literary studies. Texts are more and more conceived of as dynamic and protean, rather than static, shaped in the ever changing dynamics of their reception. Yet the shift to considering reception and comprehension as subjective and dynamic processes tends to stress the relative autonomy of written texts, and this emphasis makes it even more likely that in its production or composition the written text will not be perceived as a process. More precisely, the actual production of written texts does not show in the text, the art of written composition being directed to hiding the processes related to production and to the presentation of a finished product.15
Speech is a process evolving in time, and not just because it takes time to utter words and sentences. The processlike quality of speech, less trivially, has to do with what happens in the minds of the speaker and his audience. As an act of socialization, speech may be public, an observable event occurring in the world, and words may, in the formulation of Mikhail Bakhtin, be “performed outside the author”;16 but this observable, public event is nonetheless closely associated with what is inherently private and non-observable: the consciousness of both the speaker and his or her listeners.
In introducing consciousness, not only as the source of speech, but also as a constraint on speech that determines important aspects of spoken discourse as it is publicly performed, I am following the ideas on the relation between language and consciousness of Wallace Chafe, whose earlier Pear Film project I mentioned above.17 Spoken discourse, Chafe argues, represents the consciousness of the speaker in a more direct way than written discourse does the consciousness of writers, and hence the observable, physical properties of the speech process are best explained as reflecting some characteristic properties of the flow of consciousness of the speaker. This flow, as Chafe notes, is a matter of activation, the flow of successive ideas through the mind: “Although every human mind is devoted to modeling a larger reality within which it (or the organism it inhabits) occupies a central place, only one small piece of that model can be active at one time. At any given moment the mind can focus on no more than a small segment of everything it ‘knows.’ I will be using the word consciousness here to refer to this limited activation process. Consciousness is an active focusing on a small part of the conscious beings self-centered model of the surrounding world.”18 Hence consciousness can be understood in terms of vision, a particularly important aspect of Homeric poetics.19 Consciousness is like vision, according to Chafe,20 in that both can focus on no more than a very limited amount of information at any one moment. The eye s area of foveal vision, where visual acuity is greatest, is small compared to the total field of vision and is continuously shifting. In the same way, a focus of consciousness, as Chafe calls it, containing the information that is activated in a persons mind at a given moment, is small in comparison with the huge amount of inactive information, of which one could be but is not conscious at any given moment.21 This focus of consciousness is also much smaller than the amount of information that is of immediate relevance for it, and that other information is analogous to the background in a field of vision: in order to be meaningful, any conscious experience, be it visual or introspective, needs a certain amount of information of which one is half-conscious (or which constitutes the periphery of the field of vision).22 And like vision, consciousness is continuously moving, not in an uninterrupted, smooth flow, but owing to the seemingly discrete nature of a focus of consciousness, in small jerks from one focus of consciousness to the next.23
Consciousness, moreover, is closely related to time and memory. Not only is the movement from one focus of consciousness to the other a process evolving through time; the notion of focus is itself a matter of time, in that its short duration can be related to the essence of time: the experience of a now, a present moment. Human experience of a now has to cover a time span that is short enough to be adjusted to the rapid changes in the environment, but long enough to make possible a reaction to what might happen during this now. Cognitive and neurophysiological research suggests that this time span is a two- to three-second period.24
The activation and conscious experience of a small amount of information during this period, furthermore, appears to be the proper domain for what in psychology has been called short-term memory, the evanescence of which matches the restless movement of conscious experience itself. Short-term memory can either be visual, retaining for a few seconds what was on the eye’s retina a few moments before, or nonvisual, keeping strings of digits or other elements of information that, once focused on consciously, can be held in memory for a short period after the activation. And finally there is the short-term remembering of sound, by which acoustic signals (e.g., words spoken) can be retained for a limited duration after the moment of hearing.25 So our consciousness, with its successive shifts in focus, seems to be limited both in capacity and in the duration of each active state, limitations that also apply to foveal vision and short-term memory, as is suggested by intuition and confirmed by experimental research.
Now when we consider focal consciousness and short-term memory in terms of linguistic expression, we have to think of short strings of words, conveying the information that is in focal consciousness and not exceeding the capacity of short-term memory. Furthermore, we must assume that when a consciousness is engaged in turning its successive foci into speech, these foci are apparent in the flow of speech produced by a verbalizing consciousness in order to be processed by a listening consciousness. And this appears to be in fact the case.
If we see speaking as a turning of what is private into public speech, or in Bakhtinian terms, as the creation of an object or the objectification of consciousness,26 then listening to and making sense of speech is necessarily the reverse of this process: it involves turning what is public into private sense, or alternatively, the subjectification of speech. But it is also possible to listen to speech as an object of study in its own right. In adopting this empirical stance, Chafe is one of the relatively few linguists who have noted that spoken discourse comes in a series of “spurts of vocalization” beginning every two or three seconds and often (though not always) preceded by a pause which may last from a slight break up to several seconds. These units are above all characterized by a coherent intonation contour, that is, they are uttered as integral wholes and end with a pitch contour that signals a sense of closure.27
On the basis of the last property, Chafe calls these units “intonation units,” emphasizing their physical, empirically observable quality as units of speech.28 It is intonation units that are mainly responsible for what might be called the fragmented style of spoken discourse, as opposed to the more fluent and integrated quality of written discourse. But this public, readily observable reality is intimately connected with the private consciousness that drives the speech: the intonation unit is the linguistic equivalent of the focus of consciousness, the amount of information that is active at any one time in a speaker’s consciousness. The intonation unit is the largest linguistic unit that is still available in its entirety to consciousness, the typical sequence of speech sounds that is within the grasp of the speaker’s, and listener s, echoic memory: any stretch of discourse that is longer will have to be processed as more than one of these basic chunks.29
It is in terms of the segmentation of spoken discourse into intonation units that we have to view the characteristics of the Pear Film passage presented above. We present it here again, this time with each intonation unit displayed as a separate “line:”
- And he rides off,
- [1–55] with the basket,
- [-3] in front of his handlebars,
- [.8] balanced,
- . . and he hits some bumps,
- and a few pears spill out.
- [.85] He goes down the road,
- [.75] and he passes a girl on a bicycle,
- [.35] and in passing her, his hat comes off,
- [.55] and he turns his head, and he looks back,
- [.5] then his bicycle hits a rock in the road,
- and he falls over,
- [.2] spills his pears,
- hurts his leg.
Most of the intonation units in this stretch of spoken discourse are short clauses, consisting of a verb with a subject and/or object (such as units a or m), and most of these clauses are introduced by the particle “and.” Intonation units may also be something other than a clause and are in principle not predetermined by any kind of linguistic structure. In terms of syntax, intonation units can be anything from complete clauses to all kinds of nonclausal elements: prepositional phrases as in units b and c, phrases involving participles as in unit d, or even separate noun phrases when they are the verbalization of the idea on which the speaker focuses.
The little jumps from one focusing to another, furthermore, are mainly responsible for what Aristotle would have called the unperiodic or strungon style of the passage: there is no inherent end point in the discourse of this speaker. The transcript is punctuated to suggest intonation: the comma transcribes a rising intonation, which conveys the idea that more is to come, whereas the period represents the falling pitch that signals closure in spoken English, the sense that something has been completed.30 In the composition and comprehension of written discourse, we conceive of such moments of closure in terms of the end of a sentence, and we tailor our syntax in such a way as to make that moment a meaningful and a pleasing one for the reader.
Such planned organization is quite absent from the above fragment, whose speaker merely passes from one scene to another in recalling the story. Syntactically, the beginning of the second “sentence” (unit g) is marked only by the absence of the particle “and” in the middle of a whole string of clauses introduced by that particle. The concept of sentence, then, the primary stylistic unit of written discourse and the principal domain for the operation of written syntax, is much less relevant in spoken discourse. Speakers may regularly produce sentences by intonational means (sentences that may or may not correspond to what is for us a finished, syntactically correct sentence), but the syntax of their speech is the syntax of the intonation unit as it reflects the flow of ideas through their consciousness.31
The same kind of nonsentential segmentation becomes apparent when we re-present the Iliad passage that was cited above, changing the hexametric lines of our modern printed text into the lines of cognitive production:32
And first Antilokhos,
b. Tρώων ἕλεν ἄνδρα κορυστὴν
of the Trojans he took a helmeted man,
c. ἐσθλὸν ἐνὶ προμάχοισι,
vaHant among the foremost fighters,
d. Θαλυσιάδην ’Eχέπωλον·
Thalusias s son Ekhepolos.
e. τόν ῥ’ ἕβαλε πρῶτος
He first struck him,
f. κόρυθος φάλον ἱπποδασείης,
on the crest of his horse-haired helmet,
g. ἐν δὲ μετώπῳ πῆξε,
and he planted [it] in his forehead,
h. πέρησε δ’ ἄρ’ ὀστέον εἴσω
and it pierced right through the bone,
i. αἰχμὴ χαλκείη·
the bronze spearpoint,
j. τόν δὲ σκότος ὄσσε κάλυψεν,
and darkness covered his eyes,
k. ἤριπε δ’ ὡς ὅτε πύργος,
and he fell as when a tower [does],
l. ἐνὶ κρατερῇ ὑσμίνη.
in the tough battle.
My argument, here and in later chapters, rests on the assumption that the lines into which I divide Homeric extracts must have been a prosodic, intonational reality. Our text obviously does not record that reality as such, but it provides some cues. First of all there is meter. The metrical dimension of Homeric speech will not concern us until Chapter 6, but we can already note that boundaries between the units in this fragment coincide either with the end of the hexameter line or with the middle caesura (penthemimeral or trochaic). The coincidence of intonation with metrical units is a universal characteristic of performed poetry in oral traditions, and in the study of Homer it seems justified to use the latter as evidence for the former.33
Furthermore, the content of phrases can be used as criterion for division—in fact, this is the very reason of being for intonation units in Chafe’s analysis: each unit represents a single focus of consciousness. In an analysis of the Homeric text, this semantic criterion is to a certain extent arbitrary, but in practice the often observed coincidence of metrical units and semantic units in Homer can guide us, each metrical colon being the verbalization of a single idea in a satisfyingly large number of cases.34 Thus each of the phrases that in the above analysis of the passage were called appositional (here units c, d, and f) can now be seen as the verbalization of a focus of consciousness. Besides meter and semantic content, syntax is an important cue for the division of Homeric discourse into speech units. As noted above, intonation units are in themselves not determined by a syntactic structure, but the reverse does frequently occur. Like many languages, ancient Greek has enclitic particles that occupy, as postpositives, the second position within a “domain” that is intonational rather than syntactic.35 Consequently, we may consider these particles as textual evidence of an intonational boundary. In our fragment the particle rh’ (unit e), a phonetically attenuated form of ára, is an example. Even more significant is the conjunctive postpositive particle dé (units a, g, h, j, k). This particle can be described as a boundary marker, setting off discourse segments against each other.36 In written prose, those segments tend to be larger, with textual structure and cohesion being the rationale for the boundary. In Homer, on the other hand, the segments marked by dé appear to be much shorter, the size of Chafe’s intonation units, so that dé becomes an important feature of the Homeric text considered as speech. Dé plays a role similar to “and” in English speech, and is the prime feature of Homeric parataxis, as we shall see in more detail in Chapter 4.37
It appears, then, that metrical and syntactic cues may be of help in determining the way in which Homeric discourse is segmented according to the cognitive principle noted by Chafe and others. Many readers of Homer will find this segmentation intuitively satisfying. Yet not all of Homeric discourse is as clear-cut as the extract presented above, with meter, syntax, intonation, and cognition all in perfect agreement. There are passages that are more complex and sophisticated, with deliberate, rhetorical mismatch between the various levels. I discuss some of those in Chapter 6 below. In other cases, a division into foci of consciousness is simply not as clear as in the fragment just cited, or less well supported by meter or syntax. The reader will encounter such examples in the chapters below, and may or may not agree with the division I propose in each case. Instead of invalidating the general principle, however, such less-than-prototypical cases are inevitable, as in any investigation of an observable phenomenon. Even Chafe’s acoustic data are not always clear-cut.38
To return to our Homeric passage, each unit specifies the image or idea evoked by the preceding unit in a way that is no different from what we see in Chafes transcribed stretch of ordinary speech. And just as in that fragment, the notion of syntactic closure as a criterion for “sentencehood” is a precarious one. Indeed, the editorial business of punctuation, in this passage and throughout the Homeric corpus, is no more than arbitrary. Especially when it comes to using the full stop, the modern editor is making decisions not unlike the ones the conversational narrator makes in uttering units with a falling intonation. What matters much more is the kind of relation between units in Homer that a discourse analyst editing a transcript would mark with a comma.
In the next two chapters we shall be concerned with more details of the progression of Homeric discourse through time and with the nature of the relations between Homeric speech units, the syntax of epic discourse. Here we simply observe that Homeric speech is amenable to the analysis of speech that is demonstrated in Chafe’s work. The segmentation of Homeric discourse, as evidenced by the length of the linguistic units of which it consists, can be seen as the manifestation in speech of the flow of the speaker’s consciousness, each unit being the verbalization of a focus of consciousness. The length and duration of the units fits the acoustic short-term memory of the performer, or in other words, the ability to process linguistic expressions as wholes, which is determined by an “auditory buffer” of two to three seconds.39
No less important than the similarities, however, are the differences between the ordinary and the epic passage. Chafe’s text is the transcript of a recording, made in order to study the evanescent sounds of his speaker’s narrative. The Homeric text is a transcript, too, but one that was obviously made for purposes very different from linguistic analysis. The speech which it records is not a onetime event; it will be repeated by future speakers, and heard by future listeners. In fact, these speakers are listeners themselves, insofar as each of them receives the discourse from a consciousness other than his own. The segmentation displayed by the Homeric passage, then, is not merely a matter of production and cognition, but also of re-production and re-cognition.
The reuse of the Homeric passage and of the narrative to which it belongs, as well as the special nature of this speech, has important conse-quences for the status of its constitutive units: to bring out the resemblance between them and the intonation units of ordinary spoken discourse may be an important step, but it does not exhaust the complexities of the phenomenon. Homeric discourse is obviously not an ordinary discourse, and its constitutive units are better known as formulas than as intonation units. In fact, formulas are stylized intonation units, and the cognitive approach to Homeric narrative is incomplete without some idea of how it might serve as a step toward a psychology of the Homeric formula. The Homeric formula and its metrical dimension will occupy us in Chapters 6 through 8. But first, in the next two chapters, we continue the discussion of Homeric speech units as a sequence that progresses through time. A picture will emerge in which Homeric discourse is much more bounded and goal-oriented than Aristotle’s concept of unperiodic style or Parry’s account of adding style might lead us to believe.
1 On medium and conception, see Chapter 1 above.
2 Parry 1971: 252. Immediately before the quoted passage, Parry discusses Dionysius on enjambement in Od. 14.1–7 (De comp. verb, xxvi, in which the phrase occurs to which Parry’s term “unperiodic” goes back: ὀ ἑξῆς νοῦς ἀπερίοδος ἐν κώλοις τε καί κόμμασι λεγόμενος ‘the sequence of ideas is unperiodic, uttered in clauses and phrases’). In Parry’s description of periodic and unperiodic sentences, the passage quoted from Dionysius is, oddly, from De comp. verb, xxii, a discussion not of Homeric unperiodic style but of the so-called grand, emotional style as a possible register in public oratory. See also Parry 1971: 262. Parry uses the term “unperiodic” to designate an important type of enjambement in Homeric verse, in which the end of the verse falls between two constitutive parts of an unperiodic sentence (see further Chapter 6 below).
3 The translations “strung-on” and “turned-down” are borrowed from Kennedy 1991: 239. Aristotle’s use of the term λέξις lies somewhere between “style” in the philological sense and “presentation strategy” as the phrase might be used by modern sociolinguists or discourse analysts. On the one hand, the term is used by Aristotle to designate properties of language as text; on the other, λέξις never quite loses the meaning that is inherent in its morphological status as an action noun.
4 See Kennedy 1991: 239. See also Chapter 6 below.
5 It is not clear whether this quality applies to the relation between periods or to the internal structure of periods. See further Chapter 6.
6 Parry 1971: 262.
7 See, for example, Notopoulos 1949; but cf. Austin 1966: 295–96.
8 Kirk 1976: 152; Kirk’s notion of cumulation is criticized for imprecision in its relation to enjambement in Higbie 1990: 13.
9 Meillet 1937: 358–59; Meillet and Vendryes 1968: 598–99; Chantraine 1948–53: 2:12. See also Ammann 1922: 9; Schwyzer 1947: 8; and esp. Thornton and Thornton 1962, an account of apposition in terms of temporal experience. See also Chapter 5.
10 Translation is from Lattimore 1951, except for the insertion of “him, on” in 459 and different transliterations of the proper names.
11 The literature on parataxis in early Greek literature is vast and covers not only parataxis as the stylistic appearance of single passages (e.g., Perry 1937: 410–12; Chantraine 1948–53: 2:351; Frankel 1968: 40–96), but also parataxis as the primary compositional principle behind whole works (e.g., Van Groningen 1937, 1958; Notopoulos 1949). Notopoulos (1949: 5–6) proposes parataxis as the core element of a non-Aristotelian poetics: “The Iliad and Odyssey have a unity; but unlike that of the drama it is inorganic and, moreover, the digressions far from being, like Homer’s similes, for purposes of relief, are actually the substance of the narrative, strung paratactically like beads on a string.” A more recent account of parataxis as compositional principle is Thalmann’s (1984: 1–6), who argues that parataxis by no means excludes planned design. For a more detailed discussion, see Chapter 4 on parataxis as movement.
12 Cf. Norden 1909: 37 n: “Daß die λέξις είρομένη überhaupt das wesentliche Kennzeichen primitiver Rede 1st, weiß jeder aus der Sprache der Kinder und Naturvölker.” Similar prejudices exist against repetition in discourse, as noted by Finnegan 1977: 131–32.
13 Kirk 1962:169. See the introduction in Kühner and Gerth 1898–1904:1–2, where the task of grammar is defined as the description of how the Greek language has developed from the primeval form of the simple clause to the elaborate perfection of the Attic periodic sentence. See also Kühner and Gerth 1898–1904: 2:224–26, where parataxis is described as a primitive stage of linguistic expression, the precursor of more sophisticated stages in the development of text and the human mind. See also Denniston 1952: xl: “As expression develops, subordination largely replaces co-ordination, the λέξις κατεστραμμένη the λέξις εἰρομένη.” Notice that Kirk’s phrase “logically subordinate ideas” testifies to the widespread tendency in classical philology to treat style as the form of a text in a binary opposition to content or what the text represents, the underlying idea being that the text “contains” logical, subordinating relationships but that the style used by its author is still too primitive to express them. This way of thinking applies also to suprasyntactic levels, as appears from Immerwahr’s characterization (1966: 307) of the structure of Herodotus’s Histories.
14 Chafe, ed. 1980: 319. Numbers in brackets indicate pauses measured in seconds. Two dots indicate brief pauses up to one-half-second long. Punctuation marks show intonation contours: a period indicates falling pitch, and a comma indicates rising pitch.
15 Of course, this distinction between product and process is not a clear-cut one, just as the orality of a discourse is not a matter of “yes” or “no.” As pointed out at the beginning of Chapter 1, discourses may, according to their conception, be oral to a greater or lesser degree. The processual features that result from orality may even be imitated; see Oesterreicher 1997.
16 Bakhtin 1986: 122. The public nature of words and utterances is the basis of Bakhtins ideas on the “joint creation of meaning” by the speaker or author, the listener, and “those whose voices are heard in the word before the author comes upon it” (121).
17 Cf. also Bakker 1990b; 1993b: 5–8c;1993: 278–80. Chafes ideas are now most fully presented in Chafe 1994.
18 Chafe 1994: 28.
19 See further Chapters 4 and 5 below, as well as Bakker 1993b: 15; 1997a.
20 Chafe 1994: 31, 53; see also Chafe 1980: 12–13.
21 On activation and inactivity considered in terms of presence and absence, and in the context of the Greek epic tradition, in terms of kleos and forgetfulness, see Bakker 1993a.
22 Such a background (both in visual and in cognitive terms) is what Chafe (1987: 28–31; 1994: 71–80) calls semi-active information. Further details, in connection with the particularities of Homeric discourse, in Chapter 5 below.
23 Chafe 1994: 29–30. See also Flanagan 1992: 159.
24 In Turner 1992: 86–88, the three-second period is called a “fundamental parcel of experience.”
25 On short-term memory, see Miller (1956), who states that short-term memory typically can hold seven items (plus or minus two). The short-term memory of sound is sometimes called echoic memory (e.g., Chafe 1994: 55). Rubin (1995: 69) speaks of “working memory.”
26 See also the psycholinguist Goldman Eisler (1958) on what she calls subjective and objective speech.
27 See Chafe 1980: 13–16; 1987: 22–25; 1988: 1; 1990: 88–89; 1994: 56–62. See also the general discussion in Brown and Yule 1983: 153–69, with special reference to the work of Halliday (1967); as well as Turner 1992: 88; Devine and Stephens, who list (1994: 412–14) other prosodic features as well.
28 Other researchers making the same observation have proposed different terms, such as “tone unit,” “information block,” or “idea unit.” As often in linguistics, there is a proliferation of terminology that does not seem to be entirely justified by the researchers’ differing interests and points of view. Chafe uses “idea unit” in earlier publications (1980, 1982, 1985), a term that seems to apply more to units of consciousness than to units of speech. Devine and Stephens (1994: 411) speak of “major phrases” as intonational, prosodic units that are “important not only as phono-logical cues to syntactic, and consequently semantic, structure, but also as cues to processing units. Our brains seem to process the utterances we hear in clausal chunks.” Hymes (1981: 309–41) speaks of “lines” in this connection, a term stressing the importance of “intonation units” in poetry, which is also one of the main themes in Turner 1992: 61–108. Intonation units and meter are the subject of Chapter 6 below.
29 Chafe 1994: 55. Cf. Rubin 1995: 68–69, 104–5. See also Turner 1992: 88: “A human speaker will pause for a few milliseconds every three seconds or so, and in that period will decide on the precise syntax and lexicon of the next three seconds. A listener will absorb about three seconds of heard speech without pause or reflection, then stop listening briefly in order to integrate and make sense of what he has heard.” Turner’s connection between the brain and poetic meter will prove important in Chapter 6.
30 Chafe 1988: 6–10; 1994: 143.
31 On “sentences” see more in Chafe 1980: 20–32, where the frequent discrepancies between intonational and syntactic closure are noted; 1987: 45–47; 1994: 139–44. See also Devine and Stephens 1994: 418–19.
32 In this and every similarly presented example in this book, I have retained the punctuation of the OCT for the Greek text in the left column and have used Chafe’s intonational punctuation of transcripts (see above) for the English translation in the right column. I have sought to strike a balance between a literal rendering enabling the Greekless reader to follow the discussion of each passage, and English idiom as required for an independently readable translation.
33 See Bakker 1993b: 8; 1997b; Devine and Stephens 1994: 410, 424–28; Rubin 1995: 86, 206. Notice that the significantly greater tolerance as regards hiatus and brevis in longo at the middle caesura, which since Parry’s (1971: 197–221) seminal discussion has been discussed in terms of the modification and juxtaposition of formulas (e.g., Hoekstra 1965), is primarily an intonational phenomenon, because formulas are stylized intonation units. See further Chapter 6.
34 See Parry 1971: 256–62; Lord 1960: 145; Edwards 1966: 121–25; Turner 1992: 7379־· In Bakker 1990b: 17–18, I show that certain ideas may be so strongly associated with a given metrical contour as to influence the surface realization of an adjacent unit.
35 See Ruijgh 1990; Devine and Stephens 1994: 422–23; H. Dik 1995: 31–38. More on the details of this phenomenon, known as Wackernagel’s Law, in Chapter 5 below.
36 See Bakker 1993 c.
37 Other postpositive particles whose occurrence signals an intonational boundary are mén, dḕ, gàr, and οûn, to be discussed in Chapters 4 and 5 below.
38 See Chafe 1994: 58: “The physical manifestations of psychologically relevant units are always going to be messy and inconsistent. If one breaks eggs into a frying pan, it may or may not be easy to tell where one egg leaves off and another begins. It may be similarly easy or difficult to read off the boundaries of intonation units directly from displays of acoustic data.”
39 Notice that my position with regard to the relation between speech and thinking—speech is closely related with mental processes—is the reverse of the usual conception of mental processes in Homer—thinking is closely related with interactive speaking—noted by Russo and Simon 1968 and by the classic writers on Homeric psychology (Snell 1975: 27; Dodds 1951: 16; Onians 1951: 13–22). The two possibilities complement each other, of course; together they point to the fact that in a speaking culture, speech will be conceived of in terms that are very different from our own.