The Construction of Orality
L’oralité est une abstraction; seule la voix est concrète, seule son écoute nous fait toucher aux choses.
—Paul Zumthor, La lettre et la voix
In our culture, speaking and writing are distinct activities, with one often happening in the absence of the other, or even to the exclusion of the other. So we telephone instead of writing a letter, and increasingly we e-mail instead of speaking on the phone. It would seem that the two activities—speaking and writing—constitute a symmetrical contrast, in which either term can be used to define the other; but in practice this is not so. Often we use the one term, “writing,” as vantage point for our conception of the other, in ways that betray a distinct cultural bias. Our use of the term “oral” and related expressions is a case in point. We may speak of “oral” simply when a discourse is spoken; but more often we endow the term with a cultural value. So we routinely speak of “oral poetry,” not to characterize a given poem as “spoken” but to oppose it to the dominant form of poetry in our culture. And we speak of “orality” not to describe what happens when someone talks, but to label a period or a culture as different with respect to our own writing culture. Oral poetry and orality, in short, are abstractions derived from the property of not writing or being written, and as such they are literate constructs: they define speech as the construction of a writing culture that uses its own absence to define its opposite.
It may be useful, at the outset of this study, to put this cultural bias in perspective by distinguishing between two dimensions in which something can be oral. We may use “oral” in a medial sense, meaning simply that something is spoken and as such is a matter of sound and the voice of the speaker. In this sense, a neutral, medial opposition between spoken and written discourse obtains, the one being “phonic,” the other “graphic.” This neutral, medial opposition gets compounded by the fact that each medium comes with its own set of associations, and even its own mentality: speaking and writing are different activities that call for different strategies in the presentation and comprehension of a discourse. Thus “oral” may also be a matter of conception and may enter into an opposition—not so much with “written” as with “literate”—that is far less neutral than the medial one: whereas no one would question the simple existence of medial orality (speech), our acceptance of conceptional orality as a phenomenon in its own right is much less obvious. For it is here that we have to become aware of the inbuilt biases of our writing culture in order to arrive at real understanding.1
In the conceptional sense, “oral” can designate the mental habits of persons who do not participate, or who do not participate fully, in literate culture as we know it, a phenomenon we associate with societies other than or earlier than our own. When applied to texts, “oral” in this sense implies that a given piece of writing does not display the features that are normal and expected in a writing culture: it came into existence without the premeditation that is usually involved in the production of written texts. Such a discourse has been written down and is “graphic” as to its medium, but it may be called “oral” as to its conception. In the conceptional sense, then, “oral” may denote the absence of characteristics of written language, whether a discourse is spoken or written. Thus even though the two senses of “oral” have a certain affinity to each other, it is important not to confuse them. A discourse that is conceptionally oral (such as a conversational narrative) is often medially oral as well, but it is also possible for such a discourse to be written. And a medially oral (phonic) discourse is often conceptionally oral, but instead it may be fully literate as to its conception (as in the case of an academic paper read out loud).
Finally, a third sense may be distinguished. The term “oral” has been used to refer not to a set of mental habits or to a mode of communication, but to a property of literary language and hence of literary texts. Parry is one of the first to speak systematically of orality as a property of literature, in opposition to the property of being written, a distinction that serves as the basis for a classification of literature in general: “Literature falls into two great parts not so much because there are two kinds of culture, but because there are two kinds of form: the one part of literature is oral, the other is written.”2 In this sense, “oral” denotes not so much an absence as a presence: that of the formula as the prime feature of the oral or traditional style. We shall turn later in this chapter to the notion of the formula. Here we simply note that the concept of oral poetry as a class of literature is insensitive to the distinction between medium and conception just made: it applies to both. The oral poem is considered a spoken discourse, but at the same time a text with properties that make it very different from texts in our culture. We shall have occasion later to consider the problems that this may cause.3
The relation between orality and its opposite seems to be unproblematic and straightforward in the conceptional and medial senses. “Oral” and “literate” in the conceptional sense can be considered prototypes, or opposite end points on a continuum: as properties that come in degrees, they need not exclude each other. Someone’s mental habits may be oral to a greater or lesser extent, and that person’s degree of literacy will be inversely proportional. Likewise, societies as a whole may be oral or literate in various degrees, and since the transition from a preliterate to a literate society in which writing is institutionalized is never an abrupt one, the notions of orality and literacy, though distinct, do not exclude each other, either diachronically or synchronically.4 The same point applies to orality as a textual feature: texts may be oral to a greater or lesser degree, depending on the nature of the conception underlying them.5 The notion of an oral medium, however, seems to exclude its opposite on first sight, in the sense that whoever speaks does not write, and vice versa. The exclusion is superficial, of course, since speech and writing as media are coexisting strategies, both of them being available to most members of a literate community and chosen according to the communicative needs and purposes at hand.
But what about “oral” in the sense used by Parry in the quotation above? As a subclassification of poetry, Parry evidently means “oral” to exclude “written,” thereby turning a mere opposition into what seems a contradiction in terms: for how can poetry like Homer’s be “oral” and therefore “not written,” if we experience it as a written text? By “oral” Parry means, as is well known, “orally composed” as against “composed with the aid of writing,” yet that does not really alter the picture, and by conceiving of “oral” in terms of form and style (i.e., formula) Parry and many researchers after him believed that “oral” could be made to apply to written texts, provided they have certain oral (-formulaic) properties.
Traditional Style and Homeric Kunstsprache
The dichotomy between oral style and literate style is not the original form in which Parry cast his conception of Homeric discourse. First came the notion of traditional style as opposed to the individual expression of later poetry, an idea that remained closely attached to orality throughout Parry’s writings. Owing to its formulaic nature, Homeric diction is traditional, by which Parry means that it cannot possibly have been the individual creation of any one single poet. “The nature of Homeric poetry,” he writes, “can be grasped only when one has seen that it is composed in a diction which is oral, and so formulaic, and so traditional.”6 And again: “Oral poetry is formulaic and traditional. The poet who habitually makes his poems without the aid of writing can do so only by putting together old verses and old parts of verses in an old way.”7
For Parry, the traditional nature of Greek epic diction is reflected in the distinctly systematic character of the Homeric formula.8 Going well beyond the usual treatment of Homeric diction in terms of gratuitous cosmetics or poetic style tout court, Parry introduces two key concepts that are central to his conception of the traditional nature of Homeric style: extension and economy.9 Economy is the one-to-one correlation of a given formula and a metrical form, or in Parry’s own words, the absence of “phrases which, having the same metrical value and expressing the same idea, could replace one another.”10 Extension, on the other hand, is the degree to which such a unique pairing of a meaning and a form applies to a whole class of formulaic expressions (a formula type), for example, metrically interchangeable expressions for a range of gods and heroes. Thus, according to Parry, a system of formulas is economical in that all expressions with the same meaning are different from one another in their metrical form (e.g., polútlas dîos Odusseús ‘much-suffering godlike Odysseus’ vs. polúmētis Odusseús ‘many-minded Odysseus’, both expressions having as their meaning “Odysseus”). And the system is extended, conversely, in that all formulas with the same metrical form are different as to their meaning (e.g., polúmētis Odusseús vs. pódas ōkús Akhilleús ‘swift-footed Achilles’).
The arrangement of formulas in systems is treated by Parry as an indication that the whole of Homer’s diction is organized in this way. An analysis in terms of economy and extension, in Parry’s view, can show the way to a better understanding of why and how Homeric diction in general has the specific form it does. For example, the systems of formulaic expressions for gods and heroes in the nominative case are for Parry not an isolated observation; they merely offer a clearer and more striking case of systematic economy and extension than do other areas of Homeric diction,11 a case yielding proof that Homeric diction as such, not just some parts of it, is schematized, systematic, and hence traditional.12
This proof has been controversial,13 with much discussion, both inside and outside oral-formulaic theory, devoted to defining the formula. The debate, which took as leitmotiv Parry’s own definition of the concept (“a group of words which is regularly employed under the same metrical conditions to express a given essential idea”14), tended to start from the “intension” of the concept (What is a formula?) and move to its extension (How much of the Homeric text is formula?). It was hoped that clarity as to the essence of the basic unit of oral composition would clarify the concept of oral composition as such.15 But this strategy made the definition of “formula”—in itself already more a definiendum than a definiens—dependent on each scholar’s position regarding the wider issue of oral-formulaic composition and created circularity problems that oral-formulaic theory has never satisfactorily solved.16 The question of systematicity and of the essence of the formula will be addressed in Chapter 8 below, where I will attempt to view Parry’s insights from the perspective developed in this book.
The participants in the debate on formulas are divided about how traditional and formulaic Homeric diction really is: from entirely or almost entirely traditional and formulaic, as Parry and Lord claimed, all the way to mostly or entirely nontraditional, except for the core of formulaic systematicity that Parry had established irrefutably for noun-epithet phrases.17 But whatever one’s position on this continuum, there was one thing that bound most participants in the discussion together. Most scholars took the traditional and formulaic as the phenomenon to be accounted for, the figure against the ground. Such a position assumes that whereas the definition of the formulaic (and so the traditional) may be problematic and the subject of inquiry, the definition of the nontraditional and nonformulaic is not.
We notice here the same perspective as the one with which this chapter began: Homeric poetry, whether under the aspect of oral or of traditional, is seen in terms of an opposition in which the one member (oral or traditional) is defined with respect to the other (literate) which functions as framework and norm. Language, we almost unconsciously assume, is written and individual by default, and it takes some special conditions for it to become the “other” with respect to these notions. This same perspective, again, is apparent in the third and last dichotomy we will discuss, that between artificial and natural.
Before Homeric style became traditional or oral, it was artificial. Philological criticism found in Homer an artificial diction that could never have been spoken in ordinary discourse at any time or place, a Kunstsprache, as it came to be called in the dominant publications on the subject.18 The notion of Kunstsprache originated in nineteenth-century historical and descriptive linguistics, the backbone of linguistic thought in philology, and was based on a thorough investigation of the morphological, phonological, dialectal, and lexicographic features that distinguish Homer from other authors. These features include the geographic and temporal mismatch of Homeric forms and the use of inherently “artificial” forms, words that do not exist elsewhere. As a simple example we may cite the accusative phrase euréa pónton ‘wide sea’, created on the analogy of the dative formula euréï póntōi, to replace the “natural” but metrically undesirable eurùn pónton;19 or the accusative form hēniokh21
Parry is a direct heir to this approach: his oral-formulaic analysis amounts to a continuation of the research of Witte, Karl Meister, and their predecessors. The essential dimension that Parry adds is the perception that the dependence of language on verse is not merely an issue of aesthetics, the result of the hexameter functioning as a poetic generative principle, but a matter of functional motivation. Parry shows that the bewildering variety of epithets and morphologically heterogeneous dialectal “forms” is not an arbitrary feature of “epic style” but conforms to a system designed to facilitate oral composition in performance. It is within the context of this system that the “artificial” element in Homeric style, or generally the characterization of Homeric language as Kunstsprache, comes to be reinterpreted as that which exists in service of the traditional, and ultimately, the oral: “A whole new word no poet could make, since no one would understand him if he did, but he may make a form like another. That is to say, he may make the artificial by analogy with the real The reason for such a creation is of course the same which leads the singers to keep the old and foreign forms, namely the need of a formula of a certain length which can be gotten only by this means.”22
It appears, then, that in order to account for the Homeric formula various interrelated concepts have been used and are still in use: artificiality, traditionality, and orality, differentiating Homeric poetry respectively from the naturalness, the individuality, and the writtenness of ordinary poetic language. The phenomenon thus created, however, may be more a product of the perspective used than an objective reality. For when we change our perspective, shifting our focus from poetic language to language, and particularly language as it actually functions in speech, the three opposites of natural, individual, and written are not such distinctive features anymore.
From Poetry to Speech
We may wonder how “natural” ordinary language is when we realize that analogy, the creation of one form on the basis of another, is not at all confined to the creation of formulas in the oral style. Analogy is a regular feature of language change in general, part of a general drive of languages to make grammatical paradigms more “regular” when phonetic changes or unproductive morphology have made certain forms obsolete and “irregular.”23 A simple example in Greek is the “artificial” but regular form oídamen ‘we know’, created on the analogy of the “natural” and regular form oîda ‘I know’ to replace the “natural” but irregular form ísmen.24 Natural and artificial, then, do not seem to be viable concepts in the study of language change, where creating the artificial by analogy with the real is a very natural phenomenon. Nor does the artificial as such seem useful as a concept to characterize Homeric language. True, the motivation for analogical formation in Homeric diction is metrical, and therefore artistic; but that does not mean that the resulting form is inherently any more artificial than analogous formations in ordinary language.25
Analogical change in language, whether ordinary or poetic, is traditional in that no individual speaker or singer can impose a given newly made expression on the language community and make it enter the system. The very concept of system, in fact, which Parry sees as the key element in the traditionality of Homeric style, is not left unaffected by the shift in perspective from poetry to language. The system consists of the systematic differences between formulas as forms. A very similar account, however, can be found in structuralist linguistics, the approach to language introduced by Ferdinand de Saussure.26 Just as Parry himself does in the case of formulaic language, the structuralist linguist sees language in general as a system or code, a coherent set of differences and similarities between linguistic forms that serve the purpose of the efficient transmission of their content or meaning. An just as in the case of Parry’s formulas, the form of linguistic expression is seen as determined by an autonomous structure or system that is distinct from the content of the message transmitted (Parry calls the latter the “essential idea” of a given formula).27
But the problem of how to define Homeric traditional diction is not confined to the way in which the Homeric formulas relate to one another; it affects the concept of formula itself. The study of language has long been concerned with creativity and originality, with how speakers are the individual composers of expressions within the system of rules that makes up their language. In recent years, however, very different voices have been heard, drawing attention to the traditionality of speakers, their use of phrases of which they are not the original makers or authors. Indeed, repetition is a pervasive phenomenon when language is conceptionally oral. In the third part of this book we shall consider some implications of this observation. For now, we may simply reconsider the concept of traditional or oral formula, the exclusive feature of oral epic as proposed by Parry.28
If formulas as traditional “prefabs” also occur in ordinary language, we might do better than opposing Homeric poetry to ordinary poetic Ianguage, viewing it instead in connection with ordinary spoken language. True, the formular nature and function of expressions in Homeric poetry is determined, in ways to be explored further in Chapters 6 and 8, by metrical factors that do not, as such, play a role in ordinary speech. But the difference between metrical and nonmetrical is not as clear-cut as it seems, and even a predominantly metrical function for a Homeric expression does not preclude its also having a function akin to that of formulaic expressions in ordinary speech.
All this is not to say, of course, that there are no important differences between Homeric discourse and everyday speech. The point is rather that the phraseology of Homeric poetry may not be most fruitfully characterized by calling it chiefly and inherently artificial or traditional. And when we add orality to the mix, we risk running into outright paradox. In explaining artificial by means of traditional, and traditional by means of oral, Parry gave the notion of Kunstsprache an entirely different conceptual load, causing a revolution in Homeric scholarship. This revolution, how-ever, was in a sense a conservative one. Earlier analysts of the Homeric Kunstsprache focus on poetic style; Parry keeps that perspective and analyzes oral poetic style as a special case. In explaining the artificial, poetic element in Homer’s diction, Parry comes to equate the oral with what distinguishes Homer’s speech, treated as poetic style, from ordinary speech. In the oral-formulaic analysis, stylistic features like formulas, repetitions, and rhythmical patterns are studied by explicitly confronting them as a special case, isolated as a research object sui generis, a traditional oral style and a separate formulaic language.29 But this research project ignores, and has continued to do so till the present day, that the oral is all around us, not as a special poetic style but as everyday speech.
It is with respect to this pervasive but therefore neglected and illunderstood phenomenon that I attempt to define Homeric poetry in the chapters that follow. My starting point and main assumption in this book is that orality is not so much a phase in the development of culture, or literature, to be overcome in due course, as it is the primary manifestation of language. Orality is language in the spoken, phonic medium along with the conceptional process that it implies: a conception that has to be studied on its own terms, rather than with respect to writing. The study of speech as a phenomenon that can be observed all around us is an empirical discipline. In the chapters that follow, we will draw on some of this research, highlighting those approaches and methodologies that provide the best opportunities for the study of Homer.
We are concerned, then, not with oral as the special case of poetry, but with poetry as the special case of oral, in other words, with poetry in speech. We will view Homeric discourse not as oral poetry but as special speech.30 A general catchword for the poetry in Homeric speech is the term “stylization,” which implies, just as imitation or parody, two discourses: the discourse that is the object of stylization and the stylizing discourse itself. The stylizing discourse is meant to be distinctive, but in order to be recognizable as a stylizing discourse, it has to display some essential features of its model.31 In the same way, Homeric discourse can be said to stylize ordinary discourse by departing from it and yet retaining, or even highlighting, its most characteristic features. An important part of the program of this book will be to identify those most characteristic features, and to present them as the basis for an account of Homeric poetics.
Our main task, then, will be to ask what it means for language to be spoken, so that we can define what it means for speech to be special. The first question will concern us in Part 2; in Part 3 we will then be in a position to deal with the second one. But let us first finish our preliminary discussion of perspectives: if it is of interest to ask, for the benefit of the interpretation of a text, what it means for language to be spoken, then we may also want to ask what it means for speech to be written.
1 The notion of “medial orality” as opposed to “conceptional orality” derives from Koch and Oesterreicher 1985 (see also Oesterreicher 1993, 1997)· For more discussion of the various ramifications of the opposition between spoken and written, see also Brown and Yule 1983: 4–19; Tannen, ed. 1982 (esp. Chafe 1982); Olson et al., eds. 1985 (esp. Chafe 1985); Chafe 1994: 41–50; Givón 1979: 207–33.
2 Parry 1971: 377.
3 See also Bakker 1997b.
4 The sociological or conceptional sense of “orality” often implies a historical perspective, focusing on the arrival of literacy and its implications (e.g., Goody and Watt 1968; Ong 1982; Stock 1983); on the residue of orality left after this arrival (e.g., Havelock 1963, 1986); or on the coexistence of the two (e.g., Thomas 1989). Thomas (1992: 19) warns against approaches in which each and every aspect of progress and innovation is attributed to literacy.
5 Oesterreicher 1997.
6 Parry 1971: 328.
7 Parry 1971: 377.
8 See Parry 1971: 276: “It is the system of formulas, as we shall see, which is the only true means by which we can come to see just how the Singer made his verses.”
9 Parry originally spoke in terms of “extension” and “simplicity” (e.g., Parry 1971: 6–7, 16–17); later he came to speak of “length” and “thrift” (e.g., Parry 1971: 276).
10 Parry 1971: 276.
11 “In the case of this system, as in that of other formulas, such as those of the types πολύμητις ’Οδυσσεύς and δΐος ’Οδυσσεύς, the length and the thrift of the system are striking enough to be sure proof that only the very smallest part of it could be the work of one poet. But for the greater number of systems which are found in the diction of the Homeric poems we cannot make such sure conclusions, since their length is rarely so great and their thrift never so striking. This does not mean that the proof by means of the length and thrift of the system is possible only in the case of the noun-epithet formulas. It is clear without need of further search that the greater part of the system quoted above must be traditional.” Parry 1971: 277–78.
12 See Parry’s analysis (1971: 301–14) of the first twenty-five lines of the Iliad and Odyssey.
13 For the argument against Parry’s systematic treatment of noun-epithet formulas, see Vivante 1982: 158–59, 164–67. The most recent and extended argument against economy is Shive 1987. For a general discussion of the reception of Parry in current Homeric studies, see Martin 1989:1–5. The term “formalist,” although sometimes used with hostile intent (e.g., Fenik 1986: 171, cited by Martin 1989) can, if used in a more neutral sense—as in “formal linguistics”—quite adequately characterize Parry’s conception; see below, as well as Bakker 1995.
14 Parry 1971: 272. Cf. Parry 1971: 13; Lord 1960: 30. Discussion of Parry’s definition in Bakker 1988: 152–59.
15 For discussion of the development in Parry’s thinking about the formula, see Hoekstra 1965: 8–18; Bakker 1988: 152–64; 1990a. Well-known critical statements of Parry’s notion of formula include Hainsworth 1964; Minton 1965; Austin 1975: 11–80; Mueller 1984: 14–21. Just how large a portion of the debate on the formula in the 1970s was devoted to problems of definition can be gleaned from most of the papers in Stolz and Shannon, eds. 1976.
16 As is well known, the number of clear-cut noun-epithet formulas (conforming to a system and as such the core of Parry’s theory as well as of any theory of Homeric formulas) is almost negligible compared to the bulk of the Iliad as a whole, and to make Homeric diction acceptably formulaic, Parry had to work with a vague concept of similarity (1971: 313), which later scholars would develop into the notion of the structural formula (e.g., Russo 1963, 1966). But statements to the effect “x is like y” are not very useful, in that anything uttered in hexametric rhythm is as such like anything else uttered in that rhythm. The statistics in O’Neill 1942 are revealing in this respect. Discussion and criticism of structural formulas in Minton 1965; Bakker 1988: 159–64. See also my remarks below on analogy.
17 See, for example, Bassett 1938: 15–19; Hoekstra 1965: 15–205 Austin 1975: 11–81.
18 Witte 1913; Meister 1921. These works can be seen as the culmination of extensive nineteenth-century research on the decisive influence of the verse on Homers language. See also Ruijgh 1971: 106–12. Up-to-date bibliography in Janko 1992: 8–19.
19 Witte 1912: 113 (=Latacz, ed. 1979: 112).
20 Chantraine 1948: 95.
21 Witte 1913: 2214: “ein Gebilde des epischen Verses.”
22 Parry 1971: 339, emphasis added; cf. Parry 1971: 68–69.
23 This process is sometimes called leveling; see Hock 1986: 168–71.
24 The oldest attested uses of the new form seem to occur in Herodotus (4.46, 7.214) but it is not until the New Testament that the form has become normal; the older form, however, continues to be used.
25 After Parry, the issue of the relation between Homeric formulaic language and ordinary language has occasionally been raised; see, for example, Lord 1960: 36; Wyatt (1988: 29), who in criticizing Schein’s treatment (1984: 2–13) of Homeric language as artificial and never spoken by anyone as an ordinary language, submits that “Homer’s language was as natural for him as is English to us.” For the acquisition of poetic language as a second language, see Rubin 1988; 1994: 136–44·
26 Structuralist linguistics conceives of language, partly in opposition to nineteenth-century diachronic and historical linguistics, as a synchronic system, consisting of functional differences between its constitutive elements. For elementary expositions of Saussure’s work (1972) see Jackson 1991: 20–56; Holdcroft 1991.
27 Parry 1971: 13, 272. More on the affinity of Parry’s account of the Homeric formula with structuralism and other formalist linguistic paradigms in Bakker 1995: 118–22 (see also Lynn-George 1988: 55–68). Notice in this connection that Antoine Meillet, Parry’s teacher, was a pupil of Saussure.
28 “It is the nature of an expression which makes of it a formula.” Parry 1971: 304. On formula as a natural category of language vs. formula as a function of language, see Bakker 1988: 153–57.
29 See, for example, Parry 1971: 328: “Clearly a special language for the hexameter could come into being only when poetry was of a very different sort from that which we ourselves write, and which we know to have been written throughout the history of European literature.”
30 Cf. Nagy’s concept (1990a: 5–6, 29–34) of marked speech, or “song,” as an anthropological notion.
31 See Bakhtin 1981: 362; 1984: 185–204, on “double-voiced discourse,” which “is directed both toward the referential object of speech, as in ordinary discourse, and toward another’s discourse, toward someone else’s speech” (1984: 185).