The Writing of Homer
Let us say that a text is any discourse fixed by writing. According to this definition, fixation by writing is constitutive of the text itself. But what is fixed by writing? We have said: any discourse. Is this to say that discourse had to be pronounced initially in a physical or mental form? that all writing was initially, at least in a potential way, speaking? In short, what is the relation of the text to speech?
—Paul Ricoeur, “What Is a Text? Explanation and Understanding”
The orality-literacy dichotomy is not so much an objective distinction as a reflection of a particular perspective. It creates orality as what happens in the absence of writing. But it may also create orality as what happens in the presence of writing, diminishing its strange alterity in favor of a conception that is compatible with what we call literature. A case in point is Ruth Finnegan’s monograph Oral Poetry. This book argues against a “deep gulf” between oral and literate poetry, and tries to correct some of the “dubious” and “romantic” generalizations about oral poetry in our modern literate culture.1 This knowledgeable contribution to the subject of oral poetry considered as a cross-cultural social phenomenon has had quite a surprising effect on its classical readership. Finnegan’s insistence that there are no hard boundaries between the oral and the literate has blurred the distinct senses of oral discussed in the previous chapter and has led to a new conception of orality as a phenomenon not essentially different from literacy. Instead of helping readers gain greater insight into the various aspects of oral poetry, Finnegan’s book has actually contributed to the new spread of a literate and antioralist stance in Homeric studies. The feeling that no special categories of criticism and analysis, like those of oral-formulaic theory, are needed for the study and appreciation of Homer has been reinforced, paradoxically, on the authority of an acknowledged specialist in oral poetry.2
Finnegan’s attempts at bridging the chasm between oral and literate poetry may indeed yield some insights into the nature of both that have been obscured by the oral-formulaic analysis of epic poetry and by the study of modern written narrative. In Homeric studies, however, they cause confusion by evading what I take to be the real issue. Homerists have drawn support from Finnegan’s work for their position in what, since Parry, has been a major controversy: the question whether the Homeric poems were composed orally or not.3 Yet this question is less central and productive than has been assumed, in that either answer to it misses the point by excluding the other. The mere acknowledgment of the existence of the text of the Homeric poems and of its importance in antiquity, for example, might in the dichotomized orality-literacy debate be taken as a statement to the effect that the Homeric poems have not been composed orally; or at least that they were composed by a poet who had liberated himself from the constraints of oral composition or was in the process of doing so.4
More interesting than either opposing oral and literate or blurring the distinction between them, however, seems to be the question—even if it does not lead to a clear answer—what it might have meant for the Iliad and Odyssey to have been written in their historical context. And if reading is the intended reception of a written text, we have to ask what it means for the texts of the Homeric poems to have been read in that historical context. Instead of manipulating oral and literate, then, as immutable primary elements that either coexist with or exclude each other, we should try to endow these terms with cultural value. Such an attempt requires of us a readiness to question the assumption that our conception of writing is a norm for writing in all times and places. In other words, we will have to make the effort to defamiliarize our culturally and professionally determined habits and preconceptions, to make them more meaningful, or rather, less self-evident. The resulting sense of relativism is what we surely need when it comes to the study of Homeric writing.
What are the possibilities when we try to bring speech and writing into closer contact with each other than they are in the binary contrast between written and oral composition? The most exciting possibility imaginable would be that the text of the Homeric poems is directly related to the real event, as the faithful recording of an epic oral performance, with traces of this actual event still present in our text. This would reconcile the Iliad and the Odyssey as performances with the existence of a written text and leave orality in its strong sense of textlessness intact. But this possibility is highly improbable. While it is true that text and performance have much to do with each other (in fact I will argue later on in this chapter that they are interdependent), the notion that the text is a faithful recording does not seem a viable one. It would lead to blatant anachronism and turn the Homeric text into the published version of the field notes of some ancient anthropologist, an interested scribe who was present at the performance of the master, but who had no part in the dynamics of the speech event, and in whose writing dynamics the performing poet had no part.5
One could also separate in time the composition of the Iliad and the Odyssey from their writing (and thereby save the notion of oral composition): the Homeric poems could have been written down long after their composition. This is the position that G. S. Kirk defends.6 Kirk argues for a Homer who made no use of writing, a truly oral poet whose highly atypical, colossal poems were transmitted orally until they were committed to writing. Kirk’s views rest on the common assumption that a creative phase in the development of the oral tradition (that of the aoidoí ‘singers’, of whom Homer was allegedly the last and the greatest), in which new poems were composed, was followed by the reproductive phase of the rhapsōidoi ‘rhapsodes’, who merely recited the creations of their great predecessors.7 But this is a very literate, not to say literary, presentation of the matter; Kirk assumes that repetition is the reproduction of a first, normative, and superior original, and hence the creation of something derivative.8 In his conception of Homer’s creative orality (as opposed to the rhapsode’s repetitive orality) Kirk thinks in solidly textual categories. What he neglects is the notion of performance, in which the concepts of repetition and reproduction are not as appropriate as those of reinstantiation, recreation, reiterability.9 The insight that repetition across performances is crucial, not reprehensible, not only for the epic performance but also for speech in general, makes Homer somewhat less original than we like him to be, but the reciting rhapsodes become a great deal more creative; in fact the distinction between the two might well collapse.10
Which brings us to the position of Gregory Nagy and his elaborate argument in defense of the anonymous rhapsode. In Nagy’s view, the single genius poet, whether composing orally or not, recedes in favor of a concept of Panhellenism, a “cumulative process, entailing countless instances of composition/performance in a tradition that is becoming streamlined into an increasingly rigid form as a result of ever-increasing proliferation.”11 This increasing rigidity, according to Nagy, eventually led to the textual fixation of the Homeric tradition, but the final stage of fixation was not reached until the performance tradition of Greek epic had completely vanished.12 In Nagy’s account, writing is seen as inherent in the fixation of the tradition, not as an external phenomenon superimposed on the tradition from outside. In what follows we will elaborate on a similar conception of writing, focusing more than Nagy does on the precise interrelationships between speaking, writing, and text.
We are dealing, then, with the role played by a poet who was capable of recomposing the Iliad in the actual production of the written text. The best-known scenario here is Lord’s conception of “oral dictated text,” in which a poet and a scribe engage in a deliberate communication, with the production of the text as intended outcome.13 Owing to the changed communicative situation in which it was produced, the oral dictated text, the reasoning goes, can be longer, covering more detail, than any single normal performance. Taken by itself and viewed from the right perspective, dictation remains a plausible conception for the fixation of the Iliad and the Odyssey in writing and for the initial stages of their text. To the extent, however, that the idea of dictation arose from a conception of orality that is opposed to literacy in the binary contrast that we are discussing here, the proposal must be viewed as born out of necessity.
As we have seen, in the original conception of Parry and Lord, oral and literate are always mutually exclusive concepts: a poet is either oral or literate, and an oral poet has an entirely different technique from the literate poet. Lord puts the dichotomy as follows: “The written technique … is not compatible with the oral technique, and the two could not possibly combine, to form another, a third, a ‘transitional’ technique. It is conceivable that a man might be an oral poet in his younger years and a written poet later in life, but it is not possible that he be both an oral and a written poet at any given time in his career. The two by their very nature are mutually exclusive.”14 Statements such as this one reflect the direction of the discussion of oral composition and the climate in which it has been conducted; oral composition has often been conceived of as a fragile flower that is immediately killed off when it comes into contact with its mortal enemy, literacy.15 Lord’s conception of the oral dictated text, in the context in which it was proposed, amounts to an effort to account for the existence of a text while leaving the much cherished idea of a monolithic oral tradition intact. Oral dictation in Lord’s sense is the tangential contact between the oral tradition and writing, in the form of a special, abnormal performance of an unspoiled oral poet.
Historical and sociological research outside oral-formulaic theory, how-ever, suggests that the picture is not so simple. Even if we adhere to orality vs. literacy as a meaningful opposition (a policy which I try to question in this chapter), it is clear that the advent of literacy in a society by no means kills off the phenomena that we conceive of in terms of orality: as we saw in the previous chapter, conceptional orality is a matter of degree. Eric Have-lock, for example, proposes an oral residue in archaic and classical Greek societies, which are well beyond the phase of primary orality, the age of complete textlessness in all respects and on all levels. An even better picture of the interaction of orality and literacy is provided by the Middle Ages, another period of early writing, which has yielded much more evidence than archaic Greece on what people actually did with written texts. The medieval record has made it clear on a large scale that people can live graciously with written texts while retaining degrees of orality in the reception or transmission of those texts that would surprise the doctrinaire oralist.16
A first step toward bringing more nuance into the one-dimensional dichotomy between orality and literacy might be to fraction it into dichotomies of composition, reception, and transmission. This would allow texts to be both oral and written, according to the parameter chosen. For example, texts may be written and still be oral in their reception (i.e., in performance / recitation), an oral text may be written down so as to be distributed and read, and so forth. On this basis we could argue that pure orality, in which neither the poets, nor their audiences, nor the tradition has any contact with writing, is a phenomenon much more limited in its occurrence than the earlier oralists would allow. In practice there is always some aspect of the oral tradition that is no longer oral. And the very existence of texts pertaining to an oral tradition, as in the case of Homer, seems to prove that the tradition does not belong to the pure type.17
The fractioning of the orality-literacy contrast into composition, transmission, and reception may seem to have an advantage over the one-dimensional dichotomy, but in reality it amounts to nothing more than a multiplication of the contrast. In themselves the three components still yield absolute, binary distinctions between the written and the oral: a text is conceived of as orally transmitted or not, orally composed or not, with our literate notion of nonoral writing still intact as a cultural universal. The fractioning of the orality-literacy contrast, we may note, is Finnegan’s way of bridging the “great divide” between orality and literacy.18 Yet the result of this solution is the reconciliation of literacy with orality as a concept defined in literate terms, and a conception of orality as entirely encapsulated within a literate universe: with orality being confined to certain aspects of an oral poem, it is possible to continue studying it as if it were a written poem, without questioning or rethinking any of our common literate assumptions. In order to avoid this anachronism, we need to view writing and written composition from the perspective of speech, instead of looking at oral composition from the point of view of writing. This shift will mean reexamining our notions of writing and oral composition. These two concepts have been maintained as mutually exclusive alternatives by an anachronistic approach to writing and an idiosyncratic treatment of orality. Both the anachronism and the idiosyncrasy are due to the perspective discussed in Chapter 1: a tendency to see oral poetry as a special case whose characteristic properties must be studied in terms of production. This emphasis on composition has led scholars to neglect the importance of reception in the creation of the Homeric poems.
Lord’s writings, and indeed most writings in the oral-formulaic tradition, leave no doubt as to what is meant by “oral composition”: for oral theorists, oral composition is the use of the formulaic system provided by the tradition. The use of this system, or better, the dependence on it, is for Lord the quintessential feature of oral composition, and since written composition is not dependent on the formulaic system, there is a hard distinction, formulated by Lord as follows: “The oral singer thinks in terms of… formulas and formula patterns. He must do so in order to compose. But when writing enters, the ‘must’ is eliminated.”19
This treatment of the formula as the necessary condition for oral composition has led analysts to make a sharp distinction between such composition and writing.20 But it has also led to a distinction between the oral style and ordinary speech; in fact, these two oppositions are related, and if one is invalid the other is too. At the end of the previous chapter I hinted that, rather than separating Homeric discourse from ordinary language, the formula helps create that discourse by stylizing everyday speech. This relation between Homeric discourse and ordinary language (indeed between special speech and speech) necessarily implies a certain amount of listening knowledge on the part of the poet’s audience within the tradition. Beside composition, then, reception is an important factor in the function of formulas, and to the extent that this is the case, the formula ceases to be irreconcilable with writing. For the intended effect of formulas as speech units stylizing the units of ordinary speech was not in any way changed by their being written down, especially since the text produced was meant not to be read but to be heard.21
The allegedly destructive effect of writing on the capabilities of the poet is an artifact not only of the formulaic view of oral composition, but also of an anachronistic view of written composition. Lord may seem to see writing as the absence of the dependence on formulas, but in reality his conception is much broader: it is based on what writing means in our culture. Lord’s use of the terms “writing” and “written composition,” in fact, is underspecified and ignores a distinction that derives directly from the differentiation in the semantics of “oral” presented in the previous chapter: there is an important difference between “writing something down” and “composing something with the aid of writing.” In the former case, writing is a medium transfer, the transcoding of a phonic discourse into a graphic discourse. One could do this without having anything to do with writing as composition, the production of a text that is written as to its conception.22 In other words, one could write in a technical and physical sense, without one’s thought processes being governed, or even touched, by writing as a compositional process. Lord, in his discussion of the opposition between oral and written composition, restricts “writing” to our notion of compositional activity or process, ignoring the fact that this notion is the highly developed end point of a long, evolutionary process of interiorization, and applying it to the very earliest stages of that process, where a quite different sense of “writing” is much more appropriate.
It is essential to realize that for us writing is a way of organizing thought and texts, and that its underlying technology has become entirely transparent and taken for granted. Earlier stages in the development of the technology were very different. The technology and psychology of medieval writing has been investigated by M. T. Clanchy, with results that should be of great interest to classicists.23 In the writing culture of the eleventh century, composition in writing and writing technology were entirely different fields, with their own specialists. As Clanchy notes, “Just as reading was linked in the medieval mind with hearing rather than seeing, writing (in its modern sense of composition) was associated with dictating rather than with manipulating a pen. Reading and writing were not inseparably coupled with each other as they are today.”24 In other words, writing in the sense of composition was a form of speaking, a matter of voice, separated from writing in the technological sense, which was the transcription of the sound produced by the voice.25 A person, as Clanchy notes, might be able to “write” and still be illiterate by our standards: the writing that we couple with reading (in the visual sense) was an art in itself that was practised by highly trained scribes.26 But even on the relatively rare occasions when the writer took pen in hand, the concept of dictation and voice did not lose its relevance, because writing then amounted to dictating to oneself, writing down what one heard one’s own voice saying.27 The central insight to be gained, then, from Clanchy’s detailed investigation of early writing is that our distinction between speaking and writing is anachronistic and artificial when applied without further consideration to earlier stages of the literacy evolution. In particular, it appears that the distance between speech and text need not be so large as it is in our culture, either from the point of view of production, or from the point of view of reception, as we shall see in the next section.
Returning to the dictation of the Homer text, we now face the paradoxical situation that whereas Lord’s conception of writing has to be rejected, his scenario for the writing of the Iliad, at least in the earlier stages of the transmission of the text, is highly plausible, if not compelling, provided that the dictation of the Homeric poems is viewed in the right perspective. What I am suggesting is that the dictating bard “wrote” by his own standards and “spoke” by ours. The Iliad is real speech: in recomposing it, the poet actually produced every sound of which the poem consists and his thought processes, and hence the presentation and structure of his discourse, were not in any way governed by writing in our conceptional sense. And yet he produced a written text that is the basis—at first probably in competition with other, similar variants of the same discourse28—of the tradition that has eventually yielded our Homeric text.
So I agree with Lord that a poet (who might well have been illiterate by our standards) dictated the Iliad to an amanuensis. But to distinguish this dictation event from writing, as Lord does, is to draw lines that do not seem relevant for the earliest stages of literacy, when speaking and writing form a continuum that runs counter to our own cultural routines. The relation between the earliest Homeric texts and their speakers must have been one for which no close parallel exists in our culture. The speaker was neither a disinterested party in the dictation event (and hence external to the text produced there), nor the very “maker” (the literal sense of poiētḕs ‘poet’) and hence the author or owner of the text; in the textual fixation of his activity, he was rather a link in the transmission of Homeric discourse, and hence an agent in its survival. A possible scenario for this survival I consider in the final part of this chapter.
The Interdependence of Text and Performance
Texts are usually not created as self-contained objects; they have to be read.29 Being read, in fact, is no less than the completion and fulfilment of any text.30 The manner of this completion quite crucially depends on the nature of the text, its function and its manner of composition. A discussion of the writing of the Homeric text, then, is incomplete without a treatment of how it might have been read. In fact, a discussion of reading has the potential to deepen our understanding of writing, insofar as nothing is likely to be produced without some idea of the use that will be made of the product. We are concerned, then, with how the Homeric text might have been used in its original historical context.
Even without the preceding discussion, it would hardly be controversial to label as anachronistic the notion that the first text(s) of the Homeric poems were read by silent readers, or that the ancient reception of the poems was somehow a “written reception.”31 Less evident, but still unmistakable, is the anachronism in the suggestion that the first text of the Homeric poems was kept by its commissioner as a valuable possession, in principle not to be read at all.32 Such a conception separates the text from the oral tradition, just as it separates the oral poet from the production of the text; it reinforces the awkward coexistence of orality and literacy described earlier and leaves us with the idea of the Homeric text as an anomaly in an oral milieu, waiting to be discovered by someone who would treat it in the way in which we would treat it, by reading it or copying it.
In the context of archaic Greece, the idea of a text as a thing, or rather the idea of a thing as text, is more suitable for inscriptions, texts inscribed on stone or pottery. As Jesper Svenbro has argued, the term “inscription,” denoting the activity of inscribing, may be used to refer not so much to the text as such in its physical medium (e.g., stone), as to the speech situation created by the earliest archaic Greek inscriptions: it is the inscribed tomb that “speaks” in the first person (e.g., “I am the tomb of so-and-so”), thereby functioning as substitute for the speech of the author. Contrasting with this speech situation are inscriptions in the third person (“This is the tomb…”), presenting not so much a speech act that takes the place of speech as a transcription of an earlier speech.33 The transcript inscriptions, according to Svenbro, do not seem to appear before the middle of the sixth century. Whether the text on stone is an “inscript” or a transcript, however, its effect is always one of addressing a reader, in the presence of the text and in the absence of its author.34
The idea of texts as an address to a reader and as a substitute for speech is applied by the philosopher Paul Ricoeur not to the earliest stage of alphabetic writing, but to our writing. Like Svenbro, Ricoeur opposes the notion of inscribed discourse to the notion of transcription: “What is fixed by writing is … a discourse which could be said, of course, but which is written precisely because it is not said. Fixation by writing takes the very place of speech, occurring at the site where speech could have emerged. This suggests that a text is really a text only when it is not restricted to transcribing an anterior speech, when instead it inscribes directly in written letters what the discourse means.”35 Inscription is here opposed not to a subsequent stage (as in Svenbro’s opposition) but to an earlier stage of the development of writing; it may be characterized as the fixation of the meaning of the speech event, whereas the transcript is the fixation of the speech event itself.
Transcripts, then, find themselves wedged between two types of inscription, one ancient and one modern. The transcribed text differs from either type in a number of ways. It does not come in the place of speech, but rather crucially depends on speech. It does not speak itself, as a physical inscribed object does. Nor does it speak in the metaphorical, fictional way of the modern written text. Rather, it is the recording of speech, as the transformation of speech into a different medium.36 Finally, it does not directly address a reader in the way the inscriptions do: instead of directly transmitting a message to its recipient, it defers this activity to the speaker that gave rise to its existence, or to the speaker who acts as his substitute.37 And in doing so it has recourse to a reality outside: not a context that is constituted by the text as a physical object, nor a reality contained in the text,38 but rather a reality that must precede the text in time.
Now what does it mean to read a transcript, a text that according to Ricoeur is not really a text? The question cannot be seen in isolation from another question: what does it mean to write a transcript? Both questions lead us back to the most important transcript the archaic Greek period produced, the Homeric text. Being neither the physical embodiment of the speech act it represents, nor its autonomous container, the transcribed text needs voice to become physical, to accomplish the speech act that led up to its existence. In other words, in the case of the Homeric text the medium is different indeed from the message, to invert McLuhan’s well-known dictum. An analogy between the ancient transcript and the transcripts of our time, the fruits of advanced technology, might help here: like the musical compact disk or computer diskette, the Homeric text in its original transcript stage preserves an original message in a format that is quite alien to the message as such. To read such a text is not to receive the information transmitted but to restore the medium of the original message, to convert it to a format with which the user is familiar, or rather, which is understandable to the user at all. And the writing of the text is not the act of transmitting the information it contains, but saving it in a medium that is more suitable for that purpose.39
Writing and reading as “medium shifts”: this is the original use of the Homeric text that is proposed here. If this text, a transformation of speech into a different medium, owed its existence to speech and voice, then its reading was nothing other than the reversal of this process: the transformation of text back into the medium of speech, the reenactment of the speech represented by the text. It appears, then, that not only the distinction between speaking and writing, but also the one between writing and reading, begins to break down: if speaking is a matter of cognition, of the activation of ideas in a speaker’s consciousness, as I shall argue in detail in the next chapter, then reading is a matter of the re-cognition and re-activation of those same ideas, both in the reader’s and in the listeners’ consciousness.40 The reader of the Iliad, then, as well as of the other transcribed texts of early Greek literature, was no less a speaker than the writer of the text was, which amounts to saying that writing and reading were related to each other as performance and re-performance.
That the Iliad is a text to be heard, not read, may have anyone’s consent, but the present discussion aims to show that such a claim does not go far enough. Whoever wrote the Iliad, or gave orders for it to happen, did not merely write down a poem that was meant to be heard rather than read, still engaging in what we would call literary communication. Nor did the writing of the Iliad, whether as a unique event or as a series of events, a process, put an end to the public performance tradition of the Homeric epics. On the contrary, the writing of the Iliad was a masterful attempt—and a successful one, we may say—to secure this tradition by regulating the ongoing flow of performances and supplying it with a firm basis, a model for the rhapsode’s act. This model was a written text that was authoritative precisely because it derived from, and was meant to give rise to, authoritative speech (speech that, in its turn, derived its authority from being stylized and special).
The writing of the Iliad did not constitute the first literary text, with a strong footing in the oral tradition; nor did it constitute the often mentioned culmination (i.e., the death) of the oral tradition, as the Homeric poems were transformed into the higher medium of writing. The writing of the Iliad was not an oral tradition becoming a literate one, unless one sees this process as indissolubly connected with the oral conception of a written text: the original text was meant to represent the Iliad in its essential quality of speech and performance, and to be as such a normative model for reenactment. As the fixation of an ideal performance, the original text of the Iliad was an attempt to establish a canon, a means to exert power over future performances in the Homeric tradition.41 Or in other terms, the text of the Iliad is a transcript of the previous performance, and hence a script for the next one. In other words, the reading of the Iliad is the reenactment of its writing, with the actual text as nothing more (but also nothing less) than the preservation of the actual wording.42
Returning now to the orality-literacy contrast, we may assume that the contact between the Greek oral epic tradition and writing did not result in a mixture of what we conceive of as oral and literate features, with the almost inevitable outcome that those features that are oral are accommodated to the literate framework. Rather, the mixture goes deeper, resulting in a text the study of which is crucially bound up with the analysis of speech, considered as a phenomenon in itself, rather than as a concept defined from a literate point of view as the absence of writing. Accordingly, the body of this study will be a reading of the Iliad as a transcript, in an attempt to arrive at the speech that must once have been its primary referent.
To the extent that the orality-literacy contrast turns Homer into either an oral poet whose poems have somehow become texts, or a literate poet who in having liberated himself from the limitations of orality is one of us, it has to be rejected. Homer is not an oral poet, one of them, whose poems have been overheard, as it were, to be transcribed into the text which we possess. And because his writings need voice to become alive and do what they are meant to do, he is not one of us. It is the crucial importance of the human voice in the production, transmission, and reception of poetry whose essence lies in performance that led the medievalist Paul Zumthor to coin the term “vocality.”43 His coinage is gaining in popularity among medievalists as a way to characterize phenomena that are misrepresented by the orality-literacy dichotomy in its classical form.44 More important than terms, however, are the perspectives they reflect. And the importance of perspectives lies in our being aware of them.
1 See, for example, Finnegan 1977: 2, 16–24, 30–41.
2 See in particular Griffin 1980: xiv; Lloyd-Jones 1992. For a similar critique of Finnegan, see Nagy 1992b: 318–19.
3 For the lingering importance of this question, see most recently Erbse 1994.
4 See now Nagy 1996, a forceful antidote against this view.
5 Lord himself admits (1960: 125, 149) that written texts of actual performances are extremely rare.
6 Most recently Kirk 1985: 10–14.
7 See, for example, Kirk 1976: 126–27: “In Greece there are reasons for believing that the Iliad and Odyssey came very near the end of the creative oral tradition. The introduction of writing, and the supremacy of the monumental poems themselves, no doubt hastened the decline of creative oral poetry.”
8 More on repetition as a literate concept in Chapter 7 below.
9 See, for example, Foley 1991: 56–57; Hymes 1981: 82–83; Lord 1960: 99–102; Zumthor 1990: 49. The transmission of songs is not so much a matter of verbatim recall per se as of constraints on the memory of the performers; see Rubin 1995.
10 See also Bakker 1993a. For another critique of Kirk’s position, from a different perspective, see Parry 1966.
11 Nagy 1979: 8; see also Nagy 1990a: 52–56.
12 See now Nagy 1996, esp. 109–11, on the “five ages of Homer,” and the transition from genuine oral poetry via “transcript” to “scripture,” a process involving various centuries of transmission, that is, fixation in (re-)performance. See also the Conclusion below.
13 Lord 1953; 1960: 126–28, 148–57; 1991: 38–48; see also Jensen 1980: 81–95; West 1990; Janko 1990; 1992: 37–38.
14 Lord 1960: 129.
15 See, for example, Kirk 1985: 15–16: “For the oral tradition, which would have been killed off by any immediate and serious extension of literacy… . Writing had spread too far by the early years of the seventh century B.C. for the creative oral genius to flourish much longer.”
16 On medieval literacy and illiteracy see Bäuml 1980, 1984, 1993; Clanchy 1993; Schaefer 1991, 1992; Zumthor 1987.
17 See Bäuml 1984 on medieval “oral” texts and the problems those texts pose for the theory of oral composition, given the social context in which they have to be situated.
18 See, for example, Finnegan 1977: 17–18.
19 Lord 1960: 130. Yet formulas do persist after the introduction of writing, and in a later publication, Lord (1975: 18) denies these phenomena the status of formulas because they evidently had not facilitated oral composition. Convincing criticism in Smith 1977: 141–44.
20 Consequently, the alleged weakening of the concept of formula has been thought to bring Homer closer to writing. See Shive 1987.
21 This liberation of the formula from production in the oral-formulaic sense stems mostly from the work of medievalists; see, for example, Schaefer 1988; 1992: 59–87.
22 Oesterreicher (1993: 269–70) brings out this distinction by means of the terminological opposition between Verschriftung (writing in the medial sense, the transcoding of a discourse into text) and Verschriftlichung (writing in the conceptional sense, the textualization of a discourse). See also the Conclusion below.
23 Clanchy 1993. I cite from this second edition.
24 Clanchy 1993: 270–71.
25 On transcription and inscription, see below.
26 Clanchy 1993: 125–26; 131–32. But even literates would make use of scribes; see Zumthor 1987: 111; Clanchy 1993: 126.
27 This is the suggestion made by Magoun in a discussion (1953: 460) of the Anglo-Saxon poet Cynewulf. The ultimate consequence of this idea, writing as performance and the scribe as performer, has been recently pursued by Doane (1994)·
28 Medieval philology again provides suggestive, if indirect, evidence for what must have preceded the Alexandrian recension and redaction of Homeric poetry. See, for example, Zumthor 1987: 160–62, on mouvance, the dynamism of a poetic tradition, with textual variants resulting from its expansion in time and space (on which see now also Nagy 1996: 7–38). See also Cerquiglini 1989; Fleischman 1990b: 24–25; and O’Brien O’Keeffe’s 1987 study of the coexisting textual traditions of Caedmon’s Hymn in Old English, Latin, and vernacular. The manuscripts in the last-mentioned tradition provide less spatial organization (resulting in scriptio continua and rudimentary punctuation) and more textual variants than the manuscripts in Latin, and thus constitute a case of “the accommodation of literate to oral” (20) rather than the reverse.
29 The remainder of this chapter is based on Bakker 1993a: 15–18.
30 Cf. Ricoeur 1991: 151–52; Svenbro 1993: 45–46.
31 Harris 1989: 49; see also Taplin 1992: 37: “Not even the most zealous of the vigorous new school of scripsists would claim, I think, that Homer wrote his poem for a reading public. The important thing is that it was created to be delivered orally and to be heard.” Contrast Lord 1960: 131: “Writing as a new medium will mean that the former singer will have a new audience, one that can read.” See also Kirk (1985: 12), who indirectly makes the same assumption: “It [the Iliad] was still designed for a listening audience, since the spread of literacy cannot possibly have been such, by say 700 B.C., as to allow for a proliferation of copies and readers.”
32 See, for example, West 1990: 48.
33 Svenbro 1993: 28–43; see also Nagy 1992a: 35.
34 The presence of the text is discussed by Svenbro as the presence of the reader, who has to lend his voice to the speech act of the text during sonorous reading, being dispossessed of his own “I” (1993: 44–63), a conception of reading that runs counter to “animistic” interpretations of early writing (1993: 41). The magical or minatory use of early writing, however, cannot be completely discarded. See for example Harris 1989: 29 and Thomas 1992: 59, both citing the well-known text (seventh century B.C.E.) inscribed on an aryballos from Cumae: Ταταίες ἐμì λέkυθος · hòς δ’ ἂν με κλέφσει θυφλòς ἔσται, “I am the jug of Tataie; whoever steals me will be blind,” cf. Jeffery 1961: 238. Magical writing occurred in the Middle Ages as well; see Clanchy 1993: 314–17, 333–34·
35 Ricoeur 1991: 44, emphasis added.
36 Cf. Oesterreicher 1993: 269–70, on Verschriftung, the medial transformation of speech into text.
37 For functions of early texts other than transmission, see Forster 1989; Goody 1987: 54.
38 On the closure and autonomy of the written text as opposed to the contextual openness of speech, see Olson 1977: 264; Bäuml 1980: 251–52; Schaefer 1991: 119–20; 1992: 52–54. Ricoeur (1991: 47) speaks of “interception” to characterize the modified “referential contract” a written text entertains with the world.
39 In referring to CD-players and computer drives as lecteurs and to “saving” as enrégistrer, French usage for the management of computer files brings out the transcript status of “texts” in the new technologies more clearly than English.
40 I owe the play on “cognition” and “recognition” to the meaning of one of the Greek verbs for “reading” (άναγιγνώσκειν ‘to know again’, ‘recognize’), as well as to Svenbro’s discussion (1993: 165–66) of it. See also Nagy 1990a: 171. Svenbro repeatedly emphasizes the practice of writing in scriptio continua which made the reader’s voice necessary for the sounds of the text to be “recognized.”
41 On power in the relation between a writer and a future reader who is forced to lend his voice to the vocalization of words that are not his own, see Svenbro 1993: chap. 3. On canons, value, and power, see Most (1990), who stresses the close link between literacy and canonization. Outside classical philology see Smith 1983; 1988: 30–53. Among the possible motives for estab-lishing the Homeric tradition as canonical are Panhellenism (Nagy 1979: 1–11; 1990a: chap. 2); propaganda for aristocracy (Morris 1986; Latacz 1989: 85–90; Janko 1992: 38); and competition with new performance genres (Burkert 1987).
42 See Herington’s comment (1985: 45) on the absence of all mention of writing and texts, not only in the Homeric poems, but also in archaic poetry in general: “The texts were no part of the performed poem as such, but merely a mechanical means of preserving its wording between performances. You could hardly expect the archaic poem to allude to its own written text any more than you could expect a violinist in the concert hall to interrupt the music with an allusion to his printed score.”
43 See Zumthor 1987: 21: “C’est pourquoi je préfère, au mot d’oralité, celui de vocalité. La vocalité, c’est l’historicité d’une voix: son usage.” See also Zumthor 1987: 19: “Lorsque le poète ou son interprète chante ou récite (que le texte soit improvisé ou mémorisé), sa voix seule confere à celui-ci son autorite. Le prestige de la tradition, certes, contribue à le valoriser; mais ce qui l’intégre a cette tradition, c’est l’action de la voix.”
44 Schaefer 1992: 5–20.