Chapter 1

{|7} Let us begin with a passage from epic, where the epic is representing lyric, not epic. Specifically the lyric form is a song of lament. Penelope is at the moment comparing herself to a nightingale, the typical songbird of lament in ancient Greek traditions, who in a previous life had been a woman who suffered the ultimate grief of ‘inadvertently’ killing her own child:

This form of the story, where the unfortunate woman is daughter of Pandareos and wife of Zethos, is different from the better-known variant native to Athens, where Procne the daughter of Pandion and wife of Tereus deliberately kills her child Itys. [3] For now, however, the focus is on the variations not in the myth but rather in the actual wording of the passage. At verse 521, a variant reading πολυδευκέα (poludeukéa), the meaning of which is unclear, is reported by Aelian De natura animalium 5.38, in place of what we see in the text as quoted, πολυηχέα (poluēkhéa) ‘with many resoundings’.

In a book about the textual history of the Homeric poems, one critic notes that poludeukḗs (the nominative) is “rarer” than poluēkhḗs, adding: “we have seen, however, that many conjectures were introduced by the ancients into Homer and that sometimes the original was replaced by a rarer and more difficult word.” [4] In his footnotes, he offers this opinion: “in a poet such as Homer the simpler and less sophisticated expression is likely to be the original one.” [5]

Let us juxtapose this opinion with a general formulation offered by Albert Lord in response to modern cultural preconceptions about oral poetry: {8|9}

If we apply this line of thinking to the passage about the nightingale, we may ask whether the variant readings poluēkhḗs and poludeukḗs may both be “original,” if indeed they stem ultimately from variant performances in oral poetry. But how do we square variation in performance with variation in text? This question brings me to consider two concepts, mouvance and variance. As we will see, neither of these concepts provides an immediate answer to the question at hand, but together they help shape an ultimate answer.

The term mouvance was suggested by Paul Zumthor as a way of coming to terms with his perception that a medieval literary production like the Chanson de Roland is not so much a finished product, un achèvement, as it is a text in progress, un texte en train de se faire. [7] Viewing mouvance as a widespread phenomenon in medieval manuscript transmission, Zumthor defines it as a quasi-abstraction that becomes a reality in the interplay of variant readings in different manuscripts of a given work; he pictures mouvance as a kind of “incessant vibration,” a fundamental process of instability. [8] He links mouvance with the workings of oral tradition: for example, he suggests that certain textual variations in the Carmina Burana reflect the potential for actual variations in performance. [9] {9|10}

Zumthor’s idea of mouvance is not far removed from the idea of variance, which is a second concept that I apply to the general question of variation in performance and variation in text. This term variance was formulated by Bernard Cerquiglini in his influential Eloge de la variante. [10] For Cerquiglini, “medieval writing does not produce variants; it is variance.” [11] Unlike mouvance, however, Cerquiglini’s model of variance is not to be viewed in terms of oral tradition as reflected in written tradition: his definition insists that the written tradition itself is a matter of variance. [12]

While there is much to be learned from Cerquiglini’s far-reaching insights concerning the fact of variation in medieval manuscript traditions, it is more useful for now to pursue the implications of Zumthor’s term mouvance. There are two reasons.

First, since the term mouvance is predicated on a link with oral traditions, it seems apt for describing a wide variety of situations where we do indeed observe a distinct degree and even a distinct kind of textual variation: there is a genuine distinction, it can be argued, between variant manuscript readings stemming from errors or deliberate changes in the mechanical process of writing copies of previous manuscripts on the one hand and, on the other, variant manuscript readings reflecting a performance tradition that is still alive in a given culture. This observation about the writing down of poetic wording extends also to the writing down of melodic patterns that may accompany the wording, as the research of musicologists has suggested: in the case of the medieval French chansonniers, for example, “there was not only a scriptless culture next to a literate one, but also a notationless culture side by side with a very small notated one.” [13] It has even {10|11} been argued that “for each chanson there existed probably as many versions as there were jongleurs who performed that particular chanson.” [14]

Second, mouvance is not just a word coined by Zumthor: we are about to see a word meaning ‘to move’ which actually designates the process of mouvance and which is being used by a given songmaking tradition in referring to its own capacity for variation. Moreover, this word meaning ‘to move’ is used in this given tradition to refer to the song of both the nightingale and the poet.

It should be noted in advance that the tradition in question—even the culture in question—differs is in many ways from that of the ancient Greeks. We must therefore recognize from the start that any parallels we may find between the two traditions about to be compared are merely typological ones, and the implications of such parallelisms will have to be re-examined at length in terms of the available Greek evidence—to which we will turn in the two chapters that follow this one. Still, the poetic and even philological problems that we are about to see are in some respects strikingly similar to those faced by specialists in the ancient Greek Classics.

The key word in question is Provençal mover, the equivalent of French mouvoir and meaning, like the French word, ‘to move’. The textual tradition in question involves the medieval Provençal chansonniers—in this case a sub-set of songs or lyric poems attributed to a twelfth-century troubadour named Jaufré Rudel, prince of Blaye. The edition in question is a 1978 publication by Rupert {11|12} T. Pickens. The methodology that is adopted in this edition, as we will see, is particularly relevant to the questions at hand.

Pickens, as editor of Jaufré Rudel, is candid in telling his readers that he had originally undertaken his project in hopes of improving on an earlier edition of this troubadour by Alfred Jeanroy (1924): In abandoning one solution, where the goal is to reconstruct a given manuscript tradition back to one textual archetype, following methods established by Karl Lachmann, [16] this particular editor adopts an alternative solution by invoking the concept of mouvance as formulated by Zumthor, [17] and he explicitly connects “the poetics of mouvance” with the factor of “performance” in oral tradition. [18] More than that, this editor has discovered a {12|13} remarkable detail, to which I have just now drawn attention: that the Provençal word mover in the sense of French mouvoir and English move is actually used in Provençal songmaking to express an idea of mouvance.

The passage in question is the ending of Jaufré Rudel’s Song VI, version 1a: [19]

In other attested manuscript versions of Song VI of Jaufré Rudel, it is made clear by the composer that the intermediary transmitter of the song, described in the passage just quoted as ‘the one who will learn it from me’, must deliver it unchanged to two noblemen, who must in turn hear it. In version 1a of Song VI, the composition ends as just quoted. In other versions, however, the references to the destined audiences are followed by further references, resulting in a longer song. Of these other available versions, let us take as one example the last eight lines of Song VI version 1 (as distinct from 1a): [20] As another example, let us take the last eight lines of Song VI version 1b: [22]

Whereas according to Pickens the intermediary must deliver the song unchanged to two noblemen, “those who are destined to receive it must, on the contrary, introduce changes.” [23] It is {14|15} probably enough to say that the destined audience may “move” the song while the intermediary must not.

As we see from the wording of the variations in Song VI, the chance to move the song is equivalent to the chance to change it into a version different from that of the composer, even though the intermediary must keep the composer’s version the same until it reaches the destined audience. The noblemen who are to be the song’s audience are in turn to become the song’s new performers—and thereby the song’s recomposers.

The remarkable thing, moreover, is that mover ‘move’ can designate not only the recomposing of a song through reperformance but even its archetypal composition through its archetypal performance. As Pickens also notices, mover ‘move’ can refer not only to the changing of the song, as here, but also to the actual singing of the song, even to the actual beginning of the singing, as when the poet starts his song by picturing a nightingale as it sings, that is, as it moves its song. [24] Here is the beginning, for example, of Jaufré Rudel’s Song I version 1: [25]

Here in Song I of Jaufré Rudel, the symbol of the singing nightingale is drawn into a parallel with the singer who is the {15|16} poet. Just as the nightingale moves his song by singing—that is, by performing—so also the poet implicitly moves his own song by composing it. [26] This symbol provides an opportunity to compare a model developed by Lord, centering on composition-in-performance, with the model of Zumthor centering on mouvance. The symbol of the nightingale, deployed as it is to launch the poet’s composition, superimposes the medium of performance on the act, the fact, of composition. By comparing himself to the nightingale, the poet presents himself as one who performs as he composes. Just as the nightingale goes about his performance, so also the poet goes about his composition by performing it, by moving it. Just as the poet composes his song, so too his model, the nightingale: in Pickens’ translation, the songbird “composes (moves) his song rejoicing and joyous” (e mou so chant jauzen joios). We may recall the image in Aeschylus (Suppliants 60–67) of the mythical figure Procne, who has been transformed by the gods into a nightingale (aēdṓn, verse 62: ἀηδόνος): the songbird is pictured as literally “composing” (suntíthesthai, verse 65: ξυντίθησι) the sad song of her murdered child’s fate (verse 65: παιδὸς μόρον). [27]

Just as the nightingale’s song in Song I of Jaufré Rudel is an implicit model for the poet who hears him and makes his own song, so also the poet is the model for the noblemen who in turn hear him and make their own song by performing the song of the poet. To perform the song, however, is to recompose it, to change it, that is, to move it. In this light, mouvance is the same thing as recomposition-in-performance. The nightingale who is “composing” his song in Song I of Jaufré Rudel may serve as the model, the archetype, for the song of the poet, but even the songbird is in fact recomposing his own song by virtue of performing it. So it is apt for the nightingale to move his song, which is “original” for the poet but which is at the same time inherently recurrent and recomposed, much as every new season of spring is {16|17} a joyous event of inherent recurrence and recomposition, even re-creation.

In applying the idea of a “poetics of mouvance” to his edition of the songs of Jaufré Rudel, Pickens confronts a set of problems. Even without a single holograph, to be reconstructed according to the principles of Lachmann, [28] the question remains whether it is possible to explain the variations in Jaufré Rudel’s manuscript tradition simply in terms of the poet’s own lifetime activity: “a stemma,” Pickens explains, “could represent the career of a poet just as easily as a two-hundred-year tradition of manuscript transmission.” [29] It is in any case impossible to exclude the author himself from the process of varying his own composition. [30] Accordingly, in considering all the variations attested in the manuscript tradition of Song V, Pickens is willing for the moment to entertain the idea that all these variations may be “by” Jaufré Rudel; after all, Jaufré “was a troubadour who constantly reworked his material.” [31] Pursuing the question, Pickens concludes after an intensive analysis of the manuscript variations in both Songs I and V of Jaufré Rudel:

Even allowing for this much participation by the author himself in the process of variation, we are reminded of his own references to other participants, such as the noblemen in Song VI who are imagined as not only hearing the song in performance but also reperforming it themselves afterwards and thereby recomposing it. Pickens explicitly argues for the reality of such participation by invoking “the dynamic condition of the medieval poem,” with specific reference to Zumthor’s idea of mouvance. [33] Here he links his observations about the medieval Provençal troubadour with those of Harry H. Lucas about the medieval French trouvère: the song, Lucas argues, is not only created by the trouvère but also re-created by any number of fellow trouvères, as well as amateurs, before it is ever incorporated into the manuscript tradition of the chansonnier, and most likely even afterward, so that the song of a trouvère can truly be said to be a work of collaboration. [34]

By now, Pickens has brought into play two “solutions”: according to the first, Jaufré Rudel is the only contributor to the “creative acts”; according to the second, these creative acts “are seen as operating in transmission as well.” [35] Moreover, “Jaufré himself affirms the principle of change as esthetically proper to his genre, so that it might be said that mouvance is an aspect of the intention of his songs.” [36] Precisely in this context Pickens introduces the passage {18|19} from Song VI where the poet urges that the noblemen must hear and then reperform the song of Jaufré Rudel, with the implicit assumption that they will thereby recompose it as well. [37] In this same context we have seen the poet urging the intermediary transmitter not to “move” the poem, that is, not to recompose it.

This detail is essential for coming to terms with questions of authorship in this culture. If indeed Jaufré Rudel is not the only contributor to the “creative acts,” then how exactly is he an author? Here we may recall the striking formulation of Rupert Pickens: “The conventions and traditions of the courtly lyric have conspired to efface the author and to create at least as many Jaufré Rudels as there are medieval anthologies.” [38]

Let us go beyond the problems raised by this particular example in this particular culture and ask a more general question: what is it, in any case, to be an author in any tradition where performance is needed to make a song come to life? [39] Applying the observation of Parry and Lord that composition and performance are aspects of the same process in oral traditions, I suggest that authority in performance is a key to the very concept of authorship in composition. [40] In the present example from Jaufré Rudel as well, it is authority in performance that is crucial: the poet’s song does not become authoritative until it is performed in an authorized setting. Only then does the song become real, authentic. Thus the intermediary transmitter is enjoined not to “move” the song of Jaufré Rudel because he is as yet unauthorized to do so. This injunction by the poet is presented not as a statement of fact so much as a stylized gesture to his intended audience. Thus I doubt that this unnamed intermediary is really understood by the poet to be a “mechanical” transmitter who is not a “re-creative” singer. [41] I propose instead that the authorization of the composer is implicitly not enough because the transmitter as performer must also be authorized by his audience, who are presumed to be authoritative members of the song culture. [42] {19|20}

The intermediate transmitter of the troubadour, who is potentially a jongleur, becomes an authoritative performer through the dual authorization of composer and audience. [43] It would be misleading, though, to generalize the jongleur as a mechanical performer who repeats the song of the authoritative composer, the troubadour; it would also be misleading to set up a strict dichotomy between a “creative” troubadour and a “re-creative” jongleur. [44] In the troubadour traditions, the transmitter of song becomes a potential troubadour by virtue of reperforming the song. It all depends on the circumstances of reception: in Jaufré’s song, for example, it is implied that the transmitter of his song must have approval from both the composer and the audience which is to hear the transmitter’s performance. The transmitter is to be authorized on the grounds that both the composer and the audience of noblemen are authoritative. The noblemen may reperform and thereby recompose the song precisely because they are presumed, by the song, to be authoritative. So also the nightingale “moves” the song because he has the authority to do so. Only in this case, the authority is not social but elemental—even archetypal. We will return to this theme of archetypal authority in Chapter 2, where I will explore the idea that repetition is archetypal.

We see another example of performative intermediacy in Song II of Jaufré Rudel, where the song is pictured as being transmitted from the troubadour to a nobleman through an intermediary, named Fillol or ‘Godson’ in some versions. [45] Here is the text of Song II, version 1, strophe v: [46]

Even in the sort of situation where the composer allows himself to express a concern that his song may be exposed to unauthorized performance, as if there were a danger that someone will “move” it in a negative sense, this concern turns out to be a rhetorical way of seeking reassurance from the audience that the performance really is authoritative, so that those who heard the song and learned to perform it can thus implicitly “move” it in a positive sense, much as the nightingale “moves” his own song. {21|22} We come back to the case of Song VI, which is predicated on the poet’s satisfaction with his composition: bos es lo vers can no.i falhi ‘the poem is good, since I did not fail in it’ (version 1b strophe v). The anonymous transmitter is enjoined to learn the song from the poet exactly as it was composed: in the different versions of the song, we hear that the transmitter must therefore ‘beware lest he fail in it and break it up’ (gart se no.i falha ni.l pessi, version 1b strophe v) or ‘beware lest he fracture it or break it up’ (gart no.l fran[ha] ni [no.]l pezi , version 2a strophe vii) or ‘beware lest it move or change’ (gard si non mueva ni camgi, version 1a strophe iv) or even perhaps ‘beware lest anything changes me’ (guart si que res no mi coambi, version 1 strophe vii). [50]

The fact that even this poetic injunction against variation survives by way of variants is a striking example of a paradox that is characteristic of a wide variety of oral traditions: a tradition may claim unchangeability as a founding principle while at the same time it keeps itself alive through change. Outsiders who are looking in, as it were, on a given tradition can be objective about any change that they do observe. Insiders, however, are apt to be subjective. Participants in a given tradition may of course choose to ignore any change whatsoever. [51] If they do recognize change, however, either it must be negative or, if it is to be positive, it must not really be change after all. In other words, positive change must be a “movement” that leads back to something that is known, just as negative change leads forward to something that is unknown, uncertain, unpredictable. And yet, even if positive change is a moving back toward whatever is known, certain, and predictable, all the more will it be deemed to be an ongoing process of improvement, not deterioration, by those who participate in the tradition. In fact, it will be an improvement precisely because such positive ‘movement’ aims at the traditional, even the archetypal.

From this point of view, the noblemen who are to hear the song of the poet are described in Song VI of Jaufré Rudel as {22|23} destined to improve that song by way of a presumed authoritativeness inherent in their reperformances. To quote from one of the variants that we have already seen, ‘The melody is good, and they will do there | whatever things from which my song will grow more noble’. [52] When the nightingale ‘moves’ his song, it is a matter of positive change because tradition is reactivated. If, however, a jongleur ‘moves’ the song of a troubadour in an unauthorized situation, it is a matter of negative change because tradition breaks down. For a performer of a song to ‘move’ it in a negative sense is to ‘change’ it, even to ‘break’ it.

Just as the idea of ‘moving’ a song can be turned from a negative to a positive sense, however, even the idea of ‘breaking’ a song can be made positive in the poetics of mouvance. The negative poetics of the verb franhar ‘break’, as deployed in the poet’s injunction to the transmitter not to ‘break’ the song, are echoed by the positive poetics of the verb refranhar, to be interpreted in another poem of Jaufré Rudel as ‘refract’ in referring specifically to the singing of the nightingale and, in response, the singing of the poet. [53] Before we reflect on the meaning of the metaphor inherent in the image of ‘refraction’, let us consider its precise context in the song:

The metaphor inherent in the Provençal verb refranhar can be explained as an auditory equivalent of a visual metaphor, the ‘refracting’ of light (as in Latin refringere). The driving image of refraction also accounts for two Provençal nouns: refrins, meaning ‘echo’ (as a part of sound that repeats itself), and refrim, meaning ‘birdsong, sound, refrain’. [55] The verb refranhar can also refer to the musical process of modulation in song: much as light is refracted through glass or a prism, so also the musical sound of song is modulated. [56] When the nightingale ‘turns and refracts [refranhar] and polishes’ his song, the songbird is being envisaged as a craftsman who is constantly engaged in the process of improving the work of his craftsmanship, in principle coming ever nearer to the finished product. The poet echoes the songbird as he reaches the end of the strophe just quoted, and so also by implication the other singers must echo the poet, as they too must ‘turn and refract and polish’ the song, refining it and ‘bringing it to an end’. The end of one singer’s ‘refinement’, however, is the beginning of another’s, and each beginning, each new ‘movement’, is a return to tradition. In this theme of ‘refinement’, we see the ultimate image of improvement as an eternal return to the traditional, which is envisaged as an eternal musical modulation.

Let us return here once again to the troubadour’s image of the songbird, with one more example of the word mover ‘move’ in the archetypal sense of referring to birdsong. In the sole extant version of Song IV, the first four verses of the last strophe run as follows: [57]

It is time to propose a reformulation of the idea of mouvance, supplementing it with the usage of mover as we have seen it operate in a troubadour’s poetics—and in light of one given editor’s detailed and patient work on the texts attributed to Jaufré Rudel. I propose, then, that mouvance is the process of recomposition-in-performance as actually recognized by a living oral tradition, where the recognition implies the paradox of immediate change without ultimate change.

On the basis of his editorial experience, Rupert Pickens concludes that “at least in the case of Jaufré Rudel, mutation is appropriate to the lyric genre.” [58] According to this line of thinking, the courtly lyric of Jaufré is not “authoritative” in the same sense as Scripture, in that the work is freed to be “re-created and re-generated.” [59] I agree, though I stress that the authoritativeness of Jaufré’s tradition is as real as that of Scripture, with the basic difference being that the troubadour’s words do not insist on the idea of unchangeability, typical of the claims of scriptures in a wide variety of cultures. At a later point, we will consider in further detail the very notion of “scripture.”

Pickens observes about the patterns of mutation in the lyrics of Jaufré Rudel that “it is impossible to determine at what points his personal interventions ceased and his transmitters continued the process of perfecting beyond his personal intentions.” [60] Which leads to this conclusion: “Given these conditions, under which it is impossible to rediscover Jaufré’s intentions (i.e., the extent of his personal involvement in the creation and regeneration of his works), each manifestation of a song must be considered to be, in its own right, as valid a whole, complete poem as any other versions.” [61] In light of these findings, this editor of medieval texts differs from the approach of previous {25|26} scholars like Gaston Paris and Alfred Jeanroy, whose goal was to recover a given author’s archetypal text. [62] But he also differs from the approach of other editors like Joseph Bédier, whose own findings about textual variation led him to abandon the idea of recovering the archetypal text, but whose goal remained nevertheless the idea of recovering at least the closest thing to an archetype. [63] Pickens offers the following critique:

While offering his own adjustments in line with the hermeneutic principle of mouvance, Pickens has this to say about the criteria of editing proposed by Lucas: “to my knowledge, they have never been observed in an edition of medieval lyric poetry.” [65] In a retrospective work, Pickens with good reason describes his own 1978 edition of Jaufré Rudel, with its “multitext format,” as “the first widely recognized edition attempting to incorporate a procedure to account for re-creative textual change.” [66]

The need for a multitext format in editing texts is most evident in the case of manuscript traditions where the phenomenon of phraseological variation seems to reach all-pervasive proportions. A striking example is the textual transmission of the medieval French epic, the Chanson de Roland. [67] As Ramón Menéndez Pidal observes, three of the earliest manuscript versions of the Chanson {26|27} have not a single identical verse in common with each other. [68] He concludes that such a degree of textual variation is symptomatic of an ongoing oral tradition, and that in fact an oral tradition stays alive through its variations and reworkings. [69] Following Menéndez Pidal, Michael Zwettler has concluded that early Arabic poetry is likewise the product of a vigorous oral tradition, as reflected in the extraordinary wealth of variants transmitted in the textual tradition:

Following Lord, Zwettler emphasizes not only the multiformity inherent in the oral tradition, as in the observation just quoted, but also the futility of attempting to establish an “original” text on the basis of attested variants. If indeed oral poetry lives through its variants, then it is ironic, Zwettler finds, “that scholars of Arabic poetry have so often cast doubt upon the ‘authenticity’ or ‘genuineness’ of this or that verse, poem, or body of poems or, sometimes, of pre-Islamic poetry in general, because they have found it impossible to establish an ‘original version’.” [71] Following Zwettler, Olga Davidson argues that the degree of textual variation in the medieval Persian manuscript transmission of the epic Shāhnāma of Ferdowsi likewise reveals the product of an oral tradition. [72] Using as a test-case a randomly-selected passage from the Shāhnāma, Davidson shows that “every word in this {27|28} given passage can be generated on the basis of parallel phraseology expressing parallel themes.” [73] She finds, moreover, that “the degree of regularity and economy in the arrangement of phraseology” is “suggestive of formulaic language.” [74] She advocates the need for a monumental new edition of the Shāhnāma that would account for all attested variants above and beyond the verifiable instances of scribal error, in order to come to grips with “the full creative range of the Shāhnāma tradition.” [75]

Davidson too, like Zwettler before her, stresses the futility of trying to recover the archetypal fixed text from a mass of textual variants that can all be judged “genuine” in terms of the poetic tradition that had generated these variants. There are many other studies that focus on variation in textual transmission as a mark of oral tradition, but it will suffice for the moment to mention just one more example, the work of Joseph Nagy on medieval Irish traditions: he concludes that “the bewildering proliferation of variants which often characterizes the medieval literary transmission of Irish narrative takes on new meaning when viewed as the imprint of an ongoing oral tradition.” [76]

In all three of these studies just mentioned, as also in the study of troubadour poetry that we had considered in detail earlier, it is the degree of multiformity in the textual tradition that leads to the conclusion that an oral tradition is at work backstage, as it were. In each case, the nature of the given oral tradition may be quite different, but the effect of variation may be strikingly similar. The question arises, then, whether we can find cases of a comparable degree of variation in ancient Greek—or for that matter in Latin—textual traditions.

From a survey of Martin West’s handbook on textual criticism, we see from the precise wording of his descriptions that the most likely candidates are (1) Greek tragedies, “which suffered extensively from interpolations by actors (or at any rate for their use), probably more in the fourth century BCE than at any later {28|29} time,” [77] (2) the comedies of Plautus, “which may have undergone something of the sort on a smaller scale in the second century,” [78] and (3) the Homeric poems, through the “embellishment” of rhapsōidoí ‘rhapsodes’ as reflected in quotations by authors of the fourth century and in the attested papyri, especially those dated before the middle of the second century. [79]

In West’s descriptions here, as also in most accounts written by Classicists, the textual traditions of ancient tragedy, comedy, and epic are not organically related to the performance traditions of these three forms. It is as if the performance traditions were impositions on the text, rather than historical antecedents of the text. Thus textual variations are explained in terms of textual rather than performative conditions even when the medium in question is overtly performative. It is not the performance itself that is held accountable for textual variation, but the text that is used by the performer. Thus the so-called “actors’ interpolations” are invoked to explain textual change in drama. Where epic, too, is concerned, West explains the existing patterns of textual variation in terms of textual causes: the “embellishment” of rhapsodes, he notes, is primarily characterized by “additional lines of an inorganic nature (often borrowed from other contexts).” [80]

Where the Classical text stems not so much from performance traditions as from “practical learning,” the role of ongoing recomposition as a direct cause of textual variation is more easily understood. [81] With reference to works of practical learning, West makes the remark that “commentaries, lexica and other works of a grammatical nature were rightly regarded as collections of material to be pruned, adapted or added to, rather than as sacrosanct literary entities.” [82] A similar description applies to more exalted compositions such as treatises on rhetoric, perhaps even to such canonical works as the essay on style attributed, rightly or wrongly, to Demetrius of Phaleron. An ideal example of a literary {29|30} form in which massive textual variation results from active ongoing recomposition is the Hippocratic corpus: as West puts it, parts of this corpus “were subject to revision or rearrangement.” [83]

James Zetzel is exceptional among contemporary Classicists in comparing explicitly the textual variations resulting from traditions of practical learning with those resulting from traditions of performance. [84] Zetzel’s list of “works that provide practical learning,” subject to “wholesale alteration,” includes such other diverse examples as the cookbook of Apicius, magical prescriptions, and the Digest of Roman Law. [85] Adopting for all such works the operative metaphor of the cookbook, Zetzel goes on to describe them as “unprotected texts.” [86] By contrast, he notes that “wholesale alteration of the sort just described rarely occurred in the transmission of standard texts of the school curriculum ..., precisely because they generally stayed out of the kitchen.” [87] A notable exception among standard works, however, is a text that stems from a performance tradition: Zetzel cites Plautus’ comedies, “the only major Latin literary text that I know in which wholesale alteration has taken place, precisely because our manuscripts largely descend from actors’ versions.” [88]

Zetzel makes clear his sense that the ultimate sources of textual variation in such cases as Plautine comedy are the traditions of performance, even if we posit the texts of the performers as an intermediate source. Zetzel’s emphasis on performance is subtle but clear, as we see in his assessment of textual variations in Greek tragedy: “similarly, our texts of Greek tragedy incorporate actors’ interpolations that must have been added at a time when the plays were still being performed.” [89]

Where textual variations stem from outright performance traditions, as in the case of drama, it seems justified to compare the phenomenon of mouvance. In cases of textual variations stemming from the vicissitudes of practical learning, as in the case of, say, a {30|31} cookbook, the broader term variance, as coined by Cerquiglini, seems a more appropriate point of comparison. It must be recognized, however, that the prescriptive traditions of a cookbook may also be to some extent “performative.” As for instances of textual transmission following the schoolbook mentality, in sharp contrast with the “cookbook” mentality, we may expect the minimum degrees of variation. [90] In such instances, it seems unnecessary to insist on any point of comparison with either mouvance or variance.

Let us focus on the subject of variations in the textual traditions of Greek drama and epic. Unfortunately, historical circumstances have prevented our access to an ample range of attested variants in the performance traditions of drama. As for epic, there is relatively more attested evidence, but the history of Homeric transmission is in any case far more complicated than what we have seen so far in the case of medieval traditions. If indeed a multitext format is needed for editing medieval texts like the songs of Jaufré Rudel, then perhaps the need is even greater in the case of ancient Greek drama and epic. But the difficulties are greater as well.

Here we must confront a major intellectual and aesthetic obstacle for Classicists. It is clear to us that the actor in a Sophoclean drama is not another Sophocles, the rhapsode of epic is not another Homer. How, then, could an actor’s so-called “interpolation” or a rhapsode’s “embellishment” be comparable to the ipsissima verba of a Sophocles, of a Homer? Any answer must be formulated in relative terms, and varying degrees of comparison are to be applied. To start at one extreme of the spectrum, we might say that the compositions attributed to, say, Sophocles, are relatively less adaptable to the process of recomposition because the phraseology itself is less capable of variation in the first place. At the other extreme, the compositions attributed to Homer were in the earliest recoverable periods doubtless far more adaptable to {31|32} ongoing recomposition because the phraseology traditionally operated on the very principle of variation. In the later periods of Homeric transmission, however, this adaptability becomes drastically reduced, for reasons that we will later on examine in some detail.

Concentrating on the principle of variation in earlier phases of Homeric transmission, let us return to the passage with which we started this chapter, the Homeric image of the aēdṓn ‘nightingale’ in Odyssey xix (518–523), who is pictured as singing her beautiful song at the onset of yet another new season of springtime (519: ἔαρος νέον ἱσταμένοιο), perched in the dense foliage (520) and pouring forth her voice as she keeps changing it around (521: τρωπῶσα), thick and fast (521: θαμά)—a voice that is described as poluēkhḗs ‘with many resoundings’ (521: πολυηχέα) or, according to the variant reported by Aelian (De natura animalium 5.38), as poludeukḗs (521: πολυδευκέα), the meaning of which word we have not yet examined. Here, then, is Aelian’s report:

In considering these two variants poluēkhḗs and poludeukḗs in Odyssey xix (521), I am ready to argue that both are legitimate, both ultimately generated from the multiform performance tradition of Homer. I will also argue that the variant reported by Aelian has an archaic meaning that even he could not fully understand. This meaning, as we will see in the next chapter, captures the very essence of continuity in variation.

We have already considered one critic’s analysis of these two variants. [91] His view is that, although poludeukḗs is “rarer” than poluēkhḗs, the editorial principle of lectio difficilior should not be applied in this case: “we have seen ... that many conjectures were introduced by the ancients into Homer and that sometimes the original was replaced by a rarer and more difficult word.” [92] Arguing that the reading poludeukḗs could not be “original,” this critic goes on to say: “If this reading were original, it would be very surprising that the leading critics of antiquity ignored it, while it was preserved by a later author like [Aelian], who had far fewer resources at his disposal.” [93]

Let us contemplate for a moment the resources at Aelian’s disposal. Born in the third quarter of the second century CE, he seems to typify the kind of scholars who populate the compendium that we know as Athenaeus’ Deipnosophistai. As for his specific source in this case, Aelian says that he ‘had heard tell’—that is to say, that he had read or had someone read to him—about the nightingale, and that his source was a man called Charmis of Massalia. [94] At the very mention of Massalia, a question may arise {33|34} in our minds whether this man from Marseille may also have been Aelian’s actual source for the variant reading in the Homeric passage about the nightingale—and whether this variant may even have stemmed from the so-called Massaliōtikḗ, the city text of Homer that was often cited by the Homer critic Aristarchus. [95] While we cannot with any certainty answer such a question, we can say at least this much: that Aelian in De natura animalium 5.38 uses the report of Charmis about nightingales as a pretext for offering what he implies is his own scholarly interpretation of the Homeric variant poludeukéa describing the sound the nightingale makes.

‘I hear tell’, says Aelian, ‘from Charmis of Massalia’ (Χάρμιδος ἀκούω τοῦ Μασσαλιώτου λέγοντος...) that the nightingale is a creature that loves the tradition of the Muses: it is philómouson. It also loves the fame that songmaking brings with it: it is philódoxon. When the nightingale sings in the wilderness, as Charmis reports, her song is simple and unvaried; when she sings in captivity, however, she becomes aware that she now has an audience, and so she now shifts to a different style, that is, ποικίλα ... ἀναμέλπειν καὶ τακερῶς ἑλίττειν τὸ μέλος ‘she sings in a varied [poikíla] way and meltingly changes the song around’. Then Aelian quotes the passage from the Odyssey, xix 518–521 in the same form that we have in our latter-day editions, featuring poluēkhéa ‘with many resoundings’ as epithet of phōnḗn ‘voice’ at verse 521.

It would seem, at first sight, that the reading poluēkhéa ‘with many resoundings’ has already aptly illustrated the report of Charmis about the more patterned voice of the nightingale who has become aware that she has an audience. Then, if it were an afterthought, Aelian adds at this point the scholarly observation that ‘some even write’ (ἤδη μέντοι τινὲς ... γράφουσι) not poluēkhéa but poludeukéa, which he glosses as τὴν ποικίλως μεμιμημένην ‘making imitation [mímēsis] in a varied [poikílōs] way’, and he compares the formation poludeukéa with the negative adeukéa (ἀδευκέα), which he glosses as μηδ᾿ ὅλως παρατραπεῖσαν ἐς μίμησιν ‘not at all adapted for imitation’. The use of the word mímēsis, which can be translated for the moment as {34|35} ‘imitation’, is crucial for the interpretation of the variant reading poludeukéa, as we will see in Chapter 2.

Whether this variant reading is Aelian’s own additional piece of information or, as is more likely, it stems ultimately from Charmis of Massalia, it seems that Aelian’s own interpretation does not do full justice to the archaic meaning of poludeukéa. Aelian is interested mainly in the nightingale’s versatility as an imitator, whereas the epithet poludeukéa draws attention also to the continuity of the singer’s performance, as we will see in the next chapter.

We may be dissatisfied also with other facets of Aelian’s interpretation. He seems to be saying that the expanded variety of the songbird’s repertoire as a musician (philómouson) is a singer’s desire for fame (philódoxon) before an expanded audience, now that she finds herself in captivity. We may note the fact that the nightingale in the ancient Greek songmaking traditions is generically female, singing a song of lament, unlike the male nightingale of the medieval troubadour traditions, which is conventionally singing a love-song. [96] This is not to say that songs of lament and songs of love are incompatible: they are in fact regularly interchangeable in Greek traditions, for example. [97] The point is, rather, that the nightingale sings from the heart, as it were, the afflictions of love or death. As a captive, the nightingale may sing her heart out not only because of any pride in her songmaking virtuosity but also, more basically, because of her sorrows. [98] In light of the {35|36} tragic story underlying the nightingale’s lament in the Greek traditions, we may wish to reinterpret the variety of the songbird’s song not only as a performer’s response to an ever-widening audience but also as a victim’s response to an ever-widening threshold of pain and suffering. We may note in passing the title of the autobiography of the poet Maya Angelou, I Know Why the Caged Bird Sings. [99] (The wording of the title is taken from the exquisite poem “Sympathy,” by Paul Laurence Dunbar [1872–1906], which needs to be read in its own right.)

Surely the nightingale in the Odyssey sings her beautiful song of lament about the fact of her own suffering. The woman who had killed her own child—inadvertently, the Odyssey claims—had suffered much, but now she suffers more, transformed into a songbird singing her own mournful song. Of course the song is as beautifully varied as it is mournful, and then it can become even more so with her new misfortune, her loss of freedom and her new identity as a captive singer. We may recall the words in Toni Morrison’s Beloved, a novel centering on the twin tragedies of slavery and infanticide: “it was not a story to pass on.” [100] And yet, as we will see in the next chapter, the song of the nightingale has a continuousness, a continuity. It is indeed passed on.

The continuity is implicit in the variety. We may note the meaning of trōpôsa (τρωπῶσα) in Odyssey xix (521), describing the nightingale as she changes around or literally turns the sound of her beautiful song. Let us recall the song of the henightingale in Song I of Jaufré Rudel, where birdsong serves as model for the song of the poet, and where the model itself is that of recomposition, not just composition, in that even the songbird is in fact recomposing his own song by virtue of performing it. The nightingale moves his song, which as we have seen is inherently both recurrent and recomposed, much as every new season of spring is a joyous event of inherent recurrence and {36|37} recomposition, even re-creation. To the extent that mouvance in the poetics of the troubadour conveys the idea of continuity through variation, we can see in the usage of the word trōpôsa (τρωπῶσα) in Odyssey xix (521), the first attestation of this idea in European traditions.

The usage of the root *trep- as in trōpôsa, with the basic meaning of ‘turn’ or ‘change around’ and with the implied meaning of continuity through variation, lives on in the songmaking traditions of ancient Greece, taking on the form of the noun-derivative trópos. This word refers in practical terms to a given pattern of modulation in the singing voice, corresponding to a given pattern of accordatura or tuning of the accompanying lyre (an ideal example is Pindar Olympian 14.17). More generally and theoretically, trópos is a ‘style’ of melody (“Aristotle” Problems 19.38). Such a ‘style’, in the usage of ancient Greek music theory, “may be determined by any combination of scale-structure, pitch, and melodic shape.” [101] Eventually, trópos becomes a word for ‘style’ in general, especially rhetorical style (Plato Republic 3.400d). In the technical language of rhetoric, it takes on the meaning of ‘trope, figure of speech’, and it is at this stage of semantic development that Greek trópos is borrowed into Latin as tropus (Quintilian Institutio oratoria 8.6). From the high language of the schools, the noun tropus is borrowed into the popular language, which creates on its own a derivative verb *tropāre. This “Vulgar Latin” verb *tropāre is actually attested in late Latin authors as the compound verb contropāre (also adtropāre), [102] and it lived on as an inherited form in the Romance languages or as a borrowed form in the Celtic languages, witness modern French trouver and medieval Irish trop. [103] Most important for now, *tropāre also lived on as the medieval Provençal verb trobar, meaning ‘find, invent’ or ‘compose poetry’, and in the noun-derivative trobador, later spelled as troubadour. This meaning is pertinent, as we shall see, even to the modern French verb trouver, in the everyday sense of ‘find, invent, discover’.

This is not the time to attempt a systematic retracing of the semantic route of continuity in meaning from *tropāre to trobar in {37|38} the language of the troubadours. We may simply review the intuitive summary of Peter Dronke concerning medieval musical traditions during the century and a half following the death of Notker in the year 912:

The concept of lyrical dialogues, which we may picture in ancient Greek terms as the mímēsis—let us translate the word for now as ‘imitation’—of speech by way of song, lies at the heart of the medieval troubadour traditions, where one side of a dialogue, the side of the lover, is highlighted as if it were a monologue. As a performance, such a monologue is of course implicitly a dialogue with the audience who is being addressed, as also with the beloved, real or imaginary.

Here the chapter comes to a halt. But the subject of the nightingale’s song, sung again and again in all its varieties through time, from the aoidoí of ancient Greece to the troubadours of medieval Provence, will continue. {38|39}