Chapter 3. Homer and the Evolution of a Homeric Text

In order to find a historical context for the writing down of the Homeric text, the most obvious strategy is to look for a stage in ancient Greek history when the technology of writing could produce a text, in manuscript form, that conferred a level of authority distinct from but equivalent to the authority conferred by an actual performance. As we have seen in the previous chapter, the opportunity for a text to become the equivalent of a performance already exists in the case of early poetic inscriptions from the eighth century BCE onwards. But it is another matter when it comes to manuscripts as distinct from inscriptions. It is only at a later period, after 550 BCE or so, that we begin to see actual examples of the use of writing in the form of manuscripts. As we will now see, some of these examples involve the use of a manuscript for purposes of a transcript, that is, in order to record any given composition and to control the circumstances of any given performance. [1]

One such example comes from the era of Peisistratos and his sons, tyrants at Athens in the second half of the sixth century BCE: from various reports, we see that this dynasty of the Peisistratidai maintained political power at least in part by way of controlling poetry. [2] One report in particular is worthy of mention here: according to Herodotus, the Peisistratidai possessed manuscripts of oracular poetry, which they stored on the acropolis of Athens (5.90.2). [3] I draw attention to a word used by Herodotus in this context, kéktēmai {65|66} ‘possess’, in referring to the tyrants’ possession of poetry. As I have argued elsewhere, “the possession of poetry was a primary sign of the tyrant’s wealth, power, and prestige.” [4] We may recall in this context the claim in Athenaeus 3a, that the first Hellenes to possess “libraries” were the tyrants Polykrates of Samos and Peisistratos of Athens.

For Herodotus, the control of poetry by tyrants was a matter of private possession, a perversion of what should be the public possession of the city-state or polis. [5] An example of public possession is evident in Herodotus’ description of a consultation of the Oracle at Delphi that took place after the era of the tyrants, at a time when the Athenian state was already a democracy: we see from Herodotus’ account of the consultation that there was a conventional procedure for the use of oracular poetry, and this procedure can be divided into three stages: (1) the poetry could be transcribed by delegates that had been sent to Delphi in order to hear the actual delivery of the oracular poetic message; (2) these delegates were to bring home the transcript, from Delphi to the people of the polis of Athens; and (3) these same delegates would proclaim to the people, on the basis of the transcript, the poetic message of the oracle (ἀπήγγελλον ἐς τὸν δῆμον 7.142.1). [6] This procedure, which Herodotus describes as if it were a regular practice in Athens during the era that followed the fall of the tyrants, stands in marked contrast to the earlier practice of the Peisistratidai, which he describes as if it were a usurpation of the public possession of poetry: to have private possession of poetry as a text is to control the occasion of its performance and the contents of its composition. [7]

It is, then, in this era of the tyrants, the Peisistratidai, that we may imagine a plausible historical occasion for the transcription of the {66|67} Homeric poems in manuscript form. As we will see further below, the plausibility seems enhanced by various reports from the ancient world about an event that some Classicists have described as the Peisistratean recension of the Homeric poems. Before we can take up the whole question of such a “recension,” however, we must first examine further what exactly it means to speak of a transcript in the era of the Peisistratidai and thereafter.

It is easiest to start with a negative consideration: a transcript is not the equivalent of a performance, though it may be an aid to performance. In other words, a transcript need not be a speech-act. Aside from the testimony that I have already considered from sources like Herodotus, there is also the evidence of pictures, in vase paintings, representing the use of manuscripts. As we examine the representations of people using manuscripts, that is, books or papyrus-rolls in these vase paintings, we can see that “books seem to have a mainly mnemonic role supplementing oral recitation.” [8]

By the fifth century, however, there are indeed cases where something that was written in a manuscript form could indeed become the equivalent of a performance. In other words, we now start seeing clear traces of an impulse to re-enact the performative dimension by way of a manuscript’s written word. For example, in the case of Herodotus’ own large-scale composition, the Histories, it is clear that the writing down of this composition in manuscript form was indeed meant to be the equivalent of a performance, a genuine speech-act. [9] For Herodotus, to say “I write something” is deliberately made equivalent to saying “I say something” in solemn, public situations (e.g. 2.123.3, 4.195.2, 6.14.1, 7.214.3) - because whatever he is saying in this solemn way at the moment that the reader reads it has already been written down. [10] In other words, whatever is staged as being said by the speaker of the Histories of Herodotus is predicated on the fact that it has already been framed by the medium of writing, as if the staging itself were a creation of the writing. [11] “I am saying this now in what I have written, therefore I am writing this now.” {67|68}

The main difference between this kind of stance in early literature and similar effects in modern literature is that, for someone like Herodotus, the situations in the text where it is written “I write this now” instead of “I say this now” still match conventional situations in public life where one would normally say “I say this now.” Thus, even as late as the second half of the fifth century, the era of Herodotus, the actual writing down of any given text could still be viewed as tantamount to the production of yet another performance, to the extent that the technology of writing could produce a text that conferred a level of authority parallel to that conferred by an actual performance. [12] In the Histories of Herodotus, the written text is not only an equivalent to performance: it is considered the authoritative equivalent.

By contrast, a transcript is not an equivalent to performance but merely a potential means to achieve performance. To that extent, a transcript in the era of the Peisistratidai may be viewed as a prototypical “script.” In what follows, I will argue that whatever poetry might have been transcribed in this era still has to be defined in terms of oral poetics, that is, it has to be viewed as resulting from a fundamental interplay between the dimensions of composition and performance. Further, I will continue to argue that there is no evidence for assuming that the Iliad and Odyssey, as compositions, resulted from the writing down of a text. The point remains that the writing down of a composition as text does not mean that writing was a prerequisite for the text’s composition—so long as the oral tradition that produced it continues to stay alive. Moreover, the writing down of any kind of composition that could otherwise be produced in performance will not necessarily freeze the process of recomposition-in-performance. [13] There are numerous parallels in European medieval literature, as we see for example in the following description, with {68|69} reference to fifteenth-century English manuscript production: “the surviving manuscripts of a poem like Beves of Hamptoun make it clear that each act of copying was to a large extent an act of recomposition, and not an episode in a process of decomposition from an ideal form.” [14] Paul Zumthor describes as mouvance the process whereby the act of composition, so long as this composition belongs to a living tradition of composition-in-performance, is regenerated in each act of copying. [15]

So the question is: if indeed a transcript could have been made of the Homeric poems in the era of the Peisistratidai, how exactly are we to imagine the use of such a transcript? As a parallel to the pattern that we have seen reflected in the account of Herodotus, where he describes the Peisistratidai as establishing control over oracular poetry, we may suppose that this dynasty sought to control epic poetry as well. We will return later to this aspect of the parallelism. The problem for now is, the parallelism cannot be extended in other respects. Oracular poetry is visibly occasional, responsive to the ad hoc requirements of time and place. [16] The epic poetry of the Iliad and Odyssey, on the other hand, is distinctly non-occasional and at least notionally unchanging, to be performed again and again on a seasonally recurring basis at formal occasions like the Feast of the Panathenaia. As I have already suggested, the Homeric poems reveal a high degree of text-fixation or textualization, and again I am using the concept of text without the implication that writing is a prerequisite. [17] So the question still remains: what use is there for a transcript of such a text?

Another way to approach the question is to consider the textuality of the Homeric poems. Although I will continue to argue that no writing had been required to bring about this textuality, I propose now to rethink the question in terms of a later era when written texts were indeed the norm. Even in this later era, I insist, any written text {69|70} that derives from an oral tradition can continue to enjoy the status of a recomposition-in-performance—so long as the oral tradition retains its performative authority. [18] In such a later era, where written text and oral tradition coexist, the idea of a written text can even become a primary metaphor for the authority of recomposition-in-performance. As I will now argue, the very concept of a “Peisistratean recension” can be derived from such a metaphor.

The intrinsic applicability of text as metaphor for recomposition-in-performance helps explain a type of myth, attested in a wide variety of cultural contexts, where the evolution of a poetic tradition, moving slowly ahead in time until it reaches a relatively static phase, is reinterpreted by the myth as if it resulted from a single incident, pictured as the instantaneous recovery or even regeneration of a lost text, an archetype. In other words, myth can make its own “big bang” theory for the origins of epic, and it can even feature in its scenario the concept of writing.

A particularly striking example is a myth about the making of the Book of Kings in the classical Persian epic tradition: There is a comparable myth in Old Irish traditions, concerning the recovery of the “lost” Cattle Raid of Cúailnge. [20] There are also similar themes in Old French traditions. The work known as Guiron le courtois, {70|71} for example, composed around 1235 CE, lays the foundations for its authority by telling of the many French books that were produced from what is pictured as an archetypal translation of a mythical Latin book of the Holy Grail. [21]

We can find further examples in the living oral traditions of India. In Telugu society, there is an aetiological myth explaining why the Palnāḍu epic is now sung by untouchable Malas: “the epic, it is claimed, was first written by a Brahmin poet, torn into shreds, discarded, and then picked up by the present performers.” [22] Another example comes from the Pābūjī oral epic tradition of Rajasthan: “a bhopo [= bhopā or medium, folk-priest] of Pābūjī like Parbū will insist that the epic he performs ‘really’ derives from a big book composed by high-caste Cāraṇ poets and kept in Pābūjī’s native village of Kolū: for him it is the written word that carries authority.” [23]

I have saved till now two examples from ancient Greece. Both involve myths, but in the second case the myth in question seems at first to be a report based on historical events. This is not the place to explore at length the role of myth as a reflex of institutional history in ancient Greece. [24] For the moment, it is enough to say that both myths about to be examined concern the institutions of the respective communities to which they belong, and that the two communities in question are Sparta and Athens. Since the myths may not seem like myths at first sight, I will for the moment refer to both of them by way of the more neutral term “story.”

The first story is from Sparta, centering on the topic of a disassembled text, scattered here and there throughout the Greek-speaking world, and then reassembled in a single incident, at one {71|72} particular time and place, by a wise man credited with the juridical framework of his society, Lycurgus the lawgiver. According to this story, as reported by Plutarch, Life of Lycurgus 4.4, Lycurgus brought to Sparta the Homeric poems, which he acquired from a lineage of epic performers called the Kreophyleioi, descended from Kreophylos of Samos. [25] In archaic Sparta, it appears that the Kreophyleioi of Samos were more authoritative than the epic performers elsewhere credited with the transmission of Homeric poetry, the Homeridai of Chios: as Aristotle reports (F 611.10 Rose), the Homeric poems were introduced to Sparta by Lycurgus, who got them from the Kreophyleioi when he visited Samos. [26] With reference to the Homeric poems, Plutarch reports that Lycurgus, having received them from the Kreophyleioi, ‘had them written down’, ἐγράψατο (Life of Lycurgus 4.4), and that he then ‘assembled’ them (ibid.). What follows in Plutarch’s account is worth citing verbatim: ἦν γάρ τις ἤδη δόξα τῶν ἐπῶν ἀμαυρὰ παρὰ τοῖς Ἕλλησιν, ἐκέκτηντο δὲ οὐ πολλοὶ μέρη τινά, σποράδην τῆς ποιήσεως, ὡς ἔτυχε, διαφερομένης· γνωρίμην δὲ αὐτὴν καὶ μάλιστα πρῶτος ἐποίησε Λυκοῦργος ‘for there was already a not-too-bright fame attached to these epics among the Greeks, and some of them were in possession [verb kéktēmai] of some portions, since the poetry had been scattered about, carried here and there by chance, and it was Lycurgus who was the first to make it [= the poetry] well-known’ (Life of Lycurgus 4.4).

In this passage, I have highlighted the word kéktēmai ‘possess’, with reference to the “ownership” of Homeric poetry. The same word is used by Herodotus in referring to the “ownership” of oracular poetry on the part of the Peisistratidai, the dynasty of tyrants at Athens (5.90.2). [27] Elsewhere, Herodotus refers to the manipulation, by the {72|73} Peisistratidai, of oracular poetry with the help of one Onomakritos, described in this context as diathétēs ‘arranger’ of this poetry (7.6.3). [28]

This detail about a diathétēs ‘arranger’ of poetry brings us to the second of the two ancient Greek examples of the kind of myth that we are presently considering. This second story is from Athens. Even more than the first story, it seems at first to be not a myth but a straightforward account of a historical event. As I will argue presently, however, it can be explained as a myth that happens to account for a historical process. This myth, like others we have already examined, accounts for the evolution of a poetic tradition which, moving slowly ahead in time until it reaches a relatively static phase, is reinterpreted by the myth as if it resulted from a single incident, pictured as the instantaneous recovery or even regeneration of a lost text, an archetype. Again, myth is offering a “big bang” theory for the origins of epic. As I will argue, what makes the Athenian version of this type of myth more distinct than other versions is that we know more about the historical circumstances of its ultimate political appropriation.

For now, a summary of the Athenian story will suffice. According to Tzetzes (Anecdota Graeca 1.6 ed. Cramer), a certain Onomakritos, the same person whom we have just seen described by Herodotus as a diathétēs ‘arranger’ of oracular poetry (7.6.3), was the member of a group of four men commissioned in the reign of Peisistratos to supervise the ‘arranging’ of the Homeric poems, which were before then ‘scattered about’: διέθηκαν οὑτωσὶ σποράδην οὔσας τὸ πρίν. [29] There is a convergent report in Aelian, Varia Historia 13.14, where the introduction of Homeric poetry to Sparta by Lycurgus the lawgiver is explicitly compared to a subsequent introduction of the Iliad and Odyssey to Athens by Peisistratos. The most explicit version of the story can be found in Cicero De oratore 3.137: Peisistratos, as one of the Seven Sages (septem fuisse dicuntur uno tempore, qui sapientes et haberentur et vocarentur), [30] was supposedly so learned and eloquent that “he is said to be the first person ever to arrange the books of Homer, previously scattered about, in the order that we have today”: qui primus {73|74} Homeri libros confusos antea sic disposuisse dicitur, ut nunc habemus. [31] In these accounts of the supposedly original Athenian reception of Homeric poetry, reinforced by the story in “Plato” Hipparchus 228b claiming that it was Hipparchus, the son of Peisistratos, who introduced the Homeric poems to Athens, we confront the germ of the construct that has come to be known among Classicists as the “Peisistratean recension.” [32]

On the basis of the other narrative traditions that we have examined concerning the topic of an archetypal text that disintegrates in the distant past only to become reintegrated at a later point by a sage who then gives it as a gift to his community, the story of a “Peisistratean recension” can be explained as a myth that bears clear signs of political appropriation by the Peisistratidai. Particularly striking is the parallelism in the accounts of Plutarch and Cicero between Lycurgus, lawgiver of Sparta who gives his community the Homeric poems, and Peisistratos, described as one of the Seven Sages, who likewise gives his community of Athens the Homeric poems. I repeat an observation made earlier: Greek myths about lawgivers, whether they are historical figures or not, tend to reconstruct these figures as the originators of the sum total of customary law. [33] Traditions about the Seven Sages, the most prominent of whom is Solon the lawgiver of Athens, are closely linked to those about lawgivers in general. [34]

The distinction between historical tyrants on the one hand and mythical lawgivers or sages on the other is oftentimes blurred. [35] In {74|75} the account of Aelian, the parallelism between the lawgiver par excellence and the tyrant Peisistratos is explicit: just as Lycurgus gives the Homeric poems to Sparta, so also Peisistratos gives the Homeric poems to Athens. The parallelism may possibly be extended: just as Lycurgus is reputed to have brought the Homer performed by the Kreophyleioi of Samos to Sparta (Plutarch Life of Lycurgus 4), so also the Peisistratidai seem to have taken credit for bringing the Homer performed by the Homeridai of Chios to Athens. [36]

The Homeric poems took shape, according to Athenian versions of the story, in the context of what is now called the Panathenaic rule, where the performance of the Iliad and Odyssey by rhapsōidoí ‘rhapsodes’ was not allowed to favor some parts of the epic narrative over others, in that the narrative had to be performed by one rhapsode after another in sequence (“Plato” Hipparchus 228b and Diogenes Laertius 1.57). [37] I will have more to say at a later point about this story and about its pertinence to the question of rhapsodes. For now I simply draw attention to the fact that this “Panathenaic rule” is attributed by the sources either to the Peisistratidai (“Plato” Hipparchus 228b) or to Solon himself (Diogenes Laertius 1.57). The parallelism linking the Peisistratidai with Solon, lawgiver of Athens, can be compared to the parallelism linking Peisistratos with Lycurgus, lawgiver of Sparta. Again we see indications of the appropriation of a myth by the Peisistratidai.

The politics involved in the attribution of this Homeric institution to the Peisistratidai are to be expected. Also to be expected, I suggest, is that this attribution to the tyrants would in time be ousted by an attribution to Solon, once the tyrants themselves were ousted: it makes sense for the credit that they once could claim as would-be lawgivers to be retrojected to an earlier figure, Solon, whose status as primary culture hero of the State, originator of a wide variety of institutions, makes him the ideal recipient of any credit taken away from others who came after him. {75|76}

These stories about the fixation of Homeric performance traditions will help provide an answer to a basic question about the dimensions of Homeric composition: how, after all, are we to account for the sheer length of these epics? How was it possible for the Iliad and Odyssey to reach the monumental proportions of over 15,000 and 12,000 verses respectively? [38]

In order to appreciate the mythological answer to such questions, let us first consider a common feature of oral poetic traditions, which is, the potential for expansion or compression of a given topic, either way. Analysis of this phenomenon in living oral epic traditions makes it clear that neither the relatively more expanded nor the relatively more compressed versions need necessarily be considered basic from the internal standpoint of the given tradition. [39] Such an absence of standardization in length is of great importance for coming to terms with the Homeric tradition, where we find a great variety of compression as well as expansion of themes. Of these two features, expansion and compression, the more noticeable is of course expansion, in that the impact of an overall composition may keep getting augmented with the expansion of size, whereas any instances of compression, even if they happen to be miniature feats of artistic skill, will have to be contained within an expanding composition. In the aesthetics of Homeric poetry, multiple marvels of compression are fated to be contained by the singular marvel of ultimate expansion, such as the monumental composition of the Iliad. [40] It is much harder for us to appreciate compression, enclosed as it is within the expansive monumentality of the whole Iliad, the whole Odyssey. {76|77}

Aside from instances of bravura in compression and expansion, however, we should also expect to find in living oral traditions the more ordinary levels of these phenomena, where the context of a given occasion leads to shortening or lengthening by default. Even in such default situations, however, it appears that relatively longer versions of a given epic performance have more to say about their given occasion than do shorter versions. We have considerable evidence about the potential monumentality of Indian epic performance, in both size and scope, and how that monumentality is managed in terms of actual performance. A key element is the subdivision of monumental epic performance into performance segments—which may be called “episodes”: Drawing attention to the principle of unevenly weighted episodes in this description, I propose that the evolution of ancient Greek epic involved a progression from uneven weighting toward even weighting. Let us take as our point of departure the example of uneven weighting {77|78} that we have just considered in the Indian evidence. We find a striking analogy in the following description of Homeric poetry at an early stage when it was supposedly divided into separate narrative portions, which have actually been described by one commentator as “episodes”: [42] For earlier stages of Homeric poetry, we may link the principle of uneven weighting with the preeminence, let us say, of the Achilles theme in the narrative traditions about the Trojan War—at the expense of themes magnifying the epic deeds of other heroes at Troy. {78|79} This preeminence or even popularity of Achilles is surely still reflected by the Iliad that we have. As for the later stages of Homeric poetry, however, we see an integration of epic themes that had been sloughed off, as it were, by the driving theme of Achilles, so that the Iliad in the end has something to say about practically every epic theme connected with the Trojan War: it re-stages, in the final year of the war, a Catalogue of Ships—which would be more appropriate, like the Catalogue of the Cypria, to the very beginning of the Trojan War; it re-introduces Helen of Troy—as if for the first time, re-matching Menelaos and Paris to fight over her as if she had just been abducted; it even re-tells, toward the end of its own narrative, the Judgment of Paris—which had ultimately started it all. [44] Such feats of narrative integration, I suggest, exemplify an impulse of even weighting.

I propose now to refine this notion of even weighting by considering the actual sequence as well as the content of what is being performed. Let us start with a comparative example, taken from the description by Keith H. Basso of a girls’ puberty ritual or na ih es as performed by a group of Apaches living at Cibecue on the Fort Apache Indian Reservation in Arizona. [45] This ritual is made up of eight distinct parts or “phases”: I draw attention to the positioning of the songs within a preordained sequence. There is a set of 32 or so of these songs that are sung at the na ih es, and it is believed that the whole set, collectively called go jon sin’ ‘full-of-great-happiness songs’, was “originally” sung by an archetypal female known as Changing Woman. [47] The totality that is {79|80} realized every time its “parts” are performed in song is merely notional. Moreover, there is a correlation here of meaning and sequence, where part of the meaning is the sequence: In this case, the option of free variation, a function of meaning, is subordinated to the non-option of fixed order, which is also a function of meaning. Such a pattern of subordination, I suggest, is a feature of even weighting.

In light of these considerations, let us reconsider the final and definitive stages of Homeric poetry, marked by a tightening-up of epic conventions. My focus is on a story that explains the institution, in Athens, of a customary law applying to the festival of the Panathenaia, where the performance of the Iliad and Odyssey by rhapsōidoí ‘rhapsodes’ was not allowed to favor some parts of the epic narrative over others, in that the narrative had to be performed by one rhapsode after another in sequence: According to another version, this law about fixed narrative sequence in Homeric performance was introduced not by Hipparkhos of the Peisistratidai but rather by the lawmaker of Athens himself, Solon: τά τε Ὁμήρου ἐξ ὑποβολῆς γέγραφε ῥαψῳδεῖσθαι, οἷον ὅπου ὁ πρῶτος ἔληξεν, ἐκεῖθεν ἄρχεσθαι τὸν ἐχόμενον ‘he [Solon the Lawgiver] wrote a law that the works of Homer were to be performed rhapsodically [rhapsōidéō], by relay [ex hupobolēs], so that wherever the first person left off, from that point the next person would start’ (Diogenes Laertius 1.57). [51] We have already observed that the story is appropriate to either Solon or Peisistratos in the role—deserved or undeserved—of lawgiver. More important for now, in any case, is that fact that these stories attempt to explain the unity of Homeric composition as a result of sequencing in performance.

As we have seen, Classicists conventionally refer to this customary law about Homeric performance as the Panathenaic rule. [52] I {81|82} suggest that this “rule” is actually a Greek reflex of the principle of even weighting, indicative of a communalization of repertoire. I also suggest that an even more appropriate term might be equalized weighting.

Once the sequencing of Homeric “episodes” becomes a tradition in its own right, it stands to reason that any cross-referencing from one episode of the sequence to another will also become a tradition. It is from a diachronic perspective that I find it useful to consider the phenomenon of Homeric cross-references, especially long-distance ones that happen to reach for hundreds or even thousands of lines: it is important to keep in mind that any such cross-reference that we admire in our two-dimensional text did not just happen one time in one performance—but presumably countless times in countless reperformances within the three-dimensional continuum of a specialized oral tradition. The resonances of Homeric cross-referencing must be appreciated within the larger context of a long history of repeated performances. [53]

It is also from a diachronic perspective that we can appreciate the institution and even the concept of rhapsōidoí ‘rhapsodes’, who are the performers associated with the pattern of equalized or even weighting in the Homeric narrative tradition. In my earlier work on the rhapsodes, I concluded: “it is simplistic and even misleading to contrast, as many have done, the ‘creative’ aoidós [‘singer’] with the ‘reduplicating’ rhapsōidós.” [54] In terms of my evolutionary model for the making of Homeric poetry, the figure of the rhapsode is the very embodiment of an evolving medium that continues, in the course of time, to put more and more limitations on the process of recomposition-in-performance. The succession of rhapsodes linking a Homer in the remote past with Homeric performances in the “present” of the historical period—as extrapolated from such accounts as Plato’s Ion—is a diachronic reality. This reality can only be distorted by any attempt to arrive at a synchronic definition of rhapsodes, meant as some kind of foil for an idealized definition of Homer. {82|83}

The diachronic reality of the rhapsōidoí ‘rhapsodes’ is expressed indirectly by the various myths that link the fixity of Homeric composition with the fixation of rhapsodic performance. According to the myths that we have considered so far, the reintegration of a prototypical text causes both the fixity of Homeric composition and the fixation of rhapsodic performance. But there are other myth patterns that are even more radical, making the concept of a sequence of rhapsodes more basic than the concept of a prototypical text. As we have seen, the evolution of a poetic tradition, moving slowly ahead in time until it reaches a relatively static phase, can be reinterpreted by myth as if it resulted from a single incident, a “big bang,” pictured as the instantaneous recovery or even regeneration of a lost text, an archetype. As we will now see, the “big bang” can also be pictured as the actual sequencing of rhapsodes.

Among the explanations given by the scholia for Pindar Nemean 2.1d for the concept of rhapsōidós, one version tells of a reintegration of Homeric poetry by way of rhapsodic performance, which is equated with a process of sewing the disintegrated parts back together again: οἱ δέ φασι τῆς Ὁμήρου ποιήσεως μὴ ὑφ᾿ ἓν συνηγμένης, σποράδην δὲ ἄλλως καὶ κατὰ μέρη διῃρημένης, ὁπότε ῥαψῳδοῖεν αὐτήν, εἱρμῷ τινι καὶ ῥαφῇ παραπλήσιον ποιεῖν, εἰς ἓν αὐτὴν ἄγοντες ‘but some say that—since the poetry of Homer had not been brought together under one thing, but rather had been scattered about and divided into parts—when they performed it rhapsodically [rhapsōidéō], they would be doing something that is similar to sequencing or sewing, as they produced it into one thing’. In the scholia for Dionysius Thrax, Codex Venetus 489, it is reported that the Homeric poems were “sewn together” (συνερράφησαν) by Peisistratos himself. [55]

The scholia for Pindar Nemean 2.1d proceed to offer yet another version, which explicitly links the term rhapsōidós with the innovation of an equalized distribution of “parts” assigned to the performers of Homeric poetry: {83|84} οἱ δέ, ὅτι κατὰ μέρος πρότερον τῆς ποιήσεως διαδεδομένης τῶν ἀγωνιστῶν ἕκαστος ὅ τι βούλοιτο μέρος ᾔδε, τοῦ δὲ ἄθλου τοῖς νικῶσιν ἀρνὸς ἀποδεδειγμένου προσαγορευθῆναι τότε μὲν ἀρνῳδούς, αὖθις δὲ ἑκατέρας τῆς ποιήσεως εἰσενεχθείσης τοὺς ἀγωνιστὰς οἷον ἀκουμένους πρὸς ἄλληλα τὰ μέρη καὶ τὴν σύμπασαν ποίησιν ἐπιόντας, ῥαψῳδοὺς προσαγορευθῆναι, ταῦτά φησι Διονύσιος ὁ Ἀργεῖος ‘others say that, previously—since the poetry had been divided part by part, with each of the competitors singing whichever part he wanted, and since the designated prize for the winners had been a lamb—[those competitors] were in those days called arnōidoí [= lamb-singers], but then, later on—since the competitors, whenever each of the two poems [56] was introduced, were mending the parts to each other, as it were, and moving toward the whole composition—they were called rhapsōidoí. These things are said by Dionysius of Argos [between 4th and 3rd centuries BCE; FGH 308 F 2]’.

The metaphor inherent in the word rhapsōidós itself is pertinent to these myths. The compound noun rhapsōidós means, etymologically, ‘he who sews together [rháptō] the song(s) [aoidḗ]’. [57] This metaphor is actually attested in the syntax of a song composed by Pindar, referring to the very beginning of a Homeric performance by the Homērídai ‘Sons of Homer’: ὅθεν περ καὶ Ὁμηρίδαι ῥαπτῶν ἐπέων τὰ πόλλ᾿ ἀοιδοὶ ἄρχονται, Διὸς ἐκ προοιμίου ... ‘starting from the very point where [hóthen] the Homērídai, singers [aoidoí] of sewn-together [rhaptá] utterances [épē], most often take their start [= verb árkhomai], from the prelude [prooímion] of Zeus ...’ (Pindar Nemean 2.1–3).

The point of all departures, as this song claims, is the ultimate god, Zeus. [58] As such, Zeus is invoked in the prooímia ‘preludes’ for the ultimate songs, the songs of Homer. It is precisely within the framework of this form, the prooímion ‘prelude’ (plural prooímia), that the author of a given song conventionally identifies {84|85} himself. [59] In the Homeric Hymn to Apollo, to which Thucydides explicitly refers as a prooímion (3.104.4–5), the first-person speaker identifies himself as the blind singer of Chios, whose songs will win universal approval in the future (Hymn to Apollo 172–173); the singer of this hymn claims to be none other than Homer, “author” of the universally approved Homeric poems. [60] According to this prooímion, the performer who speaks these words in the first person is not just representing Homer: he is Homer. [61]

The prooímia or ‘preludes’ are represented in Pindar’s song as performances of the Homērídai ‘Sons of Homer’; this name applies to a lineage of rhapsodes in Chios who traced themselves back to an ancestor called Hómēros, or Homer (scholia for Pindar Nemean 2.1, Plato Phaedrus 252b, Strabo 14.1.33–35, Contest of Homer and Hesiod p. 226.13–15 Allen). [62] Pindar’s representation of the Homeric prooímion is pertinent to the etymology of this word, which I have up to now translated conventionally as the ‘prelude’ of a song. It stems from oímē ‘song’, so that the pro-oímion is literally the front or, better, the starting end of the song. [63] Further, pro-oímion is the starting end of the thread of the song, if indeed the noun oímē stems from a verb-root meaning ‘sew’. [64] The metaphor implicit in this etymology of oímē, where making songs is equated with a process of sewing together or threading songs, is explicit in Pindar’s reference to the Homeric rhapsodes at the beginning of Nemean 2, where rhaptá ‘sewn together’ is applied to épē in the sense of poetic ‘utterances’. The same metaphor is implicit, as we have seen, in the etymology of the {85|86} actual word for rhapsode, rhapsōidós, ‘he who sews together [rháptō] the song(s) [aoidḗ]’. [65]

This metaphor of sewing together the song(s) must be contrasted with a related metaphor in archaic Greek traditions, that of weaving the song(s), which is in fact so old as to be of Indo-European linguistic provenience. [66] An example is this phrase of Pindar, F 179: ὑφαίνω δ᾿ Ἀμυθαονίδαισιν ποικίλον ἄνδημα ‘I weave [huphaínō] a patterned [poikílos] headband [that is, of song] for the Amythaonidai’. [67] As we see from such passages, song is being visualized as a web, a fabric, a textile (Latin textilis, from texō ‘weave’), or—to use only for the moment an English word that no longer retains its metaphorical heritage—even a text (Latin textus, again from texō). [68]

As we juxtapose these two metaphors for songmaking in archaic Greek traditions, weaving and sewing, we discover that the second of the two is more complex than the first. [69] The idea conveyed by rhapsōidós, ‘he who sews together [rháptō] the song(s) [aoidḗ]’, corresponds to an idea conveyed by the myths: that many and various fabrics of song, each one already made, that is, each one already woven, become re-made into a unity, a single new continuous fabric, by being sewn together. The paradox of the metaphor is that the many and the various become the single and the uniform—and yet there is supposedly no loss in the multiplicity and variety of the constituent parts. In effect, this metaphor conveyed by the concept of rhapsōidós amounts to an overarching esthetic principle, one that may even ultimately settle the ever-ongoing controversy between advocates of unitarian and analytic approaches to Homer.

Eustathius, in his Commentary on the Iliad (vol. 1 p. 10), quotes the Pindaric description (Nemean 2.1-3) of the Homērídai ‘Sons of {86|87} Homer’ as ῥαπτῶν ἐπέων ... ἀοιδοί ‘singers [aoidoí] of sewn-together [rhaptá] utterances [épē]’, interpreting these words as a periphrasis of the concept inherent in the word rhapsōidoí ‘rhapsodes’. Eustathius goes on to offer what he considers a second interpretation (again, 1.10), claiming that this concept of sewing together can be taken either in the sense that we have seen made explicit in Pindar’s wording or in a more complex sense - a sense that I think is actually implicit in the same Pindaric wording - which emphasizes the characteristic unity of the Iliad and the Odyssey: ῥάπτειν δὲ ἢ ἁπλῶς, ὡς εἴρηται, τὸ συντιθέναι ἢ τὸ κατὰ εἱρμόν τινα ῥαφῇ ὁμοίως εἰς ἓν ἄγειν τὰ διεστῶτα. σποράδην γάρ, φασί, κειμένης καὶ κατὰ μέρος διῃρημένης τῆς Ὁμηρικῆς ποιήσεως, οἱ ᾄδοντες αὐτὴν συνέρραπτον οἷον τὰ εἰς ἓν ὕφος ᾀδόμενα ‘sewing together [rháptō] either in the simple sense, as just mentioned, of putting together or, alternatively, in the sense of bringing different things, in accordance with some kind of sequencing [heirmós] in sewing, uniformly into one thing; for they say that Homeric poetry, after it had been scattered about and divided into parts, was sewn together by those who sang it, like songs sung into a single fabric [húphos]’.

Following up on what he considers two different interpretations of Pindar Nemean 2.1-3, Eustathius (again, p. 10) offers a third one as well: that the concept of sewing together songs is parallel to the concept of rhapsōidía, a word that he uses to designate any one of the twenty-four books of the Iliad or Odyssey. Although this interpretation still invokes the esthetic principle of sewing songs together into a unified whole, the songs are now visualized textually, as separate rhapsōidíai or ‘books’ of Homer. Eustathius contrasts this usage of rhapsōidíai as ‘books’ of Homer of with what he describes in the same context as earlier conventions of “the ancients,” the majority of whom had referred to the totality of Homeric poetry as rhapsōidía ‘rhapsody’ and to those who sing it, as rhapsōidoí ‘rhapsodes’ (p. 10): οἱ δὲ πλείους τῶν παλαιῶν τήν τε ὅλην Ὁμηρικὴν ποίησιν ῥαψῳδίαν λέγουσι καὶ ῥαψῳδοὺς τοὺς αὐτὴν ᾄδοντας ‘but the majority of the ancients refer to the totality of Homeric poetry as rhapsōidía and to those singing it as rhapsōidoí ’.

There is, to be sure, an ongoing debate about the origins of the eventual division of the Iliad and the Odyssey into twenty-four books each. Some would argue that these book-divisions are derived from {87|88} earlier patterns of performance-segmentation, [70] while others think that they are merely editorial superimpositions deriving from the era of Alexandrian scholarship. [71] What is needed to supplement this debate, I submit, is a diachronic perspective. What may be a performance break in one stage of the performance tradition may not be at another. [72] In other words, I hold open the possibility that the eventual division of the Iliad and Odyssey each into twenty-four Books results from the cumulative formation of episodes in the process of equalized or even weighting. It is from a diachronic point of view that I emphasize the cumulative formation of episodes in the process of even weighting. The point remains, in any case, that the concepts of rhapsōidós and rhapsōidía are compatible with myths about Homeric origins.

In fact, the concept of rhapsōidós can be applied by myth to Homer himself as prototypical poet, as also to his counterpart, Hesiod. For example, the scholia for Pindar Nemean 2.1 (the source is Philochorus FGH 328 F 212) quote the following verses attributed to Hesiod, who speaks of performing, in competition with Homer, hymns to Apollo: So Homer and Hesiod are models of rhapsodes by way of performing like rhapsodes. [73] Even for Plato (Republic 600d), Homer and Hesiod can be visualized as performing like rhapsodes (rhapsōidéō). For Plato, a figure like Phemios, represented as a prototypical poet in the Odyssey, is likewise a rhapsōidós (Ion 533c).

The mythical view of the poet as a rhapsode implies not only that he is a performer. The metaphor of sewing, as conveyed by the word rhapsōidós, refers also to the poet’s powers as a composer. Moreover, this metaphor of sewing is closely related to another metaphor, that of woodworking, which refers to the process of poetic composition in a strikingly analogous way.

The key word in this metaphorical world of woodworking is the name of Homer himself, Hómēros. In order to understand the traditional force of the metaphor at work, let us begin by reconsidering the traditional status of Homer as a prototypical author. [74] The further back in time we reconstruct this figure, the greater the repertoire attributed to him: in the pre-classical period, it seems that he is credited with all the so-called Cycle, all the Theban epics, and so on. [75] As we have already noted, the very concept of “Cycle”—that is, kúklos—had once served as a metaphor for all of Homer’s poetry. [76] But now we discover that this same word kúklos, used as a metaphor for the sum total of Homeric poetry, is attested with the meaning of ‘chariot-wheel’ in Homeric diction (Iliad XXIII 340, plural kúkla at V 722). This meaning will help explain the name of Homer himself.

In the poetic traditions of Indo-European languages, we find a direct attestation of a metaphor that compares a well-composed song to a well-crafted chariot-wheel: in the oldest Indic poetic traditions, we see the verb takṣ- ‘join, fit together’, regularly used to designate {89|90} the handiwork of a carpenter, combined in one passage (Rig-Veda 1.130.6) with the direct object vāc- ‘poetic voice’ (cognate of Latin vōx); in the same passage, this combination is then made explicitly parallel to that of takṣ- plus the direct object rátha- ‘wheel’ (in the metonymic sense of ‘chariot’; cf. the Latin cognate rota ‘wheel’). [77] The Indic root takṣ- ‘join, fit together’, designating the craft of a carpenter, is cognate with the root of Greek téktōn, meaning ‘carpenter’, which is applied in Pindar Pythian 3.112–114 as a metaphor for the poet as master carpenter or “joiner” of words (épos plural; cf. the cognate vácas-, direct object of takṣ- in Rig-Veda 6.32.1). [78] In the Greek poetic traditions, the specific image of crafting a chariot wheel is implicit: the root ar- of ararískō ‘join, fit together’ (the verb refers to the activity of the carpenter in the expression ἤραρε τέκτων ‘the joiner [téktōn] joined together [ar-]’ at Iliad IV 110, XXIII 712) is shared by the word that means ‘chariot-wheel’ in the Linear B texts, harmo (Knossos tablets Sg 1811, So 0437, etc.); in another dialectal form, hárma (ἅρμα) becomes, metonymically, the word for ‘chariot’ (Iliad V 231, etc.). I submit that this same root ar- is shared by the name of Homer, Hómēros, the etymology of which can be explained as ‘he who joins together’ (homo- plus ar-). [79] If this etymology is correct, then the making of the Cycle, the sum total of epic, by the master poet Homer is a metaphor that pictures the crafting of the ultimate chariot-wheel by the ultimate carpenter or “joiner.”

To be sure, the parallelism between aoidós ‘singer’ and téktōn ‘carpenter, joiner’ exists not only on the level of metaphor. Both professions belong to the category of dēmiourgós or ‘itinerant artisan’, as we see from Odyssey xvii 381-385. [80] Moreover, the carpenter is not the only craftsman who is comparable to the poet in the poetic traditions of Indo-European languages. To pursue this point, let us consider the root *tek(s)- in the Greek noun téktōn ‘carpenter, joiner’, also attested in tékhnē ‘craft, art’. This root, which we have already seen in Indic takṣ- ‘join, fit together’, does not survive as a verb in Greek, but we {90|91} find it in Latin, where texō is attested with the meaning ‘join, carpenter’ (as in Virgil Aeneid 11.326). [81] Ordinarily, however, Latin texō means ‘weave’ (as in Ovid Metaphorphoses 6.62). The parallelism in craftsmanship between carpenter and weaver, implicit in the semantics of the Latin verb texō, is even more pervasive: in Indo-European languages, the metaphor of carpentry as songmaking is actually paralleled by the metaphor of weaving. [82]

Let us return to the image of Hómēros as a primordial rhapsōidós. We now see that the parallelism between carpenter and weaver as metaphors for the poet corresponds to the association of Hómēros and rhapsōidós.

The key to this parallelism, I suggest, is the idea of a specialist. In the case of woodwork, we may say that only a master carpenter will have the skills required to put together, say, a chariot-wheel. Let us hereafter consistently refer to such a specialist as a “joiner,” someone who joins the pieces that other woodworkers have already fashioned. In the case of textiles, on the other hand, we have already seen that the word rhapsōidós implies a specialist. It implies an ability to sew together, into an artistic whole, pieces that other textile-workers have already woven. In other words, I propose a proportionality of metaphors: the carpenter of song is to the joiner of song as the one who weaves the song is to the one who sews together or stitches the song, that is, to the rhapsōidós. [83] Just as a joiner is a master craftsman, capable of special feats like the making of a chariot-wheel out of pieces of woodwork already made by himself or by other carpenters, so also the stitcher, one who sews together pieces of fabric already woven, is a master craftsman in his own right, fashioning something altogether “new” that is tailor-made to suit a given form. Thus the metaphor of a joiner or a stitcher, as distinct from a carpenter or a weaver, conveys the idea of a master singer. [84] {91|92}

Whichever way myth figures the creation of Homeric poetry, whether it be a joiner’s chariot wheel or a stitcher’s perfect fit, the actual creation is viewed as happening at a remote point in time, not over time. From the standpoint of the myth, it is as if there had been a “big bang” that produced a fixed pattern of composition, which led to a fixed pattern of performance or both. [85]

Moreover, as I already argued in Chapter 1, Homer is not just the creator of heroic song: he is also the culture hero of this song. To repeat the essence of what I said earlier: ancient Greek institutions tend to be traditionally retrojected, by the Greeks themselves, each to a proto-creator, a culture hero who gets credited with the sum total of a given cultural institution, and it was a common practice to attribute any major achievement of society, even if this achievement may have been realized only through a lengthy period of social evolution, to the episodic and personal accomplishment of a culture hero who is pictured as having made his monumental contribution in an earlier era of the given society. [86] So also with Homer: he is retrojected as the original genius of heroic song, the proto-poet whose poetry is reproduced by an continuous succession of performers. Conversely, each successive performer of Homer is one step further removed from this original genius: in Plato’s Ion, for example, Socrates envisages the rhapsode Ion as the last in a chain of magnetized metal rings connected by the force of the original poet Homer (533d–536d). In Plato’s mythical image of Homer and his successors, the magnetic force of the poetic composition weakens with each successive performer. Pictured as the last or at least the latest replicant of Homer, Ion becomes the weakest of all replicants. [87]

From the standpoint of an evolutionary model for the fixation of Homeric poetry, by contrast, the reality is altogether different from the myth: “even if the size of either the Iliad or the Odyssey ultimately defied performance by any one person at any one sitting, the monumental proportions of these compositions could evolve in a social {92|93} context where the sequence of performance, and thereby the sequence of narrative, could be regulated, as in the case of the Panathenaia.” [88] In quoting this formulation, I have highlighted the idea that an evolving fixity in patterns of performance leads to a correspondingly evolving fixity in patterns of composition, given that performance and composition—or, better, recomposition—are aspects of the same process in this medium.

By now we have seen a variety of myths offering “big bang” explanations for the creation of Homeric poetry, and we have noted in each case a variety of metaphors that articulate these myths. [89] First there is the myth of Homer’s making the Homeric poems himself. A metaphor associated with this myth is that of a master craftsman who produces a masterpiece of craftsmanship. Then there are the myths of a post-Homeric remaking of the poems. Among the metaphors used in these myths is that of an integral fabric produced by the “sewing together” of different parts of fabrics, corresponding to a total song that rhapsodes put together by singing parts of songs in sequence. But perhaps the most salient metaphor of them all comes from later stories about a prototypical written text, disintegrated into separate parts that are then all at once reintegrated at the initiative of a culture hero. As I have argued, this metaphor is the germ of the concept that we now know as the “Peisistratean recension.” I see no need, in short, to defend the concept of a “Peisistratean recension”—as a historical event.

The concept of a “Peisistratean recension” has been generally attacked or ignored by unitarians, supported by analysts. [90] Since such an event - and the concept of a recension surely requires that it be viewed as an event - supposedly took place sometime after the middle of the sixth century, it is almost two centuries removed from the era assigned to Homer by many experts. The concept is therefore not congenial to those who see a need to recover the presence of a “text” composed by a Homer who lived in the eighth century - I refer to them for the moment under the more general heading of “unitarians” - since there is no way of bridging the gap between this “Homer” and a written text that supposedly first came into being {93|94} only some two hundred years later. By contrast, “analysts” who do not care about singular authorship can afford to be less concerned about the prospect of moving the date of Homeric composition forward by two centuries. After all, we would expect them to view this composition as a matter of patchwork—in a negative sense of the word. In his Homerische Untersuchungen, for example, Wilamowitz describes the Odyssey as the end product of “a not very gifted patchwork-poet” (“ein gering begabter Flickpoet”). [91]

Particularly influential in questioning the concept of a “Peisistratean recension” has been an article by J. A. Davison, [92] whose negative views are restated in a widely-read chapter dealing with the transmission of the Homeric text in A Companion to Homer. [93] Although Davison explicitly rejects the concept of a “Peisistratean recension,” [94] he speaks of a “Panathenaic text,” [95] with reference to the evidence indicating that the “text” of the Iliad and Odyssey was regularly performed, as we have already seen, at the Athenian festival of the Panathenaia (Lycurgus Against Leokrates 102; “Plato” Hipparchus 228b; Diogenes Laertius 1.57). [96] From his point of view, such a “text” is a script, as it were, for the seasonally recurring performance of the Iliad and Odyssey at the Panathenaia. Presumably, such a “Panathenaic text” eventually became available for private ownership by way of the book-trade at Athens, which we see already flourishing at the end of the fifth century BCE. [97] Davison goes on to offer this warning: “any attempt to speak of a single ‘pre-Alexandrian vulgate’, and still more to create out of it a version of the Panathenaic text by arguing back to sixth- or {94|95} fifth-century Athens from the conditions which existed in Egypt after the establishment of the Alexandrian library, is doomed to failure from the beginning.” [98] Even in this context, however, we note that he speaks of “the Panathenaic text” as if it were a given.

While I agree with Davison to the extent that any attempted reconstruction of such a Panathenaic text presents major difficulties, I disagree with his argument that the story suggesting a “Peisistratean recension,” as reflected by the testimony of “Plato” Hipparchus 228b and Cicero De oratore 3.137, was invented by a scholar from Pergamon, perhaps Asclepiades of Myrlea (around 100 BCE), in order to undermine rival scholars from Alexandria. [99] As Davison puts it, such a scholar would have intended to discredit “the ‘authentic’ text which his Alexandrian rivals were so successfully imposing on the reading public.” [100] In disagreeing here with Davison, I follow in part the reasoning of Raphael Sealey, who argues that we have no grounds for assuming the successful production of an Alexandrian standardized text of the Homeric poems. [101] Referring to the earlier work of T. W. Allen, [102] which antedates that of Davison, Sealey stresses that the Alexandrian editorial adjustments made on the Homeric poems, as we can ascertain especially from the scholia for the Iliad, have had “singularly little effect” on the Homeric text as preserved in the medieval “vulgate” manuscript tradition. [103] Sealey goes even further, paraphrasing Allen: “either ... there was no Alexandrian edition or, if Alexandrian scholars did publish editions of Homer, ... these did not become popular with the reading public.” [104]

Sealey also objects to Davison’s formulation of the motive behind the Pergamene scholar’s purported “invention” of the story about a {95|96} Peisistratean recension. [105] Davison had put it this way: “without challenging the Athenian origin of the new text,” the “invented” story supposedly strips away the authority of this “new text” edited by the Alexandrian scholars because it lowers the date of the Homeric text’s creation from about 1050 BCE, as conjectured by Aristarchus of Alexandria (Proclus F a 58-62 ed. Severyns) to about 550, the era of the Peisistratidai. [106] Disputing the notion that an Alexandrian edition of Homer could assert such authority over text-production, Sealey concludes: “No one ever successfully imposed an Alexandrian text on the reading public.” [107]

There are other experts who stop short of such a conclusion. Let us begin with what may well be, at the time that I write this, the most widely-read account of the Homeric textual tradition, a chapter by Stephanie West entitled “The Transmission of the Text,” to be found in the opening pages of a new commentary on the Odyssey. [108] While conceding that the editorial work of the earlier Alexandrian scholars Zenodotus and Aristophanes “had little if any effect on the book trade,” West draws a line at the next step in the succession of Alexandrian scholars, the era of Aristarchus, whose scholarly activity is dated around the middle of the second century BCE: Even if the papyri dated after the era of Aristarchus “offer a text which differs little from that of the medieval manuscripts,” we need not necessarily connect this fact with the activity of Aristarchus. No one, in my opinion, has yet been able to refute successfully the observation of T. W. Allen that Aristarchus’ editorial prescriptions exerted practically no effect on the Homeric text as preserved in the medieval “vulgate” manuscript tradition. [110] What West has called the eventual “standardization” of the Homeric text after around 150 BCE can be explained in other ways, without recourse to the editorial authority of Aristarchus. One factor, it seems, is the nature of the book-trade during the period in question.

West herself stresses the role of the book-trade and “a general rise in standards of book-production at this period.” [111] Conceding that the editorial judgments of Aristarchus were ignored by the booksellers and proprietors of scriptoria, West nevertheless makes an exception in the case of Homeric lines deemed non-Homeric by Aristarchus, arguing that such lines were leveled out in the process of commercial copying, thanks to the authority of an Aristarchean text that featured special notations for supposedly non-Homeric lines: In the end, then, West’s model does not differ all that much from Allen’s, which rejects altogether the idea of a standard Alexandrian edition. In West’s own words, “the Alexandrian scholars did not impose a single specialist’s version on the tradition, but effected a {97|98} general purge of extraneous material and an increase in knowledge which afforded some permanent protection.” [113]

Still, I do not even see any compelling reason to infer, as does West, that such a “purge” depended on the authoritativeness of a given edition promulgated by Aristarchus. The very technology of the scriptorium, I submit, could easily promote the kind of leveling process where additional lines found only in some manuscripts but not in others tended to be omitted. The editorial minimalism espoused by Aristarchus, whose practice was to question the authenticity of lines that were missing in those manuscripts that he specially valued, could be matched by a pragmatic minimalism in the scriptorium. As West concedes, even the papyri dated after 150 BCE “offer too wide a range of variants to allow the hypothesis that they might all be copies of a single edition.” [114]

It seems to me, then, that the new degree of textual “standardization” in the era after 150 BCE reflects not the authority of Alexandrian scholarship but other factors—including the advances being made in the kind of minimalist quasi-editing techniques that would be needed for large-scale commercial copying of manuscripts. [115] In this connection, we may note Sealey’s observation that “one could achieve multiple production on a small scale by setting one slave to read a text aloud while many slaves sat around him and wrote down what they heard.” [116] A successful publisher in the Roman era, T. Pomponius Atticus, is said to have employed men described as anagnostae optimi et plurimi librarii ‘the best readers and the greatest number of scribes’ (Nepos Life of Atticus 13.3.). [117] This mode of manuscript production may be appreciably different from that of earlier times, if we accept the following description of manuscript production in the era before 150 BCE or so: {98|99}

The point remains that, even for West, the textual “standardization” of the Homeric poems after 150 BCE is due to developments in the book-trade, and the Homeric text of this era and thereafter can hardly be described, even in terms of her argument, as an Aristarchean text, let alone an Alexandrian one. Conversely, we may infer that the greater degree of variation in the papyri attested before 150 BCE is due not to the vagaries of a more old-fashioned sort of book-trade but to the absence of even the limited kind of textual standardization that we see taking place after that date.

West uses the term “second standardization” in referring to the era of Homeric textual history after 150 BCE. [119] For her the first standardization takes place in the era of Peisistratos, as she speaks of “this sixth-century standardization of the text.” [120] She also speaks of “this sixth-century recension,” which “must be regarded as the archetype of all our Homeric manuscripts and of the indirect tradition represented by ancient quotations and allusions.” [121] West thus basically accepts the concept of a “Peisistratean recension,” citing for support the arguments advanced in favor of this concept by Reinhold Merkelbach and Minna Skafte Jensen. [122] She is in effect also accepting the concept of a “Panathenaic text.” {99|100}

Jensen’s model of a “Peisistratean recension” requires an outright dictation that was supposedly commissioned in the era of the Peisistratidai, [123] which it turn leads to a standard “Panathenaic text.” [124] Her model differs from the one developed here mainly in the fact that she thinks of the hypothetical dictated text not as a transcript but as a new archetype. I can agree with the general notion of a “Panathenaic text” as the main source of the Homeric papyri in the Alexandrian era and of the later Homeric “vulgate” tradition in general. [125] But I do not go so far as to posit a single archetypal written text, preferring instead an evolutionary model that allows for the eventual textualization of the Homeric poems in the process of seasonally recurring performance at the Panathenaia. [126] As I have already argued, this textualization can could have taken place without the intervention of writing, but it could indeed yield a transcript, or a variety of transcripts, at various possible stages of the performance tradition of Homer at the Panathenaia, starting from around 550 BCE and proceeding towards the middle or even the end of the fifth century. [127] {100|101}

We have seen that Davison too assumes a “Panathenaic text,” though he does not go so far as to accept the concept of a “Peisistratean recension.” [128] As for Sealey, his disbelief in a standard Alexandrian edition of Homer is matched by a disbelief in a standard Panathenaic edition: he goes only so far as to say that the Panathenaic version of the Iliad and Odyssey could have been written down any time between 550 BCE and 450 BCE. [129] He connects the possibility of a more precise dating for any writing down of the text with the need to come up with a more precise dating for the rise of the book-trade in Athens. [130] By implication, then, there is for him no standard Panathenaic archetype on which the manuscripts of the incipient book-trade are based. In terms of Sealey’s model, I infer that any writing down of the Iliad and Odyssey in this period between 550 BCE and 450 BCE would amount to a mere transcript, not some standard of reference for future performances. [131]

I find that my position is closest to that of Sealey, to the extent that I too see no proof for the existence of an archetypal Panathenaic manuscript of Homer, any more than there seems to be any proof for an archetypal Alexandrian manuscript. There is, however, room for positing an archetypal Panathenaic form for performing the Iliad and Odyssey, as embodied in a Greek development that we have already compared with similar developments attested in living oral traditions. As we have seen, that development is the Panathenaic rule, attributed either to the Peisistratidai or to Solon himself (“Plato” Hipparchus 228b and Diogenes Laertius 1.57 respectively). [132] It is instructive to consider the following formulation by Sealey: I disagree with Sealey’s formulation to the extent that the arrangement of the narration is viewed here as a historical event, corresponding to an event in the story that told about the Peisistratidai and how they produced a standard text of the Homeric poems. I propose instead an evolutionary model for both “events,” that is, for both the arrangement of narration and the textualization of the poems.

I must stress again that my goal is not to revive the case for positing a “Peisistratean recension,” where recension is obviously to be understood in the conventional sense of a critical revision that takes into account the basic available sources of a text. Rather, I have approached the problem in a different way by pointing out that the details of reports leading to the very idea of a “Peisistratean recension” happen to match the details of myths explaining the composition, performance, and diffusion of epic. The emphasis of these myths on the ultimate unity or integrity of any given epic, as we see most dramatically illustrated in the classical Persian example, corresponds to the reality of a unified and integrated text, such as the Homeric Iliad and Odyssey. It also corresponds to the narratives, already analyzed above, concerning a customary law in effect at the Athenian festival of the Panathenaia, where it was ordained that the performance of the Iliad and Odyssey by rhapsōidoí ‘rhapsodes’ had to follow the sequence of composition, and that the entire composition had to be performed by one rhapsode after another, likewise in their own sequence. Our two clear references to this customary law, “Plato” Hipparchus 228b and Diogenes Laertius 1.57, disagree about the identity of the initiator of this practice, the first source indicating the Peisistratidai and the second, Solon the lawgiver. For our purposes, the question of determining the originator of this custom is irrelevant to the more basic question of the significance of the custom itself. [134] The narratives about this customary law, I submit, serve as a clear indication that unity or integrity of composition was itself a tradition, and was venerated as such. {102|103}

If, then, the Peisistratean recension is a myth, whose myth is it? The answer is, surely, that the Peisistratidai owned it, or, better, appropriated it as an instrument of propaganda for their dynasty. We may note that Cicero’s account, which is most explicit about the recension, portrays the tyrant Peisistratos as one of the canonical Seven Sages in the context of crediting him with the arrangement of the Homeric poems. Other narratives, as we have also seen, draw an explicit parallel between Peisistratos and a venerable lawmaker like Lycurgus of Sparta. [135] In short, the historicity of the Peisistratean recension is to be found not in the actual story of the recension but in the appropriation of the story, the myth, as a source of propaganda for the Peisistratidai.

Reinhold Merkelbach argues that the Peisistratean recension was a genuine historical event, though he treats the story itself as an extrapolated invention, probably to be dated to the fourth century. [136] The central point for him is that there was a Homeric text, in manuscript form, which took shape in the sixth century, the era of Peisistratos. He also thinks that the Lycurgus story is based on the Peisistratos story. The comparative evidence that I have already adduced suggests otherwise, that it was in fact the Peisistratos story that was based on an appropriation of earlier narrative patterns concerning sages and lawmakers. I also disagree with Merkelbach when he argues that the Homeric poems would have disintegrated through the repeated performances of “improvising” rhapsodes had it not been for the primacy of a written manuscript. [137] In positing a disintegration through re-performance, Merkelbach is appealing to the concept generally known as zersingen, which has been successfully challenged by folklorists. [138]

Merkelbach also argues that the relative stability of the Homeric textual transmission, in comparison with the far more pronounced manuscript variations of other epics like the Song of Roland, proves {103|104} the archetypal existence of a written Homeric text. [139] But I have already argued, by re-applying the comparative perspectives applied by Merkelbach, that the alternative model of a relatively static phase in the evolution of the Homeric poems can account for such textual stability. In the meantime, Merkelbach seems to leave out of consideration the fact that the Homeric poems were performed. I repeat what I said earlier: the fact that Homeric poetry was meant to be performed live, and that it continued to be performed live through the Classical period and beyond, remains the primary historical given. [140] A Panathenaic written text cannot be the primary medium of the Homeric poems. I find even less plausible Merkelbach’s supplementary thesis of a “reading public” for these poems. [141] In this regard, I also find it difficult to reconcile Stephanie West’s acceptance of Merkelbach’s overall thesis [142] with her own special emphasis on the performance traditions of Homeric rhapsodes. [143]

These objections are not meant to slight the importance of Merkelbach’s contributions, especially when it comes to the actual dating of the narrative traditions about the so-called Peisistratean recension. [144] According to Merkelbach, our sources go at least as far back as the fourth-century BCE, the era of Dieuchidas of Megara, who claimed that it was Solon, not Peisistratos, who “interpolated” into the Iliad verses favorable to the Athenians (FGH 485 F 6, by way of Diogenes Laertius 1.57): if Dieuchidas has to go out of his way to claim that the “interpolator” was Solon, not Peisistratos, then there must have been a pre-existing version featuring Peisistratos. [145]

Clearly the act of “interpolation” is viewed as a fraud, as we see from a story about Onomakritos, who is caught red-handed in the act of inserting his own verses into a body of oracular poetry (Herodotus {104|105} 7.6.3). [146] This is the same Onomakritos whom we have already seen described as a diathétēs ‘arranger’ of the oracular poetry possessed by the Peisistratidai (7.6.3), [147] and whom we see described elsewhere as actually performing oracular poems on behalf of the Peisistratidai (Herodotus 7.6.5) [148] This is also the same Onomakritos who is reputedly one of the four “arrangers” of the Homeric poems (Tzetzes in Anecdota Graeca 1.6 ed. Cramer). [149] I have already suggested that, once the Peisistratidai were ousted, the positive stories about their “recension” of the Homeric text could be reshaped by way of transferring the credit for achieving an Athenian version of Homer from the Peisistratidai to Solon the Lawgiver. [150] Meanwhile, the motive of Dieuchidas of Megara is to undermine any standard Athenian version of Homer, since the Athenians had a long history of using citations from Homer - especially from the Catalogue of Ships in Iliad II, in their ongoing territorial claims against Megara. [151] It therefore suits the purposes of Dieuchidas to undermine Solon, who is viewed positively by the Athenians, instead of Peisistratos, who is now viewed negatively by them. What is a matter of “recension” in a positive version of the myth can become a matter of “interpolation” in a later negativized version—as we have just seen in the case of the Peisistratidai and their agent Onomakritos. What Dieuchidas is trying to accomplish, I suggest, is to extend such a negativized version of the myth from Peisistratos to Solon, who would be at that given moment the current culture hero of the positive version.

Let us return to the basics of what we have explored so far on the subject of Homeric textualization. We have concentrated on a relatively static phase of Homeric performance traditions extending roughly from the middle of the eighth century BCE all the way to {105|106} the middle of the sixth, at which point I posit the reaching of a near-textual status for the Iliad and Odyssey in the specific historical context of the Feast of the Panathenaia at Athens, as reorganized under the régime of the Peisistratidai.

This relatively static phase in my evolutionary model for Homeric poetry, lasting almost two centuries and culminating in a near-textual status for the Iliad and Odyssey at Athens under the Peisistratidai, can be correlated with a relatively static phase in the iconographic representations of “Iliadic” and “Odyssean” themes in the archaic period, and the convergences linking epic and iconographic treatments of epic themes become increasingly pronounced as we approach the middle of the sixth century BCE.

Let us survey a few examples from the earliest attested iconographical evidence. In the case of “Odyssean” themes, we may note in particular the story of Odysseus and the Cyclops, well attested in the seventh century. [152] For “Iliadic themes,” we may turn to the dossier assembled by Friis Johansen. [153] I draw attention to a bronze relief from Olympia which he dates to the second half of the seventh century BCE and which represents an Embassy to Achilles that is comprised of Phoenix, Odysseus, and Ajax (corresponding to the narrative that we find in Iliad IX). [154] We may note also a plate from Rhodes, dated to the last quarter of the seventh century and featuring a representation of the Death of Euphorbos (corresponding to the narrative in Iliad XVII 1–113). [155] Pointing out analogies between the style of this plate from Rhodes and the style of painting attested for archaic Argos, Friis Johansen observes other connections between Argos and Rhodes, including the Rhodian claim of descent from Argos (cf. Pindar Olympian 7.19). [156] In this context, I merely record the possibility of connecting a relatively early proliferation of Iliadic narrative traditions in Rhodes with the extraordinary highlighting of Rhodes in Iliad II 653–670. {106|107}

The evidence of such examples adduced by Friis Johansen makes clear that we are dealing with iconographical references to Iliadic narrative traditions, not to the Iliadic text as we know it. Still, I suggest that a relatively static phase in the development of Iliadic narrative traditions is what makes it possible for us to recognize as distinctly Iliadic whatever correspondences we find in iconographical evidence that is contemporaneous with this posited phase.

If indeed we are dealing with a lengthy static phase of Iliadic narrative traditions, not with the Iliadic text as we know it, we may still expect considerable degrees of variation. If we take as an example the François Vase, dated around 570 BCE and of Attic provenience, we see there a representation of the Funeral Games for Patroklos, converging with the narrative of Iliad XXIII in the following details: (1) five chariot-teams, (2) Achilles as president of the games, (3) the participation of Diomedes in the chariot race. [157] There are also the following narrative details in the vase painting that diverge from details in Iliad XXIII: (1) each chariot is drawn by a team of four horses, not two as in the Iliad; (2) besides Diomedes, the participants are Odysseus, Automedon, Damasippos, and Hippothoön rather than Eumelos, Menelaos, Antilokhos, and Meriones, as in the Iliad; Diomedes comes in third, while he is the winner in the Iliad. [158] Still, the relative stability of narrative traditions in archaic Greek iconography is illustrated by the similarity between the painting on the François Vase and another painting, dated almost a hundred years earlier, on a proto-Corinthian aryballos: In this context, we may note in general Friis Johansen’s own frequent observations of variations between the corresponding narrative details in the attested artifacts and in the attested epic of the Iliad. Still, if we choose to emphasize the continuity that is manifested in the phenomenon of these variations, then Friis Johansen’s terminus post quem of 630 BCE or so for the inception of distinctively “Iliadic” themes in iconographical representations need not be deemed too early. [160] As we turn to later developments, we see that significant variations persist until the middle of the sixth century BCE or even as late as 530 BCE, which can serve as a terminus post quem for the textualization or quasi-textualization of the Iliad and Odyssey. [161] It may also serve as a terminus post quem for reforms of the Homeric performance traditions during the régime of the Peisistratidai.

It is in this context that I am ready to ask the question for the last time: when was it that the Iliad and Odyssey were recorded as written texts? On the basis of linguistic criteria, Richard Janko has proposed 750–725 BCE and 743–713 BCE as definitive dates for the text-fixation of the Iliad and Odyssey respectively. [162] On the basis of historical and archaeological considerations, Ian Morris agrees, to the extent that the contents of the Homeric poems may reflect a social context datable to the eighth century before our era. [163] Both these assessments require the “dictation theory” for establishing such {108|109} an early date. [164] As I have already pointed out, however, the evidence of the earliest poetic inscriptions suggests that the very concept of a poetic transcript is not likely to have evolved until around 550 BCE. [165] Thus I continue to resist the arguments for an early dating of Homeric poetry as a text.

Given the strong parallelisms between textuality and certain patterns of evolution in oral poetic traditions, I have been arguing that the fixation of Homeric poetry as a text can be viewed as a process, not necessarily an event. Text-fixation becomes an event only when the text finally gets written down. But there can be textuality—or better, textualization—without written text. [166] I have been arguing further that the Homeric tradition of epic provides an example of such textualization: in the process of evolution in composition, performance, and diffusion, the Homeric tradition of epic became increasingly less fluid and more stable in its patterns of recomposition, moving slowly ahead in time until it reached a relatively static phase. [167] We may refer to this static phase as an era of rhapsōidoí ‘rhapsodes’. [168] {109|110}

As we have seen, the static phase of Homeric performance traditions could easily have lasted about two centuries. If we were to make cross-sections at either end of this static phase, I would picture at one end a relatively more formative stage starting with the middle of the eighth century and, at the other end, an increasingly definitive stage towards the middle of the sixth century, by which time I can imagine the achievement of a near-textual status of the Homeric poems in the context of performance by rhapsodes at the Panathenaia. [169]

My evolutionary model differs from that of G. S. Kirk, who posits a sequence of oral transmission starting with a monumental composer in the eighth century BCE, to be defined as an individual Homer, and proceeding from there into the historical period of sixth-century Athens. [170] Either model takes us down to 550 BCE or so. [171] Significant variations of Homeric themes in the iconographical evidence of vase paintings—especially variations of “Iliadic” themes—persist until around 530 BCE. [172] In sum, the approximate date of 550 BCE— {110|111} or perhaps a few decades later—seems to me the most plausible one as a terminus post quem for a potential transcription of the Iliad and Odyssey.

This evolutionary model of Homeric poetry culminating in a static phase that lasts about two centuries, framed by a relatively formative stage in the middle of the eighth century and an increasingly definitive stage in the middle of the sixth, is comparable to the model that I have worked out for the body of poetry attributed to Theognis of Megara, where the external dating criteria applied to the contents suggest a span of evolution exceeding a century and a half. [173] There are also other points of comparison. With reference to Hesiod, Archilochus, and Tyrtaeus, Stephanie West has dated the compositions attributed to all three before the last third of the seventh century, adding that “their precisely worded compositions could not long have survived their authors without a written record.” [174] I prefer to apply an evolutionary model to all three, noting the rhapsodic traditions explicitly attested in the case of Hesiod [175] and Archilochus; [176] similar arguments can be made in the case of Tyrtaeus. [177]

The time has come to reach conclusions. The comparative evidence of living oral epic traditions goes a long way to show that unity or integrity results from the dynamic interaction of composition, performance, and diffusion in the making of epic. Such evidence, added to the internal evidence of the Iliad and Odyssey as texts, points to an evolutionary process in the making of Homeric poetry.

And yet, this envisioning of Homer in evolutionary terms may leave some of us with a sense of aching emptiness. It is as if we had suddenly lost a cherished author whom we could always admire for the ultimate achievement of the Iliad and Odyssey. But surely what we have really admired all along is not the author, about whom we never did really know anything historically, but the Homeric poems themselves. To this extent, the evolutionary model may even become {111|112} a source of consolation: we may have lost a historical author whom we never knew anyway, but we have recovered in the process a mythical author who is more than just an author: he is Hómēros, culture hero of Hellenism, a most cherished teacher of all Hellenes, who will come back to life with every new performance of his Iliad and Odyssey. {112|113}