How many ways are there to tell the story of Troy? A passage from Iliad 20 makes me wonder just how flexible the Homeric tradition might be. At the beginning of book 20, Zeus calls the gods to an assembly. He tells them that they may now join the battle taking place before the walls of Troy on whichever side they wish, something that he had expressly forbidden them to do at the beginning of book 8. The reason he has changed his mind, he explains, is that Achilles is now preparing to return to battle for the first time since his withdrawal in book 1, and Zeus is afraid that the Trojans will not be able to withstand him for even a little while:

καὶ δέ τέ μιν καὶ πρόσθεν ὑποτρομέεσκον ὁρῶντες·
νῦν δ᾽ ὅτε δὴ καὶ θυμὸν ἑταίρου χώεται αἰνῶς
δείδω μὴ καὶ τεῖχος ὑπερ μόρον ἐξαλαπάξῃ.

Iliad 20.28–30 [1]
Even before now they would tremble before him when they saw him.
And now when he is terribly angry in his heart because of [the death of] his companion
I fear lest the wall [of Troy] will be sacked beyond [i.e., contrary to] fate.

Apollo, the god of prophesy and the one besides Zeus most often associated with seeing into the future, likewise fears that the Trojan walls will come down too soon at Achilles’ hands: μέμβλετο γάρ οἱ τεῖχος ἐϋδμήτοιο πόληος/μὴ Δαναοὶ πέρσειαν ὑπὲρ μόρον ἤματι κείνῳ (“For he was concerned about the wall of the well-built city, lest the Danaans destroy it on that day beyond fate,” Iliad 21.516–517). Zeus’ and Apollo’s fear in these passages is remarkable, and begs questions that anyone who has read the Iliad with undergraduates will be familiar with. If the walls of Troy are destined to fall at a particular moment, how could they fall before that? Is fate something that can be changed? Is Zeus subject to fate or can Zeus alter it?

The question of whether an individual’s fate can be changed creates a tension that runs throughout the Iliad. In Iliad 16 and 22 Zeus contemplates saving Sarpedon and then Hektor from death. Both times he is met with outrage (on the part of Hera and Athena respectively) and is rebuked via the formula ἔρδ’· ἀτὰρ οὔ τοι πάντες ἐπαινέομεν θεοὶ ἄλλοι at 16.443 and 22.181 (“Do it, but not all the rest of us gods will approve”; see also 4.29), and he does not follow through on the threat. Despite what the formula implies (i.e., that Zeus can change the fate of these mortal men), the fact that he never acts on this suggests either that he cannot or that the consequences would be so grave that he will not risk it.
In Iliad 2.155 we are told that the Argives would have returned home “contrary to fate” (ὑπέρμορα) if Hera had not intervened. At the beginning of Iliad 4, Zeus contemplates ending the war with Menelaus’ apparent victory over Paris (thus saving Troy), but he meets with strong resistance from Hera and agrees to accept a deal whereby he has the right to destroy Argos, Sparta, and Mycenae in the future. Athena then restarts the Trojan War by provoking Pandaros to shoot Menelaus with his bow. In Iliad 20.290ff. Poseidon saves Aeneas from death at the hands of Achilles, because it is his destiny (μόριμον 302) to survive. In none of these cases, however, is fate actually changed, and we are left to wonder what would have happened if the gods had not intervened. [2]
We could also reframe these questions in terms of narrative. If the Iliad tells a traditional story, shouldn’t Zeus and Apollo know how the story ends? [3] Would it really be possible to change the story now and have Troy fall while Achilles is still alive, and indeed at his hands? In fact all of book 20 seems preoccupied with the possibility that the story could unfold in the wrong way. Hera says that the gods have come down to fight in order to make sure that Achilles does not suffer anything before what is fated (20.125–128). As I have already noted, Poseidon saves Aeneas from his encounter with Achilles (20.288–339), lest he not go on to continue the family line of Dardanos, as fated (μόριμον), and Apollo (with no stated explanation) protects Hektor from dying at the hands of Achilles two books too soon (20.419–454).
In his 1979 book The Best of the Achaeans, Gregory Nagy argued that the first song of Demodokos in Odyssey 8, in which the Phaeacian bard narrates a quarrel between Odysseus and Achilles, is in fact a compressed reference to an epic tradition in which Achilles and Odysseus quarreled over whether Troy would be taken by cunning (mētis) or by force (biē). [4] Nagy reads the scholia preserved in the manuscripts at Odyssey 8.75 and 77 as likewise pointing to such a tradition, which is otherwise not attested in our surviving sources (Nagy 1979:46). Might we find here in the fears of Zeus and Apollo another glimpse of these two rival possibilities for the fall of Troy? If so, we have to wonder if the Iliadic tradition is indeed so multiform, so flexible, that such a radically different ending could be possible. Is there (or was there) an alternative epic universe, in which Achilles really did take Troy by force? And if not, why does Zeus entertain the idea?
As it turns out, ancient commentators on the Iliad were concerned about these same questions. And so the scholia in the margins of the so-called Townley manuscript (Burney 86, folio 220v) record for us a fascinating variation on these verses from book 20:

τινὲς γράφουσιν ἀντὶ τοῦ <δείδω, μὴ καὶ τεῖχος>
οὐ μέντοι μοῖρ’ ἐστὶν ἔτι ζῳοῦ Ἀχιλῆος
Ἰλίου ἐκπέρσαι εὖ ναιόμενον πτολίεθρον.
πέρσει δουράτεός [θ’] [5] ἵππος καὶ μῆτις Ἐπειοῦ.
πῶς γὰρ ὁ εἰδὼς “μοῖράν τ’ ἀμμορίην τε” [= Odyssey 20.76] νῦν διστάζει;
Some write instead of “I fear lest the wall”
It is not fated, however, with Achilles still alive
to sack the well-inhabited citadel of Ilium.
A wooden horse will destroy it and the craftiness [mētis] of Epeios.
For how is he [= Zeus], the one who knows “what is fated and not fated” [= Odyssey 20.76], now in doubt?

These alternative verses make clear that Troy is not going to fall at the hands of Achilles, but rather as a result of the mētis of the wooden horse. Problem solved. But the commentator, in seeking to solve a narratological, mythological, and indeed existential problem, now presents us with a textual one. What is the source of these verses that “some write,” and how do we reconcile them with our received text?

A similar interpretive crisis can be found in the scholia on Iliad 19.108. [6] Here Agamemnon tells the story of how Hera made Zeus swear an oath that the descendent of his born on that day would be king of all those around him. It is in this way, through Hera’s machinations, that Eurystheus becomes king of Argos instead of Herakles. A comment preserved in the Venetus A manuscript asks:

εἰ δ’ ἄγε νῦν μοι ὄμοσσον: διὰ τί ἡ Ἥρα ὀμόσαι πρὸς τὸν Δία; ἢ δῆλον ὡς οὐ ποιοῦντα, ἃ ἂν φῇ. εἰ δὲ τοῦτο διὰ τί οὐ κατανεῦσαι ἀλλὰ καὶ ὀμόσαι ἠξίωσεν ὡς καὶ ψευδομένου ἂν μὴ ὀμόσῃ; ὁ δὲ ποιητής φασιν ἀληθεύειν “ὅ τί κεν κεφαλῇ κατανεύσῃ” [~ Iliad 1.527]. τὸ μὲν οὖν ὅλον μυθῶδες· καὶ γὰρ οὐδ’ ἀφ’ ἑαυτοῦ ταῦτά φησιν Ὅμηρος οὐδὲ γινόμενα εἰσάγει, ἀλλ’ ὡς διαδεδομένων περὶ τὴν Ἡρακλέους γένεσιν μέμνηται … οὕτως Ἀριστοτέλης.
“Come now and swear to me”: Why does Hera make Zeus swear an oath? Certainly it is clear that he is not doing what she says. But if this is the case, why does she want him not only to nod in assent, but also to swear, as if he would be lying if he did not swear? But the poet says that he tells the truth with respect to “whatever he nods in assent to.” Therefore the story is mythological. For Homer doesn’t invent things nor does he introduce the things that happen, but he recalls the birth of Herakles as these things have been handed down.… This was the view of Aristotle.

This particular comment derives from the ancient tradition of scholarship known as the “Homeric Questions,” best known from the surviving writings of Porphyry but which, as we can see here, goes back at least as far as Aristotle. [7] The commentator wonders why Hera required Zeus to swear an oath, when elsewhere in the Iliad (notably in Iliad 1.527) Zeus asserts that whatever he assents to will come true. Underlying the question seems to be anxiety similar to that which we find in the Iliad 20 scholion. Is it possible that what Zeus agrees to won’t come true? Could the story of Herakles have turned out another way if Hera hadn’t made him swear the oath? But the commentator has a solution and goes on to excuse Homer here. He is working with a traditional story, a myth (τὸ μὲν οὖν ὅλον μυθῶδες): “Homer doesn’t invent things nor does he introduce the things that happen, but he recalls the birth of Herakles as these things have been handed down.”

I submit that this comment has deep implications not just for an understanding of the relationship between myth and poetry in the Homeric epics but also for our understanding of the textual tradition that transmitted these poems over two and a half millennia. In my earlier work (especially Dué 2002 and Dué and Ebbott 2010) I have argued that the multiformity of the mythological tradition from which epic singers wove their tales and the textual tradition of Homeric epic go hand in hand. The Iliad and the Odyssey are synoptic representatives of an entire system of traditional songs that developed over many hundreds of years. In its earliest phases, this system included the song traditions of the Epic Cycle and still further epic traditions to which the Iliad and the Odyssey sometimes allude, such as the voyage of the Argo, together with the mythological traditions on which those songs were based. [8] As we will see, these song traditions were multiform. They did not exist in a fixed form until very late in their evolution. But at the same time they were traditional, in that they told the story as it had been handed down.
As Milman Parry showed in his analysis of Homeric diction in the 1920s and 30s, any innovations that are present in the text as we now have it were introduced by means of a complex process over time and in the context of performance, and cannot be attributed to any one poet. [9] The same is true for the content expressed by that diction. Myth is by nature multiform, as were the epic song traditions that narrated such myths as the story of the Trojan War and the exploits of the hero Achilles, but it is simply not possible to call any one character or plot development the “invention” of a particular poet (Dué 2002:21–36). As the commentator remarks in connection with Zeus’ oath, “Homer” does not invent the myth but rather recalls what has been handed down, using the formulaic diction that has evolved over centuries to relate these stories.
So too the textual variations that we find attested in the multitude of ancient sources—quotations from ancient authors, fragments on papyrus, and the scholia in the margins of medieval manuscripts (which derive ultimately from the work of Alexandrian and Roman scholars)—are just as likely to have been generated by the system of oral traditional poetry in which the Iliad and the Odyssey were composed, and therefore can be considered every bit as “Homeric” as those that we find in our modern printed editions. Modern editions are primarily based on the texts found in a handful of medieval manuscripts, which postdate the oral tradition by more than a millennium. In this book I propose to take these far more ancient sources seriously and to treat the variations, or multiforms, that they preserve as traditional variations generated in performance whose poetic implications are worthy of exploration.
It is this nexus of multiformity and tradition that I want to explore. As we will see, the mythological and narratological questions being grappled with by the scholars of Alexandria and the later authors whose comments survive in the scholia of our medieval manuscripts can sometimes have profound implications for the textual transmission of the Iliad and the Odyssey. The scholion on Iliad 20.30 with which I began is a perfect example of the interdependence of the two. In this one comment not only can we possibly catch a glimpse of a now-lost epic tradition in which Achilles and Odysseus contend to be the sacker of Troy and the “best of the Achaeans,” but we also learn about three verses that do not survive in our medieval manuscripts of the poem. All we are told, in the typically compressed way of the scholia, is that “some” (presumably editors) write these verses (presumably in their editions). They are not a seamless replacement for Iliad 20.30, however. If we replaced 20.30 with the verses that “some write,” we would get this:

[20.28] καὶ δέ τέ μιν καὶ πρόσθεν ὑποτρομέεσκον ὁρῶντες·
[20.29] νῦν δ᾽ ὅτε δὴ καὶ θυμὸν ἑταίρου χώεται αἰνῶς
[20.30a] οὐ μέντοι μοῖρ’ ἐστὶν ἔτι ζῳοῦ Ἀχιλῆος
[20.30b] Ἰλίου ἐκπέρσαι εὖ ναιόμενον πτολίεθρον.
[20.30c] πέρσει δουράτεός <θ’> ἵππος καὶ μῆτις Ἐπειοῦ.

[20.28] Even before now they would tremble before him when they saw him.
[20.29] And now when he is terribly angry in his heart because of [the death of] his companion…
[20.30a] It is not fated, however, with Achilles still alive
[20.30b] to sack the well-inhabited citadel of Ilium.
[20.30c] A wooden horse will destroy it and the craftiness [mētis] of Epeios.

If we assume an ellipsis here (as sometimes occurs, as at Iliad 1.135–136), we can make it work, but it is more likely that the scholia are quoting from an edition in which the entire passage was substantially different from what we find in the medieval manuscripts of the Iliad. [10] But these verses are in no way objectionable beyond the fact that they do not survive elsewhere (Edwards 1991 ad loc.). [11] There is nothing “un-Homeric” about them—they are simply an attested multiform of the verses transmitted by our medieval manuscripts.

I hope to show that variations like the one discussed here are not aberrations but are in fact attested in many different ancient historical sources. In other words, I will highlight and call attention to the surviving examples of the multiformity of the epic tradition, rather than try to explain them away, dismiss them, or hide them in the dark recesses of an apparatus criticus, as is usually done. [12] Building on my previous work on this subject (especially Dué 2001a and 2002 and Dué and Ebbott 2010), I will explore these examples of multiformity as a necessary reflex of an oral tradition in which epic tales are composed in performance, and I will show how awareness of multiforms leads to a greater understanding of the poetics of the tradition in which the Iliad and the Odyssey were composed. By treating attested multiforms as manifestations of a traditional system, I am able to explore what kinds of multiformity were natural to this system in different historical eras and how ancient audiences may have understood them.
In so doing I will be drawing on nearly two decades as coeditor with Mary Ebbott of the Homer Multitext, a digital project that seeks to make the full complexity of the textual transmission of the Iliad accessible by means of high-resolution images of the historical documents that transmit the poem together with digital, diplomatic editions of their contents. I and my principal collaborators on this project (which include, in addition to Mary Ebbott, the project architects, Christopher Blackwell and Neel Smith, and the associate editors, Douglas Frame, Leonard Muellner, and Gregory Nagy) assert that a multitext edition is a methodologically superior way of representing the transmission of an oral tradition, in that it does not seek to establish a single original from which all other surviving texts derive but rather endeavors to make accessible as many historical instantiations of the text as possible together with their historical contexts. [13] In this belief we have been profoundly influenced by the fieldwork and scholarship of Albert Lord, who wrote that “the word multiform is more accurate than ‘variant,’ because it does not give preference or precedence to any one word or set of words to express an idea; instead it acknowledges that the idea may exist in several forms” (Lord 1995:23; see also Lord 1960:100).
Albert Lord’s 1960 book The Singer of Tales was the culmination of decades of close observation of a living oral epic tradition in Yugoslavia and careful application of that work by analogy to the Homeric epics. Lord went to Yugoslavia initially in the early 1930s, as the undergraduate assistant of Milman Parry. Parry’s 1928 doctoral dissertation on the traditional epithet in Homer had been a brilliant demonstration of the economy and traditionality of Homeric diction, but even Parry himself did not fully grasp the implications of this work until Parry and Lord went to Yugoslavia to observe the still-flourishing South Slavic oral epic song tradition (Parry 1971:439). It was in the context of this fieldwork that Parry came to understand that Homeric poetry was not only traditional but oral—that is, composed anew every time in performance, by means of a sophisticated system of traditional phraseology and diction. For Parry, witnessing the workings of a living oral epic song tradition was a paradigm shift. Through the analogy with the South Slavic tradition, the workings of the Homeric system of composition became clear to him.
In two expeditions to the former Yugoslavia, in 1933–1935, Parry and Lord together collected songs, stories, and conversations from singers of the South Slavic epic song tradition. Albert Lord took additional trips in the 1950s and 1960s, and in many cases was able to rerecord the singers he and Parry had captured decades earlier. No two of the songs collected were exactly alike, nor did any two of the singers have exactly the same repertoire. These singers composed extremely long epic poems in performance. In order to do this they drew on a vast storehouse of traditional themes and phrases that worked within the metrical rules that give the poetry its rhythm. That is to say, they created and used what we call formulas to build each verse as they went along, instead of employing static, individual words or words memorized in a fixed order. Just as formulas are the building blocks of a line in performance, themes are the larger components that make up songs. The poets observed by Parry and Lord moved from one theme to another as they sang; themes were connected in the oral poet’s mind and his plan for the song followed from their habitual association in the tradition. This performance method resulted in each song being a new composition, which is why no two songs that Parry and Lord recorded were exactly the same. Each song was a multiform of a notional song that never existed in a single “original” form.
Parry and Lord applied the findings of this fieldwork to the Homeric poems by analogy, and they were able to show that the workings of the South Slavic system reveal a great deal about how the Iliad and the Odyssey were composed. Parry planned a series of publications based on his observations and subsequent analysis of Homeric poetry, but it was never completed. His surviving writings have been incredibly influential, but he died at the age of 33, long before he had a chance to pursue the many implications of his fieldwork. It became the work of his young undergraduate assistant, Albert Lord, to brings these ideas to the world. Of course, scholars before Parry and Lord had proposed that the Iliad and the Odyssey were composed orally, but never before had the system by which such poetry could be composed been demonstrated, nor were the implications of the creative process truly explored. The Singer of Tales does just that.
Lord asserts in The Singer of Tales that in an oral tradition like that in which the Iliad and the Odyssey were composed, “singing, performing, composing are facets of the same act” and that the implications of this process are “both broad and deep” (1960:13). Indeed they are. The work of Parry and Lord and the scholars who have built on their efforts suggests that, rather than there being a single master version of each poem from which all others descend, in its earliest stages of development there was a great deal of fluidity in the Greek oral epic tradition. Countless variations on the story of the Trojan War and the episodes within it—the anger of Achilles, the returns of the heroes, and any number of traditional tales—are known to have been current in different times and different places in antiquity, and were likely composed in performance by countless poets whose names are now lost to us. At the same time, because Greek oral epic poetry was already traditional in content in ancient times, any audience on any given occasion of performance would already have known the story and the characters. There would have been nothing about the story, the language, the rhythm of the song, or the characters that was new for that audience. A poet in a traditional song culture like that of the ancient Greeks could compose poetry in performance using techniques, plots, characters, and an ever-evolving corpus of formulaic language that he had inherited from many previous generations of singers. The material and techniques were traditional, but each performance was a new composition—a recomposition, in and for performance.
This dynamic necessarily affects how ancient epic poetry is understood and appreciated. To make an even more modern analogy than Parry and Lord’s South Slavic one, for an audience that knows Anakin Skywalker will become Darth Vader, Star Wars is no less full of drama and tension as he proceeds down that path. Many a Star Wars fan has been content to watch that story replay itself again and again upon subsequent viewings. But for the Homeric epics there was an additional layer of tension. Certainly Achilles was always going to choose to return to battle and die at Troy, but the song that narrated that choice was always being composed anew. Not only might different singers arrange the song differently, and not only might there be variations on what episodes were included in any given performance, but there was always the lurking possibility, however unlikely, that the current performance would turn out differently.
Once we understand the Iliad and the Odyssey to have been composed this way, we can no longer attempt to stake rigid claims upon any one version of the text, as if the poem were fixed and unchanging. We have no reason to necessarily privilege one formulaic variation over another, even if one is well attested in our medieval manuscripts and one is known only from another source. Both are at least potentially authentically generated performance multiforms, and both have something to teach us about the compositional process and the poetics of the system in which they were generated. [14] Not all surviving multiforms would have been known to all singers at all times and in all places, yet each has the potential to reveal something about the poetics of the tradition in the time and place in which that multiform is attested.
But just how fluid, how multiform, was the Greek epic tradition? If the tradition was, as I have claimed, quite fluid in its early phases, why do our medieval manuscripts present us with a relatively uniform text? How do we get from the creative and vibrant oral epic song tradition that Albert Lord describes to the seemingly fixed text of the roughly two hundred manuscripts of the Iliad that survive from medieval times? Can’t we simply regard the examples of multiformity such as we find in the scholia at Iliad 20.30 as anomalies or interpolations, perhaps the work of rogue singers, scholars, and scribes, and conceive of the “true” Iliad as having existed in a primarily static form for two and a half millennia or more?
This will be the subject of chapter 1, and indeed it is a central question for this book as a whole. The short answer is that we could imagine the Iliad to have become fixed in the eighth, seventh, or even sixth century BCE, and that certainly seems to be the prevailing approach taken by editors of the twentieth century, [15] but I will argue that our surviving evidence supports a different model, namely an evolutionary model for the Homeric text, like the one developed by Nagy. Nagy identifies five stages of evolution of the Homeric poems, which move from relatively most fluid (and most multiform) to relatively most rigid (and far less multiform) over a span of more than fifteen hundred years. Such a model accounts for a great deal of multiformity in the epic tradition as it developed over the course of many centuries. [16] Nagy’s basic assumption, based on both internal and external evidence, is that the Greek epic tradition evolved over time, and therefore we are required to take a diachronic perspective.
As Mary Ebbott and I note in our 2010 book (Dué and Ebbott 2010:19–20), and as I will explore in more depth in chapter 1, the implications of Nagy’s model are many and significant. He fundamentally rejects the model which posits that an oral tradition came to an abrupt halt sometime in the eighth century BCE, when the new technology of writing was used to record the monumental epics of a single singer, who was able to transcend the limits of said tradition, which he effectively ended, and whose works we (for the most part) have in our textual sources dating only from the tenth century CE onward. Replacing that outdated and untenable paradigm, [17] Nagy’s evolutionary model offers a framework for an understanding of how an oral tradition and the technology of writing coexist and influence one another for a long time before writing becomes the dominant means of transmission.
Nagy’s model has many points of contact with that offered by E. S. Sherratt (1990). Sherratt analyzes various passages of the Iliad as examples of the difficulties of relating material culture to the poem for the purposes of “solving” questions of when and how it was composed, noting that many passages of the Iliad seem to present “the juxtaposition or super-imposition of more than one chronological reflection” (1990:810). She proposes to explain these archaeological layers as reflexes of the oral tradition in which the Iliad was composed, arguing that over a long period of evolution some historical periods would have been more generative than others, with corresponding changes in the formulaic diction and its representation of material culture.
While there are differences between the two models (Sherratt’s arguments are primarily centered on eras that fall into period 1 in Nagy’s model), both Nagy and Sherratt conceive of the Iliad as a work that evolved (with no teleology implied) over many centuries in a song tradition that dates at least as far back as the early palatial period of the Mycenaean Bronze Age, if not earlier. The work of both scholars makes clear that the process by which earlier and later material became incorporated into and integral to the oral formulaic diction resulted in a system from which it would be impossible to separate out and isolate the poetic contributions of different eras. Sherratt vividly illustrates that even within a single passage of the Iliad (for example the encounter on the battlefield between Achilles and Aeneas in Iliad 20) different eras of material culture are inextricably intertwined. Likewise the diction of Homeric poetry cannot be separated into distinct layers, even though it is clear to linguists that some formulas are earlier than others and were composed in different dialects in different eras. [18] And finally both scholars allow for a great deal of multiformity within the system up until the point that the Iliad and the Odyssey come to be prized Panhellenic possessions, at which point multiformity comes to be screened out.
Building on the work of earlier scholars who have demonstrated why we should expect the Iliad to be multiform (especially Lord 1960 and Nagy 1996a), this book asks two basic questions: First, what kinds of multiformity are attested in our surviving sources? And second, what are the implications of multiformity for our interpretation of the reception and transmission of Homeric poetry? The answers to these questions that I formulate in the following chapters have emerged from twenty years of collaborative work and discussion with numerous colleagues on the Homer Multitext. [19] I have often joked that the aim of the Homer Multitext is to “unedit” the Iliad, but indeed a central goal of the project is to present the historical witnesses of the Iliad unmediated by the interventions of editors seeking to reconstruct a hypothesized “original.” Only in this way can we gain a clear picture of the multiformity with which the Iliad has been transmitted to us. In our experience, it can be incredibly difficult, sometimes impossible, to ascertain what the historical sources actually transmit if one relies solely on existing publications of the scholia in print or the cryptic reporting of an apparatus criticus. [20] The Homer Multitext allows each document to be viewed and considered on its own terms.
It is this desire to peel back centuries of editorial intervention that leads me to title this book Achilles Unbound. I want to remove the bindings, so to speak, from the medieval manuscripts and fully examine their contents, [21] I want to study the surviving papyrus fragments in all their multiform messiness, and I want us to try to visualize without judgment the Iliad known to Plato and Aeschines. But my title also, I hope, conveys a sense of possibility. The attested multiforms of the Iliad give us an opportunity to know and appreciate a wider range of performance traditions for this remarkable poem than most of us have been taught to do. Although our attested multiforms derive from the later stages of the evolution of the poem, even so I submit that they give us a glimpse of the very long history of the text, access to even earlier Iliads, and a greater awareness of the mechanisms by which such a poem could be composed in performance.
I do not plan to show that Iliads radically different from our own were known in antiquity. There is no Iliad in which Achilles goes home after quarreling with Agamemnon, and no Iliad in which Achilles himself sacks Troy. There is no Odyssey in which Odysseus decides to stay with Calypso after all. The Iliad and the Odyssey were traditional poems whose stories were deeply ingrained in the culture that produced them, and though the poems unquestionably evolved over time, they did so within a conservative and highly sophisticated system of composition in performance that claimed to be inspired by the Muses, who were believed to have witnessed the events recounted. [22]
Here again we can see that myth and narrative are inextricably intertwined, and that fate, at least as it is presented in epic, is in many ways a reflex of both. There are many places within the Iliad where fate is suggested to be at least potentially multiform. Most famous is Achilles’ stated choice between fates in Iliad 9, where he implies that there are two possibilities for how his life will turn out: either he will die at Troy and have kleos (“glory in song”) that is “unwilting” (Iliad 9.413) or he will return home and have a long but unremembered life. Could Achilles really have chosen to return home at this point? Certainly in terms of narrative, it cannot be a viable variation. There is no Iliad if Achilles goes home. Likewise in terms of myth Jonathan Burgess (2009:43–55) has convincingly argued that Achilles does not really have a choice of fates in the Iliad (though he may have traditionally had such a choice before the war). The Iliad consistently presents Achilles as fated to die at Troy at the hands of Apollo: “It is clear that no other passages in the Iliad support Achilles’ assertion in book 9 that he can choose to live. Achilles is never unaware that he will die at Troy, nor does he ever really think that his fate is avoidable” (51–52). Indeed Thetis is said to have foretold his death repeatedly to Achilles (πολλάκι, Iliad 17.406–408). Instead the proclaimed “choice” is part of a larger struggle on the part of Achilles to accept his mortality, a major theme of the poem: “The dishonor of Achilles by Agamemnon and the death of Patroklos are major events that provoke Achilles to undergo contemplation and eventual (re)acceptance of his fate” (55). Here again then, as in Iliad 20, we find contemplation of a narrative alternative that probably was never a genuine multiform in myth, nor did it ever actually come into being in song. It is the contemplation of the alternative that is part of the tradition. [23] The alternative was never in fact chosen, nor could it be, because the poet and his audience already know the tradition; they know what Achilles “chose.”
Indeed if any character is “bound” by fate it is Achilles, as Laura Slatkin has shown in The Power of Thetis (1991). Zeus, the third ruler of the universe, does not get overthrown. In every case, he thwarts the next generation from taking over. A perfect example is Achilles, who is the son Zeus never had, according to Pindar, Isthmian 8:

ταῦτα καὶ μακάρων ἐμέμναντ᾽ ἀγοραί,
Ζεὺς ὅτ᾽ ἀμφὶ Θέτιος ἀγλαός τ᾽ ἔρισαν Ποσειδᾶν γάμῳ,
ἄλοχον εὐειδέ᾽ ἐθέλων ἑκάτερος
ἑὰν ἔμμεν: ἔρως γὰρ ἔχεν.
ἀλλ᾽ οὔ σφιν ἄμβροτοι τέλεσαν εὐνὰν θεῶν πραπίδες,
ἐπεὶ θεσφάτων ἐπάκουσαν: εἶπε δ᾽
εὔβουλος ἐν μέσοισι Θέμις,
οὕνεκεν πεπρωμένον ἦν φέρτερον γόνον ἄνακτα πατρὸς τεκεῖν
ποντίαν θεόν, ὃς κεραυνοῦ τε κρέσσον ἄλλο βέλος
διώξει χερὶ τριόδοντός τ᾽ ἀμαιμακέτου, Δί τε μισγομέναν ἢ Διὸς παρ᾽ ἀδελφεοῖσιν.—‘ἀλλὰ τὰ μὲν παύσατε: βροτέων δὲ λεχέων τυχοῖσα
υἱὸν εἰσιδέτω θανόντ᾽ ἐν πολέμῳ,
χεῖρας Ἄρεΐ τ᾽ ἐναλίγκιον στεροπαῖσί τ᾽ ἀκμὰν ποδῶν.
This the assembly of the Blessed Ones remembered,
When Zeus and glorious Poseidon
Strove to marry Thetis,
Each wishing that she
Should be his beautiful bride.
Love held them in his grip.
But the Gods’ undying wisdom
Would not let the marriage be,
When they gave ear to the oracles. In their midst wise-counseling Themis said
That it was fated for the sea-goddess
To bear for son a prince
Stronger than his father,
Who shall wield in his hand a different weapon
More powerful than the thunderbolt,
Or the monstrous trident,
If she wed Zeus or among the brothers of Zeus.
“Put an end to this. Let her have a mortal wedlock
And see dead in war her son
With hands like the hands of Ares
And feet like the lightning-flashes.” (trans. C. M. Bowra)

Zeus prevents his own overthrow by forcing Thetis to marry a mortal, Peleus, and their son of course turns out to be Achilles. Instead of becoming the supreme ruler of the universe, Achilles will be merely supreme among mortals, and doomed to die young, in battle. This same myth plays an important role in Aeschylus’ tragedy Prometheus Bound. In that play there is a secret that is a crucial part of the plot: Prometheus knows the key to Zeus’ downfall, and it is Thetis. As Slatkin has written (1991:101): “If Themis had not intervened, Thetis would have borne to Zeus or Poseidon the son greater than his father, and the entire chain of succession in heaven would have continued: Achilles would have been not the greatest of heroes, but the ruler of the universe. The price of Zeus’ hegemony is Achilles’ death.”

Achilles is bound by fate and by narrative tradition, but Achilles’ poem, the Iliad, was not fixed and monolithic in antiquity. It was multiform. And the wider epic tradition from which the Iliad emerged was more multiform still. It is my aim in this book to explore the traditionality and multiformity of the Iliad in a way that gives us a greater appreciation of what has been handed down to us.


[ back ] 1. Throughout this book the Greek text of the Iliad I cite is from the Venetus A manuscript (with only superficial editorial intervention, such as the addition of quotation marks for quoted text or punctuation, which the scholia of the Venetus A frequently omit). Discrepancies from printed editions often point to interesting examples of multiformity in the textual transmission of the Iliad, as here. For example, there was debate in antiquity about the reading of τέ in line 20.28 (the Venetus A scholia tell us that Aristarchus read τί, whereas most manuscripts and papyri, including the Venetus A, read τέ). Likewise ancient scholars were divided as to whether ὑπερ μόρον should be read as one word or two. In the Venetus A (folio 260v) there appears to be a space here between ὑπερ and μόρον (no accent on ὑπερ), even though the scribe has clearly used a hyphen to join ὑπέρμορα at Iliad 2.155. (The correct reading is also debated in the main scholia at 20.30 and in both the interlinear and main scholia surrounding 2.155 [folio 27r in the Venetus A].) Boreel and Yi (2017) have argued that the scholion at 20.30 in the Venetus A is a first-person extract from the On Prosody of Herodian.
[ back ] 2. Conceptually, such incidents seem related to the frequent appearance of contrafactual statements in Homeric epic (if x had not happened, then y would have), on which see De Jong 2004:68–81, with further bibliography ad loc. On the phrase ὑπερ μόρον (or ὑπὲρ μοῖραν or κατὰ μοῖραν) see, in addition to De Jong, Pestalozzi 1945: 40; Kullmann 1956; Fränkel 1962: 62–64; Matthews 1976, and note 23 below.
[ back ] 3. Cf. Slatkin 1991:111 on Hera’s seduction of Zeus as an unraveling of the plot of the Iliad, over which Zeus must reassert control at the start of Iliad 15: “Zeus’ omniscience fails in the face of his own desire. Invincible and all-knowing, he is nevertheless baffled by eros. In the [deception of Zeus] he is unable to see beyond his desire for Hera … and the consequence is that the plot of the Iliad is temporarily out of his control. Thus when he awakens to find what has happened, his response has less to do with punishing Hera than with reclaiming control over the narrative: he declares what the plot of the rest of the poem will be, and goes beyond.” On moira (“destiny”) as a function of tradition, see also Nagy 1979:40–41 and 268 (§3n9) and Scodel 2002:68–69.
[ back ] 4. Nagy 1979:45–48 and passim; see also Wilson 2005.
[ back ] 5. The [θ’], which is not present on the T manuscript but seems necessary for metrical reasons, is supplied by Erbse in his edition of the scholia.
[ back ] 6. I am grateful to Grace Anthony, Adam Beckwith, and Ashley Nickel for calling my attention to this particular scholion.
[ back ] 7. For more on the scholia derived from the Homeric Questions, see the Homer Multitext research blog:
[ back ] 8. On the wide-range ancient hexameter poetry that was being composed and performed contemporaneously with the Iliad and the Odyssey see Marks (forthcoming). Burgess (2001) demonstrates the antiquity of Cyclic epic traditions, even though they likely reached the textual form summarized by Proklos later than the Iliad and the Odyssey did. Burgess makes careful distinction between Homeric epic, Cyclic epic, and Cyclic myth. I too wish to observe this distinction, but my emphasis in this book is on how the multiformity of the mythological tradition comes to be reflected in the multiformity of the poetic tradition. Where there is multiformity in our attested sources for the text of the Iliad and the Odyssey, often, though not always, we can uncover an underlying multiformity in the mythological tradition.
[ back ] 9. See Parry 1932:7–8 (= Parry 1971:330): “A single man or even a group of men who set out in the most careful way could not make even a beginning at such an oral diction. It must be the work of many poets over many generations. When one singer…has hit upon a phrase which is pleasing and easily used, other singers will hear it, and then, when faced at the same point in the line with the need of expressing the same idea, they will recall it and use it. If the phrase is good poetically and so useful metrically that it becomes in time the one best way to express a certain idea in a given length of verse, and as such is passed on from one generation of poets to another, it has won a place for itself in the oral diction as a formula. But if it does not suit in every way, or if a better way of fitting the idea to the verse and the sentence is found, it is straightaway forgotten, or lives only for a short time, since with each new poet and with each new generation of poets it must undergo the two-fold test of being found pleasing and useful. In time the needed number of such phrases is made up: each idea to be expressed in the poetry has its formula for each metrical need, and the poet, who would not think of trying to express ideas outside the traditional field of thought of the poetry, can make his verses easily by means of a diction which time has proved to be the best.”
[ back ] 10. West (2001:254) argues similarly: “The three lines quoted in schT [= the scholia in the Townley manuscript, Burney 86] are a gloss on 30 ὑπὲρ μόρον. The scholiast says that they were read by some in place of line 30, but this cannot be right: they must either have been substituted for 28–30 or (more likely) appended after 30.”
[ back ] 11. Bolling (1925:187) argues that the verses derive from a Cyclic epic. Implicit in his argument is a conception of the Iliad as a fixed text that can be “interpolated” with verses from another poem. Such a view is incompatible with my own understanding of not only the oral tradition in which the Iliad was composed but also the methodology of ancient editors, on which see further below.
[ back ] 12. On the scholarly contempt with which variation, including and especially additional verses, is typically greeted by scholars, see Dué 2001a and chapters 2 and 3 below. An exception is Bird 2010.
[ back ] 13. Cf. Nagy 1996a:113. Nagy was in fact the first to assert the need for a multitext edition of Homer, for which his evolutionary model, discussed below and in chapter 1, forms a basis: “The ultimate purpose in drawing up this scheme is to lay the groundwork for an eventual multitext edition of Homer, one that would be expected not only to report variant readings but also relate them wherever possible to different periods in the history of textual transmission.” For more on the goals and theoretical underpinnings of the Homer Multitext see the project website ( as well as Dué and Ebbott 2010: 153–165, Bierl 2015:193–194, and Dué and Ebbott 2017.
[ back ] 14. Cf. Nagy 1996a:33 on the multiform πολυδευκέα at Odyssey 19.521 (attested in Aelian De natura animalium 5.38): “In considering these two variants…I am ready to argue that both are legitimate, both ultimately generated from the multiform performance tradition of Homer.”
[ back ] 15. On the twentieth-century desire to attribute the text fixation of the Iliad and the Odyssey to a single Archaic master poet see Dué 2006b and the conclusion. Martin West’s 2017 edition of the Odyssey is the most recent example of this approach to editing the Homeric epics.
[ back ] 16. For an early formulation, see Nagy 1996a:107–110 and chapter 1 below. For a more recent and further-refined formulation see Nagy (forthcoming).
[ back ] 17. For the recent demonstrations of the untenability of this model (and a survey of various dictation theories for the creation of the Homeric texts), see González 2013:15–70 and Ready 2015. See also chapter 1, pp. 42–44.
[ back ] 18. See especially the classic treatment by Milman Parry, “Studies in the Epic Technique of Oral Versemaking. II. The Homeric Language as the Language of Oral Poetry” (= Parry 1932; reprinted in A. Parry 1971:325–364).
[ back ] 19. I wish to thank in particular those who have regularly participated in an annual seminar held since 2005 at the Center for Hellenic Studies: Tazuko van Berkel, Christopher Blackwell, Eric Dugdale, Mary Ebbott, Douglas Frame, Madeleine Goh, Olga Levaniouk, Leonard Muellner, Gregory Nagy, Corinne Pache, Ineke Sluiter, and Neel Smith.
[ back ] 20. Before we had access to the images of the Venetus A manuscript that were taken in 2007, we attempted to reconstruct the texts of six of the oldest manuscripts of the Iliad using scholarly editions. The attempt was a failure because all scholarly editions report readings selectively and are prone to human error. One of the primary advantages of the Homer Multitext is that we provide not only diplomatic editions of the documents that transmit the Iliad but also links to high-resolution images of those documents, which allow for verification of our editions.
[ back ] 21. Collaboration with the E-codices project of Switzerland and the Bibliothèque de Genève allowed for the unbinding and photography of the thirteenth-century manuscript of the Iliad known as the Genavensis 44. Likewise the extremely tight binding of the eleventh-century manuscript known as the Townley (Burney 86) made it largely inaccessible to scholars until recent high-resolution images of the manuscript were captured and made publicly available by the British Library (
[ back ] 22. See Nagy 1979:271–272.
[ back ] 23. Cf. Burgess 2009: 52: “Counterfactual musings are part of Homeric poetry.” Cf. Iliad 21.275–278, in which Achilles (in his fight with the river) speculates that his mother lied to him about his fate: ἄλλος δ᾽ οὔ τις μοι τόσον αἴτιος oὐρανιώνων / ἀλλὰ φίλη μήτηρ ἥ με ψεύδεσσιν ἔθελγεν / ἥ μ᾽ ἔφατο Τρώων ὑπὸ τείχεϊ θωρηκτάων / λαιψηροῖς ὀλέεσθαι Ἀπόλλωνος βελέεσσιν (“No other of the Olympians is so responsible as my dear mother, who used to enchant me with lies, who claimed that I would be destroyed beneath the wall of the Trojans by Apollo’s swift arrows”). For other examples, see Macleod 1982:151. See also De Jong 2004:68–81.