The Homer Multitext and the System of Homeric Epic

The work of Milman Parry and Albert Lord demonstrated that the oral tradition of the Homeric epics is a system that is good and useful for composition-in-performance and that also creates meaning for the audience in a special way. While some oral traditions for which Parry and Lord’s work has been illuminating can still be studied in performance, for the Homeric epics we rely on written texts to investigate and interpret this system. The question of how these texts should be understood, represented, and used is the inspiration for the Homer Multitext. This digital project, of which we are editors, builds upon the theory and fieldwork of Milman Parry and Albert Lord and expands their findings by presenting the surviving textual witnesses to the Homeric poems in a critical framework designed to account for and highlight the fact that they were composed and transmitted within an oral tradition of composition-in-performance. How can a digital scholarly edition provide easier access to these texts and help us to understand the system of Homeric epic better? In this essay we will give examples from two kinds of primary sources: (1) medieval manuscripts that contain not only the text of the poetry but also accompanying scholarly commentary on the poetry, called scholia; and (2) papyri, more fragmentary texts from a much earlier time period than the manuscripts. These examples illustrate the kinds of evidence that these texts provide for understanding and interpreting the system of Homeric poetry. The Homer Multitext is a multi-dimensional, long-term, collaborative effort of many scholars that it is now starting to bear fruit, and we anticipate that it will continue to grow for years to come. We submit that our project embraces the full implications of the theory and fieldwork of Parry and Lord and extends them by taking advantage of the possibilities offered by digital editing, computational processing, and electronic publishing to present the Homeric texts in a way in which the oral, traditional nature of the epics can be better appreciated and investigated. [1] The Homer Multitext also seeks to present the textual transmission of the Homeric poems in a historical framework. As we have argued elsewhere (see especially Dué and Ebbott 2009 and 2010), such a framework is needed to account for the full reality of the complex medium of oral composition-in-performance in which these epics evolved over time and for the different historical contexts in which texts of these epics were produced. The Homer Multitext seeks to provide ways to explore these contexts both synchronically and diachronically. Using technology that takes advantage of the best available practices and open source standards that have been developed for digital publications in a variety of fields, the Homer Multitext offers free access to a library of texts and images that are licensed for scholarly re-use and to tools that allow readers to discover and engage with the Homeric tradition.

The question that inspired the creation of the Homer Multitext is this: “How do you make a critical edition of a now past oral tradition, like that of the Homeric Iliad and Odyssey, that spanned a thousand years or more? What is the best way to represent the textual history of songs that were created in and for performance, but survive only in textual witnesses from later eras as oral-derived texts?” As we have noted in our previous work, [2] the practice of textual criticism, as applied to most humanistic texts, has conventionally had the goal of recovering the original composition of the author. The work of Albert Lord has made perfectly clear, however, that in an oral tradition there is no original composition to recover and that seeking one for a song within an oral tradition “simply makes no sense.” [3] Thus the fundamental goal of textual criticism must be different in the case of the Homeric epics. At the same time, we do not have the opportunity to witness or record performances within the tradition. Thus our goal must differ also from those who collect records of living oral traditions. Our criticism must take into account both the shift from the oral medium to the written that happened over a long span of time and the fact that the surviving textual artifacts are from a still later period of time.

We will begin this discussion of the goals for our project by addressing the multiformity of the Homeric epics, as first articulated by Lord. This multiformity, which is natural to an oral tradition in which poetry is composed in performance, calls for a different way to think about the textual witnesses that preserve evidence of that multiformity. A digital medium still does not allow us to “hear” directly the flourishing oral tradition in which the Iliad and Odyssey were composed. But the change from printed texts and conventional textual criticism to digital media and what we describe below as a new critical approach to our texts does afford the opportunity to think about these oral-derived texts in new ways. That is, we are investigating what new ways of seeing a digital medium can provide, and how those changes in visualization as well as access can enable new perspectives and new questions. As we will be arguing here, this shift in medium opens up new paths for discovering and new means for evaluating the evidence for the oral tradition preserved in the textual record.

Parry and Lord’s foundational work demonstrated that the fluidity and multiformity of the Homeric textual tradition are a necessary product of the process of oral recomposition-in-performance. [4] One of the most important revelations of the fieldwork of Parry and Lord is that every time the song is performed in an oral composition-in-performance tradition, it is composed anew. Parry and Lord intentionally recorded the same singer composing and performing the notionally “same” song on multiple occasions, and likewise the “same” song composed and performed by different singers. Lord’s additional fieldwork in subsequent decades allowed him to go back and record those same singers again much later in their careers. The fieldwork of Parry and Lord therefore was inherently multitextual and provides a model for how to think about multiple texts of the “same” song.

The multitextual songs that Parry and Lord recorded reveal variation, or multiformity, to be a natural feature of the compositional process. Just as the idea of the “original” makes no sense in an oral tradition, the term “variant,” as employed by textual critics when evaluating witnesses to a text, is not appropriate for an oral compositional process in which the singer composes as he performs. Lord has explained the difference this way: “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.” [5] Our textual criticism of Homeric epic, then, needs to distinguish what may genuinely be copying mistakes from what are performance multiforms: that is, what variations we see are very likely to be part of the system and the tradition in which these epics were composed. [6]

Parry and Lord’s work, then, asks us to consider whether the variations we find in the textual record are valid parts of the system rather than mistakes or corruptions; doing so has meant changing the practices for textual criticism of the Homeric texts. Digital editing offers the opportunity for doing just that, for creating a new kind of textual criticism, one that we have called “digital criticism.” [7] One way in which a digital edition can change how we study and investigate the Homeric epics is by privileging our primary sources, while simultaneously preserving them and providing greater access to them. By providing access to primary sources—that is, the surviving medieval manuscripts and their scholia, the surviving papyri, and other texts that quote or comment on the Homeric epics—“digital criticism” redefines a critical edition as one in which readers have the same access to these materials as the editors do, with each witness and its historical context accurately represented.

Once readers have access to the witnesses just as we editors do, the nature of criticism changes. Instead of choosing just one from among multiple attested readings or even emending texts in an attempt to return to the author’s “original” composition, the editor creates a framework in which the textual witnesses are easily compared, placed within their historical contexts, and connected to yet further related witnesses. This new criticism still involves meaningful comparison of witnesses, but with fundamentally different aims. [8] The goal is not to choose one multiform as “right” or “original” but to show the reader that both exist, while explaining what the witnesses are, and pointing the reader to other witnesses that are related: for example, that contain a particular line, that contain the same formulaic expression, or that have some other relationship to the sources that the reader has chosen so that she can further her investigation. One challenge that this new criticism presents for editors is that they must take into account possible differing levels of expertise among readers. Such differences in expertise mean that we can no longer rely on a practice of reference to and abbreviation of primary sources that, although previously necessary in some ways because of the limitations of space in a print medium, serves mostly to obscure what the editor is reporting to have seen in the textual witnesses she has examined. Rather than devaluing the role of editors, then, this new concept of critical digital editions makes the role of editors even more important and necessary in representing precisely what each witness contains, creating the critical framework for presenting those witnesses, and putting the primary sources in historical context and in relation to one another in ways that users with different levels of expertise can understand. Below we examine particular multiforms that appear in Homeric papyri and in the scholia of the Venetus A manuscript of the Iliad, and with those examples we explain in more detail how this criticism works in practice.

Of course, not every variation that survives in our medieval manuscripts, papyri, or ancient quotations is rightly called a “multiform” or “performance variant.” Indeed, from these many types of witness, we find a number of different kinds of variation. It is easy to find in the textual transmission of the Homeric poems examples of multiforms in which different, but equally formulaic, words and phrases are used. [9] But there are also cases of a single word change, differences in word division or accent, and other matters of orthography. We believe that these kinds of differences, which have not been traditionally of great interest to textual critics, are important for what they can reveal about the textual tradition and the editorial practices of earlier stages of transmission.

We resist, therefore, the notion that by refusing to screen out even orthographical or other sorts of variation we might somehow be providing too much information, or not being critical enough. Even what might be considered outright scribal error—an inevitable source of variation when dealing with handwritten material—might have something interesting to tell us about how these texts were assembled and their scholia compiled. For example, an accidentally omitted verse, later copied in the margin, might reveal something about the sources from which the scribe was drawing, especially in comparison with other manuscripts. Just such an example can be found in the Venetus A at Iliad 5.57. This verse is not present in the manuscript’s main text, but it has been written in the margin. Closer study suggests that the omission is not an accident, but a carryover from the exemplar on which this manuscript is based. If our interpretation is correct, the verse in the margin is most likely the result of collation against a second source. [10] Other types of “errors” can also be revealing. Holy Cross undergraduate researcher Christine Roughan noticed that in the Venetus A, a running count of the similes in the Iliad is kept in the margins, but for a significant section of the poem the numbering is abandoned, only to be resumed at a much later point with the correct count. Her further investigation showed that one long sequence of missing numbers corresponds to a particular quire. That quire happens to be the one that contains the Catalogue of Ships in book 2, for which the scribe has also noted in the margins the count of each contingent’s number of ships (as reported in the poetry). [11] Such details about the construction of the manuscript will no doubt be of interest to codicologists, and indeed to anyone interested in the possible exemplars of the Venetus A.

The ancient and medieval witnesses to the Iliad and Odyssey that survive are themselves quite multiform in their presentation of the Homeric text. The tenth-century Venetus A manuscript of the Iliad just mentioned is an excellent example. A typical folio in the Venetus A contains 25 lines of the Homeric text, surrounded on three sides by a body of marginal notes, all written in the same minuscule hand. Individual comments in the main body of scholia are generally but not always preceded by lemmata in semiuncial script. In the gutters of each page and in the margins between the text and scholia are written an additional set of scholia, in the same semiuncial script as the lemmata. Outside the main column of scholia in the far outside margins of each page are sometimes additional semiuncial scholia; on a few folios these scholia are extensive. These additional scholia may be written in the shape of a cross, column, or another object. Still more semiuncial scholia may be found between lines and very near the text in various places around the page. At the far edge of many pages are the traces of (very likely) two correcting hands, [12] and from the beginning of the poem through verse 188 of Book 2 an interlinear paraphrase appears in a later (thirteenth century) hand. Throughout this commentary, which derives from the work of the Alexandrian editors of the Homeric poems, alternative readings are adduced and debated and thereby preserved. [13]

In the following sections we will examine first some examples of multiformity from papyri, those much earlier witnesses that exhibit the greatest multiformity. Then, in a later section (“Homeric scholia, a window onto the tradition”), we will provide an example from the scholia of the Venetus A manuscript. Within our discussions of these examples, we will continue to consider multiforms and scribal choices and errors. The digital diplomatic editions in the Homer Multitext report exactly what the witnesses contain because knowing precisely what each witness records has value for understanding the whole of the historical transmission of the epics and can provide the very evidence that is needed to continue to investigate these questions and many more. In addition to providing basic access to primary sources, then, the Homer Multitext seeks to make these sources more useful for research—research that then could ask questions of these sources that we have not been able to ask before. The Homer Multitext can also be used to understand better what Parry and Lord’s work has already shown. Let us begin by examining examples of the general multiform nature of the oral tradition and the ways in which it shows up in our earliest textual sources.

In the earliest witnesses—papyri from the third and second centuries BCE or quotations in Classical authors, such as Plato or Aeschines—we find differences on the level of entire lines of poetry. There are numerous verses in the papyri that are seemingly intrusive from the standpoint of the medieval transmission. These additional verses, the so-called plus verses, are not present in the majority of the medieval manuscripts of the Iliad. [14] Other verses that are canonical in the medieval manuscripts are absent from the papyri—these may be termed minus verses. Students of Parry and Lord know that composition-in-performance allows for the expansion or compression of the theme or episode that the singer is performing. These plus and minus verses are evidence of what any given performance might have included, the operation of the system underlying the performance, and what the epic tradition as a notional totality comprised. [15] Far from being “banal” [16] or “inept” [17] (and therefore not worth our consideration) we assert that these examples of compression and expansion—coming late as they do in the performance tradition yet at the same time early from the standpoint of transmission of the textual tradition—provide unique insight into the system in which the Homeric poems were composed, as well the means by which they were transmitted.

For the past few years we have been developing a collection of XML-encoded editions of Homeric papyri that users will be able to search and compare against other historical witnesses. The Homeric papyri are, with the exception of some ancient quotations, the oldest surviving witnesses to the text of Homer. Some papyrus fragments predate the medieval tradition by as many as 1200 years. They give us an otherwise irrecoverable picture of the Iliad and Odyssey as they were performed and recorded in ancient times. These fragments show that the poems were performed and recorded with a considerable amount of fluidity in antiquity, as indicated by the plus and minus verses. Also prevalent is variation in the formulaic phrasing within lines, such as the presence of one noun-epithet formula in place of another. It is not until about 150 BCE that the papyrus texts begin to present a relatively more uniform text.

Iliad p609 (Mertens-Pack 864.1; P. Mich. 6972) is one such early papyrus that displays these kinds of multiformity. It is a palimpsest and dates to the second century BCE, making it among the earliest of the witnesses to the text of Iliad 10. [18] The papyrus has two columns, containing 10.421–434 and 10.445–460, in which Diomedes and Odysseus interrogate and kill the Trojan spy Dolon. The papyrus exhibits both horizontal multiformity on the level of individual words and even whole lines, and also vertical multiformity, with a so-called “plus verse” between 10.432 and 10.434. [19] Many of these multiforms are not supported by any other extant manuscript or witness, and yet the variations themselves are demonstrably formulaic and, therefore, “Homeric.” [20]

For example, at 10.423, the Venetus A has the formulaic speech introduction τόνδ’ ἀπαμειβόμενος προσέφη πολύμητις Ὀδυσσεύς (“Odysseus who is crafty in many ways answered him and said”). The papyrus clearly has πολύτλας δῖος Ὀδυσσεύς (“much-enduring radiant Odysseus”) at line end and the first editor of this papyrus, Anthony Edwards, reconstructs the line as the equally formulaic [τὸν δ’ ἠμείβετ’] ἔ̣[πει]τ̣α πολύτλας δῖος Ὀδυσσεύς. [21] Such multiformity often appears within the textual tradition at this level of formulaic, frequently used lines. Both πολύμητις and πολύτλας are equally traditional and, indeed, distinctive epithets of Odysseus. Each formula also has a connection with ambush poetics, as we have explored in our recent book on Iliad 10. [22] Either is appropriate in this line, and so we need not reject either possibility. A multitextual approach to the Homeric poems allows both.

In our digital critical edition, then, whenever a reader indicates an interest in 10.423, she would be able to request all the witnesses that contain this line currently included in the HMT (with more being added as we continue to build it). She could then easily see on the screen and compare what each witness contains. The viewing options for this papyrus include displaying just the text the papyrus contains (that is, a diplomatic edition) and also a display with the missing portions filled in with an editor’s conjecture, clearly identified as such. For some expert readers, that alone will give them the information they are seeking as well as multiple pathways to follow for further investigation. For a less expert reader, however, the editor must give further context by providing easy access to the historical context, including the information that the papyrus is from the second century BCE and is a palimpsest, as well as a guide to where more information can be found, such as the photographs of the papyrus available elsewhere online. Similarly for the Venetus A, we will need to specify that it is a tenth century CE Byzantine manuscript and guide these less expert readers to our introduction to the manuscript and other publications about it. [23] Readers will also be able to search for other uses of the formula πολύτλας δῖος Ὀδυσσεύς in the Homeric corpus, and we will point them to commentary on the traditional meaning of the epithets (as we do in the footnote above). In sum, then, our edition is critical because it brings together multiple texts and offers tools for the meaningful evaluation of multiple witnesses and their multiform readings. As we continue to examine this papyrus for the kinds of multiformity it exhibits, the comparisons and commentaries we provide here give an idea of the kinds of context that the edition will make possible, and eventually will make explicitly available.

If we look at 10.425 on the papyrus, we find further horizontal multiformity. The papyrus reads: ]ι νημερτὲς ἔνισπε. Edwards (1984) reads this line as [ μο]ι νημερτὲς ἔνισπε (“tell me and don’t miss the mark”). There have been various suggestions for the restoration of the beginning of the line, based on what we find in other witnesses, but the metrical pattern of this text differs from the equivalent phrase we find in the Venetus A (δίειπέ μοι ὄφρα δαείω, “tell me so that I may know”). Because of this difference, Edwards notes the possibility that the entire line was different. [24] We should not assume that we know what was on the papyrus based on what appears in much later witnesses. The phrasing and placement is similar to what we find at the end of Iliad 3.204: ὦ γύναι ἦ μάλα τοῦτο ἔπος νημερτὲς ἔειπες (“Woman, you have spoken this word without missing the mark”). For the imperative of ἐννέπω, we can compare also Odyssey 4.642, νημερτές μοι ἔνισπε, which is also in a context that leads to plotting an ambush. Finally, p609’s version of the line matches the endings of lines such as Iliad 14.470, Odyssey 3.101, Odyssey 4.314, Odyssey 4.331, and Odyssey 22.166. δίειπέ μοι ὄφρα δαείω (the phrase attested in the Venetus A here) is not attested elsewhere in our Iliad or Odyssey, and ὄφρα δαείω appears only three times. [25]

Substitutions of one formula for another are typical of oral composition-in-performance, and we should not be surprised to find these kinds of multiforms preserved in the textual tradition. Other more complex kinds of performance variation are also attested in the papyri, however, as p609 demonstrates once again. At 10.432–434, the papyrus has a so-called plus verse. The Venetus A has the following sequence:

The lines on the papyrus have similar endings to the manuscript’s lines 10.432 and 434, [26] but on this papyrus there are two lines between these:

Edwards argues that “there can be no doubt that the sense … is the same as that of K 433.” [27] If this is so, we can see here an example of expansion (or, conversely, of compression in the Venetus A version): the singer in performance can expand episodes, scenes, or speeches with more lines if he so chooses. Even a relatively short and simple expansion such as this one seems to be evidence of the performance tradition. The language of these lines is formulaic, and they are just as likely to have been generated by a traditional singer as those found in the majority of witnesses.

The antiquity of this papyrus fragment and the attestation of these particular multiforms elsewhere in Homeric epic should give these variations weight. We resist approaches that dismiss them (or similar multiformity) in the early Homeric papyri as “eccentric.” P609 is no isolated example. The design of the Homer Multitext, then, provides access to these witnesses in full diplomatic editions and makes clear to readers both their early date and the formulaic nature of their language.

The Ptolemaic papyri as a group have been labeled “wild” and “eccentric” by papyrologists and editors of the Homeric poems alike. But as Dué has argued elsewhere, they are only “wild” from the perspective of the medieval transmission. The variations themselves present little that is surprising when considered from the viewpoint of the system as a whole. [28] P40 (P. Hibeh 19) is another such “wild” papyrus, dating to the first half of the second century BCE. [29] It is a lengthy papyrus with quite a few plus verses as well as multiformity within individual lines. The following is an excerpt from fragment L of this papyrus:

At verse 302, the papyrus seems to read ὣς ἔφαν εὐχόμενοι, μέγα δ᾽ ἔκτυπε μητίετα Ζεὺ̣ς (“So they spoke praying, and Zeus the deviser thundered loudly”) in contrast to the medieval manuscripts, which read ὣς ἔφαν, οὐδ’ ἄρα πώ σφιν ἐπεκραίαινε Κρονίων (“So they spoke, but not yet did Zeus bring it to fulfillment for them”). Following that verse, there are four plus verses that are not attested in the medieval manuscripts. At verse 303, the papyrus reads πρὸς where the medieval manuscripts have μετὰ. At verse 304, the papyrus reads Δάρδανοι ἠδ᾽ <ἐ>π̣ί̣κ̣<ουροι,> (“Daradanians and allies”) where the manuscripts read ἐϋκνήμιδες Ἀχαιοί (“well-greaved Achaeans”). After 304 there is another plus verse. At verse 306, the papyrus appears to read <ο>ὐ̣ γάρ κεν τλαίην <ποτ᾽ ἐν ὀφθα>λμοῖσιν ὁρᾶ<σθαι> (“for I would never dare to watch with my own eyes”) whereas the manuscripts have ἄψ, ἐπεὶ οὔ πω τλήσομ’ ἐν ὀφθαλμοῖσιν ὁρᾶσθαι (“back, since I will not yet dare to watch with my own eyes”).

As we can see in this small segment, this particular papyrus text differs from medieval texts of the Iliad in its number of lines and often within lines (with one verse almost entirely different at 302), but the variations are formulaic in nature, with the kind of compression and expansion and differences in choice of formula that we would expect of oral-derived texts. Verse 302 on the papyrus does not happen to survive in any other manuscript or papyrus in this particular location within the epic, but it is found at Iliad 15.377, indicating that it is a perfectly Homeric line. Verse 302 in the manuscripts (οὐδ’ ἄρα πώ σφιν ἐπεκραίαινε Κρονίων) is also attested elsewhere in the poem (at Iliad 2.419). Here is another case where a multitextual approach can accommodate both without excluding one or the other.

There are five so-called plus verses in a span of fourteen lines, but four of these have close parallels elsewhere in the Iliad. 3.302b–302c on the papyrus resemble closely Iliad 2.39–40:

Verse 3.302d on the papyrus can be found at Iliad 14.280:

And finally, verse 2.304a on the papyrus can be found at Iliad 19.102. In fact, this particular passage from Iliad 19, in which Zeus addresses the other gods, is contextually similar to the one we are exploring in Iliad 3:

We can see that verse 3.304 on the papyrus is likewise parallel to 19.101, with the substitution of the contextually appropriate Τρῶες καὶ Δάρδανοι ἠδ᾽ ἐπίκουροι for πάντές τε θεοὶ πᾶσαί τε θέαιναι.

These differences are not a case of a rogue scribe inserting lines from elsewhere in the Iliad into Iliad 3. Rather, each of the variations presented by p40 are the kind of multiforms that are natural to oral poetry composed in performance. Milman Parry demonstrated the significance of this for the text of Homer:

Following Parry, we, the editors of the Homer Multitext, view the early Homeric papyri as the vestiges of a once vibrant performance tradition of the Iliad and Odyssey. [32] The variations preserved in the Homeric papyri are a unique channel into the oral tradition that we have lost. We are not seeking to privilege the papyri in any special way over the medieval transmission; rather we seek simply to make the readings they contain readily available to scholars and anyone interested in the transmission of the Homeric poems, and to suggest that they have great historical value in the picture they present of the state of the Homeric texts in the earliest stage at which we have it.

More than that, we believe that the more complete our evidence, and, as a result of that more complete evidence, the better our understanding of the system in which the Iliad and Odyssey were composed, the better in turn we are able to understand the poetics of those compositions. Our multitextual edition and commentary on Iliad 10, [33] from which we have taken some of the examples of multiformity discussed here, was intended as an extended application of the work of Parry and Lord to an episode of the Iliad whose authenticity as Homeric poetry has long been questioned. One of our central assumptions in that work is that the more evidence that we take into account, the more the theories of Milman Parry, who died quite early in his scholarly career, are borne out, and the more illuminating they are as a result.

Parry’s work on the Homeric epics famously examined noun-epithet formulas as a way to demonstrate and understand the oral-traditional nature of the Homeric epics. This aspect of Parry’s work has been frequently challenged in subsequent scholarship, and, we believe, misunderstood in important ways. [34] The Homer Multitext, and a multitextual approach in general, can provide important evidence for expanding our understanding of these and other formulas. In our commentary on Iliad 10, [35] we examined Parry’s work on fixed epithets in detail by focusing on the epithet βοὴν ἀγαθός.

βοὴν ἀγαθός is a fixed epithet. Fixed epithets, as Parry describes them, exhibit four interdependent characteristics: they are used in accordance with their metrical value and not in accordance with their signification; they are traditional; they are always ornamental; and they are often generic. [36] βοὴν ἀγαθός is fixed because it is always used in the same position in the line: it follows the weak caesura, a common placement within the line for epithets, but it can be (and is) used with more than one hero’s name. Parry argues that these kinds of “rigorously” fixed epithets (which as a category can be seen as a subset of the more general kind of ornamental epithet), “clearly must have had, for the poets who used them, an existence independent of any particular type of noun-epithet formula.” [37] More controversial is Parry’s contention that the poet uses these epithets for metrical reasons rather than their signification, a point often targeted as the vulnerable underbelly of Parry’s work. [38] For example, an argument has been put forward, from a stated position sympathetic to Parry, that when this epithet is used at Iliad 10.283 it is actually contextually appropriate as its final appearance in Iliad 10 because, once the Achaeans have departed on their mission, they will have to be quiet. [39] In other words: no more shouting, we’re on an ambush.

In our commentary, we go into greater detail about this argument and the multiple ways in which its particular challenge to Parry’s work on epithets can be answered. But here we want simply to emphasize that a multitextual approach exposes a major methodological problem with such an argument, since it is predicated on this line containing the “final” instance of this epithet formula in the book. When we consult modern editions of Iliad 10, it can easily appear to be the case that βοὴν ἀγαθός is no longer used after verse 283. But when we consider the full range of evidence, including papyri and the scholia, and when we fully grasp what composition-in-performance means for the multiformity of the epic tradition, we see that such arguments about epithets in particular can be refuted on the basis of our textual evidence. On the papyrus that both Allen (1931) and West (1998) refer to as p90 (Oxyrhynchus 6.949 from the second–third century CE), there is a possible use of this epithet again at 10.446: [τον δ ημειβετ επειτα βοην αγα]θος Διομηδης, which would put it squarely in the middle of the mission, after Diomedes and Odysseus have captured Dolon. And from the scholia in both the Venetus A and Townley manuscripts, we have evidence of a version of 10.349, known to the Homeric critic Aristophanes of Byzantium, that also uses this epithet: ὣς ἔφατ’, οὐδ’ ἀπίθησε βοὴν ἀγαθὸς Διομήδης. The Homer Multitext allows easier access to such textual variation and thus changes the notion of what “the text” shows and does not show when it comes to the uses of these formulas. [40]

Since epithets like βοὴν ἀγαθός were so useful, flexible, and traditional—qualities that Parry himself identifies—it is difficult to prove the argument that the absence of one in a particular passage is significant: even with our limited evidence we can see that this epithet could indeed be used in more lines within this episode. Instead, a thorough understanding of the oral traditional poetics of these formulas requires us to take a larger view of their meaning, just as a true understanding of the textual tradition requires us to take into account all the evidence we have and to recognize that a singer might have sung this same episode with the multiformity natural to composition-in-performance. These are some of the differences between Homeric epic and modern poetry that Parry enjoined us to acknowledge and to appreciate with a different aesthetics. The Homer Multitext is being designed to allow us to see more easily the multiformity still present in the textual record.

We saw in the previous example how both papyri and scholia from medieval manuscripts offer important evidence for formulaic diction and therefore for the system of Homeric epic. Now we would like to take that one step further to look at how such evidence can also help us to understand the poetics of that system better. The multitextual manuscripts with scholia such as the Venetus A described earlier provide further evidence of multiformity in the system of Homeric epic. Easy access to and an accurate presentation of both the poetic texts and scholia of various manuscripts, then, also create an opportunity to consider what the multiforms in them can teach us about the poetry of the Iliad.

To get a sense of this opportunity from just the briefest of examples, let us look at a scholion in the Venetus A on Iliad 3.100. [41] This line is part of Menelaos’s reply to Hektor’s proposal of a duel between Alexander and Menelaos over the woman both men claim, Helen. Menelaos agrees to the duel and explains why. The main text of the Venetus A—as well as those of other manuscripts in the Homer Multitext, namely, the Venetus B and the two Escorial manuscripts (Υ.1.1 and Ω.1.12, E3 and E4 in Allen’s notation), and also most modern editions—reads: εἵνεκ᾽ ἐμῆς ἔριδος καὶ Ἀλεξάνδρου ἕνεκ᾽ ἀρχῆς· (“because of my conflict and because of the way it started with Alexander.”). The line is marked in the Venetus A with a sign called a diplē periestigmenē (>:), or a dotted diplē, which Aristarchus used to indicate that he was disagreeing with a reading of Zenodotus on that line. [42] The note itself reads (we have represented it here as it appears in the manuscript, which at times omits accents or other details we might expect to see): ὅτι Ζηνόδοτος γράφει ἕνεκ’ ἄτης. ἔσται δὲ ἀπολογούμενος Μενέλαος ὅτι ἄτη περιέπεσεν ὁ Ἀλέξανδρος. δια μέντοι τοῦ ἕνεκ’ ἀρχῆς· ἐνδείκνυται ὅτι προκατῆρξεν:~ (“[it is marked] because Zenodotus writes ‘because of [his] error’. This would be Menelaos saying in his defense that Alexander [was the one who] had fallen into error. With ‘because of his beginning’ however, it reveals that he began the hostilities.”).

In the Homer Multitext, as we have seen above with the examples from the papyri, if a reader develops an interest in Iliad 3.100, or in folio 43v of the Venetus A, she will be able to easily discover this scholion in a full diplomatic text. The names Zenodotus, Menelaos, and Alexander that appear in the scholion are also designated in the diplomatic edition with TEI-XML mark-up and unique identifiers, so that you know it is Alexander, the son of Priam, and not Alexander the Great or any other Alexander the scholia might mention. Searching for any one of those names within the HMT would also allow a reader to find this scholion. Connections to commentary on the line and scholion would then provide some of the information such as we offer below, about comparable lines and what other witnesses read or offer as alternatives.

The disagreement between Aristarchus and his predecessor Zenodotus contained in the scholion at 3.100 provides valuable information about textual criticism in an earlier age, from which survives only partial written evidence. In terms of understanding the passage, however, our first reaction might be that the explanation about the difference between the two versions is not all that illuminating: is there an important distinction in meaning between saying it was Alexander’s error and saying that he started it? And so this brief note, similar to so many others about the choices that Aristarchus made in contrast to those Zenodotus made, could be dismissed: such textual differences, modern scholars have said, are banal and unimportant, unworthy of consideration.

But to dismiss this different reading is to take a rather narrow view of this potential multiform. Why should we not consider Zenodotus’ reading a true multiform, or, we might also say, a performance variation, a potential end to a line in a performance when this was an oral poem? One piece of evidence in support of doing so is that the phrase Ἀλεξάνδρου ἕνεκ’ ἄτης appears in this same position, after the weak caesura and completing the line, in at least two other lines: (1) Iliad 6.356, in which Helen says to Hektor that Hektor has had to take the brunt of the suffering caused by herself and Alexander’s error, and (2) Iliad 24.28, where the narrator most famously alludes to the Judgement of Paris. For the latter line, the intermarginal scholion in the Venetus A records ἕνεκ’ ἀρχῆς as an alternate reading. The line at Iliad 6.356 likewise reads ἀρχῆς in a papyrus and several manuscripts, according to Allen’s (1931) apparatus. West (1998) adds two more papyri that contain the reading ἀρχῆς as either the main reading or an alternate one. So in each of these three lines, we have evidence for both multiforms, and this evidence tells us something about how oral composition-in-performance can work, when a singer either by training or under the pressure of performance might choose one or the other phrase in any of these places.

But this evidence should also make us pause before making any pronouncements about how often the Judgement of Paris is alluded to in the Iliad. Indeed, commentators often say that there is only one reference to it in the whole epic, the one in Book 24, [43] but if Menelaos could use the same phrase in Book 3 that we see expanded on in 24 to indicate the Judgement and could in this way encapsulate that traditional sequence of events, our understanding of the frequency and thus our interpretation of that frequency, would have to change. [44] It cannot be denied that the multiforms in the scholia have much to teach us about the composition of the poetry. In turn, they also provide substantial food for thought about how we then interpret the poetry that has been transmitted. The profit to be gained from this particular scholion, and many others like it, is not that it dramatically alters the sense of that one particular line, but that it makes a significant difference in our understanding of the system of Homeric epic as a whole. [45]

Our editorial rationale for the Homer Multitext as informed by the work of Parry and Lord underpins the creation of the digital editions of surviving witnesses to the Iliad and Odyssey within the Homer Multitext. These digital editions in turn allow for research into the texts in ways not possible before in terms of both natural language processing and data analysis of the poetic language. Even before our editions are complete, however, the Homer Multitext has published high-resolution digital images of five of the oldest multitextual manuscripts of the Iliad. These include the manuscripts in Venice known to scholars as the Venetus A (Marciana 822), the Venetus B (Marciana 821), and Marciana 841 (Allen’s U4), as well as two manuscripts in the Escorial in Spain, Escorial Υ.I.1 (294) and Ω.I.12 (513). Our collaboration with the e-Codices project of Switzerland has resulted in the publication of images of the Iliad manuscript with scholia known as Genavensis 44 that is similarly available for scholarly work under a Creative Commons License. The collaborators in the Homer Multitext are pursuing multiple lines of research simultaneously: historical, philological, and scientific research based on the content of the artifacts represented by this digital library, and research into theories, techniques, and practices of massive, scalable, rigorous, and highly distributed digital scholarship.

The Homer Multitext has benefitted immeasurably from the technological infrastructure and assistance provided by the University of Houston’s Research Computing Center. As we have noted, the project is a collaborative one, combining research in digital humanities with more traditional philology, and our work would not be possible without the support of the Research Computing Center. The HMT uses two servers on the University of Houston network: amphoreus.hpcc.uh.edu and beta.hpcc.uh.edu. The project’s information architects are Christopher Blackwell and Neel Smith, both Classicists who have acquired in the course of their own research considerable expertise in computational methods. Under their direction, the Homer Multitext has implemented a distributed, modular approach to integrating diverse digital objects into a research library. In broad terms, we deal with texts, images, and data-objects. Each example of each type of data has a unique, canonical identifier.

The amount of text and scholia contained in the five Homeric manuscripts already photographed is vast, and the project now publishes thousands of images that have not yet been fully transcribed or studied thoroughly. The focus of our current work in 2016 is a complete, web-based, digital scholarly edition of the entire contents of the Venetus A manuscript, the oldest complete witness to the poem, together with its scholia, which transmit scholarship about the poem from intellectuals in Alexandria, in Rome, and in Byzantium from the third century BCE until the Middle Ages. [46] There does not yet exist a comprehensive edition of this manuscript and its scholia, and thus no systematic way to evaluate the evidence it offers. Our earliest work on the manuscript suggests that a significant portion of its scholia has not been published in the most recent edition by Erbse. [47] We have therefore undertaken to produce a complete scholarly edition: its text, scholia, and all other elements on its 654 pages. The text and scholia are being transcribed (as a diplomatic edition, representing faithfully the text of the manuscript, including accents and spelling that are not “standard” from our point of view), and marked up with TEI-XML encoding for several key features. Each portion of the digital text, that is, each individual line of poetic text and each individual scholion, is then mapped to its location on the published digital images by means of an inventory procedure developed by Blackwell and Smith. Thus the edition that we are creating will feature “one-click” navigation from the transcription to the image of the primary source so that a user can see for herself what the manuscript says. This work, which has been funded by a National Endowment for the Humanities grant, is being published in installments with a projected completion date of 2017.

With the online publication of these images, freely accessible online and licensed for scholarly re-use, we expect scholars, professional and amateur alike, to contribute new discoveries about these manuscripts, their content and construction, and indeed the Iliad itself. The process of discovery began already when we were in Venice and Spain and Geneva, but it has never been our intention to limit the investigation of these pages to a select group of people. By making the manuscripts available in this way, we hope to encourage new and collaborative ways of exploring the Iliad, and new methods of scholarship. The Iliad has the ability to bring together scientists and historians and literary scholars, those interested in physical objects and the technology of creating them, preserving them, or capturing them digitally, those interested in the history of ideas, and those hoping to better understand the poetry and the tradition in which it was composed. Future research and publications on these manuscripts will further our own collaborative work on the Homeric poems, the system of oral poetry in which they were created, and the transmission of this oral tradition through time, but they will also allow us to learn from scholars from a wide variety of fields and thereby shed new light on a very ancient poem.

In a 2001 article, [48] Dué argued that the multiformity of the Homeric texts, as evidenced by the earliest quotations of Homer and the Ptolemaic papyri, calls for a new approach to editing the texts of Homer. Building on the work of Gregory Nagy, [49] who was himself building on the insights of Parry and Lord into the oral traditional nature of Homeric poetry, she suggested that a web-based, multitext edition would be truer to the complexity of the transmission of the Homeric poems, which are oral-derived texts composed in performance. The texts as we now have them are the product of many singers over the course of many generations. What Milman Parry and Albert Lord’s work shows us most essentially is that there is not one original text that we should try to reconstruct. Instead of reconstructing an “original text,” the aim of the Homer Multitext, now at last becoming a reality after many years of research and planning, is to present a series of complete, historically contextualized texts, together with images of the artifacts that transmit them, and a variety of tools with which users can compare and analyze these historical documents.

The work of Parry and Lord is the inspiration for and foundation of the Homer Multitext, and we believe that the Homer Multitext as it continues to evolve will be useful for new research that will expand their findings and therefore also expand our understanding of the oral and textual traditions of the Homeric epics.