BMCR 2000.09.12

Homeri Ilias

, Homeri Ilias. Vol. 1. Rhapsodias I-XII continens, vol. 2. Rhapsodias XIII-XXIV et indicem nominum continens.. Bibliotheca scriptorum Graecorum et Romanorum Teubneriana. Stuttgart and Leipzig: B.G. Teubner, 1998-2000. 2 volumes.

1 Responses

This edition of the Iliad is not, and cannot be, the last word. Still, it serves its purpose in presenting a reconstruction of what the editor deems to be the definitive text. The question remains, though: how do you define what exactly is definitive when you set out to reconstruct the text of the Homeric Iliad ? How you edit Homer depends on your definition of Homer.

Martin West, the editor of this latest version of the Homeric Iliad (volume I covers the first half of the overall text), argues that Homer did not exist (West 1999b). In denying the existence of a Homer, West is not arguing that Homer the poet is a mythical construct (as I have argued in N 1996b.111-112). For West, only the name of Homer is mythical (again, West 1999b). The Praefatio of West’s edition makes it explicit that the poet of the Iliad was not a mythical but a real historical figure, even if we do not know his name; this poet was the “primus poeta,” and he was “maximus” (p. v).

Here is West’s scenario for the Iliad of this master poet. The poem was written down in the course of the poet’s own lifetime ( Praefatio p. v). Even during his career, the poet had the opportunity to make his own changes in his master poem: there were major interpolations, says West, that the poet himself introduced into his written text from time to time (p. v). After the master’s death, the scrolls ( volumina) of his Iliad were abandoned to the whims of rhapsodes ( rhapsôidoi), who kept varying the text in their varied performances, much like the actors of a later era who kept varying the text left behind by Euripides (p. v: “rhapsodorum … qui Iliadem nihilo magis sacrosanctam habebant quam histriones Euripidem”). The opportunities for introducing more and more interpolations kept widening. Meanwhile, the master’s composition eventually made its debut at the Panathenaic Festival in Athens toward the end of the seventh century, but only in bits and pieces at the start (p. vi). In the late sixth century, the era of the tyrant Hipparchus of Athens, the text was formally adopted for Panathenaic recitations and divided up into 24 rhapsodies; in other words, this system of division had nothing to do with the “primus poeta” himself (p. vi n3; cf. West 1999b.382). The Athenian phase of transmission was consolidated in the sixth through the fourth centuries, with teachers playing a particularly significant role (p. vi). Throughout this period of Homeric transmission, the text suffered from Athenian accretions (p. vi).

Here ends my summary of West’s scenario, which serves as the premise for his edition of the Iliad. The editor’s task, in terms of this premise, is relatively straightforward: West sets out to reconstruct the seventh-century Ionic text of the master poet, which needs to be purged of its Athenian accretions, its rhapsodic variations, its editorial interpolations. I will now proceed to evaluate West’s premise, and his edition, against the historical background of previous editions of the Homeric Iliad and Odyssey.

The very idea of establishing the definitive text of Homer was destabilized over two centuries ago when Friedrich August Wolf published his Prolegomena to his editions of the Homeric Iliad and Odyssey ( Prolegomena 1795; Iliad 1804; Odyssey 1807).1 The ground has been shifting ever since. For Wolf, there was no Homer to recover. He argued that the Homeric text had evolved out of oral traditions, which could not be traced all the way back to some “original” single author. How far back, then, in terms of Wolf’s argumentation, could we trace the oldest of ancient Greek texts, the Iliad, if not all the way back to Homer? In other words, if we work backward from the surviving medieval textual tradition of Homer, just how far back can we go in reconstructing that tradition? To paraphrase the sarcastic response of a rival editor of the Iliad, Pierre Alexis Pierron, the Iliad edition of Wolf takes us no farther back than around the third century CE, the era of Porphyry’s Homeric Questions (Pierron 1869.cxl; cf. N 1999c.71).

For the likes of Pierron, we could indeed go farther back, at least as far back as the era of Aristarchus of Samothrace, head of the Library of Alexandria around the middle of the second century BCE, who produced what was thought to be a definitive diorthôsis — let us translate it for the moment as “edition” — of the Homeric Iliad.2 Pierron’s own edition of the Iliad (1869) was meant to be the closest thing to an edition of the Iliad by Aristarchus. The great Alexandrian’s diorthôsis, Pierron imagined, would be the closest thing to the Iliad of Homer.

So what was so special about the Homer edition of Aristarchus? As far as the ancient world was concerned, Aristarchus’ diorthôsis of Homer represented an authoritative attempt to restore the original text of the Iliad and Odyssey. As far as Aristarchus himself was concerned, such a restoration depended on finding the best surviving textual evidence, which he thought would ultimately lead back to whatever it was that Homer himself had written down sometime around 1000 BCE — if we convert the Aristarchean chronological reckoning to our own ( Πρόκλου περὶ Ὁμήρου 59-62 Severyns: τοῖς δὲ χρόνοις αὐτὸν οἱ μὲν περὶ τὸν Ἀρίσταρχόν φασι γενέσθαι κατὰ τὴν τῆς Ἰωνίας ἀποικίαν, ἥτις ὑστερεῖ τῆς Ἡρακλειδῶν καθόδου ἔτεσιν ἑξήκοντα, τὸ δὲ περὶ τοὺς Ἡρακλείδας λείπεται τῶν *TRWI+κῶν ἔτεσιν ὀγδοήκοντα. οἱ δὲ περὶ Κράτητα ἀνάγουσιν αὐτὸν εἰς τοὺς *TRWI+κοὺς χρόνους).3

In attempting to reconstruct the Iliad of Aristarchus, Pierron was following a tradition established by the Iliad edition of Jean Baptiste Gaspard d’Ansse de Villoison (1788), whose earlier attempt at recovering Aristarchus’ edition was based primarily on the text and scholia of a tenth-century CE manuscript of the Iliad commonly known as Venetus A (codex Marcianus 454), which Villoison himself had discovered in Venice. The text and textual apparatus of the Venetus A, according to Villoison, could lead us back to the text and textual apparatus of Aristarchus’ very own edition of the Iliad. For Villoison too, as for followers like Pierron, the recovery of Aristarchus’ Iliad would have been the closest thing to the recovery of Homer’s own Iliad.

At this point, a major question emerges: how are we to deal with the historical chasm separating Homer from Aristarchus? Villoison anticipated Wolf by imagining an oral tradition, perpetuated by rhapsôidoi or “rhapsodes,” which must have “corrupted” an original composition of Homer. For Villoison, as for Wolf later, the wording of Josephus Against Apion 1.12-13 could be interpreted to mean that this “original” composition of Homer had been oral, not written.4 For Villoison, the “original” oral composition of Homer had been eventually rescued from the “corruptions” of rhapsodic transmission, thanks largely to the research of scholars at the Library of Alexandria, especially Aristarchus. The text of the Venetus A codex of the Iliad was for Villoison the eventual result of this evolving rescue operation. Even if we could never recover an original Iliad, we could at least reconstruct the next best thing, that is, a prototype of the Venetus A text of the Iliad. For Villoison, such a prototype represented the recovery of the Iliad through the editorial efforts of Aristarchus in Alexandria and, secondarily, of such Alexandrian predecessors as Aristophanes of Byzantium and Zenodotus of Ephesus. The “original” composition of Homer seemed within reach.

Wolf’s Prolegomena changed all that. He challenged the authority of the Venetus A text and its scholia by questioning the credibility of Zenodotus, Aristophanes, and even Aristarchus as sources that could lead to the recovery of the “original” Homer text. For Wolf, most of these Alexandrian editors’ readings, wherever they differed from the readings that continued into the medieval text traditions, were mere conjectures. The historical chasm separating Homer from Aristarchus had dramatically reasserted itself.

Wolf’s pessimistic views on the validity of the Alexandrian editions of Homer led to a prevailing “default mentality” that relied primarily on the post-Alexandrian textual traditions (N 1997b.112). Wolf’s pessimism was most unsettling for the likes of Pierron, as we have already seen from that editor’s sarcastic remark: the Iliad of Wolf is the Iliad of Porphyry.

The recent editions of Homer by Helmut van Thiel ( Odyssey 1991, Iliad 1996) represent an extreme case of such pessimism. This editor systematically privileges the readings attested in the medieval manuscripts at the expense of variant readings attributed to the Alexandrians, which he generally dismisses as editors’ conjectures. A more moderate case is the Homer edition used up to now by most English-speaking Classicists, the Oxford Classical Text of T. W. Allen (with D. B. Monro).5

Yet another case of pessimism is the Teubner edition of the Iliad by West. He too is generally pessimistic about the Alexandrian editions of Homer. On the other hand, West is relatively optimistic about his own edition. He seems confident that he has recovered the closest thing to the putatively original Iliad. To that extent, he is like Villoison and his followers. Unlike Villoison, however, West has attained his own version of the “original” without relying on Aristarchus as a primary source of support. West’s Iliad shares with van Thiel’s version a general stance of diffidence about the textual variants that go back to Aristarchus — not to mention Alexandrian variants in general.

Granted, West is not as extreme as van Thiel is in this regard. Occasionally, he goes out of his way to defend a variant that goes back to Alexandrian sources, even when that given variant is weakly or not at all attested in the existing manuscript traditions. A notable case in point is Iliad IX 394, where West opts for the Aristarchean reading γε μάσσεται instead of γαμέσσεται, which is the reading transmitted by the existing manuscripts (cf. N 1998a [= BMCR 98.7.14]). As West remarks, with reference to this case and a handful of others ( Praefatio p. vii n9): “de bona traditione agitur, non de coniecturis” (also with reference to IX 397, XII 218, 412). Still, in most cases where a given Aristarchean variant is weakly or not at all attested in the existing manuscripts, West opts for a non-Aristarchean variant, as we will see later.

What makes some Aristarchean readings, but not most of the others, a matter of “bona traditio” and not “coniectura”? West’s approach to deciding what is good or bad tradition, what is traditional and what is conjectural, does not seem to me systematic. That is, his decisions about good or bad textual traditions are not based on external evidence. I can find no unambiguous instance where West prefers an Aristarchean variant for the simple reason that it stems from one manuscript tradition or another — or at least on the grounds that it derives from an ancient editorial source that guarantees pre-existing manuscript traditions no longer known to us. In the end, it all comes down to whether West believes that any given variant — Aristarchean or otherwise — happens to recover the right wording. It does not matter for him, ultimately, whether such a variant happens to be a “coniectura,” putatively deriving from some Alexandrian editor, maybe even from Aristarchus himself, or rather an authentic reading that the Alexandrians might have found preserved in a “bona traditio” no longer known to us. Ultimately, the goodness of the given tradition depends on whether West thinks that the given reading is right in the first place. He is not concerned whether a reading comes from an ancient source or from a conjecture, ancient or modern, as long as it is right. For him, a conjecture offered by, say, Richard Payne Knight (1820) can in theory be just as right as a reading found by Aristarchus in some ancient source.6 In taking a generally non-Aristarchean stance, West bypasses the neo-Aristarcheans, as represented by the so-called Königsberg school of Karl Lehrs (1st/2nd/3rd eds. 1833/1865/1882). The work of Lehrs challenged the Prolegomena and Homer editions of Wolf by undertaking a large-scale rehabilitation of the editorial methods of Aristarchus. Earlier followers of Lehrs included Pierron, who acknowledged the indebtedness of his Iliad (1869) to the neo-Aristarcheans. Among later followers, Arthur Ludwich stands out: his research culminated in a Teubner edition of the Iliad (1902; reissued 1995, three years before the appearance of West’s Teubner).7 In order to understand why West bypasses the neo-Aristarcheans, it is instructive to examine the basics of Ludwich’s editorial method.

Ludwich’s edition of the Iliad explicitly follows the editorial method of Aristarchus himself. Essentially, Aristarchus sought to balance the internal evidence of Homeric poetry, as a system of composition and diction, with the external evidence of the Homeric texts, as a sampling of manuscripts. The Aristarchean method is most clearly documented by Ludwich himself (1884, 1885), who went on to apply that method in his own edition of the Iliad (1902). Like Aristarchus, he tried to balance two different kinds of evidence. On the one hand, he made editorial decisions based on his own sense of the internal evidence of Homeric composition and diction. On the other hand, such decisions were regulated by the external evidence of the Homeric manuscripts — to the fullest extent of their availability. Here is where we begin to see a major difference between the editorial methods of Ludwich and West.

For the neo-Aristarcheans, the question of the availability of manuscripts is essential. Obviously, the external evidence available to Aristarchus was different from what was available to Ludwich — and what is now available to West. But the essential question remains, what exactly was in fact available to Aristarchus? The general conclusion reached by the neo-Aristarcheans is that Aristarchus himself had access to a wealth of manuscripts, containing a wealth of variant readings, and much of this evidence is no longer available to us. For the neo-Aristarcheans, the central source of this information must be Aristarchus himself.

Here we come back to the question: why is it that West chooses to bypass the neo-Aristarcheans? An answer that now emerges is this: it is because he has also chosen to bypass the authority of Aristarchus as a reliable guardian of the Homeric textual transmission. West’s reasons for this radical departure have to do mainly with his theory about Didymus. For West, as we will now see, it was not Aristarchus but Didymus who must be recognized as the central source of our surviving information about ancient Homeric manuscripts and editions. The implications of this stance are far-reaching. At stake here is the authority of Aristarchus himself as an editor of Homer.

Before we consider the details of West’s theory, we need to review some basic information about the methodology of Aristarchus and about the role of Didymus in reporting on that methodology.

Didymus was an Alexandrian scholar who flourished in the second half of the first century BCE and the beginning of the first CE. For posterity, the primary mediator of the Aristarchean method has turned out to be this man. Didymus’ commentary on the Aristarchean diorthôsis of the Iliad, as excerpted mainly in the scholia to Venetus A, has become the central source for reconstructing the editorial method of his Alexandrian predecessor, Aristarchus. Ludwich’s 1884 edition of the surviving fragments of Didymus’ commentary serves as an essential foundation for his 1902 Teubner edition of the Iliad (as Ludwich observes at pp. vii-viii of his Praefatio). This foundation has now been challenged by West, who offers a radically different theory about Didymus—a theory that he applies pervasively in his 1998 Teubner edition of the first half of the Iliad.

West’s theory, as we will see, depends on how you answer this question: what facts can we learn from Didymus about Aristarchus? In order to introduce the theory, I will now highlight two of these facts.

The first fact is the more obvious of the two: the methodology of Aristarchus, as mediated by Didymus, insists on adherence to the internal evidence of Homeric poetry, as a system of composition and diction. This editorial policy is indicated by the scholia A to Iliad XVI 467c, where Didymus observes that Aristarchus would not leave anything aparamuthêton‘uncontextualized [in the mythos]’, in other words, that Aristarchus’ goal was to make contextual comparisons with all available internal evidence. Aristarchus’ rigorous analysis of Homeric poetry as a system was monumentalized by his reputation as an “analogist,” in opposition to an “anomalist” like his contemporary, Crates of Mallos, who was head of the Library of Pergamon (Varro, De lingua latina 8.23). Crates was an editor of Homer in his own right, and his editorial judgments were frequently contested by Aristarchus (cf. Broggiato 1998.41; N 1999b.260). The antithesis between Aristarchus the analogist and Crates the anomalist is liable to various exaggerations, but the basic contrast is valid to this extent: Aristarchus as editor of Homer was more likely than Crates to reject as incorrect a given variant reading that does not fit the rest of the system as he saw it (N 1998b.219-223).

Now I confront a second fact — less obvious but more important — about the methodology of Aristarchus as mediated by Didymus: in producing his edition of Homer, Aristarchus did not overprivilege the internal evidence of Homeric diction at the expense of the external evidence of Homeric manuscript transmission. Here I strongly disagree with West, who claims that Aristarchus did indeed prefer to concentrate on the internal evidence of Homeric diction at the expense of the external evidence of Homeric manuscripts (p. viii).

It does not follow, just because Aristarchus was an analogist, that he would sacrifice the evidence of the manuscripts to his own sense of analogy. There is ample evidence to show that Aristarchus, although he was indeed an analogist, cautiously avoided mechanistic appeals to analogy at the expense of the manuscript evidence (N 1996a.129n99).

Ludwich’s Iliad edition consistently relies on the testimony of Aristarchus wherever an editorial choice has to be made between variant readings, on the grounds that Aristarchus adopted readings based on the external evidence of manuscripts even when he thought that some other reading was “right.” Aristarchus’ definitive statements on whatever he judged to be right or wrong, better or worse, were originally to be found not in the actual text of his edition but in his hupomnêmata (which can best be described as a combination of a modern apparatus criticus and a modern commentary; cf. Lührs 1992.10).8 We may compare the editorial policy of Origen of Alexandria (late second to mid-third century CE), who formatted the received text of the Septuagint as the fifth selis or “column” of his six-column Hexapla edition of the “Old Testament”; the critical signs in the margins of the fifth column (obelus, lemniscus, hypolemniscus, asterisk) would refer the reader to Origen’s hupomnêmata, where the editor offered his own judgments about the available variants (N 1996a.194-195, following Allen 1924.315-320). Similarly, Aristarchus’ hupomnêmata kept track of variants that were signaled by critical signs in the text proper of his edition.9

West does not see things this way, as we discover from a closer examination of his theory about Didymus. He seems to be implying that Aristarchus placed his own choices of variant readings into the text proper of his edition; more important, West says explicitly that it was not Aristarchus but Didymus who collected a mass of manuscripts to be collated for purposes of tracking down any non-Aristarchean variant readings (p. vi). According to West’s theory, it was Didymus, not Aristarchus, who developed such criteria as αἱ πλείους, πᾶσαι, αἱ Ἀριστάρχου… ἡ δὲ κοινή (cf. scholia to XII 404), αἱ χαριέστεραι, and so forth; it was he, not Aristarchus, who made use of “city” editions from Chios, Argos, Cyprus, Sinope, Massalia (Marseille), and so forth (p. vi).

West’s theory about Didymus leaves out of consideration—and is refuted by—the explicit testimony of Didymus himself, as mediated by the Homeric scholia, concerning the methodology of Aristarchus. A most explicit statement of Aristarchus’ editorial policy comes from the scholia A to IX 222, where Didymus says: ἄμεινον οὖν εἶχεν ἄν, φησὶν ὁ Ἀρίσταρχος, [εἰ] ἐγέγραπτο “ἂψ ἐπάσαντο” ἢ “αἶψ’ ἐπάσαντο”,… ἀλλ’ ὅμως ὑπὸ περιττῆς εὐλαβείας οὐδὲν μετέθηκεν, ἐν πολλαῖς οὕτως εὑρὼν φερομένην τὴν γραφήν. The last part of this statement, as underlined, is quoted in the apparatus criticus of Ludwich but not in that of West. The wording is crucial: we see here the most explicit testimony, coming from Didymus himself, concerning Aristarchus’ practice of comparing variant readings by examining a wide range of manuscripts. I have written elsewhere about this testimony (N 1998a): “The wording assumes that some of the texts did indeed feature ἂψ ἐπάσαντο instead of ἐξ ἔρον ἕντο. I infer that Aristarchus ‘changed nothing’ ( οὐδὲν μετέθηκεν) even though he could have made a change on the basis of manuscript attestations of a variant reading. Moreover, he is quoted as considering the variant reading as a contrary-to-fact proposition. Accordingly, it seems unjustified to describe such readings as his own editorial conjectures.” I should add that the verb metatithénai in the Homer scholia means ’emend’ not ‘conjecture’ when applied to the editorial activities of Aristarchus (Ludwich 1885.97). Also, whenever Aristarchus did make conjectures, he did not put them into his edited text (Ludwich 1885.92).

There is further evidence against West’s Didymus theory, provided by the ipsissima verba of Aristarchus. Here I cite a quotation from Aristarchus himself, as preserved in the scholia A to I 423-424. The quotation, designated as the lexis of Aristarchus, is introduced this way: λέξις Ἀριστάρχου ἐκ τοῦ Α τῆς Ἰλιάδος ὑπομνήματος. The expression λέξις Ἀριστάρχου here and elsewhere seems to convey the idea that the hupomnêmata of Aristarchus were not only a commentary written down in papyrus scrolls but also, at least notionally, a commentary delivered as lectures by Aristarchus, as if they were meant to be transcribed by his students.10 As the quotation proceeds, it is difficult to determine exactly where the words of Aristarchus himself leave off, to be picked up by the words of Didymus.11 This much is certain, however: the person who is being quoted, after expressing his preference for the variant reading κατὰ δαῖτα instead of μετὰ δαῖτα, immediately goes on to say: οὕτως δὲ εὕρομεν καὶ ἐν τῃ.12 As Ludwich (1884.194-196) points out, the context of οὕτως δὲ εὕρομεν makes it clear that the subject of this verb is Aristarchus, not Didymus: the first person of εὕρομεν comes from the direct quotation of words “spoken” (notionally and I would say perhaps even literally) by the master teacher. In other words, the rhetoric of the quotation is set in the mode of a master’s ipse dixit.

The distinction between what was said by Aristarchus and what defaults to Didymus is collapsed in the apparatus criticus of West, who reports (and chooses for his text proper) the variant κατὰ δαῖτα at I 424 on the authority of the following (I list them in the order given by West): Antimachus, Aristophanes, Aristarchus, Callistratus, Dionysius Sidonius, Ixio, four papyrus fragments, Apollonius Sophista, manuscripts “V” (=V1 Allen) and “Z” (=Ve1 Allen). In other words, Aristarchus is represented here and elsewhere as a source, not as a collator of sources. West juxtaposes (and rejects) the variant μετὰ δαῖτα, reported on the authority of three papyrus fragments, various ancient Homer-quotations, and the “Omega” family of manuscripts (more on which later; this family approximates Ludwich’s concept of “vulgate”). A casual reader of West’s apparatus is left with the impression that it was Didymus, not Aristarchus, whose collation of manuscripts set the framework for choosing between the variants, and that it was Didymus who chose κατὰ δαῖτα while all along being fully aware of μετὰ δαῖτα — to which variant West applies his ubiquitous formula: “novit Didymus.”

In his apparatus criticus at III 405, West applies the tag “omnes boni libri Didymi” on the basis of the following report of Didymus in the A scholia: καὶ οὐ μόνον ἐν ταῖς ἐκδόσεσιν, ἀλλὰ καὶ ἐν τοῖς συγγράμμασιν ἁπαξάπαντες οὕτως ἐκτίθενται. προσθήσειν μοι δοκῶ καὶ τὴν Ἀριστάρχου λέξιν οὕτως ἔχουσαν. [what follows here is the quotation of the ipsissima verba from Aristarchus’ exegesis]. I submit that the tag “omnes boni libri Aristarchi” would be more apt. From the wording of the scholia here at III 405, I interpret the report of Didymus this way: he first refers to the external evidence adduced by Aristarchus, and then he follows up by quoting the master’s exegesis of the internal evidence.

West’s theory about Didymus also leaves out of consideration another basic aspect of Aristarchean methodology. It concerns the criterion of numerus versuum, which has been studied extensively by Michael Apthorp (book 1980, articles 1990 and following). West’s general bypassing of Alexandrian editorial methodology is most strikingly exemplified by his lack of engagement with this criterion and with the work of Apthorp, who is nowhere cited in volume I of West’s Iliad. Apthorp builds on Aristarchus’ criteria for establishing an “authentic” numerus versuum, that is, a fixed number of verses that the Alexandrian editor had deemed genuine in the text of the Homeric poems. These criteria have to do with the external evidence of stronger vs. weaker attestations in the available manuscripts of the Homeric text, not just with the internal evidence of Homeric diction and composition.

The far-reaching implications of Aristarchus’ principle of numerus versuum can be explored further by rethinking (and moderating), as Apthorp has done, the extreme formulation of George Melville Bolling (1950.1-16). For this particular neo-Aristarchean, the canonical length of the “Pi Text,” which was the putative “archetype” of our Iliad and which supposedly dates from the sixth century BCE, was around 14,600 verses.13 Over the next couple of centuries — so goes Bolling’s formulation — the Iliad grew in length to 15,600 verses; this was the accretive “Alpha Text” with which Aristarchus was forced to contend and which corresponds roughly to the Iliad that survives by way of the medieval manuscript traditions. In terms of Bolling’s formulation, Aristarchus reduced the text of the Iliad from ca. 15,600 to ca. 14,600 verses by way of athetesis (that is, marking a given verse in the left margin with the critical sign known as the obelus, to indicate the editor’s doubts about the appropriateness of that verse) or outright deletion. Such verses were thought to be textual interpolations, the result of “corrupting” accretion.

Bolling’s analysis of Homeric “interpolations” needs to be refined in terms of two related considerations: (1) the principle of numerus versuum and (2) the distinction between the editorial procedures of athetesis and deletion. As Apthorp argues (1998.187), the Homer edition of Aristarchus became the standard source for subsequent applications of the editorial principle of numerus versuum, and literary authorities like Plutarch were well aware of this principle. Apthorp emphasizes that Aristarchus in his Homer edition not only athetizes some verses (that is, marks them with an obelus but keeps them in the text proper): he also deletes (“omits”) some other verses altogether. The criterion for deletion (“omission”) was based on manuscript evidence. To quote Apthorp (1980.xv), “Aristarchus … omitted only lines which he found very weakly attested.”14 Such lines are “plus verses.”15 Apthorp (1980.xvi) goes on to argue that “the numerous lines absent from all our mss. which we know to have been pre-Aristarchean but absent from Aristarchus’s edition — some cited by the scholia, some present in extant Ptolemaic papyri, some included in ancient quotations or discussions of Homer — stand condemned as interpolations alongside the weakly-attested lines of the mediaeval mss.” Whether or not we choose to think of these “plus verses” in terms of “interpolations,” the point for now is simply this: Aristarchus’ foundational research in determining the editorial principle of numerus versuum must have required a large-scale collection and collation of manuscripts — as many and as varied as he could find.16

The question of an “original” Homeric numerus versuum now brings us to an essential point that we have not yet considered about Homeric textual variants. In fact, there were not one but two dimensions of variation in the history of Homeric textual transmission. Before I introduced the topic of numerus versuum, we were speaking exclusively in terms of “horizontal” variants, that is, where the ancient editor had to choose between different wordings that make up a single line of Homeric poetry. The point is, the ancient editor also had to contend with “vertical” variants, where he had to choose between fewer or more lines that make up a given sequence of lines (N 1996a.139-140).

The vertical dimension of textual variation can best be understood by coming to terms with one of the most basic — and elusive — concepts in the history of Homeric scholarship, the so-called Homeric “vulgate.” Elsewhere, I have studied this concept by focusing on the Aristarchean usage of the Greek term koinê in the combined sense of a “standard” and a “common” text of Homer; this combined sense, I argued, corresponds to Jerome’s usage of the term vulgata (N 1997b.118-122). I confine myself here to stressing a single point in my earlier argumentation: modern editors of Homer tend to skew the meaning of the term “vulgate” by overemphasizing the sense of “common” and underemphasizing the sense of “standard,” so that the expression “Homeric vulgate” is commonly understood to mean nothing more than the default text of Homer. The usage of West is illustrative: he highlights the readings of Aristarchus, as reported by Didymus, against the backdrop of the “vulgata traditio” (p. vii).

I have no space here to develop my thesis that the Aristarchean mentions of the koinê or koinai texts refer to a standard Athenian version of Homer (N 1996a.117, 133-134, 152-156, 185-190, 193-195). The point to be made, for the moment, is simply that the concept of a Homeric vulgate can be applied systematically to the study of Homeric variants, and that the vertical variations in the history of the Homeric text provide a particularly rich source for such systematic study. This point is linked to my ongoing argument that Aristarchus’ editorial policy concerning vertical variants reveals his active interest in the collecting and collating of manuscripts.

In order to test the concept of a Homeric vulgate, let us turn to the editorial criteria of a neo-Aristarchean like Ludwich, who systematically distinguished between the vulgate and the Aristarchean versions of Homer. For this editor, the vulgate Homer was both pre-Aristarchean and post-Aristarchean.17 If we apply Ludwich’s distinction between vulgate and Aristarchean readings to the vertical dimension of textual variation, then we can say that the pre-Aristarchean vulgate represents an accretive text that exceeded the length of the putative Homeric “original.” We can also say that the work of the Alexandrian editors, especially Aristarchus, helped re-establish the “original” numerus versuum by way of deleting (“omitting”) the accretive verses. We can even say, finally, that the post-Aristarchean vulgate represents a newly accretive textual tradition that unsystematically re-absorbed accretive verses deleted by Aristarchus.

As the discussion proceeds, it will be evident that I do not agree with Ludwich’s views concerning the inherent superiority of Aristarchean readings over their vulgate counterparts. I focus here simply on his systematic application of the actual distinction between Aristarchean and vulgate readings. This distinction continues to be most useful, as we will see, for studying the phenomenon of vertical — and horizontal — variation in the history of the Homeric textual tradition.

In this context, it is instructive to contrast Ludwich’s neo-Aristarchean view of the “vulgate” (1898) with such anti-Alexandrian views as represented by Marchinus van der Valk (1963.609), for whom a pre-Aristarchean “vulgate” had “preserved the authentic text,” and this text “was also transmitted by the vulgate of the medieval manuscripts.” The contrast has to do with the validity of the Alexandrian editions of Homer (N 1997b.114-115):

“For both Ludwich and van der Valk, this ‘vulgate’ is distinct from the Homer ‘editions’ of the Alexandrians, especially that of Aristarchus. For van der Valk, however, the readings of the ‘vulgate’ are generally more authentic than the variant readings attributed by the Homer scholia to scholars like Aristarchus, which he generally takes to be ‘conjectures’; for Ludwich, by contrast, such variants are not ‘conjectures’ but authentic readings preserved by the scholia from the Alexandrian editions of Aristarchus and others [cf. N 1996a.185]. For Ludwich, the Alexandrian ‘edition’ of Aristarchus represents a quantum leap beyond the pre-Alexandrian ‘vulgate’; for van der Valk, by contrast, the pre- and post-Alexandrian ‘vulgate’ text is relatively superior to the Alexandrian ‘edition’ of Aristarchus, which may not even be deserving of the term ‘edition.'”

The negative position of van der Valk in questioning the validity of the Alexandrian editions has been influential in shaping the views of Homerists about the variants reported by Aristarchus (N 1996a.136-137). Haslam remarks (1997.87): “A newly fashionable attitude, owed to van der Valk, is to revere the vulgate and condemn the Alexandrians for tampering with it.” A prominent defender of van der Valk’s anti-Alexandrian position has been Richard Janko (1992.21n6; cf. N 1996a.137). I should stress, however, that Janko’s own negative position applies mainly to Aristarchus’ treatment of what I call horizontal variants. When it comes to vertical variants, Janko’s position shifts, at least in part: like the neo-Aristarcheans, he seems to approve whenever Aristarchus deletes (“omits”) plus verses on the grounds that they are interpolations (1992.21n6; cf. N 1996a.139-140), though he tends to disapprove whenever Aristarchus athetizes (1992.27). I think that Janko is applying a double standard here. Further, as we will see later, I generally think that there is no need to decide whether the vertical — and horizontal — variants transmitted by Aristarchus are inferior or superior to those transmitted by the vulgate. For the moment, though, I simply note the importance of the variations themselves, and the conflicting inferences they inspire.

There are important precedents for van der Valk’s and Janko’s negative positions about the Alexandrian editions. A most notable example is the work of Hartmut Erbse (1959). Some have interpreted Erbse’s negativity about the Alexandrian editions as a revolution against neo-Aristarcheans like Lehrs and Ludwich, who in turn had revolted against the 1795 Prolegomena of Wolf. Rudolf Pfeiffer (1968.215) puts it this way: “it looks to me as if by a sort of unconscious counter-revolution Wolf has now been put back on the throne from which Lehrs had driven him.” For now I leave aside the question whether the work of Lehrs and Ludwich represents a revolution — or a counter-revolution, from the standpoint of those for whom Wolf’s work was the real revolution in the first place. Instead, I continue to focus on the far-reaching consequences of choosing either an Aristarchean or a “Wolfian vulgate” model in the editing of Homer.18 Pfeiffer’s formulation may conceivably apply to the Odyssey and Iliad of van Thiel (1991, 1996), but surely not to the Iliad of West, who generally bypasses Wolf and the model of the “Wolfian vulgate.” More important for now, we have seen that West generally bypasses the model of the neo-Aristarchean editors as well. Most important of all, he even bypasses the editorial method of Aristarchus himself.

For West, there is no need to go back to the edition of Aristarchus in order to recover the text of Homer. He thinks that the closest thing to an original Iliad is his own reconstruction of the text composed by the “primus poeta.” West’s belief that the poet of the Iliad produced a written composition helps explain this editor’s optimism about reconstructing the ipsissima verba of a prototypical poet. Aristarchus’ own belief was probably similar to West’s, to the extent that he too posited a written text produced by the prototypical poet. He too aimed to reconstruct the poet’s ipsissima verba, relying on the external evidence of available manuscripts as well as the internal evidence of the composition and the diction. West relies on the same two kinds of evidence, but he differs from Aristarchus in how he uses that evidence. Further, he differs from Aristarchus (and the neo-Aristarcheans) in how he responds to this basic challenge confronting an editor of Homeric poetry: how is one to judge the variants of the overall manuscript tradition? In general, West’s version of the Iliad has less flexibility in allowing for variation, while that of Aristarchus has relatively more. It comes as no surprise, then, that West’s version of the master poet’s ipsissima verba frequently differs from the version of Aristarchus.

I submit that Aristarchus’ ancient edition of the Iliad, if it had survived in its original format, would in many ways surpass West’s present edition. It would be a more useful — and more accurate — way to contemplate the Iliad in its full multiformity.

The multiformity of Homeric poetry is an aspect of its prehistory as an oral tradition. This essential observation emerges from the research of Milman Parry (collected papers 1971, = MHV) and Albert Lord (1960, 1991, 1995), who explained Homeric poetry as a system derived from oral poetic traditions of composition-in-performance (see Dué 2000b). Their explanation is validated by the editorial work of Aristarchus. Although this ancient editor hardly thought of Homer in terms of oral traditions, his objective study of the Homeric texts provides crucial evidence for an inherent multiformity, which is indeed typical of oral poetic traditions. As Parry asks, in criticizing the “neo-unitarian” Homerists of his day (1930 [ MHV ] p. 268): “How have they explained the unique number of good variant readings in our text of Homer, and the need for the laborious editions of Aristarchus and of the other grammarians, and the extra lines, which grow in number as new papyri are found?”19 In the Praefatio to his Iliad, West ignores altogether the work of Parry and Lord. Throughout his edition, moreover, there is a noticeable lack of engagement with oral poetics. West believes that the master poet wrote the Iliad, and that is all there is to it.

In the Homeric textual transmission, we can find many signs of oral poetry, but West consistently prefers to explain such signs in terms of written poetry. For example, with reference to the mechanics of “movable nu” in the Homeric text, West adduces the evidence of the mid-seventh-century Nikandre inscription ( CEG 403): the poetry of this inscription, composed in the Ionic dialect, shows that movable nu was already being used in this early period for the sake of preventing hiatus caused by the loss of digamma in this dialect. Instead of considering the argument that the technology of writing involved in such early inscriptions was used not for the actual composition but only for the recording of poetry (cf. N 1996b.34-37), West simply assumes that the use of movable nu in such an early poetic inscription is proof that the analogous use of movable nu in Homeric diction was likewise a matter of written poetry (West 1998a.101):

“This [= the use of ‘movable nu’ in the Nikandre inscription] shows that by that time Ionian poets were already using movable nu to cure digamma hiatus. If one believes, as I do, that the Iliad was composed and written down at about the same period, one will take this as adequate justification for leaving such nus in the text. It is not unreasonable to suppose that the insertion of particles such as γε, κε, τε, or ῤα for the same purpose began as early as the addition of the nu.”

West assumes that the Nikandre inscription is a case of written rather than oral poetics, and he implies that the dating of the Nikandre inscription therefore gives him a terminus ante quem for the writing down of the Iliad. I disagree. The justification for leaving such nus in the text is to be found in the internal and comparative evidence of oral poetics, not in the fact that we can find such nus in a poetic inscription that West thinks is contemporaneous with the composition of the Iliad. It can be argued that the “insertions” of such particles and of morphophonemic elements like movable nu are part of an overall formulaic system and have nothing to do with the technology of writing. (For parallels to be found in the living traditions of South Slavic oral poetry as recorded by Parry and Lord, see the discussion of hiatus-breakers [= “bridging consonants” that prevent the occurrence of glottal stops] in Foley 1999.73-74, 85, 88, 293n28.)

As I will now argue, West’s lack of engagement with oral poetics has a direct effect on the actual text of his Iliad edition, especially with respect to his judgments about horizontal and vertical variants in the textual evidence.

In its own way, the factor of oral poetics has a direct effect on the editing of Homeric poetry, as I have previously argued in several publications (especially N 1996a, 1996b, 1997b), none of which is cited in West’s Praefatio. In terms of oral poetics, there is an important reason for keeping track of all horizontal and vertical variations in the Homeric textual transmission as attested in the medieval manuscripts, the ancient quotations, and the papyrus fragments. The reason is simple: in principle, any surviving variant in the Homeric textual transmission may represent an authentic form generated by Homeric poetry.

From the standpoint of oral poetics, Homeric poetry is a system that generates the forms that survive in the texts that we know as the Homeric poems. Homeric poetry is not the same thing as the Homeric texts that survived in the medieval manuscripts, the ancient quotations, and the papyrus fragments. Still, the Homeric textual tradition stems from the performance tradition of Homeric poetry. It can be shown, by way of formula analysis, that many of the variants we find in the Homeric texts result directly from the variability inherent in the poetic system itself (see N 1998a). In other words, the textual variants in the Homeric poems stem from formulaic variations in Homeric poetry.

Even from the standpoint of West’s theory of a written text produced by the “primus poeta,” there is an important reason for keeping track of all vertical and horizontal variations in the Homeric textual transmission as attested in the medieval manuscripts, the ancient quotations, and the papyrus fragments. Again, the reason is simple: in principle, any surviving variant in the Homeric textual transmission may represent the authentic form that goes back to the prototypical text of the prototypical poet. In terms of West’s editorial stance, there can be ultimately only one authentic form in each case of variation, that is, the form that goes back to the “primus poeta.” For West, the horizontal and vertical variants in Homeric textual transmission are important only insofar as one given variant in each case may be the right form, while all other variants would have to be wrong forms. Even where West is not sure whether a given variant is right or wrong, he needs at least to determine whether it has a better or worse chance of being the right reading: thus there can be better or worse variants, even where the difference between right and wrong variants cannot be determined.

Obviously, this editorial stance is necessary for those like West who espouse the theory of an original written Iliad and Odyssey by an original poet. As West says (1998a.95), “We may assume that there existed a complete and coherent Urtext of each epic, the result of the first writing down.” Maybe less obviously, this same editorial stance is also necessary for Richard Janko, even though he espouses the theory of an orally composed Iliad and Odyssey. Janko actively uses the criteria of right or wrong, better or worse, in judging the variants surviving in the Homeric textual tradition (cf. Janko 1992.26). Like West, Janko needs these editorial criteria because he too needs to assume original texts from which our Iliad and Odyssey can be derived. These poems, he thinks, were written down from dictations made by Homer himself, sometime in the eighth century BCE (see especially Janko 1992.22, 26). Thus, in terms of Janko’s dictation theory, the “original” Iliad and Odyssey are not only oral compositions but also, at the same time, textual archetypes. In a review of West’s edition of the Iliad, Janko (2000) expresses general agreement with West’s construct of an original text of the Iliad, and most of their disagreements are limited to specifics, except for the general distinction between a writing poet (West) and a dictating poet (Janko). In a review of my work (N 1996a), Janko (1998a) complains about my resisting his theory about the origins of Homeric texts. At the conclusion of his review of West, he recalls this complaint of his (Janko 2000.4): “Given that it has been claimed, by some who would have us know less than we used to, that there was no original text of Homer [here he refers to his review, Janko 1998a, of N 1996a], these disagreements [with West] are minor.”

I would never want to say it that way, that there was no original text of Homer. Obviously, at some point in history, there had to be a first time for textually recording what we know as the Iliad and Odyssey. But the basic problem remains: such an Iliad and Odyssey were not the same texts that we know as our Iliad and Odyssey.

So I prefer to say it this way: there is no original text of the Iliad and Odyssey that Janko or West or anyone else can reconstruct on the basis of the existing textual tradition. The variations that survive in this textual tradition, many of which are transmitted by Aristarchus, prevent such a monolithic reconstruction. The significance of these textual variations has been dismissed by Janko, who follows van der Valk in claiming that most Aristarchean readings, especially those that are weakly attested in the surviving manuscript traditions, are “conjectures” and should be rejected (1998a). In his review of West’s Iliad, Janko (2000.1) commends him for following this same editorial stance: “Following van der Valk, he [West] holds that Zenodotus was arbitrary and the majority of Aristarchus’ unique readings are wrong.”20 Elsewhere, I have called into question Janko’s editorial stance by critically re-examining each one of his entries in his sample list of some thirty-three cases of Aristarchean “conjectures” in the Iliad (N 1998a).

In general, I have also called into question the overall editorial stance of judging variants in terms of a hypothetical Homeric archetype. As I have argued in an earlier work (N 1996a.153), “we cannot simplistically apply the criteria of right or wrong, better or worse, original or altered, in the editorial process of sorting out the Homeric variants transmitted by Aristarchus or by earlier sources”; instead, we can ask “whether a variant is authentic or not — provided we understand ‘authentic’ to mean in conformity with traditional oral epic diction.” If indeed Homeric poetry, as a system, derives from traditional oral epic diction, then we can expect such a system to be capable of generating multiform rather than uniform versions, and no single version can be privileged as superior in and of itself whenever we apply the empirical methods of comparative philology and the study of oral tradition (N 1996a.117-118).

In terms of oral poetics, the multiformity of variations applies to vertical as well as horizontal variants in the textual history of the Homeric poems. With specific reference to vertical variations, it needs to be emphasized that whatever appears to be an interpolation or an omission of a verse in terms of Homeric textual history may in fact be a matter of expansion or compression respectively in terms of oral poetics. (On the mechanics of expansion and compression in oral poetics, see N 1996b.76-77.)

In terms of oral poetics, moreover, the compositional phenomena of expansion and compression correspond respectively to the lengthening and shortening of the number of verses in expressing a given essential idea. In terms of textual history, such cases of lengthening and shortening could be reinterpreted as corresponding respectively to the scribal phenomena of interpolation and omission. From a diachronic point of view, however, the actual adding or subtracting of verses is basically a matter of variation, and we may think of the longer and shorter versions as vertical variants.

Thus the textual phenomena of (1) interpolation/omission and (2) varia lectio are comparable to the oral poetic phenomena of (1) expansion/compression and (2) intralinear formulaic variation. In terms of oral poetics, we can even expect situations where the two textual phenomena are found together. In terms of textual criticism, we actually have an Aristarchean formulation for just such a co-occurrence. Apthorp highlights for me ( per litteras 1/26/99) this Aristarchean editorial criterion, as mediated by Didymus: “an indication of interpolation is the fact that the given line is also transmitted differently” (scholia A to XIX 327: τεκμήριον δὲ τῆς διασκευῆς τὸ καὶ ἑτέρως φέρεσθαι τὸν στίχον, “εἴ που ἔτι ζώει γε Πυρῆς ἐμός, ὃν κατέλειπον”). In terms of textual criticism, I offer this restatement: cases of vertical variation typically coincide with cases of horizontal variation. In terms of oral poetics, the varia lectio at Iliad XIX 327, εἴ που ἔτι ζώει γε Πυρῆς ἐμός, ὃν κατέλειπον, is an intralinear formulaic variant of the vulgate reading εἴ που ἔτι ζώει γε Νεοπτόλεμος θεοειδής.

In this light, I offer a reformulation of Bolling’s formulation concerning Aristarchus’ editorial work on the “Alpha Text” of the Iliad (N 1997b.116n48):

“Aristarchus reduced ca. 15,600 lines [= his received ‘Alpha Text’] to ca. 14,600 [= Bolling’s reconstruction of the prototypical ‘Pi Text’] by systematically adding in the left margin, next to suspect lines in the ‘Alpha Text’ that he used as his point of departure, the editorial sign of the obelus. The obelus marks the judgment of athetesis; when a line is athetized by Aristarchus, it is not deleted from his ‘Alpha Text’: rather, it is simply marked as textually suspect, the result of ‘corrupting’ accretion. The fundamental problem with the methodology of Bolling is that he allows only for expansion, never for compression, in the evolution of Homeric poetry. If we apply the perspective of diachronic studies in oral poetics, Bolling’s assumptions about a tradition that can only add, never subtract, are unjustified.”

Pursuing further the phenomenon of vertical variation, I stress again the fact that Aristarchus not only athetized verses he also deleted (“omitted”) verses. And it was only when a verse was weakly attested in the manuscripts known to Aristarchus that he would actually consider the deletion (“omission”) of such a verse from his edition (cf. again Apthorp 1980.xv).

The distinction between athetized and deleted (“omitted”) is relevant to my resisting West’s theory that it was Didymus, not Aristarchus, who collected and collated the Homer manuscripts that yielded the variant horizontal readings tagged as “novit Didymus” in the apparatus criticus of West’s Iliad. Aristarchus’ procedure concerning deleted (“omitted”) verses, as distinct from athetized verses, was not subjective: it depended on external evidence, and on the application of formal criteria to that external evidence. As we have just seen from my brief review of Apthorp’s findings, Aristarchus systematically contrasted the stronger and weaker attestations of textual variants. In other words, the editorial criteria of Aristarchus must have required an active policy of collecting and collating manuscripts. An editorial practice of distinguishing between athetesis and deletion (“omission”) simply could not work without such a policy. I submit, then, that the cumulative testimony of the scholia concerning Aristarchus’ editing of vertical variations provides additional evidence in favor of rejecting West’s “novit Didymus” theory — and editorial practice.21

What I have just argued about Aristarchus’ editorial policy on vertical variants applies also to his policy on horizontal variants. His evaluation of these variants must have required an active policy of collecting and collating manuscripts. In this case, however, it is more difficult for us to intuit the degree of variation, from manuscript to manuscript, since any discussion of variants emerging from Aristarchus’ study of the manuscript evidence would have been reserved for his hupomnêmata. Such Aristarchean discussions of horizontal variants, abridged by Didymus and further abridged by the Homer scholia, have reached us in a most unsatisfactory state of incompleteness. The testimony of the scholia to I 423-424, which we have considered earlier, is exceptional in showing a relatively less abridged version of Aristarchean commentary. Moreover, the actual forms of the horizontal variants, as mentioned in the Aristarchean hupomnêmata, could easily be ignored by copyists if their primary task was simply to transcribe the ipsissima verba featured in the text proper of the Aristarchean edition. From the standpoint of copyists who had no stake in the complete Aristarchean editorial legacy, hupomnêmata and all, vertical variation would have mattered far more than horizontal variation: there was far greater or less labor involved in the copying of more or fewer verses (cf. Apthorp 1980.9-10).

Throughout this discussion, we have seen how the testimony of the Homer scholia concerning Aristarchus’ editorial policy on vertical variants helps us better understand his corresponding policy on horizontal variants. To summarize both editorial policies, I offer the following template:

1) Vertical variations

1a) Aristarchus kept in his edited text the “vulgate” variant verses that he athetized, while he reserved his judgment of the wrongness of these verses for discussion in his hupomnêmata (marginal marks in the text proper would signal such a discussion).

1b) It was only when the manuscript attestation for a given verse was weak that he could delete (“omit”) it rather than athetize it, perhaps justifying the deletion in his hupomnêmata.

2) Horizontal variations

2a) Aristarchus kept in his edited text the “vulgate” wording, while he reserved his judgment on the rightness or wrongness of corresponding variant wordings for discussion in his hupomnêmata (marginal marks in the text proper would signal such a discussion).

2b) It was only when the manuscript attestation for a given wording was weak that he could delete (“omit”) it and allow the substitution of a variant wording, perhaps justifying the substitution in his hupomnêmata.

In terms of the second part of this posited template, concerning horizontal variations, let us take another look at Iliad IX 394, where West chooses for his own text the Aristarchean reading γε μάσσεται instead of the “vulgate” reading γαμέσσεται. As I have remarked earlier, such a choice is exceptional for West. In most other cases where Aristarchean variants are reported, he tends to choose a variant reading that is non-Aristarchean if the corresponding Aristarchean variant is only weakly or not at all attested in the medieval manuscripts and in the papyri (e.g. III 406). He is likely to opt for an Aristarchean variant only if he finds it strongly attested in the medieval manuscripts and in the papyri.

Here we have an opportunity to compare West’s general editorial stance with that of Aristarchus as I have just outlined it. Ironically, West comes closest to an Aristarchean editorial principle precisely in those situations where he has to make a choice between an Aristarchean and a non-Aristarchean variant. In other words, he resorts to Aristarchean methodology in order to reject Aristarchean variants. When he is forced to choose between an Aristarchean and a non-Aristarchean variant, he tends to rely more than usual on the external evidence — and less than usual on the internal. In other words, in such situations he is more likely to allow the relative strength or weakness of manuscript attestation to influence his choice. If the Aristarchean variant is weakly attested, he tends to reject it. If it is strongly attested, he tends to adopt it. That is because, for West, an Aristarchean variant is like any other variant.

By contrast, if we follow the neo-Aristarcheans, then any variant reported by Aristarchus has a special status, because it may indeed come from ancient manuscripts or manuscript traditions that were known to Aristarchus and are no longer known to us. In those ancient traditions, the relative strength or weakness of manuscript attestations may indeed have been markedly different from what is available to modern editors like West.

I should emphasize that West in these special situations merely tends to make choices on the basis of external evidence. Even here, as elsewhere, he is not systematic in applying the criterion of stronger vs. weaker attestation. Confronted with a choice between a weakly attested Aristarchean variant and a strongly attested non-Aristarchean counterpart, he does not always opt for the second alternative. Again I point to such salient exceptions as his choice of the Aristarchean reading γε μάσσεται instead of the “vulgate” reading γαμέσσεται at IX 394. Exceptions like this help account for the qualification built into the wording of Janko (2000.1): “[West] holds that … the majority of Aristarchus’ unique readings are wrong.”

With regard to choices to be made between Aristarchean and non-Aristarchean variants, Ludwich’s editorial principle is more straightforward than West’s: he generally sides with Aristarchus wherever a choice is available on the basis of available manuscript evidence, that is, wherever Aristarchus’ own choice between variants is backed up by manuscript evidence available to Ludwich. In the apparatus criticus of Ludwich’s Iliad, the most striking illustration of this principle is the ubiquitous formula “Aristarchus + Omega,” that is, where the choice of Aristarchus is backed up by “Omega,” designating a relatively strong attestation in the medieval manuscript tradition. West has adopted Ludwich’s “Aristarchus + Omega” formula in the apparatus criticus of his own Iliad, and we find it applied with the greatest frequency.22 On the surface, then, it seems as if Ludwich and West are applying the same criterion in such situations. Underneath the surface, however, the criteria differ: for Ludwich, the “Aristarchus” component of the “Aristarchus + Omega” formula has special status, but for West it has merely equal status. Correspondingly, whenever the “Omega” drops out, that is, whenever the manuscript support is lacking or weak, the Aristarchean variant tends to be kept by Ludwich but dropped by West. To repeat, West’s adoption of γε μάσσεται instead of γαμέσσεται at Iliad IX 394 is for him exceptional.

From the standpoint of oral poetics, I should emphasize, such choices do not have to be made. In the case of the two variants γε μάσσεται and γαμέσσεται, for example, both actually fit the formulaic system of Homeric diction, as I have argued elsewhere (N 1998a). In other words, the internal evidence indicates that both forms are functional variants in the formulaic system of Homeric poetry. The external evidence gives further support, if we agree with Ludwich that γε μάσσεται is a variant that Aristarchus had authenticated on the basis of manuscript evidence known to him, just as γαμέσσεται is a variant authenticated on the basis of manuscript evidence known to us.

Although West has a generally negative policy about Aristarchus’ reliability concerning horizontal variants, he is more positive about the vertical variants. He tends to approve whenever Aristarchus athetizes or deletes (“omits”) verses on the grounds that they are interpolations (as in the scholia to III 144a; cf. West 1999a.186-187). Such a double standard toward vertical and horizontal variations as mediated by Aristarchus is also evident, as I have noted earlier, in the work of Janko (1992.21n6; cf. N 1996a.139-140).23 For West’s Iliad, in any case, the final judgment on both vertical and horizontal variations does not depend on the external evidence of manuscript attestations. Whatever this evidence may have been in the time of Aristarchus, and whatever it is now as we contemplate the texts and quotations that do survive into our time, the ultimate criterion for West is whether a given variant does or does not fit the model built on his assumption that our Iliad and Odyssey go back to “originals” written around the middle of the seventh century BCE. I have already quoted him as saying: “We may assume that there existed a complete and coherent Urtext of each epic, the result of the first writing down” (West 1998a.95). But can we actually restore this Urtext ? West’s answer is no: “we are not in the position to restore the original version” (ibid.). He goes on to ask: “Should we then be aiming to recover the Athenian text of the epics, as they were recited at the Panathenaea after Hipparchus, or at any rate as they were known to Thucydides, the orators, and Plato?” (ibid.). The answer, again, is no: “the text at this period was in a far from settled state” (ibid.). West explains that “there was no single ‘Athenian text'” (ibid.). And “the text” becomes only more unsettled thereafter, in the era of the early Ptolemaic papyri: now it is “characterized by so-called ‘wild’ variants (think of wild flowers rather than wild poets, perhaps), diverging from the medieval vulgate not in narrative substance but with substitution of formulae, inorganic additional lines, and so forth” (ibid.). So, finally, we have come all the way down to the Alexandrian editors. West now asks the inevitable follow-up question: “Should our aim, then, be to reconstruct the Alexandrian text?” (ibid.). One last time, the answer is no: “But again, there is no single Alexandrian text, even if some agreement was reached in terms of Versbestand” (ibid.). West explains: “We know of many variant readings current in the Alexandrian period, from which we must make our choice” (ibid.).

So it seems that there has never been any single text, not since the Urtext. In face of all this multiformity sketched by West in his survey of Homeric textual history, we may well ask: how and even why must we “make our choice”? What kind of uniformity should an editor try to reconstruct? West’s answer is guarded (1998a.95):

“The answer must be some kind of compromise. Let us state our aim to be the establishment, so far as our means allow, of the pristine text of the poems in the form they attained following the last phase of creative effort. We must concede that, as the tradition passed through several centuries of ‘wildness’, it may be impossible to establish exactly what lies on the far side. But let that be our objective.”

West is eager to cross the fence and move on to that far side, where the wild flowers grow — and far beyond that, far beyond the horizon. He leaves behind whatever we can still see on this side of the fence, the ruins of gardens once teeming with cultivated flowers tended by the Alexandrian editors. As we have already noted, West has not much use for Aristarchus, and even less for Aristophanes of Byzantium; as for Zenodotus of Ephesus, that Alexandrian editor is practically of no use at all ( Praefatio pp. vi-vii, though there are a few exceptions listed at n6).24 He concludes by asking: “Why limit our task to reconstructing an Alexandrian text, stopping five hundred years short of the originals, when we have at least a modicum of evidence that takes us further back?”

The problem is, trying to go further back to “the originals,” to a single text, is an impossible task, when all you have is a “modicum of evidence.” To wander off on such a quest is to leave untended the evidence that is closer at hand, available from the Alexandrians. In terms of West’s own reconstruction, as we have seen, there is no evidence for a “single text.” For him, the last time there was any single text was the time of the hypothetical Urtext. Similarly for Janko, we need to go back to an Urtext — this time, by way of a prototypical dictation, in terms of his own theory.

As we come back to the realities of multiformity described by West in his sketch of Homeric textual history, I must repeat my question: what kind of uniformity should an editor try to reconstruct? The answer, I submit, is that the evidence of textual multiformity precludes a uniform reconstruction, a “unitext” edition of Homer. Instead, the editor of Homer needs to keep coming back, I submit, to the facts of textual multiformity. The basic source for these facts, however incompletely preserved they may be, remains Aristarchus. Revenons à cultiver notre jardin.

We cannot afford to lose sight of the facts known to Aristarchus and no longer known (or known as well) to us, even if we think that his editorial judgment was impaired because he in turn did not know some other facts that we do indeed know. In comparing our own cumulative knowledge with that of Aristarchus, it is tempting for some to feel superior to him, as we see from the explanation given by Janko for his general disapproval of Aristarchean atheteses (1992.27): “The ethical and probabilistic criteria he [Aristarchus] applies are not those of Homer’s society; his knowledge of epic usage is less complete than ours (based on sophisticated indices and concordances); he was unaware of Indo-European and Near Eastern philology, archaeology, oral poetry, ring-composition and Linear B; and, as for literary insight, he is often outshone by the later scholarship seen in [scholia] bT.” So also in the Iliad of West, as we have seen, the editorial judgments of Aristarchus are often overruled on the basis of considerations derived from current Homeric scholarship, much as they are being described by Janko. Still, there remain three fundamental facts that remain unaffected by such considerations.

One, Aristarchus knew far more about the ancient manuscript evidence than we do.

Two, on the basis of all his manuscript evidence, Aristarchus could not and would not produce a single unified text of Homer. He left room for choices among variants — both horizontal and vertical — in his hupomnêmata.

Three, on the basis of all the manuscript evidence available to us, which is less extensive than the manuscript evidence available to Aristarchus, editors like West are forced to admit that they cannot produce a single unified text of Homer. The unity of their “unitext” editions is achieved by way of reconstructions and conjectures based on considerations of chronology, dialect, historical provenance, and so forth.

In light of these three facts, I propose an alternative to the concept of a “unitext” edition of Homer. Instead, I advocate the concept of a multitext edition (cf. Bird 1994). Such an edition needs to account for Homeric multiforms attested as textual variants, recovered mostly through the research of the Alexandrian editors, especially Aristarchus.

As I have argued extensively elsewhere, the textual multiformity of the Homeric poems indicates a heritage of oral poetics. Although the Alexandrians did not think in terms of oral poetics, their editorial work on textual variants provides evidence of this heritage (N 1996a.151-152):

“Even though Aristarchus, following the thought-patterns of myth, posited a Homeric original, he nevertheless accepted and in fact respected the reality of textual variants. He respected variants because, in terms of his own working theory, it seems that any one of them could have been the very one that Homer wrote. … That is why he makes the effort of knowing the many different readings of so many manuscripts. He is in fact far more cautious in methodology than some contemporary investigators of Homer who may be quicker to say which is the right reading and which are the wrong ones. Aristarchus may strike us as naive in reconstructing an Athenian Homer who ‘wrote’ around 1000 [BCE], but that kind of construct enables him to be more rigorous in making choices among variants. … What, then, would Aristarchus have lost, and what would we stand to lose, if it really is true that the variants of Homeric textual tradition reflect for the most part the multiforms of a performance tradition? If you accept the reality of multiforms, you forfeit the elusive certainty of finding the original composition of Homer but you gain, and I think this is an important gain, another certainty, an unexpected one but one that may turn out to be much more valuable: you recover a significant portion of the Homeric repertoire. In addition, you recover a sense of the diachrony.”

The need for a diachronic perspective in analyzing Homeric poetry has led to my developing an evolutionary model for the making of this poetry (N 1996b.30-112). This model was designed to account for all variations that stem from the performance traditions of Homeric poetry. The multiformity of variations in the oral poetic context of composition-in-performance cannot be viewed exclusively from a synchronic perspective (N 1999a). A multitext edition of Homer could provide the needed diachronic perspective on this multiformity.

A multitext edition of Homer needs to be designed in a format that displays most clearly all the surviving textual variants, both vertical and horizontal. It should have a base text ( texte à base) that is free of arbitrary judgments, such as the choosing of one variant over another on the basis of the editor’s personal sense of what is right or wrong, better or worse. In other words, the base text needs to be formatted to show all locations where variants are attested, and all the variants that can be slotted into those locations — without privileging any of these variants. In the format of ” hupomnêmata,” the editor of the base text may then proceed to analyze the variants from a diachronic perspective, making his or her own considered judgments about differences in the chronology, dialect, historical provenance, and so forth. For such a multitext edition, the most convenient base text would be the relatively most standard and common manuscript tradition. For Aristarchus, that base text was essentially the koinê version of Homer — what neo-Aristarcheans call the “vulgate” (N 1997b.118-122).25 For us, the closest thing to such a base text is the Homer of van Thiel (1991, 1996). Something much closer to an ideal, however, would be the edition of Aristarchus, if only it had survived. In fact, Aristarchus’ edition of Homer would have been the closest thing to what I am describing here as an ideal multitext edition.

If van Thiel’s Homer were chosen as the most convenient base text for a multitext edition, where could we go from there? What models could we find for the ” hupomnêmata” to accompany such a base text? The apparatus criticus of West’s Homer seems a most elegant model, at least as a starting point. But there are problems, as we see from the interplay of his apparatus criticus with his testimonia, that is, his display of the ancient quotations of Homer.

The testimonia take up a separate middle band on each page of West’s Iliad (his edited text is of course in the upper band, while his apparatus criticus is in the lower band). This middle band turns out to be of limited value to the editor — let alone the editor’s readers. After all, West thinks that most ancient quotations are of limited value in the first place (p. x). In his 1898 book, Die Homervulgata als voralexandrinisch erwiesen, Ludwich had already published an exhaustive collection of ancient quotations, down to the Augustan period. Ludwich’s mass of cross-references, now further supplemented, has been assimilated by West into the middle band of his Iliad (see again West 1998a.97). Ludwich’s original motive — and premise — in making his 1898 collection had been simple: he thought that these ancient quotations would help further recover the pre-Alexandrian vulgate. Since West does not operate on this premise, his updating of Ludwich’s repertoire merely exemplifies the law of diminishing returns. The ubiquitous references to “t” and “tt” (“testimonium” and “testimonia”) in the lower band containing the apparatus criticus, drawing our attention upward into the middle band containing the testimonia, hardly make a difference whenever it comes to the question of what reading is finally chosen by West in the upper band containing the edited text proper. Typically, “t” or “tt” simply join the chorus of “Omega” readings in privileging the vulgate version of Homer. For West, they seldom make a difference (he lists a handful of exceptions at p. x n14; see also West 1998a.97). By contrast, wherever an ancient quotation does make a difference in Ludwich’s Iliad, that editor simply confronts it directly in his apparatus criticus (his Iliad does not feature a separate middle band of testimonia, which he published separately, as we have seen, in 1898). In this respect, Ludwich’s policy seems more economical — and elegant.

Although the “lower band” of West’s three-band Iliad, his apparatus criticus, is a most valuable contribution to our knowledge of the Homeric textual tradition, the “upper band,” his base text, is of questionable value. The text of West’s Iliad contains many editorial judgments that go beyond the manuscript evidence and that flatten out the textual history of Homer. Here is where his “unitext” edition differs most radically from a multitext edition.

A salient example of differences between “unitext” and multitext editorial approaches to Homer is the case of Athenian forms and themes that we find embedded in the Homeric textual tradition. From a diachronic point of view, we may expect a multitext edition of Homer to accommodate even those textual variants that are dated as late as the sixth through the fourth centuries BCE, the era of Athenian performance traditions (N 1999b.271-272). In fact, there are numerous Homeric textual variants that reflect Athenian accretions in both form and content (West p. vi; cf. N 1998a). For a “unitext” Homer edition like West’s, however, the goal is to reconstruct the prototypical text of the prototypical poet. Accordingly, for all practical purposes, West’s Iliad has screened out all the textual variants of Athenian provenance. In the case of horizontal variants, West consistently chooses the non-Athenian form over the corresponding Athenian one (Ionic τεσσερ – vs. “Attic” τεσσαρ -, passim; cf. Praefatio p. xxxv). Wherever an Athenian form occurs without the attestation of a non-Athenian textual variant, he will simply emend the transmitted Athenian form and conjecture a corresponding non-Athenian form. In the case of vertical variants, he will bracket or simply omit from his text any line that suggests any kind of Athenian cultural agenda stemming from the sixth century or thereafter (a prime example is II 547-551; cf. in general West 1999a).

At one extreme, then, West’s edition tends to underestimate the chronological diversity of the Homeric tradition. At the other extreme, however, it occasionally lapses into overestimations that defy credibility. In particular, the relatively older forms tend to be treated without sufficient regard for the operative system of Homeric diction. For example, West (p. xxix) claims that the Luvian particle tar (which he translates as ‘usquam’) is a cognate of the Homeric particle ταρ, and that the Luvian usage of tar is key to understanding the Greek usage of TAR in the Iliad, as at I 8, 65, 123, and so forth.26 In printing ταρ instead of τἄρ in these contexts, West is following the testimony of Apollonius and Herodian, according to whom the particle ταρ counts as one word, by contrast with τἄρ. The particle τἄρ counts as two words, τ’ (from τε) plus ἄρ. If Homeric TAR is really cognate with Luvian tar, then it cannot be related etymologically to Homeric τἄρ. And yet, I see no convincing way of separating completely the Homeric contexts of this ταρ from Homeric contexts of τἄρ where the constituents τ’ (from τε) plus ἄρ are transparently functional, as at I 93, III 398, and so forth. Moreover, the etymology of γάρ, another particle that counts as one word, can serve as evidence against the etymological separation of ταρ from τἄρ : in the case of γάρ, the constituents are transparently γ’ (from γε) plus ἄρ (Chantraine DELG p. 210). Just as some usages of γάρ preserve the syntax of the constituent γε (as in the case of O( γάρ via ὅγε plus ἄρ : see Schywzer/Debrunner 1966.560), so also some usages of τἄρ preserve the syntax of the constituent τε (as in the case of τἄρ at III 398, where West’s apparatus shows a syntactically analogous textual variant D’ ἄρ actually attested in the papyri). Further, just as other usages of γάρ no longer preserve the syntax of the constituent γε, so also it is possible that other usages of τἄρ no longer preserve the syntax of the constituent τε; whence the special status of Homeric ταρ.27

Here and elsewhere, there are problems with West’s application of linguistics in the process of rewriting the received text of the Iliad.28

Concluding, I should stress that my criticisms of West’s Iliad are not meant to detract from the praise he deserves. I gratefully acknowledge all his valuable contributions to Homeric scholarship. Even the size of this review can be taken as a tribute to the editor’s industry and learning. West’s Iliad is a most useful and important edition. Still, it cannot be considered an authoritative replacement of previous editions of the Iliad. Many of those — including the 1902 Teubner text of Ludwich and the 1920 OCT of Monro/Allen — will continue to have their uses.29

Bibliography

Allen, T. W. 1924. Homer: The Origins and the Transmission. Oxford.

—-, ed. 1931. Homeri Ilias I-III. Oxford.

Apthorp, M. J. 1980. The Manuscript Evidence for Interpolation in Homer. Heidelberg.

—-. 1990a. “Some Neglected Papyrus Evidence Against the Authenticity of Iliad 16.381.” ZPE 81:1-7.

—-. 1990b. “Papyrus Evidence in Favour of Some Suspected Lines in Homer.” ZPE 82:13-24.

—-. 1992. ” Nochmals the Authenticity of Odyssey 10.475-9.” Classical Quarterly 42: 270-271.

—-. 1993. Review of van Thiel 1991. Classical Review 43:228-230.

—-. 1995a. Review of R. D. Dawe 1993. The Odyssey: Translation and Analysis (Lewes: The Book Guild). Classical Review 45:1-2. N.B. p. 2, on Dawe’s ignoring “the vital distinction between omission and athetesis.”

—-. 1995b. Review of J. R. Tebben 1994. Concordantia Homerica, Pars I: Odyssea. A Complete Concordance to the Van Thiel Edition of Homer’s Odyssey 2 vols. (Zurich and New York). Classical Review 45:221-222.

—-. 1995c. “Did Homer Give his Nereids Names? A Note on the Ancient Manuscript Evidence.” Acta Classica 38:89-92.

—-. 1995d. ” Iliad 14.306c Discovered in the Syriac Palimpsest.” ZPE 109:174-176.

—-. 1996a. “New Evidence from the Syriac Palimpsest on the Numerus Versuum of the Iliad.” ZPE 110:103-114.

—-. 1996b. ” Iliad 18.200-201: Genuine or Interpolated?” ZPE 111:141-148.

—-. 1998. “Double News from Antinoopolis on Phoenix’s Parricidal Thoughts ( Iliad 9.458-61).” ZPE 122:182-188.

—-. 1999. “Homer’s Winged Words and the Papyri: Some Questions of Authenticity.” ZPE 128:15-22.

Bird, G. D. 1994. “The Textual Criticism of an Oral Homer.” In Nile, Ilissos and Tiber: Essays in honour of Walter Kirkpatrick Lacey (ed. V. J. Gray), Prudentia 26:35-52.

Bolling, G. M. 1925. The External Evidence for Interpolation in Homer. Oxford.

—-, ed. 1950. Ilias Atheniensium: The Athenian Iliad of the Sixth Century B.C. Lancaster PA.

Broggiato, M. 1998. “Cratete di Mallo negli scholl. A ad Il. 24.282 e ad Il. 9.169a.” Seminari Romani di Cultura Greca 1:137-143.

CEG. See Hansen.

Chantraine, P. 1968, 1970, 1975, 1977, 1980. Dictionnaire étymologique de la langue grecque I, II, III, IV-1, IV-2. Reissued 1999, with a Supplement (eds. A. Blanc, Ch. de Lamberterie, J.-L. Perpillou). Paris.

Citti, V. 1966. “Le edizioni omeriche ‘delle città’,” Vichiana 3:227-267.

Cook, B. F. 1984. The Elgin Marbles. 2nd ed. 1997. British Museum Press. London.

DELG. See Chantraine.

Dué, C. 2000a. “Poetry and the Dêmos : State Regulation of a Civic Possession.” Stoa Consortium, ed. R. Scaife, http://www.stoa.org/demos/camws-casey.html.

—-. 2000b. “Achilles’ Golden Amphora and the Afterlife of Oral Tradition in Aeschines’ Against Timarchus.Classical Philology, forthcoming.

Erbse, H. 1959. “Über Aristarchs Iliasausgaben,” Hermes 87:275-303.

—-, ed.. 1969-1988. Scholia Graeca in Homeri Iliadem I-VII. Berlin and New York.

Foley, J. M. 1999. Homer’s Traditional Art. University Park PA.

Hansen, P. A., ed. 1983. Carmina epigraphica Graeca saecularum viii-v a. Chr. n. Berlin and New York.

Haslam, M. 1997. “Homeric Papyri and Transmission of the Text.” Morris and Powell 1997:55-100.

Janko, R. 1982. Homer, Hesiod and the Hymns: Diachronic Development in Epic Diction. Cambridge.

—-. 1990. “The Iliad and its Editors: Dictation and Redaction,” Classical Antiquity 9:326-334.

—-. 1992. The Iliad: A Commentary IV (general ed. G. S. Kirk). Cambridge.

—-. 1998a. Review of Morris and Powell 1997. Bryn Mawr Classical Review 98.5.20.

—-. 1998b. “The Homeric Poems as Oral Dictated Texts.” Classical Quarterly 48:1-13.

—-. 1998c. Review of Nagy 1996a. Journal of Hellenic Studies 118:206-207.

—-. 1998d. “Corrigendum.” Bryn Mawr Classical Review 98.6.17.

—-. 2000. “West’s Iliad.” Classical Review 50:1-4.

Katz, J. 1998. ” Αὐτάρ, ἀτάρ, ταρ : The Poetics of a Particle in Homer.” American Philological Association Abstracts 128:81.

Keaney, J. J., and Lamberton, R., eds. [Plutarch] Essay on the Life and Poetry of Homer. APA American Classical Studies 20. Atlanta.

Knight, R. Payne, ed. 1820. Carmina Homerica, Ilias et Odyssea: a rhapsodorum interpolationibus repurgata, et in pristinam formam, quatenus recuperanda esset, tam e veterum monumentorum fide et auctoritate, quam ex antiqui sermonis indole ac ratione, redacta. London.

Leeuwen, J. van, and Mendes da Costa, M. B., eds. 1906. Ilias. 3rd ed. Leiden. Lehrs, K. 1882. De Aristarchi Studiis Homericis. 3rd ed. Leipzig; 1st / 2nd eds. 1833 / 1865.

Lord, A. B. 1960. The Singer of Tales. Harvard Studies in Comparative Literature 24. Second edition 2000, with Introduction (vii-xxix), by S. Mitchell and G. Nagy. Cambridge MA.

—-. 1991. Epic Singers and Oral Tradition. Ithaca.

—-. 1995. The Singer Resumes the Tale (ed. M. L. Lord). Ithaca.

Ludwich, A., ed. 1884. Didymi commentarii qui inscribebatur Περὶ τῆς Ἀρισταρχείου διορθώσεως fragmenta. = Ludwich 1884.175-631.

—-. 1884 / 1885. Aristarchs Homerische Textkritik nach den Fragmenten des Didymos I / II. Leipzig.

—-. 1898. Die Homervulgata als voralexandrinisch erwiesen. Leipzig.

—-, ed. 1902. Homeri Ilias. I / II. Leipzig. Reissued 1995, Stuttgart.

Lührs, D. 1992. Untersuchungen zu den Athetesen Aristarchs in der Ilias und zu ihrer Behandlung im Corpus der exegetischen Scholien. Beiträge zur Altertumswissenschaft 11; Hildesheim.

MHV. See Parry 1971.

Monro, D. B., and Allen, T. W., eds. 1920. Homeri Opera ed. 3 Oxford.

Montanari, F. 1998. “Zenodotus, Aristarchus and the Ekdosis of Homer. In Most 1998:1-21.

Morris, I., and Powell, B., eds. 1997. A New Companion to Homer. Leiden.

Most, G. W., ed. 1998. Editing Texts / Texte edieren. Aporemata II. Göttingen.

N = Nagy.

Nagy, G. 1996a. Poetry as Performance: Homer and Beyond. Cambridge.

—-. 1996b. Homeric Questions. Austin.

—-. 1997a. “An inventory of debatable assumptions about a Homeric question.” Bryn Mawr Classical Review 97.4.18.

—-. 1997b. “Homeric Scholia.” In Morris and Powell 1997:101-122. This piece is replete with printing errors. For a list of corrigenda, please write the author (gnagy@fas.harvard.edu).

—-. 1998a. “Aristarchean Questions.” Bryn Mawr Classical Review 98.7.14.

—-. 1998b. “The Library of Pergamon as a Classical Model.” Pergamon: Citadel of the Gods (ed. H. Koester) 185-232. Harvard Theological Studies 46.

—-. 1999a. “Homer and Plato at the Panathenaia: Synchronic and Diachronic Perspectives.” Contextualizing Classics (eds. T. M. Falkner, D. Konstan, N. Felson Rubin) 127-155. Lanham MD.

—-. 1999b. “Irreversible Mistakes and Homeric Poetry.” Euphrosyne: Studies in Ancient Epic and its Legacy in Honor of Dimitris N. Maronitis (ed. J. N. Kazazis and A. Rengakos) 259-274. Stuttgart.

—-. 1999c. “Les éditions alexandrines d’Homère.” Homère en France après la Querelle (1715-1900) (ed. F. Létoublon and C. Volpilhac-Auger) 63-72.

—-. 2000. “Epic as Music: Rhapsodic Models of Homer in Plato’s Timaeus and Critias.” The Oral Epic: Performance and Music (ed. K. Reichl) 41-67. Berlin.

Nickau, K. 1977. Untersuchungen zur textkritischen Methode des Zenodotos von Ephesos. Berlin and New York.

Parry, A., ed. 1971. The Making of Homeric Verse: The Collected Papers of Milman Parry. Oxford.

Parry, M. See MHV.

Pfeiffer, R. 1968. History of Classical Scholarship from the Beginnings to the End of the Hellenistic Age. Oxford.

Pierron, P. A., ed. 1869. L’Iliade d’Homère : texte grec, revu et corrigé d’après les documents authentiques de la recension d’Aristarque. 2 vols. Paris.

Porter, J. I. 1992. “Hermeneutic Lines and Circles: Aristarchus and Crates on the Exegesis of Homer,” Homer’s Ancient Readers: The Hermeneutics of Greek Epic’s Earliest Exegetes (ed. R. Lamberton and J. J. Keaney) 67-114. Princeton

Rengakos, A. 1993. Der Homertext und die hellenistischen Dichter. Hermes Einzelschriften 64. Stuttgart.

Schmidt, M. 1997. “Variae lectiones oder Parallelstellen: Was notierten Zenodot und Aristarch zu Homer?” ZPE 115:1-12.

Schwyzer, E., and Debrunner, A. 1966. Griechische Grammatik II: Syntax und syntaktische Stilistik. Munich.

Sherratt, E. S. 1990. “‘Reading the Texts’: Archaeology and the Homeric Question.” Antiquity 64:807-824.

van Thiel, H., ed. 1991/1996. Homeri Odyssea/Homeri Ilias. Zürich and New York.

van der Valk, M. 1963/1964. Researches on the Text and Scholia of the Iliad I/II. Leiden.

Villoison, J. B. G. d’Ansse de, ed. 1788. Homeri Ilias ad veteris codicis Veneti fidem recensita. Venice.

Watkins, C. 1995. How to Kill a Dragon: Aspects of Indo-European Poetics. New York.

West, M. L. 1990. “Archaische Heldendichtung: Singen und Schreiben.” Der Übergang von der Mündlichkeit zur Literatur bei den Griechen (ed. W. Kullmann and M. Reichel) 33-50. Tübingen.

—-. 1998a. “The Textual Criticism and Editing of Homer.” In Most 1998:94-109.

—-, ed. 1998b. Homeri Ilias I. Stuttgart and Leipzig.

—-. 1999a. “Frühe Interpolationen in der Ilias.” Nachrichten der Akademie der Wissenschaften in Göttingen I. Philologische-Historische Klasse 3-11. Göttingen.

—-. 1999b. “The Invention of Homer.” Classical Quarterly 49:364-382.

Wilson, N. 1990. “Thomas William Allen, 1862-1950.” Proceedings of the British Academy 76:311-319.

Wolf, F. A. 1795. Prolegomena ad Homerum, sive de operum Homericorum prisca et genuina forma variisque mutationibus et probabili ratione emendandi. Halle.

—-, ed. 1804/1807. Homerou epe. Homeri et Homeridarum opera et reliquiae. 4 volumes. Leipzig.

Wyrick, J. D. 1999. “The Genesis of Authorship: Legends of the Textualization of Homeric Epic and the Bible.” Ph. D. dissertation, Harvard University.

Notes

1. Wolf’s Iliad and Odyssey editions are not nearly as well known as the separately published Prolegomena : cf. Janko 1998d (= BMCR 98.6.17) correcting Janko 1998a (= BMCR 98.5.20), where he had denied the existence of these editions in the context of attempting to correct what I had to say (Nagy 1997b) about the influence of Wolf’s editions on later Homer editions like the Oxford Classical Text of Monro and Allen 1920.

2. On the concept of Aristarchus’ “edition” of Homer, as expressed by words like diorthôsis and ekdosis, I find most useful the summary of Montanari 1998.11-20.

3. Further references in N 1996a.151; cf. Pfeiffer 1968.228; also Keaney and Lamberton 1996.67n2. On the rivalry of Aristarchus and Crates as editors of Homer, see below.

4. The relevant passage from Josephus, Against Apion 1.12-13, needs to be interpreted in its own right, aside from the interpretations of Villoison and Wolf; as far as Josephus himself is concerned, the point he is making is that the ancient Greeks, unlike the Jews, did not have a continuous written tradition going all the way back to a prototypical authorship (see Wyrick 1999). Josephus leaves it open whether there had been an “original” Homeric text, and his wording points to a variation on the theme of the Peisistratean Recension, which I view as a cultural construct based on the mythical idea of a prototypical text that had disintegrated on account of neglect, only to be reintegrated later by an enlightened ruler (N 1998b.227-228; in this discussion I include references, with bibliography, to the alternative view that the Recension was a historical event).

5. Monro and Allen 1920. For a defense of Allen’s overall editorial work, which has been severely criticized by e.g. Wilson 1990, see Haslam 1997.89-91.

6. A typical emendation by Payne Knight, accepted by West, is τεόν for τὸ σόν (mss) at I 185, on the analogy of τεόν at I 138; also at I 207 (in this case, both τεόν and τὸ σόν are attested as manuscript variants). The problem is, the syntax of Homeric τό etc. as article — even though it is less archaic than the syntax of τό etc. as a demonstrative pronoun — pervades the formulaic system of Homeric poetry and thus cannot simply be eliminated everywhere by emendation (on the “article” in Homeric diction, see Schwyzer-Debrunner 1966.20-22). In general, Homeric diction is linguistically multi-layered, allowing older and newer phenomena to coexist (for example, the older forms τοί and ταί function only as demonstrative pronouns, while the newer forms οἱ and αἱ function either as articles or, by default, as demonstrative pronouns). Thus the sum total of phonological, morphological, and syntactical functions in the formulaic system of Homeric poetry cannot simultaneously be reduced to their oldest phase. Such is the problem with the efforts of Payne Knight (and West) in emending Homeric phrases into a state of conformity with a pre-existing article-free phase of their existence. This kind of effort reveals a general sense of dissatisfaction with the anomalies of existing systems — a dissatisfaction so strong that it verges on an impulse to recover what is irrecoverable. Such an impulse can lead to an overextended idea about what is genuine — and what cannot be so. Payne Knight is the man who told Elgin about the Acropolis Marbles: “You have lost your labour, my Lord Elgin. Your marbles are overrated: they are not Greek: they are Roman of the time of Hadrian” (Cook 1984.80).

7. On Pierron and Ludwich, see Ludwich 1885.82, 91, 168; for more on the Königsberg School, see Ludwich 1885.199.

8. For a persuasive reconstruction of the history of Aristarchus’ Homeric hupomnêmata, see Montanari 1998.11-20. First, Aristarchus produced a set of hupomnêmata that were keyed to the Homer edition of his predecessor, Aristophanes of Byzantium: cf. the wording in the scholia to II 133a, ἐν τοῖς κατ’ Ἀριστοφάνην ὑπομνήμασιν Ἀριστάρχου. Then, after he published (= made an ekdosis of) his own edition ( diorthôsis) of Homer, he produced a new set of hupomnêmata that were keyed to this new edition. Relevant are the titles of two monographs produced by Aristarchus’ direct successor, Ammonius: (1) Περὶ τοῦ μὴ γεγονέναι πλείονας ἐκδόσεις τῆς Ἀρισταρχείου διορθώσεως and (2) Περὶ τῆς ἐπεκδοθείσης διορθώσεως. Montanari argues persuasively that the two titles are not mutually contradictory. The ekdosis or “edition” of Aristarchus can be viewed as an ongoing process of diorthôsis or “editorial work.” Such is the idea behind the reference to a single ekdosis or “edition” of Homer rather than a multiplicity of editions, as expressed in one of the two titles of Ammonius; the same idea accounts for the reference to a “re-edition,” as expressed by the wording ἐπεκδοθείσης in the other title. It seems that Ammonius preferred to think of two distinct phases in the history of Aristarchus’ editing of Homer, while others thought — wrongly, according to Ammonius — that there were two distinct editions. Perhaps the first and the second phases of Aristarchus’ diorthôsis (= D1 and D2) may be correlated respectively with the first and the second sets of hupomnêmata (= H1 and H2). According to Montanari’s schema, the sequence would be H1 D1 H2 D2. But I am not sure that we need to infer, as Montanari does (p. 19), that the contents of D2 were written into the same “copy” that contained the contents of D1. In the Homeric scholia, we find ubiquitous references to sources labeled αἱ Ἀριστάρχου and ἡ ἑτέρα τῶν Ἀριστάρχου (the wording ἑτέρα makes it explicit that there were two), and it may be easiest to think of D1 and D2 in the sense of first and second copies. In West’s apparatus, “Ar ab” is the equivalent of αἱ Ἀριστάρχου, while “Ar a” or“Ar b” are equated with ἡ ἑτέρα τῶν Ἀριστάρχου — without specification of which one is which.

9. As Montanari (1998.10) argues with reference to the procedures of Aristarchus, the margin of his edited text would not have been suitable for displaying variants; it would have been suitable only for critical signs that refer to the relevant discussion in the hupomnêmata. To put it another way: when it comes to the transmission of editorial judgments, marginal notes would not be as useful as marginal signs that refer to an authoritative discussion in the hupomnêmata.

10. Here are other survivals of quotations of Aristarchus by Didymus, introduced by the tag ” lexis of Aristarchus”: scholia to II 125, Ἀριστάρχου λέξεις ἐκ τῶν ὑπομνημάτων…; II 245, οὕτως αἱ Ἀριστάρχου λέξεις ἐκ τοῦ Β τῆς Ἰλιάδος…; III 406 (already cited above), προσθήσειν μοι δοκῶ καὶ τὴν Ἀριστάρχου λέξιν οὕτως ἔχουσαν…; also II 111, ἔν τινι τῶν ἠκριβωμένων ὑπομνημάτων γράφει ταῦτα κατὰ λέξιν… (again, a quotation from Aristarchus follows; earlier in this context, we read in the discussion of Didymus: παρ’ ὃ δὴ καὶ κατά τινα τῶν ὑπομνημάτων μετειλῆφθαι τὸ μέγα ἀντὶ τοῦ μεγάλως. τὸ δὲ οὐκ ἔχει τἀκριβὲς οὕτως. εἰ γὰρ τὰ συγγράμματα τῶν ὑπομνημάτων προτάτ[τ]οιμεν, ἕνεκα γοῦν τἀκριβοῦς γράφοιμεν κατὰ Ἀρίσταρχον…). In this last example, the expression γράφει ταῦτα κατὰ λέξιν implies that Aristarchus is “speaking” (hence κατὰ λέξιν = verbatim) the exegetical remarks that are about to be quoted, but he is also “writing” (hence γράφει) these remarks in the sense that they are now being read in the quotation. We could paraphrase in English: “he writes, and I quote from what he says.” My point is, the expression κατὰ λέξιν insists on the idea of ipse dixit.

11. See Ludwich 1884.194-196; cf. also Erbse I p. 119 footnote.

12. The Homer edition of Antimachus seems to have been a specially prized possession of the Library of Alexandria, treated as a counterweight to the Homer edition of Euripides, a prized possession of the Library of Pergamon: see Bolling 1925.38-39.

13. I focus here only on Bolling’s reconstruction of the length of this hypothetical “Pi Text”; his reconstruction of the actual diction of the text seems to me far less useful.

14. I use the term “delete” instead of “omit” in order to reinforce Apthorp’s emphasis on the deliberateness of the editorial cancellation of such plus verses. See also Ludwich 1885.132-143 (“Athetierte und ausgestossene Verse”), especially p. 142 on the remark ἔνιοι ὑποτάσσουσι [followed by the quotation of a plus verse] in the scholia to Iliad IX 140 and 159. West’s comment on both these lines, “add. quidam ante Ar” (cf. Monro-Allen, “quidam ant.”), does not convey the editorial implications of the expression ἔνιοι ὑποτάσσουσι.

15. Unlike predecessors like Bolling, Apthorp avoids using the term “plus verse” with reference to verses athetized but not deleted (“omitted”) by the Alexandrian editors.

16. For more on “plus verses,” see especially Apthorp 1998 on Iliad IX 458-461, with new papyrus evidence for deletion (“omission”); cf. my discussion N 1996a.139n135, where “a papyrus” should have been specified as P.Ant. III.158. Apthorp adduces P.Ant. III.158, P.Ant. III.160, and the Leiden glossary. West’s apparatus does not mention P.Ant. III.160; but it adduces a new Oxyrhynchus papyrus, “1139” (West’s apparatus indicates that the plus verses are omitted here as well).

17. Ludwich 1885.192-199 (“Aristarch und die Vulgata”).

18. For more on the term “Wolfian vulgate,” see N 1997b.118-122.

19. When Parry says “unique” in this context, I interpret it to mean “uniquely characteristic of oral poetics.” For a fuller application of this quotation, with more context, see Nagy 1998a. For another application of this same quotation, see the important discussion of Apthorp 1980.110n67.

20. On both points, Janko refers to West p. vii. In a separate work, I hope to dispute West’s general dismissal of Zenodotus, as he phrases it on this page. On the value of variant readings that go back to Zenodotus, in terms of oral poetics, see N 1996a.133-138; see also my p. 144n160 on Aristophanes of Byzantium (with a discussion of the numerus versuum of Aristophanes’ Homer edition, on which see also Apthorp 1980.3).

21. This is not to say that Didymus did not collate Homer manuscripts in his own right or that Aristarchus was the only collator (cf. Bolling 1925.39 for a post-Aristarchean dating of the Krêtikê). It is only to say that the primary collator of Homer manuscripts was Aristarchus himself and that Didymus may not have had access to all the sources still available to Aristarchus (cf. again Bolling, ibid.).

22. I should note one important difference between West’s and Ludwich’s usage: for West (see his p. xiii), but not for Ludwich, readings from “Omega” are regularly to be contrasted with readings from “Z” (=Ve1 Allen).

23. For more on Aristarchus’ editorial methodology in using external manuscript evidence for determining whether or not to apply athetesis, see the extended discussion in N 1996a.146-147n169. Janko 1998c.147 has asserted about my discussion: “N. also believes (146, n. 149 [should be 169]) that the Alexandrians athetised suspect verses because of MS evidence, whereas in fact they athetised such verses, rather than omitted them, only when they lacked external evidence against them.” This assertion is criticized by Apthorp 1999.19n22: “As M. Haslam correctly states, ‘It is … clear that Aristarchus did at least on occasion have manuscript authority for his atheteses’ [1997.76]. The evidence for this is assembled in [Apthorp 1980] pp. 49-53, with notes on pp. 102-9. R. Janko is misleading when he writes that ‘the Alexandrians athetised suspect verses, … rather than omitted them, only when they lacked external evidence against them’ [1998c.207, in Janko’s review of Nagy 1996a].” Elsewhere, Janko tones down his assertion by inserting the qualification “largely” [1992.28]: “for him [Aristarchus], athetesis was largely based on internal evidence.”

24. On Zenodotus, I refer again to my defense of his editorial methods in N 1996a.133-138.

25. See especially Ludwich 1884.11-16 (“Die alte Vulgata”).

26. West is here following Watkins 1995.150, who refers at p. 151 to an unpublished 1994 work of Joshua Katz on Homeric formulas beginning αὐτὰρ ἐπεί (for an abstract, see Katz 1998).

27. If indeed Homeric ταρ is derived from τ’ (from τε) plus ἄρ just as γάρ is derived from γ’ (from γε) plus ἄρ, we still need to account for the enclitic status of ταρ in some Homeric contexts and the non-enclitic status of Tἄρ in others. It may be relevant that γάρ is fully lexicalized in all its attestations, that is, it has become a single word, whereas neither ταρ nor ἄρ / ῤα (vs. ἄρα) have achieved that status in Attic/Ionic, outside of poetic diction. With reference to the arguments of Katz 1998, I see further evidence for counter-arguments. For example, not only are αὐτάρ and ἀτάρ syntactically parallel in a variety of Homeric contexts: so also are ἀτάρ and (δ’ / (δ’τ(ἐ. Note too the collocations AU)TAR ἄρA at II 103 and αὐτὰρ ὃ αὖτε at II 105, 107.

28. For examples of questionable rewritings in West’s Iliad, see also the criticisms of Janko 2000.1. I hope to discuss further examples in a separate piece.

29. I thank the following for their valued advice (any mistakes that remain are my own): Michael Apthorp, Egbert Bakker, Graeme Bird, Timothy Boyd, Miriam Carlisle, Olga Davidson, Casey Dué, Mary Ebbott, David Elmer, Douglas Frame, Albert Henrichs, Carolyn Higbie, Alexander Hollman, Olga Levaniouk, Richard Martin, Leonard Muellner, Jed Wyrick, Dimitrios Yatromanolakis.