4. Editing the Homeric Text: Different Methods, Ancient and Modern*

4§1 Martin West published in 2001 a book that supplements his 1998/2000 edition of the Homeric Iliad. [1] It serves as a vehicle for explaining (1) his overall editorial stance and (2) the differences between his edition and those of his predecessors. My essay here concentrates on debating those two aspects of the book. [2] Such debate serves as a point of re-entry into an ongoing question, basic to Homeric studies: how was the text of Homer edited in the ancient world, especially in Alexandria, and how were the editing methods of the ancients different from those of the moderns?

4§2 West thinks of his own text of the Iliad as the “version” produced by Homer himself—or, in West’s terms, by the poet of the Iliad (as we saw {75|76} earlier, he denies that the poet’s name was Homer): “We have to choose one version that we are aiming to establish, and clearly it should be the poet’s last version.” [3] I focus on a detail in this statement. West’s assumption, and his wording makes it clear, is that he is working with something other than oral poetry.

4§3 There is a serious problem with this assumption. West is claiming that he can isolate one distinct and integral “version” of the Iliad, sorting it out from a mass of multiple versions contained in the Homeric textual tradition. He claims further that this “version” was produced by a single poet. Even further, he claims that he can distinguish between earlier and later “versions” stemming from this same poet.

4§4 As I have been arguing all along, the study of living oral poetic traditions shows that different available versions cannot be reduced to a single basic version from which all versions are to be derived. For West’s editorial method to succeed, it must be assumed, as a starting point, that the Iliad is not a matter of oral poetry. [4]

4§5 Whether or not it is true that the Iliad is oral poetry, one thing is for sure: West has simply bypassed the available comparative evidence provided by the study of oral poetics. Especially valuable, for comparison with Homeric poetry, is the evidence of oral poetry as attested in literate—not illiterate—societies. There are many examples from over the world, but the textual traditions of medieval vernacular poetry in Western Europe offer perhaps the most relevant and telling points of comparison. Detailed descriptions can be found in the work of medievalists like Paul Zumthor, whose descriptive model of mouvance is most useful to Homerists, as I have argued elsewhere. [5] Also useful is Bernard Cerquiglini’s alternative model of variance. [6] West chooses not to engage with any such models.

4§6 A basic lesson to be learned from studies of oral poetry as transmitted through texts has to do with the actual variation that we find in such texts. Such variation, in and of itself, reveals the workings of oral poetry. So long as oral poetry persists in a given literate society, each written instance of a “version” will be different from each succeeding instance—even in traditions where the audience at large assumes that each new instance is simply a repetition of previous instances. The degree of variation is itself a variable. {76|77} Studies of oral poetry as transmitted though texts have made it clear that the degree of variation—and even the audiences’ awareness of it—will itself vary from culture to culture, or even from phase to phase in any one culture. [7] The actual fact of variation, however, remains a constant. Variation is an essential feature of oral poetry, which is a process of composition-in-performance, and this feature is reflected—in varying degrees—by the written records of oral poetry. In terms of ancient Greek, a word that approximates such a model of variation is poikiliā (ποικιλία) ‘variation, variety’. [8]

4§7 Beyond the work of Zumthor and others on oral traditions in literate societies, it is essential to keep in mind the more fundamental work of Milman Parry and Albert Lord on oral traditions in general. [9] In terms of Lord’s analysis, a suitable word for describing the variation to be found in Homeric poetry is multiformity. [10] West not only bypasses this concept: he ignores altogether the work of Parry and Lord. In West’s 2001 book, as also in the introduction to his 1998 / 2000 edition of the Iliad, he does not even mention the names of Parry and Lord anywhere. [11]

4§8 When West dismisses various arguments about the multiformity of Homeric poetry as a sign of its oral poetic heritage, he backs up his dismissal by citing an article that purports to show how the Iliad is not multiform but “uniform.” [12] Unfortunately, this whole article is based on a misunderstanding of Lord’s concept of multiformity in oral poetry. [13] West’s citation of the article is not only derivative—since he ignores Lord’s concept of multiformity—but also inadequate.

4§9 In this context, West also dismisses a related concept, “multitext.” [14] The model of a multitext edition of Homer, developed for the purpose of addressing the multiformity of oral poetry and of Homeric poetry in particular, is designed as an alternative to West’s hypothesis of recovering an “original” version of Homer. [15] A multitext edition offers an alternative to {77|78} West’s reconstructing a single “version” as the basis for his explanations of all other “versions.” West misreads the concept of multitext when he claims that a multitext edition of Homer promotes an attitude of indifference toward the critical evaluation of variant readings in the history of the Homeric textual tradition. Contrary to West’s claim, a multitext edition of Homer does indeed allow for the privileging of one variant over others—but only in relative terms, since the editor may find that different variants became dominant in different phases of the Homeric tradition. In terms of a multitext edition, the editor of Homer needs to adopt a diachronic perspective—as an alternative to a pseudo-synchronic perspective. The term “pseudo-synchronic” will be explained further below.

4§10 From a diachronic and even historical point of view, it is indeed possible to think of a single given variant as the definitive variant at a single given time and a single given place. But the privileging of any given variant by any given audience is itself a matter of variation, and a diachronic perspective makes it clear that different variants were perceived as the “right” version at different points in the history of the Homeric tradition.

4§11 It is not that a diachronic perspective avoids “passing any value judgments,” as West claims. [16] The point is, rather, that the modern value judgments of the editor need to be responsive to changes in the ancient value judgments about Homeric poetry throughout the historical continuum of the Homeric tradition. West applies the term “value judgments” only to the critical stances of modern editors and their readers. But what about the “value judgments” of the ancient world? In this case, a more suitable term is “reception.” The problem is, West does not take into account the history of Homeric reception. If indeed Homeric poetry—as recorded by the Homeric textual tradition—reflects a system derived from oral poetry, then the value judgments of an editor need to be responsive to the multiple value judgments represented by that system as it evolves through time. An empirical analysis of the textual evidence reveals an underlying system capable of generating a multiplicity of versions, and it is methodologically unsound for an editor to assume that only one of these extant versions was basic while the others were derivative. Such an assumption exemplifies what I call a “pseudo-synchronic” point of view. I define such a point of view as one that treats irregularities within a given traditional system as if they could never have been regularities in other phases of that same system.

4§12 A related problem is West’s uncritical usage of the term “original.” In terms of such an absolutizing concept, any attested variant that happens to {78|79} be different from the “original” variant will be dismissed as not genuine, not authentic, not worthy of the “original” author, and so on. If the “original” version involves fewer verses than other versions, the additional verses in the other versions will be judged to be a textual interpolation. If the “original” involves more verses, then the absence of these verses in the other versions will be judged to be a textual lacuna. [17] Such an approach, looking at Homeric texts without considering the system that produced these texts, loses sight of what we know about the workings of oral poetry. If indeed Homeric poetry is at least derived from oral poetry—and even West concedes such a derivation—then the privileging of any one Homeric variant is merely a hypothesis, and the rejection of other variants depends on that hypothesis. [18] Then the judgments of privileging or rejecting one or another variant can be exposed for what they are—the product of an exercise in circular reasoning.

4§13 Even in terms of this circular reasoning, West’s book is inconsistent. On the one hand, he is certain that the poet of the Iliad had his own “version,” on the grounds that there must have been an “original text.” On the other hand, West is also certain that this text was original on the grounds that it must have been the only truly “oral” version, while the other versions were somehow “non-oral” on the grounds that they stem from rhapsōidoi (ῥαψῳδοί) ‘rhapsodes’. West wants to have it both ways—not only an original text but also a truly oral version. The problem is, these two concepts are mutually exclusive. There is a further problem with his model of the “oral” version: West’s distinction between Homer (or West’s “Iliad poet”) and the “rhapsodes,” representing respectively “oral” and “non-oral” traditions, is merely an assumption. Several works, applying the “diachronic perspective” that West rejects, have demonstrated the arbitrariness and artificiality of this distinction, but West does not engage with any of these works. [19]

4§14 Sometimes a model, even if it happens to be flawed, turns out to be a useful approximation of reality. Let us assume, for the moment, that West’s model is the closest thing to reality. Even if that were so, it is unjustifiable for him to claim from the very start that his explanation is the absolute truth. Such a claim is evident from this statement of his: A casual reader may be led to think that West’s “original text” is some kind of established reality. It is not. Nor can it be asserted as a fact that the business of rhapsodes was to perform “excerpts” from “specific poems.” Such an assertion assumes not only an “original text” but also a textual procedure of “excerpting” parts from a totality. While asserting this assumption, West ignores alternative arguments, including my argument for rhapsodic “relay mnemonics,” a performative procedure of sequencing a totality from start to finish. [22]

4§15 In West’s “Notes on Individual Passages,” he consistently makes a distinction between the “original” poet and the “interpolating” rhapsodes. [23] Here are examples from the first half of the Iliad, and in each case I will highlight the words “original,” “interpolate,” and “rhapsode.” At II 258, “rhapsodes had difficulty in remembering this unusual phrase (νύ περ [at] [Odyssey ii] 320 v.l.; not elsewhere, at least in Homer), and improvised substitutes.” [24] The verses at Iliad II 491–492 “must have been added by a rhapsode who misinterpreted πληθύν in 488, not realizing that the mass of the army was being distinguished from the leaders.” [25] On Iliad II 535: “perhaps it was added by a rhapsode who thought … East Locrians should be unequivocally distinguished from the Ozolian Locrians.” [26] On Iliad II 547–551: “I have therefore bracketed these five lines, assuming them to have replaced the original list of towns.” [27] On Iliad III 41: “but it [= an idiomatic usage containing an implicit comparative] may have prompted a later rhapsode to interpolate a line {80|81} containing the explicit comparative.” [28] On Iliad III 144: this verse “was added by someone who knew the Aithra story.” [29] On Iliad IV 177: “This verse may be a rhapsode’s addition. It adds a gratuitous crudity to a sentence which is perfect without it.” [30] On Iliad V 313: “this inorganic line ... may be a rhapsode’s addition.” [31] On Iliad V 398–402: “Hades’ withdrawal to Olympus ... cannot have been an essential part of the original myth.” [32] On Iliad V 778: τώ not αἵ “is the original, guaranteed by the principle ‘utrum in alterum abiturum erat?’” [33] On Iliad V 820–821: “Mechanical repetition is not foreign to Homeric technique, but it may also result from a rhapsode’s or copyist’s zeal.” [34] On Iliad VI 354, in conjunction with Iliad III 100 and XXIV 28, with reference to the variants ἀρχῆς and ἄτης in all three passages: “The explanation is probably that each noun is original in one or two of the three passages and was then introduced in the remainder by contamination in some sources.” [35] On Iliad VI 388: “the interpolator presumably felt that a nearer main verb [= nearer than the one at [Iliad VI] 386] should be provided.” [36] On Iliad VII 135 as an “interpolation”: “It might have been due to a rhapsode who had some particular interest in Pheia, or in pleasing an audience in Pheia (cf. [Odyssey xv] 297?).” [37] On Iliad VIII 359: “suspicion of a rhapsodic interpolation may well be founded. Not only is the line otiose as a whole, but ἐν πατρίδι γαίῃ is especially feeble and irrelevant.” [38] On Iliad IX 320: “The rhapsode responsible failed to follow the tenor of the argument.” [39] On Iliad IX 523: this verse “better fits the logic” of the context of Odyssey xxii 59 and is “surely a rhapsode’s addition.” [40] On Iliad XI 662: “It is an ill-considered interpolation from [Iliad XVI] 27” (similarly on [Iliad XI] 830). [41] On Iliad XII 175–180: “But it is hard to see why a rhapsode should have interpolated lines that seem so at variance with the situation. I suspect that we have here a fragment of composition originally {81|82} meant for another setting.” [42] Iliad XII 372 “seems to be an Attic interpolation” and Iliad XII 426 is “surely a gloss by a rhapsode,” [43] while Iliad XII 449 “was interpolated by a rhapsode who missed the point.” [44]

4§16 Here are further examples, from the second half of the Iliad. Iliad XV 551 is a “concordance-interpolation.” [45] Iliad XVIII 39–49 may be a “casual omission by a rhapsode.” [46] On Iliad XIX 76, with reference to a variant adduced by Zenodotus: “if this were the original text, it would be difficult to understand the logic of 79–82.” [47] On Iliad XXIII 870–871, I refer to my extended comments further below. [48] On Iliad XXIV 616: “Some later rhapsodes or scribes, being unfamiliar with the name [Ἀκελήσιος], assimilated it to the well known [Ἀχελώιος].” [49]

4§17 I have gone to all the trouble of collecting these examples because they show the pervasiveness of West’s attitude. They give a valuable insight into his overall approach in establishing the text of the Iliad. There is no reason, of course, to question West’s attempts to pass editorial judgment on variations in the Homeric textual tradition. But there is indeed reason to question his setting up an arbitrary dichotomy between “original” poet and rhapsodes, “original” text and interpolation. Such a dichotomy flattens the history of Homeric reception. I see here a “unitext” explanatory model of Homeric poetry, based on the pseudo-synchronic perspective of imagining an “original” poet. Such a model requires, ideally, the editorial elimination of all variants from a once-uniform text. But the textual variations noted by West—and he could have adduced hundreds and even thousands of further examples from the Iliad—have a history of their own, which cannot be extricated from the compositional variations inherent in Homeric poetry itself. This history—which is the history of Homeric reception—can be studied systematically by applying what I have described as a multitext model of Homeric poetry. Such a model allows an editorial evaluation of variants from a diachronic perspective encompassing the entire history of Homeric performance traditions.

4§18 In a few isolated cases, West’s response to textual variants goes beyond his simple distinction between “original” poet and “interpolating” rhapsodes. It is in such cases that I see a better chance for establishing a scholarly {82|83} dialogue that could lead to making more complex distinctions between relatively earlier and later phases of a tradition that drew its life from public performance. A prime example is Iliad XV 64–71: these verses, says West, were composed: We can see in this formulation, if we subtract from it West’s speculation about textual tampering, the bare outlines of distinct phases in the development of the Iliad as a composition subject to ongoing recomposition-in-performance.

4§19 I offer some further examples. On Iliad XV 263–268, “repeated from” Iliad VI 506–511, West remarks: “There is no reason why the poet should not have repeated at least a part of the simile in Iliad VI. But the truism that ‘repetition inheres in the oral style’ (Janko ad loc.) should not be used as a defence of the whole passage, since it is equally true that repetition from other contexts is typical of interpolations.” [51] At least we see here a concession about the realities of “oral style.” Then there is Iliad XIX 365–368, which is parallel to verse 164 of the Hesiodic Shield of Herakles; this parallelism, says West, “puts the author in bad company, and it may be suspected that he was a rhapsode of the same period (early sixth century) who saw fit to add what he thought was a dramatic touch to the account of Achilles’ arming.” [52] A more positive way of looking at this parallelism is to investigate further the implications of the relative dating to the sixth century. On Iliad XX 213–241: “The genealogy ... looks like an addition to the original speech, hardly anticipated in 203–209. But it may be an addition by the first poet, after he received the impulse to honour the [Aineiadai] as the latter-day royalty of Troy.” [53] West’s category of “first poet” is a roundabout way of admitting that the wording of the “interpolation” rivals in importance the wording of the “original.” On Iliad XX 485–499: “Andromache’s lament is one of the emotional high points of the Iliad. We are in the presence of great poetry here, but also of a passage irreconcilable with the view that our text is the product of a straightforward {83|84} process of dictation by an oral poet who never altered or added to his primary effusions.” [54] The “dictating oral poet” seems to me a “straw man”; what is of primary importance, rather, is West’s concession about the “great poetry.” On Iliad XXIV 662–663: these verses are “genuine lines, but not part of the passage as originally composed. They were inserted later by the poet, after he had related the domestic laments for Hector (723–776) and found himself with nine days still to fill.” [55] Once again, West’s speculation about textual tampering can be reinterpreted in terms of distinct phases in the history of the Iliad as a composition subject to ongoing recomposition-in-performance.

4§20 We have by now seen in some detail how West’s notion of an “original text” of the Iliad pervades his editorial treatment of surviving textual variants in general. These variants fall into the following categories of transmission: the medieval textual tradition, [56] ancient quotations, [57] papyri, [58] and ancient editions. [59] Of special relevance is West’s treatment of the variants reported by the three major editors of Homer in Hellenistic Alexandria—Aristarchus of Samothrace, Aristophanes of Byzantium, and Zenodotus of Ephesus. {84|85}

4§21 For background, it is important for me to review here the essence of my discussion in the first chapter, where we examined what we can learn from the Homeric scholia about the actual procedures of ancient editorial work on the text of Homer. My general term “editorial work” is meant to encompass a wide range of different methods employed by ancient scholars, especially by the Alexandrians. I use this term to accommodate the meaning of Greek diorthōsis (διόρθωσις), for which I have developed the working translation ‘corrective editing’. [60] Two other Greek words are most relevant: ekdosis (ἔκδοσις) and hupomnēma (ὑπόμνημα), corresponding respectively to the modern concepts of ‘edition’ and ‘commentary’.

4§22 The semantic overlaps and distinctions between ekdosis (ἔκδοσις) and diorthōsis (διόρθωσις) in the reportage of the Homeric scholia concerning the editorial work of Aristarchus can be clarified in light of the observations made by Franco Montanari concerning the history of Aristarchus’ Homeric hupomnēmata (ὑπομνήματα) ‘commentaries, annotations’. [61]

4§23 Montanari points out that Aristarchus’ first publication of commentaries or hupomnēmata to the text of Homer was keyed to the text established by his predecessor, Aristophanes, as we see from the wording in scholia A at Iliad II 133a, where Didymus says: ἐν τοῖς κατ’ Ἀριστοφάνην ὑπομνήμασιν Ἀριστάρχου ‘in the commentaries [hupomnēmata] of Aristarchus according to the text of Aristophanes’. After that, there was an ekdosis—that is, a ‘publication’—of Aristarchus’ own diorthōsis ‘corrective editing’ of Homer. After that, according to Montanari, there followed a new set of hupomnēmata that were keyed to the editorial work of Aristarchus himself. Relevant are the titles of two monographs produced by Aristarchus’ direct successor, Ammonius: (1) περὶ τοῦ μὴ γεγονέναι πλείονας ἐκδόσεις τῆς Ἀρισταρχείου διορθώσεως ‘About the fact that there did not exist more ekdoseis of the Aristarchean diorthōsis’ (via Didymus in scholia A at Iliad X 397–399a) and (2) περὶ τῆς ἐπεκδοθείσης διορθώσεως ‘about the re-edited diorthōsis’ (via Didymus in scholia A at Iliad XIX 365–368a). 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 diorthōsis, which, to repeat, I have been translating as ‘corrective editing’. Such is the idea in the first of the two titles of monographs attributed to Ammonius, where this Aristarchean scholar refers to a single ekdosis (ἔκδοσις) or ‘edition’ of Homer as opposed to more than one edition; the same idea accounts for {85|86} Ammonius’ reference in the second title to a ‘re-edition’, as expressed by the wording ἐπεκδοθείσης. 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. [62]

4§24 In the Homer scholia, we see that Didymus refers to texts labeled αἱ Ἀριστάρχου ‘the texts of Aristarchus’ (at Iliad II 221, for example, he says ἐν ταῖς Ἀριστάρχου ἐκδόσεσι ‘in the editions [ekdoseis] of Aristarchus’); elsewhere, he says ἡ ἑτέρα τῶν Ἀριστάρχου ‘the other one of the texts of Aristarchus’ (as at II 131a, where the wording ἡ ἑτέρα ‘the other one’ makes it explicit that there were two). [63] Evidently, the stance of Didymus differs from that of Ammonius. For Didymus, there were two ekdoseis of Homer by Aristarchus, [64] though there is no reason to think that they were considered autographs of Aristarchus himself. [65]

4§25 From the cumulative evidence of the references in the Homeric scholia to the two ekdoseis ‘editions’ attributed by Didymus to Aristarchus, I conclude that these two Aristarchean ekdoseis represent a new base text dating from a post-Aristarchean era and featuring the preferred readings of Aristarchus as extrapolated primarily from his commentaries or hupomnēmata. This new base text needs to be distinguished from an older and more “standard” ekdosis, derived (but distinct) from the ekdosis of Aristophanes. It was this more “standard” ekdosis that Aristarchus himself had used as the base text for his hupomnēmata. Also, the base text of Aristonicus was this older “standard” ekdosis, featuring Aristarchean critical signs in its margins. [66] {86|87}

4§26 With this background in place, I proceed to outline West’s overall approach to the editions of Homer by Aristarchus and his predecessors. Addressing the editorial work of Aristarchus himself on the textual traditions of Homer, West says that this editor “appears to have been simply unaware of the possible benefits of extensive manuscript collation; or perhaps he realized only too well that it was likely to yield little but a bewildering variety of aberrations.” [67] He offers a similar assessment of Aristarchus’ predecessor, Aristophanes. [68] As for the earlier editorial work of Zenodotus, West dismisses its credibility altogether, slighting its value as a source for the textual history of Homeric poetry: “As for Zenodotus himself, the text associated with his name is so eccentric that it is impossible to regard it as the product of any rational process of selection of readings from alternative sources—or indeed of any rational thought at all.” [69]

4§27 Surveying the editorial legacy of Zenodotus, Aristophanes, and Aristarchus, West in the introduction to his edition of the Iliad had made the extreme claim that it was Didymus, not Aristarchus, who regularly collated manuscript evidence in search of variants: “Didymus fuit, non Aristarchus, qui exemplaria quam plurima consuluit, ut varias lectiones compararet.” [70] In making this claim, West ignored the testimony of scholia A at Iliad IX 222, where Didymus says of Aristarchus: 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 ‘it would have been better, says Aristarchus, if it had been written “ἂψ ἐπάσαντο” or “αἶψ’ ἐπάσαντο”; nevertheless, because of his extreme caution, he changed nothing, having found in many of the texts this attested way of writing it [= “ἐξ ἔρον ἕντο” instead of “ἂψ ἐπάσαντο”]’.

4§28 Not only was this testimony ignored in West’s 1998 introduction to his edition of the Iliad: the most relevant wording—as highlighted here—was {87|88} not even quoted in the apparatus criticus of his 1998 edition of Iliad I–XII. [71] Then, in his 2001 book, West finally quoted the passage and then backtracked slightly from his earlier claim by saying: “Didymus knew that Aristarchus consulted more than one text, because he cited different scholars’ readings on different occasions, but it is just his own assumption that Aristarchus systematically checked ‘many’ copies before discussing any reading.” [72] So, in his 2001 book, West was now more open to the possibility that Aristarchus “checked” some copies, if not “many.”

4§29 Even in his 2001 book, however, West was still saying: “Not once does he [Aristarchus] appeal to the authority of manuscripts.” [73] West justifies this statement by adding in his next sentence: “It is Didymus, not Aristarchus, who makes all the references.” At this point, West lists various terms used in the Homeric scholia to designate various manuscript traditions: αἱ πλείους ‘the majority [of the ekdoseis]’, αἱ πᾶσαι ‘all’, αἱ χαριέστεραι (khariesterai) ‘the more elegant ones’ or αἱ ἀστειότεραι (asteioterai) ‘the more cultivated ones’. He lists also the antithetical terms ἡ κοινή (koinē) or αἱ κοιναί (koinai) ‘the common one(s)’, κοινότεραι (koinoterai) ‘the more common ones’, δημώδεις (dēmōdeis) ‘the popular ones’, εἰκαιότεραι (eikaioterai) ‘the simpler ones’, φαυλότεραι (phauloterai) ‘the inferior ones’. He thinks that all these terms reflect the usage of Didymus, not of Aristarchus. In addition, he thinks that it was Didymus, not Aristarchus, who “cites the texts” of such early sources as Antimachus of Colophon (late fifth century) and Rhianus of Crete (second half of the third century)—though West immediately appends this qualification: “... even though his [Didymus’] knowledge of some of these sources was indirect.” [74] In fact, there is a further qualification implicit in the very fact that West says “cites the texts” here instead of “collates the texts.” West is forced to say it this way because, as it turns out, he has no proof that Didymus had direct access to the Homer text of either Antimachus or Rhianus. [75] As we will see later, the question of Didymus’ access, especially to the text of Antimachus, is crucial for determining the validity of West’s theories about the sources used by Didymus. In what follows, I maintain that these theories are not supported by the evidence of the Homeric scholia.

4§30 No doubt, Didymus was a direct source for the information transmitted by the A scholia about variant readings in the Homeric textual tradition. Nor is there any doubt that Didymus spoke with authority about these variant {88|89} readings. This authority, however, came not from Didymus’ searching for variant readings in texts available to him but from Aristarchus himself. Didymus, as a successor of Aristarchus, spoke with the authority of Aristarchus because he situated himself as a continuator of that authority. When Aristarchus said εὕρομεν ‘it is our finding’, the authority of his finding could be passed on to the latest of Aristarcheans. In this case, that man was Didymus, who flourished over a century after Aristarchus. When Didymus said εὕρομεν ‘it is our finding’ in his own turn, he too was speaking with the authority of Aristarchus, just as previous Aristarcheans had spoken before him. Such is the dominant mentality of the reportage we read in the scholia of Homer, especially in the Venetus A. My formulation in this paragraph will now be supported by a selection of examples, with commentary.

4§31 My premier example is a quotation taken directly from Aristarchus himself, as preserved in scholia A at Iliad I 423–424. In my 2000a review of West’s 1998 edition of Iliad I–XII, I had cited this quotation, not mentioned in West’s introduction to his edition, as evidence against his theories about Didymus. [76] The quotation, designated as the lexis ‘wording’ of Aristarchus, is introduced this way: λέξις Ἀριστάρχου ἐκ τοῦ Α τῆς Ἰλιάδος ὑπομνήματος ‘here is the lexis of Aristarchus, from his hupomnēma on volume I of the Iliad’. The expression λέξις Ἀριστάρχου ‘the lexis of Aristarchus’, as I noted earlier, seems to convey the idea that the hupomnēmata of Aristarchus were not only commentaries written down in papyrus scrolls but also, at least notionally, commentaries delivered as lectures by Aristarchus, as if these commentaries were meant to be transcribed by his students. [77] 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. [78] I give the full sequence here, as printed by Erbse: [79] As I stressed in my 2000a review, the person who is being quoted here, after he expresses his preference for the variant reading κατὰ δαῖτα instead of μετὰ δαῖτα, says the following: οὕτως δὲ εὕρομεν καὶ ἐν τῇ Μασσαλιωτικῇ καὶ Σινωπικῇ καὶ Κυπρίᾳ καὶ Ἀντιμαχείῳ καὶ Ἀριστοφανείῳ ‘this is the way we found it in the Massaliōtikē and the Sinōpikē and the Kupria and the Antimakheios and the Aristophaneios’. [81] I argued that this wording provides evidence against West’s claim that Aristarchus did not collate manuscripts. Agreeing with Ludwich, [82] I argued further that the subject of οὕτως ... εὕρομεν ‘this is the way we found it’ 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.” [83]

4§32 For confirmation of this way of interpreting οὕτως ... εὕρομεν ‘this is the way we found it [attested]’ in scholia A at Iliad I 423–424, I note the wording οὕτως εὑρών ‘having found it [attested] this way’, applied to Aristarchus {90|91} himself by Didymus, in the scholia A at Iliad IX 222. 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 ‘it would have been better, says Aristarchus, if it had been written “ἂψ ἐπάσαντο” or “αἶψ’ ἐπάσαντο”; nevertheless, because of his extreme caution, he changed nothing, having found in many of the texts this attested way of writing it [= “ἐξ ἔρον ἕντο” in place of “ἂψ ἐπάσαντο”]’. I highlighted the pertinent wording, as underlined here, in my 2000a review of West’s 1998 edition of Iliad I–XII. [84] As I noted, the wording is quoted in the apparatus criticus of Ludwich but not in that of West. More important for now, the usage of οὕτως εὑρών ‘having found it this way’ shows once again that the ‘we’ of Didymus represents the Aristarchean tradition writ large, of which Didymus considers himself the extension and even culmination—whether or not we agree with this self-assessment. The magisterial οὕτως ... εὕρομεν ‘this is the way we found it’, as ‘spoken’ by Aristarchus, is perpetuated by Didymus. [85]

4§33 Responding to the criticisms of his theories about Didymus, West in his 2001 book is now forced to reckon with this expression οὕτως ... εὕρομεν ‘this is the way we found it’ in scholia A at Iliad I 423-424. West claims that the subject of οὕτως ... εὕρομεν here must be Didymus, not Aristarchus. [86] To back up his claim, he lists five instances in the Homer scholia where Didymus is the subject of the verb heuriskein ‘find’: scholia at Iliad II 131a, II 517a, IV 3a, XV 469–470a, XVI 636c. [87]

4§34 I note that the first three of these five instances listed by West involve explicit references to what Didymus ‘found’ in one or the other of the two ekdoseis ‘editions’ that he attributed to Aristarchus himself, which as I argued earlier date from a post-Aristarchean era. [88] To that extent, such instances can even be used as evidence to support my argument that οὕτως ... εὕρομεν ‘this is the way we found it’ is an Aristarchean usage, faithfully continued by Didymus. As for the last two of the five instances, the fourth and the fifth, we are about to see further indications of Aristarchean usage. [89]

4§35 In the fourth instance, the wording suggests that Didymus uses εὕρομεν {91|92} ‘we found’ in the sense that he is participating in the authoritative report of Aristarchus concerning what variants are attested where: Aristarchus here is reported by Didymus to be saying that both variants at verse 470, πρῴην ‘just now’ and πρώϊον ‘this morning’, are attested or ‘written’ in the manuscript evidence, γράφεσθαι, and then it is further reported by Didymus that the variant ἐΰστροφον ‘well-twisted’ is also found to be ‘written’, γεγραμμένον, alongside the variant νεόστροφον ‘newly-twisted’ at the previous verse, 469. The train of thought conveys a unified discovery procedure, not a separation of Didymus’ authority from that of Aristarchus. The speaker who reports the procedure has just quoted a sequence of Homeric phraseology; this quotation is the lēmma, the basis for the speaker’s discussion. The speaker then proceeds to discuss the manuscript variants that he finds within the whole lēmma. In this case, the lēmma spills over from one Homeric verse to the next: νευρὴν δ’ ἐξέρρηξε νεόστροφον, ἣν ἐνέδησα | πρώϊον ‘he [= the god] snapped my newly-twisted bowstring, which I strung to my bow this morning’. The speaker begins the discussion by focusing on the last word of the lēmma, taking note of an attested variant—πρῴην ‘just now’ instead of πρώϊον ‘this morning’—and then adds, working backwards, that there is another case of variation earlier on in the lēmma, since ‘we find’ an attested variant ἐΰστροφον ‘well-twisted’ instead of νεόστροφον ‘newly-twisted’, just as πρῴην ‘just now’ was found instead of πρώϊον ‘this morning’.

4§36 In the case of finding the different readings πρῴην and πρώϊον in different manuscripts, the authority is explicitly Aristarchus. In the case of the {92|93} different readings ἐΰστροφον and νεόστροφον, the authority for finding the variant ἐΰστροφον is the ambiguous ‘we’ of the statement εὕρομεν ‘we found’. Despite its ambiguity, this statement about the variants ἐΰστροφον and νεόστροφον is semantically and even syntactically integrated with the earlier statement about the variants πρῴην and πρώϊον. The authority that supports the comparison of different readings in different manuscript traditions is presented as the same. That authority is Aristarchus.

4§37 Not only do we see here that Didymus explicitly attributes to Aristarchus the finding of the variant reading πρῴην instead of πρώϊον. We can see also, from both scholia A and scholia T at Iliad XV 470a, that Aristonicus is another witness to this reading. Moreover, Aristonicus specifies that Aristarchus found this reading in the Homer text of Zenodotus: ὅτι Ζηνόδοτος γράφει “πρῴην.” The word ὅτι ‘because’ is used here by Aristonicus in order to explain why Aristarchus put a sign in the margin to the left of the verse (Erbse thinks that the sign here at verse Iliad XV 470 was originally a διπλῆ περιεστιγμένη ‘double mark with dots on the side’, not the plain διπλῆ ‘double mark’ shown by the Venetus A). In this context, it is essential to keep in mind that Didymus did not have access to the Homer text of Zenodotus. [90] Further, Didymus says explicitly that his information about the variant reading πρῴην comes from Aristarchus. [91]

4§38 Now we come to the fifth of the five instances adduced by West to support his claim that Didymus is the subject of the verb εὕρομεν ‘we found’ in the scholia A at Iliad I 423–424. The wording is as follows: In this case, we see a collocation of εὕρομεν ‘we found’ with a direct quotation from Aristarchus himself. Moreover, it could be argued that the subject of εὕρομεν is likewise Aristarchus, since the statement that starts with the {93|94} phrase ἐν δέ τισιν εὕρομεν ... ‘but in some texts we found…’ may be a direct quotation in its own right.

4§39 Having surveyed West’s five examples of usages involving heuriskein ‘find’, let us return to the wording that begins with οὕτως δὲ εὕρομεν καὶ ἐν τῇ Μασσαλιωτικῇ καὶ Σινωπικῇ καὶ Κυπρίᾳ καὶ Ἀντιμαχείῳ καὶ Ἀριστοφανείῳ ‘this is the way we found it in the Massaliōtikē and the Sinōpikē and the Kupria and the Antimakheios and the Aristophaneios’ in scholia A at Iliad I 423–424. From what we have seen so far, there is no support for West’s claim that οὕτως δὲ εὕρομεν ‘this is the way we found it’ is the precise point where the “verbatim quotation” from Aristarchus stops and the discourse of Didymus begins. [92]

4§40 West goes even further in his discussion of this passage. Not only the wording but also the content, he claims, points to Didymus and not Aristarchus. As I argue in what follows, there is no support for this additional claim.

4§41 The wording that immediately follows in scholia A at Iliad I 423–424 is unproblematic: Καλλίστρατος δὲ ἐν τῷ Πρὸς τὰς ἀθετήσεις ὁμοίως, καὶ ὁ Σιδώνιος καὶ ὁ Ἰξίων ἐν τῷ ἕκτῳ Πρὸς τὰς ἐξηγήσεις ‘Callistratus in his work Πρὸς τὰς ἀθετήσεις has it in a similar way; also [Dionysius] Sidonius, and [Demetrius] Ixion in the sixth volume of Πρὸς τὰς ἐξηγήσεις’. West is not alone when he says that Didymus must have been the originator of this statement highlighting Callistratus, Dionysius of Sidon, and Demetrius Ixion. [93]

4§42 It is quite another thing, however, for West to say that the man who consulted the Massaliōtikē, the Sinōpikē, the Kupria, the Antimakheios, and the Aristophaneios was not Aristarchus but Didymus himself. In order to justify such a claim, West would need to prove that Didymus had direct access to each of these five sources mentioned in the statement introduced by οὕτως δὲ εὕρομεν ‘this is the way we found it’. As I hope to show in an overview of the five sources, West produces no such proof.

4§43 Viewing the list of these five sources in reverse order, I start the overview with the fifth, the Aristophaneios. West asks: “Was it still accessible to Didymus?” [94] He responds immediately: “His [Didymus’!] wording at Α {94|95} 423–4, οὕτως δὲ εὕρομεν [...], would suggest autopsy, as I remarked in connection with Antimachus’ copy.” [95] The reasoning is circular in the case of the Antimakheios as well as the Aristophaneios. The wording at Iliad I 423–424, οὕτως δὲ εὕρομεν ‘this is the way we found it’, is in fact the only passage that West adduces to support his theory that Didymus had direct access to the Antimakheios. [96] The reasoning is also circular in the case of the Aristophaneios because, once again, the same wording in the scholia at Iliad I 423–424 is the only passage West adduces to support his theory of an autopsy by Didymus. In the case of the Aristophaneios, moreover, I note that West himself observes: “In the final analysis, then, it remains doubtful whether Didymus had direct acquaintance with Aristophanes’ text. All his knowledge of it may have come from Aristophanes’ pupils Aristarchus and Callistratus.” [97]

4§44 Continuing the overview, I turn to the fourth source in the listing of the scholia at Iliad I 423–424, the Antimakheios. In this case, West expresses less doubt than in the case of the Aristophaneios, although the lack of evidence for direct access is just as pronounced. [98] In the case of the Antimakheios, moreover, there are indications that Aristarchus himself did indeed have direct access to it (perhaps by way of Hermesianax of Colophon, who may have brought the text to Alexandria). [99] As West himself observes, “[Aristarchus] did use Antimachus’ text, at least sporadically.” [100] The most justifiable inference, then, is that the Antimakheios is mentioned in the passage of the scholia at Iliad I 423–424 starting with οὕτως δὲ εὕρομεν ‘this is the way we found it’ simply because Aristarchus did indeed consult the Antimakheios. [101] The same inference applies to the mention of the Aristophaneios in this same passage. In the case of the Aristophaneios, moreover, we are dealing with the most basic relevant fact of them all: that Aristarchus regularly consulted the Aristophaneios as a base text.

4§45 I conclude the overview by turning to the three other mentioned sources in the passage starting with οὕτως δὲ εὕρομεν ‘this is the way we found it’: the Massaliōtikē, the Sinōpikē, and the Kupria. All three are representatives {95|96} of the so-called ‘city editions’, the politikai (πολιτικαί). West describes as “a serious misconception” the idea that it was Aristarchus who collated the politikai (he gives a list of those who share this idea, including Ludwich, Erbse, Pfeiffer, Janko, and myself). [102] He claims that the politikai—that is, Homeric texts from Massalia (= ἡ Μασσαλιωτική / Massaliōtikē), Chios (= ἡ Χία / Khia), Argos (= ἡ Ἀργολική / Argolikē), Sinope (= ἡ Σινωπική / Sinōpikē), Cyprus (= ἡ Κυπρία / Kupria), and Crete (= ἡ Κρητική / Krētikē)—were collated not by Aristarchus but by Didymus (in the case of the Massaliōtikē, Khia, Argolikē, and Sinōpikē) and by Seleucus (in the case of the Kupria and the Krētikē). His claim is not based on any attempt to establish a relatively later date for the Homeric variants contained by the politikai—a date that would somehow suit the era of Didymus better than the era of Aristarchus. It is evident from his discussion that he can find no proof for any such redating. [103] Rather, the claim is based on West’s examination of the patterns of associations linking the politikai with other sources named in the Homer scholia. [104] I propose here to re-examine these patterns.

4§46 To begin this re-examination, I offer a formulation that anticipates my conclusion. West’s claim about the politikai seems to me impossible to sustain in view of the fact that the Homer scholia, especially those of Venetus A, refer consistently to these politikai as sources directly comparable with other sources that Didymus himself could not possibly have consulted. Such other sources include not only the Homer texts of Antimachus of Colophon and Aristophanes of Byzantium, as we see them mentioned in the scholia A at Iliad I 423–424 and elsewhere, but also the Homer text of Zenodotus. The scholia are in fact so consistent in comparing the texts of the politikai with the texts of Antimachus, Zenodotus, and Aristophanes—all three—that it becomes impossible or at least near-impossible to escape the conclusion that this consistency reflects a sustained system of collating choice texts. To accept West’s theories about Didymus is to deny the existence of any system, let alone any system devised by Didymus, since there is no evidence that Didymus had any direct access to the Homer texts of Antimachus, Zenodotus, and Aristophanes. By contrast, Aristarchus did indeed have such direct access. As the testimony of the Homer scholia shows, Aristarchus had the opportunity to compare systematically the texts of Antimachus, {96|97} Zenodotus, and Aristophanes, whereas Didymus did not. Moreover, the testimony of the Homer scholia shows an ongoing systematic comparison of these texts with the texts of the politikai. On these grounds, I infer that it was Aristarchus and not Didymus who collated the texts of the politikai just as he collated the texts of Antimachus, Zenodotus, and Aristophanes.

4§47 For a test case, I turn to the scholia A at Iliad XXI 870–871. Here the received text of Didymus is evidently this: The scholia A first quote the variant verses found in the Massaliōtikē: The scholia A then quote the variant verses found in the Antimakheios (the reading seems to have been damaged in scholia A, but scholia T have preserved the undamaged reading, which is given below). Then the scholia A continue with a paraphrase of what Aristarchus had said in his hupomnēmata: This paraphrase suggests that Aristarchus prefers the reading of the received text (and of the Antimakheios), which amounts to what survives as the received text of Didymus. There is further reinforcement in scholia T, where {97|98} the comment οὕτως Ἀρίσταρχος ‘Aristarchus has it this way ...’ is followed by this further comment: ἡ δὲ Μασσαλιωτικὴ οὕτω· ... ‘but the Massaliōtikē has it this way ...’ At this point the scholia T quote the same variant verses from the Massaliōtikē that we have already seen quoted by the scholia A. Then the scholia T follow up with this further comment: Ἀντίμαχος δὲ ... ‘but Antimachus...’ At this point the scholia T quote the variant verses found in the Antimakheios (the reading of which seems to have been damaged in scholia A), which agree with the sense of the received text and which disagree with the sense of the Massaliōtikē:

4§48 Throughout this discussion of variant verses at Iliad XXI 870–871, I have spoken in terms of quotations and comments by “scholia A” and “scholia T,” as if the scholiasts were speakers on their own. Of course the immediate speaker is notionally Didymus himself, even if we cannot reconstruct his exact words on the basis of the dual testimony provided by scholia A and T. As for the ultimate speaker, however, I maintain that it must be Aristarchus: what Didymus is saying is simply a reaffirmation of what Aristarchus had said before him concerning the master’s preference of the wording provided by his received text as opposed to the wording provided, in this case, by the Massaliōtikē. The wording of scholia T indicates that Aristarchus in his hupomnēmata had expressed his preference in the context of juxtaposing the reading of his received text with the reading of the Massaliōtikē and, secondarily, with the reading of the Antimakheios. To repeat, in this case the sense of the reading of the Antimakheios agrees with the received text, whereas the sense of the reading of the Massaliōtikē disagrees with it. The reference in the T scholia to the editorial choice of Aristarchus, οὕτως Ἀρίσταρχος ‘Aristarchus has it this way...’ is being juxtaposed with the reference to the reading of the Massaliōtikē. The οὕτω ‘this way’ in the expression ἡ δὲ Μασσαλιωτικὴ οὕτω ‘but the Massaliōtikē has it this way...’ marks the variant reading, while the οὕτως ‘this way’ of οὕτως Ἀρίσταρχος ‘Aristarchus has it this way...’ marks the judgment of Aristarchus.

4§49 West discusses these passages from scholia A and T at Iliad XXI 870–871 after having finished his discussion of the passage introduced by οὕτως ... εὕρομεν {98|99}‘this is the way we found it’ in scholia A at Iliad I 423–424. [105] He dismisses without justification Erbse’s inference that Didymus took the testimony of scholia A and T at Iliad XXI 870–871 from the commentaries of Aristarchus. [106] West now says about the same οὕτως δὲ εὕρομεν passage that he had discussed earlier: “I allowed earlier that he [Didymus] may not have had access to Aristophanes, but his εὕρομεν should be valid with respect to the items that head the list. The Massaliotica in particular has a strong claim to be among his direct sources in view of the frequency with which he cites it.” [107] Here West has reminded himself of one of his allowances, that Didymus probably had no access to Aristophanes, and I think it should be linked directly with another allowance that West had made elsewhere: that Didymus probably had no access to Antimachus, either.

4§50 If Didymus had no access to either Antimachus or Aristophanes, then the systematic comparisons of their manuscript evidence with that of politikai like the Massaliōtikē must be Aristarchus’ own. I infer, then, that Aristarchus did indeed collate manuscripts, and his procedure of collating is made explicit in expressions like ὁ μέντοι Ἀρίσταρχος... ‘but Aristarchus…’ in scholia A at Iliad XXI 870–871. That Didymus mediated such information can be granted, but the point remains that the judgments and sources reported by Didymus go back to Aristarchus. Moreover, Aristarchus is the authoritative source not only where a judgment—stylistic or otherwise—is explicitly linked with him, but even where we find only bare statements of ancient variant readings. [108]

4§51 Besides Antimachus and Aristophanes, the Homer text of Rhianus was another source that seems to have been inaccessible to Didymus. West leaves open the possibility that Didymus could have consulted Rhianus, but without positive evidence. He tries to build a case for negative evidence by saying: “Aristarchus, so far as we can see, ignored Rhianus altogether. Aristonicus never mentions him. All the references come from Didymus.” [109] Here it becomes essential to point out that Aristophanes of Byzantium was in fact interested in Rhianus, and that the Homer edition of Aristophanes did in fact contain Homeric variants that could be traced back to Rhianus. [110] West is not justified in inferring that these Homeric variants would have been ignored by Aristarchus. After all, Aristarchus’ own Homer edition was {99|101} directly linked with that of Aristophanes. [111] Aristarchus’ hupomnēmata or ‘commentaries’ were based initially on the Homer edition of Aristophanes, which was only later superseded by the Homer edition of Aristarchus himself. [112] In this connection, it is easy for modern scholars to forget a central fact about the reputation of Aristophanes in the ancient world: he was considered to be the ultimate editor in his own time, while his student Aristarchus was primarily a commentator and only secondarily an editor. [113]

4§52 Here it is relevant to mention another student of Aristophanes, Callistratus, who made explicit references to the variants derived from Rhianus. It is understandable that Callistratus, who produced his own ekdosis ‘edition’ of Homer, would have named his teacher Aristophanes, not his peer Aristarchus, as the editorial source. It is also understandable that Didymus, who referred to both Aristarchus and Callistratus, would likewise have named Aristophanes, not Aristarchus, as the editorial source for variants derived from Rhianus. But the point is, the systematic consultation of Rhianus as a source of Homeric variants did not start with Didymus, nor even with Aristarchus, but rather with Aristarchus’ mentor, Aristophanes.

4§53 In this connection, I note with interest what West says about Aristophanes and Callistratus taken together. [114] West’s conclusion about them is linked with his conclusion about the inaccessibility of Aristophanes’ text to Didymus: “In the final analysis, then, it remains doubtful whether Didymus had direct acquaintance with Aristophanes’ text. All his knowledge of it may have come from Aristophanes’ pupils Aristarchus and Callistratus.” [115] In this context, West revisits the wording of scholia A at Iliad I 423–424, οὕτως δὲ εὕρομεν καὶ ἐν τῇ Μασσαλιωτικῇ καὶ Σινωπικῇ καὶ Κυπρίᾳ καὶ Ἀντιμαχείῳ καὶ Ἀριστοφανείῳ ‘this is the way we found it in the Massaliōtikē and the Sinōpikē and the Kupria and the Antimakheios and the Aristophaneios’. [116] We have already seen that West is uncertain about the Antimakheios as a direct source for Didymus, and now we see that he is also uncertain about the Aristophaneios as a direct source. And yet, West persists in taking the {100|101} subject of εὕρομεν ‘we found’ as Didymus, not Aristarchus. He says: “his [= Didymus’] wording ... would suggest autopsy, as I remarked in connection with Antimachus’ copy.” [117]

4§54 In the case of the Zenodotean edition of Homer, the Zēnodoteios (Ζηνοδότειος), it is not just a probability but a certainty that Didymus had no direct access to it. When Didymus is reported as saying in scholia A at Iliad IV 3a that the variant reading ἐνῳνοχόει comes from the edition of Zenodotus, he says explicitly that he gets this information from the hupomnēmata of Aristarchus; he also says explicitly that he gets the reading ἐῳνοχόει from the ekdoseis of Aristarchus. [118] On this basis, West acknowledges that Didymus had no direct access to an ekdosis of Zenodotus and instead consulted the two ekdoseis of Aristarchus and his hupomnēmata. West concludes: “It must be accepted that Didymus’ knowledge of Zenodotus’ text—like Aristonicus’—was indirect. Most of it probably derived from Aristarchus, partly indirectly (e.g. through Aristonicus), partly directly.” [119]

4§55 West states that Didymus regularly consulted the two ekdoseis of Aristarchus and his hupomnēmata. [120] This statement requires a further observation: these two ekdoseis or ‘editions’ that Didymus attributes to Aristarchus may be described as modified base texts dating from a post-Aristarchean era and featuring the master’s preferred readings as extrapolated from his hupomnēmata or ‘commentaries’, which had originally been published in volumes separate from the volumes of the base text. Aristarchus himself in his hupomnēmata must have used as his point of reference a relatively unmodified base text—a more “standard” version, so to speak, than the versions represented by the two later ekdoseis ‘editions’ attributed to him by Didymus. The left margin of the earlier standard version was marked by critical signs that referred to variant readings as discussed by Aristarchus in his hupomnēmata, where he indicated his preferred readings and the reasons for his preferences. This earlier standard {101|102} version used by Aristarchus was similar to, though slightly different from, the still earlier standard version used by Aristophanes. Aristarchus’ system of critical signs and his verse-count of “genuine” verses were likewise similar to, though slightly different from, the corresponding system and verse-count of Aristophanes. The reasoning of Aristarchus was an exercise in balancing the evidence: he weighed the external evidence of variations in manuscripts against the internal evidence of Homeric diction as a system. Aristonicus must have used this same “standard version” as his own point of reference, systematically discussing variants signaled by the Aristarchean critical signs in its margins; these variants would be found not in the base text of the “standard version” but in the discussion provided by Aristarchus in his hupomnēmata.

4§56 With this observation in place, I proceed to one other passage cited by West as evidence to support his theories centering on Didymus. West draws attention to scholia A at Iliad III 406a, where Didymus comments on the lēmma θεῶν δ’ ἀπόειπε κελεύθους at Iliad III 406. [121] West speaks about this lēmma as the wording that Didymus had read in “his own text” of Homer. Here is the passage in its entirety: I interpret as follows: Didymus read ἀπόειπε in his received text, but he read ἀπόεικε in the two ekdoseis that he attributes to Aristarchus. Aristarchus referred to ἀπόεικε in his monographs, though he indicated somewhere else in his writings that he had manuscript evidence in favor of the reading ἀπόειπε. I infer that this indication was expressed in Aristarchus’ hupomnēmata, though even there he must have discussed the merits of choosing ἀπόεικε on the basis of the internal evidence of Homeric diction. Aristarchus paraphrases the given Homeric passage in terms of the variant ἀπόεικε. His paraphrase is quoted here directly by the scholia, which refer to the quotation as the lexis of Aristarchus.

4§57 As West reports, what Didymus “has”—ἀπόειπε—is also attested for us in “five papyri, a quotation by Aristonicus, a late papyrus commentary, and the whole medieval tradition.” [124] What, then, is exceptional about this reading? It is not that the lēmma itself is something out of the ordinary for Didymus. The reading that the lēmma gives, ἀπόειπε, is the received reading. Rather, what is exceptional in this case is simply that Didymus is not sure about the manuscript source of this received reading ἀπόειπε as given in the lēmma (θαυμάσειε δ’ ἄν τις, ἡ ἑτέρα διὰ τοῦ π πόθεν παρέδυ ‘one could wonder about where the other reading with the π came from’). Didymus is surprised because he had expected to find ἀπόειπε in the base text of at least one or the other of the two ekdoseis attributed to Aristarchus, or in one of Aristarchus’ surviving monographs. What is exceptional is that ἀπόειπε can be found ‘neither in the Aristarkheioi [= the Aristarchean ekdoseis] nor in any other of the moderate ones [ekdoseis] at any rate’ (οὔτε γὰρ ἐν ταῖς Ἀρισταρχείοις οὔτε ἐν ἑτέρᾳ τῶν γοῦν μετρίων ἐμφερόμενον πέφυκεν). [125] It seems that the surprise here for Didymus is simply that Aristarchus in this case, exceptionally, cannot be used as a source of information—either indirectly by way of the two ekdoseis or directly by way of his own writings—about the actual manuscript source of this received reading ἀπόειπε.

4§58 West draws attention to the wording καὶ οὐ μόνον ἐν ταῖς ἐκδόσεσιν, ἀλλὰ καὶ ἐν τοῖς συγγράμμασιν ἁπαξάπαντες οὕτως ἐκτίθενται ‘and not only in the ekdoseis but also in the monographs [sungrammata] all sources feature it this way [= ἀπόεικε in place of ἀπόειπε]’. [126] His interpretation of Didymus’ reasoning, as reflected in the Greek wording, runs like this: not only does everybody in their editions read it this way—ἀπόεικε—but also everybody in their monographs. For West, such wording is proof that Didymus was {103|104} regularly “consulting” his own collection of Homeric manuscripts, rather than reporting on Aristarchus’ survey of manuscript sources.

4§59 No doubt, Didymus regularly consulted the two Aristarchean ekdoseis as well as the monographs or sungrammata of Aristarchus—and most likely the ekdoseis of other scholars such as Callistratus. But I see no evidence here for a “consultation,” as theorized by West, of alternative Homeric manuscripts unknown to Aristarchus. Let us look at the wording again: καὶ οὐ μόνον ἐν ταῖς ἐκδόσεσιν, ἀλλὰ καὶ ἐν τοῖς συγγράμμασιν ἁπαξάπαντες οὕτως ἐκτίθενται ‘and not only in the ekdoseis but also in the monographs [sungrammata], all sources feature it this way [= ἀπόεικε in place of ἀπόειπε]’. The ekdoseis to which Didymus is referring here are I think primarily the two ekdoseis attributed to Aristarchus and, secondarily, to such ekdoseis as that of Callistratus. We may note that the emphasis in the Greek is on the monographs or sungrammata, in any case, and not on the ekdoseis.

4§60 Now let us look more closely at the wording that follows: προσθήσειν μοι δοκῶ καὶ τὴν Ἀριστάρχου λέξιν οὕτως ἔχουσαν· “†τί δὲ† εἰς τὰς θεοὺς ὁδοῦ εἶκε καὶ παραχώρει, μὴ βαδίζουσα εἰς αὐτούς.” I translate this way: ‘I [= Didymus] think the lexis of Aristarchus will also add [to the evidence], which has it this way’. Then, Didymus quotes Aristarchus’ paraphrase of the Homeric wording. [127] I infer that Aristarchus was showing the merits of the variant reading ἀπόεικε on the basis of contextual evidence—hence his paraphrase of the relevant Homeric passage—while allowing for ἀπόειπε on the basis of the manuscript evidence. What was out of the ordinary in this case, and what surprised Didymus here, is that the manuscript source of the reading ἀπόειπε, which I think was being implicitly supported by Aristarchus in his hupomnēmata, was left unspecified in the two Aristarchean ekdoseis. I further infer that Aristarchus—that is, Aristarchus as represented by the two post-Aristarchean ekdoseis used by Didymus—in this case exceptionally failed to provide something that Didymus ordinarily expected him to be providing. [128] In this exceptional case, Didymus could not rely on the authority of Aristarchus concerning the manuscript sources for the reading ἀπόειπε; all he could report was the master’s opinion that ἀπόεικε was preferable to the reading ἀπόειπε, at least in terms of the internal evidence provided by Homeric diction. [129]

4§61 Ordinarily, Aristarchus in his hupomnēmata would have specified whether {104|105} a reading like ἀπόεικε is attested in the manuscript sources known as the ‘more elegant ones’, khariesterai (χαριέστεραι), also known as the ‘more cultivated ones’, asteioterai (ἀστειότεραι). Further, he could have specified, by way of expressions like αἱ πλείους ‘the majority’ and αἱ πᾶσαι ‘all of them’, whether the majority or all of the khariesterai (χαριέστεραι) showed this reading. By contrast, he could have specified that ἀπόειπε is the attested reading in the manuscript sources known as ἡ κοινή (koinē) or αἱ κοιναί (koinai) ‘the common one(s)’ or κοινότεραι (koinoterai) ‘more common ones’ or δημώδεις (dēmōdeis) ‘the popular ones’ or εἰκαιότεραι (eikaioterai) ‘the more simple ones’ or φαυλότεραι (phauloterai) ‘the more base ones’. What is exceptional in this case is that the two Aristarchean ekdoseis used by Didymus had failed to specify either of these readings. [130]

4§62 I take special interest in West’s discussion of Didymus’ references to αἱ πᾶσαι ‘all of them’. [131] This discussion can actually be used to support the argument that Didymus was simply reporting on ‘all’ the privileged textual sources—that is, on ‘all’ the khariesterai, which were regularly cited by Aristarchus. I infer that even the scholiasts who quote and paraphrase Didymus understand this rationale of the expression αἱ πᾶσαι ‘all of them’. The references to the khariesterai ‘more elegant ones’ that disagree with the edition of Zenodotus, interchangeable with references to αἱ πᾶσαι ‘all of them’ that disagree with the edition of Zenodotus (scholia A vs. T at Iliad II 196c), show that the scholiasts who transmitted the work of Didymus understood their role: they were reporting on the sources of Aristarchus, not on Didymus’ own sources. [132]

4§63 As West acknowledges, the term khariesterai (χαριέστεραι) ‘more elegant ones’ is a criterion of “support” for Aristarchean editorial preferences. [133] In the scholia A at Iliad II 12a, for example, where Didymus speaks of Aristarchus’ reading being “supported” (so West) by αἱ πλείους καὶ χαριέστεραι τῶν ἐκδόσεων ‘the majority of the ekdoseis that are khariesterai’, it is evident that the tag khariesterai is an ex-post-facto validation of Aristarchean criteria. To put it another way, khariesterai—as this criterion is attested by way of Didymus—is a re-validation of Aristarchus’ earlier validations in terms of the same word. {105|106}

4§64 There is a comparable contrast made by Didymus in the scholia A at Iliad II 53a, where the category αἱ πλείους καὶ χαριέστεραι ‘the majority that are khariesterai’, in combination with ἡ Ἀριστοφάνους ‘the text of Aristophanes’ is being contrasted with the category αἱ κοιναὶ καὶ ἡ Ζηνοδότειος ‘the koinai and the text of Zenodotus’.

4§65 Here is yet another comparable contrast, where once again we see Didymus highlighting a parallelism involving the khariesterai and the Aristophaneios as a point of contrast with less privileged Homer manuscripts: We can see from the wording here that the actual combining of manuscript evidence from the khariestatai (χαριέσταται) ‘most elegant texts’ with manuscript evidence from the Aristophaneios (Ἀριστοφάνειος) ‘text of Aristophanes’ is not the work of Didymus himself. The wording that I have highlighted makes it explicit, I think, that the information provided by the khariestatai and by the Aristophaneios has been mediated for Didymus by the authority of Aristarchus himself. Didymus is saying primarily that the relevant manuscript information is to be found in the Aristarchean editions (diorthōseis must refer to the two editions that Didymus attributes to Aristarchus) and in the commentaries (hupomnēmata). In saying this, Didymus focuses on the actual reading that Aristarchus preferred: κἀν ταῖς διορθώσεσι καὶ ἐν τοῖς ὑπομνήμασιν ‘both in the diorthōseis and in the hupomnēmata’. For Didymus, it is of first priority that Aristarchus himself preferred the variant readings found in the khariestatai and the Aristophaneios. For Didymus, Aristarchus’ specific choice of a reading is of first priority, while its corresponding source is only secondary. Accordingly, Didymus gives as his first category of evidence the actual judgment of Aristarchus as reflected in his preferred reading. Only after that does Didymus give, as his second category of evidence, the testimony of the choice manuscripts: καὶ αἱ πλείους δὲ τῶν χαριεστάτων οὕτως εἶχον, καὶ ἡ Ἀριστοφάνειος ‘the majority of the most elegant texts [khariestatai] had it {106|107} this way, as also the Aristophaneios’. And only after that does Didymus add, as his third category of evidence, the supplementary testimony from (in this case) Dionysius Sidonius and Demetrius Ixion: καὶ ὁ Σιδώνιος δὲ καὶ ὁ Ἰξίων οὕτως γράφουσιν ‘also, [Dionysius] Sidonius and [Demetrius] Ixion write it this way’.

4§66 This third category is analogous to what we have already seen in scholia A at Iliad I 423–424: Καλλίστρατος δὲ ἐν τῷ Πρὸς τὰς ἀθετήσεις ὁμοίως, καὶ ὁ Σιδώνιος καὶ ὁ Ἰξίων ἐν τῷ ἕκτῳ Πρὸς τὰς ἐξηγήσεις ‘Callistratus in his volume Πρὸς τὰς ἀθετήσεις has it in a similar way; also [Dionysius] Sidonius, and [Demetrius] Ixion in the sixth volume of Πρὸς τὰς ἐξηγήσεις’. Just as Didymus must have been the originator of the statement highlighting Callistratus, Dionysius Sidonius, and Demetrius Ixion, the same can be said about the statement in scholia A at Iliad II 192b highlighting, again, Dionysius of Sidon and Demetrius Ixion. Conversely, I argue that Aristarchus must have been the originator of the preceding statement in scholia A at Iliad II 192b: καὶ αἱ πλείους δὲ τῶν χαριεστάτων οὕτως εἶχον, καὶ ἡ Ἀριστοφάνειος ‘and the majority of the most elegant texts [khariestatai] had it this way, as also the Aristophaneios’. So also I ascribe to Aristarchus the correspondingly preceding statement in scholia A at Iliad I 423–424: οὕτως δὲ εὕρομεν καὶ ἐν τῇ Μασσαλιωτικῇ καὶ Σινωπικῇ καὶ Κυπρίᾳ καὶ Ἀντιμαχείῳ καὶ Ἀριστοφανείῳ ‘this is the way we found it in the Massaliōtikē and the Sinōpikē and the Kupria and the Antimakheios and the Aristophaneios’.

4§67 Working my way backwards in the scholia A at Iliad I 423–424, I now focus on the even earlier statement preceding the one I have just quoted about the Massaliōtikē, the Sinōpikē, the Kupria, the Antimakheios, and the Aristophaneios. In this earlier statement, which even West identifies as a direct quotation from Aristarchus, we see that the master adduces Sophocles (after having adduced “Homer” himself): The wording οὕτως ‘this is the way’ in this statement, which signals the adducing of Homer and Sophocles, is picked up by the wording οὕτως δὲ εὕρομεν ‘this is the way we found it’ in the next statement, which signals the adducing of the Massaliōtikē, the Sinōpikē, the Kupria, the Antimakheios, and the Aristophaneios. Elsewhere too in the Homeric scholia, the {107|108} adducing of evidence signaled by οὕτως εὕρομεν ‘this is the way we found it’ and related expressions conveys primarily the findings of Aristarchus, and only secondarily the findings of the Aristarchean speaker. Here are two telling examples where the speaker is the Aristarchean scholar Herodian, who flourished about 200 years after Didymus:

4§68 Having examined the evidence of the Homeric scholia concerning the criteria of Aristarchus and Didymus, let us return one last time to the wording of scholia A at Iliad I 423–424, οὕτως δὲ εὕρομεν καὶ ἐν τῇ Μασσαλιωτικῇ καὶ Σινωπικῇ καὶ Κυπρίᾳ καὶ Ἀντιμαχείῳ καὶ Ἀριστοφανείῳ ‘this is the way we found it in the Massaliōtikē and the Sinōpikē and the Kupria and the Antimakheios and the Aristophaneios’. On the basis of what we have seen, the communis opinio is confirmed: the subject of οὕτως δὲ εὕρομεν ‘this is the way we found it’ is Aristarchus. It is not Didymus, as the theory of West would have it. [136]

4§69 I conclude that it was Aristarchus, not Didymus, who developed a system for the collating of Homeric texts. And it was Aristarchus who led the way in searching for significant variations in the history of the Homeric textual tradition. [137] {109|110}