More than a year has passed since Antonios Rengakos reviewed two chapters of my book Studies in the Text and Transmission of the Iliad (Munich & Leipzig 2001). I have been waiting to see if he or someone else was going to review the rest. In the meantime there has appeared a twenty-page review by Gregory Nagy in Gnomon (75, 2003, 481-501), the greater part of which is focused on the same two chapters of my work. Seeing that his criticisms have a certain amount in common with R.’s, and the two of them have colluded (N. showed R. his review before publication, and R. leans on it), perhaps I may take the liberty of responding to both in this forum.
As it is the first duty of a reviewer to give readers an idea of what is contained in the book under review, and neither R. nor N. has done this adequately, let me begin by doing it on my own behalf. In chapter I.1, “The Pre-Alexandrian Transmission”, I discuss the archetypal text of the Iliad, the form it took, the circumstances of its early diffusion, rhapsodic interpolation and other sources of textual instability, the role of the Homeridai, the Panathenaic recitations, the Athenian book trade, the question of μεταχαρακτηρισμός, the Homeric criticism of the Sophists, and the exposure of the text to Atticism and modernization. For the next two chapters see below. Chapter I.4, “The Papyri”, contains a catalogue of the 1356 Iliad papyri and 185 accessory ancient documents used in my Teubner edition, which is more than twice the number listed in any other book. Chapter I.5, “The Early Medieval Transmission”, gives an account of the medieval manuscripts used and their mutual affinities. Here too I can claim to have taken the subject some way forward. Chapter I.6, “The New Teubner Edition”, explains my aims and methods, the objectives of a modern critical edition, and the principles followed in collecting, assessing, and presenting the evidence. Part II consists of over a hundred pages of detailed discussions of textual matters. Some are to do with the authenticity of certain lines or passages, which sometimes raises questions of the narrative structure. Others are concerned with readings and so with particularities of Homeric language and usage. There is a rich buffet here for future commentators.
The chapters with which my two critics have taken issue at length are I.2 “Zenodotus’ Text” and I.3 “Didymus and his Sources”. It must be some time since a contribution to the study of Alexandrian Homeric scholarship has caused such consternation. What are my outrageous theses? In brief: (1) that the great majority of the readings attributed to Zenodotus were neither his conjectures nor his considered choices but the readings of a fourth-century Ionian rhapsode’s copy that Zenodotus owned and annotated; (2) that neither Zenodotus nor Aristophanes nor Aristarchus set about his editorial work, as everyone since Wolf has assumed, by assembling and systematically collating numbers of manuscripts in order to compare their readings. It was Didymus who compiled the “critical apparatus” reflected in our scholia, with citations of half a dozen “city texts” and half a dozen other individually identified exemplars. He did not have personal knowledge of all these copies but took information about some of them from his contemporary Seleucus or from earlier commentators.
Both R. and N. give slightly misleading accounts of my views. R. says that I “cautiously side with those who believe that Alexandrian Homeric criticism was limited to conjectures”; I have not said and do not believe any such thing. He says that I “also suggest that the Aristarchean readings that differ from both the vulgate known to Didymus and the medieval vulgate are rarely correct (1998: VII).” What I wrote in the place cited (the preface to my edition) was that such readings are sometimes superior but more often inferior. No inquirer without an Aristarchean bias, I think, would argue otherwise. N. (488) says that in refusing to see much sign of rational thought or choice in Zenodotus’ text, I slight its value as a source for the textual history of Homeric poetry. On the contrary: precisely because Zenodotus’ revision of his base text was so desultory, it remains a valuable reservoir of older variants that are frequently interesting and occasionally convincing. In more than one place N. implies that I have modified my positions in response to his previous critique of the first volume of my edition (BMCR 2000.09.12). I have not. I have simply explained and justified them more fully.
As regards Zenodotus, R. writes that my hypothesis sounds plausible but “is not borne out by the available evidence”. This is odd, as the available evidence (as distinct from the scholarly fable convenue) is exactly what my hypothesis is based on. What contrary evidence does he have in mind? Firstly he refers to the Homeric allusions in the Hellenistic poets, which sometimes agree with variants found in early papyri or anticipate Alexandrian scholars’ readings. “All this cumulative evidence points to the conclusion that the Alexandrians must have compared different copies of the Homeric text available to them and must have chosen among variae lectiones.” No it doesn’t; it only shows that many of their readings were previously current in the tradition, which is unsurprising. Secondly R. states that “there is direct evidence that the great Alexandrian philologists relied on manuscript authority”. In support of this claim he adduces sch. Il. 9.222, a passage which I discuss below, and then appeals to the fact that in some Ptolemaic papyri variants have been inserted. Of course it was commonplace for manuscripts after copying to be checked for error against another source. But this is not evidence, let alone “direct evidence”, that scholarly editors of the third and second century BC systematically collected and compared multiple copies in a search for the best readings, any more than did those of the sixteenth century AD. Aristarchus of course studied Aristophanes’ text, and Zenodotus’. He cited these two often, and others hardly at all. But the concept of manuscript authority nowhere appears in the quotations from his writings, nor in Aristonicus’ paraphrases of his teaching. He argues for particular readings on the grounds of their intrinsic merit, never because of the number or quality of the sources in which they appeared. It is Didymus who does the listing of sources.
There are a number of scholia, mostly from Didymus, that have been, and are still being, claimed as evidence of Aristarchus’ manuscript studies. Let me fly them by and shoot them down. (1) Sch. Il. 1.423-4. The first part is an explicit quotation from Aristarchus (introduced as λέξις Ἀριστάρχου, an expression misinterpreted by N. 490); the last part is undeniably Didymus’ own. Is the intermediate sentence, “And we have found this reading in the Massiliotica”, etc., part of the quotation, or the start of Didymus’ supplementary comment? On pp. 70-1 of my book I show that Bekker and Bergk were right in taking it as the latter, and I quote five Didymean parallels for the “we have found”. Nagy uses up more than two pages of Gnomon (491-3, small type) in a vain attempt to transfer this Didymean idiom back to Aristarchus. (2) Sch. Il. 6.4a: Aristonicus (copied by Didymus) reports that, after assuming one reading in his Hypomnemata, Aristarchus came across a different one and approved it (probably in his monograph περὶ τοῦ ναυστάθμου). This sounds like a casual discovery; at any rate it is not evidence of any systematic study of manuscripts. (3) Sch. Il. 9.222b: I dealt with this on p. 37 n. 19 of my book. Aristarchus had remarked that ἂψ ἐπάσαντο or αἶψ’ ἐπάσαντο would have been better than ἐξ ἔρον ἕντο; but, says Didymus, “out of exceeding caution he made no change, as he found the latter current in many copies.” This may be entirely Didymus’ inference, or Aristarchus may have said something like “but I have not encountered such a variant in any text I have seen”. Again, it is not evidence of systematic collation of numerous exemplars. (4) Sch. 9.401b: “Aristarchus says that some wrote ἐμῆς ψυχῆς, which is not to be commended.” “Some” need not imply more than a single manuscript; it might have been Zenodotus’. It might have been not a manuscript at all, but a quotation that Aristarchus knew from somewhere. (5) Sch. Il. 18.10-11a (“that the best of the Myrmidons, with me still alive, would depart from the sunlight at the Trojans’ hands”): “the two lines were not in Rhianus’ text, perhaps because Patroclus was not a Myrmidon but a Locrian from Opous. Aristarchus says that the phrase is to be taken as ‘best of the Myrmidons after (Achilles) himself’.” According to Rengakos, “this is clearly Aristarchus’ retort to Rhianus’ assumed objection”, and therefore Aristarchus made direct use of Rhianus. No: Aristarchus is answering the difficulty that “the best of the Myrmidons” was indisputably Achilles, not Patroclus. His interpretation has no connection with the view that Patroclus was not a Myrmidon. Moreover, Rhianus’ omission of the lines, like many peculiarities of his text, was shared by Aristophanes (sch. 10c), who may well have mediated the information; see my p. 57. (6) Sch. Il. 19.386a/b: Didymus reports that Aristarchus first wrote εὖτε but later preferred αὖτε on poetic grounds. He also says that the latter reading was found in the city texts (that is, probably, in one or two of them; see my p. 67). R. makes this into “Aristarchus even cited a reading from the city editions … Didymus himself states that the second reading came from the city editions.” But Didymus does not say that it came to Aristarchus from that source, only that it was to be found in that source. All we are entitled to say is that it was current in some Hellenistic texts; we cannot assume that Aristarchus found it just where Didymus found it.
R. further criticizes my characterization of Zenodotus’ text. He takes me to task for calling it “abbreviated”, saying that Nickau “has successfully refuted this very widespread misconception: according to his calculations Zenodotus’ text was at most 138 lines shorter than the ‘Aristarchean vulgate’ … very little in a poem of approximately 15,700 lines”. What I wrote was (and it is the truth) that “it more often had an abbreviated text [than an expanded one]: in over fifty places it omitted lines or passages which are present in the rest of the tradition known to us … Most of the omissions are of just one or two lines, or not more than five, but there are five or six omissions of over ten lines and one of over twenty” (33). Then R. accuses me of not distinguishing between athetesis and omission, when in fact I take pains to emphasize the distinction in n. 5 on the same page. On the next page I mention that “other features of Zenodotus’ text were the appearance in it of various neo-Ionic forms and other late morphological phenomena; at least ten cases of dual verb forms or participles used as plurals; and a recurrent tendency to use irregular forms of pronouns.” In the course of the chapter I argue that these were not personal quirks of Zenodotus but reflected tendencies of post-Homeric hexameter poets that had affected the practice of Homeric rhapsodes and that Zenodotus’ was essentially a rhapsode’s text, the neo-Ionic forms pointing to Ionia, where Zenodotus himself came from. With these reasonable and not altogether original inferences R. is determined to quibble, even calling in aid the ridiculous arguments of Van der Valk against the existence of Ionian texts. But after all his quibbling he offers a summary of “the consensus today” which largely agrees with my position: that “Zenodotus made his textual interventions on a base-text to which he introduced obeloi and other signs to mark omissions or to introduce selected readings. This base-text may very well have been of Ionian extraction …” We agree that Zenodotus introduced his own conjectures in places, though R. probably thinks he did so on a larger scale than I see evidence for. The major difference between us is his insistence on the conventional doctrine that Zenodotus collated manuscripts.
Nagy continues to be flustered by my insistence on distinguishing between a tradition of oral poetry and the transmission of the written Iliad, two things that it is his life’s mission to confuse. I keep trying to put him straight (BMCR 2001.09.06; Studies 159 n.2, 160 n. 5), but he does not get the point. The first six pages of his review are directed against my notion that we can speak of an original author of our Iliad and distinguish this author from later rhapsodes who amplified or otherwise altered the text. He implies that this is an outlandish claim, based on an ignorance or wilful neglect of Oral Poetics that will horrify the pious readers of Gnomon. “Those who study living oral poetic traditions know from experience that different available versions cannot be reduced to a single basic version from which all versions are to be derived. W. claims 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” (482).
This is not what I claim to do, since in my view (and that of most Homerists) what the tradition gives us is not a “mass of multiple versions” but a single version that has suffered trivially from interpolation and corruption. That single version may indeed bear the imprint of differing versions that its author, an oral poet by profession, had recited on different occasions. But he laboured to produce a single, coherent written version, and, the more deeply one studies the Iliad, the more one is impressed by the measure of his success in doing so. Those who corrupted and interpolated his work between 600 and 300 BC were certainly, in the main, oral performers. But if they sustained a live tradition of “oral poetry” as understood by Parry and his school, it went unrecorded and is irrelevant to the transmission of the Iliad and Odyssey, since there is no evidence for any materially “different versions” of these epics, only for trivial variation and amplification at the verbal level. It is apparent that from an early date the public required these performers to recite, not their own poetry, but the poems of “Homer”. They were expected to know and reproduce — not recompose — classic texts that were also available in written form; not to be word-perfect, but to relate the familiar episodes pretty well as they stood in the books, with no material deviation. Naturally they produced oral/mnemonic variants of phrasing, and occasionally added a few lines of their own. But this is not oral poetry in the sense of composition in performance. N. really must try to get his capacious head round the difference.