We gain two great benefits from this new editio maior of Aristotle’s Poetics : a careful working- through and discussion of the Arabic witness to the text, and, consequently, an excellent apparatus criticus. 1 The Greek text itself is good and the detailed textual discussions are of lasting value. The text they edit against, if you will, is Kassel’s OCT, which, Tarán argues, fails on three counts: it does not provide an adequate apparatus, holds too closely to the text of A, and does not take sufficient account of the Arabic. Tkatsch’s edition with commentary was also deeply flawed. Tarán and Gutas have collaborated to remedy these deficiencies. I should note here that I do not know Arabic. Gutas takes care to make his presentation of the material clear to classicists (he is almost always successful), but some points of Arabic philology are necessarily opaque to the uninitiated.
The Poetics, more than any other work of Aristotle, has been the object of numerous editorial strategies. Tarán is fairly conservative. He completely rejects, for example, the transpositions of de Montmollin and Else, based as they were on elaborate hypotheses about the composition and revision of the text, and he tends to be sparing with emendations. The text may not be satisfactory to everyone, but even more liberal critics will value it as a clear and well-supported statement of what the archetype contained. The apparatus is the most detailed ever published. 2
The book is divided into an introduction, the text, and two textual commentaries. The first section of the introduction is Tarán’s “History of the Text of the Poetics.” He is as full as possible on the history of the text from Aristotle through the 19 th century, and casts doubt on the idea that Aristotle’s works were unavailable in the Hellenistic era, by argument from probability rather than by doctrine as Sandbach does.3 Most of the history of the editorial work is handled here. Tarán rightly criticizes the unmethodical textual work of earlier scholars and traces how some misunderstandings or unwarranted assumptions have wreaked havoc with the text and its interpretation.
The second section of the introduction is Gutas’ “The Poetics in Syriac and Arabic Transmission,” which is very enlightening for a classicist. He collects all of the Syriac and Arabic testimonia to the Poetics and writes a detailed history of the text. According to Gutas, the first translation into Syriac was in the ninth century and was revised. Abū-Bišr Mattā ibn-Yūnus translated this revised Syriac version into Arabic. Finally, Avicenna indicates that the Arabic translation was itself revised twice. There are some additional quotations of it in Averroes, who quotes liberally, probably from the first revision, and is therefore occasionally useful. The second, more thoroughgoing revision of Abū-Bišr’s version implies the use of a second Greek MS. On these grounds, Gutas posits a MS Σ, the basis for the original Syriac translation, as well as Ψ, the MS used for the second revision of the Arabic. The stemma on page 110 reveals a world of Near Eastern interest in Aristotle which most, I think, would not suspect existed. Unfortunately, only one MS (Parisinus Arabus 2346) of the original Arabic translation, sloppily copied and damaged, exists.
Gutas provides a history of the scholarship on the Arabic text and discussions of previous publications of the Arabic. His comments on the Greek text show the gains from his reflection on the Arabic and point out the difficulties of using the Arabic translation, but also the possibilities; examples of misuse of the Latin translations provided by Margoliouth and Tkatsch are informative of the difficulties. In spite of all that, he manages to elicit a great deal of information relevant for the editing of the Greek text. In place of the three sigla (Σ, [Syr], [Ar]) in Kassel’s edition for the Syro-Arabic tradition, there are now five sources of text (Σ, [Syr], [Ar], Ψ, Σ q) plus a much more subtle and detailed understanding of the “behind-the-scenes” of the Arabic MS, which bears fruit in his “Greco- Arabic Critical Apparatus and Commentary.”
Tarán’s “Prolegomena to the Edition of the Text” makes up the third part of the introduction. The MSS used are the same as in Kassel’s edition; nevertheless, their selection is put on a more rigorous basis. Tarán also provides the evidence for his assertions about the MSS, including A, which is the oldest MS of the text extant and must on those grounds be a primary MS. The MSS are given very full descriptions and histories. Some of his assertions may go too far: for example, he claims (p. 139) that A and Φ descend from different transliterations into minuscule on the basis of six errors, four of which are errors of word division. But three of these also include phonetic errors, and further, early minuscule did not divide words obviously nor were breathings and accents regularly written.
The text is not radically different from Kassel’s. There are, to be sure, cumulatively many small changes: several changes of punctuation, different variants chosen, two new emendations by Tarán, and different choices among emendations, as well as new evidence from the Arabic, but the overall thrust of the investigation of the MS tradition was to confirm the basic uniformity of the tradition.
The MS basis for the text has not changed: it relies on the two Greek MSS, A (Parisinus 1741, 10 th century) and B (Riccardianus 46, which has been redated to the 12 th century), William of Moerbeke’s Latin translation, the Arabic version, and a few quotations. Further, Parisinus Graecus 2038 (15 th century) is cited several times for good readings (which Tarán thinks are due to good conjectures on the part of the scribe), as well as the anonymizing “rec.” for the undifferentiated mass of recentiores. I disagree with Tarán about the quality of Par. 2038, which I consider the result of contamination between a descendant of A and a lost MS. The readings of Par. Gr. 2038 are thus traditional and very important, as both Kassel and Tarán recognise by adopting many of them.4
The commentary (chapter 4) is dedicated to explaining the textual decisions made. The individual notes are brief and intended to fill out the spare information which the apparatus gives us. It is a reference of first resort for users of this, or any, edition of the text, because Tarán justifies his own text and handles objections. Gutas’ work on the Arabic text is digested and used to edit the Greek (his notes are referred to whenever a difficulty appears).
Generally, Tarán tries to explain errors palaeographically, especially errors from misreading majuscules, and has faith in the scribes’ sincerity. His reliance on these misreadings to explain variants is notable; at 54b37 fore in the Latin translates ἔσεσθαι, erroneously for the correct αἰσθέσθαι: ΑΙϹΘΕϹΘΑΙ was misread as ΑΙϹΕϹΘΑΙ and then mispronounced as ἔσεσθαι. The initial vowels had become identical in pronunciation by the Byzantine era. Many errors are blamed on homoioteleuton which other scholars might blame on homoiarchon, haplography, or sauts du même au même.
Gutas’ “Greco-Arabic Critical Apparatus and Commentary” (chapter 5) will be of greatest use to those who know Arabic, but it is still useful for the traditional classicist, if for no other reason than to register Gutas’ confidence (or lack thereof) in the Greek he reconstructs and his discussion of other possibilities. The Arabic is usually presented in transliteration, but occasionally in Arabic script, usually to show how misunderstandings could have arisen through orthographic similarity. Some entries merely reproduce the apparatus (e.g. “50a18-19 οὐ ποιότης non vert. Ar.”). Others add the Arabic which supports the reading of the Greek (e.g. “50a19 εἰσὶν Π, Σ ( wa-hum) : ἔστιν B”), which will be useful for Arabists working on the text. The variety of qualifiers ( non vertit, ut intellexit, ut interpretatus est, ut legit, dubitanter/dubitavi and ut videtur) and detailed philology give nuance to what has been previously used as a blunt instrument. This sometimes requires a leap of faith on the part of the reader – a case in point is Gutas’ discussion of ὥσπερ ποτὲ καὶ ἄλλοτέ φασιν (51a9), for which the Arabic gives “as is our custom to say at some time and when.” Though the phrase ποτὲ καὶ ἄλλοτέ (“sometimes, on some occasions”) was misunderstood, it seems clear enough to a layman that “at some time and when” renders that phrase, though in reverse order; for the rest, note M. Schmidt’s emendation of φασιν to εἰώθασιν, which, as Gutas notes, is the Greek word expected for the Arabic we have. He nevertheless argues that the Arabic reflects φασιν, though his “dubit.” in the apparatus is telling. On the other hand, his commentary on 55a14ff. is perfectly clear, easy to follow, and compelling, even though the discussion is about a slightly corrupt Arabic translation of problematic Greek.
Briefer notes to specific passages: 48a16: I cannot understand μιμήσαιτο ἄν τις (the rest of the sentence is “similarly with dithyrambs and nomes, like Timotheus and Philoxenus in their Cyclopes”. We want a verb like ἐποίησαν (“just as T. and Ph. composed their Cyclopes ”) or ἐμιμοῦντο (“just as T. and Ph. were imitating cyclopes”) in its place. As Richard Janko suggests (pers. comm.), this phrase should be deleted as an anticipation of the same phrase below at 20.
52a31: Tarán rightly corrects Kassel’s recording of B, but the entry in his apparatus is itself possibly misleading: “ἢ εἰς φιλίαν ] = B.” But this implies that the other MSS have something different, which is false; it is the reading of all MSS.
53a19: The apparatus is misleading about Parisinus 2038’s reading by suggesting that it only had τινα without ᾖ; it actually reads τινὰ φαῦλον μὲν ἐὰν φαύλην (without ᾖ). Tarán’s discussion in the commentary requires knowledge of the full reading to be understood.
The text and apparatus are free from typos as far as I noticed.5
In sum, this is a valuable tool for those working on the Poetics and should become the edition of reference. Gutas’ treatment of the Arabic text and Tarán’s work on the MSS are exemplary. Hopefully a less expensive version will be made available.
1. Gutas is responsible for chapter 2, the history of the Syro-Arabic tradition and chapter 5, the Graeco-Arabic Critical Apparatus and Commentary. Tarán is responsible for the rest.
2. A version of the Leiden conventions is used for reporting the readings in B where it is physically damaged. N dashes stand, evidently, for a single letter. Tarán’s arguments about the relationship of the MSS extend only to the primary witnesses; for the later tradition, we have to see Lobel and Harlfinger-Reinsch.
3. Aristotle and the Stoics (Cambridge Philological Society, 1985).
4. M. Centanni ( Il testo della Poetica nel Par. Gr. 2038 in Bollettino dei classici 7:37-58, 1986) already made this point; I am preparing an article now to reinforce it.
5. There were a handful of slips elsewhere: Add dub. or dubit. to sigla.
47a25: τυγχάνουσιν is given to “rec.” in the apparatus; it is found in Parisinus Graecus 2038.
p. 105 al-Farabi for a-Farabi, middle of page
p. 231, note on 47b14: Tarán states that Janko proposed the addition of τοὺς μὲν; it is Gryphius’.
p. 244, note on 49a28-30, for ἔσται, read ἔστω.
p. 251, note on 50a29, for ῥέσεις, read ῥήσεις.
p. 266, note on 54a19, for χρεστόν, read χρηστὸν.
p. 270, note on 54b37, for “weaved,” read “woven.”
p. 276, note on 55b6, for ἱεροσύνη, read ἱερωσύνη and for “sanctuary,” read “priesthood.”
p. 281, note on 56a16, for similii, read sinilii (sic, translated from a misdivided word συνιλίου).
Centanni 1986: “ bollettino ” is misspelled in the bibliography.