This is the first volume of a two-volume project undertaken by Silvia Fazzo, which will include, together with a new edition of the Greek text and its translation into Italian, a detailed commentary on book Lambda of Aristotle’s Metaphysics. Fazzo has long been working on the Metaphysics and its interpretation through Antiquity and the Middle Ages and is most suited to producing such a commentary. Her edition of the Greek text undoubtedly constitutes progress with regard to previous editions in that it records readings of two manuscripts that had not been collated by previous editors, namely Ambrosianus F 113 sup. (M) and Taurinensis VII. B. 23 (C), and adduces more carefully than its predecessors the testimony of the Arabic translations by Ustāth (9 th cent.) and by Mattā ibn Yûnus (10 th cent.), and of the Byzantine commentary of Michael of Ephesus (11 th /12 th cent.). Fazzo says, and this seems true,1 that her edition of Lambda is the first to be made on a completely stemmatic basis (‘condotta su base interamente stemmatica’, 19). It remains to be seen, however, whether the stemma codicum upon which her edition rests is correct.
Before going into the substance, it would be useful to give a brief outline of the manuscript studies on which previous editions of the Metaphysics have been based, notably those of David Ross (1924) and of Werner Jaeger (1957) that are generally used. Fazzo helpfully discusses in her introduction (35-110) all editions of the Metaphysics from Christian Brandis (1823) onwards. It was Alfred Gercke who in 1892 called attention to an old Aristotelian manuscript owned by the National Library in Vienna and who thus elevated the number of the authoritative manuscripts of the Metaphysics to three: next to Parisinus gr. 1853 (E) and to Laurentianus plut. 87,12 (A b), already known to previous editors, one should henceforth also collate Vindobonensis phil. gr. 100 (J).2 Gercke also deserves credit for showing that these three manuscripts represent two independent textual traditions, one (α) being represented by E and J and another (β) by A b, and to have thus established the ‘règle d’or’ for any subsequent editor of the Metaphysics : accordance between E and A b or between J and A b guarantees in principle the reading of the archetype. The editor’s skill must therefore be applied primarily to cases in which the two traditions of the text diverge. Ross hesitated as to the superiority of one or other tradition, but Jaeger saw in the common ancestor of J and E a Byzantine ἔκδοσις and thus privileged in his edition what he took to be the more ancient testimony of A b. Things changed profoundly about a century after Gercke’s research, when Dieter Harlfinger identified two further independent witnesses of the β-family, namely M and C, and showed that from Λ 7, 1073a1 onwards A b ceases to represent the β-tradition and becomes a copy of a lost manuscript closely related to J, namely δ.3 Therefore, it is only M and C that allow us to reconstruct β for this part of the Metaphysics. These manuscripts have been collated by Concetta Luna for books Mu and Nu.4
Fazzo gives now an additional reason for extending the interest of the testimony of M and C not only to Λ 7-10 but to the whole of Lambda. She claims that the change of model in A b, i.e. from β το δ, did not occur at Λ 7, as Harlfinger believed, but at K 8, 1065a26, where the Metaphysics becomes more or less a collection of extracts from the Physics.5 The novelty, then, that Fazzo’s edition of Lambda ventures to introduce is the downgrading of the testimony of A b, which in her opinion has been overestimated by previous editors (‘la differenza indubbiamente più importante riguarda il codice A b il cui peso è stato sopravvalutato dagli editori del XIX e XX secolo’, 19). Indeed, admitting that A b belongs to the α-tradition and that β is solely represented by M and C changes profoundly one’s editorial choices: the accord of E and J against the derivative A b must always be preferred and, a fortiori, the accordance of the two independent families EJ and MC must also always be preferred to the testimony of A b. In practice, Fazzo does away in her edition with A b and records its readings in the critical apparatus mainly for historical reasons.
Fazzo does not seem aware of the difficulty posed by admitting a change of model in A b at K 8. Contrary to her thesis, Harlfinger’s rests on safe palaeographical grounds: δ has been used as a model for A b by a scribe who continued the copy of the Metaphysics around the middle of the fourteenth century, because this copy stopped abruptly at Λ 7, 1073a1 at the end of the eleventh. To accept Fazzo’s thesis, one has first to accept that δ was re-used as a model for A b after approximately 250 years had elapsed. Even if one is willing to accept this rather implausible idea, one will still need a justification for Fazzo’s ungrounded assumption that M and C are faithful copies of β. Had Fazzo undertaken collations of other manuscripts of the Metaphysics, she would have come to the conclusion that M and C are copies of a nowadays lost manuscript which has been corrected according to a manuscript of the α-family, namely Marcianus gr. 214 (H a). The following cases in Lambda, which in Fazzo’s edition can be assessed as genuine β-readings, derive from H a : 1069b2 κοινὴ ΕJA b : κινήσεως H a MC
1069b23 ἡμῖν ΕJA b C : ὁμοῦ H a M
1070a14 συνθέτην] σύνθετον H a M
1070a32 ὡς EJ, ὥς A b C : ὥς οὔ H a M
1072a11 εἶναι ἀεὶ ἐνεργοῦν A b : ἐνεργοῦν εἶναι EJ : ἐνεργοῦν H a ΜC
1073a1 τὸ] om. H a M, add. s.l. C
Granted that M and C are copies of a Byzantine edition in which both α- and β-readings have been taken into account,6 Fazzo’s editorial rationale falls apart: the concord between ΕJ and MC has in itself no value for establishing the reading of the archetype and A b regains its authority as β-manuscript. Unfortunately, Fazzo has not rightly appreciated the relatively frequent concord between the Arabic translations and A b before the palaeographical break at 1073a1 (cf. 1070a36, 1070b7, 1070b 21, 1072a26-27, 1072a30), which confirms the non-Byzantine origin of its readings, nor has she further exploited the collations made on E by a scholar in the 12 th century (E Σ), whose accordance with A b confirms in all probability the β-reading. Another manuscript essential to establishing the text of β is Vaticanus gr. 255 (V d), which has not been taken into account by Fazzo. While this manuscript belongs to the α-family, a later hand (V d2) has corrected its text according to a β-manuscript, which was presumably the model of A b. I quote here some examples from Lambda : 1070a1 τι] τι καὶ ἐξ οὗ A b V d2
1070a18 δὴ] om. A b, del. V d2
1072b2- 3 καὶ τινός V d, πρόσκειται ἔν τισι τὸ καὶ τινός E Σ : τινός A b : om. ΕJH a C (in lacuna M)
Against such a background, it is fortunate that Fazzo, following a suggestion by Michael Frede and Günther Patzig that has been lately confirmed by Oliver Primavesi,7 has privileged in her edition the α-tradition against the ‘smoothed’ text of the β-tradition. This, however, does not mean that A b must always be in the wrong. I can here provide only one example of A b ’s pointing to the right reading (as was seen by both Ross and Jaeger) but necessarily dismissed by Fazzo in virtue of her stemma:
1072a11-12 δεῖ ἐνεργοῦν εἶναι ΕJ : δεῖ ἐνεργοῦν H a ΜC : δεῖ εἶναι αἰεὶ ἐνεργοῦν A b
It seems clear that the archetype read ΔΕΙΑΕΙΕΝΕΡΓΟΥΝΕΙΝΑΙ,8 the ἀεὶ having been omitted in α(EJ) due to a confusion of the triangular letters Δ and Α and a subsequent haplography, while it was conserved but mutated to the more poetic αἰεὶ in A b, as is usual in the β-tradition (εἶναι has been also transposed).
I have been concerned so far with Fazzo’s assessment of the β-tradition. With regard to the α-tradition, Fazzo notes (151) that previous editors made no distinction between what she calls J, i.e. the original copyist of J, and J 2, a scholar who collated another manuscript and added variant readings some 150 years later, according to Fazzo’s dating. I assume that previous editors have not made that distinction because J 2 is identical with J. This can be clearly seen, for instance, in 1069b34-35, where what Fazzo describes as “τὸ δὲ τρίτον ἡ ὕλη habuit J erasit et in mg. fecit J 2 | post ὕλη (a34) μετὰ–ὕλη om. J supplevit J 2 ”, is only a self-correction of J in f. 185 v, who had to use some space in the margin in order to copy a portion of text that he had for once omitted. As always, correct palaeographical assessment is highly relevant to textual criticism. In this particular case, it allows us to see that two models were used in order to produce J. Had Fazzo recognised this, she would perhaps have nuanced her strong commitment to A b ’s following the model of J; quite to the contrary, it is J that seems to occasionally follow the model of A b.
All in all, the stemmatic basis upon which Fazzo’s edition of Lambda rests is not tenable.
Let me finally note some omissions and factual errors: In 1069b3 Fazzo reads ἡ δὴ αἰσθητὴ but all manuscripts have unanimously ἡ δ’ αἰσθητὴ. The variant that she records in the apparatus concerns the beginning of the next clause: εἰ δὲ/δὴ μεταβολὴ.
In 1070a14 συνθετὴν is not attested in any manuscript; lege συνθέτην.
In 1070a32 C has ὥς and not ὥς οὔ.
In 1070b22 the variant τὸ ἐκτὸς for τῶν ἐκτὸς, which M shares with another manuscript not taken into account by Fazzo, namely Marcianus gr. 211 (E b), is not recorded in the apparatus.
In 1071b9 E pc reads not ἀλλ’ ἢ, as M and C read, but ἀλλ’ ἡ together with J and A b; E ac read ἀλλη.
In 1072a35 E ac had συστοιχεία, which is also the reading of A b (not recorded by Fazzo).
Fazzo notes (131-132) that the β-text is transmitted in parallel with Michael’s commentary. This is manifestly false. The so- called fragment Y (Parisinus Suppl. Gr. 687), which is perhaps the most ancient surviving manuscript of the Metaphysics, exhibits solely the β-text, without any commentary. Nor is it true that β had not acquired its actual form by the time of Michael. Since Michael had access both to α- and to β-text, as is made clear by his discussions of variant readings, his commentary is not an absolutely safe guide for reconstructing β.
[For a response to this review by Silvia Fazzo, please see BMCR 2013.08.17.]
1. A new edition of Lambda has also been published, albeit in a limited number of exemplars, by S. Alexandru, Aristotle’s Metaphysics Lambda. Annotated Critical Edition Based upon a Systematic Investigation of Greek, Latin, Arabic and Hebrew Sources, Athens, 2011(and now also published by Brill in late 2013).
2. A. Gercke, ‘Aristoteleum’, Wiener Studien 14 (1892), 146-8.
3. D. Harlfinger, ‘Zur Überlieferungsgeschichte der Metaphysik ’ in P. Aubenque (ed.), Études sur la Métaphysique d’Aristote (Paris, 1979), 7–36.
4. C. Luna, ‘Le texte des livres M-N de la Métaphysique ’, Documenti e studi sulla tradizione filosofica medievale 16 (2005), 553- 93.
5. Fazzo reproduces in her book the thesis presented in her ‘Lo stemma codicum dei libri Kappa e Lambda della Metafisica: une revisione necessaria’, Aevum 84 (2010), 339–59. I deal with the flaws of this thesis in my ‘Collation but not contamination: on some textual problems of Aristotle’s Metaphysics Kappa 1065a25sqq.’ (forthcoming).
6. C is only an indirect copy of that edition, having been directly copied from a lost manuscript that bore supplementary α-readings.
7. M. Frede and G. Patzig, Aristoteles Metaphysik Z: Text, Übersetzung und Kommentar (Münich, 1988), vol. 1, 16; O. Primavesi, ‘Aristotle, Metaphysics A: A New Critical Edition with Introduction’ in C. Steel (ed.), Aristotle’s Metaphysics Alpha (Oxford, 2012), 387-464, esp. 439.
8. See 1050b22, 1050b29 and 1071b32 for similar wording.