I appreciate Professor Moreschini’s taking the trouble to reply to some points raised in my review of his new edition of Cicero, De finibus. In doing so, he reaffirms the value of scholarly dialogue; in that same spirit it may be worth my trying to clarify a few matters as well.
M. writes that he is “persuaded … that conjectures and emendations often do not aim at correcting a really corrupt text but simply are introduced to improve it — or at least, this is the result.” Even when a self-evident “correction” of the text is not achieved, conjecture can be worthwhile in helping to diagnose what ails the text and may permit another scholar some day to arrive at a fully convincing correction.1 M. himself must believe this, or he would not have printed conjectures (including his own) in the apparatus that he does not accept into his text. There would thus seem to be no difference in principle between us on the value of conjectural activity. In practice, however, a great gap opens up, a gap well illustrated by the one concrete disagreement between us in a reading that M. cites, namely 2.12.142-43; and he presents his case, characteristically, purely in terms of attestation, without reference to sense or ease of corruption.
Another topic of his riposte is the relation of his work to the edition that the late L.D. Reynolds published in 1998. I had written that I thought M.’s description of the merits of Reynolds on p.XIV as consisting in a careful register of earlier editions, diligent selection of conjectures and a clear stemma did less than full justice to his achievement. M. now claims to have done justice to Reynolds by taking his work “continuously into account.” It is true that M. very often cites Reynolds’ position in his apparatus, even when Reynolds was not the first author of the reading in question (which, of course, M. did not have to do). In this way M. attributes to Reynolds’ edition in practice a weight and authority beyond what one would have expected based on M.’s Praefatio. This fact underlines the point that the characterization of Reynolds’ edition in the Praefatio is inadequate.
There is no substantive disagreement between M. and me on the relations of the codices M, N, and V. My point was simply that in citing these three witnesses side by side M. was proceeding as if he had shown that the younger N and V were independent of M, even though he had not — another gap in the argumentation of his Praefatio (which, however, he fills in his reply).
I welcome M.’s willingness now to argue explicitly for his different view of the tradition from Reynolds’, in particular his placement of A and BE in the same branch. I am not sure, however, that the passages M. adduces are conjunctive errors, i.e., in this case, have the power to place A and BE together in one branch and the rest of the tradition in the other. First, the archetype itself contained errors, as M. himself has shown (p.VII). Some of the errors claimed by M. as conjunctive for ABE are, in fact, errors of the archetype (which one would expect to be shared by the leading representatives of the the two branches, as A and the exemplar of BE are in Reynolds’ stemma). Thus at 2.6.52 ABE present quis quasi, R has quis qua sit, PLY quasi quis, when the true reading was quis quaeso (Goerenz); similarly, 3.52.583 primorie ABEN, primore YMV etc. (Heine’s primo is accepted) or 4.6.64 de spinas ABE, de spinis cett. (M. prints the conjecture non spinas). This may also be the explanation for 3.11.115, where word-division alone separates ABE from the true reading found in MNV, perhaps a happy conjecture by the scribe of their exemplar. In some other cases A and the exemplar of BE could have fallen independently into the same error: cf. the haplographic ut for aut after qua at 2.53.625 or the omission of e after esse at 2.106.1213. The same may also be true at 1.66.720, where the tradition is essentially divided between confirmatur and confirmetur: the preceding quibus may have tempted inattentive scribes to expect an indirect question with independent corruption to subjunctive in both A and the exemplar of BE. Finally, the other passages M. cites, 1.7.76 and 2.88.1021, do not even involve shared errors of A and BE, and in the former passage A’s reading, apart from word-division, is the true reading. On p.X of his Praefatio Reynolds gives the reasons for his alternative classification; as I said in my review, one would have welcomed a discussion of these by M.
M.’s final observation is puzzling: would he really now throw consistency to the winds and read optumis at 2.23.298 and optimus etc. everywhere else in the text? Surely orthographical variants should be relegated to an appendix (as is done in some recent Teubner editions).
1. See further P. Maas, Textual Criticism, tr. B. Flower (Oxford, 1958) 53-54.