¹ See Reynolds, L. D. (ed.), Texts and Transmission (Oxford, 1983), 125Google Scholar.

² Ibid., p. 130 n. 31.

³ Marci Tulli Ciceronis Topica edidit, commentario critico instruxit Georgius Di Maria (Palermo, 1994). I use his manuscript sigla (not that they are very convenient), and cite the Topica by section and Di Maria's line, adding in parentheses a line-reference to A. S. Wilkins's shoddy but more easily accessible Oxford Classical Text (1903).

⁴ I cite the commentary from M.Tullii Ciceronis Scholiastae, ediderunt Io. Casp. Orellius et Io. Georgius Baiterus (Zürich, 1833) (= Orelli-Baiter V.i), 270–388, citing by page and line. The commentary as we have it extends only to Top. 76.

^4a The supplementing folios B _A cannot, if only on grounds of date, have been copied from a, though they could have been copied from a's exemplar, a is not a copy of B as completed by B _A (see Di Maria's list of errors unique to AB on p. xvii).

⁵ A parallel list of errors common to B _A (in the absence of β) and the other integri would not be instructive, because such errors might go back to the archetype that lies behind α and β.

⁶ They do not seem to come from Top. 26, Di Maria's alternative explanation

⁷ B ^sl is said to have elia, but this will not have been in the exemplar. That e has aelia sancia before rather than after lex could be a sign of the contamination. a ^mg's giving the Boethian reading is not untypical of the behaviour of a ^mg's corrector; a's original text is, as I shall show, consistently sincere.

⁸ The reading is perfectly sensible after ab effectis rebus at the start of the section and similar phrases earlier (it could even be right); and this, like some of the other readings given, could be a correction made independently of Boethius.

⁹ And F has fortuitarum (rather than the neuter), without rerum. O's phrase may be right: cf.rerum fortuitarum at 73 (the right order, as in Boethius?).

¹⁰ Perhaps a coincidence.

¹¹ But the constructions differ.

¹² Rightly?

¹³ eγCFO 38, 25–1(9) coniugatione (coniunctione B _AaB _lb^ac), cf. 337, 26–28 (and Top. 12 above); 44, 15(19) venisset (venissent BAaac), cf. 341, 19 (as well as the context); 54, 5(7) igitur (autem B _Aa), cf. 361, 12; 54, 12(14) ut (aut B _A^ac [?]a), cf. 364, 18; 64, 26(13) magis quam (magisque B _A^ac and perhaps C) = 375, 4; 66, 17 (30) quid (before socium) (quis B _A^ac [?]s), cf. 378, 43.

eFO 43, 9(13) <ex eodem similitudinis loco= sic B _AaγC: the corrector saw that this comes from 44 below, a good perception eased by the Boethian discussion at 341, 1 seq.

C 43, 9(12) adigere (adicere B _AaeγF, addicere O) = 341, 10.

CO 23, 24(22) huius (eius )(B _AaeγF), cf. 307, 37 ‘hoc modo’; 39, 10(18) propiore (propriore B _AaV ₂ B ₁₆^ac F: e has a further error), cf. 338, 42 ‘genus proximum’.

At 75,4(26) for Staieno β has statiaeno. Only eC have the right reading, but a has staleno, and this (or staieno) will probably have been the reading of α. But the correction could have been made from 388, 7.

¹⁴ It is a striking testimony to the sincerity of B _Athat before correction it had the accusative.

¹⁵ That is, passages where Boethius does not hand a corrector the reading on a plate, or where he does not help at all: I mention some where he might have helped somewhat. I naturally exclude passages like 86, 8(27)propositi, where the correction was only made in modern times. 16

¹⁶ As will be seen, the correction is often made in only one or two (often later) manuscripts; this perhaps makes the likelihood of conjecture (rather than contamination from some unknown source) stronger.

¹⁷ Boethius 287, 31–32 tune …quando rests on a misunderstanding of Cicero, for whom turn definitio correlates with turn partium enumeratio in 10.Google Scholar

¹⁸ The converse of my hypothesis that coincidence with Boethius weakens the authority of readings of eγCFO against β(B _A)a is that when β(B _A)a and Boethius agree against the others they should be giving us the truth.

¹⁹ But Professor Adams, J. N., whose monograph Wackernagel's law and the placement of the copula esse in classical Latin (Cambridge Philological Society Suppl. Vol. 18, 1994) inspired this remark, points out to me that his examples 153–159 (pp. 30–31) support the order putat esse. He also remarks on the likeness of 53, 19–20(28–29) ‘reperiendi argumenti locus est simplex, tractandi triplex’ to his example 320 (p. 62, from Cic.de orat. 3.67). That order (printed by Di Maria) is found only in C and one of the γ manuscripts, simplex seems to have been omitted in the transmission (om. B _Aa ^ac), but to have been added in different places in the MSS by conjecture or from Boethius 360, 21–22.Google Scholar

²⁰ Boethius in fact gives quod, apparently with the meaning ‘because’.

²¹ It is interesting that Boethius 364, 17 confirms unum, without adding aut plura.