¹ Cf. Bett, R., ‘Carneades' Distinction between Assent and Approval’, Monist 73 (1990), 3–20CrossRefGoogle Scholar; Frede, M., ‘The sceptic's two kinds of assent and the question of the possibility of knowledge’ in Rorty, R., Schneewind, J., Skinner, Q. (edd.) Philosophy in History (Cambridge, 1984), 255–78.CrossRefGoogle Scholar

² And considerations of the same kind are also behind the charge later levelled against the Pyrrhonists that, by adopting the appearance as a standard, they have deprived themselves of the ability to decide between conflicting appearances (cf. D.L. 9.107).

³ Cf. H. von Arnim, s.v. ‘Karneades’, R.E. vol. X.2 cols. 1964–85, col. 1969–70; Goedeckemeyer, A., Die Geschichte des griechischen Skeptizismus (Leipzig, 1905), pp. 62ff.Google Scholar; Striker, G., ‘Sceptical Strategies’ in Malcolm, Schofield et al. (edd.) Doubt and Dogmatism (Oxford, 1980), 54–83, p. 70Google Scholar; von Staden, H., ‘The Stoic theory of perception and its “Platonic” critics’ in Machamer, P. K. and Turnbull, R. G. (edd.) Studies in Perception (Columbus, 1978), 96–136, n. 102Google Scholar; Tarrant, H., Scepticism or Platonism (Cambridge, 1985), pp. 14ff.Google Scholar; Richard, Bett, ‘Carneades' Pithanon: a Reappraisal of its Role and Status’, OSAP 7 (1989), 59–94.Google Scholar

⁴ Cf. Mutschmann, H., ‘Die Stufen der Wahrscheinlichkeit bei Karneades’, Rheinisches Museum, 66 (1911), pp. 191ff.Google Scholar

⁵ Cf. Weische, A., s.v. ‘Karneades’, R.E. Suppl. vol. XI, cols. 853–6.Google Scholar

⁶ Antiochus is mentioned at M 7.162 and his Canonica is quoted at M 7.201–2, which suggests that Sextus is dependent on the Canonica for a fair bit of M 7's survey of epistemology, almost certainly for the discussion of the Academy. Cf. Hirzel, R., Untersuchungen zu Ciceros philosophischen Schriften, vol. III (Leipzig, 1883), 493ff.Google Scholar; H. Tarrant, op. cit., pp. 94–6. On, Sextus' doxographical sources in M 7 also Sedley, D. N., ‘Sextus Empiricus and the Atomists' Criteria of Truth’, Elenchos, 13 (1992)Google Scholar. A cautionary note has been sounded by Barnes, J., ‘Antiochus of Ascalon’ in Griffin, M. and Barnes, J. (edd.) Philosophia Togata (Oxford, 1989), pp. 51–96.Google Scholar

⁷ I am indebted to M. F. Burnyeat's unpublished but widely circulated paper ‘Carneades was no Probabilist’ for his detailed discussion of this issue and for much else besides.

⁸ Cf. Heintz, W., Studien zu Sextus Empiricus (Halle, 1932), pp. 100ff.Google Scholar

⁹ Cf. D.L. 7.78; Origen, comment, in Mattaeum = SVF 3.170.

¹⁰ Cf. H. von Arnim, op. cit., cols. 1969–70.

¹¹ Cf., e.g., Ac. pr. 2.33, 59, 99, 105, 107, 109.

¹² The example is introduced ‘and the impression which is also unimpeded is of such a kind’ ( δ κα περσπαστος ϕαντασα τοιδε ρτν), apparently referring to the example to follow. If we adopt the reading of the Latin translation (T), which omits the κα, a perfectly good point is made: because it has been carefully considered in connection with related impressions without encountering opposition, the impression that this is a coil of rope is unimpeded.

¹³ Cf. H. von Arnim, op. cit, col. 1970.

¹⁴ Talk of seeming false seem a bit odd here. What one expects is a reference whose apparent truth is somehow in tension with the truth of the impression at issue, but perhaps it is the apparent falsity of what is to be expected that is meant.

¹⁵ See Mette, H. J., ‘Weitere Akademiker heute: von Lakydes bis zu Kleitomachus’, Lustrum 27 (1985), 39–148, p. 76 ad loc.Google Scholar

¹⁶ This is an overstatement; what is sought is that they be apparently true.

¹⁷ Perhaps ‘manner’ (τρπος) would be better. This is not the only place in Sextus where the sense of a passage would apparently be improved by the substitution of τρπος for τπος (cf. M 7.424–5, 437).

¹⁸ E.g., does the impression that there is a snake in the room change from being probable to improbable? Or would it be better to require a great deal more than sameness of propositional content if impressions are to count as identical and say that the observer no longer has the same impression, his first probable impression having been displaced by another? Can one have a perceptual impression which is not at all probable? It may be significant that none of the Stoa's examples of improbable or non-probable impressions is perceptual (cf. M 7.243).

¹⁹ The latter is the preferred rendering of Long, A. A. and Sedley, D. N., The Hellenistic Philosophers, vol. 1 (Cambridge, 1987), pp. 452–60 passim.Google Scholar

²⁰ Sextus gives περισκοπω as an alternative for περιοδεω at PH 1.228.

²¹ Arthur, Kochalsky, De Sexti Empirici adversus logicos libris quaestiones criticae (Marburg, 1911), p. 62Google Scholar; cf. on M 7.437, p. 73. Kochalsky's supplement is strongly endorsed by Werner Heintz, p. 100 ad loc. and adopted by the text's latest editor, H. J. Mette, op. cit., p. 73, but not by Mutschmann, who, however, believes this is what Sextus ought to have said though he carelessly neglected to do so.

²² Galen's brief allusion to the Academic view does not require a division of the contested type, though it could be emended to do so, albeit with more effort (De placitis Hipp, et Plat. 586, 16ff., De Lacy): ‘the more recent Academics refer judgement to the impression which is not only probable, but also thoroughly examined (περιωδευµνη) and unimpeded as well.’ The result of emendation would, unfortunately, align the present passage with PH 1.227–8.

²³ I have followed Mette, op. cit., in labelling the three varieties of probable impression distinguished at M 7.184 as (a), (b) and (c). He too can find no place for (b) and proceeds straight to (c).

²⁴ For the argument in full, see Striker, G., ‘The problem of the criterion’ in Everson, S. (ed.) Epistemology (Cambridge, 1990), 143–60, 152–3 with n. 14Google Scholar; also Frede, M., ‘Stoics and Skeptics on Clear and Distinct Impressions’ in Burnyeat, M. (ed.) The Skeptical Tradition (Berkeley, 1983), 65–93, 84Google Scholar. For a dissent, see Sandbach, F. H., ‘Phantasia kataleptike’ in Long, A. A. (ed.), Problems in Stoicism (London, 1971), 9–21, 15Google Scholar. The younger Stoics' claim that the cognitive impression all but takes us by the hair drawing us to assent is not an innovation of theirs. Rather, they seem to have departed from their predecessors by qualifying this view which had earlier been held in an unqualified form. See Görler, W., ‘Ἀσθενς συγκατθεσις, zur stoischen Erkenntnistheorie’, Würzburger Jahrbucher für die Altertumswissenchaft N.F. 3 (1977), 83–92, 91 with n. 29.Google Scholar

²⁵ This is especially clear in the passage from Cicero: ‘meo autem judicio [Lucullus is speaking] ita est maxima in sensibus veritas, si et sani sunt ac valentes et omnia removentur quae obstant et impediunt. itaque et lumen mutari saepe volumus et situs earum rerum quas intuemur, et intervalla aut contrahimus aut diducimus, multaque facimus usque eo dum aspectus ipse fidem faciat sui judicii’ (Ac. pr. 2.19).

²⁶ We can see how difficult the relation between reasoning and the cognitive impression was thought to be from Cicero, who reports that some philosophers objected to giving any answer at all to the Academy's arguments because to do so would be to imply that impressions than which nothing can be more evident stand to benefit from support and justification (Ac. pr. 2.17). Nonetheless, Cicero maintains that a defence of the evident can be mounted without undermining itself, and he represents Lucullus' speech as just a defence of the evident from the sophistical captions of the Academy (Ac. pr. 2.45–6, 105).

²⁷ As Sextus sometimes fails to do (cf. M 7.424). Cf. J. Annas, ‘Stoic Epistemology’ in S. Everson, op. cit., 184–203, 201.

²⁸ Translated in accordance with Bekker's emendation, οὐκ εἶχε δ αὐτν 〈πιστν〉. Cf. W. Heintz, op. cit., 118ff.

²⁹ Those troubled by the supernatural element in the two examples might want to substitute a recognition scene from one of the more implausible but nonetheless remotely possible Shakespeare plays such as the Winter's Tale or especially Pericles. The wife of its hero has to all appearances died in childbirth in the middle of a storm at sea. She is promptly put overboard in a chest, which, however, quickly washed ashore near the home of a wise and kindly physician who is able to revive her. Many years later, when Pericles is confronted with the wife he thought he had lost, he refuses to believe his eyes until the whole story is laid out for him.

³⁰ Striker, Pace G., ‘The problem of the criterion’, (op. cit., n. 24), 152, with n. 14.Google Scholar

³¹ I take it that the Stoics' notion of weak assent includes all acts of assent by those who are not wise, even with regard to their cognitions, but the testimony on this point, and on the closely connected issue of the relation between knowledge, ignorance, opinion and cognition, is not uniformly consistent and is subject to controversy. See W. Görler, op. cit., and A. A. Long and D. N. Sedley, op. cit., 256ff.

³² See n. 28.

³³ I am grateful to N. Belnap, M. F. Burnyeat, J. M. Cooper, G. E. R. Lloyd, D. N. Sedley, R. J. Wallace and audiences at Berkeley, Columbia and Princeton for giving me the benefit of their reactions to earlier versions of this paper.