1. PH is Sextus Empiricus, Outlines of Pyrrhonism; M is Sextus, adversus Mathematicos.

2. I shall use ‘Sceptic’ and ‘Pyrrhonist’ interchangeably; I have nothing to say about the Academic Sceptics.

3. Modern sceptics customarily reject knowledge and they may allow themselves a full measure of belief. Ancient Sceptics reject belief: they also, of course, reject knowledge, but that is only a trivial consequence of their rejection of belief.

4. E.g. PH 1. 34 ; D.L. 9.81 ; Timon, apud Aristocles, apud Eusebius, P.E. 14.18.3 .

5. E.g. PH 1.59 , 1.78 ; 1. 89 .

6. The point needs stressing: unless it is firmly grasped we cannot begin to understand the Pyrrhonist's bizarre attitude to his own arguments (PH 3.280–1).

7. Hence the Stoics may consistently indulge in selective : Cicero, Ac. 2.29.94., PH 2.253;, cf. D.L. 3.52 (on Plato).

8. See esp. Burnyeat, Myles, ‘Can the Sceptic Live his Scepticism?’, in Doubt and dogmatism, edd. Schofield, M., Burnyeat, M. F., Barnes, J. (1980)Google Scholar, and Frede, Michael, ‘Des Skeptikers Meinungen’, Neue Hefte fur Philosophie 15/16 (1979) 102–29Google Scholar (Cf. Burnyeat, M. F., ‘Idealism and Greek Philosophy: what Descartes saw and Berkeley missed’, Philosophical Review 91 (1982) 3–40, esp. pp. 23–32CrossRefGoogle Scholar.) My paper is indebted on every page to the work of those two scholars and friends.

9. See esp. Brochard, Victor, Les sceptiques grecs (1923 ²)Google Scholar.

10. Galen, diff. puls. 7.711K; praenot. 14.628K.

11. Sextus' extant writings were probably composed in the order: PH - M 7–9 - M1–6 (see esp. Janáček, K., ‘Die Hauptschrift des Sextus Empiricus als Torso Erhalten?’, Philologus 107 (1963) 271–7CrossRefGoogle Scholar). Janaček's various philological studies have shown in detail how Sextus' style altered in the course of his career I think it plausible to suppose developments in his thought too - but the topic awaits detailed investigation.

12. There may, of course, be no determinate answer to that question either - PH may, in the end, turn out to offer no coherent view on the extent of See further below, pp. 15–18.

13. see the passages cited above, n. 10.

14. Myles Burnyeat has called this the country gentleman's Scepticism, in honour of Montaigne. (I take this from an unpublished paper on ‘The Sceptic in his Place and Time’, which he has kindly allowed me to read.) Burnyeat suggests that urbane Pyrrhonists ‘insulate’ their philosophy from the rest of their life, and that only a rustic treats his Scepticism as a philosophy to live by. But ‘insulation’ may be taken in either of two ways. (a) Some modern Sceptics claim that their doubts are ‘philosophical’ doubts, not ordinary doubts ‘Philosophical’ doubt is allegedly compatible with ordinary belief: a man may believe, with everyone else, that roses are red and violets blue - and at the same time he may doubt, philosophically, that violets are blue and roses red. A Scepticism which limits itself to philosophical doubt ‘insulates’ itself from real life, inasmuch as a Sceptic may share in all the beliefs - and hence join in the normal activities - of his fellow men. The distinction between philosophical doubt and ordinary doubt is scarcely to be found in ancient Scepticism. (But some scholars find it at M 11.165: and Michael Frede has in effect suggested that it underlies the theorising of the Methodical School of medicine: see his ‘The Method of the so-called Methodical School of Medicine’, in Science and speculation. edd. Barnes, J., Brunschwig, J., Burnyeat, M. F., Schofield, M. (1982).Google Scholar) The ‘insulation’ which ‘philosophical’ doubt introduces was no part of normal Pyrrhonism. (b) The urbane Pyrrhonist directs his to philosophico-scientific matters; although he never doubts and believes the same things, his doubts are still, in a sense, ‘insulated’ from ordinary life - for they touch only on the concerns of professionals. But that is not to say that his doubts have no practical manifestations. For, first, in some cases at least he may well part company with ordinary beliefs and practices (see below, p. 12). And secondly, his professional doubts may have a profound effect on his professional practices. One ancient example may illustrate that point. The Empirical doctors were urbane Sceptics, and their Scepticism had a notable effect upon their approach to medicine - it affected their research, their classification of diseases, their diagnoses and prognoses, their therapy. See, most strikingly, the remarks on anatomy and vivisection at Celsus, , prooem. 40–3 (with 23–4)Google Scholar.

15. Timon, frag. 74 Diels = D.L. 9.105.

16. D.L. 9.104; cf. M 7.197–9; Galen, simp. med. 11 380K.

17. Two troublesome side-issues should be mentioned (a) Very many sentences in the text of PH appear to commit the Pyrrhonist to beliefs of various sorts: Sextus says that men's eyes are differently structured from those of cats (PH 1.47), and his account of the Ten Tropes is largely composed of such observations, he says that Plato was not a Sceptic (PH 1.222), and his writings are full of such doxographical remarks Surely all that indicates a mass of ordinary beliefs? (So already the ancient critics of Pyrrhonism see Aristocles, apud Eusebius, PE 14.18.11) It does not, and the passages will bear no weight sometimes we should plainly understand a (cf. PH 1.80, 85) - Sextus is not speaking in propria persona; sometimes an must be read ‘catachrestically’ as (cf. PH 1.135, 195, 202; cf. M 11.18–19) - Sextus is not saying how things are, sometimes, no doubt, we should simply suppose an understandable carelessness on Sextus' part If PH is urbane, then (some of) those passages may be taken to express Pyrrhonian beliefs, but the passages cannot be adduced as evidence for urbanity (b) Sextus is a Pyrrhonist attempting to describe Pyrrhonism the attempt, as Sextus is acutely aware, is always close to incoherence - how can someone who purports to have no philosophical beliefs describe his own philosophical position? I am not here concerned with that problem, or with Sextus' efforts to surmount it For it is a problem independent of the dispute between rustic and urbane interpreters (it arises for the urbane no less than for the rustic) My question is this how should we, who are probably not Pyrrhonists, describe the philosophy which Sextus advocates in PH?

18. Burnyeat, , ‘Can the Sceptic’, 43–6Google Scholar, is convincing on this point.

19. See further Barnes, J., ‘Aristotle's Methods of Ethics’, Revue Internationale de Philosophie 133/4 (1980) 490–511Google Scholar, at 491 n. 1.

20. The canonical form of the Pyrrhonist's sentences is ‘x appears F to me now’ (see e.g. PH 1.196 ; cf. e.g. 1.4, 193, 197) Sextus says little about what appears to others or to us at other times, but I assume that the conclusions of the Ten Tropes, at least, are implicitly limited to what appears to me now.

21. Burnyeat, , ‘Can the Sceptic’, 25–6Google Scholar, takes a different line He argues, in effect, that sentences were not regarded by the Pyrrhonists - or, in general, by the Greeks - as being true (or false), for truth was, for them, a matter of correspondence with external reality, and sentences say nothing about external reality Now since belief is tied to truth (believing something is believing it true), sentences do not express beliefs at all I am not happy with that argument, but I have no room to examine it here The argument I produce in the text may be regarded either as an alternative or as a complement to Burnyeat's.

22. At M 1.269, 272, is merely a synonym for Note that is the technical term in Christian writings for ‘confess’ (e.g. Tertullian, paen. 9.2, and see Lampe, Patristic Lexicon, s.v.) also has the sense of ‘confess’ (e.g. Bion F 30 Kindestrand = Plutarch, superst. 168D, Ptolemy, tetrab. 154); but I have found no occurrences of the word in a Pyrrhonian context.

23. PH 1.187 , 197, 201 .

24. See PH 1.4, 15, 197, 200, 203. (At M 1.255, 258, means no more than ) I have not found any clear parallels to this usage outside Sextus But there is something close in Plotinus, who frequently uses for the ‘reports’ made by, or on the testimony of, the senses (e.g. Enn. 4.4.18.35, 19.6, 23.28, 5.4.24) - i.e. for reports of .

25. See PH 1.188–91: some Sceptics construed as a question, Sextus himself says that ‘although the phrase has the form of an assertion or denial, we do not use it in this way; rather, we employ it , either in lieu of a question or instead of saying “I do not know whether …’” (1.191).

26. For details and discussion see e.g. Hacker, P. M. S., Insight and illusion (1972) ch. 9Google Scholar.

27. Do they also by-pass truth? There is no need to suppose so, pace Wittgenstein When I say ‘It hurts’ it may be true that it hurts, even if I am not stating that it hurts (If I say ‘Suppose it's raining in London’ it may be true that it's raining in London, though I am not stating that it's raining in London) A Pyrrhonist who is committed to avowals does not require a metaphysically loaded concept of truth (see above, n. 21).

28. (a) Why does he limit his verbal repertoire to sentences? Instead of uttering ‘x is F’ to make a statement he utters ‘x appears F’ to make an avowal - why not retain ‘x is F’ but use it to make an avowal? Not everything can be avowed an avowal is an expression of your , and sentences of the form ‘x appears F’ were taken by the Pyrrhonists as canonical formulae for expressing (b) Does the Pyrrhonist hold that ‘x appears F’ is always used to make avowals? He need not: he uses it to make avowals, but he need not claim that other men do or must use it so, nor that the formula characteristically functions in ordinary speech as an expression of .

29. The Cyrenaics held that (e.g. PH 1.215; M 7.191; Anon. in Tht. 65.30) Like the Pyrrhonist of PH, they assent only to sentences of the form ‘x appears F’ (for their curious neologisms - - are merely verbal variants on ), unlike the Pyrrhonist, they apparently used such sentences to make statements and express beliefs (Hence, incidentally, their notion of truth was not the one mentioned in n. 21 above.) - Galen says of certain people influenced by the Pyrrhonians that (diff. puls. 7.711K.) Galen does not mean that rustics do not assent to sentences: he means that they do not use such sentences to make assertions or to express knowledge of their own .

30. These remarks are an elucidation, not a defence, of Sextus Sextus means the Pyrrhonist's utterances to be construed as avowals, and that shows that, in his view, the Pyrrhonist is not thereby committed to belief, i.e. it shows that the PH Pyrrhonist is rustic so far as his sentences go In order to defend Sextus' account from a philosophical point of view, we should require a decent analysis of avowing One element in that analysis would presumably be the claim that the Pyrrhonist's utterances are produced as a direct and natural response to external stimuli - just as a child's cry is a direct and natural response to the stimulus of pain.

31. This is vague - intentionally and harmlessly so For a more rigorous definition see below, n. 86.

32. The survey is impressionistic: I have not conned every occurrence of and its cognates in Greek. In addition to the authors mentioned in the text, I have consulted concordances or indexes to all the major prose-writers from 400 BC to AD 250: the general conclusions I reach in this section would doubtless be refined by further study but I hardly think that they would be overthrown.

33. For verbal nouns in -μα see Buck, C. D. and Petersen, W., A reverse index of Greek nouns and adjectives (1944) 221Google Scholar they suggest that the -μα termination was an intellectual's favourite. See also Pollux, onom. 6.180.

34. See D. L. 3.51 [ =Suda, s.v. ] [i.e. a man's are either the things he believes or his believings].

35. E.g. Lysias, 6.43 (399 B.C.); Andocides, 4.6 (c. 395); Xenophon Anab. 3.3.5 (c. 375); IG II ²96Google Scholar (375/4), 103 (369/8) 123 (357/6).

36. E.g. Laws 644D3, 797C9, 926D2; Rep. 403A2, 506B9, 538C6; cf. Minos 314BE; Def. 415B8, 11, C2.

37. See, e.g. Mauersberger's Lexicon to Polybius or Rengstorf's concordance to Josephus; cf. = to decree (e.g. Josephus, Ant. 14.249; LXX, 2 Mac. 108.15.36).

38. At Heraclitus, B 50 DK ( = 26 M, from Hippolytus (?), ref. haer. 9.9.1), the MSS read editors generally accept Bernays's (see Marcovich, M.Heraclitus (1967) 113Google Scholar), but has recently been defended by Holwerda, D., Sprünge in die Tiefe Heraklits (1979) 9–10Google Scholar.

39. Cf. Rep. 506B8, , picking up B6, occurs some 30 times in the Platonic corpus usually in political contexts (see Brandwood's concordance).

40. LSJ s.v. offer ‘notion’ for at Tht. 158D3, and the Supplement s.v. discovers a new sense for the word, viz. ‘thought, intention’, for which Tim. 90B and Laws 854B are cited. But at Tht. 158D and Tim. 90B the word is used in the same way as in Tht. 157C; and at Laws 854B the is a decree or resolution.

41. The distinction between sense and colour (Färbung) is due to Frege: see Dummett, M., Frege - Philosophy of language (1973) 83–9Google Scholar.

42. E.g. Laws 791D5, 798E2, 900B4; Phlb. 41B5; Tim. 48D6, 55D1; Soph. 265C5.

43. See Phys. 209b15 (Plato's ), Met. 992a21, 1076a14 (But at Top. 101a31–2 the word appears to have a broader denotation) See also Met. 1062b25; M.X.G. 974b12; Rhet. ad Alex. 1430b1, 1443a25 (and Bonitz's Index).

44. For see frag. 562 Us =D.L. 10.121 (cf. Burnyeat, , ‘Can the Sceptic’, 48 n. 50Google Scholar) For see esp. frag. 29 Arr., at 28.5, 6, 10, 12 (with Arrighetti's note, 602–3); cf. frags 30 (31.1), 31 (2, 4, 6), 36 (10.3), and Arrighetti's index Note also the title of Colotes' pamphlet (Plutarch, adv. Col. 1107E) For in later Epicurean texts see the index to Philodemus by Vooys; and cf. Diogenes of Oenoanda, frag. 27 Ch, 1.8.

45. See Leisegang's index (vol. VII of the Cohn-Wendland edition of Philo). -I say ‘almost invariably’ only because Philo occasionally uses of decrees.

46. Leg alleg. 2.25.100; migr. Abr. 21.119.

47. See Wyttenbach's index.

48. See the indexes to the relevant volumes in CIAG; e.g. de fato 164.16; 165.1; 177.6, 187.9; 12, 27, 188.17, 22; 190.6, 12; 192.21; 199.23; 205.23; 212.2; in Met. 40.31; 78.2; 24; 197.1, 8; 652.33. See also, e.g. Atticus, frags 2 (83, 113, 149), 4 (33, 60), 7 (10, 12, 35) des Places; Lucian, vit. auct. 17, bis acc. 21.

49. The way was prepared by the LXX (e.g. 3 Macc. 1.3; 4 Macc. 10.2) and the NT (e.g. Col. 2.14, 20). See further Kittel, G. (ed.), Theologisches Wörterbuch zum neuen Testament II (1933–1935) 233–5Google Scholar.

50. See e.g. Stählin's index to Clement, Wendland's to Hippolytus, ref. haer., Koetschau's to Origen, c. Cels.

51. For in this metaphorical sense see Lampe's Lexicon s.v. sense A.

52. See, e.g. Galen, in Hipp. vict. acut. 15 728K (those who construe Hippocrates as a think he is referring to and , those who make him an hold that he is talking about , etc.), in Hipp. art. 18A. 735K (Heraclides advances his views ) cf. opt. sect. 1.146K (the say that when ) - Note that Galen may supply a new term from the family, viz. or (see subfig. emp. 65.15: the Latin has in dogmatibus, emended by Schöne to indogmaticus).

53. Compare also the use of in the stock definition of a PH 1.16; D.L. 1.20; Clement, strom. 8.5.16.2 (p. 89.24 St); [Galen], hist. phil. 7; def. med. 13, 19.352K; Suda, s.v. .

54. See D.L. 7.199 (a title of a work on ethics by Chrysippus: ); Origen, c. Cels. 8.51 (from Chrysippus' ); Stobaeus, ecl. 2.62, 112; Philo, om. prob. lib. 97 (6.28.5–9).

55. Compare also Seneca's frequent use of decretum (see below, n. 58).

56. e.g. diss. 2.22.37; 3.7.20–29, 16.7.

57. The same is true for Marcus - see Dalfen's index For Epictetus see the index to Schenkl's edition.

58. See Acad. 2.9.27… de suis decretis, quae philosophi vocant (cf. 29; 34.109; Tusc. 2.11; fin. 2.28, 99). Seneca uses decretum frequently in this sense (see the Concordance of Busa and Zampolli). See esp Ep. 95.12 decreta sunt quae muniant, quae securitatem nostram tranquillitatemque tueantur, quae totam vitam totamque rerum simul contineant; cf. ib 45 persuasio ad totam pertinens vitam - hoc est quod decretum voco. See further TLL s.v. - the word dogma was itself used by Cicero (it had already been Latinised by the poet Laberius), and it is common in later authors, always with reference to principles or tenets: see TLL s.v.

59. See, e.g., his worries over the translation of , where he is explicitly concerned to get the colour right ad Att. 13.21.3.

60. That conclusion may seem pretty unexciting But it is not uncontroversial. Burnyeat, , ‘Can the Sceptic’, 48 n. 50Google Scholar, concludes that in Hellenistic usage ‘is a broader and more nearly neutral term than , not a term for a more stringently defined type of belief’; it means ‘“belief” or “judgement” in the broad sense in which it is a component of knowledge’.

61. See Janáček's index.

62. Note that over half (c. 140) of those occurrences are in PH, though M is three times the length of PH. I detect no difference in Sextus' use of between PH and M.

63. Cf. D.L. 9.102–4 (see below, n. 70).

64. With what follows compare Frede, , ‘Des Skeptikers’, 120–6Google Scholar.

65. See Lampe's Lexicon s.v.; Kittel, 's Theol. Wört, II 736–48Google Scholar; Mauersberger's Lexicon to Polybius. Typical texts: Polybius, 2.38.7; 3.8.7; 4.22.7; 8.14.8; cf. Suda, s.vv. , etc..

66. So at NT, Mark 1.11 (‘Thou art my only begotten son: in thee I am well pleased’), the Greek is .

67. See Bekker, , Anec. Gr. II 260Google Scholar, where I take καί to be epexegetic. Note that , outside its Stoic use to mean ‘assent’, regularly means ‘accept’, ‘acquiesce in’; see e.g. Polybius, 21.30.8, where and appear in the same sentence as synonyms.

68. Why does Sextus think that gives a sense of I have found no texts outside PH 1.13 where or its cognates are used in that weak way. I can only suppose that the ‘broad’ sense of is a dialectical concession by the Pyrrhonists (who do not indulge in : PH 1.195, 297). An opponent urges: ‘Of course you Pyrrhonists dogmatise - after all, you avow your ’. The Pyrrhonist retorts: ‘If you like to use “dogmatise” in that sense, we do indeed dogmatise - but that does not imply that we also dogmatise in the normal, narrow sense’.

69. [Galen], def.med. 14, 19.352–3K, should be quoted: . The text is hardly sound. for ἐνεργείᾳ is easy enough; but I suspect the corruption is more extensive. E.g. . If something like that is right, then [Galen] may be recognizing ‘belief’ as the general sense of (i.e. he may be allowing that, in one sense, any belief may be called a ). Then [Galen] is close to D.L. 9.102–4 (see below, n. 70) and his distinction of senses is not the same as the one in PH 1.13.

70. That urbane suggestion may appear appropriate to D.L. 9.102–4. Replying to the charge that they dogmatise, the Pyrrhonists there are made to concede that . In other words, they allow that, if may cover ordinary beliefs, then they do dogmatise. Of course, if that is the meaning of D.L.'s Pyrrhonists, it does not follow that the same is true of PH. And in any event, the meaning of D.L.'s Pyrrhonists is by no means clear-cut. For the sentence I have just quoted is introduced by the remark that , and followed by the assertion that . Thus and the like are apparently to be constructed as expressions of . D.L.'s Pyrrhonists accept ordinary beliefs - but only because they reconstrue them as beliefs about their own . Hence they are not exactly urbane (though they are not rustic either, if we insist on the claim that they know - - their ). It must be said, however, that the text of this passage in D.L. is very confused, and it would be unwise to rely upon it for the interpretation of any piece of Pyrrhonism.

71. E.g. PH 1.16, 193, 198, 200, 202, 208, 210, 219, 223; 2.9; cf. 1.18, 201.

72. See Janáček's index, s.v. ; cf. [Galen], opt. sect. 1.175–6 K.

73. See esp. PH 2.95; M 7.25; 8.140–1.

74. There is another connected inconsistency in the same stretch of argument. Sextus plainly states that the Pyrrhonist attack on undermines belief in (PH 2.95; M 7.25); he also expressly defines a as (PH 2.15; M 7.33). I see no escape from that inconsistency - except the appeal to a systematic and unexpressed ambiguity in such terms as , .

75. Myles Burnyeat has suggested to me that anything which depends on a must itself be a . Hence ordinary beliefs are in the Pyrrhonists' eyes.

76. Again (see above, n. 30), I am concerned to explain Sextus, not to defend him Against the argument advanced m the text it might be objected that, although in order to judge that p I must possess a criterion, it is not true that in order to judge that p I must believe that I possess a criterion. Thus the Pyrrhonian may possess a criterion even if he himself does not believe that he does; and in that case he is in a position to judge that p. That is perhaps true; but could a Pyrrhonist judge that p after reflecting on the existence of a criterion and reaching ἐποχή on the matter? Sextus might plausibly argue that, having reached ἐποχή on the δόγμα of the criterion, a Pyrrhonist will naturally find himself in a state of ἐποχή vis-à-vis ordinary judgements.

77. Something must be said about the word ἀδόξαсτοс, which occurs 16 times in PH, all but once in its adverbial form The word is rare outside PH (it does not appear in M) It is found in a fragment of Sophocles (fr. 223, where it means ‘unexpected’), at Phaedo 84A (where το ἀδόξαсτοс is joined with τὸ ἀληθέс and τὸ θεῖον to characterize the objects of the soul's proper study), at D.L. 7.162 (Ariston [Scaliger: δοξαсτον codd.] εἶναι), at Aristocles, apud Eusebius, P.E. 14.18.3 (according to Timon we should be cf. ib. 16 - ,) In PH the adverb ἀπαγγέλλει usually qualifies either a verb describing the Pyrrhonist's way of life (βιοῦν: 1.23, 231, 2.246, 258; 1.226, 3.235; etc.) or a verb describing the Pyrrhonist's utterances (ἀπαγγέλλει 1.15; φαμέν 1.24, 3.151; сυγκατατιθέμενοι: 2.102). The word may be part of the Pyrrhonist vocabulary adopted by Sextus, but it is not clear to what extent Aristocles is citing Timon's own words, and could well be his own gloss on (which are presumably genuine Timon) What does mean in PH? Plainly, it means ‘having no ’; but that is capable of three importantly different glosses, according to the colour we see in here. [α] ‘Having no mere opinions’: that is the word's meaning in D.L. 7.162 (and in the Phaedo - ‘not an object of mere opinion’) If the word was used by Timon, then it might well bear that meaning in his sentence: ‘having no mere opinions’, i.e. ‘fixed’, ‘firm’ (cf. ). In many - but not all - the passages in PH a sense like ‘fixedly’, ‘unwaveringly’, fits perfectly well [β] ‘Having no ’: that meaning is hardly suggested by the word's etymology or by its history; but is frequently contrasted with vel sim, and such a contrast could well give the word that particular colouring. (And some might see a neat polemical point: the Stoic Sage lives , with but without , and so in tranquillity; the Pyrrhonian lives , without , and so in tranquillity ) All the PH passages will readily accept that meaning [γ] ‘Having no belief of any sort’ that is surely how Aristocles intends the word at 14.18.16 - and therefore how he intends us to understand it in Timon That sense is, I think, compatible with most of the occurrences in PH if not with all. (The coupling at 2.246 does not sit easily with [γ] inasmuch as normally is supposed to involve beliefs, and [γ] does not have any obvious intelligibility at 1.239 and 240, where Sextus talks of using technical terms .)

If sense [γ] is correct for PH then there are two corollones of immediate relevance to my theme. First, we have Sextus explicitly stating that the Pyrrhonist's avowals do not involve him in any beliefs: (1.15) Secondly, we have Sextus explicitly claiming that the (below, pp. 13–18) does not require belief in the Pyrrhonist who follows it: (3.235) (See further, below nn. 96, 98). Indeed, if [γ] is right, then that alone virtually makes PH rustic. Unfortunately, I can see no way of determining the sense of without presupposing the rustic/urbane dispute solved; hence I have relegated to a footnote and shall not rest any argument upon its interpretation.

78. E.g. PH 1.165; M 7.322, 8.355, 9.138, 1.232.

79. E.g. PH 2.105, 258, 3.235; M 9.50.

80. Cf. at NT, 1 Cor. 6.3–4.

81. E.g. Apollonius Dyscolus, adv. 130.6; conj. 245.21, 246.10; synt. 40.1, Galen, meth. med. 10.269 K. For the various locutions for ‘ordinary usage’ see Schneider's, note in Grammatici Graeci II i 2, 45Google Scholar.

82. E.g. Plutarch, mor. 25C, 1033A, 1116C, Epictetus, diss. 1.15.2, 26.1, 3, 7, 17, 2.3.3, 5; frags 1, 2; Galen, subf. emp. 68.7; diag. puts. 8.78K; Soranus, gyn. 1.4.1; 3.3.1 See Epictetus, frag 16 .

83. See Lampe's Lexicon s.vv βιοс (6), βιωτικόс (cf. e.g. NT, Luke 21.34; 2 Tim. 2.4) In Christian writers βιοс is often contrasted with (e.g. Eusebius, P.E. 7.8.41); but that is only verbally comparable to what we find in Sextus: the Christian contrast is between deeds and words, between works and doctrine.

84. See M 9.50, 138.

85. Cf. D.L. 9.88 , Here as often, s synonymous with βιοс.

86. The contrast between βιοс and , like the term itself, is vague. I do not think the vagueness is harmful (see above, n. 31), but a little precision can readily be supplied For Sextus' remarks enable us to define as follows: A sentence expresses a iff(i) it expresses a proposition and (ii) it contains at least one term which denotes something ἄδηλον. Most ordinary beliefs will not bed , most philosophico-scientific tenets will be . But βιοс will include some , notably (a) involving reference to the Gods, and (b) those involving moral concepts (for, in the Pyrrhonist's eyes, terms like ἀγαθόν and κακόν denote ἄδηλα), If an urbane Pyrrhonist defends the beliefs of βιοс, he does so only for the most part.

87. See also M 9.165 A similar respect for βιοс was ascribed to Pyrrho himself by Galen (subf. emp. 62.20), by Aenesidemus (D.L. 9.62), and perhaps by Timon (frag 81 Diels =D.L. 9.105 - but see Caizzi, Fernanda Decleva, Pirrone - Testimonianze (1981) 236–41)Google Scholar. It was a commonplace among the Empirical doctors: e.g. Galen, diff. puls. 8.783K; Med. Exp. 18.5 Walzer.

88. The argument had a long history and went through different forms; see e.g. Burnyeat, , ‘Can the Sceptic’, 22Google Scholar n, 4; Striker, Gisela, ‘Sceptical Strategies’, in Doubt and dogmatismGoogle Scholar.

89. The word τήρηсιс has the same ambiguity as the English ‘observation’ - observation of rules etc. (i.e. obedience), or observation of objects and events (i.e. perception etc.) Sextus generally uses the word in the latter sense (see Janàček's index), but the former is more appropriate at PH 1.13.

90. ‘But surely “Because he was hungry” will not by itself explain why men eat? We need, in addition, some reference to beliefs “Why did he eat that tough steak?” - “Because he was hungry, and thought that the steak was the only food available” The πάθη by themselves are not sufficient to explain even our simplest actions’ But that objection misses the point: Sextus is not implying that ‘Because he is hungry’ explains, in general, why a man eats; he may properly allow that in all normal cases an explanation will invoke beliefs as well as πάθη. His point rather is that such actions can be explained even if the agent has no beliefs strike a man on the knee and his foot will kick, by a sort of natural necessity; similarly, if a Pyrrhonian is thirsty he will drink, by a sort of natural necessity. Non-Pyrrhonian drinking is no doubt only explicable via beliefs but, according to Sextus, drinking can be explained even in the absence of belief - and that is all a Pyrrhonian requires.

91. See the list of Pyrrhonists at D.L. 9.115–6 (Menodotus, Sextus, Saturninus); add, e.g., Cassius (Galen, subf. emp. 40.15), Dionysius of Aegae (Photius, bibl. codd. 185=codd. 211).

92. M 5.1–2 accepts farming, seamanship and astronomy as legitimate professions.

93. There is in any case a tension within PH, for Sextus argues at PH 3.252–73 (cf. M 11.216–56; 1.9–18) that is impossible, and his argument does not appear to make any exceptions for the which PH 1.23 accepts (Nor will the distinction between transmitting beliefs and inculcating skills help: many of the arguments against are equally applicable to each sort of teaching).

94. PH 1.237 () refers back explicitly to 1.23–4.

95. Again (see above, n. 90), Sextus does not imply that other men's conventional actions are explicable without invoking beliefs: his point is simply that a Pyrrhonian may act conventionally, ‘because it's the custom’, without subscribing to any beliefs.

96. [cf. 1.24] . I incline to construe with rather than with the three finite verbs. See below n. 98.

97. may certainly indicate adoption without any implication of belief (see e.g. PH 1.191, 195, 240) But is more difficult to construe in a belief-neutral way (See e.g. M 1.201, where means ‘to accept ordinary usage as reliable’, i.e. to believe that it is reliable).

98. If is construed with (see n. 96) and if the adverb means ‘without belief” (see n. 77), then PH 3.2 actually asserts this; for Sextus then expressly argues that the Pyrrhonist will say ‘There are gods’ but will not believe that there are gods.

99. Again, Wittgenstein might be invoked: see, e.g. his Lectures and Conversations on … Religious Belief esp. 53–9. (But according to Wittgenstein, all churchgoers are playing the language game which in the text I prescribe for the Pyrrhoman).

100. For with the sense ‘instruction” (not ‘guidance’) see PH 1.6, 2.120, M 7.22, 8.300, 11.47; 1.35, 172, 258, 3.18, 5.3.

101. This is a rough characterisation; for a detailed and subtle treatment see now Burnyeat, M. F., ‘The origins of non-deductive Inference’, in Science and speculationGoogle Scholar.

102. Cf. M 8.156–8, which makes the same point in similarly forthright terms.

103. ‘But it does not look like fire over there The whole point of the sign is that it allows us to grasp that there is fire there even when we cannot see or otherwise perceive the fire: the fire is , not and if it were not, we should have no need of a sign’. The rustic may say, in reply, that when he experiences the normally reported by ‘It looks like smoke’, he also experiences the normally reported by ‘It looks like fire’ - i.e. he experiences the which he normally experiences when (as a non-Pyrrhonist would put it) he is actually looking at the fire itself. Naturally, he reports the second in the standard way, saying ‘It looks like fire’ - there is no reason why all cases of its looking like fire should be qualitatively indistinguishable (But is that really coherent? Suppose a rustic looks at an oar in water; why shouldn't he say ‘It looks straight’? For there is no reason why all cases of oars looking straight should be qualitatively indistinguishable. Maybe a Pyrrhonist will sometimes say ‘It looks straight’: he is, after all, simply reporting his , and there is nothing in Pyrrhonism which demands that the caused (as a non-Pyrrhonist would put it) by a submerged oar should always be the normally reported by ‘It looks bent’) - The second example of a recollective sign at PH 2.102 introduces an important point which Sextus nowhere develops For the ‘conclusion’ of the second sign is ‘He looks as though he has been wounded’, and that contains a reference to the past The Pyrrhonist's sentences are always present-tensed: he reports his present (see above, n. 20) But the contents of those may themselves advert to past - or to future - times A Pyrrhonist may say - to make the point fully explicit - “The man now appears to me as having been wounded’, ‘The clouds now appear to me as being about to produce rain’. In that way a rustic Pyrrhonist may have some purchase on the past and the future, and plainly some purchase on the future; is necessary if his actions are to be given any adequate explanations.

104. My standard of coherence is pretty low: I mean only that this account of recollective signs is at least as plausible as, say, Sextus” account of Methodical medicine.

105. A certain amount of is inseparable from the human condition there the best the doctor can do is produce (PH 1.25; 3.235–6).