I

«Relations between the Sceptics and Peripatetics were definitely atarm’s length. Sextus Empiricus mentions Aristotle and his followersquite often and attributes an elaborate doctrine of the “criterion” ofknowledge to them, but his work shows no signs of a deep study of theirwritings and his knowledge of them seems to come from handbooks. Inmany places he writes as if their teaching hardly differed from that ofthe Stoics, and when he does acknowledge a difference in order to exploitthe contradictions between dogmatic schools for his own ends, he tendsto state their position in a terminology strongly influenced by Stoicism».

Thus H.B. Gottschalk in his magisterial account of Aristotelianphilosophy in the Roman world H. B. Gottschalk, Aristotelian philosophy in the Roman world from the timeof Cicero to the end of the second century A.D., (“Aufstieg und Niedergang derrömischen Welt” ii, 36.2), Berlin 1987, pp. 1079-174.. And Gottschalk’s opinion is but-tressed by Luciana Repici Cambiano, who points out that Sextus’ treat-ment both of the Peripatetic school as a whole and of particular Peripatet-ics is shaped by his tendency to play the different schools of dogmatistsoff against one another, and in particular to play off other schools againstthe Stoics, the school which most dominates his philosophical agendaL. Repici Cambiano, Sesto Empirico e i Peripatetici, in Lo scetticismo antico (Attidel Convegno organizzato dal Centro di Studio del Pensiero Antico del C.N.R., Roma5-8 Nov. 1980) ed. by G. Giannantoni (“Elenchos” vi), Napoli 1981, pp. 689-711. .

With Sextus, however, things are rarely simple. For Sextus’ work,dense and thorough as it is, is not intended to inform or to entertain us;it is meant to turn us into sceptics. If we feel, as we likely do, safefrom this intended effect, it is tempting to explore Sextus’ work witha view to the doxographical information which we can extract. Butthis can be unsafe; Sextus’ work is always, even at its most tediousand scholastic, driven by the desire to shape his material to the formmost likely to produce in the reader the reaction which the scepticaims at: reaching the position where arguments for and against pullwith equal force, and so finding oneself suspending judgement on thesubjectThere is an apparent problem, of course, in the sceptic’s self-conception bothas someone who has shed the beliefs which trouble other people and as a personconcerned to practice on others the therapy of ridding them of their beliefs. I shallnot pursue this question further here; for discussion see the chapters on the Sceptics,especially in Part 4, in my forthcoming book on Greek ethics, The Morality ofHappiness (Oxford).. Sextus’ strategies of opposing various schools and thinkersto one another obviously derive from this drive. But they also, I think,complicate the picture when it comes to his sources. In what followsI shall set out the case for a more complex picture of Sextus’ relationto the Peripatetic school.

As Sextus uses the phrase “Peripatetic”, it mostly serves to include,rather than to exclude, the distinctive views of Aristotle, the school’sfounder and by far its most dominant figure. Thus he once refers to«Aristotle the Peripatetic» (PH iii 31), and at M x 30-33 a beginningreference to «the philosophers from the Peripatos» is shortly followedby three explicit references to Aristotle, one to “Aristotle’s followers”and another to “the Peripatetics”. Where Aristotle is played off againstother Peripatetics, such as Strato, this is clearly signalled. I shall, then,proceed by first examining material which clearly relates to Aristotleor to Peripatetic theses clearly deriving from Aristotle, and only thenturn to Sextus’ treatment of the distinctively different contributions ofother PeripateticsJanáček’s Indices reveal a wide variety of phrases: hoi peri (ton) Aristotelen,hoi Aristotelikoi, hoi Peripatetikoi, hoi apo/ek tou Peripatou..

II

It is obvious that in some places in his treatment of Aristotle andhis school Sextus is drawing on sources quite distinct from their ownwritings — on Hellenistic doxographies, which reshape Aristotle’s ideasto fit a Hellenistic philosophical agenda, and which often recast themin specifically Stoic terms. This is most obvious in the long account ofperception in M vii, but also seems to be the case in the ethical section.Gottschalk and Moraux have stressed the continuing influence of suchHellenistic doxographies even after the revival of Aristotelian textualstudy initiated by AndronicusH. B. Gottschalk, op. cit., pp. 1172-4; P. Moraux, Diogène Laërce et lePeripatos, in Diogene Laerzio storico del pensiero antico, «Elenchos», vii (1986)pp. 245-94..

Sextus’ references to Aristotle in the ethical sections of his workare brief and glancing. At first sight they seem unexceptionable. He men-tions at M xi 45 the view of «the Academics and the Peripatetics» thatthere are three kinds of good, as opposed to the Stoic view, and thiscertainly suggests that such views as he has about Aristotle’s ethics come,not from Aristotle’s own ethical works, in which the distinction doesnot play an important roleThe distinction appears at Politica, H 1 as a generally accepted intuitiveview, and at eth. nic. A 8. 1098 b 12 ff. Aristotle claims that his own theory accountsfor this intuitive view, but it is not part of Aristotle’s own theoretical accounts,though it seems to have become part of the Peripatetic ethical theory as early asTheophrastus., but from the Hellenistic doxographical tra-dition, in which “three kinds of good” is the slogan for the Peripateticview of our final end, by contrast with the Stoic viewWe find this in the section on Peripatetic ethics in Arius Didymus ap.Stobaeus, as well as in the brief section on ethics in the Diogenes Laertius sectionon Aristotle.. And this issupported by M xi 51 and PH iii 180-181, where the Peripatetics are saidto hold that health is a good, though not the chief good.

However, although Sextus is to some degree drawing on the stan-dard Hellenistic doxographical view of Aristotle’s ethics, there is a greatoddity at M xi 77, where Sextus is discussing disputes about what is really

good. He compares Zeno, Epicurus and Aristotle as examples of dog-matists who try to establish this by argument. Zeno argues that virtueis good, Epicurus that pleasure is — and Aristotle, that health is! Sextushas inserted two examples of final goods into a discussion of what isgoodBury in the Loeb translation tries to save Sextus’ credit somewhat by insert-ing a definite article into the translation, and so having Sextus recognize that forZeno virtue is the good; but there is nothing corresponding in the Greek.. One hopes that we do not have to ascribe to Sextus the com-mitted view that health had the role for Aristotle that virtue had forZeno and pleasure for Epicurus. This would be an unbelievably grosserror. But how do we explain the passage? It is not plausible to accountfor it in terms of sceptical strategy — that is, of supposing that Sextusassumes that his audience might make the mistaken assumption. Sextushas no reason to suppose his audience capable of such an error. It simplyseems that Sextus has been careless; he is not paying proper attentionto the argument, because he does not have a serious interest in ethicsDavid Runia has suggested to me various points which may serve to mitigateSextus’ error somewhat. Sextus is here indicating ethical options, and his centralinterest does not lie in getting the positions correctly attached to authors. His exam-ples are given as examples of what is good, but in a way which suggests the finalgood, not just good in general. Sextus thus simply appeals to Aristotle as an exampleof someone for whom health is the chief good; he falls into this mistake becausehe is thinking of a contrast to the Stoics, who hold that virtue is the good, buthealth is not a good. Aristotle comes to mind as someone whose ethics contrastssharply with this in giving health a prominent place. This still, of course, leavesSextus making an error; the best we can do is to find a diagnosis of the error..This is indeed, unfortunately, what PH iii and M xi confirm. Sextus justshirks the task in ethics, analogous to that in the logical and physicalsections, of examining the major theories about major issues in ethics,such as the nature of our final end. Instead he gives us a few generalarguments about ethics, and a few sets of arguments about Stoic andEpicurean theses, most of which are not only feeble but go off into irrele-vant non-ethical points. The ethical section is pathetic by comparisonwith the long and knowledgeable sections on logic and physicsThe inadequacies of Sextus’ account of ethics are further discussed in myScepticism about Value (forthcoming in the Proceedings of Scepticism: A Pan-AmericanDialogue, a conference held at the Center for Ideas and Society, University of Cali-fornia, Riverside, February 1991)..

Thus Sextus’ failure to consider Aristotle’s ethics with the care ac-corded to the logic and physics seems to derive simply from a lack ofcompetence in ethics, and we cannot infer anything from it about thenature of Sextus’ ethical sources. This is a major disappointment, con-sidering the extent and richness of the Hellenistic sources for Peripateticethics, and the importance here of Antiochus, whom Sextus ignores inethicsSextus couples the Peripatetics with the Academics in holding that there arethree kinds of good, suggesting the Antiochean conflation of ethical traditions.However, he shows no awareness of the argument that the Stoic views are only ver-bally distinct from the Peripatetic ones. Moreover, in drawing the contrast with theStoic view he seems to get the latter wrong (M xi 46, PH iii 181). There is nothingto suggest that Sextus has seen the point of Antiochus’ procedure. Indeed at Mxi 173 Peripatetics, Stoics and Epicureans are said to have “different” accounts ofthe "techne of life”. And at M xi 3 seems to accept the Antiochean conflation ofAcademic, Peripatetic and Stoic theories, only to ascribe to this tradition a divisionof good, bad and indifferent things — a division which makes scant sense for thePeripatetics, and suggests again that Sextus has not understood his ethical theories.As Jonathan Barnes has suggested to me, Sextus is temperamentally opposed to syn-cretistic accounts such as Antiochus’, which downplay the differences betweenphilosophers of which Sextus makes so much. although he seems to be using him as a main source in M viifor epistemology. We must just accept that Sextus shows no knowledgeof Aristotle’s ethics other than slight acquaintance with some ethical dox-ography. Moreover, he lacks interest to the point of being careless onan important point.

III

Sextus’ account of Aristotle on perception is less depressing. Theaccount, in M vii 217-226, very clearly comes from a Hellenistic doxo-graphy, probably Antiochus’ Canonica, as is suggested by the referencesto Antiochus in this long stretch of M vii, though not in the Peripateticsection itselfM vii 162, 201. But for arguments against finding Antiochus to be Sextus’source here, see J. Barnes, Antiochus of Ascalon, in J. Barnes and M. Griffin (eds),Philosophia Togata; Essays on Philosophy and Roman Society, Oxford 1989, pp. 51-96,esp. pp. 64-8.. «Aristotle, Theophrastus and the Peripatetics in general»,Sextus informs us, have a twofold criterion, «perception for perceptible

things and thinking (noesis) for thinkable things, and common to both,as Theophrastus used to say, is the evident (to enarges)». It is obviousat once that Hellenistic concern for “the criterion” has created a laterframework into which Aristotle’s concerns in the De anima and Parvanaturalia have been forced to fit. There is of course nothing objectionablein principle about writing history of philosophy from a thematic point ofview, and giving an account of past philosophers in terms of their contribu-tion to the chosen philosophical theme; that is in fact what much doxo-graphy consists of. The present passage, however, gives us an interestingexample of what happens when a theory is interpreted in terms of concernswhich are absent or peripheral in it. Thus, epistemological concern about“the criterion” is read into Aristotle’s theory, in which epistemologicalconcerns are far less prominent than they become in the Hellenistic peri-odSee C. C. W. Taylor, Aristotle’s epistemology, in Epistemology Companionsto Ancient Thought, ed. by S. Everson, Cambridge 1990, pp. 116-42.. The role of phantasia is greatly expanded, is made far more cen-tral and is read into the account of perception itself. It is used as a basisfor giving Aristotle a theory of representation based on similarity (seeespecially 220)See D. Glidden, Aristotelian Perception and the Hellenistic Problem ofRepresentation, «Ancient Philosophy», iv (1984) pp. 119-31. The perspective of An-tiochus (if he is the source of this passage) is becoming newly fashionable; see S. Ever-son, Aristotle on Perception (forthcoming), especially chapters 3 and 4; V. Caston,Aristotle on Mental Representation, forthcoming. And Alexander of Aphrodisias alsoexpands the role of phantasia in giving an account of the Aristotelian theory.. Thus the Peripatetic school is given fashionable newtheories of representation and the criterion, and cast in the mould whichSextus uses to put together all Dogmatic theories of perception beforeproducing his own systematic objections. It is interesting that, despitethe explicit reference to Theophrastus, the material used seems to comefrom AristotleThis is the only mention of Theophrastus in Sextus, apart from a purelyhistorical reference in M i 248. However, see P. Huby, Theophrastus and the Criterion,in P. Huby and G. Neal (eds), The Criterion of Truth, Essays written in honourof G. Kerferd, Liverpool 1989,pp. 107-22 for a defence of the claim that the Sextuspassage derives largely from Theophrastus. She adduces interesting parallel The-ophrastean material from Clement of Alexandria.. We find use of the potentiality/actuality distinctionand an account of the genesis of techne and episteme that recalls

Aristotelian texts Cfr. metaph. A, Analytica Priora, B 19.. In fr. 16 of Arius Didymus’ physical doxographywe find a short passage which seems to derive from the same sourceas the Sextus passage, since it ascribes to Aristotle the “double criterion”,and which repeats from the De anima the derivation of phantasia fromphaos H. Diels, Doxographi Graeci, Berlin 1976 (repr.), p. 456, 5-12.. Antiochus, if he is the source, is working from a knowledge ofthe psychological works. He is producing a theory-driven account whichelevates phantasia and problems of representation from marginal and frag-mentary status to centre stage, and frankly imports a concern with “thecriterion”. The account seems to owe nothing to Theophrastus exceptthe use of the phrase to enarges, which is employed to provide a linkbetween Aristotelian psychology and Hellenistic epistemology. As for thelater Peripatetics, their contribution to either is negligible, and Sextusor his source presumably brings them in as merely carrying on Aristotle’sideas in this areaOn the paucity of later Peripatetic contributions to these problems, see myHellenistic Philosophy of Mind, Berkeley 1991, Part I..

While in the ethics Sextus seemed to use doxographical accountsmerely out of lack of interest, in this section it is clear why he usesa Hellenized version of Aristotle. For the Hellenistic perspective on is-sues of perception and thinking was distinctively different from Aristo-tle’s, in just the two respects in which we find Sextus’ Aristotle reshaped.An account of perception had to include a response to the problem of“the criterion”, and it had to have some account of phantasia and theproblem of representation. This is Sextus’ own perspective, and we canreadily see why he uses an account of Aristotle which slots him into placein a familiar Problematik, rather than turning to the psychological worksthemselves. From his perspective they would have seemed not only dif-ficult but maddeningly elusive on the main issues. Even had Sextus hadthe De anima and Parva naturalia handy on his shelves, he would havehad little reason to peruse them. To the extent that he could have real-ized that what they contained was rather different from the Antiocheanaccount, Sextus would have found them marginal and irrelevant to whathe saw as the issues of philosophical importance in the area of perception.

Should we fault Sextus all the same? It seems at least somewhatlazy to assume that all philosophers are well represented by their contri-butions to a single set of currently fashionable philosophical problems.But here it is salutary to recollect Sextus’ sceptical perspective. The scep-tic wants to free people of the dogmatic beliefs that they are unfortunateenough to suffer from. Where philosophical beliefs are concerned, it islikely that what is making most people suffer are beliefs formed in the con-text of current philosophical debate; and this is what Sextus accordinglyconcentrates his attention on. He is not obliged to hunt out exploded doc-trines or to strive to recover outdated perspectives; this would be a wasteof time, since few or none are likely to be suffering from the effectsof belief in them, and thus the sceptic has no motive to argue againstthem, and therefore no motive to disentangle them from later recasting Note that what Sextus finds to be currently debatable is very different fromwhat we find in Alexander of Aphrodisias, for example. Sextus is writing for anaudience whose philosophical interests are determined not by Peripatetic specialists,but by philosophical debates shaped by the earlier Hellenistic handbooks. The im-portance of this point was emphasised to me by Jonathan Barnes..

IV

Sextus also clearly had access to another source of philosophical in-formation about Aristotle apart from the school treatises, namely the “ex-oteric” works, which continued to circulate even after Andronicus’ edi-tions of the school works became widely availableH. B. Gottschalk, op. cit., p. 1172..

In the second mode, for example, (PH i 84) Sextus says explicitlythat «Aristotle tells us about» a man from Thasos who continually saw«an image of a person» going in front of him. Aristotle does discussthe case, at Meteorologica, 373 a 35-b 10 — but without the detail thatthe man came from Thasos. It is possible, of course, that this detail hasjust dropped out of our text of the Meteorologica, or that Sextus or anintermediary source has put it in. However, earlier in the paragraphSextus has described the case of the amazingly thirst-free Andron ofArgos, and, although he does not ascribe it to Aristotle, we know fromparallels that Aristotle mentioned this in his work On DrunkennessFr. 4 of the Symposium, Ross..

And one can readily imagine many contexts in a work on drunkennessin which the man from Thasos could have figured in a less scientificway than he does in the MeteorologicaThese points were made earlier in J. Annas and J. Barnes, The Modes ofScepticism, Cambridge 1985, in our treatment of the second mode, p. 61. It wasJonathan Barnes who noticed the relevance of Andron of Argos.. It seems reasonable to concludehere, then, that Sextus had access to Aristotle’s work On Drunkenness.

TwiceM iii 57-59; M ix 412-413.. Sextus tells us explicitly that Aristotle held a certain viewas to how we come to conceive of the length without breadth that geome-ters make use of. In the passages he makes use of some later terminology,for example the term ennoia, but seems at any rate to be applying himselfclosely to a particular passage of Aristotle. «He bases the argument on anobvious and clear example» says Sextus at M iii 57, namely, that we cangrasp the length of a wall without thinking in addition of the breadth ofthe wall. In the same passage he adds that we infer from this example tothe conceivability of length without breadth on the basis of the principlethat « apparent things are a glimpse (opsis) of what is not evident»; it is notclear whether the use of this tag, and with it the inference, are to beascribed to Aristotle or not. (Sextus himself is quite partial to quoting it.)In both passages Sextus adds his own criticism, namely that this showsonly that we can conceive of the length of the wall minus the breadth ofthe wall, not minus any breadth, which is what was required.

This passage is not to be found in any of Aristotle’s extant works,and has no close analogue. At first sight it seems akin to the argumentsof Metaphysica, M 3, where mathematical objects are said to be “sepa-rated in thought” from physical objects, and where Aristotle’s solutionto the problem of the status of mathematical objects is, roughly, thatmathematics studies physical objects, but not qua physical objects, rather,qua length, breadths and so on. This is not a theory of “abstraction” ifsuch a theory implies that the existence of mathematical objects is in anyway dependent on the mental activity of mathematicians in mentally ab-stracting lengths, breadths etc. from the objects whose lengths, breadthsetc. they areIn my commentary on this chapter in J. Annas, Aristotle’s ‘Metaphysics’ Mand N , Oxford 1976, I did make this claim. I defend the correct view in my DieGegenstände der Mathematik bei Aristoteles, in Mathematik und Metaphysik bei Aristo-teles, ed. by A. Graeser, Bern/Stuttgart 1987, pp. 131-47.. However, it does seem to commit Aristotle to some kind

of psychological theory which explains how we can separate lengths,breadths etc. in thought from the objects whose lengths, etc. they are.This passage is a contribution to such a theory, as is De memoria,449 b 31-450 all. Neither passage is very developed, although Aristotleclearly needs some such theory to explain how it is that we can cometo consider physical objects but not qua physical.

As for the source of the passage Sextus has in mind, we have, asfar as I can see, absolutely nothing to go on. Ross assigns it as fr. 3of On the Good, and this seems reasonable, supposing Aristotle’s accountof Plato’s lecture to have also included criticisms of it; On the Goodis the obvious place for discussions of mathematical matters. But it couldequally well come from On PhilosophySee fr. 11 Ross for discussion there of philosophical problems aboutmathematics. There was confusion between the two works in antiquity; see Simpl.in de an. 28. 7-9 (fr. 11 Ross). or a discussion of thinking insome other works. In any case, it clearly comes from one of the “exoteric”works.

At the beginning of M vii 6-7, in the context of disagreements asto the divisions of philosophy, Sextus comments that, «Aristotle says thatit was Empedocles who was the first to start up rhetoric, to which dia-lectic is antistrophos, that is, isostrophos, since they are concerned with(strephesthai) the same matter (just as Homer calls Odysseus antitheos,which means isotheos). Parmenides would seem not to have no experienceof dialectic, since Aristotle again supposed his friend Zeno to be the lead-er (archegos) in dialectic». This passage is intriguing. Some passages inDiogenes LaertiusDiog. Laert. viii 57, 63; ix 54. tell us that in his dialogue The Sophist Aristotlepresented Empedocles as the discoverer of rhetoric and Zeno of dialectic.It seems, then, that Sextus had access to this dialogue. But the Diogenespassages say nothing about the celebrated contrast of rhetoric and dialec-tic, familiar to us of course from the opening lines of the Rhetoric. Sextusclearly wants to put this in, if only for the sake of adding his own expla-nation of the difficult antistrophos. Was The Sophist the original homeof this contrast? Or is Sextus referring to the opening lines of the Rhe-

toric? At M ii 8-9 Sextus mentions several definitions of rhetoric fromAristotle «in the first book of his Rhetorical Technai». Sextus seems tobe like many since who show knowledge of the opening chapters ofBook i of the Rhetoric, but no knowledge of, or possibly no interest in,the rest. An any rate, Sextus seems here to be acquainted with Aristotle’sSophist as well as the opening of the RhetoricThe point about dialectic and rhetoric seems to have become a common-place; Alex., in top. 3, 25 ff. makes it, with an explanation of antistrophos similarto Sextus’..

At M ix 20-22Fr. 12 a of On Philosophy, Ross, who prints 26-27 as fr. 12 b, althoughAristotle is not named there. Sextus tells us that Aristotle used to say that theconcept (ennoia) of the gods had two sources: from «what happens withregard to the soul» and from the phenomena of the heavens. In states likesleep the soul is “by itself” and takes “its own nature” and produces in-spired states and prophecies. This happens also at death; witness theprophetic last words of Patroclus and Hector in Homer. This leads peopleto conceive of the existence of god as an intelligent being like the soul. Se-condly, the orderliness of the progressions of the heavenly bodies leadspeople to think of god as the cause of this orderly motion.

The second of these grounds reminds us of a familiar passage in theMetaphysicsMetaph. A 8. 1074 a 38-1074 b 14. But Enrico Berti has pointed out thatMetaphysica A does not seem to have been well-known in antiquity, and this mightindicate that the reference is rather to On Philosophy.; but the first is alien to anything in the school treatises, andin particular to the kind of account that we find in the work on divinationin dreams in the Parva naturaliaCarlo Natali has pointed out the relevance of the end of eth. eud. Θ 2; butI do not think that the similarity is close enough to suggest that that passage isthe actual source of the reference.. We need not suppose, of course, thatin what Sextus reports Aristotle was endorsing these reasons for comingto conceive of god; he is presumably reporting endoxa, in his customaryway. However, we nowhere in the school treatises find him treating thefirst kind of endoxon with such respect. We seem to find something at leastdifferent from the school treatises, even though we should be cautious ininferring that in the “exoteric” work Aristotle was committed to aphilosophically very different view from the one we find in the treatises.

Sextus, then, appears to be acquainted with a selection of Aristotle’s“exoteric” works. It is not plausible to suppose him acquainted onlywith an intermediate source containing selections from The Sophist, OnDrunkenness and On Philosophy. And I have hyper-cautiously stuck onlyto contexts where Aristotle is explicitly named. Many think that a pas-sage at M x 248-283 contains material from On the Good, and they areprobably right, although we should be cautiousThe Tübingen school of Plato interpretation takes this whole passage to bea report of Plato’s lecture On the Good, mediated by Aristotle’s account. Thus itfigures in its entirety as Testimonium 32 in K. Gaiser, Platons Üngeschriebene Lehre,Stuttgart 1968. Sextus is, however, explicitly talking about “the Pythagoreans”, andat 258 there is an actual reference to Plato, making it in my view unlikely that thewhole passage goes back to a Platonic source via Aristotle. For a more cautious view,which uses Sextus but relies more heavily for Aristotle’s On the Good on Alexanderof Aphrodisias, see the Introduction to my Aristotle’s ‘Metaphysics’ M and N, cit.. So Sextus has accessnot just to doxographies, but also to the more popular sources forAristotelianismIt is probably in this context that we should mention that he has access alsoto chatty gossip, such as that Pythias, Aristotle’s daughter, had three husbands(M i 258)..

V

What of passages which plausibly suggest that Sextus is using someof the school treatises? He seems to know of them, since he characterizesAristotle as someone who uses rebarbative technical termsM i 315; his examples are entelecheia and to ti en einai., but doeshe actually make use of them? Our best bet here is the “physical” divi-sion of philosophy. For, while Aristotelian ideas do not dominate in dis-cussions such as that of cause, and Aristotelian concepts such as naturereceive scant attention, we do find two striking things, which will bediscussed in this section and the next. One is that Sextus sometimesmakes what look like references to particular passages in Aristotle’sPhysics and Metaphysics.

At M ix 7-8, for example, Sextus in listing earlier proponents ofthe view that there are “material” and “efficient” principles mentionsAnaxagoras’ Nous as an example of an efficient principle, and adds

that Aristotle says that Hermotimus of Clazomenae and Parmenidesand much earlier Hesiod shared this view, since they all introducedLove as a moving and uniting factor among the elements. He adds aquotation from Parmenides and one from Hesiod, both introducing Erosas a cosmic factor. This is a fairly close reference to Metaphysica, A3-4. 984 b 15-31, where we find Anaxagoras mentioned as someone whorecognizes the importance of a kind of factor hitherto ignored, andHermotimus, Hesiod and Parmenides are cited as predecessors, with thesame two quotations.

We might wonder whether Sextus has actually read Aristotle whenwe notice that in Aristotle this group of philosophers is mentioned asintroducing not the efficient cause, which has been already discussed,but the final cause; what unites these philosophers from Aristotle’s pointof view is not their recognition of the need for something to get thingsmoving, but their recognition of a factor like cosmic Love, which aimsat something. Possibly, then, Sextus is citing Aristotle at second hand.But would a doxographical summary include Hermotimus, a somewhatpointless reference to those no longer familiar with his work? Would itinclude the quotations from Hesiod and Parmenides? And we can easilyexplain Sextus’ procedure another way. Sextus is reading Aristotle froma perspective in which Aristotle’s four causes schema is out of date, andin particular in which teleology is not seen as a problem of the form itseems to be to Aristotle. Sextus begins M ix by dividing the archai ofthe physicists into the drasterioi and the hulikai. In terms of this scheme,Aristotle’s comments can reasonably be seen as falling under Sextus’heading of drasterioi causes, even if their original home was in a morefine-grained schema.

We may well object to this, of course, on grounds of historical inac-curacy, or unfairness in argument. But Sextus, to repeat, is not concernedwith these matters. He is concerned to produce in the opponent a stateof equipollence, and thus epoche, in this case about physical principles.This will be achieved only by attacking beliefs and attitudes which theopponent is likely to hold. No purpose is served, indeed effort is wasted,by digging up past issues and perspectives which are no longer foundcurrent. Thus Sextus adapts Aristotle’s words to a current perspectiveon the basis of “physics”, because that is the only way of making them

relevant to his task in writing the book. Thus we do not have to sup-pose an intermediate source which had already simplified or distortedAristotle’s original scheme. Sextus may just as well have read theMetaphysics text for himself and adapted it to his own sceptical purposes.

There is another example at M x 45-46, the introductory discussionof kinesis. In the corresponding passage PH iii 65 Sextus mentions Par-menides and Melissus as philosophers who deny the existence of kinesis;in the M passage he adds the point that Aristotle calls them stasiotai ofnature, from stasis, and aphusikoi, because nature is an arche of kinesis,and by removing kinesis they abolish nature.

We turn at once to Physica, A 2-3, especially 184 b 15-185 a 20, thepassage where Aristotle gives three good reasons why the student of na-ture need not bother with those whose arche is akineton, like Parmenidesand Melissus, before going on to argue against them anyway. It is strikingthat Aristotle uses neither the term stasiotai nor aphusikoi there. Sextusis correct in his main point, that Aristotle claims that these philosophersare irrelevant to a study of nature, since by denying change and move-ment they are attacking the first principles of the subject, and the practi-tioners of a subject do not, as practitioners, have to refute attacks onthe first principles which establish their subject. Further, the point thatattacking kinesis abolishes nature is a peculiarly Aristotelian point, whichSextus retails without his own strategy being committed to it; in generalhe ignores nature as a basic concept in physical philosophy. Why,however, has Sextus added the colourful terms? Perhaps, of course, hemay have found them in a work by Aristotle now lost, the contents ofwhich, minus the colourful terms, we now find in the PhysicsBury ad loc. suggests that Aristotle derives the phrase stasiotai from Plato’sTheaetetus, 181 a, where the Eleatics are called tou holou stasiotai. However, theplay on words is obvious enough to occur to a philosopher without benefit of theTheaetetus, and in any case this explanation leaves aphusikoi unaccounted for.. Butperhaps Sextus produced the colourful terms himself in order to sum upAristotle’s discussion.

In both these cases we have what looks rather like a citation froma passage of Aristotle. In both cases it comes from very near the begin-ning of the work; in neither case is it the kind of point which we wouldexpect to find in a doxographical summary. In both cases there is some

distortion of the original intent of the passage; but both times this canadequately be accounted for in terms of Sextus’ own sceptical purposein using these citations. This does not amount to a strong case thatSextus had read the original; but likewise it does not straightforwardlyconfirm the thought that he had not; we are given, I think, groundsfor equipollence and thus for epoche.

VI

We also find, in the physical sections of Sextus, long stretches ofargument which at least appear to derive from passages in Aristotle’sPhysics. The most striking are the discussions of place (topos) atPH iii 119-135 and M x 6-36On these see M. Burnyeat, The sceptic in his place and time, in Philosophyin History, ed. by R. Rorty, J. B. Schneewind, Q. Skinner, Cambridge 1984, pp.225-54. Burnyeat is concerned with the philosophical status of Sextus’ arguments,but his discussion brings out how closely these depend on Aristotelian arguments..

Sextus mentions Aristotle and the Peripatetics in these sections, butin the PH version he spends longer attacking the Stoics than the Peripa-tetics, and he nowhere suggests that his whole approach derives fromthe Aristotelian one. Still, this appears clearly to be the case.

The PH version opens with a distinction between a loose and anexact usage of “place”. In the loose use, a thing’s place is just whereit is intuitively said to be, like “my city”; but in the exact use, it is thething’s “exact enclosure”. Sextus thus begins from the Aristotelian con-ception of place as the limit of the surrounding body (Physica, Δ 4.212 a 5-7). The distinction between the loose and the exact usage of“place” does not itself bear any philosophical weight; in the discussionsof motion Sextus frustrates any attempt to evade problems about placein the exact usage by appeal to the loose usage PH iii 75-76; M x 95-96, answered at 108-110.. The distinction sim-ply seems to establish the point that, while there are loose intuitive usesof “place” in which someone’s place is «in Alexandria or in the gym-nasium or in the school» (M x 15), the object of philosophical interestis something more exact, and this Sextus simply identifies, in both ac-

counts, with Aristotle’s account, produced at Physica, Δ 1-5 after discus-sion, of place as the innermost boundary of what encloses the thingI agree with Burnyeat (note 15) that Aristotle’s addition at 212 a 20-21 ofthe point that the boundary must be static is not problematic (contra E. Hussey,Aristotle’s ‘Physics’, Books III and IV, Oxford 1983, pp. 117-8)..

Sextus follows this with a list of arguments to establish the existenceof place. Place must exist if its parts do, these being right and left, upand down, before and behind; we see one person coming into the placethat another has left; things have natural places, the naturally light goingup and the naturally heavy down; Hesiod, in saying that Chaos camefirst, lends authority to the idea that things must have places to be in;the existence of body presupposes the existence of place; if we give senseto one thing’s coming about because of something, and one thing’s com-ing from something, then we can give sense to one thing’s being in some-thing. All but the last two of these points come straight from Physica,Δ 1, where Aristotle is, as usual in the Physics, collecting reputable opin-ions on the topic, and here collecting the endoxa which support the exis-tence of place.

When it comes to the countering opinions, however, which suggestthat there is no such thing as place, Sextus does not follow the rest ofPhysica, Δ 1 — unsurprisingly, since Aristotle is setting up the considera-tions pro and con in a way which will lead to a positive solution, whereasSextus is concerned rather to have the two sets of general considerationscancel each other out at the start, before going on to the specific argu-ments. He therefore just tries to counter each of the considerations hehas mentioned (often rather feebly).

The specific arguments against the Peripatetic position are ratherbrief and feeble (131-134). Sextus brings against the Aristotelian accountof place a set of arguments which do not rely on special features of thataccount, but just recycle generic sceptical strategies. As far as PH is con-cerned, we can see that Sextus knows the Aristotelian account, and thathe bases his own initial general arguments on it as well as arguing brieflyagainst it; but the only text which he seems to know well is one givingthe content of the first part of Physica, Δ 1.

The M account is fuller and more interesting — and it gives the

impression that Sextus is working from the whole Aristotelian accountof place. He begins again by giving the considerations mentioned inPH iii in favour of place; they are given more fully, and in a way whichrefers more to the text of Physica, Δ 1. Thus the point about replacement,illustrated both times by a non-Aristotelian example, is here illustratedalso by the example of water poured out of a jar, recalling Aristotle’sown exampleM x 8; Physica, Δ 1. 208 b 1-8. The example is not exactly the same;Aristotle says that water goes out “as if” from a jar, and is replaced by air; Sextususes a simplified example of liquid poured out of a jar and replaced by other liquid.. The point about natural places is expanded by refer-ence to fire being naturally light and water being naturally heavy; again,though Aristotle mentions earth rather than water as an example of thenaturally heavy, the details seem plausibly lifted from the discussion inPhysica, Δ 1. Even the point about our making sense of various locutionsabout “because of” and “from” something is expanded in an Aristotelianway: the options are labelled as as matter (hule), cause (aition) and end(telos)At 12 Sextus adds another point which is not in Aristotle: we can imagineaway body, but not the place that the body is in..

As in PH, Sextus proceeds to counter the arguments we have seen,as he did in PH and for the same reason. He then adds (19-23) a versionof the feeble arguments we have seen in PH. But from 24 to 36 weget arguments that have no analogue in PH, but seem to draw from laterparts of Aristotle’s discussion. The argument at 24-29 starts by claimingthat if place contains body, then it must be one of four things: matter,form, the extension between the body’s limits, and those limits them-selves. Sextus claims at 29 that he has shown that it can be none ofthese; therefore, there can be no such thing. Manifestly, Sextus has takenover the schema of Physica, Δ4.211b5-212a30, where Aristotle givesus exactly these four alternatives — except, of course, that Aristotleargues for the fourth alternative whereas Sextus tries to knock it outof the running also. Even though Sextus does not take over Aristotle’sarguments for the first three cases, his objections seem to be based onclaims elsewhere in Aristotle’s discussionPerhaps Aristotle’s own arguments seemed too tied to Aristotle’s own dis-cussion, and insufficiently compelling as independent arguments. Sextus ignoresAristotle’s argument that matter, but not place, is inseparable from the object(211b 36-212 a 2; cfr. earlier, 209 b 22-24), but he does argue that form cannot beplace because it is inseparable from matter (26).; and the fact that he gives

exactly the four Aristotelian alternatives suggests fairly close dependenceon Aristotle’s text.

And finally, from 30 to 36, Sextus again sets up the Aristoteliandefinition (which he has just supposedly destroyed along with the otherthree) and proceeds to use it to reduce to absurdity one of the mainconclusions of Physica, Δ 5, namely that the world as a whole is notin place. Sextus brings three arguments against this. Two are not veryinteresting (if the world is not in place, it is not anywhere; it is absurdfor the world to be its own place)The objection at 34 just repeats a point already made about any place: thereis an alleged absurdity whether it is body or non-bodily. This does not seem to belongparticularly well here.; but one (33) brings the Physicsconclusion into absurd collision with the Metaphysics conclusion that Godis outside the heavens. Sextus, of course, does not mention any of thearguments about a first cause which give sense to the latter claim; hejust argues that God will have to fulfil the role of containing boundaryto the world, and thus be the world’s placeI think that this stretch of argument is Sextus’ reason for saying in theethical section (PH iii 218) that Aristotle’s god is bodiless and is the peras tououranou, rather than reliance on a lost source or, as Bury suggests, over-interpretationof De caelo, A 9..

These two last stretches of argument, and the way the second con-tinues from the first, strongly suggest that Sextus had access to whatwe call chapters 4 and 5 of Aristotle’s discussion of place. Indeed, thewhole M discussion strongly suggests this also. Of course, this does notamount to proof that Sextus had read a text of what we call Physica,Δ. But, supposing him to have access only to an intermediary account,this would have to be far different from the kind of doxographical ac-count from which he derived the information that Aristotle thinks thathealth is a good and that there is a double criterion. His supposed sourcewould have to contain something corresponding exactly not just to Aristo-tle’s definition of place but to entire stretches of argument in what wecall chapters 1, 4 and 5 of the discussion. It is really more economicalto suppose that Sextus had read Physica, Δ 1-5 for himselfKeimpe Algra has suggested that at M x 28 Sextus seems to miscontrue theAristotelian account, suggesting that he might be dependent on a handbook ratherthan the original text. However, I think that it is also possible that Sextus’ miscon-strual may be quite deliberate, intended to facilitate criticism; he does not, afterall, accept the account himself..

VII

Sextus’ apparent reliance on Aristotle’s actual text for the discussionof place is even more striking when we bear in mind that he seems notto have the same relation to the rest of Physica, Δ. There is no discussionof the dogmatists on the voidThe void turns up in Sextus’ discussion of the Stoics on place, in PHiii 124-130, a passage to which nothing corresponds in the M version. Sextus con-trasts Peripatetics, as rejecters of a void, with Epicurus at M viii 332, but saysnothing as to their reasons for rejecting it., and when it comes to time, Sextustreats it very differently from place. The discussions of time in both PHand M show no special interest in Aristotle’s view; the Epicurean viewof Demetrius Lacon, and that of Aenesidemus, occupy more of Sextus’attention. Further, at PH iii 136-137 we find, among the definitions oftime that Sextus lists, «Aristotle, or as some say Plato, [define it as]the number of the before and after in movement, and Strato, or as somesay Aristotle, as the measure of movement and rest» (cfr. M x 228). Onthe other hand, at M x 176-180 Sextus does not hesitate to ascribe toAristotle the definition of time as the number of the before and afterin movement; after bringing an objection to it he says that this is whyStrato rejected it and introduced the definition of time as the measureof movement and rest. But later at 229 Sextus regards the two definitionas sufficiently similar to each other and to Plato’s to be lumped togetherfor purposes of argument.

It is possible to explain Sextus’ procedure here by genuine puzzle-ment over Aristotle’s account of time in Physica, Δ 10-14. Aristotle doescall time the number of movement in respect of before and after (Physica,Δ 11. 219 b 1-2), but he also calls time a measure of movement(220 b 32-221 a 1); the idioms of number and measure are used confus-ingly in the passageSee my Aristotle, Number and Time, «Philosophical Quarterly», xxv (1975)pp. 97-113, for some discussion of this., and one can take there to have been genuinedisagreements of interpretation, in the course of which it might seem

helpful to align one of Aristotle’s formulations with Plato and anotherwith Strato. So we do not have to assume that Sextus used a carelessor confused doxography. However, there is nothing on the other sideto compel us to assume that he was working from Aristotle’s actual text,as in the case of place; conspicuously absent, for example, is any discussionof the nun or “now”.

We get the same frustrating result from some of Sextus’ other refer-ences to Aristotelian ideas in the discussion of “physics”. The discussionsof movement, for example, open with Aristotle’s classification of sixkinds of kinesisAscribed to Aristotle at M x 37, and at PH iii 64 to «those who have madea quite complete distinction among kinesis». In M the division is itself the subjectof some discussion., but after this Aristotle quietly drops from view;none of his more interesting ideas about kinesis figure as Sextus’ targets,for example the definition of kinesis as incomplete actuality, or Aristotle’sviews on the identity of kineseis. Sextus seems to have picked up theidea as a handy way to begin his own discussion, but not to give it itsstructure. It is the kind of capsule information that we would expecthandbooks to contain, and we need not suppose that he has read Physica,Δ 2 or Categoriae, 14 for himself.

The same is true of Sextus’ report of Aristotle’s position as to“material principles”: Aristotle holds, he says, that the elements are fire,earth, air, water and to kuklophoretikon soma PH iii 31; M x 316, where Aristotle is said to share this view with OcellusLucanus, and it is made clear that the “fifth element” explains the constitution ofthe heavenly bodies.. This is again capsuleinformation; and Aristotle’s views are not prominent, or dealt with indetail, in the following discussion; nothing pushes us to assume thatSextus had read the De caelo.

VIII

Sextus’ discussion of Aristotelian logic is short PH ii 163-166, 193-198, 204; nothing corresponds in M. and problematic.

The amount of attention that Sextus gives to Aristotelian logic is mi-nuscule compared with the amount he devotes to Stoic logic; and, although

he twice makes references which imply that he is familiar with PeripateticterminologyPH ii 163, 198., there are in fact strikingly few occurrences of charac-teristically Aristotelian logical terms. Repici Cambiano accuses Sextus ofassimilating Aristotelian logic to StoicL. Repici Cambiano, op. cit., p. 693. She takes Sextus to be attributing tothe Peripatetics a theory of the conditional (on the basis of M viii 329 ff.) and ofthe anapodeiktoi (on the basis of PH ii 198). For the latter see note 54 below. Mviii 329 ff., however, does not ascribe to the Peripatetics any interest in the formof the conditional itself, merely a set of beliefs which will lead them to allot a certainset of truth values to the conditional that Sextus is currently discussing., and, although I take the pic-ture to be more complex than this, we certainly do not find in Sextusmuch independent interest in this aspect of Aristotle’s philosophy.

The passage PH ii 163-166 follows a discussion of the Stoic “in-demonstrable” or anapodeiktoi, which Sextus has concluded by demon-strating that all five of the Stoic indemonstrable argument forms (modusponens and so on) fall to a dilemma. In each case a premise is either“evident” (prodelon) or “not-evident” (adelon). However, if the premiseis evident, it is redundant, and so the argument, though formally valid(by our standard) will be asunaktos, that is “non-concludent”, since theconclusion will not be inferred from all and only the relevant premises.And if the premise is not-evident, it will not be granted, and so theargument will still fail to be concludent, since it will fail the definitionof demonstration or apodeixis given earlier at 143: an argument which“reveals” a non-evident conclusion from agreed premisesThe entire passage, and the issues it raises, are discussed by J. Barnes,Proof Destroyed, in M. Schofield, M. Burnyeat, J. Barnes (eds), Doubt and Dog-matism, Oxford 1980, pp. 161-81.. Sextus’ ar-guments against the Stoics are powerful and damaging, as Barnes haspointed out. But when we turn to the Aristotelian section, we find somethings which are disconcerting.

Sextus gives us two examples. Firstly (163) we get the argument:«The just is fine, the fine is good, therefore the just is good»; then (164)«Socrates is human; every human is a living thing; therefore Socratesis a living thing». In each case Sextus makes an objection to one premise.In the first case he has to restate it before declaring his objection, puttingit in the form, «everything which is fine is good». In the second case

it is «every human is a living thing». In each case the objection is: Eitherthe premise in question is “not evident” (adelon, ou prodelon), in whichcase it will not be granted, and the argument will not go through. Or,it is “evident”; but if it is evident it is redundant, since the conclusionwill follow from the other two premises alone.

There are three obvious points to be made about this. Firstly,neither of these examples looks like a standard Aristotelian syllogism.The first example, interestingly enough, is found in Alexander in an. pr.46, 17 ff., where he is commenting on Analytica priora, A 4, on the crucialrole of the middle term, in a part of the Prior Analytics in which generalissues are being debated before the characteristic forms of Aristotelian syl-logism are brought in. The second example is not a regular syllogism be-cause of the singular term “Socrates”Aristotle does have some examples of syllogisms containing singular terms;see G. Patzig, Aristotle’s Theory of the Syllogism, Reidel 1968, pp. 4-8. The pointwas drawn to my attention by Jonathan Barnes.. It is not likely that Sextus in-vented either examples, and they most likely come from a logic textbook.But they are hardly the kind of example we expect in what is supposed tobe an attack on specifically Aristotelian forms of argument.

Secondly, Sextus merely repeats, against the Aristotelian arguments,the dilemma he has already used against the Stoic anapodeiktoi. In eachcase, he claims that a premise is either evident, in which case it is redun-dant, or not-evident, in which case it will not be granted. Later (193)he refers back to this passage and claims that he has shown systematicallythat all the Stoic and Peripatetic apodeiktikoi logoi fail. But he fails toconsider that Aristotelian arguments are not meant to be measuredagainst the definition of Stoic apodeixis; Aristotle does not appeal to theidea that proof starts from premises that are granted and “reveals” aconclusion which is less evident than they are. Sextus fails to show thatthese arguments are faulty by Aristotelian standards.

Thirdly, Sextus does, at 195-197, provide another argument againstthe offending premise «Every human is a living thing», but he does soin a way which is rather odd. He objects to the way it is establishedby epagoge, raising the objection that establishing a generalization on thebasis of particular instances is circular, since it requires the assumption

of the truth of the very generalization in questionSextus gives, as an example of an exception to an apparent generalization,the fact that only the crocodile moves its upper jaw; all other animals with jawsmove the lower one. The example comes from Aristotle’s hist. an. 516 a 23-25, butit is a standard example, and Sextus probably got it from a logic textbook, ratherthan from the Aristotle’s text, which uses different vocabulary.. This has, however,nothing to do with Sextus’ “official” argument against epagoge at 204,where the objection is different, and is based on the claim that eithersuch generalizations will be insecure, since established on the basis onlyof some instances, or establishing them securely, on the basis of all in-stances, will require completing an infinite review of them, which is ofcourse impossible.

Thus Sextus’ only foray against Peripatetic, as against Stoic logicis disappointing; it displays only superficial konwledge, and it tries,without success, to squeeze Aristotelian material into a sceptical argu-ment designed for Stoic material.

It is certainly hard to believe that Sextus had any direct acquain-tance with Aristotelian logic itself; but this is not very surprising. It isquite standard to do logic out of the current logic textbook; indeed, thereis no strong reason for anyone, sceptical or otherwise, to do anythingelse unless they are specializing in logic. What is more striking is thatSextus gets so little even out of his textbooks on Aristotelian, as op-posed to Stoic logic, and pays it so little careful attention. Presumablyin logic the main dogmatic challenge is taken to come from the Stoics,and, while the sceptic needs to know that there is such a thing asPeripatetic logic, he does not need to know much about it. Dogmaticunhappiness, it appears, comes from prepositional rather than from predi-cate logicPresumably Sextus takes his reference to Peripatetic terminology («what arecalled categorical syllogisms» at 163 and «what are called anapodeiktoi by the Peripa-tetics» at 198) from some logic textbook. (Although the term anapodeiktoi is notAristotelian, it may be that the textbook Sextus is drawing on made a connexionbetween the function of the Stoic anapodeiktoi and Aristotelian “perfect” syllogisms:both stand in no need of demonstration, and have a role in other demonstrations.I am grateful for this point to Vincenza Celluprica, who has also made other helpfulsuggestions about the logic section.).

IX

When it comes to individual Peripatetics after Aristotle, I can bebrief, given the valuable work on this in the article by Repici Cam-biano. Sextus mentions seven later PeripateticsTheophrastus gets a couple of incidental mentions (see above, note 15), butSextus does not retail any of his ideas.: Strato PH iii 33 = fr. 43 Wehrli; PH iii 137 = fr. 79 b; M vii 350 = fr. 109; Mviii 12 = fr. 115; M x 155 = fr. 82; M x 177 = fr. 79 a; M x 228 = fr. 79 c., Dicaearchus PH iii 31 = fr. 8 b Wehrli; M vii 349 = fr. 8 a; M iii 3 = fr. 78.,Heracleides PH iii 32 = fr. 119 b Wehrli; M x 318 = fr. 120., Critolaus M ii 12 = fr. 32 Wehrli; M ii 20 = fr. 34., Ariston the younger M ii 61 = fr. 2 Wehrli., Aristoxenos M vi 1; not in Wehrli., and theobscure PtolemyM i 60-61; see L. Repici Cambiano, op. cit., pp. 697-8 note 26..

None of these references suggest a deep or even extensive acquain-tance with Peripatetic works. The reference to Aristoxenus merely refersto him as one kind of “musician”, i.e. a student of musical theoryPossibly this comes from the source from which Sextus gets the informationthat “the Peripatetics” have many arguments to show that phone is not physical(M vi 54)..Ptolemy is mentioned as raising an objection to Dionysius Thrax ondefining grammar; the point most likely comes from a handbook on gram-mar. Critolaus is twice cited for a view of general hostility to rhetoric,and Ariston the younger for the stunning cliché that rhetoric aims atsecuring persuasion; these are very general points about rhetoric, whichSextus duly uses in piling up arguments against rhetoric and again mostplausibly come from a handbook on the subject in question.

Even the references to Strato, Dicaearchus and Heracleides, whichare strewn somewhat more thickly througout PH and the logical andphysical parts of M, do not amount to anything very impressive. Thetwo main references to DicaearchusThe reference at M iii 3 merely notes that, in a sense of hypothesis irrelevantto the argument, Dicaearchus collected hypotheses (plot summaries) of Euripides andSophocles. merely retail the point that he

denied the existence of the soul or identified it with a state of the body.This is kind of capsule or slogan claim which recurs frequently aboutDicaearchus, with no supporting argumentThe eleven passages collected by Wehrli as fr. 8 all show this feature: weare given a simple slogan, but no grounds for holding it to be true.. Sextus merely uses thepoint in a mechanical way, as part of the arguments to shown that humanbeings are not the criterion.

The references to Heracleides are likewise standard; in both cases heis coupled with Asclepiades of Bithynia as holding a theory of anarmoionkoi, which differ from the atoms of the Atomists in being breakable.Nothing more is said of Heracleides and nothing more is made of the con-nexion; although Asclepiades is elsewhere referred to in Sextus fairly copi-ously the theory of the curious onkoi is nowhere discussed in detailFor the theory see J. T. Vallance, The Lost Theory of Asclepiades of Bithynia,Oxford 1990..

The references to Strato are more copious and range wider. Threeof them (PH iii 137; M x 177, 288) we have already seen; “some people”ascribe to Strato the Aristotelian definition of time as the measure ofmovement and rest. The other references are varied; Strato made qualitieshis elements (PH iii 33); he and Aenesidemus hold that the mind “looksout of” the sense-organs (M vii 350); he and Epicurus deny the existenceof a third item, between the words and the world, to carry meaning (Mviii 12); he holds that there are time-atoms but infinite divisibility of mat-ter and space (M x 155). It is notable how many of these references arecomparative: Strato and somebody, or Strato or somebody, holds such andsuch a thesis. This is presumably what we would expect from a doxographi-cal source listing philosophers by their contribution to some philosophicalproblem.

On individual Peripatetics after Aristotle, then, we have no reason totake Sextus to have access to any but standard doxographical sources, andoften he seems to be concerned with sources on the topic in question, ratherthan with the Peripatos itself. And on references to “the Peripatetics” asa school Repici Cambiano has also highlighted Sextus’ lack of specific in-terest. The Peripatos is not stressed among the Dogmatic schools whenSextus talks about them in general. And references to the school standardly

put them in contrast or comparison with other schools, particularly theStoics. Thus at M vii 388 the Peripatetics are put together with the Stoicsand Academics as schools which accept that some phantasiai are true andothers false, in contrast to the alleged extreme views of Protagoras (theyare all true) and Xeniades (they are all false)Cfr. also M vii 369.. At M viii 185 we like-wise find Stoics and Peripatetics put together as holding a more reasona-ble, “middle” view, that not all perceptions are true, as opposed to theextreme views of Democritus and Epicurus. These are clearly mechanicalcomparisons, and their schematic nature is underlined by fact that whenthe same contrast with Democritus and Epicurus crops up again at Mviii 355 the Peripatetics are omitted. It is clear that “the Peripatetics”figure for Sextus as a school which fits into various of his argumentativeschemata, usually in terms of a contrast with other schools which isdrawn from his thematic concerns, not from a direct concern to under-stand the Peripatetics in their own termsCfr. also H. B. Gottschalk, op. cit., pp. 1139-40..

X

Sextus, then, seems to draw on a wide variety of sources for histreatment of Aristotle and the Peripatetics. As has always been clear,he draws on doxographical accounts, particularly for ethics and psycholo-gy. And he also has access to some at least of the more popular “exoter-ic” works. Both of these types of sources continued in currency after therevival of Aristotelian textual study initiated by Andronicus’ edition, sowe get no help here, unfortunately, towards narrowing down possibilitiesfor Sextus’ elusive dates. Given that the Peripatos is not a school whichsets Sextus’ philosophical agenda, it is tempting to conclude that Sextususes this type of source, so much less subtle and difficult than Aristotle’sown texts, just because he is not particularly interested in the Peripateticsin their own right, and so, when he brings them in, he just turns tothe easiest kind of source.

I have suggested, however, that matters are not quite so simple. Forthere are some passages where Sextus at least seems to be referring toa particular passage, or in one case to a quite extensive stretch of argu-ment, in Aristotle’s school treatises, notably the Physics and Metaphysics.It is true that in none of these cases do we have what amounts to demon-strable proof that Sextus had a text of “our” Aristotle in front of him.However, close consideration of the passages makes this quite likely, Ithink, at least for the treatment of topos. In any case, even if Sextusis not reading Aristotle first hand, he is, for these passages, using asource which is very different from the kind of source he uses for theethics and psychology. His sources for the Physics and Metaphysics pas-sages retail not just Aristotle’s definition of the topic in question, butAristotle’s distinctions and some of Aristotle’s arguments. And the topicsdo not fit neatly into the standard Hellenistic schemata so marked in,for example, Sextus’ account of “Aristotle on the criterion”. So at thevery least we have grounds for thinking that Sextus used a third, morescholarly kind of source for Aristotle alongside his other two.

Why would Sextus proceed in such an uneconomical and confusingway? His procedure, however, will seem uneconomical and confusing onlyif we persist in thinking of him as concerned with historical accuracy,and ignore his sceptical purpose, which I have claimed is ever-present.If we bear in mind that the point of Sextus’ works is to turn us intosceptics, then the variety of his procedures makes complete sense. Sextusis arguing against the dogmatic beliefs which people have, which makethem unhappy. And (whether correctly or not we are not in a positionto say) he takes the Peripatetics not to be a school which is prominentlyresponsible for much unhappiness in this regard; he takes the Stoics tobe the real Dogmatic enemies. If Sextus is right about this, then thereis no need for him fo examine the Peripatetics closely in their own right.For Aristotle’s views on the psychology of abstracting mathematicalobjects, for example, it will suffice to mention a passage in an exotericwork, rather than examine his philosophy of mathematics more carefully;for the point about abstraction is just one point in a schema of argumentto which a full account of Aristotle’s views is not relevant. And, as wehave seen, Sextus can fit Aristotle into a general schema of “thecriterion” because he has arguments pro and con which fit current

philosophical interests, and a careful examination of the De anima couldshow only that Aristotle did not fit into this schema; and all that thiscould do would be to demonstrate Aristotle’s irrelevance for Sextus’purposes.

So if we sometimes find Sextus paying attention to Aristotle’s owntexts, this is explained in exactly the same way. Sextus follows Aristotleclosely on place because he takes it that this is the account which thesceptic needs to take seriously. On time, however, the details of Aristo-tle’s account do not matter, because it is theories like those of DemetriusLacon which are taken more seriously and thus provide the real Dogmaticproblem, and which thus are the theories which the sceptic needs tocombat.

Sextus’ procedure is thus wholly pragmaticSextus’ knowledge of and attitude to Aristotle can usefully be comparedwith that of Hippolytus; see C. Osborne, Rethinking Early Greek Philosophy, London1987, pp. 35-67. Osborne stresses that Hippolytus’ treatment of Aristotle is heavilyinformed by his polemical purposes in confuting heresy, and also that he uses bothhandbook accounts and closer study of the texts where each is appropriate. In thecase of Sextus, however, his sceptical purpose provides a further and stronger basisfor his pragmatic attitude to his sources.. To remove Dogmaticbeliefs successfully, he needs to argue against the Dogmatic beliefs whichin fact bother people and have a hold on them. These, he takes it (andwe are not in a position to correct him) will very seldom be Aristotle’sown theories in their original form. So mostly the Aristotle he needsto argue against will have to be the up-dated, Hellenistic Aristotle hefinds in doxographical sources. But occasionally, as with place, Aristotle’sis still the theory with a Dogmatic hold on people. And in such a case,I have suggested, Sextus may quite well have read a scholarly editionof the Physics. As often (and as is appropriate) with Sextus, we cannotachieve certainty. But we can go some way, I think, towards makingsense of what he doesAt the conference, Jacques Brunschwig objected that my view of Sextus’procedure is unfalsifiable: there is no systematic access to the philosophical profilesof Sextus’ dogmatic patients which is independent of Sextus’ own priorities. DavidSedley, on the other hand, objected that it is in fact falsifiable; independently ofSextus, we do know that the Aristotelian account of topos was uninfluential, andAristotelian logic influential, in the period when Sextus is most plausibly to belocated. My thesis is not, however, that from Sextus’ priorities we can reconstructa comprehensive view of the philosophical scene contemporary with him. Sextus’own philosophical selections and emphases have seemed to many to be those ofHellenistic handbooks probably much earlier than Sextus himself. What I claim isthat Sextus attacks only ideas which, whether currently fashionable or not, couldplausibly have an intellectual hold on his contemporaries. Even this attitude is verydifferent from interest in history of philosophy for its own sake. (Think of the promi-nence of Hume in contemporary analytical philosophical debates, and the very remoterelationship that this bears to genuine historical interest in Hume.).

Sextus’ procedure is open to criticism even so. His most notablefailure is in the ethics section, where he fails to engage with, probablybecause he fails to understand, the kind of Stoic-Peripatetic ethicaldebate over our final end which is so marked in our Hellenistic sources.This is simply part of his generally disappointing performance in ethics.In general, however, his treatment of Aristotle and the Peripatetics showsan intelligent use of the potentialities for sceptical attack. From our pointof view, of course, Sextus is a confusing source in the history ofphilosophy. But, as I have stressed, this is not a failure on his part, forhe never aimed to inform us about the Peripatetics, but rather to loosenthe hold on us of any Peripatetic ideas which we might find ourselvescommitted toI am extremely grateful to Jonathan Barnes for helpful comments. In sectionVIII he has saved me from gross errors about ancient logic, and his criticisms havegreatly improved the paper as a whole. I am also grateful to members of the confer-ence for discussion and questions, particularly Vincenza Celluprica and David Runiafor written comments, and also Keimpe Algra, Jacques Brunschwig, David Sedley,Malcolm Schofield, Carlo Natali, Enrico Berti, Walter Leszl and Cristina Rossitto..