david sedley - epicurus' refutation of determinism

8/13/2019 David Sedley - Epicurus' Refutation of Determinism

1/22

EPICURUS REFUTATION OF DETERMINISM

David Sedley

The SwerveA few facts are, I hope, uncontroversial enough to b e set

out without defence. Epicurus inherited Democritus atomicsystem, but modif ied i t in a numbe r of respects. In particular,he so vehemently objected to its rigidly deterministic laws as topostulate a minimal. swerve 1 t X p t I X A t a t ~ in the motion ofatoms, occurring at no fixed place or time - a doctrine whichdoesonot feature. in his meagre surviving writings but is nonetheless amply attested as his; and defended 011 his behalf by Lucretius II 216-93). The swerve a) enables atoms falling throughspace at equal speed in parallel lines to collide occasionally andinitiate cosmogonic patterns of motion; and b somehow or other serves as a necessary condition for the behavioural autonomyof animate beings - a power oft en identified as free will .The latter function of the swerve has been widely debated,but no general agreement has been reached on its interpretation. 1 The overriding puzzle is perhaps this. The swerve is

1 Notably].M. GUY U La morale d picure Paris 1886); C. GIUSS Nl Studilucreziani Torino 1896), pp. 125-69; C. BAILEY, The Greek atomists and picurusOxford 1928); C. DI NO articles collected in his Scritti epicurei Firenze 1974), pp.129-280 orig. pub . 1939-42); D.]. FURLEY, Two studies in the Greek atomists Prince-ceton 1967), Second study ; E. ASMIs, The picurean theoryof free will and its ori-gins in Aristotle Ph. D. dissertation, Yale University, 1970); A.A. loN Hellenisticphilosophy London 1974), pp. 56-61; 1. AVOTINs, N ote s o n Lucretius 2.251 293


2/22

12 DAVID SEDLEYEPICURUS R E FUTAT ION OF DETERMINISM 13

a minimal degree o f i nd eterm in is m, and i nd et ermi ni sm is thenegation of determinism, which is widely held to be incompatible w ith t he existe nce of free will. 2 Yet it is unclear how thecause of free will can be benefited by the introduction ofi nd et ermi ni sm . A ran do m element i n t he b eh av io ur o f o ur cons ti tu en t atoms might make us all thoroughly eccentric, butrandom eccentricity is h ardl y t he same t hi ng as free will or selfdetermination, and might even be held to militate against them.The puzzle is of more than antiquarian interest, for as luckwould have it Epicurus was more or less right. Since 1927,when He ise nb er g publishe d his Un ce rt aint y Principle, mostphysicists have accepted that there is an element of indeterminacy i n t he b eh av io ur o f su b-atomic particles. T he i nn ov at io nwas w elcomed b y some as eliminating t he clash o ft en felt toexist b et ween N ew to ni an physics and t he w id es pread belief i nfree will. Ye t i t has be come no clearer in t he subsequent halfcentury just how th e champions o f free i will are supposed toexpl oi t t his w in dfall . E pi cu ru s certain ly h ad some ideas o n t hematter, and one of our tasks will be to hunt them out.HSCP 84 1980), pp. 75-9; K. KLEVE Id facit exiguum clinamen SO 151980), pp. 27-31; W.G. ENGLERT, Aristotle and EpicutUs on voluntary action Ph. D.dissertation, Stanford Un iv ersity , 1 98 1); P. A. CONWAY, Epicurus theory of freedomof action Prudentia 13 19 81 ), p p. 8 1-9; D. FOWLER, Lucretius on the clinamenand free will II 2.51-93), in this volume. I should also mention the highly unorthodox interpretation of M. BOLLACK, Momen mutatum: la deviation et le plaisir cf. appendix, Histoire d un probleme by and M. BOLLACK and H. WISMANN), in ].Boll.ACK and A. LAKs eels.), Cahiers d e p hilo lo gie 1 Lille 1976), pp. 161-201.FURLEY S interpretation has won little su pp ort, and the upshot of t hi s p ap erwill be a v ind icatio n of h is b as ic hunch o f t he s ep ar at en es s of t he swerve fromvolition, bu t on rather dif ferent g ro un ds f ro m t ho se w hi ch he offered. ENGLERT Sdissertation should be c on su lt ed as a n o ut st an di ng ly i mp re ss ive d ef en ce o f t hemain o pp osing lin e o f interpretation.

To avoid possibly misleading implications, I shall wh ere p ossib le exclu de th eexpression C free will fro m d iscussio n o f the ancient debate. For differences between C free will t t ~ p and t txoo


3/22

14 DA VID S EDLE Y EPICURUS REFUTATION OF DETERMINISM 15That it should occur to him to kill a second bird with thesame st one, by in vok ing t he swerve i n ex pla nat io n o f cosmogony, is not in itself surprising. Anaxagoras, whom Epicuruspronounced his favourite Presocratic,6 had similarly postulatedMind Nous a as the initial cause of cosmogony, and b as af actor sti ll pre se nt as t he controlling power in all ensouledbeings,7 presumably to account for th e autonomy whichdistinguishes them from inanimate objects. It is not inconceivable that Epicurus thought of the swer ve as an atomistic substi

tute for Anaxagorean Mind. I say substitute and not, e quivalent beca use no inter pr etation of the swerve c an makeit do the entire job envisaged by Anaxagoras. The possession ofMind, the most power ful of all things , appe ars to be intendednot just as a necessary but also as a s uffi cient c on di ti on o fautonomy; at any rat e, i t is n ot clear what further conditionsAnaxagoras might require. But the swerve, which may occur unrest rictedl y i n any a to m at any time, an d hence presumably intables and stones as much as in animals, c an har dly be thoughtof as a sufficient condition of autonomy, but a t best as a necessary condition.What sort of nec essar y c ondition? Is the re a dir ec t causalconnexion between individual swerves in the soul atoms andindividual autonomous acts, or some more tenuous relation? Itis cust omary at this p oi nt t o turn to detailed analysis of Lucretius defence of the swerve. My proposal is to leave Lucretiustill last. F or one t hin g, his account belongs to a discussion notof psychology or ethics but of atomic motion. The existence ofpsychological autonomy is his premiss, the existence of theswerve his conclusion. It is not clear that we should expect inthis con tex t any explanation of how t he swerve functions inanimal a nd hum an psychology - indee d, Lucr etius imm edia tesource may not e ve n have supplied the answe r to that question.For another thing, excessive concentration on the Lucretius texthas tended to exaggerate the centrality of t he swerve and t o ob-

6 D.L. X 12.7 59 B 12 DIELS-KRANZ.

scure the fact that Epicurus had a gre at deal m ore of inter est tosay in opposition to determinism. I prefer therefore to examinethe other evidence first,8 starting with Epicurus own writings.If an adequate interpretation can be arrived at i n this way, itwill then be enough if Lucretius t ext proves able to bear areading consistent with it.

g Unlike a number of recent studies, I do not propose t o a pproac h this questionvia Aristotle. I have no doubt whatsoever that many of Epicuros adjustments to earlie r a tomi sm a rose i n response t o Ari st otl e s c ri ti ci sms of Leuci ppus a nd Democ ri tusin the Physics whi ch Epi curus c le arly studi ed wi th c are see e spec ia ll y PURLEY,Two studies Fi rst study ; In Aristotle and the atomists on motion in a void in P.K.MACHAMER and R.G. TURNBULL eds.), Motion and time sp ce and matter: interre-lations in the history philosophy and science Columbus 1976), pp. 83-100; Bra dINWOOD The origin Epicurus concept void CP 76 1981), pp. 273-85, a t pp.282-4; ENGLERT, Ope c it ., p p. 6 4- 71. B ut I am q ui te o ut o f s ym pa th y w it h o n e cu rrent tendency, t ha t o f picturing Epicurus, and Hellenistic philosophers in general,as so in awe of Aristotle that t he y never dared put a foot forwa rd wit hout l ookingup a nd t aking a cc ount of t he gre at man s vie ws on t he t op ic i n q ue st io n. T hi s app ro ac h se ems t o me not o nl y t o b e u nw ar ra nt ed b y t he e vi de nc e b ut also to killmost of the excitement that can be found i n Hel l eni st i c phi losophy. The re i s sti ll aserious question outstanding about which of Aristotle s school treatises were evenr ea di ly a vai lab le i n t h e t h ir d a nd s ec on d c en tu ri es B .C . A nd a t all events there is

l ~ t l to suggest that Aristotle was as widely discussed as Socrates, Plato, the CynICS, or even Heraclitus. Of c ourse Ari st ot l e wrot e on many of t he same t opic s asthe Hellenistic philosophers, and their i deas c an oft en be fruitfully compared. Oncein a while it may b e r ig ht t o su spect d ir ect o r i nd ir ec t in fl uen ce. But w e mu stalso make some allowance for the possibilityof two philosophers arriving at similarviews i nd ep en den tl y - t he h is to ry o f p hi lo sop hy is ful l of suc h c oi nc idence s.When it comes to the swerve, I do not bel ie ve that t here a re i n Ari st ot le a ny sui table criticisms of Democritean necessity to which it is likely to be a response butfor an i nte rpre ta t ion a long t he se l ines, see ASMIs, Ope c it .), a nd I hope to showthat on this occasion Epicurus own criticisms of t he Democ ri t ea n t radi ti on werethe decisive factor. FURLEY and ENGLERT, i n parti cula r, have made i nte re st i ng a ttempts to relate the swe rve t o c ert ai n posit ive views of Ari stot le s, but whi le t hec onne xi ons t hey draw a re ofte n e nl ight ening I am n ot p er sua ded t ha t t he y aree nough t o e st abli sh a di re ct i nfluence .


4/22

16 DA VID SEDLE Y EPICURUS R EFUTATI ON OF DETERMINISM 17

2. Epicurus on DeterminismIn the Letter to Menoeceus (133-5)9 Epicurus stresses theneed t o b e aware o f t he di ffe renc e b et we en necessity, chance,and to 1tcx.p' 1 } ~ ~ . What characterises the third of these is t ha t i thas no ma ste r, and that the blameworthy and the praiseworthygo naturally hand in hand with it. This suggests that 1tcx.pacarries something not unlike the causal sense, d ue to foundfrequently in the Letter to Pythocles so that the class .of t ~ g spicked out as 1tcx.p' is whatever results from ,our own mdiVldualagency, that for which we are responsible, that which dependson us. lO Far from connecting it directly with the swerve a t nofixed place or time Epicurus emphasises, through the contrastwith unstable c hance, t he reli abili ty of to 1tcx.p' 1 } ~ a ~ . That

which depends on us is reliable precisely, one supposes, becauseit lies within our power to bri ng i t about.The same notion, more commonly under the designation to1 } ~ w v cx.u twv, o ur own a ge nc y , recurs i n t he t ex t which willbe central to my discussion. It is a passage which, althoughavailable t o scholars for thi rt y-f ive years, has never yet b ee nread in a textually accurate form or recognised as containing acontinuous and coherent argument. It comes from a book possibly Book XXXVll - of Epicurus major work, nature.The book was apparently one held in particular esteem, becauset he li brar y o f c arbo ni se d pa pyri found a t H er cul ane um i n t he

18th century has provided no fewer than three copies of it

I propose av ytAwvtOC; for yytAWV tOC; in Ep. Men. 133 a nd a cc ep t t he n ee dfor some s up pl eme nt i mmed ia tel y a ft er it. I suggest


5/22

18 DAVID SEDLEY EPICURUS' REFUTATION OF DETERMINISM 19

no means impairs its intrinsic interest as a piece of philosophicalreasoning. It offers a fascinating opportunity to see Epicurus'mind at work on the philosophical challenge of determinism,and his ability at every stage to antic ipate and counter his imaginary opponent' s next move. However , allowance must bemade for the fact that the n natureis no polished handbook ofdoctrine bu t an exploratory work, wri tten for oral delivery in aloose and colloquial style which Epicurus himself is honestenough to call eXoOAE.ax,(a. 13

In my discussion I shall try to exploit a striking parallelwith a passage in Book IV of Lucretius (469-521), whose sourcecan hardly be doubted to be Epicurus. In the first par t of BookIV Lucretius gives the Epicurean's own positive account ofsense-perception. His entitlement to do so presupposes that thesceptic's doubts about knowledge of the external world areinvalid. But his actual refutation of scepticism is reserved for adigression comparatively late on in the discussion. So too in the n naturebook Epicurus' refutation of determinism comes in adigression slotted in near the end of a positive causal account ofhuman motivation. It is striking, too, that both digressions seekto refute the offending doctrine by showing it to be inherentlyself-defeating and untenable. Indeed, both run through more orless the same sequence of tactics: first a self-refutationargument, then a l inguistic argument , and finally a pragmaticargument. All this will be developed shortly. Here, first, is thetext: 14

13 Epic., Nat. XXVIII, fro 13 col. XIII 1, 9-10 sup. (in my edition, CEre 3,1973, pp. 5-83).

14 The corresponding text in the standard edition of Epicurus' papyrus fragments, ARRIGHETTI's Epicuro. Opere (Torino 19732 ; references in edt 1, 1960, donot correspond), is 34.26-30. But a proper comparison can only be made with DI -NO'S text Epicuri ethica pp. 39-47), whose readings ARRIGHETIl follows in themain, or with tha t o f H ] METTE, Epikuros 1963-1978, Lustrum 21 (1978), pp.45-116, 63 f. , s ince ARRIGHETTI himself does not amalgamate the readings of thethree papyri but simply switches from one to another, so that many letters insidehis square brackets are in fact attested in one or both of the other papyri. I plan intime to produce a complete new edition of the book, and hope that I may be forgi-

Ex] [ t i 1 ~ 1 t p ] w ~ ~ p x i i ~ O 1 t E . p l : t [ a ~ a 1llJ.rv eX]IWlcX IJlv~ a o [ e . ] ~ a o e . 0' alJ.cpw [ ~ c x G ] ~ c X [ l]CT tLV e.t

[ x ~ ] t 1tpeX[e]e.WV [ ~ c x t ] O ~ c x v o 1 } a e . w v xext oLcxge[ae.]wv xcxt1tAE.[W] xat l A e X ~ W ~ . wate. 1tCXp ~ l J . a , ~ [ p W ~ O v ] c X 1 t A W ~ eX1t0'YE.'YE.vvlllJ.ivov 7101) y e . ( v e . G 9 c x ~ , [ ~ ] o r c x ~ o r a , xcxt lx 51 t E . p ~ E X O V ~ o ~ x [ a ] ~ ' e X v e X ' Y x ~ v O ~ c X ~ o U ~ 1 t 6 [ p o u ~ ] E . L a p i o [ v ] ~ a

~ a . p ' l l l J . a ~ 1 t [ 0 ] ~ E . y e . [ v E . ] a 9 a ~ xcxt 1tCXPcX ~ [ c X ~ ] ~ l J e [ tE-p a ] ~ [l]e lllJ.WV a u ~ [ w v ] 0 6 e [ a ~ .. . ] e . ~ [1t]cxpcX ~ 7 } V ~ [ u ] -GL[v]a[ ]ua ]ux l9[ ]e.L [ ]

e . ~ ~ a [ ] ~ e x [ ]a[ ]Xe.[ ]v[ ]u[ .]wae.[ l 10V E . ~ e . [ LACUNA APPROX. 60 WORDS) ~ O l J . c p u a ~ x o l

x a t p e x ] ~ ~ T i p a O l J . O ( w ~ xext ~ o r ~ ~ [ ] o ~ ~ x ~ [ v o ] r ~ ~ 6 P O L ,[ WV a u[ ]v l0 0 1} [ wv ]0[ 1t t 1t eXv[ v : . ] ~ w [ : 1te.LO[ (LACUNA APPROX. 12 WORDs)le.CTt1}XE.L, wv ou[... ] 1 t O A e . ( 1 t E . ~ 1ta91l I(Ve.[aaCXL .] 15vou9E.[ ~ ] E . r V A A 1 } A O U ~ xa.t l J . a X E . [ a ] 9 c x ~ xcxt l J . e . ~ c x p u 9 l J . ~ e . ~ v

x . o v ~ e x ~ xat lv e . a u ~ o r ~ -c1}v a l ~ a v xcxt ouXt tV t i l ~ p x . 1 i ~~ 6 v o v a u c . T t a a e . ~ xcxt 'lv t i i ~ 1 t E . p ~ i X O ~ o ~ xcxt 1 t E . ~ a L 6 ~ o ~ X C X ~ ~e x u ~ 6 1 J . C X ~ o v cXvcXrxllL. L 'Yap xext v o u 9 e . ~ e . r v xcxt ~ ( ; ) ~ vou-

9 E . ~ e . r a 9 a ~ -c1}v x a ~ c X ~ 6 ] l J . c x [ t]ov veXIXT)v 1tpOc.TtL[9e.11l 20eXE.t [ ~ 6 ] 9 ' e . e x u ~ w [ ~ ] u 1 t a p x o [ v ~ o ~ ] , I : t ~ ou[xJ ~ o [ ~ e . ]~ u v [ 1 ) ~ a ~ ~ a u ~ ] O [ a u ] V ~ i v e x L [. ]aiA[. :

~ e . L [ v ]1tOVE.[ (LACUNA OF A FEW WORDS)~ E . ~ ] c p 6 l J . e . v o ~ 1} l 1 t a ~ v w v eXAA e l] I:ttv ~ O U ~ o 1tpa-r toL,

[ c x u ~ o ] tpy[o]v v e.i 11 [ ~ ] e x ~ a [ A ~ ( ] 1 t W V 0 lcp' ~ l J . W y 25c x ~ : c F > v [ ~ O ~ t r . J 't7jv t i i ~ c x l d c x ~ 1 t p 6 [ A 7 1 1 ~ L V , 06 [YlJ. cx ] lJ. e. e. 9e. [ ]Illv 0 [ ] lJ.1l 1tp [

: ]1t[ .. ]o[ (LACUNA APPROX. 45 WORDS)]~ o ~ [ C X U ~ ] T ) ~ 1 t A e X ~ T ) ~ . 1 t e . P L X c X [ ~ W ] I e ~ O ~ O U ~ o ~ A 6 r o ~ ~ p t 1 t E , -~ a ~ , xat o u O i 1 t O ~ e . o u v c x ~ e x L ~ e . ~ a ~ w a a L c.Tttv ~ O L e x U ~ C X 1 t c X ~ a 30OLcx x c x ~ ' cXvcXrxllv XCtAOUIle.VCX llcX 1 J . C X x . e . ~ c x ( ~ ~ v ~ 1te.et cxirtou

~ o u ~ o u E . c x U ~ O U ~ e . A ~ e . p e . U O ~ v w ~ . x v a1te.Lpov 1tcXALV, , , - , 6 .1 -exvexYXT)V ~ u ~ 1 t p a ' t t e . ~ v CX1tO A leuv CXe.L oux ~ 1 t L A O r ~ -ven on this occasion for failing to supply an exhaustive apparatus criticus or adequatelinguistic commentary (e.g. for leXv indicative, at 43 . on which see art . c it . note13 above, p. 69 .).

I should note that, in addit ion to the columns printed by DI NO my text incorporates three portions of P 1191 missing from his edition.


6/22

20 DAVID SEDLEY EPICURUS REFUTATION O F D ET E RM IN I SM 21lv 'tWL E


7/22

22 DAVID SEDLEY EPICURUS REFUTATION OF DETERMINISMbility. A nd i n t hat h e w ould at o ne p oi nt b e altering his t heory,[at another ] such error. For this sort o f account isself-refuting, and can never prove that everything is o f th e ki ndcalled necessitated ; bu t he debates this very question on the assu mp tio n t hat his o pp on en t is himself responsible for talkingn o ~ s e n s e ~ v e ? if he goes on to infinity saying th at thisactIon of hIs IS In t ur n necessitated, always appealing to

a r g ~ m e n t ~ he IS ~ o t reasoning it empirically so long as he goeso n ImputIng t o hImself t he resp ons ibi lity for having reason edcorrectly and to his opponent that for having reasoned incorrectly. Bu t unless he were t o st op att ri bu ti ng his action t o hi mselfand to pin it on necessity, he would not even [be consistent.....:.. On the other hand,] if in using the word necessity ofthat w h l ~ h we call our own agency he is merely changing a name,and won t prove t ha t we have a pre conception of a k ind whichhas f a ~ t y d e I i n e a t i o ~ s when we call own agency responsible, neither hIS ow n [behavIour n or t hat o f o th ers will b e affected . .. .. .. .:.] b ut even t o call n eces si tati on emp ty as a result of yourclaIm. If someone won t ex plain th is, and has no auxiliary element or. impulse in us . .dissuade from actions which we per

f ? r ~ callIng the responsIbilIty for them o ur own agency , but isgIv.Ing t he name o f foolish necessity to all t he things whi ch wec l a I ~ to do calling th e .responsibility for them our own agency,he will be .mere.ly changmg.a name; he will not be modifying anyof our actIons In t he way In which in some cases the man whosees w hat s ort o f act ions are n eces si tated regularly dissuadesthose who desire to do something i n t he face o f force. A nd th eintellect will be inquisitive to learn what sort of action it shouldthen consider that one to be which we perform in some way

t h r o u g ~ o ur own agency b ut w it ho ut desiring to. Fo r he has noalternatIve but to say what sort of action is necessitated [and whatis not : } supr.emely unintelligible. But unless someone per-v e r ~ e I y m a l ? t a I n ~ t?lS, or makes it clear what fact he is rebuttingor I n t r o d u c ~ n g It IS merely a name t ha t is being changed, as Ikeep repeatIng.. 59-69) T he first m en to give a satisfactory account o f causes,men not only much greater that their predecessors but also, many

times o ver, t han t heir successors, t urned a b lin d eye to themselves although in many matters they had alleviated great ills) inorder to hold necessity and accident responsible for everything.Indeed, th e actual account promoting this view came to griefwhen i t le ft t he gr eat man bli nd to t he fact t ha t in his actions hewas clashing w ith his doctrine; and t ha t i f i t were no t that acertain blindness to the doctrine took hold of him while acting hewould be constantly perplexing himself; and that wherever thedoctrine prevailed he would be falling into desperate calamities,while wherever it did not he would be filled with conflict becauseof the contradiction between his actions and his doctrine. 59-71) It is because this is so that the need also arises to explainthe matter which I was discussing when I first embarked on this digression, lest some similar evil [befall us ... .

I have s ta rt ed my quo tation of t he t ex t at what I ta ke t o bea point shortly before the digression begins. Epicurus is arguing1-8 for our ability to shape ou r lives. A t b ir th we have a widerange of alternative potentials seeds )1 for character development; therefore the way we do in fact develop is dependent onus. T he p oin t is no do ub t t ha t our futu re development is n otalready built into our initial make-up, so that nothing pre,vents usfrom determining it ourselves. We are able, particularly by thebeliefs which we form, t o cont rol t he impressions t hat o ur immediate surroundings make on US: 16 it is not the surroundingsthat control us.

15 It may be suggested that the cmtPJL[cx tcx a new re ad ing o f t he papyri) a reatoms. I do not myself believe that prior to Lucretius poetic use of semina anyEpicurean or other Greek philosopher including Anaxagoras: see e.g. M. SCHOFIELD, An essay on Anaxagoras Cambridge 1980), p. 121 ff.) can be s ho wn t o us et he w or d i n t he s ens e C elements or C elementary particles . Epicurus use of it atEp. Pyth. 89 carries some biological connotation, as his use o f i 1 t c x p 8 t u < 7 ~ in connexion with it shows: he must mean either literally C seeds or at any rate complexseedlike particles of some sort: BAILEY, The Greek atomists and Epicurus Oxford1928, p. 343 f.; A A loN Chance and natura law in Epicureanism Phronesis 22,1977, pp. 63-87, e sp . 77 . . The one possible exception is Plato, Tim. 56 b.

16 For the technical Epicurean explanation of our power to contr ol our intakeof simulacra see Lucretius IV 777 f. on th e correct ord er of t he lines, see E.AsMIs, Lucretius explanation moving dream figures at 4.768-76, AJPh 102,1981, pp. 138-45).


8/22

24 DAVID SEDLEY EPICURUS REFUTATION OF DETERMINISM 25It is clear that the deterministic h llenge is already in Epicurus sights, and presumably the determinist s countermovewas en visaged i n t he lines lost b et ween 8 and 15, since Epicurus is replying to it when the text resumes. The determinist wasp erhaps made t o obj ect t hat we in t he e nd develop in only onedirection, and th at there must be some cause, either presentfrom t he o ut set o r imp osed by o ur env iron ment , t o s teer us thatway. Epicurus reply 16-19) is t ha t every time we t rea t othe rsas answerable for th ei r behaviour we imply t ha t they themselves, and not t heir congenital make-up o r t he con st rain t o ftheir environment, are responsible. It is an implicit premiss ofthis t hat we wo uld n ot blame o th ers for any characteris ti c already i n ~ r a d i c a b l y b ui lt i nt o t heir cong eni tal make-up,17 andthis is in fact a point brought out slightly earlier in the text,w here E pi cu ru s ex plo its a con tras t b et ween moral agents andwild animals. I8 The reason we do n ot rep ro ve wild animals for

their behaviour, he says, is that we draw no distinction betweentheir congenital make-up and their subsequent developments _meaning, presumably, that their entire future character is det ermi ned at b irth and t herefo re b ey on d t heir cont ro l.T he appeal t o prais e and blame as conflicting with determinism is a familiar one,19 and Epicurus is ready for the determi-17 C f. Arist., EN 1 11 4 a 2 3- 5.18 l-rt tVLO-r[t X ] ~ X ~ O ~ t v , tv vou9tnr:[LX]wL ~ i V - r O L -rP01tW(L], X XL oux

WO 7ttp [ r]Ot [ r1ptcx -rwv ~ w t w v X C X 9 ] C X L P O ~ t v ~ E . V o J J . o w ~ exu rOt rOt cX1tOrtrt[vv] r 1 ~ i v c x[x]cxi til[v] aUcrtexaw lv -rt a u J J . 1 t [ A i ] x o v - r t ~ ou JJ llv o [ ~ ~ t r6>t vou9t[-r]1)-r[L]xWt -rP01twtX i i m M ~ ~ L oUu rWL [v-nfttotbil ttXi;)t XPWJJ t9cx [. ..] [34.25] 21-34AIm. = DIANo p. 38 f. . . .We sometimes vilify it [sc. a self-determining animal] all t he more, b ut more in anadmonitory way - and not in the way in which we exonerate those animals whichare wild by c onf la tin g t he ir de vd op me nt s a nd t he ir make-up alike into a singlething, and indeed do not use e it he r t he a dm oni to ry a nd re forma tory mode o r t hesimply retaliatory mode. Three passages in particular have blocked the recognition of -r

as devdopments . One is [34.17] 1-3 AM.; b ut there I h ave foun d that thetermination wv in line 3 has b ee n c orr ec te d by t he scribe to ex The ot he r twoare i id 2 0.11 -1 2 and 2 4.4-5, o n which see ARJuGHETTI s commentary, p. 269; b ut Id o n ot b el ie ve that w e a re o bl ig ed t o c on st ru e t he g en it iv es t he re as ARRIGHETTIdoes. Fuller discussion of t hi s m us t a wa it my p la nn ed e di ti on of th e b oo k.

19 Cf. Ep. Men. 133; Diogenes of Oen oan da 3 2 .3 .9 ff.

nist s reply: yes, of course we re buke miscreants, b ut that isbecause the rebuking process is itself necessitated (19-21).20 Thebeginning of Epicurus next move is lost in the ensuing lacuna,and t he fragment at 24-28 is t oo s ho rt t o give us much help. Itappears t o b e re-applying t he earlier p oi nt 16-19) t hat t o issu epraise or blame is t o imply t hat t he reci pient is responsible forhis behaviour. 21 I suspect that he is already here embarkingupon a self-refutation argument: the determinist cannot indulgein philosophical debate without apportioning the credit betweenhimself and his o pp on en t, and i n doing so h e will b e adop tin gprecisely the kind of critical attitude which leads us to conceiveof th e self as responsible.A t any rate, from 29 the self-refutation tactic becomeexplicit. This is the earliest survivor in a long line of Hellenistic1tEPL tp01t f} arguments. T he ir charac ter has b ee n explored i n apioneering study by Myles Burnyeat,22 to which my discussion isdeeply indebted. The starting point 29-32) is that the determini st refut es hi mself b y arguing his case, since t o do so presupposes w ha t h e seeks t o deny, t ha t his o ppone nt is an autonomous agent responsible for his own foolish views. The determinist of course has an easy retort to this - that his own actionin arguing with his opponent is itself causally necessitated. ButEpicurus is ready for him 32-35). He envisages an infinite regress in which. the determinist, challenged once again for continuing t o argue, goes o n to explain away his new ret ort as beingitself causally necessitated; and so too with th t retort, etc. Epicurus p oi nt is n ot t hat t here is anything vicious about the regress - indeed, there is not - but rather that the determinist

20 Alternatively, the determinist may be arguing that he blames people in order to necessitate certain changes in their behaviour. The drawback of that interpretation is that accidental necessity would be a rather inappropriate expression.C f. also R. SORABJI Necessity cause and blame L o nd on 1 98 0) , p. 8 7.

21 The fragment, omitted a lt og et he r f ro m p re vi ou s e di ti on s, c on ta in s a n ewinstance of 1 t p 6 A 1 ) ~ t . ~ , on whose use as a criterio n see b elow. c x l t ~ must, I su pp ose,be closely linked in sense with lcp cxu t(;)v our own responsibility).

22 M.F. BURNYEAT, Protagoras and seii re/utation in later Greek philosophy Philoso ph ical R ev iew 75 1 97 6), p p. 4 4-69 .


9/22

26 DAVID SEDLEY EPICURUS R EFUTATI ON OF DE TE R MI NISM 27

c ann ot save himself b y resort ing t o i t, since at v y step it iscontradicted by his behaviour: at every step he is both assertinguniversal necessitation and apportioning the philosophical creditbetween himself and his opponent, as i f b ot h were agents capable of making u p t he ir ow n minds. H en ce he fails t o use empirical reasoning ( l 1 t t A O i L a t J . 6 ~ , 33 f. ;23 his account never finds empirical confirmation i n his own practical attitudes. T he finalsentence of this section 35-37) seems t o b e embarking on t heargument t ha t if t he d et er mi ni st is to restore consistency hemust eventually ha lt t he regress and cease to claim the creditfor reasoning correctly. And once he does that it is indeed hardt o see w ha t grounds h e could have left for believing his thesi sto be correct.The opening challenge, 1 t E P t X c X ~ W A 6 i o ~ ~ P 1 t E ~ ( X t , ispeculiar Epicurean jargon for what other philosophers came toexpress as 1 t E P t ~ P 1 t E ~ ( x L . Determinism refutes itself . As Burnyeathas beautifully demonstrated, there is a close parallel at the startof Lu cre ti us ant i-sce pti c argument , wher e t he sceptic whoclaims that nothing is known is a man who IV 472) has stoodwith his own h ea d w he re his fe et b el ong - clearly Lucr et iusimaginative rendering of t E P ~ X c X ~ W ~ P 1 t E ~ ( x ~ . 24 Indeed, turnasomersault could conceivably be the metaphor underlying Epicurus choice of term. In both passages it is a pragmatic self-refutation that is envisaged.25 There is no internal logical inconsis-23 F or t he me an in g of t hi s t er m, see my Epicurus On nature Book XXVIII

CE rc 3 1973), p p. 5 -8 3, pp . 27 -3 4.24 M.F. BURNYEAT, The upside down back to front sceptic of Lucretius 472Philologus 122 1978), pp. 197-206.25 I once felt that f or this r eason the anti- determinist passage would wor k better the self-refuting AOjOt; w er e n ot j us t a t he si s, as BURNYEAT takes i t, b ut a n

argument. Argume nts i n He ll en is tic philosophy do no t just exist, they Careasked , so that an argument which could not be argued - dialectically maintained- would in dee d b e a c on tr ad ic ti on i n terms . But a t SV 40 our passage is pithilysummarised as follows: /) A{jWV 1tav-tcx XCX t cXvajx1}v jtvta6cxt ou8v tjXCXAttv lXtt tiiALroV tt 1J.1 1tcXV tCX XCX t vcXYX1}v jtvtaacxt. cxu to r ~ Lucretius notiti and word meaning.The sceptic s ability to use the words k now and no t knowsignificandy presupposes that he possesses the preconceptions of

true and false , doubtful and cert ain , since t nl e and,certain feature in the definition of know . This conflicts withhis claim to have had no e x p e r i e n c ~ of truth and certainty, forwithout that experience he could not have acquired t he pre-conceptions.The On nature argument also appeals to preconception andword meaning 38-41), but on different grounds. One is reminded of t he Epicurean dictum t ha t some inquiries are aboutthings while others are about mere words.26 Since the determinist refuses to behave in accordance with his doctrine, Epicurosn ow fe els entided to conclude that he is not making any claimabout the world b ut is merely renaming as necessity the verysame thing which we normally call our own agency . This isthen expressed in terms of preconception. Our p r e c o n c e ~ t i . o n ofour own agency is of something with its own causal efficacy.The determinist, to make his claim a substantive one, wouldhave to show that the preconception is a faulty one.

This is an excellent new example of the Epicurean appealto preconception as a criterion. Underlying our use of eachword is a preconception, embodying a generic delineation (-c6-

26 D.L. X 34.


10/22

28 DA VID SEDLEY EPICURUS REFUTATION OF DETERMINISM 29 t O ~ .D.L. X 33) o f th e type of thing named, and, being the cumulative p ro du ct o f exp erience o f things o f t hat k in d, a primesource o f empirical inf or ma tio n a bout them. If, t hen , th e determinist is to succeed in altering our preconception of our ownagency, he is obliged t o show us precisely how its delineationhas faile.d to convey the facts. Elsewhere Ep. Men. 123-4) Epicurus hImself undertakes a similar task in arguing tha t t healleged p reconception of the gods as concerned with human affairs is defective, and this requires a) showing that it conflictswith s o ~ e more fundamental or secure preconception, that of thegods as blessed Ep. Hdt. 76-7; Plutarch, Stoic. repugn. 1051 e,

~ o m m not. 1075.e), and b) explaining how the faulty conceptIon arose LucretIus V 1183 ff.). No d ou bt t he p oi nt o f Epicurus remark in 39-41 is that the determinist has no comparablearguments at his disposal.In t he ne xt stage 43-51 , I take i t t ha t the auxiliary element or impulse in us 44 is th e cause within us 17

~ h i c h E p i c u r ~ s .uph?lds but the determinist denies. By denyingIt, the determInIst gIves up the hope of ever dissuading us fromany of our actions, and so once again his argument is about merewords, without any hold on reality. In this he is contrasted withsomeone who makes t he correct d is ti ncti on b et ween necessit at ed and u nn ecess it ated actions, and wh o con sequent ly candissuade us from resisting compulsion. An example may makesome sense o f this. It is necessary that everybody should die,and someone who understands this can dissuade an old manfrom v ~ y seeking immortality. But he can only do this becausehe appreciates t ha t c er ta in o th er things are up t o us. If he attributed everything to necessity, he could not consistently eventr y t o dissuade t he old man, since he would have to regard hisdesire for immortality as itself necessitated and not up to him atall.

In this argument Epicurus underestimates the resilience oft he d et ermi ni st , wh o mi gh t regard his fellow men as automataand still, without inconsistency, try to modify their behaviourregarding his own exhortation as just a noth er in t he set of cau:sal fact ors t hat will d et ermi ne i t. B ut the next point is a rat her

more penetrating one 51-54 . The d e t e ~ m i n i s t s blanket application of necessity to all kinds o f event leaves him un ab le t ocharacterise those mixed actions which we reluctantly chooseto p erfo rm i n avoidance of some g reat er evil. These, heto mean, can only be understood by someone able to distinguishthe voluntary from the compulsory elements in them. Here, asin t he p receding lines, we can glimpse Epicurus reasons forstressing in the etter to Menoeceus 133-5 t he ethical value o fproperly grasping the differences between necessity, chance, that which depends on us .A ft er a b ri ef rei terati on of t he second sect io n s main p oi nt56-59 ,27 the third section opens 59-63 with a most interestingcomment on the founders of atomism, Leucippus and DemocrituS. 28 They are revered as the greatest, exponents of C X l ~ t O A O r C X 29but are nevertheless criticised for their thesis of universal

necessitation, which they were only able to maintain by turninga blind eye to themselves. This last phrase implies a recognitionon Epicurus pa rt of the value of introspection as a counterweight to determinism. It is perhaps comparatively easy to27 tt i n l in e 57 looks indef ensible, and should probably be emended to which has hitherto been reported as P s actual reading). In line 58, w he re I r ea d

~ U v f } Marcello GIGANTE who presents a section of this text at p. 8 . ofhis important new book Scetticismo e epicureismo Napoli 1981), reads~ U v [ i U t and translates risponde solo a parole. Both restorations, according to mynotes on the papyri, are palaeographically possible although I would print cpv[iOt ,but my own choice is governed by Epicurus comment that this is a po in t h e hasa lr ea dy mad e mo re t ha n o nc e - an a llusion to line s 39 and 48.

28 There is a longstanding debate as to the correct construal of 61-63; see nowA. LAKs, Une eg ete Dhnocrite CEre 11 1981), pp. 19-23. In my translation Ifollow the e n t i r ~ l y convincing interpretation of G. ARRIGHETII, Un passo dell operaSulla Natura di Epicuro Democrito Colote CE rc 9 1979), pp. 5-10. F or t heabs olute use o f AClv9avtw tClu t6v he cites Pl ato, Philebus 19 c. My only alt er at io ns t o t he c ur re nt rea dings o f t he t ex t are 1tPO ClU tW[V] for npo ti]p[v] 60)and t [ l ] ~ for lv Iv 101 . The latter renders GIGANTE S very ingenious handling ofthe text op. c it ., p. 56 ff.) unnecessary; a final choice between the two readingscannot, in the present state of the papyr us, be made on palaeographical groundsalone, b ut I h ope t ha t my version will commend itseH on linguistic grounds too.

29 Cf. Democritus remark 68 B 118 DIELS-KRANZ) that he would r ather f indone Cll ttOAOr a than win the kingdom of the Per sians.


11/22

30 DAVID SEDLEY EPICURUS REFUTATION OF DETERMINISM 31think of others as automata, but our power to control our nextaction and to falsify predictions about our behaviour makes ithard to see how determinism can be true of ourselves. 30Leucippus and Democritus are certainly not spared altogether. But neither do they bear the full brunt of the attack. Themain target down to 59 has been an unspecified deterministwho consciously places all human behaviour, his own includedwithin the scope of his theory. Leucippus and Democri tusmentioned in primarily historical terms as great forerunners whounfortunately failed to notice their deterministic theory s implications for their own conduct.

As the argument proceeds 63-69), the spotlight falls moredirectly onto Democritus, for must be he who is meant by t OVv8pIX 64. There is a well established Epicurean use of cXV P(or v r the sense gre at man . Epicurus uses of his philosophical predecessors,31 though not of contemporary opponents,and later Epicureans extended its application to the founders ofthe Epicurean school. 32 Democritus appears as Democritus virboth times that his views are cited by Lucretius. 33 So the singular ;OV vopa appears to indicate that he now has Democritusprincipally in mind (after all he is supposed to have doubtedLeucippus existence).

In what follows, the practical consequences of seriouslymaintaining the determinist doctrine are spelt out: a clash between theory and practice, leading to continual perplexity. Onceagain there is a close parallel in Lucretius IV 500-21 wheret?e disastrous practical consequences of scepticism for e ~ e r y d a ylife are spelled out . Similarly here, obedience to the doctrinewould have led Democritus into one calamity after another (nodoubt because he would have given up making decisions and let

JO Cf. G . RYLE The concept mind (1949), p. 196 fH [26.44J 22; [29.18] 1 (?), [26] 14 ARR.32 See Francesca L o N GO A U RI C CH IO La scuola di picuro CEre 8 (1978),pp. 21-37.H Lucr. III 371; V 622. Cf. A.E. HOUSMAN Classical papers (ed. ]. DIGGLEand F.R.D. GOODYEAR, Cambridge 1972), III , p. 906.

I

necessity have its way), while disobedience to it would have lefthim bewildered by his inability to match the deed to the word63-69).In Book XXVII of the On nature Epicurus had outlined his

ideas for a brand of empirical reasoning l 1 t L A O r L a I J . 6 ~ notentirely unlike some modern pragmatist theories of t rut h,whereby the truth or falsity of a doctrine was to be establishedby examining its practical consequences. Our passage could beseen as an application of this method, with the peculiarity thatthe consequences listed are purely hypothetical. Democritus, ashe stresses with two uses of Aav9cX JtL J and one of A1)9T}, was unaware of the conflict because he conveniently forgot the doctrine when i t came to acting.W ear e now ready to turn to the all-important quest ion:who is Epicurus target? I have spoken throughout of the,determinist opponent, who is said to hold that all things arenecessitated. But that might, on the face of i t, be interpreted asa physical, a logical, a psychological, or even a theological thesis,or as some combination of these. Indeed, the determinismwhich Chrysippus later developed on behalf of the Stoicsamalgamated all four types. Hence there has been a frequenttendency to see the early Stoics as Epicurus determinist target.But I think this must be mistaken. Zeno and Cleanthes werecertainly enthusiastic part isans of destiny (tLIlCXPIJ. VT)), butthere is no evidence that to them this was ever anything morethan the popular view of destiny as a divine force predetermining the eventual outcome of every story, while leaving it up tothe individual whether to resist or submit. They viewed thescheme as a providential one which supplied the proper contextfor man s exercise of his moral faculties. And Cleanthes, a tleast, was a declared opponent of the Master Argument, whichwas being used in his day as a formal proof of universal necessitation.34 Even allowing for the possibility of misrepresentation

}4 SV 1.489. Zeno s simile of man as the dog t ied to a car t who must followwilly-nilly SV 2.975), later adopted by Chrysippus, implies the view of destinyas determining outcomes alone. So does Cleanthes, SV 1.527. It was ChrySiPPU5


12/22

32 DAVID SEDLEY EPICURUS REFUTATION OF DETERMINISM 33on E pi cu ru s p art, t here seems no way t hat h e could have conjured up out of th e Stoa his ha rd determinist foe who is readyto attr ibu te everything, his own states of mind included, toaccidental or foolish necessity. The thesis envisaged isclearly one based on physical causation;35 and when Epicurus attacks destiny in the Letter to Menoeceus 134 he specifies it asthe d octrin e n ot o f t he Stoics b ut o f t he n at ural p hi lo so ph ers .T,he so rt o f d et ermi ni sm b es t s ui ted to b e his t arget is thekind thesis sometimes called mechanism , the entirely serious view that human behaviour can, like everything else, be exhaustively accounted for in terms of material changes chemicaland neuro physiolog ical changes i n t he mo dern th eo ry , atomi cchanges in the ancient), and tha t talk of intention desire ,etc. is superfluous, with no additional descriptive or explanatoryforce. Norman Malcolm has criticised such theories as self-refuting w it h con si derati on s very like th os e advanced by Epi cu rus, for example that a mechanist would have to deny tha t hehad any rati on al grou nd s for asserting anyt hin g, including hisown th eory. Su ch a d octrin e seems to b e exactly w hat Ep icurushimself is grappling with, especially in his own self-refutationargument, in f. where he takes i t for gr an te d t ha t the oppon en t has d en ied h im self t he rig ht to appeal t o any pu rp osi ve element w it hi n us, and i n his disparaging talk o f accid en tal andfoo li sh necessi ty 18-20, 47 f., cf. 62 f.).

Who could b e t he p erpetrat or o f th is mechani stic th eory?We h ave seen D em ocri tu s exo nerated fro m t he main b urden o fguilt. Readin g b et ween t he lines, one m ig ht t ak e t he real oppo-who at some stage added to this picture a theor y of inter nal causal deter minationo f a tt it ud es and ev en h e may have h es it at ed to call this de sti ny - Cicero, Fat41-3). Ancient doxography makes Chrysippus the author of Stoic determinism, andas far as I c an t el l i t is r ight. Cf. the mater ial collected by R.W. SHARPLES, Neces-sity in the Stoic doctrine of fate so 56 1981), pp. 81- 97.

}, An explicit allusion to the self-refutation argument, in an unnumbered bookof On nature [35.11] 1-5 ARR. , has the opponents making everything be caused bythe preceding motion. N. MALCOLM, The conceivability of mechanism Philosophical Review 77 1968), pp. 4.5-72.

n ent s t o b e t he l at er followers w it h whom t he first atomists arefavourably contrasted in 59-61.7 The same possibility seemsimplicit in some remarks of Diogenes of Oenoanda, who attacksnot Democritus himself but those who adopt Democritus thesiso f u ni ve rs al n ec es si ta ti on u nm od if ie d by t he E pi cu re answerve.8In con fi rm at io n of th is, we can ret urn t o t he close and apparently conscious parallelism between Epicurus treatments ofdeterminism and scepticism. The sceptics refuted in LucretiusIV must be, o r prominently include, those fourth- centuryDetilocriteans like Metrodorus of Chios, Anaxarchus, and evenEpicurus own reviled teacher Nausiphanes, who had played upt he sceptical side o f D em ocri tu s t ho ug ht, and against w ho m

Epicurus was eage r to marshall th e positive empiricistarguments which Democritus had also bequeathed. 9 This scepticism was the result of what I shall call reductionist atomism.Because phenomenal objects and properties seemed to reduce tomere configurations of atoms and.void, Democritus was inclinedto suppose that the atoms and void were real while the phenomenal objects a nd propertie s were no more t ha n ar bitr aryconstructions placed upon them by human cognitive organs. 40 Inhis more extreme moods Democritus was even inclined to doubtth e power o f h uma n judgment, since judgment was it self n omore t han a reali gnm en t o f atoms i n t he min d E . 1 t L p U G ~ l l . 417 T he c om me nt h as a ve ry close pa ra lle l a t L uc r. I 734 ff. , wher e t he g re atEmpedocles is contrasted with h i s inferior pluralist successors. The parallel con

firms that the inferior successors of Leucippus and Democritus are their philosophical heirs, not just later philosopher s in gener al. 8 Diogenes of Oenoanda 32.2.3 - 31.9. I mean tha t Diogenes Epicurean sourcemay we ll h av e b ee n a tt ac ki ng s pe ci fic D em oc ri te an ph il os op he rs , not thatDiogenes himself is consciously doing so. FURLEY, Two studies see note 1 above),p. 174 f ., has alr eady suggested that Nausiphanes r ather than Democritus may beEpicurus target.39 Especially Democritus 68 B 125 DIELs-KRANZ.40 Id./ap. Plut., dv Col 1110 e cf. 68 B 9, 125 DIELS-KRANZ .41 Id. 68 B 7; cf. B 9. T he argument is essentially that of J.B.S. HALDANE,Possible worlds London 1927) , p. 209: If my mental processes are deter mined

wholly by the motion of atoms in my b ra in , I h av e no r ea so n t o suppose that mybeliefs are true ... and hence I have no reason for supposing my brain to be composed

. J l

F


13/22

34 DAVID SEDLEY EPI URUS R EFUTATI ON OF DE TE R MI NISM 35

Epicurus response to this is perhaps the least appreciatedaspect of his thought. It was to reject reductionist atomism. Almost uniquely among Greek philosophers he arrived at what isnowadays the unreflective assumption of almost anyone with asmat te ri ng of science, t ha t t he re are t ru th s at t he microscopiclevel of elementary particles, and further very different truths att he phenome na l level; t ha t t he former m ust be capable o f explaining the latter; bu t t ha t nei ther level of description has amonopoly of t ruth. The truth t ha t sugar is sweet is notstraightforwardly reducible to t he t ru th t ha t i t has such andsuch a molecular structure, even though the latter truth may bere qui re d i n o rd er t o explai n t he former). By establishing thatcognitive scepticism, th e direct o ut co me o f r ed uc ti on is tatomism, is self-refuting and untenable in practice, Epicurusjustifies his non-reductionist alternative, according to which sensations are t ru e a nd t he re are t her ef or e bona fide t ru ths at t hephenomenal level accessible through them. The same will applyto t he 1ta rj which Epicurus also held to be veridical. Pleasure,for e x a m p l e ~ is a direct datum of experience. It is commonly assumed that Epicurus must have equated pleasure with such andsuch a kind of movement of soul atoms; b ut although he willhave taken it to have some xpl n tion at the atomic level, Iknow of no evi denc e t ha t he, any more t ha n most moral philosophers or psychologists, would have held t ha t an adequateanalysis of i t could be found at that level. Physics are strikinglyabsent from Epicurus ethical writings, and it is curious that interpreters are so much readier to import t hem the re t ha n theyare when it comes to the moral philosophy of Plato orAristotle. 42

of atoms . I think it was Bob Sharples who f ir st drew my attention to this passage .42 Put technically, the point is that pleasure is a a U ~ t t w ~ ~ of a per son s consti

tuent atoms Demetrius Lacon ap. Sext. Emp., Adv. Math. X 225-6), and that a u ~ t t ~ J L C l t ~ t he ms elv es e xi st o nl y a t t he p he no me na l le vel , n ot a t t he mic ros co picEpIC., Ep. Hdt. 70-1 . Cf. loN art. cit. note 15 above , for an objection tophysical interpretations of Ep. Men. 133 5.

The metaphysical status of phenomenal properties, states ofmind, e tc. is that of accidental properties of groups of atoms.43T ha t is, t he y c ann ot e xi st i nde pe nde nt ly of t he atoms. Butcommon assumption that t h e ~ are j ~ s t pa tt er ns of ~ t ~ i ~motion does not follow from this, and ruled out by Ep1curusepistemology. Sensations, which come out always true on his account, never report patterns of atomic motion, since atoms areimperceptibly small.Cognitive scepticism is not the only outcome of reduc-tionist atomism. Another is determinism. If human judgments,impulses and emotions just are the mechanical bouncing aroundof atomic billiard balls i n t he mind, t here seems no room forindividual self-determination or responsibility. T h i ~ is preciselythe mechanistic brand of determinism which we have alreadyseen reason to identify as Epicurus target. And his treatmentof i t, we have also seen, consciously parallels his t re at me nt ofscepticism. By showing it to be both self-refuting and u n t e n a b ~ ein practice, he justifies the level of description used e s e w h e ~ e mthe same book of On nature one which allows for a r at lona lautonomous self with its own judgments and impulses, none ofthese being straightforwardly reducible to motions of a person sconstituent atoms.This parallelism makes the conclusion irresistible that Epi-eurus, primary motivation was i n b ot h cases the need to rescuethe atomistic tradition with which he had aligned himself fromt he i nt er na l r ot of reductioni sm. I know of no di re ct evi dencefor mechanistic determinism in fourth-century Democriteans,but i t is only t oo easy t o see how, as reductionist atomists, theywould have arrived at it, and an argument from silence is hardlyadmissible when dealing with men whose voluminous writingshave l ef t scarcely a trace i n ou r sources. It is indeed possiblet ha t t he view was still mai nt ai ned by some o f Epi cur us owndisciples. For in one unfortunately fragmentary passage of theanti-determinist argument 4 . he can be seen using the

43 Lucr. I 449-58; Epic., Ep. Hdt. 68 73.


14/22

36 DAVID SEDLEY EPICURUS REFUTATION OF DETERMINISM 37

44 u l l t f ~ also occurs in the fragmentary finaI sentence of the book (not reportedin DIANa s text).Nat. xxvrll fro 13 col. XIII 7-8 sup. in ed. cited note 23 above).

, 46 op. c ~ t (note 23 .above), p. 43, 45 f., 48, 56. For other cases of Epicurus pupils steppmg out of l ine, see my EpicufUS and the mathematicians of CyzicusCEre 6 (1976), pp. 23-54, pp. 2 7 f. 46.47 = [34.21-2] AIm. ]aro9cxL IJ.


15/22


That is why those who cannot correctly make such distinctions confuse themselves about the adjudication of responsibilities .This passage seems to be using Epicurus contrast betweenwild animals, whose misbehaviour we hate but do not blameand self-determining animals including humans), whose misbe:haviour we both h at e and blame. 49 The l at te r s failures arecaused by themselves - selves which are t identical withtheir constituent atoms 2_4 .50 Or, as he puts it in the nextsentence, the nature of their atoms does not contribute to certainof their actions and dispositions: rather, these are caused mainlyby certain characteristics which they develop 7-10 . Atomicmake-up may be responsible for disorderly motion in their mindatoms 10-11 ,51 but i t does not follow t ha t they cannot makedecisions which override those motions. One is reminded hereLucretius insistence that even though the atomic compositionof the mind at birth determines an animal s natural temperament, nothing prevents our learning to overcome that temperament.52 E ve n t he na tur al coward, Epi curus would say, c anuse his rationality to learn courage.Shortly after the lacuna, the relationship of developmentsto the self s responsibility is amplified. Implicitly, at birth we

r just atomic mechanisms. The self becomes responsible assoon . as the animal develops a certain type of characteristic49 Cf. text quote d i n note 18 above. It is n ot clear w ha t t he hallmark o f awild animal is, but it is a good be t t ha t the term covers t he same ones as arepicked o ut in RS XXXII as incapable of forming a social contract and hence beyond therealm of justice and injustice. The comparison with RS as lso that with Lucre-tius II 251-93 supports my assumption that ~ i i x are the unspecified neuter subject.Cf. also P.M. HUBY The picureans animals and freewill Apeiron 3 1969 pp.17-19.50 F or a no th er E pi cu re an a pp ea l t o t he n on -i de nt it y o f o ur co ns ci ou s s elves

with our constituent atoms, see Lucretius III 847-6l.n Glenn Most has persuaded me that it is more natur al to take l x E ] V 1 } ~ 10 sreferring back to c p u n ~ 8 t ha n to cxl dcxv 10 , as I originally thought. On the lat

ter reading 10-12 would be making almost the same point as 18-20).52 Lucr. III 288-322.

over and above his atomic make-up 16-19 . The distinction between the character is tic and th e atoms must be of a,transcendent literally separative ) kind, not just a differenceof scale. The point, I think, is that all things have certain phenomenal properties, for example colours, over and above theirconstituent atoms, but t ha t t ha t k ind of difference is primarilyone between microscopic and macroscopic analysis; whereas thesort of moral characteristics required in a responsible agent overand above his constituent atoms are ones which differ from theunderlying atomic configurations in a much more radical way.No doubt the nature of this difference was more fully discussedi n t he earlier, l ost p ar t of t he book. But i t seems clear that whatis envisaged has much in common wi th t he modern n ot io n o f,emergence . In Epicurus view matter in certain complex statescan take on non-physical properties which in turn bring genuinely new behavioural laws into operation.

Once the agent has achieved responsibility, he adds 19 f.),this does then leave it s mark up on his c onsti tu ent atoms, a ndbecomes a yardstick {the reading is very conjectural, however}.This, I suppose, would be a pattern of atomic motion regulatinghis future behaviour. When the natural coward learns throughrational reflexion to adopt brave attitudes, this will in turn havea stabilizing influence on his disorderly soul atoms, and he mayno longer even suffer the physical sensations of fear.Many details remain obscure, but the non-reductionist approach is both plain and deliberate. The developed self is nott he playt hi ng of i ts c on st it ue nt atoms, b ut a power capable o fcontrolling them. And as he says 20-22 , without this distinctionbetween atomic and psychological levels of truth the issue of responsibility. becomes hopelessly confused.

It may be objected that the self and its mental states, beingsecondary properties parasitic on configurations of atoms, couldnot possibly control the motion of those configurations. Epicurus might offer two answers. The first is t ha t somehow histhesis must come o ut t rue , because its negat ion is mechanisticdeterminism, which is self-refuting. The other is that there is noreason in principle why secondary properties of configurations


16/22

40 DAVID SEDLEY EPICURUS REFUTATION OF DETERMINISM 41o f atoms shou ld n ot affect t he ir motions. Aft er all the densityof a sto ne is a mere secondary property of its constituent atoms,yet is capable of det er mining t ha t th e st one will move downwhen it is dropped in water, where a leaf or a piece of woodwould. remai n a t t he surface. Alth ou gh t he m ot ion o f a single

IS fully accountable for in terms of primary physical laws, t seems p erfectl y co rrect t o invoke seco ndary p rop erti es ascauses of the corporate behaviour of atomic compounds.

The Swerve AgainThe swerve is not mentioned in the surviving fragments ofthe n n ture book. And, given th e above un der stand ing o fEpicuru s purpose in t ha t book, i t is easy to see why. Theswerve, established, could not so much as dent the armour of thered uct io ni st a to mi st. Eve n i f h e were persuaded to reduce human behaviour to the undetermined, instead of the determined,motions o f atoms, t he re would be no more room for an aut ono mous self over and above a person s constituent atoms th anthere had been on the fully deterministic account.Indeed, the standard reductionist interpretation is preciselywhat has always ma de i t h ard t o see how Epi cu rus e xp ect ed theswerve to help him. But once the self is recognised to b e forEpicurus, an independent causal factor, it is possible reexamine t he ev id en ce wit h new i nsight. According t o CiceroEpicurus introduced the swerve

be ca us e h e f ea re d t ha t, an atom was always c ar rie d along by n atu ral an dnecessary weig ht, n ot hin g wo uld b e up to us, since t he m in d w ould b e mov ed i nwhatever way t was c om pe ll ed by t he m ot io n o f a to ms Fa t. 23 .

On the old reductionist assumption this contrast between u sand the mind s atoms must have looked misconceived. But now i ~ r o can b e seen to be stating a thoroughly plausible explan at Io n o f t he swerve. The s to ne analogy o ffered above fails t odeal with one crucial feature of Epicurus doctrine, that the cau-

sal influence which emergent properties, like the self and its volitions, exert upon the atoms of which they are properties shouldbe o ne which c an t ran scen d t he purely physical laws of ato mi cmotion and bring about motions which these atoms would probably n ot h ave followed if l eft t o t he ir own devices. Ye t that iswhat Epicurus appears to hold when he drives a wedge betweencausation by t he self and causation by the atoms, and i nd eedany weaker interpretation of the doctrine will reduce psychological states to supervenient consequences of atomic motions andret urn Ep icurus t o t he clutches of mechanism. The d ifficu lt yabout seeing how the laws o f atomic mo tio n could b e so t ran scended arises from the supposition that they are all-embracinglaws sufficient in themselves to determine the outcome of yprocess, thus rendering extraneous explanatory factors powerlessto i nt erve ne a nd t hrea ten in g t o make us helpless spectators ofour bodies actions. What more natural, then, than that Epicurus should conclude that physical laws are not in themselves sufficient to determine outcomes? And the swerve is just t he mosteconomical realisation of that conclusion, a minimal factor ofabsolute chance,s3 introducing too little indeterminacy to upsett he o bserved regu larities o f n at ural processes, b ut eno ug h t op re ve nt t he m in d s atoms from moving along p re de termi ne dgrooves from which our volitions would be powerless todislodge them. t will not be so much the actual occurrence of swervesthat matters as the mere possibility of their occurrence. A freemoving atom at any given moment has several possible trajectories to follow - one being the continuation of its currenttrajectory, the others being immediately adjacent parallel trajectories. This might at first appear merely to widen the pre-

n I am fully p er su ad ed by LONG art. cit ., not e 15 above) t hat t he swerve hasno important macroscopic chance effects outside cosmogony and volitional acts.The association of it with chance in Philodemus, Sign 36.7-17, and possibly in Plutarch, SolI an 9 64 c, may simply b e a n o ve ri nt er pr et at io n o f t he f act that it is akind of chance.

54 For the mathematics of the swerve, see my article cited in note 46 above, p.25 and ENGLERT, Ope cit . chapt er 2.


17/22

42 DAVID SEDLEY EPICURUS REFUTATION OF DETERMINISM

determined groove within which the atom must move. Butall the way along each of the possible trajectories further possible trajectories will branch out , owing partly to the possibilityof further swerves and partly to the various possibilities for colli-sions, themselves par tly dependent on the unpredeterminedbehaviour of o ther atoms. The effect is an indefinite proliferation of possible trajectories which does not just widen but totally erases the determinist ic grooves, yet still leaves an overwhelming probability that the atom s actual trajectory betweenone collision and the next will be a close approximation to arectilinear one.

It might be wise to stop there , and to say that once the deterministic grooves have been obliterated our volitions are freeto redirect our mind atoms as they choose. But are they thenfree to exceed the rules for swerves, for example by diverting anatom at an angle of ninety degrees from its original path? Thereis no need, to suppose so. It is sufficient that a volition shouldobtain its leverage on a mind atom by determining whether theatom goes straight ahead or adopts one or other of the availableswerve trajectories. This may at first sight appear to be ru ledout by the t radi tion that swerves are uncaused. In fact, though,causelessness is expressly reported as a feature of the theory notstated by Epicurus bu t alleged by his opponents to be an unwelcome implication of i t. No doubt the implication has somet ru th in i t, for a t any rate in the swerve s cosmogonic function(on which these hostile sources usually concentrate) there can beabsolutely no cause of the atom s following this rather than thatof the available trajectories. But all that Epicurus needs to holdis that there is no physical cause of its choice between the possible trajectories. And volitions must be counted as non-physicalcauses,56 for they are never l is ted alongside blows, weight andswerves as causes of atomic motion by the sources on Epicurean

55 Cic., Fat 18, 22, 46-8; Plut. , n procr 1015 c. Only Cic., Fin 19makes causelessness part of Epicurus own formulation of the doctrine.

6 Cf [34.32] 21-5 ARR., where o ur own agency is imp li ci tl y a non physical cause.

physics. On this ~ n t e r p r e t t i o n then, Epicurus would hold thatthe laws of physics determine no more than the approximatetrajectory of an atom. At any given moment it has a number ofpossible paths to follow, and as far as physics is concerned thereis no cause of its following one rather than another. Normally,then , its choice of path will be random. But in the special caseof animate beings there are also non-physical causes of ~ o t o nvolitions, and these operate not by overriding the laws of physics but by choosing between the possibilities which the laws ofphysics leave open. At an extreme, an atom might in this waybe moved far from its original path by a long series of swerves,or a large number of atoms might be simultaneously diverted statistically improbable events according to the laws of physicsalone, bu t still intrinsically possible ones, which volition couldtherefore be held to bring about.There are enough modern parallels to show that, r ight orwrong, such theories are tempting to adopt. The basic idea thatphysical indeterminism somehow removes an obstacle to theexercise of free will has had many advocates,57 and has at timeslooked like acquiring the status of a safe position. For example:The bearing of this [physical indeterminism] on t he prob lem of freewill hasbeen widely and rather confusedly discussed, but perhaps ll that can be said withany certainty at present is that there can now be no objection on the part of physics to t he belief that human actions are not mechanically determined (M.B.Hesse, Science and the human imagination 1955, p. 78).

. More specifically, the v ery theory which I have attr ibuted toEpicurus was proposed by the physicist Sir Arthur Eddingtonwithin a ye ar of Heisenberg s publication of the UncertaintyPrinciple:

57 Notably A H COMPTON, The freedom of man (1931 lecture, pub . New Haven 1935); K. POPPER, Of clouds and clocks (St. Louis 1966); D. WIGGINS, Towardsa reasonable libertarianism in Ted HONDERICH (ed.), Essays on freedom of action(London and Boston 1973), pp. 31-62, esp. 52 f.

58 A.S. EDDINGTON, The nature of the physical world (Cambridge 1928), p. 311.


18/22

44 DA VID SEDLE Y EPICURUS R EFUT AT ION OF DET E RMI NI SM 45Let us now look more closely into the problem of how the mind gets a gripo n m at er ia l a to ms so that moyements of t he b od y a nd l imbs c an b e co nt ro ll ed b yi ts v ol it io n. I t hi nk we may n ow feel q ui te s at is fi ed that t he vol i ti on is genuine.The mat eri al i st vie w was that t h e m ot io ns w hi ch a pp ea r t o b e c au se d b y o ur vo lit io n a re r ea ll y r ef le x a ct io ns c on tr ol le d b y t h e m at er ia l p ro ce ss es i n t he b ra in , t h ea ct of wil l bei ng a n i ne ssenti al side phenome non occ urri ng simul ta ne ousl y wi th t hephysi ca l phenomena. But t hi s a ssumes that t he resul t of applying physical laws tot he brai n i fully determinate. It is meaningless to say that t he b eh av io ur o f a c on scious brain is precisely the same as that of a mec ha ni ca l brai n i t he behavi our ofa mechanical brain is left undetermined. If t he l aws o f p hy si cs a re n ot s tr ic tl y c au sal t he m os t t ha t c an b e sai d is that the behaviour of the conscious brain is o ne o fthe possible behaviours o f t he mec ha ni ca l bra in. Pre ci se ly so; a nd t he dec ision bet ween the possible behaviours is what we ca ll voli t ion.And the same idea was later independently formulated {thoughnot as his own view} by Bertrand Russell:59 An d since, according t o q ua nt um physics, t he re are no physical laws to det er mi ne w hi ch o f s ev er al p os si bl e t ra ns it io ns a g iv en a to m wi ll u nd er go , we mayi ma gi ne t hat , i n a brai n, t he c hoic e be twee n possi bl e t ransit i onsis de te rmi ne d by apsychological cause called volition

Th e inte rpretation also fits in comfortably with the remaining ancient evidence. It is higWy significant that Epicurusadopted, in addition to the swerve, a closely parallel thesiswith precisely t he o bject o f p revent in g fut ure outcomes b ei ngdetermined independently of our wishes. Cicero s De lato showsthat physical determinism was from t he s tart o f t he Hell eni st icage p ar tne re d by logical de ter minism, t he thesis t ha t f utu reevents are n ecess itated by its already b ei ng t ru e t hat th ey willcome about. This b ra nd o f de ter minism h ad first surfaced i nAristotle s celebrated sea-battle argument at De interpretatione 9D io do ru s C ro nu s M as ter A rg ument , whi ch became t he chi efvehicle of logical determinism in the Hellenistic age 60 perhaps9 B RUSSELL uman Knowledge N Y 1948), p. 41.6 So I maint ai n i n Diodorus Cronus and Hellenistic Philosophy Proceedingsof t he Cambri dge Phi lologi ca l Soc ie ty, 23 1977), pp. 74-120, 96 ff. Diodorus was

proba bl y not himse lf a de te rmi nist , bu t his Mast er Argume nt wa s regul arly use d asa d ef en ce o f d et er mi ni sm . I h av e n ow g iv en u p t he s ug ge st io n m ad e i n that articlethat Diodorean arguments prompted the attack in the On nature hook: even i logi-

originated as an att em pt t o iso late and d efen d its several premisses. Now Epicurus, we are told, adopted Aristotle s solution ofdenying the principle of bivalence with regard to future-tensedstatements. No one I imagine will suggest that Epicurus, anymore t han t he many o th er ph ilosophers from A ris to tle o n whohave felt obliged to resist arguments of the sea-battle type,thought that his chosen logical thesis would have a positive partt o play i n t he analysis o f auton om ous volitio n. H is obj ect wasmerely to remove an obstacle to the possibility of self-determination, namely a thesis which. would render deliberation aboutfuture action null and void. In Cicero s De lato the swerve theory is regularly treated as parallel to, i nd eed as virtually int erch ang eable w ith , t he logical th esis, and t he same negativ epurpose, the avoidance of necessitation, is specified for both. 61An equally negative description of the swerve s purpose is offered b y P lu tarch62 to free the voluntary from everlastingmotion, so as not to leave vice irreprehensible, or so that thatwhich is up to us sho uld n ot b e eli min ated. A nd I kno w o f noo th er rep ort t hat implies a more i nt eg ral role for t he swerve i nthe psychological analysis of volition. 63All t his evid en ce s upp orts t he following story. E pi cu rusdismissed the reductionist psychology of earlier atomism as selfrefuting, and thus justified a non-reductionist psychology whichp ermi tt ed t he att ri bu ti on o f responsibility t o an auton omo usself with volitions, beliefs, impulses, etc., none of these beingstraightforwardly reducible to patterns of atomic motion. Thatwas, in my view, his most significant contribution to thecal a nd physi ca l det ermi nism were see n as mutually entailing, there is no evidencethat Diodorus himself supplied physical arguments for determinism. Butit may stillbe that t he i nfl uence of t he Maste r Argument ha d hel pe d some of Epi curus pupi lsdevel op t he ful l hard determinist consequences of Democritean physics).

61 T he re is n ev er any su ggesti on i n t he so urces t ha t e it he r o f t he t wo t hes eswas s ub or di na te d t o, o r a c on seq ue nc e o f, t he o th er . T he y a re t re at ed as parallel atCic. Fat 21-2 and Nat dear I 69-70, and as interchangeable at Fat18-19. Cf. al so ibid 37-8, cad II 97, for t he l ogic al t he si s.

6 Plut., Stoic repugn 1050 c; Sol/ an 9 64 e.63 Cf. Phil od., Sign 36.7-17; Diogenes of Oenoanda, 32.1-3.

~ j 1


19/22

46 DAVI D SEDLE Y EPI URUS R EFUTAT ION OF DE TE R MI NISM

crusade against determinism. But his atomic and logical theoriesstill had to be so constructed as not to pre em pt th e self s decisions b y d et er mi ni ng t he an ima l s b eh av io ur i nd ep en den tl y o fthem. Hence the indeter ministic swe rve, a nd, par alle l to it inlogic, t he denial o f bivalence - both theories being designedn ot t o ex pl ai n what volition is bu t to guarantee its efficacy bykeeping alternative possibilities genuinely open.It r em ains to ask whe ther the evidence of Lucretius can beinter pr eted c onsistently with this story. It reads as follows II25193):

Mor eover, if all motion is always linked, and new motion arises out of oldin a fix ed o rd er , a nd a to ms d o no t by s we rvi ng make a c er ta in b eg in ni ng o f m ot io nto break the decrees of destiny, so that cause should not follow cause from infinity,from w he re do es t hi s f re e v ol it io n e xi st fo r a ni ma ls t hr ou gh ou t t he wo rl d? F ro mw he re , I as k, c omes t hi s v ol it io n w re st ed away fr om t he fa te s, t hr ou gh w hi ch wep ro ce ed w he re ve r e ac h o f us is l ed b y h is p le as ur e, a nd l ike wis e s we rve o ff o urmotions at no fixed ti me or f ixed r egion of space, but wherever the mind itselfcarries us?

B. For without doubt it is volition that gives these things their beginning f oreach o f us, a nd it is f rom volition that motions are spr ead thr ough the limbs. Don tyou see how also when a t an ins tant the s tarting gates are opened the eagerstr ength of hor ses can never theless not sur ge f or ward as suddenly as t he m in d i tself wishes? For all the mass of matter has to be stirred up throughout the body, sot ha t s ti rr ed u p t hr ou gh all t he limbs i t may i n a c on ce rt ed e ff or t follow t he m in d sde sire . T hu s you may see that the beginning of motion is c re at ed f ro m t he h ea rta nd p ro ce ed s i ni ti al ly f ro m t he m in d s v ol it io n, a nd f ro m t he re is spread furthert hr ou gh t he e nt ir e b od y a nd l imbs .

C. Nor is i t t he same w he n we move f or ward impelled by a blow, thr ough ano th er p er so n s g re at s tr en gt h a nd g re at c oe rc io n. F or t he n i t is plain that all them at te r o f t he w ho le b od y mov es a nd is driven against our wish, until volition hasr ei ne d i t b ac k t hr ou gh ou t t he li mb s. So d o you now see t ha t, a lt ho ug h e xt er na lf or ce propels many along and often obliges them to proceed against their wishesand to be driven headlong, never theless there is something in our chest capable off ighting and r esisting, at whose decision the mass of matter is also f or ced at timesto be turned throughout the l imbs and f rame, and, when hur led f or ward, is reinedback and settles down.

D. T he re fo re in a toms too you mus t ad mi t t he s ame thing, t ha t t he re is another cause of motion besides blows and weight, f rom which this power is b or n i nus, since we see t ha t no thing can come int o being ou t o f n ot hi ng . F or w ei gh tp re ve nt s all t hi ng s f ro m c om in g a bo ut b y bl ow s, by a s or t o f e xt er na l fo rce . B ut

t ha t t he t hi ng [or the mind ]64 should not itself possess an internal necessity in alli ts behaviour , and be overcome and, as it we re , fo rc ed t o s uf fe r a nd t o be actedupon - that is b ro ug ht a bo ut by a t iny s werve o f atoms at no fixed re gi on ofspace or fixed time.

This text has been thoroughly r ansa cked for clues m anytimes in the past. It may well be c onsistent with a num be r ofi nt er pr et at io ns q ui te d if fe re nt fro m my own. My only o bj ect isto show that it fits in no less co mfo rt ab ly w it h w hat I haveargued on independent grounds to be Epicurus view.

Th e conclusion formally stated in D is that since th e tw opreviously established principles of an atom s motion, blows andweight, are deterministic, ,there must be a third, indeterministicprinciple, th e swerve. This conclusion is derived from observedfact - th e power o f animals to initiate unpredeterminedcourses of action A , and more specifically, the power of thesource of that indeterminacy, volition,65 to redirect large number s o f at oms in defiance of their purely mechanical patterns ofmotion B, C .

I c an no t ma ke u p my mind between the res of t he mss. and the f avouredemendation mens in 289. I am impressed by the linguistic arguments f or mens seeD. FOWLER i n thi s volume), b ut also by I. AVOTINS obser vation that, since theblows a nd weight i n 288 must still be as i n 28 5 t ho se m en ti on ed as causes ofatomic mo tio n, t he r efe re nce in 28 9 s hould b e to t he a tom, n ot t he m ind Thequestion mens in Lucretius 2.289, Q 29, 1979, pp. 95- 100) . On the latterview res might be retained meaning perhaps th ething , viz. the atom, but I confess thatit mak es p oo r Latin. mens is preferred the overall structural clarity of th e passagesuffers, but t he p oi nt made in 289-91 will fit i n closely wi th Cic. , Fat 23 a ndw it h my i nt er pr et at io n o f t he s werv e d oc tr in e. H ow ev er , I p re fe r n ot t o r es t mycase on such r ~ g i l evidence.

65 The r ef er ences to t he mind at 260 and 265 must be understood asequivalent to what Lucretius more accur ately calls C the mind s volition 270) orjust volition 261, 276; cf. 268, 281). Str ictly, the mind is the corporeal organ ofconsciousness, volition an emergent property of it. But the precise significance of, mind will n ot be established until Book III. That volition is not itselfstraightforwardly analysable in atomic terms explains why it is t he o ne s ta ge l ef tunanalysed in Lucretius IV 877-906, where the mechanics of animal motion ared es cr ib ed - a n o ld s ou rc e o f p er pl ex it y).


20/22

8 DA VID SEDLE Y EPICURUS R EFUTATI ON OF DE TE R MI NISM 9

This seems fully consistent with the interpretation offeredabove, that swerves are n ot in vol ved i n v ol it io n i ts el f but aret he element o f i nd et ermi nacy i n atomic m ot io n w hi ch en ab lesvo li ti on , an emerg en t p ro pert y o f t he mi nd , t o exercise con trolover th;e mind s physical processes, and thereby over the body,in such a way as to exemp t animal b eh av io ur from t he lawlikeregularity of natural phenomena. y interpretation must facethe difficulty that Lucretius nowhere makes the precise contribution of the swerve explicit. It is enough, after all i n t he cont ext of t he laws of atomic motion, for h im t o argue as so manyhave that the familiar phenomenon of volitional indeterminismcould not co-exist with physical determinism; and hence that theswerve must exist). But an advantage of the proposed interp retati on is t hat i t fully explains why Lucretius should so emphasise in B and C th at the power of volition to act on matterdepends on the swerve.

One plainly deliberate effect achieved in this passage is theimplicit analogy between A, where we swerve off our motionsat no fixed time or fixed region of space, and D s tiny swerve ofatoms at no fixed region of space or fixed time. It mig ht b eprude nt to grant Lucretius t he persuasive force of th e echowithout squeezing it t oo h ard for preci se t heoret ical con tent .But for those who feel t ha t t he echo is meant to hint at a directcorrelation between volitions and swerves,66 can p oi nt o ut thaton my interpretation there will indeed be such a correlation. Itwill not be one in which the existence, or the efficacy, of thevolition de pends o n t he happy coincidence of an u npr ede te rmin ed event like the swerve chancing to occur just whe n a ndw here i t is needed. Rather, as exp lain ed above, i t will b e o ne

66 NGL RT o p. c it ., p . 104) p oi nt sout a s im il ar e ch o o f the certain beginning of motion made b y the swer ve 253-4) a t 2 61 -2 , w her e v ol it io n gives abeginning t o a ni ma l moti on. If t hi s is t ho ug ht t o be a del i be ra t e ploy on Lucret iu s p ar t , my c om me nt s a pp ly t o it equally.

7 I o ffer this as a n o bj ec ti on t o mo st o f t he c ur re nt i nt er pr et at io ns of thes we rv e t he or y b ut f or a c ou nt er ar gu me nt , see NGL RT Ope c it ., p p. 197-9). Itmay also pose a t hr ea t t o FURLEY S view that o ne o r m or e p as t s wer ve s a re n ee de dt o b re ak the c ausa l c ha i n whi c h would ot he rwi se make your present c ha rac t er theinevitable consequence o f t he s ta te o f m at t er a t y ou r birth Perhaps he should have

in which the intervention of volition itself actualises in a groupof mind atoms the ever-present possibility of swerving.It may be suggested that the description of the swerve in Aand D, as that from which unde o ur p ow er arises to b eh av ein an unpredetermined way, implies a stronger relation between

it a nd volition t ha n am proposing. But as I observed earlier,no i nt erpret at io n o f t he swerve can make i t more t han a necessary condition of animal autonomy. Once it is granted that undeexpresses no stronger a relation than that,68 it becomes hard tosee what more the mere use of the word can be expected toreveal. It would, at any rate, be perfectly intelligible to say thato ur p ow er to determine our own behaviour arises from thefact tha t the laws of physics leave a large number of alternativepossibilities open. A nd t ha t fact is, o n my analysis, just whatt he s w; rv e d octrin e amo un ts to.In short, Lucretius argument can quite coherently be read

as an appeal to the evident power of volition to alter mechanicalprocesses as requiring that those processes be less t ha n fullyd et ermi ned b y physical laws, and hence as confirming the existence of the indeterministic swerve.

A riticismThe Academic sceptic Carneades proposed a highly intelligent and constructive modification to the Epicurean account.69Arguing entirely with premisses and analogies which the Epi-

stressed the mere possibility of a swerve s intervention, as being sufficientto render the actual outcome non-necessary: given the completely undeterminednature ofswerves, no one could count on actually having had one of his own . B ut a t all events,a lt hough I fol low FURLEY in a ssigni ng a pri ma ri ly negat ive rol e to t he s we rv e, Ibelieve that the precise function which he gives it is now rendered superfluous bythe non-reductionist psychology. What makes your present behavi our unde t ermined is the fact that the self which is responsible for it is more than a m er e b un d leo f a to ms , a nd t he re fo re n ot r ed uc ib le to a l in k i n a p hy si ca l c au sal c ha in .

68 For unde e xpre ssi ng a nec essa ry c ondi t ion, c f. Luc r. I 382.69 Cic., Fat 23 5


21/22


cureans could be expected to endorse 7 he observed that the autonomous character of voluntary motions of the mind need n omore have an e xter na l a ntec edent cause that did the downwar dmotion of the atom due to weight. It was the nature of theatom that in itself c ause d it to move downwar ds a nd likewisethe nature of the voluntary motion of the mind that in itselfcaused it to be up t o u s Hence he maintained the Epicureandoctrine that there could be a voluntary motion of the mindwas quite sufficient to answer determinism without the swervetheory s implicit aban do nmen t of th e principle that everymot ion has a cause.

It is clear from this t ha t Carneades recognised t he Epicurean analysis of the voluntary motion of the mind to be anon-reductionist one. If he had under stood it as no more than achain o f atomic motions he could not have at tr ib ut ed t o i t aself-determin ing n at ur e wh ich t he m in d s c on st it ue nt atomsthemselves lacked. But if my overall interpretation of t he Epi curean position is correct Carneades advice is very much tothe point. P er ha ps Epicurus could h ave r es te d c on te nt w it h hisnon-reductionist analysis of voluntary motions of the mind asa utonom ous e ve nts ca pa ble of initiating new courses of a ctionand t hu s bely ing d et erm ini sm . H is a dd it io n o f t he swerve doctrine was a very understandable attempt to ensure that thechanges w hi ch t hes e volit ions w er e r eq ui re d t o b ri ng ab ou t i natomic trajectories should be physically possible ones. But whyneed the ir being physically possible results of volitions require

70 In particular the analogy of th e causes of atomic motion ibid 24-5 , but also ibid 24 the void example where quibus inane esse nihil placet w ll mean not whohold that no vacuum exists RACKHAM; likewise YON , b ut who hold t ha t th e empty is nothing at all an attempt to characterise the atomist s technical sense of void . Carneades dialectical technique required that he use only the premisses ofother philosophers never his own see, for example Gisela STRIKER, Sceptical stra-tegjes n M. ScHOFI l et a [eels.], Doubt and Dogmatism Oxford 1980 . His obje ct in this case, as Cicero makes clear was to strengthen the Epicurean positioninto an adequate counterweight to Chrysippan determinism. This feature of hismethod sometimes gave the false impression that he himself endorsed the view thathe was defending Cic., Acad II 78, 139 ; Fin V 20 .

tha t the y also be capable of happening independently volitions?Per hap s h er e a t l as t t he i mp act o f t he swerve s secondary rolebecomes apparent for E pi cu rus could n ot a t any cos t a ff or dt o give v ol it io n a causal role n cosmogony. In th e co nte xt ofthe ancient debate on determinism at least it might have beenmore politic to insist as a datum of experience that volitions dosomehow exert an independent causal influence on matter andt o leave i t a t t ha t. B ut from t he van tage p oi nt o f t he t we nt ie thcentury. it is still t oo soon t o conclude t ha t E pic ur us mo re adventurous treatment was altogether wrong.71

Ancestors of this paper and versions of it at various stages of completionwere read to meetings at Lille Cambridge Baltimore Berkeley and Harvard. Inaddition the n n ture text which is central to it was extensively discussed in a series of seminars at Princeton University in 1981-2 and the penultimate version waspresented to the Princeton Classical Philosophy Colloquium in December 1981. Iam grateful for the enormously helpful comments I received from far more peoplethan I can list here. Among them special mention must be given to Richard Sorabji Myles Burnyeat Don Fowler Walter Englert Alan Code Don Morrison Nicholas White Tony Long John Cooper Glenn Most Michael Frede and aboveall, David Furley my commentator at the Princeton Colloquium whose own workfirst inspired my interest in this topic and continues to serve as a model. Finally Iowe great debts of gratitude to the Humanities Council of Princeton University forthe award of a visiting fellowship in the Fall Semester and to the Institute for Advanced Study Princeton for membership in the second term of 1981-2 when.the current version of the paper was written; and to Marcello Gigante to whose efforts and personal example it is due that the Herculaneum papyri are at last receiving the serious attention that they deserve.


22/22

I

k Y Z T ~ I ~ - STUDI SULL EPICUREISM OGRECO ROM NO_

OFFERTI MARCELLO GIGANTE . . . .

GAETANO MACCHIAROLI EDITOREh ~ t?Y

david sedley - epicurus' refutation of determinism

Documents