Oxford Studies in Ancient Philosophy: Volume XXIV: Summer 2003 0199263434, 9780199263431

416 60 16MB

English Pages 416 [410] Year 2003

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

Oxford Studies in Ancient Philosophy: Volume XXIV: Summer 2003
 0199263434, 9780199263431

Citation preview

OXFORD STUDIES IN ANCIENT PHILCTSOPHY

VOLUME XXIV SUMMER 2003

Oxford Studies in Ancient Philosophy is a volume of original articles on all aspects of ancient philosophy. The articles may be of substantial length, and include critical notices of major books. Aristotle and the Stoics receive particular attention in this volume. Editor: David Sedley, Laurence Professor of Ancient Philosophy, University of Cambridge

‘standard reading among specialists in ancient philosophy’ Brad Inwood, Bryn Mawr Classical Review

OXFORD STUDIES IN ANCIENT PHILOSOPHY EDITOR: DAVID SEDLEY

VOLUME XXIV SUMMER 2003

OXFORD UNIVERSITY PRESS

OXPORD UNIVERSITY PRESS Great Clarendon Street, Oxford 0x2 6dp Oxford University Press is a department of the Univershy of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education byj^ublishing worldwide in Oxford New York Auckland Bangkok Buenos Aires Cape Town Chennai Dar es Salaam Delhi Hong Kong Istanbul Karachi Kolkata Kuala Lumpur Madrid Melbourne Mexico City Mumbai Nairobi Sao Paulo Shanghai Taipei Tokyo Toronto Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries Published in the United States by Oxford University Press Inc., New York © Except where otherwise stated, O.xford University Press, 2002 The moral rights of the authors have been asserted Database right Oxford University Press (maker) First published 2003 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this book in any other binding or cover and you must impose the same condition on any acquirer British Library Cataloguing in Publication Data Data available Library of Congress Cataloging in Publication Data Oxford studies in ancient philosophy.— Vol. xxiv (2003).—Oxford: Clarendon Press; New York: Oxford University Press, 1983— V.; 22 cm. Annual. I. Philosophy, Ancient—Periodicals. /

B1.O9

180. 5—dc.19 AACR 2

84-645022

MARC-S

ISBN 0-19-926343-4 ISBN 0-19-926344-2 (Pbk.)

13579 10 8642 Typeset by John Was, Oxford Printed in Great Britain on acid-free paper by T. J. International Ltd., Padstow, Cornwall

ADVISORY BOARD

Professor Jonathan Barnes, Universite de Paris-Sorbonne, Paris IV Professor Michael Frede, Keble College, Oxford Professor A. A. Long, University of California, Berkeley Professor IMartha Nussbaum, University of Chicago Professor Richard Sorabji, King’s College, University of London, and Wolfson College, Oxford Professor Gisela Striker, Harvard University Contributions and books for review should be sent to the Editor, Professor D. N. Sedley, Christ’s College, Cambridge,

CB2 3BU,

UK.

He can be contacted by e-mail on [email protected]. Contributors are asked to observe the ‘Notes for Contributors to Oxford Studies in Ancient Philosophy’, printed at the end of this volume. Up-to-date contact details, the latest version of Notes to Con¬ tributors, and publication schedules can be checked on the Oxford Studies website: www.oup.co.uk/philosophy/series/osap



' ‘VfORn

H

q>V/;Oa ■/

t

'■_

^

tr.

f'tj'f***

'•

N*' J !..■»■« I-

.‘‘

«- ->

'■» i-

■ ‘ *•

ir'jfiti. i

Ht# «

«

.

rv^.«

4 .**

’f«ir *

.

n^AvtWiii V

loassloi'T

,vi>\V6'j Hot\!ia'H Wtft

^

» :!)l}>fi« jvr**.—

.Kiiib'J ^

’ I'*!'•

'*■

bl4^< vW9iv«i

.){!/ t IVk '

*4

,f

«*ifi io

'

.1



'iJ

*Mlj r -«»

o aiBb-oJ-qU

rto brt93i^b. jrt ,1^

nrul*»»Jdtfq hiw ^loiudhn

^

^

...

ft B>». I

;aitiKb»w v*dutlZ

* • j

I. * vrfqrxol I rfq vrfi»r>>4| uo .jrww *

’ -

* i% 4 ^

»■•'»♦ •

«- ?

•j i-j»

it

4

■?K

1 V • * *•♦
ii iin^3«\.

T^E

'V

■’A:. . jUs

.'ll I'.'.i,••- i’'

I

EMPEDOCLES ON THE ULTIMATE SYMMETRY OE THE WORLD SIMON TREPANIER

I.

Introduction; L’Empedocle de Strasbourg

The modern study of Empedocles entered a new era in 1999, with

the publication by Alain Martin and Olivier Primavesi of a sensa¬ tional new text, the Strasbourg papyrus. ‘ The papyrus, dated to the first or second century

ad

and assembled from many smaller pieces,

comprises six continuous sections, of which four are of significant size. These sections, called ensembles by the editors, preserve a total of 74 full or partial new hexameter lines, while a further 20 coin¬ cide with lines previously known, leaving no doubts concerning the identification of the work as Empedoclean.^ The papyrus is also, to all evidence, a text of Empedocles, not a quotation, and is thus the first instance of direct textual transmission of that author.^ {a) The three thetas Among the issues raised by this new material, that which so far has generated the most scholarly controversy centres on a difficult © Simon Trepanier 2003 I would like to express my gratitude to Professors Brad Inwood and David Sedley for their helpful comments on and constructive criticisms of earlier drafts of this paper. ‘ A. Martin and O, Primavesi, L’Empedocle de Strasbourg (P. Strasb. gr. Inv. 16651666): introduction, edition et commentaire [M-P] (Berlin and New York, 1999). ^ The first 5 lines of ensemble a overlap with lines 31-5 of fr. 17 (H. Diels and W. Kranz, Die Fragmente der Vorsokratiker, 6th edn. [DK] (3 vols.; Berlin, 1951), vol. i, no. 31); ensemble b 2 and 4 = B 76, 3 and 2 DK; ensemble c 2-8=B 20 DK; and ensemble d 5-6 = B 139 DK (see n. 7 for details). In M-P’s notational system the bold letters a to h stand for different ensembles, (i) or (ii) for column numbers, the lines being then numbered by column. ^ This new material now brings the number of extant verses in the Empedoclean corpus to a fairly substantial total of 524 lines and a few phrases, but if one disregards IS of the papyrus’ most severely mutilated new lines, where less than two words are visible, the total is closer to around 509.

2

Simon Trepanier

textual problem. The debate turns on what to make of a previ¬ ously unknown textual variant revealed by the papyrus in two, per¬ haps three, instances of an otherwise familiar Erppedoclean turn of speech, which I will call.the ‘unification formula’. This formula, a recurring poetic phrase in Empedocles’ poetry, is one of a number of such formulae coined by Empedocles in a creative adaptation of epic tradition to his own didactic ends. As found most conspicu¬ ously in fragments 17, 20, 21, and 26, the formula is one part of a frequently deployed poetic motif, usually found as two paired lines, symmetrically contrasting the unification formula (A) in the first line with a separation formula (B) in the second.'* Most commonly, the two formulae depict the contrasting influence that Empedo¬ cles’ two main cosmic/psychological powers, Love and Strife, exert upon the four elements. Love causing them to unite or come to¬ gether, and Strife driving them apart. Almost without fail, in the (A) portion of the motif Empedocles employs the verb owepxofjiaL to describe the process of unification, e.g. 31 B 17. 7-8 DK: aXXore p.ev ^iXorpTL avvepxofiev’ els iv airavra aXXore S’ av hlx’ eKaara (ftopevp^eva NeiKeos exOet.

at one time through Love all coming together into one, at another in turn each carried apart by the hatred of Strife.

The novelty of the papyrus is an intriguing variant in that verb in two, perhaps three, places: a (i) 6, c 3, and then, less certainly, at a (ii) 17. In these new passages, instead of the neuter plural par¬ ticiple, the only known form throughout the entire indirect textual tradition, the copyist has written a 6 instead of a v at the end of the verb, transforming it from a neuter plural participle to a first person plural. But in two of those three cases, a (i) 6 and c 3, a second hand has corrected the copyist’s 9 back to a v, turning the verb into the more familiar (to us) participial form.^ M-P, presented with such a choice, in both cases prefer the first hand’s text. In their view, the variant is too systematic to allow the

“ For a fuller account of this poetic motif and its philosophical significance, see D. W. Graham, ‘Symmetry in the Empedoclean Cycle’ [‘Symmetry’], Classical Quarterly, ns (1988), 297-312. ^ In the first, a (i) 6, this correction is unmistakable, whereas in c 3 the papyrus is too damaged to reveal the actual letter, but the traces of a correction above the still legible 8 indicate the summits of two vertical lines, suitable for N or H. For c 3 see p. 142 and pi. V.

Empedocles on the Ultimate Symmetry of the World

3

possibility of a random error.* In the case of a (ii) 17, moreover, which was not revised by the second hand, they do not even posit an instance of the unification formula, and so restore the verb with eta- rather than ow- as the prefix, to the imperfect eta7^]pxd(U.e0(a) ‘we were coming together to’. The doctrinal upshot of this abstruse philological point, as M-P see it, is to bring ‘us’, i.e. humanity, directly into the cosmologi¬ cal picture Empedocles is developing throughout fragment 17, of which ensemble a is the direct continuation.’ This is so, they con¬ tend, despite the rather conspicuous absence of any other mention of ‘us’ before, after, or in between these points, according to their own edition of the sequence formed by B 1"] +ensemble a. This becomes even more astounding when M-P, pp. 83-6 and 90-5, proceeding from such debatable evidence, purport to found upon their restoration of a (ii) 17 an identity of the migrating daimon and the Empedoclean first principle Love. Eor if the identity of the daimon with Love has been canvassed before, it is no more than a modern scholarly conjecture.* If we omit it as a possibility here, it has no direct support in the fragments, nor any ancient pedigree whatsoever."^ To the extent that it can be measured, the initial scholarly re¬ action to this editorial decision, along with its purported doctrinal implications, has been mixed.'® My own view, arrived at indepen‘Nous excluons en tout cas que les formes de la le personne du pluriel puissent resulter, par une extraordinaire coincidence, de trois fautes survenues de maniere independante dans le texte’ (91). ’’ The join is assured by a five-line overlap, linking the end of fr. 17, quoted by Simplicius at In Phys. 157. 25 Diels, to the opening of ensemble a; see M-P 103-10 for discussion. Thanks to Simplicius, we also know that the passage came from the first book of Empedocles’ Physics, and now a stichometric mark at the end of ensemble a, P, indicates that this very line was the 300th line of the first book, i.e. that fr. \-] + ensemble a spanned roughly lines 233-300 of book i of the poem read by Simplicius. * See F. M. Cornford, From Religion to Philosophy (London, 1912) 233 ff.; C. H. Kahn, ‘Religion and Natural Philosophy in Empedocles’ Doctrine of the Soul’, Archiv fiir Geschichte der Philosophie, 42 (i960), 3-35, repr. in A. P. D. Mourelatos (ed.), The Pre-Socratics, 2nd edn. [Pre-Socratics] (Princeton, 1992), 426-56; D. O’Brien, Empedocles’ Cosmic Cycle [ECC] (Cambridge, 1969), 325-34 (but the latter two advance the view much more tentatively than Cornford did). ’’ No ancient source ever tells us that Empedocles thought that the soul =the first principle Love. On the contrary, see e.g. fr. 109 and Aristotle’s discussion at De anima 404*. The reviews known to me are: B. Inwood, Classical Review, NS 50 (2000), 5-7, who endorses the thetas; M. L, Gemelli Marciano, in Gnomon, 72 (2000), 389-400,

4

Simon Trepanier

dently of the reviews since published, certainly registers at the negative end of the spectrum." k

{h) From red herring to old chestnut \

The downside of this, of course, is that, if the negators are correct, and the three thetas are a red herring, then the papyrus is that much less significant. Or is it? The answer is not that obvious, since until now the dissenting side has not had occasion to provide an alternative reconstruction of the debated passages, not to mention other issues raised by the papyrus. To be sure, one way of testing the three-thetas reading is to ask whether, when we do consider this material without the ‘we’, a better text results. But one does not get very far down this road before becoming entangled in some much more extensive problems. The question of the three thetas, it turns out, cannot be tackled directly, for it goes deeper than a single debatable philological decision. In the first place, that decision is predicated upon a deeper set of interpre¬ tative problems vitiating M-F*’s reconstruction of ensemble a. And these problems, by way of a second, still more extensive, ramifi¬ cation of complications, are generated by M-P’s debatable views about the cosmic cycle as a whole. For if one were to unravel these strands all the way to the end, it would become clear that Primavesi (the Empedocles specialist of the two) has undertaken the inter¬ pretation of this new material with a preconceived idea of what it ought to contain, namely O’Brien’s reconstruction, as propounded in ECC (above, n. 8)." And while I am less hostile to O’Brien’s who does not take a side; C. Osborne, ‘Rummaging in the Recycling Bins of Upper Egypt: A Discussion of A. Martin and O, Primavesi, L’Empedocle de Strasbourg’ [‘Rummaging’], Oxford Studies in Ancient Philosophy, i8 (Summer 2000), 329-56, who is quite sceptical (345-6), as is S. T. Mace, Classical World, 95 (2002), 195-7; Jaap Mansfeld and Keimpe Algra reject them, in ‘Three Thetas in the “Empedocle de Strasbourg’” [‘Three Thetas’], Mnemosyne, 54 (2001), 78-84. " See my abstract in the abstract book of the American Philological Association 131st Annual Meeting, 27-30 Dec. 1999, p. 97, and in particular my study cited below, p. 5. Thus I would go even further than Osborne’s scepticism (‘Rummaging’, 349), to side with Mansfeld and Algra, ‘Three Thetas’, 81: ‘And the 0s in the papyrus fragment discussed above are simply wrong. The slightly bizarre interpretation based on them may be abandoned.’ A more concise formulation of his position is also available in his ‘Empedocles Revisited’ [‘Revisited’], Ancient Philosophy, 15 (1995) 403-70. Osborne, ‘Rum¬ maging’, 336-7, arrives at the same diagnosis of Primavesi’s excessive reliance on O’Brien, while Inwood, in the preface to the second edition of The Poem of Empe-

Empedocles on the Ultimate Symmetry of the World

5

reconstruction than others, and agree with the editors that the new evidence does confirm it in part, this in turn seems to have blinded M-P to those sections of the new evidence which do not provide such support. Thus, there is much basic work still to be done. First, that great old chestnut of Empedoclean studies, the question of the cosmic cycle, has yet to be re-examined in the light of the new material from the papyrus, in a manner that recognizes the possibility that this material may overturn, rather than merely confirm or refine, earlier suppositions about the cycle. Equally pressing is the need to provide an alternative reconstruction and interpretation of en¬ semble

a, in particular a (ii) 3-17. Elere, to anticipate somewhat, I

think that M-P’s identification of these lines with the phase of the cycle known as the reign of Strife is mostly wrong, and without it their reading of the ‘we’ as particles of Love becomes untenable. Eor a full discussion, including my suggestion for an alternative reconstruction of the text, I refer the reader to my study ‘“We” and Empedocles’ Cosmic Lottery: P. strasb. gr. Inv. 1665—1666, en¬ semble

a’, forthcoming in Mnemosyne. In the present study I will

concentrate upon the first question, an assessment of the new evi¬ dence’s contribution to the debate about the general nature of the cycle.

(c) Note on the cosmic cycle: old chestnut and red herring? The centrality of the cosmic cycle to Empedocles’ doctrine may seem so obvious as not to require any defending, but it has re¬ cently come under fire, and I here append a few words to justify my approach. In particular, Osborne has attacked the central posi¬ tion of the cycle in Empedocles’ thought in two articles, first in her important CQ article, apd more recently in her OSAP reviev/ of M-P’s edition (‘Rummaging’).” Her scepticism about the impor¬ tance of the cycle, back in 1987, and now with even more vigour. dodes [Poem] (Toronto, 2001), explains his reluctance to follow many of M-P’s editorial decisions because of ‘unease about the degree to which their conjectural supplements reinforce a view about the poem, which in turn justifies further con¬ jectural supplements’. ” C. Osborne, ‘Empedocles Recycled’ [‘Recycled’], Classical Quarterly, NS 37 (1987), 24-50. R. M. Wright, Empedocles: The Extant Fragments, Edited with In¬ troduction, Commentary, Concordance [Extant Fragments] (New Haven, 1981), is another critic of the importance given to this issue, but hardly as vehement.

6

Simon Trepanier

flows from her belief in a single original Empedoclean poem, rather than the traditional division between two works, and through that, the resulting unity of Empedocles’ thought.Byilding on the re¬ newed importance of demonology and eschatology within this uni¬ fied scheme, Osborne seeks to discount the importance or centrality of the cosmic cycle. More specifically, she suggests that such an ap¬ proach reflects a continued confidence in familiar preconceptions established in the classic work of Denis O’Brien, preconceptions which have guided much tradi¬ tional work in the twentieth century, but which belong to a period when Empedocles was read in compartments, when the Physics was taken to be a work of natural philosophy whose sole aim was to set out a sequence of mechanical cosmic events, and when there could be no point to Empedo¬ cles’ message unless it settled the details of an elaborate pluralist response to Parmenides. (‘Rummaging’, 337)

Osborne’s main point here, as in her 1987 critique, is that Empedo¬ cles probably never sought to spell out in so many words the exact phases of the cycle and their attendant chronology, or perhaps, more drastically, that he never even devoted much serious thought to the question.” Either possibility, in her view, would help to explain the amount of interpretation advanced by commentators, both ancient and modern, in order simply to formulate a view of the cycle. Against the more drastic of these claims is the evidence from Aristotle and Plutarch, who both advance quite definite views on the cycle, some of which I discuss below, and more directly, the degree of detail Empedocles allows himself in questions of natural history.” Even more to the point, Osborne’s own alternative read¬ ing of Empedocles, in particular her exclusion of the response to This is in my view one of the most important questions raised by the new material, but I cannot consider it here. The most recent full case in print for the single-work hypothesis is the second edition of Inwood, Poem, while David Sedley, in the first chapter of his Lucretius and the Transformations of Greek Wisdom (Cam¬ bridge, 1998), makes a case for the two-work hypothesis; both write in cognizance of the new material. The first is the cautious middle position of W. K. C Guthrie, A History of Greek Philosophy, ii. The Presocratic Tradition from Parmenides to Democritus {HGP ii] (Cambridge, 1965), 180-1. One thinks among others of fragments frr. 84 and too, devoted respectively to the functioning of the eye and lungs, or his interest in plants, animals, embryology, etc. Given such detail, it seems unreasonable to maintain that no further specifics on the cycle were advanced, beyond the broad pattern of alternation given in frr. 17, 21, 26, and 35, or that these specifics could not be inferred from other passages.

Empedocles on the Ultimate Symmetry of the World

7

Parmenides, strikes me as so deflationary as to risk banishing him from early Greek philosophy and turning him into an ancient Sicil¬ ian Savonarola.” Yet perhaps the most potent objection of all to her counter-position is that, even following the single-work hypothesis she herself has championed, the fragments concerned with the cos¬ mic cycle—fragments 17, 26, and 35—will nevertheless remain the core doctrinal passages of the work. This is shown, first and most directly, by the contrast in scope between fragments 17, 26, and 35, which present an account of the universe or ‘everything’, and the more circumscribed tale of the exiled daimones in fragment 115. This, especially on the hypothesis of a single work, indicates that the story of the daimon would not have stood alone, but must somehow have been accommodated within the former. Such a verdict, coming from the fragments themselves, is reinforced in turn by the atten¬ tion paid to fragments 17, 26, and 35 by Aristotle, Theophrastus, and Simplicius, to the exclusion of the ‘demonological’ material, even though that, at least if the single-work hypothesis is correct, may well be a distortion of the original. In short, Osborne is not wrong to raise these methodological issues, but she overstates them, and the result is an arbitrary cur¬ tailing of the work. The questions of how far one can go in recon¬ structing the structure of the cycle, and what concerns the cycle was meant to address, should and no doubt will continue to be asked. The outcome of this issue does play an important part in our at¬ tempt to make philosophical sense of Empedocles. In the words of A. P. D. Mourelatos: one might judge the . . . question [the number of stages in the cycle] to be of merely antiquarian interest, and esoteric. To do so would be a mistake. The four-stage and two-stage interpretations'® respectively place ” ‘Recycled’, 48—9: ‘Empedbcles’ preoccupation is with motivation; his observa¬ tion that mortals perversely choose a life of hostility, strife and bloodshed when their lot would be much improved if they lived in love and unity leads him to postulate an explanation in terms of an opposition between Love and Strife, a degree of delusion on the part of those who live in strife and an element of compulsion to explain why things ever went wrong in the first place. Clearly he seeks some sort of symmetry between the forces . , . but to seek precise measures and equal periods, times of increasing Strife matching increasing Love, and to enquire about the duration of “total Strife” or “total Love” would be misguided.’ Once again, what is to prevent Empedocles from worrying about both, rather than one or the other? '* By ‘four-stage’, Mourelatos designates the traditional view: an endless cycle with two separate worlds like ours between a-cosmic periods, where the complete domination of Love or Strife precludes the existence of individual mortal crea-

Simon Trepanier

8

Empedocles in. two distinct contexts of Greek thought . . . What the fourstage interpretation emphasizes is the theme of isonomia, equal balance between polar and distinct powers ... The two-stage interpretation, by contrast, emphasizes the superiority of Love over Stril'e. For it holds that although there once was, and may a^ain be, a period of the complete sway of Love, there never was, nor will there ever be, a period of the complete sway of Strife. The two-stage interpretation has affinities not with egalitarian but with hierarchical conceptions of reality, such as those of the Pythagoreans, Heraclitus, and of Plato. (Pre-Socratics, 17)

2. Modern scholarship on the cycle The modern debate over the nature of the cycle can be character¬ ized broadly as a series of independent challenges to a view of the cycle first advanced by Panzerbieter in 1844,'“' hereafter referred to as the traditional view. Relying in part only on the evidence of Aristotle and Plutarch, it holds that between two opposed, a-cosmic phases, characterized by the complete domination over the elements of Love or Strife, there come into being two separate worlds, con¬ taining mortal creatures, wherein the two forces vie for control over the elements. Both of these worlds come to an end, along with their mortal inhabitants, when one of the two forces has complete sway over the elements. Under the full sway of Love, all the elements are harmoniously fused together into ‘One’, also described by Empe¬ docles as the Sphairos (fragments 27-9). Strife meanwhile, having retreated outside the elements (?), then reasserts itself by destroying this unity (fragments 30-1), thereby creating ‘Many’, and contin¬ ues to assert itself until it has separated the four elements into pure or homogeneous bodies. This state is the reign of Strife. Then once more it is Love’s turn to take the initiative (fragment 35), reintro¬ ducing mixture into the cosmos, and blending the elements in ever increasing amounts until it has reconstituted the Sphairos. Having come full circle, the process begins anew, and so on ad infinitum. This view, endorsed by Zeller,met with a few challenges late in tures. By ‘two-stage’ he means a linear progression, with Love as sole creative power. For clarifications, see the next section. F. Panzerbieter, Beitrdge zur Kritik und Erkldrung des Empedokles (Meiningen, 1844). E. Zeller, Die Philosophie der Griechen in ihrer geschichtlichen Entwicklung, 6th edn. by W. Nestle (3 vols.; Leipzig, 1919-20; repr. Hildesheim, 1963), i/2. 969-79.

Empedocles on the Ultimate Symmetry of the World

g

the nineteenth century,^' but was reinforced once more by the work of Bignone,^^ w'hich remained for over fifty years the most detailed account. Disregarding for now smaller but not inconsequential dif¬ ferences among its adherents, the salient feature of the traditional view is its conception of a double creation and destruction of life within one single cycle. According to this view of the cosmic cycle, these tw'O worlds arise, and then disappear, in the intervals between the a-cosmic phases, when both forces still vie for control. As I will argue more fully below, this broad cosmic alternation is best understood as a symmetrical process, a sort of pendulum swing, each half of which operates in the opposite direction: in one half. Strife waxes while Love wanes, and vice versa in the other. Thus, for any given cosmic phase, one may speak of Love or Strife as having the initiative, or overcoming the other, or again of one world coming into being under Strife, the other under Love. But at no time in these cosmic phases, qua cosmic phases, does either force rule without opposition; that, as I will try to show, is the distinguishing mark of the a-cosmic phases. This picture, or even more precisely, the cosmogonic and zoogonic role of Strife it implies, was the main focus of criticism in three independent studies by Bollack, Holscher, and Solmsen, all pub¬ lished in 1965.“ While there is no single counter-position to the traditional view, Bollack’s critique is by far the longest, and as a slogan to this counter-tradition, one might well apply a sentence of his which encapsulates its basic spirit: ‘Neikos ne construit jamais rien.’^"^ The traditional view, however, was strongly reasserted in 1969 by O’Brien in ECC, whose heroic labours in amassing and P. Tannery, ‘La cosmogonic d’Empedocle’, Revue philosophique de la France et de I’etranger, 24 (1887), 285-300; H. von Arnim, ‘Die Weltperioden bei Eimpedokles’. Festschrift Theodor Gomperz (Vienna, 1902), 16—27. “ E, Bignone, Empedocle: studio critico (Turin, igi6), in particular his appen¬ dix II, ‘II ciclo cosmico’. ” J. Bollack, Empedocle, i. introduction d Tancienne physique [Bollack, i] (Paris, 1965), of which see in particular ‘Le faux probleme’, 95-124; ii. Les Origines: Mition et traduction des fragments et des temoignages, and iii. Les Origines: commentaires [Bollack, ii and iii] (Paris, 1969); U. Holscher, ‘Weltzeiten und Lebenskyklus: eine Nachpriifung der Empedokles-Doxographie’ [‘Nachpriifung’], Hermes, 93 (1965). 7-33; p' Solmsen, ‘Love and Strife in Empedocles’ Cosmology’ [‘Love and Strife’],

Phronesis, 10 (1965), 109-48, repr. in R. Allen and D. Furley, Studies in Presocratic Philosophy [Studies], vol. ii (London, 1975), 221-64. P’or a critique of Holscher’s analysis of the doxography, see now D. O’Brien, ‘Hermann Diels on the Presocratics: Empedocles’ Double Destruction of the Cosmos, (Aetius ii 4. 8)’, Phronesis, 45 (2000), 1-18.

Bollack, 1. 114.

lO

Simon Trepanier

analysing all the secondary sources remain unsurpassed to this day. Despite his contribution, dissent remained, in particular because the key primary evidence for the traditional viev^ namely B 17. 35, as emended by Panzerbieter, is expressed in a highly condensed form. The difficulty in establishing the precise reference of these lines means that one can still appeal to Empedocles’ own words, over the heads of Aristotle, Theophrastus, and much of the doxography, to build a case for a single creation within the cycle. The most important attempt of this kind was made by A. A. Long in ‘Empedocles’ Cosmic Cycle in the Sixties’.^’ After that work, a certain fatigue naturally set in, and while interpreters continued to take sides, few mustered the energy to go over all the evidence in print, and instead, perhaps more wisely, devoted their energies to less worked-over problems.^* The question thus fell somewhat by the wayside, and Long’s interpretation did not receive a full re¬ ply until D. W. Graham’s ‘Symmetry’ (above, n. 4), a study whose importance has not yet been fully appreciated. Still more recently, despite some relative cooling of the debate, the traditional view of the cycle had been undergoing a revival, and now, with the publi¬ cation of the papyrus, I believe it to be vindicated.^’

In Mourelatos, Pre-Socratics (ist edn. 1974), 397-425, henceforth ‘Sixties’. Thus e.g. J. Barnes, The Presocratic Philosophers (London, 1982), 308 ff., and Wright, Extant Fragments, both support the traditional view. Wright, 41 ff., gives the matter more attention, but does not make it the centre of her study (see her notes to the 1995 reprint, p. 310). The dissenters at this period include J. C. Liith, Die Struktur des Wtrklichen im empedokleischen System iiber die Natur (Meisenheim, 1970); J. Mansfeld, ‘Ambiguity in Empedocles B 17 3-5: An Interpretation’ [‘Ambi¬ guity’], Phronesis, 17 (1972), 17-39; and N. van der Ben, The Proem of Empedocles’ Tlepl fvaews: Towards a New Edition of the Fragments [Proem] (Amsterdam, 1975), criticized at length by D. O’Brien, Pour interpreter Empedocle [Pour interpreter'] (Paris, 1981). Other dissenters worthy of note include G. S. Kirk, J. E. Raven, and M. Schofield, The Presocratic Philosophers [Presocratic Philosophers] (Cambridge, 1983). 288-9, D. Furley, The Greek Cosmologists, vol. i [Greek Cosmologists] (Cam¬ bridge, 1987), 98-104; and Osborne, ‘Recycled’, 38-49. M-P 75-86, 96. The revival of the traditional view was noted by O’Brien in ‘Revisited’, 403-4. More recent defences of the traditional view include K. Steiger, ‘Die Kosmologie des Parmenides und Empedokles’, Oikoumene, 5 (1986), 173-236; A. Stevens, ‘La physique d’Empedocle selon Simplicius’, Revue beige de philologie, 67 (1989), 65-74; Inwood, Poe?n, 41-52. A case for the traditional interpretation, centring on the doxographical evidence for zoogonic phases, is made by J, Wilcox, ‘Whole-Natured Forms in Empedocles’ Cosmic Cycle’ [‘Whole-Natured Forms’], in A. Preus (ed.). Essays in Ancient Greek Philosophy, vi. Before Plato (Albany, 2001), 109-22.

Empedocles on the JJltimate Symmetry of the World

11

3. The traditional view in the light of the new evidence The reality of a cycle of some sort has been openly denied by only a few scholars.^* I therefore take it for granted, and assume that at the very least it consisted in an alternation between ‘One’ and ‘Many’, as the most superficial reading of fragment 17 allows; and I rely upon the very word

kvkXos

applied by Empedocles to the process.^''

What the various challenges to the traditional view have repeat¬ edly called into question is the creative, and hence in some regards positive, role of Strife in the creation of the cosmos and its inhabi¬ tants. For it is on the basis of a negation of these capacities that they wish to deny a cosmogony or a zoogony, or both, under the influence of Strife, and thereby eliminate one of the four phases postulated by the traditional interpretation. Bollack, apparently taking as his cue line 20 of fragment 17, where Strife is introduced as owAd/xevov ‘destructive’, allowed it neither a cosmological nor a zoogonic role. He limited its function to an initial pulverization of the Sphairos, which Love then undertook to repair, first by sorting and collecting the elements together (the reign of Strife in the traditional inter¬ pretation), and from there combining them into mixtures. In point of fact, it is unclear whether or not Bollack envisaged any actual return of the Sphairosd'' Thus, if there is a cycle according to him, only in a very limited sense is Strife the author of the world, for it does no more than create the need for its reconstruction by Love. Solmsen’s account is somewhat more generous towards Strife, The strongest outright negations of the cycle are by Holscher ‘Nachpriifung’, and van der Ben, Proem, amply refuted by O’Brien, ECC (for Holsher and Bollack) and Potir interpeter (for van der Ben). Bollack at his most obscure and Neoplatonic sometimes seems to deny the reality of the cycle, e.g. Bollack, i. 151-2, but at other places (e.g. i. 97, no) seems more positive. B 17. 3 aKivTjToi Kara kvkXOv. This is reinforced by the depiction of ‘One’ and ‘Many’ as the end products of the cosmic process: cf B 17. 1-2, and now a (i) 6-7 and perhaps a (ii) 18—20. Bollack, i. 95-124. It is possible that Bollack wanted to see in Empedocles a parallel to 1 leraclitus and the debate as to whether or not that thinker was committed to the idea of a cyclical ekpyrosis (but see Bollack, i. 151) If, as most hold, Heraclitus did not believe in that doctrine, then it is perhaps easier to view Empedocles also as resolving the ‘one’ vs. ‘many’ opposition within the world, not via a succession of worlds. Cf Bollack, iii. 149: ‘le cercle s’accomplit au moyen des instants de repos qui compensent a chaque instant I’elan du mouvement.’ Against this is the contrast between the two thinkers sketched by Blato at Sophist 242 c-d. See also O’Brien, ‘Heraciite et I’unite des opposes’. Revue de metaphysique et de morale, 2 (1990), 147-67.

12

Simon Trepanier

seeking to refuse it only a zoogonic role. Basing his interpreta¬ tion on the contention that Love or Kypris is always portrayed as the fashioner of living creatures,^' and accepting the testimonia concerning the cosmological function of Strife, he proposed to correlate each power with a single function. In his scheme, the opposition between separation and unification is maintained, but a complete parity, in which each power is the active cause of the destruction and creation of a world, is denied. This allows Solmsen to make more sense of Aristotle’s testimony and the idea of com¬ plete separation of the elements under Strife, and to have a cycle of sorts, but without a creation of mortals under Strife. There would be a separating out of the elements following the (instantaneous?) disruption of the Sphairos,^^ but this process would produce no mortal creatures, at least until the tide turned back towards in¬ creasing Love, whose fashioning of mortal creatures prepares the way for the Sphairos. Long’s account is closer to Bollack’s, with an instantaneous destruction of the Sphairos, albeit with a firm com¬ mitment to the reality of the cycle. His arguments for it, however, are more interesting, and will be considered separately. Before the publication of the Strasbourg papyrus, upholders of the traditional view had to make their case for the most part from within the confines of cosmological passages. While this did not prove an insurmountable problem (see especially Graham), it limi¬ ted the available supporting evidence, in the first instance, to frag¬ ments of the Physics, as then conceived, and in the second, to the larger-scale statements of doctrine such as fragments 17, 26, and 35. For in the case of Strife, it is in those passages alone, in particu¬ lar the crucial B 17. 3-5, and less directly at B 26. 7, that we find a counterbalance to the overwhelming preponderance of zoogonic activity elsewhere ascribed to Love alone. But now, according to M-P, a number of its readings offer us unambiguous statements of Strife’s cosmogonic function, in particular

a (ii) 3-17, which

they interpret as a depiction of the world under the rule of Strife. Unfortunately, most of these new passages are not as clear-cut as M-P contend, even though I am in agreement with them in hold¬ ing to the traditional view of the cycle. Consequently, I will forgo “ On Love as seemingly the exclusive cause of life see frr. 20, 35, 71, 73, 75, 84, 86, 95, 96, 98. Against all of these are B 17. 3-5, as emended by Panzerbeiter, B 26. 7, B 62, and now ensemble d: see discussion below and sect. 4. In this Solmsen, ‘Love and Strife’, 230, is close to Bollack, but remains ratber obscure on the issue; see further n. 91.

Empedocles on the Ultimate Symmetry of the World

13

discussing here much of this new evidence, with the exception of ensemble

d, in order to concentrate on what I consider to he much

stronger evidence of the cosmogonic and zoogonic role of Strife. This is evidence which M-P do not overlook, but which they fail in my opinion to assess at its proper worth. It operates on a more general level than the individual new readings they cite, most of which are either partially restored or controversial on philological grounds, and instead flows from the unity of Empedoclean doctrine, which—I will argue more fully below—is now decisively shown by ensemble

d.

(a) Fragment 124 For the moment, taking this much for granted, the new develop¬ ment allows us to adduce as direct evidence of Strife’s zoogonic function, against the views of Bollack and Solmsen in particular, lines that few previously would have ventured to take at face value, fragment 124: oj TTOTTOL, a> SeiXov 6vrjTv yeVo?, w SvadvoX^ov, roluiv

eK

t’ iptScuv €k re arovaxdiv iyeveaOe.

O wretched, O unhappy race of mortals, from such quarrels and such cries of woe were you born!

The dramatic power of the passage is undeniable, and its theme had long justified its placement among The Purifications, as a lament for the fallen daimon. But as Plutarch points out, Empedocles does not indulge in metaphor for mere decoration.“ It is entirely character¬ istic for him to bring a dead metaphor back to life by re-establishing its reference, in his view perhaps revealing its true meaning in the spirit of the ancient etymologists. Here, therefore, if we take the lines sensu stricto, as I bdlieve we now can, we have a direct attes¬ tation of the zoogonic role of Strife, for men are unambiguously called the products of eridesd* The abject misery that colours the observation, moreover, does not allow us to take the line as express¬ ing, at the time of utterance, our freedom from the rule of Strife. " QC 683 E 1—7: ‘nor is it his habit, for the sake of a beautiful style, to deck out his subject with the most specious adjectives, like some garish flowers, but rather he gives a specific illustration of every substance or power’. As well as ‘of laments’, arovaxoiv, probable compatriots of the edvea Krjpdiv, fr. 121.

Simon Trepanier

14

In fact, worse is to come, according to Aristotle’s testimony that we live in a world of increasing Strife/’ k (b) Ensemble

d

and the unity of Empedocles’ thought

Such a claim, however, as well as its corollary about the future direction of the world, remains contingent upon the actual integ¬ ration of the physical and ‘demonological’ in Empedocles. Let us now consider how the two intersect in ensemble d, and then how that intersection provides new support to the traditional view of the cycle. I quote the first fifteen lines of M-P’s edition and translation, minus their most debatable restorations (since it is a first edition, many textual questions remain open for debate, most of which I ignore for now); [avJSi^’

o-tt’

dAA'j)Aai[v] TTecre[et]v xo-l

tt[6t]ixov imOTreLv

[770]AA’ de/[Xoy]iJ,6s, draw¬ ing along a ‘mixture of much woe’, presumably from the subjacent region, and the result in d 13 is the production of creatures or mor¬ tals (the word itself, a neuter plural, is lost) which are (/)LiTaA^ta, either nourishing or, as M-P translate, ‘capable of reproduction’. As correctly discerned by the editors, p. 308, the events described are comparable to those in another passage, fragment 62, where the origins of humanity are outlined, and this allows us to fill in some of the details. Specifically, in B 62. 6 we learn that it is the fire remaining in the earth that draws out the ‘whole-natured forms’, as Empedocles there calls these first organisms, while it seeks to ‘join its like’, OeXov irpos opuolov iKeadai. At d 15, however, M-P suggest, p. 314, that this narrative breaks off once more to return to the topic of demonology. This is highly debatable, not least since lines 15 ff. are very poorly preserved, and certainly no direct indication of a transition survives. Though they De abst. 2. 31: ‘All that remains is for them to make amends thereafter through

purihcations [Sid. toiv Kadapixaiv] for their previous sins in nutrition [i.e. for having eaten meat].’

Empedocles on the Ultimate Symmetry of the World make heavy weather of their restored phrase ]efV

17

tottov iaxdTig[v

went to the uttermost place’, at the end of

d 15,

and the

apparent change of subjects that implies, considerations can be ad¬ vanced for doubting their restoration of the text. First and most importantly, the size of the lacuna seems too narrow to accommo¬ date ONB, the three letters postulated by their restoration.If we doubt a three-letter restoration, then there might be the possibility M-P themselves raise, p. 317 n. i, of filling the lacuna with the two letters (l) T, and thereby restoring the line-end as a single word, the very rare adjective eaxaTiajT'i7?.‘FIere it would mean ‘further¬ most’ or ‘on the border’. M—P, however, dismiss this alternative on the ground of its rarity. This has some weight, but then again Empedocles did coin many strange neologisms, to the delight of later grammarians. Moreover, such a supplement would not only suit the metre, but is also a better fit for the lacuna; it might even explain the supralinear note added by the second hand, perhaps as a simple correction to the accusative. As for

d 15

itself, a simpler explanation of its general meaning is

to hand when we observe that fragment 62, used to fill in the back¬ ground of the previous lines, seems to have as its own background a more general cosmological event, known to us from an Arme¬ nian commentary upon Philo of Alexandria’s On Providence, 31 A

Basing myself on pi. v, comparison with two other instances of the letter se¬ quence

ON on

the same page shows that insufficient space is left for their postulated

beta. In the last word of d 3 the OWfrom the extant

ONTAl measures

7 mm. across

(allowing for i mm. to round out the partial omicron), whereas the OAf from d 15’s

TOnON measures

5 mm. Even if we apply the shorter of the two to the lacuna, this

would leave only 2 or at most 3 mm. for the beta and the spacing between it and its two adjacent letters. The beta itself usually measures at least 2 mm. across, e.g. d 6 and d 8, with an average spacing of i mm. on either side between it and the adjacent letters. Here the postulated beta would overlap with the extant eta. Even if one grants M—P’s choice of positing some form of the verb ^aww ‘to go’ for the last syllable of the verse, then it is still far from clear which of two possible readings, both attested in the papyrus, is the better one. If the line ended with such a verb, then the first hand probably wrote ji]-^ ‘[sc. something or someone] was going’, third person singular. But to this the second hand then added, over the line, an eta and a nu between two dots, indicating either a variant reading or a correction; cf. M-P’s discussion of textual annotations, pp. 20-5. Thus, only if we first posit a verb, despite the obvious bad fit in the lacuna, and then choose the second hand’s reading, can we arrive at /Sjij'v' ‘I was going’. As should be evident by now, this is hardly solid evidence for a change of subjects. Finally, M-P’s decision to follow the second hand here seems all the more capricious in view of the fact that they elsewhere go out of their way to reject the second hand’s corrections, many of which turn on a nu, as in the case of the three thetas (see sect. i).

Simon Trepanier

i8 49a

There we are told how, in the formation of the cosmos,

first the aither, then fire, sprang out of the primordial mixture, leav¬ ing earth and water below. Basing myself upon that, I suggest that we can better understand

d 15

by positing that the passage remains

cosmological, more specifically,that fire is still the line’s subject, and that the reference ‘to the uttermost place’, perhaps reading ]eiV TOTTov €axaTi(n[T]ri'v', describes how the fire remaining in the earth is seeking to join its like, most of which has already ascended into the heavens.'^'* Finally, while Strife is no more named in this second half of enseynble

d

than in the first, given that what drives the emergence

of mortal creatures in

d 13

is most likely a process of separation

(cf. B 62. 2 KpLvojjLevov TTvp), it is no great leap from there to infer that Strife must be the force at play behind that process. Further still, I think that this new passage, with its link to fragments 124 and 62, not only confirms Strife’s activity as a zoogonic power, but also provides us with a fuller characterization of that creation as a sorrowful event: fragment 124 tells us in so many words that men are products of Strife; the mixture that emerges from the primeval mire at

d 12

is iT[o]XvTnppi[ov]a ‘much-suffering’; and the opening

line of fragment 62 describes men and women (taking the adjective diTo KOLvov) as ‘wretched’ or ‘miserable’, dvSpojv re rroXvKXavTcov re yvvaLKiov.

Thus, to return to the question of how to relate the two halves of ensemble

d,

it is in this last regard especially that we can see

a connection: the passage, though articulated, is one continuous meditation upon Strife. This is what accounts for the general pes¬ simism we find in both halves, and even hints at what must have been Empedocles’ general train of thought, or at least its dramatic portrayal, in the missing broader context. Ftaving perhaps initially undertaken to describe a zoogony of Strife, and thus finding himThe passage now appears to have been based on a quotation rather than a doxographic account: see Inwood Poem, fr. 40. This is not the place for a full discussion of the text, but to further support my interpretation, I note the close verbal similarity of d 15 to B 35. 10, which describes Strife’s retreat in’ iaxara ripfiara kukXov ‘to the uttermost limits of the cycle’ (or again, fr. 36). Once more, that this ‘uttermost place’ is the vault of heaven is directly stated in A 49a (cf. also Lucr. i. 73 ‘flammantia moenia mundi’), and I remain to be convinced that the form of the verb as written by the second hand is either correct or all that significant (see n. 42). As for the remains of lines d 16—19,

printed

above, they are slim indeed, and offer little evidence either way; if anything, ‘earth’ in the nominative at d 19 supports a cosmological context.

Empedocles on the Ultimate Symmetry of the World

19

self brooding upon its work as an agent of separation, Empedocles was led on by such reflections to rail against his own thraldom to Strife. The emotional power of this cri de coeur is second only to the lament of the ‘exile from god’ in fragment 115, and indicates the strong dramatic nature of the work. For my immediate purpose, the most interesting aspect of this link between the two parts of ensemble d is the way in which it lays bare the underlying unity of Strife and its activities. For, if I am correct that the whole passage is concerned with Strife, it displays Empedocles passing seamlessly from the physical to the ‘demonological’, or even more telling, to the personal. The transition is so seamless, in fact, that it reveals the extent to which such distinctions, implied by the terms ‘physi¬ cal’ and ‘demonological’, are impositions upon this material, rather than something felt by Empedocles.

(c) The zoogony of Strife, Empedoclean pessimism, and early Greek eschatology From these considerations I move back to the general debate over the nature of the cycle, and the attempt to negate a zoogonic func¬ tion for Strife. First, as has just been seen, such a view is directly denied by fragment 124, and implicitly denied by fragment 62, while ensemble d links the two, intermingling the mechanical with the moral. By its own ties to fragment 62, ensemble d refines the physical picture of how Strife’s influence causes living things to arise, which in turn demonstrates how the pessimism of fragment 124 can be more than a figure of speech. Second, on an even wider level, that of the general ethos projected by the work, it is possible now to see that the dual cosmogony is required to make sense of Empedocles’ pessimism about our current situation. How is that so? If Strife’s function were limited only to an initial destruction, as in Bollack, for instance, we must take it that the tide of destruction is presently receding, and that, qua living, you, I, and all living creatures are in the clear for now, even probably back on the path to the reign of Love. If one follows Solmsen in turn, the contradic¬ tion is equally acute, since we ‘come into being’ from Strife only as opposed to the Sphere or the One, and the more proximate cause of our being is Love. Our present condition might be relatively bad, in comparison with what it may become, but unless we imagine Strife regaining the ascendant, which amounts to reverting to the

20

Simon Trepanier

traditional view, there does not seem to be cause for such despon¬ dency/^ Against this possibility, once again, there is the testimony of Aris¬ totle, who quite plainly tells us that we live, now, under increasing Strife, evrt

tov

NeUovs, GC 334^5 (a reading I will defend below in

Section 4) and, even more significant, the fact that this testimony, which Aristotle advances on an unrelated point, squares perfectly with the pessimism characteristic of Empedocles’ depiction of our current situation. Such pessimism simply does not seem compat¬ ible with the idea of humanity’s imminent return to the Golden Age of Love, any more than Hesiodic pessimism allows for the possibility of putting all the world’s ills back into Pandora’s box. As correctly discerned by M-P 346-7, the whole importance of en¬ semble

d

for the general interpretation of Empedocles is that it shifts

the burden of proof onto those who would maintain that the two poems—if there were in fact two poems, O heresy!—display two difTerent world-views and, I would add, that Empedocles confined his pessimism to one of them."** To widen the perspective still more, if one considers the broader protreptic and missionary goal of Empedocles’ message, then the picture of a rising tide of Strife is what must have given it its particu¬ lar urgency and rhetorical force. Empedocles, I take it, in teaching mankind to cultivate Love, was offering to the elect few their only chance of salvation, not prophesying the inevitable salvation of all and sundry. Indeed, if, as the asymmetrical reading of the cosmic cycle implies, we are already on our way back to the reign of Love, what grounds could there be for the pessimism outlined above, or what need could there be to turn mankind away from Strife? Sooner Of course, the doctrine of the cycle necessarily ensures the eventual return of Strife, as it does all other phases, so that it would be possible to speak of any of them in the future tense, or in the past for that matter. On such grounds it could still be possible to stretch the meaning of the passage to accommodate Solmsen and Bollack’s views, but only by way of the odd omission of the world of Love which would precede it. By far the more straightforward reading is the one that takes increasing Strife as the frightening and unavoidable future of this, our current world. Solmsen’s remarks at ‘Love and Strife’, p. 242, are unfortunately typical of this earlier stage of the debate. After considering fr. 62 in support of the dual zoogony, he says: ‘The atmosphere of the fragment is not particularly cheerful, and TToXvKXavros . . . evokes associations with Strife rather than Love.’ He then pulls himself short: ‘However, in On Nature Love and Strife are essentially physical agents. The idea that under the rule of Love there must be a life of perfect bliss is not to be found in this poem (but is imported from the other).’

Empedocles on the Ultimate Symmetry of the World

21

rather than later, everyone would be coming round to Love. But once again, only if things stand to get worse than they now are does Empedocles’ message have the desperate urgency its tone betrays. As I imagine it—but it would take more than I can say here fully to substantiate the claim—Empedocles’ wider message was that by cultivating Love one could ascend the scale of being, and so ultimately become a god. However, it is crucial to note that these Empedoclean gods will have remained part of creation, and would not escape the world’s hnal ruin: they are merely long-lived, hoXiyalovcs: cf. B 21. 11, 23. 8, and now a (ii) 2/’ Thus, becoming an Empedoclean long-lived god might offer us a salvation of sorts, but it would be at best a reprieve, until the end of this world. Against this one might still object that the picture I present is too pessimistic, since it commits Empedocles to a belief in the eventual destruction of everything except the six first principles, and thus undercuts any possibility of eternal salvation. This is not the place for a full exploration of Empedocles’ eschatology, but I would nevertheless offer the two following considerations to disarm that objection. First, the promise of eternal salvation is by no means the only form of afterlife survival entertained in early Greek literature and philosophy—witness Cebes’ position in the PhaedoW If anything, it was rather Plato who made this an issue. The objection that the culmination of Strife’s work denies eternal salvation is thus true, but of quite limited force. Second, to consider later philosophical views, two prominent later cyclical accounts of world history also seem to understand that cyclical time renders salvation, if at all possible, then temporary at best. Plato, for instance, in his own variation on the theme at Phaedrus 248

E,

does no better, since at the end of the ten-thousand-year cycle, no matter how good or bad (immortal) souls have been, everyone begins again at square one,. Among the Stoics, in turn, can be found the belief that only the souls of the virtuous survive the separation of the body, but even they can do so only until the ekpyrosisW All told, then, the unity of Empedocles’ thought, as confirmed by ensemble d, enables one to assemble, out of fragments 124, 62, and For a different reading of Empedoclean pessimism, as it relates to the cosmic cycle, see Osborne, ‘Recycled’, 34—48. See Inwood, Poem, 52-65, on the affinity between the Pythagorean views of Cebes on the soul and Empedocles’ description of the daimon. A. A. Long and D. Sedley, The Hellenistic Philosophers (2 vols.; Cambridge, 1987), vol. i, no. 53W (Eusebius, PE 15. 20. 6).

Simon Trepanier

22

ensemble

d

itself, a single and coherent picture of the cosmogony

and zoogony of Strife. Both of these are required ex hypothesi on the traditional view of the cycle, but in the previous state of our evidence they could still be denied, since opponents could set some of the evidence aside, e.g’. fragrpent 124, while refuting one by one the other passages claimed in their support. Now, however, the unity of Empedocles’ thought disallows the first, while inviting a reconsideration of the passages at the centre of the old debate, to which I now turn.

4. Other passages: B 17. 3-5, B 26. 7, etc. The majority of these passages have been left until now because a number of them, occurring as they do in the context of more general statements of doctrine, have a density of expression verging upon the obscure and, considered in isolation, they often prove incapable of yielding a single obvious meaning. (a) B 17. 3-5 The most important of these by far is B 17. 3-5: Soil)

§€ 6vrjTd)v yeveoLS, Solti

S’

aTroXeuliLS'

Trjv (ixev yap rravrcov aiivoSos' tIktsl t’ oXeKSL re, rj

Se

TrdXiv

Sia^uopieVojv

OpeL^delaa Sie-nTT].

And double is the generation of mortals, double their falling-away; for the communion of all things both begets and kills [mortal life] and as [all things] grow apart again, [mortal life] is nurtured, then vanishes.

These lines, as first emended by Panzerbieter in 1844, are the cor¬ nerstone of the traditional view of the cycle. 8L€TTTr] for the manuscripts’

By restoring Opecjideiaa

SperrTrj (SteVxTj Scaliger), Pan¬

zerbieter not only gave much greater clarity to the passage, but he also became the first modern scholar to advance a view of the cycle containing a double cosmogony and zoogony. Other readings have since been suggested, but none with success, and I will accordingly base my discussion on the lines as they stand in DK.®' Before the publication of the Strasbourg papyrus, the best case O’Brien, £CC 157 ff. Holscher ‘Nachpriifung’, tried to revive

SieVrij, advanced by von

Empedocles on the Ultimate Symmetry of the World

23

in print for the alternative line of interpretation was that of Long, and in what follows I will focus my criticisms on his interpretation. Although Graham has already capably refuted it on more formal grounds, the new evidence makes Long’s interpretation, and the tradition for which it stands, simply untenable. Long endorses Bollack’s general scheme for the cycle: (i) the Sphere; (2) its instantaneous shattering by Strife; (3) the weaken¬ ing of Strife; (4) the increasing reign of Love. Most controversially, phase 3 would be characterized by Love’s gradual reunification of the scattered particles of matter into the four internally homoge¬ neous elements.’^ Phase 4 would see the first mixtures of elements into living beings, culminating in the Sphere. Long produces this general scheme through two key interpreta¬ tive moves. His first and most important is to question the actual subject under discussion in B 17. 3-5. He argues, p. 403, that the essential point of the whole 35-line fragment 17, as it then stood, was to describe a universal rule of alternation between ‘One’ and ‘Many’. This rule has an equally universal domain of application; it refers to ‘all things’. And what are these? For Empedocles, Long ar¬ gues, they amount fundamentally to the six first principles, namely the elements plus Love and Strife. All that is consistent enough with my own understanding of the passage. The question then is whether or not the general message of the passage also applies in lines 3-5. According to the traditional interpretation, the lines form something of an aside. Their subject is mortals, dvqrihv, so named by way of contrast, Se, to the first exposition of the cycle in lines I and 2. Those lines, as can be ascertained by reading back into them the content of the subsequent lines of fragment 17 (except, of course, lines 3—5), do have the six first principles for their subject; see e.g. B 17. 18-20. Long, however, says that this distinction is merely apparent. Since ‘all things’ are in the end Love, Strife, and the elements, these ‘mortal things’ are not to be contrasted with an Armin in ‘ Weltperioden’. Bollack proposed

8pv(f>drjeta’ aTrohpvTTTei

‘dispersant, se dis¬

perse’: see his discussion at iii. 52-6. Wright, Extant Fragments, fr. 16. 3-5, obelizes the manuscript reading and discusses various proposed emendations, without taking a side; Inwood, Poem, fr. 25. 3-5, follows DK’s text. For a longer discussion and critique see O’Brien, ECC 164-7, arid M-P 76-7. ” Long, ‘Sixties’, 407, following Bollack (?), seems to understand the unification of the elements as driven by Love; in the traditional view this is done by Strife, and is an act of separation. For some further criticisms see Furley, Greek Cosmologists, 102-3.

Simon Trepanier

24

immortal order, for they are in fact part of it. The six basic entities are thus also these mortal things, and the true subject of the lines is once again the alternation of one and many. k

(i) Reductionism. Is such a reading, which I think it not unfair to brand reductionist, true to the text? I think not, and now the con¬ tinuation of fragment 17 into ensemble

a

quite clearly refutes it. Of

the new information it contains, the most telling in this regard is the last section,

a (ii) 24

and

30,

where Empedocles promises to show

the disciple the ^vvoSov re SiaTTTv^lv re yevedXrjs ‘the coming-together and development of birth’, or rather, taking yevedXrjs as a genitive of definition, ‘the coming-together and development that are birth’. Specifically, Empedocles says he will demonstrate this definition by means of examples taken from the natural world, namely plants and animals.

These new lines thus seem to confirm two things at once.

First, they confirm the accuracy of understanding B 17. 3-5 as a reference to mortal creatures such as plants, animals, and men, not to first principles. Second, they also reinforce a more philosophical point in 17. 3-5, the ambivalence of Love and Strife with regard to mortals. The first point seems beyond debate. The terms employed are very close: ^vvoBos is identical with its predecessor, and SidirTv^ts is a good match for Sta^uo/revcov. Once more, the unification-separation motif, mentioned above, is maintained, this time within a single line. Reading

a (ii) 24-30

back into B 17. 3-5, as we now can, vin¬

dicates the view that Empedocles’ point in the earlier passage is to compare and contrast the first principles and the world of mortals. Now of course Long is correct to observe that the world of mor¬ tals, or rather mortals themselves, are made up of first principles.®^ Where he goes wrong is in supposing that this legitimizes reading Empedocles as a kind of reductionist avant la lettre, or commits Empedocles to the view that the first principles are the only real or stable objects for scientific discussion, perhaps along the lines ” So understood by M-P 247,1 think correctly. For a similar genitive of definition, compare the expression NetKeos exdei at B 17. 8, i.e. ‘the hatred that is Strife’. The biological term ‘development’ seems preferable to ‘unfolding’ as a translation for hiaTTTV^is, as when we speak of ‘the development of life’.' On ^/avvoSos see M-P 161. In ‘Sixties’, 403, Long wrote: ‘It should be stressed that in fr. 17 Empedocles says nothing specific about “the world” or animals, plants, men, etc.’ ” That all things come from the first principles is asserted in a formula, with variations, at B 21. 9-12 and a (i) 8-(ii) 2, and most memorably in the simile of the painters, fr. 23.

Empedocles on the Ultimate Symmetry of the World

25

of the ancient atomists, where such a move would be much more valid. The point may seem unduly recondite at the moment, but I will show below how it has been put to work by Long and others in unpacking the meaning of B 17. 3-5. At the time it was made, moreover, this identification seemed all the less problematic be¬ cause fragment 17, as it then stood, could arguably be understood as dealing solely with a more abstract level of speculation. But the entire sequence B ly + ensemble a now demonstrates how its central theme of alternation was pursued beyond that initial level of high abstraction, into an account of cosfnogonic phases and zoogonic activity. At the end of that sequence, terms similar to those used to describe the universal alternation of the cosmos are employed to introduce a zoogonic account. This shows that the formal expres¬ sion of the motif has no e.xclusive application to any one subject or level of generality, nor does Empedocles makes any implied con¬ trast between the reality of the elements and the non-reality of their mortal compounds. In contrast to Parmenides, for Empedocles the world of living creatures is just as real as the elements, and can be referred to alongside them, or contrasted with them.^® (ii) The ambivalence of Love and Strife. This leads to my second point, the ambivalence of Love and Strife with respect to mortal life. With it I leave behind Long’s interpretation to work my way back to the exact meaning of B 17. 3-5 according to the traditional view. The new terms ^vvoBov

re StdvTTu^tV re yevcdXrjs,

as I take them,

refer to, or rather programmatically announce, the zoogonic effect of both unification and separation. Although to some it may be tempting to see in the second term a reference to destruction rather than a creation of life, by simply equating unfolding or develop¬ ment with death, and coming-together with life, I think that such a temptation resides in an over-simplified equation of mixture with life and separation with d'eath. Once again, line 3 specifies that the double production and falling-away is that of Ovrjrd. If we simply equate mortals with mixture, it is difficult to see how the avvoSos TTo-vTcov can destroy this condition; indeed it can only enhance it. Long’s second move is to argue that the aorist of line 5, SUnTt), which on the traditional view is understood as an atemporal gnomic aorist, should in fact be taken as indicating the priority in time of the process it describes, by way of contrast to those described in the present tense in line 4. Accordingly, there would be a single sequence, rather than two alternating worlds: the rise and demise of Strife would be past events, whereas Love would be active at the moment. But since the new evidence denies his first move, there is no need to refute the second.

26

Simon Trepanier

The answer must be that Empedocles understood mortals not as mixture alone, nor as solely an interaction of Love and the elements. What else, then, must they also be? What else but Strife? Only thus must the coming-together of all'things kill: it destroys the existence of individual creatures By sub.suming them into itself.’’ In doing so, it simultaneously dispels Strife, that which had given them their separate existence, when it drove them apart from one another and, ultimately, from the unity of the Sphairos. (iii) The syntax of B ly. j-5. There remains the question of the actual wording of the lines, and how the general scheme I have defended on the basis of more explicit passages can be had from them. The three lines, according to the traditional view, are some¬ thing of an excursus within the broader exposition of the cycle. If that is the case, as I believe it is, then it is not unfair here to speak of Empedocles as overtaxing the intelligence of the most willing and capable listener. For he has hardly begun to develop his gen¬ eral thesis, in B 17. 1-2, concerning what we only later learn are the six first principles, before turning to the corollary of its effect on mortals. What appears to have set him on this track is the an¬ nouncement of the double tale of the cycle, at the very outset: StVA’ ipeoj. For to that double or two-fold story he apparently wants to contrast the double coming-into-being and falling-away of mortals. To my mind, the lure in that was the opportunity it afforded him to repeat the exposition of the cycle, through the term rravrcov avvoSos and the participle StafoopievoDv, while also explaining the nature of mortal life. This seems as likely a candidate as any to be one of those ‘splendid poetic conceits’ Aristotle attributes to Empedocles.’* The main difficulty of the passage is exactly how to understand the antecedents of rrjv p-ev of line 4 and 7) Se of line 5. Perhaps the more natural way of construing the syntax, but not the one I will defend, is to correlate Trjv pev with yeveois and 17 Se with aTToXenjns, and this is chosen by some.’'' But such a reading proSee Simplicius, In Phys. 33 = Inwood, Poem, ctxt 45c. Such a view seems im¬ probable to us, as it did to Aristotle in the case of limbs, GA 722*’6—30, but it has something of a parallel in the modern theory of the evolutionary appearance of multi-cellular organisms from previously autonomous single cells. The translation is O’Brien’s rendering (ECC 322) of Arist. fr. 70 Rose

Kat

rotj

aAAoiy rots Trepl noirjTiK-rjv eTnTevyp.aaL xpcup-evos, which more prosaically reads ‘and making use of the many other inventions of poetry’. Bollack, iii. 52-7; for a good review of previous interpretations see Mansfeld, ‘Ambiguity’.

Empedocles on the Ultimate Symmetry of the World

27

duces a curious statement, which would run something like this: ‘for the coming-together of all things gives rise to the coming-intobeing of mortals and then kills the coming-into-being of mortals, whereas as [all things] grow apart, the falling-away of mortals, be¬ ing nurtured, then dies.’ Thus, in line 4 we would be told that the coming-into-being of mortals has its beginning, then ends: that is, the mortals themselves, after coming-into-being, do not cease being, they only cease coming-into-being. Line 5, by contrast, would de¬ pict the cessation of their existence. The term ‘the falling-away of mortals being nurtured’, rj 8U . . OpdfOelaa, would describe the on¬ set of their march towards death, i.e. the coming-into-being of their coming into not-being, while SieWry, lit. ‘flew away’, would denote the cessation of that process, the end or ‘death of death’. Taken that way, the two second deaths described, the object of line 4’s oAewret and the subject of line 5’s

are purely abstract or relational.

Neither involves the actual destruction of anything. We can see thereby why such a reading was attractive to some opponents of the double cosmogony: it is one way of reducing the second set of alternations to semantics. But the reading has further serious problems. For one thing, it implies that the process into death or separation is of equal duration to that of coming-intobeing, which is a serious complication for those adherents of the single-cosmogony view, who assume that Strife does all its work of destruction in an instant.*" More damaging to it is the need, in order to make it intelligible, to impose a strong distinction between the diflferent kinds of life and death meant by terms which appear, allowing for some degree of imagery, to have been employed as synonyms by Empedocles.*' The alternative construction, which has found more adherents on both sides of the controversy concerning the cycle, and which I also adopt here, is to correlate each

yeveai?—aTroAeti/a?

pair with the

verbs in each sentence.*^ This makes for a better formal responsion: For Bollack in particular, who proposes (iii. 55) arbitrarily to interpret BvTjTwv as the subject of

Given his dismissal of Panzerbieter’s emendation,

analysis of his general interpretation would require more discussion. For Solmsen as well (‘Love and Strife’, 231), the death of the Sphairos is short and the power of Strife already now past its apex, hence there is no symmetry between the growth and decay of the Sphairos. Mansfeld especially (‘Ambiguity’, 28-9) takes the discussion to the level of semantics rather than cosmology. Upholders of the single cosmology who construe it thus include: Solmsen ‘Love and Strife’, 244; Ffolscher, ‘Nachprtifung’, 202; Long, ‘Sixties’, 405; van der Ben,

Simon Trepanier

28

the double yiveais would correspond to TtVret and dpecfyOeiaa and the double a-TToXenjiis to oXeKei and 8i€77Trj. The difficulty then is in spec¬ ifying the actual reference of the two demonstratives rrjv fjcev and rj 8e. The favoured solution has been to consider the accusative of the pair as a cognate, th'e nonjinative being a passive form of the same construction.*^ r-pv and rj would then have as antecedents both yeveoLs and aTToXenjjiS, depending on the verb, the important point being that the destruction and creation in each case would be that of actual mortal creatures, or ‘mortal life’ in general, not some abstract concept like the coming-into-being of coming-into-being, as assumed in the construal considered above. Indeed, taking the construction as cognate allows a healthy reduction in the passage’s level of ab¬ straction. Thus TLKTet and oXeKCL describe effects upon actual mortal creatures, or upon the totality of such creatures in each zoogony. By construing the passage in this manner, moreover, we not only achieve a better formal integration with line 3, but we also have a clearer life-death polarity: instead of contrasting two types of genesis and two types of falling-away, as above, each is paired with its more conventional opposite. The simplest and most elegant understand¬ ing of the lines, therefore, is that which supports the traditional view of the cycle. {b) B 26. y The same scheme also offers the best understanding of the prob¬ lematic line 7 of fragment 26. DK’s text of B 26. 3-7 is as follows: avTO. yap eariv ravra, St’ dXXrjXajv Se Oeovra yLyvovT(ai) dvOpw-noi re teat dXXojv edvea Orjpwv, dXXore p,€v ^tAoTTjrt avvepyop-ev’ els eva Koapov

5

dXXore S’ av Sty’ eKaara (j)opovpeva NelKeos dydei, elaoKev ev avvv Travr(a) oaa r' rjv r’ oaa r’ iarll) oaa r’ earai

OTrtaaw. The papyrus now confirms the specific form of Aristotle’s quotation by

reproducing it at a (i) 8-(ii) 2, and proves the independence of the two citations as original Empedoclean variants of the same formula: see M-P 175-86. ” The question in fact pertains to all of Empedocles’ statements involving the generative powers of the first principles. The debate reaches back into antiquity, where from Aristotle on commentators have been tempted to treat Love and Strife as formal or moving causes. More recently, see DK’s note on B 26. 2, or Wright’s comments. Extant Fragments, 182-3, who argues that only the elements are meant here.

Empedocles on the Ultimate Symmetry of the World

35

to us from Simplicius, but which is identical to that in lines a (i) 8a (ii) 2 (for details see n. 71), establishes that its use in a (i) 8-a (ii) 2 applies only to a cosmogony under Strife. But this is only ap¬ parent to them because they believe that a (ii) 3-17 describe the dissolution of the world under the reign of Strife. In ‘“We” and Empedocles’ I criticize the latter claim at greater length, but here I can at least take issue with the dubious manner in which Aristotle’s remark is pressed into service. In the first place, it is only the narrow context of that section of Aristotle’s discussion which establishes a putatively exclusive link to Strife, whereas the verses quoted by Aristotle use a plural antecedent to refer to the things out of which all else but the Sphairos arises. By no means need this commit Empedocles to the view of Strife as sole creator of the cosmos. In fact, the continuation of Aristotle’s discussion shows that his intent throughout that passage (iooo‘*24-^2o) is to make the same point I have been pursuing throughout these pages, namely that Strife, despite being Strife, is also in a sense creative, while Love, despite being Love, also destroys. M-P are entirely right to recognize that Aristotle’s discussion supports the dual-creation interpretation of the cosmic cycle, but wrong to suggest that when Aristotle there quotes the zoogonic formula, he reproduces a passage where Empe¬ docles meant us to understand that Strife alone was the active agent of creation. To reinforce my general claim about the zoogonic formula, I turn now to fragment 23, the simile of the painters, and another instance of the formula. This passage, which Simplicius, In Phys. 159.27 Diels, tells us Empedocles gave as a ‘clear illustration’ of the capacity of the first principles to generate all things, describes two craftsmen preparing various colours from a limited set of primary ones. To obtain the variety of colours they wish, the two painters mix the colours, but not all of them in the same proportions, B 23.4: apfj.ovlrj /xel^avre ra jjuE nXeai, aXXa 8’ iXdaao).

both [painters] mixing them in harmony, some more, some less.

If the first part of the line, describing mixture through harmony, must be understood as the agency of Love, the last three words ad¬ mit a function, not to say a creative role, for Strife, since it specifies the need, in mortal creatures, for certain parts to be less mixed. Still more interesting is the manner in which the two craftsmen are portrayed as equal partners in production, simultaneously exerting

36

Simon Trepanier

their powers ‘to set down [/cTt^^ovre] trees and men and women and beasts and birds’, etc. k

(e) A last complication: fragment 22 While those passages show or imply Strife’s active presence among mortals, a much more direct assertion of its generative capacity is perhaps to be had from the last line of fragment 22, comparable to the directness of fragment 124. For the sake of intelligibility, and because of other relevant matter it contains, I quote the whole fragment: apdiJLia nev yap ravra iavrcbv ndvra p-epeoaLV, rjXeKTCjp re

re Kal ovpavos rjSe daXaaaa,

oaaa (j)iv iv dvrjrolaiv aTTOTTXax9evra rrefvKev. CO? S’ avrojs oaa Kprjaiv inapKea p,dXXov eaaiv dAAiyAot? earepKrai opLOiojdivr' AfpoSlrr]-

5

ix^pd. piaXiaA {oaa) -nXelarov an’ dXXrjXwv Siexovai yevvrj re Kp-qaei re Kai eiSeaiu €Kp.aKToiat, ndvTrj avyylyveadaL arjOea Kai p.dXa Xvypd fveiKeoyevve'arTjaLvf on ac^iLOL fyevvav opydf

For all of these are well fitted to their own parts, sun and earth and sky and sea, as many as, having strayed from themselves, have grown in mortals. Similarly, those which are more inclined to mixture are attracted to one another, made alike by Aphrodite.

5

But those which are most different from each other are hostile to birth and mixture and fashioned shapes, being entirely unaccustomed to come together and sorrow’ful at the bidding of Strife, in that it brought about their birth (?).

The passage is problematic, not only because of its uncertain word¬ ing and syntax, but also in terms of its reasoning, wFich seems pe¬ culiarly circular.’'* All the same, the pertinence of the last line to my general thesis, even if it is particularly ill-preserved, is that it directly applies the vocabulary of birth and creation to Strife. The garbled text of line 9, which cannot stand as transmitted, DK tentaText from Wright, Extant Fragments, fr. 25. ’■* On one level, Empedocles seems to want to account for the differences between the elements’ varied natural capacity for mixture as the result of their being produced by Love or Strife. This is unhappy because it might deny the elements a nature independent of Love and Strife. Conversely, and this is where confusion sets in, he also appears to assume their independent nature in order to explain their resistance to or acceptance of mixture. See Stokes’s discussion. One and Many, 168-73.

Empedocles on the Ultimate Symmetry of the World lively restored to:

NeiKeo^ irveoLrjaiv,

ort

oftac yivvav eopyev

37 (?). But

note now that this ‘birth’, which regardless of textual corruption can certainly be had from the last line, is that of the elements. Does this not drive us back to the kind of dual aspect or semantic ana¬ lysis I sought to refute in my analysis of B 17. 3-5? Is Empedocles not at bottom just echoing the Heraclitean theme mors alterius, vita alterius'E^ I think not, or not simply. An Empedoclean mortal, according to the interpretation de¬ fended above, requires the presence of all six first principles. Love mixes and holds the elements in us together, but a certain amount of Strife is still required: otherwise mortals would be internally homogeneous, which we are not, and nothing would maintain the boundary between one mortal and the next. That must also be how Love ultimately ‘kills’ mortals, by driving away the residual Strife in our constitutions, that which maintains us as discrete individ¬ uals.'^'^ Conversely, when Strife breaks up a previous ‘unity’ into several parts, however these be understood, it is also bringing these parts ‘to life’, or at least into some kind of separate existence. Mor¬ tal life for Empedocles, once again, is a middle state between two boundaries: separate elements | mortal life | complete unity of the elements. Keeping that general scheme in mind, I suggest that fragment 22 is concerned not with the issue of whether Strife destroys or creates mortals, but with a deeper question, namely whether or not such destruction is an altogether bad thing, at least considered from the eternal standpoint of the elements. As he does openly in fragment 9, I think that Empedocles here is questioning mortals’ exclusive association of life with mixture, and the consequent folly of understanding separation as an evil. There is also, one might say, a life of the elements, which are ‘born’ through separation. And despite the language of hatred and reluctance at the end of the fragment, that we can understand such a birth as a creation of ” ‘The death of one is the other’s life.’ Compare Heraclitus’ much terser formu¬ lation at 22 B 62 DK, ‘Immortals [are] mortal, immortals mortal, living the other’s death, dead through the other’s life.’ The notion that Strife bounds us is not explicitly stated, but seems assumed at e.g. fr. 36, frrjm Stob. i. 10. 11 WH, tojv be ovvepxoixivov e’f eax'iTov lararo NeiKOs ‘and from them as they came together Strife stood out upon the outer edge’. As usual in Empedocles, this could be appropriate as a description of either a cosmic or a microcosmic biological process. If the latter, it would be clear support for the idea that Strife bounds us off from each other.

38

Simon Trepanier

sorts is reinforced by the application to the elements of the phrase (f>iv . . . aTTOTrAax^eWa earlier in the fragment, at line 3. To speak of the elements, when in a mortal, as ‘having strayed from themselves’ is a symmetrical inversion of the image at B 20. 5, where the limbs ‘wander each separately on the^shore of life’, uXa^eraL av8ix’ e/caara vepl prjyjXLvi jSioio. This, however, goes beyond the ambiguity of

Strife as both destructive and creative of mortal life, and questions rather the applicability of the life-death opposition to the eternal elements. Accordingly, if one should recognize that Empedocles in this passage is indeed developing a theme from Heraclitus, it is crucial to see that its application concerns only the elements. For the elements themselves, gods that they are (cf. fragment 6), there is only one boundary, that between mixture and separation. This, once more, is the basic point of the distinction Empedocles makes between them at B 17. 3-5: the coming-into-being and falling-away of mortals is double, by contrast, that is, to the single boundary the elements cross. When he speaks of the elements, Empedocles consciously cultivates a systematic equivocation in his application of the life—death polarity.’’ (/) Fragment 2 of Philolaus To round off this account of the general nature of Empedocles’ cosmic cycle, I invoke a comparison with the system of his near¬ contemporary Philolaus.’* Doing so, I hope, will help illustrate the viability of the interpretation I have defended, both as a plausible historical account and as one that allows Empedocles more philo¬ sophical insight than he is usually granted. By the standard chronology only a generation younger than Empedocles, and confronted by many of the same problems be¬ setting Greek cosmologists after Parmenides, Philolaus proposed a philosophical system that bears a number of points of resemblance ” There is a parallel here to the equivocation found in the doxa of Parmenides: see A. P. D. Mourelatos The Route of Parmenides: A Study of Word, Image and Ar¬ gument in the Fragments (New Haven, 1970), ‘The Deceptive Words of Parmenides’ Doxa', 222-63, abridged and reprinted in Mourelatos, Pre-Socratics, 313-49. In Parmenides, this equivocation is the tell-tale mark of the t/ojca’s falseness, whereas, on my reading of Empedocles, it has the more positive function of denoting the inapplicability of the life-death polarity to first principles: no piece of fire, water, etc., qua eternal, can truthfully be said to live or die, but they do enter and leave mixture. ’* The basic treatment in any language is now that of Carl Huffman, Philolaus of Croton, Pythagorean and Presocratic [Philolaus] (Cambridge, 1993).

Empedocles on the Ultimate Syynmetry of the World

39

to that of Empedocles, although these have not generally been much noticed. Perhaps the most significant difference between the two is that the fragments of Philolaus betray a more abstract and method¬ ologically explicit approach.Fragments i-6, the most interest¬ ing of the corpus, lay out his basic principles and epistemology. In these fragments Philolaus argues for the thesis that the most basic principles which underlie the world are what he calls ‘lim¬ iters’, vepalvovTa, and ‘unlimiteds’, aneipa (and some third thing, harmony, which brings them together), and from these the cosmos is generated. Perhaps the most significant difference between the two systems is the ontological primacy that Philolaus grants his ‘limiters’ and ‘unlimiteds’; in Empedocles the boundaries and con¬ tinuities found in nature are produced by the interaction of Love, Strife, and the elements. Nevertheless, this very reversal shows in what way Love and Strife do bear a certain resemblance to Philo¬ laus’ unlimiteds and limiters. Love being the cause of homogeneous and continuous masses, i.e. ‘unlimiteds’, and Strife bringing out the internal divisions and distinctions of the world, i.e. the ‘limiters’. Beyond that, Philolaus’ harmony, the thing that binds ‘limiters’ and ‘unlimiteds’ together, also bears an obvious resemblance to Empedocles’ Love, but in other regards the two are not so readily comparable. Most important, and the point of my comparison, is that, just as Philolaus openly argues that both ‘limiters’ and ‘unlimiteds’ are required to account for the world order and the things it contains, so should we understand Empedocles to have assumed the necessity of both Love and Strife to produce the cosmos. I quote the first three sentences of fragment 2 of Philolaus, followed by Huffman’s translation: (l) avayKa to. eovra eipev ndvra rj -nepaivovTa rj dneLpa rj Trepalvovrd re Kal dneipa' (2) dneLpa 8i povov ovk aei’"^ (3) errel rolvvv palverai ovr’ Ik nepaivovTOjv eovra ovr’ e^ arrelpojv rravrojv, SpXov rdpa on e/c rrepaLvovrcov re Kal drTelpcjv o re Koapos Kal rd ev avrip avvappoxBrj.

(i) It is necessary that the things that are be all either limiting, or unlimited, or both limiting and unlimited, (2) but not in every case unlimited alone. (3) Well then, since it is manifest that they are neither from limiting things alone, nor from unlimited things alone, it is clear that the world-order

” Huffman,

51-2.

*“ DK and others prefer to supplement the line to read arreipa Se povovly nepatvovra povov) ov Ka etrp, but see Huffman’s discussion, Philolaus, 102-9.

40

Simon Trepanier

and^the things in it were fitted together both from limiting and unlimited things.

Here, I think that Philolaus’ explicitness helps tokbring out features that remain implicit in Empedocles’ account. Specifically, it makes clear the main feature of the interpretation defended above, viz. that Love and Strife are both necessary conditions for the existence of a cosmos and of mortals. Indeed, both counterfactual states en¬ visaged and rejected by Philolaus resemble nothing more than the a-cosmic states of the Empedoclean cycle. Under the reign of Love all internal divisions are banished, so that we get one single, giant homogeneous continuity, an unlimited thing alone, while under the reign of Strife there is nothing but division, limiting things alone. By contrast, Empedocles’ cosmic phases, along with the mortals they contain, contain both continuities and divisions, as does the cosmos of Philolaus. The difference is only that where Philolaus insists that if there is to be a cosmos, it must contain both types of things, Empedocles states that when there is a cosmos, it is because the causes for both types of thing are present, in other words, that the cosmic phases are the product of the joint action of Love and Strife. Of course, inasmuch as Empedocles describes the world as a process, one can rightly speak of either Love or Strife as having the initiative at any given time, and thus being the ‘moving cause’ of any particular event. Yet this convenient shorthand does not in fact tell the whole story, and is especially misleading if we imagine that Empedocles did not assume such a constant interaction of Love and Strife during the cosmic phases of the cycle. But as I have tried to show, a close reading of numerous fragments supports just such a view.

5. Conclusion: the disproportion in the depiction of the activity of Love and Strife If all of this is correct, it will not be out of place here to offer a few general reflections on the significance of the preceding mate¬ rial. Eor Empedocles, I suggest, the ambiguity of Love and Strife turns largely on the difference between a common-sense and a more sophisticated world-view, or in terms closer to his own, between a mortal and a god’s-eye view of reality. While on one level Empe¬ docles remains committed to a more intuitive view of Strife as

Empedocles on the Ultimate Symmetry of the World

41

destructive and negative (B 17. 19 ovXojxevov) and Love as creative and positive (B 17. 21-6), on another he is intent on exposing their more profound ambiguity in the life of mortals, where both forces participate in the creation and destruction of life.** On that univer¬ sal scale, both kill and nourish (B 17. 3-5). Even the complete sway of Strife may be said to result in a birth of sorts, the birth of the elements (B 22. 9), just as in Love and mixture these same elements ‘pass away into one another’ (B 26. 2). Love, in turn, creates various forms of life (B 35. 7 and 16) until the completion of her work puts an end to the separate existence of rriortal creatures (B 26. 7). Since mortals occupy a middle ground between the complete fusion of the Sphairos and the complete separation of Strife, they can find themselves in that middle position either through a fall from greater harmony and integration, i.e. via Strife, or through an ascent from the separated elements, thanks to the constructive work of Love. There will be a sense in which one of the two is always the more proximate cause of becoming, depending upon the direction of the cosmic pendulum-swing, but both are also necessary for mortal life, not only for the cycle. This, as I have insisted above, is the point of the expression ^vvoSov re BidrrTvUr re yeveOXrjs ‘the unification and development that are birth’.

(a) Asymmetrical explanations: cosmology vs. biology Yet, while it remains to Empedocles’ credit to have recognized the ultimate symmetry of the world sub specie aeternitatis, this also complicated his task as a cosmologist and biologist, especially when it came to co-ordinating the two views with any degree of coher¬ ence. As a cosmologist, he prefers an Anaxagorean, developmental narrative, one that depicts separation and articulation from a more ‘confused’ state, a point already noted by Aristotle.*^ And as it Huffman, Philolaus, 47 n. i, also points out that Aristotle does not appear to align Philolaus with those Pythagoreans who, in constructing their table of oppo¬ sites, placed the ‘limited’ under the good, the unlimited under the bad: ‘Indeed, throughout the fragments of Philolaus, limiters and unlimiteds are presented on completely equal terms and it would appear that Philolaus, at least, saw both as necessary for the world-order to arise and did not consider either category as good or bad.’ “ Decaelo 30i‘‘i4-i8: ‘But to make the world process start from things in motion and separate is irrational. That is why Empedocles passes over the process of [world] formation in the period when Love is prevailing [Std Kal 'EixTreSoKXrjs TrapaAeiVet r-qv {yevtaiv) iirl rrjs (/nAoT-pros]; he could not have built up his universe by making it out

42

Simon Trepanier

happens, that developmental pattern into ever greater articulation and separation conforms to the current direction the world is tak¬ ing, a fall from the Sphere and higher integration, towards complete separation under the reign of Strife.

4 seem that Empedocles’ biology, in

But given that much, it'woul

order to be at least historically correct, would have to limit itself to describing the emergence of the various forms of life in our present world as ‘developments’ or ‘unfoldings’ from some more perfect blend. In fact, with the exception of fragment 62 and now the comparable material in ensemble

d, Empedocles rarely has recourse

to such a scheme. Instead, we find that the majority of the biological passages depict Love constructing organs from apparently pure or unmixed—and seemingly inert—elements, e.g. fragment 73: cos Se Tore

Kvirpis, enel t’ iSlrjvev ev Ofu-Ppcp

el'Sea noLTTVvovaa docp nupi ScoKf Kparvvai.

So then Kypris, when she had drenched earth with water, fashioned forms and gave them to swift fire to harden.

As in that example, drawn from pottery or baking, most of Love’s craftsmanship is described in the past tense. Once more, this raises the question of where in the past we should situate these ‘labours of Love’. Do they belong in the distant past, in the world of increasing Love before the current cosmos, or did they occur in the past of our world, yet still within the current phase of the increasing Strife? That Empedocles should be discussing our world certainly seems a more obvious assumption than the alternative, but the appar¬ ent availability of separate elements for Love to assemble clashes with the dual-creation view. According to that view. Strife and the separated elements lie ahead of us, not behind. As such, this incon¬ gruity has been claimed by advocates of the single-creation view to be strong evidence against the double creation. Yet since I have now given my grounds for rejecting that general scheme, some explanation is still required for what I would concede is a com¬ plication for the traditional view, but certainly no insurmountable objection.

of separate elements and combining them by Love’ (trans. W. K. C. Guthrie, Loeb edn., 1939)-

Empedocles on the Ultimate Symmetry of the World

43

(b) The timing of Love’s work and fragment yi If the traditional view of the cycle is correct, and we are to imagine a double creation, then for the specific position of Love’s work as artifex, there are two possibilities. First and least satisfactory would be to place Love’s work in the past of the present phase of increasing Strife. This is plausible enough, but doing so without falling back upon the single-creation view involves a little special pleading. At most. Love’s assembling of the already separated elements could be no more than a temporary counter-offensive in which she reclaims what she has already lost.*^ The other alternative, which I will defend here, is that most if not all of the descriptions we have of Love assembling the elements took place in the opposite phase of the cycle, after the reign of Strife, when Love first reasserted herself and reintroduced mixture to the elements. Assuming for now the correctness of the dual-creation view, as well as our position in a world of increasing Strife, this is the only phase to which fragment 35 can be assigned. But if that seems to make the best sense of fragment 35, some have felt that it is directly denied by fragment 71: el he TL aoL irepl raivhe Xltto^vXos eVAero ttlotis TTcjs vharos yaiijs re Kat aWepo'S rieXlov re KipvapLevoiv eihrj yevolaro ypocd re dvrjTwv Toaa’, oaa vvv yeydaai avvappLoaOevr’ Affipohlrri . . .

But if you are not fully convinced of these matters, how water, earth, and aither and fire in mingling became (?) the shapes and colours of mortals, as many as have now come into being, held together by Aphrodite . . .

The fragment marks a transition in the work, introducing an addi¬ tional proof or example, one offered to reinforce the disciple’s belief in the capacity of the first principles to generate the great variety of mortal life. Its point seems to be much the same as B 23. 9-11, and it would even be tempting to speculate that it introduced that very fragment, if we did not know from Simplicius, In De caelo 530. 5 Heiberg, that it actually introduced fragments 73 (quoted above)

So Wright, Extant Fragments, 221, on her fr. 6i (=71 DK): ‘In the present world there is an increasing separation of the elements, but Aphrodite is able to counteract this for a time by bringing together parts of the separating elements into temporary compounds.’

44

Simon Trepanier

and.75, both of which depict Love fashioning organs from unmixed elements. Line 4 is what has been thought to show that Love is currently, vvv, the sole demiurge at work, and that this work consists in mixing

separated elements.*'^ Does it irpply as much? I think not. The line merely makes the point that Love is now, as ever, what holds mortals together. Strictly, it does not claim that the colours and forms of mortals produced by the mixture of the elements are the very same ones now held together by her, only that they are as many as, roao", oaa according to DK’s text, which follows Karsten’s 1838 edition.*®

In this regard, it is noteworthy that MSS A and F of Simplicius record the opening of the line as rota (sic), oaa, which opens up the possibility of correcting the line to

toV,

oia ‘of the same kind

as’, or to Toi’, oaa ‘such, as many’ (see O’Brien, ECC 184 n. 3). Both options would make the line’s meaning as a comparison much more explicit. But even without changing DK’s text, I doubt that the creation of mortals considered at B 71.3 is anything more than a hypothetical situation being put before the hearer. Notably, yevoiaro in that line, an aorist optative, surely indicates a contrast with the perfect indicative yeydaai in line 4; the implication of the perfect indicative is that the mortals held together by Aphrodite in line 4 are to be considered an accomplished fact, whereas the optative in line 3 denotes the obscurity of the process of origin, yet to be illuminated. The original sense must have been; ‘how . . . water, earth, etc. might have become the shapes and colours of mortals, as many as have now (actually) come to be, held together by Aphrodite’.** As such, it does not commit Empedocles to any specific chronological ** Solmsen, ‘Love and Strife’, 222—30. ** Nowhere have I found in Empedocles an instance where tooos . . . ocros implies a complete identity of the two things compared; see Wright’s index, Extant Fragments, s.v. Toaos, and more widely, LSJ s.v. Although the aorist optative yevotaro in line 3 can be explained as the result of an indirect question following an imperfect verb, nothing prevents it from also indicating potentiality, since in Homer and earlier poetry the particle av is not always found, even when potentiality is meant: see H. W. Smyth, Greek Grammar (Cam¬ bridge, Mass., 1920; rev. edn. 1956), § 1821. Note, by contrast, that the indicative perfect yeydacn is not ruled by the indirect question. The indicative mood might stand for vividness, but the contrast in moods rather reinforces the idea of a contrast between the actuality of the mortal forms available to the hearer for observation and the potential or obscure status of their origin. DK’s translation of the verbs, ‘entstehen konnten, als jetzt entstanden sind’, conveys this nuance better than Wright, Extant Fragments, 221: ‘how from the combining of water . . . came the forms and colors of mortal things which have now arisen’ (emphasis added).

Empedocles on the Ultimate Symmetry of the World

45

frame of reference, but merely to an explanation of the origin of mortal life through mixture of the elements. Lastly, to quell any lingering worries about

vvv,

I point out that the type of creation

intended, as shown by fragments 73 and 75, which fragment 71 introduced, appears limited to organs; it does not produce whole creatures. Without delving too deeply into the complexities of the doxographic evidence for the zoogonic phases, 31 A 72 DK (Aetius 5. 19. sff., Diels, Dox. Gr.), it should be clear that this is not what currently occurs, but that at present most mortals reproduce through one another. Si’ aWrjXojv in the doxographer’s phrase. Love still holds mortals together, but she does not now build them from scratch.

(c) Why the world of Love? Fragment 71 aside, I would venture to say that the most serious ob¬ stacle to the scholarly success of the double cosmogony and zoogony has been the difficulty, especially in the fragmentary state of the work, of discerning a clear motive for such a fantastical essay in pure speculation as the depiction of an entire parallel world. Even supporters of the dual cosmogony have felt that the burden involved is so heavy as to render the decision to undertake it incomprehen¬ sible.*’ Yet if the traditional interpretation is sound, and I have argued that it is, then we must also suppose that Empedocles had his reasons for choosing to describe Love’s zoogonic activity in the anti-world of increasing Love. What can these have been? My first answer is to suggest that Empedocles may have judged that his audience would more readily accept an account of the ori¬ gins of mortals in terms of composition of the elements, rather than remnants of the Sphere-god. Accordingly, he chose to locate his exposition in that phase of the cycle in which Love first built up mortals from the elements. Below I will attempt to produce some further philological grounds for thinking that this is what he in fact did, but perhaps here it may suffice to wonder what difficulties Empedocles would have faced if he had attempted to The modern locus classicus for this occurs in the first edition of G. S. Kirk and J. Raven, The Presocratic Philosophers (Cambridge, 1957), 347; ‘Empedocles, by his introduction of the cosmic cycle, has set himself a task which might well overtax the most fertile imagination. He has imposed upon himself the necessity of describing a cosmogony and a world that are the exact reverse of the world we know and of the cosmogony that brought it into being.’

46

Simon Trepanier

explain mortal life solely as varying degrees of decline or remove from the more perfect mixture of the Sphairos. One could, for ex¬ ample, attempt to dehne a marble statue in terms only of what has been removed from the original stone block, or seek to describe plants as ‘animals without locpmotion’. But for such an indirect and negative account to give us even a vague idea of the definiendum, we would need an accurate account of the starting point, the marble block or animal nature. In the case of the One or the Sphairos, the hearer cannot have been expected to have any famil¬ iarity with it.** My second answer also takes the form of a counter-question. Given the symmetry of the cycle, why need we assume that an account of the world of Love would tell us nothing about our own world? In ethics especially, a long Western tradition of political idealism has made use of alternative or potential worlds, be it Plato’s Politicus or various early modern conceptions of the ‘state of nature’, placing theses before our eyes as a means of analysing our actual predicament.*'’ In this specific, proto-scientific case, why need a zoogony under increasing Love not also have explanatory force now, in the world of increasing Strife, especially since both powers would be active in the cosmic phases? To describe mortals as elements plus a certain amount of Love, or as mixture plus a certain amount of Strife, would be in some sense equivalent, since in any case the existence of elements implies the co-presence of Strife, and mixture the co-presence of Love, and in giving an account of the cosmos and of mortal beings one would have to invoke the same six explanatory principles. If that is so, then ultimately it may not matter where one starts.

** Since mortals are (now) mostly dominated by Strife and mistake parts for the whole (cf B 2. 6), the natural place to begin in explanations is from parts to the whole, since they would have less familiarity with such wholes, e.g. the Sphairos, than with the parts. The protreptic aim of Empedocles’ message as a whole is still largely unex¬ plored. But might not the description of a world ruled by Love serve as a powerful counter to the claim that Strife is the only way of the world, or that ‘war is the father of air (Heraclitus fr. 53)? Aristotle puts his finger on this at GC 315“ 19-20 when he asks: ‘Nor is it clear whether one must posit that for him [Empedocles] the principle is the one or the many [i.e. the Sphairos or the elements].’

Empedocles on the Ultimate Symmetry of the World

47

{d) Aristotle’s evidence To reinforce this possibility, I invoke the statement reported by Aristotle, GC 334'‘5-7, that ‘[Empedocles] also says that the world now under Strife is similarly arranged as it was previously under Love ,

ajLta

Se

/cat tov Koojxov ojxoLcos eyeiv frjGip irrl rov veiKovs vvv

/cat TTporepov ivi

rijs fiXlasd^ I wonder if such a statement, in the

original, may not have been less of a general point about the sym¬ metry of the cosmic cycle than a justification for, depending on what was convenient, either adopting a more congenial mode of ex¬ position, e.g. construction vs. development in biology, or avoiding a repetition of a subject already discussed in one phase by claiming ‘and vice versa’ for the same phenomenon in another phase. To be sure, Aristotle’s specific chronological reference, that the rule of Love occurred before the present dispensation, coincides with the past tenses used to describe Love’s activity in the fragments, and is otherwise difficult to explain, since strictly speaking the rule of Love belongs no more to the past than to the future. It must correspond to something in the original. A final advantage of this scheme is the way in which it disarms the complaint that it was needlessly complicated on Empedocles’ part to have two cosmogonies and zoogonies, where one of each would suffice. As argued above, the assumption of symmetry already ex¬ plains why Empedocles, upon adopting a theory of dual creation, need not thereby have also felt obliged to tell his tale twice. But is there any evidence that he did in fact eschew recounting one whole zoogony and one entire cosmogony? There is, although the total is perhaps closer to one and three-quarters. From Aristotle, De This passage was much discussed in the heyday of the controversy over the cycle. My own views follow O’Brien, ECC 167—77, who offers there some lucid pages on matters left obscure by others. Solmsen in particular rejected this testimony (‘Love and Strife’, 231—3), but/without, so far as I can tell, giving an explanation of what Aristotle misunderstood, and simply claimed that evn' NeUovs denotes every phase but the Sphairos. Against this, Aristotle must be talking about the lead either principle takes in cosmogony and zoogony, i.e. between two opposed cosmic phases. The antithesis ‘under Strife’ and ‘under Love’, as I take it, can denote either (i) the two extreme and non-cosmic phases, i.e. the Sphere vs. the reign of Strife, or (2) the opposite cosmic processes, from Many to One, or increasing Love, and One to Many, increasing Strife. But since (i) is not a symmetrical contrast, he has to be talking about (2), which, especially at the mid-point, will produce a cosmos in more or less the same state, albeit headed in opposite directions. Otherwise, Aristotle seems to be confusing the two, and contrasting the Sphere with the cosmic phase under increasing Strife. But how can the Sphairos be in any way ‘similarly disposed’ (opolais e'xetv) to any of the cosmic phases?

48

Simon Trepanier

caelo 301*14, quoted above in note 82, we know that Empedocles ‘skipped’,

napaXeiTTei,

the cosmogony under Love, which I take to

mean that he gave no account of it, even though he assumed it.'’^ I

In biology, if the majority of our extant passages describe events occurring in another phase of the cycle, under rising Love, minus a few lines in fragment 62 and ensemble d, then Empedocles not only evaded the above criticism of needless duplication, but cannily exploited this situation to his rhetorical advantage. The double creation freed him to select those phases more suited to the needs of each exposition: the world of Love for biology, rising Strife for cosmology.

(e) The world of Love and book i of Empedocles’ Physics Can this be taken any further than a mere possibility? Is there any further evidence bearing on the original context of this pu¬ tative zoogony in the distant world of Love? A full discussion is not possible here, but to the extent that it may render my proposed doctrinal reconstruction more plausible, I end this study by sketch¬ ing out how I think the new evidence from the papyrus affects the placement of these zoogonic passages, including fragment 71, in the original literary work. In the current state of our evidence, there is sufficient material to form some idea of the general structure of book i. Thanks to Simplicius, we know that fragments 17, 21, 23, and 26 occurred in that order in book i of the Physics, and now, by a fortunate accident of preservation in the Strasbourg papyrus, we also know that fragment ly + ensemble a spanned lines 233-300 of that first book.''^ Eurther, it seems very likely that fragment ly

ensemble a

stood as the initial presentation of Empedocles’ doctrine, following what must have been a rather long proem or preamble, even though the exact nature of that proem is a highly contentious topic. Be that as it may, the importance of fragment 17 as the core doctrinal statement of the Physics has never been doubted, and its program¬ matic nature is conspicuous in at least two regards. First, the last ten lines of ensemble a promise the hearer examples or proofs of its general thesis (cf. a (ii) 29), one of which is perhaps to be linked For discussion, see O’Brien, ECC 175-7. ’’ See n. 7 above for further details, and M-P 103-14 for discussion and citations of source material.

Empedocles on the Ultimate Syrntnetry of the World *

to fragment ^blensemble

49

and second, fragments 21, 23, 26, and

35 all rehearse and develop material from fragment 17. Fragment 21 points out the presence of the four elements behind the great world masses, while fragment 23, the simile of the painters, shows how a limited set of four basic elements can underlie the full spec¬ trum of diversity found in the world (for fragment 26, see above, pp. 28-30). Lastly, fragment 35 openly announces that it is a re¬ turn to the account of fragment 17, and repeats verbatim a (ii) 18-20 from fragment

+ ensemble a, which, once again, Simpli¬

cius places in book i. The book in which fragment 35 figured is not known from external evidence, but its explicit reference to frag¬ ment 17 gives prima facie grounds for also placing it in book i. Although the evidence is hardly as complete as we should like, it seems to follow that a substantial part of book i, from lines 300 ff., was devoted to reinforcing and refining the thesis advanced in frag¬ ment 17. How, then, does this relate to fragment 71 and the world of Love? I begin with the evidence linking fragment 71 to the world of Love. From Simplicius, In De caelo 529. 28 and 530. 5 and 8 Heiberg, we know that the explanation which fragment 71 introduced involved the zoogonic activity of Love, and there Simplicius further tells us that fragments 73 and 75 are but two of many examples he could produce of Love acting as craftsman. If we collect the parallels to fragments 73 and 75 in the corpus, by which I mean those passages specifically concerned with Love as artifex, and not those where she is simply associated with mixture or mortal creatures (e.g. B 20. 2-3 or B 35.7), this produces the following list: fragments 73, 75,84, 86, 87, 95, 96, 98.“^’ The general similarity between these passages, in which Love is shown fashioning only organs, not whole creatures, when combined with Simplicius’ remark, noted above, suggests that most if not all of them occurred together in the original work. And if that is correct, then pari passu they probably all belong to the same phase of the cycle, the world of increasing Love, whose specific Fr. 76 seems to be making the point that the elements can be arrayed in patterns that differ from their current stratification in the macrocosm. Its identification as a proof rests on a verbal similarity between fr. yh/emewWe ensemble

a:

b

and the last lines of

see M-P 247-64.

More difficult to place in this group is fr. too, the clepsydra simile. Its purpose is to describe respiration, but it contains no clear reference to Love. Frr. 33 and 34, however, which discuss gluing or fitting, are certainly appropriate for Love, and could also have figured in this context.

50

Simon Trepanier

chronological context will have been introduced and explained in fragment 35. Further indirect evidence reinforces the case for taking them all together, and shows that the whole set, or a‘good majority of them, most likely occurred in^ book i of the Physics. We know from Simplicius, In Phys. 300. 19 Diels, that fragment 96, on the elemental composition of bone, occurred in the first book of the Physics. Fragment 98, in turn, on the ratio of flesh and blood, is said by him {In Phys. 32. ii) to have occurred after fragment 35. For reasons advanced in the previous paragraphs, fragment 35 is likely to have belonged to book i, and if that is so, then fragment 98 also belongs there (conversely, fragment 98’s similarity to fragment 96 is additional evidence for placing fragment 35 in book i). While not every one of these points is conclusive, their cumulative force inspires'reasonable confidence that the work of Love, or at least her ‘organogony’, was described in book i of the Physics, with fragment 71 introducing that section of the work, perhaps followed closely by fragment 35, which will have fixed the chronological context of Love’s work within the cycle. But that cannot be all. For once it is supposed that these pas¬ sages occurred in book i, then this also makes book i the most likely location for the further stages of Empedocles’ zoogony, the linking of the organs and limbs by trial and error into various larger, mostly monstrous, assemblages: cf. fragments 57-61. According to Simplicius, In Phys. 371. 33-372. 14 Diels, it was only after that second period of trial and error, when some parts happened upon appropriate counterparts, that a third age or generation of mortals arose, comprising the species with which we are familiar today. This leads to a further observation. If a substantial amount of text after lines 300 ff. of book i of the

was devoted toproving

or reinforcing the thesis of fragment 17, and book i also included a description of Love as the artifex of various organs, followed by yet more passages retailing the combination of these organs into monsters and ultimately known species, then unless we posit a rather high number of lines for book i, say beyond 600 or 700, this leaves little room for cosmology.Could cosmology have been left out until book 2, or at least until after the zoogony of Love? Nor do I think it at all likely that these zoogonic passages could have preceded fr. 17, as they seem to presuppose an acquaintance with the six first principles, most likely first unveiled in fr. 17.

Empedocles on the Ultimate Symmetry of the World

51

Once again, the evidence is severely limited, but fragment 62, which Simplicius places in book 2, does seem to presuppose a cosmology based upon differentiation and division, i.e. Strife. It does not seem outlandish, therefore, to suppose that the cosmology of Strife, the story of our world, came after the world of Love, and perhaps was even left for book 2. This relative order meshes well with the evidence from Aristotle quoted above, according to whom the world of Love preceded the present cosmos,"*’ and otherwise conforms to Empedocles’ general habit of putting Love before Strife, as in his AB motif of alternation in the poetic presentation of his doctrine. The significance of the latter, as Graham has shown, is likely to be more than ornamental. But there is yet another implication to be drawn from all of this. If most of Love’s work was described before the cosmology of Strife, then from the hearer’s standpoint there will not have been any immediate, glaring contradiction between that process of construction and our current situation, in which Strife’s increasing power ‘unfolds’ mortal forms. That process will not yet have been explicitly depicted in the work.

(/) The puzzle of emergence and the zoogony of increasing Love From these considerations I return to the link between fragment 71 and the didactic aims of book i, in order to offer one last sug¬ gestion as to Empedocles’ motive for the order of presentation reconstructed above. If, then, it is in fact the case that Empedocles depicted Love’s creation of mortals first, in the anti-world of rising Love, before undertaking to describe the corresponding zoogonic process in our world of Strife, what can have motivated that de¬ cision? Above I suggested that the choice may have turned on a rhetorical consideration, and that his audience especially might have found explanations based on composition of the elements, or ‘from the bottom up’, mote intuitive or easier to grasp. But perhaps there is more to it than that. It is notable that the opening lines of fragment 71 promise to reinforce the hearer’s confidence that mixture of the elements can explain the full diversity of life. Based on Simplicius’ evidence for the relative sequence 17, 21, 23, 26 of the fragments, we can infer Quoted above, p. 47. This picture is also confirmed by Simplicius in several places, among which one may quote In De caelo 590. 19-21 Heiberg: ‘he [Empedo¬ cles] says that the elements, having been previously united under Love, [when] later divided by Strife, produced the present world [rdrSe

tov

«:dapor].’

52

Simon Trepanier

that, in the original work, this had been asserted at least four times before now. The same idea is advanced via the zoogonic formula, at a (i) 8-a (ii) 2 of fragment i'] + ensemble a, as well as in the other occurrences of that formula, fragments zf and 23, and, in a more succinct fashion, ^t B 26. 3—4. This time round, however, Empedocles does more than simply assert the fact, he promises to give further grounds for believing that the mixture of the elements could generate mortal forms. As such, I think that fragment 71, probably for the first time in the work, problematizes the relation between mixture and the diversity of life. The opening line, ‘But if you are not entirely convinced . . .’, concedes that mixture alone may seem insufficient to explain the full diversity of life. In par¬ ticular, and this can be inferred from the fact that fragments 73 and 75 followed fragment 71, the main puzzles that Empedocles wants to elucidate are, allowing for some reformulation: (i) how various tissues could arise from elements, i.e. non-tissues, and (2) that mix¬ ture of the elements alone seems incapable of accounting for the emergence of more complex and functional or tool-like structures. His reply combines the two problems by invoking the figure of Love as craftsman. Through her knowledge of the elemental ratio underlying each tissue or material. Love can produce tissues with different properties, and by artfully recombining these into larger structures she can create organs, or functional ‘tools’. The first part of this reply, proportional combination as an ex¬ planation for emergent qualities, has an interest all its own, which I will not dwell upon here. The second problem, however, whose an¬ swer seems an adumbration of the argument from design, became a crucial issue in later debates about natural science. Hands, teeth, eyes, all of these display an aptitude to perform certain tasks, one predicated upon the assembly of specific properties into complex structures, that seems to defy explanation in terms of chance ag¬ gregation. This appears already to have been enough of an issue for Empedocles to have to account for it, although in proposing his an¬ swer, the purposive design of Love, he does not give it anything like the importance later teleological thinkers did. If he had, it seems unlikely that Aristotle would have so readily classified him, along with the early Atomists, as an anti-teleologist.'’* The two most illuminating discussions that I know of Aristotle in this regard are (i) D. Furley, Greek Cosmologists, 177-200, and (2) M. Furth, ‘Aristotle’s Biological Universe: An Overview’, in A. Gotthelf and J. G. Lennox (eds.). Philosophical

Empedocles on the Ultimate Symmetry of the World

53

This becomes particularly conspicuous when one considers the later stages of this same zoogony, something which I have already briefly touched upon above. Without once again entering into too much detail, I think that it will nevertheless be uncontroversial to recognize that, on the w'hole. Love as the agent of rational design plays a fairly circumscribed role in Empedocles’ general account of life.'^^ Indeed, the most notable common feature of the ‘craftsman’ fragments is that they predominantly portray Love fashioning tis¬ sues or organs, not whole creatures, and doing so out of pure elements. When Empedocles describes the further assemblage of these parts into various monsters and animals, he tends to invoke chance, as in fragments 57-61. There we find individual organs at first stand¬ ing alone, and subsequently assembled haphazardly into monsters. Properly, therefore, Empedocles’ fuller answer to the problem he raised in fragment 71 shows that he postulates what we would recognize as two additional explanatory factors beyond mixture. First, he imputes to Love the capacity to design, but this does not rule out the importance of chance. Certainly, Love alone does not dictate how any of the diverse creatures found or obtained their particular forms. Thus, despite his use of Love as a purposive agent, Empedocles hardly intended a full-scale teleological account of life, or the uni¬ verse for that matter. If anything, the craftsmanship of Love here seems almost a stopgap, granted to the end of reinforcing the

ttlotis

of the general picture presented in fragment 17.'°° It seems only intended to get over the initial but crucial hurdle involved in the transition, not from element to mixture, but from element to the

Issues in Aristotle's Biology (Cambridge, 1987), 21—52, although Furth in places is somewhat unfair to Empedocles. ” The principal sources are collected in 31 A 72 DK (=Aetius 5. 19. 5) and frr. 57-61, including especially Sim’plicius’ comments introducing fr. 61, InPhys. 371. 33-372. 14 Diels. For a recent discussion see Wilcox, ‘Whole-Natured Forms’. Solmsen’s analysis of the relation between fr. 71 and the zoogony of Love in ‘Love and Strife’, 225—9, if taken in abstraction from his general thesis about the cycle, remains perceptive; cf. p. 228: ‘However it is probably unwise to postulate that Empedocles gave an exhaustive account covering a large variety of biological forms. He may well have thought it sufficient to be specific about the most important tissues and a few organs, and after this to suggest that the tissues entered into a variety of combinations and that items seemingly heterogeneous were yet identical in substance. A few striking examples would create ttIotls', and since exact proofs could not be given, it was important to inspire the imagination of the audience and turn it in the right direction.’

54

Simon Trepanier

simplest organic or functional unities Empedocles could conceive of, tissues, otgans, and limbs. From this one can safely conclude that Aristotle is mostly right, and that Empedocles’ overall strategy remains in large part mech¬ anistic. Even more specifically, his goal is anti-emergentist, in the sense of attempting to deny that generation, be it of the cosmos, or in this case of mortals, is an absolute beginning, rather than one more phase in an endless and ongoing process.This of course is what one would expect of a cosmologist working in the wake of Parmenides. As the doctrine of the cycle holds, there is no absolute beginning, nor final end. Either notion, as Empedocles points out in fragments 8 and 11, is equally absurd. Yet here, I suggest, perhaps Empedocles appreciated that the problem of emergence, as posed by the zoogony of Love, and involving an account of life from the bottom up, in the end presented a greater difficulty than the re¬ verse, the developmental account. In the latter, the ‘unfolding’ of the world by Strife, including the emergence of animal forms, could be understood as merely a development of latent properties. This is not so in their construction by Love, where these features have to be designed and built into the assemblage of the elements by Love. They are not already there, but in Love’s mind, as it were.'®^ At the same time, by undertaking to describe precisely that point in the cycle at which the difficulty will have been the most conspicuous, the return of Love to the elements, might we not also suppose that Empedocles recognized in it the most serious challenge to his posi¬ tion, the one he had to meet first? If so, this choice may also indicate that he at least had an inkling of the difficulty which his commitA valuable account of these questions is A. P. D. Mourelatos, ‘Quality, Struc¬ ture, and Emergence in Later Pre-Socratic Philosophy’, in J. Cleary (ed.), Proceed¬ ings of the Boston Area Colloquium in Ancient Philosophy, 2 (Lanham, Md., 1987), 127-94. This, I think, is precisely why Aristotle in Physics 2. 8, where he spends so much time criticizing Empedocles’ zoogony of Love, spends only one dismissive line on the zoogony of Strife. At

after making the case that the whole animal

or creature has to be ever present as form, he adds, quoting B 62. i, ‘and the whole natures “first” toere seeds'. What Aristotle implies by suggesting that these ‘whole natures’ are seeds, as Simplicius helps explain {In Phys. 381. 29-382. 18 Diels), is that, if taken as such, then that part of Empedocles’ account is much more compatible with Aristotle’s own views about the eternity of natural species, and need not be criticized at length. The process of generation by Strife, because it posits that everything required for generation is in a sense already present, only latent, is far more plausible to Aristotle than the building up of mortal creatures out of pure elements, followed by a period of trial and error in which all sorts of combinations are attempted, mostly in vain.

Empedocles on the Ultimate Symmetry of the World

55

ment to symmetry set him; explanations from the ‘bottom up’ and from the ‘top down’ may not be mirror images of one another.

{g) The ultimate symmetry of the zvorld If that is correct, then perhaps it may serve as an explanation for what I would call a disproportion in the depiction of the zoogonic and cosmogonic activity of Love and Strife, as opposed to their more fundamental but philosophically problematic equality. The recognition of such a distinction, should it become generalized, may help to put an end to the various distortions of the symmetrical cycle found in Empedoclean scholarship. University of Massachusetts, Boston

BIBLIOGRAPHY Allen, R., and Furley, D. (eds.). Studies in Presocratic Philosophy [Studies] vol. li (London, 1975). Barnes, J., The Presocratic Philosophers (Tondon, 1982). Bignone, E., Empedocle: studio critico (Turin, 1916). Bollack, J., Empedocle, i. Introduction d Vancienne physique [Bollack, i] (Paris, 1965); ii. Les Origines: edition et traduction des fragments et des temoignages; iii. Les Origines: commentaires [Bollack, ii and iii] (Paris, 1969; repr. 1992). Cornford, F. M., From Religion to Philosophy (London, 1912). Diels, H., and Kranz, W., Die Fragmente der Vorsokratiker, 6th edn. [DK] (Berlin, 1951). Furley, D., The Greek Cosmologists, vol. i [Greek Cosmologists] (Cambridge, 1987). Furth, M., ‘Aristotle’s Biological Universe: An Overview’, in A. Gotthelf and J. G. Lennox (eds.). Philosophical Issues in Aristotle’s Biology (Cam¬ bridge, 1987), 21-52. Gemelli Marciano, M. L., review of M-P, in Gnomon, 72 (2000), 389-400. Graham, W. D., ‘Symmetry in the Empedoclean Cycle’ [‘Symmetry’], Classical Quarterly, ns 38 (1988), 297-312. Guthrie, W. K. C., A History of Greek Philosophy, ii. The Presocratic Tra¬ dition from Parmenides to Democritus [HGP ii] (Cambridge, 1965). Holscher, U., ‘Weltzeiten und Lebenskyklus: eine Nachpriifung der Empedokles-Doxographie’ [‘Nachpriifung’], Hermes, 93 (1965). 7-33Huffman, C., Philolaus of Croton, Pythagorean and Presocratic [Philolaus] (Cambridge, 1993).

Simon Trepanier

56

Inwood, B., The Poem of Empedocles: A Text and Translation atid Intro¬ duction [Poerw] (Toronto, 2001; ist edn. 1992). -review of M—P, in Classical Review, 50 (2000), 5-7. Kahn, C. H., ‘Religion and Natural Philosophy in Empedocles’ Doctrine of the Soul’, Archivfiir Qeschichte der Philosophic, 42 (i960), 3—35; repr. in Mourelatos, Pre-Socratics, 426—56. Kirk, G. S., Raven, J. E., and Schofield, M., The Presocratic Philosophers [Presocratic Philosophers] (Cambridge, 1983; ist edn. by Kirk and Raven, 1957)Long, A. A., ‘Empedocles’ Cosmic Cycle in the Sixties’ [‘Sixties’] in Mourelatos, Pre-Socratics, 397—425. -and Sedley, D. N., The Hellenistic Philosophers (2 vols.; Cambridge, 1987). Liith, J. C., Die Struktur des Wirklichen im empedokleischen System iiber die Natur (Meisenheim, 1970). Mace, S. T., review of M—P, in Classical World, 95 (2002), 195—7. Mansfeld,J., ‘Ambiguity in Empedocles B 173—5: An Interpretation’ [‘Am¬ biguity’], Phronesis, 17 (1972), 17—39. -and Algra, K., ‘Three Thetas in the “Empedocle de Strasbourg’” [‘Three Thetas’], Mnemosyne, 54 (2001), 78-84. Martin, A., and Primavesi, O., L’EmpMocle de Strasbourg {P Strasb. gr. Inv. 1665—1666): introduction, edition et commentaire [M—P] (Berlin and New York, 1999). Mourelatos, A. P. D., The Route of Parmenides: A Study of Word, Image and Argument in the Pragments (New Haven, 1970). -‘Quality, Structure, and Emergence in Later Pre-Socratic Philo¬ sophy’, in J. Cleary (ed.). Proceedings of the Boston Area Colloquium in Ancient Philosophy, 2 (Lanham, Md., 1987), 127—94. -(ed.). The Pre-Socratics, 2nd edn. [^Pre-Socratics] (Princeton, 1992). O’Brien, D., Empedocles’ Cosmic Cycle [ECC] (Cambridge, 1969). -Pour interprher EmpMocle [Pour inteprher] (Paris, 1981). -‘Heraclite et I’unite des opposes’. Revue de metaphysique et de morale, 2 (1990), 147-67-‘Empedocles Revisited’ [‘Revisited’], Ancient Philosophy, 15 (1995), 403-70. -‘Hermann Diels on the Presocratics: Empedocles’ Double Destruc¬ tion of the Cosmos, (Aetius ii 4. 8)’, Phronesis, 45 (2000), 1-18. Osborne, C., ‘Empedocles Recycled’ [‘Recycled’], Classical Quarterly, ns 37 (1987), 24-50. -‘Rummaging in the Recycling Bins of Upper Egypt; A Discussion of A. Martin and O. Primavesi, L’Empedocle de Strasbourg’ [‘Rummaging’], Oxford Studies in Ancient Philosophy, 18 (Summer 2000), 329—56.

Empedocles on the Ultimate Symmetry of the World

57

Panzerbieter, E, Beitrdge zur Kritik iind Erkldrung des Empedokles [Kritik] (Meiningen, 1844). Sedley, D. N., Lucretius and the Transformations of Greek Wisdom (Cam¬ bridge, 1998). Smyth, H. W., Greek Grammar (Cambridge, Mass., 1920; rev. edn. 1956). Solmsen, R, ‘Love and Strife in Empedocles’ Cosmology’ [‘Love and Strife’] Phronesis, 10 (1965), 109-48; repr. in Allen and Furley, Studies, 221-64. Steiger, K., ‘Die Kosmologie des Parmenides und Empedokles’, Oikoumene, 5 (1986), 173—236. Stevens, A., ‘La physique d’Empedocle-selon Simplicius’, Revue beige de philologie, 67 (1989), 65-74. Stokes, M. C., One and Many in Presocratic Philosophy [One and Many] (Washington, 1971). Tannery, P, ‘La cosmogonic d’Empedocle’, Revue philosophique de la France et de I’etranger, 24 (1887), 285—300. Trepanier, S., ‘“We” and Empedocles’ Cosmic Lottery: P Strasb. gr. Inv. 1665—1666, ensemble a’, Mnemosyne (forthcoming). Van der Ben, N., The Proem of Empedocles’ TJepl ^vaeojs: Towards a New Edition of the Fragments [Proem] (Amsterdam, 1975). -‘Empedocles’ Cycle and Eragment 17. 3—5’, Hermes, 18 (1984), 97lOI.

Von Arnim, H., ‘Die Weltperioden bei Empedokles’, Festschrift Theodor Gomperz (Vienna, 1902), 16—27. Wilamowitz-Moellendorff, U. von, ‘Lesefriichte’, Hermes, 65 (1930), 24550; repr. in his Kleine Schriften, iv (Berlin, 1962), no. 269, pp. 513-18. Wilcox,

J.,

‘Whole-Natured

Eorms

in

Empedocles’

Cosmic

Cycle’

[‘Whole-Natured Forms’], in A. Preus (ed.). Essays in Ancient Greek Philosophy, vi. Before Plato (Albany, 2001), 109—22. Wright, R. M., Empedocles: The Extant Fragments, Edited with Introduc¬ tion, Commentary, Concordance [Extant Fragments] (New Haven, 1981; repr. 1995). Zeller, E., Die Philosophie der Griechen in ihrer geschichtlichen Entwicklung, 6th edn. by W. Nestle (3 vols.; Leipzig 1919-20, repr. Hildesheim, 1963) vol. i/2.

^ '*' Vrtcrwm •^^•P^.miliitt'H lv,*!'vl.‘

- V

*^111. ^

,yi .O .v^li/9^

li,, HtUttMUi )U»'I S'. 'U-'ll i'hli-w^PiV ittFl-’

f*nti

J

l>ai**l (cyb ■WjfttPirrMd^ aiu' ..'A x*V^^


(»^.< Pt’'v-'\•.

•!c,\ii:^T

■Tc'

,o!>tml.4t

' 'f- i

h^;at*m»l»*iu*.A hi «cnn'jt U^lhl*i«5;^>♦t(4'Jy•-, A>iv^ w »*l X4j^n'?i •vWt\yv\ ii**!!** -• *

twu^r^ tiAX..M , r^»uwr«nti\%U4W4^ ;,'•'< s'/J' /*/. '

• '

.,ma»

W< •': r .'> Cl\»r,-

(^'V* ^lj»(-.firXhR »njt lO-^^iifi ^fxqi*kl i,*kr/ til w^y V VO .ilh» rilr>

I,.!-•-, 1^. >

Jl

* .'' Ml *Mh OirU >n> I - < •-’i .% !.»■ .,f

i*oifc*.rrt. u» .1, ^ •^>..>a»\^l ■ I'l.f.j#.*, .. f />♦ , JU [’HuimiijS'fnw’J, ; '^‘w. .. Af'ftIVf '•< .fiv-u»tl ll C^^VrialM* 9090'^,

SOCRATES AND PROTAGORAS ON VIRTUE DENIS O’BRIEN

CONTENTS PART I

1. The three theses of the Protagoras 2. Penner and Vlastos 3. The list of virtues PART II

4. Knowledge as the ‘whole’ of virtue 5. The image of gold in the Protagoras and the Timaeus 6. ‘Parts’ of virtue in the Meno PART III

7. Aristotle’s Socrates 8. Menedemus of Eretria and Eucleides of Megara 9. Ariston of Chios

PART

I

I. The three theses of the Protagoras {a) The opening questions and Protagoras’ replies When

Socrates has shaken himself free from the spell of Pro¬

tagoras’ oratory (328

D

4-7), he fires oflF three seemingly simple

questions in quick succession (329 c 6-e 4).' Each question is cast © Denis O’Brien 2003 I am most grateful for comments and advice to Luc Brisson, Jean-Fran^ois Pradeau, David Sedley, and Suzanne Stern-Gillet. I also learnt a great deal from discussion of the Protagoras at a colloquium organized by Giovanni Casertano, ‘II Protagora di Platone; struttura e problematiche’ (Naples, 27-30 Sept. 2002). ‘ I adopt J. Burnet’s text and lineation, Platonis Opera, vol. iii, 2nd edn. (Oxford, 1909). I refer to two other editions by the editor’s name alone: Henricus Stephanus

Denis O’Brien

6o

in the form of a choice between two distinct theses. In each case, the twd theses are introduced by the same pair of disjunctive particles {noTepov . . .

7]

. . h

Are justice, temperance, piety, and the like parts of one thing, which is virtue [la], or are their names ‘all names of one and the same thing’ [ndvra ovopara

tov

aiirov evos

ovtos]

[ib]? (329

C

6—D l)

Are the individual virtues like the parts of a face (mouth, nose, eyes, and ears), where each part is different from every other part and different from the whole [2a], or are they like ‘parts’ or pieces of gold, which differ from one another and from the whole only ‘in largeness and smallness’ [p-eyeOei Kal ap-LKporpri] [2b]? (329 D 4—8; cf. 349 C 4—5)

Can people have only one part of virtue and not have another part [3a], or, if they have one part of virtue, do they necessarily have all the others as well [3b]? (329

E

2-4)

In his successive replies, Protagoras chooses the first thesis in each pair (329

D

3-E 6).

The individual virtues are parts of virtue [la]. (329 n 3—4) They differ as do the parts of a face [2a]. (329

d 8—e

2)

We can have one part of virtue and not have another part [3a].^ (329

e

5—6)

(b) A rival thesis From this it does not follow that Protagoras’ opponent can adopt the second thesis in each pair to form a cohesive series. On the contrary, (Henri Etienne), Platonis Opera Quae Extant Omnia . . ., vol. i ([Geneva], 1578), 309-62 {=Protagoras), and I. Bekker, Platonis . . . Scripta Graece Omnia, vol. i (London, 1826), 249-372 {= Protagoras). ^ Here and throughout, I would have preferred to write of successive alternatives, using that word with the ‘strict’ meaning given in the Oxford English Dictionary s.v.: ‘A proposition containing two statements, the acceptance of one of which involves rejection of the other.’ But I am assured that that use is liable to be confused with the ‘loose’ meaning of the same word (ibid.: ‘Either of the two “sides” or members of the alternative proposition’), and is therefore better avoided. The reader should however bear in mind that, throughout my comments on this passage (326 c 6-E 6), the various periphrases employed (‘choice’, ‘option’, ‘pair of theses’) refer to three successive alternatives sensu stricto. ^ In the list of virtues, I adopt ‘temperance’ as a purely conventional translation of the untranslatable acafpoavvT). The opening pages of Tuckey’s study of the Charmides provide a useful collection of texts designed to illustrate the occurrence of the word in writers contemporary with and earlier than Plato. See T. G. Tuckey, Plato’s Charmides (Cambridge, 1951), 5—9.

Socrates and Protagoras on Virtue

6i

both options in each of Socrates’ second and third questions stem from the assumption that the individual virtues are parts of virtue. The opponent who wishes to argue, against Protagoras, that the individual virtues are like pieces of gold and not like the parts of a face (2b), and that if we hav'e one part we must have them all and that we cannot therefore have one part and not another (3b), has to start off by agreeing with Protagoras that the individual virtues are parts of virtue (la). Initially therefore we may distinguish two theses with a common starting point. Protagoras claims thatthe individual virtues are like the parts of a face (2a), and that we can have one part without another part (3a). A rival thesis would claim that the individual virtues are like pieces of gold (2b), and that we cannot have one virtue without having all the virtues (3b). The two theses share a common premiss (la), namely that the individual virtues are parts of virtue and that their names are not simply so many names of one and the same thing.'*

(c) A second rival thesis The opponent who disagrees wdth that premiss, and who claims that the names of the individual virtues are indeed only so many names of one and the same thing (ib), cannot be called upon to choose between the comparison of the individual virtues to parts of a face (2a) and the comparison with pieces of gold (2b), since he will have started off by refusing the thesis that the individual virtues are parts of virtue, so that the question what kind of ‘parts’ they might be (whether like the parts of a face or like pieces of gold) simply does not arise. The same opponent cannot be asked to choose between the final pair of theses, whether we can have one part and not another part (3a) or whether our having one part entails our having all the parts ■* I here assume (too arbitrarily?) that the comparison with parts of gold (2b) and the claim that we cannot have one virtue without having all the virtues (3b) are intended to make up a single thesis, in the same way that the comparison with parts of a face (2a) leads ‘clearly’ (cf. 330 B i: S^Aa) to the opposite thesis, namely that we can have one part of virtue without having all the other parts (3a). 1 return to this point in sect. 5 below. For the moment, I note only that, in passing from (2a) to (3a), Socrates does not ask us to imagine a face so mutilated that it would have no ears or no mouth or no eyes. His point is that each ‘part’ of a face carries with it a specific faculty (cf. 330 A 4-B 3). We may therefore be deaf, but not dumb, or dumb, but not blind.

Denis O’Brien

62

(3b), because, again, he would not agree that the individual virtues wete ‘parts’ of virtue in the first place. On the other hand, it does not follow that someone who refuses the thesis of the ‘parts’ of virtue (of whichever kind) has no opinion on the relation between the virtues. On the contrary, someone who adopts the second of the two thhses that Protagoras is initially called upon to choose between (ib) will be committed to the view that if we have any one virtue then we have the whole of virtue. This will not be because he thinks that the parts of virtue are so similar (cf. 2b) that we cannot have one without having all the others (cf. 3b), but for the yet more radical reason that he has al¬ ready claimed that the names of the individual virtues are only so many names of one and the same thing (ib). If therefore some¬ one possesses justice, for example, then he already possesses, so to speak, all that there is to be had in the way of virtue. In possessing any one virtue he already possesses that ‘one and the same thing’ (329D

i: Tov avrov evos ovtos) which is named by all and any of the

individual virtues.

(d) Protagoras’ thesis and the two rival theses Socrates’ three questions therefore allow us to distinguish three theses in all. The thesis that Protagoras subscribes to in adopting the first option in each of Socrates’ three successive questions is opposed to either one of two theses, both of which make use of the terms that Protagoras has discarded. Protagoras’ thesis. The individual virtues are parts of virtue (la). They are like the parts of a face, each differing from the other and from the whole (2a). We can have one part without having another part (3a). A first rival thesis. The individual virtues are parts of virtue (la). They are like pieces of gold, differing one from the other and from the whole only in ‘largeness and smallness’ (2b). We cannot have one virtue without having all the virtues (3b). A second rival thesis. The individual virtues are not parts of virtue; their names are only so many names of one and the same thing (ib). As soon as we have one virtue we have the whole of virtue.

Socrates and Protagoras on Virtue

63

{e) ‘Either the same, or as similar as can be’ {331 b 4-5) The distinction between the three theses that stem from the initial session of question and answer (329 c 6-e 6) explains the form taken by the conclusion to Socrates’ opening argument against Protagoras (331 A 6-b 8). Socrates argues that piety is just and that justice is pious. He concludes that those two virtues cannot therefore be dif¬ ferent from one another in the way that Protagoras had claimed that they were. Instead, he says, ‘justice is either the same as piety, or as similar as can be’ (331

B

4—5: -qrot ravrov y’ eGTir Sucaiorr/? ooiOTrjTi

17 OTL OjXOlOTaTOv).^ The two terms of that disjunction look back to the options on of¬ fer in Socrates’ opening series of questions. The claim that justice is ‘the same’ as piety looks back to the second option in Socrates’ first question (ib), where the names of the individual virtues are so many names ‘of one and the same thing’ (ravrov at 331 back to rov avrov at 329

D i).

B

4 looks

The claim that justice is ‘as similar as

can be’ to piety looks back to the second option in Socrates’ second question (2b), to the comparison between the individual virtues and dififerent pieces of gold which differ from one other and from the whole only ‘in largeness and smallness’ (on ofioiorarov at 331 b 5 is a restatement in positive form of ov8ev 8ia(f)€pei . . . dAA’ ij /xeyedei Kat ajjUKporrjri at 329 D 6—8).*’ ' At 331 B 4, I depart from Burnet’s text in writing iarw, preceded by an elision, as orthotone (and so y’ iartv). D. Sedley, ‘The Ideal of Godlikeness’, in Gail Fine (ed.), Plato 2: Ethics, Politics, Religion, and the Soul [Plato 2] (Oxford, 1999), 309— 28 at 313 n. 7, rightly corrects the common error whereby on dp-otoTaror (331 b 5) is translated as a plain superlative, ‘very like’. See e.g. W. K, C. Guthrie, Plato, Protagoras and Meno, Translated with an Introduction (Harmondsworth, 1956), 63. The error perhaps arises from taking on at B 5 to be a repetition of the otl at B 4. The neuter pronoun (on) is here used as an adverb, and serves to intensify the

superlative: on og-OLorarov, ‘as similar as can be’. For this use of on, see LSJ s.v. III. Although Denniston does not say so (cf. J. D. Denniston, Greek Particles, 2nd edn. (Oxford, 1959), 553 =roiV\ (*7)), it is difficult not to believe that the combination of -roi and ye (the latter particle omitted here by Stephanus, Bekker, and other early editors) gives added force to the first arm of the disjunction, 331 b 4-5: f/Toi ravrov y' eorlv hiKai6r'r)s ooiorrjn rj on ofxoLorarov. If SO, the point will be, not that ‘the same has Socrates’ personal preference, but simply that sameness is a stronger relationship than similarity. For the same combination of particles strengthening the first arm of a disjunction, see Ap. 27 C io-D i: tous Se Sai/aoraj

proi deovs ye rjyovfxeda rj

decov TratSaj; and Theaet. 189 E 3: we have to think of two things, rjroi ajxa ye rj ev jiepei. As so often, such shades of meaning can be brought out in English only by a change in the tone of voice. In the absence of any possibility of a literal translation, the distinction is most simply conveyed by adding an adverb to emphasize the first arm of the disjunction: ‘Justice is either actually the same as piety, or as similar

Denis O’Brien

64

The two theses whose presence is implied at this moment in the text (331

B

4-5) are therefore the two theses that I distinguished

earlier. If justice is ‘the same’ as piety (331

B

4: ravTov), it is because

the names of the individual virtues are only so many names of ‘one and the same thing’ (ib). If justice is ‘as similar as can be’ to piety (331

B

5: OTL o/aotoTarov), it is "because, although justice and piety

are parts of virtue (la), they are parts only in the way that different pieces of gold are parts (2b); we cannot have one virtue without having all the others (as in 3b). In the conclusion to Socrates’ first argument (331

A 6-b 8),

those

two theses are presented as the two terms of a disjunction (331 4—5: rjroL ravTov y’ ...

rj on ofxoioTaTOv).

B

For if either thesis is true,

Protagoras’ claim that the individual virtues are different from one another in the way that the parts of a face are different will be refuted. But the two counter-theses are not therefore the same. Protagoras is refuted by either one of two distinct theses (331

B

4-5; T]TOl . . . rj . . .).

(/) ‘Virtually the same’ {333 b 6) The implication, I would suggest, is no different when, two pages later, Socrates summarizes his earlier argument as the claim that justice and piety have been shown to be ‘virtually the same’ (333

B

6:

ayeSov n ravrov ov). The qualification (ayeSov n) is all-important.

‘Virtually the same’ straddles the difference between ‘the same’ as can be.’ Alternatively, we might think of adding a heavy-handed qualification (‘at least’/‘failing that’) to the second arm of the disjunction. Although the result is perhaps too paradoxical to rank as a translation, it does bring out the meaning of the Greek: ‘Justice is either the same as piety, or at least as similar as can be.’ Cf Ap. 27 c lo-D i: ‘Don’t we think of daimones as gods, or at least as the offspring of gods?’; Theaet., 189 E 3: we have to think of two things, ‘either simultaneously or, failing that, in [quick] succession’. ’ The conclusion ‘either the same, or as similar as can be’ (331

B

4-5) is of course

also determined by the nature of the argument that Socrates puts forward (330 b 6-331

B

8), where the criteria for identity and for similarity have not been clearly

distinguished. (Protagoras’ protests, 331

B

8-332

A

i, only make matters more con¬

fused.) But, if the distinction is not provided, it is also because it is not required, in so far as Protagoras will be shown to be wrong, whether the conclusion is couched in terms of identity (cf. 331

B

4: ravrov) or whether the,argument leads only to an

assertion of extreme similarity (cf. 331

b

5: on ofxoioTaTov).—Tempting though it

would be to enter into the details of the argument, I have restricted my remarks in the present essay to the relationship of the arguments within the structure of the dialogue as a whole. To have attempted to do otherwise would have more than doubled the length of the essay.

Socrates and Protagoras on Virtue

65

and ‘as similar as can be’. The two theses can be joined in this way simply because they are both alike opposed to Protagoras’ belief that the individual virtues are as different as are the parts of a face and that we can therefore have one part and not have another part. Taken in this way, ‘virtually the same’ (333 b 6) is a mere con¬ venient shorthand. The expression is not meant to obliterate the distinction that had earlier been marked by the disjunction ‘either the same or as similar as can be’ (331

B

4-5; fjTot . . . f) . . .). If

Socrates now summarizes that distinction by an expression that is not cast in the form of a disjunction, in claiming to have shown that justice and piety are ‘virtually the same’ (333 b 6: oxcSov tl ravTov ov), it is simply because he does not have to come down on the side of either of the two distinct theses, since either thesis will convict Protagoras of error. If justice is ‘either the same as piety, or as similar as can be’ (331

B

4-5), then justice and piety are ‘virtually the same’ (333

B

6). Contrary to Protagoras’ claim, neither one therefore can be had without the other.

{g) Socrates’recapitulation After the long diversion on Simonides, Socrates and Protagoras agree to reopen the debate on the parts of virtue (348 c i ff.). Socrates therefore recapitulates the earlier discussion (349 a 6-d i). The way in which the successive theses are arranged is now significantly different. Socrates first puts a single question in the form of a distinction between two theses, using the same pair of disjunctive particles that he had used in the earlier passage

{iroTepov

. . . rj . . .).

Are wisdom, temperance, courage, justice, and piety five names for a single object [rroTepov ravra, ir^vre qVra ovofMara, erri ivi npayiJiaTL iariv], or under¬ lying each of those names is there an individual nature and object [17 eKdaru) TOiv ovopLaTwv TovTwv VTTOKeLTal TLS tStos ovola /cat TTpdypia], SO that each has its own peculiar capacity [eyov iavrov Svva/uiv c/caarov], no one of them being like the other [ovk ov olov to erepov avrdiv to eVepov;]? (349 B 1—6) To this question, Protagoras’ reply (349 b 6-c i) is said to have been a firm choice in favour of the second thesis, explained as a choice between the same two images employed earlier of how it is that the individual virtues are parts of virtue.

Denis O’Brien

66

All the names that have been listed are parts of virtue, not in the way that parts of ^old are like each other and like the whole of which they are parts, but as it is with parts of a face, which are unlike the whole of which they are parts and unlike each other, since they have each c^f them an individual capacity [tSiat' eKaara Svvafxiv e'xovta]. (349 C 2—5) S

(h) The order of arguments in Socrates’ recapitulation Inevitably, since this is a summary, Socrates abbreviates the original discussion. He does so by running together what had earlier been distinguished as question and answer. This he is able to do by taking account of the choices that Protagoras had already made when he was first interrogated on the nature of virtue and the relation of its ‘parts’ (329 c

6-e 6).

Thus the single question of Socrates’ recapitulation (349

b

1-6)

presents an opposition between two theses: first, the thesis that the names of the individual virtues are only so many names for one and the same object (what had earlier been ib), and secondly, what had earlier been spelt out as the consequence of Protagoras’ choosing the comparison of the individual virtues to the parts of a face (2a), namely that each of the virtues has its own peculiar capacity (349

B

4-5: cyov iavTov hvvayav eVauTov, in the question that Socrates asks in his recapitulation, and 349 c 5: Ihlav cKaara hvvayav exovra, in Protagoras’ reported reply, both look back to 330 A 4: Surap-tv avroiv eKaoTov ISlav ex^O-

The second thesis therefore takes for granted that the individual virtues are parts of virtue (what had earlier been la) and passes directly to the consequences of Protagoras’ choice of how it is that they are parts. However, that choice is still given in terms of a distinction between parts of gold (what had earlier been 2b) and parts of a face (what had earlier been 2a), presented no longer as a question, but as a summary of Protagoras’ reply (2b: the image that Protagoras has discarded, 2a: the image that Protagoras agrees to). What had earlier been Socrates’ third question, whether we can have one virtue without another (3a) or whether our having one virtue entails our having all the others (3b), will be the bone of contention in the debate that follows (349

D 2

ff.). Protagoras fires

the opening shot by reaffirming his contention that people can have one virtue without the others (what had earlier been 3a), and

Socrates and Protagoras on Virtue

67

specifically that extremes of injustice, impiety, wantonness, and ignorance may accompany extremes of bravery (349

d

2-8).*

(z) The three theses: Protagoras and the opposition I return therefore to my initial conclusion. Despite Socrates’ tele¬ scoping the various possibilities in the later passage (cf. 349

b i-d

8), we may still distinguish three separate positions. Protagoras’ claim that the individual virtues are diflferent from each other, and that we can therefore have one without the other, is opposed to either one of two distinct theses. The more radical thesis is that the individual virtues are not parts of virtue at all; their names are only so many names of one and the same thing. The less radical thesis is that the individual virtues are indeed parts of virtue, but parts that differ from one another and from the whole only in the way that different-sized pieces of gold differ from one another and from the whole. On either thesis, we cannot have one virtue without having all the virtues (or the whole of virtue). In aiming to refute Protagoras’ claim, Socrates does not therefore have to state a preference for either the less radical or the more radical of the two theses, since Protagoras will be wrong on either count.

2. Penner and Vlastos From this outline of the structure of the debate it will be seen that I part company with both Penner and Vlastos. (I single out Penner and Vlastos for comment, since theirs seem to be the in¬ terpretations most frequently referred to in recent literature on the Protagoras.) Penner attributes to Socrates the thesis that the names * ‘Wantonness’ (cf. 349 D 7: akoXaaTOTaTovs) is the vice that corresponds to ‘tem¬ perance’ {aa)(j>poavvri), here as in the Gorgias (cf 507 C 6: d evavriws

Tip

awifipovi,

6 oiKoXaaTos) and often elsewhere, including Aristotle’s Nicomachean Ethics (3. 10, iii7’’23-iii9'’i8). The five virtues named at the beginning of Socrates’ recapitula¬ tion (349 B 1—2: wisdom, temperance, courage, justice, and piety) therefore reappear, in Protagoras’ opening restatement of his thesis, as four vices (349 D 6-8: injustice, impiety, wantonness, and ignorance), all of which, so Protagoras claims, even in their most extreme forms (hence the list of superlatives, 349 d 6—8: dSiKoiraTov^, dvoaiojTarovs, dKoXanroTaTOVs, dpLadcardrovs), are compatible with the exercise of courage, in even its highest form (hence, again, the use of a superlative, reinforced, this time round, by an adverb, 349 D 8: di/Spcwrarovs Se SiaiftepovTOJs). Protagoras has firmly nailed his colours to the mast.

Denis O’Brien

68

of the individual virtues are so many names of one and the same thihg (ib), but not the thesis that the individual virtues are like so many pieces of gold (2b).® Ylastos takes the opposite view. He attributes to Socrates the comparison of the individual virtues to pieces of gold (2b), but- not (or not initially) the thesis that the names of the individual virtues are so many names of one and the same thing (ib).‘”

(a) Benner’s interpretation Penner’s interpretation (Socrates’ position is given by ib and not by 2b) is presumably the cause or consequence of his taking Pro¬ tagoras’ thesis to be in part compatible with the image of pieces of gold (2b).“ That is not at all how Socrates presents Protagoras’ position. ’ T. Penner, ‘The Unity of Virtue’. I shall quote the pagination of the reprinted version of this article in Fine (ed.), Plato 2, 78-104. The article was originally published not, as the reprint claims (p. 78), in the Philosophical Review, 80 (1971), but in the Philosophical Review, 82 (1973), 35-68. Penner takes what I have called ib (329 c 8-d I, repeated in Socrates’ recapitulation at 349 b 1-3) to be a statement of Socrates’ own opinion. See ‘The Unity of Virtue’, 90-1, and especially 90 n. 19. Penner’s account of the comparison between the individual virtues and pieces of gold has to be extracted from a footnote, 90 n. 18. G. Vlastos, ‘The Unity of the Virtues in the Protagoras’, first printed in the Review of Metaphysics, 25 (1972), 415-58, reprinted ‘with substantial additions and corrections in both text and notes’ in the same author’s Platonic Studies (Princeton, 1973), 221—65. What is commonly referred to as a second edition of Vlastos’s Platonic Studies, and what is indeed described as such on the outer cover of the paperback edition, is none the less stated by its publisher, on the verso of the title page, to be a ‘Second printing, with corrections’ (Princeton, 1981). We are told (xi—xiii: ‘Preface to the second printing’) that the new material consists of three short essays and a large number of ‘starred notes’, where Vlastos mainly takes issue with the criticisms that had been made of the essays published in his 1973 edition. There are, however, other significant changes to the text not acknowledged in the ‘Preface to the second printing’ (see n. 92 below), although these have been made in such a way that the pagination of the main body of the original text remains the same in the two publications (1973 and 1981). For the comparison of the individual virtues to pieces of gold (what Vlastos calls the Similarity thesis), see Platonic Studies, 228—31 and 246—52. For Socrates’ initial statement of the thesis that the names of the individual virtues are so many names of one and the same thing (what Vlastos calls the Unity thesis), see ibid. 224—8. " Penner, ‘The Unity of Virtue’, 90 n. 18. I hope that I have disentangled aright the kaleidoscopic list of references in this footnote. In less than six lines of text we are given eleven references to the text of Plato variously correlated with fourteen references to Penner’s own numbered propositions. As if that were not difficult enough, the overlapping line references sometimes seem to stop and start at random, so that the reader is left wondering whether, as would be expected, they follow the lineation of Burnet’s edition, or whether perhaps they have been taken from the

Socrates and Protagoras on Virtue

69

Protagoras claims that the individual virtues differ from one an¬ other, in such a way that we can have one without another (3a), because he agrees with Socrates that this follows ‘clearly’ (330 i: rj hrjXa

8rj

otl

b

. . .) from the comparison he has already adopted

between the parts of virtue and the parts of a face (2a). In choosing that image to represent in what way the individual virtues are ‘parts’ of virtue, Protagoras has discarded the alternative image which would compare the individual virtues to pieces of gold (2b), just as, in claiming that we can have one virtue without another (3a), he has discarded the rival thesis by which we cannot have one virtue without having all the others (3b). At no point does Protagoras suppose that he might be able to adopt both items in the two successive choices that he is faced with (2a as well as 2b, 3a as well as 3b). On the contrary, he is summoned to choose between the two images (329

d

4-8), and does so

(d 8-e

2). He is summoned to choose between the two theses that would follow from them (329

e

2-4), and again he does so

(e

5-6). There is

no indication that the two images (2a and 2b) might be compatible, nor that the two theses (3 a and 3 b) might be compatible, nor that either image is thought to be compatible with the thesis attached to its rival. lineation in some other (unspecified) edition of the dialogue. See also the footnote following this. The only truth I can see in Fenner’s claim of ‘compatibility’ would lie in the point (already made in sect, i above) that both images (aa and ab) and both theses (3a and 3b) share a common premiss, namely that the individual virtues are so many ‘parts’ of virtue (la) and that their names are not only so many names of ‘one and the same thing’ (ib). But that common premiss cannot be what Fenner intends as the feature of Frotagoras’ theory which he claims is ‘compatible’ with the image of pieces of gold. For in the footnote that I have quoted (‘The Unity of Virtue’, go n. 18) Fenner states that ‘(P2) would be compatible with the “parts of gold” model’. (Fa), in Fenner’s notation, is one of four propositions, numbered (F[rotagoras] i) to (P[rotagoras] 4), which are said to represent ‘progressive articulations’ of Protagoras’ position (cf. ‘The Unity of Virtiie’, 90). (Pi), in Fenner’s notation, contains both the statement that wisdom, temperance, justice, piety, and bravery are parts of virtue, and the comparison with parts of a face (aa in my numbering), as opposed to the comparison with pieces of gold (my ab). (Pa), in Fenner’s notation, is stated in the form ‘Wisdom^ttemperance^tjustice^ipietyjtbravery’. Since Fenner’s series is a ‘progressive’ series and since (Pa) is listed after (Pi), (Pa) must be supposed to fall somewhere in the text following the choice between the parts of a face and pieces of gold (329 d yff.). Fenner’s ‘compatibility’ cannot therefore be supposed to link the image of pieces of gold (my ab) with the premiss that has already been given (329 C 6-8) as common to both aa and ab, namely that the individual virtues are ‘parts’ of virtue (my la).—But what precisely does Fenner’s (Pa) refer to? In the first paragraph of this section, I have written that, on Fenner’s interpretation.

Denis O’Brien

70

(b) The origin of Benner’s error If Penner thinks that Protagoras’ thesis is in some way ‘compat¬ ible’ with the image of pieces of gold, it is presumably because he has chosen to identify Socrates’ position exclusively with the second item in the first choice, that Protagoras is presented with, and therefore with the claim that the names of the individual virtues are only so many names of one and the same thing (ib). That thesis (i b) is opposed to both versions of how the individual virtues are parts of virtue. But from that joint opposition it does not at all follow, as Penner apparently thinks it does, that the two different accounts of how the individual virtues are parts (2a and 3a, as distinct from 2b and 3b) are intended to be in some way ‘compatible’. In particular, there is no indication in the text that the image of pieces of gold (2b), which Protagoras refuses, is intended to be taken as at all compatible either with the image of parts of a face (2a), which Protagoras adopts, or with the thesis that ‘clearly’ (cf. 330 B i: hrjXa) follows from that image (3a), and which Protagoras also adopts.”

(c) Vlastos’s interpretation Vlastos’s interpretation is, initially, the obverse of Penner’s. Vlastos sees very clearly that the second item in the choice that Protagoras is initially presented with (ib), which he calls the Unity thesis, ex¬ cludes both versions (2a and 2b) of how the individual virtues are Protagoras’ thesis is ‘in part compatible’ with the comparison of the individual virtues to pieces of gold, hoping, by this catch-all phrase, to include whatever part of Protagoras’ theory is represented by (P2). I have sought to show that, on the contrary, the image of pieces of gold cannot be taken to be ‘compatible’ with either of the two main planks of Protagoras’ position (2a: the image of parts of a face, 3a: the conclusion that is drawn from that image). But the reader who wants to know just where Penner’s (P2) is situated in the text, and therefore just which part of Protagoras’ thesis Penner himself holds to be ‘compatible’ with the image of pieces of gold, will have to wrestle with the references given in n. 34 below. ” Will Penner perhaps reply that the image is ‘compatible’, even if Plato does not present it as being so? Penner is one of those authors who think to do Plato a service by stringing together series of numbered propositions, culled from different places in the text, and then looking to find their logical coherence or lack of it. Such a procedure almost inevitably blurs the distinction between what might, or might not, have seemed cogent enough to Plato (or what Plato might, or might not, have wanted to present as persuasive at any one point in the course of his dialogue) and what the modern critic judges to be logical or illogical.

Socrates and Protagoras on Virtue

71

parts of virtue. Since he attributes one of those versions (2b: pieces of gold) to Socrates, Vlastos therefore exonerates Socrates from any responsibility for the initial statement of the Unity thesis (ib: 329 c 8-d I),

which in any case he holds to be ambiguous, if not ‘prepos¬

terous’.*'* But that is only a first step. Later in the dialogue (349

b

1-3), Vlastos distinguishes a new and different version of the Unity thesis, which he does not hold to be ambiguous or preposterous and which he claims that Socrates does endorse.*^ Vdastos’s distinction turns on the difference between the form of words that Socrates uses in his opening question and the form of words adopted when Socrates and Protagoras take up their debate again after the long interlude on Simonides. When Socrates first asks whether the names of the individual virtues are so many names ‘of one and the same thing’, he uses a simple genitive (329 c 17

ravV ioTiv a vvvhrj eyoj eXeyov rravra ovog^ara

tov

avrov ivos

In the later text, he uses a preposition with the dative (349 aocfiia Kal aoj(f>poavvri kcil duSpeta Kal hiKaioovvq Kal

ooiottjs,

8-d i : ovto?).

B

1-3:

rroTepov

Tavra, rrevre ovra ovopLara, cttI eVt TrpdypbaTi eartv).

Vlastos holds that, on the first construction (329 c

8-d i),

the

names are used as ‘proper’ names, so that the virtues so named are in reality all one and the same virtue and as such indistinguishable. On the second construction (349

b

1-3), Vlastos claims that only

one name at a time is taken as a ‘proper’ name and that, whenever one virtue has been taken as a proper name, the remaining names act only as so many ‘descriptive predicates’, so that bravery (as a proper name) is just, wise, moderate, and pious, justice (as a proper name) is brave, wise, moderate, and pious, and so on, each of the virtues therefore retaining its own character, but being ‘interpredicable’ with all the others.*'’ Relying on that distinction, Vlastos claims that the earlier text (329c

8-d I

=ib) has been deliberately recast to accommodate a

version of the Unity thesis which the original wording of Socrates’ question had excluded and which Socrates can now (349

b

1-3)

subscribe to.*’ Vlastos, Platonic Studies, 224-8. ‘Preposterous’: see ibid. 227 (in the 1973 printing only). How well-founded or not that judgement may be will be seen in sect. 8 below. ” Vlastos, Platonic Studies, 234-46. See also the anticipation of Vlastos’s distinc¬ tion, ibid. 225—6. Ibid. 234-46. I repeat Vlastos’s own terminology (‘proper name’ and ‘descrip¬ tive predicate’), ibid. 225 n. 9.

” See again ibid. 234-46.

Denis O’Brien

72

{d) A curious conclusion i



I hope to have presented Vlastos’s interpretation as simply and as persuasively as I can, for we have only to probe beneath the surface to see that his account of the matter is far from simple and very far from persuasive. ' ^ Vlastos definitely states, of the first version of what he calls the Unity thesis (ib in my numbering), that, ‘if put in this ultra-strong form, Socrates would not have championed it’.' * Vlastos also claims, no less definitely, that his ‘reinterpreting’ the Unity thesis and what he calls the Similarity thesis (the comparison with pieces of gold) ‘rids each of their unacceptable features’, with the result that both theses become ‘complementary expressions’ of Socrates’ own be¬ lief.So far (perhaps) so good. But quite how far is Vlastos’s ‘rein¬ terpretation’ meant to take us? The theory that the individual virtues are ‘parts’ of virtue, so Vlastos tells us, was ‘standard Socratic doctrine’

Does it then fol¬

low, as Vlastos would seem to imply at one point, that the denial of that theory (the denial that the individual virtues are parts of virtue) was one of the ‘unacceptable features’ in the original statement of the Unity thesis (cf. 329 c

6-d i)?^'

If it was and if, in the later version of the Unity thesis (349

b

1-3), that feature has somehow been got rid of, then it will follow that the so-called Unity thesis, in its later manifestation, counts the individual virtues as so many ‘parts’ of virtue. But, if that is so, then what remains of the original form of the Unity thesis in Socrates’ later statement of the theory (at 349

b

1-3)?

When Socrates first asked Protagoras (329 c

6-d i)

whether the

individual virtues were so many parts of virtue (la), or whether their names were only so many names of one and the same thing (ib), the implication was clearly that, with the latter option (ib), the individual names, since they were only so many names of ‘one and the same thing’, were therefore no longer to be counted as ‘parts’ of virtue. If, following Vlastos’s revised version of that same thesis (ib), the individual virtues are to be reintroduced as so many ‘parts’ of virtue, shall we still be able to distinguish two theses (la ** Ibid. 225. Ibid. 252 (cf. 234ff.: ‘Reinterpreting the Unity and Similarity Theses’). Ibid. 225 (in italics in the original). For the seeming implication, see the first sentence in the main text of Vlastos, Platonic Studies, 226.

Socrates and Protagoras on Virtue

73

and ib) in the question that Socrates first puts to Protagoras (329 c 6-d i)?

If not, then we end up with the curious, and surely unpalatable, conclusion that Vlastos’s supposed ‘reinterpretation’ of the Unity thesis collapses the very distinction on which the original formula¬ tion of that thesis was founded.

(e) A strained interpretation Not only is the logic of Vlastos’s distinction between the two pas¬ sages suspect; even grammatically,' his revision of the so-called Unity thesis (the distinction between proper names and descrip¬ tive predicates) places an impossible strain on the expression IttI evl TTpayfiari in the later text. (Socrates asks whether wisdom, temper¬ ance, courage, justice, and piety are five names ‘for a single object’, 349

B

3: iirl evl TrpdypLaTL. See also

B

6: eTrl ivV)

On Vlastos’s interpretation of the passage, the ‘single object’ (cf. 349

B

3:

i-nl €vl TTpdypiaTi.)

will not be the same single object for all

five names (the five virtues listed at 349

b

1-3). On the contrary, we

shall have to envisage a succession of single objects (each one of the individual virtues taken in turn as a ‘proper name’), and therefore in effect many things and not a single thing.^^ Conversely, and even more awkwardly, the ‘five names’ (cf. 349

B

1-3: -n-evre ovra ovopLara), since they have to take it in turn, so to speak, to provide the ‘single object’ of which the remaining names will be so many ‘attributive predicates’, are never able, all of them, to fulfil the same role at the same time. They can never all be used, at the same time, as attributive predicates, unless we add to the list a virtue whose name is not included among the five. Plato can hardly have intended so contorted a reading of his text. On the contrary, Plato clearly means that, if we adopt this thesis, then there is ‘one object’,, and only one, which all five names ‘apply to’ (349

B

2-3), in exactly the same way that, in the earlier passage

(ib), the names of the individual virtues are ‘all names of one and the same thing’ (329

D

i).

I conclude that the difference in wording between the two for¬ mulations of the so-called Unity thesis (329

C 8-d i

: TavV iariv . . .

See again ibid. 234-46 (summarized above). Each of the individual virtues is taken in turn as a ‘proper name’, to which the remaining four names are attached as so many ‘attributive predicates’: bravery (as a proper name) is just, wise, moderate, and pious, justice (as a proper name) is brave, wise, moderate, and pious, etc.

Denis O’Brien

74 Tov avTov

ivos

?, and 349

ovto

3-

npaynaTL

eariv) cannot bear

the

At the time, Socrates had carefully checked that his interlocutor did indeed intend to include both bravery and wisdom as ‘parts’ of virtue

(329 E 6-330 A 2),

‘different’ therefore ‘from one another

and from the whole’ (to repeat the wording implied by Protagoras’ endorsement of the comparison between the parts of virtue and the parts of a face:

329 D 4-E 2,

and spelt out in Socrates’ recapitulation:

c 4-5). The preliminary statement of Protagoras’ position had even

349

Denis O’Brien

76

ended with a solemn roll-call of the five new ‘parts’ of virtue, and' with the ringing declaration that no one of the five, accord¬ ing to Protagoras, is ‘like’ any other (330

B

3—6: oOScv apa iorlv

twv

Trjs dperrjs pLOpLCJV dXXo otov eTTiaT^pLrj, ov8’ oiov Suv rrjs dperrj? • ■ •)>

was like any other

(a 5~6: ov8ev ecjirj

[sc.

a

5:

one of which

d npcvrayopas] etvai ro i'repov

otov TO i'repov) and each one of which had its own peculiar dynamis (a

6—7: I8tav Se avrov eKaarov eyetv 8vvapi.iv). This, after all, is exactly what Protagoras had agreed to when he

first included wisdom and bravery in the list of virtues (359 a 5—6: ouSev . . . etvai ro erepov otov ro erepov

looks back to 330 A 8—B i: ovk

eariv ro erepov otov ro erepov, while 359 A 6—7; I8tav 8e avrov eKaarov eyetv 8vvapiiv

looks back to 33°'^ 4- 8vvap.iv avriLv eKaarov I8tav eyei,

already summarized at 349 and at

C

5 as:

B

4-5 as: eyov eavrov 8vvap.Lv eKaarov,

I’Siav e/caara 8vvap.iv eyovra).

In repeating, once again,

towards the end of the dialogue, Protagoras’ claim that there are five ‘parts’ of virtue, each of which is different from the other and each of which has its own peculiar dynamis (359

a

5-7; cf. 349

B

i—c

5), Socrates is keeping scrupulously to a re-statement of precisely the points that Protagoras had already agreed to. ‘Wisdom’ (aoijyta) and ‘knowledge’ {iTnarp^y]) appear to be used more or less in¬ terchangeably, here (330 A i and 2: aota, B4: iniaT-qpri) as not infrequently elsewhere in the Protagoras (and in other early dialogues).

Socrates and Protagoras on Virtue

77

(d) The fifth virtue None the less, the addition of wisdom as a fifth virtue is not at all the simple move that commentators have taken it to be. For ‘wisdom’ {aofia), in the guise of ‘knowledge’ {iTTiaTruirj), so we learn on the final page of the dialogue, is what Socrates has been aiming all along to establish, not as a part of virtue, but as the whole of virtue. (The Argument, personified, addresses Socrates with the words, 361

B

5~7-

fav-fjoCTaL [sc. rj apeTrj] imoTrjiJL'q oAov,

EojKpares .

d)?

cru arrevheis, d)

. .)

When therefore ‘wisdom’ is listed as a ‘part’ of virtue alongside all the other ‘parts’, as it is in Socrates’ recapitulation of the earlier argument (at 349

A

8-c 5 and at 359

2-7), we know that whatever

A

thesis is said to include the ‘five parts’ of virtue cannot be intended as a statement of Socrates’ own belief. No matter that Protagoras had earlier said that wisdom was the ‘greatest’ of the parts of virtue (330 pLopi(jjv)d*

A

2: p-eyiarov ye rj aofia tcov

Protagoras’ assumption that wisdom is a ‘part’ of virtue

(however ‘great’) drives an immovable wedge between his concep¬ tion of virtue and the thesis that Socrates is here said to be ‘aiming’ at (cf 361

B

6: arrevheLs), where wisdom or knowledge is not a ‘part’

of virtue, but the ‘w’hole’ of virtue.

{e) Is Socrates ‘unnecessarily fussy’? In his younger days, Vlastos had glimpsed that something signi¬ ficant was afoot at the moment when bravery and wisdom were added to the virtues (329

E

5-330 A 2). Socrates, he wrote, ‘makes a

special point of the later addition of wisdom and courage to the list’ of the initial three virtues. However, the glimpse is no more than a glimpse, for Vlastos continues: ‘But he is unnecessarily fussy, for Protagoras never said that the first triad are the only virtues.’^® It is true enough that Protagoras’ ‘first triad’ of virtues is tagged as an incomplete list; Protagoras, so Socrates tells us, spoke of justice, temperance, piety, ‘and all the rest’ (329 c 3-6: . . . /cat Tidvra ravra). None the less, the ‘fuss’ that Socrates makes (at 329 5-330

A

E

2) is not at all ‘unnecessary’. On the contrary, Protagoras’

seemingly light-hearted introduction of wisdom as on a par with ^'* Repeated in the Epinomis (977 d 3: Traaps dperijs

to

jjLdyiaTov /xepos [sc. (jo(/ita]).

Plato’s Protagoras, trans B. Jowett, rev. M. Ostwald, ed. with introduction by G. Vlastos (New York, 1956), xi n. 15.

Denis O’Brien

78

the other ‘parts’ of virtue marks irrevocably the gap between his conception of virtue and that of Socrates.^* In his later years, Vlastos lost even the glimmerings of the aware¬ ness that he had shown in his earlier publication. He now writes: ‘One would certainly have expected Socrates to join Protagoras in maintaining that each of the five'names applies to a distinct “thing” or “essence”.’” On the contrary, we have no reason at all to expect that Socrates would have agreed to a formulation that put the parts (justice, temperance, piety and bravery) on a par with the whole (wisdom/knowledge).

(/) Pieces of gold Protagoras’ assumption that wisdom is a part of virtue marks a significant difference between the first and the second occurrence of the comparison of the individual virtues to pieces of gold. When that image is first introduced (329

D 6-8),

wisdom has not yet been

included as one of the parts of virtue (cf. 329 c 2—6). In the later passage (349 (349

B

b

1-6) wisdom heads the list of the individual virtues

i). Its inclusion is therefore implied in the comparison with

pieces of gold (349 c 2-3) no less than in the comparison with parts of a face (349 c 4-5). How can we explain such a discrepancy? How is it that the same image (pieces of gold compared to the individual virtues) can be presented without knowledge as one of the ‘parts’ of virtue in the earlier passage (329

d

6-8), and with knowledge (or wisdom) in¬

cluded as one of the ‘parts’ of virtue in the later passage (349 c 2—3; cf. 349

B i)?

An additional reason for the ‘fuss’ (at 329 e 5-330 A 2) is, of course, that the relationship of wisdom and courage will be taken up as the main bone of contention in the second part of the debate (349 e i ff.), triggered by Protagoras’ claim that, however closely related the other virtues may be, courage stands out as compatible with extremes of injustice, impiety, wantonness, and ignorance (349 d 2—8; repeated by Socrates: 359 b 2—6), But it is Protagoras’ obviously unconscious downgrading of wisdom to being a ‘part’ of virtue (albeit the ‘greatest’ part, 330 A 2) that is the fundamental obstacle to his sharing Socrates’ conception of virtue, and in particular to his appreciating Socrates’ recognition of knowledge as the ‘whole’ of virtue (cf. 361 B 5-6).

VX&stos, Platonic Studies, 2z6

R. Duncan comments on the difference between the two lists (329 C 4—5 and 349 B 1-2), in his article ‘Courage in Plato’s Protagoras’, Phronesis, 23 (1978), 21628. He writes (p. 220) that the second list ‘differs in one respect from the previous formulation: dvSpela is right up front’. But Duncan fails to note that wisdom is also a newcomer and that it is presumably even more ‘up front’ (whatever that curious expression might mean) in so far as it heads the second list (349 b 1-2).

Socrates and Protagoras on Virtue

79

The difficulty is insurmountable if, as Vlastos supposes, the im¬ age of pieces of gold is intended, throughout the dialogue, to repre¬ sent Socrates’ own conception of the relation between the parts of virtue. If Socrates is aiming to show that knowledge is the ‘whole’ of virtue (cf. 361

B

5-7), then the second occurrence of the image

(349 c 2—3), where knowledge has been given pride of place among what are explicitly said to be the ‘parts’ of virtue (349

B

1-2 and

c 2), cannot be taken as a straightforward statement of Socrates’ own belief. Do we therefore have to suppose that it is only the first occurrence of the image (329

D

6-8).that Socrates would have been

willing to endorse? But, if that is so, then how can we explain that, when the image makes its second appearance (349 c 2-3), Socrates claims to be repeating the very question (cf. 349

B i: to ipcoT-q^a

roSe) that he

had asked in the earlier passage? Is Socrates leading us up the garden path, lulling us into the false belief that the addition or the omission of knowledge as a part of virtue is a merely incidental feature of the comparison?^'’ {g) A changing image Those questions can be answered satisfactorily only if we abandon Vlastos’s assumption that, of the two images put forward to ex¬ emplify the relation of the parts of virtue (2a: parts of a face, 2b: pieces of gold), Socrates must be supposed to approve the image that Protagoras is led to reject (2b). If that were so, then Socrates would contradict himself, since the second occurrence of the image that he is supposed to endorse (2b) includes knowledge as one of the parts of virtue, contrary to The distinction between the two versions of the image (without ‘knowledge’: 329 D 6-8, with ‘knowledge’: 349 c 2-3) is particularly awkward if, like Vlastos (Platonic Studies, 234—46), we also believe that there are two versions of the Unity thesis, and that Socrates accepts’ the second version (349 b 1—3) and not the first (329 c 8-d i). For, if we suppose that Socrates accepts one or other version of the image of

the gold pieces, in Vlastos’s nomenclature the Similarity thesis, then we shall have to suppose that he accepts the first version (329 D 6—8: without ‘knowledge’) and not the second (349 C 2-3: with ‘knowledge’). In the opening series of questions and answers (329 C 6—E 6), Socrates would therefore endorse in petto the Similarity thesis (2b), but not what Vlastos calls the Unity thesis (ib), whereas in the recapitulation (349 A 6-d i) it would be the other way round: Socrates would endorse in petto the new version of what Vlastos calls the Unity thesis (ib), but not the new version of the Similarity thesis (2b). Something has obviously gone badly awry with a reading of the text that would require Socrates to play cat and mouse with the reader in this way.

8o

Denis O’Brien

what we later learn is Socrates’ own belief (349 c 2-3; cf. 349 B i). But if the image of pieces of gold is intended primarily as a foil to Protagoras’ choice of image, and not, or not primarily, as a statement of Socrates’ own belief, then the addition of knowledge as one of the parts of virtue becomes not merely tolerable, but indispensable. Protagoras counts knowledge (or wisdom) as one of the parts of virtue. Knowledge therefore has to be included in the comparison between the individual virtues and the parts of a face, since that is the comparison which Protagoras accepts as illustrating the relation between part and whole. Inevitably, therefore, knowledge has also to be included in the comparison that Protagoras rejects, where the parts of virtue are like so many pieces of gold. Only so will the two images match, respectively, the thesis that Protagoras accepts (2a; the comparison with parts of a face) and the thesis that he rejects (2b: the comparison with pieces of gold). Unless the parts of virtue are the same for both images, then the second image (2b) cannot be held to deny what the first image (2a) asserts.

{h) Wisdom as a ‘part’ of virtue Seen in this light, the presence of wisdom or knowledge in the second occurrence of the comparison with pieces of gold (329

d

6-8) and not in the first (349 c 2-3), far from being a disreputable legerdemain, designed to conceal from the reader Socrates’ own ultimate intentions (cf. 361

b

5-7), becomes instead a necessary

consequence of Protagoras’ putting forward knowledge as a candi¬ date for one of the ‘parts’ of virtue. Once knowledge has been listed as one of the ‘parts’ of virtue, then it must be included both in the image that Protagoras chooses to illustrate his belief (2a: parts of a face) and no less in the image which serves to illustrate the relation of part to whole which Pro¬ tagoras rejects (2b: pieces of gold). The two images jointly illustrate (the first image, 2a, by its acceptance, the second image, 2b, by its rejection) the thesis on the relation of part to whole adopted by Protagoras. But from this it does not follow that Socrates’ own personal belief can be simply ‘read off’ from the negation of the thesis that he sets out to refute (and therefore taken to imply an acceptance of 2b consequent upon the refusal of 2a). For both the image that Protagoras accepts (2a: parts of a face) and the later version of the

Socrates and Protagoras on Virtue

8i

image that he rejects (zb;’pieces of gold) share a feature which Socrates himself would not have agreed to, namely that knowledge is one of the ‘parts’ of virtue, and not the whole of virtue (cf. 361B 5-7)d°

(f) Plato’s ‘muddled thinking’? I return to the conclusion of Socrates’ earlier argument, to his claim that ‘justice is either the same as piety, or as similar as can be’ (331

b

4-5), repeated as the claim that justice and piety are ‘virtually the same’ (333

b

6). Vlastos and Penner'wrestle in vain with that form

of w'ords. Vlastos writes of Plato as ‘prey to the muddled thinking which surfaces in the phrases “either the same thing or extremely similar” and “almost the same”’d‘ He even writes of ‘the residual unclarity of his [Plato’s] thought which shows through the cracks when he says that Piety and Justice are “either the same or extremely similar” (331

b)

and are “nearly the same” (333

b)’.

An ‘unclarity’, Vlastos

adds, ‘which wdll dog his steps, threatening to trip him up sooner or later’ Penner sees, in the same form of words, an ‘admission’ that the argument put forward can prove only similarity, and not identity/^ The implication of Penner’s judgement is almost more devastating than the words I have quoted from Vlastos. If, as Penner believes, Socrates knows that he has proved only similarity, and not identity, then why does he claim to have proved either identity or similarity (33

I B

4-5: T7Tot . . . T/ . . .)? Is this an example of muddled thinking

or, worse still, of bad faith?^^ I perhaps need to point out that refusal of both aa (parts of a face) and of the later version of ab (pieces of gold), in so far as both images include wisdom as one of the ‘parts’ of virtue, does not therefore make the second image ‘compatible’ with the first (cf. sect, a aboveThe fact that, certainly in the later passage (349 c 3-5), Socrates would not choose either image does not in any way imply that his interlocutor has to (or is able to) choose both. Vlastos, Platonic Studies, 43a (taken from the ‘starred notes’ of the 1981 pub¬ lication). Ibid. 433 (again taken from the ‘starred notes’ of the 1981 publication). ” Penner, ‘The Unity of Virtue’, 91 n. ai. But see the footnote following. Once again, I hope I have interpreted aright Fenner’s numbered propositions, built up, or so at least it seems to me, from a miscellany of disparate references. Thus in the footnote I have quoted (‘The Unity of Virtue’, 91 n. ai), Penner refers to the proposition that he calls (Pa), and which he has earlier (p. 90) stated in the form ‘WisdomT^temperanceTtjustice^ipiety^ibravery’. That proposition has

82

Denis O’Brien

{]) Penner and Vlastos: the false assumption Fortunately, that is not a question that we have to answer. For both Penner and Vlastos share a common assumption, which is false. Both Penner and Vlastos think that the opening session of question and answer (329 c

6-e 6)

contains a straightforward statement of

the thesis that Socrates himself is intended to adopt. Admittedly, the choice of thesis is different. Penner chooses as a statement of Socrates’ own belief the thesis that the names of the individual virtues are so many names of one and the same thing (ib). Vlastos chooses (at least to start with) the alternative thesis: the individual virtues are parts of virtue (la), but they differ only as do so many different-sized pieces of gold (2b). In both cases the error lies, not in the choice that is made, but in the assumption that Socrates must be supposed to endorse one or other of the two theses that Protagoras has discarded. On the contrary, in the opening series of questions and answers, Socrates does not state a preference for either of the two theses that are opposed to the position that Protagoras adopts. Socrates therefore quite properly presents the conclusion to his opening argument in the form of a disjunction (‘either . . . or . . .’, 331

B

4-5:

rjTOL

. . .

i] . . .). Protagoras has rejected both theses. An argument that will lead to either one of the two theses that Protagoras has rejected will therefore prove Protagoras to be wrong. Far from being an example of ‘muddled thinking’, Socrates is following the logical structure of the arguments that had been to be recovered (‘The Unity of Virtue’, go n. i8) from the following three texts; (i) 330 A 3-B 6; (2) 332 A 2-5; (3) 349 B 3-6. Those three texts contain respectively: (i) a comparison of the individual virtues to parts of a face and the conclusion that we can have one part of virtue without another; (2) a statement of the opposition between aota and a(j>poavvr)', (3) in Socrates’ recapitulation, the choice between taking the names of the virtues as five names ‘for a single object’ and supposing that, underlying each one of the names, there is an individual nature and object. In the first of those three texts, we are told that Fenner’s (P2) is joined to another numbered proposition (P3), stated as ‘The parts of virtue, like the parts of a face, each have their own peculiar power (tStav SvvapLiv)', both the parts themselves and their powers are unlike each other’, while the two remaining texts are cited to show ‘the distinctness of (P2) and (P3)’. As if all that were not already confusing enough, we are told a page later, in the footnote I originally quoted (p. 91 n. 21), that the same proposition, (P2), is ‘at issue’ at 331 A i-b 4, a reference which corresponds to the concluding lines of the argument leading to the claim that ‘justice is either the same as piety, or as similar as can be’ (words which in fact run over from b 4 to B 5). I hope that others have less difficulty than I do in picking their way through

Penner’s maze.

Socrates and Protagoras on Virtue

83

presented at the beginning of the discussion (329 c

6-e

6). The

‘muddle’ is to be laid at the door of those modern readers of the dialogue who have added their own assumptions and their own pre¬ suppositions to the carefully structured arguments that are to be found in Plato’s text.

{k) Hartman’s thesis For the sake of simplicity and economy, I have limited critical comments in the body of my text to two well-known essays by Vlastos and by Penner. Whatever their faults may be, those two writers do at least choose to associate what they suppose to be Socrates’ own theory with theses that are clearly opposed to the position that Protagoras is made to adopt. Not so Barbara Hartman.^® Hartman attempts to explain in detail how the two images (2a: parts of a face, 2b: parts of gold) are related to Socrates’ own conception of virtue. But, very curiously, Hartman favours for this purpose the first image (2a: parts of a face) over the second (2b: pieces of gold). In particular, she takes the assertion that each part of the face has its particular dynamis (330 A 4-B 3), and more especially Plato’s explanation of what she claims to be the dynamis of sight in the Timaeus (45

b

2-46 c 6 and 67 c 4-68

d

7, where the word dynamis does not occur), to be intended as a blueprint for what she argues is Socrates’ own belief in virtue as a single dynamis to which the individual virtues are subordinated (in the way that colours are subordinated to the dynamis of vision). In pursuing the detail of this interpretation, Hartman appears to have lost sight of the point that the image of parts of a face illustrates the theory that Socrates seeks to oppose. In that image, each part of the face is taken to have its own dynamis, and the multiplicity of the different dynameis is taken to show how it is that the individual virtues ar^ different from one another and different from the whole. This is so, when the dynameis are first introduced (330 A 4-B 3), when they reappear in Socrates’ summary (349 and c 4-5), and again at the end of the dialogue (359

A

b

3-6

5-7). In each

of those three passages, the account of the dynameis is introduced to illustrate exactly the feature of Protagoras’ conception of the relation of part to whole which Socrates is at such pains to disprove, ‘How the Inadequate Models for Virtue in the Protagoras Illuminate Socrates’ View of the Unity of the Virtues’,

16(1982), 110-17.

84

Denis O’Brien

namely that we can have one part of virtue without having all the other parts. If Hartman fails to see that point it is because, without the slight¬ est encouragement from the text, she has singled out one of the specific dynameis, the dynamis of sight, to represent not, as it does in the dialogue (330 A 4-B 3), o-ne part of virtue, but the whole of virtue. As soon as that crucial move (that crucial error) has been made, the die is cast. The image that Socrates introduces to il¬ lustrate the theory that Protagoras will adopt is wrenched from its original purpose and recast as an illustration of the thesis that Protagoras is said to reject.

(/) Traditional interpretations of the Protagoras I conclude that the use that Hartman would make of the image of the parts of the face and their dynameis, as an illustration of how it is that, if we have one part of virtue then we have all the other parts as well, is just the opposite of the conclusion that in its context the image is intended to sustain. Hartman’s interpretation, all unwittingly, sounds the knell of traditional interpretations of the Protagoras. Not only has either one of the two distinct theses opposed to the position that Protagoras is led to adopt been pressed into service as an exclusive statement of Socrates’ own belief (Penner, Vlastos). Socrates even ends up saddled with the image designed to illustrate the very thesis that he had set out to oppose (Hartman). Please note: in criticizing Hartman’s thesis, my point is not that ‘seeing’ could not be used to illustrate the theory whereby if we have one virtue then we have all the virtues, but that in the text of the Protagoras it does not do so.—It so happens that Ariston of Chios (as reported by Plutarch, Virt. mor. 2, 440 f) does include sight as one illustration among several of how virtue that is ‘one in essence’ is diversified by its relationships (see the summary of Ariston’s theory given in sect, g below). But Hartman should not take false comfort from Ariston’s use of the comparison. Ariston’s comparison of virtue to seeing and Plato’s comparison in the Protagoras do not at all serve the same purpose. Ariston’s point is that, although ‘one in essence’, virtue may be diversified, in different relationships, as prudence, temperance, or justice, in the way that vision might be called ‘white-vision’ (AevKoOear), when it is exercised in relation to white things, or ‘black-vision’ (/neAarde'av), when it is exercised in relation to black things. Ariston’s comparison will therefore be most naturally taken to imply that we cannot have one virtue and not have all the others (in so far as it would be difficult to imagine that we might be able to see black things and not be able to see white things). Not so the Protagoras (330 a 4-B 3), where the comparison with vision is designed to lead to exactly the opposite conclusion. The dynamis of sight, as included in the comparison of the individual virtues to parts of a face (2b),

Socrates and Protagoras on Virtue

PART

85

II

4. Knowledge as the ‘whole’ of virtue {a) Does Socrates answer his own questions? But perhaps I have strained my reader’s patience with this account of what Socrates does not believe. For Socrates, in the Protagoras, does not aim only to prove that Protagoras is wrong, with no indi¬ cation of what he himself holds to be true. His approach therefore is neither sceptical, nor merely adversarial. On the contrary, Socrates seeks to win Protagoras’ acceptance of at least two major points. Before the interval on Simonides, Socrates aims to show that tem¬ perance and wisdom ‘are one’.^’ After the interval on Simonides, in the second round of the debate

(349 d 2

ff.), he aims to prove

that ‘wisdom is courage’.^* Admittedly, the argument gets off to a false start

(349 E 1-351 b 2).

But Socrates does eventually succeed

in winning Protagoras’ grudging admission that courage should be defined in terms of knowledge or wisdom

(360 D 4-5),

and that

he had therefore been wrong to suppose that those who are ‘most ignorant’ can be ‘most brave’

(360 d 4-E 5,

looking back to

349 d

6-8). Is it possible to relate those two conclusions to the initial set of questions and answers on the ‘parts’ of virtue (329 c

6-e

6)? Can we

tell, can we at least guess, what Socrates’ own answers would have been to the questions that he puts to Protagoras at the beginning of their debate? is put forward to show that we can have one virtue and not have all the others (in so far as we can have one dynamis and not another: we may be blind, but not deaf, deaf but not dumb, in the same way that, according to Protagoras, we may be brave but not wise). The similarity between Hartman’s thesis and Ariston’s theory therefore serves merely to confirm, if confirmation were needed, that Hartman’s comparison of virtue and seeing cannot be v^hat is intended in the Protagoras. Socrates concludes his argument by asking, 333 B 4—5: oukovv ev av etij 17 aoxfipoavvrj Ka'i rj aopia', The question is designed to elicit an affirmative answer. Before the

interval on Simonides, Socrates also sets out to prove that justice and temperance are to be in some way associated (333 b 8 ff.). But the argument is aborted by Protagoras’ turning to makrologia (334 A 3 ff.) Socrates asks Protagoras, 350 c 4—5: ovkovv . . . t) aoia av dvSpcta d-p; I’he interrogative particle is ‘understood’ from 350 c i. The question is again designed to elicit an affirmative answer. In the lines following, Protagoras has no hesitation in telling Socrates ‘you think that courage and wisdom are the same’ (3500 4-5: o’Ui TTjV aoLav Kal rpv dvdpetav ravrov elvai).

Denis O’Brien

86

'If'he Protagoras is not designed to satisfy our curiosity. But it is difficult not to believe that Plato intends the reader to ask some such question for himself, whether or not he means him to find the answer. This is particularly so when, on almost the final page of the dialogue, Plato lifts a corner ocf the veil by having the Argument itself tell us that Socrates is aiming to make knowledge the ‘whole’ of virtue (361

B

5-7). If knowledge is the ‘whole’, then what are the

‘parts’? Or is virtue perhaps a ‘whole’ that has no ‘parts’?^’

(6) How can we distinguish gold from gold? To explore, however briefly, the (possible) answer to that (unasked) question, I suggest we return to the comparison of the particular virtues to pieces of gold (329

d

6—8 and 349 c 2—3). As it stands

in the later passage (i.e. including wisdom as one of the ‘parts’ of virtue), the comparison cannot be taken to represent what Socrates himself is supposed to believe. But what of the earlier passage? Can the initial comparison of the individual virtues, excluding wisdom, to pieces of gold possibly be endorsed by Socrates, not only by way of refutation of Protagoras’ thesis, but as a reflection of Socrates’ own belief? The two references given above (333 b 4-5: temperance and wisdom are ‘one’, 360 D 4—E 5: the association of courage and wisdom) suffice to qualify Kahn’s claim that ‘Socrates (as a dramatic persona) hardly expresses his own opinion at all: he sim¬ ply argues against Protagoras’. See C. H. Kahn’s uncharacteristically diffuse article, ‘Plato on the Unity of the Virtues’, in W. H. Werkmeister (ed.). Facets of Plato’s Philosophy {Phronesis, suppl. 2; Assen and Amsterdam, 1976), 21-39 at 25. Twenty years later, the point is expressed a trifle more cautiously, C. H. Kahn, Plato and the Socratic Dialogue: The Philosophical Use of a Literary Form {Carahridge, 1996), 221: ‘The unity of virtue is never expressly defended by Socrates in the Protagoras', he only argues against the claims of plurality and diversity maintained by Protagoras’ (Kahn’s italics). It is true that the position we might think Socrates would adopt cannot be simply read olT as at each point the obverse of the thesis attributed to Protagoras, and Kahn’s remark may be taken as a useful corrective to those who think that it can be. But to claim that Socrates’ arguments are purely adversar¬ ial (Socrates ‘only argues against the claims of plurality and diversity maintained by Protagoras’), and therefore devoid of positive content (‘Socrates [as a dramatic persona] hardly expresses his own opinion at all’), is excessive, and Kahn himself elsewhere recognizes as much (Plato and the Socratic Dialogue, 211: in the final pages of the dialogue, ‘Socrates presents a sustained argument to prove that courage is a form of knowledge’, 215: in the Protagoras, ‘Socrates is ultimately arguing . . . for a conception of virtue as knowledge or wisdom’). Whether or not the arguments put into Socrates’ mouth are successful, whether or not they are even intended to be successful (c£ Kahn, ibid. 242: in the last argument of the Protagoras, ‘much of Socrates’ reasoning ... is manipulative and insincere’), is (or should be) a quite different question.

Socrates and Protagoras on Virtue

87

Vlastos is convinced that it can be. If we adopt Vlastos’s second interpretation of the Unity thesis (each of the virtues has all the other virtues as ‘attributive predicates’), then, he claims, we ‘purge’ the image ‘of its most obnoxious feature: the suggestion that the virtues would be as destitute of individuating qualitative character¬ istics as are bits of a gold bar’.**® We do so because, ‘as each of its parts has all of the qualities characteristic of gold (each is yellow, ductile, malleable, and so forth), so each of the virtues has all of the qualities characteristic of virtue (each is wise, temperate, and so forth)’. Has the purge been successful? I fail to see how. If I tell you that I have in my hand a piece of gold that is yellow, ductile, and malleable, can you possibly tell me which of two or more pieces of gold I have in my hand? If I tell you that my behaviour this afternoon displayed to a heroic degree the virtues of justice, temperance, and piety (not to mention wisdom and courage), have you any idea, any inkling even, of what kind of virtuous behaviour I claim to have engaged in? But if Vlastos’s account fails to give meaning to the comparison, what other account can we possibly put in its place? How can we distinguish one piece of gold from any other or from the whole, since they all display precisely the same qualities, and differ only ‘in largeness and smallness’?

(c) The Protagoras and the Laches The quandary we are faced with in the Protagoras is closely ana¬ logous to the puzzle on which the Laches ends. Setting out to dehne bravery as only one part of virtue, Socrates and Nicias end up with a definition of the whole of virtue. In the Laches, as with the image of gold pieces in the Protagoras, how are we intended to distinguish one part of virtue from the whole of which it is a part? The Laches perhaps points to the answer. Nicias, in the Laches, offers a definition of courage as knowledge ‘of things to be feared and things not to be feared’ (195

D

8:

rwr Setvwv Kal fx-rj Setvcov),

‘of things to be feared and things to be dared’ (198

b

or

2-3: twv SeLvwv

Kal OappaXiojv).*^ From this Socrates and Nicias pass to courage as

knowledge of things good and evil in the future (198 ““ Vlastos, Platonic Studies, 248.

b

2-c 8), and

■" See again Vlastos, ibid. 247-8.

See also 199 A 10—B i. The earlier form of words occurs in the Protagoras, when courage is defined as ‘knowledge of things to be feared and things not to be feared’ (360 D 4-5). The definition is alluded to in the Republic (4, 429 C 5-8).

88

Denis O’Brien

then to courage as requiring knowledge of things good and evil in the past, the present, and the future (198c 9-199 c 2). It is at this point that Socrates declares them to have found the definition, not of a part of virtue (courage), but of the whole of virtue (199 c 3-E 5). (I am obviously having to cpncertina the steps in the argument.) Socrates claims that he and Nicias have therefore failed to find ‘what courage is’ (199 E ii:

ovk apa 7]vpT]Kap,€v,

cS Nikm, dvSpeta

oTi eanv). Nicias agrees: ‘We seem not to have done’ (e 12: ov (f>aiv6p.€0a). Commentators commonly take those words at their face

value. They conclude that the search for a definition of courage has indeed failed, and that the failure must imply a fault either in the initial definition of courage as knowledge of things to be feared and to be dared, or in the concluding definition of virtue as knowledge of things good and evil, or in the reasoning that links the two. Such at least is the range of scholarly opinion recorded by L.-A. Dorion, in a long and learned Annexe added to a recent translation of the Laches into French.But is that necessarily the implication of the passage? Dorion appears not to take account of the distinct possibility that the fault is nowhere to be found, that the final form of words (199c 4-D I: courage is knowledge of ‘things to be feared and things to be dared’ and at the same time knowledge of ‘all things good and of all things evil’) is a definition both of courage as a part of virtue and of virtue as a whole. {d) The double definition The paradox is less startling than it may seem. Socrates and Nicias agree that knowledge of things good and evil cannot be limited to past, present, or future time, i.e. we cannot know good and evil only in part. We cannot therefore distinguish good and evil in the future without being able to distinguish good and evil in the present and in the past (198 b 2-199 c 2). But from that it does not follow that the distinction between past, present, and future time has been obliterated. For the specific knowledge that courage displays, even if it draws on the present and the past, does so as a guide to what has not yet been accomplished, and therefore as a guide to what lies in the future. So much at least is implied when, in distinguishing courage from L.-A. Dorion, Platon, ‘Laches’, ‘Euthyphron’: traduction inedite, introduction et notes (Paris, 1997), 171—8 (‘Annexe: L’argument final du Laches (1976—1996) et I’unite des vertus’)-

Socrates and Protagoras on Virtue

89

mere animal boldness or fearlessness, Nicias links ‘courage’ to ‘fore¬ sight’ (197

B

2—3: avhpeias . . . vpopirjOias), and so to the future/"*

Socrates’ definition of fear as ‘an anticipation of future evil’ (198

b

8—9: Seo? yap etVai npoaSoKLav p^eXXovros KaKov) serves only to make the same point more apparent. If fear is ‘an anticipation of future evil’, and if courage is defined as knowledge ‘of things to be feared and things not to be feared’ (195 and things to be dared’ (198

B

d

8-9) or ‘of things to be feared

2—3), then again the knowledge in

question cannot but relate to the future."*’ Is that perhaps the answer to the conundrum in the Laches? Courage will be courage, in the sense that Socrates would give to the word, only if ‘the things that are to be feared’ and ‘the things that are to be dared’ are recognized as good or evil. But the recognition of good and evil requires a knowledge that is not limited to present or to future time. If we are to have the knowledge of good and evil that is needed for the exercise of true courage, then we must have knowledge of all goods and of all evils, past, present, and to come, even if, in the exercise of courage, our ‘knowledge’ is brought to bear on future dangers, on fears that by definition are fears for the future. That is why, in arriving at the definition of courage, we inevitably arrive at the definition of virtue as a whole. The knowledge of things to be feared and things to be dared (the proper object of courage) has at the same time to be a knowledge of things good and bad (the knowledge that is virtue), since only so can courage claim to be a virtue. But we do not therefore have to suppose that courage has so to speak ceased to be courage, swallowed up as it were in the knowledge of all good and of all evil. Knowledge of good and evil, even if it is

Vlastos, Plato’s Protagoras,, Introduction, xlviii-1, makes the interesting sug¬ gestion that the association of courage and ‘foresight’ has been taken from Prodicus. The definition of fear (^o/3oj or hios) as ‘an anticipation of evil’ is repeated in the Protagoras (358 D 6-7: npoaSoKlav nva . . . KaKov). In the Protagoras the definition of courage and cowardice as respectively knowledge and ignorance ‘of things to be feared and things not to be feared’ (360 D 4-5) therefore again points to the future, albeit with a shift in terminology, in the Protagoras as in the Laches, from Seoj or describing the subjective feeling of fear, to Setvd, describing the objects that inspire our fear. Cf. La. 198 B 5—7: Xyod/J-eda S’ ij/iet? Seird pciv eivat d Kai Seos Trapexci-, dappaXea d fir) Sios rrapixcL. Fear as ‘anticipation of evil’ is taken up by Aristotle in his sustained critique of Plato’s account of courage in the Nicomachean Ethics (3. 6, III5‘‘9: Tov (f>6^ov opl^ovrai vpoaSoKiav KaKOv).

Denis O’Brien

90

knowledge of all good and of all evil, when it is exercised as courage, relates Specifically to the future. i

{e) The plurality of virtues It is true that, in the Laches, Socrates objects that, if courage is the knowledge of good and evil in the past, the present, and the future (198 B 9—c 2),

then the man who is courageous cannot be lacking in

the other virtues, listed as temperance, justice, and piety 2).

(199 c 3-E

But is that perhaps the point?

Justice will require knowledge of things good and evil, specifically in relation to our fellow men.'** Piety will be knowledge of good and evil, specifically in our behaviour towards the gods.*’ In each case, knowledge of things good and evil, in order to be specified by the names of the individual virtues, will need to be circumscribed, not by what we know, but by the application of what we know. Hence the relation between the virtues. Piety will be knowledge of good and evil in our relationship to the gods. But the knowledge that piety requires is none other than the knowledge that we have already displayed, in a different context, by being courageous. We cannot therefore have the one virtue without the other, since that would imply a restriction on the nature of our knowledge, whereas the knowledge that both virtues require (the knowledge of good and evil) is knowledge of the whole. If courage and piety are parts of virtue, they are not parts that can be had independently one of the other.** Cf. Prot. 323 A 5—c 2; Gorg. 507 B 1—2. I am deliberately choosing at random from conventional descriptions.

Cf. Euthph. 12 E 5-8; Gorg. 507 B 1-3.

I have deliberately dodged the issue of how we might hope to delimit the domain of temperance {ao}(j>poavvrj). Following the very simple line of argument adopted above, we might suppose temperance to be knowledge of good and evil as related, perhaps, to our bodily needs and desires or to our personal ambitions (cf. Phaedo 68 c 8-d i; Symp. 196 c 4-5; Rep. 4, 430 e 6-7). But the Charmides is a warning of how elusive this ‘virtue’ proves to be (the seven successive attempts at a definition are conveniently listed by G. Santas, ‘Socrates at Work on Virtue and Knowledge in Plato’s Charmides’, in E. N. Lee, A. P. D. Mourelatos, and R. M. Rorty (eds.). Exegesis and Argument: Studies in Greek Philosophy Presented to Gregory Vlastos (Phronesis, suppl. i; Assen, 1973), 105-32 at 108).—There is, furthermore, the possibility that aw(j)poavvri might perhaps, after all, be not a part of virtue, but like ao(f)la or eTnaTpfjL-r] (cf. 361 B 5—6) the whole of virtue. (See e.g. Gorg. 507 A i-c 7. The ambivalence arises in part from aw(j>poavvrj sharing with CToc/ita and (jipovrjais a common opposite, dcfipoavvrj. The fluctuation in Plato’s own usage, at least as evidenced by Socrates in the Meno, 88 a 7-0 i, and by Diotima in the Symposium, 202 A 2-10, is turned to account in the trap laid for Protagoras

Socrates and Protagoras on Virtue

91

(/) The Protagoras and the Sophist If that is Socrates’ conception, can we perhaps see that the very notion of ‘part’ becomes ambivalent? Courage is restricted in its scope to ‘things to be feared and things to be dared’, and so to the future. But the knowledge of good and evil which courage calls upon (in order to decide whether ‘the things to be feared’ and ‘the things to be dared’ are good or bad) cannot be a partial knowledge. Is courage therefore a part or is it at the same time necessarily also the whole of virtue? If we retain the language of ‘parts’, we have to recognize that, on the conception of virtue that is here proposed, the knowledge of good and evil cannot be divided into ‘parts’ in the way that knowledge, in the Sophist, is divided into parts. In the Sophist, the parts of knowledge (like the parts of otherness) have each a separate object (257 c

lo-D

3; the examples that follow are not in the text).

Medicine is the knowledge of health and disease. Astronomy is knowledge of the heavenly bodies. Arithmetic is the knowledge of odd and even. But the doctor is not therefore an astronomer nor even a mathematician. In the Sophist, we can therefore have one ‘part’ of knowledge, and not have other parts. That simple distribution of knowledge into ‘parts’ is no longer possible when the knowledge in question is knowledge of good and evil, and when the ‘parts’ of that knowledge are individual virtues, each of which requires, not a partial knowledge, but the whole of knowledge (the knowledge of all good and of all evil).

{g) The ambivalence of ‘part’ and ‘whole’ To take the measure of so paradoxical a conception of the relation of part to whole, we may try turning the tables, and treating each of the parts of knowledge in the Sophist as at the same time knowledge of the whole. That is of course not so. In the Sophist, ‘knowledge’ exists only as ‘chopped up’ into all the different parts of knowledge

at 332 A 4—333 B 6.) Were that to be so, then the claim that aco^poovvri and aota are ‘one’ (333 B 4-5) would be quite different from the claim that ‘wisdom is courage’ (350 c 4—5). Even more troubling, the original list of virtues (329 C 4—5; C 7-8) would already have presented the anomaly that in sect. 3 of my essay I have associated with the inclusion of aota among the ‘parts’ of virtue. But I intend to leave aside those possible complications in the present essay.

Denis O’Brien

92

(cf. Soph. 257 c 7—8: KaTaKeKepixariodai).*'^ But suppose there were, instead, a science that was not restricted to a single object, a masterscience, a science of the whole. Its owner would be omniscient. Possession of this one science would bring with it possession of all the individual sciences. As if that were not startling enough, we should also have to suppose, to complete the parallel with the theory of the virtues, that no single science could be had independently of any and every other science. It would no longer be possible to have a knowledge exclusively of medicine or astronomy or arithmetic. Possession of any one of the individual sciences would inevitably be accompanied by possession of all the others. No one would any longer be a doctor without also being at the same time an astronomer and a mathematician. Omniscience, and the impossibility of mastering any one dis¬ cipline without mastering all the others, would not obliterate the knowledge of medicine, astronomy, or arithmetic. But how would one distinguish any longer the parts from the whole?

(h) An almost impossible question To answer that almost impossible question, we might be tempted to claim that there were no longer any ‘parts’ of knowledge, that the names of what are now regarded as the particular sciences are only so many names of one and the same thing (the knowledge that is knowledge of the whole). Alternatively, and a trifle more moderately, we might allow that the different branches of knowledge are ‘parts’ of knowledge, but insist that no one part can be had without every other part, and that they are ‘parts’ therefore only in a very attenuated sense, since the division of the sciences into different areas, each one covered by a separate discipline (medicine, astronomy, arithmetic), no longer corresponds to any limitation in the nature of our knowledge. On this alternative hypothesis, the division into ‘parts’ will con¬ tinue to be an objective feature of what it is that we know, but the ‘parts’ will no longer be a measure of how much, or how little, we know. For the division of the field of knowledge into ‘parts’ will no longer imply that we can have knowledge of any one part without For Socrates’ use of the same verb in the Meno (79 c 2 (cf. a 10); Socrates criticizes Meno for chopping virtue up into parts), see sect. 6 below.

Socrates and Protagoras on Virtue

93

having knowledge of every other part. The knowledge we have of any single ‘part’ will be at one and the same time a knowledge of the whole. But before we could even begin to explore which of those two dif¬ ferent descriptions might answer the better to our new conception of omniscience, we should have an uphill task persuading the likes of Protagoras to abandon the belief that medicine, astronomy, and arithmetic are so many independent disciplines. Protagoras will in¬ evitably cling to the conviction that each ‘part’ of knowledge can be known independently of all the others, and that someone can be a perfectly competent doctor without having to be an astronomer and a mathematician to boot, in just the way (to revert to the descrip¬ tion of the virtues) that someone can be brave, without necessarily possessing any of the other virtues, and certainly without needing to be ‘clever’.^®

{i) Plato’s two theses That (imaginary) perspective may conceivably give an indication as to why Plato should have put forward two distinct theses in opposition to Protagoras: both the radical thesis that the individual virtues are names of ‘one and the same thing’, and the less radical, but still paradoxical, thesis that the individual virtues are ‘parts’ of virtue, but parts that differ from one another only in the way that different-sized pieces of gold differ from one another. Those two theses certainly serve a common purpose in being united in their opposition to Protagoras’ concept of the virtues. But do they perhaps also answer to the ambivalence in the concept of virtue underlying the criticisms that are made of Protagoras? Are the two theses perhaps intended to point up the difficulty and the paradox of a theory whereby the knowledge displayed by each ‘part’ of virtue is at the same time a knowledge of the whole, so that each part of virtue is therefore also, in a sense, the whole of virtue?

Protagoras is not alone in singling out bravery and wisdom as the two virtues least able to be assimilated (cf. 349 D 2-8). The underlying reaction is the same in the Laches, where Nicias fairly snorts with disapproval when he is first told that, in order to be brave, his men need to be clever (see the exchange at 194 e i 1-195 A 6).

Denis O’Brien

94 ’

5. The image of gold in the Protagoras and in the Timaeus

In that relationship of the ‘parts’ of virtue ten each other and to the whole do we see a possible interpretation of Socrates’ initial comparison of individual virtues to different-sized pieces of gold in the Protagoras?

(a) An impossible paradox? The use to which I have put that image, in my interpretation of the three theses of the Protagoras (Section i above), may seem, at first sight, impossibly paradoxical. We can obviously have one piece of gold, and not have another piece. How then, it may be objected, can the image possibly be intended to illustrate a theory where we cannot have only one part of virtue, without having all the others? If the image is taken in isolation from its context, that objection may well seem overwhelming. But, once we return the image to its context, the force of the objection fades. With this objection, are we to suppose that the lesson to be drawn from the image, in the context of the Protagoras, is that, just as we can have one piece of gold and not have another piece, so too we can have one part of virtue and not have another part? If that were the meaning the image is intended to convey, then the choice between the two images would no longer be a real one. For the conclusion drawn from both images would then be the same. Whether the individual virtues are like parts of a face (za) or like pieces of gold (zb), in either case the lesson would be that we can have one virtue and not have another. As it is, the two images, it seems to me, are clearly intended to illustrate the two opposed theses that follow in Plato’s text. The thesis that we can have one part of virtue and not have another part (3a) is explicitly said to follow ‘clearly’ (330B i) from the image of parts of a face (za). The implication is plainly that the image of pieces of gold (zb) will lead to the opposed thesis, that we cannot have one virtue without having all the others (3b). I note particularly that Socrates’ third question begins, 329 E 2: TTorepov oSv . . . Whether we can have one virtue and not have another (3a), or whether our having one virtue necessarily entails our having all the virtues (3b), is presented as a consequence (cf. ovv) of the choice that has already been made between taking the individual virtues to be like parts of a face (2a) or like pieces of gold (2b). The particle (ovv)

Socrates and Protagoras on Virtue

95

(b) The two images Quite how and why that should be so will be clearer if we allow ourselves to interpret the difference between the two images in the light of the theory of good and evil outlined in the section preceding this. On the theory illustrated by the parts of a face, the acquisition, so to speak, of a new ‘part’ brings with it a new capacity. We are able to see, as well as to hear and to smell, for each ‘part’ of a face brings with it a specific dynamis (330 A 4—7). Conversely, we may have one capacity and lack another: we may be blind, but not deaf, deaf but not dumb (just but not pious, pious but not brave). Not so with pieces of gold. ‘Gold’ is the knowledge of good and evil. Pieces of gold are the individual virtues. Each piece has all the qualities of every other piece and of the whole, because we cannot know good and evil only in part and because each virtue therefore requires knowledge of all good and of all evil. In so far as the individual virtues circumscribe our knowledge, it is not by any intrinsic limitation to or diminution of our knowledge, but because each of the virtues requires our knowledge of the whole to be adapted to some specific range of action or behaviour. The individual virtues are in that sense ‘parts’ of virtue, compar¬ able to so many pieces of gold, each of which retains the properties of the whole but is limited, extrinsically as we might say, merely by being larger or smaller than some other piece. The individual virtues require a full knowledge of good and evil, even if they dis¬ play that knowledge in only a restricted area. Pieces of gold are all fully gold (all equally yellow, ductile, malleable), even when they have been divided into pieces that differ, not in possessing the prop¬ erties of gold to any greater or lesser degree, but only ‘in largeness and smallness’ (329

D

8).^^

therefore anticipates the explicit link (cf 330 B i: 17 hrjXa S17 on . . .) that will be made between Protagoras’ choice of the first term in either disjunction (2a and 3a). The obvious implication is that the relationship between the second term in either disjunction (2b and 3b) is the same. ” I need to warn my reader that this interpretation of the image is possible only if ‘knowledge’ is not included as one of the ‘parts’. On the interpretation I have outlined, knowledge (for Socrates, the ‘whole’ of virtue, cf. 361 B 5-7) is gold (all gold, any gold), as distinct from the different-sized pieces of gold that are the individual virtues (justice, temperance, piety, courage, and the like). When therefore the image is repeated (349 C 2-3) in a context where knowledge is included as one of the five (Protagorean) ‘parts’ of virtue (as it is in Socrates’ recapitulation, 349 B 1-6),

gb

Denis O’Brien

(c) The Timaeus and the Protagoras If that (simple) application of the image has been misunderstood, it is in part because the image itself has been misunderstood. In par¬ ticular, commentators hpve been led astray by Socrates’ characteri¬ zation of the pieces of gold as differing ‘in largeness and smallness’ (329

D

8). Kahn tells us that the specification of differences of size

is ‘crude’, Taylor that it is ‘an error’.” Both scholars fail to see that the differences ‘of largeness and smallness’ (329

D

8) are essential

to the use that is made of the image in its context. To see why that is so, we shall do well to compare the image, in the Protagoras, of gold divided into pieces that differ ‘in largeness and smallness’ (329

D

6-8; cf. 349 c 2-3) with the image of gold

moulded into a variety of different shapes, in a well-known passage of the Timaeus

(50 A 5-B 5).

Differences of size in the Protagoras

will have as their counterpart differences of shape in the Timaeus. In the Timaeus Plato uses the image of gold to illustrate the nature of the mysterious ‘third kind’ that for later philosophers will be either space or matter. If, of any one of the many shapes made out of gold/space/matter, someone asks: ‘What is it?’, then the interpretation proposed above cannot be intended as the immediate meaning of the text. At this point in the dialogue (349 c 2-3), the image of pieces of gold (2b) has to be taken primarily as a foil to Protagoras’ choice of the alternative image (2a: parts of a face), and not, or not primarily, as a possible statement of Socrates’ own belief (see sect. 3 above).—Is it none the less perhaps legitimate to seek to read a secondary meaning into the text by distinguishing between the first and the second occurrence of the image, separated as they are by Protagoras’ addition of knowledge to the parts of virtue (329 E 5—330 A 2)? The first occurrence (329 D 6—8). The image is put forward by Socrates himself, before knowledge has been included among the ‘parts’ of virtue (cf. 329 c 2—6). Protagoras is invited to choose between this image (2b) and its rival (2a). The second occurrence (349 c 2-3). The image is repeated in Socrates’ recapitulation of the earlier discussion, in a context where knowledge is included among the ‘parts’ of virtue (cf 349 b 1—3). The image is no longer one of two apparently equal terms that Protagoras has to choose between (cf 329 D 4—8). Its role is now purely negative: the comparison with pieces of gold (2b) is no more than the image that has to be cast aside if Protagoras is to adopt the rival image (2a: parts of a face).—Although the second passage is presented as a repetition of the first (349 A 8-b i), I would not myself be averse to seeing, in the difference between the presence of knowledge in the second occurrence of the image and its absence in the first, an implication of how the image would have to be corrected if it were to be taken as a possible representation of Socrates’ own belief But I refrain from pursuing so delicate a point (a meaning which at best could be only implicit in the text), when even the more immediate purpose of the image, in its context, is still the subject of controversy. ’’ Kahn, Plato and the Socratic Dialogue, 220 n. 12. C. C. W. Taylor, Plato, Protagoras.' Translated with Notes \Plato, Protagoras] (Oxford, 1991), 108.

Socrates and Protagoras on Virtue

97

the ‘safest answer by far’, so Timaeus tells us, will be to say that it is gold. Suppose a man had moulded figures of all sorts out of gold [ei yap -navra tls ax^p-ara nXdaas (k xpvoov], and were unceasingly to remould each into all

the rest [pr/Sev perauXaTTiov navoiro e/caara els aTravra]: then, if you should point to one of them and ask what it was [SeiKvvvros Sij rtvo? avrojv ev Kat ipopevov TL ttot’ eart], much the safest answer in respect of truth would be

to say ‘gold’ [paKpw Wpos dX-rjOeiav da(f>aXeaTaTov elrreiv otl ypuad?]- {Tim.

50 .A

5-B

zy*

When Protagoras declares that the individual virtues are so many parts of virtue, Socrates asks him whether they are like parts of a face or like pieces of gold which differ from one another and from the whole only ‘in largeness and smallness’. (1) Protagoras declares that the individual virtues that Socrates is asking about ‘are parts of the single thing that is virtue’ (cf. 329

D

4: evd? ovTO?

TTjs dpeTrjs pi.6pid Igtlv d ipcnras).

(2) Socrates then asks him ‘whether they are so in the way that parts of a face are parts’(d 4—5: TO. piopia piopid ioTLv),

vorepov, €rjv, coa-ncp TTpoocdnov

‘mouth and nose and eyes and ears’

(d 5—6: ardpia re Kal pis Kal 6(j)daXpLol Kal cSra), ‘or in the way

that parts of gold are no different’ (d 6—7: Xpvoov pidpLa ovSeu 8ta(j>ep€i),

rj (Loirep rd tov

‘one lot from another lot’ (d 7’

rd irepa twv Irdpojv), ‘from each other and from the whole’ (d 7: dXXrjXujv Kal tov dXov), ‘except in largeness and smallness’ (d 7—8: dAA’ 7] pLcyedei Kal apnKpoT-pTL;).^^

In his recapitulation of the earlier discussion at 349 a

8-c

5,

Socrates repeats the contrast between the parts of gold and the parts of a face, but the other way round (beginning therefore with the parts of gold), and no longer stating the relationship between the pieces of gold negatively, in terms of their lack of difference (cf. 329

D

7; ouSev Sta(/>epet), but positively, in terms of their similarity.

(i) Socrates states that Protagoras had declared the individual I repeat the translation given by F. M. Cornford, Plato’s Cosmology: The Timaeus of Plato Translated with a Running Commentary (London and New York, 1937), 182. ” I have translated a direct question in Greek (329 D 4-8: rrorepov, ef-qv . . . apuKporrjTi;) as an indirect question in English, simply so that I can bring out the force of TTorepov ...

17

... as ‘whether . . . or . . .’ (since ‘whether’ can no longer be

used, in modern English, to introduce a direct question).

Denis O’Brien

98

virtues ‘to be all of them parts of virtue’ (cf. 349 c 2: '

T.avra

^opia elvai

dperyj^

ndura

Se

[sc. e UcoKpares, rd Xeyopueva.

There follows a moment when all is sweetness and light. Meno thinks that he has found the answer to his original question

(70 A

1—2: dpa SiSa/cTov 7] dper'p;), the question that he had insisted on returning to after the demonstration of anamnesis

(86

c

7-D

2).

Socrates and he had agreed that if virtue is knowledge it would be teachable

(87

c

5-7).

Meno is now all ready to assert that virtue is

knowledge and is therefore teachable (cf.

89

c 2-4).

But the brief moment of optimism soon fades. Aporia closes in again (from 89 c 5 to the end of the dialogue). If virtue is teachable, there must be teachers of it. But those who profess to teach virtue are charlatans. If there are no teachers of virtue, can virtue be teachable? (e) Virtue and knowledge in the Meno and in the Protagoras It is obviously no mere chance that the expression ‘either virtue in its entirety or some part of it’ (89 A 4) is brought in to cater, not for any one of the individual virtues named in the earlier discussion The occurrences of eViaTij/xr; and ^pov-qoii in the three pages that follow the adoption of the ‘hypothesis’ (86 c 11-89 c 4) are too frequent to need individual references.

Socrates and Protagoras on Virtue (78 c 3-79

E

107

4), but specifically for knowledge. The use of that

expression at precisely this point in the dialogue clearly marks a distinction between the individual virtues that had been named earlier (78 c 3-79

E

4: justice, temperance, piety ‘and the rest’), all

of which are counted as ‘parts of virtue’ without qualification, and knowledge, in the guise of phronesis, for which there is provided a careful and emphatic disjunction (89

A

3-4): phronesis is ‘either

17x01

has to be taken

virtue in its entirety or some part of it’. Once again, the disjunction (89

A

4;

seriously. The remark that phronesis is ‘either virtue in its entirety or some part of it’ (89

A

3-4) is not, as might perhaps be thought at

first glance, a mere cautionary aside. On the contrary, the first arm of the disjunction, the possibility that phronesis should be counted as ‘virtue in its entirety’, points directly to the Protagoras. The role of phronesis coupled with episteme in the Meno (from 86 c 4 to 89 c 4) is obviously comparable to the role given in the Protagoras to episteme coupled wdth sophia. At the end of the Protagoras, the Argument, personified, tells Socrates that he is aiming to establish knowledge, not as a part of virtue, but as the ‘whole’ of virtue (361

B

5—7: el (f)avrjoeTaL [sc.

t) dpeTrj]

oXov, cu? ov aiTevSeis,

d) ScoKpares ■ ■ ■)■ That give-away expression has as its obvious

counterpart, in the Meno, the possibility that knowledge, in the form of phronesis, might be ‘virtue in its entirety’ (cf. 89

A

4:

rjxot

avpLTraaav [sc. dperrjv . . .]).

In the Meno as in the Protagoras, Socrates’ interlocutor is blithely unconscious of any radical difference between knowledge and any of the other virtues as ‘parts’ of virtue (even Protagoras says no more than that knowledge is ‘the greatest’ part,

330 a 2).

But, in both

dialogues, Plato has gone out of his way to warn the reader that ‘knowledge’ {phronesis, episteme, sophia) cannot be simply tacked onto the list of individual virtues as one more ‘part’ of virtue (as distinct from the whole of virtue). That is why, in the Pro¬ tagoras, we are told that the identification of knowledge with the ‘whole’ of virtue is what Socrates had all along been ‘aiming at’ (cf.

361 B 6: o-nevheLs).

That is why, when knowledge appears on

the scene in the Meno, Socrates opens up a choice: phronesis is ‘either virtue in its entirety or some part of it’ (89 V ■ ■

•)•

A

4: rjroL . . .

Denis O’Brien

io8

(/) A repeated error Vlastos has been deaf to that warning. Neither in the Protagoras nor in the Meno can Socrates’ talk of the individual virtues as ‘parts’ of virtue be taken as evidence of ‘standard Socratic doctrine’ {Platonic Studies, 225), unless we exempt knowledge from being one of the ‘parts’. That exemption at once undermines Vlastos’s claim that the Meno can be brought in as additional evidence for his interpretation of the Protagoras. On the contrary, the error in Vlastos’s interpreta¬ tion of the Protagoras is repeated in his interpretation of the Meno. In the Protagoras, the theory of the parts of virtue which Vlas¬ tos would attribute to Socrates (as illustrated by pieces of gold) includes, at least in its later formulation, ‘knowledge’ as one of the parts (349 c 2—3; cf. 349 B 1-3), despite our being told, at the end of the dialogue, that Socrates had all along been ‘aiming’ to make knowledge, not a part of virtue but the ‘whole’ of virtue (361 b 57).®^’ Knowledge as the ‘whole’ of virtue is precisely the possibility acknowledged in the Meno, when we hear that knowledge is ‘either virtue in its entirety or some part of it’ (89 A 3—4). In the Meno as in the Protagoras, if Socrates aims to show that knowledge is the ‘whole’ of virtue {Protagoras 361 B 5-7), if he allows the possibility that knowledge is virtue ‘in its entirety’ {Meno 89 A 3-4), then he cannot be supposed to endorse a theory where, as in the theory that Vlastos would attribute to Socrates in the Protagoras, ‘knowledge’ is counted, not as the whole of virtue, but as a part of virtue. Far from supporting Vlastos’s interpretation of the Protagoras, the Meno reinforces the very point in the Protagoras which proves Vlastos’s interpretation of that dialogue to be impossible.

{g) The language of the Meno But can even the premiss of Vlastos’s argument be counted as sound? Is the language of ‘parts’ in the Meno (from 78 e i to 79 D 7) intended to carry an ideological commitment? Does Socrates’ talk of ‘parts’ of virtue prove, as Vlastos claims it does, that belief in the individual virtues as so many parts of virtue is—even for the majority of virtues—‘standard Socratic doctrine’P’’ See sect. 3 above. " For the expression ‘standard Socratic doctrine’, see again Vlastos, Platonic Studies, 225.

Socrates and Protagoras on Virtue

109

Protagoras finds the choice between treating the individual vir¬ tues as so many ‘parts’ of virtue (la) and seeing in them only so many ‘names of one and the same thing’ (ib) an ‘easy’ one (cf. 329

d

3; paSiov). It is tempting to read the Meno in the same light. The language of ‘parts’ (78 c 3-79 E 4) is introduced as obvious, as the natural way of talking for someone who wants to lead an aspiring adolescent to recognize the distinction between each of the virtues taken piecemeal and virtue as such. To help Meno to see the point, Socrates leads him to distinguish ‘shape’ as such from any particular shape (cf. 74 b 4-c 4), ‘colour’ as such from any individual colour (cf. 74 c 5-0 2). The modern com¬ mentator may wish to remark that colours, shapes and virtues are not fully analogous. The same shape cannot be round and square. The same patch of colour cannot be black and white. Does any one instance of a virtue exclude any other? But such distinctions are beside the point. Different shapes are all shapes. Different colours are all colours. Different virtues are all virtues. In so simple a context, are we entitled to suppose that the de¬ scription of the individual virtues as ‘parts’ of virtue is any more than an obvious and convenient way of referring to a plurality of virtues?

{h) ‘Chopping up ’ virtue The verb that Socrates uses when he remonstrates with Meno for ‘chopping virtue up into parts’ may serve as a warning against being too ready to read a doctrinal overtone into Socrates’ choice of language. The verb used at that moment in the Meno (79 c 2: KaraKeppLartl^rjs) is the very verb used for knowledge being ‘chopped up’ into parts in the Sophist (257 c 7—8: KaTaKeKeppiariodaL)^^ But, if the image is the same, thfe philosophical content which accompanies and underlies the image is entirely different in the two dialogues. The ‘parts’ of virtue, in the Meno, are not at all the same as the ‘parts’ of knowledge that are to be found in the Sophist—unless we are to suppose that, in the Meno, Socrates has abandoned his opposition to Protagoras and now believes that any one ‘part’ of virtue can be had without any other. For that is the relationship See also, in the Meno, the non-compounded verb used in the lines immediately preceding, 79 A 10: Kepij-arlCi-^-

Denis O’Brien

no

between the ‘parts’ that is implied when we are told of knowledge being ‘chopped up’ into ‘parts’ in the Sophist', any one ‘part’ of knowledge (medicine, astronomy, arithmetic) can be had without any other.'’’

>



Clearly we would be wrong to suppose that the same image (‘chopping up’) has to carry the same implication in two quite different dialogues. But is it any different with the ‘parts’ of virtue? Can Socrates’ use of that expression in the Meno be used to eluci¬ date the doctrine of the Protagoras?

(i) An unfair comparison? Vlastos would perhaps have objected that the comparison is unfair. ‘Chopping up’ knowledge or virtue into ‘parts’ is so patently an image. It would obviously be foolish to w'ant to read into so concrete an image any serious ideological commitment. Not so (Vlastos may say) the language of ‘parts’. If Socrates encourages Meno to talk and think of the individual virtues as ‘parts’ of virtue, it must be because Socrates himself believes that the individual virtues are indeed so many ‘parts’ of virtue. From that it will follow, so Vlastos would have us believe, that Socrates must be supposed to have refused in petto the thesis in the Protagoras which denies that the individual virtues are ‘parts’ of virtue, and which asserts instead that their names are all only so many ‘names of one and the same thing’ (cf. ib). Let us therefore allow that that is so. Let us grant argumenti causa that the language of ‘parts’ in the Meno does exclude Socrates’ endorsement of a theory of the unicity of virtue as formulated in the Protagoras (ib). Even so, that concession will not entitle us to adopt as a statement of Socrates’ own belief the second theory opposed to that of Protagoras, namely that the individual virtues are so many ‘parts’ of virtue, but as similar as are different-sized pieces of gold (zb). For the elaboration of that theory in Socrates’ recapitulation (349 c 2-3; cf. 349 B 1-3) includes knowledge as one of the parts of virtue, and that is precisely the position which, in the Meno as in the Protagoras, Socrates does not endorse.

69

See sect. 4 above.

Socrates and Protagoras on Virtue

PART

III

III

7. Aristotle’s Socrates {a) Aristotle’s allusion to the theme of the Protagoras The passages I have quoted from the Meno (78 c 3-79 E 4 and 86 c 11-89 c 4) are very possibly among the texts Aristotle has in mind when he tells us, in the Nicomachean Ethics, that Socrates thought of the individual virtues as so many fpov-qacis or cTTiarrgiaL. Aristotle writes, NE 6. 13, ii44'^i9—20: fpov-poas cocto [sc. ZtoKpaTris] civai /

\

5

^70

rraoas ras aperas^

Although Aristotle repeats that claim in various forms elsewhere, the interest of the passage from book 6 of the Nicomachean Ethics is that, in the following lines of his treatise, Aristotle alludes to the ‘dialectical’ argument by which it would be shown that the individual virtues are ‘kept apart from each other’, ii44'^32-4: d Adyo? . . . cp SiaXexdelrj ns av on ycupt^ovrat dXX-pXojv al dperal. That is precisely the theme of the Protagoras. (b) Aristotle’s own answer to the problem of the Protagoras Aristotle gives his own recipe for how to ‘undo’ the argument by which the virtues would be ‘kept apart’ (1144 32-3; d Xoyos ravT-rj SiaXvoLr’ dv . . .). We have to distinguish ‘natural virtues’, by which Aristotle means innate dispositions to virtue, from ‘moral virtue’, or what Aristotle calls ‘virtue in the full sense’, the virtue which will make a man ‘good’.’' The natural virtues, Aristotle tells us, may indeed be ‘kept apart’, in so far as we may have developed one without yet having devel¬ oped another, but not so the virtue by which someone is ‘good absolutely’.” ‘For as soon as thephronesis that is one is present,’ so See also II44’’28—9: HwKpdTrjs pLev ovv Aoyovs rds dperd? wsto ecvat {iTnoTrjfjLas yap eivai Trdaas). The ‘natural virtues’, rds cjivatKcis dperds, so Aristotle has explained at the beginning of the chapter, are present ‘from the moment of our birth’, ii44'’6: evdiis ck yeverijs. I have therefore paraphrased them, perhaps too boldly, as innate dispositions to virtue. The virtues by which someone is ‘simply good’ or ‘good absolutely’, Ka8’ as Se [sc. dperds] airXdis Xeyerai dyaBos, also referred to in the singular as ‘moral virtue’, ii44'’32: T-fjs pdtKrjs dperrjs, is what Aristotle has called at the beginning of the chapter ‘virtue in the full sense’, ii44'’4: r-pv Kvplav [sc. dperrjv]. ” Aristotle writes, 1144'’35-1145“!: rovro yap [sc. xatplt^eadaL dXXrjXojv rds dperds] Kard pLev rds vaiKds dperds evSeyeraL, Kad' as Se anXcds Xeyerai dyaBos, ovk evSeyerai.

Denis O’Brien

I 12

Aristotle concludes, ‘all the others will be present’, ii4S‘‘i-2: yap'Trj (fspoviqaei pud vrrapxovor]

Trdaai v-ndp^ovoLV.

Commentators who have not recognized the Socratic, or Pla¬ tonic, ring to that conclusion have been (needlessly) troubled by Aristotle’s claim and have even (wrongly) gone so far as to suggest that Aristotle is perhaps not fully consistent/"* The point to appre¬ ciate is that, in the passage quoted, Aristotle aims to provide his own solution to the problem of the Protagoras. He does so on the basis of a distinction (‘natural virtues’ and ‘moral virtue’) that is fully Aristotelian/^

Gauthier and Jolif translate the latter part of the passage as follows, NE 6. 13, I i44*’30-i i4S‘‘2:

‘II ressort clairement de tout ce que nous venons de dire qu’il n’est

pas possible d’etre bon, au sens propre du mot, sans la sagesse, ni d’etre sage sans la vertu morale. Mais par la se trouve refute I’argument par lequel on voudrait prouver que les vertus peuvent exister separement les unes sans les autres; le meme homme, dit-on, n’est pas egalement done pour routes et par consequent il aura deja acquis I’une qu’il n’aura pas encore acquis I’autre. Ceci en effet peut bien se faire tant qu’il s’agit des vertus naturelles, mais s’il s’agit des vertus qui nous valent d’etre appeles bons purement et simplement, cela ne se peut pas. Car supposee presente la sagesse qui est une, routes seront presentes.’ See R. A. Gauthier and J. Y. Jolif, L’Ethique d Nicomaque [L’Ethique], i. Introduction, traduction', ii. Commentaire (Louvain and Paris, i9s8-9)> i- 183. See, most recently, D. Bostock, Aristotle’s Ethics (Oxford, 2000), 51—2, and Sarah Broadie’s comments in Sarah Broadie and C. Rowe, Aristotle, Nicomachean Ethics; Translation, Introduction, and Commentary (Oxford, 2002), ad loc. ” On the ‘natural virtues’, see the references given by Gauthier and Jolif, L’Ethique, ii. 552 (ad iiq^S—9). On ‘moral virtue’, and its relation to phronesis, see the same authors’ long note on p. 555 (ad iiqTih—17).—Broadie, apparently as proof of Aristotle’s inconsistency (see the footnote preceding this), appeals to an earlier passage of the Ethics, 3. 6, iii5‘'20—2, where, she claims, ‘Ar[istotle] him¬ self allows a less than ideally strict use of excellence-terms when he says that some people are both open-handed about wealth and cowardly’. The point of the earlier passage, read in its context, is to deny that being ‘open-handed about wealth’ can be counted as an example of courage, and therefore to deny that one and the same person can be lacking courage in battle and yet be counted as courageous in the face of poverty {NE 3. 6, 11 i5‘'i7-22). In avoiding inconsistency on one point (in denying that the same person can be courageous and not courageous), has Aristotle stumbled into inconsistency on a related point, by allowing that someone who lacks one virtue (courage in battle) may none the less possess another (being ‘open-handed about wealth’), contrary therefore to the theory advanced in bk. 6 of the Ethics (the presence of one virtue implies the presence of all the others)? I very much doubt it. For it is not at all clear that fearlessness in the face of poverty (cf 111 s^iS—19), and therefore being ‘open-handed about wealth’, would qualify as an example of what Aristotle, in the later passage (6. 13, i iq^i-i iqs^z), calls ‘moral virtue’. Un¬ less it does so, Broadie’s charge of inconsistency vanishes like morning dew. It is only for ‘moral virtue’ that Aristotle claims as his own the (Socratic) theory ‘if one, then air.

Socrates and Protagoras on Virtue

113

(c) An argumentum ex silentio? Despite its Aristotelian cast, is the passage from the Ethics possibly none the less of significance for our enquiry? Aristotle presents his own theory as a direct answer to the ques¬ tion raised in the Protagoras, whether the virtues can be ‘kept apart from one another’ (1144^32-4). To express his theory, he uses the very word (phronesis) that he has chosen, in this context, to summa¬ rize Socrates’ account of the virtues. But very curiously Aristotle makes no attempt to summarize how Socrates arrived at what was, in its practical outcome, Aristotle’s own conclusion, namely that the presence of one virtue (for Aristotle a ‘moral virtue’) implies the presence of all the others. Is Aristotle’s silence possibly significant? Should we possibly conclude that, at least on this point, Aristotle knew no more of ‘Socrates’ than we do? In that case, has Aristotle possibly recog¬ nized, as we should do, that in the dialogues, although Plato makes clear Socrates’ opposition to the view adopted by Protagoras, and therefore his belief in the co-presence of all the virtues, none the less he does so without spelling out fully Socrates’ own explanation of how the virtues come to be so related?

8. Menedemus of Eretria and Eucleides of Megara From Aristotle I turn to a pair of theses summarized by Plutarch in the opening chapters of his treatise On Moral Virtue, and attributed respectively to Menedemus of Eretria (counted by Diogenes Laer¬ tius among the successors of Socrates) and to Ariston of Chios (a first-generation follower of Zeno).’^ In this and in the following section of my essay, I use the following abbrevia¬ tions. DPhA =R. Goulet (ed.),‘Dictionnaire des philosophes antiques, vols. i ff. (Paris, 1989— ). RE =Paulys Real-Encyclopddie der classischen Altertumswissenschaft. SSR = Socratis et Socraticorum Reliquiae, ed. G. Giannantoni (4 vols.; Naples and Rome, 1990). SVF = Stoicorum Veterum Fragmenta, ed. H. von Arnim (4 vols.; Leipzig, 1903-24).—For Menedemus of Eretria, see SSR i. 503-18 (lll f), and iv. 129-33 (‘Nota 12: Menedemo di Eretria’). Denis Knoepfler provides a text of Diogenes’

Life of Menedemus (D.L. 2. 125-44), with commentary, in La Vie de Menedeme d’Eretrie de Diogene Laerce: contribution a I’histoire et a la critique du texte des ‘Vies des Philosophes’ [La Vie de Menedeme] (Basel, 1991). For information on Ariston of Chios, see C. Guerard, ‘A 397: Ariston de Chios’, in DPhA i (1989), 400—3. For a translation of Plutarch’s De virtute morali, with commentary, see D. Babut, Plutarque, ‘De la vertu ethique’: introduction, texte, traduction et commentaire

114

Denis O’Brien

{a) ‘A man’and‘a mortal’ Plutarch (with no mention of the Protagoras) presents the first thesis as one whose author (MenedemUs) ‘did away with both the multi¬ plicity of the virtues and their differences’ (Virt. mor. 2,440 e: avrjpei Ttov dperoiv «rai to TrXrjdos Kat rds; Btacfyopds). He did so, Plutarch tells

us, ‘on the grounds that virtue is one, although it goes under many names’ (co? pads ovarjs /cat ypcojaevTy? ttoXXoIs dvofraat)-” ‘The same thing’, so the theory eontinues, ‘is spoken of as tem¬ perance and courage and justice’ (Virt. mor. 2, 440 e: to yap avro aojcfypoavvrjv Kat dpSpelav Kat SiKaioavvjjv Xeyeadai). ‘Just as it is’, SO

the theory concludes, ‘with “mortal” and “man”’ (KaOdnep ^porov Kat dvdpcoTTOv).

Helmbold, in his Loeb edition of Plutarch’s treatise, translates TO aoTo XdyeadaL as ‘the same thing is meant by’.’* With that trans¬

lation, the names of the three virtues sound suspieiously as though they are being put forward as synonyms. The more natural trans¬ lation of TO avrd Xeyeadai, ‘the same thing is spoken of as’, requires only an identity of reference, as in the illustration offered (Kaddnep jSpoTor Kat dvdpojTTov), where the two words ‘mortal’ and ‘man’ have

the same reference (our mortal selves), but not the same sense or meaning. The identity of reference is perhaps clearer in the Greek example than it is in English, sinee in Greek ^pords is used exclusively for [Plutarque, ‘De la vertu ethique’] (Paris, 1969). There is an English translation by W. C. Helmbold, Plutarch’s Moralia (LCL vol. 6; London and Cambridge, Mass., 1939)-

’’ The passage is included in SSR i. 515 (m F 17). I have given Kal (in the ex¬ pression Kal xptv/aeVijj ttoXXoIs ov6fj.aai) a concessive value, as Helmbold also does (Loeb translation, 19-21). Babut contrives to give a translation where there is no obvious equivalent to the word at all {Plutarque, 'Dela vertu ethique’, 90: ‘Menedeme d’Eretrie entendait supprimer la pluralite des vertus et leurs differences, estimant qu’il n’en existe qu’une seule, qui prend de nombreux noms’), although he adopts a concessive meaning in the paraphrase he gives of the same sentence in his Introduc¬ tion (ibid. 3—4: ‘Menedeme d’Eretrie . . . croyait que la vertu est unique, malgre les nombreux noms qu’elle regoit’). A concessive use of /rat might seem to be inevitable, an obvious implication of the opposition between ‘one’ (cf. /xidj) and ‘many’ (cf. TToWols). But, although I have adopted that translation (partly because, as will be seen in a moment, I attach the same concessive value to the participles employed by Diogenes, at 2. 106 and 7. 161, in reporting the same or a similar theory), I am not convinced that it is essential. We might suppose that, as fjiias answers to to nXrjdos, so TroAAoij answers to ras Siacjiopds, and that the Kal which links puds and noWols therefore does no more than echo the Kal . . . Kal . . . linking to ttXt}9os and rdshia(f)opds.

Helmbold, Loeb translation, 21.

Socrates and Protagoras on Virtue

115

mankind, and not as a generic term that would include animals (or things) other than man.^'^ The addition of an article (‘a mortal’, with ‘mortal’ used therefore as a noun) is needed to bring out the same restriction in English. In Greek as in English, the reference of the two words ‘a man’ {avOpatTros) and ‘a mortal’ {^poros) is therefore the same: both words ‘refer’ to our own (mortal) selves. However, the sense or meaning of the two words is not the same. In Greek as in English, it would be contradictory to speak of ‘an immortal mortal’. But there is no contradiction in speaking of ‘an immortal man’, even if, as is commonly supposed, all men have so far proved to be mortal. The reason is a simple one. ‘Mortals’ are opposed to ‘immortals’ and so to the gods, whereas, in Greek as in English, ‘man’ is conventionally opposed to ‘beast’

(drip).

(b) Genus and species To savour the opposition between an identity of reference and a difference of sense or meaning, we need to recognize that the reference in Plutarch’s example (Virt. mor. 2, 440

e: Kadarrep ^porov

Kal avOpojTTov) is not to a generic unity, as though we were to say

that a shape is round or square or diagonal. For there is no genus of which ‘a man’ and ‘a mortal’ would be independent species. Nor indeed could it be said that the presence of a single genus ‘did away with both the multiplicity of the virtues and their differences’ (cf. Virt. ynor. 2, 440

e).

On the contrary, the relation of genus to species one might almost say is designed to allow for a plurality of species, different from each other and only ‘generically’ related to the whole. Independently therefore of any difference of genus or species, the theory recorded by Plutarch claims simply that, whichever of the two names we call it by (‘man’ or ‘mortal’), the object that we name is the same. So too with temperance, courage, and justice. Whichever of the three names we call it by, the object that we name is the same, and can be called equally well by either of the two remaining names.*” ’’ See LSJ s.v. ^poroj. ‘Whichever’/‘it’/‘the object’: I follow scrupulously Plutarch’s text, where the bearer of the three names (temperance, courage, justice) is designated by a neuter pronoun (Virt. mor. 2, 440 e:

to

. . . avro). In what follows, I shall allow myself

to write, less scrupulously, of ‘virtue’ as the bearer of the three names (at the end of the section following this: ‘The virtue that is spoken of as “temperance” is the same virtue that is spoken of as “bravery” and as “justice”’), since, in the sentence

ii6

Denis O’Brien

(c) ‘Reference’and‘sense’ Hence the point of Plutarch’s comment that the theory does away with ‘both the multiplicity of the virtues and their differences’. The modern reader, aceustomed to a crisp distinction between reference and sense or meaning, will obviously be tempted to take the absence of ‘multiplicity’ to imply that the three names have a single refer¬ ence, and the absence of ‘differences’ to imply that the names of the virtues have the same sense or meaning, with the implication that they are therefore synonyms. The example of ‘man’ and ‘mortal’ shows that this is not what is intended. On the contrary, the point will be simply that the use of any one of the three names (temperance, courage, justice), no matter which, justifies the use of either of the two remaining names, without the need, or the possibility, of our bringing in any fresh object as the bearer of the name. Anyone who is ‘a mortal’ (with the Greek use of ^poros) is ‘a man’. The virtue that is spoken of as ‘temperance’ is the same virtue that is spoken of as ‘bravery’ and as ‘justice’.** immediately preceding in Plutarch’s text, the noun that is understood with the feminine numeral adjective cannot but be dperij (cos ficds ovarjs [sc. rrjs dperijs] Kal XpcO|U.er7;s iroXXots ovofxaat).—Plutarch’s use of a neuter expression {to . . . avro) is not, however, gratuitous. It serves, in the context, to avoid the implication that what is being spoken of is either (a) the genus ‘virtue’, of which the individual virtues are so many species, or (b) a specific ‘virtue’, somehow or other masquerading as all the others. The implication would be false, for the individual virtues are neither (a) species of ‘virtue’, since, if we follow the example of ‘mortal’ and ‘man’, each one names the whole of virtue, nor are they (b) substitute names for some specific virtue, since they are all equally and jointly applicable as names of ‘the same thing’ (to

. . . avTo). At the same time, the use of a neuter expression

(to

.

.

. avTo) helps

ease the transition to the parallel with ‘mortal’ and ‘man’. ‘Virtue’, provided it is not understood as a genus, can be supplied as an object for the names of the individual virtues. But what word could be used for the object designated as ‘mortal’ and as ‘man’? There is no obvious candidate, other perhaps than to repeat the word ‘man’, but with the proviso that it should be understood as a universal and not as a particular (a distinction for which nothing in the context has prepared us). Both for the virtues and for ‘man’, the use of a neuter expression at this point

(to

.

.

. avro)

helps therefore to keep the bearer of the ‘names’ from being identified with any one of them, while at the same time it obviates the need for ‘virtue’ seeming to be used as a genus, or for ‘man’ being used as a universal. *' I regret having to clutter my account with the repetition ‘sense or meaning’, but there seems to be as yet no settled convention for the translation into English of the distinction between Sinn and Bedeutung. The classic example, first given by Frege, for the distinction between sense (or meaning) and reference is that of the morning star and the evening star. The two expressions have a diflFerent ‘sense’, but ‘refer’ to the same heavenly body. The neophyte risks finding the example confusing, since it

Socrates and Protagoras on Virtue

117

{d) Eucleides of Megara Where has that theory of what we may perhaps call the unicity of virtue come from? Plutarch’s attribution of the theory to Menedemus is only the beginning of a doxographical trail. Plutarch’s two theses reappear in Diogenes’ Life of Ariston; however, the first the¬ sis is now no longer attached to the name of Menedemus, but is given simply as belonging to ‘the Megarians’.*^ From a doxographical point of view, the difference of attribu¬ tion is perhaps of minor importance, since Menedemus is else¬ where presented as a product of the ‘school’ allegedly founded by Eucleides of Megara.*^ What is significant is the form of words used by Diogenes in summarizing the theory. ‘The Megarians’, so Diogenes tells us, ‘presented virtue as one, though called by many names’ (D.L. 7. 161: fjiiav ttoXXols ovofxaai KaXovjjievrjv [sc. dpeTrjv etCTTjyor]).*'* Exactly that form of words had appeared earlier in the Vitae, in Diogenes’ Life of Eucleides of Megara, though no longer explicitly in relation to virtue.** Eucleides, so Diogenes tells us, ‘declared that the good is one, though called by many

is almost Parmenidean in its implication. The same heavenly body (or so at least one may imagine) acquired two names only because those who scanned the skies failed to realize that the star they saw in the morning was the same as the star they saw in the evening, and therefore supposed that each ‘name’ had a separate reference. There was as yet no Parmenides to tell them that the seeming duality was an illusion.—A simpler example, for our purposes, is mother and daughter. Of every woman (except Eve) who has borne a child, the two words can be used with a single ‘reference’ (since every mother was someone’s daughter); but, even when the ‘reference’ is the same, the ‘meaning’ of the two words is obviously different. Admittedly, my e.xample does not match Plutarch’s parallel in respect of ‘extension’. Every mother is a daughter; not every daughter is a mother; the extension is not therefore the same for ‘mother’ and for ‘daughter’. In Plutarch’s example, ‘a man’ and ‘a mortal’ have the same extension: every man is a mortal, and every mortal (with the Greek use of jSporoj) is a man. The simple point remains that, for every woman who has borne a child, the two words ‘mother’ and ‘daughter’ have a different meaning, but can be used with the same reference. The same distinction is necessarily true, in Greek, of ‘a mortal’ (^poTos) and ‘a man’ (avOpanros). The two words have a different meaning, but the

same reference. D.L. 7. 161. See SVF i. 79 (no. 351), and SSR i. 387 (ii a 32). Cf Plut. Virt. mor. 2, 440 E-F (quoted in part above). See D.L. 2. 125—6 (from his Life of Menedemus), and Knoepfler’s notes, La Vie de Menedeme, ad loc. (pp. 170-3). Cf Giannantoni, SSR iv. 133. The point is put in a negative form in the original context, D.L. 7. 161; dperds t’

ouTe TToAAdi elariyev [sc. Ariston of Chios], ws d Zrjvcuv, oure pLiav noAAois dvop-aai

KaAovpLe'vTjv, d)? 01 MeyapiKOi . . .

D.L. 2. 106. See SSR i. 386 (11 a 30).

Denis O’Brien

ii8

names’ (D.L. 2. 106: eV

to ayaBov

d7T€(f)alv€TO ttoXXols ovojxaci KaXov-

fxevov).**

That new attribution is crucial, for Eucleides was a contemporary of Socrates. His name is included, in the Phofdo, among those who were present at Socrates’ death, and he is portrayed as the narrator of the TheaetetusP

it possible therefore that, in the

words recorded by Diogenes in his Life of Eucleides, we find the original of the thesis which Socrates introduces in the Protagoras? Socrates asks Protagoras (ib: the second of the two theses Pro¬ tagoras is first asked to choose between) whether the names of the individual virtues are ‘all names of one and the same thing’ (329 i: irdvra dvdfxara tov avrov evd?

ovtos).

D

‘One’ thing ‘called by many

names’ is precisely the form of words that Diogenes tells us was used by Eucleides (D.L. 2. 106: e'v . . . voXXol? ovoyiaai KaXov^evov), and that, according to Diogenes, ‘the Megarians’ had given spe¬ cifically as their description of ‘virtue’ (D.L. 7. 161: (xiav noXXots 6v6g.aai KaXovjjievrjv [sc. dpeTrjvJ).

(e) Diogenes’ ‘many names’ If we are able to recover a contemporary reference from Diogenes’ account of Eucleides (joined to his account of ‘the Megarians’), then a further possibility arises. Can the more elaborate version of what would appear to be the same theory, as recorded by Plutarch in the De virtute morali, be taken as a guide to what is meant, in the Protagoras, by the first of the theses that Protagoras refuses? If we take the evidence in Diogenes (relating to the Megarians and Eucleides) as a guide to what Socrates has in mind when he asks whether the individual virtues are ‘all names of one and the same thing’ (329

D

i; rravra ovojxaTa

tov

avrov ivos

ovtos),

then the

possibility he envisages is simply that virtue should be ‘one, al¬ though called by many names’ (cf. D.L. 7. 161: p-lav noXXots ovojxaaL KaXovfxevTjv [sc. dpeTrjvJ). If we add the evidence from Plutarch (relating to Menedemus), the ‘many names’ will not be so many definitions of virtue, nor yet In both passages I make explicit the concessive force that would seem to be implied in the context by the use of a participle (7. 161: Ka\ovfx4vr)v, and 2. 106; KaXovfxevov, ‘though called’). Phaedo 59 c 2: MeyapoBev EvKXetSrjs. For further details, see SSR i. 377-88 (iia), and iv. 33-9 (‘Nota 3; Euclide Megarico’), and R. Muller, ‘E 82; Euclide de Megare’, in DPhA iii (2000), 272—7.

Socrates and Protagoras on Virtue

iig

will they be synonyms. The ‘many names’ are, quite simply, variant names for one and the same thing, in the way that ‘a man’ and ‘a mortal’ are variant names for our own mortal selves (cf. Plut. Virt. mor. 2, 440 e).

(/) A 'preposterous’theory? Simple, even obvious though it may be, that conclusion would put an end to Vlastos’s wrestlings over the meaning to be given to what he calls the Unity thesis (our ib). Vlastos states the view opposed to his own as implying, first, the claim that ‘the five virtues are the same virtue’ and secondly the claim ‘that their names are synonyms’ The theory of ‘synonymy’ with which Vlastos charges his oppo¬ nent is not that anyone who acts justly or piously or bravely acts virtuously, in the way that (to take our example from Aristotle) a man or an ox is ‘synonymously’ an animal.*’ The theory that Vlas¬ tos both invents and condemns is apparently that any individual act of bravery is also an act of justice and piety. Anyone who acts bravely may therefore be said, synonymously and indifferently, to have acted bravely or justly or piously (or indeed all three). It is as though, with Aristotle’s example, any individual animal were to be called, synonymously and indifferently, an ox or a man (or indeed both). That use of ‘synonymy’ leads Vlastos to claim that, on the inter¬ pretation opposed to his own, the names of the individual virtues would have both the same reference and the same meaning. They would therefore be interchangeable, or as Vlastos prefers to call it, ‘inter-substitutable’. Vlastos rightly finds such a view ‘prepos¬ terous’, since it would mean that courage could be substituted for piety in the definition given in the Euthyphro (cf. 6

D 10—11):

‘Piety

is that eidos in virtue of which all pious actions are pious. >

Vlastos, Platonic Studies, 227 (Vlastos’s italics). Cf. Arist. Cat. i, i“6-i2. ™ Vlastos, Platonic Studies, 227 and 233 (a ‘starred note’). On Vlastos’s use of ‘preposterous’, see n. 92 below. I hope I have uncovered aright the theory of ‘syn¬ onymy’ underlying Vlastos’s use of the word. True synonyms are exceptional in any language. A (rare) example of true synonymy is given by the use of the two words ^cXtlovs and dixeivovs in the Gorgias. When Socrates asks, Gorg. 489 E 3-4: TLvas Aeyets

tovs

IScXtIovs eivat; Callicles’ reply, E 5:

tovs

diJ.ewovs eyaiye, is deliber¬

ately unhelpful, a mere play on words. Otherwise, expressions that would usually be counted as synonymous nearly always carry a difference of tone or feeling, as when we say of someone that he ‘has died’, that he ‘is deceased’, or that he ‘has passed

Denis O’Brien

120

{g) A multiplicity of names If we take our lead from the passages recorded by Diogenes and by Plutarch, we can avoid that conclusion by steering a course between the Scylla of Vlastos’s ‘the same virtue’ and the Charybdis of his ‘synonyms’.’' In giving Socrates’ words their plain meaning (329 ovofjiaTa Tov avrov ivos ovtos,

D

i: navTa

the names of the individual virtues are

‘all names of one and the same thing’), we do not have to suppose that the names of the individual virtues are synonyms (any more than ‘a man’ and ‘a mortal’ are synonyms in Plutarch’s text). Nor do we have to suppose that ‘the five virtues are the same virtue’, if by that we mean that one of the virtues has to be singled out as somehow ‘the’ virtue, with the remaining names void of individual content. The possibility which Socrates offers, and which Protagoras re¬ fuses, will be simply that there is ‘one and the same thing’ (329 i: TOV avTov ivos ovtos),

‘a single object’ (349

B

D

3: ini ivi TTpayfiaTi),

for which the individual virtues provide so many names, in just the way (if we may draw on Plutarch’s Menedemus) that ‘a mortal’ and ‘a man’ are both names for members of the human race. The theory may seem over-simple. Its consequences may be paradoxical. But it is not a view that can be described as ‘pre¬ posterous’.''^ away’. (There is of course still a difference of tone in the 5th- and 4th-cent. use of jBeAriou? and d/xetvouj, it is simply that the difference has worn so thin that it can on occasion be discounted altogether, as in the passage that I quote from the Gorgias.) But Vlastos is clearly not concerned with the nature and extent of synonymy as such. If I have interpreted his remarks aright, Vlastos’s point is the logical point that I have outlined above (whereby any individual animal would be ‘synonymously’ an ox or a man), which is why Vlastos calls the theory that he has invented for his opponent ‘preposterous’. See again Vlastos, Platonic Studies, 227 (quoted above). Vlastos obviously got cold feet over his use of the word ‘preposterous’. In his original article, ‘The Unity of the Virtues in the Protagoras’, Review of Metaphysics, 25 (1972), 415-58 at 419-20, he wrote that ‘to make either of these claims [sc. of identity or of synonymy] would be nothing short of preposterous’. That form of words is repeated exactly in the 1973 printing of Platonic Studies, 227. But in the ‘second printing’ of 1981 we read simply that Socrates ‘could hardly have wished to make either of these claims’ (p. 227). The two claims that Vlastos attributes to his adversaries are also significantly different in the two publications. In his 1973 printing, Vlastos gives the view opposed to his own as ‘implying the claim that (i) the five virtues are the same virtue and perhaps even (ii) that their names are synonyms’ (p. 227). In his 1981 printing, the ‘perhaps even’ linking the two claims has been

Socrates a7id Protagoras on Virtue

I2I

9. Ariston of Chios (a) The historical Socrates? When he wrote as he did, Vlastos was pretty clearly unaware of the passages in Diogenes and Plutarch and of their possible repercus¬ sions for the interpretation of the Protagoras. But I am of course not the first to suggest the parallel. In his monumental work Socratis et Socrahcormn Reliquiae, Gabriele Giannantoni goes much further than I have done. Giannantoni takes the second of the first two theories put forward in the Protagoras (ib) as representing, in the dialogue, Socrates’ own belief, and he sees, in the coincidence be¬ tween that theory and Diogenes’ account of Eucleides, proof that the common theory was taken from the historical Socrates.'*^ That happy coincidence (Eucleides and the Socrates of the Pro¬ tagoras lay claim to the same theory), and the conclusion drawn from it (the theory has been taken from the historical Socrates), will be less compelling if we do not share Giannantoni’s supposi¬ tion that Socrates, in the Protagoras, is made to adopt the thesis that the names of the individual virtues are so many names of one and the same thing (ib), as distinct from their being different parts of virtue, even if ‘parts’ that are as similar as are different-sized pieces of gold (2b). Strangely enough, it is the passages already quoted from Dio¬ genes and Plutarch that should put us on our guard. Eor both Diogenes (writing of ‘the Megarians’) and Plutarch (writing of Menedemus) introduce what I have called the thesis of the unicity of virtue in contradistinction to Ariston’s theory.The place given to Ariston’s theory by those two authors should warn us against beomitted. Vlastos now writes of the view of his adversaries as ‘implying that (i) the five virtues are the same virtue and (ii) that their names are synonyms' (p. 227). I find it regrettable that, in the ‘Preface to the second printing’ {Platonic Studies, 1981, xi—xiii), the reader should not have been warned that such significant changes have been made to the main body of the text. (Changes of such magnitude can obviously not be tucked away under the rubric ‘correction of misprints’, acknowledged in the ‘Preface to the second printing’, p. xiii.) Giannantoni, SSR iv. 57-8 (from ‘Nota 5: Euclide e la cosiddetta filosofia megarica’). The potential significance of Diogenes’ account of Eucleides (D.L. 2. 106) for reconstructing the ideas of the historical Socrates had been mooted by earlier writers, as noted by Giannantoni. See e.g. J. Stenzel, ‘Sokrates [Philosoph]’, iniiiA (1929), 811-90 at 876-7. D.L. 7. 161. Plut. Virt.mor. 2, 440E-441 a.

Denis O’Brien

122

ing too quick to share Giannantoni’s assumption that the Socrates of the Protagoras is an unconditional supporter of the unicity thesis. {b) Ariston, Menedemus, and‘the Megarians’

»

Ariston, so Plutarch tells us, ^‘made virtue out to be one in its essence’ {Virt. mor. 2, 440

f; tj]

fxev ovaia ybiav . . . dperr^v eTTOiei),

‘but different and more than one by being relative in some way’ {rep 8e rrpos ri rrajs 8La

Ii|t .

. . a/< t, ' w ' M ■ ' I :,'.■ I > , ■,

■' '

,



.



.

•/,

I'V

,1 ' J } i

>- j

». M -

\ k

I**- •:H)

KaOit-jo.i

‘ ••

-iirr'. ,i| /, jj,

i^\iU‘ ■* '• >

•vn.iiV*'^ ivW‘»■» >'Ci«''.»kA"'

1.;

, -

-< '•

H/,

i

• i'

no f

tn,\

»►--

.silitu/. lU!^!

. ,..

i ■'

i

\ I . . -

1 •..

-tV

I

'

r ., I ■

I V

f r

• i|

c.

..'.1

-p’-

,

,

:r\,rt

; ,

ARISTOTLE’S NATURAL TELEOLOGY AND METAPHYSICS OE LIEE MARC PAVLOPOULOS

Over

the past thirty years, two stances have dominated interpre¬

tations of Aristotle’s teleology. Martha Nussbaum and Hilary Put¬ nam, in arguing that Aristotle was a functionalist, have barely con¬ ceded that the notion of a final cause is more than a mere epistemic concept.’ While they focus their reading on conceptual analysis, other commentators, chiefly David Balme and Allan Gotthelf, have relied on Aristotle’s scientific views on the constitution of living matter to interpret his arguments for final causation.^ Balme and Gotthelf both claim that this teleology rests on ontological, and not merely semantic, non-reductionism about forms, souls, and ulti¬ mately life, to physical matter. They further maintain that in the domain of life at least, such non-reductionism is largely an empir¬ ical claim. This paper proposes an alternative to either line of interpreta¬ tion. Its central claim is that Aristotle’s teleology is grounded in his metaphysics of living beings. As such, his teleology is almost

© Marc Pavlopoulos 2003 I am extremely grateful to John Richardson for his inspiration and support through¬ out the elaboration of this paper and to David Sedley for enabling me, thanks to his careful reading and incisive comments, to improve my argument significantly. I would like to thank Vincent Descombes, Pierre Livet, and Pierre Pellegrin for their perceptive criticism of an earlier draft. My thanks also go to Sarah Linford for her invaluable help in the revision of this paper. ‘ See M. Nussbaum, ‘Aristotle on Teleological Explanation’, in Aristotle’s De Motu Animalium [De Motu] (Princeton, 1978), 59-106, esp. 81-5 and 88-93; M. Nussbaum and H. Putnam, ‘Changing Aristotle’s Mind’, in M. Nussbaum and A. Rorty (eds.). Essays on Aristotle’s De Anima (Oxford, 1992), 27-56. ^ See D. Balme, ‘Teleology and Necessity’, in A. Gotthelf and J. Lennox (eds.), Philosophical Issues in Aristotle’s Biology [PIAB] (Cambridge, 1987), 275-85; A. Gotthelf, ‘Aristotle’s Conception of Final Causality’, in PIAB 204-42. Though they share the premiss of an empirical basis for teleology, these two interpretations are mutually incompatible.

Marc Pavlopoulos

134

wholly compatible with modern physics and biology, and should fully remain of interest to us todayd In Section i I develop a fourfold model of teleology that endeav¬ ours to comprehend the varying uses of teleologioal explanations in Aristotle’s biological writings. In Section 2 I set forth two jointly necessary and sufficient require'ments for an Aristotelian teleolog¬ ical explanation to be truly explanatory. I then proceed to show how Aristotle fulfils the first of these requirements on the basis of his metaphysical distinction between change and activity. In Sec¬ tion 3 I proceed to show that Aristotle’s teleology does not engulf or cancel other modes of causal explanation of natural beings, i.e. material and efficient causalities. On the contrary, teleology would not be able to explain anything were these two other causalities not also at work in the natural world. I further argue that Aris¬ totle’s teleology is compatible with the evolutionary hypothesis of the origin of species, and even with empirical emergentism about life. Finally, in Section 4 I come to the metaphysical bedrock of Aristotle’s teleology: his analysis of life.

I.

Domains, logical pattern, and kinds of teleology

Setting aside his small treatises on specific issues,'* Aristotle’s pri¬ mary writings on the subject of living beings are four in number: History of Animals \HA], Parts of Animals [PA'\, Generation of Ani¬ mals [GA], and On the Soul, or De anima \_DA^. The beginning of GA suggests that the first three each apply one

’ The only position that I know of which may be germane to the one I am advo¬ cating here is Sarah Broadie’s insistence, contra Gotthelf, that the arguments against Empedocles in Physics 2. 8 do not rest on purely physical concerns, and her attack on the excesses of ‘Mr. Gotthelf’s defense of “empiric Aristotle’”. Though she is in line with Gotthelf’s contention that Aristotle’s teleology implies an empirical claim, she also writes: ‘To Aristotle, the notion of intrinsic direction [of movement] is of logical, not merely teleological, significance; or we might say rather that the teleol¬ ogy is grounded on the logic’ (S. Waterlow [Broadie], Nature, Chajige and Agency in Aristotle’s Physics [Nature] (Oxford, 1982), 91, 128). Two essays have particularly stimulated my thinking in the course of the mat¬ uration of this paper: Sarah [Waterlow] Broadie’s Nature and John Richardson’s unpublished typescript on ‘Aristotle’s Teleologies’. “ I am referring to the treatises gathered in Parva naturalia; De motu animalium and De incessu animalium easily find their place in the fourfold partition I am proposing. See also n. 7.

Aristotle’s Natural Teleology

135

of the four types of causality to the explanation of living beings.^ Indeed, HA is a factual report of the similarities and differences between parts of animals classified according to which species they belong to; at most, explanation amounts to noting universal corre¬ lations between the animal parts, i.e. between the material causes of animals. PA takes up the same topic from an explicitly teleological point of view: Aristotle tries to explain the presence of each of the parts already mentioned in HA according to the end or function they serve in the organism as a whole.GA, in turn, is explicitly devoted to the study of moving, or efficient, causality in animals, which is identified, here, with the study of generation.’ Finally, DA examines living beings (this time including plants) according to their elSos, seeking to answer the question typical oi formal causal¬ ity: what makes this matter (these parts) be an animal?

I. I.

Basic analysis of teleology as hypothetical necessity; material and kinetic teleologies

One would therefore expect teleological explanation proper to inter¬ vene only in PA. This work does open with a polemical defence and methodological exposition of teleology that puts teleological expla¬ nation in terms of what Aristotle calls ‘hypothetical necessity’. Aristotle’s main point in distinguishing hypothetical from ‘absolute’ [aTrAcas'] necessity is that nothing actually occurring can fully guar¬ antee that a particular living thing is going to exist or be generated. Therefore, any necessary condition for the existence or generation of a living thing is never a sufficient condition to bring it about. On the contrary, the study of a living thing must begin with the supposi¬ tion that the thing is to exist and deduce from there the actual occur’ GA yis^i-iS. However, in his translation of GA i David Balme suggests that this passage might be an interpolation. See Aristotle, De Partibus Animalium I and De Generatione Animalium I {iuith Passages from II i-j), trans. D. Balme, 2nd edn. (Oxford, 1992), 127. ^ I follow Pierre Pellegrin’s take on the relationship between HA and PA and the idea that the first is devoted to the study of material causality, the second to that of final causality. See P. Pellegrin, La Classification des animaux chez Aristote [Classifi¬ cation] (Paris, 1982), 171-9. However, Pellegrin leaves On the Soul out of Aristotle’s zoological work: Classification, 170. This may not be entirely consistent with the idea that each treatise has a ‘proper aetiological purpose’ since DA is obviously de¬ voted to the study of the formal cause of life. Further, if natural teleology rests on the analysis of life as a formal cause, as I shall argue, DA should not be set apart from PA. ’ One should also classify On the Movement of Animals and On the Progression of Animals with GA as studies of the moving cause in animals.

136

Marc Pavlopoulos

rence of its necessary conditions. In PA 1. i this notion of hypothet¬ ical hecessity appears to be Aristotle’s main argument for the exis¬ tence of teleology in nature.* Any teleological explanation must take the form of reasoning by hypothetical necessity.* Below I provide arguments in favour of what I will simply suppose for now: if ‘E’ stands for the explanans-end and ‘x’ for the explanandum, a teleo¬ logical explanation can be analysed as follows into a Modus Ponens: (TE) X because of £■ (i) If E is to be, then x is present. But (ii) E is to be. So (hi)

X

is present.

When applied to parts of animals, as is the case in PA, such a tele¬ ological explanation amounts to no more than explaining the pres¬ ence of a material constituent x—an organ, a homoiomerous part, or even its elemental constituents—by the whole E of which it is a constituent. Since Aristotle analyses organic life on several levels of complexity,^ it appears that a teleological explanation going from the hypothetical existence of the whole organism to the presence of its elemental constituents can be expanded into a chain. Each medi¬ ating link of this chain is itself modelled on (TE), except that in each the conclusion (iii) will read ‘x is to be’ instead of 'x is present’. In so far as this model is valid, it is only at the elemental level that a tele¬ ological explanation can conclude as to the presence of the required constituents. In such a chain, the justification for any of these links’ second premiss (ii) is provided by the next highest link. For ex¬ ample, the elements constituting an anhomoiomerous part such as flesh are present in such-and-such proportion if there is to be flesh. But why ‘is there to be’ flesh? Because ‘there is to be’ an organ that needs to be made of flesh, such as the heart. Why, then, is there to be a heart? Because there is to be a blooded animal and every blooded animal must have a heart. Starting with the hypothetically necessary * See PA 642‘‘2—13 for a clear statement of the link between the admission of final causality and the need to distinguish between absolute necessity and necessity ex hypothesi. ’’ From the highest level of complexity, which is the whole animal, down to the lowest level, that of the elements. Cf. GA i. i, 7i5“9-ii: ‘The material of animals is their parts—of the whole animal, the anhomoiomerous parts, of these again the homoiomerous, and of these last the so-called elements of bodies.’ Unless otherwise noted, I use the Oxford Revised Translation [ORT] edited by Jonathan Barnes: Aris¬ totle, The Complete Works (Princeton, 1984). When I depart from the ORT, I bracket the Greek term or phrase, except for ‘homoiomerous’ and ‘anhomoiomerous’, which I always keep in their transliterated form, given their technical significance.

Aristotle’s Natural Teleology

137

existence of the whole animal, we can expect one or several chain(s) of reasoning by hypothetical necessity. From the animal’s existence, we proceed to the explanation of the organs until, ultimately, we arrive at the explanation of the presence of such-and-such elements in such-and-such proportion in the homoiomerous parts. Despite the fact that teleological explanation is ostensibly the

PA, a great number of explanations on the same general model are also to be found in GA. For instance, to explain specific object of

why the heart comes first in the maturation of the foetus, Aristotle argues that an organism’s future organs all need an external nu¬ trient (blood) in order to be maintained and to grow. This implies the need for an organ that can process the blood even before other organs are formed, namely, the heart. To paraphrase a more com¬ plete example, the semen’s motion is explained by considering that it will (hypothetically) impart its motion to the foetal matter upon generating the foetus. This particular motion is necessarily to be imparted to the foetal matter if it is to generate a foetus. In turn, a foetus is to be animated by certain specific motions because those are the ones required for every organ to be formed. Finally, these organs are required because together they produce an animal. Like

PA, not only does GA use teleological explanation, it also links teleological explanations into chains of reasoning." Taken in its entirety, generation extends from the formation of the semen in the male body to the maturation of the foetus in the womb until, finally, there emerges a new and viable organism that is to grow through to adulthood. Perhaps, in fact, we should say that generation ter¬ minates only upon reaching adulthood. This complex process is divided into intervening motions and stages—such as the semen’s imparting its own motion to the foetal matter and, in turn, the for¬ mation of each organ. Each of these requisite motions can then be explained teleologically by the stage it arrives at. So if ‘E” stands for any of those stages and ‘x’ stands for the motion that terminates Though Aristotle does not explicitly mention the possibility of building such chains of teleological explanation in PA i. i, he clearly models the order of ex¬ planation by hypothetical necessity on that of the practical reasoning at play in housebuilding or medicine; cf. PA i. i, 639'^i5-i9 and 64o‘‘i6-i9. Further, Aristotle explicitly allows for chains of practical reasoning in his analysis of deliberation in NE 2- 5,1112*" 17-20: ‘If it [the end] is achieved by one [mean], they [doctors, orators and statesmen] consider how it will be achieved by this and by what means this will be achieved, till they come to the first cause, which in the order of discovery is last.’ I owe the idea of chains of teleological explanations to John Richardson (cf. n. 3). " I owe this point to John Richardson (cf. n. 3).

138

Marc Pavlopoulos

when ‘E" is reached, we obtain explanations that clearly conform to the above model (TE). It could be objected that the processes’ end-stages are more often than not motions themselves and that the whole‘process of gener¬ ation could be viewed as being a single continuous process—in which case the explananda of UA would exist only in the mind of the natural scientist. But the case here is different from that of distinguishing abstract moments of time in the course of a single continuous movement. Something other than the motion of the semen is really coming into existence at the moment the semen im¬ parts its motion to the foetal matter: before that moment, the foetal matter was at rest; after, it is in motion. The motion of the foetal matter really is different from the motion of the semen, though both are specifically identical.*^ The distinction is not only in thought here: there is a real being actually terminating the semen’s move¬ ment. This new being is itself a motion, that of the foetal matter. Provided that the motion to be considered as a stage in a complex process is not numerically identical with the motion that engenders it—provided, in other words, that the two motions differ either in subject matter, quantity, place, or form—there is no obstacle to considering that a motion can be an end-stage for another motion. There are, then, at least two basic kinds of teleology: one is ma¬ terial or functional; the other, kinetic or genetic.'^ ‘That for the sake of which’ can explain on the one hand precisely why these constituents are present and, on the other, why this precise move¬ ment occurs. In each case, one can, and probably should, expand a teleological explanation into a chain: in each case, for any link in such a chain of final causation, the hypothetically necessitatmg condition (end) at one level of explanation is in turn necessitated by the end of the upper level of explanation. At any level of explana¬ tion, the proposition of type (iii) that concludes the Modus Ponens is identical to the proposition of type (ii) at the next level down. Therefore, each embedded Modus Ponens provides justification It is evident here that, unlike modern mechanics, Aristotle does not consider movement to be something (such as a quantity, for example) that passes from one body to another. Each movement starts at the point at which the thing in movement is set in motion and stops when it comes to a standstill. Therefore, when one thing sets another in motion, we should actually say that there are two different move¬ ments. This does not mean that there is a special inner impetus of force in things that sets each one into motion. This is a strictly metaphysical consideration. I owe this distinction to John Richardson (cf. n. 3).

Aristotle’s Natural Teleology

139

for the existence of the end mentioned in a proposition of type (ii) at the next level down. Of course, at the highest level of explanation this justihcation is missing: the first E in the chain of explanations cannot be justified in the same fashion. Finally, there is a remarkable commonality between the ultimate ends of kinetic and material teleologies: in both cases, the ultimate explaining premiss is that there ‘is to be’ a living being, plant or animal, of a certain species. When ‘is to be’ is taken in the sense of generation, we have kinetic teleology; when ‘is to be’ is taken in the simple sense of existence, we have material teleology. In both cases, the ultimate and commanding end of any chain of teleological explanation is a mature, living organism.

1.2. Formal equivalence between kinetic and material teleologies There is yet another important relationship between these two kinds of teleologies: their mutual dependency. It stems from Aris¬ totle’s definition of nature as ‘inner principle of motion and rest’ (Phys. 2.

I,

i92*"i4). This definition should not be taken to mean

that natural beings can start a motion by themselves (ex nihilo, as it were), as though they had some sort of inner impetus for motion. Otherwise, as Sarah Broadie points out in Nature, 205-7, it would be difficult to understand why each and every natural being can¬ not change by itself. Indeed, for Aristotle, in the sublunary world only animals are self-changers and then only with respect to loco¬ motion. The definition of nature as ‘inner principle of motion and rest’ means that, provided it has been set in motion, any natural being is by virtue of some ‘inner tendency’ (d/a/x?) 2.

I,

Phys.

192^18-19) responsible for the fact that its change is typi¬

cally of a certain determinate form (fire goes upwards, roots grow downwards, the heart beats, etc.)—unless it is prevented by some external factor. The nature of a thing is responsible for the form, or typical manifestation, of its motion and not for its starting to change, or for its changing at any particular time. Explaining why a motion starts, or why it occurs at a particular moment, is the job of efficient causality. The typical sort of motion that a thing tends to undergo whenever it moves (and which we might, today, prefer to explain by the universal laws of mechanics) is in Aristotle solely due to the ovala of natural beings—to what they are. Under normal circumstances, then, there is equivalence between

140

Marc Pavlopoulos

the form of motion that is typical of a natural being and what the being is (what we would call its nature). When movement is not pre¬ vented or violent,'■* if ‘x’ stands for some natural being and ‘m(x)’ for its typical form of movement we have: (a) If

X,

then m(x).

i

'

(b) If m(x), then x. Moreover, there is no a priori reason to suppose that being embed¬ ded in a whole would change the nature of a natural being. This is plain enough for all parts of animals because these are usually only to be found in the context of whole organisms, and Aristotle holds that they cannot retain their nature (what makes them what they are) when separated from the body they belong to. Being embed¬ ded in a whole simply is their normal condition. The equivalence between the form of motion typical of a natural being and what the being is also holds for elements that participate in the con¬ stitution of homoiomerous parts, at least in the attenuated sense that these elements potentially remain after ‘mixture’ (/lu'Ii?) is per¬ formed and inasmuch as they also potentially retain their natural movement.If the homoiomerous part dissolves, the elements that were formerly part of it will regain their natural movements. The equivalence therefore still holds when natural beings are embedded in wholes, with the restrictive clause that it only potentially holds, so to speak, for homoiomerous parts. Now let ‘x’ stand for a natural being, 'E’ for the natural whole to which it belongs, ‘m(x)’ for the type of movement that x typically exhibits, and ‘m(£')’ for the typical movement of E. The Modus Ponens of material teleology is: (i) If E is to be, then x is present. (ii) E is to be.

.'. (iii) X is present. Now, if something is to be, it is a prima facie reasonable contention that all its essential features are also to be.'* As the propensity to '*

And for (a); provided that

x

moves.

See

GC

i. lo,

.

I write ‘prima facie’ not because, as modern moral philosophy and modal logic teach us, the conclusion of an ‘is to be’ Modus Ponens would be defeasible by another such Modus Ponens reaching an opposite conclusion but rather because the presence of matter is never wholly explainable from the sole standpoint of final causality. Therefore, though it might be essential to an animal of a determinate

Aristotle’s Natural Teleology

141

produce a certain typical form of movement essentially belongs to a given species of natural beings, it seems legitimate to infer by {b) that if m{E) is to be, E is also to be (which means that only E can perform m{E) by nature); next, by (i) that x is present; finally, by {a)

that

X

will perform movement m{x) whenever it moves. This

gives us the first premiss of the kinetic teleology Modus Ponens: if m{E) is to be, then m{x) occurs, provided x is actually moving. Likewise, considering that m{E) is an essential feature of E, by (a) we can infer from (ii) that m{E) is to be; this is the second premiss of the kinetic teleology Modus Ponens. Therefore, any material teleological relation between a part x and the whole E to which it belongs provides grounds for a kinetic tele¬ ological relation between the movement of the constituent x and the movement of the whole E. A similar line of reasoning would show that the reverse relation also holds between material and kinetic teleologies: a kinetic teleological relation between two movements provides grounds for a material teleological relation between the performers of these movements. Aristotle’s theory of nature guar¬ antees that there is a formal equivalence between both kinds of teleologies.

1.3. Diachronic and synchronic teleologies And yet, because of temporal considerations, we cannot replace the teleological explanations we found in GA with those typical of PA. In GA the end—be it an intervening stage in the complex process of generation or the new organism that terminates the process—is temporally distinct from the movement and the being that were hypothetically required for its production. In PA, however, both end movements and things are contemporaneous with their mate¬ rial and kinetic conditions. This suggests a second divide between teleologies, not according to the ontological status of the terms of explanation (movements or things), but according to their temporal relationship. Therefore, I propose to distinguish diachronic from synchronic teleology. This distinction cuts across the previous divide between kinetic species to be made of such-and-such elemental matter, it might not be the case that this constitutive matter ‘is to be’ given—even if the animal itself ‘is to be’. In sect. 3 I will argue that for Aristotle a teleological explanation is never a complete explanation in the sublunary world. I would like to thank Pierre Livet for calling my attention to this point, which deserves greater development than the scope of this paper allows.

Marc Pavlopoulos

142

and material teleologies, yielding four models of teleology in all. Kinetic diachronic teleology is straightforward: the seminal mo¬ tion is for the life of the mature organism to be begotten. Material synchronic teleology is equally plain: the heart us for the whole organism. Kinetic synchronic teleology is at work when the oc¬ currence of a movement is explained by its inclusion in a wider simultaneous movement that ‘is to be’. ‘Selachian’ cartilage’s typ¬ ical motion, for example, enables the whole animal to undulate; and respiration occurs ‘for the sake of’ cooling the animal.'^ Last, there is diachronic material teleology when, for instance, semen is present in the male body for the sake of the mature organism that it is ultimately to generate.** In any case, it can be argued that this diachronic-synchronic dis¬ tinction is more apt than the material-kinetic divide to capture the difference between the teleologies respectively exemplified in PA and GA. Whereas in PA the end is always contemporaneous with the means, in GA the end always comes about when the means have ceased to perform their function. On the other hand, material teleology can systematically justify kinetic teleology, or vice versa. Aristotle keeps interweaving both teleologies, and this is why we so often find (synchronic) kinetic teleology in PA, for instance. Although this formal equivalence allows Aristotle to go both ways, he far more often justifies a particular material teleological relation by its corresponding kinetic relation than the other way around—in PA as well as in GA.^^’ The reason seems intuitively ” See n. 19. The semen, however, is not in any sense the matter of the mature organism since it does not enter into its constitution. Some cases of diachronic material teleo¬ logy such as this might better be called cases of ‘agentive’ teleology. Still, ‘material’ is adequate to characterize the foetal matter’s role for the foetus. ’’’ See e.g. PA 2. 9, 655‘‘23-7, on why some fish (Selachia) have cartilage: to allow for ‘supple motion’. Here, the presence of cartilage in fish (a material teleological relation) is clearly explained by the sort of movements that it enables the whole fish to make. Also, in PA 3. 6, 668‘’33-669’‘i, the presence of lungs in blooded animals is justified by the consideration that cooling is ‘necessary’ to them. Evidently, this necessity is of the hypothetical sort (and not anXws), and its explanans-end is simply keeping the whole animal alive. Inasmuch as life can be considered a movement, here too a material teleological relation is justified by a synchronic kinetic relation. Incidentally, this path is exactly opposite to the one. Robert Cummins takes in his ‘functional analysis’. Cummins reduces functional explanations to consider¬ ations of part-whole relationships, whereas in his scientific practice Aristotle very frequently explains the constitution of living beings by what we would call their functions. See R. Cummins, ‘Functional Analysis’, JoMrwa/ of Philosophy, 72 (1975), 741-65-

Aristotle’s Natural Teleology

143

plain. To pursue Aristotle’s ongoing analogy of natural teleology with housebuilding, saying that a solid, heavy material such as brick is needed because the house must resist wind, storm, and rain seems more evidently self-explanatory than saying that the house’s constituents will have a strong principle of rest because houses are square, multiple-room roofed constructions. After all, what explains why something is constituted of specific materials, if not the fact that only these allow the whole to behave as required? Conversely, asserting without further justification that something must be made of a certain material can often appear to be merely verbal and arbitrary. True, if a house were not made of bricks but of, say, paper or feathers, we would hardly call it a house except perhaps by analogy; but that would not explain why the thing that we mistakenly call a house is not really one. We can find an echo of this preference in Aristotle’s contention that ‘no part of a dead body, such as its eye or its hand, is really an eye or a hand’ (PA i. i, 64i’’4-6). Aristotle maintains this claim even though the dead parts retain their ‘configuration’ (oxrjiJ,a) for a short while (it does not in any case matter how long). It seems that what makes something what it is, here, is more its activity than its mere static principle of composition: cases of dead body parts all share the fact that although their configuration might be intact, they can no longer perform their function. But if it is their activ¬ ity rather than the blueprint of their composition that defines the parts of animals and even whole organisms, it should not surprise us that the pre-eminence of activity over configuration also holds with regard to their explanatory power in teleological explanation. In the following sections I will argue that the explanatory power of teleological explanation—and hence the reality of final causa¬ tion—rests on the assumption that living beings are primarily to be defined by their activities and, above all, by the most inclusive of ac¬ tivities: life. We first nee^ to ask under what conditions Aristotelian teleology can be explanatory.

2. Teleology and ontology: the explanatory power of teleology I will now return to the general schema of a teleological explanation as taken from PA i. i, and ask under what conditions the Modus

144

Marc Pavlopoulos

Ponens of hypothetical necessity is explanatory. There seem to be two tequirements, each concerning one of the premisses: I.

Non-redundancy requirement. The first re•*♦ r' r; ..‘.i

'

'•

.

f

. ■j'. '

■'

'

r

.dm ♦!('

\. **A( .'tuirf-.jitioj’ .

snro/

) .>■'*

'

tis

voi-vi . i'swi



:
,

't :«*

^

.rest

,

I

>

u.« nil rf i •iiyii-r. I

. ifjjy n

I \sr*i,}i/ lifr-it,

v, •-, >

' .

THEMISTIUS AND SPONTANEOUS GENERATION IN ARISTOTLE’S METAPHYSICS DEVIN HENRY

I.

The problem of spontaneous generation

At the outset of Generation of Animals Aristotle announces that the generation of living things is to be divided into two kinds, biogenetic and abiogenetic: Some animals are produced from the union of male and female, among such kinds as have male and female—for not all them have this division. Among the blooded kinds, in all but a few the male on the one hand and the female on the other are perfected. But among the bloodless, while some have male and female in such a way that they generate the same kind, others generate offspring that are not in fact the same kind: such are those that do not come from the union of living animals but from decaying earth and residues. {GA i. i, yis^iS—26, trans. Balme with modifications)

The latter type of coming-to-be, abiogenesis, is later identified as the ‘spontaneous activity of nature’ (avroiJLarL^ovGTj? tt)? fvaeojs, 715*^27).' The first commentator to recognize the implications of © Devin Henry 2003 I wish to thank Richard Sorabji, David Sedley, Peter Adamson, and Verity Harte for their invaluable comments on earlier drafts of this paper. I have also benefited greatly from the comments of those who participated in a seminar on Metaphysics Zeta at King’s College London'in the spring of 2002, where I had the opportunity to test my interpretation of Z 7-9; I likewise benefited from suggestions made by those who participated in Professor Sorabji’s seminar on the Ancient Commentators at the University of Texas at Austin in March 2002. Finally, I would like to thank Robert Todd and James Lennox for their helpful correspondence. * In addition to Aristotle’s canonical account of spontaneous generation at GA 3. II, see alsoyisL ff., 721 “2-10, 723’’3-9, 73i‘’8-i8, 732‘’ii-is, 743''35-6, ysST-S, '’21-4, 7S9“7; HA (the discussions of abiogenesis in the HA are too numerous to list individually here); see also the pseudo-Aristotelian works On Plants and Problems for other references. Abiogenesis’ is the generation of life from non-life, or, as it is more commonly known, spontaneous generation. I will use ‘abiogenesis’ (and

184

Devin Henry

abiogenetic kinds for Aristotle’s metaphysics was Themistius. In his commentary on Metaphysics A he targets Aristotle’s argument against the need for setting up Forms as separately existing paradeigmata after which substances come to be. Aristotle had argued that ‘it is obvious that there is' no need at all, on these grounds [sc. loyo^g-ab] at least, to posit the existence of Forms, for a man is begotten by a man who is a particular man; and similarly with art, for the art of medicine is the formula of health’ (i 070^26-30). Themistius points out that Aristotle’s argument here is inconsis¬ tent with his belief in the existence of abiogenetic kinds, since these creatures are generated from decaying corpses and other putrefying materials and not from their likes in form: This argument would be sufficient as a refutation of the Forms, if its author [Aristotle] had not overlooked the many animals which are not born from their likes, in spite of their great number; for we see that a kind of hornet is born from the bodies of dead horses, bees from the bodies of dead cows, frogs from putrescence when it becomes sour. We see that nature does not generate these things from their likes in form and we are convinced that there are in the sperm and the seed of each kind of animals and plants proportions^ proper to it by which are begotten these animals and plants ‘abiogenetic kinds’) interchangeably with ‘spontaneous generation’ (and ‘sponta¬ neous kinds’). Two preliminary points must be made here. First, in the passage quoted above, Aristotle is expressing his view that some abiogenetic kinds repro¬ duce but not after their own kind (for example, flies are generated spontaneously but have the ability to produce grubs, which Aristotle takes to be /ri)

to

avro ylvos,

‘not the same kind’). In the main, however, abiogenetic kinds are infertile. Second, and more importantly, Aristotle’s use of the word ‘spontaneity’

(to

avro^iarov) in

this context is not the concept employed in Physics 2. 4—6 to denote chance pro¬ ductions in the sphere of nature that only appear to come about for the sake of something but are in fact accidental coincidences. The main (but certainly not the only) reason is that abiogenesis is far too regular in Aristotle’s view to be a mere coincidence: Aristotle thinks, for example, an entire class of bloodless animals re¬ ferred to as testaceans

{tcov

oarpaKoSepficov) come into being spontaneously. See also

David Balme, ‘Development of Biology in Aristotle and Theophrastus: Theory of Spontaneous Generation’, Phronesis, 7 (1962), 91-104; A. Gotthelf, ‘Teleology and Spontaneous Generation in Aristotle: A Discussion’ [‘Spontaneous Generation’], in R. Kraut and T. Penner (eds.). Nature, Knowledge, and Virtue: Essays in Memory of Joan Kung {Apeiron, special issue 22/4; Edmonton, 1989), 181-93; and G. E. R. Lloyd, Aristotelian Explorations [Explorations^ (Cambridge, 1996), ch. 3. For an alternative view of Aristotle’s concept of

to

avropLarov which attempts to locate its

biological application within the context of Physics 2. 6 see J. G. Lennox, ‘Teleo¬ logy, Chance, and Aristotle’s Theory of Spontaneous Generation’ [‘Spontaneous Generation’], Journal of the History of Philosophy, 20 (1982), 219-38, repr. in J. G. Lennox (ed.), Aristotle’s Philosophy of Biology: Studies in the Origins of Life Science [APB] (Cambridge, 2001), 229-49. ^ I presume that Themistius’ Greek would read logoi here. For the most part

Thernisttus and Spontaneous Generation in Aristotle

185

which are begotten from it specifically and not from another, so that a horse is not begotten from the sperm of a man, nor a man from the sperm of a horse, nor a plant from the seed of another plant. Where are the models of these proportions in that from which these (spontaneously generated) animals are born if they have not been formed previously in nature as proportions prepared and ready to produce any possible species of animal and made suitable for the production from it of some animals? (Themistius, In Metaph. 12 ap. Averroes, Long Commentary on the Metaphysics, iii. 1492-4 Bouyges, trans. Genequand)^

We can give Themistius the benefit of the doubt here and take his objection that Aristotle ‘overlooked’ the abiogenetic kinds as a suggestion that he overlooked their metaphysical implications, for of course Aristotle was well aware of spontaneous generation.'* The problem that abiogenetic kinds raise for Aristotle’s rejection of the Forms is supposed to follow from the theory of reproduction that underlies that rejection expounded in e.g. Zeta 7. The central tenet of this theory is the idea that offspring inherit their forms from their parents. What accounts for this fact, Aristotle proposes, is that the moving cause of reproduction (that by the agency of which the offspring comes to be:

ov) is the formal nature transmitted in the

parent’s seed, which is an internal principle of motion and change according to the same form as the generating parent. By saying that I will retain the English ‘proportions’ throughout, though occasionally I will use ‘formative principles’ to cover ‘proportions and forms’ (to which Themistius also refers: see below) and the Greek logos to cover simply ‘proportions’. ^ All references to Themistius come from Themistius, In Metaph.

I2

ap. Averroes,

Long Commentary on the Metaphysics, ed. M. Bouyges (3 vols.; Beirut, 1961—3), iii. 1492—4, translated from the Arabic by Charles Genequand, in R. Sorabji, Four Hun¬ dred Years of Transition: Philosophy of the Commentators {200—600

ad):

A Sourcebook

[Sourcebook], ii. Physics (London, forthcoming), § i(b)(i6)). (I have added as a con¬ jecture the reference to spontaneous generation set off by angular brackets; ‘Where are the models of these proportions in that from which these (spontaneously gener¬ ated) animals are born . . .?’). I,decided to use Genequand’s translation of Averroes’ testimony over Brague’s French translation of Themistius’ original text {Themistius: Paraphrase de la Metaphysique d’Aristote. Livre Lambda (Paris, 1999)) on the ad¬ vice of Sorabji, who suggested that Genequand’s translation (which coincides with Themistius’ text) is more reliable than Brague’s. Unfortunately I became aware of Brague’s commentary only after writing this paper, and so I have not been able to comment on his short discussion of spontaneous generation here. ^ See n. i above. Note, however, that Aristotle did not include among his abio¬ genetic kinds hornets, bees, or frogs. In GA 3. 10, for example, he rejects the commonly held belief that bees are spontaneously generated, both on empirical and on a priori grounds. In what follows I will use sea urchins as the paradigm example of one of Aristotle’s spontaneous generations.

186

Devin Henry

this form (the form according to which {Kara

to

efSo?) the nature in

the seed fashions the embryo) is ‘in another’ (io32‘‘25), Aristotle is making the point that the model after which a new member of the species comes into being is another natural substance which already contains that form in actuality, riot a separately existing Form (cf. GA 735'’2-4).' Themistius argues that the fact that many species of animals and plants are generated in the absence of parents cannot be reconciled with this, since these offspring do not inherit their forms from a particular generator which already bears that form in actuality. In these cases reference to the parental form is not necessary to ex¬ plain the form of the offspring that comes to be. Thus, Themistius concludes, Aristotle’s argument against the Forms is insufficient because it fails to cover all cases. We must have recourse to Platonic Forms as the models after which living things come to be; otherwise we will be at a loss to explain spontaneous generation. Themistius’ own solution to the metaphysical problem presented by the spontaneous occurrence of species-specific forms (e.g. the continuous generation of sea urchins from decay) is twofold. First, Themistius argues that there exist everywhere in nature formative principles (‘proportions and forms’) proper to each kind of ani¬ mal and plant, including those that come from decay and not from ’ Metaph. Z 7, I032“24—5:

Kal

v’ ov -q

Kara to

etSos

(j>vais p

{avTTj

Se iv aXXw)- avOpunros yap avBpwnov yevva. The theory of natural generation being expounded in Zeta is packed into Aristotle’s phrase ij

Kara to

eiSos (f>vais

q

op.oeiSijs'

(cf. GA 77o'’i7). Since Aristotle is talking about the moving cause of generation here (that by the agency of which an organism comes to be), •>) fvais should be read in the technical sense of Physics 2 (hence XeyopLev-ql) referring to an internal source of motion, the motion in question being growth and development (I will translate kIvtiols interchangeably as ‘motion’ and ‘change’). Thus, taking ‘internal source of motion’ for ‘nature’, the organism’s formal nature will be an internal source of motion according to the form, where ‘according to the form’ (/card to etSoj) modifies ‘motion’: it specifies the kind of motion for which the animal’s nature is an internal source. As I understand Aristotle’s point, r) opLoeiSqs further modifies the kind of motion initi¬ ated by the nature in the seed: the moving cause of natural generation is an internal principle that initiates motion or change according to the same form as the generating parent, i.e. the form ‘in another’ (e’r dXXw). By saying this form is ‘in another’ Aris¬ totle avoids the Third Man objection: he does not mean that the parent and child are generated after some third individual (the Form Human); rather, he is saying that the agency by which a new member of the species comes into being is the nature transmitted in the parent’s seed, which is an internal principle of change that orga¬ nizes the embryonic material according to the parent’s form. For a further discussion of this theory as expressed in GA see n. 8 below. See also J. G. Lennox, ‘Material and Formal Natures in Aristotle’s De Partihus Animalium’, Proceedings of the Boston Area Colloquium in Ancient Philosophy, ii (1995), 217-40, repr. in Lennox,182-204.

Themistius and Spontaneous Generation in Aristotle

187

kindred stock. And whereas the proportions, or logoi, according to which most species come into being pre-exist in the seeds from which they grow, those corresponding to the forms of the sponta¬ neous kinds must have already existed in the putrefying ooze be¬ forehand, ‘prepared and ready to produce’ any one of the organisms that might arise from it.*’ Of course, the suggestion that each new member of a given species comes into being according to certain fixed logoi transmitted in the generator’s seed is consistent with the Aristotelian theory of sex¬ ual reproduction.’ Indeed, the idea that the development of an offspring is a ‘continuous change originating from an internal prin¬ ciple in the seed and which arrives at some definite end’ (Phys. 2. 8, I99*"i5~i7) cf. PA 64i'’24-6) is grounded in the idea that the nature in the seed regulates the process according to a logos {GA 74o'’323). What Themistius denies, however, is the further suggestion that what determines the specific character of that change (the source of the logos that nature follows in producing the new individual) is the form of the generating parent, which is a particular.* This, he says, is most obvious in the case of spontaneous generation, where we actually observe nature producing various animal kinds in the

This seems to be the point of Themistius’ (rhetorical) question, ‘Where are the models of these proportions in that from which these (spontaneously generated) animals are bom if they have not been formed previously in nature as proportions prepared and ready to produce any possible species of animal and made suitable for the production from it of some animals?’ ’ GA 74o’’32-3, 767‘‘i7-i8, '’15-20; PA 663'’22-5; cf. PA 64o‘‘i9-33. * For Aristotle, the parent determines the character of the offspring’s development in virtue of being the producer of the seed (e.g. GA i. 18-21, esp. 726'’5 ft'., 766'’7-is; PA 640*22-7, 64i'’27-642“2). The theory behind this, and thus behind Aristotle’s ‘man begets man’ mantra, is fully explicated in GA. That theory is much too complex to treat properly here. The basic idea is that the complex changes involved in forming the embryo ultimately derive from the activity of the generating parent’s nutritive soul (cf. DA 2. 4), which Aristotle tells us is the nature of each thing (74o'’35-74i‘‘2). It is this nature that is transmitted in the parent’s seed during reproduction (see n. 5 above). Aristotle explains this idea by saying that the parent’s seed contains a set of genetic ‘movements’ that are derived from the various ‘potentials’ that make up the animal’s formal nature (734'’8 ff., 734'’20 ff., 738'’io-i8, 743*27-34); each potential in the animal’s formal nature corresponds to an inheritable feature of its sensible form, including those that belong to it qua individual and qua member of a particular species and genus (768*34—7, '’12-14). The seminal movements that code for each of these features are acquired during the production of the seed and then imparted to the embryonic matter at the moment of conception (737*19-23). The result is that the embryo’s material is organized according to the parent’s shape and form, since those movements were originally initiated by the parent’s nutritive soul, i.e. its nature (767'’i6—21). This theory helps shed light on Metaph. io32'’24-5.

i88

Devin Henry

absence of parents. In these cases we still find a continuous gener¬ ation of things of the same determinate form, which suggests that there are certain fixed proportions {logoi) proper to the abiogenetic kinds as well, proportions according to which ev^ery new member of a given species spontaneously comes to be. What we do not find, however, is any parent organism standing as the model for these proportions; these creatures are generated from putrefying ooze and not from something that already bears the form in actuality. Thus, Themistius concludes, we must have recourse to separately existing paradeigmata: for example, we must posit the Form Sea Urchin over and above the individual sea urchins in order to ex¬ plain how the particular instances of that Form could spontaneously come to be in various pockets of putrefying material stuff. Themistius’ next move is to generalize this conclusion to cover all cases of generation, even those where offspring come to be from parents of the same kind, as when man begets man and horse begets horse: Do not be deceived by contempt for such animals (as are generated spon¬ taneously), but reflect that we have more admiration for the skill of an artisan who makes something frorri clay than for what he makes from gold and ivory; if you examine carefully what happens with animals bigger than that, you will find that nature proceeds in exactly the same way. Although man is only begotten from man, the father exerts no craftsmanship in his composition of this, which cannot be in another state better than the state he is in. He comes to be in this state only because of the proportions and forms which have been put into the nature of each of the substances, not by reason of any craftsmanship on the part of the father, but as the result of the proportions.’ Themistius’ point here is that when we turn our attention from the spontaneous cases to the rest of the animal kingdom we will see more clearly that even in the central cases of generation, while it is true that a horse does not come to be from the human seed nor a human from equine seed, this is not because the producer of the seed is itself a particular human or horse. For Themistius, ‘because its ’ I have added the words in angle-brackets (and the emphasis) to Genequand’s translation. This has the ring of the doctrine of ‘pre-existence’, as it became known in the late 17th cent., according to which the seed or “'germ’ of the preformed parts is not produced by the parent organism but is rather ‘created by God at the beginning and is conserved in that state until the moment of its development or revolution' (J. Farely, The Spontaneous Generation Controversy from Descartes to Oparfn (Baltimore, 1974), 12). Seen. 10.

Themistius and Spontaneous Generation in Aristotle

189

parent was a human’ has no more explanatory value when it comes to understanding why this particular offspring has the human form than ‘because it was generated from putrefying ooze’ explains why t/zat particular organism has the sea-urchin form rather than that of any one of the other animal forms that could have come from it. In both cases we must have recourse to a Platonic Form. It will be the part of the explanation that relates the logos of a thing’s generation to its corresponding Form that ultimately explains why one kind of animal comes to have one form while another kind comes to have another form, since not all offspring come from something that already bears that form in actuality. What spontaneous generation shows, then, is that the parent is causally inefficacious in determining the form of its offspring in the sense that it neither contributes to the character of the ‘proportions’ that exist in its seed nor houses the model after which the particular offspring comes to be. The parent’s form, like that of its offspring, is simply an effect or expression of those formative principles which have been put into its nature as pre-existing proportions and forms ‘prepared and ready to produce’ whatever is begotten from it spe¬ cifically. For Themistius, nature is an irrational agent which, like the irrational part of the soul in Aristotle’s Ethics, follows a rational plan (logos) in producing what it produces. Yet, as the example of spontaneous generation makes clear, the logoi which it follows in producing new organisms have been previously put into the natures of each substance by a higher cause (the Demiurge?) and modelled after Platonic Forms—and not, as Aristotle proposed, after the enmattered forms of particular generators (‘for not everything that is begotten has a like from which it is begotten’). Themistius says that the nature in the seed produces things in an orderly way because it is ‘somehow inspired by a cause nobler, worthier, and higher in rank than themselves’ in the same way that a person who speaks in tongues is inspired by the gods. Compare Simplicius’ suggestion that nature is an instrument of the divine (B. Fleet, Simplicius: On Aristotle Physics 2 (London, 1997), 314. 9-24); see also Sorabji Sourcebook, vol. ii, § i(b). Sorabji and Genequand originally disagreed as to whether Themistius’ higher cause is the ‘soul in the earth’ (Sorabji) or the ‘World Soul’ (Genequand), I suggested to Sorabji that if we are to take Themistius’ argument in the context of spontaneous generation, then he must be referring to Plato, Statesman 271 A ff., where the generation of the so-called earthborn race (spontaneously generated humans) during the time of Kronos is attributed to ‘souls planted in the earth [presumably by the secondary gods] like seeds’, in which case ‘soul in the earth’ would be correct. If I am right, then Themistius is saying that the formative principles that were originally responsible for the spontaneous generation of human beings during the time of Kronos are the very same ‘proportions and

190

Devin Henry

While Themistius’ solution to the metaphysical problem that spontaneously generated organisms presents may be less than sat¬ isfying, the force of his objection should not be underestimated. For it threatens to remove one of the load-bearing p^remisses of Aris¬ totle’s metaphysics, namely, the idea that particulars are sufficient to explain the coming-to-be of other particulars. Themistius’ chal¬ lenge is plain enough. Whence do the many spontaneously gener¬ ated organisms derive their forms if not from pre-existing parents? What is impressive is the general strategy Themistius employs in exploiting this challenge in order to undermine Aristotle’s attempt to remove the need for separately

paradeigmata by turning

his argument on its head. According to Themistius, Aristotle tries to show that Forms are unnecessary when it comes to explaining the generation of particulars on the grounds that ‘a man is be¬ gotten by a man who is a particular man’ (1070*26—30). However, Themistius argues, spontaneously generated animals show us that Aristotle got things the wrong way round: it is the parent’s form that is superfluous. For Themistius, the weakness of Aristotle’s ar¬ gument is that it overestimates the role of the particular parent in the generation of its oflFspring, since it is obvious that living things do not need parents to come into being. What they do need, how¬ ever, is something similar in form to stand as the model after which they come to be (‘since the generation of something requires some¬ thing similar’: see n. 27 below). And this, he says, is supplied by the Forms. What I want to do now is evaluate Themistius’ objection by turning to Aristotle’s own account. As we shall see, Themistius was apparently unaware of the fact that Aristotle did not overlook the philosophical implications of ‘the many animals which are not born from their likes’ at all, for he had already anticipated his objection in Metaphysics Z 9.

forms’ that presently exist in the seeds of animals which reproduce sexually while they still remain in the earth as the moving cause for all those that are generated spontaneously. In each case, what Aristotle takes to be the organism’s formal nature is actually the proportions and forms that have been previously put into us by a higher cause, namely the soul in the earth, and which are modelled after the Forms. (Genequand eventually agreed with Sorabji, and so ‘soul in the earth’ appears in Genequand’s translation in the Sourcebook, which is the translation I have used here.)

Themistiiis and Spontaneous Generation in Aristotle

191

2. Why Platonic Forms are unnecessary The argument in Metaphysics A 3 (the target of Themistius’ attack) is in fact a condensed version of the argument Aristotle presents at the end of Z 8: Obviously, then, the cause which consists of the Forms—as some are ac¬ customed to speak of the Forms, if they are something over and above the particular things [Tiapa rd /ca0’ eWaara]—is useless [ov9ev ypT^atp,?]], at least with respect to coming-to-be and being. Nor for this reason should these Forms be regarded as self-subsistent substances [ovalai Kad' avras]. Indeed in certain cases it is even obvious that the generator is of the same kind as that which is generated, not, however, identical, nor one in number, but one in form, e.g. in the case of natural things (for man begets man). . . . Thus, it is obvious that it is not necessary to set up a Form as a pattern [oiidev Set ws TrapdSeiyfjia etSoj Karaa/ceud^eiv], for we should have looked for

Forms in these cases especially, since (natural things) are substances in the strongest sense, but the thing which generates is sufficient to produce and be a cause of the form in the matter [iVavdi'

to yewwv Troirjaai /cat tov eiSovs

aiTiov elvai ev tt) vXt]]. (io33*’26—1034^5)

Themistius’ objection can be seen to follow from a certain interpre¬ tation of this argument. According to this interpretation, the central premiss in the argument against the Forms is the idea that partic¬ ulars are always generated by other particulars of the same form-, in other words, the fact that ‘man begets man’. It is this feature of biological generation—the making of another like the producer— that Aristotle thinks allows us to dispense with separately existing paradeigmata." And yet, at the same time it is also what Themistius sees as rendering Aristotle’s argument insufficient as a refutation of them, for not everything is born from its like in form. It is easy to see how Themistius’ objection follows when Aris¬ totle’s argument is interpvreted in this way. If what makes the appeal to Forms unnecessary in the sphere of coming-to-be is the fact that the producer is formally identical with what it produces (as when man begets man), then spontaneously generated animals are obvi" This interpretation (which I am here suggesting is the same interpretation behind Themistius’ objection) is put forth by J, G. Lennox in ‘Are Aristotelian Species Eternal?’, in Lennox (ed.), APB 131-59 (see esp. § iv). According to Lennox, the analysis of biological generation offered in Metaph. Z 7-8, according to which ‘Individuals come to be sorts of things by becoming such as the generator: and this is to become one in form with it’ crucially underlies the final argument of Z 8 (p. 151).

192

Devin Henry

ous counter-examples. However, this line of interpretation faces a serious problem. After concluding in Z 8 that there is no need to set up a Form as a model or pattern after which particular sub¬ stances come to be, Aristotle immediately procee^is to the question of spontaneous generation in Z 'g. Clearly the reason is that Aris¬ totle, like Themistius, sees this as a threat to that conclusion. This is extremely significant for interpreting the argument of Z 8, and thus A 3, because it is equally clear that Aristotle thinks he can show that instances of spontaneous generation do not escape its conclu¬ sion. As such, whatever we take to be the reason Aristotle thinks one can dispense with the Forms, it must be equally applicable to those cases he recognizes in Z 9 as being spontaneous.*^ This pre¬ empts Themistius’ objection and shows that the interpretation of the argument in /I 3 that motivates it is in fact misguided. For, as we have said, if what makes the Forms unnecessary is the formal re¬ semblance between the generator and the thing generated (as when man begets man), then spontaneously generated animals become unavoidable counter-examples: sea urchins do not come to be from anything even remotely like them in form. In this sense, then, the project of Z 9 would have been, in vain. The problem with Themistius’ interpretation of Aristotle’s ar¬ gument in Z 3 is that it overestimates the role of ‘man begets man’ in establishing its conclusion. A careful reading of the expanded version of the argument in Z 8 (translated above) reveals that it is not because man begets man that Aristotle thinks the Forms are unnecessary: that is, it is not because he thinks every particular organism that comes to be comes to be from its like in form, but because in each case of coming-to-be the source of the form can be shown to pre-exist in the particular, and thus particulars are suf¬ ficient to produce other particulars. The fact that man is begotten by man is not offered as the reason why particulars are sufficient to produce and be the cause of the form in the matter; Aristotle only says that ‘in certain cases’ {IttI

. . . tlvcdv)—as

when man begets

man—‘it is even obvious’ {koI (f)avep6v) that this is so. For when man begets man we can actually see that the source of form pre-exists This same argument will also be relevant with respect to the last section of Z 9 (starting at io34*’7), which states that in the case of substances another substance ‘must’ pre-exist to produce it, ‘for example, an animal if an animal is coming to be’ (io34*’i6-I9). At the very least I034*’i6-i9 must be questioned on the grounds that it cannot possibly be reconciled with the discussion of spontaneous generation that precedes it (or Aristotle’s belief in spontaneous generation in general).

Themistius and Spontaneous Generation in Aristotle

193

in the generator who is a particular man (cf. GC 2. 9, 335^20-4). Thus, the reference to man begetting man, both in the version of the argument in A 2 and here in Z 8, serves only as an example of those ‘obvious’ cases where the source of form pre-exists in the particular; however, they are not the only cases where this holds. On this reading, instances of spontaneous generation like selfhealing and abiogenesis will be problem cases for Aristotle, not because they escape the argument against Platonic Forms as The¬ mistius had supposed—that is to say, not because they do not come from their likes in form—but because it is simply not obvious in these cases (contrasting irrl

fiev 8rj tcvcdv kul

(f)avep6v at 1033^29—30)

that the source of the product’s form pre-exists in the particular. And yet, the fact that Aristotle immediately confronts this problem in Z 9 suggests that what he wants to do there is show that even when things do not come to be from anything formally identical with themselves—for example, when a patient is healed of his own accord without the aid of medical treatment or when the members of a particular species of organism all arise spontaneously from decay without the help of parents—the product can still be shown to derive its form from some pre-existing particular or other, and so there is in fact no violation of the rule. It is to this that we now turn.

3. Spontaneous generation in Metaphysics Zeta The general structure of Z 7-9 is as follows. Aristotle discusses three kinds of coming-to-be. First, there are those cases of natural generation where the product is generated from something of the same name (e^ ojjiojvvpiov, io34‘’22), as when man begets man. In this case it is the complete substance, the form-in-the-matter (e.g. Socrates), that pre-exists in actuality. The second type of coming>

” Compare Lennox, ‘Are Aristotelian Species Eternal?’, 153: ‘One might respond that this analysis [where the feature of biological generation that allows us to dispense with the Forms is the idea that the producer and product are formally identical] works fine for one small class of sensible objects, but that is all.’ Lennox replies that Aristotle’s account of artificial production is ‘self-consciously parallel to his model of biological generation’ in this respect. ‘Thus, this model has far greater generality than might at first be supposed, at least as an account of substantial generation.’ I think the objection Lennox considers here has much more force than he allows, for his defence of the interpretation in question does not extend far enough to cover the spontaneous cases. As we shall see, Aristotle extends the synonymy principle beyond formal identity to include synonymy with respect to matter.

Devin Henry

194

to-be is artificial production. Here the product is generated from something that bears the same name as a part of itself

{Ik ixepovs

oficnvviiov, 1034^23), as when a carpenter builds a house or a doctor heals a patient. In these cases the product is not generated from the complete substance pre-e?:isting in actuality—for example, a house is not generated from an actual house (the form-in-the-matter); rather, it is the substance without the matter, i.e. the essence, that pre-exists.With artificial productions the source of form is the art, which is the form of the product in the artist’s soul. ‘Therefore it follows that in a sense health comes from health and a house comes from a house, that which has matter from that which has not, for the art of medicine is the form of health and the art of building is the form of the house’ (1032'’! 1-14). We can note here that when Aristotle says of artificial productions that the product is generated from something which bears the same name as a part of itself (e^r fiipovs), he does not mean a part in the sense that the door is a part of the whole house or an arm is a part of the whole body; these are parts of the whole in the sense of matter (Z 10). Rather, by ‘part’ he means one of the two metaphysical parts into which a hylomorphic compound can be analysed, the form vs. the matter.'^ With artificial productions, producer and product are synonymous with respect to form only, not with respect to matter: the producer’s matter (e.g. the artist’s flesh and bones) does not match the product’s matter (e.g. stones and mortar, if a house is coming to be). On the other hand, when man begets man the product and producer are synonymous in both respects: the product and producer are the same in form and matter together. However, Aristotle also recognizes a third mode of generation in which the product arises spontaneously and of its own accord {dno TavToptaTov),

as when a body heals itself or an animal comes

into being from decay.** The project of Metaphysics Z 9 is thus to show that even in cases of spontaneous generation the source of the product’s form pre-exists in the particular in some sense and so does not escape the conclusion of the argument in Z 8 (and thus A 3). When an organism is generated spontaneously, there is a sense in which Aristotle thinks it can be said to come from some pre'■* Cf. 1032*14: ‘by “substance without matter” I mean the essence’. Cf. Metaph. A 25, 1023*19-23. We can notice here that the example of self-healing will stand to the doctor healing the patient as abiogenesis stands to man begetting man.

Themistius and Spontaneous Generation in Aristotle

195

existing substance bearing the same name as itself, not, however, with respect to form, but with respect to matter (for the matter is just as much a pre-existing part of the product as the form: cf. 1032^32-1033“!). Here synonymy exists at the material level.*’ In this w'ay the three types of synonymy match up the three types of coming-to-be with the three ways in which substance or being is analysed in Zeta: (i) natural generation proceeds from the com¬ posite (form-in-matter); (2) artificial production proceeds from the form or essence (form without matter); and (3) spontaneous gener¬ ation proceeds from the matter (matter without form).'* Thus, by extending the notion of synonymy to cover spontaneous generation in Z 9, Aristotle is also able to show that in every case of comingto-be there is an identity (i.e. synonymy) between the product and that from which the product comes. Most of the work done to show how the source of form in spon¬ taneous generation pre-exists in the particular (and so is covered by the argument against the Forms) is taken over by the model of art and the example of self-healing. In Z 7 Aristotle introduces us to the difference between artificial processes that begin from the essence (the form in the artist’s soul) and those that arise sponta¬ neously. The example of health is chosen to illustrate this difference because it is among those artificial products that come to be in both ways (contrast houses at i034“io and statues at PA 64o“3i). A dis¬ eased body is such that it can be healed either by means of medical treatment or of its own accord without the help of medical treat¬ ment. When the coming-to-be of health originates from the art of medicine the healing process is divided into two stages, reasoning and production proper. The reasoning begins from the doctor’s knowledge of the essence and proceeds through a series of inter” On the idea of synonymy at the level of material causes see GC 2. 8, 33S°923 (esp. ‘for all of them are fed by the substances which are the same as their constituents’), and HA 8. 2, 59049-27 (esp. ‘Of testaceans that are incapable of local motion, some subsist in fresh water . . . just as they originally came into being from the same’). Note, however, that material-level causes in the biology refer to a much wider class of causes than simply that out of which an organism is formed. For example, at PA 64o'’4-i6 the material ‘source’ is taken to include the movements initiated by things like heat, cold, and pneuma (see also Phys. i98'’io-i6; GA 13; Metaph. 984''5-io). By ‘matter without the form’ I mean that a sea urchin comes from material stuffs devoid of the sea urchin form (analogous to menstrual fluid). We might also notice that these three modes of generation match up with the three levels of being in DA-. first potentiality (matter); second potentiality/first actuality (the carpenter’s art); complete actuality (the ensouled body, e.g. Socrates).

196

Devin Henry

mediate steps until it reaches a conclusion about what particular motion or change is required to bring about the desired end, e.g. heating the patient (the efficient cause of health): ‘Now a healthy subject is produced as a result of this reasoning, ^ince health is soand-so, if the subject is to be healthy, then such-and-such a quality must be present, e.g. homogeneity; and if this, then heat (must be produced)’ (io32'’6—9). The conclusion of this deliberation will be something in the doctor’s power; for instance, the requisite heat can be produced by means of a rubbing motion, and this is some¬ thing in his power. The part of the process that proceeds from this conclusion to the actuality of the product (a healthy body) is called the ‘production’ of health (^’lo). Thus, when an artificial process proceeds from the artist’s knowledge of his craft, the starting point of the reasoning is the form or essence of the product, while the starting point of the production is the conclusion of that reasoning, and this conclusion will be something in the artist’s power. After concluding in Z 8 that there is no need for the cause that consists of the Forms because the source of form pre-exists in the particular substance from which the product comes, Aristotle im¬ mediately turns to the question of spontaneous generation. The reason, I have suggested, is that he sees such processes as a threat to that conclusion precisely because it is not obvious in these cases that the source of form does pre-exist in the particular. Zeta 9 opens with a statement of the aporia to be solved: why is it that some things, e.g. health, can come to be and exist of their own accord as well as by art, while other things, e.g. houses, cannot? The reason (aiVior), Aristotle says, is that the matter of a spontaneous generation—which is both the starting point of the coming-to-be and that in which there exists to begin with some part of the thing produced (io34‘*io-i3; cf. i032'^32-i033“i)—has the capacity to be set in motion by itself {KLveladai

avTrjs). And yet, while all

natural bodies have in themselves some elementary source of change (e.g. fire qua fire can initiate upward motion by itself), the matter of a spontaneous generation is such that it can move itself in the particular way required to realize a particular form (cf. 1034“! 516). Those products whose matter cannot do this (those that lack the capacity to move themselves in the particular way required) can only acquire their forms from some other pre-existing substance that already has that form in actuality:

Themistius and Spontaneous Generation in Aristotle

197

For this reason some things will not exist apart from the one who possesses the art, while others will. For motion will be initiated by those things which do not possess the art but are themselves capable of being moved either by another thing that does not possess the art or by a part (of the product). (io34"i8-2i)

The first part of this text ('‘18-19) expresses Aristotle’s negative point. Those things that cannot come to be of their own accord must derive their forms from some other pre-existing substance that already has that form in actuality. Houses are like this. Stones and mortar (the matter of a house) are not the sorts of material that can spontaneously organize themselves into the form of a house, for they do not contain the source of that form in themselves. The builder’s art is the source of the housebuilding motion in the builder. This is why we do not find houses without housebuilders. The second half of the text (“19-21) supplies Aristotle’s positive point. When it comes to producing, and being a cause of, the form in the matter, some things are entirely se//-sufficient. These things can come into being of their own accord because the requisite mo¬ tion can be supplied by a pre-existing part of the product itself, namely the matter. A living body is like this. In some cases the body can restore the form of health of its own accord without the need for medical treatment simply by setting up the same motion that the doctor would produce by rubbing, i.e. the heating motion. And this, we have seen, is the starting point of the process towards health (1032*^21—7). The analysis of self-healing is supposed to show that sponta¬ neous productions of this sort are not actually counter-examples to Aristotle’s argument against the Forms after all. They do not require setting up a Form as the cause of the form in the matter, since even here the source of form pre-exists in the particular. It is this analysis that Aristotle wants to transfer to cases of abiogenesis, the spontaneous generation of organic substances in the absence of parents: The matter of those things which come to be of their own accord . . . can bring about by itself that kind of change which the seed brings about; the matter of those things which cannot bring about that kind of change cannot be generated otherwise than from themselves. (io34'’4-7)

(By ‘from themselves’ Aristotle means intraspecies reproduction.) What Aristotle is attempting to do in this text is extrapolate from

Devin Henry

198

the analysis of self-healing given earlier in the chapter to the spon¬ taneous generation of living things. His point, then, is that when living things come to be of their own accord, the material source of the organism is also its source of form. Aristotle n^ed not be saying anything more here than that the fnaterial cause of some organisms, taken in the broad sense to include certain environmental condi¬ tions such as the place where the foetus develops (e.g. GA oi TOTTOL atViot), is capable of effecting the same kind of changes that a seed (or the nature in the seed) brings about in natural generation. The best illustration of this is afforded by the case of the sea urchin. At PA byp’’15-31 Aristotle offers a functional account of the shells of testaceans. There he tells us that the sea urchin has the best defence mechanism of all the testaceans, because ‘it has a good, thick shell all around it fortified by a palisade of spines’. In GA 5. 3 we are given an account of the formation of this structure in terms of various material-level causes surrounding the animal’s develop¬ ment, such as its relative lack of natural heat and the congealing effects of its surrounding environment: Thus, with wild animals the reason (they have hard hair) is that they live outdoors in the open air, but in other cases it is the environment’s being of such a kind that is responsible for this [atViov ...

6 tottos tolovtos

clSv].

This is shown by what occurs in the case of the sea urchins, which are used as a remedy for cases of strangury. These creatures, because the sea water in which they live is cold on account of its depth (60 fathoms or even more is the depth at which they are found), although they are themselves small, are covered by long, hard spines: long, because the growth of the body is diverted there (for as they possess but little heat and do not (properly) concoct their nourishment they contain a great deal of residue, and it is out of residue that spines and hair and things of that sort are formed); hard and petrified, because of the cold and its congealing effect. {GA 5. 3, 783‘‘i8-29)

In this case, the same limit and structure that would otherwise have been imposed on the growth of the organism by the nature transmitted in the parent’s seed are here supplied by the surround¬ ing environment: the cold water congeals the streaming residue of inconcocted (i.e. partially concocted) nourishment into long, hard spikes as it protrudes from the urchin’s body.’’ ''' In GA 3.

II,

762*19-34, Aristotle says that the place where a testacean devel¬

ops may act as a kind of mould and thus can be seen as the source of its shape (in particular, its shell). With a human being, on the other hand, the place where the foetus develops (the womb) is not ‘human-shaped’; here Aristotle regards the em-

Themistius and Spontaneous Generation in Aristotle

199

We could also take Afistotle’s point in 1034^4-7 to be much stronger than this. For an analysis of the general mechanism of spontaneous generation in the biological works reveals a much closer affinity to sexual reproduction than the Metaphysics account suggests. To see this we need to turn to the Meteorologica and the process of putrefaction, for spontaneous generation does not involve just any kind of material stuffi hut putrefying material. In the Meteorologica Aristotle tells us that putrefaction (4 arjfif) is a special type of destruction in which a body undergoes a pro¬ cess of becoming increasingly dry until eventually ending up as soil or dung (379^*8-9, 22-3; cf. HA 569T0-15, 57o“7-i2). What explains this process, Aristotle says, is the fact that when a body is decomposing its internal heat gets expelled, which in turn causes its natural moisture to evaporate along with it (379‘‘23-4). And since there is nothing left to draw in any new moisture (attracting mois¬ ture and drawing it in being a function of a body’s internal heat, 379*25—6), the putrefying thing eventually dries out. It is this loss of heat and moisture that comes to form the basis of putrefaction. However, Aristotle is careful to point out in the GA that it is not so much the putrefaction of the body that is essential for the spon¬ taneous generation of a kuema (an embryo); putrefaction and the thing putrefied are only a by-product of this event {GA 762*13-15). Rather, what is important is the interaction between the heat and moisture being released into the atmosphere that happens in con¬ nection with putrefaction. It is here, outside the putrefying ooze, that Aristotle thinks the real excitement takes place. At 379*^7-9 he tells us that an embryo is spontaneously generated when ‘the natural heat that is expelled {from the putrefying subject) fuses to¬ gether the substance that is being thrown off along with it’.^° It is the amalgamation or fusion (awLordvai) of the putrefying thing’s natural moisture (the passive factor) by its natural heat (the active factor) as they are released together into the atmosphere that forms the arche of abiogenesis.^' When we combine this with the canonical account of spontaneous bryo’s structures and shapes as being ‘programmed’ into its formal nature. Compare Lennox’s commentary on PA

'\n Aristotle: On the Parts of Animals (trans.

with notes) (Oxford, 2001), 287. The substance being thrown off here is the body’s natural moisture, which is elsewhere identified as the ‘sweet’ {to gluku, 762“i3). The ‘sweet’, i.e. sugar, is the useful residue extracted from nourishment during digestion {De sensii 44i'’24442‘“io, “26-30).

Cf.

translated in n, 23 below.

Devin Henry

200

generation in GA 3. 11, the following picture begins to emerge. At GA 762‘‘9-i3 Aristotle says that ‘all things formed (by sponta¬ neous generation) manifestly come into being in connection with putrefaction [ixera arnpews] and a mixture of rainyvater’. From the Meteorologica account we .can infer that what happens next is that the rainwater mixing with the materials on the one hand, and the proactive effects of the heat from the surrounding environment on the other, combine to initiate putrefaction. The process that ensues (the thing that happens ‘in connection with putrefaction’) includes two important events: evaporation, whereby the ‘sweet’ (the useful material) is extracted from the putrefying materials in the form of a moist vapour (762“:2-13); recombination, whereby, as the sw'eet is being thrown off from the putrescent mixture, it is immediately reorganized and concocted into a new substance by the heat being expelled along with it (379’’7-9, 743'“35-6, 762‘‘i2-i3, ‘’14-16) or, alternatively, by the heat from the surrounding air.^^ This same basic process of extracting and amalgamating useful material that we find occurring in spontaneous generation is also said to underlie the formation of a kuema out of menstrual fluid by the action of semen.The evaporation/recombination process that takes place in spontaneous generation can thus be seen as corresponding to the conception of a sexually generated embryo: in both processes a kuema is formed through the action of some heat that works by separating out the useful components and fusing them together to form a new substance with the capacity to grow (cf. 762‘‘35ff., and see also lAoyd, Explorations, 116-17). This, then, is another way in which the material causes of a spontaneously generated organism may be said to be capable of bringing about that same kind of change which the seed (i.e. male semen) brings about. Whatever Aristotle means by io34‘’4-7, one thing is clear: that statement is supposed to show us that abiogenesis, like self-healing, is covered by the conclusion of the argument in Z 8 (and thus yl 3) Cf. De sensu 4, 442“7-8. Aristotle tells us that the action of external heat in external bodies (putrefaction?) will bring about the same basic changes that an animal’s nutritive system brings about, namely the extraction of nutrients from rav/ materials and the working up of these nutrients into viable organic compounds. " Compare 739'’20-9 with 762‘’6-9: ‘We need to understand that even in the case of animals that reproduce (sexually) it is the heat in the animal that makes the starting point of the embryo [rijr apxrjv tov /ruT^/aaros] out of the incoming nourishment by separating out and combining the residue.’ (This reading drops the comma at ‘’8 and takes the subject of Troief to be rrjv dpxrjv tov Kvqp.aTos rather than to TrepiTTcu^a.)

Themistius and Spontaneous Generation in Aristotle

201

and so does not require setting up a Form as a model after which these products come to be. Thus, contrary to what Themistius sug¬ gests, instances of spontaneous generation do not show Aristotle’s argument to be insufficient as a refutation of the Forms simply because ‘nature does not generate these things from their likes in form’. For it does not follow from the fact that sea urchins do not derive their forms from other sea urchins that they must derive them from separately existing paradeigmata. All spontaneous generation shows is that the source of their form is not a parent organism of the same kind. For Aristotle, plants and animals that are generated spontaneously derive their forms from their material causes, which may equally be considered a pre-existing part of the product itself^'* To close this paper, I would like to offer what is (as far as I know) the only other attempt to resolve Themistius’ dilemma, that of Aquinas. While several scholars have attended to various problems surrounding Aristotle’s theory of spontaneous generation, includ¬ ing the problem of synonymy (the fact that sea urchins do not come from anything formally identical with themselves), not one of them has addressed Themistius’ problem, namely the fact that Aristotle’s belief in abiogenesis appears to be inconsistent with his rejection of the Forms.

4. Aquinas’ reply to Themistius In his commentary on Metaphysics Zeta, Aquinas discusses an ob¬ jection raised by an unnamed source who appeals to the existence of spontaneously generated animals to expose a weakness of Aris¬ totle’s attack on the theory of Forms: Now it must be observed from what has been said here that it is possible to solve the problems facing those who claim that the forms generated in these lower bodies do not,derive their being from natural generators but from forms which exist apart from matter. For they seem to maintain this position chiefly because of those living things which are generated from decay, whose forms do not seem to come from anything that is similar to them in form. . . . Nor does it seem (to them) that the argument which the Philosopher used against those who posited separate exemplars, viz. that Of course, synonymy will only apply to the materials out of which the organism is formed (e.g. sea water); however, the source of its form is to be traced to its ‘material origin’ in the wider sense that includes these materials as well as the other material-level causes mentioned above.

202

Devin Henry

the forms in things causing generation are sufficient to account for the like¬ ness of form in the things which are generated, holds in all cases. (Aquinas, In Arist. Metaph. 7. 12. 8, § 1455, trans. Rowan, slightly modified)”

Aquinas claims that the problem.can easily be resblved by a careful examination of Aristotle’s text. His solution begins with an attempt to reconcile the doctrine of spontaneity from Physics 2. 6 with the fact that ‘animals which are generated without seed from decay do not seem to be produced by chance but from some definite agent’ (§ 1402). And the way he does this is by arguing that spontaneously generated animals are in one sense accidental and in another sense for the sake of something. On the one hand, Aquinas says, when the cause is referred to par¬ ticulars operating in the sublunary world, e.g. heat, the generation of these animals is accidental and the result of chance, for ‘the heat, which causes decay, is not inclined to have as its goal the generation of this or that particular animal which results from decay, as the power in the seed has as its goal the generation of something of a particular type’ (§ 1403). Yet, on the other hand, when the cause is referred to ‘the powers of the heavens’, by which Aquinas pre¬ sumably mieans the rotation of the sun and moon, whose motion regulates generation and corruption in the sublunary world, we say that spontaneous generation comes to pass for the sake of what re¬ sults. For in this case the animal that comes into being will not be accidental but ‘directly aimed at’, since it is the goal of the celes¬ tial rotation that ‘all forms existing potentially in matter should be brought to actuality’ (§ 1403). However, Aquinas continues, there is also a formative power in the matter out of which the organism comes that corresponds to the active principle in the father’s seed. For Aquinas, this is Aristotle’s point at Metaph. io34‘’4-7: the matter of a spontaneously generated organism moves itself just as the seed would move it, namely ‘with the aim of generating an animal’ (§ 1453). In the non-spontaneous cases the power to induce the form that is present in the seed comes from the sire and from the celestial body; in the spontaneous cases ” Thomas Aquinas, Commentary on the Metaphysics of Aristotle, trans. J. P. Rowan, vol. ii (Chicago, 1961). All translations of Aquinas are from Rowan. Al¬ though Aquinas does not mention Themistius by name here, it is clear, both from the content of the objection itself and from the fact that he has just finished discussing Averroes’ commentary on the Metaphysics (the work in which we find Averroes at¬ tributing this very objection to Themistius), that this is whom he has in mind.

Themistius and Spontaneous Generation} in Aristotle

203

this power derives from the celestial body alone: for in this body ‘all forms of things which are generated are present virtually as in their active principle’ (§ 1457).^'’ W e can summarize Aquinas’ solution to the problem of abiogenesis by setting it out in two parts. First, the forms of spontaneously generated animals exist potentially in the matter from which they arise (the passive power corresponding to the female?); the actu¬ alization of this potentiality is the goal of the celestial rotation (so that ‘all forms existing potentially in matter should be brought to actuality’). Second, Aquinas says that the formative power that determines the specific character of the spontaneously generated animal is derived from the celestial body alone ‘in which all forms of things which are generated are present virtually as in their ac¬ tive principle’. In sexual reproduction the active power is derived both from the celestial body and from the nature of the generat¬ ing parent. It is because of this second proposition that Aquinas thinks spontaneous generation is covered by the argument against the Forms: everything Aristotle needs is provided by the power of the divine body. W’hatever the merits of Aquinas’ solution, it clearly fails to meet Themistius’ challenge. In the first place, Themistius would agree with Aquinas that spontaneously generated organisms arise from forms pre-existing potentially in the matter from which they come. Themistius’ point is simply that Aristotle is wrong to suppose that where animals and plants are begotten from seed the parent organ¬ ism supplies these forms. As the example of spontaneous generation shows, the forms proper to each kind of animal already exist in na¬ ture ‘prepared and ready to produce any possible species of animal’. In those animals that reproduce sexually, the parent’s form must be taken only as an expression of these same forms and not as the Aquinas may be referring here to Aristotle’s statement at Phys. igf’iy that ‘man is begotten by man and by the sun’. Aquinas’ point, then, would be that spontaneously generated animals are begotten by the sun only. However, Aquinas goes well beyond anything Aristotle means by saying man is begotten by the sun in claiming that all the forms of things in the sense of their formative principles are contained ‘virtually’ in the celestial bodies. (All Aristotle means is that the timing of an animal’s gestation period and lifespan are ‘measured’ by the periodic motions of the sun and moon: GA 4. 10, 777'’i6 ff.) Anthony Preus sees Aristotle’s belief in spontaneous generation as forcing us to accept a similar interpretation: ‘The souls of heavenly bodies must contain all these species eminently [Aquinas’ “virtually?”], for they do not contain them formally, in the Cartesian sense of these words’ (A. Preus, Aristotle’s Biological Works (New York, 1975), 89, emphasis added).

204

Devin Henry

model after which they are patterned. The models for all forms are the forms that exist apart from matter (Platonic Forms). Moreover, Themistius could equally allow that the goal of the celestial rota¬ tion is to actualize the forms that are somehow latent in the matter, just as Aquinas proposes, without having to concede his point. As for Aquinas’ second’propqsition, the idea that in the central cases of reproduction the formative principle contained in the seed derives from the sire and from ‘the powers of the heavens’, whereas in the spontaneous cases it comes from the heavens alone, this sim¬ ply ends up impaling Aristotle on the second horn of Themistius’ attack. Recall that Themistius’ strategy was to generalize from the spontaneous cases where we find forms coming to be without par¬ ents to the conclusion that in the core cases the parents themselves are causally redundant, since parents are clearly not necessary for generating the form in the matter. Thus, Themistius would say, if a living organism can be generated by the powers of the divine body alone, then clearly the parental form is superfluous. And yet, on the other hand, if we say that the divine body can generate some life forms on its own but needs the help of the parent organism to gen¬ erate others (e.g. humans), then we end up with the consequence that the divine is deficient in some respect and dependent on the powers of these lower bodies to bring things into being.

5. Conclusions Looking back at Aristotle’s account of how particular sea urchins come to be fortified by a palisade of spines, it becomes clear that Aristotle does not need to appeal to Platonic Forms or divine powers to reconcile the argument from Metaphysics Z 8//1 3 with his be¬ lief in spontaneous generation. For it is easy to imagine Aristotle giving a similar account for the other abiogenetic kinds, namely by locating the source of each form in a complex set of material fac¬ tors, including the types of material undergoing putrefaction (e.g. HA 547^18 ff., 55 iT flf.), the changes brought about through vari” Themistius says: ‘In short, there are assuredly in nature proportions and forms since the generation of something requires something similar and not everything that is begotten has a like from which it is begotten. But when we need any form, we act in such a way that we know that this form cannot be produced by this act alone; this form, then, is produced as if it had been latent in something else, and it was indeed latent in the begetting nature.'

Thernistius and Spontaneous Generation in Aristotle

205

ous ‘chemical’ reactions {Meteor. 4), and a myriad of environmental factors such as the place of development {GA Still, would this succeed in vindicating Aristotle’s argument completely? While showing that the source of form is always some particular substance or other may allow Aristotle to dispense with separately existing paradeigmata (Forms), does it succeed in an¬ swering Thernistius’ further objection that the example of spon¬ taneous generation renders the parent’s form causally redundant? What of the significance of Aristotelian teleology and the central role it affords to the organism’s formal nature in determining the character of the informing movements in biological reproduction? If organisms can come into being in the absence of such causes, why should we think that parents have any genuine efficacy at all? For Thernistius, the fact that we see that nature generates some animals in the absence of parents shows that when parents are in¬ volved they play no significant role at all (as he says, ‘the father exerts no craftsmanship’). Aristotle could respond to this charge by pointing out that a generalization from the absence of parents in the spontaneous cases to their redundancy in the sexual cases is unjustified, for it depends on the assumption that nature proceeds in the same way in all cases of generation.^* This is a move Aristotle would certainly resist on empirical grounds. Cases of spontaneous generation do not disprove the causal efficacy of the parent’s formal nature in sexual reproduction any more than cases of self-healing disprove the efficacy of medical treatment in the sphere of health.^’ In other words, Aristotle could respond to Thernistius here by say¬ ing that it does not follow from the fact that parents are not neces¬ sary for generation taken universally, because not all living things are generated from parents, that where they are present they are not causally responsible for the form of the offspring.^" Aristotle’s See Thernistius’ argument above: ‘if you examine carefully what happens with animals bigger than (spontaneously generated ones), you will find that nature pro¬ ceeds in exactly the same way’. Balme takes a similar line in his commentary on PA 64o“27-35. Balme (unaware of Thernistius’ objection) takes Aristotle to be anticipating a similar objection to the statement that the seed contains a species-specific dunamis derived from the sire. Aristotle offers a similar example in GA i. 4 when he argues that the testicular system is not a necessary condition for reproduction, ‘for serpents and fish repro¬ duce, but they do not have testes’ (717“ 17-19). Notice here that it does not follow from the fact that testes are not universally necessary for reproduction, because not all animals that reproduce have testes, that in those species where testes are present they play no role in generation. For however true it may be that serpents and fish

2o6

Devin Henry

biology divides the generation of living things into two kinds, those that are naturally generated (i.e. from a nature) and those that are generated spontaneously. In the one case Aristotle takes the parent as sufficient to produce and be the cause of the offspring’s form in virtue of being the producer of the seed from which it comes (see n. 8 above). In the sp'ontanpous cases, it is the matter, taken in the broad sense described above, that supplies this cause. But in both cases, the source of the product’s form can always be shown to pre-exist in some particular or other, and this is what renders separately existing paradeigmata unnecessary.^’ King’s College London

BIBLIOGRAPHY Aquinas, Thomas, Commentary on the Metaphysics of Aristotle, trans. J. P. Rowan, vol. ii (Chicago, 1961). Balme, D. (trans. with notes), Aristotle’s De Partibus Animalium I and De Generatione Animalium / {with Passages from II. 1—3) (Oxford, 1972). -‘Development of Biology in Aristotle and Theophrastus: Theory of Spontaneous Generation’, Phronesis, 7 (1962), 91-104. are able to reproduce without testes, humans and horses cannot. If we were to apply Themistius’ reasoning to this case, then it would follow that castrated horses could still reproduce simply on the grounds that serpents and fish do. And this is clearly not a valid inference. A solution having been offered to the problem of synonymy created by the fact that sea urchins do not come from anything/orma/fy identical with themselves, the central problem of spontaneous generation becomes the fact that it not only appears to breach Aristotle’s defence of teleology in Physics 2. 8, it also seems to violate a first principle of Aristotle’s biology, ‘Nature makes the organ to suit the work it must perform, not the work to suit the organ’ {PA 694’’i3—15). Ideally he would want to answer the question of why sea urchins are covered in a palisade of hard spines by citing the function of that structure. However, this seems to be inappropriate here, since the cold water does not solidify the urchin’s nourishment into hard spines because of the function they must perform in the mature organism. In other words, the urchin does not come to be covered in a palisade of spines/or the sake of protection, but only on account of the material-level causes surrounding the development of that particular individual. The breach of Physics 2. 8 follows from this; the fact that those parts function in the mature organism as a means of defence (and thus confer a good on the creature) would seem to be coincidental and the result of how the particular urchin came to be. No one, in my opinion (with the exception, perhaps, of Lennox, ‘Spontaneous Generation’, in APB 243), has sufficiently appreciated the threat that this problem poses for Aristotle’s philosophy of biology, let alone offered any kind of strategy for solving it.

Thernistius mid Spontaneous Generation in Aristotle

207

Brague, R. (trans.), Themistilis: Paraphrase de la Metaphysique d’Aristote. Livre Lambda (Paris, 1999). Farely, J., The Spontaneous Generation Controversy from Descartes to Opa¬ rin (Baltimore, 1974). Fleet, B. (trans.), Simplicius: On Aristotle Physics 2 (London, 1997). Genequand, C. (trans.), Thernistius, In Metaph. 12, ap. Averroes, In Metaph. 12, in Sorabji, Sourcebook, ii. Physics, § i(b)(i6). Gotthelf, A., ‘Teleology and Spontaneous Generation in Aristotle; A Dis¬ cussion’ [‘Spontaneous Generation’], in R. Kraut and T. Penner (eds.), Nature, Knoivledge, and Virtue: Essays in Memory of Joan Kung (Apeiron, special issue 22/4; Edmonton, 1989), 181—93. -and Lennox, J. G. (eds.), Philosophical Issues in Aristotle’s Biology (Cambridge, 1987). Lennox, J. G. ‘Are Aristotelian Species Eternal?’, in Lennox, APB 131-59. -(trans. with notes), Aristotle: On the Parts of Animals (Oxford, 2001). -(ed.), Aristotle’s Philosophy of Biology: Studies in the Origins of Life Science [APB'] (Cambridge, 2001). -‘Material and Formal Natures in Aristotle’s De Partibus Animalium’, Proceedings of the Boston Area Colloquium in Ancient Philosophy, ii (1995), 217—40; repr. in Lennox, APB 182-204. -‘Teleology, Chance, and Aristotle’s Theory of Spontaneous Genera¬ tion’ [‘Spontaneous Generation’], Journal of the History of Philosophy, 20 (1982), 219—38; repr. in Lennox, APB 229—49. Llo^'d, G. E. B., Aristotelian Explorations [Explorations] (Cambridge, 1996). Preus, A., Aristotle’s Biological Works (New York, 1975). Sorabji, R., Four Hundred Years of Transition: Philosophy of the Commen¬ tators {200-600 forthcoming).

ad),

a Sourcebook [Sourcebook], ii. Physics (London,

0

I*

"v

1*

'-1^

14n>i*t n4/Hiru; V kM&

• • ;v»-' . .
,4*>

Kt )«r Miv ‘«{A 4 ««u* %•

" M-**^^** I

.« < I.

. ’

inrrjf*

it

WHAT DOES ARISTOTLE MEAN BY PRIORITY IN SUBSTANCE? STEPHEN MAKIN

I.

Introduction

IN Metaphysics 0 8 Aristotle discusses three different ways in which actuality is prior to potentiality. First, actuality is prior in defini¬ tion, Xoyco, to potentiality (1049^12-17). For example, in defining a capacity I cite its exercise rather than vice versa: to be visible is to be capable of being seen. Second, actuality in one way is temporally, Xpovco, prior to potentiality, and in another way is not temporally prior (i049'^i7-io5o‘‘3). For example, take something which is ac¬ tually a human being: an adult. That was preceded by something which w'as potentially a human being (a fertilized egg, or an in¬ fant). In that way potentiality is temporally prior to actuality. On the other hand, something else which is actually a human being (an adult parent) temporally preceded the fertilized egg: and in that way actuality is temporally prior to potentiality. My concern in this paper is not with these first two types of priority. Whatever the complexities of Aristotle’s treatment of ac¬ tuality and potentiality, it is fairly easy to understand what priority in definition and temporal priority amount to. But he then pro¬ ceeds to consider a third type: priority in substance, ovoia.' And it is less clear what is at issue when we consider whether one type of item is or is not prior in substance to another.^ Now there are some >

© Stephen Makin 2003 An earlier version of this paper was read to the B Club at Cambridge University, and I am grateful to those present for points raised. I would also like to thank Charlotte Witt for written comments on an earlier draft, and David Sedley, whose editorial comments considerably improved this paper. ' The discussion falls into two parts, ioso‘‘4-‘’6 and i05o'’6-ios Ta. Eor different views on the significance of this split for understanding Aristotle’s account of pri¬ ority in substance, see further the appendix below. ^ I will sometimes drop the qualification ‘in substance’ from now on, and abbre¬ viate to ‘priority’ when no misunderstanding will result from doing so.

210

Stephen Makin

'

textual grounds for thinking that Aristotle means us to understand the claim that Fs are prior in substance to Gs as the claim that there can be Fs without Gs but not vice versa. I will refer to this as the Independence Condition [IC]: Fs are prior in substance to Gs if there can be Fs independently of there bein^ Gs, while there cannot be Gs independently of there being Fs. The most important passage is from Metaphysics d 11: Some things, then, are called prior and posterior in this way, others in re¬ spect of nature and substance [rd 8c /card cfivaiv Kal ovaiav], i.e. those which can be without other things, while the others cannot be without them. (ioi9h-4)'

While the textual evidence is not decisive—if only because it is not unanimous”*—the point of this paper is to see how far we can go with [IC] as a charitable account of priority in substance. Aris¬ totle does not offer an alternative explicit account of priority in substance elsewhere, so we should pause before turning our backs on the material provided in Metaphysics A 11. And a number of commentators do appeal to Metaphysics d 11 as relevant to the dis¬ cussion in Metaphysics

0 8 of priority in substance. An excellent

example is Charlotte Witt, who also identifies the main difficulty in bringing [IC] to bear on Metaphysics

0 8: namely, that it is hard

to see how [IC] accommodates the examples given in Metaphysics 0 8—an adult as prior in substance to a child, and a human being to a fertilized egg (io5o'‘5-6): I have suggested that we interpret the priority in being of actuality to mean that actualities can exist independently of potentialities but poten¬ tialities cannot exist independently of actualities. Unfortunately, when we look more closely at Aristotle’s examples of priority in being, and his ex¬ planations of them, it is at first sight implausible to interpret the notion of priority of actuality as ontological priority. . . . [In io5o‘‘4-ii] Aris¬ totle explains the priority in being of actuality in relation to potentiality not in terms of ontological priority, but rather in terms of a teleological relation between the two. And there is no obvious connection between

^ See also Metaph. M z, lO’j’f'z—T,- Translations from Aristotle follow the Revised Oxford Translation (Princeton, 1984), with occasional deviations, except in the case of Metaphysics &, where the translations are wholly my own. “ For example, at Cat. 12, I4“29-3S and i4'"io-22, priority in nature {cj>vaei.) is explained in causal terms, rather than in terms of existential independence; and at Phys. 8. 7, 26o*’i6-I9, priority in substance is contrasted with [IC].

What Does Aristotle Mean by Priority in Substance?

211

the teleological relationship which unites potentiality to actuality, and the idea that potentialities are ontologically dependent upon actualities. Fur¬ ther, when we consider Aristotle’s examples, it is not easy to see how ontological priority applies to them. What sense can we make of the idea that a man can exist without the existence of a boy, but a boy cannot exist without the existence of a man? And what sense can we make of the parallel notion that the activity of seeing can exist without the ca¬ pacity for sight, but that the capacity for sight cannot exist without see¬ ing?'

An initial motivation for this paper is to see how far we can answer the questions posed by Witt. One might respond to them denying that [IC] is relevant to Metaphysics & 8, and appeal¬ ing to other ways of understanding priority which would render Aristotle’s examples more plausible. For example, one could de¬ velop the idea that for Fs to be prior in substance to Gs is for Fs to be better exemplars of a substance kind than Gs; and in that case it would be quite plausible to say that adults (as ac¬ tual instances of the substance kind human) are prior in substance to children (as potential instances of the kind)—that adults more fully manifest the properties in virtue of which something is hu¬ man, and so are better exemplars of the kind, than children. Aris¬ totle’s explication at

would then be perfectly clear: the

adult specimen possesses the form fully, the immature specimen does not, and to be an instance of the kind F is to possess the form F. Alternatively one might agree that [IC] is relevant to Metaphysics 0 8, but hold that there is no great difficulty in fitting [IC] with the examples in that chapter. One could appeal to the plain biolog¬ ical facts. There cannot be children without adults because there have to be adults in order that reproduction take place, and with¬ out reproduction there can be no children. But the same biological facts also establish thati there cannot be adults without children either, since every adult was a child once. And that would suggest that according to [IC], neither adults nor children are prior to the other. Or one could try to bring [IC] and the examples in Metaphysics 0 8 together by appeal to some conceptual facts: that children are ' C. Witt, ‘The Priority of Actuality’ [‘Priority’], in T. Scaltsas, D. Charles, and M. L. Gill (eds.). Unity, Identity, and Explanation in Aristotle’s Metaphysics (Ox¬ ford, 1994)215-28 at 217-18.

212

Stephen Makin

'

understood as being young-adults, while adults are not understood as being aged-children. That does make children definitionally de¬ pendent upon adults in a way that adults are not definitionally dependent upon children. But it threatens to collapse priority in substance into priority in definition, and the disciission at 1050^4— ^6 of priority in substance is supposed to add something to the material on priority in definition which is found earlier in Meta¬ physics @ 8. Or, finally, one could, with Witt, rethink the examples at i05o’*5— 6 in order to accommodate [IC]. According to Witt, for example, Aristotle’s point concerns not individual children and adults (e.g. young Candy, adult Candy), nor children and adults in general: rather, it is the plausible claim that young Candy cannot exist with¬ out the human species existing, while the human species can exist without young Candy existing.® I am going to recommend a different approach. I will argue that a closer examination of Metaphysics d ii, ioi9'‘2-i4, will lead to a more nuanced understanding of [IC] which will fit better than might at first sight appear with the examples at Metaphysics

0 8,

io5o'"5-6. Further, that understanding of [IC] will sustain a chari¬ table treatment of Aristotle’s general discussion of priority in sub¬ stance at Metaphysics

0 8, io5o‘‘4-io5o'’6, allowing us to see that

difficult passage as a plausible and well-structured series of argu¬ ments. While there will not be room in this paper to offer a detailed exegesis of Metaphysics

0 8, i05o‘’4-io5o‘’6, I hope I will be able

to say enough about that passage to bring out the advantages of my proposed understanding of [IC].’

2. Metaphysics A ii, ioi9“2-i4 I divide the discussion of priority in (nature and) substance at Metaphysics Zl 11 into three main sections:** ^ See Witt, ‘Priority’, 224: ‘What is important for Aristotle is not a particular telos—Sally as an adult—but rather the idea that the child exists for the sake of being a mature person. Sally’s telos is the type or species which she will realize, and not the token or individual she will become. On this view the end or actuality in question is the species. And, if this is right, we can make some sense of the onto¬ logical dependence of the potential on the actual. For the human species, unlike the adult Sally, exists now, and so it is possible that Sally’s existence might be dependent upon it.’

’ See the appendix below.

* See n. 18 below on the connecting sentence at 1019“! i—12.

What Does Aristotle Mean by Priority in Substance?

213

{a) 1019^2-4: a formulaic statement that priority in substance is to be understood in terms of [IC]; {b) ioi9“4-ii: an appeal to the claim that ‘being is said in many ways’; (c) ioi9‘‘i2-i4: some points concerning possibility.

{a) ioig°2-4 At ioi9‘^3—4 Aristotle states a formula; those things are prior in (nature and) substance which can be without other things, while"the other things cannot be without them [oCTa ASeyerat eivai avev aXXcov, eKCiva Se avev iKeivtov /xi)]. Suppose I seek to discover, as this formula directs, whether it is possible for there to be Fs without Gs or not. There are two points at which complexity might arise; first as regards what it is for there to be

(eivai)

Fs or Gs, second as regards its being possible (erSeyerai)

for there to be Fs without Gs. The material which follows in Meta¬ physics Zl 11 picks up each of these points in turn. Closer attention to some of this material—in particular, to 1019*12-14—will lead to an explanation of priority in terms of [IC] which is more nuanced than the original formula; and it is that more nuanced account which will sit more easily with Aristotle’s discussion of priority in substance in Metaphysics & 8.

{b) ioig‘‘4-ii If we consider the various senses of being, firstly the subject is prior (so that substance is prior); secondly, things according to potentiality and according to actuality are prior in different ways [dAAa;? to koto Svvafj.Lv Kal /cot’ evreXeyeiav]-, for some things are prior in respect of potentiality, others in respect of actuality; for example, in potentiality the half-line is prior to the whole line and the part to the whole and the matter to the substance, but in actuality these are posterior; for it is when the whole is dissolved that they will exist in actuality [SiaXvdevTO? yap Kar’ evreXey^iav eoTOi].

Commentators have typically been unimpressed with Aristotle here. Both Ross and Jaeger punctuate these lines as parentheti¬ cal, and subsidiary to Aristotle’s main line of argumentation. And both Ross and Kirwan remark unfavourably on Aristotle’s discus-

214

Stephen Makin

,

sion.'' This passage is certainly not easy, but I think it is more perceptive and impressive than is usually allowed. However, the significance of that verdict on the passage is limited. For while it is pleasing to get a clearer view of the point Aristotle is mak¬ ing in ioi9‘‘4-ii, it will turn out that it is not this passage, but the later ioi9“i2-i4—on which see (c) below—that is most im¬ portant for the purpose of fittitig [IC] to the examples of Meta¬ physics 0 8. I pass over without comment the initial remark on the sub¬ ject as prior, and concentrate instead on the distinction between actuality and potentiality. Why does Aristotle draw attention to this distinction in particular? As Ross recognizes, it is tempting to see [IC] as according priority to components over complexes: to segments over lines, parts over wholes, and matter over sub¬ stances. For since a complex is made up of its components, it seems that the existence of the complex requires or depends on the existence of the components; and since a complex can be bro¬ ken into or built out of its components, it seems that the ex¬ istence of the components does not require or depend on the existence of the complex. But that result will be unwelcome to someone of Aristotle’s anti-reductionist and anti-atomist inclina¬ tions. The distinction between actuality and potentiality is used to show that [IC] does not give rise to that consequence. How does it do so? Ross allows that Aristotle’s meaning is hard to seize and is not very satisfactorybut it is instructive to look at the sort of analysis Ross provides, since it suggests an important general moral. Start with the question whether the existence of a whole (e.g. a line) does depend on the existence of its parts (e.g. seg¬ ments). Aristotle’s answer is that it depends on the parts exist¬ ing potentially (since when combined together into the whole, the parts exist only potentially). And so we will say that the parts are prior in potentiality to the whole (ioi9‘’8-9)—which is sup¬ posed not to be such an unwelcome consequence. Do we have rea¬ son to say that the existence of the whole depends on the parts existing actually, with the result that the parts would be prior in some stronger way too? No, quite the opposite: the existence ’’ For more on Ross’s verdict see below. For Kirwan’s discussion see C. Kirwan, Aristotle’s Metaphysics Books F, A, and E [Metaphysics F A E\ (Oxford, 1971), 155-6-

Ross, i. 318.

What Does Aristotle Mean by Priority in Substance?

215

of the whole does not depend on, but is inconsistent with, its parts existing actually (since the parts exist actually only when the whole is dissolved into them, ioi9‘‘io-i i). So the parts are not prior in actuality to the whole—and we might conclude rapidly from that that the parts are posterior in actuality to the whole (ioi9**9-io)d' The trouble with this gloss is that it does not really connect priority with [IC], [IC] explains the priority of Fs over Gs in terms of an asymmetric relation between Fs and Gs. Fs can ex¬ ist without Gs while Gs cannot exist without Fs; Gs depend on Fs w'hile Fs

not depend on Gs. However, the current sugges¬

tion is that we should, for example, understand the segments’ be¬ ing prior in potentiality to the line by reference to the claim that the segments are (only) potentially in the line. But the relation between segments’ existing potentially and a line’s existing ac¬ tually is entirely symmetrical: a line’s existing actually comes to the same thing as its segments’ existing potentially. Grant that a line’s existing actually does depend on its segments’ existing po¬ tentially (and not, as would be more troublesome, on its segments existing actually). Still, it is equally the case that the segments’ existing potentially also depends on the line’s existing actually (loip^io-i i: if the line no longer existed actually, the segments would either not exist at all or exist not merely potentially, but actually). So there is as much reason to say that the segments are posterior in potentiality to the line (since the segments’ existing potentially requires or depends on the line’s existing actually) as there is to say that the segments are prior in potentiality to the line (since the line’s existing actually requires or depends on the segments’ existing potentially). Given that there is no asymmet¬ rical dependence of either on the other, [IC] gives no reason to count the segments as either prior or posterior in potentiality to the line. The gloss offered on posteriority in actuality suffers from the '* See ibid.: ‘in considering a whole we should naturally say “the whole cannot exist without the parts, but they can exist without it, and therefore (according to 1. 3) they are prior”; but that when we reflect we find that in the whole the parts do not exist actually. The half-line does not exist till the whole has been cut in two; the matter does not exist till the concrete thing has been resolved into its components. Actually, therefore, the parts will exist only when the whole has ceased to exist; “actually they are posterior to it”. But the existence of the whole presupposes the potential existence of the parts; “in respect of potentiality they are prior to it.” ’

216

Stephen Makin

same flaw. The posteriority of Fs to Gs involves, according to [IC], an asymmetric relation: Fs depend on Gs, while Gs do not de¬ pend on Fs. But while the line’s existence does not depend on (but is inconsistent with) the segments’ existing actually, it is equally the case that the segments’ existing actually dcves not depend on (but is inconsistent with) the line’s existing. Since nothing de¬ pends asymmetrically on anything here, the verdict of [IC] should be that segments are neither posterior nor prior in actuality to the line. Of course, Ross allows that Aristotle’s meaning is unclear. But the important point is the general one: that we could give a better ex¬ planation of Aristotle’s position in this passage if we were to explain the notions of priority in potentiality and posteriority in actuality in terms of some asymmetric relation concerning items which ex¬ ist potentially and actually. For it might then be that that relation could map on to the asymmetric dependency relations which [IC] introduces. The asymmetry to which [IC] points is derived from an asymme¬ try in the way in which the notion of something’s being potentially F is explained. When that asymmetry—about which I shall say more presently—is manifest in the account of what it is to be po¬ tentially F, then and only then can we draw the distinction which Aristotle here introduces, between priority in respect of potentiality (/card Swa/xiv) and priority in respect of actuality {Kar' evreXeyeLav).^^

Consider the account given in Metaphysics Q 7, io48^37-io49‘‘i8, according to which something is potentially F so long as it is a suitable starting point for the production of something actually F by the exercise of an appropriate capacity. For example, someone is potentially healthy just so long as they can be made actually healthy by an exercise of medical skill (they do not first of all need to grow up, or lose weight . . .); something is potentially hot so long as bringing it into contact with something with the capacity to heat (e.g. a fire) results, under normal conditions (e.g. in the absence of chilling breezes), in its becoming actually hot; and fer¬ tilized spawn is potentially a frog just so long as the exercise of

The point has to be made in this qualified way because Aristotle allows for cases which are symmetrical, in which each of a pair of items is potentially the other, in that each can turn into the other. There is an explicit statement at GC 2- 7. 33T21-3; the actually hot is potentially cold and the actually cold is potentially hot.

What Does Aristotle Mean by Priority in Substance?

217

its own life capacities will, under normal conditions, result in an actual frog. Suppose there is a relation between one type of item and an¬ other, in virtue of which the first is potentially the second: for example, healing takes us from a diseased bodily condition to a state of health and in consequence that bodily condition is po¬ tentially healthy. The relation (healing) defines the first of its relata (the terminus a quo: a certain type of bodily condition) as potentially the second (the terminus ad quern: health). Now in at least many cases there will not also be another converse relation which serves symmetrically to define the second item as potentially the first. Candy asleep and Candy awake do not stand symmet¬ rically to one another: Candy asleep stands to Candy awake as something potentially perceiving to something actually perceiving. And Aristotle would not also say that Candy awake is potentially slumbering, appealing to the fact that there is a converse relation (falling asleep) which takes us from awakened Candy to sleeping Candy.” Now look at Aristotle’s remark at the end of the present pas¬ sage: The example of someone awake/someone asleep is one of those given at Metaph. & 6, io48‘‘37-io48'’6, in the course of explaining by analogy the actuality-poten¬ tiality distinction. The significance of perception for understanding the difference between being asleep and being awake is emphasized at De somno i, 453'’24-4S4’‘i9, 454*’23-455'‘3; and De somno 3 offers an account of what is involved in the processes of waking and falling asleep. Cases like these might remind one of Metaph. H 5, I044'’29-I045“6. There are two questions raised there: (1) i044*’2g-34. Is an underlying matter, e.g. water, both potentially wine and potentially vinegar, is a body both potentially healthy and potentially ill? Aristotle approaches that question armed with the distinction between being matter kuO’ e^iv Kal Kara to eiSoy and Kara orepeaiv Kai (jidopav rpv -rrapa vaiv, ‘in virtue of its positive

State and its form’, ‘in virtue of the privation of its positive state and the corruption of it contrary to its nature’. (2) io44'’34-io4s‘‘6. Is wine potentially vinegar, the living animal potentially dead? Here the issue is whether’there being some underlying matter via which F can turn into G establishes that F is potentially G. On this Aristotle wants to say that it is the matter of the animal or the wine which is kotol (j)6opav the hvvapas and vX-iq for the dead body or the vinegar. Neither (i) nor (2) is quite what is at issue in the body of this paper. Both (i) and (2) concern cases in which there is an underlying material level which can give rise to contrary conditions (e.g. water: wine or vinegar). The cases I am concerned with are those in which we can go from one item to another (e.g. water to wine, so that water is potentially wine) and back again (from wine to water: does it follow also that wine is potentially water?).

Stephen Makin

2i8

for it is when [the line, the whole, the substance] is dissolved that [the segment, the part, the matter] will exist in actuality. (loip^io—11)

There is a relation (dissolution, destruction, division) which takes us from one type of item (e.g. a line) to another (e.g. segments), and it is in virtue of that relation that the first item is potentially the second, or contains the second potentially.''^ The line contains segments potentially in that dividing will take us from an actual line to actual segments. Suppose, however, that there were a con¬ verse relation (combining) in virtue of which we could define the segments as potentially a whole line. In that case we could parody loip^io-ii: ‘for it is when the segments [the parts] are combined that the line [the whole] will exist actually’. And then there would be as much reason to say that the whole line is posterior to the segments in respect of actuality (since it is when the segments are combined that the line will exist actually) as there is to say—what Aristotle in fact says—that the segments are posterior to the whole line in respect of actuality (since it is when the line is divided that the segments exist actually).'^ Does it matter whether I say that the relation which takes us from (actual) Fs to actual Gs establishes (i) that Fs are Gs potentially (e.g. lines are segments poten¬ tially), or (2) that Fs contain Gs potentially (e.g. lines contain segments potentially)? Different instances of these locutions will sound more or less natural. For example, it sounds more natural (i) to say that flour, water, and yeast are potentially bread than to say that they contain bread potentially; whereas it sounds better (2) to say that a line contains segments potentially—though saying that a line is potentially a collection of segments sounds fine too. It may be that (i) sounds more natural when it is components which are poten¬ tially complexes (e.g. flour, water, and yeast as potentially bread: where the defining relation is constructive, taking us from parts to wholes). And (2) is more natural when it is the complex which is potentially the components (e.g. lines as poten¬ tially segments: where the defining relation is destructive, taking us from wholes to parts—as at loip^io—ii SiaXvdevros yap kot’ AreAe^etav earai). lam being selective in the examples I use from Metaph. A 11. The general point I want to make is that [1] we can make sense of the claim that Gs are prior (card Svvapuv to Fs but posterior fcar’ ivTcXe^etav only so long as [2] there is a relation that takes us from Fs to actual Gs, in virtue of which Fs are (or contain) Gs potentially; whereas there is no relation which takes us from Gs to actual Fs in virtue of which Gs are (or contain) Fs poten¬ tially. The case of matter/substance is difficult. It does not seem true that destruction of a living body takes us to items which are actually flesh/blood, or actually eyes/ hands—quite the opposite: such functional matter and functional parts when sep-

What Does Aristotle Mean by Priority in Substance?

219

We can now see that Aristotle’s discussion in ioi9‘’4-ii makes a point which is more cogent than commentators generally allow. Consider a case in which we characterize what it is for some type of thing (Fs) to be potentially Gs in terms of a relation which takes us from Fs to actual Gs, and where there is no converse actualizing relation which takes us from actual Gs to actual Fs and which defines Gs as potentially Fs; a case in which Fs are potentially Gs, while Gs are not potentially Fs. According to [IC], priority involves an asymmetric relation of dependence: it is possible for there to be one type of item without 'another, but not vice versa. But when we speak of there being items of a certain type we can mean either there actually being or there potentially being such items. And in the sort of case we are now considering the application of that distinction—between being actually and being potentially— will line up pretty w'ell with the verdicts Aristotle reports. First, an explication of priority in respect of actuality (/car’ irreXeyeiav) in line with [IC]:*'’ [AJ The whole line is prior in respect of actuality to the segments in that (i) it is possible for there actually to be whole lines without there actually being segments, but arated from the living organism are then only potentially what they previously were actually (cf. Metaph. Z 16, i04o‘’5-8, on the separated parts of animals). It may be that Aristotle has in mind a lower type of matter, e.g. elemental mat¬ ter. Kirwan, Metaphysics F A E 155-6, comments on the tension between (a) the claim that matter survives (indeed, is actualized through) the dissolution of a sub¬ stance, on which Aristotle’s discussion in ioi9‘‘4-ii rests, and (b) the claim ap¬ pealed to in Metaph. Z lo to show that material parts are posterior to the wholes they compose, namely that when separated they are only homonymously eyes, hands, etc. The whole/part example is difficult to work with because it is entirely inde¬ terminate what sort of whole and part we have in mind. Aristotle emphasizes at Metaph. Z 10 that the question'whether wholes are or are not prior to their parts— precisely the question being considered here in Metaph. A 11—has no straight¬ forward answer. There are different notions of a part (Metaph. Z lo, io34'’32 rj TToXAdycos

XeycTCLL

to

—and see Metaph. A 25). ^Ve should answer the ques¬

tion whether something is prior to its parts by saying that it is not a simple mat¬ ter (io36‘‘i6 OTt ovx aTrAcDj); the question must not be answered simply (io36“23 a-rrXdis 8’ ov (jiareov). The best strategy here is simply to specify types of part and

whole which do illustrate Aristotle’s point: hence my concentration on lines and segments. In fact Aristotle says at loig^g-io that the segment is posterior kot' evreXeyeMv to the line; but I take it that that is equivalent to saying that the line is prior kot evreXeyeiav to the segment.

220

Stephen Makin (ii) it is not possible for there actually to be segments with¬ out there actually being whole lines.

[A] points to the sort of asymrnetric relation of idependence with which [IC] is concerned:-(Aii) there actually being segments de¬ pends on there actually being whole lines, while (Ai) there actu¬ ally being whole lines does not depend on there actually being segments. And each clause of [A] is plausible. On the one hand, (Ai) since actual segments result from the dissolution of actual lines (which are thereby defined as potentially segments), it is possible for there to be actual lines without there ever being ac¬ tual segments—so long as no lines are ever divided. On the other hand, (Aii) it would not be possible for there sometimes to be ac¬ tual segments without there sometimes being actual lines, since actual segments are just the result of an actualizing operation (di¬ viding) on actual lines (which are thereby defined as potentially segments). Second, the explication of priority in respect of potentiality (/card Svvajjiiv):

[P] The segments are prior in respect of potentiality to the whole line in that (i) it is possible for there potentially to be segments without there potentially being whole lines, but (ii) it is not possible for there potentially to be whole lines without there potentially being segments. [P] explains priority in respect of potentiality in terms of [IC]. And, in the sort of case we are concerned with, the first clause (Pi) is plausible. For there potentially to be segments is for there to be actual lines which are not divided: and so if there are no divi¬ sions then there are potentially segments while there are only ever actually (and so not potentially) whole lines. The second clause (Pii) is harder to assess. We define actual segments as the result of an operation on actual lines (which are thereby potentially seg¬ ments), whereas we do not define actual lines as the result of an operation on actual segments (which would thereby be potentially whole lines). But as a result we have not given much content to what is meant by there potentially being whole lines. To repeat, we had better not say that there are potentially whole lines if there are actual segments which could be combined—since in that case

What Does Aristotle Mean by Priority in Substance?

221

we would lose the asymmetry between potentiality and actuality which the account of priority in terms of [IC] requires. I might say that (Pii) is, as it were, trivially plausible: since we have little idea what it would be for there to be potentially whole lines, then it certainly is not possible for any outcome to obtain which in¬ volves there being potentially whole lines. But that is not a very attractive option. And it is preferable to allow that while I have not made Aristotle’s discussion in loiq^’q-ii completely clear, I have got more from it than have other commentators—in par¬ ticular, I have given an account of priority in respect of actuality and in respect of potentiality which accommodates the point that they can each be understood in terms of, and are variant forms of, [IC]. However, my final comment on the distinction between priority in respect of actuality and in respect of potentiality is deflationary and may seem disappointing. For it is unlikely that the distinction drawn at loig’^q-i i between priority in respect of actuality and in respect of potentiality will be of much use as regards the discussion in Metaphysics 0 8 of priority in substance. We are told in that chapter that something actually F (e.g. an adult human) is prior in substance to something potentially F (e.g. a child). But it is not, for example, a very promising way of elucidating that claim to ask—as [A] above would recommend—whether it is or is not possible for there actually to be items which are actually human (adults) with¬ out there actually being items which are potentially human (chil¬ dren).'’ Far from that question clarifying things, it barely seems to make sense. It is far more promising to turn to another distinc¬ tion Aristotle draws in the present chapter at 1019''12-14, where I think we will find a line of thought which will help us as regards Metaphysics

0 8.'*

” It is important to see that it is this—strange—question which, according to [A], we should be asking. For if we ask just whether it is possible for there to be (unqualifiedly) items which are actually human without there being (unqualifiedly) items which are potentially human, then we are back to where we started: namely, an unqualified reading of the examples in Metaph.

0

8 in terms of [IC], with the

consequent worry that it is then hard to make the examples from Metaph.

0

8 seem

plausible. The sentence ioig“ii-i2, which links (a) ioi9'‘4-ii and (b) loig’iz-iq, is puzzling: ‘In a sense, therefore, all things that are called prior and posterior are so called according to this fourth sense’ {rpoTTOv Sij riva irdvra rd nporepov Kal varepov \ey6p.eva Kara ravra Ae-yerai). The majority of commentators take Kara ravra Xeyerai

to refer to the formulaic statement of [IC] at loig^a-q. That is the reading sug-

222

Stephen Makin

(c) ioig'‘i2-i4 Aristotle’s purpose in these remarks is not immediately appar¬ ent:

^



For some things can exist without; others in respect of generation [Kara yeveoLv], e.g. the whole without the parts, and others in respect of dissolu¬

tion [/card (j)9opdv], e.g. the part without the whole. And the same is true in all other cases.

I suggest the following. According to [IC], if I wish to discover whether Fs are prior to Gs I need to enquire whether it is pos¬ sible for there to be Fs (actually, potentially?) without there being Gs (actually, potentially?), or whether the reverse is possible. But possibility is itself a complex notion, as Aristotle recognized. Re¬ call the account of possibility provided at Prior Analytics i. 13, 32'‘i8-2o:

I use the terms ‘to be possible’ and ‘the possible’ of that which is not necessary but, being assumed, results in nothing impossible.

I establish that something is possible by citing a coherent way in which the putative possibility can come to be actual. Consider Candy, a trained builder, presently inactive but surrounded by bricks and wood, in daylight, healthy and sober, etc. Is it possible that she build? Well, suppose she takes the bricks and wood, which she can see, and lays them on top of one another . ..—then she would be building. And there is nothing impossible in that occurring. So, yes, it is possible that Candy build. The important point to note here is that I do not establish a possibility simply by describing a gested by the expanded translation ‘according to this fourth sense’, along with the numbering provided, in the Revised Oxford Translation. The decision to translate Kara, ravra Aeyerai in this way dovetails with Ross’s editorial decision to punctu¬ ate as parenthetical ioi9‘‘4-ii, which intervenes between the statement of [IC] and Kara ravra. According to this reading, Aristotle’s point is that all the senses of priority so far distinguished within Metaph. A ii (ioi8'’io-30 nearer to some dpx’?; ioi8'’30-7 priority rfj yvaiaei; ioi8'’37-ioi9’‘2 the rradr) of things which are prior are themselves prior—though the structure of the chapter is not clear) are to be understood in terms of [IC]. In contrast, Alexander {In Metaph. 387. 32-388 7 Hayduck) takes the heterodox position that Kara, ravra refers to the immedi¬ ately preceding material on priority xar’ evreXeyeiav and Kara Svvafiiv. I have not found an entirely satisfactory way of understanding the sentence. While my in¬ clination is to side with Alexander against the majority, Ross’s objection to this reading—that it is utterly unclear how the other notions of priority and posterior¬ ity are to be understood in terms of priority Kar’ ivreXey^iav and Kara Svvajuip—is strong.

What Does Aristotle Mean by Priority in Substance?

223

state of affairs which would be the actual obtaining of the putative possibility, if there is no route by which that state of affairs can come about. Suppose Candy is unskilled and blind. Is it possible that she build? There is certainly nothing impossible in this state of affairs: Candy being trained and able-bodied, and manipulating bricks and wood appropriately. But given Candy’s state of health and her ignorance, there is no way for that state of affairs to come about. And so it is not possible that Candy build. The upshot is that we establish that something is possible by showing that there is a way or a route by which that putative possibility can come to be actual. It follows from this that we can characterize a possibility by spec¬ ifying the sort of route by which the outcome is to become actual, the existence of which route establishes the possibility at issue. Suppose I am in Edinburgh at 9 a.m. Is it possible for me to be in London at noon? If I take an aeroplane, then it is consistent to suppose that the outcome—my being in London at noon—should become actual. If I take a train then that supposition is not consis¬ tent. We could say that it is possible in respect of (/card) flying for me to be in London at noon, but not possible in respect of (Kara) surface travel. To put this more naturally: it is possible, travelling by air, to be in London at noon, but not possible by rail. It further follows, then, that there need be no unique answer to the ques¬ tion ‘Is such and such possible or not?’ It is possible, travelling by air, to be in London at noon, it is not possible by rail. The bare question ‘But is it possible or not?’ does not have any determinate answer. When Aristotle distinguishes at loi 9*12-14between what is pos¬ sible in respect of generation

(kuto.

yeveoLv) and what is possible in

respect of destruction (Kara (f>dopdv), we should understand him along these lines. Sometjiing is possible in respect of generation if it is the outcome of a process of generation, possible in respect of destruction if it is the outcome of destruction—here we have two different ways of qualifying possibility. And so there need be no unique answer to the question which [IC] poses: is it possible for there to be Fs without Gs or not? If the outcome of one type of process is Fs and is not Gs, then it is possible in that respect for there to be Fs without Gs, and [IC] will adjudge Fs prior to Gs in that respect. But if the outcome of another type of process

224

Stephen Makin

is Gs and is not Fs, then [IC] will count Gs prior to Fs in that respect. The point can be illustrated by means of the example Aristotle provides, of parts and whole. Once again the example is schematic: it is not clear what sort of parts and wholes he has in mind.*** But we can take an example he would find favourable: the four elements (earth, air, fire, water) as parts or components of living organisms. Is it possible for there to be cats without the elements, or for there to be the elements without cats, or neither, or both? The upshot of generation (i.e. feline reproduction and growth) is cats, and not the elements, and so it is possible in respect of generation for there to be cats without the elements; while it is not possible in respect of gen¬ eration for there to be elements without cats, since a process which results in elements and not in cats is not an instance of reproduction and growth—and so according to [IC] cats (wholes) will be prior in respect of generation to elements (parts). On the other hand, the upshot of destruction (i.e. death and decay) is some elemental stuff and not cats, and so it is possible in respect of destruction for there to be elements without cats; while it is not possible in respect of destruction for there to be cats without elements, since a process which results in cats and not in elements is precisely not an instance of death and decay—and so, in line with [IC], we should adjudge elements (parts) as prior in respect of destruction to cats (wholes).^® The question to consider now is how this material will be helpful in understanding the discussion at Metaphysics 0 8 of priority in substance in terms of [IC].^' See the references at n. 15. This example rests on Aristotle’s views on the composition of animals: see PA 2.

I,

646T3-24, on the relation of the simple elements to the homoiomerous parts

(e.g. flesh and bone), and of those tissues to structured parts (e.g. faces and hands); and GC i. 10 for detail on

(especially 327*’22—31, constituents are present only

potentially and not actually in a mixture). Perhaps a more homely example would be the relation of milk to curds and whey. Bovine lactation gives rise to milk, and not curds and whey, and so it is possible Kara, yeveaiv for there to be milk without curds and whey. But on the other hand, allowing milk to separate out produces curds and whey, and not milk, and so it is possible Kara (j>dopdv for there to be curds and whey without milk. A final comment on Metaph. A ii. lam adopting a deflationary reading of ioi9“i4

and the same is true in all other cases’ {ofxotojs

Kal raAAa), accord¬

ing to which this remark is a generalization over the previous logical point, that there can be Fs and Gs such that it is possible Kara ydv^oiv for there to be Fs without Gs, while possible /card (f>6opdv for there to be Gs without Fs. Perhaps

What Does Aristotle Mean by Priority in Substance?

225

3. Back to Metaphysics 0 8: the first set of examples at 1050^5-7 Let me rehearse the problem driving this paper. The discussion at Metaphysics

0 8 of priority in substance starts like this:

But indeed (actuality is prior) in substance too, first because things poste¬ rior in coming-to-be are prior in form and in substance (e.g. adult to boy and man to seed; for the one already has the form, the other does not). (1050V7)

If we think of priority in substance in terms of [IC], these examples are difficult: it is not obvious in what sense there can be adults with¬ out children whereas there cannot be children without adults. But there is reason to try to understand priority in substance in terms of [IC]. What should we do? I think the distinction we have found in Metaphysics Zi 11, ioi9‘“i2-i4, between what is possible in respect of generation and in respect of destruction, makes it easier to see these initial examples in the light of [IC]: in eflfect, it helps us to answer the question posed by Charlotte Witt in the passage quoted earlier. Now it is clear that Aristotle himself has a teleological view of natural generation and development, and sees the teleological structure of generation as of great significance for his claims about priority in substance. After giving the examples above, he con¬ tinues: . . . and because everything that comes to be proceeds to a source, i.e. an end (for that for the sake of which is a source, and the coming-to-be is for the sake of the end), and the actuality is an end and the potentiality is acquired for the sake of this. (io5o‘‘7-io)^^

What is the connection between, on the one hand, taking a teleolog¬ ical view of generation and, on the other hand, using the distinction drawn from Metaphysics A 11 between what is possible in respect raXXa are just the other cases presented earlier in discussion of the Kar’ eVreAfyetar/ /card Swa/iir distinction, viz. ioig‘'8-g: half-line/whole line, part/whole, matter/ substance. For other statements of the view that the final item in a process of genera¬ tion is prior in substance (or, as it is sometimes put, in nature) see also Phys. 8. 7, 26P13-14.; 8. g, 265‘22-y; GA 2. 6, 742‘‘ig-22; Metaph. A 8, 989“i5-i8; Rhet. 2. 19,

226

Stephen Makin

of generation and in respect of destruction in order to accommo¬ date the examples at 1050^5-7? Aristotle’s view is that generation proceeds from a potential F to an actual F; the actual F is the goal or end of the process; the process is for the sake of that actual F; and the upshot of generation i^ that something acquires the rele¬ vant form. Now, given this teleological view, it follows that there is a privileged stage in a process of generation, a stage to which the process will normally run so long as it is not interfered with or hindered. That is to say that a process of generation is not just a temporally extended series of later stages following earlier stages, like a pool of water under a leaking pipe getting bigger and big¬ ger. We can distinguish between processes of generation which are completed or successful (viz. which reach that privileged stage) and those which are interrupted or interfered with (viz. those which do not reach that stage). In contrast, it is because there is no privi¬ leged stage as a pool of water increases under a leaking pipe that it makes little sense to say that the oozing has been completed or interrupted. And it is because it is sensible to talk of processes of generation being interrupted or hindered that the distinction be¬ tween what is possible in respect of generation and in respect of destruction can elucidate an account of priority in substance based on the independence condition [IC]. Further, the fact that it is such notions as interference and hin¬ drance which are crucial to the account of priority in substance enables us to understand the rest of Aristotle’s discussion in Meta¬ physics & 8, io5o®4-*’6, sympathetically.^^ He starts his discussion with an instance of the potential-actual schema—immature and mature specimens of a kind as potentially and actually F respec¬ tively (1050*4—9)—for which the teleological perspective is the most plausible, and he bases his conclusion that actuality is prior in sub¬ stance to potentiality on an appeal to that teleological perspective (1050*6-9). He then tries to extend the teleological perspective from that initial case to other instances of the potential-actual schema; to present something’s possessing a capacity as being for the sake of its exercising the capacity (1050*10-14); pre-existent matter as being for the sake of the form it comes to possess (1050*15-23); and " A detailed defence of this account of the structure of Metaphysics & 8, and of the more general treatment of actuality and potentiality in Metaphysics 0, would not be appropriate in this paper. I can only hope that what I say does not seem too implausible.

What Does Aristotle Mean by Priority in Substance?

227

a capacity itself (as opposed to something’s possessing it) as being for the sake of its exercise (1050*23-1050^2). Since the teleological perspective justifies the conclusion that actuality is prior to poten¬ tiality in the initial case (e.g. adult humans as prior to children), the hope presumably is that it will justify the same conclusion in the later cases too (e.g. an animal’s seeing as prior to its being able to see, a substance which is actually F as prior to the pre-existent matter which is potentially F, seeing or building as prior to vision or building skill). But the teleological perspective on these cases is more contentious. So we are in a better dialectical position once we have identified that feature of the teleological view of generation in virtue of which the priority of adult over immature specimens can be accommodated by [IC]. For we can then better appreci¬ ate, for example, Aristotle’s claim that an exercise of a capacity is prior in substance to the capacity, and see the relevance to that claim of earlier material in Metaphysics Q. It is to the point, when thinking about the relation of an exercise to a capacity, to contrast a normal or unhindered outcome with what results from interfer¬ ence. And the fact that that contrast is to the point explains why Aristotle goes through the discussion of rational and non-rational, and two-way and one-way, capacities in chapters 2 and 5 of Meta¬ physics 0. I will go through the first set of examples Aristotle presents (1050*5-6: adult/child, human/fertilized egg) in some detail, ex¬ plaining why the account of [IC] derived from Metaphysics A 11 1019*12-14, should lead us to conclude that, for example, adults are prior in substance to children, and summarizing a general ac¬ count of what it is for Fs to be prior in substance to Gs. Then in the following section I will say something more briefly about Aristotle’s treatment of other cases in the remainder of 1050*10io5o‘’2. The two thoughts I am trying to render jointly plausible are first (1050*5-6) that adults are prior in substance to children, and second {Metaphysics A ii, 1019*2-4) that according to [IC] that claim should somehow be understood along the lines that it is possible for there to be adults without children but not vice versa. The upshot of Metaphysics A ii, 1019*12-14, is that the claim about possibility can be taken in different ways: in particular as either

Stephen Makin

228

(i) it is possible in respect of generation for there to be adults without children, while it is not possible in respect of gener¬ ation for there to be children without adults, (ii) it is possible in respect of destruction for there to be adults without children, while it is not possible in respect of de¬ struction for there to be children without adults. Claim (i) is in fact extremely plausible. It amounts to the asser¬ tion that (human) generation results in adults and not in children, while a process which resulted in children and not in adults would not be a process of generation—on the contrary, it would be an interrupted or incomplete generative process, and interrupting a generative process is a kind of destruction. So (i) provides a way of rendering the two thoughts mentioned jointly plausible. It gives a gloss on the claim that adults are prior in substance to children which both renders it plausible and interprets it along the lines of [IC]—the claim is plausible in that it is possible (in respect of generation) for there to be adults without children while not in the same way (in respect of generation) possible for there to be children without adults.^'* The significant feature of natural generation which renders (i) plausible is that there is a privileged stage in generation to which the process normally runs, unless interfered with. For if that were not the case, then the second conjunct of (i)—that it is not possible in respect of generation for there to be children without adults— would not be plausible. That is because it would then be implau¬ sible to insist that any process resulting in children rather than in adults would be an incomplete or interrupted generation, and thus a process not of generation but (in effect) of destruction. For if there were no privileged stage in generation, it would make little sense to talk of a process of generation being incomplete or in¬ terrupted. Generation would, in that case, be like water leakage. Water leaking out of a pipe often forms a puddle, but sometimes it just makes the ground damp. But there is no reason to think that in the latter case abnormal conditions have interrupted or interIn contrast it would not be a good idea to understand the claim about possibility as (ii), since (ii) is extremely implausible. Contrary to what (ii) asserts, it is possible in respect of destruction for there to be children without adults, since the only way to get children without getting adults is to go in for some killing, some destruction of human life, some infanticide.

What Does Aristotle Mean by Priority in Substance?

229

fered with the process of leakage, nor is there reason to deny that it is a process of leakage. In contrast, we will say that when gen¬ eration ends with a child it has been interfered with by abnormal conditions. We can now move to a general statement of my account of pri¬ ority in substance. Fs are prior in substance to Gs so long as it is possible for there to be Fs without Gs, but not vice versa: that constitutes the connection between priority in substance and [IC]. What establishes the possibility of there being Fs without Gs is that there is some process which in normal conditions re¬ sults in Fs and not in Gs—a process of which the unhindered outcome is Fs and not Gs; whereas in contrast the way to get Gs and not Fs is to interfere with, interrupt, or hinder that pro¬ cess. What that process is will be different in different cases. For the examples at 1050^*5—7 the process is generation. But gener¬ ation will not be relevant to the cases which turn up later in Metaphysics Q 8, since it is not clear that generation has any¬ thing to do, for example, with the relation of a capacity to its exercise. Further, it is not just any process that is appropriate to ground a possibility of there being Fs without Gs in virtue of which Fs w'ould be prior in substance to Gs. It would be disastrous to sug¬ gest that Fs are prior in substance to Gs just so long as there is some route or other which results in Fs and not in Gs, even though the existence of such a route will indeed establish that it is possible (in respect of that route) that there be Fs with¬ out Gs, while not being possible (in respect of that route) that there be Gs without Fs. For just as there is a route which re¬ sults in adults and not in children (namely generation), so too there certainly is a route which results in children and not in adults—infanticide or destruction. But priority in substance is an asymmetrical relation, and so it should not be the case that children are prior in substance to adults (in virtue of what is possible in respect of destruction), just as adults are prior in sub¬ stance to children (in virtue of what is possible in respect of gener¬ ation). The account I am offering secures the asymmetry required. The route by which one gets children and not adults—killing or destruction—is an interruption of or interference with some other process (generation), while that other process is not an interfer¬ ence with it. Juvenile demise is an interruption of an adult life-

Stephen Makin

230

span, while an adult lifespan is not an interference with juvenile demise.^^

4. The other examples at Meta¬ physics 0 8, io5o'‘io-i05o'’2 According to the structure I propose for this part of Metaphysics

0 8, Aristotle considers three further types of case in which the pri¬ ority in substance of actuality over potentiality is exhibited. First (1050^10—14), the relation of capacity possession to capacity ex¬ ercise; second (1050“*!5-23), the relation of pre-existent matter to substance; third (io5o‘‘23-io5o^2), the relation of a capacity to its exercise. Consider the first type of case. Aristotle’s view that the exercise of a capacity is prior in substance to its possession is sup¬ ported by the following plausible claim: the normal or unimpeded outcome is that something not merely possesses, but also exer¬ cises, its (natural) capacities—whereas, in contrast, if a capacity is only possessed and not exercised, then that is the result of interfer¬ ence. Aristotle has prepared the ground for this point earlier in Meta¬ physics 0 in his discussion of capacities and their exercise in chap¬ ters 2 and 5. In particular, he argued in chapter 5 for the following principle concerning the exercise of capacities (stated at 1048^5-7; argued for on the basis of 1047*^35-1048*2, 1048*2-5, 8-10): [EP] As regards one-way capacities: necessarily, if agent and pa¬ tient are in the right condition and related in the right way, action results. According to [EP], if an agent possesses a (one-way, non-rational) capacity to and there is the right sort of patient present and con¬ ditions are normal, then the agent acts, the capacity is exercised, ^-ing occurs. If nothing interferes with or hinders the agent, then the normal outcome is that the agent not merely possesses but also exercises the relevant capacity. In contrast, if the agent merely pos¬ sesses, but does not exercise, the capacity, then some interfering “ The same asymmetry shows up if we consider the disintegration which follows on an adult lifespan. Death—at any point—is an interruption of adult life; adult life is not an interference with the process of dying. So while it is true that dying results in corpses and not in humans, it does not follow—as it should not follow—that corpses are to be accounted prior in substance to adults.

What Does Aristotle Mean by Priority in Substance?

231

factor is preventing its exercise. For example, according to [EP], if a fire is hot, but is not currently heating, then something is pre¬ venting it heating: perhaps a fan is blowing, or there are no cold objects in the vicinity. Now the general moral which emerged from discussion of Metaphysics Zl 11 and [IC] was that Fs are prior in substance to Gs so long as there is some process of which the nor¬ mal outcome is Fs and not Gs, while one can obtain Gs without Fs only by interrupting or interfering with that process. So, since what obtains in normal conditions is that a capacity is not merely pos¬ sessed but also exercised—since a capacity which is possessed but not exercised is being interfered with—it follows that the exercise of a capacity is prior in substance to its possession. That conclu¬ sion goes some way towards filling out Aristotle’s general claim in Metaphysics 0 8 that actuality is prior in substance to potentiality. And support accrues to my proposed gloss on the argument for that conclusion from the fact that it renders the conclusion plaus¬ ible. The second type of case Aristotle considers concerns the relation of (pre-existent) matter to substance (1050T 5-23). As it stands, the crucial opening sentence is highly compressed: Again the matter is potentially because it may go to the form [eVt rj vXrj ean Swa/xet OTL eXdot av eis to etSoj]; and whenever it is actually then it is in the form. (io5o''i5-i6)

It would be less neutral but more informative to expand on the translation in the way suggested by Burnyeat, Notes on Eta and Theta, and render the optative on eXdoi dv els to etSos as ‘because it would normally go to the form unless interfered with’.^*’ Aris¬ totle’s point is to be understood in the light of the discussion in Metaphysics 0 7, 1048*^37-1049^18, of the conditions under which something is potentially F. That chapter rejects a permissive ac¬ count according to which something is potentially F just so long as it could end up being F.^’ The account offered is rather that some¬ thing is potentially F so long as it is a suitable starting point for “ See Notes, 142: ‘eXBoi av: not “may come” (Ross, Reale), nor “might come” (Apostle). For the optative as “would normally except in odd cases” . . . [then follow some cross-references] . . . There is a definite form which the vXrj is programmed to acquire, so that one can say that a potential F is what it is in a sense much stronger than that in which one says of the table that it can (might) be green. The stronger sense is what is required for the argument to show that actuality is prior ovala, not just Xoyw.’ ” See Metaph. ©7, 1049T-5.

232

Stephen Makin

the production of something actually F by means of some appro¬ priate single capacity. Quite what the appropriate capacity will be is different in different cases: we explain what it is to be potentially healthy by reference to medical skill, what it is*to be potentially hot by reference to the capacity to heat. In general the account will need to be supplemented by an account of capacity identity, which will in its turn connect what it is for something to be a single capacity with its normally resulting in the manifestation of a single form. Roughly: it is because (human) health is a single form that medical skill is a single capacity; and it is because med¬ ical skill is a single capacity and this body can be made healthy simply by an exercise of medical skill that it is potentially healthy. But as we saw above, according to Metaphysics 05a capacity will be exercised in normal conditions (in the case of a rational capac¬ ity we have to add: so long as the agent wants to exercise it in a certain way). So we can conclude that something which is poten¬ tially F will, when acted on appropriately in normal conditions, result in something actually F. For example, if these materials are potentially a house, then an exercise of building skill upon them under normal conditions will result in something which is actually (and not potentially) a house. In contrast, if I interrupt or interfere with a builder exercising her skill on the materials, I will obtain something which is still potentially (and not actually) a house—e.g. half-built walls and piles of bricks and timber. But according to [IC], for Fs to be prior in substance to Gs is for there to be some process of which the normal outcome is Fs and not Gs, while Gs without Fs result only from interference with that process. So we can now draw the conclusion we want about the actually F items which result from an exercise of the appropriate capacity under normal conditions on what is potentially F: that they are prior in substance, according to [IC], to the matter which was poten¬ tially F. Finally, Aristotle returns at 1050^3-1050*^2 to the case of a ca¬ pacity and its exercise. The focus here is different from that of the earlier io5o'‘io-i4. The concern there was with the relation of capacity possession to capacity exercise: animals have sight in order that they may see (1050T0-11). But here Aristotle is inter¬ ested in the relation of the capacity itself to its exercise—which is unsurprising, since that relation was explicitly identified as an in¬ stance of the potential-actual schema in Metaphysics

0 6 (io48'’8-

What Does Aristotle Mean by Priority in Substance?

233

9). Further, Aristotle’s discussion here is complicated by the fact that there are two significantly different types of case to accom¬ modate. In the one, the exercise of a capacity results in a product distinct from the exercise (as the exercise of a builder’s skill re¬ sults in houses); in the other, no distinct product results (as the exercise of vision results in nothing besides seeing). Now Aristotle himself is working towards his overall conclusion—that actuality is prior in substance to potentiality—on the basis of a teleological perspective on the cases he is considering. And that means that he wants to show that there is in both cases a teleological relation between the capacity and its exercise, sufficient to justify the con¬ clusion that the exercise (the actuality) is prior in substance to the capacity (the potentiality). But the fact that there are these two different types of case raises a problem for him. For in the former type, where a product results, there are two candidate items which could stand in a teleological relation to the capacity (e.g. building skill): on the one hand the exercise (e.g. the act of building), on the other hand the product (e.g. the house). So he has to argue that the relation of the capacity to the exercise (as opposed to the product) is sufficiently teleological to justify the conclusion that in this case too the exercise (the actuality) is prior in substance to the capacity (the potentiality). This is the task he announces at 1050*23-8: And since in some cases it is the exercise that is final (for example see¬ ing is the exercise of sight, and no product [e'pyov] in addition to this comes to be from sight), but from others there does come-to-be some¬ thing (for example, from the building craft a house in addition to the act of building), it is nevertheless no less (a fact that the exercise is) in the one case the end, and in the other more the end than the potential¬ ity is.

Aristotle’s point is that, even in cases in which there is a product distinct from the exercise of a capacity, still the exercise is more of an end than the capacity is. The material which follows (1050*28•^2) is intended to support that claim, by appealing to facts about the relation between the exercise of a capacity and the product (if any) in which the exercise results—in particular, the location of the exercise in the product. The argument there has its complexi¬ ties. But it is a further advantage of the approach I am advocating that we can, to some extent, finesse these complexities. For we can

234

Stephen Makin

appeal again to the claim established in Metaphysics & 5, that a capacity to (/> results in ^-ing in normal conditions, in the absence of hindrance or interference. That claim seems equally appeal¬ ing whether we are considering capacities such‘as vision, which do not result in a distinct product, or capacities such as building skill, which do. When the exercise of someone’s visual capacities does not result in sight (but in blurs and shadows), that will be due to interference or prevention: the lights are off or dim. And similarly if the exercise of someone’s building skill does not re¬ sult in acts of building that produce houses (but in attempts at building, resulting in half-finished walls and piles of bricks)—that too will be due to prevention or interference: perhaps the winds are blowing strongly or the ground is too soft. That last point re¬ tains its plausibility even if it is acknowledged that the product (the house) has a better claim to be the end at which the capacity (building skill) is directed than does the exercise (the act of build¬ ing). And the point is sufficient to give us Aristotle’s conclusion. If a capacity, whether vision or building skill, is manifested not in the activity which is its proper exercise—not in seeing or in build¬ ing, but in half-glimpsing or attempted building—then that will be the result of interference or preventative factors. The outcome in normal conditions will be the proper exercise: seeing or build¬ ing. Since the capacity is properly exercised in normal conditions, while it is only through interference that we will get the capacity without its proper exercise, we can once again conclude that ac¬ cording to [IC], the exercise is prior in substance to the capacity itself.

5. Summing up The best way to understand what Aristotle means by priority in substance is this. Fs are prior in substance to Gs so long as there is some process of which the normal outcome is Fs and not Gs, whereas one can obtain Gs without obtaining Fs only by interrupt¬ ing or interfering with that process. That account has a number of advantages. It is in line with [IC]—the idea that the priority in sub¬ stance of Fs over Gs is something to do with Fs being ontologically independent of Gs—which is offered in Metaphysics d 11 as a gloss on priority in substance. But by making use of a distinction derived

What Does Aristotle Mean by Priority in Substance?

235

from Metaphysics Zl ii, loig'^ia—14, between different respects in w'hich something can be the case (e.g. in respect of generation and in respect of destruction), it allows us to accommodate some dif¬ ficult claims which Aristotle makes in Metaphysics & 8, e.g. that an adult is prior in substance to a child. Further, the important feature of those examples, in virtue of which they can be accom¬ modated by [IC] as I understand it, is that the relevant process (in this case, generation) has a privileged stage to which it will run in normal conditions, so long as it is not interfered with. It then follows that we can respond sympathetically to other claims Aris¬ totle makes in Metaphysics & 8: that the exercise of a capacity is prior in substance both to the possession of the capacity and to the capacity itself, and that the same priority relation holds for sub¬ stance and pre-existent matter. And finally, as a result of this we obtain a clearer view' of the internal structure of Metaphysics 0 8, io5o‘*4—1050^6; we see how' that discussion is dependent on earlier material in Metaphysics

0; and Aristotle’s general conclusion—that

actuality is prior in substance to potentiality (io5o'’3-4)—gains in plausibility. The University of Sheffield

APPENDIX Alternative Accounts of Metaphysics

0 8 on Priority in Substance It may be helpful to compare my recommended account of [IC] and Metaph.

0

8 with that of other commentators. Many commentators on the

discussion of priority in substance in Metaph.

0

8 adopt a ‘splitting strat¬

egy’, claiming that the first part of Aristotle’s discussion (io5o‘‘4-io5o‘’6) should not be understood in terms of [IC], whereas [IC] is relevant to the discussion of the Kvptwrepwg type of priority in substance in the second part (io50*’6—i05i“2; there is a clear reference to [IC] at io5o'’i9 Kalroi ravra TTpcora- el yap ravra /xi) tjv, ovdev av rjv ‘and yet these things are pri¬

mary; for if these were not, nothing would be’). For further detail see the following: (i) W. D. Ross, Aristotle’s Metaphysics.- A Revised Text with Commen¬ tary [Ross] (2 vols.; Oxford, 1924). In connection with the first part of the discussion at 9. 8, io5o‘‘4-‘’2, ‘prior in substance’ is glossed as ‘more

236

Stephen Makin

real or more substantial’ (ii. 262), and Ross does not cross-reference Metaph. A ii. On the second part of the discussion (io5o‘’6-io5i’‘2), he does, however, refer to zl ii (Kara ravra Xiyerai at loig^ii—12 re¬ ferring back to [IC] at ioi9’“2-4) in explaining why ithe second discus¬ sion concerns priority in substance ‘more properly’, KvpicjTepuis; see ii. 265: ‘Now the eternal can exist without the perishable and the perishable cannot exist without the eternal, and though Aristotle does not explic¬ itly put the matter in this way (except incidentally in 1. 19), it seems that this is the “stricter” sense of “prior in substance” which he has in mind.’ (2) R. M. Dancy, ‘Aristotle and the Priority of Actuality’ [‘Actual¬ ity’], in S. Knuuttila (ed.). Reforging the Great Chain of Being (Dor¬ drecht, 1980), 73—115, esp. 88—9, takes much the same line. He distin¬ guishes natural priority {Kara fvaw, fvaei) and substantial priority (rfi oiiala, Kar’ oiioLav). Natural priority is the type of priority established by the Principle of the Priority of Ends {Metaph. @ 8, io5o‘‘4—5, cf. e.g. Phys. 8. 7, 26i‘‘i3-i4); Dancy makes that the topic of the first part of the discussion (1050^4—*^6, cf. the section-heading at p. 89). Substantial priority is understood in terms of [IC] (ibid. 89: ‘one thing is some¬ thing on which the other depends for its existence’); that is the topic of the second part of the discussion (io5o'’6—24: note the section-heading at p. 95). Dancy allows (p. 89) that ‘we do not have Aristotle’s per¬ mission for this [distinction]’; but also comments that ‘it is not at all clear that substantial priority and natural priority are the same prior¬ ity. It is not even absolutely clear that Aristotle himself thought they were.’ (3) At M. Burnyeat et al.. Notes on Eta and Theta [Notes], recorded by Myles Burnyeat and others (Sub-Faculty of Philosophy, Oxford, 1984), 133—4, Owen’s introduction to the discussion of priority in 0 8 starts with an analysis of Metaph. A i\. There is no explicit reference to [IC] when he turns to the discussion of priority ovatai in io5o'‘*4—*^6. Owen suggests ten¬ tatively that the opening, io5o‘‘4-7 (dAAd pi.rjv ...

to

5’ ov), which includes

the examples on which much of my discussion focuses, may be no more than a conjunction of priority Adyo; and priority xportp. (4) Notes, 144, comments on 1050’’2-4, where Aristotle, summing up the first part of his discussion, claims that his account makes it clear not only that {a) rj ovaia

Kai to

eiSo? ivepyeia iariv, ‘the substance and

the form are actuality’, but also {b) nporepov rfj ovaia evepyeia 8vvdp.ea)s, ‘actuality is prior in substance to potentiality’. Notes states: ‘We could not find a satisfactory answer to the question why, in advance of the KvpLwrepcvs argument of [i05o'^]6 ff, it is already clear that actuality is prior rfj ovaia to potentiality. No doubt, as Ross says, what makes it clear is the whole section [io5o]''4-‘’2. The trouble is that the emphasis of the

What Does Aristotle Mean by Priority in Substance?

237

section is on ways in which form as riXos is prior in explanation, not prior TT7 ovaia, in the sense at issue [in io5o’’]6 fF, viz that X must ex¬ ist if Y is to exist but not vice versa ([io5o]'^i9, Ross ad loc).’ There is tacit endorsement here of the splitting strategy, since the thought— as with Ross and Dancy—is that [IC] is relevant to the KvpumlpcDs ma¬ terial (as strongly signalled at 1050'" 19)—but not to the earlier 1050^41050^6. I am offering an alternative to the splitting strategy, by suggesting a way in which [IC]—correctly interpreted—is relevant to 1050^*4—*’6, and thus explaining how the teleological slant present in io5o“4-'^6 can be accom¬ modated within [IC]. This obviates the need for the splitting strategy, the main motivation for which is a desire to insulate the second part of the discussion io5o'^6—1051^2, to which [IC] plainly is relevant, from the first teleologically informed part, io5o'^4-’’6. The discussion in the body of this paper of the distinction at Metaph. d

II,

loip^ia—14, between priority

Kara

yeveoLv and

Kara cfiOopdv

also

suggests a way to relate the two parts of the material at Metaph. 0 8, 1050*4—'’h and loso^'h-iosi^a. [IC] directs us to think about possibilities; when we have as examples of what is actually or potentially F items which are changeable (e.g. children turning into adults), then when we enquire whether something is possible we will have to consider the way in which the actuality could come about—and it is in considering that issue that the difference between possibility /card yeveaiv and Kara cjiOopdv opens up. But when the issue is the priority of eternal to perishable things (io5o’’6 onwards), it is not appropriate to think about the way in which an actual outcome could come about (there is no coming about for eternal things), and so the distinction between possibility

Kara yeveaiv

and /card cj)6opdv does

not arise; thus, [IC] will apply to those cases in a more straightforward way (as signalled at io5o'"i9). Witt, ‘Priority’, differs from those who adopt the splitting strategy, since she allows that [IC] is relevant to the first part of the discussion at Metaph. 0 8 (io5o*4-'*6). But my approach diverges from hers in that I do not try to accommodate 1050*5-6 to [IC] by reinterpreting the ex¬ amples.

BIBLIOGRAPHY Burnyeat, M., et al., Notes on Eta and Theta [Notes], recorded by Myles Burnyeat and others (Sub-Faculty of Philosophy, Oxford, 1984). Dancy, R. M., ‘Aristotle and the Priority of Actuality’ [‘Actuality’], in S.

238

Stephen Makin

Knuuttila (ed.), Reforging the Great Chain of Being (Dordrecht, 1980), 73^115-

Kirwan, C., Aristotle’s Metaphysics Books r,A, and E [Metaphysics D J £] (Oxford, 1971). k Ross, W. D., Aristotle’s Metaphysits; A Revised Text with Commentary [Ross] (2 vols.; Oxford, 1924). ' Witt, C., ‘The Priority of Actuality’ [‘Priority’], in T. Scaltsas, D. Charles, and M. L. Gill (eds.). Unity, Identity, and Explanation in Aristotle’s Metaphysics (Oxford, 1994), 215-28.

CHRYSIPPUS’ PUZZLE ABOUT IDENTITY JOHN BOWIN

Philo of Alexandria, in De aeternitate mundi 48 {SVF ii. 397), gives the following brief and notoriously cryptic account of a puzzle about personal identity created by the Stoic philosopher Chrysippus: (1) Chrysippus, the most distinguished member of their school, in his work On the Growing [Argument], creates a freak of the following kind. (2) Having first established that it is impossible for two peculiarly qualified individuals [§uo iSicay

ttolovs]

to occupy the same substance jointly, (3) he

says: ‘For the sake of argument, let one individual of as whole-limbed, the other

[t6v

[t6v

^.eV] be thought

Se] as minus one foot. Let the whole-

limbed one be called Dion, the defective one Theon. Then let one of Dion’s feet be amputated.’ (4) The question arises which one of them has perished, and his claim is that Theon is the stronger candidate. (5) These are the words of a paradox-monger rather than of a speaker of truth. For how can it be that Theon, who has had no part chopped off, has been snatched away, while Dion, whose foot has been amputated, has not perished? (6) ‘Necessarily’, says Chrysippus. ‘For Dion, the one whose foot has been cut off, has collapsed into the defective substance of Theon. And two peculiarly qualified individuals cannot occupy the same substrate. Therefore it is necessary that Dion remains while Theon has perished.”

Perhaps the most widelv accepted interpretation of this passage is that offered by David Sedley in 1982.^ In this paper, I will offer an interpretation that leaves the most important features of Sedley’s © John Bowin 2003 I would like to thank Elizabeth Asmis, R. James Hankinson, David Sedley, Stephen White, and especially Richard Sorabji for their comments on this paper. ' Trans. A. A. Long and D. N. Sedley, The Hellenistic Philosophers, i. Trajislations of the Principal Sources with Philosophical Commentary [Hellenistic Philosophers] (Cambridge, 1987), 171-2. ^ D. N. Sedley, ‘The Stoic Criterion of Identity’ [‘The Stoic Criterion’], Phronesis, 27 (1982), 255-75.

240

John Bowin

account intact, chief among which is his view on the basic purpose of the puzzle. Like Sedley, I take the fact that the puzzle appears in a work called On the Growing Argument to indicate that it is a rejoinder to, and indeed a reductio ad absurdum of,Khe Growing Ar¬ gument. Where I diverge from Sedley’s approach, I do so to shore it up against certain objections to which I think it is vulnerable. My chief concerns are to achieve a better fit with the text, and to ensure that since we view the puzzle as a reductio ad absurdum, we do not take Chrysippus to be deducing a contradiction by means of premisses extrinsic to the Growing Argument. Otherwise, Chry¬ sippus’ reductio ad absurdum would fail in its purpose of showing that the Growing Argument is internally inconsistent.^ I also follow Sedley on two other significant interpretative points. First, I agree that since, from at least Chrysippus’ point of view, the puzzle runs up against the principle that ‘two peculiarly qualified individuals cannot occupy the same substrate’, we must suppose that we are dealing with one body at the outset and that Theon is a part of Dion.'^ Second, the justification for Dion’s survival that Sedley supplies on behalf of Chrysippus seems right. The amputee who is grieving over his severed foot must be Dion since ‘Theon cannot have lost a foot that was never part of him in the first place’ Here is a very preliminary paraphrase of how Chrysippus’ argu¬ ment appears to run that incorporates these points. At the outset we have one living, anatomically complete human being named Dion, a region of whose body has been named Theon—the whole body except one of its feet. The foot just mentioned is then amputated, with the result that either Dion or Theon must perish because, as Chrysippus tells us (and as Philo apparently agrees), ‘two pe-

^ By an extrinsic premiss, I mean a premiss that is neither explicit in the argument nor plausibly ascribed to the arguer as common sense. ■' Otherwise, when the foot is chopped off, the resulting state of affairs would not run up against this principle, and it is apparent from the text that it must. Besides, as Sedley also points out, Philo essentially tells us that Theon is a part of Dion several pages later in the same text {Aet. 49-51). Prior to Sedley’s 1982 article, the consensus was ‘that [Dion and Theon] are supposed to be two numerically distinct individuals who are qualitatively identical except for the fact that Theon has a foot missing: hence when Dion’s foot is amputated the two are made completely indis¬ tinguishable.’ Sedley cites M. E. Reesor, ‘The Stoic Concept of Quality’, American Journal of Philology, 75 (1954), 40-58; J. M. Rist, ‘Categories and their Uses’, in J. M. Rist (ed.). Stoic Philosophy (Cambridge, 1969), 152-72, repr. in A. A. Long (ed.). Problems in Stoicism (London, 1971), 38-57; J. B. Gould, The Philosophy of Chrysippus (Leiden, 1970).

* Sedley, ‘The Stoic Criterion’, 269.

Chrysippus’ Puzzle about Identity

241

culiarly qualified individuals cannot occupy the same substrate’.* A dispute arises about who should perish. Chrysippus claims that Dion should survive and Theon should perish, since it cannot be Theon who is grieving over his severed foot. But Philo claims, on behalf of the Academics, that Theon must survive and Dion per¬ ish, ‘for how can it be that Theon, who has had no part chopped off, has been snatched away, while Dion, whose foot has been am¬ putated, has not perished?’ I will argue in the sequel that the re¬ sult favoured by Philo is congenial to what the Growing Argu¬ ment would predict—that Theon should survive and Dion should perish—while the result favoured by Chrysippus is not. This, I be¬ lieve, supports Sedley’s claim that Chrysippus’ puzzle is a reductio ad absurdum of the Growing Argument.

I Although the name ‘Growing Argument’ (av^avojxevos Adyo?) was coined by the Academics, the argument itself originated with Epicharmus, the comic playwright of the fifth century BC. The ar¬ gument turns on the assumption that the personal identity of an individual is a strict function of its material composition. Since the material composition of our bodies, so the argument goes, is in a state of constant flux, and since our identities are a strict function of this material composition, our personal identities are also in con¬ stant flux. Epicharmus seems to have meant the argument as a joke, since he exploited its humorous consequences as a stratagem for evading one’s creditors. If, as seems probable, it was the Academy of Arcesilaus that revived the argument in the third century BC, then it also seems likely that the Academics meant it to be a reduc¬ tio ad absurdum of the very notion of personal identity.’ The con¬ clusion that our identities are in constant flux obviously conflicts Philo, who is on the side of the Academics, seems just to assume this principle when he speaks as if the only problem at issue after the amputation is how to determine who has died. ’ .^t Comm. not. 1059 b ff. Plutarch lays out the dialectical context for the dialogue in which the account of the Growing Argument is given (at 1083 A-1084 a). The interlocutor of Diadoumenos has just come from a group of Stoic friends who have been denouncing the ‘older Academics’. The interlocutor says that one of his friends had opined that it was providential that Chrysippus had come after Arcesilaus and before Carneades, because by means of his rejoinders to Arcesilaus, Chrysippus had left many aids to sense perception. Given this background, I follow Sedley (‘The

242

John Bowin

with the common-sense view that personal identity is continuous over time. And given Plutarch’s testimony that the Academy ‘sus¬ pended judgement about everything’ {Against Colotes, 1120 c 9), we should probably assume that the Growing Argument is meant to be aporetic: that is, instead of taking the Academics to be committed to one or the other of the conflicting views—either that matter is the sole principle of identity, or that identity is continuous over time— we should take them to be exposing a conflict between these views and then suspending judgement about its resolution. In the light of this, then, if Chrysippus’ puzzle is itself a reductio ad absurdum, it is a reductio ad absurdum of a reductio ad absurdum, where Chry¬ sippus exposes unintended absurdities in the Academics’ Growing Argument. Sedley says that the target of Chrysippus’ reductio ad absurdum is the Growing Argument’s assumption that matter is the sole prin¬ ciple of identity*—that the personal identity of an individual is a strict function of its material composition. Even though no such principle is expressed in the puzzle, this view makes good sense of a premiss that would otherwise be quite baffling—the fact that Theon and Dion are apparently related to each other as part to whole. Chrysippus’ reductio ad absurdum reduces to absurdity the assumption that matter is the sole principle of identity by means of reducing to absurdity the premiss that Theon and Dion are re¬ lated as part to whole, because the latter is validly deduced from the former. Thus, Sedley says that Chrysippus ‘borrows from the Growing Argument’s own presuppositions’ to ‘concoct’ a premiss in which Theon and Dion are related to each other as part to whole. ‘According to the Growing Argument’, he says, ‘every material addition to or subtraction from an individual results in his re¬ placement by a new individual; and since in such cases the old and the new individual are related as part to whole or whole to part, the Academic argument does indeed imply that whole and part constitute distinct individuals—the very premiss that Chry¬ sippus’ own paradox presupposes.’"' The material additions and subtractions that Sedley has in mind are, no doubt, the changes in bodily bulk caused by the ingestion and excretion of food. I think, however, that Chrysippus has something a bit more bizarre Stoic Criterion’, 272 n. 17) in ascribing the Academic formulation of the Growing Argument given at 1083 A-1084 A to Arcesilaus. * Sedley, ‘The Stoic Criterion’, 270.

^ Ibid.

Chrysippus’ Puzzle about Identity

243

in mind.‘° We are meant, first, to imagine an individual named Theon, who happens to lack a foot. Then we suppose that for a cer¬ tain restricted period, Theon’s body experiences only one material (and quite miraculous) fluctuation—he grows a new foot. Accord¬ ing to the Growing Argument, since the material composition of Theon has changed, we now have a new individual. Let us call him Dion. But since the personal identity of each individual is a strict function of its material composition, and since all of the flesh that constituted Theon is still present in a particular region of the individual that we now call Dion, we must still view this region of Dion as a numerically distinct individual that is related to Dion as part to whole. Therefore, Theon is related to Dion as part to whole. But at first sight, there appears to be a problem. Although this seems to be a valid deduction from the principle that personal identity is a strict function of material composition, prima facie it is in direct conflict with the conclusion of the Growing Argu¬ ment that growth is actually ‘generation’ and ‘destruction’. Ac¬ cording to Plutarch, the Growing Argument concludes that ‘the prevailing convention is wrong to call these [material fluctuations] processes of growth and decay: rather they should be called gen¬ eration and destruction, since they transform the thing from what it is into something else’ {Comm. not. 1083 a 8-c i). Likewise, the Epicharmus fragment concludes that as a man grows, his former self ‘withers’ (fragment 2 DK). Therefore, since old individuals allegedly ‘wither’ when new individuals come into being as a result of growth, Theon should have perished when he grew the foot, rather than becoming part of Dion. Moreover, Plutarch, who is a spokesman for the Academics, seems to think that the notion of two people being in one body is ridiculous. Plutarch, in fact, criticizes the Stoic notion of the peculiarly qualified individual precisely be¬ cause he says it implies the view that each of us is composed of a multiplicity of entities—a parcel of matter, and a peculiarly quali¬ fied individual. For comic effect, Plutarch even likens the Stoics to Pentheus, the deranged king of Thebes, who in seeing double was ‘going crazy in his arithmetic’ {Comm. not. 1083 F 2-3). One may cavil that what follows is too bizarre, and that if Chrysippus meant something like this, he would have had Theon grow a mole instead of a foot, I will show in the sequel, however, that bringing in the growth of a discrete new part makes better overall sense of the puzzle, even though it produces a scenario that is biologically impossible.

John Bowin

244

These considerations raise two questions. First, are there re¬ sources within the Growing Argument, as it is transmitted to us, to resist the conclusion that Theon is a living part of Dion? If there are, then we will clearly need to rethink our position that Theon is a living part of Dion and perhaps even our claim that Chrysippus’ puzzle is a reductio ad absurdum'. The reason, of course, is that we cannot imagine a reductio ad absurdum of the Growing Argument to include premisses that no proponent of that argument would accept. Second, if there are no such resources and the contradiction we have just discussed is unavoidable, can one of the conflicting claims be rejected, and if so, which one? Can one reject the con¬ tention that growth is actually ‘generation’ and ‘destruction’, or must one reject the conclusion that Theon is a living part of Dion? These questions can only be answered by taking a closer look at the texts. Fragment 2 of Epicharmus frames the argument as follows: DEBTOR. If you like to add a pebble to an odd number—or to an even one if you like—or if you take one away that is there, do you think it is still the same number? CREDITOR. Of course not. DEBTOR. And if you like to add some further length to a yard-measure, or to cut something off from what’s already there, will that measure still remain? CREDITOR. No. DEBTOR. Well, consider men in this w'ay too—for one is growing, one declining, and all are changing all the time."

It seems fairly clear that this version of the Growing Argument permits the same inference that allows Sedley to conclude that Theon is a part of Dion. If I add one pebble to a set of say eight pebbles, the number of pebbles would now be nine but the original eight pebbles would still be present as a subset of the new total. Plutarch’s most extended description of the Growing Argument seems to allow precisely the same inference. He lists the premisses of the argument as follows: All particular substances are in flux and motion, releasing some things from themselves and receiving others which reach them from elsewhere; the numbers or quantities which these are added to or subtracted from do not remain the same but become different as the aforementioned arrivals *' Trans. J. Barnes, The Presocratic Philosophers (London, 1979), 106-7.

Chrysippus’ Puzzle about Identity

245

and departures cause the substance to be transformed. {Comm. not. 1083 b 308, trans. Long and Sedley)'^

And from these premisses Plutarch concludes that ‘the prevailing convention is wrong to call these [material fluctuations] processes of growth and decay: rather they should be called generation and destruction, since they transform the thing from what it is into something else’ {Comm. not. 1083

b

8—c

i).’^

It seems clear that

if we imagine a case where something ‘receives some things from elsewhere’ while not at the same time ‘releasing some things from itself’, nothing in Plutarch’s account would block the inference that ‘the old and the new individual are related as part to whole or whole to part’. Just as in the Epicharmus fragment, every change in the ‘number or quantity’ of material parts in an individual results in a change in its identity.And if growth is simply the augmentation of an existing set of material parts, then clearly the unaugmented set will persist as a subset of the augmented set. Granted, an Aca¬ demic might insist that an individual must be a discrete body, which would defeat the line of argument that I am attributing to Chry¬ sippus. But this would amount to introducing a new premiss that appears now'here in our sources and does not strictly follow from the view that matter is the sole principle of identity. If the Grow¬ ing Argument did not contain this premiss, then there would have been no reason for Chrysippus to recognize it in On the Growing Argument. The Academics may well claim in a rejoinder to Chry¬ sippus that an individual must be a discrete body, but this should have no effect on how we interpret the text at hand. One might also object that since the Growing Argument envisages diminution, the set/subset relationship that we have been considering would be dis¬ rupted when diminution occurs at the same time as growth. But the Growing Argument does not say that growth and diminution acting in concert constitute generation and destruction. Rather, the claim is that growth and diminution each constitute both genera¬ tion and destruction, and for this reason it is perfectly legitimate to consider the case of growth in isolation. This feature of the Growing Argument is quite clear in the Epicharmus fragment. Whether Epicharmus is describing addition or subtraction, the alHellenistic Philosophers, 166.

” Trans. Long and Sedley, ibid.

Note that this is different from saying that the ‘number or quantity’ of material parts alone is criterial for identity, which would yield the absurd consequence that all equally numbered sets are identical.

246

John Bowin

leged outcome is the same: the old number perishes when the new one,comes into being. Plutarch’s language is more ambiguous, but since he ascribes his argument to Epicharmus without signalling any disagreement, I see no reason to interpret PJutarch’s account of the Growing Argument differently. It appears, then, when we consider the case of growth in isolation, that there is nothing in the Growing Argument to block the inference from matter being the sole principle of identity to the possibility that Theon could be a part of Dion. Since this is the case, and a contradiction is unavoidable between this result and the view that growth is actually ‘generation’ and ‘destruction’, we can now turn to the question of which of the two, if either, can be rejected. The fact that the claim that growth is actually ‘generation’ and ‘destruction’ appears as a conclusion of the Growing Argument helps us here. That is, the foregoing analysis seems to show us that Plutarch and Epicharmus are wrong to claim that the premisses of their argument establish that growth is actu¬ ally ‘generation’ and ‘destruction’. The supersession of successive individuals undergoing growth results in the incorporation of the su¬ perseded individuals instead of their destruction. Thus, it appears that one must reject the claim that growth is actually ‘generation’ and ‘destruction’ because it has not been validly inferred from the premisses of the Growing Argument. This result, I think, implies that Sedley’s view that Theon is a living part of Dion need only be modified to recognize that Chrysippus must have undertaken a certain sort of argument in On the Growing Argument prior to the passage that Philo summarizes—one that convicts the Growing Argument of the logical error that I have just described, and/orce5 this premiss on the Academics against their will.

II There also appears to be a problem with including the premiss that ‘two peculiarly qualified individuals cannot occupy the same sub¬ strate’ in a reductio ad absurdum of the Growing Argument. I doubt Sedley’s claim that a proponent of the Growing Argument would accept this as a ‘common-sense principle’, chiefly because I find it incredible that the Academics would even acknowledge, much less think it common sense, that there is such a thing as a peculiarly

Chrysippus’ Puzzle about Identity

247

qualified individual. This is because the ‘peculiarly qualified indi¬ vidual’ was a Stoic invention intended to neutralize the Academics’ Growing Argument. Sedley says that Plutarch, arguing on behalf of the Academics, implicitly accepts the existence of peculiarly qual¬ ified individuals in his treatise On Common Conceptions. But if we consider the nature of the cited passage, it seems that this cannot be true. The passage that Sedley refers to {Comm. not. 1077 c-e) is it¬ self a reductio ad absurdum of another Stoic doctrine—that Zeus and Providence come to occupy the same aether during the Conflagra¬ tion. Arguing on behalf of the Academics, Plutarch supposes that Zeus and Providence are peculiarly qualified individuals so that he can infer the unwelcome conclusion for the Stoics that their story about Zeus and Providence requires two peculiarly qualified indi¬ viduals to occupy the same substrate. As in any reductio, Plutarch entertains premisses that he need not accept—that peculiarly qual¬ ified individuals exist—in order to bring out inconsistencies in a contested Stoic theory. Plutarch, in fact, seems to think that the idea of a peculiarly qualified individual is manifestly absurd, on the ground, as I mentioned above, that it implies the non-evident, if not obviously false, claim that each of us is composed of a multiplicity of entities. It is clear from this that Plutarch does not countenance the existence of peculiarly qualified individuals, and it is even clearer that he would not think that any proposition about them could qualify as common sense. So how should we view the premiss that ‘two peculiarly qualified individuals cannot occupy the same substrate’ in the light of this difficulty? I think that this obstacle can be overcome as long as we consider that the very definition of personal identity is in dispute. When a Stoic or a proponent of the Growing Argument confronts a puzzle like this, each will construe the term ‘individual’ according to his own definition (granted, of course, that an Academic would take such a definition dialectically). Consequently, we must keep in mind two points of view as we run through the argument. The Growing Argument defines the individual as a particular collection of material parts. Thus, when a proponent of the Growing Argu¬ ment is told that Dion and Theon are individuals, he will argue that they are collections of material parts. The Stoics, on the other hand, hold that if Dion and Theon are individuals, they must be pe¬ culiarly qualified individuals. So they, of course, will think of Dion and Theon as such when they consider the puzzle. It is important

248

John Botvin

to note that Philo does not explicitly state that Dion and Theon are peculiarly qualified individuals. And it is also telling that he reports in indirect speech that ‘it is impossible for two peculiarly qualified individuals

[Suo

ISicos

ttoiouj]

to occupy the same^substance jointly’

but then switches to a direct quotation as follows: ‘[Chrysippus] says, “For the sake of argument, let one individual [rdv juev] be thought of as whole-limbed, the other

\t6v Se]

as minus one foot.’”

This leads me to suspect that the two premisses just stated are not part of a continuous quotation, and that the rdv line 3 need not refer back to the 8vo ISiojs

ttolovs

ixev

and

t6v Se

in

in line 2. Certainly,

at the end of the passage Chrysippus does say that one of the two must perish because two peculiarly qualified individuals cannot oc¬ cupy the same substrate. But this just reflects the Stoic diagnosis of the problem, and there is nothing to prevent a proponent of the Growing Argument from interpreting this stipulation in an en¬ tirely different way—that two (not necessarily peculiarly qualified) individuals cannot occupy the same substrate. We can also take comfort in the fact that Chrysippus’ reductio ad absurdum still works, even if we assume that Dion and Theon are not peculiarly qualified individuals. When Dion’s foot is amputated, the Growing Argument requires that we call the amputee Theon, because we again have the same collection of flesh that we initially attached this name to. But, as Sedley suggests, there is a good prima facie reason to call the amputee Dion, since why would Theon be grieving over a foot he never had? Thus, the Growing Argument says that the amputee is Theon but common sense says that it is Dion. The amputee cannot be both Dion and Theon because of the principle that two individuals cannot share all of their material parts. So since Dion is alive, then Theon must be dead just as Chrysippus claims, and the Growing Argument is contradicted without making use of any propositions about peculiarly qualified individuals. At first sight, it seems somewhat puzzling that the Academics would accept the stipulation that ‘two (not necessarily peculiarly qualified) individuals cannot occupy the same substrate’, since they might still have escaped the conclusion that Theon is dead by saying that the amputee is both Dion and Theon—that Dion and Theon are still numerically distinct individuals, but their spatio-temporal histories have converged. The stipulation that ‘two (not necessarily peculiarly qualified) individuals cannot occupy the same substrate’

Chrysippus’ Puzzle about Identity

249

is designed to rule out this’ possibility, and this is why Chrysippus would want it in the puzzle. But it is unsatisfying simply to claim, as Sedley does, that the Academics should accept it as common sense, because at this point in the argument, the Academics would have already been forced to accept that Theon and Dion have some of their material parts in common. And this is a strange thing to admit indeed, since Theon and Dion are not related as Siamese twins, for instance, but as part to whole. In this context—being already so far beyond the pale of common sense—it seems like a perfectly reasonable strategy for the Acadernics to bite the bullet and say that Dion and Theon can share all of their material parts, if by doing so they can forgo the additional embarrassment of admitting that Theon is dead. I think that the Academics’ acceptance of this principle makes more sense if we consider the fact that by arguing that matter is the sole principle of identity they seem to propose a criterion of identity. The relevant property of a criterion is that it allows one to make unequivocal ']\idgements. An underlying assumption of the Growing Argument is that given a sufficiently precise specification of an object’s material composition, one should be able to determine that object’s identity unequivocally. If this were not the case, then some additional prineiple would be required and one could not hold that material composition is the sole principle of identity. The requirement that ‘two individuals cannot occupy the same substrate’ seems just to reflect the view that one should assign at most one identity to any collection of matter, which follows from viewing material composition as a criterion of identity. I tried to reflect this earlier by saying that the Growing Argument assumes identity to be a strict function of material composition, since when we call a relation a function, we typically mean that every element in its domain maps onto at most one element in its co-domain. Thus, the requirement that ‘two individuals cannot occupy the same substrate’ is simply a uniqueness requirement that says that if we assign two names to the same collection of matter, they both refer to a single individual. I have argued that Chrysippus’ puzzle is a reductio adabsurdum of the Growing Argument that can be understood without any refer¬ ence to ‘peculiarly qualified individuals’. Why, then, are peculiarly qualified individuals mentioned at all if the concept seems to serve no purpose in Chrysippus’ reductio ad absurdum} I suspect that it is

250

John Bowin

because Philo is drawing from a summation of Chrysippus’ attack on the Growing Argument that includes a Stoic diagnosis of why the argument fails—a summation in which Chrysippus first tells the Academics that the Growing Argument fixes^on an apparently commonsensical, but none the less misguided, concept of personal identity. This, in Chrysippus’ view, is really the root of all of the trouble. An entity must be ‘peculiarly qualified’ to count as an indi¬ vidual because, as we have seen, attempting to define an individual solely in terms of its matter does not work. Next in the summation, Chrysippus reminds us of the general principle that he ‘established in advance’—that ‘it is impossible for two peculiarly qualified in¬ dividuals to occupy the same substance jointly’. Chrysippus has made it clear that the Growing Argument, so far from implying that growth and diminution are really generation and destruction, implies instead that growth results in a multiplicity of individuals that are related as parts to wholes. Moreover, even if we allow, per impossibile and just ‘for the sake of argument’, that Dion and Theon are peculiarly qualified individuals sharing the same matter, it will turn out that one of them must perish, not because an individual is identical to its matter, but because of a metaphysical limitation on peculiarly qualified individuals—that they cannot occupy the same matter jointly. It may be that, according to Chrysippus, hav¬ ing no material parts in common is a necessary condition for two entities to be numerically distinct. But at any rate, it is clear that for Chrysippus, material composition cannot be a sufficient criterion of identity. The irony will not have been lost on Chrysippus that he had convicted the Academics of the very absurdity that they claimed the doctrine of peculiarly qualified individuals implies—that in¬ dividuals consist of a multiplicity of entities. It is the Academics, rather, who have ‘gone crazy in their arithmetic’ by taking up the Growing Argument. Of course, the peculiarly qualified individual is not a multiplicity, but rather a single individual under different descriptions—as a substrate, and as a peculiarly qualified substrate. This is not the way that the man in the street thinks about iden¬ tity, but the man in the street is often wrong. When the Stoic talks of ‘common conceptions’ he does not mean ‘common opinions’, and indeed, the common opinion that matter is the sole principle of identity is, on this showing, incoherent. The Academics set out to show that the very notion of personal identity is incoherent

Chrysippus’ Puzzle about Identity

251

by exposing a conflict between two venerable items of common opinion—that matter is the sole principle of identity and that iden¬ tity is continuous over time. What Chrysippus’ puzzle shows is that one of those common opinions is incoherent by itself, and this resolves the Academic aporia. The University of Texas at Austin

BIBLIOGRAPHY Barnes, J., The Presocratic Philosophers (London, 1979). Gould, J. B., The Philosophy of Chrysippus (Leiden, 1970). Long, A. A., and Sedley, D. N., The Hellenistic Philosophers, i. Transla¬ tions of the Principal Sources with Philosophical Commentary [Hellenistic Philosophers] (Cambridge, 1987). Reesor, iM. E., ‘The Stoic Concept of Quality’, American Journal of Phi¬ lology, 75 (1954), 40-58. Rist, J. M., ‘Categories and their Uses’, in J. M. Rist (ed.). Stoic Philo¬ sophy (Cambridge, 1969), 152—72; repr. in A. A. Long (ed.). Problems in Stoicism (London, 1971), 38-57. Sedley, D. N., ‘The Stoic Criterion of Identity’ [‘The Stoic Criterion’], Phronesis, 27 (1982), 255-75.

jr/.'fMJ-,■"•1‘j ♦

» « /!> if'* rrad

'lyj •-'^ Y"* '‘'' ’*-'■ ' ' V\

.-oH' |H.!i t .

! ,

Wsr;'.'-', w M

■!aisr:''i f3":f!jr:»..n.jiJiJj-' • '’f'"^;.

■ ■ -stf-'n.! ..,. ^ ,^ai*o ji-H;

vr

5

C^.

,

ilk»

u

^■ -. "



— x’-«rfr»o*f .« 't.V.viaW'|

. 1^.

•s*iptA-.-AV( .'.•■.■i:>r(jirtw*.' ■ -r ‘..')

,i.

\

'1

w-'. ;’v'4

•■’• .



MJ

«J: .

j V*U>

' : 1.. icv H'Op

fa-?',

- .-ci

(ipic.

''f' .

j

•’’V'* ' 'j

4{

) *n7'iji. '’l W

‘".r'-

,

'

is-

V r-rSn-s •• •\,|- 'nrt ’. It J.'.iil •

( \iC1 •f«|

M.
aivovrai, Trip t^^vikov 6S

vaiv to pick out any determinate class of options, and play a meaningful role in deliberation, it must have a special sense—a sense I take to be marked in these contexts by its association with ‘selection’ (eKAoyr)).

A Puzzle in Stoic Ethics

307

dance with nature are to-be-taken, and all things contrary to nature are not-to-be-taken’ (Stob. ii. 82. 20-1 (LS 58c)). Now these categories also seem to be equated with those of the ‘indifferents’ (d8id(f)opa) W'hich are ‘preferred’ {Trpo-pyfxdva) and ‘dispreferred’ {dTTOTrporjypieva)—the clumsy terminology is said to be Zeno’s own coinage (Cic. Fin. 3. 51; Stob. ii. 84. 21-4). As Dio¬ genes Laertius explains, some things are ‘indifferent’ inasmuch as they ‘activate neither impulse nor repulsion, as in the case of having an odd or even number of hairs on one’s head, or stretching or con¬ tracting a finger’ (7. 104 (LS 58A)). But these absolute indifferents, as w'e might call them {Kaddna^ d8Ld(f>opa, Stob. ii. 79. 9), must be distinguished from other things which are ‘indifferent’ in a weaker sense. The latter make no difference to our happiness or unhap¬ piness; yet they are still capable of moving us to action, and may indeed do so rationally and appropriately. These are the objects of our selection and disselection (D.L. 7. 104-5). So selection and disselection are applicable to objects which, without being genuinely good or bad, are still capable of appropriately arousing ‘impulse’ (6pp.ri), i.e. the assent to a motivating impression, which is, in Stoic psychology, the cause of human action.*^ Various confusions and refinements apart,the position so far I will have little to say in this paper about impulses and impressions. One would expect these concepts to be the crucial framework for Stoic thinking about deliberation: however, so far as I can see, all the very different models of deliberation I consider in this paper are compatible with what we know about the Stoic theory of action. I cannot here enter into all of these complications, but will note a few; cf. the discussion of A. F. Bonhoffer, The Ethics of the Stoic Epictetus [Epictetus], trans. W. Stephens (New York, 1996; first pub., Stuttgart, 1894, cited by original pagination), 172-7). (i) Diogenes Laertius and Stobaeus both distinguish three senses of value (dfia) (D.L. 7. 105; Stob. ii. 83. 11-84. 3); they do not correspond very exactly, however, and Stobaeus’ are further compjicated by some comments attributed to Diogenes of Babylon (ii. 84. 4-17). In both texts, one of the three senses relates to the appraisal of an expert, which seems unilluminating and perhaps irrelevant here (though of course the Sage is an expert on the values of things). Of Diogenes Laertius’ other senses, one applies to all goods; the other is ‘a certain intermediate power or use’ ifjLearjv riva Svva^iiv 17 xpctav) which contributes to the life in accordance with nature;

the examples of this are health and wealth, so this is clearly the sense (or one of the senses) in which value is characteristic of the preferred indifferents. Of Stobaeus’ other senses, one, credited to Antipater, is ‘selective’ (iKXeKTiKri) value; this too is marked by the examples of health and wealth as being the value belonging to the preferred indifferents. The other sense is ‘a thing’s contribution and rank in itself’ (T-qv T€ Soaiv Kal TifjLriv Kad’ avTo) (ii. 83. 12 (LS 580)); it would be natural to take

this as corresponding to Diogenes Laertius’ sense reserved for goods. However, at

308

Rachel Barney

seems to be something close to a set of equations. Things in accor¬ dance with nature are things which have value, and things which have value (or at any rate sufficient value) are preferred indifferents, to be selected or taken; things, contrary to natdre have disvalue, and things which have disvalue are dispreferred indifferents, which should be disselected or rejected. But these equations are uninfor¬ mative until we have the answers to two questions; (i) what things are in accordance with nature, valuable and preferred? and (2) what is the theoretical pomt of these categories—or, to put it another way, why should we select and disselect these things? Fortunately our texts offer quite explicit, if not exhaustive, answers to both quesStob. ii. 84. 4—13 Diogenes of Babylon seems to claim that the ‘contribution’ sense appropriately applies to preferred indifferents; LS read Stobaeus’ prior distinctions accordingly, arguing that the same thing (e.g. health) might have value in all three senses, and that ‘health is something valuable per se' (ii. 351). On that reading, the distinction between this ‘contribution’ value and ‘selective’ value is perhaps between preferability in general, at the level of types, and the preferability of a token, all things considered, in a particular ‘selective’ context (cf. (2) following).) (2) Several passages indicate that not all indifferents with value count as preferred: only some have

value to qualify (S.E. P/if 3. i9i;M. 11.62; Stob. ii. 80. 14-

21, ii. 84. 18-22; Cic. Fin. 3. 51, cf LS ii. 352). It is odd that no text explains what the other, mysteriously underweight, indifferents might be. I would suggest that they are either {a) tokens of the usual preferred indifferents in too small a quantity to be rationally motivating (wealth is preferred, but the Sage need not pick up a penny from the pavement); or {b) tokens of the usual preferred indifferents which fail to be rationally preferable in some given situation, because they are outweighed by other more valuable ones. In the latter case, this distinction, too, simply gestures, clumsily, towards the distinction already noted under (i), between what has value and is ‘preferred’ in the weak sense of being a generally preferred type (or being a token of such a type), and what turns out to be ‘preferred’ in the stronger sense of being the correct object of selection in a particular situation. (3) Some texts distinguish between what is preferred for its own sake (Si’ avrd), what is preferred for the sake of something else, and what is both (D.L. 7. 107; Cic. Fin. 3. 56; cf. Stob. ii. 82. 21—83. 9)- Little seems to hang on this: per se preferability should not be confused with the way in which one might, incorrectly, view the preferred indifferents as desirable for their own sake. Since the indifferents are not even a means to the end of virtue, the correct attitude is not to desire them at all, but simply to take or select them with detachment, for their own sake or for the sake of other indifferents as the case may be. (4) Several texts speak of ‘primary’ things in accordance with nature, but they seem to pick out rather different things, and do not add up to a clear picture of how these are related to ‘things in accordance with nature’ simpliciter (prima or principia naturae, Cic. Fin. 3. 17-23; rd irpcPra Kara (jivuiv, Stob. ii. 80. 6-8, ii. 82. 11-12; Plut. Comm. not. 1071 Aff.; Galen, Plac. Hipp. et Plat. 5. 6. 10-14). The principal distinction between what is ‘primary’ and what is not, I would suggest, is both logical (overlapping with the ‘for its own sake’/‘for the sake of something else’ distinction) and chronological. Wealth is preferred because we learn that it can help us to obtain other indifferents, and our impulse towards it is therefore subsequent to our impulse towards the latter (cf. LS i. 357).

A Puzzle in Stoic Ethics

309

tions. Diogenes Laertius gives several lists of preferred and dispreferred indiflferents: ‘life, healths pleasure, beauty, strength, wealth, reputation, noble birth, and their opposites, death, disease, pain, ugliness, weakness, poverty, low repute, ignoble birth, and the like’ (D.L. 7.

102

(LS

58A)).

Or, according to another of Diogenes’ lists,

‘among things of the soul, natural ability, skill, [moral] progress and similar things; among bodily things life, health, strength, good condition, soundness, beauty and the like; among external things w'ealth, reputation, noble birth and similar things’ are all preferred; and the dispreferred indifferents are-their opposites (7. loh).'"* The rationale for these classifications goes back to Plato. At Euthydemiis 278E-281

e

Socrates argues that ‘goods’ like wealth,

health, beauty, noble birth, power, honour, and even useful psy¬ chological qualities such as self-control can be used well or badly. When used well they are beneficial to their possessor, but when used badly they are harmful; and the cause of right use is know¬ ledge or wisdom. He concludes that these ‘goods’ are not good in themselves or by nature, but only if put to use by wisdom; if ig¬ norance controls them, they are greater evils than their opposites. Only wisdom is good in itself, and ignorance bad; and so the key to happiness is the possession of w'isdom. Essentially the same argu¬ ment is attributed to the Stoics by Diogenes Laertius. Only what reliably benefits or harms us is good or bad respectively: ‘For just as heating, not chilling, is the peculiar characteristic of what is hot, so too benefiting, not harming, is the peculiar characteristic of what is good. But wealth and health no more do benefit than they harm. Therefore wealth and health are not something good’ (D.L. 7. 103

(LS58a)).'^ But why, then, should some indifferents count as ‘preferred’? The designation does not simply express the claim that the preferred and dispreferred indifferents standardly do arouse our impulses of pursuit and avoidance. A preferred indifferent is one which it is rational for us to prefer; and what makes it rational is the norma'■* Trans. B. Inwood and L. Gerson, Hellenistic Philosophy, 2nd edn. (Indianapo¬ lis, 1997). Similar enumerations are given in other sources: according to Stobaeus, the preferred indifferents include ‘health, strength, well-functioning sense organs, and the like’ (ii. 79. 20—80. i (LS 58c)). ” Hence also, in some texts, an ancillary schema of terminological distinctions: the preferred and dispreferred indifferents provide not ‘benefits’ {ernolumenta) and ‘harms’ (detrimenta), but ‘advantages’ (commoda) and ‘disadvantages’ (incommoda) respectively (Cic. Fin. 3. 69).

310

Rachel Barney

tive standing of nature to which I alluded earlier. We are endowed by nature with an inborn orientation to our own constitution and whatever promotes it: like animals, we have a natural tendency to pursue health rather than sickness. So the preferred indifferents are in accordance with nature in the important sense that we are constituted by providence (orX5od, or ‘universal nature’), to have an impulse towards them. Strictly speaking, everything fate sends us is in accordance with nature; for it serves the interests of the whole and with it our interest as parts of that whole.But since we do not know what is fated, we should act on the basis of our innate tendency to select the preferred indifferents—a tendency, after all, which nature or God has itself beneficently instilled in us.” Thus when the advanced Stoic forms the impulse to pursue some indif¬ ferent, he does not simply give in to a habit or ingrained tendency, but recognizes a demand of nature and reason. There is also a second line of reasoning behind the attribution of value to the indifferents. This can best be seen by considering the challenge posed to Stoic orthodoxy by the views of Aristo of Chios. Aristo notoriously rejected the categories ‘preferred’ and ‘dispreferred’ and insisted that everything between virtue and vice was absolutely indifferent—and indeed that the telos consisted pre¬ cisely in indifference to everything except virtue and vice (S.E. M. II. 64—7; D.L. 7. 160; Cic. Fin. 3. 50). The correct interpretation of Aristo’s position is a matter of some controversy, but two of his arguments are worth noting here. One is that under special circum¬ stances, it is rational to disprefer the ‘preferred’, and vice versa. If a tyrant is drafting all healthy men into his army, where they can expect to be killed, while the sick are exempt, the wise man will chose sickness over health (S.E. M. 11. 64—7). This argument makes a point which the Stoics could and indeed did accommodate: they need only distinguish between the claim that some indifferent type is in general ‘preferred’ and the fact that tokens of it may or may not be preferred (i.e. be rational objects of selection) in particular contexts, all things considered.'* But Aristo adds to it an analogy which cannot be accommodated by the orCf. Plut. Stoic, repugn. 1050 c-d; Gelliusv. i. ” Cf. Chrysippus ap. Epict. 2. 6. 9-10. Aristo’s argument still scores a decent ad hominem point, however, if only regarding terminological hygiene. For he shows that the Euthydemus argument relied on by the Stoics can be made to tell against their own position: after all, if nonmoral ‘goods’ should not be deemed good because they are not beneficial in every

A Puzzle in Stoic Ethics

311

thodox theory: ‘Just as in writing people’s names we put different letters first at different times, adapting them to the different cir¬ cumstances ... so too in the things which are between virtue and vice no natural priority for some over others arises but a priority which is based rather on circumstances’ (S.E. M. 11. 67 (LS

58F)).

How the Sage would determine what circumstances require seems to have been left obscure by Aristo; and the Stoics charged that there could be no rational basis for action if everything other than virtue and vice were absolutely indifferent. Cicero thus claims that on Aristo’s view, ‘the whole of life would be thrown into chaos . . . Wisdom would have no role or function, since there would be no difference whatsoever between any of the things that pertain to the conduct of life, and so no method of choosing could properly be applied’ {Fin. 3. 50). This response to Aristo brings out an important point which we might already have suspected from the telos formulae of Diogenes and Antipater: ‘the selection and disselection of indifferents’ is evi¬ dently a description of any and all rational action. For only in that case are considerations of value necessary for practical reason to have any ‘role or function’ at all. (I will return to the implications of this shortly.) Thus the intermediate standing of the ‘indifferents’ is fixed by powerful convergent pressures within Stoicism. On the one side is the Platonic argument that such things cannot be genuinely good or bad—an argument not just passively inherited from Plato, but necessary to the central Stoic project (shared with their Epicurean rivals) of showing that happiness is always within our power. On the other are two powerful reasons to ascribe genuine value to the indifferents: first, the need to endorse the natural drives and dispo¬ sitions granted to us by providence; and second, the need to stop short of the arguably irrationalist position reached by Aristo. Hence the delicate and terminolbgically laborious system of balances and qualifications which make up the Stoic account of the indifferents. On the one hand, indifferents may be genuinely and objectively natural, valuable, preferred and selected; on the other, they cannot be good, beneficial, constitutive of happiness, or legitimate objects of choice and desire.’’’ situation, neither should they be called ‘preferred’ when they are not preferable in every situation. To note in passing another important aspect of the indifferents’ role, the mis-

Rachel Barney

312

In most of our sources, the contrast between genuine goods and the indifferents typically appears as an ethical topic in its own right. But the indifferents also play a crucial role in Stoic accounts of human maturation and moral development. Th^ human infant, like other animals, naturally.seeks what will preserve it, and this forms the necessary starting point for its performance of ‘appropriate actions’ {KadrjKovTa). In the important developmental account of Definibus 3, Cicero explains: With this established, the initial ‘appropriate action’ [officium] (this is what

I call the Greek

KadrjKov)

is to preserve oneself in one’s natural constitution.

The next is to take what is in accordance with nature and reject its opposite. Once this method of selection (and likewise rejection) has been discovered, selection then goes hand in hand with appropriate action. Then such selec¬ tion becomes continuous and, finally, stable and in agreement with nature. (Fin. 3. 20)

This might suggest that the ‘things in accordance with nature’ are merely a starting point, to be transcended once our practice of appropriate actions has got off the ground. And indeed, our attitude to the indifferents is, Cicero explains, transformed when we recognize the infinitely greater beauty and importance of rational and appropriate action itself, in something approaching a moment of conversion or transcendent insight {Fin. 3. 21-3); when one ‘sees an order and as it were concordance in the things which one ought to do, one then values that concordance much more highly than those first objects of affection’ (3. 21). This concordance or consistency (Cicero notes the Stoic use of the Greek term OjaoAoyia, reminiscent of Zeno’s telos formula) is the true location of the human good; it is ‘the only thing to be sought in virtue of its own power and worth, whereas none of the primary objects of nature is to be sought on its own account’ (ibid.). But this transformation, and the Stoic agent’s new commitment to the supreme importance of rational action, does not have the implications we might assume. Cicero insists vehemently that an agent who has fully grasped the indifference of the natural ‘starting points’ will continue to select them. Nor is their role reduced to an instrumental one, as a means to the now recognized real good, take of taking them to be genuinely good or bad is the hallmark of those pathological states, the emotions {-rradr)). This helps to explain why it is worthwhile to put some¬ thing like ‘noble birth’ in the list of indifferents (D.L. 7. 106): I can hardly pursue it, but I might be mistakenly pained at not having it, envious of those who do, etc.

A Puzzle in Stoic Ethics

313

despite the somewhat misleading language just quoted.^** Neither is there any indication that the considerations on the basis of which the advanced Stoic selects indifferents differ in kind from those applied previously. True, the advanced Stoic’s deliberation is characterized by a new detachment, owing to his recognition that no indifferent is genuinely good or bad; and he now values homologia, consistency, in action more than any particular indifferent—just as he now re¬ cognizes that the telos (according to Diogenes of Babylon, anyway) is a matter not of selecting (let alone obtaining) the indifferents but of ‘reasoning well’ in doing so. But it is not clear that norms such as ‘reasoning well’ and ‘being consistent’ will figure in the Stoic’s deliberations as, so to speak, considerations in their own right— indeed, they could not do so unless she can give them a more con¬ crete sense than Cicero provides. Rationality and consistency may be better understood as deliberative virtues which supervene when selection, understood as before, is carried out correctly and in the right spirit. (Of course, just what ‘correctly’ might mean remains to be clarified: see the remainder of this section and Section II.) And it seems likely that we are to understand the ‘transformation’ of selection in just this way. The indifferents remain objects of our agency in their own right, as Cicero flatly affirms: What I have called ‘appropriate actions’ originate from nature’s starting points, and so the former must be directed towards the latter. Thus it may rightly be said that all appropriate actions are aimed at our attaining the natural principles [principia naturae]. It does not mean, however, that this attainment is our ultimate good, since moral action is not included among our original natural attachments. (Fin. 3. 22, my emphasis)^'

Otherwise, he notes, the Stoic Sage would risk ending up, like Pace G. Lesses, ‘Virtue and the Goods of Fortune in Stoic Moral d'heory’, Oxford Studies in Ancient Philosophy, 7 (1989), 95-128. An instrumental role (in any normal sense) for the preferred indifferents would actually give them a much closer connection to the good of virtue than is compatible with the profound separation between the two on which both the Stoics and their critics insist. ‘In any normal sense’, because the ‘hierarchical’ relation I elaborate later in this section, using the model of motivations in a game, could (rather misleadingly) be described as making the selection of goods (though not the goods themselves) a means to a higher end. Does Cicero (or his source) simply assume that whatever is the ‘primary’ object of our impulses in the temporal sense must also be logically primary, by forming the raw material of all deliberation? The assumption (for which cf Galen, Plac. Hipp. et Plat. 5. 6. 10-14; Cic. Fin. 5. 17-20) is obviously questionable. But it can perhaps be explained by the Stoic commitment to preserving continuity with our natural tendencies as seamlessly as the Epicurean could claim to do (cf. sect. vi).

Rachel Barney

314

Aristo’s wise man, with no material on which his wisdom might be exercised. In fact, even the decision of a Sage to commit suicide will, according to the De finibus account, be determined by the balance of indifferents in his situation: ‘It is the^appropriate action to live when most of what one has is in accordance with nature. When the opposite is the case, or is envisaged to be so, then the appropriate action is to depart from life’ {Fin. 3. 60).^^ In deciding whether or not to commit suicide, both a Sage and a non-Sage may avail themselves of such a principle: so far as we can tell from De finibus 3, the deliberations of the two will differ only in spirit (the Sage will be calm, detached, and fearless, knowing as he does that life and the other indifferents are neither good nor bad) and in accuracy (for only the Sage’s decisions are informed by the craft or science of living). Like the Stoic response to Aristo and the telos formulae of Dio¬ genes and Antipater, Cicero’s insistence on the continued central¬ ity of the indifferents for decision-making strongly suggests that all rational actions may be parsed as selections of indifferents, a principle I will refer to as the exhaustiveness of selection. Indeed, it seems that all the considerations involved in deliberations are evidently considerations regarding the indifferents as such, i.e. as bearers of value and disvalue. I will call this closely related (but arguably stronger) principle deliberative sufficiencyThe upshot of this pair of principles is that, as John Cooper has put it, virtue turns out to be a ‘formal’ condition: ‘All the specific, substantive content of this state of mind—everything that determines what the virtuous person wants, cares about, makes an object of pursuit or avoidance in his actions, etc.—is drawn from the list of “preferred” and “avoided” (or “rejected”) things.’^* Cf. Plut. Stoic, repugn. 1042 D (with an attribution to Chrysippus), The position is of course an attempt to deal with an awkward question: given that the Sage is, ex officio, perfectly happy, and every non-Sage unhappy, is not suicide always irrational for the former and always rational for the latter? Other sources give what seem to be very different criteria for determining when suicide is an appropriate action. See

5FFiii.

757-68, and J. M. Cooper, ‘Greek Philosophers on Euthanasia and Suicide’

[‘Euthanasia’], 'm id.. Reason and Emotion {FrinceXon, 1999), 515-41. “ In addition to the evidence for these principles provided by the telos formulae and the response to Aristo, it is worth noting that both seem to be assumed by the critics of the Stoic telos, including Alexander (‘it is surely absurd [for the Stoics] to say that virtue applies only to selecting’, De anima 2. 164 (LS 64B)) and Plutarch {Comm. not. 1069 c-D, 1071 a-b, etc.). Cooper, ‘Euthanasia’, 534. Cf. also T. Brennan, ‘Demoralizing the Stoics’ [‘De¬ moralizing’], Ancient Philosophy, forthcoming.

A Puzzle in Stoic Ethics

315

So to progress as a Stoic is to learn to care and not to care. We are not to desire the preferred indifferents with the urgency appropriate to goods; but we are to opt for them with a tranquil diligence. As Seneca puts it in an oddly debonair moment: Of course I will [select preferred indifferents]. Not because they are good, but because they are in accordance with nature, and because they will be taken on the basis of my good judgement. ‘What, then, will be good in them?’ Just this—being well selected. For when I put on the right sort of clothes, or walk as I should, or dine as I should, neither the dining nor the walking nor the clothes are good, but the intention I display in them by preserving a measure, in each thing, which conforms to reason. ... So it is not elegant clothes which are a good in themselves, but the selection of elegant clothes, since the good is not in the thing but in the quality of the selection. {Ep. 92. 11-12 (LS 64;))

Unsurprisingly, this delicate position was a focus for attacks by the Stoics’ philosophical rivals. A standard criticism was that the status the Stoics wish to assign to the indifferents is unstable: if they are not matters of pure, Aristonian indifference, the preferred indifferents must really be a rival good. I cannot here work through the complex dialectic between the Stoics and their critics on this question. I will briefly argue, however, that the intermediate status which the Stoics assign to the indifferents is in fact quite defensible; seeing how this is so will help to bring out the point at which, I will suggest, the Stoic account does become problematic. As a starting point, consider the Stoic claim that the virtues are possessed and exercised as a unity, collectively constituting a craft or art {techne) of living. The Stoics seek to clarify aspects of that craft by way of analogies with various other skills—acting and dancing, navigation and medicine, dice-playing and archery. The last of these is the most suggestive here: the Stoic account of the relation of the indifferents to the telos is supported by comparison with an archer whose goal is not to hit the target (a sudden gust of wind might make that impossible, even if he executes the shot impeccably), though all his efforts are directed towards that reference point, but simply the correct exercise of his craft.Cicero elaborates: Here, though, one must immediately avoid the error of thinking that the theory is committed to there being two ultimate goods. Take the case of Cf. also Plut. Comm. not. 1071 B-c, 1072 e-f; LS i. 406-10; A. A. Long, ‘Carneades and the Stoic Telos’, Phronesis, 12 (1967), 59—90.

Rachel Barney

316

one whose task it is to shoot a spear or arrow straight at some target. One’s ultimate aim is to do all in one’s power to shoot straight, and the same applies with our ultimate good. In this kind of example, it is to shoot straight that one must do all one can; none the less, it is to do all one can to acomplish the task that is really the ultimate airn. It is just the same with what we call the supreme good in life. Actually to hit the target is, as we say, to be selected [seligendum\ but not sought [expetendurri]. (Cic. Fin. 3. 22)

For a perhaps clearer analogy, consider, as a counterpart to the pro¬ ficient Stoic, the deliberations and attitudes of an advanced tennis player, meaning one who plays tennis both skilfully and for the right reasons.^® Ask such a person why she hit the particular shot she did in any given situation and the answer will be a strategic one: given the particular circumstances, it was the most likely to win the point of any shot she could play, or the most likely to set up such a shot. For particular actions within the game of tennis are determined by the aim of winning points, and with them games and matches. But if we reiterate the question at a higher level (or, we might say, ‘externally’ to the game), asking the player why she cares about winning points, a reasonable answer would be that in a sense she does not care—it is only a game and the points mean nothing—but that she has various reasons for playing tennis, and to play is to play to win. We can imagine a child who would be un¬ able to articulate any such ‘higher’ motivations, and who might not This account is intended to build on Gisela Striker’s in ‘Following Nature: A Study in Stoic Ethics’, Oxford Studies in Ancient Philosophy, g (1991), 1—73; cf also Adam Smith, The Theory of Moral Sentiments, 7. 2. i. 24 (presumably following Epict. 2. 5), and M. Frede, ‘On the Stoic Conception of the Good’, in K. lerodiakonou (ed.). Topics in Stoic Philosophy (Oxford, 1999), 71-94 at 91-2. Striker notes the crucial point that the goal or end of a game (chess, soccer, archery) is not to be identified with what constitutes the intended result within it (31): checkmating one’s opponent is not the end of chess, even if it is the aim which governs every move the chess-player makes. However, Striker concedes in the end that ‘the comparison of virtue with the skill or craft of a player is still misleading’ (32). This is apparently because in a game ‘the performance of players is evaluated—just as in non-stochastic crafts—in terms of their success. The best player is the one who wins most often, even though she may occasionally lose. . . . But this shows that the analogy between virtue and the skill of a good player breaks down, since moral evaluation, as was emphasized before, is not based upon success’ (33). As I suggest above, this is not quite right: the best player is rather the one who exercises to the highest degree those abilities which are, in general and all else being equal, most likely to produce victory. The problem is rather, I will try to show, with a more specific feature of the Stoic position, viz. their attempt to identify the craft for which the obtaining of preferred indifferents serves as reference point with virtue of any recognizably moral sort.

A Puzzle in Stoic Ethics

317

even experience them, being simply motivated by a kind of blind competitiveness; but it seems natural to think of that as a starting point to be transformed by experience and reflection. So it is wrong to treat winning points at tennis as the end of playing tennis. The motivational structure is rather the reverse: we play to win in order to play tennis. And an advanced player should have no difhculty in explaining the higher ends attained by playing tennis: to become more fit, for exercise, for the benefits of the sociability and competition involved, for the ‘love of the game’ and the satisfaction of playing it well. Winning is not even a means to these ends; but trying to win is, because it is constitutive of playing the game. Thus the ‘internal’ aim of winning points remains explanatorily central, in two respects. First, as I have noted, for any player it is the deliberatively sufficient ‘reference point’ used to arrive at decisions within the game. If I am exercising the craft of tennis, my decisions about what shot to make are entirely governed by considerations about winning points: something has gone wrong if we need to invoke the ‘higher-order’ considerations behind my playing tennis at all. (‘Why did you play that shot that way?’—‘To get more exercise’—‘Oh,

properlyV) The other significance of

this reference point is that the correct exercise of the craft must be defined in relation to it. This relation cannot be direct: it is not the case that the best player is the one who wins the most often. (If it were, then an Agassi who usually loses to a Sampras would be a poor player, and indeed the most important part of skill in tennis would be skill in choosing opponents one could beat.) Playing well is rather playing in a way which in general—all else being equal and circumstances aside—is likely to result in winning; and skill in tennis is the set of physical and strategic abilities which are actualized in playing well. Likewise with Stoic virtue. We live happily when we live skilfully, exercising the art of living. The advanced practitioner of that art treats the indififerents as a deliberatively sufficient ‘reference point’: it is in terms of attempting to obtain them that each of her decisions is arrived at and can be explained. But her goal in doing so is simply the exercise of her skill, though as a child she might well have taken obtaining those indififerents to be desirable in itself. And corresponding to the higher-order ends served by tennis, the advanced Stoic will be able to explain the value of the exercise of her art in terms of rationality, consistency, agreement with nature.

318

Rachel Barnev

conformity to divine will, and the fulfilment of human nature—the whole rich and complex content provided to the idea of virtuous agency by the apparatus of Stoic theory. On this model of deliberation a hierarchical superstructure of reasons for action may'lurk behind any particular decision, just as it does (or should) behind any shot in tennis. The Sage buys lunch to promote his health—but also in order to do what is ap¬ propriate, natural, rational, prudent, and so on. And it seems at least possible that all of these descriptions have motivating force. But so long as the latter simply supervene on and reinforce the impulse deliberation has already produced, the principle of delib¬ erative sufficiency—and with it the understanding of virtue as a ‘formal’ condition, rather than one which figures as salient within Stoic deliberation itself—still effectively holds. On this reading, the Stoic distinction between ‘selection’ (eKAoyij), which takes for its objects the indifferents, and ‘choice’ (atpeai?), which properly is applied only to the good, is one not between actions, but between these ‘first-order’ and ‘higher-order’ descriptions and impulses.^* An act of selection may also be a ‘choice’: and, on the interpretation I have been developing, it is by deciding on selections that the Sage discovers how to choose. He seeks to perform actions which are appropriate (KadrjKov); he hopes to reason well, so as to achieve a harmonious agreement with nature; and his grasp of nature, human and cosmic, directs him to identify happiness with virtue, to fulfil his social duties, to treat his fellow humans as kin, and to embrace whatever Fate sends his way. But, like the complex reasons which may lie behind the practice of playing tennis, none of these conIn particular, we are told that there is a class of impulses which have real goods as their object, namely ‘wish’, or rational desire, /SouATjau (D.L. 7. 116); and since virtue and right action are good things, the Sage presumably acquires a new wish whenever he recognizes that some action on his part would be right. Since impulses cause us to act, this rational wish must be a cause of the subsequent action, so the Sage’s actions seem to be overdetermined: one impulse causes his purchase of lunch qua selection of indifferents, while another, quite different, impulse causes the same purchase qua right action. (For right action as a good—ontologically awkward though that may be on Stoic assumptions—cf. Stob. ii. 71. 15-72. 3 (with the canonical example of prudent walking, cf. ii. 96. 18-97. 5); Plut. Stoic, repugn. 1042 E-F (citing Chrysippus); Clem. Strom. 6. 12 (SVF iii. no); D.L. 7. 94; and Cic. Fin. 3. 55.) Cf T. Brennan, ‘The Old Stoic Theory of Emotions’, in J. Sihvola and T. Engberg-Pedersen (eds.). The Emotions in Hellenistic Philosophy (Dordrecht, 1998), 21-70, and ‘Demoralizing’. For the careful distinction drawn by the Stoics between ‘choice’ (al'peais) and ‘selection’ (exAoyTj), cf Stob. ii. 75. 1-6, ii. 78. 7-12, ii. 79. 1-4; Plut. Stoic, repugn. 1042 d-e; Comm. not. 1060 c, 1061 a; Inwood, Ethics., 238-40.

A Puzzle in Stoic Ethics

319

siderations need figure in his first-order deliberations about how to behave in any given situation. d'he game analogy shows that there is nothing formally impos¬ sible about a craft whose reference point is different from its end, and whose immediate end consists in its own exercise. Nor is it structurally impossible for a craft with a low or trivial reference point to serve more exalted ends. The life in agreement with nature could have that structure. But in the Stoic case there is, I will argue, something substantively puzzling about the content to be fitted to this model. For if the ‘reference point’ of the craft of living is the obtaining of the preferred indifferents—if this is what it is to win points at the game of life—then the skills which make up that craft seem most unlikely to resemble human virtue in any recognizable sense. To put it another way, the principle of deliberative sufficiency implies that any ‘higher-order’ reasons for action acquired by the Stoic are, we might say, non-revisionist: they supervene on and rein¬ force the reasons for action already provided by the reference point of selection. But if the higher-order reasons we acquire by making moral progress involve the whole machinery of Stoic ethical the¬ ory, then they are hardly likely to leave our tendencies to select the preferred indifferents where they were; and if they did, they could hardly issue in the actions of which the Stoics in fact approve.

II For consider some of the results which Stoic deliberation is ex¬ pected to reach. Standard instances of ‘appropriate actions’ (Kad-^Kovra), for instance, include fulfilling one’s social roles and obliga¬ tions to kin and country, returning a deposit, discovering the truth, not harming others except in response to injustice, and return¬ ing favours (D.L. 7. 108-9; Cic. Fin. 3. 59; Off. i. 15-20, 47-8). And actions which according to the Stoics are wrong, under nor¬ mal circumstances, include the following: betraying one’s country, showing violence to one’s parents, and stealing from temples (Cic. Fin. 3.32); also stealing—even food from another person when you are starving—if you do it simply for your own sake (Cic. Off. 3. 29), or grabbing someone else’s life-raft in a shipwreck (Off. 3. 89). On the contrary, you should let another more socially useful person have the life-raft (Off. 3. 89-90); you should behave with the ring

320

Rachel Barney

of Gyges as you would without it {Off. 3. 38-9); and if need be you should offer yourself as a hostage to face torture for the good of your country {Off. 3. 99-115). This last case provides the peroration of Cicfero’s De officiis, by far our richest source of reasonably concrete information on what actions the Stoics approve. Here Cicero discusses the story of Regulus, who as consul was taken prisoner by the Carthaginians and sent back to Rome, having sworn an oath that he would return to Carthage unless some Carthaginian hostages were returned. When Regulus got to Rome he followed the Carthaginians’ instructions; but he also argued in the Senate against returning the hostages, as being against Rome’s best interest. His authority prevailed and the captives were retained; Regulus himself kept his oath and returned to Carthage. According to Cicero, he realized when he did so that he was going ‘to a very cruel enemy and most sophisticated tor¬ ture’ {Off. 3. 100).^'^ Why, then, did he go, and rightly according to Cicero? Because justice demands that even oaths to an enemy be kept. And why did he recommend retaining the hostages, dooming himself to a grisly death? Because it would not have been beneficial to his country to hand the hostages back. Therefore for Regulus to recommend against doing so was honourable; and since the hon¬ ourable is always beneficial, Regulus himself was, all appearances to the contrary, better off for his decision. So my puzzle is this, and an embarrassingly crude one it is: how can Regulus’ actions be parsed as instances of the selection of indifferents? For that matter, precisely what consideration, figuring in the deliberation of a practising Stoic, would preclude her select¬ ing the preferred indifferent of wealth by robbing temples? The doctrine of the indifferents can explain why the Sage buys lunch, and a nutritious lunch at that; but it seems to remain utterly silent about the dimension of Stoicism which enjoins law-abidingness, justice, philanthropy, resistance to tyranny, and, in general, what from a non-Stoic point of view looks like selfless behaviour. Yet in Section I we saw good reasons for attributing to the Stoics the principles of the exhaustiveness of selection and the deliberative sufficiency of the indifferents: and in that case, whatever actions the doctrine of selection cannot account for, it excludes. To see just how bad the problem is, it will help to bring out Translations from De officiis are from Cicero, On Duties, ed. M. T. Griffin and E. M. Atkins (Cambridge, 1991).

A Puzzle in Stoic Ethics

321

some assumptions left lurking in the background of the account I gave in Section I—and, so far as I can see, lurking in the back¬ ground of most expositions of Stoicism, ancient and modern alike. To begin with, recall Aristo’s first argument, based on the case of the person who rationally prefers sickness to being drafted and killed. As I noted, the Stoic response is that in particular cases we must consider ‘preferred’ and ‘dispreferred’ indifferents as tokens, not types, and allow that some particular token of a preferred type may be dispreferred under special circumstances and vice versa. In other words, some indifferents mlay outweigh others: in the case in question, the overall outcome ill-health-and-continued-life is clearly preferable to the outcome good-health-and (or, more pre¬ cisely, up until) early-death. And the need to weigh one package of indifferents against another must be a pervasive, even a univer¬ sal, feature of deliberation. It is difficult to think of prospective actions in which only a single indifferent is involved: every time the Sage buys lunch, he must judge that the advantage to his health in doing so outweighs the diminution of his wealth. Selections are necessarily ‘all things considered’: what is rationally preferred or dispreferred can only be assessed in context. Hence the Stoic doc¬ trine of appropriate actions which depend on circumstances: under special circumstances it may be appropriate to mutilate oneself or give away one’s fortune (D.L. 7. 109). Hence too Cicero’s claim that the Sage’s decision to commit suicide will depend on the balance of preferred and dispreferred indifferents in his life {Fin. 3. 60). Such reasoning implies the adoption of a calculus of value, en¬ abling us to weigh the indifferents which may figure in any given deliberation. That must be at least one respect in which correct se¬ lection requires ‘reasoning well’ {evXoyiaTeiv), as per Diogenes’ telos formula: it is at least in part because he has mastered this calcu¬ lus, and can correctly recjcon all the features of a complex situation against each other, that the Sage’s selections are perfectly rational. How this calculus might work is left obscure by our sources: but that all preferred indifferents have ‘value’ (d^ia) and all dispreferred indifferents ‘disvalue’ {arra^la) confirms the existence of a common denominator in terms of which various packages of indifferents may be compared. (And clearly different amounts of value must accrue Cf. B. Inwood and P. Donini, ‘Stoic Ethics’, in K. Algra, J. Barnes, J. Mansfeld, and M. Schofield (eds.). The Cambridge History of Hellenistic Philosophy (Cam¬ bridge, 1999), 675-738 at 695-7.

322

Rachel Barney

to tokens of a single type of indifferent which differ in quantity or degree: to say that wealth is preferred means that more of it is, all else being equal, preferred over less.)

^

So, on this reading, selection requires weighing against each other the quantities of value and disyalue likely to be obtained in action. And this suggests that a rational selection will be one in which the agent selects an option whose value is not outweighed by any other, i.e. one which maximizes value—or expected value, since the Stoics emphasize that it is not the outcome of an action which matters for our assessment of it, but the skill or lack of skill exercised in our decisions. Talk of maximization is not to be found in our texts, and may sound suspiciously anachronistic. However, the ideal of an art of deliberation which would consist in the ability correctly to reckon value, so as to maximize it, goes back to Plato’s Protagoras. In a famous and influential passage, Socrates there argues that ‘no one who knows or believes there is something else better than what he is doing, something possible, will go on doing what he had been doing when he could be doing what is better’ (3580);^' and he sketches a vision of a deliberative art of measurement {rj fierpgTiKrj rexvr], 356 D ff.) as ‘our salvation in life’. For by rationally reckoning harms and benefits against each other, this art will enable us to select the value-maximizing course of action on a systematic and reliable basis (353 c-358 e). And there can be little doubt that, like the Euthydemus and Meno arguments discussed earlier, this passage had a profound influence on the Stoic conception of virtue as a rational art or skill of living. A final respect in which the notion of ‘selecting indifferents’ is usually left ambiguous is in the relation of the indifferents to the agent. The question here is, again, an obvious and crude one: whose health etc. does the Sage select? Is ‘selection’ a procedure which aims at getting its object into one’s own grasp (‘agent-relative’, as I will term it), or something more impartial or ‘agent-neutral’? The point is one on which accounts of Stoicism, ancient and modern alike, tend to be bizarrely vague.But there is considerable pieceTrans. S. Lombardo and K. Bell, in Plato, Complete Works, ed. J. Cooper, assoc, ed. D. S. Hutchinson (Indianapolis, 1997). A descendant of this principle seems to be adopted by the Stoics, though their distinction between the good and the valuable significantly complicates matters: cf. Cic. Tusc. 4. 12-14; Epict. 3. 3. “ We might be tempted to infer ex silentio that ‘selection’/’disselection’ need not be on behalf of or in the interest of anybody at all: that it is simply a way of characterizing the manipulation of indifferents which almost any agency in the world

A Puzzle in Stoic Ethics

323

meal evidence that we are to understand selection as the selection of indifferents for the agent himself. (i) Though ‘selection’ {iKXoyrj) and cognates are rather colourless terms, the Stoics also talk about the preferred indifferents as ‘to-be-taken’

{Xtqvtov,

from Aa/x/3ava»

‘take’), w'hich strongly suggests getting something into one’s own grasp; and Antipater seems to paraphrase ‘selection’ by saying that we should ‘do everything to obtain [ruyxavetv]’ the things in accor¬ dance with nature. (2) The opponents of the Stoics argue that it is absurd to hold that the end consists in ‘selection’ rather than in the actual obtaining of what is selected; and the Stoic defences of their conception of the telos accept both the relevance of the select¬ ing/obtaining contrast and, it seems, the understanding that this is roughly the contrast between impulse and successful outcome. (3) In the seminal Platonic arguments at Meno 87 D-89 A and Euthydemus 278 E-281 e, it is clearly an agent’s own possession of the ‘goods’ W'hich is in question (note Meno 77 c 7-8, where Socrates stipulates that in these contexts ‘desire’ means ‘desire to secure for oneself’). (4) Recall Aristo’s argument that health should not be deemed ‘preferred’: as the Stoics are expected to agree, I should weigh my health against my life in selecting sickness over the draft.

(5) When the theory of the indifferents is put to work in the devel¬ opmental exposition of De finibus 3, the initial appropriate actions which they ground are presented as agent-relative selections, and this is never revisited or revised. Indeed, Cicero says, ‘since all people by nature love themselves . . . the foolish no less than the wise will adopt what is in accordance with nature and reject what is contrary’ (3. 59). (6) Accordingly, as already noted, agent-relativity is also assumed in the De finibus account of when suicide is reason¬ able (3. 60-1). These clues add up, I believe, to good prima facie grounds for us to take ‘selection’ as an impulse to get the object selected into

is bound to involve. In that case, the telos formula of Diogenes, for instance, would use the language of ‘selection’ to emphasize that the Sage’s action will involve some redistribution, so to speak, of items in the world which are themselves indifferent to happiness, while what is not indifferent is the ‘reasoning well’ with which that redistribution (whatever shape it may take) is performed. However, it is hard to make sense of the distinction between selection and disselection on this reading; and the evidence above for an agent-relative understanding of ‘selection’ seems to me to rule it out. " Cf LS §

64

passim-, SVF iii.

polemics of Plutarch, Comm. not.

190-6

passim-, and in particular the extended

1068 F-1072

F.

Rachel Barney

324

one’s own grasp. And with this and the other specifications I have noted in place, we now have a reasonably full picture of what the ‘selection of indifferents’ seems to involve: k

(1) Some equations. For deliberative purposes, the ‘things in ac¬ cordance with nature’ aVe the bearers of (sufficient) ‘value’, i.e. the preferred indifferents; the ‘things contrary to nature’ are the bearers of (sufficient) ‘disvalue’, i.e. the dispreferred indifferents. (2) Exhaustiveness. All actions can be described as selections of indifferents. (3) Deliberative sufficiency. Deliberation incorporates only con¬ siderations about indifferents qua indifferents, i.e. as bearers of value and disvalue. (4) Maximization. A correct deliberation is one which maximizes (reasonably expected) value. (5) Agent-relativity. Selection is selection for oneself, an impulse to get what is selected into one’s own grasp: so (4) should be understood in terms of maximization for the agent. I shall call (i)-(5) the Maximization Model of Stoic deliberation. On this model, the Stoic’s deliberative task in any situation is to determine which selection, of those open to him, will provide him with as much overall value as possible. His deliberations will thus be governed by a sort of ersatz egoistic consequentialism—ersatz, because of course his aim in selection is to maximize that rather mysterious entity ‘value’ rather than happiness or anything which directly contributes to it. This sounds alarmingly unlike the views we usually associate with Stoicism, and it is hard not to suspect that something has gone badly wrong. The starkest difficulty is of course that, like the vaguer account I outlined in Section I, the Maximization Model cannot account for the results the Stoics expect deliberation to reach: a wide range of deliberative procedures could ground Regulus’ heroic decisions, but this is not one of them.^'^ The difficulty is to see precisely where the Maximization Model goes off the rails. As I have tried to show, each of its constitutive propositions is decently The problem is not that the Maximization Model is in a general way ‘too egoistic’ to be authentically Stoic—Cicero insists that as a right action, Regulus’ action was beneficial, i.e. happiness-promoting, to himself—but more precisely, that such heroic actions cannot be parsed as agent-relative maximizations of value.

A Puzzle in Stoic Ethics

32s

grounded in the texts; and there is nothing structurally unstable in the relation it asserts between the indifTerents and the end. If Maximization none the less cannot be the right model of Stoic deliberation, what is?

Ill Here and in the following two sections I consider some alternative interpretations of ‘selection’, ones more easily made compatible with heroic or philanthropic action. The simplest solution leaves intact the structural features of the Maximization Model, including propositions (i)-(5) listed above, and simply reconsiders the scope of the preferred and dispreferred indifTerents. Some of the Stoics themselves seem to have been tempted by this solution. Though scholars have not made much of it, ‘moral progress’ itself, TTpoKo-n-ri, appears as a preferred indifTerent in several of our lists (Stob. ii. 81. i;D.L. 7. 106-7). To endorse Regulan behaviour we need only stipulate that any amount of progress has so great a value as to outweigh any other indifTerents; so the Stoic deliberative calculus will always give the morally ‘right’ sort of results. However, this is clearly the wrong kind of solution. Moral pro¬ gress cannot really figure as a ground-level consideration in delib¬ eration: for we progress morally by performing appropriate actions, and which action would be appropriate is what deliberation seeks to find out. Worse, moral progress is no longer available to the Sage: so it could not figure in his deliberations as preferred (nor could the virtue which replaces it, since this is a genuine good), with the ab¬ surd result that his deliberations would be more dominated by the other preferred indifTerents, such as wealth. Finally, moral progress cannot really have the same kind of normative role as health and wealth, for it has a different relation to the genuine good of virtue. It is perhaps not quite right to say that progress is a means to virtue, but we can at least say that (for human beings) it is a precondition for becoming virtuous; not so the indifTerents.^’ Some other ways we might try to ‘moralize’ the list of the preferred indifferents: (i) Cicero reports that some Stoics took reputation to be preferred, for its own sake; this would affect a promisingly wide range of deliberations, but Cicero emphasizes that it is not the original or (in his view) correct Stoic position {Fin. 3. 57). (2) Outsources often describe at least some appropriate actions as ‘intermediate’ or ‘middle’ {yiiaov, Stob. ii. 86. 2; Plut. Stoic, repugn. 1037 Eff.; medium, Cic. Fin. 3.

326

Rachel Barney

That moral progress figures in some lists of the indifferents shows, I think, that the Stoics themselves felt at least a flicker¬ ing discomfort about the puzzle I have identified. Still, it has the distinct air of an afterthought, not to mention a category mistake. A more elegant and profnisin^ alternative, which also leaves much of the Maximization Model in place, is suggested by Adam Smith’s exposition of the Stoic theory in The Theory of Moral Sentiments. (I say ‘suggested by’ because my presentation will somewhat exagger¬ ate and oversimplify Smith’s account: my interest here is primarily in the menu of alternatives.)’* Smith’s approach is to incorporate other-regarding concerns into deliberation by conceiving of ‘selec¬ tion’ as, ideally, an impartial procedure: Among those primary objects which nature had recommended to us as eligible, was the prosperity of our family, of our relations, of our friends, of our country, of mankind, and of the universe in general. Nature, too, had taught us, that as the prosperity of two was preferable to that of one, that of many, or of all, must be infinitely more so. That we ourselves were 58—9, etc.). These descriptions are puzzling, but the: principal idea seems to be that under any reasonably informative description (‘returning a deposit’), appropriate actions will be ‘between’ or common to both Sages and non-Sages (cf. Cic. Fin.

3-

59; Off. 3. 14-15), and that appropriate action as available to the non-Sage is not a good. Some passages, notably Cic. Fin. 3. 58—60, tie this ‘intermediate’ status to the indifferents in a way that suggests that appropriate actions should themselves be understood as indifferents, presumably ‘preferred’ ones. However, as such they, like progress, would not be available to the Sage (since his appropriate actions, being ‘perfected’, are genuinely good)—again with absurd results, if the point is to ‘moralize’ deliberation. Moreover, Stoic lists of preferred indifferents are consistently of states or objects (health, wealth) to be obtained through action; action-types themselves do not seem to be the right kind of item to be included here. In sect. V I consider a more promising strategy for incorporating considerations of the KadfjKov into deliberations about selection. Smith’s exposition occupies chapters 7. 2. i. 15—47 of the Theory: it is by far the longest exposition of any earlier moral system, and makes clear the profound influence of Stoicism on Smith’s own views. Given the depth of that influence, his account, which draws principally on De finibus 3 and Epictetus, is remarkably accu¬ rate and still helpful; but there are certainly moments of philosophical projection. Epictetus’ demand that we see our own interests in terms of the good of the whole is expounded in terms of viewing our interests ‘with the eyes of others’, and with reference to Smith’s own account of justice in terms of an impartial spectator (7. 2.

I.

19, note k). Of the Sage, Smith says: ‘All his affections were absorbed and

swallowed up in two great affections; in that for the discharge of his own duty, and in that for the greatest possible happiness of all rational and sensible beings’ (7. 2. i. 21). And he continues: ‘For the gratification of this latter affection, he rested with the most perfect security upon the wisdom and power of the great Superintendant of the universe.’ This understanding of Stoic holism and providence is surely a major source of Smith’s own notorious faith in the ‘invisible hand’ of capitalism.

A Puzzle in Stoic Ethics

327

but one, and that consequently wherever our prosperity was inconsistent with that, either of the whole, or of any considerable part of the whole, it ought, even in our own choice, to yield to what was so vastly preferable. (‘Of Systems of Moral Philosophy’, The Theory of Moral Sentiments, 7. 2. I.

18)

Smith here leans heavily on the Stoic picture of moral progress as requiring an expanding circle of oikeiosis, ‘appropriation’: the identification of others and their interests as my own. There is far more to be said about this central Stoic doctrine, and the Stoics’ concomitant insistence on the naturalness of human sociability and philanthropy, than I can enter into here: for our purposes, the im¬ portant question is exactly how the associated motivations are to be integrated into deliberation. The extreme difficulty of parsing considerations grounded on oikeiosis and community in terms of the Maximization Model is in effect what the example of Regulus brought out. The great advantage of Smith’s model is that it pro¬ vides a clear procedure for explaining how such considerations can indeed figure in Stoic deliberation: in making selections, we are to prefer the greater ‘prosperity’ of the greater number. The Maxi¬ mization Model held that Stoic talk of selection comes with, as it were, the agent in an implicit dative of interest: my selections are selections/or me. On Smith’s model, this is replaced by something along the following lines: (sO Agent-neutrality. Different selections can be made on behalf of different interests, and in so far as I have made moral progress, I select what maximizes value impartially. Smith’s model, on which Stoicism turns out to be a kind of quasi¬ utilitarianism, has some tremendous advantages. It retains the wellsupported propositions (i)-(4) of the Maximization Model, but shows how they may b^ reconciled with central Stoic doctrines about oikeiosis, justice, and human fellowship: thus, unlike Maxi¬ mization, it can yield the behaviour which the Stoics actually want. It shows how even the Sage’s refusal of the life-raft could be parsed as a rational ‘selection of indifferents’, and how virtue might consist precisely in an art of making such selections correctly. So it is no surprise that Smith’s model seems to be at work in a number of recent interpretations. As Julia Annas puts it, ‘a person who has developed towards virtue and extended the circles of social oikeiosis will realize that from the moral viewpoint she has

328

Rachel Barney

reason to prefer these things impartially, that is, at the least without arbitrary limitations to particular people’.^’ What is odd, of course, and should give us pause, is that the Stoics themselves never say anything quite like this. That-tantalizingly sirhple ‘for everyone’ is never supplied. Rather, when the language of oikeidsis or justice enters, the language of selection generally departs. The passages which come closest to being evidence for Smith’s model come from Cicero. In Definibus he says (in what may w^ell be the passage Smith has in mind): ‘The Stoics hold that the universe is ruled by divine will, and that it is virtually a single city and state shared by humans and gods. Each one of us is a part of this universe. It follows naturally from this that we value the common good more than our own [communem utilitatem nostrae anteponamus]’ (3. 64). Just what this means is rather unclear: utilitas is not Cicero’s of¬ ficial term for ‘value’ [aestimatio], or for the ‘benefits’ [commoda] provided by what is ‘preferred’ [praeposita] (cf. 3. 50-3, 3. 69). Still, the general commitment to impartiality seems clear, and is ringingly reaffirmed in De officiis, where utilitas seems to be ‘benefit’ in an ambiguous sense applicable to both the preferred indifferents and the genuinely good (cf. Section V below). Here Cicero even suggests that to determine what is appropriate a quasi-utilitarian form of deliberation may be necessary: ‘promises should not be kept if they are disadvantageous to those to whom you have made them. Nor, if they harm you more than they benefit the person whom you have promised, is it contrary to duty to prefer the greater good to the lesser’ {Off. i. 32). Thus prima facie duties are only that; what determines our real duties, it seems, are their consequences, with the ‘harms’ and ‘benefits’ impartially calculated. So, despite the mysterious failure of our texts to supply the cru¬ cial ‘dative’, and the contrary evidence I noted in Section I for the agent-relative reading of ‘selection’. Smith’s model is not wholly without textual support. And it provides what looks like a suitably central and direct role for the ‘other-regarding’ concerns involved in the central Stoic doctrines of oikeidsis, philanthropy, and justice.^** ” The Morality of Happiness (Oxford, 1993), 307, cf. 173-4. Cf. also Richard Sorabji’s discussion of the attitudes of a Stoic employer: things in accordance with nature ‘would include health and money both for yourself and for your workers’ {Animal Minds and Human Morals (Ithaca, 1993), 139). I do not mean to imply, though, that either Annas or Sorabji intends to adopt Smith’s model to the exclusion of the others I consider here. ‘Other-regarding’ may not be quite the right term here; but I take it that it

A Puzzle in Stoic Ethics

329

However, Smith’s model also raises some philosophical complica¬ tions worth bearing in mind. For one thing, there is perhaps some¬ thing not quite right about his talk of ‘two being preferable to one’, and Annas’s of ‘impartiality’.^*’ It is not just that complete impar¬ tiality is attained, if at all, only by the Sage, and that there seem to be some (perhaps many) contexts in which it is not called for at all (recall Seneca’s blithe selection of elegant clothes for himself, not to mention the elaborate Stoic theory of the differing duties we owe to kin, country, etc.). For I advance in oikeidsis when I come to think of others as oikeioi, adopting them into my house. It is not that the Stoic is impartial, counting himself as one and not more than one: rather, he identifies the interests of others with his own. So though Smith’s model manages to give oikeidsis-h^ised considerations the centrality they deserve, it seems to me to get them subtly wrong. Smith’s model gains a specious plausibility from its kinship with certain central notions of modern ethics: that morality is essen¬ tially linked to universalization and so to a kind of impartiality, that justice regulates the distribution of goods, that moral progress involves recognizing that there is no magic in the pronoun ‘my’. What standing these thoughts have in ancient philosophy, if any, is a difficult and complicated problem. But one particular respect in which we should be wary of importing modern assumptions is clear. We might be tempted to read ideas of impartiality into the crucial account in De finibus 3 of the evolution of appropriate ac¬ tion, according to which I move from an attachment to particular ‘things in accordance with nature’ to an attachment to reason as what is most natural for me, so that my interest is redirected from those objects of pursuit to the exercise of rational agency itself (3. 20-3). It is difficult for us not to imagine that sort of transformation as involving a powerful move towards impartiality: in grasping that I am above all a rational agent, surely part of what I must acquire

is not an objection to Smith’s model that it renders Stoic deliberation formally non-egoistic. As now seems to be generally accepted, ancient eudaimonism can incorporate a genuine taking into account of the interests of others; and I see no reason why this could not be expressed in formally non-egoistic deliberation, like that of a craftsman at work. For a systematic reading of Stoic virtue in terms of this (ultimately Platonic: cf. Rep. 345 E-347 a; Tim. 28 B-30 c) conception of craft moti¬ vation, see S. Menn, ‘Physics as a Virtue’, Proceedings of the Boston Area Colloquium in Ancient Philosophy, ii (1995), 1-34. Cf. B. Inwood, review of J. Annas, The Morality of Happiness, in Ancient Philosophy, 15 (199s), 647-65 at 661-3.

330

Rachel Barney

is a recognition that, objectively, as a rational agent I am no dif¬ ferent from any other, and that reason itself gives my interests no special standing. However, it is dear that none of this is in play in the transformation depicted in. De finibus] rather, what I attain in coming to value rational agency more highly than the indifferents is simply the right conception of my own happiness. So exactly where and how philanthropic and heroic motivations operate within Stoic deliberation remains, I think, to be identified.

IV A popular alternative solution is what I will call the Dualist Model. This involves the rejection of propositions (2) and (3) of the Max¬ imization Model above: not all actions are selections, if this im¬ plies that they take into consideration only indifferents qua indif¬ ferents. Rather, the Stoic’s deliberation is a two-tiered business: he must first decide whether ‘moral’ considerations demand any action under the circumstances, and only if not—if he turns out to be literally off duty—does he go through the deliberations of the Maximizer. Some of our texts do talk as though the selection of indifferents can at times be superseded by a different kind of agency. As Stobaeus explicates ‘selective value’, it is according to this that, 'when circumstances permit, we choose these particular things instead of those, for instance health instead of disease, life instead of death, wealth instead of poverty’ (ii. 83. 14-84. i (LS 580), my em¬ phasis). The Maximization Model can see in this merely an al¬ lusion to the all-things-considered nature of selection. But we could also read the relevant circumstances as including the ab¬ sence of any contrary ‘duty’—i.e. any of the appropriate actions prescribed by our social roles, human fellowship and oikeiosis, or other ‘moral’ considerations. The point of doing so, of course, is that the Dualist Model seems sufficiently elastic to give the right range of results. It allows what must be right if Stoicism is to be seen as a coherent system: that in some situations it is ap¬ propriate to select elegant clothes, and in others to die for one’s country. Something along the lines of the Dualist Model figures in a num¬ ber of scholarly interpretations. As Long and Sedley put it, the

A Puzzle in Stoic Ethics

331

locutions ‘to-be-taken’ and ‘to-be-selected’ express ‘the attitude a Stoic should adopt towards AN things [things in accordance with nature] which happen to be available . . . and which he can take or select without compromising his moral principles’ (LS i. 358)/° Likewise Inwood: ‘An adult continues to pursue those things which are preferred, but always in such a way that in case of a conflict with his pursuit of the good the impulse to the good will override his selection of the preferred thing’ {Ethics, 210)/’ Like Adam Smith’s reading, the Dualist Model has a strikingly modern air. In this case the prophetic aspect of Stoicism is its dis¬ covery of the dualism of practical reason: duty and inclination (or something resembling them) have their separate spheres, and the former speaks with authority. Though this may raise suspicions of anachronism, these should not, I think, be sufficient to disqual¬ ify this model. For the performance of ‘duty’ is still, in the Stoic case, governed by the eudaimonistic framework of the theory as a whole. Only by performing appropriate actions can I promote my ow'n happiness: so the Stoics, even so read, have arguably come no closer to treating the two spheres of practical reason as fun¬ damentally autonomous than Plato or Aristotle. Just as the ‘Adam Smith’ Stoic is only quasi-utilitarian, the ‘Dualist’ Stoic is at most pseudo-Kantian. So the charge of anachronism is not in itself a powerful ob¬ jection to this reading. A more serious problem is that, as Tad Brennan has convincingly argued, it is hard to see how a contrast between overriding ‘duty’ and the inclination to select preferred indifferents could be presented within any coherent and recogniz-

Cf. also Bonhoffer, Epictetus, 43: we must ‘weigh the values against one an¬ other . . . and, in the event a specific moral good is not at stake . . . prefer what is according to nature to what is contrary to nature’. ■" Presumably the ‘impulse to the good’ here includes a progressing Stoic’s ef¬ forts to perform appropriate actions. More recently, Inwood has presented this ‘overriding’ relation as being on a continuum with the way in which the value of one indifferent outweighs another: ‘They [the indifferents] are generally the object of an agent’s efforts and activities, although the value of pursuing preferred things can be overridden. ... It could be that the pursuit of wealth in a given case will turn out to impair other interests, such as the preservation of one’s health or the development of virtue. . . . Most important of all, some indifferents will tend to promote the acquisition of virtue and some will (at least sometimes) tend to hinder it; keeping in mind the ultimate importance of the good will aid with such choices’ (‘Stoic Ethics’, 694-5). This is perhaps closer to the Degrees of Nature model I consider in the next section.

Rachel Barney

332

ably Stoic model of deliberation/^ For the word we need to translate ‘ddty’ in this context is simply KaO-qKov, and correct selections of the indifferents are instances—indeed central, canonical instances—of KaOriKovra. So it is unclear under what descriptfon the Sage would perceive overriding duties as such. Nor do our sources hint at any such category. On the contrary; as I noted earlier, we have signifi¬ cant textual evidence for the claims that selections are exhaustive of actions and that considerations about indifferents are sufficient for all deliberations. To recall only one obstacle to the Dualist Model, the telos formulations of Antipater and Diogenes, which speak only of selection, would turn out to be radically incomplete, as descrip¬ tions of only one class of action—and the less salient one, by be¬ ing less distinctive of the Sage and the progressing Stoic, at that. Bonhoffer for one seems at times prepared to bite the necessary bullet and conclude that Diogenes and Antipater have left ortho¬ doxy behind: ‘What the Middle Stoa in a one-sided manner made the sole end of the human being, namely the rational selection of the things according to nature, in Epictetus has its correct position as a sphere of moral action beside others’ {Epictetus, 42-3).'*^ But it seems to me unthinkable that the doxographic tradition (including the polemic of the Stoics’ keen-eyed enemies) would fail to mark this heresy as such—unthinkable that Cicero could at Fin. 3. 31 treat the telos formulae of Chrysippus, Diogenes, and Zeno as if they were all one and the same. And in that case, the Dualist Model cannot be right either.

V One option remains. This is to reject the equations I introduced at the outset as thesis (i) of the Maximization Model: that is, to deny that ‘the things in accordance with nature’ (rd Kara cfivacv), as they figure in contexts related to ‘selection’, are to be identified with For Brennan’s arguments against the Dualist (or as he puts it, the Salva Virtute) Model see ‘Demoralizing’. At the same time, Bonhoffer seems to realize the implausibility of ascribing such a fundamental heterodoxy to these scholarchs: ‘even those older Stoics, when they . . . defined the telos one-sidedly as conduct according to reason in the selection of what is according to nature, still, exactly like Epictetus, must have considered the rational operation of the hegemonikon in fact as the supreme goal and must have delimited and regulated that rationality (evXoyiaTia) by means of the duties that are moral in and by themselves’ {Epictetus, 44).

A Puzzle in Stoic Ethics

333

the preferred indiflFerents. Perhaps we should say instead that the preferred indifferents are only a subset of the things in accordance with nature, and, in particular, that the appropriate actions enjoined on us by justice, by our social roles, and by the naturalness of human fellow'ship and oikeiosis are all more ‘in accordance with nature’ than any preferred indifferent/^^ Thus we may incorporate oikeiosis-based considerations directly into deliberation, in what I will term the ‘Degrees of Nature’ model: (i') For deliberative purposes, the preferred indifferents are a subset of the ‘things in accordance with nature’ and the dispreferred indifferents are a subset of the ‘things contrary to nature’. Other things in accordance with nature include all appropriate actions. {2') All actions can be described as selections among ‘things in accordance with nature’ and ‘things contrary to nature’. (3O All deliberations are about selections, and weigh things in accordance with nature and contrary to nature as such. (4'j A correct selection is one which produces the action most in accordance with nature. This improves on the Dualist Model in that it does not introduce a fundamental dichotomy between selections and other kinds of ac¬ tions, and can thus make sense of the telos formulae of Diogenes and Antipater. On the Degrees of Nature model, there is a unifor¬ mity to the Stoic’s deliberations: all aim at selecting what is most in accordance with nature given the circumstances. In some cases that will be the wearing of elegant clothes, in others the sacrifice of one’s life for one’s country; and moral progress will generally be a matter of coming to recognize the greater accordance-withnature of the weightier, ‘moral’-looking considerations at the top of the scale. Inasmuch as these weighter motivations derive from Stoic doctrines about social responsibility, human fellowship, and oikeiosis, this model can reach the same desirable results as the “ We might or might not want to allow that such things are, like preferred indif¬ ferents, bearers of ‘value’; I will hedge my bets on that question in what follows. Obviously for the Degrees of Nature model to be viable, we (and the delib¬ erating Stoic) must have sufficient information about what actions are appropriate, independent of the doctrine of the indifferents and their selection, for this to be a substantive and independent deliberative principle. A starting point would be Epictetus’ deduction of appropriate actions from our ‘names’ (i.e. identities and social roles) in 2. 10.

334

Rachel Barney

‘Adam Smith’ one. But formally it is more like the first strategy I considered, of treating moral progress itself as a preferred indif¬ ferent: like ‘value’ there, ‘accordance-with-nature’ here is used to provide a common denominator which will, it is hoped, always give the ‘right’ deliberative result. The Degrees of Nature model is proposed with reasonable clarity by Cicero in De officiis 3: Indeed, when the Stoics say that the greatest good is to live agreeably with nature, this means, in my view, the following: always to concur with virtue; and as for other things that are in accordance with nature, to choose them if they do not conflict with virtue. {Off. 3.13)

In spirit, this sounds rather like the Dualist Model, but it dif¬ fers crucially in the reference to other things {cetera) in accordance with nature; for this implies that action which ‘concurs’ with virtue is itself in accordance with nature in the same way as, albeit to a greater degree than, other potential objects of selection. Thus Cicero also claims that anyone who wants to live in accordance with nature ‘will never act so as to seek what is another’s, nor to appropriate for himself something that he has taken from someone else. For loftiness and greatness of spirit, and indeed, friendliness, justice, and liberality, are far more in accordance with nature than pleasure, than life, than riches’ (3. 23-4, my emphasis). Since we are by nature social and co-operative beings, nothing can be more contrary to nature for us than to commit injustice (3. 21, 26, 28, 35). We are not required to be relentlessly impartial: ‘It is permitted to us—nature does not oppose it—that each man should prefer to secure for himself rather than for another anything connected with the necessities of life’ (3. 22). However, we are forbidden to harm others in pursuing our own advantage, or to neglect the common good (3. 30). This position, as I have so far presented it, seems to me clear and coherent; and it has many advantages as a model of Stoic deliberation. Above all, it solves the problem of how to endorse Regulan behaviour, and incorporate oikeiosis-hdLsed considerations into deliberation, while upholding the orthodoxy of Diogenes and Antipater. However, there are several reasons to hesitate before tak¬ ing the Degrees of Nature model as a generally happy solution to our puzzle. First, it is explicit here that Cicero does not speak with the weight of a well-attested Stoic tradition behind him. On the

A Puzzle in Stoic Ethics

335

contrary, De officiis claims to be following Panaetius, and Cicero complains that Panaetius’ uncompleted work left off at precisely the crucial point, on the question of how we are to deliberate when the apparently honourable and the apparently beneficial conflict (3. 7-11,33-4)- Hence the rather unnerving ‘in my view’ {ut opinor) at Off- 3.13 above: Cicero here is largely on his own, working his way through problems on which the earlier tradition had not reached any clear verdict. A second striking feature of Cicero’s discussion is what seems to be, from the point of view of earlier Stoic doctrine, its considerable conceptual confusion. For Cicero seems to want the crucial con¬ cept of ‘benefit’ (utilitas) to mediate between moral and non-moral ‘goods’. In fact, his vehement and repetitive identifications of the beneficial and the honourable seem intended to make two claims at once: only what is honourable is beneficial for me (in the strong sense of contributing to my happiness) and only what is generally beneficial (in the weak sense of promoting non-moral advantages) is honourable (3. 19, 33—4, 40). These claims are not incompatible; but Cicero seems to have lost sight of the fact that they are quite independent. Likewise Cicero’s quotation of Antipater as saying, ‘Your benefit is the common benefit, and conversely, the common benefit is yours’ (Off. 3. 52). This is simply not true if ‘benefit’ is taken consistently in the weak sense, as the kind of advantage which an individual reaps from the possession of health or wealth; nor is it true if ‘benefit’ is consistently understood as the advantage which a Sage reaps from virtue, i.e. happiness. The use of the slogan can only be to conflate the two senses, the better to insist on the bonds between my happiness and the prosperity (as Smith put it) of my community. Cicero’s equivocation on the ‘beneficial’ is not just suggestive of conceptual confusion; it points to a fundamental problem with the Degrees of Nature model. That model is designed to ensure that considerations about the ‘honourable’ (what we would identify as ‘moral’ considerations) will always be decisive in the Stoic’s de¬ liberations, since to obey them is more in accordance with nature than anything else. So Cicero’s ambiguous use of the ‘beneficial’ simply tracks ‘accordance with nature’, which is, on this model, the salient normative property for deliberation, and common both to the preferred indifferents and to appropriate actions (cf. Off. 3. 35). Now this conception of the relation between moral and non-

336

Rachel Barney

moral ‘goods’ is not without its philosophical strengths, and is well attested in Hellenistic philosophy. Unfortunately, it is attested as the position of the Peripatetics-. In De finibus 3-4 the battle-line between Stoicism and the Peripatetic-Academid position champi¬ oned by Antiochus is drawn over this very question: whether the value of non-moral ‘goods’ is commensurable at all with the value of virtuous agency, so that the former can be seen as making some contribution, however small and easily outweighed, to the telos. And here Cicero, speaking for Antiochus, criticizes the Stoics for going to absurd lengths to preserve a merely terminological distinc¬ tion between themselves and the Peripatetics. I argued in Section I that the De finibus 4 critique of Stoicism can be answered; but it hits the mark against the post-Panaetian Stoicism presented by Cicero himself in De ojficiis 3.'*'^ For if ‘beneficial’ and ‘in accordance with nature’ apply to the indifferents and to moral action alike, without any distinction except one of degree, then the hierarchical model of the game or craft, in which only the indifferents figure as first-order objects of pursuit, while moral considerations operate on a different order altogether, has been given up. It is indeed hard to see how the Stoic position can be held apart from the Peripatetic in the absence of that model; and so it is easy to see why Stoic orthodoxy never avowed this solution to our puzzle.

Cf. Cic. Fin. 4. 39: ‘The inconsistency of the Stoics here causes me endless amazement. They determine that natural desire—what they call horme-—and appro¬ priate action, and even virtue itself are all things that are in accordance with nature. Yet when they wish to arrive at the supreme good, they skip over everything else and leave us with two tasks instead of one—to ‘adopt’ \sumamus\ some things, and ‘seek’ [expetamus^ others, rather than including both of them under a single end.’ Here Ci¬ cero, speaking for Antiochus, seems to describe the Stoics as including appropriate action and virtue among the ‘things in accordance with nature’: but his complaint is precisely that the Stoics none the less refuse to integrate them into deliberation, as the Degrees of Nature model would require. The complaint confirms that for the orthodox Stoa, the important conceptual task is precisely to hold apart the morally worthy and the merely preferable, insisting that no such comparisons of value can be made, no matter how reliably they favour virtue. The Stoics could perhaps avoid this collapse into the Peripatos by stipulating that only what in any situation is most in accordance with nature is genuinely in accordance with nature at all. The accordance-with-nature of any alternative option is only prima facie: given that theft is unjust, the consideration that it would increase my wealth is not just outweighed but rendered null and void. However, it is notable that even Cicero does not claim this, instead relying on the comparative language (‘more in accordance with nature’) I cited earlier in this section; and it would distort the concept of ‘accordance with nature’ almost out of recognition.

A Puzzle in Stoic Ethics

337

VI So how does a practising Stoic decide what to do ? All of the models I have considered here have their advantages: and so long as we do not bother to distinguish them clearly, it is easy to suppose that together they give us a reasonably good picture of how Stoic deliberation should proceed. But the models are not in fact compatible. Either every action is also a ‘selection’ or it is not; either the Sage selects the indiflferents for himself or he does so in some more agentneutral way; either motivations to select indifferents are trumped by motivations different in kind or they are not; either the phrases ‘preferred indifferents’ and ‘things according to nature’, in contexts relating to selection, denote the same class of objects or they do not. Until we have answers to these questions, I for one have little sense of what it would mean to act as a Stoic; and I do not think that clear answers are to be found in our texts. This unclarity can, I would speculatively suggest, be traced back to a fundamental conflict between two central strategies of Stoic ethics. One is the Stoics’ deep commitment to arguing that what matters for happiness is not the possession of non-moral ‘goods’ but the correct exercise of reason and virtue in relation to them. This is the thesis which the Stoics take from Plato as the central building-block for their argument that happiness is always within our power. The other is the Stoic determination to give no quar¬ ter to their Epicurean rivals in the battle to appropriate ‘nature’ as an ethical norm, particularly by way of the inborn self-benefiting tendencies brought out in the ‘cradle argument’."** The latter com¬ mits the Stoics to presenting their conception of virtuous agency as continuous with more obviously ‘natural’ behaviour, by depicting appropriate action as developing out of inborn drives and prefer¬ ences. The better to do so, however, they make a fatal shift from Plato: for the most part,"*’ they locate the task of rationality not in On which see J. Brunschwig, ‘The Cradle Argument in Epicureanism and Stoicism’, in M. Schofield and G. Striker (eds.), The Norms of Nature (Cambridge, 1986), 113-44. ■*'* Of course, the Stoics do sometimes talk of the correct use of the indifferents (e.g. D.L. 7. 103—4; S.E. M. ii, 61; Epict. 2. ^—6 passim), and this may well be the way to understand Chrysippus’ talk of the ‘material of virtue’ {vXrjv T-ijs aperrjs, Plut. Comm. not. io6g E, cf. 1069 D and 1069 f). But the relation of use to selection is never clearly stated, and it is the latter which comes to be the more prominent, above all in the telos formulae.

338

Rachel Barney

the correct use of the indifferents but in their correct selection—in our pursuing with insight objects continuous with those pursued spontaneously by the infant. As I argued in Section I, the Stoics have ari adequate defence against the standard charge that the resulting role assigned to the indifferents is unstable or incoherent. But this charge does land in the vicinity of a real, and indeed insoluble, problem. The game model I outlined in Section I makes good sense of the side of Stoicism engineered to compete with Epicurean naturalism, and shows how this could in principle be reconciled with a Platonic de¬ tachment from non-moral goods. For it shows how my ‘external’ or higher-order reasons for acting might become increasingly rich and authoritative while being non-revisionist in relation to action: my expanded insights into the ends served by a game do not revise the goals I pursue within it. The problem is that in the ethical case, the Platonic insight into the indifference of the non-moral goods surely does lead to their playing a radically different role in my life from the one they played before: and the discontinuity becomes very stark when we look at how the advanced Stoic is actually expected to behave. In practice, the serious adoption of Stoic higher-order ends must revise how I play the game of ‘selection’: a Regulus or Cato is no longer keeping score like everyone else. A closing suggestion. If we think of other ethical systems which, like Stoicism, have attracted sophisticated advocates and practition¬ ers over a number of generations, it is clear that the indeterminacy of Stoicism is far from exceptional. How, after all, does a utilitar¬ ian deliberate? There are almost as many answers as utilitarians, and with them comes a corresponding range of answers to the sub¬ stantive question of how a utilitarian should act—so much so that ‘utilitarianism’ has really come to name a whole genus of philosoph¬ ical positions. Hellenistic philosophical systems, with their revered founders, scholarchs, and other institutional pressures towards or¬ thodoxy, are sociologically very different creatures from modern ones.’® So where a modern philosophical movement fragments into theoretical diversity, we can expect the ancient one to be held to¬ gether by a kind of indeterminacy—what we might call constructive ambiguity. In the present case, the sheer complexity of Stoicism, with its tectonic plates of doctrine engineered in response to difCf. D. N. Sedley, ‘Philosophical Allegiance in the Greco-Roman World’, in M. Griffin and J. Barnes (eds.), Philosophia Togata (Oxford, 1989), 97-119.

A Puzzle in Stoic Ethics

339

ferent problems and traditions, serves to give it the practical versa¬ tility and breadth of appeal which can belong to modern doctrines only at the generic level. There is a Stoicism to help you wear ele¬ gant clothes and a Stoicism to help you give up the life-raft: best of all, you need never choose between them. University of Toronto

BIBLIOGRAPHY Annas, J., The Morality of Happmess (Oxford, 1993). Bonhoffer, A. E, The Ethics of the Stoic Epictetus [Epictetus^, trans. W. Stephens (New York, 1996; first pub., Stuttgart, 1894). Brennan, T., ‘Demoralizing the Stoics’ [‘Demoralizing’], Ancient Philo¬ sophy (forthcoming). -‘The Old Stoic Theory of Emotions’, in T. Engberg-Pedersen and J. Sihvola (eds.). The Emotions in Hellenistic Philosophy (Dordrecht, 1998), 21-70. Brunschwig, J., ‘The Cradle Argument in Epicureanism and Stoicism’, in M. Schofield and G. Striker (eds.). The Norms of Nature (Cambridge, 1986), 113-44. Cooper, J. M., ‘Greek Philosophers on Euthanasia and Suicide’ [‘Euthana¬ sia’], in id.. Reason and Emotion (Princeton, 1999), 515—41. Frede, M., ‘On the Stoic Conception of the Good’, in K. lerodiakonou (ed.). Topics in Stoic Philosophy (Oxford, 1999), 71-94. Inwood, B., Ethics attd Human Action in Early Stoicism [Ethics] (Oxford, 1985)-

-review of J. Annas, The Morality of Happiness, in Ancient Philosophy, 15 (1995), 647-65. -‘Rules and Reasoning in Stoic Ethics’, in K. lerodiakonou (ed.). Top¬ ics in Stoic Philosophy (Oxford, 1999), 95-127. -and Donini, P, ‘Stoic Ethics’, in K. Algra, J. Barnes, J. Mansfeld, and M. Schofield (eds.). The Cambridge History of Hellenistic Philosophy (Cambridge, 1999), 675-738. -and Gerson, C., Hellenistic Philosophy, 2nd edn. (Indianapolis, 1997). Kidd, I., ‘Stoic Intermediates and the End for Man’, in A. A. Long (ed.). Problems in Stoicism {Condon, 1971), 150—72. Lesses, G., ‘Virtue and the Goods of Fortune in Stoic Moral Theory’, Oxford Studies in Ancient Philosophy, 7 (1989), 95—128. Long, A. A., ‘Carneades and the Stoic Telos', Phronesis, 12 (1967), 59-90. -and Sedley, D. N., The Hellenistic Philosophers [LS] (2 vols.; Cam¬ bridge, 1987).

340

Rachel Barney

Menn, S., ‘Physics as a Virtue’, Proceedings of the Boston Area Colloquium

in Ancient Philosophy, ii (1995), 1—34. Sedley, D. N., ‘Philosophical Allegiance in the Greco-Roman World’, in M. Griffin and J. Barnes (eds.), Philosophia

(Oxford, 1989), 97—

119. Smith, Adam, The Theory of Mqral Sentiments, ed. D. D. Raphael and A. L. Macfie (Indianapolis, 1984; this text first pub. 1976; 6th edn. first pub. London, 1790). Sorabji, R., Animal Minds and Human Morals (Ithaca, 1993). Striker, G., ‘Following Nature: A Study in Stoic Ethics’, Oxford Studies in

Ancient Philosophy, 9 (1991), 1-73.

SEXTUS AND EXTERNAL WORLD SCEPTICISM GAIL FINE

I In

his Lectures on the History of Philosophy, Hegel argues that

ancient scepticism is more profound and all-encompassing than Cartesian scepticism and the scepticism of his day.' Lately, however, several commentators have argued to the contrary that ancient scep¬ ticism is less all-encompassing than Cartesian and post-Cartesian scepticism.^ One way in which ancient scepticism is thought to © Gail Fine 2003 Earlier versions of this paper were read at Dartmouth College, at the 35th Chapel Hill Colloquium, and at St Andrews University. I thank the audiences on these occasions for very helpful and friendly discussion, especially David Sedley, who was my commentator at Chapel Hill and who, as editor of Oxford Studies, provided yet further comments. Ancestors of parts of this paper were read at Corpus Christi College, Oxford; at a symposium at the Pacific APA; at MIT; at Stanford Univer¬ sity; at the University of Michigan; at a Keeling Conference at University College London; at the University of Canterbury in Christchurch, New Zealand; and at a conference on Hellenistic and early modern philosophy held at the University of Toronto (at which Alan Kim was my official commentator, and Myles Burnyeat my unofficial commentator). I thank the audiences on these occasions for helpful comments. Thanks too to Charles Brittain, Jacques Brunschwig, Benj Flellie, Terry Irwin, Sydney Shoemaker, and Nicholas Sturgeon for helpful discussion and/or written comments. ‘ Lectures on the History of Philosophy, ed. and trans. E. S. Haldane and F. FI. Simson (2 vols.; London, 1968)' ii. 331—2; cf. 347. In a similar vein, Reid, in his Inquiry into the Human Mind, characterizes Hume as a ‘semi-sceptic’ and ‘half¬ sceptic’, because he does not question the existence of ideas or impressions (ch. 5, § 7). He thinks the same is true of Descartes, for which see also ch. i § 3. Thanks to Sydney Shoemaker for the Reid references. I shall follow convention and use ‘Cartesian scepticism’ for the sort of scepticism that Descartes considers in, for example. Meditation I and the beginning of Meditation H. The phrase is strictly a misnomer, since Descartes is not a sceptic. ^ See e.g. M. F. Burnyeat, ‘Can the Skeptic Live his Skepticism?’ [‘Can’], in M. F. Burnyeat (ed.). The Skeptical Tradition (Berkeley, 1983), 117—48 at 118—19, and his ‘Idealism and Greek Philosophy: What Descartes Saw and Berkeley Missed’ [‘Idealism’], Philosophical Review, 91 (1982), 3-40 at 35—6. Cf his ‘The Sceptic in

342

Gail Fine

be less all-encompassing than Cartesian and post-Cartesian scep¬ ticism is that it allegedly does not question whether there is an external world. Myles Burnyeat, for example, says that ‘the Greeks never posed the problem of the existence of ap external world in the general form we havp known it since Descartes’.^ Indeed, it has been argued that ancient sceptics, so far from being External World Sceptics, are no more than what I shall call Property Sceptics:'' they assume that there is an external world housed with familiar sorts of objects; they even assume that I can identify what object I am confronted with on a given occasion. They question only whether such objects have the properties they appear to have. Burnyeat, for example, says: Ask Sextus what he means when he claims to suspend judgment about everything, and he will typically reply, ‘Well, take honey: it appears sweet to me but bitter to people with jaundice, and there is no criterion for deciding his Place and Time’, in M. F. Burnyeat and M. Frede (eds.), The Original Sceptics: A Controversy {InAi&VLdipoWs, 1997), 127-51. (This is an expanded version of a paper with the same title that was originally published in R. Rorty, J. B. Schneewind, and Q. Skinner (eds.), Philosophy in History (Cambridge, 1984), 225-54. I shall use the pagination of The Original Sceptics', this volume also contains Burnyeat’s ‘Can’.) See also M. Williams, ‘Descartes and the Metaphysics of Doubt’, in A. Rorty (ed.). Essays on Descartes’ Metaphysics (Berkeley, 1986), 117—39 at e.g. 118; S. Everson, ‘The Objective Appearance of Pyrrhonism’, in S. Everson (ed.). Companions to Ancient Thought, ii. Psychology (Cambridge, 1991), 121—47; R. J. Flankinson, The Sceptics (London: Routledge, 1995); and J. McDowell, in ‘Singular Thought and the Extent of Inner Space’ [‘Singular Thought’], in P. Pettit and J. McDowell (eds.). Subject, Thought, and Context (Oxford, 1986), 137-68. ^ ‘Idealism’, 23; cf 19: ‘Greek philosophy does not know the problem of proving in a general way the existence of an external world. That problem is a modern invention.’ '' My phrase ‘Property Scepticism’ is an adaptation of what Hankinson calls Es¬ sential, in contrast to Existential, Scepticism {The Sceptics, 25-6). I prefer my terminology since, at least on some versions of Property Scepticism, all properties, not merely essential properties, are at issue (though see n. 5). Notice that in this context being honey, for example, is not a property, though being sweet is. If some¬ one doubts, or suspends judgement about, whether she is tasting honey, or about whether the predicate ‘is honey’ is true of anything, she is what we might call an Object Sceptic, doubting whether a given object is present, or whether a certain kind of object exists. (These are themselves two different versions of Object Scepticism, a less and a more extensive one.) Hence Property Scepticism and External World Scepticism are not exhaustive options; moreover, there are different versions of each of them. I discuss some of these issues in ‘Scepticism, Existence, and Belief: A Discussion of R. J. Hankinson’s The Sceptics’, Oxford Studies in Ancient Philosophy, 14 (1996), 273-90 at 276. Various ‘grades of sceptical involvement’ are usefully dis¬ cussed by J. Barnes, ‘Ancient Scepticism and Causation’, in Burnyeat, The Skeptical Tradition, 149-86 at 159-60.

Sextus and External World Scepticism

343

which it really is. Likewise the tower appears round from a distance and square from close by. And so on. That’s how it is with everything.’ It is one and the same external thing, honey or the tower, which appears thus and so, and which has a real nature that the skeptic is unable to determine. (‘Idealism’, 29)®

And Stephen Everson writes: The subject-matter of the [ancient] sceptic’s investigation—and thus what he will find himself suspending judgement about—is what the honey is really like. What he does not question is whether the honey is really there. Whereas Descartes will acknowledge, indeed be certain of the fact, that his experience is such that honey appears sweet, but doubt that there is anything beyond that experience, the ancient sceptic merely suspends judgement as to whether the honey is sweet. Although he points to a distinction between appearance and reality, he does not, as the Cartesian sceptic does, argue that since knowledge is limited to the first half of the divide, we cannot say w’hether there is anything at all on the second. His doubt is just whether reality is like its appearance. (‘The Objective Appearance of Pyrrhonism’, 127)*

Is it true that Sextus is (typically) no more than a Property Sceptic? Or is he an External World Sceptic? Or is he neither of these sorts of sceptic? Is Hegel right to think that Sextus’ scepticism is more ex¬ tensive than Cartesian scepticism is? Or are modern commentators right to think that it is less extensive?’ * Burnyeat says that Sextus typically speaks in this way; others omit ‘typically’, and simply characterize Sextus as (what I call) a Property Sceptic. See e.g. the passage I quote next (from Everson). Though Burnyeat says that Sextus typically adverts to no more than Property Scepticism, he thinks that Sextus is committed to a more extensive scepticism. But he thinks the practical nature of Sextus’ scepticism prevents him from seeing what he is committed to: see ‘Idealism’, 29-31. I discuss this issue in ‘Descartes and Ancient Skepticism: Reheated Cabbage?’, Philosophi¬ cal Review, log (2000), 195-204. In saying that Sextus suspends judgement only about the ‘real nature’ of honey and the tower, Burnyeat makes it sound as though he thinks Sextus suspends judgement only about the essential properties of things, which would leave him free not> to suspend judgement about their non-essential properties, However, I do not think Burnyeat intends to limit Sextus’ alleged Prop¬ erty Scepticism in this way. For one thing, he says that Sextus suspends judgement as to whether the tower is round or square; but its particular shape is presumably not an essential property of it. But some of those who say that Sextus suspends judge¬ ment only about the natures of things sometimes seem to mean that he suspends judgement only about their essential properties. See e.g. M. Frede, ‘The Skeptic’s Beliefs’, in id.. Essays in Ancient Philosophy (Minneapolis, 1987), 179-200 at 187, repr. in Burnyeat and Frede, The Original Sceptics, 1-24, However, this is not his only or even, I think, his main view. I discuss Frede further in ‘Sceptical Dogmata: PH I 13’, Methexis, 13 (2000), 81-105.

‘ Lf, 136.

’ 1 do not have the space here to ask about all ancient sceptics, so I shall limit my

344

Gail Fine

Even if Sextus is not himself an External World Sceptic, he could none the less advert to it in, for example, describing someone else’s position; or it could be a view he entertains. If he does so, then Ex¬ ternal World Scepticism was at least discussed j^y ancient sceptics: in which case it was not invented by Descartes. Accordingly, I shall sometimes ask whether a given passage adverts to External World Scepticism, even if it does not describe Pyrrhonism.

II I have already said something about Property Scepticism: it as¬ sumes that the world is housed with familiar objects such as honey and towers, and that I can identify such objects on particular oc¬ casions; it suspends judgement only about what properties such objects have. Property Scepticism may be an unstable position. Suppose, for example, that I take myself to be confronted by an apple, but suspend judgement about whether it is sweet, juicy, red, round, or edible; I suspend judgment about all the properties it appears to have. How, then, can I be justified in thinking I am con¬ fronted by an apple? And if I suspend judgement about whether anything is sweet, juicy, red, round, or edible (or green or tart . . .), how can I be justified in thinking that apples exist? An extreme form of Property Scepticism—one that suspends judgement about all the properties of a thing—seems to lead to Object Scepticism, which suspends judgement about whether a given object exists, or about whether certain sorts of objects exist.* It might even lead to External World Scepticism: though whether it does so depends, among other things, on precisely how we characterize External World Scepticism, an issue I take up shortly.But even if someone scope to Sextus Empiricus, who is the main e.xponent of Pyrrhonian scepticism, one of the two main schools of ancient scepticism. In asking about Sextus’ scepticism, I am asking how he characterizes Pyrrhonism. I take no stand on what his epistemic stance to it is. In particular, I do not assume that he claims to know, or even believe, that scepticism of any particular variety is true: perhaps he suspends judgement on the matter. * See n. 4. Cf Hankinson, The Sceptics, e.g. 301-3. I discuss Hankinson’s views on this issue in ‘Scepticism, Existence, and Belief’. “ Hankinson seems to think that Property or Object Scepticism could lead to External World Scepticism only by means of a fallacious any/all slide; see The Sceptics, 302. I am not convinced that this is so; be that as it may, I largely focus here on other possible routes to External World Scepticism.

Sextus and External World Scepticism

345

who holds an extreme version of Property Scepticism is committed to a stronger sort of scepticism, it does not follow that she sees that she is so committed. And my question here is not whether Sextus is committed to more than Property Scepticism; my question is whether he in some sense explicitly endorses, or adverts to, a more extensive sort of scepticism. Let me now say something about External World Scepticism. More than one position has been housed under this rubric.'® The version I shall focus on asks whether there is anything external to one’s mind or to one’s present states of being appeared to." That, In his entry under ‘the problem of the external world’ in J, Dancy and E. Sosa (eds.), A Companion to Epistemology (Oxford, 1992), 381-6, for example, G. Pappas asks: ‘What, then, is the problem of the external world . . .? Certainly it is not whether there is an external world; this much is taken for granted’ (381). He goes on to describe two different problems of the external world: ‘(PEWi) Can one have knowledge of propositions about objects and events in the external world based upon propositions which describe how the external world appears, i.e. upon appearances?’ (381—2); and ‘(PEW2) Can one have knowledge of propositions about objects and events in the external world based upon propositions about directly perceived sensa?’ (383). If we answer ‘no’ to Pappas’s two questions, we get two versions of scepticism, each of which is less extensive than the sort of External World Scepticism that I shall focus on. Modern brain-in-vat scenarios are also less extensive than the sort of External World Scepticism I shall focus on. For according to these scenarios, for all I know I am a brain in a vat manipulated by, say, a mad scientist. Were this alleged possibility realized, I would be, or have, a brain (in which case something physical would exist); and there would be vats and scientists (and so something physical and external to me would exist). It is important to bear in mind that there are different formulations, or kinds, of External World Scepticism: for even if Sextus is not an External World Sceptic of one stripe, he might be an External World Sceptic of a different stripe. "

I take it that ‘external’ here means ‘independent’. It cannot mean ‘spatially

external’ since the mind is not being conceived as a material object: which is not to say that dualism is being assumed. It is worth bearing in mind that not all modern accounts of subjectivity and scepticism assume dualism. For a lucid explanation of this, see McDowell, ‘Singular Thought’, esp. §§ 6-8. In addition to, or rather than, asking whether there is anything other than one’s mind or one’s present states of being appeared to, one might ask whether there is anything other than one’s body; here ‘external’ presumably does mean ‘spatially external’ (or at least includes that notion). Even if ancient sceptics do not ask whether there is anything other than one’s mind or present states of being appeared to, they might ask whether there is anything other than one’s body. (Cf Everson, ‘I’he Objective Appearance of Pyrrhonism’, 146.) That would not be External World Scepticism of the sort I shall focus on here, though it would none the less be reasonable to count it as a sort of External World Scepticism; be that as it may, it certainly goes beyond Property Scepticism. Though External World Scepticism is often formulated in the way suggested in the text, in practice it often focuses on whether we can know, or justifiably believe, that there are external physical or material things. If, for example, Platonic Forms exist, they are external to—independent of—my mind. But External World

Gail Fine

346

however, is only part of the relevant ‘problematic’, as we might call it4(‘^ In addition, the mind must be viewed in the right way, as the locus of subjective states; alternatively, states of being appeared to must be conceived as subjective states. External World Scepticism also involves a ‘radically,first-person stance’.” The idea is that one has some sort of privileged a'ccess to one’s mental or subjective states, but thinks there is ‘an epistemological barrier between soul and body’, or between how one is appeared to and everything else, such that one suspends judgement as to whether there are any jus¬ tified inferences from claims about the former to claims about the latter.” External World Scepticism suspends judgement not only about whether any inferences from claims about one’s mind (or about one’s states of being appeared to) to claims about an external world are justified, but also about the existence of other routes to knowledge or justified belief about an external world. It does not, for example, mean to allow that we have knowledge of, or justified belief about, external things by direct inspection of them or through revelation. The idea is that we could know, or justifiably believe, that there is an external world only by way of an inference from claims about our own minds or states of being appeared to; since we must suspend judgement about whether any such inferences are justified, we must also suspend judgement about whether there is an external world. External World Scepticism therefore involves some sort of epistemological asymmetry; it privileges claims about the first per¬ son over other claims. I ask later precisely what sort of privileged access and epistemological asymmetry are required for External World Scepticism. But for now, this preliminary account will do.

Scepticism does not typically ask whether there are such things. On the other hand, when Descartes, in Meditation III, argues that God exists, he is arguing against something properly called External World Scepticism, even though he still suspends judgement about whether there is a corporeal world. Thanks to Calvin Normore for getting me to think about this issue. Thanks to Michael Martin for this way of putting the point. “ Burnyeat, ‘Can’, 145 n. 32. '■* The quotation is from Burnyeat, ‘Idealism’, 30 n. 39 (emphasis in the original). Some External World Sceptics, rather than suspending judgement about whether there is an external world, claim flat out that one cannot know whether, or justifiably believe that, there is an external world. Eor Sextus, this is a form of dogmatism rather than genuine scepticism: see e.g. PH i. 1-4. Hence if he is in some sense an External World Sceptic, he is not the sort who claims that knowledge or justified belief about the existence of an external world is impossible; he would suspend judgement.

Sextus and External World Scepticism

347

III I shall now consider five objections to the view that Sextus counte¬ nances External World Scepticism.In subsequent sections I ask whether these objections can be overcome. Objection i. Sextus’ phainetai statements are of the form ‘x ap¬ pears F' rather than of the form ‘it appears that x is F\ And, according to Stephen Everson, there is an important difference between ‘x appears F’ and ‘it appears that x is E’ Whilst the former presupposes the existence of x, the second does not. Framing their argument in this way [as ‘x appears F’], the sceptics challenge merely the attribution of properties to objects and do not question their existence. (‘The Objective Appearance of Pyrrhonism’, 136)'"

That is, the (alleged) fact that Sextus uses the locution ‘x appears F’, rather than ‘it appears that x is F’, shows that he assumes that there are external objects. Objection

2.

Sextus does not conceive of appearances in the right

way to be an External World Sceptic. In her seminal paper ‘The Ten Tropes of Aenesidemus’, for example, Gisela Striker writes: I suspect that it is only when we begin to think of appearances as entities intervening between the observer and the object of his observation that we are tempted to ask how we can say that there is one object, or any object at all. But while the skeptics obviously distinguished between the way a thing appears and the way it really is, I see no reason to attribute to them the view that this distinction must be made in terms of special entities—images or otherwise—mediating between observers and observed objects. ” These objections are far from exhaustive. But they are among the most common and important ohjections, and they fit together in ways that make it appropriate to focus on them in a single paper. , Cf. Hankinson: ‘Sextus is, indeed, quite happy with locutions of the form x appears to be F, which apparently at least entail x’s existence’ {The Sceptics, 25; emphasis added). See also Burnyeat, ‘Idealism’, 29 (quoted in sect. i). ” ‘TheTen Tropes of Aenesidemus’ [‘Ten Tropes’], in ead.. Essays on Hellenistic Epistemology and Ethics (Cambridge, 1996), 116-34, at 124-5 (originally published in Burnyeat, The Skeptical Tradition, 95-115). Striker uses ‘appearance’ to translate aiv6fj,€vov. I follow J. Annas and J. Barnes in using ‘appearance’ for (f>avraaLa, and ‘what is apparent’ for aiv6ixevov; see their The Outlines of Scepticism (Cambridge, 1994). (My translations of Outlines of Pyrrhonism [P//] generally follow those to be found in Annas and Barnes, though I have sometimes modified their versions without comment.) Burnyeat, Striker, and Frede, by contrast, all use ‘impression’

Gail Fine

348

Objection 3. External World Scepticism asks whether there is anything external to the mind. But Sextus does not use ‘external’ to mean ‘external to the mind’. Rather, according to Burnyeat, he uses it to mean

^

external to oneself, the cognitive^ subject, i.e. a man (cf. adv. Math. VII 167)—and the question is, ‘What does that come to?’ Sextus can contrast the external thing with the bodily humors which affect one’s perception of it (PH I 102) or with the medium through which it is perceived (tb. 12427), so it seems plain that the line is not drawn in Cartesian fashion between the mind and everything else outside it, including the skeptic’s own body. By the same token, ‘external’ in Sextus’ use of it imports no Cartesian (Augustinian) break between things outside and an inner (subjective) world of things apparent. (‘Idealism’, 29)**

Objection 4. According to Burnyeat, Sextus is not ‘afflicted with radical Cartesian doubt as to whether he has a body’ (‘Idealism’, 30).'’ But doubting, or suspending judgement about, whether one has a body is necessary for being an External World Sceptic. Objection 5. Sextus does not view the mind, or states of being appeared to, in the way necessary for being an External World Sceptic. He lacks a concept of subjectivity or, at least, a sufficiently robust one. And in so far as he thinks there are subjective states, he does not accord us the right sort of privileged access to them, or favour the right sort of epistemological asymmetry.^® for (f)avTaata\ but that misses the connection with aLveo9ai. Many authors agree with Striker that Sextus does not conceive of alveTaL {fJ.ev) rffiLv yXiiKa^eiv to fjceXt (tovto avyxcopovfxev yXvKa^6fj,e0a yap alaOrjTLKO)?), el Se Kal yXvKV eoTiv oaov irrl rw Xoycu, ^yroCp-eu' o ovk eari TO atvoju.evov aAAa (to) irepl tov cf>aivop,evov Xeyopievov.

Translation i. For instance, honey appears to us to

sweet. We allow this,

since we are perceptually sweetened. But we doubt if it is sweet as regards its definition; this is not the appearance, but something said about the appearance.

Translation z. For example, it appears to us that honey sweetens (we concede this inasmuch as we are sweetened in a perceptual way); but whether (as far as the argument goes) it is actually sweet is something we investigate—and this is not what is apparent but something said about what is apparent.

This important passage raises many issues. The crucial point here is just that Translation i takes the first sentence to be of the form ‘x appears F’, whereas Translation 2 takes it to be of the form ‘it appears that x is F”. Both translations are justifiable.^^ However, it is true that, in many of the relevant passages, phainetai statements are most naturally taken to be of the form ‘x appears F’. But that, by itself, does not show that Sextus presupposes x’s existence. Macbeth asks: Is this a dagger which I see before me. The handle toward my hand? . . . ... or art thou but A dagger of the mind, a false creation. Proceeding from the heat-oppressed brain? I see thee yet, in form as palpable As this which now I draw.^^

(n. i. 33-41) He is not assuming that there is a real dagger there; nor, of course, is there a real dagger there. The ‘dagger’ he mentions is what we translation is controversial); the second is Annas and Barnes’s (emphases added to both translations). This is because ‘honey’ (to ;aeAi) is neuter, and so it has the same form in both the nominative and accusative. This makes it unclear whether it falls within, or outside, the scope of (f>alveTaL. I am not suggesting that either translation is completely satisfactory in every respect. The crucial issue here is just the translation of (f/alverai. statements. " Accompanying Le Rive, a painting by Puvis de Chavannes in the Musee d’Orsay, are the following words: ‘II voit dans son sommeil I’Amour, la Gloire et la Richesse lui apparaitre’. These too are presumably ostensible objects only.

Sextus and External World Scepticism

351

might call an ‘ostensible object’, in something like the sense defined by H. H. Price: The object of any one act of perceptual consciousness is as such ostensible only, and has, as it were, aprmia facie character. It ‘claims’ to be real and to have certain characteristics, and it may in the end turn out to have them; but equally it may not.^'‘

Moreover, it is far more natural to say ‘I see a dagger’ than it is to say, for example, ‘I am appeared to as though confronted by a dagger’ or ‘It appears to me that I am confronted by a dagger’.“ And Sextus tells us that he does not like to fuss about words. Contrary to Objection i, then, it is not clear that Sextus’ phainetai statements should always be taken to be of the form ‘x appears F’ rather than of the form ‘it appears that x is F’. But even if they should always be so taken, that by itself does not commit him to the existence of external objects, for x might be an ostensible object. Of course, even if Sextus’ use of the locution ‘x appears F' does not, on its own, show that he assumes that x exists, his use of that locution in a given context might show that he does so. And it has been argued that the Ten Modes of Aenesidemus are such a context.^’ In a fuller paper, I would argue that the Ten Modes H. H. Price, Perception (London, 1932), 148. In a note, he adds; ‘I here mean by “object” only “object-of”, i.e. that which is before the mind in a particular sort of consciousness. There is no reason why this should not be unreal—unless “consciousness” be equated with “knowing”, which seems an unfortunate use of language. So also we speak of the object of hope or of desire; plainly, only that which is not real can be desired or hoped for’ (148 n. i). One might also mention here the notion of an intentional object. See e.g. G. Harman, ‘The Intrinsic Quality of Experience’, in id.. Reasoning, Method, and Mind (Oxford, 1999), 244-61 (originally published in Philosophical Perspectives, 4 (1990), 31-52). Ostensible objects are a subclass of intentional objects; the ones that, as Price puts it, ‘claim’ to be real. Cf. R. Chisholm, Theory of Knowledge (Englewood Cliffs, NJ, 1966), 34; ‘Let us consider another way of describing these self-presenting states. In our examples, “appears” requires a grammatical subject and thus requires a term that purports to refer not merely to a way of appearing, but also to a thing that is said to appear in that way. However, we may eliminate the reference to the thing that appears if we convert our appear-sentences. Instead of saying “Something appears white to me,” we may say, more awkwardly, “I am appeared white to by something.” We may then eliminate the substantival “something” by merely dropping the final clause, saying, “I am appeared white to.”’. ‘We say too that we do not use phrases strictly, making clear the objects to which they are applied, but indifferently and, if you like, in a loose sense—for it is unbecoming for a sceptic to fight over phrases’ [PH i. 207; cf i. 191); cf. Plato, Theaet. 184 c. See e.g. Everson, ‘The Objective Appearance of Pyrrhonism’, 136; Hankinson,

Gail Fine

352

do not in the end assume that there are external objects; thinking they do so ignores their concessive strategy. The Ten Modes are, in any case, just one weapon in the Pyrrhonist’s arsenal; even if, as in the Ten Modes, Sextus assumes that there qre external objects, it does not follow that he always does so. There is also the question of whether the Ten Modes are purely ad hominem. If they are, then even if they assume that there are external objects, it would not follow that Pyrrhonists ever assume that there are external objects.

V I now turn to Objection 2, which says that External World Scep¬ ticism would not be tempting unless one viewed appearances as intermediate entities. I turn in a moment to the question whether this assumption is correct. But first it will be helpful to ask whether Sextus ever countenances intermediate entities. He may seem to do so in PH 2. 72-5:^* Even if we grant that appearances [^avraatat] are apprehended, [external] objects cannot be judged in virtue of them. For the intellect, as they say, sets itself upon external objects and receives appearances of them, not through itself, but through the senses, and the senses do not apprehend The Sceptics, 157. Contrast L. Groarke, Greek Scepticism: Anti-Realist Trends in Ancient Thought (Montreal, 1990), 129-30. In this passage Sextus uses ‘appearances’ ((/lavraaiai) and ‘affections’ (irddr]), but not ‘what is apparent’ {(iyaivo^eva); see n. 17. (^avraalai are a subclass of Trddrf, and I take it that they are the only sort of 170.67] that are relevant here. Though appearances can, and in Sextus often do, range outside the perceptual sphere, in the present passage he is discussing only sensory appearances. I use ‘being sweetened’ and ‘being bittered’ to translate ykoKo^eadaiand 7rucpd^ear0at respectively. As Sextus’ next sentence makes clear, these terms refer to affections, in particular, to states of being appeared to: being affected sweetly (bitterly) is seeming to taste something sweet (bitter). The terminology is Cyrenaic. My understanding of it, though not unusual, is controversial. (See e.g. n. 63 below, where I mention Frede’s alternative understanding.) Cf M. 7. 352—3: ‘Furthermore, since, according to most philosophers, there is in us not only a thinking but also a perceiving part, the latter, being set in front of the thinking part, will necessarily prevent thought from grasping external objects. P'or just as the body which lies between sight and the object of sight prevents sight from grasping the object of sight, so if the non-rational sense of sight intervenes between thought and the external object of sight, sight will prevent thought from grasping the external object of thought . . . Thought, then, being locked away inside, and being kept in the dark by the senses, will not be able to grasp any of the external objects.’ The whole of M. 7. 352-8 is worth looking at in this connection.

Sextus and External World Scepticism

353

external objects but only—if anything—their own affections [TraOrj]. An appearance, then, will actually be of the affection of a sense—and that is different from an external existing object. For honey is not the same as my being sweetened, nor is wormwood the same as my being bittered; they are different. And since this affection is different from the external existing object, an appearance will not be of the external existing object, but of something else different from it. So if the intellect judges in virtue of appearances, it judges badly and not in accordance with the existing object. Hence it is absurd to say that external objects are judged in virtue of appearances. Nor can we say that the soul apprehends external existing objects through its sensory affections [-nddr]] inasmuch as the affections of the senses are similar to the external existing objects. For how will the intellect know whether the affections of the senses are like the sense-objects, given that the senses make clear to it not the nature of these [external] objects, but their owm affections, as we deduced in the modes of suspension? Just as someone who does not know Socrates but has looked at a picture of him does not know whether the picture is like Socrates, so the intellect, study¬ ing the affections of the senses but not observing the external objects, will not know whether the affections of the senses are like the external existing objects.

Sextus’ comparison of appearances, or sensory affections, with pic¬ tures suggests that he is treating them as intermediate entities. The idea is that we have access to pictures of things, but not to the things themselves. Pictures are entities that interpose themselves between us and the external world.So, contrary to Objection 2, a case can be made for the claim that Sextus sometimes describes appearances as intermediate entities. But I do not want to press this suggestion. For one thing, in addition to comparing appearances to pictures. The similarity with Berkeley is striking. Just as Sextus suggests that we have direct access only (or at most) to our sensory affections, so Berkeley says that we immediately perceive only our own ideas. Like Sextus, Berkeley then argues that we do not know that there are external material objects through perception, since each of us perceives only our own ideas. He also argues that neither do we know that there are external material objects through reason. In the course of arguing this, he considers but rejects the suggestion that we can infer that there are external objects from the (alleged) similarity they have to our ideas. See Principles of Human Knowledge, Principles 18 and 19 (cf. 8). There are also, however, significant dif¬ ferences between Sextus’ argument and Berkeley’s. For example, unlike Sextus, Berkeley argues in his own right that ideas cannot resemble external objects, and that material things cannot exist. M. 7. 352—3, cited in the previous note, also seems to conceive of appearances as intermediate entities: for they are compared to bodies that interpose them¬ selves between sight and external objects, thereby preventing us from perceiving the latter.

Gail Fine

354

Sextus also speaks of being sweetened (bittered). This seems to treat appearances as states of being appeared to: to be sweetened (bittered) is to be appeared to as though tasting something sweet (bitter). i Even if Sextus means to describe appearances as states of being appeared to, rather than as entities (in a narrow sense), he clearly describes them as ‘things’ (in a broad sense) that block our access to external objects. For he says that the senses do not apprehend external objects, but only (‘if anything’, he carefully says in 72, but the qualification is omitted in 74) their own affections. He also says that any access the intellect has to external objects will be not direct, but through the senses. But, he argues, no inferences from one’s sensory affections to external objects are justified. So sensory affections, whether they are states or entities, are viewed as blocking, rather than as enabling, cognitive access to external objects. One might argue that in order to find External World Scepticism tempting, one must view appearances as entities; viewing them as states will not do. In my view, however, it is not necessary for being an External World Sceptic that one view appearances as intermedi¬ ate entities rather than as states: I reject the assumption on which Objection 2 rests. Adverbial theorists, for example, do not posit intervening entities; they posit states of being appeared to.^° But an adverbial theorist can be just as robust an External World Sceptic as someone who views appearances as intervening entities. All she needs to do is argue that one is aware only of how one is appeared to, and that there are no justified inferences from claims about how one is appeared to, to claims about anything else. What is crucial is not the ontological status of appearances—whether they are en¬ tities or states—but whether one has direct access only to them, so that all other claims can be justified only by inference from claims about them. So even if Sextus does not view appearances as in¬ tervening entities (as opposed to states) that block our access to external things, he might none the less find External World Scep¬ ticism tempting.^' For a classic statement of an adverbial theory, see R. Chisholm, Perceiving: A Philosophical Study (Ithaca, 1957). It is true that 2oth-cent. adverbialists have not typically been External World Sceptics; indeed, the position is often advocated in reaction against sense-data theories that seemed to give rise to External World Scepticism. But my point is just that an adverbialist is not prevented from being an External World Sceptic; one could consistently hold both views. ” Not only is countenancing intermediate entities not necessary for being an

Sextus and External World Scepticism

355

One might put my argument in the form of a dilemma: if one thinks that, in order to be an External World Sceptic, one needs to view appearances as intervening entities as opposed to states, one is wrong; hence even if Sextus never so describes appearances, that would not show that he is not in a position to consider External World Scepticism. If, on the other hand, one thinks states will do, one is right—but Sextus so describes appearances, and so he is in a position to find External World Scepticism tempting. To be sure, PH 2. 72—5 is at least partly ad hominem\ hence it is not clearly evidence of what views Sextus himself in some sense endorses or finds tempting. But it shows that he has the conceptual resources for considering External World Scepticism, in so far as he is aware of a view on which appearances function as intermediate ‘things’ (whether entities or states) that block our access to external objects. So far I have argued that Sextus sometimes describes appearances {(^>avTaalai) as intermediate ‘things’ (whether entities or states) that block our access to external objects. It is worth noting that he is also aware of a view on which (/>atvO(U.em are not external objects, but affections. In M. 7. 193-4, in the course of discussing the Cyrenaics, he writes: Hence we must posit as what is apparent either affections [TraOrj] or what produces affections. And if we say that affections are what is apparent, then we must say that everything that is apparent is true and apprehensible; whereas if we call what produces affections what is apparent, everything that is apparent is false and inapprehensible. For the affection which occurs in us reveals nothing more than itself.

Here Sextus contrasts two ways of conceiving of rd ^atrdp,eva: as Trddrj (which in this context I take to be subjective states of being External World Sceptic, but neither is it sufficient. For one might countenance such entities, but think one has a knock-down argument for the claim that there are external objects, or think it is obvious that there are such objects. For the view that positing intermediate entities (as opposed to states) is neither necessary nor sufficient for being an External World Sceptic, see J. Greco, ‘Modern Ontology and the Problems of Epistemology’, American Philosophical Quarterly, 32 (1995), 241—51; and id., Putting the Sceptics in their Place (Cambridge, 2000). “ Sextus’ target in PH 2. 72-5 is, or at least includes, the Stoics. They seem to view (j>avraalaL as states that enable (rather than preclude) perception of external objects; they favour a form of direct realism. Sextus argues that they are committed to treating (j>avTaaLai as intermediate ‘things’ that preclude apprehension of external objects.

Gail Fine

356

appeared to),” and as what produces them (presumably external objects—or, perhaps better in this context, ostensible external ob¬ jects). According to the Cyrenaics, -nddri are the only (fyaivofxeva; in their view, external objects are not (j>aLv6^^va. Now, if x is a (^atvojuevov, it is clear, and something we are aware of non-inferentially. So in arguing that Trddr] are the only

(^aivofieva,

the Cyrenaics

imply that we do not have direct access to external objects. This fits with the view expressed in PH 2. 72-5, according to which the most we can apprehend directly are (j^avraalaL or nddr]: that is, at most they are (/>atvoju.eva. On both views, we lack direct access to external things. One might argue that we perceive them indirectly. But the Cyrenaics argue, more strongly, that Trddrj reveal only them¬ selves. Similarly, we have seen that in PH 2. 74 Sextus says that the senses make clear to the intellect only their own nddr] (though cf. 72: ‘if anything’). In both places, ^avraalai or ndd-q block our access to external things; they do not enable us to perceive them.

VI I now turn to Objection 3, according to which Sextus does not use ‘external’ to mean ‘external to the mind’, and so does not view his body as external. An interesting passage in Against the Ethicists {=M. ii) 79-89 (cf. PH 3. 183-90) undermines this objection.” Sextus is in the course of arguing {ad hominem) that nothing is good or bad by nature. He argues that if anything is worth choos¬ ing for its own sake (as the good is alleged to be), it is either the act of choice itself or the object of that act. In 81-2 he argues against the first alternative. He then divides the second alternative into two: any object of choice that is worth choosing for its own sake must be either ‘separate from us’—that is, external (e’/crd?) to us”—or some¬ thing ‘relating to us’ (vrept rjiJids, 83). He then argues that if a separate object of choice is intrinsically choiceworthy, it either has an effect on us {avyL^alvei ti nepl riij.ds) or it does not. If it does not, it is not " I defend this interpretation, and discuss the Cyrenaics in more detail, in ‘Sub¬ jectivity, Ancient and Modern’, and in ‘The Subjective Appearance of Cyrenaic Pathe . Thanks to Charles Brittain for calling my attention to this passage. It is well discussed by R. Bett, Sextus Empiricus: Against the Ethicists (Oxford, 1997), 10713 (cf appendix A, pp. 260-1). Both my translation and my discussion are indebted to his.

” I assume that Kat in 83 is epexegetic.

Sextus and External World Scepticism

357

intrinsically choiceworthy after all; if it does, it is the effect, not the object, that is intrinsically choiceworthy. He describes the relevant effect as ‘a civilized motion

[dareiov

Klv-qiiaY

‘a welcome condi¬

tion [oiTToSeKTov KardoTrifMaY, and ‘an agreeable affection [dyaarov nddos]’. He seems to be describing subjective states under, as we

might put it, a subjective mode of presentation: being civilized, welcome, and agreeable are subjective or psychological properties; they are not wholly and only physical or objective properties.” By 87, Sextus has argued that nothing external to us is chosen for its own sake. He then argues that neither is anything relating to us intrinsically choiceworthy: Nor, however, is that which is to be chosen and good among the things relating to us. For this belongs either just to the body, or to the soul. But it could not belong just to the body. For if it belongs just to the body, and is no longer also an affection [irddos] of the soul, it will escape our awareness [yvaicu?] (for all awareness is on the part of the soul), and it will be equivalent to [t'ffov] things which exist externally and have no affinity with us. (87)

First Sextus says more about what it is for something to ‘relate to us’: either such things belong just to the body (that is, they are not identical to and do not constitute any psychological state); or else they are psychological states (even if such states are identical to or are constituted by physical ones). Here, then, he views the body as something relating to me, in which case it is not external. However, he goes on to qualify the remark: he says that if the thing (allegedly) relating to me is just bodily, and not at all psychological, then it is equivalent

(loov)

to things separate from us and so has no affinity

with us. That is to say, every purely bodily state of myself is, for present purposes, external to me. Sextus also explains why purely bodily states of myself are equi¬ valent to something external to me: because ‘it will escape our awareness [yvcuai?] (for all awareness is on the part of the soul)’. “ I follow Bekker (who, in turn, is followed by Mutschmann and Bury) in adding aaretov before Klvr]fi.a; cf. PH 3.

184. Contrast Bett. I agree with Bett that the mere

fact that PH 3. 184 has da-Tcior does not by itself mean that dareiov should be included here. However, since (rardaTijiu-a and tt6,9os are both qualified by adjectives, it seems reasonable to think that so too is Ktvrjfxa. Be that as it may, nothing I say depends on including aareLov. ” Which is not to say that the states at issue are not also physical: one and the same state can be both subjective and physical. I discuss this further in ‘Subjectivity, Ancient and Modern’ and in ‘The Subjective Appearance of Cyrenaic Pathe'.

358

Gail Fine

This seems to say that since states of awareness are psychological, one can only be aware of—and so can know only^*—what is psy¬ chological. That is a peculiar argument. But whether or not that is his argument, he seems to conclude that all one pan be aware of are one’s own mental states, under a subjective mode of presentation; one is aware of them only in so far as they are mental states.^’ In 88 Sextus says: if the pleasing effect which it has extends as far as the soul, it will be something to be chosen and good as far as that is concerned, but not in so far as it is a movement merely of the body. For everything which is to be chosen is judged to be so by way of sensation or thought, not by way of an unreasoning body. But the sense or intelligence which grasps that which is to be chosen belongs by its very definition to the soul; so none of the things which happen to the body is to be chosen for its own sake and good, but, if any, those which happen to the soul.

Sextus seems to assume here that the relevant affections are not only psychological but are also bodily. None the less, the bodily component or aspect is viewed as external. He seems to be distin¬ guishing between his affections conceived as psychological states and conceived as physical states; and he argues that it is at most the former that are choiceworthy. This suggests that, at least at this stage, he leaves certain forms of materialism open;'^'’ but he insists that the material constitution of one’s mental states, if there is one, is irrelevant to their being choiceworthy. For he is at the moment interested only in subjective states under their subjective mode of presentation. Something is intrinsically worth choosing, at this stage of the argument, only in so far as it is a psychological state of oneself, so described. Even if my subjective states are identical David Sedley (in conversation) objects that the mere fact that we lack yvdiai? of something does not mean that we do not know it; to lack yvwais of x is to lack direct acquaintance with it; but we might know x in some other way. However, Sextus— here as elsewhere—seems to assume (if only ad hominem) that we must suspend judgement about any inferences from what is directly accessible to anything else. Hence if we do not have yvcoCTij of x, we suspend judgement about x altogether. Similarly, we saw that in PH 2. 72 Sextus says that if the senses apprehend anything, they apprehend only their own -rradr]. These claims are reminiscent of the Cyrenaics, who argue that each of us has KaTaX-qijiis only of our own affections (which, in my view, are, and are viewed by the Cyrenaics as being, subjective states); see e.g. PH

I.

215; M. 7. 190-200. I discuss the former passage briefly below; I discussed

part of the latter passage briefly at the end of the previous section. What Sextus says is, for example, compatible with non-reductive materialism, but not with eliminative materialism.

Sextus and External World Scepticism

359

to or are constituted by material states, the latter are (or should for present purposes be viewed as being) external to me/‘ In M. II, then, Sextus explicitly uses ‘external’ for being exter¬ nal to the mind; and he explicitly counts his body as external. We should therefore be open to the idea that that is also how he under¬ stands externality elsewhere. To be sure, Sextus does not (so far as I know) elsewhere explicitly say that one’s body is external. This can, however, easily be explained. For in M. 11 Sextus is work¬ ing with a familiar tripartite division of goods into external goods, bodily goods, and goods of the soul. In this context, one’s body would not normally be classified as external; hence Sextus needs to make his understanding of externality clear.But that special eth¬ ical understanding of externality is not relevant in epistemological contexts; so perhaps in such contexts Sextus does not feel the need to spell out his understanding of externality. None the less, a given context might make it clear that he takes his body to be external, even if he does not make the point as explicitly as he does in M. 11. Let us, accordingly, look at some further passages. In PH 1. 15 (an important passage that I explore further below), Sextus says: TO Se fiiyiarov, iv Trj Trpoac/iopa twv rfxjjvdtv tovtcov to iavTW (j)aLv6p.evov Xeyei Kal TO

irddos dirayyeXXeL

to eavTOV

dSo^daTCos, pLrjSiv irepi

tu)v

viroKeipievcov

StajSe/SatoOjixevo?.

The main point is this: in uttering these [sceptical] phrases, he [the sceptic] says what is apparent to himself and reports his own affections without holding opinions, affirming nothing about external objects.'*^ This is very like Descartes’s position in Meditation II. Descartes says he is certain of his psychological states so described; for example, he is certain that it seems to him that he sees light. But (in Meditation II) he suspends judgement as to whether these states are physical. He says, for example: ‘And yet may it not perhaps be the case that these very [bodily] things which I am supposing to be nothing, because they are unknown to ipe, are in reality identical with the “I” of which I am aware? I do not know’ (AT vii. 27/CSM ii. 18). Hence at this stage Descartes takes himself to be certain of a state under a subjective mode of presentation, while suspending judgement as to whether that state is physical. Thanks to David Sedley here. As my account in the text makes clear, I take Kal to be epexegetic: to report what is apparent to oneself just is to report one’s affections. One’s affections, that is, are precisely what is apparent to one; cf i. 203. I take it that the relevant af¬ fections here are states of being appeared to. Hence, what is clear to one is how one is appeared to. In sect, vill I present further reasons for thinking that this is Sextus’ view. PH I. 15 uses

e^wdev;

M. ii uses

Iktos.

I assume that this is mere stylistic

variance, i. 15 is the first place in PH in which Sextus explicitly mentions what is

Gail Fine

360

To say what is apparent to oneself is to report one’s affections: in particular, how one is appeared to. How one is affected—that is, how one is appeared to—is contrasted with ‘external objects’. The contrast is surely meant to be liot only exclusive but also exhaus¬ tive. Presumably here, then, ‘external’ means ‘external to how one is appeared to’. Now, being e^fternal to a given state of being ap¬ peared to is different from being external to the mind: the class of things that is external to a given state of being appeared to is if anything larger than the class of things that is external to the mind. In particular, as we shall see in Section VII, one can suspend judge¬ ment about whether there are minds or souls without suspending judgement about whether there are states of being appeared to. Be that as it may, since Sextus says that everything other than how one is appeared to is external, he is committed to viewing his body as external. One might argue that Sextus does not see what he is committed to: perhaps he did not have the conceptual resources for treating his body as external; or perhaps the idea of treating it as external was for some other reason just not a salient option for him. We have seen, however, that Sextus not only has the conceptual resources for viewing his body as external, but also explicitly describes it as external. It therefore seems reasonable to assume that he sees what I.

15 commits him to. In Sections VII and VII we will find further

evidence for this view. Sextus tells us that i. 15 makes his main point. He is telling us, at the beginning of his enquiry, how he is to be understood. If externality here means being external to a given present state of being appeared to, there is a presumption that that is his of-

external. With i. 15, compare i. 208: ‘Besides, you must remember that we do not use these [sceptical] phrases about all objects universally, but about what is unclear and investigated in dogmatic fashion, and that we say what is apparent to us and do not make firm assertions about the nature of externally existing things.’ I assume that states of being appeared to are not being conceived as physical: which, again, is not to say that dualism or any form of immaterialism is being assumed. See the first paragraph of n. 10. In Skepticism (London, 1968), A. Naess suggests that Sextus standardly understands externality as being external to a given present state of being appeared to (17). In The Skeptic Way (New York, 1996), B. Mates says that Sextus ‘plainly’ uses ‘external’ for what is ‘external to our minds or souls, and not necessarily external to our bodies’ (i8). He adds, however, that Sextus is sceptical about the coherence of the concepts of soul and body, and suggests we should perhaps speak instead of a distinction between reporting ‘that something appears to one to be the case and flat-out asserting that it is the case’ (21).

Sextus and External World Scepticism

361

ficial understanding of externality. If other passages understand externality differently, we should infer that Sextus is being loose or colloquial, or explaining someone else’s view.^^ In fact, though Sextus does not always adhere to his official usage,'*'’ he often does so. Consider, for example, the conclusion of the Third Mode of Aenesidemus: If the senses do not apprehend external objects, the intellect is not able to apprehend them either (since its guides fail it), so by means of this argument too we shall be thought to conclude to suspension of judgement about external existing objects

[rrepl

rchv

.Iktos

inTOKeLfjLevwv]. (PH i. 99)

If the senses do not apprehend external objects, what do they ap¬ prehend? Sextus addresses this question in PH 2. 74, which we looked at earlier. He says there that he deduced in the Modes (i.e. in the Ten Alodes of Aenesidemus) that ‘the senses make clear to [the intellect] not the nature of [external] objects but their own affections’. Here ‘external’ again means ‘external to a given pre¬ sent affection (i.e. appearance)’. We apprehend (if anything) only how we are appeared to at a time. We do not apprehend anything In this connection, it is worth noting that post-Cartesian philosophers do not always clearly use ‘external’ for being external to the mind rather than for being external to the body. Berkeley, for example, sometimes seems to conceive of exter¬ nality (he uses ‘without’) in terms of physical distance. See e.g. Prin. 42-4, and near the end of the First Dialogue (Philosophical Writings, ed. D. Armstrong (London, 1965), 164—6). There are also passages where he speaks of perception as being done, not just by the senses, but by specific bodily sense organs such as the eyes (and so by bodily parts); see e.g. First Dialogue, p. 165 Armstrong; Prin. 29 and 35. (Of course, he might say in reply that sense organs are nothing but ideas. The point is just that he is not always careful to speak in that way.) Thanks to Nick Sturgeon for the point and the references. Flume also uses ‘external’ not only for being external to one’s mind, but also for being external to one’s body. See A Treatise of Human Nature i. 4. 2, entitled ‘Of Scepticism with regard to the senses’. For example, at pp. 190—1 Selby-Bigge (Oxford, 1967), in asking whether the senses account for the belief that objects are external, he seems to use ‘external’ for being external to one’s body. But the existence of one’s body is itsplf in question. So he should be using ‘external’ for being external to the mind, where ‘external’ means ‘independent’. Yet Hume speaks as though externality and independence are distinct. Here I am indebted to Sydney Shoemaker. Just as Berkeley and Hume, if read out of context, might not be thought to have the relevant notion of externality, so too with Sextus. But attention to their official usages makes it clear what notion they have in mind. We should not focus on their casual uses of ‘external’ in order to determine their official understanding of it. See e.g. PH 1. 102 (where bodily humours seem to be treated as internal) and 126 (where the membranes and liquids in our eyes are treated as internal). Cf Burnyeat, ‘Idealism’, 29 (cited above in connection with Objection 3). Interestingly, in

I.

125 the colour of our skin is treated as external.

Gail Fine

362

external to such affections; and so we suspend judgement about everything external to such affections. At 2. 74 Sextus says that this is the general moral of the Ten Modes; again, then, it is not an isolated use of‘external’.^ I conclude that Sextus has the conceptual resources for using ‘external’ to mean both' ‘external to the mind’ and ‘external to a given present state of being appeared to’; indeed, he uses it in both ways: in the first way at M. 11, and in the second way at PH i. 15, in explaining his official understanding of externality. He also has the conceptual resources for viewing his body as external: indeed, he explicitly so views it in M. 11. Nor is that passage aberrant. On the contrary, our discussion of PH i. 15 suggests that on his official position, one’s body counts as external.

VII According to Objection 4, Sextus never suspends judgement about whether he has a body. So far as I know, there is no passage in which Sextus explicitly says ‘I suspend judgement about whether I have a body.’ What, however, should we make of this interesting empirical fact? Should we conclude that Sextus does not suspend judgement about whether he has a body, either because he lacks the conceptual resources for doing so or because, though he has the If this is the moral of the Modes, we should hesitate before concluding (as some people conclude: see n. 27) that they assume the existence of external objects. Is this understanding of externality one that Sextus accepts in his own right? If not, it is not evidence of his own views, though it would still be evidence of views he is aware of. On the one hand, we have seen that PH 2. 72—4 is at least partly ad hominem. But PH i. 15 is not. I have argued that PH 1. 15 uses ‘external’ for being external to a given present state of being appeared to. Externality also seems to be understood in that way in the Modes and in PH 2. 72—4. Hence that aspect of these latter passages is not merely ad hominem, even if other aspects of them are. One might object that Sextus would be inappropriately dogmatic were he to accept this division—between how one is appeared to, on the one hand, and everything external to that, on the other—in his own right. In ‘Appearances and Impressions’, Phronesis, 37 (1992), 283-313, for example, Rachel Barney says that the distinction between ‘internal subjective experience and external world’ is ‘theory-laden’ (304), and so should not be attributed to Sextus. She is more sanguine than I am about there being ‘ordinary’ views that are not in any sense theory-laden. Nor is assuming some sort of distinction between how one is appeared to and everything external to that as divorced from ordinary life as she takes it to be. (Not that I think Sextus is as wedded to /3(o? as he is sometimes taken to be.) Certainly operating with some such distinction does not require one to have a detailed theory of subjectivity; and I have not ascribed the latter to Sextus.

Sextus and External World Scepticism

363

conceptual resources, he none the less does not do so? Or does he in some way indicate that he suspends judgement about whether he has a body, even though he does not say this explicitly? We have already seen some reason for thinking that the latter is the case. For in the previous section we saw that Sextus seems to view his body as external; and he says that he suspends judgement about everything that is external. If so, he suspends judgement about whether he has a body. I shall now reinforce this suggestion from another angle. \r\PH 3. 38—46 Sextus considers several concepts (irrlvota, evvoia) of body, and argues that all of them are defective; he concludes that (so far as the concept of body is concerned)"** body is inapprehen¬ sible. The underlying thought is that if we do not have a clear account of what body is, bodies are not apprehensible. That is, we cannot know, or justifiably believe, anything about bodies, in¬ cluding whether there are any. In 3. 47-9 he pursues a somewhat diflferent strategy. He says that according to his dogmatic oppo¬ nents, everything that exists is an object either of perception or of thought; so if there are bodies, they are objects either of perception or of thought. But they cannot be either. Hence (so far as this ar¬ gument goes) there are no bodies (48). On the other hand, ‘bodies seem to appear real’

{tm (fyalveadai [SoKetv] vTrdpxov to acojjia)

(49).

Hence there are conflicting appearances as to whether there are any bodies. Sextus tacitly assumes here, as he often does elsewhere, that the appearances are, or seem to be, equipollent. And so sceptics suspend judgement about whether there are any bodies (3. 49).*’ If Sextus suspends judgement about whether there are any bodies, he presumably suspends judgement about whether he has a body. ■** For an excellent discussion of how to understand this and related phrases, see J. Brunschwig, ‘The hoson epi t6{i) logd{i) Formula in Sextus Empiricus’, in id., Papers in Hellenistic Philosophy (Cambridge, 1994), 244-58, originally published as ‘La formule hoson epi top) logo{i) chez Sextus Empiricus’, in A. J. Voelke (ed.), Le Scepticisme antique (Cahiers de la Revue de theologie et de philosophie; Geneva, Lausanne, and Neuchatel, 1990), 107-21. According to Everson, PH 3. 38-55 and M. 9. 359-440 are ‘concerned mainly with causing problems with definitions and accounts of matter’ (‘The Objective Appearance of Pyrrhonism’, 126 n, ii). And, one might think, even if one cannot offer a satisfactory definition or account of matter or body, one need not suspend judgement about whether there are any bodies. Cf. Striker, ‘Ten Tropes’, 124. Of course, Sextus may not in his own right accept the view that body is apprehensible only if there is a satisfactory account of body. But he seems to rely on that view here, if only ad hominem. Cf. Plato, Meno 71 a: if one does not know what virtue is, one does not know anything at all about virtue. In any case, at PH 3. 46—9 Sextus does not rely (just) on difficulties about accounts of body. He deploys an argument from conflicting appearances, in which the inadequacy of such accounts is at most a part.

Gail Fine

364

One might argue that PH 2- 49 shows only that Sextus is commit¬ ted to suspending judgement as to whether he has a body; it does not show that he is aware that he is so committed. However, Sextus is well aware that if one suspends judgement abjout a general prin¬ ciple, one likewise suspends judgement about everything that rests on it. In M. 9. 1-3, for example, he says that instead of engaging in unduly prolonged counter-argument against the Stoics, he will attack the most important and most comprehensive of their doctrines, since in the doubts cast on these we shall find the rest also included. For just as, in a siege, those who have undermined the foundation of a wall find that the towers tumble down along with it, so too in philosophical investigations those who have routed the primary assumptions on which the theories are based have virtually abolished the apprehension of every particular theory.’”

PH 3 also opens with this very point (3. i). It therefore governs the discussion about body that we have just considered. And thus it seems reasonable to assume that Sextus sees that in suspending judgement about whether there are any bodies, he is committed to suspending judgement about whether he has a body. He chooses, reasonably enough, to focus on the general and stronger claim (sus¬ pending judgement as to whether there are any bodies) rather than on the weaker and more particular claim (suspending judgement as to whether he has a body). This explains why he does not trouble to mention his own body as an example of something about whose existence he suspends judgement. One might argue that the passages just discussed concern only scientific concepts of body, and that one can suspend judgement about the coherence of such concepts without suspending judge¬ ment about whether there are any bodies in an everyday, ordinary sense. This line of argument might appeal to those who think that Cf. PH 2. 84. The similarity with Descartes is striking.

In Meditation I

Descartes says that ‘it will not be necessary for me to show that all my opinions are false, which is something I could perhaps never manage . . . So, for the pur¬ pose of rejecting all my opinions, it will be enough if I can find in each of them some reason for doubt. And to do this, I will not need to run through them all individually, which would be an endless task. Once the foundations of a building are undermined, anything built on them collapses of its own accord; so I will go straight for the basic principles on which all my former beliefs rested’ (AT vii. 18/ CSM ii. 12). Though Descartes and Sextus are making the same point, and though both mention foundations—of a building and of a wall, respectively—Descartes’s metaphor is architectural, whereas Sextus’ is military.

Sextus and External World Scepticism

365

Sextus ‘insulates’ ordinary claims from scientific ones/' This is a large issue that I cannot discuss here. So I shall just say that I agree with Burnyeat when he says: Every statement making a truth-claim falls within the scope of scientific investigation because, even if the statement itself is not at a theoretical level, it will still use concepts which are the subject of theoretical spec¬ ulation: concepts such as motion, time, place, body. If these concepts are problematical, which Sextus argues they all are, and no line is drawn be¬ tween philosophical and empirical doubt, the original statement wall be equally problematical. (‘The Sceptic in his Place and Time’, 115; empha¬ sis added)”

A diflferent but related objection is that Sextus is not discussing body as such, whether scientific or everyday, but only a special, limited concept (or a limited range of concepts) of body, as that notion figures in the distinction between corporeal and incorporeal entities. Surely one can suspend judgement about whether there are any bodies in this special limited sense, while still thinking one has a body in some other sense?” It is true that, for at least much of the discussion, Sextus focuses on extended corporeal body.” But that is the sort of body whose existence is relevant to External World Scepticism about bodies. If one says that one has a body, but then adds that bodies are just collections of ideas, one has not asserted that one has a body in a sense that is incompatible with External World Scepticism about bodies.” Moreover, in PH 2. 29-30 Sextus says: See e.g. Frede’s ‘The Skeptic’s Beliefs’; and his ‘The Skeptic’s Two Kinds of Assent and the Possibility of Knowledge’ [‘The Skeptic’s Two Kinds of As¬ sent’], in id.. Essays in Ancient Philosophy, 179-200, originally published in Rorty, Schneewind, and Skinner, Philosophy in History, 255-78, and reprinted in The Ori¬ ginal Sceptics, 127—51. I use the pagination of his Essays. ” I discuss insulation briefly in ‘Descartes and Ancient Skepticism’. ” Thanks to David Sedley for raising this objection; the discussion that follows is indebted to him.

»

At PH 3. 38, however, he says that ‘Some say that body is that which can act and be acted upon. But so far as this concept goes, bodies are inapprehensible.’ This notion of body is quite broad. ” It is true that Berkeley, for example, both takes bodies (in one sense of the term) to be collections of ideas and claims not to be a sceptic. But the plausibility of his claiming not to be a sceptic about the existence of body is much disputed. If sceptics either suspend judgement about whether something exists, or claim that it cannot be known whether something exists, then there is a sense in which Berkeley is a sceptic about body, since he says in Principle 20 that ‘if there were external bodies, it is impossible we should ever come to know it’. However, he also argues that body, in the sense of matter or corporeal substance, cannot exist. So there is a sense in which

366

Gail Fine

Even if we grant by way of concession that humans can be conceived of, we shall find that they are inapprehensible. For they are composed of soul and body; but neither bodies nor souls perhaps are apprehended; nor, therefore, are humans.

This suggests that sceptics suspend judgement, not just about whether there are bodies in some limited special sense, but also about whether there are any human bodies. In PH 2. 29-30 Sextus suspends judgement not just about whether human body is apprehensible, but also about whether soul is apprehensible. Similarly, in 3. 186 he suspends judgement about whether soul exists, just as, in PH 3. 49, he suspends judgement about whether there are any bodies. This might seem problematic for two reasons. First, we now seem to lose any asymmetry be¬ tween soul and body. Yet did we not say that some such asymmetry is required for External World Scepticism? Secondly, if Sextus sus¬ pends judgement about whether there are any bodies or souls, he might seem to suspend judgement about whether he exists; and that might seem problematic. Let us consider the second problem first. Contrary to what it suggests, one can suspend judgement about whether there are any bodies or souls without suspending judgement about whether one exists. Descartes does just this at the beginning of the Second Medi¬ tation: I have convinced myself that there is nothing in the world, no sky, no earth, no minds, no bodies. Does it now follow that I too do not exist? No; if I convinced myself of something then I certainly existed. (AT vii. 25/CSM ii. 16-17)

He goes on to say; But I do not yet have a sufficient understanding of what this ‘F is, that now necessarily exists. (AT vii. 25/CSM ii. 17) he thinks it can be known that body does not exist. That is no longer scepticism, in the sense in which I just explained it, though it is scepticism if we expand the label to include denying that something exists. (Berkeley himself asks how to define scepticism in, for example, the beginning of the First Dialogue.) Even if Berkeley should be counted as a sceptic about the existence of body, it does not follow that he is an External World Sceptic tout court. For he thinks there are other spirits, including God (though whether he is entitled to think this is another question). Flere, however, my main concern is scepticism about the existence of body. See also the last paragraph of n. ii. Thanks to Nick Sturgeon for helpful discussion about Berkeley.

Sextus and Eixternal World Scepticism

367

That is, Descartes is certain thathe exists, even though he claims not to know what he is. And since he says he does not yet know whether there are any minds or bodies, he does not yet know whether he is or has a mind or a body. Hence he suspends judgement about whether there are any minds or bodies without suspending judgement about whether he exists.'^* Perhaps Sextus does so as well. Let us now turn to the first objection: that if one suspends judge¬ ment about whether there are any souls or bodies, one loses the sort of asymmetry that is required for External World Scepticism. Here my reply is that it is not true that one cannot be an External World Sceptic if one suspends judgement not only about bodies but also about souls. For even if one suspends judgement about whether there are any bodies or souls, one need not suspend judge¬ ment about how one is appeared to; and if one does not suspend judgement about that, but does suspend judgement about every¬ thing else, that provides the sort of asymmetry that can sustain the External World Sceptic. Consider, for example, Hume, who, in a famous passage in the Treatise on Human Nature (1.4. 6), writes: For my part, when I enter most intimately into what I call myself, I always stumble on some particular perception or other, of heat or cold, light or shade, love or hatred, pain or pleasure. I never can catch myself at any time without a perception, and never can observe any thing but the perception. When my perceptions are remov’d for any time, as by sound sleep; so long am I insensible of myself, and may truly be said not to exist. And were all my perceptions remov’d by death, and cou’d I neither think, nor feel, nor see, nor love, nor hate after the dissolution of my body, I shou’d be entirely annihilated, nor do I conceive what is farther requisite to make me a perfect non-entity. If any one, upon serious and unprejudic’d reflection thinks he has a different notion of himself, I must confess I call reason no longer with him. All I can allow him is, that he may be in the right as well as I, and that we are essentially different in this particular. He may, perhaps, perceive something simple and continu’d, which he calls himself; tho’ I am certain there is no such principle in me.

Hume does not doubt that he has perceptions, though he doubts whether there is anything ‘simple and continu’d’ which constitutes himself. In just the same way, Sextus can suspend judgement about To be sure, he goes on to argue, or discover, that he is a thinking thing; and he equates that with being a mind: ‘I am, then, in the strict sense, only a thing that thinks; that is, I am a mind, or intelligence, or intellect, or reason’ (AT vii. 27/CSM ii. 18). However, he does not affirm this until after he has discovered that he exists; he claims to discover that he is before he makes any claims about what he is.

Gail Fine

368

whether he is or has a soul, without suspending judgement about whether he exists and has perceptions or states of being appeared to. And, interestingly enough, though he questions whether there are any souls, he never seriously questions whether there are states of being appeared to.” This may be why, in PH i. 15 (which we looked at in Section VI), as often elsewhere, he distinguishes, not between the mind and everything external to it, but between how one is appeared to and everything external to that. If this latter asymmetry is spelt out in the right way, it is adequate for External World Scepticism.” This in turn gives us a further reply to Objection 3, according to which an impediment to viewing Sextus as an External World Sceptic is that he does not use ‘external’ to mean ‘external to the mind’. We have seen that he does so use ‘external’ in M. ii. But we can now' also say that even if he never used ‘external’ in this way, that would not necessarily be an impediment to viewing him as an External World Sceptic. For even if he did not speak of be¬ ing external to the mind, he could speak of what is external to a given present state of being appeared to; and, as we have seen, he does speak in this way. Notice, in this connection, that I initially formulated External World Scepticism disjunctively, in terms of suspending judgement about what is external either to the mind In PH

I.

19 Sextus says: ‘And if we do propound arguments directly against

what is apparent, it is not because we want to reject what is apparent that we set them out, but rather to display the rashness of the dogmatists. For if reasoning is such a deceiver that it all but snatches even what is apparent from under our very eyes, surely we should keep watch on it in unclear matters, to avoid being led into rashness by following it?’ Sextus is here speaking in propria persona, and he tells us that he does not question what is apparent {^aivofxeva). Now as we have seen (n. 17), it is often thought that he generally uses ^au'd/xeva for external objects. But I do not think that is how he is using it here; for if he were, he would be saying he does not question whether there are external objects (contrast Hankinson, The Sceptics, 302). In ‘Sceptical Dogynata’ I argue that in i. 19-22 he takes i^in5>A !y r3K?*5'\‘'J5 5'V'i'\\ bc•' ■

mV .i-o-'f

.luiJ 'f hjdzfldiw-; -r .■

itnib'', siijjAiH

/vnt

■ ■■■'!

1 n'E

VI-Tf 1 ■-

'■

■ -

■ ’-,

'

■ .‘, ■



'

'

J

5 -

T

,

' ■ ■ ■ ** '

t.iin'i-T

,1/ , .-t ,:i;;V/

ri«

-9^ ■■

: J

t.

.-;'• '

*•'

--

.

f- ■ .



; (.. .

I ■•

,

.1’I- '*•-« ■' »• ifii** r»i ^

I

'

.

.

'•

,

. .

; , .

*i' J I.' ■ ‘.* 'T'>,

; I





- '

. tt

INDEX LOCORUM

Aetius 31 A 72=5.

De anima 19. 5 ff. Diels, Dox. Gr.:

45, S3 n. 99

,b

404 : 3 n. 9 404|’4-6: 33 n. 70 411 29:

Alcinous

154-5 n. 37

4I2'’i8-i9:

Didaskalikos

154 4i3“3i-2: 165 n. 47

28. 3: 281-2 n. 19

413 24-32:

156 n. 39

4i3*’24-3i:

154-S n- 37 4i3'=27-9: 146

Alexander of Aphrodisias

4i3‘’32-4i4‘‘i:

De anima 2.

168 n. 55

4i4*’20-5: 178 n. 64

164; 314 n. 23

4iS‘“23-5:

De fato, ed. Bruns

177

4iS‘‘27: 178

ch. 13: 254, 268

4i6“i3-i8:

181. 21-2: 268 n. 38

4i6*’28-9:

181. 26-7: 268 n. 39

4i6’’29:

181. 26: 270 n. 42

168 n. 55

4i7‘’5-9:

181. 28-30: 268 n. 40 In Aristotelis Analytica Priora, ed.

4i7‘’7:

151 n. 33

158

420*2-9:

Wallies 180. 33—6: 259 n. 14, 259-60 n. 18 In Aristotelis Metaphysica, ed. Hayduck

432*8:

148 148 n. 27

168 n. 55

168 n. 55

433*18:

168 n. 55

433‘’28:

168 n. 55

434*30: 168 n. 55

387. 32-388. 7: 221-2 n. 18

434“32-434‘’i: "79 De caelo

Alexander Lycopolis

279*17-18: 260-1 n. 22

Contra Manicheorum opiniones disputatio

301*14-18: 41 n. 82, 48 De generatione animalium

19. 2-4 Brinkmann: 257

Aquinas

.

In Aristotelis Metaphysica

715*1-18:

135 n. 5

715*9-11:

136 n. 9

7i5‘’4fT.: 183-4 n.

i

717*17-19: 205-6 n. 30

7.

202

721*2-10:

7.

202

12. 8, § 1402 12. 8, § 1403 7. 12. 8, § 1453 7. 12. 8, § 1455

202

722'*6-3o: 26 n. 57 722'’3o-3: 149

201-2

723‘’3-9: 183-4

7. 12. 8, § 1457

203

726‘’5 IT.:

183-4 n. I

n.

I

187 n. 8

73i‘’8-i8:

183-4 n. I

Aristotle

731'’28-31 :

Cate go ries

732‘’ii-i5 : 183- 4 n. I

177

1*6-12 : I 19 n. 89 14*29- 35: 2 10 n 4 14*^10- 22: 2 10 n 4

732'’is: I 67 n. 50 734'’8 ff.: 187 n. 8 734'’20 ff.: 187 n. 8

388

Index Locorum

Aristotle: De ■ generatione animalium (cont.): 735“2-4: 186 737“iQ-23: 187 n. 8 738'’io-i8: 187 n. 8 739'^20-9: 200 n. 23 \ 74o‘’32-3: 187 74o‘’35-74i‘‘2: 187 n. 8 742^19-22: 225 n. 22 742'"i7-743'‘i; 169-70 n. 56 743^27-34: 187 n. 8 743“35-6: 183-4 n. i, 200 744“8-i8: 164, 169-70 n. 56 744'’16-27: 164 7S8‘’6-8: 183-4 n. i 7S8'’2i-4: 183-4 n. i 759‘7- 183-4 n. I 762“9-I3: 200 762“ 12-13: 200 762‘'i3-i5: 199 762*19-34: 198-9 n. 19 762*26: 198, 205 762*35 ff.: 200 y62%—g: 199 n. 21, 200 n. 23 762'’8: 200 n. 23 762’’i4-i6: 200 766‘’7-i5: 187 n. 8 767*17-18: 187 n. 7 767^15-20: 187 n. 7 767'*i6-2i: 187 n. 8 768*34-7: 187 n. 8 768'’i2-I4: 187 n. 8 77o'*i7: 186 n. 5 777'’i6 fF.: 203 n. 26 783*18-29: 198 789‘*3-i3: iQS n. 17 De generatione et corruptione 315*19-20: 46 n. 90 327'*22—31: 224 n. 20 327‘’23-7: 140 n. 15 47 334*5: 14 n. 35, 20 334'*2i-3: 216 n. 12 33S‘‘9-23: I9S n. 17 335'’20-4: 193 De partibus animalium I. i: 137 n. 10 639'*I5-I9: 137 n. 10 640*14: 168 n. 53 640*16-19: 137 n. 10 640*19-33: 187 n. 7 640*22-7: 187 n. 8 640*27-35: 205 n. 29 334^5-7:

640*31: 195 64o‘’4-i6: 195 n. 17 641*4-6: 143 64i*’24—6: 187 ^ 64i’’27-642*2: 187 n. 8 642*2-13: 136 n. 8 642*14-17: 163 n. 46 643*32: 156 643*35-*’!°: 167 n. 50 646*13-24: 224 n. 20 655*23-7: 142 n. 19 658*8: 166 n, 48 662*16-23: 162 n. 44 663*33: 160 n. 42 663‘’22-5: 187 n. 7 663’’22; 171 n. 57 664*5-7: 160 668*’33-669*i: 142 n. 19 676**6-io: 198-9 n. 19 683*20-6: 162 n. 44 686*22: 166 n. 48 687*11: 166 n. 48 679*’15-31: 198 694*13-15: 206 n. 31 De sensu 441'*24-442* 10: 199 n. 20 442*7-8: 200 n. 22 442*26-30: 199 n. 20 De somno 453’’24-454“i9: 217 n. 13 454‘’23-455*3: 217 n. 13 Historia animalium 491*7-14: 168 n. 53 491*15-19: 167 547‘’i8fT.: 204 551*1 fr.: 204 569*10-15: 199 570*7-12: 199 590*19-27: 195 n. 17 Metaphysics 984'*5-io: 195 n. 17 989*15-18: 225 n. 22 1000*15-18: 165 n. 47 iooo“24-'*2o: 35 1000*26—'*3: 34 ioi8'’30-7: 221-2 n. 18 ioi8'’37-ioi9*2: 221-2 n. 18 1019*1-4: 210 1019*2-14: 212-24 1019*2-4: 213, 221-2 n. 18, 227, 236 1019*3-4: 213

Index Locormn ioi9“4-ii: 213-21, 214, 218-19 n. 15, 219, 221 ioi9“8-9: 214, 224-5 n. 21 ioi9“9-io: 215 ioi9‘‘io-ii: 215, 218 ioi9“ii-i2: 212 n. 8, 221-2 n. i8, 236 ioi9‘‘i2-i4; 213, 214, 221-2 n. 18, 222-4, 223, 225, 227, 23s, 237 ioi9“i4: 224-5 n. 21 i023'’i9-23: 194 n. 15 i032“24-5: 186 n. 5, 187 n. 8 1032*35: 186 io32'’6-9: 196 i032'’io; 196 I032‘’ii-i4: 194 I032'^i4: 194 n. 14 i032'’2i-7: 197 i032’’32-i033*i: 195, 196 io33'=26-io34*5: 191 i033'’29-3o: 193 1034*10-13: 196 io34‘‘i5-i6: 196 1034*18-21: 197 1034*18-19: 197 1034*19-21: 197 1034*22: 193 1034*23: 194 i034‘’4-7: 197, 199, 200, 202 i034'’7: 192 n. 12 io34’’i6-i9: 192 n. 12 I034'*32: 218-19 n. 15 1036*23: 218-19 ri- 15 i039‘’i-4: 149 n. 30 1040*1-5: 149 n. 30 io4o‘’5-8: 218—19 15 i04i*34-‘’9: 144 n. 21 io44‘’29-io45*6: 217 n. 13 i044'’29-34: 217 n. 13 io44‘’34-io4S’‘6: 217 n. 13 io47‘’35-io48*2: 230 1048*2-5: 230 1048*5-7: 230 1048*8-10: 230 io48*37-io48‘’6: 217 n. 13 io48'’8-9: 232-3 io48'’i8-24: 149 n. 29 1048*^20-1: 149 n. 30 io48'’29: 150 n. 31 1048*^33-4: 15° n. 31 io48'*37-io49*i8: 216, 231 1049*1-5: 231 n. 27 I049*’i2-i7: 209

389

1049*’17-1050*3: 209 io5o“4-*’6: 209 n. i, 212, 226, 235, 236, 237 io50*4-'*2: 235, 236 1050*4-11: 210 1050*4-9: 226 1050*4-7: 211, 236 1050*4-5: 236 1050*5-7: 225-30, 226, 229 1050*5-6: 210, 212, 227, 237 1050*6-9: 226 1050*7-10: 225 i050*io-'*2: 227, 230-4 1050*10-14: 226, 230, 232 1050*10-11: 232 1050*15-23: 226, 230, 231 1050*15-16: 231 i050*23-*’2: 227, 230, 232 1050*23-8: 233 io50*28-‘’2: 233 i050*’2-4: 236 io5o*’6ff.: 236, 237 io50*’6-io5i*2: 209 n. i, 235, 236, 237

io5o'*6-24: 236 i05o'*i9: 235, 237 1070*9-26: 184 1070*26-30: 184, 190 io77*’2-3: 210 n. 3 Meteorologica

354'’33-355‘‘2i: 165 n, 47 379*8-9: 199 379*22-3: 199 379“23-4:

i99

379*25-6: 199 379‘’7-9: i99, 200 388*10-20: 100 388*14: 100 388*18: 100 Nicomachean Ethics

1096*23-9: 177 n, 63 i097’*25-33: 177 n. 63 iii2‘’i7-2o: 137 1115*9: 89 n. 45 1115*17-22: 112 n. 75 1115*18-19: 112 n. 75 1115*20-2: 112 n. 75 11 i7’’23-i I i9'’i8: 67 n. 8 I i44‘’i-i 145*2: 112 n. 75 ii44‘’4: III n. 71 ii44'’6: III n. 71 ii44'’8-9: 112 n. 75 ii44'’i6-i7: 112 n. 75

Index Locorum

39°

Nicomachean Ethics (cont.):

Aristotle:

ii44'’i9-2o: ii44'^28-9:

hi

n.

112

ii44‘’32-4:

III,

113

ii44'’32-3:

III

III

n.

n. 45

n.

45

m

p.

165

Aphr. In An. Pr.

6 Wallies: 259 n.

n. 72

139

i92'*i8-i9:

II, 139

166

I94'’i4:

203

i98'’io-i6: i99’’iS-i7:

n.

26

195

n.

163

310 n.

ap.

Epict. Diss. 2.

ap.

Plut.

n.

6. 9:

26

Cicero n.

Academica

46

2. 76

380 n.

26o'’i6-i9:

210 n. 4

2.

258

225 n.

De finihus

265*22-4:

6. 9—10:

22

225 n. 22

8S

80

3- 14--15: 325-6 n. 35

Prior Analytics

3- 16 -17: 283 n. 22

32*18-20:

3- 17 -23: 307-8 n. 13

222

Rhetoric

3- 20 -3: 329

1392*19-22:

225 n.

22

Topics

3- 20

312

3- 21--3: 312

112*36-8:

287

3- 21

284,

312

Fragments, ed. Rose

3- 22

313,

315-16

17:

3- 31

305,

332

3- 32

319

26 n.

19c:

58

260-1

n.

22

3- 50 -3: 328

Arius Didymus

3- 50

310.

311

Fragments

3- 51

307,

307-8 n.

3- 55

318 n.

29:

290 n.

37

Aulus Gellius 7.

i:

310 n.

7.

I.

7-13:

262-3

n. 26

7.

I.

7-11:

262-3

r*.

7.

I.

7:

16

289 n.

26

33

307-8 n.

13

3- 57

325-6 n.

35

3- 58-60: 325-6 n. 35 3- 58--9: 325-6 n. 35 3- 59 318 n. 28, 323, 3- 60 -i: 323

iii.

314.

321

3- 63 fF.: 291 n. 39

Long Commentary on the Meta¬ ed.

1492-4:

Bouyges

184-5

3- 64

328

3- 69

309 n.

4- 39

336 n. 46

15,

328

5- 14 -15: 304

Berkeley, George Principles of Human Knowledge

5- 17 -20: 313 Dt natura deorum

18:

353 n.

29

I.

39

290 n. 36

19:

353 n. 29

2.

16

291

n.

39

20:

365-6 n.

2.

18

260 n.

21

29:

361

n.

45

2.

3611.:

35:

361

n.

45

2.

37

55

I

27

3- 56

3- 60

Averroes physics,

33—

290 n.

291

n.

306 n.

262—3 n.

Stoic, repugn. 1044

23i“i3-i4: 236 261*13-14:

180.

259-60 n.

17

17

187

200*32-3:

14,

18

112

ap. Epict. Diss. 2.

i92‘’i4:

Armstrong:

li

Chrysippus ap. Alex.

71

Physics

I93'>6:

361

73

^

n.

ii44'’3S-ii45‘’i: 1145*1-2:

361

First Dialogue,

70

ii44'’30-ii45“2:

ii44*’32:

42-4:

Three Dialogues

III

36

39,

292

d:

26

262—3

Index Locorum 2.

153:

391

284-5

7.

85 ff.:

283

n.

22

De officiis I. 15--20: 319

7.

87-9:

283

n.

21, 286

7.

87:

276,

291

n.

39,

I. 32: 328 I. 47--8: 319

7.

88:

286,

305

n.

8

7. 89:

288 n.

30

3- 7-11: 335

7. 94:

318 n.

27

3-

13:

334, 335

7.

101-3:

3-

19:

335

7.

102:

3- 21: 334

7-

103-4:

3- 22: 334 3- 23-■4: 334

7.

103:

7.

104-5:

3- 26: 334

7r 104:

307

3- 28: 334 3- 29: 319

7.

105:

307-8

7.

106-7:

3- 30: 334

7.

106:

309,

3- 33--4: 335, 335

7.

107:

307-8

3- 35: 319, 335

7.

108-9:

3- 38-9: 320

7.

109:

321

3' 40: 335

7.

116:

318

3- 52: 335

7.

135-6:

3- 89--90: 319

7.

139:

3- 89: 319

7.

156-7:

3- 99--115: 320

7.

160:

310

3- 100: 320 Tusculan Disputations

7.

161:

114 n. 77,

4.

12-14:

7.

174:

299

5.

40 ff.:

7.

201:

277 n.

12

9.

103:

380 n.

82

322 n. 283

n.

94,

31 21

261

n.

24,

n-

49

292,

304 n. 3

261-2

n. 25

309

337

309 262-3

n.

n.

26,

307

13

325 311-12 n. n.

19

13

319

n.

27

257

290 n.

37

276 n.

123

n.

10

97, n.

117,

124,

118,

125,

121

n.

126

54

Clement of Alexandria Empedocles

Stromateis 2.

21:

305

n. 8

6.

12:

318 n.

31

A 49a:

2. 6:

Descartes, Rene

6:

Meditations 1, AT vii.

ii.

12:

364 n.

II, AT vii.

25/CSM ii.

16-17:

II, AT vii.

25/CSM ii.

17:

366

II,

27/CSM ii.

18:

359

41,

367 n.

n.

56

58:

259 n.

14

Diogenes Laertius 2.

106:

114 n.

2.

125-44:

2.

125-6:

46 n.

2,

77,

117 n. 85,

113-14 n. 76 117 n.

83

118

3,

54.

n. 52,

6 n. 29,

7,

31,

II, 48,

12,

23,

24

49,

50,

51,

24,

30,

37,

53 II

17-

3-5:

10,

12,

n.

29,

n.

29

26

22,

23,

26,

27

41

17.

3:

II

17.

4:

26,

17.

5: 25

17.

7-8:

27 n.

2,

56, 29

17.

7:

17.

8-9:

29-30 n.

17.

8:

17.

14-15:

15

23

66

30

24 n.

i:

290 n.

37

17.

18—20:

7.

4:

299 n.

54

17.

19: 41

7.

40:

277 n.

II

17,

20:

7.

58:

259 n.

14

17.

21—6: 41

7.

84-131:

17.

31-5:

25

16,

30,

1-2:

7.

284 n.

44

88

17.

38,

Diogenes of Babylon D.L. 7.

50

366

>

ap.

18 n.

38

17: 18/CSM

AT vii.

17-18,

Fragments

27

S3

II

I

n.

2

Index Locorum

392

2o:

I

n.

2,

20. 2-3: 20. 4: 20.

n.

84 86

31

16 38

21;

2,

6

21.

9-12;

21. 9:

, n.

34

21.

11:

22:

36-8

22.

3:

16,

24

n.

n.

71

34,

48,

55,

49,

33,

'

51,

52

,

33-4

105:

36,

24

n.

41 55,

34,

35,

48,

49,

51,

52

23- 4: 3S 23- 5-8: 33 23.

8:

21

23. 9-11: 26:

2,

26.

1-3:

6

26.

2:

26.

3-7:

43

n.

26.

3-4: 4:

16,

7,

30,

48,

49,

51

n.

72,

34.

52

5-6: 5:

29-30

26.

29

7:

12,

n.

8,

33:

49

34:

49

35:

6

28-30,

29-30

n.

66,

30,

1-3:

35-

7:

35.

10:

95 16,

16:

36:

18

8,

12,

30,

49,

50

49

18

n.

44

44,

37

41 n.

6-11:

57-61:

7,

15

41,

3536.

95

n. n.

19

121:

13

n,

124:

13-14,

139:

I

n.

2,

109:

3

n.

9

50, n.

99

12

n.

31,

n.

54

18,

19,

n.

20

n.

46,

49,

48,

22,

36

16

ed.

Martin—

52

(i)

1-5:

I

a

(i)

6-7:

II

n.

a

(i)

6:

(i)

8-(ii)

2

n.

29

2 2:

24

n.

55,

33,

34,

35,

a

(ii)

2:

a

(ii)

3-17:

21

a

(ii)

17:

a

(ii)

18-20:

a

(ii)

21-3:

a

(ii)

24-30:

a

(ii)

24:

24

a

(ii)

29:

48

a

(ii)

30:

24

b:

49 2:

I

n.

b 4:

I

n.

5,

2,

12,

35

3 11

n.

29,

49

15 24

2

c

2-8:

c

3:

I

2 n.

2

12

2:

18,

6:

16

71:

12

n.

31,

43-5,

21

50,

51.

52,

43

n.

83,

n.

d

i-ii:

d

5-6:

d 9:

62.

2 31,

45,

15 I

n.

d

lo-ii:

15

d

10:

n.

14

14-19,

21,

ii:

16

44

d

12:

16,

d

13:

16,

d

15:

14

d

16-19:

d

19:

71-

4:

73:

12

n.

31,

42,

43,

45,

49,

75:

12

n.

31,

44,

45,

49,

52

76:

49

76.

2:

52

48

39

2,

14

n.

36

36,

15

n.

37,

16

53 d

2

n.

16

44

n.

13,

42,

102

62.

I

21, 36,

76

48, 51

3:

n.

{ensembles)

a

d:

71-

19,

14

53

53

48,

18,

31-2

62:

i:

34

Strasbourg Papyrus,

b

61:

62.

i

49 49, 50 32--3, 49, 50

52

32

n.

35-

7,

66

41

30-1:

33

115:

a

26. 26.

8

3:

a:

41

29

27-9:

32-3

109-

Primavesi

28

49 49

32

109:

29

34

26.

87 95 96 98 98

21

38

9:

23:

12

49

5:

22.

2,

3: I n. 2 12 n 31, 12 n 31, 49 12 n 31, 12 n 31, 12 n 31, i: 33

76

Fragments (cont.):

Empedocles:

18

18 18 n. 18 n.

36, n. 44

16, 44

17,

18

15

n.

38,

Index Locoruni

393

Epicharmus

Philo of Alexandria

Fragments

De aeternitate mundi

2:

39-43: 260-1 n. 22 48: 239 49-51: 240 n. 4

243,

244

Epictetus Dissertationes 1. 6.

12 ff.:

1.

20.

14:

2-

5-6:

337

2- 5-

4:

2.

9—10:

6.

283

305 n.

305

262-3

2.

8:

305

2.

10:

3.

3:

333 322

n.

3. 7-

6:

3.

7.

24-6:

3.

22.

n.

II,

310

n.

17

Plato

26

Apology

7

27 c

45

n.

977

7

305 305

lo-D

i: 63-4 n. 6

Epinomis

31

30s

76:

1-6: 39 2: 38-40

n.

n.

n.

Fragments

7

306

9:

Philolaus

21

7

49

n.

2. 6. 8.

n.

n.

n. n.

D 3:

77 n.

24

Euthydemus

7

278E-281E: 309, 323

7

Euthyphro

6d 11 E 11 E 12 E

Eusebius Praeparatio Evangelica 15.

20.

6:

21

n.

49

Gorgias

Galen De placitis Hippocratis et Platonis 4.

I,

pp.

5.

6.

10-14:

5.

243-8

169-70,

De

307-8

p.

Lacy: n.

326.

294

13,

21-7

n.

44

313

De

Lacy:

295

7-

2,

pp.

127

io-ii: 119 7-12 E 8: 103 n. 61 7-12 A i: 103 n. 61 5-8: 90 n. 47

434.

n.

31-436.

8

De

Lacy:

102

4540: 377 n. 77 489 E 3-4: 119 n. 90 489 E 5: 119 n. 90 507 A i-c 7: 90-1 n. 48 507 B 1-3: 90 n. 47 507 B 1-2: 90 n. 46 507 c 6: 67 n. 8 Laches

Heraclitus Fragments 53:

46

n.

89

62:

37

n.

75

90:

100

n.

57

Hume, David Treatise of Human Nature I.

4.

2:

361

I.

4.

6:

367

n.

>

45

Lucretius I.

73:

18

n.

44

Nemesius De natura hominis, 105.

18-19:

268

255-6

III.

14-25:

111.

26-112.

112.

1-3:

260

i: n.

ed.

n.

265 19

40

Morani

194D 194 E 195 c 195 D 197 B 198 B 198 B 198 B 198 B 198 B 198 B 198 c 199 A 199 c 199 c 199 C 199 c 199 c 199 c 199E 199 E

1-2: 129 11-195 A 6: 93 n. 50 8-9: 89 8: 87 2-3: 89 2-3: 87, 89 2-199 c 2: 88 2-c 8: 87 5-7: 89 n. 45 8-9: 89 9-C 2: 90 9-199 c 2: 88 lo-B i: 87 n. 42 3-E 5: 88 3-E 2: 90 4-D 1: 88 4—6: 129 6-d 1: 129 7-D 1: 129 ii: 88, 129 12: 88

Index Locorum

394 (cont.):

Plate)

Laws 896 D-898

c:

901A-903A;

n.

290 257

35

n.

II

329 c

4-5:

75,

n.

28,

90-1

329 c

6-d

i:

60,

72,

73,

75

329 c

6-e 6: 63,

66.

74,

75,

29,

82,

83,

78

329 c

6-e 4:

70 A 1-2: 106

329 c

6-8:

69-70

329 c

7-8:

90-1

329 c

8-d

a;

71

D 4-72 A 8:

363

n.

49 105

72 A 8-79 E 4: 105 74 B 4-c 4: 109 74 c 5-D 2: 109 77 c 7-8: 323 78 c

75, 79 329

D

i:

118,

3-79 E

4:

105,

1-79 D

7:

108

107,

109,

III

63,

120,

329

D 3—E 6:

329

D 3-4:

329

D 3:

329

D 4-E 2:

79A

3:

los

329

D 4-8:

79

A

4-5:

79

A

10:

79

B 2:

73,

79

B

4-5:

79

B 6-7:

109,

60,

D 4:

105

329

D 5-6:

105

329

D 6-8: 63,

74,

95-6

49,

109

n.

68

78,

79,

D 7fF.:

79 c

329

D 7-8: 97

329

D 7:

329

D 8-e 2:

329

D 8:

329

E 2-4:

329

E 2:

94-5

329

E 4:

128-9

329

E 5-330 A 2:

52,

105 n.

64

105

c 4 ff,: 106 86 c 4-89 c 4: 107 86 c 7-D 2: 106 86 c 11-89 c 4: 106 86 D 3-87 c 10: 106 87 c 5-7: 106 87 C 11-89 c 4: 106 87D-89A: 323 86

a 7-d A 3-4:

106,

89

A 4-5:

106

89

A 4:

104,

89 89

c c

106

98

a:

377

n.

107,

65,

n.

106,

59c 2: 118 c 8-d i;

48 108

107

78 B-80

e:

52,

97

n.

86,

95-6

n.

n. 60

n.

12

97

95,

60,

69

96,

60,

100

75

n.

51

ri.

104

77,

n.

78

26,

95-6

52

60, 69,

330 A

I:

330 A 2:

76 n. 77,

23

78 n.

3- B 6:

75

75

26,

81-2 n.

330A 3- B 3:

107

34

99 n, 4,

83,

84

330 A 4: 66, 76 330 A 8-■B i: 76 87

90-1

330 B n.

48

B; 280

Phaedrus 248

loi

69—70

4- B 3: 61 330 A 4- 7: 95

77

n.

n.

100

329 E 6--330 A 2:

Phaedo 68

98,

329 E 5- 6:

106

n.

III

96,

< 0

5:

90-1

n.

< 0

2-4:

i:

57,

97

D 6-7: 97,

89

n.

97

329

88

100

97

329

105

74,

69,

79 B 6: 105 79 c 2-3: 105 79 c 2: 92 n. 49, 109

D 5-86 C 3:

74,

127

D 4-5:

D 6-7:

73,

75

329

80

71,

60

329

n.

79

9,

55

105

105

5:

n.

60

105

92

12 48

127

i:

78

68

n. n.

29

E

78 E

n.

59

1: n.

79

48

83

Meno 71

n.

21

Protagoras

323 A 5-C 2: 90 n. 46 3280 4-7: 59 329 C 2-6; 78, 95-6 n. 52 329 c 3-6: 77, 128 n. 103

I:

61

330 B 3-■6:

n. 4,

69,

94, 94- 5 n.

51

76

330 B 4: 76 n. 23 330 B 6--331 B 8: 64 n •

7 331 A I--B 4: 81-2 n. 34 331

A 6--B 8: 63,

331

B 2--3:

331 B 4--5: 331 B 4: 331

B

5:

103

64

n.

61

63-4, 64 n • 7,

65,

81,

82

63,

64,

64 n

7, 81- 2 n. 34

63,

64,

64 n

7, 81- 2 n. 34

Index Locorum 331 B

8-332 A

332 A

2-5:

332 A

4-333 B

6:

90-1

333 B

4-5:

n.

37,

n.

i:

64

81-2

85

n.

n.

359 A

7

34 n. 86

48 n.

39,

90-1

48

333 B

6:

64-5,

81

333 B

8 fT.:

8s

n.

37

334

A

3ff.:

85

n.

37

348

C

I ff.:

65

6-d

i:

8-b

6:

75

349 A

8-c

s:

77,

97

349 A

8-b

i:

75,

95-6

349 b

i-d

8:

67

349 B

i-c

5:

74,

349 B

1-6:

65,

79

n.

29

1-3:

79

n.

349 B

1-2:

349 B

i:

349

n.

78

n.

52

76

76,

360

D 4-E 5:

360

D 4-5:

85,

87

361

B 5-7:

77,

79,

73

3-6:

81-2

349 B

3:

349 B

4-5:

349 B

6-c

5:

76

349 B

6-c

i;

65

349

6:

71,

95-6

72,

52,

28,

79

79,

n.

73,

108,

74,

no

83

120

E 6-7:

90-1

C-d:

257

C 7-8:

74,

78,

98,

79,

2;

76,

79,

349 c

3;

loi

349 c

5:

349 D

2 fF.:

n.

60,

66,

80,

86,

95-6

349 D

2-8:

67,

349 D

6-8;

67

349

D

7:

67

n.

349 D

8:

67

n.

349

349

E E

I ff.:

60 78,

80,

Aff.:

78 n.

n. 8,

26,

93

n.

4-5:

A—b;

2:

n.

38,

350 D

4-5:

8s

n.

38

n.

281

n.

351

n.

n.

a:

293

d:

289

32

B—33 a:

33

c:

289

26

n.

n. 6

42

328-9

n.

260—1 n.

b:

282

e:

290

45

b 2-46 c 6:

89 76,

359

A

5-7:

76,

359

A

5-6;

76

22

83

50

A 5-B 2:

C 4-68 D 7:

97

289

n.

83

33

90-1

n.

263

n.

n.

48

d:

90

B

91

a:

289

n.

33

29s 295

I-D 7:

279

129

6-7:

n.

297

67



90A-D:

322

2-7:

38

33

36

89

322

359 A

n.

33

34

75A-B:

3580

16

38

322

48

74

63-4

74E-75D:

5-6:

48

n.

85

8s

B

n.

90-1

SO A s-B 5: 96

i;

358 c:

90-1

376

E 3:

47A-C:

26

4-5:

357

10

8 n.

350 c

356 Dff,:

n.

50

85

8

350c

353C-358E:

91

189-90

A 2-10:

70 c: 85

30 109

85

78

I-35I B

n. 92,

98

76

66,

n. 48

no

98

75,

38

Timaeus

29 349 c 4-5:

n,

42

Statesman

28B-30C: 349 c

90-1

Theaetetus

30

108,

11

289

n.

33

45

77

Plutarch

83

Against Colotes n 20 c

9;

89

81,

n. 77

lO-D 3:

257 c

189

2-3:

n.

377

242

184 c:

n.

328-9

A ff.:

176

81

n. 26,

477

184-6:

66,

80,

107

430

202

73

96,

77,

39

Symposium

76

2-5:

52,

B 6:

n.

n. 42,

n.

86,

45 95-6

108

78

87

271

80

34,

73, 74, 75, 66,

107,

5-6;

196 c n.

26 86

C 5-8:

20 n.

52,

n. 85,

Sophist

78,

n.

78,

2-3:

349 c

9,

n.

74-5,

B

349 c

78

429

75,

95-6

349 B

B

B 2-6:

345E-347A:

66,

68

29,

76

359

Republic

52 349 B

6-7:

n.

349 A

76

A

361 65,

5:

359

361 B

349 A

395

242

27

n. 48

Index Locorum

396

Posidonius

(cont.):

Plutarch

De animi procreatione in Timaeo

Fragments,

1013 Eff.:

187:

290

n.

35

ed.

Kidd

295

\

De communibus notitiis 1060 c:

318

n.

1061 a:

318

n.

Seneca

28 28

\

1068F-1072F:

323

1069 c-d:

n.

314

337

n.

49

1069 f:

337

n.

49

307-8

1071 a-b:

314

1077 c-e:

247

33

23

1069 e:

1071 A ff.:

n.

Epistulae morales 79,

9-10:

92.

11-12:

283

21

Quaestiones naturales I.

n.

n.

13:

260

n.

21

13

Sextus Empiricus

23

Adversus mathematicos 7.

167:

8—C

i:

243

7.

190-200:

1083 B

8-c

i:

24s

7- 193-4: 355

1083 F

2-3:

1083A-1084A: 1083 a

n.

315

241-2

n.

7

348 358

n.

39

7.

248:

259

n.

De hide

7.

252:

258

n.

360 e:

7.

352-8:

352

n.

28

De Stoicorum repugnantiis

7-

352-3:

352

n.

28,

1034 c-d:

124

7.

366:

1034 c-e:

124

7.

409-10:

n.

62

243

387

1034 c:

126

103s a-b: 103s b:

276

103s c:

275

1037 Eff.: 1042 b:

n.

8-

475:

372-3

9

9-

1-3:

364

9- 359-440:

325-6

314

1042 d-e:

n.

318

1042 e-f:

n.

318

1044 d:

262-3

1047 c:

289

1050 B-D:

35

28

n.

27

n.

n.

26

33

306

1050 C-d:

n.

22

n.

310

II

n.

16

1052 a:

266

n.

33

1052 d:

289

n,

33

De virtute morali 440 E-441 b:

125

440 E-441 a:

121

440 E-F:

n.

440E:

117

114,

440 f:

84-5

441 a:

125

441 b:

125

122 n.

n.

1-7:

18-19:

II.

48:

305

n.

II.

61:

337

n.

II.

62:

307-8

II.

64-7:

II.

67:

II.

79-89:

II.

81-2:

II-

83:

356

II-

87:

357

II.

88:

358

n.

94,

123

119, n.

36,

n.

97

n.

Porphyry De abstinentia 2.

31:

16

2.

31.

5:

n. 16

40

353

16

n.

29

373

n.

n.

49

65

7 49 n.

13

310

311

I-

1-35:

I-

1-4:

I.

4:

356 356

372

I.

12-24:

346

372

n.

n.

14

61

126 374-5

n.

67

96 122,

123

I-

12:

372,

373

I-

13:

372,

372-3

I-

15:

359,

360,

99

13

363

II.

372,

Quaestiones conviviales 683 E

n.

Pyrrhoneae hypotyposes

82

115,

440F-441A:

259

258

100

n.

13,

374

8- 316: 373

275

15

n.

62,

362,

374, 375

I-

19-20:

I.

19:

368

I-

20:

349-50

373

373,

33

I-

22:

I-

33-4:

I-

99:

I.

102:

n.

57

376

370

n.

59

361 348,

361

n.

374

368,

45

369,

371,

Index Locorum 124-■7: 348 125: 361 n. 45 135: 373 n. 65 178: 374--5 n. 67 187--209: 372 187--205: 374 I9I; 351 n. 26 198: 373 n. 65 202: 373 n. 65 I. 203: 359-60 n. 43, 374 I. 207: 351 n. 26 I. 208: 359-60 n. 43, 374-5 n. 67 I. 215: 358 n. 39, 379, 380 nn. 81,

I. I. I. I. I. I. I, I. I.

83 I. 218: 380 n. 83 1. 219: 380 n. 83 2. 29-30: 365-6 2. 70; 368 n. 57 2. 72-5: 352-3, 355, 356 2- 72-4: 362 n. 47 2- 72: 354, 356, 358 n. 39 2. 74: 354, 356, 361, 362 2. 84: 364 n. 50 3- i: 364 3- 38-55: 363 n. 49 3. 38-46; 363 3. 38: 365 n. 54 3. 46-9; 363 n. 49 3. 47-9; 363

Simplicius In Aristotelis De Caelo,

ed.

Heiberg

529. 28: 49 530- 5: 43, 49 530- 8: 49 590. 19-21: 51 n. 97 In Aristotelis Physica,

26. 7-15: 291 n. 38 30. 19: 50 32. ii: 50 33: 26 n. 57 38. 8-17: 34 n. 71 157. 25: 3 r>. 7 159. 13-26: 34 n. 71 159- 27: 35 300. 19; 50

371-

33-372.

14:

50,

381.

29-382.

18:

54

53 n.

n.

99

102

Smith, Adam The Theory of Moral Sentiments n.

2.

I. 15-47: 326 I. 18: 326-7 I. 19, note k'.

2.

I.

36

7.

2.

7.

2.

7. 7.

21:

326

n.

36

326

n.

36

Stobaeus, ed. Wachsmuth-Hense i. 10. ii: 37 n. 76 i. 177. 21-179. 17: 259 n. 14 ii. 71- 15-72. 3: 318 n. 27 ii. 75- 1-6: 318 n. 28 ii. 75- 11-76. 8: 286 n. 27 ii. 75- 11-76. 6; 305 n. 4 ii. 76, 6-8; 305 n. 5 ii. 76. 9-10: 305 n. 6 ii. 76. 10—11: 305 n. 8 ii. 76. 11-15; 305 n. 7 ii. 78. 7-12: 318 n. 28 ii. 79- 1-4: 318 n. 28 ii. 79- 9: 307 ii. 79- 18-80. 13: 261 n ■ 24 ii. 79- 20—80. i: 309 n. 14 ii. 80. 6-8: 307-8 n. 13 ii. 80. I4--2I: 307-8 n. 13 ii. 81. i: 325 ii. 82. 11-12: 307-8 n. 13 ii. 82. 21-83. 9: 307-8 n. 13 2: 261-2 n. 25 ii. 83. ii. 83. lo-iI: 306 ii. 83. 11-84. 3: 307-8 n. 13 ii. 83. 12: 307-8 n. 13 ii. 83. 14-84. i: 330 ii. 84. 4-17: 307-8 n. 13 ii. 84. 4-13: 307-8 n. 13 ii. 84. 18-85. ii: 261-2 n. 25 ii. 84. 18-22: 307-8 n. 13 ii. 84. 24-85. i: 305 n. 7 ii. 86. 2: 325 -6 n. 35 ii. 96. 18-97. 5: 318 n. 27 0 1 00

3- 48: 363 3- 49: 363 3. 71: 368 n. 57 3. 183-90; 356 3. 184: 357 n. 36 3. 186: 366 3. 191: 307-8 n. 13

397

ed.

Diels

Stoicorum Veterum Fragmenta, ed.

von

Arnim

i.

49

(no.

200):

126

n.

100

i.

79

(no.

351):

117

n.

82,

i.

85-6

i.

86

i.

87:

256-7

n.

8

i.

98:

266-7

n.

36

i.

102:

(no.

(no.

374):

375):

257

127

123

n.

n. 97

123

102

n.

97

Index Locorum

398

Stoicorum Veterum Fragmenta {cont.y i.

361:

261

i.

497:

266-7

n.

24 n.

36

iii.

128:

261

n.

24,

iii.

137:

261

n.

24

iii.

190-6:

iii.

191:

757-68:

306

n.

II,

262-3

d.

26

ii.

82.

20-1:

307

iii.

ii.

84.

21-4:

307

iii Archedemus

ii.

194.

15:

290

300

257

n.

11

311

257

n.

11

n.

14

n.

37

314

iii

Boethus

7:

iii

Diogenes

257

259 239

439

258

n.

12

604

289

n.

33

624

259

n.

14

626

267

n.

37

634

290

n.

37

Theophrastus

642

290

n.

37

Fragments,

644

290

n.

37

763

289

n.

33

In Metaph. 12,

258

n.

ii.

1163:

262-3

1170:

262-3

ed. iii.

no:

318

n.

27

iii.

119:

261

n.

24,

n.

33

305 n.

259

n.

8

ii n.

14

ap.

Averroes,

Long

Metaphysics,

Bouyges

1492-4:

185

ed.

n.

3,

188,

204

n.

Fortenbaugh-Huby-

Sharples-Gutas 230: 266-7

n.

36

n.

33

291

n.

38

Zeno

26 26,

iii.

25

22

Commentary on the

12, n.

323

n.

Themistius

395

1027:

n. 21:

22:

397

ii.

261-2

289

ap.

Alex.

Lyc.

Contra Manicheo-

rum opiniones disputatio, 262-3

n.

26

Brinkmann:

257

n.

9

19.

2—4

27

Notes for Contributors to Oxford Studies in Ancient Philosophy 1. Articles may be submitted at any time of year. They should be printed, on one side of the paper only, and with double line-spacing throughout. Footnotes may at the stage of initial submission be printed at the foot of the page (however, see point 4 below). Pages should be A4 or standard American quarto (8j x ii''), and ample margins should be left. 2. Two identical copies should be submitted to the editor. Authors are asked to supply an accurate word-count {a) for the main text, and {b) for the notes. The covering letter should provide a current e-mail address, if available, as well as a postal address. 3. Typescripts will not normally be returned to authors, but will be disposed of after all dealings have been concluded. The remaining instructions apply to the final version sent for publication, and need not be rigidly adhered to in a first submission. 4. Footnotes should be numbered consecutively and located together on separate sheets at the end of the typescript, in double line-spacing. They will be printed at the foot of each page. Wherever possible, references to primary sources should be built into the text. 5. Use of Greek and Latin. Relatively familiar Greek words such as ‘psyche’ and ‘polis’ (but not whole phrases and sentences) may be used in transliteration. Wherever possible, Greek and Latin should not be used in the main text of an article in ways which would impede comprehension by those without knowledge of the languages; for ex¬ ample, where appropriate, the original texts should be accompanied by a translation. This constraint does not apply to footnotes. Greek copy must be supplied in a completely legible and accurate form, with all diacritics in place both on the hard copy and in the computer file. A note of the system employed for achieving Greek (e.g. GreekKeys, Linguist’s Software) should be supplied to facilitate file conversion. 6. In references to books, the first time the book is referred to give the initial(s) and surname of the author (first names are not usually required), and the place and date of publication; where you are abbreviating the title in subsequent citations, give the abbreviation in square brackets, thus: T. Brickhouse and N. Smith, Socrates on Trial [Trial] (Princeton, 1981), 91-4. Give the volume-number and date of periodicals, and include the full page-extent of articles (including papers and chapters of books): [The Notes continue overleaf]

Notes for Contributors

400

W. D. Graham, ‘Symmetry in the Empedoclean Cycle’ [‘Sym¬ metry’], Classical Quarterly, NS 38 (1988), 297-312 at 301-4. G. Vlastos, ‘The Unity of the Virtues in the Protagoras’ [‘Unity’], in id., Platonic Studies, 2nd edn. (Princeton, 1981), 221-65 at 228. 7. Where the same book or article is referred to on subsequent occasions, usually the most convenient style will be an abbreviated reference, thus: Brickhouse and Smith, Trial, 28—9.

Do not use the author-and-date style of reference: Brickhouse and Smith 1981: 28—9. 8. Starting from 2003, authors are asked to supply in addition, at the end of the article, a full list of the bibliographical entries cited, alpha¬ betically ordered by (first) author’s surname. Except that the author’s surname should come first, these entries should be identical in form to the first occurrence of each in the article, including where appropriate the indication of abbreviated title: Graham, W. D., ‘Symmetry in the Empedoclean Cycle’ [‘Sym¬ metry’], Classical Quarterly, ns 38 (1988), 297-312. 9. If there are any unusual conventions contributors are encouraged to include a covering note for the copy-editor and/or printer. Please say whether you are using single and double quotation marks for different purposes (otherwise the Press will employ its standard single quotation marks in all contexts). If you specifically require transliteration rather than true Greek, this must be made clear to prevent routine conversion to Greek in the course of editing. 10. Authors whose contributions are accepted will be asked to send a copy of the final version of their paper on a 3.5" high density (HD) disk (Macintosh or IBM), indicating on the disk or in an accompanying note the program in which the text is written, including the system employed for achieving Greek (see point 5 above). NB. The version

on disk must be the exact version which produced the hard copy sent in for printing.

ALSO PUBLISHED BY OXFORD UNIVERSITY PRESS Porphyry’s Introduction Jonathan Barnes Clarendon Later Ancient Philosophers Platonopolis Platonic Political Philosophy in Late Antiquity Dominic]. O’Meara Knowing Persons A Study in Plato Lloyd P. Gerson The Heirs of Plato A Study of the Old Academy (347-274 BC) John Dillon Pyrrho, his Antecedents, and his Legacy Richard Bett new in paperback

CONTENTS Empedocles on the Ultimate Symmetry of the World SIMON TREPANIER Socrates and Protagoras on Virtue DENIS O’BRIEN Aristotle’s Natural Teleology and Metaphysics of Life MARC PAVLOPOULOS Themistius and Spontaneous Generation in Aristotle’s Metaphysics DEVIN HENRY What Does Aristotle Mean by Priority in Substance? STEPHEN MAKIN Chrysippus’ Puzzle about Identity JOHN BOWIN Determinism and Recurrence in Early Stoic Thought RICARDO SALLES Cosmological Ethics in the Timaeus and Early Stoicism gAbor BETEGH A Puzzle in Stoic Ethics RACHEL BARNEY Sextus and External World Scepticism GAIL FINE

ISBN 0-19-926343-4

OXPORD UNIVERSITY PRESS

www.oup.com

9 780199 263431