The philosophers of the Hellenistic schools in ancient Greece and Rome (Epicureans, Stoics, Sceptics, Academics, Cyrenaics) made important contributions to the philosophy of mind and the philosophy of psychology. This volume, which contains the proceedings of the Fifth Symposium Hellenisticum, describes and analyses their contributions on issues such as: the nature of perception, imagination and belief; the nature of the passions and their role in action; the relationship between mind and body; freedom and determinism; the role of pleasure as a goal; the effects of poetry on belief and passion. Written with a high level of historical and philosophical scholarship, the essays are intended both for classicists and for non-specialists interested in the philosophy of mind.
PASSIONS & PERCEPTIONS
PASSIONS & PERCEPTIONS Studies in Hellenistic Philosophy of Mind Proceedings of the Fifth Symposium Hellenisticum
EDITED BY
JACQUES BRUNSCHWIG Universite de Paris i and
MARTHA C. NUSSBAUM Brown University
CAMBRIDGE
UNIVERSITY PRESS
Published by the Press Syndicate of the University of Cambridge The Pitt Building, Trumpington Street, Cambridge CB2 IRP 40 West 20th Street, New York, NY IOOI 1-42 I I, USA 10 Stamford Road, Oakleigh, Victoria 3166, Australia and Editions de la Maison des Sciences de l'Homme 54 Boulevard Raspail, 75270 Paris Cedex 06 © Maison des Sciences de l'Homme and Cambridge University Press 1993 First published 1993 A catalogue record for this book is available from the British Library Library of Congress cataloguing in publication data
Passions & perceptions: studies in Hellenistic philosophy of mind : proceedings of the Fifth Symposium Hellenisticum / edited by Jacques Brunschwig and Martha C. Nussbaum. p. cm. Collection of papers as a result of the Symposium Hellenisticum, held at the Chateau of Syam, Aug. 17-25, 1989. Includes bibliographical references and index. ISBN o 521 40202 6 (hardback) 1. Psychology - History - Congresses. 2. Philosophy of mind - History - Congresses. 3. Philosophy, Ancient - Congresses. I. Brunschwig, Jacques. II. Nussbaum, Martha Craven, 1947- . III. Title: Passions and perceptions. BF91.P37 1992 I28'.2'o938-dc2o 91-42510 GIP ISBN o 521 40202 6 hardback ISBN 2 7351 20450 8 hardback (France only) Transferred to digital printing 2004
Contents
Preface Av ant-prop os PART I
1
page ix xi
ETHICS AND PSYCHOLOGY OF HEDONISM
Epicurean hedonism Gisela Striker, Harvard University
2 Anniceris et les plaisirs psychiques: quelques prealables doxographiques Andre Laks, Princeton University PART II
ATOMISM AND EPICUREAN PSYCHOLOGY
I
3
18
51
3
Epicurus on agency Julia Annas, University of Arizona, Tucson
53
4
Democritus and Epicurus on sensible qualities David Furley, Princeton University
72
PART III
THE PASSIONS
95
5
Poetry and the passions: two Stoic views Martha C. Nussbaum, Brown University
6
Seneca and psychological dualism Brad Inwood, University of Toronto
7 Actions and passions: affection, emotion and moral self-management in Galen's philosophical psychology James Hankinson, University of Texas, Austin Vll
97 150
184
Vlll
CONTENTS
PART IV
STOIC PSYCHOLOGICAL CONCEPTS
223
8
De la 'nature phantastique' des animaux chez les Stoiciens 225 Jean-Louis Labarriere, CNRS et Centre Louis Gernet de recherches comparees sur les societes anciennes, EHESS/Paris
9
Le concept de doxa des Stoiciens a Philon d'Alexandrie: essai d'etude diachronique 250 Carlos Levy, Universite de Paris xn, Val de Marne
10 Seneca on reason, rules and moral development Phillip Mitsis, Cornell University
285
11 Chrysippus on psychophysical causality David Sedley, Christ's College, Cambridge
313
Bib liography Subject index Name index Index of passages cited
332 344 346 351
Preface
The fifth meeting of the Symposium Hellenisticum took place between 17 and 25 August 1989 at the Chateau of Syam near Champagnole in the French Jura. Following a by-now well-established tradition, to which the four previously published volumes bear witness, the papers given at the fifth Symposium underwent long and searching discussion; the texts published here have been revised in the light of those and subsequent discussions which took place in one form or another after the end of the conference. The participants in the Symposium were: Julia Annas, Jonathan Barnes, Suzanne Bobzien, Jacques Brunschwig, Pierluigi Donini, Eyjolfur Emilsson, Dorothea Frede, Michael Frede, David Furley, James Hankinson, Brad Inwood, Anna-Maria Ioppolo, David Konstan, Jean-Louis Labarriere, Andre Laks, Carlos Levy, Geoffrey Lloyd, Anthony Long, Jaap Mansfeld, Mario Mignucci, Phillip Mitsis, Martha Nussbaum, Malcolm Schofield, David Sedley, Richard Sorabji, Gisela Striker, Voula Tsouna. The organization of the conference was made possible by generous support from several institutions: le Centre National de la Recherche Scientifique; le Ministere de l'Education Nationale (Direction des affaires generates, internationales et de la cooperation); l'Universite de Paris 1 (Pantheon-Sorbonne); l'Universite de Paris x (Nanterre). We acknowledge with gratitude not only these institutions, but also those who have individually shown their interest in our project and have actively furthered its realization. We express our gratitude to Madame M. Carnot-Fernique, owner of the Chateau of Syam, the Deputy Mayor of Champagnole and his assistant, and to all those people in the region with whom we came into contact; they all worked in their various ways to make our stay in this beautiful place as pleasant as it was productive. As we send this volume to the printers, we have learned with regret of the death of Madame Carnot-Fernique. IX
Avant-propos
La cinquieme reunion du Symposium Hellenisticum s'est tenue du 17 au 25 aout 1989, au Chateau de Syam, pres de la ville de Champagnole, dans le Jura franfais. Selon une tradition deja bien etablie, et dont temoignent les quatre volumes precedemment publies, les communications presentees devant le Symposium ont fait l'objet de discussions longues et approfondies; les textes qu'on va lire ont ete revises par leurs auteurs, a la lumiere de ces discussions et de celles qui se sont encore poursuivies, sous une forme ou sous une autre, apres la fin de la conference. Etaient presents aux seances du Symposium: Julia Annas, Jonathan Barnes, Suzanne Bobzien, Jacques Brunschwig, Pierluigi Donini, Eyjolfur Emilsson, Dorothea Frede, Michael Frede, David Furley, James Hankinson, Brad Inwood, Anna-Maria Ioppolo, David Konstan, Jean-Louis Labarriere, Andre Laks, Carlos Levy, Geoffrey Lloyd, Anthony Long, Jaap Mansfeld, Mario Mignucci, Phillip Mitsis, Martha Nussbaum, Malcolm Schofield, David Sedley, Richard Sorabji, Gisela Striker, Voula Tsouna. L'organisation de la conference a ete rendue possible par l'aide genereuse de plusieurs institutions: le Centre National de la Recherche Scientifique, le Ministere de l'Education Nationale (Direction des affaires generates, internationales et de la cooperation), l'Universite de Paris 1 (Pantheon-Sorbonne) et l'Universite de Paris x (Nanterre). Nous tenons a remercier non seulement ces institutions, mais aussi les personnes qui, dans le cadre de chacune d'entre elles, ont manifeste leur interet pour notre projet et en ont favorise activement la realisation. Nous exprimons toute notre reconnaissance a Madame M. Carnot-Fernique, proprietaire du Chateau de Syam, ainsi qu'a Monsieur le Depute-Maire de Champagnole, a son assistante parlementaire, et a toutes les personnes de la region avec lesquelles XI
Xll
AVANT-PROPOS
nous avons ete en contact; ils se sont tous ingenies, chacun a sa maniere, pour rendre notre sejour dans ce bel endroit aussi agreable qu'il a ete fructueux. Au moment de mettre sous presse nous apprenons avec peine la disparition de Madame Carnot-Fernique.
PART I
Ethics and Psychology of Hedonism
CHAPTER I
Epicurean hedonism Gisela Striker
Hedonism, like pleasure, can take many forms, and its fundamental tenet, 'pleasure is the good', is notoriously open to different interpretations. Also, the advice, moral and otherwise, given to people who try to pursue this good may vary a great deal, depending on one's view of what pleasure is. To say that a certain philosopher is a hedonist, therefore, is not yet to say much about the content of his doctrine. Still, one would at least expect a hedonist's conception of happiness to be that of a recognizably pleasant life. Epicurus' form of hedonism has seemed paradoxical from the beginning, because it does not seem to meet even this modest expectation. In his own time, the Cyrenaics maintained that what he held to be the greatest pleasure was in fact more like the state of someone asleep (Diogenes Laertius n 89) or even dead (Clemens Alexandrinus Stromateis 11.21; Us. Fr. 451). Cicero was certainly not the first to argue that Epicurus' doctrine was incoherent, and his prescriptions for a pleasant life inconsistent with his principles. Plutarch devoted an entire treatise to showing that one cannot even lead a pleasant life following Epicurus' doctrine. The difficulty both ancient and modern critics have felt lies in seeing how Epicurus could present his claim that the highest good was a state of absence of pain and trouble from body and soul, as a version of hedonism. The problem is well brought out by Cicero in de Finibus n. According to Cicero, Epicurus distinguished two types of pleasures: those 'in motion', and 'static' or 'stable' ones (n. 16 hereafter, 'kinetic' and 'katastematic'). Under the first category fall Ancestors of this paper were read at the SAGP meeting in New York, the Classics Department at Harvard, and the Center for Hellenic Studies in Washington. I am grateful for criticism and suggestions arising out of the discussion at all these places, and at the Symposium in Syam, in particular to Pierluigi Donini, Andre Laks, Mary Margaret Mackenzie, and Phillip Mitsis. Special thanks are due to Mary Mothersill for reading the last version and suggesting clarifications and stylistic corrections.
4
GISELA STRIKER
all the things that ordinary people would call pleasures - 'smooth motions that affect the senses' and analogous events in the soul; the second class consists of states of painlessness or absence of trouble. Epicurus inexplicably claimed that the latter were greater pleasures than the former, and hence should be seen as the goal of life; and he did this although his own way of arguing for the thesis that pleasure is the good seemed to appeal to sensory pleasures (n. 31-2). Why, Cicero asks in exasperation, does Epicurus wish to call such totally different things by the same name (11.9)? Is it not perfectly obvious that a state in which we feel no pain, but also no 'gentle motion' — i.e. kinetic pleasure — is one that is neutral, containing neither pleasure nor pain (n. 16)? Commentators more friendly to Epicurus than Cicero have tried in various ways to show that Cicero's criticism must rest on a misunderstanding. Cicero seems to follow a line of argument that contrasts the Cyrenaics as the 'real' hedonists with Epicurus and insists that his description of the painless state as a pleasure is either a fraud or a blunder. Now Epicurus was neither naive nor stupid, and so it is unlikely that he was as incoherent as Cicero makes him out to be. On the other hand, it seems unwise to reject Cicero's report as totally misguided, since Cicero had access to many more Epicurean writings than we do and had presumably read a fair amount, even if not very carefully or charitably. Worse, since the only longish text concerned with ethics by Epicurus himself, the Letter to Menoeceus, is a protreptic, and hence uses the doctrines without providing the arguments that lead to them, Cicero's account of Epicurean ethics in Fin. 1 is the only systematic exposition we have. Because of Cicero's evident hostility, commentators have tried to reconstruct Epicurus' theory from other sources, relying on Cicero's testimony only where parallels could be found in other authors. Such caution is no doubt advisable; but because our other sources, including Epicurus' own Principal Doctrines, consist in isolated dicta, it becomes harder to see how these bits and pieces fit into what must have been the theoretical framework of Epicurean ethics. Now while we have good reasons to be suspicious about Fin. n, it seems less clear that we must treat the first book in the same way. Cicero's strategy in attacking Epicurus in book 11 is the same as the one he uses against the Stoics in book iv. He argues that a true hedonist would have to follow Aristippus, whereas a philosopher who wishes to hold that the highest good is absence of pain and
Epicurean hedonism
5
distress would have to agree with Hieronymus of Rhodes, who did not consider pleasure even desirable (cf. 11.19, n.35). Similarly, the Stoics are said to be faced with a choice between the positions of Aristo on the one hand, the Peripatetics on the other (iv.67, iv.72). This may suggest that Cicero is following an Academic pattern of argument in both cases. It does not follow, however, that books 1 and in were written to match Cicero's lines of attack. The fact that the distinction between kinetic and katastematic pleasures, so prominent in book 11, is barely mentioned, and not at all explained, in book 1 seems rather to indicate that the first book was not meant just to set up a target for Cicero's subsequent criticism. I think in fact following the line of argument in book 1 that leads up to the doctrine of the highest good criticized in book 11 may help us to see more clearly how Epicurus arrives at his paradoxical views and actually suggest some replies to Cicero's objections. This is not to say that all difficulties disappear once we use Cicero as a witness against himself, but I do think that a careful study of Fin. 1 may set the contrast between the Cyrenaics and Epicurus in a different light, as raising interesting problems about the role of pleasure as the final good. I propose, then, to take a closer look at the section of book 1 that ostensibly sets out Epicurus' arguments for his conception of the goal of life. I assume that this extends from 29 to 42: an argument about the telos is formally announced at 29, and formally concluded at 42. The intervening passage is not very well organized. It seems that Cicero has both added and omitted a few things in a continuous argument. For example, the defence of Torquatus against the suggestion that it should be embarrassing for a Roman aristocrat to profess Epicureanism (34-6) is not likely to be part of an Epicurean argument - and indeed it is marked as a digression at 37. Also, the curiously feeble appeal to the imagination in the contrast between the person who leads a life filled with pleasures and the person who suffers from all kinds of pains and misfortunes does not seem to further the argument. On the other hand, there are two very abrupt transitions, at the beginning of 39 (to the alleged joke about the statue of Chrysippus) and at the end of 41, where Cicero returns to the topic of 29 - transitions that make one suspect that some link has been left out. I do not wish to claim that Cicero was paraphrasing a particular Epicurean source; all I would want to say is that he is presenting a continuous line of thought in a condensed form while adding some unnecessary but rhetorically appealing flourishes. If
D
GISELA STRIKER
one sets aside these Ciceronian embellishments, the following structure seems to emerge: (a) 29-30: Torquatus establishes that pleasure is good, pain is bad, and that these affections are nature's only guides to action. (b) 32-3: He then deals with a possible misunderstanding of the claim that one should be guided by pleasure and pain. It turns out that one has to forego some pleasures to avoid pain, and to accept some pains for the sake of subsequent pleasures. So the aim is to attain the most pleasure and the least amount of pain in the long run. (c) 37-8: All misunderstandings about the sober and admirable character of Epicureanism are finally dispelled by an explanation of what Epicurus means by 'the greatest pleasure', namely complete absence of all pain. (d) 41-2: Finally, from the claim that all 'right and praiseworthy things are referred to living pleasurably', it is to be inferred that a pleasant life must be the highest good. Notice that this conclusion is reached only after Epicurus' conception of the greatest pleasure has been introduced. This is, I believe, as it should be. For the initial argument at 29-30 merely shows that pleasure and pain should guide our choices; it does not explain what is to count as the highest good, 'what everything ought to be 1 referred to, while it itself is not referred to anything else'. Rather, from the claim that pleasure and pain are the only guides to action we need to proceed through an account of how to apply that thesis to the conclusion that the reference-point of all action is what Cicero calls 'a pleasant life', and hence that this must be the final end. I will now discuss this line of argument in slightly more detail to show how it might help to answer some of Cicero's objections. PLEASURE AS THE CRITERION OF CHOICE AND AVOIDANCE
The first part of Torquatus' exposition (29-30) offers an epistemological argument for the two propositions (i) that pleasure is good, pain is bad; and (ii) that the feelings of pleasure and pain are our 1
Cicero is using a Stoic definition of the term TEAOS, which may well have been common property between the schools at his time. However, I doubt that it is correct for Epicurus, who seems to have been, in modern parlance, a 'psychological hedonist', maintaining that people do in fact aim at pleasure in everything they do, whether they acknowledge it or not (cf. Men. 128, 129), and not that they are in some sense obliged to do so. The statement that 'all right and praiseworthy things are referred to this, that one should live pleasurably' at
Epicurean hedonism
7
only means of deciding what to pursue and what to avoid - that is, in the ancient terminology, that they are the criteria of choice and avoidance (cf. Diog. Laer. x.31, 34). Thesis (i) is supported by pointing out that the judgments 'pleasure is good' and 'pain is bad' are self-evident truths based on perception, and hence stand in no need of any proof.2 Next, it is argued that 'what is in accordance with nature or against it', that is, good or bad 3 for a living thing, must be judged by nature herself. This means, as the following sentence shows, that good or bad must be judged by the senses. But the only perceptions that can serve to guide choice and avoidance are pleasure and pain; so these must be the goods and evils that guide our decisions. I must confess that the argument for the premise that good and bad must be judged by the senses is obscure to me. The text runs: Nam quoniam detractis de homine sensibus reliqui nihil est, necesse est quod aut ad naturam aut contra sit a natura ipsa iudicari. I do not see how the fact,
if it is one, that sense perception is necessary and perhaps sufficient for life has any tendency to show that the senses must be the judges of what is good and bad. 4 One might be inclined to believe that animals that lack reason must rely on their senses to discover what to pursue or avoid; but why should this also be true of humans? Cicero rejects this claim with contempt, and at least as far as this argument goes, he may be justified. Stronger support for thesis (ii) seems to surface at the end of 41: nee enim habet nostra mens quicquam, ubi consistat tamquam in extremo, omnesque et metus et aegritudines ad dolorem referuntur, nee praeterea est res ulla, quae sua natura aut sollicitare possit aut angere.
This might be an appeal to inconceivability, saying that it is inconceivable how anything could be bad that is not either pain or a cause of pain. Presumably the Epicureans held that it is equally
2
3
4
Fin. 1.42 sounds rather ludicrous and is, I think, misleading. The Epicureans did hold that virtue is desirable because it is necessary for the most pleasant life, but this does not commit them to the view that it is also praiseworthy for that reason. This is, of course, not presented as a proof for (i), which would be self-defeating, but as an argument to show that (i) must be true, given its status. For this part of the argument, see the excellent treatment by Brunschwig (1986). For the equation of what is according to nature with what is good, see e.g. Men. 129; Fin. 11.5, and the curious 'syllogism' criticized by Alex. Aphr. (in Top. 9, Us. Fr. 404): r\ f)5ovf) Kcrra 4>UCTIV /'TO Korra <j>Gaiv aipeTov / f] fj8ovf) <5cyaO6v. A. Laks suggests that the words 'detractis de homine sensibus reliqui nihil est' could mean, not that sense perception is necessary for life, but that if we abstract from the senses, no means of judgment will be left. Perhaps that is intended; but it seems to me to make the argument even weaker, since as far as I can see, even though reason is not independent of the senses, sense perception is not supposed to be our only means of judgment.
8
GISELA STRIKER
impossible to conceive of any good that is not either a pleasure or a cause of pleasure. The point would then be that our notions of what is good and bad derive from (sensory) pleasure and pain, and that whatever else is called good or bad must be so called by reference to these. Or, in the terminology of Diogenes Laertius' summary (x. 33), pleasure and pain are the 'first subjects' of the words 'good' and 'bad'. It is clear from many of our sources that the Epicureans did hold such a view. The point of mentioning the fundamental role of sense perception for life in this connection might be to defend the doctrine by insisting that good and bad, what is in accordance with nature and against it, are so basic that we cannot take these notions to be constructed by reason later in life. It might be interesting to remember in this context that the Stoics made a distinction between what is in accordance with nature (kata phusin) and what is good, and hence were able to say that the concept of the good is arrived at only at a rather mature age. However this may be, the two propositions that pleasure is good and that it is the only criterion of choice seem to constitute Epicurus' argument for the main thesis of hedonism, 'pleasure is the good', which can now perhaps be paraphrased more precisely as 'pleasure is the only thing that is good in itself. This is not, however, an argument that shows us what is to count as the final good, or the end of all action. For what has been established so far is only that every action must aim at some pleasure or the avoidance of some pain. It follows, indeed, that if there is a highest good, it must be pleasure. But this leaves us with a wide range of options. Are we to suppose that the final good is some specific pleasure? The greatest pleasure? Or the largest possible amount of pleasures? Saying that pleasure and pain must guide our actions will be no help before we know how to apply these criteria of choice and avoidance. Now one might think that the conception of the final good implied by the hedonist thesis is too obvious to need explicit statement. Happiness is usually assumed to be something that belongs to a whole life, 'living well' or 'acting well', as Aristotle says (i?jV 1.4. 1095a 19-20). So if the good is pleasure, surely the best life will be the most pleasant life? This seems clearly to be the view of Cicero, who feels free to refer to the Epicurean end as 'living pleasurably' (42). But things are not so simple. First, of course, we do not have an account of what counts as the most pleasant life. But secondly, we might also wonder whether the seemingly obvious step from 'good'
Epicurean hedonism
9
to 'good life' will lead us in the right direction. For one might argue, with the Cyrenaics (D.L. 11.87-8), that the good life is desired, not for its own sake, but for the sake of the pleasures contained in it. But that would show, as the Cyrenaics also recognized, that happiness is, after all, not the ultimate end desired only for its own sake. In other words, a hedonist who wants to hold that all intrinsic goods are pleasures will either have to argue that happiness is itself a pleasure, or else to give up a fundamental thesis of Hellenistic ethics, the claim that happiness is the highest good for humans. The Cyrenaics were apparently willing to abandon this claim, paradoxical as that must have seemed; Epicurus, I suspect, tried the other option by arguing that happiness is in fact identical with the greatest pleasure. Suppose one reasons, not as before that if pleasure is the good, a good life must be a pleasant life, but rather that, if pleasure is the good, the highest good must be the greatest pleasure. Then if happiness is the highest good, it must be the greatest pleasure. This sort of consideration seems to me to lie behind the next steps in Cicero's argument. By refuting the 'misunderstandings' of those who think that hedonism is an invitation to self-indulgence and luxury, Torquatus explains how the Epicureans understand the thesis that choice and avoidance must be guided by pleasure and pain. He first points out that one ought to aim at the greatest overall pleasure (rather than always choose the most pleasant option in a given situation). He then tells us what constitutes the greatest pleasure according to Epicurus - namely, absence of all pain and trouble. If we are to aim at maximizing pleasure, then this should be our goal, and a particular course of action should be undertaken only if it will lead to this end. It is this highest good, I take it, to which Cicero refers (or ought to be referring) when he arrives at the conclusion. Since this is what all action is referred to, it must be the telos. Unfortunately Cicero obliterates this point, I think, by substituting 'iucunde vivere' for the Epicurean definition of the final end, aponia and ataraxia. What we seem to get here is an inference from 'the greatest pleasure is the reference-point of all action' to 'the greatest pleasure is the highest good'. I am not sure that this rather formal transition from criterion for choice ('reference-point') to goal of life, helped along no doubt by the ambiguity of the blanket-term 'referre' (anapherein) , 5 comes directly from Epicurus, though Men. 129 comes 5
This term is used in at least three distinct senses. (i) When all evils are said to be 'referred to' pain, this means that the badness of all bad things lies in their either being pains or bringing
10
GISELA STRIKER
very close, but the doctrine seems to me to be orthodox. The criterion of choice and avoidance for Epicurus is indeed the telos, freedom from pain and trouble, and not, as one might expect on the analogy with his 'criteria of truth', individual occurrences of pleasure and pain. Once one sees this point, confirmatory evidence from Epicurean sources is not hard to find.6 Individual affections indicate the presence of a good or evil, and in this role they may serve, like sense impressions, as criteria of truth, providing irrefutable evidence of a rather limited range of facts. But they are not, strictly speaking, themselves criteria of choice and avoidance. By 'referring' everything to pleasure, Epicureans do not try to confirm or disconfirm the belief, say, that this piece of cake should be eaten by calculating the amount of pleasure (and subsequent stomach ache) it is likely to provide. Instead, they will presumably use the notorious Epicurean doctrine of necessary and unnecessary desires to determine whether they should indulge themselves on this occasion or not. It is somewhat misleading of Epicurus to describe this procedure as 'judging all good by the yardstick of [pleasure]' (Men. 129), since the decision does not really seem to depend on a comparison of the amounts of pleasure or pain involved in particular courses of action. Rather, the Epicurean's choice is based on a consideration of the best way to attain a state of freedom from pain and trouble that has independently been established to be the greatest pleasure possible. But one can understand why Epicurus kept the terminology of'measuring' if one considers that, like the calculating hedonist of Plato's Protagoras, he recommends skipping some pleasures in view of their unpleasant consequences, and accepting some pains for the sake of subsequent pleasures. Furthermore, we are formally exhorted to aim at the greatest pleasure - only it turns out that the greatest pleasure is attained not by accumulating as many smaller pleasures as possible, but by aiming directly at a state of complete pleasure that cannot be surpassed.
6
pain. (2) Beliefs are said to be 'referred to' the criteria of truth, sense impressions and preconceptions, in the sense that their truth or falsity is judged by whether they are confirmed or disconfirmed by the criteria (cf. e.g. Hdt. 37—8; 63, 72). (3) All action is 'referred to' the final end in the sense that all action ultimately aims at this goal. One might have expected pleasure and pain to be criteria in a sense analogous to (2), so that what is pleasant or leads to pleasure is to be chosen, etc. But Cicero's transition from criterion to final end makes sense only if the criterion of action is Particularly clear examples are, I think, PD 22 and Sent. Vat. 25.
Epicurean hedonism
11
THE NATURE OF PLEASURE
Let me set aside the vexed question whether a longer period of complete pleasure counts for more than a shorter one, and turn to a problem that I have deliberately left out in considering the argument about the criterion of action, namely the account of pleasure that allows Epicurus to claim that complete freedom from pain and trouble is indeed the greatest pleasure, and identical with happiness. In Cicero's text, this comes in a tantalizingly brief and obscure passage, 37-8. I have tried to establish its place in the longer argument first because when one considers this passage in isolation, one might be led to rather different conclusions about what Torquatus is trying to do. After announcing an explanation of the nature of pleasure (voluptas ipsa quae qualisque sit) that will vindicate Epicureanism against its detractors, Torquatus states that the Epicureans pursue not only sensory pleasures, but hold that the greatest pleasure is perceived when all pain is gone. The contrast here is decidedly odd, but I will come back to that later. Torquatus then produces the following argument: (1) When we are freed from pain, we enjoy the actual freedom and absence of all distress; (2) but everything that we enjoy is pleasure, just as everything that distresses us is pain; hence (3) complete removal of pain (or: every removal?; doloris omnis privatio) has rightly been called pleasure. (4) Quite generally, the removal of distress causes pleasure to take its place. (5) Hence (Epicurus held that) there is no intermediate state between pleasure and pain; for (6) whoever is aware of the way in which he is affected will necessarily be either in pleasure or in pain. (Therefore?) Epicurus thinks that (7) the greatest pleasure consists in the absence of all pain, so that thereafter pleasure can be varied and differentiated, but not heightened or increased. In the first and third propositions, it is unclear whether the subject is the process of being relieved from pain {liberation privatio) or the state of absence of pain (vacuitas molestiae). One might be inclined to opt for the process, since otherwise the step from (1) and (2) to (3) would
12
GISELA STRIKER
seem to be dubious. Even if in some cases pleasure consists in absence of pain, it is not clear that absence of pain must always constitute a pleasure. Might it not be the case that absence of pain is enjoyable only if immediately preceded by pain? On the other hand, it seems plausible to think that the process of being freed from pain is always a pleasure. So perhaps, as many commentators seem to think, Epicurus is talking only about the process of removal of pain. But then (4) becomes problematic as a generalization of what precedes, since it seems to say not that the removal of pain is identical with, but that it is followed by pleasure. The idea that pleasure must be present whenever pain is absent is also used in the following step to deny that there can be a state when both are absent. The next statement (6), also supporting the claim that there is no neutral state, would seem to require the same assumption. If we understand (6) to say that any affection will be perceived as being either pleasant or painful,7 we must take it that every case of painless affection counts as pleasant. The point of (6) would be that since we are always affected in one or the other way, there can be no state of awareness that is neither pleasant nor painful. Hence it seems preferable to take the first premise also to be about absence of pain. The argument up to this point would then proceed from the thesis that absence of pain is a pleasure to the conclusion that there is no intermediate state. So far, so good - but unfortunately, what we seem to need for the next step is not just the thesis that absence of pain is (a) pleasure, but its converse - all pleasure is absence of pain. If that had been shown, Epicurus' thesis about the greatest pleasure would indeed follow. For since painlessness does not admit of degrees, it could be greater or less only in respect of the partial or 7
If this proposition is meant to support (5), as seems to be the case (see 'enim'), it must be understood to say that living creatures are either in pain or in pleasure whenever they are perceiving anything at all. But it seems to say less than that, since one need not always be aware of the way in which one is affected. However, perhaps we can take it to make the stronger claim if we assume that being consciously affected by anything is sufficient for being aware of the way in which one is affected. This might be true in the sense that one will always be able to tell without further observation, though not in the sense that one will always consciously notice whether one is affected in one or the other way. When Aristotle says (de An. m 2. 425b 12) that 'we perceive that we see and hear', he is making a similar point - not that we may, on occasion, notice that we perceive, but rather that when we do perceive, we know that we do, although we need not explicitly notice it. I assume, then, that what is being said here amounts to 'whoever is aware of anything that affects her in some way', rather than 'whoever notices the way in which she is being affected'. We do not need to saddle Epicurus with the implausible claim that we are always conscious of being either in pleasure or in pain. Pleasure requires consciousness, as Plato pointed out (Phlb. 2iB-c), but it does not, I think, require constant self-reflection.
Epicurean hedonism
13
complete absence of pain. Perhaps, then, this should lead us to revise our interpretation of the earlier lines and read the whole passage, in accordance with Torquatus' announcement, as concerned with the nature of pleasure in general? This might be done by adopting a stronger reading for (3): 'complete absence of pain is what is rightly called pleasure' [rede nominata est voluptas). Instead of seeing the argument as an attempt to establish that absence of pain is (a kind of) pleasure, and hence that there is no neutral state, we could then take it as trying to show that complete absence of all pain is the greatest pleasure,8 on the basis of the assumption that absence of pain is what pleasure consists in. Admittedly, this only shifts the difficulty with this passage from one step to another; for while the ostensible conclusion would indeed follow, we would have to accept a rather rash generalization from one sort of case to all others in the transition from (1) and (2) to (3): since when pain is removed, what we enjoy is its absence, and whatever we enjoy is pleasure, absence of pain is [what constitutes] pleasure. But this reading of the argument has the advantage that it provides, however inadequately, what is needed to derive Epicurus' thesis about the greatest pleasure that is supposed to be the final good, whereas an argument to the effect that absence of pain is a kind of pleasure seems to lead nowhere. Or rather the result would be exactly what Cicero denounces in book n: Epicurus introduces a new kind of pleasure and then proceeds to claim that the end is the greatest pleasure of this kind. Now the thesis that pleasure consists in the absence of pain implies, of course, both that there can be no intermediate state and that the greatest pleasure consists in the absence of all pain, and it does not seem inconceivable that Cicero should have abbreviated a longer Epicurean argument to produce this somewhat incoherent passage.9 Since what is required in the 8
As A. Laks has pointed out to me, this passage seems to talk only about sensory awareness, and hence can hardly be said to deal with <5tTapa£icx as well as cnrovia. But is this true? Cicero's language seems to be neutral between sensory and mental awareness. The verb 'gaudere' can be used neutrally, as corresponding to 'voluptas' (cf. 11.41), while 'gaudium' even seems to be reserved for mental pleasures (11.13). iSentire> is not restricted to sense perception (see e.g. 1.55: 'corpore nihil nisipraesens . . . sentirepossumus; animo autem etpraeterita et future?).
9
And of course the account of pleasure is exactly parallel for body and mind. So I am inclined to think that Cicero is talking about pleasure in general, and that 'absence of all pain' covers both dTOcpo^ia and carovioc. For example, the remarks about removal of pain might be a defence of a previously stated definition of pleasure as painless affection against the objection that some pleasures involve pain. The point would be to insist that what is enjoyed in those cases is painless perception,
14
GISELA STRIKER
context of an exposition of the Epicurean doctrine about the telos is a statement about the nature of pleasure - as indeed announced by Torquatus - and not the denial of a neutral state, I am inclined to conclude that Epicurus held that all pleasure consists in absence of pain, or rather, as (6) seems to indicate, in painless affection; and that he argued from this to his definition of the end, absence of all pain from body and soul. But what, then, becomes of the distinction between kinetic and katastematic pleasures to which Cicero obviously alludes at the beginning of this section? Here we should note the strangely illogical contrast by which the distinction is introduced: 'not only' . . . 'but the greatest'. A passage in Epicurus' Letter to Menoeceus (131) seems to draw the contrast that Cicero presumably tried to capture; and it suggests that on this occasion Cicero might indeed have missed the point. When Epicurus explains that those who say that pleasure is the end do not go for the pleasures of profligates, but for absence of pain from body and soul, he is not, I think, rejecting kinetic in favour of katastematic pleasures. The contrast is not between different types of pleasures, but rather between different conceptions of the greatest pleasure - the misguided one of the luxury-seekers and the correct one of the Epicureans. Even the profligates are ultimately seeking freedom from pain and trouble, according to Epicurus; they just have the wrong idea about how this is to be achieved. This contrast, then, is consistent with the view that all pleasure is pain-free affection. But if all pleasure is held to consist in absence of pain, clearly we can no longer take this to be the distinguishing characteristic of katastematic pleasures alone. Some scholars 10 have suggested that the kinetic pleasures are the ones associated with processes of replenishment, removal of pain or, on the level of mental pleasures, desire-satisfaction, while the katastematic ones would be those of contentment or relief. The point of the distinction would be to emphasize, against earlier enemies of pleasure like the anonymous 'subtle thinkers' of Plato's Philebus, that pleasure need not come to an end when all pain is gone, since the ensuing states of contentment and relief are pleasures just as much as the removal of pain. This 10
which does not cease when the pain is removed. According to PD 3, Epicurus held that where and when there is pleasure, there is no pain. E.g. P. Mitsis (1988), 45. I cannot try to do justice here to the many ingenious attempts to make sense of this distinction. For a judicious survey, see Gosling & Taylor (1982) chs. 18 and 19.
Epicurean hedonism
15
provides no doubt a plausible explanation, but in the context of Epicurean theory, it would leave us with a classification that is clearly not exhaustive, omitting a large class of pleasures that Cicero, for one, explicitly declares to be kinetic. Eating when not hungry, admiring a beautiful statue or enjoying a surprise party are not cases of replenishment or satisfaction of antecedently felt desires, but they also do not seem to be states of relief or contentment. If these pleasures are not kinetic, then either the distinction is not meant to produce an exhaustive classification, or we must take it that they count as katastematic, and assume that this class includes all pleasures not preceded by a felt lack or a pain. This is the line that Gosling and Taylor take (1982, 37 iff.). According to their interpretation, the distinction is not one between processes or events and states, but rather between affections that go along with replenishments or desire-satisfaction as opposed to those that occur in an undisturbed state of well-functioning. Again the proposed distinction is plausible and has some noteworthy predecessors in earlier Greek philosophy.11 But it not only contradicts Cicero's account in Fin. 11, it also does not fit the one literal quotation from Epicurus in which he seems to use the terminology, Diog. Laer. x.136, in contrasting the katastematic pleasures oi aponia and ataraxia with joy {chard) and delight {euphrosune) as being 'in motion'. The text runs: 'he men gar ataraxia kai aponia katastematikai eisin hedonai he de chara kai he euphrosune kata kinesin energeiai blepontaf; Long and Sedley translate:
'Freedom from disturbance and absence of pain are static pleasures; but joy and delight are regarded as kinetic activities.' 12 There is no good reason to think that joy is necessarily tied to the removal of pain. Since there is no indication that Epicurus recognized other kinds of pleasures besides katastematic and kinetic ones, one should perhaps try to follow the lead provided by the technical terms used, and look for a distinction between temporarily limited episodes or processes of pleasure and lasting states. But what would be the point of such a distinction? 11
12
As I understand it, the distinction coincides with Plato's distinction between 'mixed' and 'unmixed' pleasures in the Philebus. But the most impressive parallel is no doubt a passage from the Magna Moralia, 11.7. i2O5b2off.: 'eirei 5'ouv ICTTIV f\ f)8ovf) Kai Ka8ioTauevr|s TT\S 4>UCTEGOS Kai Ka6e<jTr|Kuias, otov KaOiaTauevris uev ai e£ evSeias avairAripcbaeis, Ka6eaTT|Kuias 6e ai OTTO TT\S oyecos Kai TT^S OKofis' (cf. also EN vn. 1 i52b2-6). Gosling & Taylor (p. 390) want to take X a P a a n d E0<J>poauvr| as 'terms giving the positive sides of the ataraxia and aponia coins', so that Epicurus would be 'pointing out . . . that katastematic pleasures are experienced'. This interpretation seems to me to be ruled out by
l6
GISELA STRIKER
One difference that suggest itself is this: while episodes or processes of pleasure consist in enjoying or being pleased with something, an activity or an event or an object, the states of bodily freedom from pain - the healthy state of the body, as Epicurus calls it in Men. 129 - and of peace of mind do not seem to be directed at anything in particular. Nor do they seem to be like pleasant sensations (which need not be enjoyments of anything else, but of which one would at least have to be aware for them to count as pleasures), and this makes it odd to call them pleasures in the first place. But obviously, if Epicurus is right about the pleasantness of painless affection, and life consists in being affected in one or the other way, then these states must contain innumerable pleasures of the kinetic sort, since presumably anything that affects us without disturbance would count as an object of pleasure. Thus variatio might be exactly what Cicero says it should be, Fin. 11.10: taking pleasure in a variety of different things. Indeed, one might try to justify calling the painless states pleasures by appealing to the kinetic pleasures that they necessarily contain. If the text were not so desperately difficult and quoted out of context, I would be inclined to interpret Epicurus' dictum about aponia and ataraxia in this sense: these are states of pleasure, but the joy and delight (sc. that come with them) manifest themselves in motion, that is, in particular episodes of pleasure and enjoyment. The claim that states of undisturbed perception and carefree thought are pleasures in themselves is strange, and one can see why other philosophers would have wanted to insist that one could not feel pleasure unless one were enjoying something that could be specified. But if one grants Epicurus the claim that pleasure is constituted by undisturbed affection, one might find it plausible to say that a hedonist should aim at those states of body and mind that make life enjoyable regardless of what one takes pleasure in, except for the few disturbances that will arise from unavoidable pains. And if the connection between states of freedom from pain and pleasurable experiences is as close as Epicurus' theory would make it out to be, one might also be willing to accept the terminological move of calling these states pleasures as well, albeit objectless ones. If this is correct, it would seem that there were exactly two katastematic pleasures, namely aponia and ataraxia, that together make up complete pleasure. Hence it might be misleading even to the particles (jjev . . . 8e) in the sentence, which clearly mark a contrast between and onTovioc on the one hand, X a P^ a n d eO<|>pocruvr| on the other.
Epicurean hedonism
17
speak of different kinds of pleasure. It might be better to say that Epicurus proposed to extend the use of the word 'pleasure' to cover not only episodes, but also states of body and mind, on the ground that what makes experiences pleasurable is precisely absence of pain, which can be a lasting state. The distinction between kinetic and katastematic pleasures would seem to derive from Epicurus' theory, rather than from an empirical survey of the various phenomena described as pleasures in everyday language. It was this terminological move that allowed Epicurus to identify the greatest pleasure with the good life.13 So I would suggest that he introduced the notion of katastematic pleasure in order to show that happiness can be the same as pleasure after all, provided that one is willing to accept as pleasures not only episodes of enjoyment, but also those states that, according to Epicurus, make one's life enjoyable at every moment. If we return now to Cicero's objections we can see, I think, that Epicurus' theory was not incoherent in the way Cicero maintains. The trouble is rather that both the thesis that pleasure is nothing but undisturbed affection and perhaps even more the claim that all mental pleasures are parasitic upon bodily ones are highly implausible. The Cyrenaics do seem to me to be the better hedonists in the sense that they paid more attention to the complicated and varied phenomena described as pleasures. But an impoverished account of pleasure may have been the price Epicurus was willing to pay in order to fit hedonism into the framework of Hellenistic ethics, while the Cyrenaics apparently decided to opt out of the competition. 13
Aristotle envisages a similar move at EN vn. 1153D9-14, when he says that the highest good could be considered to be a pleasure, given that it consists in unimpeded activity of a sort. Obviously, 'activity in accordance with virtue' is not on the same level as the various virtuous activities that make up a good life.
CHAPTER 2
Anniceris et les plaisirs psychiques. Quelques preatables doxographiques Andre Laks
I LE PROBLEME
Les temoignages doxographiques transmis sous le nom d'un mouvement philosophique, et non sous celui d'un auteur particulier, soulevent souvent de difficiles problemes d'attribution, d'autant plus delicats a trancher que nous possedons moins de reperes independants, dans l'histoire de l'ecole, par rapport auxquels les situer. Le cas des Cyrenaiques est a cet egard exemplaire. Les donnees sont le plus souvent presentees comme un bien commun; rinformation relative aux positions specifiques d'Aristippe l'Ancien, d'Aristippe le Jeune ou d'Anniceris est beaucoup plus rare, quand elle ne se reduit pas a un seul temoignage. La tache de la critique est alors de faire la part entre le generique et le particulier, l'originaire et le remaniement, et de reconstruire les phases d'une eventuelle evolution. De ce point de vue, la tendance de Interpretation s'est considerablement modifiee depuis E. Zeller, qui, sur la base du resume de la doctrine ethique generiquement attribute aux 'Cyrenaiques' dans Diogene Laerce, 11.86-93, avait fait d'Aristippe l'Ancien le createur d'un veritable systeme ethique.1 Au terme d'un Le document de travail presente au colloque de Syam a ete elabore au cours de mon sejour au Center for Hellenic Studies, Washington D.C. en 1987/8. La discussion m'a permis de mieux cerner plusieurs points, en particulier grace aux interventions de Jonathan Barnes, Jacques Brunschwig, Michael Frede, Jaap Mansfeld, David Sedley, Richard Sorabji, Voula Tsouna. Je remercie tout particulierement Gisela Striker d'avoir formule, sur un etat intermediate, des critiques ecrites, dont la presente version a beneficie. Les temoignages sur les Cyrenaiques sont principalement cites d'apres la collection de Giannantoni (1983/85), par simple numero suivi de G. quand il s'agit des temoignages que Giannantoni edite sous le nom d'Aristippe l'Ancien (= iv A, selon le classement adopte), par l'abreviation complete, quand il s'agit d'autres membres de l'ecole (iv B = Arete et Aristippe le Jeune, par exemple). II est aussi fait reference a l'edition de Mannebach (1961) = M. L'abreviation D. L. suivie d'un chiffre renvoie au decoupage que je propose des §§86-91 a de Diogene Laerce 11 (cf. infra, pp. 25-6). 1
Zeller (1922), p. 344, avec la n. 1. 18
Anniceris et les plaisirs psychiques
19
long mouvement de revision, qui debute avec la dissertation d'E. Antoniades (1916), G. Giannantoni n'a laisse a Aristippe l'Ancien que la pratique d'un mode de vie, reservant l'elaboration theorique de positions doctrinales a un stade plus tardif,2 qu'un temoignage transmis par Eusebe permet d'associer au nom d'Aristippe le Jeune. 3 Dans Petude d'ensemble la plus recente sur l'ecole cyrenaique, 4 K. Doring developpe en substance une interpretation moyenne, dont l'edition d'E. Mannebach, reprenant les indications d'E. Schwartz et de K. von Fritz, peut aussi etre consideree comme le reflet:5 tout en recusant la position trop genereuse de Zeller, K. Doring fait remonter au premier Aristippe ce qu'il considere, dans la section generique de la doxographie laertienne, comme un noyau ancien.6 Simultanement, il repere, dans cette meme section, une serie de notices correspondant a ce qu'il juge etre un etat posterieur, en 1'occurrence hellenistique, et plus precisement annicerienne, de la doctrine cyrenai'que. De fait, la polemique 2
3
4 5 6
La demonstration de Giannantoni remonte a son recueil de 1958, pp. 74-115. Les conclusions en sont reprises, avec quelques ajustements, dans Giannantoni (1983/85), m no. 18, pp.164-170. Eusebe, Preparation Evangelique xiv. 18.31-32 = 173 et iv. B. 5 G (tres vraisemblablement un extrait du fTepi <j)iXoao<|)ias d'Aristocles = fr. 3 Heiland). Voici la traduction integrate du passage, auquel il sera plusieurs fois fait reference par la suite (cf. en particulier infra, n. 11, X 8, 56, 57, 105, et, pour la derniere phrase, non traduite ici, n. 31): 'Aristippe, celui qui a fonde l'ecole appelee Cyrenaique, d'ou Epicure a tire son point de depart pour exposer la fin, etait un compagnon de Socrate. Aristippe etait dans sa vie un parfait voluptueux et aimait le plaisir. Cependant, il ne s'entretint pas au grand jour au sujet de la fin, mais, en puissance, il affirmait que l'existence du bonheur reside dans les plaisirs [aAA' ouSev uev OUTOS £v Tcp 9ccv6pcf> Trspi TEAOUS SIEX^OTTO, 8uvau6t 8e TT\S e08cunovias TTJV C/TtoaTaaiv IXeysv Iv f)8ovaTs KeTaSai]. Car, comme il parlait toujours de plaisir, il amenait ceux qui le frequentaient a soupconner qu'il affirmait que vivre dans le plaisir est la fin. II eut pour auditeurs, avec d'autres, sa fille Arete; celle-ci, ayant donne naissance a un fils, l'appela Aristippe, qui fut surnomme "Metrodidacte", parce que c'est sous sa conduite qu'il fut amene aux arguments de la philosophic Celui-ci definit clairement que vivre dans le plaisir etait la fin, en inserant le plaisir qui est en mouvement [6s KCU aa(pcbs cbpio-cnro TEAOS eTvoa TO f|8ecos ^ v , f)6ovfiv evTorrTcov TT)V KOCTOC Kivr|atv]. Car il affirmait qu'il existe trois etats (Tpels . . . KOCTacTT&aeis) relatifs a notre constitution; le premier en vertu duquel nous souffrons, et qui ressemble a la tempete sur la mer; le second en vertu duquel nous eprouvons du plaisir, et qui est comparable a la vague douce: car le plaisir est un mouvement doux, analogue a un bon vent [sTspctv 8s KOC0' f^v f)86ueSa, TCO Xeico KUMOCTI dq>o;jioiou[ji6vr|v, slvai yap Aeicxv KIVTICTIV TT^V f)8ovfiv, oupicp TrapapaAAo|jevr|v dveiico]; le troisieme est un etat median, en vertu duquel nous ne souffrons ni n'eprouvons de plaisir, et qui est similaire a une mer d'huile.' Doring (1988). Mannebach (1961), appendice 111 (De ratione in placitis digerendis inita), p. 106-10; cf. Schwartz (1903), col. 741, et von Fritz (1934), cols. 1827s. Dans la premiere partie de son etude, Doring (1988) developpe une these parallele pour ce qui est de la theorie de la connaissance, en montrant qu'elle se laisse inscrire dans le contexte des discussions socratiques (voir en particulier pp. 27SS.). Je n'aborderai pas ici cet aspect de la question.
20
ANDRE LAKS
anti-epicurienne dont plusieurs de ces passages, et d'autres dans la Vie d'Epicure, se font l'echo7 ne saurait preceder la fin du quatrieme siecle.8 Et une notice specifiquement consacree aux Anniceriens, que transmet Clement d'Alexandrie, recoupe parfaitement une serie de doctrines que Diogene Laerce attribue aux Cyrenaiques en general.9 Alors que le role d'Aristippe l'Ancien dans la fondation de l'ecole cyrenaique a ete abondamment discute, il n'existe que peu d'etudes sur la phase 'annicerienne' de l'ecole cyrenaique. C'est sur ce dernier probleme que portent les remarques qui suivent. Contre le point de vue 'analytique' represente maintenant par K. Doring, je defendrai une perspective que Ton peut qualifier d'unitarienne, en tentant de montrer, sur la base d'un reexamen de la structure d'ensemble de Pexpose de Diogene Laerce et de l'analyse du passage 'parallele' de Clement, pourquoi ni la logique des contenus doctrinaux, ni les mentions d'Epicure dans l'expose de la doctrine cyrenaique, n'impliquent que la doxographie de Diogene Laerce resulte, comme Doring le pense, d'une combinaison entre un etat plus ancien (ante-epicurien) de la doctrine, et une version revisee, qui tiendrait compte de la situation creee par l'enseignement du Jardin et serait a mettre au compte d'Anniceris.10 II convient d'emblee de souligner que la perspective analytique adoptee par K. Doring a au premier abord quelque chose de seduisant. De nombreux temoignages anciens situent la doctrine des Cyrenaiques a l'origine de la theorie morale d'Epicure, qui en aurait adopte le principe (le plaisir est la fin de nos actions), tout en recusant la conception du plaisir qui en etait le correlat: a cote du plaisir en mouvement (cinetique), Epicure admet un plaisir en repos (catastematique).11 Cette difference, se repercutant sur l'ensemble 7 8
9
10 11
Ils'agitessentiellementdeD.L. 4, 8, 10. Cf. Diog. Laer. x.136; Clement, .ftr077zato11.21.13o, 7-8 (cite infra, n.ioo); voir aussi infra, n.90. Sur Diog. Laer. x.136, cf. Giusta (1962/3). Les indications chronologiques concernant la phase 'hellenistique' de l'ecole cyrenaique font presque entierement defaut, sauf dans le cas de Theodore, dont on fixe la naissance autour de 340, sur la base d'une serie d'indications convergentes (voir von Fritz (1934), col. 1825). On situe generalement la naissance d'Aristippe le Jeune dans les annees 60 du quatrieme siecle (non ante annum 370, Mannebach (1961), p. 89, ad fr. 136-7). Hegesias et Anniceris semblent avoir ete plus ou moins contemporains de Theodore. Nous ne pouvons done operer que sur la base de chronologies relatives et d'une reconstruction des dependances intellectuelles. Cf. aussi infra, n. 20. La notice de Clement, qui avait deja retenu l'attention de Schwartz (1903) et de von Fritz (1934), joue pour Anniceris un role analogue a celle d'Eusebe pour Aristippe le Jeune (cf. supra, n. 3). Elle est citee et analysee infra, pp. 41-5. II ne peut ici s'agir que d'esquisser la demonstration, qui exigerait, pour etre complete, une monographic analogue a celle qu'a livree Doring. La plupart des temoignages figurent sous les rubriques 178-86, 189 dans Giannantoni (1983/85). Cf. aussi 173 (cite supra, n. 3).
Anniceris et les plaisirs psychiques
21
du systeme ethique, eloigne considerablement Phedonisme d'Epicure de celui des Cyrenaiques.12 II est naturel de penser que ces derniers, face a la concurrence que representait un systeme frere, reagirent. On les voit assez bien developper, contre la notion de plaisir catastematique, des critiques du type de celles que nous a conservees Ciceron (Epicure, en nommant 'plaisir' l'absence de douleur, donne un sens arbitraire aux mots).13 Comme je l'ai deja indique, les formules doxographiques qui opposent les theses d'Epicure a celles des Cyrenai'ques sur le theme du plaisir sont certainement le reflet de cet affrontement. Comment ne serait-on pas tente d'en conclure que le resume 'generique' de Diogene Laerce, la ou il cite, ou suppose, Epicure, reproduit un etat de la doctrine cyrenai'que renouvele et modifie par la confrontation avec le Jardin? Un temoignage de Strabon semble confirmer l'hypothese, puisqu'il nous apprend qu'Anniceris passait pour avoir 'corrige' l'ecole cyrenaique et lui avoir substitue la sienne propre. 14 Cet 'amendement' ne consista-t-il pas precisement a reviser la doctrine cyrenai'que a la lumiere des objections qu'Epicure avait formulees, explicitement ou non, a l'encontre d'une version plus primitive de la doctrine? Telle est, en substance, la ligne d'argumentation developpee par Doring. 1 5 Je voudrais lui opposer (1) que la polemique anti-epicurienne d'Anniceris pouvait parfaitement se nourrir de la doctrine cyrenai'que traditionnelle, et que celle-ci est representee, de fagon consistante, dans la section generique de la doxographie de Diogene Laerce;16 (2) que par consequent la reference a la doctrine d'Epicure dans la presentation de la doctrine cyrenai'que (et inversement) est l'effet d'une construction doxographique, beaucoup plus que le 12 13 14
15 16
Au point qu'on a pu parler d'Epicure comme d'un 'hedoniste malgre lui' (M. Hossenfelder (1986)). La formule, a mon sens, indique moins une solution qu'elle ne nomme un probleme. Ciceron, de Finibus, 11.7.20 (= 183 G). Strabon, xvn. 3.22 ( = IVG, I G): (dans une liste des personalites illustres de Cyrene) 'et Anniceris, qui passe pour avoir corrige [6 SOKCOV iTravopScoaai] l'ecole Cyrenaique et avoir introduit a sa place l'annicerienne'. Sur la nuance de eTTavopSwacci et le dernier membre de phrase, cf. infra, n. 126 et 127. Doring (1988), pp. 52s. Je ne discuterai pas la question de savoir ce qui revient, dans cet etat ante-epicurien de la doctrine, a Aristippe l'Ancien et a Aristippe le Jeune. II me semble neanmoins raisonnable d'admettre, avec Giannantoni (cf. supra, n. 2), que la doctrine cyrenaique proprement dite est le fait d'Aristippe le Jeune (cf. aussi von Fritz (1934), col. 1827). Par ailleurs, le fait que le contenu des notices dans Diogene Laerce remonte a ce dernier n'empeche pas la presence dans le resume de termes hellenistiques (en particulier stoi'ciens), s'il est vrai que nous avons affaire a une construction doxographique (cf. infra, n. 17). C'est par exemple le cas de 5iao-Tpo
f|, D.L. 9 (cf. SVF 11 228ss). Pour oiKeioOcrSai (D.L. 6a) et d'autres termes, cf. Mannebach (1961), pp. 111—13.
22
ANDRE LAKS
reflet direct du debat (qui dut en tout etat de cause voir lieu); 17 (3) enfin que le renouvellement annicerien dont parle Strabon a toutes les chances de s'enraciner non dans la polemique anti-epicurienne, mais plutot dans une reaction a Pheresie, interne a l'ecole cyrenaique, d'Hegesias. Cette derniere hypothese s'appuie sur les indications que nous livrent les deux sections specifiquement consacrees, chez Diogene Laerce, a Hegesias et a Anniceris. II LA COMPOSITION DE LA SECTION CYRENAiQUE CHEZ DIOGENE LAERCE
II sera ici utile de rappeler la structure d'ensemble de la section du livre 11 que Diogene Laerce consacre aux Cyrenai'ques. La vie d'Aristippe (65~85a) est presque exclusivement composee d'anecdotes illustrant le mode de vie du disciple de Socrate (66-83a). Elle se clot sur la definition de la fin (la seule phrase theorique de cet ensemble): 'il affirmait qu'est lafinle mouvement lisse qui parvient a la sensation'.18 La doctrine cyrenai'que, du moins pour ce qui est de sa partie ethique, a laquelle Diogene Laerce se limite,19 n'est developpee qu'en relation avec les disciples. Apres un paragraphe de transition (85b-86a) qui dessine l'arbre des successions,20 sont 17 18
19
20
J e reprends en cela l'hypothese formulee par Bollack (1975), pp. 147s, a la suite de Giusta (1962/63), pp. 121, 172, et (1964/67), 1, pp. 137s. TEAOS 8'&TT6<j>aiv£ TT)V AEiav KIVT|O"IV sis aiaSriaav ava6i5o|i£vr|v. L'emploi technique du terme TEAOS indique tres vraisemblablement qu'il s'agit d'une attribution retrospective, au pere fondateur, du principe fondamental (parler avec Giannantoni (1983/85), in, Note 18, p. 166, d"erreur ou de confusion' ne suffit sans doute pas). C'est ce que confirme le temoignage d'Eusebe quand il prete a Aristippe l'Ancien de n'avoir dit qu"en puissance' (6uvd|i£i) ce que son petit-fils devait efTectivement dire (cf. supra, n. 3). Dans la formule pretee a Aristippe l'Ancien, 'mouvement lisse' doit designer le plaisir, comme le montre D.L. 1. La precision eis aiaSi^aiv <5cva8i5o|ievr|v pourrait impliquer qu'un mouvement trop faible pour parvenir a la conscience ne donne pas lieu a un plaisir. On trouve seulement, en Diog. Laer. 11. 92, un rappel du principe epistemologique fondamental en vertu duquel 'seules les affections peuvent etre apprehendees' (l'expression de Diogene Laerce est contournee: TOC TE TT&ST) KaTaAr|7TTd. iAeyov oOv ocCrrd, OUK 69' cov yivETai). Encore faut-il ajouter que la phrase se presente moins comme une these a developper que comme l'explication du fait que les Cyrenaiques se desinteressaient de la physique. Elle sert ainsi, a contrario, a souligner le primat de l'ethique. S c h e m a d e Diog. L a e r . 11. 85b—86a: ^^^-—~~ Arete
Aristippe L'Ancien ^ ^ ^ ^ Aithiops
^^~-~- Antipater
Aristippe le Jeune
Epitimides
Theodore
Paraibates Hegesias Anniceris
Anniceris et les plaisirs psychiques
23
tour a tour presentees: (1) les doctrines de 'ceux qui sont restes attaches a Pecole d'Aristippe' (86b-93a), 21 (2) les doctrines d'Hegesias et de ses disciples (93b-g6a), (3) celles d'Anniceris et de ses disciples (g6b-97a), (4) celles enfin de Theodore et de ses disciples (97b-iO3). 22 Cette derniere section se distingue des trois precedentes en ce qu'elle contient tous les elements d'une vita analogue a celle d'Aristippe, avec, en particulier, une serie d'anecdotes et d'apophtegmes conferant a Theodore une personnalite. L'ensemble peut etre
21
22
Le paragraphe comporte plusieurs confusions, cf. Antoniades (1916), pp. i8ss, ainsi que Giannantoni (1983/85), m, n. 17, p. 158. L'impression chronologique que cree le rattachement de Theodore a Aristippe le Jeune, en face des deux 'generations' representees par Epitimides et Paraibates, est corrigee au §98, qui presente Theodore comme un disciple d'Anniceris (et de Denys le dialecticien). Pour une tentative de conciliation, cf. von Fritz (1934), col. 1829. II est vraisemblable que nous ayons affaire, comme souvent, a des constructions chronologiques a visee legitimatrice. o i . . . ETTI TT\S dycoyns TT\S 'ApicrriTrrrou UEIVOVTES. ITTI est dans les manuscrits, contrairement aux indications de l'edition de Long (cf. Bollack (1975), p. 190); on retrouve l'expression au debut du livre iv, quand la fidelite de Speusippe a l'egard des doctrines de Platon est contrastee avec la difference de caractere (KOCI EUEIVE UEV ETTI TCOV OCUTGOV fTAaTcovi Soyuonxov). Le sens d'<5cycoyf| est debattu; peut-on y voir l'indice qu'Aristippe l'Ancien n'avait pas fonde d"ecole' (ce qui suppose un enseignement), comme Pa soutenu en particulier Giannantoni (1958), pp. 70-3 (cf. aussi le point de la discussion dans Giannantoni (1983/85), in, Note 17)? Doring (1988), pp. 34s, montre bien qu'il n'existe pas de distinction terminologique absolue entre dcycoyf) et ocTpecns. II est vrai que, employes dans un meme contexte, les deux termes peuvent s'opposer comme le 'mode de vie' a la 'doctrine'. Mais ce n'est pas necessairement le cas. Pour ce qui est de notre passage, il me semble que la 'fidelite' se comprend mieux s'il s'agit d'un contenu doctrinal que d'un mode de vie. Naturellement, meme ainsi, il est possible que l'attribution d'un enseignement a Aristippe l'Ancien, aussi minimal soit-il, resulte d'une construction posterieure (cf. supra, n. 18). Quant a savoir qui sont les 'fideles', il est difficile de le dire avec exactitude. Mais les candidats ne manquent pas. A cote des chefs de file, Arete et Aristippe le Jeune, nous connaissons les noms d'Aithiops, d'Antipater, d'Epitimides, de Parabaites. Diogene Laerce cite aussi en 11. 113 Aristote de Cyrene, parce que, comme Theophraste et les Dialecticiens, il fit les frais, en tant que chef d'Ecole, de la popularite de Stilpon Doring (1988), p . 34, avec la n. 56, a attire l'attention sur u n e difficulte d u §85: q>6p£ vuv 6IEASOO|JSV TOUS &TT'OVTOO Kupr|vcuKoOs, OITIVSS £CCUTOOS oi uev 'HyqcjiaKoOs, oi 5e 'AvviKep-
eious, oi 6e 0eo8cop6iovs "rcpoCTCovoua^ov. A s'en tenir a ce texte, n'est-ce pas l'ensemble des Cyrenaiques q u i se divisent en Hegesiaques, Anniceriens et Theodoriens? A u c u n e place semble n'etre reservee a u x 'fideles'. Pour la retablir, Doring suggere d e corriger omves en OIT': 'et maintenant, passons en revue les Cyrenaiques q u i derivent de lui ainsi q u e [TE d e liaison] ceux q u i [01, relatif sans antecedant exprime] se n o m m a i e n t eux-memes [en tout etat de cause, il vaut sans doute mieux traduire, en donnant sa valeur pleine au preverbe, 'se donnaient eux-memes comme surnom', cf. infra, n. 117], les uns Hegesiaques, les autres Anniceriens, les autres Theodoriens'. Doring a parfaitement raison de noter que le texte transmis ne distingue que trois categories de Cyrenaiques, et, qu'en ce sens, il ne correspond pas au plan des §§860-104. Considerons cependant la phrase qui suit le texte cite: ou |ir|V dXAd KOCI TOUS 6TTT6 OcciScovos, cov TOUS Kopu9cuoT&Tous 'EpETpiKous. Faut-il egalement corriger, sous pretexte que Diogene Laerce insere des notices sur Phedon lui-meme (105), puis sur Euclide, Stilpon, Criton, Simon, Glaucon, Simmias, et Cebes, avant de parvenir a Menedeme et a l'ecole d'Eretrie (125)? On en conclura plutot que l'annonce ne correspond pas au traitement. C'est une autre question de tenter une explication du phenomene, certainement lie au mode d'ecriture de Diogene Laerce.
24
ANDRE LAKS
figure geometriquement: sur la petite diagonale d'un losange dont Aristippe et Theodore occuperaient les extremites, se situe une triade formee de ceux qui sont restes 'fideles' a Aristippe l'Ancien, des Hegesiaques et des Anniceriens. L'analyse des contenus confirmera que ces trois groupes constituent un ensemble doxographique coherent et relativement autonome, dont je voudrais degager 1'articulation en trois temps. Je cite d'abord integralement une traduction des paragraphes 86b-gia du livre n de Diogene Laerce, accompagnee d'un sommaire analytique. Ces paragraphes contiennent l'essentiel de notre information sur l'ethique cyrenaique, et comportent la plupart des elements necessaires pour situer les doctrines specifiques d'Hegesias et d'Anniceris 23 (m). J'analyse ensuite de plus pres deux ensembles thematiques developpes dans cette section de la doxographie: la doctrine du bonheur (iv) et celle des plaisirs psychiques (v), en tentant de montrer que c'est sous l'effet de sa propre instabilite conceptuelle, plutot qu'en raison de la concurrence du Jardin, que les positions de l'ecole cyrenaique ont evolue. Nous pourrons alors revenir a la these de Doring, et reexaminer la relation entre le temoignage de Clement sur les Anniceriens (qui porte sur les points iv et v), et la doxographie de Diogene Laerce (vi). I l l ORTHODOXES ET 'DISSIDENTS'
Diogene Laerce n.
86b-gia24
Ceux qui sont restes fideles a l'ecole d'Aristippe et qu'on a appeles Cyrenaiques professaient les opinions suivantes: 23
24
Le developpement 8 6 b - g 3 a est m a r q u e p a r une premiere coupure a u niveau d u §91. A une serie d e theses entrelacant la doctrine des affects, et singulierement d u plaisir, a u n e doctrine d u bonheur, succede, selon u n ordre bien atteste dans la doxographie ethique, u n ensemble consacre a u comportement d u sage ( 9 1 - d e b u t d e 92). Bien q u e la derniere section (92b~93a) mentionne encore a deux reprises le sage, elle n'est manifestement pas sur le meme plan q u e les precedentes. Le 93a comporte, a cote d e propositions ethiques, d'autres elements, touchant p a r exemple la theorie de la connaissance (l'ordre des notices est, sur quelques lignes, difficile a justifier). Q u a n t au §92b, il ne concerne l'ethique qu'indirectement, dans la mesure ou il est question des parties d e la philosophic. O n ne suivra done pas M a n n e b a c h (1961), p . 104, q u i divise la section en deux parties, d'abord les 'opiniones eorum de voluptate et dolore' (860-90), puis les 'quaestiones ad vitae usum spectantes' (91-3). Voir aussi les remarques de Giusta (1966/67), 1, p p . 135SS (repris p a r G i a n n a n t o n i (1963/65), in, Note 18, p . 169). J e reprends ici, en la modifiant parfois, la traduction d e Bollack (1975), p p . 164SS. Le decoupage de la section et le sommaire analytique doivent b e a u c o u p a son commentaire, cf. en particulier p p . 150s. et 190SS.
Anniceris et les plaisirs psychiques
25
(A) (1) Us posaient au fondement deux affections, souffrance et plaisir, Tun, mouvement lisse, a savoir le plaisir, l'autre, la souffrance, mouvement rugueux. (2) Un plaisir ne differe pas d'un plaisir, et il n'y a pas d'objet qui soit plus plaisant. (3) Et Tun agree a tous les vivants, tandis que l'autre est rejete par eux. (4a) Mais ce plaisir est celui du corps, qui est aussi la fin, comme le dit Panetius dans son ecrit Sur les ecoles: (4b) ils nient que le plaisir catastematique, qui repose sur la suppression des douleurs et est comme une absence de maux, qu'Epicure admet, soit la fin.25 (5a) Ils sont en outre d'avis que la fin differe du bonheur. Car la fin est le plaisir particulier, tandis que le bonheur est l'ensemble compose des plaisirs particuliers, au nombre desquels sont aussi comptes tant les plaisirs passes que les plaisirs a venir. (5b) Et le plaisir particulier est choisi pour lui-meme, tandis que le bonheur ne Test pas pour lui-meme, mais pour les plaisirs particuliers. (6a) La preuve que le plaisir est la fin vient de ce que, sans decision prealable, nous sommes familiarises envers lui depuis l'enfance,26 (6b) et qu'apres l'avoir obtenu nous ne recherchons plus rien, ni n'evitons rien autant que la douleur qui lui est contraire. (7) Et le plaisir est un bien meme s'il provient des comportements les plus inconvenants, comme le dit Hippobotos dans son ecrit Sur les ecoles. En effet, meme si Faction est deplacee, il n'en reste pas moins que le plaisir est choisi pour lui-meme et est un bien. (8) Mais la suppression de la douleur, comme il est dit chez Epicure, ils ne sont pas d'avis que c'est un plaisir, ni l'absence de plaisir une douleur. En effet l'un et l'autre sont en mouvement, alors que l'absence de souffrance ou l'absence de plaisir n'est pas un mouvement, puisque l'absence de souffrance est comme l'etat de quelqu'un qui dort. 27 (9) Et ils disent que meme le plaisir, certains peuvent ne pas le choisir par egarement.
(10a) Ce ne sont cependant pas tous les plaisirs et douleurs de l'ame qui reposent sur des plaisirs et douleurs du corps. En effet, il y a une joie qui nait en nous de la pure prosperite de la patrie, comme de la prosperite personnelle. (10b) Toutefois, ils nient aussi que le plaisir se produise par la rememoration des biens ou par leur attente, ce qui etait l'opinion d'Epicure. En effet, le mouvement de l'ame s'interrompt en raison du temps.28 (1 oc) Ils disent que ce n'est pas non plus par la pure sensation de la vue ou de l'ouie que se produisent des plaisirs. Car nous ecoutons avec plaisir ceux qui imitent les lamentations, mais sans plaisir ceux qui se lamentent en verite. 25
26 27
28
(4a) f|8ovnv |J6VTOI TT]V T O O acbuocTos, f^v Koct TEAOS sTvcci, KocSa r|CTi FfavaiTios ev TCO nept Tcov afpeaecov, ( 4 b ) ou TV\V KorraoTTipiaTiKfiv f)5ovf)v TTJV €TT' dvaipEaei dAyr|66vcov KOU olov dcoxATloiav, f^v 6 'EmKOUpos <5aro86XSTai, TEAOS eTvai 9 a a i v (KCU TEAOS eTvai cprjatv, Cobet, L o n g , etc.). T O 6rn-poaip6TCOs f)|ias EK TTCUSCOV cpKEicbaSai Ttpos aOTf|v. ETTEI f| &Trovia OIOVEI KCCSEOSOVTOS E O T I
Cf. infra n. 73.
26
ANDRE LAKS
(11) Et ils appelaient l'absence de plaisir et l'absence de souffrance des etats medians.
(C)
(12) Cependant, les plaisirs du corps sont bien meilleurs que ceux de Tame et les maux du corps bien pires. C'est aussi pourquoi on chatie les coupables plutot de cette facon; car ils admettaient que la souffrance est difficile a endurer, tandis que le plaisir est familier. (13) De la vient aussi qu'ils consacraient une importante gestion au second.29 (14) C'est pourquoi, le plaisir etant choisi en fonction de cette derniere, les maux producteurs de certains plaisirs viennent souvent s'y opposer; de sorte que l'accumulation des plaisirs leur semble etre la chose la plus penible, quand elle ne produit pas le bonheur. 30 L'expose des positions cyrenai'ques orthodoxes s'ouvre sur la definition du plaisir et de la douleur, presentes comme les deux 'affects' {pathe) fondamentaux de l'ethique cyrenai'que:31 le plaisir est dit etre un mouvement lisse, la douleur un mouvement rugueux (1). Cette definition est suivie d'une double precision relative a la nature du plaisir, sous la forme de deux refus: il n'existe pas de difference (qualitative) entre les plaisirs, et les variations quantitatives - qui leur 29 30
31
evSev Koci TrAeiovoc oiKovouiav Trspi S&T6pov ETTOIOOVTO. 816 m i Korra TOOTT|V aipETfjs OOOTIS TT\S f)8ovfis TOC TTOITITIKOC evicov f]8ovcov oxAripd TTOAAOCKIS evavTioOaSai- cos SUCTKOACOTOCTOV OCUTOTS 9aiv£aSai TOV d S p o i a u o v TCOV f]8ovcov 6u8ai|ioviav uf) TTOIOOVTCC (|ifi TrotoOvTa B: \xr\ TTOIOOVTCOV P : TToioOvTa F : TTOIOUCTCOV M e n a g e ) . P o u r l a justification d u texte, cf. infra n. 6 5 . Deux temoignages, ceux d'Eusebe et de Sextus Empiricus, ont pu faire penser que les Cyrenai'ques avaient compte au nombre des TroSr|, outre le plaisir et la douleur, l'etat intermediaire dont, contrairement a Epicure, ils reconnaissent l'existence. On s'est alors interroge pour savoir s'il convenait de recuser Diogene Laerce au profit d'Eusebe et de Sextus, ou l'inverse. En fait, le probleme provient d'une lecture inattentive de la phrase qui clot le temoignage d'Eusebe (cite supra, n. 3): TOUTCOV 8f) KCU Bpocoxe TCOV -rraScov uovcov f|iias TT]V aiaSr)aiv ixeiv. 'Ces irdSri-la' (TOUTCOV . . . TCOV TTOCSCOV) peut ne pas reprendre les trois 'etats' dont il vient juste d'etre question, mais renvoyer, par-dela le developpement relatif aux deux Aristippe, au tout debut de la notice, qui precede et introduit l'ensemble du passage consacre aux Cyrenaiiques: cruyyevfi 8* OCUTOTS [il s'agit des disciples de Pyrrhon] sir) &v KOCI TOC (ScvTiAsyoueva irpos TOUS KOCT' 'ApicrrnTTrov TOV KupTjvcuov, |i6va AeyovTas eTvai TOC Trd3r| KaTaArjTTTd (suit le texte reproduit n. 3). 'Seuls' (uovcov) s'entendra alors des affections, par rapport aux objets exterieurs (comme uova dans la phrase introductive), et non par rapport aux trois etats, ce qui serait evidemment absurde. Quant au passage de Sextus (Adv. M. vn. 199 = 213c), il n'a, par lui-meme, pas suffisamment de poids pour remettre en question celui de Diogene Laerce et d'Eusebe. II debute par la description (correcte) de la doctrine cyrenai'que, et correspond aux trois 'etats' mentionnes chez Eusebe: Kcci Td uev dAyeivd Kocxd 90CCTIV eTvai, cbv TIAOS dAyr|8cbv, Td 8s f)86a dyaSd, cbv TEAOS icrriv d8idy6ucrrov f)8ovf), Td 8E U£TOC£0 OOTE dyaSd OUT6 Kaxd, cbv TeAos TO OUTE dyaSov OUTE KOCKOV. La
relative qui suit, oirep TrdSos kcrvl ueTa^O f|8ovf\s Kai dAyr|86vos, pourrait etre un commentaire (terminologiquement impropre) ajoute par Sextus.
Anniceris et les plaisirs psychiques
27
sont implicitement reconnues - ne sont pas une propriete des choses elles-memes (2). 32 Cette premiere approche (A) est suivie d'un developpement sur la 'fin', introduit par une remarque sur l'attirance que marquent tous les animaux a l'egard du plaisir, et la repulsion qu'ils eprouvent pour la douleur (3). Une double precision, amenee par la particule mentoi ('cependant'), situe une premiere fois la these cyrenaique par rapport a celle d'Epicure. Le plaisir que les Cyrenaiques prennent comme fin n'est pas tout plaisir, mais celui du corps (4a), et ce n'est pas non plus le plaisir catastematique, qu'ils ne reconnaissent pas comme plaisir (4b). 33 Meme si Epicure n'est expressement cite qu'en relation avec 4b, il n'en est pas moins aussi implicitement vise en 4a, du moins si Ton admet que le plaisir dont Epicure fait la fin n'est pas un plaisir exclusivement corporel. 34 C'est a l'enseigne de ce plaisir corporel 'cinetique', que circonscrit un double refus de la doctrine epicurienne, qu'est implicitement place l'ensemble du resume jusqu'a 89a (B), avec une serie de sujets relatifs a divers aspects du telos: non-identite de la fin et du bonheur (5a et 5b), arguments permettant d'etablir que le plaisir est la fin (la version cyrenaique de l'argument des berceaux (6a), et celui de la suffisance du plaisir (6b)), indifference du plaisir aux valeurs morales (7), negation que la suppression des douleurs constitue un plaisir, comme l'admet Epicure (8), et, en dernier lieu, explication de l'exception (cf. 3 et 6b) que constitue le refus du plaisir par une forme d"egarement' (9). 35 Un second mentoi, avec la meme valeur oppositive que le premier, introduit la deuxieme grande partie de la notice (B), qui etablit dans leurs droits les plaisirs psychiques, dont il n'avait pas ete question en (A): ce n'est pas parce que le plaisir corporel est la fin qu'il n'existe pas de plaisirs psychiques autonomes (10a); ce qui ne signifie pas qu'il faille admettre - comme Epicure - l'existence de plaisirs reposant sur la seule memoire ou sur l'anticipation (10b). II y a analogie entre ioa/iob d'une part, 4aAj.b d'autre part. Dans l'un et l'autre cas, la position des Cyrenai'ques prend le contrepied de celle d'Epicure, bien que celui-ci ne soit nomme que pour la deuxieme 32 33 34
35
C'est Interpretation qui me semble etre la plus naturelle de ce membre de phrase. Doring (1988), p. 38, tout en le rapprochant de D.L. 7, evoque Peventualite d'une corruption. Sur la relation d e D.L. 4 a et 4b, cf. infra, n. 111. C o m m e o n sait, la position d'Epicure sur ce point est complexe. L'autonomie, et la superiorite, des plaisirs d e l'ame n'excluent p a s qu'ils se referent toujours en derniere instance a u x plaisirs d u corps (cf. infra, p . 45). Cf. supra, n. 16.
28
ANDRE LAKS
proposition. La difference est dans l'existence d'un element: la reconnaissance de l'existence de plaisirs proprement psychiques, mais non du type de ceux qu'Epicure invoque (ils ne reposent pas sur le souvenir et 1'anticipation), debouche sur l'analyse d'un plaisir complexe qui, tout en etant lie a l'exercice des sensations corporelles, ne s'y reduit pas (ioc). Une remarque terminologique sur la denomination des etats intermediates (11) semble assurer la transition entre la partie B et la partie C: 36 celle-ci, egalement introduite par un mentoi, pose la superiorite des plaisirs corporels sur les plaisirs psychiques du point de vue de la fin: ce n'est pas parce qu'il existe des plaisirs psychiques autonomes, comme l'afrirme ioa, que ces derniers sont 'meilleurs' - a savoir plus intenses 37 - que les plaisirs corporels; au contraire, la douleur corporelle est toujours pire que la peine psychique (12). La valeur de ce mentoi n'est pas exactement identique a celle des deux precedents, dans la mesure ou la rectification porte ici sur l'enonce d'une these cyrenaique (a savoir l'existence de plaisirs psychiques autonomes), et non pas sur la position d'Epicure. Mais elle la rejoint par un detour, puisque Epicure, contrairement aux Cyrenaiques, ne se trouve pas moins affirmer la superiorite des plaisirs psychiques sur les plaisirs corporels que la dependance, en derniere instance, de tout plaisir psychique a l'egard de plaisirs corporels anterieurs. Une fois posee la superiorite des plaisirs corporels sur les plaisirs psychiques, la fin de la partie C revient assez naturellement a la question du bonheur, avec la version cyrenaique du calcul des plaisirs (13 et 14). II est essentiel de voir que ce corps de doctrines, attribue aux membres de l'ecole reputes orthodoxes, est aussi, pour une partie, accepte par les 'dissidents'. Le debut des notices respectivement consacrees a Hegesias et a Anniceris indique en effet que les deux groupes restent en accord avec les fideles - et par consequent entre eux - sur un certain nombre de points. 38 Les limites de cet accord seraient relativement faciles a tracer, s'il s'agissait d'epistemologie, car aucune branche de l'ecole n'a jamais remis en question le principe supreme de la philosophie cyrenaique selon lequel tous les 36
37 38
L'existence d ' u n schema directeur assez net, dans la construction d e la doxographie, n'empeche pas q u e la succession des notices presente u n e serie d e problemes locaux (redondances et desordres apparents). C e n'est pas ici le lieu d e les examiner en detail. E n effet, il n'existe pas d e difference qualitative entre les plaisirs, cf. D . L . 2. Oi 5E 'HyrjaiocKoi Aey6|ievoi CJKOTTOOS |iev eTxov TOUS OCUTOOS, f)5ovf)v KOCI TTOVOV, Diog. Laer. 11. 9 3 ; Oi 8 ' 'AvviK6p£ioi TOC \xh) ocAAcc KOCT& TOCUT& TOOTOIS, II.96 (cf. infra, n. 44).
Anniceris et les plaisirs psychiques
29
pathe sont 'saisissables', et son admission semble, inversement, avoir ete la condition necessaire et suffisante pour le port du label cyrenai'que.39 Les choses sont un peu moins claires si Ton reste, comme l'exige l'expose de Diogene Laerce, dans les limites de l'ethique. Les Hegesiaques sont dits avoir 'les memes objectifs' que les Cyrenai'ques orthodoxes, a savoir le plaisir et la peine. 40 Mais l'accord s'etend moins loin qu'il ne pourrait sembler a premiere vue. Si les divergences n'affectent pas le principe fondamental de l'ethique, la reference au plaisir et a la peine, mais les consequences qui en sont tirees, l'emploi meme du terme de skopos pour se referer au plaisir et a la douleur indique que le disaccord commenfait tres tot. II implique en effet que les Hegesiaques aient etabli une distinction entre skopos et telos etrangere a la doctrine des fideles. De fait, le telos hegesiaque est defini, en 11. 96, comme une vie depourvue de douleur et de peine. 41 II ne se confond done ni avec le telos officiel (le plaisir), ni, evidemment, avec le second des deux skopoi (la douleur). Dans la mesure ou ceux-ci se rapportent tant a la douleur qu'au plaisir, skopos doit designer, non ce que Ton vise (une vie sans douleur), mais ce que Ton prend en consideration, ou qu'on examine, quand on vise la fin.42 Quant a la formule qui ouvre la section consacree aux Anniceriens,43 elle suggere que les divergences entre les Anniceriens et les Hegesiaques se limitent a un ensemble bien defini de questions, celles-la memes que Diogene Laerce rapporte dans la suite (il s'agit de l'existence de tendances altruistes en 1'homme et de leur fonction dans le cadre de la doctrine du bonheur). II s'ensuit que, pour le reste, ils ne sont pas moins 'fideles' qu'eux a l'ancienne doctrine. 44 39
40 42
43
44
Principaux textes relatifs au principe uovee T& TYCXBT] KaTaAriTrra: Plutarque (Adv. Colot., 1120C-1121G = 211 G ) , Sextus (Adv. M., v n . 191-98 = 213G), Aristocles (dans Eusebe, PE xiv. 19, 1 = 218 G ) ; voir en outre 212, 214, 215-18 G (cf. supra, n. 31). 41 Gf. supra, note 38. TO \XT\ ETTITTOVCOS £fjv |if|6e Aimripcbs. Ce que j ' a i cherche a rendre tant bien que mal p a r 'objectif. La distinction hegesiaque ne recouvre done pas celle des Stoiciens, chez q u i TEXOS designe la fin, OKOTTOS son obtention. Le terme de OKOTTOS prend la place, chez les Hegesiaques, de l'emploi qui est parfois fait de TeAos pour designer, non la fin recherchee, mais la 'limite' (sur le double emploi de TEAOS, cf. M a n n e b a c h (1961), p p . 11 is). Ol 5 ' ' A w i K E p E l O l TOC UEV OtAAa KCXTa TCXOTOC TOUTOIS, diTEAlTTOV 5E . . . I I . 9 6 .
O n a parfois suppose q u e la datif TOOTOIS ne renvoyait pas, dans la source de Diogene Laerce (que celui-ci aurait reproduit confusement), a u x Hegesiaques, mais bien a u x 'fideles' (cf. M a n n e b a c h (1961), fr. 186, avec le commentaire p. 94). Pourtant le balancement T& UEV ocAAa . . . OCTTEAITTOV 8E . . . n'aurait guere de sens si TOOTOIS representait les 'fideles'. C a r il n'y a aucune raison de penser q u e ces derniers supprimaient 1'ami tie, la reconnaissance, etc. Cette suppression, en revanche, est caracteristique d'Hegesias. Les Anniceriens 'laissent' done 'subsister' (diTEAiTrov) ce q u e les Hegesiaques eliminent. Le referent d u demonstratif est ainsi clairement identifie (cf. Classen (1963), p . 136; la suggestion de M a n n e b a c h est egalement rejetee p a r Giusta (1964/67), 1, p . 136, n. 1;
30
ANDRE LAKS
Naturellement, comme ce 'reste' ne possede pas de contours nettement definis, on est renvoye a Interpretation pour savoir si tel point de la doctrine originelle subsiste ou non chez les Hegesiaques comme chez les Anniceriens. Toujours est-il que, en depit de la distinction explicite entre le groupe des fideles et les autres, la disposition d'ensemble de la doxographie de Diogene Laerce suggere, par dela les divergences rapportees, l'existence d'une zone d'accord. Pour ceux des aspects qui ne contredisent pas ou ne sont pas incompatibles avec les positions specifiquement attributes aux dissidents, on peut s'attendre a ce que la doctrine des 'fideles' ait aussi ete une doctrine commune. IV BONHEUR ET PLAISIR: POUR UNE LOGIQUE INTERNE DE L'EVOLUTION DE L'ECOLE
Toute analyse de l'ethique cyrenai'que doit partir de Pidentification, que Diogene Laerce attribue a Aristippe l'Ancien, de la fin de nos actions a un certain type de mouvement. 45 De cette these decoule en effet directement la remise en question de la conception traditionnelle, elaboree par Aristote dans YEthique a Nicomaque, selon laquelle la fin de nos actions est le bonheur. 46 De fait, la doxographie de Diogene Laerce nous apprend que les Cyrenaiques operaient une distinction tranchee entre le plaisir et le bonheur. Definissant le bonheur comme Tensemble des plaisirs particuliers' (to ek ton merikon hedonon sustema) que Ton aura pu gouter au cours d'une vie, les Cyrenaiques posaient qu'il n'est pas recherche pour lui-meme, mais bien pour les plaisirs particuliers (he kata meros hedone) qui en sont constitutifs.47 Seul ce plaisir particulier, etant recherche pour lui-meme, peut etre a proprement parler qualifie de fin.48 L'argumentation qui fondait cette distinction ne nous a pas ete conservee. Mais on peut tenter de la restituer. L'identification du
45 46 47
48
Doring (1988), p. 51; cf. aussi G i a n n a n t o n i (1983/85), 1, adw. G.3). O n peut bien entendu se d e m a n d e r , en u n second temps, si l'accord avec les Hegesiaques 'sur le reste' ne se reduit pas de facto a l'accord sur quelques principes relevant d e la doctrine c o m m u n e . Voir supra, n. 18. Le b o n h e u r est ce en vue d e quoi nous choisissons le reste, Aristote, EN 1.7, 1097b!. D . L . 5 (a et b ) : SOKET 5 ' CCOTOTS KOCI TEAOS euSaiuovias Siacpepeiv. TEAOS nev y&p eTvai TT)V KOCTOC
liepos i^Sovriv, e08ai|ioviav 6e TO IK TCOV |iepiKcov rjSovcbv <juaTT||ia, ais auvapiSuoOvTcu Kai a! TrapcpxT|KuTai Kai a! ueAAovaai. eTvai TE TT]V uepiKfjv f|5ovf)v 5i' auTf)v aip6TT)v TTJV S'eOSaipoviav ou 81' ocOrnv, dAAd 61a Tas KOCTOC nepos fiSovas. Voir egalement le texte d e Clement (doxographie d'Anniceris) discute plus bas, p p . 41SS. Cf. outre le texte cite a la note precedente, D . L. 6 b .
Anniceris et les plaisirs psychiques
31
bonheur au plaisir, tel qu'il est defini par les Cyrenaiques, se heurterait en effet a deux difficultes complementaires. D'abord, comment un processus comme le plaisir, condamne qu'il est a s'interrompre,49 pourrait-il etre identifie au bonheur, s'il est vrai que le bonheur s'inscrit dans la duree (en fait, la vie entiere)?50 Ensuite, s'il est vrai que le bonheur ne se confond avec aucune des fins particulieres que nous poursuivons, et qui en sont constitutives, faire du bonheur une fin ultime ne revient-il pas a confondre une fin de second ordre avec les fins de premier ordre que sont les buts immediats qui guident nos actions?51 La dissociation entre bonheur (eudaimonia) et fin (telos) semble faite pour prevenir ces objections. Elle assure du meme coup aux Cyrenaiques une position extremement originale dans le cadre de la philosophic hellenistique, qui s'en est tenue, qu'il s'agisse des Stoiciens ou des Epicuriens, a la conception aristotelicienne en vertu de laquelle la fin derniere de nos actions est le bonheur. 52 De fait, la position cyrenaique implique un declassement de Yeudaimonia: stricto sensu, les Cyrenaiques ne sont pas des eudemonistes. En EN 1.7, 1097b 1-6, Aristote dit du bonheur qu'on le choisit pour lui-meme et jamais pour autre chose, alors que les differents biens (tels que les honneurs, les plaisirs, l'intelligence, la vertu) sont choisis aussi bien pour le bonheur que pour eux-memes: c'est ce qui fait du bonheur la fin derniere. 53 Les Cyrenaiques, mettant le plaisir a la place du bonheur aristotelicien (il est la fin), ne font naturellement pas jouer au bonheur un role analogue a celui des biens aristoteliciens. Car ces derniers, s'ils sont recherches pour autre chose (a savoir, le bonheur), sont aussi recherches pour eux-memes, dans la mesure ou ils sont constitutifs du bonheur. En revanche, le bonheur cyrenaique n'est nullement constitutif des plaisirs particuliers. II n'en est pas non plus une condition necessaire. Si les Cyrenaiques disent que le bonheur est recherche a cause des plaisirs particuliers,54 49 50 51 52
53
Cf. Doring (1988), p . 40. Cf. Aristote, ENi.y, 1098a 18-20. C'est a cette intuition que se refere le TEAOS opav de Solon analyse p a r Aristote en EN 1 .1 o. C e sont, chez Aristote, la vertu, l'amitie ou la richesse, mais aussi les plaisirs. C e sont, chez les Cyrenaiques, les plaisirs. L a confrontation avec Epicure est interessante. Epicure maintient l'identite d u plaisir et d u b o n h e u r (Lettre a Menecee, 128), mais c'est a u prix d ' u n e restructuration radicale d u concept de 'plaisir' comme plaisir catastematique, qui lui permet d e perdurer a travers le temps. L a position epicurienne est reconstruite p a r Gisela Striker dans sa contribution a u present volume. Pour les Stoiciens, cf. Stobee 11.77, J 6s {SVFm 16). 54 Voir supra, n. 46. D.L. 5b (cite supra, n. 47).
32
ANDRE LAKS
cela ne peut signifier que pour obtenir des plaisirs (des nouveaux plaisirs), il faudra avoir prealablement ete heureux. Au contraire, etre heureux signifie que Ton aura accumule, au cours de la vie, un certain nombre de plaisirs. En ce sens le bonheur cyrenaique est bien une fin de second ordre. Le separation entre la fin et le bonheur est, naturellement, un point relativement subtil, et tellement singulier dans le contexte de l'eudemonisme classique, que Ton comprend que la doxographie n'en rende pas toujours clairement compte.55 Meme un temoignage aussi precis que celui d'Eusebe n'echappe pas a l'ambiguite. Quand Aristippe le Jeune est dit 'avoir clairement defini que la fin etait de vivre dans le plaisir, en inserant le plaisir qui est en mouvement', 56 cela signifie-t-il que la fin est pour lui le plaisir, ou la vie dans le plaisir (done en fait le bonheur)? 57 On pourrait naturellement se demander si la distinction entre bonheur et fin ne constitue pas un raffinement, encore posterieur a Aristippe le Jeune, de la doctrine cyrenaique. Parle contre, me semble-t-il, le fait que la dissociation decoule directement de la nature 'cinetique5 de la 'fin' cyrenaique.58 Comme le laisse deja entrevoir la distinction entre la fin (limitee dans le temps) et le bonheur (coextensif a l'ensemble de la vie), la question du temps et de son controle a dujouer un role central dans la philosophic cyrenaique. Si elle ne se distingue pas en cela des autres philosophies hellenistiques, le fait que le plaisir-fin soit defini en termes de mouvement donne a la doctrine sa tonalite particuliere. Les Epicuriens, comme les Stoiciens, soustraient la fin a la menace du temps (qui s'exprime par le fait que Phomme est soumis a la fortune 55
56
57
58
Voir p a r exemple Athenee, x n , 544a ( = 174 G ) : 6s oaroSE^auEVos TT\V i)8v/TTaSeiav TOCUTT|V TEAOS slvai Ecprj KCU EV OCUTT) TTJV EuSaiuoviav pEpAf^aSai (encore q u ' o n puisse soutenir q u e PepAf^aSai n'implique pas necessairement u n e identification); Clement, Stromates, 11.21.127, 2 = 198 G; E p i p h a n e , Adv. Haer. 111. 2. 9 = 177G. Eusebe, PE xiv. 18. 32: 6$ KCCI aocepcos cbpiaotTO TEAOS ETvai TO fjSEcos £fjv, f)8ovf)v EVTATTCOV TT\V Korra Kivr|aiv (cf. supra, n. 3). Q u e signifie 'inserer' (EVTOCTTCOV)? O n pourrait supposer que la doxographie d'Aristippe le Jeune est implicitement (et bien entendu anachroniquement) construite par rapport a la doctrine d'Epicure, selon un schema explicitement atteste ailleurs (cf. supra, n. 7). Sur le registre comptabilisant les differentes formes de plaisir, Aristippe inscrirait alors le plaisir cinetique a cote (ou plutot a la place) du plaisir catastematique d'Epicure. II se peut aussi, plus simplement, qu'Aristippe 'insere' le plaisir en mouvement parmi les candidats au titre de 'fin'. La doxa 'potentielle' d'Aristippe l'Ancien ('il affirmait en puissance q u e l'existence d u bonheur reside dans les plaisirs', cf. supra, n. 3) semble reposer sur la distinction entre le bonheur et le plaisir (cf. le pluriel EV f|8ova!s). L'adverbe TJSEGOS, dans la formule TO fiBwos ^ v qui sert a definir le TEAOS d'Aristippe le J e u n e , recouvre la pluralite des plaisirs. Mais dans ce cas, n'est-ce pas le bonheur (la vie dans les plaisirs), et non le plaisir, qu'Aristippe le J e u n e definit comme la fin? J e suivrais ici Doring (1988), pp. 40 et 57.
Anniceris et les plaisirs psychiques
33
(tuche) et finalement a la mort) en reduisant les conditions de son obtention a une disposition interieure. Quant au bonheur aristotelicien, s'il est menace par le temps, c'est de maniere pour ainsi dire extrinseque, dans la mesure ou un malheur peut toujours mettre un terme a un bonheur essentiellement fonde sur une activite dont la stabilite est garantie, autant que faire se peut, par son ancrage dans les vertus. 59 II ne parait pas douteux que les Cyrenai'ques s'inscrivent dans cette tradition, comme le montre le temoignage isole et difficile, mais instructif, de Ciceron dans les Tusculanes: la praemeditatio futurorum malorum doit nous preparer a l'irruption toujours possible du malheur, et nous eviter la peine qui en resulte, quand il n'a pas ete prevu. 60 Mais la singularity du telos cyrenaique reside en ce qu'il est essentiellement instable, et livre au flux du temps: le mouvement du plaisir ne dure en effet qu'un moment, a savoir le laps de temps dont le mouvement a besoin pour s'accomplir. C'est en ce sens qu'il est 'partiel' ou 'particulier'. 61 La fin, en raison de son caractere ephemere, fait l'objet d'une recherche toujours renouvelee. 62 D'une certaine maniere, ce n'est done pas seulement la fin, mais son retour que Ton poursuit, et done le bonheur; ceci n'introduit aucune incoherence dans la doctrine, s'il est vrai que le bonheur est precisement constitue de la somme des plaisirs particuliers. 63 On se demandera alors quel est le seuil a partir duquel un homme peut etre dit heureux, s'il est toujours possible de 59 60
61 62 63
S u r cet ensemble de themes, voir Nussbaum (1986a), en particulier le chapitre 11 {'The vulnerability of a good human life: activity and disaster'). Cyrenaici non omni malo aegritudinem effici censent, sed insperato et necopinato malo, Tusc. in. 13.28 (= 208 G); la praemeditatio est mentionnee en 29; voir l'ensemble du developpement, 28-31, ainsi que 52 et 75s. Ioppolo (1980a) est essentiellement interessee a retrouver, sous la formulation pretee aux Cyrenai'ques, la polemique anti-chrysippeenne de Carneade (cf. pp. 79s. et 87: une serie d'indices et de recoupements fanno pensare che Cicerone, aWinterno della dottrina cirenaica, abbia incluso anche I'esposizione di alcune argomentazioni di Carneade, le quali si prestavano meglio ad una discussione dialettica della dottrina stoica'). La question meriterait d'etre reprise, etant donne q u e la nature exacte de la contribution des Cyrenaiques a la these q u e Ciceron presente comme leur ne ressort pas clairement de 1'analyse. Ioppolo pourrait avoir ete prevenue contre le temoignage de Ciceron, pour autant que les Cyrenaiques sont concernes, p a r l'incompatibilite apparente qu'il presente avec celui d'Athenee (et d'Elien), q u i montre les Cyrenaiques occupes a jouir d u moment present, et p a r consequent indifferents a l'avenir comme au passe (cf. cependant Giannantoni (1983/85), in, Note 18, p p . 169s). Texte cite supra, n. 47. C'est peut-etre ce point q u e vise u n e notice obscure dans la doxographie d u sage: <5cpK£i 5E KOCV KOTOC uiocv TIS TTpoCTTriTTTOUCTav f)8ecos eTTav&yn, Diog. Laer. 11. 91. Ce point est bien a p p a r u dans la discussion a Syam. J'ajouterai que q u a n d les Cyrenaiques soutiennent q u e 'le bonheur n'est pas choisi pour lui-meme, mais pour les plaisirs particuliers' (D.L. 5b), ils ne nient precisement pas q u e le bonheur soit recherche, mais bien qu'il soit recherche pour lui-meme.
34
ANDRE LAKS
le franchir, et par quel moyen. Les Cyrenai'ques ont du elaborer ce point. On tient, dans la doxographie generate de Diogene Laerce, une trace de cette reflexion. Les Cyrenai'ques avaient en effet developpe l'idee d'une 'gestion' ou d'une 'administration' des plaisirs (oikonomia) ,64 Mais en temoignent avant tout, bien que de maniere indirecte, les positions divergentes d'Hegesias et d'Anniceris, que la suite de la notice de Diogene Laerce confronte, sur la question de savoir si le bonheur est a la portee de l'homme. C'est sans doute sur ce point que Ton pergoit le mieux le lien entre accord generique et position specifique au sein de l'ecole cyrenai'que. La doctrine originelle, partant de la definition du bonheur comme 'accumulation de plaisirs particuliers', avait insiste sur l'incertitude de Tissue, allant jusqu'a affirmer que Taccumulation des plaisirs [est] la chose la plus penible, quand elle ne produit pas le bonheur'. 65 Or les Hegesiaques etaient de toute evidence alles plus loin, en niant la possibilite effective de parvenir au bonheur, en raison de la quantite de miseres qui emplissent le corps et Fame, et des obstacles que la fortune oppose a la realisation de nos plans. 66 Cette radicalisation etait chez eux liee a un egoi'sme fondamental 67 et au refus de 64 65
D . L . 13. D . L . i 4 . J ' a d o p t e le texte d e B, TOV dSpoicTuov TCOV f)8ovcov eu8ai|joviav uf) TroioOvTa, q u i n ' a
jamais vraiment retenu l'attention des interpretes. II merite pourtant d'etre pris en consideration. L'accord partiel avec P (pour la negation) et avec F (pour le participe masculin) lui donne une certaine credibility sur le plan paleographique. Et il permet d'echapper aussi bien a Interpretation reductrice de F (et de Menage), cherchant a ramener D.L. 14 a D.L. 5a (l'accumulation des plaisirs, ou ces plaisirs eux-memes, seraient a nouveau dits constituer le bonheur) qu'aux difficultes de P (voir la defense de Bollack (1975), p. 201), a condition que Ton donne au participe une valeur conditionnelle: etant donne que l'obtention de plaisirs depend souvent de facteurs contraires au plaisir, l'agregat de plaisirs sera chose extremement penible (SuaKoAcbTorrov), dans Vhypothese ou il ne produit pas le bonheur. Si Ton suit cette interpretation, D.L. 5a definit le bonheur, tandis que D.L. 14 envisage son obtention effective. On comprendrait aussi le choix du terme oSpoiauos, qui conserve quelque chose de sa force active (activite consistant a rassembler ou a reunir) en face du neutre owTrma, pour designer le resultat final de cette activite (selon LSJ, &Spoi<7u6s equivaut a aSpoicis). 66
67
TT)V 6u8ai|ioviav oAcos dSuvorrov glvat- TO uev y a p crcb|jia TTOAXCOV dvaTreTrX'naSai TraSriMdTCOv, TT\V 6e yv/x^v auuTraSelv TCO acbuaTi Kai TaporrreaSai, TT)V 8e Tuxnv -rroXAd TCOV KOCT' sATrtSa KCOAOEIV, COOTE 8id TOUTCC dvuTrapKTOv TTIV gu8ai|ioviav eTvai, Diog. Laer. 11. 94. Gisela Striker suggere qu'Hegesias pouvait p a r la reagir a Epicure, soulignant la facilite d'atteindre le b o n h e u r (cf. Lettre a Menecee, 133). Mais oAcos d80vaTOv et dvuTrapKTOv ne rencherissent-ils pas d ' a b o r d sur la difficulte a reunir les plaisirs necessaires a l'obtention d u b o n h e u r (D.L. 14, cf. supra, n. 65)? N a t u r e l l e m e n t , la reference a Epicure pouvait se surajouter, mais la polemique anti-epicurienne, dans le cas d'Hegesias c o m m e d a n s celui d'Anniceris (cf. infra, p . 46), a peut-etre eu moins d ' i m p o r t a n c e q u ' o n ne l'admet d'ordinaire (cf. p a r exemple von Fritz (1934), col. 1828). TOV T6 cro
Anniceris et les plaisirs psychiques
35
reconnaitre la moindre realite aux mouvements qui nous portent apparemment sans profit vers autrui — reconnaissance, amitie, bienfaisance.68 Tout se passe comme si le cyrenaique hegesiaque etait irremediablement livre a la solitude, comme il Test aux affects. On comprend qu'Hegesias ait refu le surnom de Peisithanatos ('persuadeur de mort'), 69 meme si sa position doctrinale serait plutot celle d'un indifferentisme, qui lui faisait dire que la vie n'est pas moins un objet de choix que la mort. 70 Hegesias avait tire argument de l'accablement auquel la vie humaine est exposee pour soutenir que le sage cherche moins a accumuler les biens qu'a eviter les maux. D'ou une definition negative de la fin, comme une vie depourvue de douleurs et de peines.71 Par rapport a la position des 'fideles' et au telos positif d'Aristippe, le 'plaisir' que Ton associe volontiers a la philosophic cyrenaique, la rupture est nette. 72 L'interessant est cependant que ces theses, pour novatrices qu'elles soient, n'en presupposent pas moins, en arriere-plan, une parfaite orthodoxie, avec la definition du bonheur comme somme des plaisirs particuliers, la distinction entre plaisirs du corps et plaisirs de Tame (les miseres de Tun affectent l'autre), et la notion de gestion des plaisirs (nous elaborons des plans engageant l'avenir). C'est sur la base d'un accord fondamental que la dissidence est ici possible - on pourrait aller jusqu'a dire qu'elle est fondee par cet accord meme. Des lors la doctrine annicerienne, qui avait defendu, contre les Hegesiaques, la possibility du bonheur, 73 se laisse assez naturellement interpreter comme une reaction visant a restituer, contre une deviation rendue possible par la dissociation du plaisir et du bonheur, l'integrite de la doctrine primitive. Anniceris pouvait, pour ce faire, rester dans le cadre trace par Aristippe, et durci par Hegesias. II n'avait nul besoin, pour garantir la possibility du bonheur, de nier que l'homme fut un etre fragile et expose, puisque les 'fideles' eux-memes avaient insiste sur la difficulte qu'il y 68
[xryrs 8E X ^ P l v T l sTvai M^TE (JuAiav |if)TE EOepyeaiav, 8 i a TO \xr\ 81' a u x d TOCOTOC a l p e i a S a i m a u T a , dAXa 8 i a TOCS XP 6 ^ a 5 OCUTCOV, COV diTOucTcbv IJH^S* eKeTva UTrapxsiv, D i o g . Laer. n . 9 3 . 69 E p i p h a n e , Adv. Haer., m . 2 . 9 , C i c e r o n , Tusc, 1.34.83 ( = iv F 2 et 3 G ) . 70 Tf|V T e ^cof)v Kai TOV Savorrov aipETov, Diog. L a e r . 11.94. 71 TOV TECT090Vo u x OUTGO TrAeovaaeiv ev TT) TCOV dyaScov aipsaei, cos sv TTJ TCOV Kaxcov 9uyrj, TEAOS TISEIJEVOV TO |if) ITTITTOVCOS ^ V |jir|8E Au7TT|pcos, Diog. L a e r . 11.95s; cf. supra, p . 29. 72 Cf. p a r e x e m p l e le texte cite supra, n. 56. 73 Diog. Laer. 11. 96: 6CTO(|>6SO06EV fjTTOv Eu8ai|iovf|crEi... L a restriction ('pas moins') r e p o n d a d e u x concessives, s y m e t r i q u e m e n t placees d e p a r t et d ' a u t r e d u m e m b r e d e p h r a s e , KCXV OXATICEIS dva8E5*nTat [sc. 6 aocpos] et KOCV 6Xiya fiSia 7TEpiyEVT)Tai auTco, q u i se renforcent
mutuellement et qui renvoient clairement a la position d'Hegesias, au §94.
36
ANDRE LAKS
a a atteindre le bonheur. 74 Le refus du pessimisme hegesiacjue s'appuie chez lui sur une critique de leur utilitarisme, qui, ne laissant aucune place aux sentiments altruistes, 75 se privait en meme temps des seuls moyens dont dispose 1'homme pour s'assurer du bonheur, i.e. d'une quantite suffisante de plaisirs. De toute evidence, l'idee, qui eclaire en retour la position d'Hegesias, est que l'amitie et la reconnaissance, auxquels Anniceris joignait la veneration des parents et l'engagement pour la patrie, induisent des plaisirs psychiques en nombre suffisant pour compenser ce que la vie comporte de malheurs inevitables. 76 Sur ce terrain, Anniceris devait naturellement rencontrer la doctrine epicurienne, qui accorde aussi une grande place aux plaisirs psychiques. Toute la question est de savoir comment. V LA DOCTRINE DES PLAISIRS PSYCHIQUES
Conformement a un usage implicite, mais bien ancre dans la tradition de la psychologie grecque, la distinction entre les plaisirs du corps et les plaisirs de Fame ne signifie pas que le corps perfoive le plaisir aussi bien que Tame. 'CorporeP et 'psychique', dans les formules 'plaisirs corporels' et 'plaisirs psychiques', ne se rapportent pas a l'organe ou au centre de la perception, mais a ce qui, dans le sujet percevant, en est la source (meme s'il s'agit d'une source mediate, l'objet exterieur constituant la source ultime). Autrement dit, Fame est dans tous les cas l'instance de la perception; 'corporel' et 'psychique' qualifient quant a eux un certain type de plaisir. II faut done distinguer entre deux fonctions de l'ame, selon qu'elle ressent le plaisir, en tant que faculte sensorielle, ou qu'elle en est la source. 77 A suivre le resume de Diogene Laerce, les plaisirs psychiques, en ce second sens, embrassent chez les Cyrenai'ques deux categories, selon qu'ils dependent de plaisirs corporels ou qu'ils sont totalement independants a leur egard. 78 Bien que la tradition doxographique 74 76 77 78
75 D . L . 14 Cf. supra, n . 6 5 . Cf. supra, n . 68. OCTTEAITTOV SE KOCI cpiAiav ev (3icp KOCI X^piv KCU irpos yovEas TIUTJV KCU CrrrEp TrorrpiSos TI Trpa^siv. 6Sev 81a TOUTCX . . . OU6EV f]TTov EOSaiiiovrjaEi, D i o g . L a e r . 11. 9 6 . La distinction a du etre elaboree dans le prolongement des analyses platoniciennes; cf. Republique, 583egs; Theetete, i84c-e; Philebe, 33d et 43b 1-5; Timee, 64a-65b. Selon D.L. 10a, tous les plaisirs psychiques n e se produisent pas e n relation avec des plaisirs corporels (ou TT&aas JJEVTOI TOCS yuxiKas f)8ovas KOU aAyr|66vas ETTI acouaTiKous f)8ovais Kai AAyr|86ai yivEaSai), ce qui signifie que tel est le cas pour certains d'entre eux. On doit done distinguer (1) les plaisirs corporels; (2) les plaisirs psychiques relatifs au corps; (3) les plaisirs psychiques independants du corps.
Anniceris et les plaisirs psychiques
37
presente quelques flottements sur ce point, 79 le nom propre du plaisir psychique semble bien avoir ete chara (et lupe celui de la douleur psychique).80 Le plaisir etant defini comme un 'mouvement lisse', 81 la joie sera un 'mouvement lisse' specifique a Fame (tandis que le plaisir corporel serait un mouvement lisse specifique au corps). Chez les Cyrenaiques, la memoire ou Fattente des biens passes ne tombe pas, comme c'est typiquement le cas chez Epicure, dans la categorie des plaisirs de Tame (des joies), puisqu'ils niaient au contraire qu'elles puissent jamais donner lieu a des plaisirs. 82 Un exemple de plaisir psychique dependant du corps est fourni par la joie que procure a chacun son propre bien-etre, que Diogene Laerce mentionne en la distinguant de la joie, quant a elle 'purement' psychique (c'est-a-dire independante de moi, et dans cette mesure de mon propre corps), que procure le bien-etre de la patrie. 83 C'est manifestement cette seconde categorie de plaisirs psychiques, ceux dont le correlat est un objet propre a l'ame, qui a le plus retenu l'attention des Cyrenaiques. L'ame peut etre mise en mouvement par des 'objets', ou plus exactement par des quasi-objets, privilegies en ce sens qu'ils sont de nature a declencher les plaisirs psychiques sans affecter le corps. A cote de la joie que provoque la prosperite de la patrie, Diogene Laerce mentionne un second plaisir ressortissant a cette categorie. II s'agit de ce que nous appelerions le 'plaisir esthetique' - d'un terme particulierement mal venu dans l'optique 79 80
81 82
83
Cf. le temoignage de Clement sur les Anniceriens, commente plus bas. Pour la distinction X a P Q / f|8ovf| d'une part, A0TTT|/TT6VOS (OU dAyoOv) d'autre part, cf. Diog. Laer. 11. 98 (Theodore): TEAOS 8'u7TEAa|ji|3avE xocpavKOCI AuTrriv . . . ueaa 8Ef|6ovf)vKCU TTOVOV. Voir aussi la double negation dans la doxa d'Hegesias citee supra, n. 4 1 . La meme distinction se trouvait sans doute chez Epicure (cf. Maxime capitale 3). Pour u n autre usage de la difference chez les Stoiciens, cf. SVF I I I , 431SS. Cf. supra, n. 18. aAAa ufjv O08E KOTO uvf|ur|v TCOV ayaScov f\ Trpoaaiv drrroTEAeiaSai, oirep f|p6(TKEV 'EiriKoOpcp- EKAUETOU y a p TCO Xpovcp TO Tffc y^x^S Kivr||ia, D.L. 10b. A moins de supposer q u e l'argument est a m p u t e ou elliptique (Doring (1988), p . 141), l'explicative introduite p a r y a p porte aussi bien sur l'attente d u futur que sur la rememoration d u passe. J e comprends q u e le 'temps' qui nous separe de l'affect, vers l'avant comme vers l'arriere, empeche le mouvement de l'ame, en l'occurrence la memoire ou l'attente, de parvenir a son terme: l'elan d u plaisir se brise, pour ainsi dire, sur le vide temporel q u i s'intercale entre 'le bien' et rrioi-meme. Si Ton adopte cette ligne d'interpretation, EKAOETOU y a p TCO X P °V C O T O Tfjs v|A/xilS Kivnua se traduira: 'car le mouvement de l'ame s'interrompt en raison [et non 'sous l'effet'] d u temps' (TCO XP°V C P etant u n datif instrumental, et non temporel, comme on l'entend d'ordinaire, q u a n d on le rapporte au passe uniquement). xai y a p ETTI yiAf) TT) rf\s TrorrpiSos E0r|UEpia cbaiTEp TT) 161a x«pav lyyivEaSai, D.L. 10a. Le 'bienetre' ou la 'prosperite' propre (!5ia EurjUEpia) ne se reduit naturellement pas aux plaisirs corporels. O n r e m a r q u e r a que le terme de Eur|UEpia, en vertu meme d e sa formation, se distingue d u 'bonheur': c'est la prosperite d' ' u n jour'.
38
ANDRE LAKS
des Cyrenaiques, s'il est vrai que ce plaisir se caracterise precisement en ce qu'il n'est pas le fait de la sensation (meme si cette derniere en est l'instrument). En quel sens la joie patriotique et le plaisir esthetique sont-ils des plaisirs psychiques? II est difficile d'admettre que la joie patriotique soit d'ordre 'psychique' dans la mesure ou elle ne me concerne pas (je fais abstraction de moi-meme), mais autrui (Pensemble demes concitoyens, ou mes concitoyens pris comme un ensemble): ce n'est pas parce que mon corps fait partie de mon moi qu'autrui releve de mon ame. 84 'Psychique' semble en l'occurrence se rapporter a l'origine 'ideelle' de certains plaisirs, qui supposent une activite de la reflexion ou plus generalement de la pensee.85 Nous serions sans doute plus a l'aise pour comprendre ce que les Cyrenaiques entendaient par 'ame' si nous avions conserve l'analyse qu'ils donnaient du plaisir 'esthetique'. Dans un argument qui, exterieurement, presente une similitude frappante avec celui que developpe Aristote au chapitre 4 de la Poetique pour etablir la specificite du plaisir mimetique, les Cyrenaiques inferent l'origine psychique du plaisir esthetique a partir du fait qu'il nait de la representation d'objets ou de situations qui, dans la realite, ne nous procurent pas de plaisir (comme le spectacle de lamentations, par exemple dans la tragedie), ou nous sont simplement deplaisants. 86 Les Cyrenaiques expliquaient-ils le phenomene du plaisir esthetique, conformement a l'analyse aristotelicienne, comme un plaisir 84
85
86
Dans le cas ou autrui se trouve etre u n ami, on pourrait presque dire le contraire. Diog. Laer. n . 91 semble e n effet impliquer q u e les Cyrenaiques concevaient l'ami (autrui) comme u n prolongement de mon propre corps. Les Cyrenaiques d e Clement, dans u n e formule parallele a celle d e Diogene Laerce, considerent aussi les 'relations humaines' et les 'ambitions' comme d o n n a n t lieu a des 'plaisirs psychiques' (xocipeiv y a p f)uas uf) uovov e m f|8ovals, dAAd Kai ITTI 6|iiAiais Kai h r i cpiAoTiuiais); voir infra, p. 44. Aeyouai 6s ur|66 KaTa ^iXfjv TT)V opaaiv f\ TT\V dKofjv yiveaSai f)5ovas. TCOV yoOv |Jiuou|ji6vcov Spfjvous fjSscos OKo0o|jev, TCOV 6E KOCT' dAfiSeiav dr|5cbs- D.L. 10c. O n pourrait etre tente d e faire un sort a dr|8cbs, et le rapporter a l'etat intermediate des Cyrenaiques, plutot qu'a l'oppose du plaisir qu'est la douleur, d'autant que la phrase suivante rappelle precisement l'emploi technique de d^Sovia (lieaas TE KonraCTTdaEis cbvoua^ov dr|8ovicxv Kai diroviav). Le passage parallele de Plutarque, Quaest. conv. v. 1.2, 674a-b ( = 206 G), differencie peut-etre 'deplaisant' et 'non-plaisant' - cela depend de la valeur que Ton donne a Kai dans dAsKTopis ydp powaa auvexcos . . . Ainrripdv dKOua^a Kai dr|8ES ecrri (Giannantoni me semble avoir raison de considerer que l'explicative introduite par ydp correspond a 1'argumentation originelle. Mannebach, qui ecourte la citation (fr. 190M) n'y voit qu'une paraphrase de Plutarque). En tout etat de cause, rargument relatif a la specificite du plaisir esthetique ne serait pas modifie si Ton tenait compte de la nuance.
Anniceris et les plaisirs psychiques
39
cognitif, lie a la production d'un raisonnement quasi deductif?87 Pas plus que la notice de Diogene Laerce, le passage parallele de Plutarque ne permet de repondre veritablement a cette question. Du moins montre-t-il que les Cyrenai'ques avaient developpe une explication du plaisir esthetique, ou le terme de dianoia ('pensee') devait jouer un role essentiel: les cris de la poule et de la corneille nous irritent, la vue de mourants nous repugne, mais limitation des uns et la representation plastique ou picturale des autres nous rejouissent 'parce que la pensee est mise en mouvement par les imitations conformement a ce qui lui est propre'. 88 L'idee semble etre la suivante: la representation fournit a la pensee un milieu ou elle se sent chez elle (kata to oikeion),89 qui lui permet d'etre mise en branle (agesthai). Cet 'entrainement' etait-il de nature cognitif? Peut-etre. Mais l'association du plaisir esthetique avec la joie patriotique pourrait aussi pointer vers une forme de 'ravissement' de nature moins intellectuelle. Dans cette hypothese, les Cyrenai'ques auraient bien fait usage de la Poetique, mais un usage critique.90 VI ANNICERIS
Nous sommes desormais en possession des principaux elements pour examiner Interpretation 'analytique' de la doxographie cyrenaique de Diogene Laerce. La demonstration de K. Doring comprend les deux moments classiques de la mise en evidence d'une contradiction doctrinale dans cette doxographie,91 avec la distinction entre deux strates de I5argumentation (a), et de l'attribution, essentiellement sur la base d'un temoignage exterieur (celui de Clement), de l'une d'entre elle a Anniceris (b). J'examinerai successivement ces deux points. 87 88
89
90 91
Aristote, Poetique, 4, 1448b 15-17. Dans la derniere etude en date que j e connaisse, Sifakis (1986) defend u n e version forte d u 'syllogisme' esthetique. 81a TT)V Sidvoiav UTTO TCOV uiuriuaTcov a y e a S a i [KCU] Korra TO oixeTov (sur le texte, cf. infra, n. 89). Les editeurs, a la suite de Wilamowitz, suppriment KCCI devant KOTOC TO oiKeTov. Le KOU s'explique-t-il p a r u n e dittographie mal resolue? O n pourrait aussi tenter de distinguer deux temps: la pensee est mise en branle p a r les representations, et le mouvement lui est, en l'occurrence, propre. Le texte de Plutarque presente l'interet supplementaire, pour notre analyse, d'opposer les Cyrenai'ques a Epicure sur la question d u plaisir esthetique, cf. supra, n. 7. 'Die Lehren, die Diogenes d e n "eigentlichen" Kyrenaikern zuschreibt, [sind] z u m Teil miteinander unvereinbar', Doring (1988), p. 48 (avec renvoi a u x pp. 42SS). Cf. Schwartz (1904), col. 741 e t 7 5 9 .
40
ANDRE LAKS
(a) Une pseudo-contradiction
Une grande partie de la these de Doring repose sur la distinction entre deux strates heterogenes de la doctrine cyrenai'que au sein de Pexpose generique de Diogene Laerce. L'affirmation que lafinest le plaisir du corps, comme les Cyrenaiques le soutenaient au temoignage de Panetius (D.L. 4a), s'accorderait mal avec l'admission de plaisirs, comme la joie patriotique, qui ne dependent pas des plaisirs corporels (D.L. 10a). Comme D.L. 10a se situe a la jointure des deux principales sections du resume de Diogene Laerce (entre A et B) et gouverne l'ensemble de la section B, integralement consacree a Panalyse des plaisirs 'psychiques', 92 Doring en conclut que Diogene Laerce juxtapose deux etats differents de la doctrine. Alors que le premier ne reconnaitrait que le plaisir corporel, dont il fait le bien supreme, le second 'suppose une forme de la doctrine qui admet les plaisirs de Tame a cote des plaisirs sensoriels'.93 II faut cependant bien voir que, de l'aveu meme de Doring, l'admission de plaisirs psychiques n'exclut pas ipso facto que les plaisirs corporels soient consideres comme lafin.Seulement, l'insistance sur la superiorite des plaisirs corporels ne se laisse a ses yeux expliquer que comme l'effet d'un compromis passe entre deux versions divergentes de la doctrine cyrenai'que. L'analyse de Doring ne distingue done pas en fait deux etats de cette doctrine, mais trois: (1) une phase primitive, qui ne reconnait que les plaisirs corporels; (2) une nouveaute, qui accorde une place aux plaisirs psychiques; (3) une synthese, qui revient a l'orthodoxie tout en tenant compte de la modification introduite. 94 (1) correspondrait a la doctrine primitive (qu'il s'agisse d'Aristippe l'Ancien ou d'Aristippe le Jeune); (2) correspondrait a Anniceris;95 (3) enfin, correspondrait a une reaction orthodoxe, antiannicerienne mais conciliatrice. L'obligation ou se trouve Doring d'admettre cette troisieme etape constitue le maillon faible de sa demonstration. Nos sources ne 92 94
95
93 Voir supra, p . 27. Doring (1988), p . 44; cf. p . 55. ' W e n n es z u m SchluB heisst, die Kyrenaiker hatten die sinnliche Lust hoher bewertet als die geistige, d a n n liegt die V e r m u t u n g nahe, daB es sich u m eine Modification d e r zu Beginn vorausgesetzten F o r m d e r Lehre handelt, die einerseits zwar d e n V e r a n d e r u n g e n die diese inzwischen erfahren hatten, R e c h n u n g tragt, anderseits aber so viel wie moglich von ihr festhalt', Doring (1988), p . 44; cf. p . 55. ' N a c h d e r Lage d e r Dinge kann kein Zweifel bestehen, daB Annikeris u n d seine A n h a n g e r die ersten waren, die im Kreise d e r Kyrenaiker neben d e r sinnlichen auch eine von ihr unabhangige, eigenstandige seelische Lust anerkannten', Doring (1988), p . 52. Ce serait la, selon Doring, la nouveaute, a laquelle Strabon (cf. supra, n. 14) fait allusion (p. 52s).
Anniceris et les plaisirs psychiques
41
permettent d'associer aucun nom a cette phase presumee de la doctrine cyrenaique, et le seul indice de son existence reside dans le bout de proposition affirmant la superiorite des plaisirs corporels sur les plaisirs psychiques (D.L. 12). La reintroduction, dans la section C, reputee 'incompatible' avec la section B, d'un point de vue temoignant de leur parfaite compatibilite, fait evidemment probleme. De fait, l'admission du plaisir corporel commefin est parfaitement conciliable avec la reconnaissance de Vexistence de plaisirs independants du corps: il existe a la fois des plaisirs corporels et des plaisirs psychiques, mais seuls les premiers constituent la fin. Or D.L. 4 se place exclusivement dans la perspective axiologique de la fin,96 tandis que D.L. 1 o se situe dans la perspective descriptive de l'existence. Sur la base de cette distinction, D.L. 12, loin de proposer une solution moyenne, se presente comme l'explication de D.L. 4: si les plaisirs corporels sont la fin, et non les plaisirs psychiques - qui existent aussi - c'est qu'ils sont 'meilleurs'.97 Naturellement, ceci ne nous dit pas quelle pouvait etre la fonction de la reconnaissance des plaisirs purement psychiques chez les Cyrenaiques. Mais n'y a-t-il pas la une fausse question? II serait assez etrange que les Cyrenaiques aient du attendre Epicure pour avoir l'attention attiree par les plaisirs psychiques,98 alors que ceux-ci avaient ete l'objet, des Platon, d'une analyse specifique. Dans le Philebe notamment, la categorie des plaisirs relevant de Tame est tout entiere rapportee a la memoire." Avant de devenir, le moment venu, anti-epicuriens (avec Anniceris), les Cyrenaiques durent etre anti-platoniciens. Reste a savoir si le temoignage de Clement sur la doctrine d'Anniceris, decisif pour la these Schwarz-von Fritz-Doring, n'infirme pas cette assertion. (b) Clement, Stromates, 11.21.130, 7-8 = iv G 4 G Quant a ceux de la succession cyrenaique qu'on appelle les Anniceriens (1) ils n'assignaient aucun but a Pensemble de la vie, mais disaient qu'il existe un but propre a chaque action, le plaisir qui resulte de l'action. (2a) Ces Cyrenaiques recusent la limite du plaisir d'Epicure, a savoir la suppression 96 97 98
99
Pour Pinterpretation de cette phrase, cf. Bollack (1975), p . 192. 'Meilleurs' doit signifier 'plus intenses' (Doring (1988), p. 44), puisqu'il n'existe pas de difference qualitative entre les plaisirs (D.L.2). ' W e n n als sicher gelten kann, dafi es die Annikereer waren, die diese Lehre in die kyrenaische Philosophic eingefuhrt haben, d a n n liegt es natiirlich nahe zu vermuten, die Auseinandersetzung mit Epikur h a b e dazu d e n AnstoB gegeben', Doring (1988), p. 53; cf. p. 56. Voir n o t a m m e n t Philebe^ 33C5S: KCCI pf)v TO ye ihrepov eT8os TCOV f)8ovcov, 6 TT\% yvyy\% i<pa|iev eTvai, 81a pvf)|jr|S TTOV ICTTI yeyovos; cf. 34C6.
42
ANDRE
LAKS
de la douleur, en Pappelant Petat d'un cadavre; (2b) en effet nous nous rejouissons non seulement des plaisirs, mais aussi des frequentations et des emulations, tandis qu'Epicure pense que toute joie de Tame nait d'une affection anterieure de la chair.100 La doxographie de Clement comporte trois elements: (1) le concept de fin ne s'applique pas a l'ensemble de la vie, mais a chacune des actions dont la vie est composee; (2a) l'absence de douleur n'est pas un plaisir, mais comparable a l'etat d'un cadavre; (2b) il existe une joie independante des plaisirs (corporels). Les deux derniers segments font probleme, en particulier parce que la relation causale qui lie 2a a 2b n'apparait pas immediatement. Mais chacune des trois unites a son correspondant dans la section generique de Diogene Laerce. (1) = D.L. 5. Au plaisir particulier de Diogene Laerce (kata meros), correspond chez Clement le plaisir lie a une action particuliere {hekastes praxeos). Selon Diogene Laerce, ce plaisir peut seul etre appelefin,parce qu'il est le seul a etre choisi pour lui-meme. II s'ensuit que le bonheur n'est pas unefin.Comme le bonheur suppose la prise en compte de l'ensemble de la vie (holou biou),101 la notice de Clement recoupe exactement les indications de Diogene Laerce. Tout au plus pourrait-on noter une difference d'accent. Comme Tensemble' considere est la vie, et non le bonheur, les 'actions' se substituent aux plaisirs en fonction de parties constitutives. II est peut-etre significatif que le plaisir annicerien soit presente, dans la notice de Clement, comme 'survenant' {periginomeneri) a la suite d'une action; car un emploi similaire du verbeperiginomai se retrouve chez Diogene Laerce dans le paragraphe consacre aux Anniceriens.102 Plus important est que cette substitution puisse etre construite comme un element de reponse a Hegesias. S'il est vrai que c'est a nos actes qu'est attache un plaisir specifique, et qui en est la fin, nous ne sommes sans doute pas aussi livres qu'Hegesias le pretendait aux miseres corporelles ou psychiques et aux revers de la fortune. Or nous savons, par Diogene Laerce, qu'Anniceris s'etait precisement 100
ot Se 'Avvixepeioi KOcAoO|jevoi EK TT\S KvpT)vaiK"ns SiaSoxfis T O O IJIEV 6 A O U |3iou TeXos ou6ev cbplCT|iEVOV
ETOC^OV,
EK&C7TT1S
5 1 TTpO^EGOS TSlOV
UTTOCpXElV
TfiXoS
TTJV
£K T ^ S TTpd^ECOS
iTEpiyivoiiEvnv f)6ovf|v O U T O I oi Kupr|vaiKoi TOV opov TT^S f|6ovns 'E-mKoupou, TOUTEOTt TT)V TOO dcAyoOvTos OiTE^aipEaiv, <5cS£To0aiv, VEKpoO KaTaaTactv dTTOKaAovhrrES- x a ^ P 6 l v
y^p
fjlias \xr\ \xovov ETTI f)8ovaTs, dAAd Kai ETTI 6|iiXiais Kai ETTI 9iXoTi|jiais. 6 8E 'ETTIKOUPOS "rraaav Xocpav TTJS y ^ x f i s oiETai ETTI TTpcoTOTraSouCTT) TT) aapKi yEveaSai. 101
Cf. supra, p. 31.
102
Diog. Laer. 11. 96 (texte cite supra, n. 73); l'occurrence de 11. 91 s'applique a ce qui resulte de la sagesse.
Anniceris et les plaisirs psychiques
43
interesse aux 'activities' humaines capables de contrebalancer les souffrances, telles que P ami tie, la reconnaissance, la piete familiale et l'engagement pour la patrie. De ce point de vue, le premier element de la notice de Clement, tout en s'accordant parfaitement avec la doctrine cyrenaique 'generique', semble aussi porter la griffe d'Anniceris. 103 (2a) = D.L. 8. Les deux formulations ne se differencient que par le choix de la comparaison. A la suite de Grilli, on explique la substitution du cadavre (nekrou) au dormeur (Jkatheudontos) comme un effet de la polemique anti-epicurienne: alors que la comparaison de l'etat median au sommeil aurait une valeur purement illustrative, et correspondrait a un etat primitif de la doctrine cyrenaique, Passimilation hyperbolique de la suppression de la douleur a l'etat d'un mort serait specifiquement dirigee contre la definition epicurienne du plaisir.104 Cette distinction ne me parait pas legitime, car la comparaison de l'etat median avec le sommeil n'intervient pas moins, dans D.L. 8, dans un contexte anti-epicurien que Passimilation de la suppression de la douleur a l'etat d'un mort. Ce qui est vrai est que la fonction de la comparaison est differente chez Diogene Laerce et chez Clement. Chez Diogene Laerce, il s'agit d'etablir (contre Epicure) que l'absence de douleur n'est pas un plaisir, parce que le plaisir est un mouvement, tandis que l'aponie, elle, est un etat qui ne connait pas le mouvement. Le dormeur fournit ici un terme de comparaison beaucoup plus adequat que le mort, dans la mesure ou l'aponie, comme le sommeil, n'est evidemment qu'un etat d'immobilite relative; on se reveille du sommeil, comme on peut passer de l'etat intermediaire a un etat de plaisir (ou de douleur). 105 Chez Clement, en revanche, le raisonnement ne s'appuie pas sur la categorie de mouvement, mais sur P omission d'un certain type de plaisir: le cadavre, certes, ne bouge pas; mais surtout il est sans ame. Ce lien, toutefois, ne s'impose que si Pon met effectivement le 103
104 105
On pourrait aussi se demander si soutenir que la vie n'a pas de fin determinee n'a pas une portee plus radicale, a savoir explicitement antiteleologique, que la simple negation du statut de 'fin' du bonheur. Anniceris entendait-il ainsi rendre son du au pessimisme d'Hegesias? En tout cas, il est remarquable que Doring (1988), p. 57, avec la n. 119, voie dans la premiere phrase de la doxographie de Clement une doctrine commune aux Cyrenaiques 'orthodoxes' et aux Anniceriens. C'est qu'elle n'a pas directement a voir avec ce qui est, selon lui, la grande nouveaute annicerienne, a savoir l'introduction des plaisirs psychiques. Grilli (1959), p . 347s; repris p a r G i a n n a n t o n i (1983/85), in, Note 19, p . 171. Dans cette mesure, le sommeil est une illustration de meme nature que la mer d'huile, dans le temoignage d'Eusebe (cite supra, n. 3), cf. Doring (1988), p. 50.
44
ANDRE LAKS
troisieme element de la doxographie de Clement en relation avec le second - ce que les interpretes ont generalement renonce a faire.106 Sous reserve que ce lien soit celui que je suggere, on pourra dire que le couple forme par le dormeur et le cadavre reflete moins deux etats de la doctrine (ante- et post-epicurien) que deux types de critique Tun et l'autre posterieurs a la doctrine d'Epicure, mais pouvant parfaitement etre fondes sur la doctrine traditionnelle. (2b) = D.L. 10a. La troisieme partie de la notice de Clement pose un double probleme. D'abord, que doivent etre les 'plaisirs' (hedonai), pour former un groupe en face des contacts sociaux (homiliai) et des menees de 1'ambition (philotimiai)? Ces deux pratiques tracent un modele de vie oppose a l'injonction epicurienne de 'vivre cache' (lathe biosas): d'un cote, l'ouverture aux autres, d'autre part le souci de reussite personnelle, en particulier dans le domaine politique. Par contraste, les 'plaisirs' recouvrent les plaisirs du corps, qui sont chez Epicure les seuls supports de la joie psychique (comme le rappelle la derniere phrase). Mais il est inutile de supposer que Clement omet de preciser 'corporels' apres 'plaisirs'. Le contexte, et en particulier le fait que les 'plaisirs' servent d'objets a la joie (chairein) rend la chose suffisamment claire.107 Naturellement, qu'hedonais soit reserve aux plaisirs corporels n'empeche pas que nous puissions nous 'rejouir' de plaisirs corporels, a savoir 'a leur occasion' (epi, dans Clement comme dans D.L. 10a). Les Cyrenaiques de Clement ne niaient pas qu'il existe certaines joies ( = des plaisirs psychiques) qui dependent de plaisirs corporels; ils niaient seulement qu'il en aille ainsi de tout plaisir psychique. En cela, leur position correspond exactement au resume generique de Diogene Laerce (memonon I oupasas, D.L. 10a). Plus difficile est de savoir en quoi l'autonomie de certaines joies par rapport au corps peut venir appuyer (gar) le refus de la definition epicurienne du plaisir comme suppression de la douleur: car il en va 106 107
Cf. infra, p. 45. Si la notice de Clement se rapporte specifiquement aux Anniceriens - ce qui n'est pas sur, voir plus loin - cet emploi 'absolu' de f)5ovf| pour designer le plaisir corporel pourrait etre considere comme terminologique. Dans la partie de la doxographie de Diogene Laerce consacree a Anniceris, f)5eoc semble en effet bien se rapporter aux plaisirs corporels: les f)5ea sont opposes a cette serie d'activites (amitie, reconnaissance, piete filiale, engagement pour la patrie) capables d'assurer le bonheur au sage, 'meme si les plaisirs surviennent en petit nombre' (cf. supra, n. 73). Ces plaisirs ne sont done pas les plaisirs psychiques 'resultant' de ces actions memes. Le sens est bien plutot que les 'plaisirs psychiques', dont on doit supposer qu'ils accompagnent les actions mentionnees, suffiront au bonheur du sage, meme si les autres - les plaisirs corporels - sont rares. La distinction ent et fiSovrj/TTOVOS (cf. supra, n. 80) affleure ici.
Anniceris et les plaisirs psychiques
45
chez Epicure du plaisir de l'ame comme du plaisir du corps: a T'absence de douleur', dans ce dernier cas, correspond P'absence de trouble' dans le cas de l'ame. 108 Le parallelisme ne montre-t-il pas que la question de l'autonomie ou de la dependance des plaisirs psychiques a l'egard des plaisirs corporels n'a rien a voir avec la definition negative du plaisir? On a suppose que Clement, ou sa source, comprimait jusqu'a l'inintelligibilite une notice primitivement plus explicite, ou les deux questions auraient ete distinguees. 109 Ne nous trouverions-nous pas plutot en presence d'une interpretation (assez raffinee) de la doctrine d'Epicure? Car, en depit du parallele ame/corps, c'est un fait que le plaisir de reference reste toujours, chez Epicure, le plaisir du corps. Et c'est precisement, comme le rappelle la notice de Clement, parce que tout plaisir psychique se rapporte a un plaisir corporel anterieur que la definition epicurienne du plaisir mentionne la douleur corporelle (tou algountos)110 comme ce dont l'absence ou la suppression definit le plaisir. Dans cette mesure, les cas de plaisirs psychiques autonomes pouvaient parfaitement etre invoques contre le reductionnisme epicurien. Ce qui fait que le plaisir d'Epicure est assimilable - par metaphore, bien entendu - a l'etat d'un cadavre, ce n'est pas seulement qu'il definit negativement le plaisir par la suppression de la douleur, mais qu'il ne reconnait finalement pas l'existence de plaisirs psychiques independants du corps: tout se passe comme si, en reduisant le plaisir au plaisir du corps, Epicure privait l'homme de son ame, autrement dit, le transformait en un cadavre. 111 Quoi qu'il en soit des difficultes propres a la notice de Clement, les recoupements entre la notice de Clement et la doxographie de Diogene sont nets. Doring en a conclu que certains elements au 108 109 110 111
Cf. eg Lettre a Menecee, 128. D o r i n g (1988), p . 5 1 , d e v e l o p p a n t le 'sehr verstummeW d e N a t o r p (1894), col. 2260. Cf. supra, n. 80. L'ecart entre Clement et Diogene Laerce tient moins aux theses elles-memes qu'au rapprochement qui en est fait dans le cadre d'une argumentation opposant l'autonomie des plaisirs psychiques a la reduction epicurienne des plaisirs aux plaisirs du corps. Mais si cette combinaison n'est pas attestee telle quelle dans le resume generique de Diogene Laerce, elle correspond bien a ce qui en est le mouvement general: on pourrait dire que la deuxieme partie de la notice de Clement explicite le nevroi qui, ouvrant la section B de Diogene Laerce, amorce la confrontation entre les Cyrenai'ques et Epicure (cf. supra, p. 27). II est remarquable que D.L.4 etablisse une conjonction similaire entre l'aspect catastematique et l'aspect corporel du plaisir epicurien, qui a egalement arrete les interpretes: comment le plaisir catastematique, qui n'est pas specifiquement psychique, pourrait-il etre oppose au plaisir du corps (f)6ovf)v HEVTOI TT)V TOO acouotTOS . . . ou TT)V
46
ANDRE LAKS
moins de cette doxographie dependaient d'Anniceris, et ne remontaient pas a la doctrine primitive des Cyrenaiques. Mais la conclusion ne suit pas. Pourquoi ne pas soutenir, en accord avec la structure generale de la doxographie de Diogene Laerce, qui fait de la doctrine des 'fideles' une doctrine en partie 'commune', 112 que les Anniceriens partageaient les trois theses rapportees par Clement avec la tradition de l'ecole?113 Cela ne signifie naturellement pas qu'Anniceris n'ait pas ete a l'origine de la confrontation avec Epicure. II doit l'avoir ete, pour des raisons strictement chronologiques. Mais il n'a pas necessairement introduit les plaisirs psychiques, que suppose l'element (2a) dans la notice de Clement, dans une doctrine dont ils etaient primitivement absents; il a tout au plus reorganise le fonds doctrinal ancien en fonction de la position epicurienne, qui s'etait construite contre lui. Naturellement, le degre de dependance de la doxographie posterieure par rapport a Anniceris ne peut plus etre mesure. La mise en parallele des deux doctrines cyrenaique et epicurienne, dont nous possedons de nombreuses traces, parait bien etre l'effet d'une construction doxographique reprenant systematiquement les positions existantes pour les opposer dans leur resultat. 114 En tous cas, il est vain de vouloir a toute force reperer dans la doxographie, et en particulier dans le resume 'generique' de Diogene Laerce, les strates d'une discussion opposant d'abord Epicure aux Cyrenaiques, puis des Cyrenaiques a Epicure, comme Doring a tente de le faire. Car Anniceris pouvait simplement opposer, aux innovations epicuriennes en matiere de plaisir, ce que les Cyrenaiques avaient toujours deja dit. On doit naturellement des lors se demander pourquoi Clement, s'il est vrai qu'Anniceris se rattachait, sur l'ensemble des positions dont il se fit l'echo, au fonds commun de l'ecole, les presente comme specifiquement anniceriennes. Mais jusqu'ou le fait-il? D'abord, Clement pourrait dependre d'une source ou, en accord avec le role qui avait effectivement ete celui d'Anniceris, les Anniceriens repre-
112 113 114
KaTacjTrmocTiK'nv f)8ovf^v TT^V kir' avaipecrei aAyn66vcov KCU otov doxATjcriav)? On peut repondre que les Cyrenaiques opposaient, au plaisir catastematique d'Epicure, le 'vrai' plaisir du corps, a savoir son mouvement (cf. aussi Bollack (1975), p. 192). On aurait l'exemple, chez Clement, d'un schema analogue, faisant jouer contre Epicure non la cinetique corporelle effective, mais un autre type de plaisir. Cf. supra, pp. 28s. Doring (1988), p. 57, l'admet lui-meme pour la premiere partie de la notice de Clement, cf. supra, n. 103. Cf. supra, n. 17. Rien n'interdit de supposer, naturellement, qu'Anniceris soit lui-meme a l'origine de la confrontation.
Anniceris et les plaisirs psychiques
47
sentaient l'ancienne orthodoxie, et done l'ensemble de l'ecole. 115 Une autre hypothese me parait plus vraisemblable, non seulement parce qu'elle permet de preserver l'idee que la premiere unite de la notice de Clement porte la marque distinctive d'Anniceris, 116 mais surtout parce qu'elle permet de rendre compte d'un trait de la notice de Clement qui n'a jusqu'a present guere retenu l'attention des interpretes. On a en effet toujours admis que la formule 'ces Cyrenaiques', qui introduit la seconde partie de la notice, se rapportait, avec un emploi distinctif du demonstratif, aux Anniceriens. Mais comme ces derniers sont presentes, dans la premiere phrase, comme 'ceux des membres de l'ecole cyrenaique qu'on appelle Anniceriens', il se pourrait aussi que l'expression 'ces Cyrenaiques' ne renvoie pas aux seuls Anniceriens, mais, generiquement, a l'ecole dont ils sont issus.117 L'hypothese peut s'appuyer sur la fagon dont s'organise la section de Clement sur le telos qui debute en 130. Apres avoir mentionne les opinions d'Anaxagore, d'Heraclite, de Pythagore (d'apres Heraclide), Clement introduit le groupe des Abderitains en insistant sur le fait qu"eux aussi enseignent qu'il existe une fin'; 118 ainsi la doctrine explicitement rapportee aux Anniceriens apparaitelle, a la fin de l'enumeration des differentes positions abderitaines (Democrite, Hecatee, Apollodore de Cyzique, Nausiphane, Diotime), 119 comme la phrase la plus importante de la notice consacree aux Cyrenaiques, du point de vue de la construction doxographique d'ensemble. Elle nie en effet ce qui en avait ete jusqu'a present le presuppose: selon les Anniceriens, il n'existe de fin qu'a la condition de voir que le telos ne concerne pas l'ensemble de la vie. La doxographie de Clement citerait done Anniceris pour une formulation 115
116 117
118 119
La suggestion est de G. Striker. De fait, imaginons que Diogene Laerce ait expose la doctrine des Anniceriens sans la faire preceder de l'expose generique. Une partie de ce qui figure dans cet expose aurait alors ete explicitement attribue a Anniceris. Cf. supra, p. 43. On doit naturellement y renoncer, si Anniceris represente ici tous les Cyrenaiques (orthodoxes). Cf. la valeur du preverbe dans Trpoacovoaa^ov chez Diog. Laer. n. 85 (cite supra, n. 22). L'hypothese est envisagee, mais pour etre immediatement ecartee, par Grilli (1959), p. 347, n. 2. dXAoc Kcci oi 'A(35r]ptTai TEAOS Crnapxeiv 8I8&CTKOUCTI. Le nom d'Antisthene vient bizarrement s'inserer entre la descendance democriteenne et les Anniceriens. II faut egalement noter que Clement, dans cette section, oppose explicitement les opinions des physiciens sur la fin a celles d'un premier groupe de philosophes qui apparaissent par contraste comme des 'moralistes', bien qu'ils ne soient pas designes comme tels. Ni Antisthene ni les Anniceriens - qui, pas plus que les autres Cyrenaiques, ne se sont adonnes a la physique - ne semblent etre a leur place dans ce second groupe.
48
ANDRE LAKS
qui lui est propre. 12° Des lors, on voit que la suite ne se situe pas sur le meme plan que la phrase sur le telos. Elle ne fait que rappeler, pour memoire pourrait-on dire, ce que les Cyrenaiques en general, et pas seulement les Anniceriens, entendaient par 'plaisir', en situant le plaisir des Cyrenaiques - selon un schema doxographique eprouve, que Clement utilise ailleurs 121 - par rapport a celui d'Epicure. S'il est vrai que la partie anonyme de la doxographie de Diogene Laerce presente une doctrine differenciee, mais parfaitement coherente, des plaisirs cyrenaiques, qu'Anniceris n'a fait que reactualiser, on doit se demander en quoi a consiste le 'renouvellement' annicerien de la doctrine cyrenai'que dont parle le temoignage de Strabon. On remarquera d'abord que Strabon dit qu'il 'passe' pour l'avoir amende. 122 La restriction suggere que le degre d'originalite d'Anniceris par rapport a la tradition avait ete l'objet d'un debat, ou du moins d'appreciations divergentes. L'hesitation, perceptible dans la doxographie de Diogene Laerce, entre une presentation qui fait des principes de la philosophie cyrenai'que une version 'orthodoxe' et celle qui en fait un 'fonds commun' est a cet egard significative.123 De fait, la faijon dont la doxographie de Diogene Laerce articule la triade des 'fideles', des Hegesiaques et des Anniceriens reflete la nature ambigue du 'redressement' que les Anniceriens ont fait subir a l'ecole cyrenai'que. Ce contre quoi les Anniceriens semblent avoir essentiellement reagi, c'est contre Interpretation qu'Hegesias avait donnee de la philosophie cyrenai'que, et qui l'avait conduit a abandonner certaines des doctrines fondamentales de la tradition: c'est, en particulier, la definition negative de la fin,124 le declassement correlatif du plaisir en fonction de skopos,125 et la negation de la possibility du bonheur, toutes theses qui pouvaient bien sembler remettre en cause l'appartenance des Hegesiaques a l'ecole d'Aristippe. Comme nous l'avons vu, la doxographie des Anniceriens se construit chez Diogene Laerce sur le refus des deux theses caracteristiques de la position des Hegesiaques. Ceux-ci nient en efTet la realite de la reconnaissance, de l'amitie et de la bienfaisance; ils nient par ailleurs la realite du bonheur. Les Anniceriens, au 120 121
122 125
Cf. supra, p p . 42s. Cf. supra, n. 7. Clement a deja mobilise le couple dans la premiere partie de sa doxographie sur le T€Aos, celle ou sont rapportees les opinions des 'moralistes' (cf. supra, n. 119): xxi. 127.2; 128.1.
T e x t e cite supra, n . 14. Cf. supra, p . 29.
123
Cf. supra, p p . 2 8 - 3 0 .
124
Cf. supra, p p . 29 et 35.
Anniceris et les plaisirs psychiques
49
contraire, 'laissent subsister' les sentiments altruistes, et, construisant une correlation originale entre altruisme et bonheur, font reposer precisement sur l'existence de sentiments altruistes la possibilite d'un bonheur que les Hegesiaques niaient etre a notre portee. De toute evidence, les Anniceriens avaient, contre le pessimisme hegesiaque, mobilise les vertus d'une categorie particuliere de plaisirs 'psychiques', ceux qui sont lies au don de soi.126 Cela ne signifie naturellement pas qu'ils aient ete les premiers a introduire, dans la doctrine cyrenaique, le genre meme des plaisirs psychiques. Cette double opposition, sur la question d'autrui et sur celle de notre bonheur, est suffisamment tranchee pour qu'Hegesias et Anniceris apparaissent comme les deux poles d'une antithese fondamentale. On comprend que les deux doctrines viennent ainsi symetriquement flanquer, chez Diogene Laerce, un centre constitue par la doctrine des 'fideles'. Mais en s'opposant a la deviation hegesiaque, Anniceris revenait aussi a une certaine orthodoxie, quitte a en modifier les accents.127 Si je vois bien, c'est cette meme orthodoxie qu'Anniceris a du opposer a Epicure, fournissant peutetre ainsi a la doxographie ethique les elements d'un debat dont elle devait abondamment se servir. Dans sa composition complexe, le livre 11 de Diogene Laerce reflete cette position a la fois postepicurienne, anti-hegesiaque et dans cette mesure orthodoxe d'Anniceris. Qu'Anniceris ait ete amene a developper des positions originales par rapport a la doctrine des 'fideles', en particulier sur l'amitie ou la reconnaissance, cela n'est pas douteux. Ces theses ont pu meme etre suffisamment nouvelles pour que ses heritiers, les Anniceriens, aient choisi de se nommer, ou plutot de se surnommer, d'apres celui qui avait assume l'heritage ancien.128 126 127 128
V o i r supra, p . 36. Cela pourrait etre precisement la valeur de 67ravop6cbcjai, dans le temoignage de Strabon (cf. supra, n. 14). II ne semble pas que Ton doive tirer du temoignage de Strabon (cf. supra, n. 14) que les Anniceriens avaient recuse l'etiquette de 'Cyrenaiques'. Cf. Diog. Laer. 11. 85 (cite supra, n. 22).
PART II
Atomism and Epicurean Psychology
CHAPTER 3
Epicurus on agency Julia Annas
A great deal of attention has been paid to Epicurus' views on freedom in various contexts, and also to his metaphysical views on mind and body. In this paper I shall consider a topic which, though it is bound to be of relevance to these, is distinct from them, and cuts across them: Epicurus' account of agency. Any study of agency in Epicurus is bound to start from the fragments of Book xxv of On Nature,1 the understanding of which has been so greatly improved by David Sedley.2 In this paper I shall be concerned with a question which to some extent cuts across the metaphysical issues of determinism and physicalism which Sedley focusses on; nevertheless, philosophical disagreements on some points will be obvious, and I should like to open with an expression of gratitude for Sedley's pioneering studies, which have done so much to bring the important concerns of this book into the centre of Epicurean research. I am also very grateful to Simon Laursen, who is preparing a new text and edition of the whole book, for very
Some of the ideas in this paper go back to a seminar I held in Oxford with Jonathan Barnes, and derive much from him; but naturally he is not responsible for the way they have developed. Others derive from work done for my book, Hellenistic Philosophy of Mind, California, 1992. I am grateful for sustained and useful discussion at the Symposium, which clarified many points for me; from a large number of contributors I would like to mention Jim Hankinson, Geoffrey Lloyd, Jean-Louis Labarriere, Andre Laks, Carlos Levy, Jaap Mansfeld, David Sedley and Richard Sorabji. I am also grateful for comments to Paul Vander Waerdt. I have a special debt of gratitude to Simon Laursen and to David Sedley, both of whom have helped me struggle with a papyrus text which they know far better than I do. Both have been most generous and helpful, and the paper has vastly improved as a result. 1 That this is its number has now been established by S. Laursen (see note 3 below). 2 See Sedley (1983) and Sedley (1989). The first article contains new readings by Sedley of important parts of the papyrus. Portions of this passage are to be found in Long & Sedley (1987) as 20B, c and j of vol. 2 (Greek with textual notes) and 20B and c of vol. 1 (translation with philosophical notes).
53
54
JULIA ANNAS
generous help on various matters. 3 I am sure that many blunders remain, but I hope that this paper will help to encourage more work on these difficult but fascinating texts. 4 Although much remains puzzling about Book xxv of On Nature, some general points are clear: the book as a whole prominently included a discussion of how we are responsible for what we do. A passage at the end 5 refers to 'causes (aitiai) for motions', distinguishing those that are 'through us5 from those that are 'through' 'the nature' and the environment. Epicurus claims to have given an account of these that relates to the pathe and one that gives causes (pathologikos and aitiologikos tropos). There are references to our practices of praising, blaming and influencing people, and also to our 'natural end', pleasure; clearly the book focussed on responsibility as being something relevant to ethics. Epicurus' concern is commonsensical; he appeals not to some metaphysical thesis about freedom but to our notion or prolepsis that we are responsible for what we do. 6 And it is the concern we expect in an ancient ethical theorist: he asks whether the relevant cause is in or through us. Rather than focussing on actions, he is asking whether we are agents, and the discussion is a discussion of agency. Agency, I take it, is simply the capacity to perform actions. We take it that we humans are agents; Epicurus turns out interestingly to share a modern thought that some non-human animals might be agents also. Agents also, of course, perform bodily movements that are not actions; what distinguishes the actions is that they are caused in a distinctive way: they have causes which are also reasons. 7 3
4
5 6
7
See Laursen (1987) and (1988). The latter article discusses the text and interpretation of the passage printed in Long & Sedley (1987), and publishes some new readings of these parts of the text. I am also grateful to Laursen for very helpful correspondence, and for allowing me to see his new readings of portions of this book. The (painfully literal) translations in this paper are all mine. In most cases I have been able to consult the readings of Laursen, whose new edition of the work is forthcoming. In each case I refer to the most recent published version of the passage in question. Often this is Laursen (1988) or Long & Sedley (1987). Where neither is available I have referred to the older edition by Arrighetti (1973), in which the book is [34]. (For ease of reference, all references to Arrighetti are to [34].) Diano (1946) contains a text which sometimes differs from Arrighetti's, but its readings have been superseded by recent scholarship. [34] [33] 1-7 Arrighetti (1973). Sedley (1983). p. 19 lines 24-7; Long & Sedley (1987) 20c (4) (this passage is not in Arrighetti); Sedley (1983) p. 20 lines 38-41; Long & Sedley (1987) 20c (8); [28] 19-27 Arrighetti (1973). And Epicurus' discussion focusses accordingly on the kind of causation that agency involves; his solution, I shall argue, links agency with the giving of reasons and the employment of the rational soul.
Epicurus on agency
55
Epicurus' discussion has negative and positive parts. In the negative part he claims that it is self-refuting to claim that we are not responsible for our own actions and that what we call our own agency is really necessitated.8 I shall focus on the positive indications that we find, fragmentary though they are, as to what Epicurus thinks free agency consists in. The matter is complicated by the fact that Epicurus uses several key terms, and we lack any part of the papyrus where these were clearly related to one another. Firstly we have 'we' or 'ourselves'. Secondly, we have 'the constitution' (sometimes 'the primary constitution'). Thirdly, we have 'the nature'. Fourthly, we have 'the development(s)', a cumbersome and ugly term (ta apogegennemena) meaning, roughly, the way we develop.9 Finally we have 'the atoms'. Clearly, a good understanding of Epicurus' views on agency requires some examination of these terms and how they are related. What we find is at first puzzling: many of these items are first distinguished from others and then apparently run together with them. In the end, however, a coherent and surprisingly simple picture does emerge. It seems at first as though we are distinct from our developments. For we find, in an important passage, (1) From thefirstbeginning we have seeds directing us some towards some things, others towards others, others towards both - in every case seeds, which may be many or few, of actions, thoughts and dispositions.10 Thus it depends on us at first absolutely what becomes of what is already a development, whether of one or another kind, and the things which of necessity flow in from the environment through the pores depend on us when they come about at some time, and depend on our beliefs that come from ourselves.11 How we develop, it seems, depends on us; we seem to be a further factor on which the development depends, as does our reception of incoming information. (We should note that Epicurus does not 8 9
10 11
Sedley (1983) pp. 19-20: Long & Sedley (1987) 20c; [26]-[3o] Arrighetti (1973). Sedley (1983) discusses this part of the argument. See Sedley (1983) n. 18 for a defence of the translation as 'developments'. It is supported by other uses otapogennan in Epicurus (e.g. Ep. Pyth. 92; KD 18; SV81; [36], [16] 6 Arrighetti (unplaced book of the On Nature)). The Greek is syntactically ambiguous, and Sedley translates with actions, thoughts, etc. being what the seeds direct us towards, not what they are seeds of. Sedley (1983)^. 19 lines 1-8; Long & Sedley (1987) 20c (1); [26] 1-15 Arrighetti (1973). The passage continues, 'And if, depending on the nature', and 'constitution' occurs a few lines down, but the text is too fragmentary to reconstruct the sentence.
56
JULIA ANNAS
claim that these things are up to us (epi); he uses the much weaker preposition para.)12 The self or agent here seems to be the person, and the development seems to be the way the person develops, the characteristics she acquires. However, there is another problematic passage which adds complications, 13 and which will concern us in several connections. (2) Many [developments?]14 which have a nature which is capable of becoming productive of both this and that through themselves do not become productive, and it is not because of the same cause in the atoms and in themselves.15 These in particular we combat and rebuke . . . in accordance with 16 their nature which is disturbed from the beginning, as is true of all animals. For in their case the nature of the atoms has contributed nothing to some of their actions, and to the extent of their actions and dispositions, but the developments themselves contain all or most of the cause of some of these things. As a result of that nature some of the atoms' motions are moved in a disturbed way, not in every way through the atoms, but through what enters . . . from the environment into the natural . . . combatting and advising many people together, which is opposed to the necessary cause of the same kind. Thus when something develops which has some distinctness among the atoms 17 in a differential way which is not like 12
13 14
15 16 17
Usener (1977) reveals that Epicurus does not use epi with the dative with the meaning 'up to the person', an idiom common in other philosophers. (Only one uncertain example from a papyrus could have this meaning, and the phrase reports a clearly unEpicurean view.) Letter to Menoeceus 133 clearly contrasts chance with what is para hemas, not epi hemin. Further, none of Epicurus' uses of para with the accusative plausibly have a force stronger than that of'depending on'. The kind of desires one has, for example, depend on {para) one's beliefs. Some of the uses in the Letter to Pythocles are similar to the usage in the sceptical Mode from conditions, where the appearances you have depend on the state that you are in (being drunk, just having had a bath, etc.) but are not produced in a way that is 'up to' that state. This Mode is fourth in Sextus (PHi. 100-17) anc * in Diogenes (xi. 82), third in Philo (deEb. 178-80). Laursen (1988) pp. 7-9; Sedley (1983) pp. 36-8; Long & Sedley (1987) 20B; [21-2] Arrighetti (1973). Long and Sedley (1987) take the subject here to be zoia, that is, the agents or selves, whether human or animal. Laursen (1988), however, argues that the subject must be the apogegennemena themselves. The Greek is syntactically ambiguous; Sedley (1983) takes 'through themselves' not with 'becoming productive' but with 'do not become'. There is clearly a participle here, but the verb is in doubt. Sedley (1983) translates 'hating them' (misountes) but Laursen has doubts about all suggested supplements. Sedley (1983) takes the genitive in the sense 'distinctness/ro/n the atoms', which he takes to support his own claim that the development in question is non-atomic. But, apart from the fact that we are not forced to this conclusion in any case (see below), the phrase here can just as well be translated 'distinctness among the atoms'. Laursen (1988) pp. 12-13, supports this point. Sedley claims that heterotes implies non-identity and not just difference. But this is compatible with the weaker reading; it could imply distinctness from something else (e.g. prior atomic state) among the atoms. I am grateful to Andre Laks for distinguishing these construals.
Epicurus on agency
57
that from a different distance, 18 it acquires a cause from itself, then transmits it at once to the primary natures and in some way makes all of it one. 19
Here the developments are distinguished from the atoms. It is also possible that the agent is also here distinguished from the atoms, though this is not certain, since both places in the text making this point can also be read so that the distinction in fact concerns the development, rather than the agent. However, it is not necessary to take the development and the agent to be distinct items. It is clear in the above passage that the development is the focus of the concern we usually take to centre on the agent - moral concern, that is, of praise- and blame-worthiness. Further, we can easily explain how the agent can be conflated with the agent's development. I am a person who develops in certain ways. I am the whole of that person. But if I am focussing on the way I develop, it is fairly natural sometimes to speak as though I am the part or aspect of me that develops. In a context where the way I develop is being contrasted with something else about me, which does not develop, it is natural to talk as though I identify with, I really am, the way I develop. In another context, however, it is equally natural to talk as though the way I develop depends on me depends, that is, on me as a whole. I am the whole person, then, and have a development; but I can be said to be that development just insofar as I am thought of as developing. The developing self is contrasted with the other three items, which we should look at in turn. First, the references to nature. Epicurus uses 'nature' in several ways, and this is evident here. 'The nature of the atoms' 20 just seems to mean 'the atoms'; Epicurus sometimes uses 'the nature of X', or 'the X nature' as a periphrasis for 'X'. 21 And when he says that things have a nature such as to do so and so,22 he is using 'nature' in an intuitive way to imply that he is talking about the way a thing is, as opposed to its non-intrinsic properties, 23 and 18 19
20 21 22 23
I follow here the suggestion o f j . Barnes noted in Sedley (1988), n. 45. O n 'differential' see Laursen (1988), p p . 13—14. See Laursen (1988) p p . 14-15fordifficulties in identifying the subject here. Long & Sedley (1987) take it as 'he', i.e. the agent. Laursen points out difficulties in importing a hitherto unmarked ' h e ' into the text, though no reading is entirely satisfactory. Sedley (1983) p . 36, line 8; Long & Sedley (1987) 20B (3); [21] 16 Arrighetti (1973). See Ep. Pyth. 97, 113; Usener (1977), 76, 84. [4] I - 7 J atoms (or motions) have a nature such to develop something when forced together; [21] 3—5, some have a nature capable of becoming productive. See Ep. Her. 40, 48, 68, 71; On Nature [24] [48], 7, 17; [24] [49] 4, 8, 27.
58
JULIA ANNAS
also implying that this is something objective.24 But some uses in this book require a different interpretation. Thus we find a reference to a part of the nature which is not upset (aidroumenon);25 we are said to hate some people because their nature is disturbed from the beginning;26 in abusing the nature of some people we are likewise also abusing them.27 We also find the 'natural cause' contrasted with the cause coming about through us, and the cause 'through us' contrasted with the cause 'through the nature'. 28 The nature is thus frequently contrasted with the developing self. The self is also contrasted with the constitution (sustasis). For we find that people29 (3) advise, combat and reform one another, as having the cause in themselves and not only in the original constitution and in the necessity of what environs us and enters into us spontaneously. The contrasts drawn here between the self, the constitution and the information coming in from the environment make it certain that what is in question here is the constitution of the agent's body, or 'the nature', as Epicurus calls it. Sometimes, as here, the reference is not just to the constitution but to the original [prote] constitution. This could mean the agent's constitution at birth, unsullied by any information from the environment; but this is an implausible construal, and the passages make most sense if the reference is to the constitution, on each occasion, prior to the information received from the environment. Finally, the self and the development are both, as we have seen, contrasted with 'the atoms'. Sedley takes this to show that the self is a non-atomic, 'emergent' property. But in the light of all the various relations between the items in play here, it is more reasonable to take 'the atoms' here to be the atoms of the nature, just as 'the constitution' is the constitution of the nature. Apart from other objections which can be made to Sedley's construal,30 if we take 'distinct from 24
See KD 7, 25, 3 1 ; Ep. Men. 133; KS25; On Nature [37] [35], 11—15; Usener (1977), 423, 4 7 1 . [17] 5 - 6 Arrighetti (1973). 26 Sedley (1983), p . 36 lines 5-7; Long & Sedley (1987) 20B (2); [21] 11-12 Arrighetti (1973). 28 27 [ 2 3 ] 2 2 - 5 Arrighetti (1973). [32] 2 2 - 5 , [33] 3 - 4 Arrighetti (1973). 29 Sedley (1983) p . 19 lines 15-19, Long & Sedley (1987) 20c (2), [27] 1-9 Arrighetti (1973). 30 This p a p e r develops a n interpretation of these texts which differs from Sedley's a n d takes issue with some of his specific points. I d o n o t here take o n the larger task of developing objections to his overall interpretation. I shall merely note that: (1) Sedley is correct in arguing t h a t Epicurus is n o t a reductivist physicalist. B u t it does n o t follow that h e is committed to 'emergent' properties. His position could also be a n y one of a wide variety of non-reductive physicalisms. If we are to look for a m o d e r n metaphysical theory to interpret 25
Epicurus on agency
59
the atoms' here to mean 'distinct from all atoms, i.e. non-atomic3, Epicurus will be, throughout this passage, flipping back and forth without any notice between the thesis that the self and development are distinct from the nature and its constitution, and the thesis that the self and development are distinct from anything physical. Fragmentary as the book is, we can hardly suppose there to be such violent dislocation in the pieces of argument that we have. 31 We seem, then, to have three terms with the same referent, namely the person's atomic constitution or nature, which can be referred to as the atoms, the constitution or the nature. Further, we have a passage from elsewhere in the book that provides help: (4) We could talk about this particular thing not just qua (hex) assemblage, but also qua atoms and qua moving atoms or assemblage, not speaking only of the moving itself.32 'Assemblage' (athroisma) is a word that Epicurus uses for the body as opposed to the soul (as at Ep. Her. 63). The point here is that we can talk about the same thing in different ways, and if we use the ^-locution we shall not get confused. There is one thing, the body, but we can talk about it in terms of its being a body, or in terms of the atoms that make it up, for the body is atoms, functioning together in a certain way that constitutes its being a body. If we can extend this point about the qua-\oc\i\ion to other passages (and it is
31
32
these passages, then some variety of token identity physicalism provides a m u c h more plausible a n d economical candidate than emergent dualism. I t also gives Epicurus a position which is not obviously open to powerful philosophical attack. (Sedley does not discuss a n y of the standard philosophical objections to emergent dualism.) Sedley (1989) pp. 301 ff. considers token-identity theories u n d e r reductionist theories, b u t most tokenidentity theories are designed specifically not to be reductionist. (2) I t surely counts heavily against Sedley's theory that Epicurus is always criticized in the ancient world for the alleged inadequacies of his physicalism, never for breaching physicalism in the case of the self. (3) O n Sedley's view Epicurus seems committed to a n incoherent view of causality; w h a t c a n volition-atom 'causation' be, given that it is a thesis of the physics that causation holds only between physical entities? (Cf. Ep. Her. 67, Lucretius 11 161-76.) C o m p a r e also Laursen's text (a new reading) at p . 8 lines 12-13 (the passage is Long & Sedley (1987) 2OB; [21] - [22] Arrighetti (1973)). Disorderly atomic motions are said to come about not because of(dia) the atoms, b u t because of ta empiptonta ... ek touperiechontos, things entering from the environment. Here the atoms are clearly the atoms of the constitution, not all atoms as opposed to something non-atomic. [11] Arrighetti (1973). (I have not seen n e w readings for this from Sedley or from Laursen.) D . Sedley has suggested that the main point of the passage is to contrast talking of moving atoms with using the verb 'to move'. However, I d o not think that this excludes taking the passage the w a y I suggest; the context is unknown, a n d the passage seems to make several distinctions, not just one.
60
JULIA ANNAS
hard to see why we should not be able to do this; it seems a perfectly general point) then references to the atoms, the nature and the constitution can be taken to be references to the atomic constitution of nature qua atoms, qua constitution and qua nature. So far we have seen that there is only one important contrast, not several: the developing self is contrasted with the atomic constitution or nature. The ground for the contrast is that I and the way I develop have a causal capacity which is not just the same as the causality of the atomic nature. This might seem a contrast striking enough to lead to the conclusion that the self and the development are a different kind of thing from the agent's constitution and atomic structure, possibly a non-atomic kind of thing. But, apart from the problems already mentioned, there are passages which stand in the way of such a conclusion. Notably, we twice find that the development itself/WJ a constitution. Thus we find that (5) memory of this or a motion analogous to memory brings about some things at once, and increases others, containing the beginning and the cause, in one way to the primary constitution of the atoms and at the same time of what has been developed, and in another way to [the constitution] when it has been increased, by which we do all our actions.33 And also (6) And even if the primary constitution of our development is forced in some way in accordance with [?] thought, when it is not developing in such a way by necessity as far as these particular things, but is developing from such things in such a way by necessity as far as there coming to be a soul or even a soul having such-and-such a disposition and motion, but not developing in such-and-such a way by necessity as far as having a soul of such-and-such a kind; or not, when it [?] advances in age, with a development in such-and-such a way in accordance with necessity, but being able from itself... and the cause from itself.. .34 The development itself has a primary constitution - that is, as we have seen, /have a primary constitution. And this is the same as the primary constitution of the atoms, i.e. the atoms of the body, and thus is itself atomic. This passage is strong evidence that the developing selfjust is the atomic structure in one of its aspects. The contrast in the case of memory lies not between the atoms and what has already developed (these have the same constitution) but between 33 34
[20] 6-14 Arrighetti (1973). [ 2 4] 3 - I 5 Arrighetti (1973). M y translation reflects textual suggestions kindly m a d e in correspondence by Simon Laursen.
Epicurus on agency
61
both of these together and the constitution when it has been 'increased'. While this is mysterious, it clearly applies to a further change to which the distinction between the developing self and the atoms is not relevant. Further, we are told that in some cases we conflate the development with the constitution, and that this is sometimes correct and sometimes a mistake. Thus we find that sometimes (7) we do not excuse the development from causing (aitia), but, making it one with the constitution, we excuse the latter, but do not advise the former .. ,35 and (8) If because of the cause which is indeed from himself he/it goes for what is similar to the original constitution when this is bad, we sometimes abuse all the more, but more in the way of advising, not as we excuse wild animals, weaving together their developments similarly with the constitution into one thing, but not using either the way of advising and reforming or the way of simply opposing.36 Despite the frustrating nature of these texts, it is reasonably clear that in some cases, for example those of wild animals, we take it that the animal is not capable of developing; what it does is fixed by its constitution. That is, we take it that their behaviour is produced by instinct and that it is not open to the animal to develop in one way rather than another. As Epicurus points out, appreciation of this point affects our attitudes, quite extensively. We may react to a human or tame animal very negatively, 'abusing' it. But we are still in a way 'advising' it; we expect it to learn from what we say and do, unlike a wild animal which just has to be coped with because it cannot learn. Thus to conflate the constitution with the development is to regard the animal as one with fixed instinctive patterns of behaviour and no way of developing them more flexibly. In the case of wild animals this is presumably correct. In the case of humans and some animals we distinguish the development from the constitution, and regard the former as producing actions in a new way; the animal's behaviour goes beyond the fixed instinctive patterns hard-wired into 35 36
[ 2 5] 5~9 Arrighetti (1973). This is followed by several lines which I d o not understand, a n d continues as passage (8). Laursen (1988) p . 17; Sedley (1983) p . 24 n. 18; Long & Sedley (1987) 20 j (only in Greek in vol. 2); [25] 19-34 Arrighetti (1973). T h e opening conditional clause is read in Laursen b u t n o t Sedley.
62
JULIA ANNAS
its nature. But the development itself has a constitution; it is not something different in kind from the animal's bodily make-up. It is, in fact, just the animal developing - or, as Epicurus might have said, the animal qua developing. Thus, although passage (4) does not directly apply to the relation of the developing self to its bodily atomic nature, it can help us to understand that. The developing self is not something different from its constitution, any more than the assemblage is something different from the atoms. We are referring to one thing throughout — the person; but referring to importantly different aspects of it. 37 Epicurus' picture of agency is in fact a simple commonsense and attractive one. We contrast two cases. One is that of a 'wild' animal whose behaviour is brought about entirely by built-in instinctive response. Its responses are inflexible; it cannot develop differential responses beyond what is hard-wired into it, and so cannot learn. Although wasps are not very good examples of 'wild' animals, this sounds very like the classic example of the sphex wasp, described by Dennett in a similar context. 38 By contrast, some animals, and humans, are capable of learning, and thus of the appropriate application of a range ofjudgemental attitudes, because they are capable of more flexible response to their environment than is given in what is hard-wired into them. They have 'seeds', as Epicurus puts it, from the start - that is, they have potential for different kinds of action and response. As Epicurus puts it, they have a development that cannot just be identified with their primary constitution, and they or their developments are causes of a kind that does not reduce to the causality of their bodily constitution and its atomic make-up, for what they do is not determined by the impact of information from the environment on a fixed and rigidly programmed constitution; they have the potential to develop in different ways given their initial constitution, and thus they are agents in a way that the other kind are not. 37
38
Thus there is no discontinuity of the kind that would lead us to posit a breach in physicalism. But apart from this negative point there is nothing in Epicurus' account which compels us to any particular metaphysical conception of the self, and I have not set out to do this in this paper. See D e n n e t t (1973). H e describes a famous experiment in which t h e wasp performs its usual behaviour time after time, even w h e n it has been rendered pointless b y interference with the conditions. 'The experiment unmasks the behaviour as a tropism, rigid within the limits set on the significance of the input, however felicitous its operation under normal circumstances. The wasp's response lacks that free-wheeling flexibility in response to the situation that Descartes so aptly honoured as the infinity of the rational mind'.
Epicurus on agency
63
Thus the point about causality does not imply that the agent is really a combination of two things, the atomic constitution on the one hand, and the development on the other, a further item with distinct causal powers. If it did, we would have, full-blown, all the familiar problems: how to give a sense to this further causality, and, worst of all, how to give an account of what makes this two-part agent a unity. But in fact nothing in the texts compels us to such an account, and the fact that the development itself has a constitution shows that the underlying picture is a more unified one than the proliferation of jargon terms would at first suggest. I am a person, i.e. the whole person, physical and psychological. As such, I have an atomic nature and constitution. However, this nature, in certain kinds of animals (fortunately including mine), contains 'seeds5, potentials for differential response. Because of this I can develop in certain ways that are not fixed by my original constitution. And hence I can bring things about in ways that cannot be ascribed to my atomic nature alone; they have to be ascribed to the way that nature has developed. They are ascribed to me, not to my atomic nature - not because I am something mysterious over and above that nature, but because I am the whole person, the nature and also the way that can develop. I am, if you like, not a mere atomic nature but a developing atomic nature. Thus when Epicurus contrasts the 'cause through us' with the 'cause through the nature' he is not contrasting two mysteriously different kinds of causation. He is just making, in his own way, the commonsense point that I bring things about in ways that are ascribed to me, and to explain what / am we need to refer not just to my having an atomic constitution but to the fact that I am the kind of thing whose atomic constitution can develop - an agent, in fact. The Greek word is, of course, aitia; so far I have doggedly translated it as 'cause', to avoid any ambiguity between cause and responsibility, but we should note that here the 'cause through us' is just what we are responsible for. We can see the force of Epicurus' account more clearly if we put it in the following form: I am responsible for what I do (and so praised, blamed etc.) in a way that the sphex wasp is not, because I have, and it lacks, the ability to go beyond what is hard-wired into our nature and to respond in a differential and flexible way. As often in Epicurus, we find an
64
JULIA ANNAS
attempt to give an account which does justice to atomism and also saves commonsense phenomena, in this case the phenomenon of agency and responsibility. The obvious ancient analogue is Aristotle's discussion of self-movers.39 Can we, however, be more precise as to how I can develop? What is it in me that makes me an agent, in that it makes me able to respond in this flexible way? Here the fragmentary state of Book xxv is particularly frustrating. The remnants from the rest of the book that we have, seem to discuss thinking of various kinds - thinking of oneself, memory 'or a motion analogous to memory' and epilogismos or empirical reasoning. But the passages are extremely hard to make sense of.40 A passage at the end of the book41 apparently discusses concept-acquisition, 'the canon and criterion'. But the only moderately helpful passage is the one we have seen above (6) in which the primary constitution of the development is said to be forced, but only up to a point: it goes as far as having a soul, even a soul with such-and-such a disposition and motion, but not as far as having a soul of such-and-such a kind; and this is said to be a forcing that does not go as far as 'these particular things'. In other words, we cannot choose not to have a soul, or not to have a soul that functions in the general ways it does. But we can make particular choices; and hence it depends on us what kind of soul we develop. I am not free not to get hungry and thirsty; my constitution forces my development so far that I have to want to eat and drink. But that I choose to eat this food rather than another is not something my constitution is forced to do; that is a choice that I make, a way that I develop. And the results of making these choices will of course be to develop me as a person of one kind rather than another; I will turn out either greedy or self-controlled, for example. This passage certainly suggests why Epicurus says that things depend on us {para) rather than that they are up to us (epi); as agents we make choices, but there are always constraints on those choices given that our bodily nature already determines that we want certain kinds of thing, and that our bodies work in certain kinds of way and not others. This point may also help us to understand passage (2), which has 39 40
41
I agree with Sedley, however, that we do not have to regard Epicurus as responding to Aristotle in his views on agency, as m a n y do. [* 5]~~[2°] Arrighetti (1973). T h e frequent references to memory may well reflect a stress on learning from experience, as opposed to responding in a fixed way that does not learn from experience; b u t we cannot draw firm conclusions. [32] Arrighetti (1973).
Epicurus on agency
65
caused problems because it appears that with some people it is simply the development which accounts for what they do, 'the atoms' contributing nothing. We should note first of all that this passage concerns disordered people, in whom something has gone drastically wrong; it is not an account of normal agency. Furthermore, even in these people, it is only in some of their actions that the atoms contribute nothing, 42 and even then the development accounts for all or most of the cause of what they do. So this cannot be a general account, or an account of normal agency. 43 What is at stake here seems to be the condition of people who for pathological reasons have developed in a pathological way; their choices are not made within the normal limits of constraint by our bodily nature, and thus are deviant. Epicurus points out an interesting feature of such cases: on the one hand we recognize that the condition is pathological, i.e. that the disorder has bodily sources, but on the other we recognize that this does not remove responsibility, since the person is still acting as an agent, though a deviant one. 44 What kinds of condition is Epicurus talking about? This is a frustrating passage and it is difficult to come to a satisfactory solution. Epicurus says that the condition in question is disturbed from the beginning, as is the case with all animals initially. The idea seems to be that all animals' natures are open or prone to disturbance or trouble (tarache) initially, though as they mature some can develop in ways that render their natures untroubled. This suggests that the condition in question is the condition of an immature agent, who is in a state marked by disturbances. On the other hand, the problems seem to derive not from lack of development but from the particular developments that have occurred, which suggests not immature but rather deviant or perverse agents. A further difficulty lies in the reference to 'all animals' as having a disturbed nature from the beginning. Epicurus may be criticizing some humans by comparing their troubled state to that of all 42 43 44
I t is puzzling what is meant by the megethos of their actions a n d dispositions; I leave it aside, since I a m unsure what it means. This seems to me another reason to be more cautious than Sedley in basing large conclusions about Epicurus' views on agency a n d free agency on this passage. Cf. [20], 15 ff. (new readings in Laursen): there is a clause with a missing verb, a n d 'the atoms a n d what has developed' in the genitive; then 'from necessity (something) opposed (antixoun) in the case of some things to what has produced a development'. T h e passage is too fragmentary to make much of, but it does seem that there can be conflicts between what does the developing a n d the subject which it is trying to develop.
66
JULIA ANNAS
non-human animals, who are troubled because they lack the capacity to reflect and thus to achieve ataraxia. This would be odd, however, given the way that Epicureans do not draw a sharp line between humans and all non-human animals (we shall return to this). Possibly 'animals' (zoia) includes humans, as it can do, in which case the criticism would presumably be directed against humans who fail to develop beyond the initial stages that we all share, human and non-human animals alike. We can at least establish that the passage does not concern normal agency but the condition of people who are in some way defective, and that their defectiveness consists in their developing in a way which either is or is like that of immature agents. The important point is that, whatever has gone wrong, we still account these people agents, because they are making particular choices of their own, and not acting in automatic or mechanical ways. Since we develop via the making of particular choices, what enables us to do this must be our rational soul, for it must be that which enables us to appraise alternatives and decide on one of them. Agency, after all, is simply performing actions for reasons. What other than the rational soul could enable us to do this? We are reminded at once of the famous passage in Lucretius m. 288-320, where he talks about differences between kinds of animals, and between different individual animals, that are due to their atomic make-up. Lions contain more 'heat'-type atoms, and so are predisposed to fierce anger; deer contain more 'wind'-type atoms, and so are predisposed to timidity, and so on. Humans are the same, we are told, and further, there are natural differences between individuals, who may be predisposed to anger, timidity, etc. These naturae cuiusque animi vestigia prima cannot be wholly uprooted, and doctrina will not get rid of them, although it can produce outward conformity between people. Lucretius concludes that people's naturae and mores differ in many ways, which he will not try to explain, but ascribes to differences in atomic structure. But he insists that usque adeo naturarum vestigia linqui parvola quae nequeat ratio depellere nobis, ut nil inpediat dignam dis degere vitam. (320-2)
The general idea here fits well with Epicurus' book xxv. On the one hand, we are constrained by our atomic natures. These predispose us in some ways, so that our general tendencies are already laid
Epicurus on agency
67
down to some extent by our constitution. However, because we are rational we are agents; we cannot get rid of the tendencies that our nature has given us, but we are not driven by them; we can alter them and adapt to our world in ways that are not predetermined by our natural endowment. Thus even if I am predisposed, in ways I can do nothing about, to anger or fear, I can still by virtue of my rationality overcome the effects of this to a certain extent, and become brave. Indeed, Lucretius is confident that I can do this to the extent of achieving Epicurean happiness - a life, that is, in which these tendencies do not amount to 'troubles' or tarachai, because I have rationally ordered and adjusted my desires so as to achieve ataraxia.
Lucretius does not, of course, tell us just what it is about reason that does this for us. It is clear, however, that he is talking about a kind of reasoning that only humans do; there is no suggestion that deer and cows might achieve ataraxia. On Nature xxv, however, clearly treats tame animals as having some kind of agency which locates them with humans and contrasts them with wild animals. In fact the Epicurean attitude to 'the reason of animals' is in many ways a puzzle, and examining it may help us to understand reason in humans, and its role in agency. It has long been found a puzzle that in the notorious passage (Lucretius 11. 251-93) where the atomic swerve is supposed to explain something about free action, the first example concerns horses, since horses don't have Tree will', at least not in most understandings of that slippery notion.45 However, apart from the puzzles about just what the swerve is supposed to show, the passage makes it clear that a horse has a mens (265, 268) and an animus (170). And Lucretius elsewhere ascribes mind and reason to animals without apparent hesitation; deer have a mens at m. 299. However, the extensive passage deriving from Hermarchus on the origins of justice46 explicitly denies that animals have logos; that is why we can make no contracts with them.47 Hermarchus ascribes the origins of civilization to the fact that humans can either reason as to what is most useful, or follow others that can; they have epilogismos48 45 46 47 48
See H u b y (1969), p p . 17-19 w h o explicitly leaves the point as a puzzle. Fr. 34 ( = Porphyry de Abstin. 1 7-12, 26, 4 ) , in Longo Auricchio (1988). 12.5—6. It is impossible for ta me dechomena ton zoion logon to koinonesai nomou. Cf. KD 32. O n this term see Sedley (1973) p p . 14-17, 27-34. $ e e a ^ so t n e discussions of it in Asmis, (1984).
68
JULIA ANNAS
of what is sumpheron in relations, not just 'irrational memory' (alogos mneme) . 49
How can animals lack reason and yet be described as having an animus or a mens? We get some help here from the (unfortunately fragmentary) opening columns of Polystratus' On Irrational Contempt for Popular Opinions.50 Polystratus' overall position seems to be that animals share broad general features with us, but are importantly different. They take in, but do not understand, 51 certain things. These include (a) prudential concepts like healthy and expedient (b) ethical concepts likefineand base (c) religious concepts like sacred and profane and, importantly for our purposes, (d) signs (semeia). Animals cannot 'discover through signs or through another tropos\ and hence lack four capacities. They cannot use precautions before suffering something; they cannot remember the past so as not to fall into it again; they cannot provide what is expedient and they cannot reflect on their lives as wholes to make them consistent. The second claim is problematic; memory does not seem to rely on inferential capacity, and the claim as it stands looks strikingly false. The text is heavily conjectural here, though; memory seems to come in somehow, but little else is clear. Possibly we should bear in mind Hermarchus' distinction between reasoning and irrational memory; Polystratus may be accepting that animals have irrational memory, but denying them some more intellectual kind. It is a pity that this part of the text is so fragmentary, as it would be interesting to find a disagreement on this point between the two Epicureans; but we can come to no firm conclusions. Polystratus' general idea, however, is clear; like Hermarchus he denies that animals have reason. He is, though, more nuanced about it. Because animals lack value-concepts and inferential capacities 'they do not share in reasoning (logismos) or not one like ours' (vn. 6-8). But it is ridiculous to deny (as the Stoics do) that we are in general ways like animals. 52 49
50 51 52
Hence the rise of h u m a n society is frankly at the expense of animals; the formation of h u m a n groups is twice described in terms of'expelling' animals (presumably from forming part of the same group), and humans are even said to have a kind of'affinity' (oikeiosis) for each other which does not extend to animals. O n this see V a n d e r Waerdt (1988) pp. 87-106. Inlndelli (1978). This seems to be what sunoran means in this context; animals have an appearance of the relevant things, but can make no sense of it. W e have the same sumbebekota, though not the same sumptomata, as he puts it. D . Sedley has pointed out to m e that the w a y Indelli (1978) takes vi. 29 is controversial. T h e to most
Epicurus on agency
69
So for the Epicureans animals lack what we inevitably call 'higher' rational faculties, and because they lack them they cannot achieve happiness or a good life in the way that humans can, since they cannot order their desires rationally as humans can. It is 'sober reasoning3 (logismos) that produces the pleasant life by examining the causes of our choice and avoidance, 53 and reasoning which orders and organizes what is most crucial 54 and measures out the boundaries of pleasure. 55 If this is to be consistent with the way Lucretius talks about animals, and the distinction between wild and tame animals that Epicurus draws in On Nature xxv, with the ascription of responsibility to the latter, 56 then we are forced to the conclusion that the Epicurean animus or rational soul is not a single thing but a cluster of capacities, and that animals have some, and lack some, of these capacities. When one thinks of reason in general one tends to think of these higher faculties; hence there is a tendency for Epicureans to say that animals lack reason. But they also recognize that some animals are, while others are not, agents and responsible for what they do; and this can only be because they have some rational faculties. Epicureans are in general willing to see specifically human capacities as more developed forms of capacities that other animals have. Thus human language use is seen as a development of the ways that animals make different sounds in different circumstances, 57 animals and humans feel sexual desire in similar ways, 58 and both animals and humans have dream experiences that reflect their waking lives.59 It is not surprising, then, that our fundamental ability to be developing beings at all, beings whose atomic constitution contains potentials for developing in different ways, should be shared to some extent by some animals. They, of course, can develop
53 55 56
57
plausibly governs an infinitive in the lacuna after vn, 7; what Polystratus claims to be ridiculous is the opponents' lost conclusion. This does not, however, distance Polystratus from the content of the 'since' clause; it presumably expresses a position shared with his opponents; so it does not affect the conclusions we can draw from the present point of view. 54 Letter to Menoeceus 132. KD 16. KD 19. Cf. Philodemus, de Dis 1 12-15. See Diels (1915). O n this issue see D i e r a u e r (1977)
pp. 194-8.
Lucretius frequently contrasts wild animals {ferae) with tame ones (pecudes, armenta), as at 1. 15, 163,11. 343, 921, 1081-3. Sometimes the reference is to humans, tame a n d wild animals. T h e contrast m a y be a conventional one, b u t it underlines the point that the Epicureans d o not, like the Stoics, naturally see the world in terms of a sharp line between humans a n d all other animals. 58 59 Lucretius v 1056-90. Lucretius iv. 1192-207. Lucretius iv. 986-1010.
70
JULIA ANNAS
only in limited ways, and, as the contrast of wild and tame animals suggests, perhaps mainly in ways that are imposed by us; whereas we can develop in ways that lead to reflection on our entire lives, and grasp of the philosophical understanding that we require if we are to be happy. But there is nowhere any clear Epicurean account of the differences between human and animal rationality, and this may not be due merely to the state of our sources, since unclarity here would explain the unclarity we find on the issue of whether animals are capable of entering into the relationship of justice with humans. 60 We are agents, then, because we are rational; and while different species show themselves to be rational in different ways, what overlaps in humans and other animals and makes them agents seems to be the ability to make different particular choices within the general constraints imposed by the animal's or human's atomic constitution. Rationality is associated with flexibility of response and ability to learn, and contrasted with dispositions which are fixed or tropic. The rational animals are those that have learnt to deal with their environments in ways that are responsive to changes and flexible rather than automatic; that is what makes them agents, beings whose developments are not just identical with their constitutions. (Humans go beyond this, to grasp what is useful; hence the start of social life.) Can we go further as to what it is about rationality that produces this striking result? So far I have said little about it because we glean little about it in the context where Epicurus is talking about agency. He does say that the information we take in depends on our beliefs, and that these 'come from ourselves'. This suggests that belief plays a crucial role in agency. Now if agency shows itself in the having of beliefs, then it would seem that some animals have beliefs, or at least belief-like structures. Polystratus insists that animals do not have inferential or reflective beliefs, but this need not exclude their having something that we would be prepared to call beliefs. It is especially frustrating that we know so little about the role of belief in Epicurean agency, given that the Stoic account makes assent and belief so central. 61 60
61
H e r m a r c h u s ' swift dismissal of justice between h u m a n s and animals on the grounds that animals lack reason seems to contrast with KD 32, which excludes some animals only. O n this see Goldschmidt, (1977) ch. 11 p a r t 1. Goldschmidt sees that even those animals to which justice does apply do not have to be supposed to understand this; their being useful to h u m a n s and h u m a n s protecting them amounts to a kind of implicit contract. See the account in Annas 1992.
Epicurus on agency
71
To these and other problems we would have better answers if Epicurus had given a clearer account of how perception and thinking involve the reception and understanding of content and information. Nonetheless, Epicurus' account of agency, while incomplete and jargon-ridden, turns out to be relatively simple and attractive. We are agents; and so are those animals that share the relevant rational capacity. And this turns out to be the capacity to develop in ways that are not fixed by our original constitution. Agents are those beings that can respond to their environment in flexible and discriminating ways, and can learn from their interactions with it. As accounts of agency go, it is rather convincing, and it avoids implausible corollaries (such as a sharp line between humans and all other animals) which accrue to accounts which start from a more ambitious metaphysical basis.
CHAPTER 4
Democritus and Epicurus on sensible qualities David Furley
The puzzle that provoked this study lies in the apparent difference between Democritus and Epicurus about the value of sense-perception. Epicurus took over the atomic theory from Democritus; he gave the same account, in essence, of the mechanics of sense-perception. Yet Democritus is said by many sources to have been sceptical about the ability of the senses to reveal truth, whereas Epicurus is said to have declared that 'all aistheta are true'. 1 It happens that we have two lengthy criticisms of the Atomists' theory of qualities from the point of view of the Peripatos and the Academy: I mean the long fragment of Theophrastus' book On the Senses, and Plutarch's essay Against Colotes, with its sequel, On the Impossibility of a Pleasant Epicurean Life. What should give these two an exceptional value as historical documents is that the first was written before the author knew anything about the philosophy of Epicurus, or at least at a time when he could take notice only of Democritus among the Atomists, whereas Plutarch aims to criticize the Epicurean Colotes precisely for his misunderstandings and mistreatments of Democritus, as well as for the inadequacies of his own Epicurean theory. A study of the criticisms of Theophrastus and Plutarch might perhaps give us a better understanding of the differences between Democritus and Epicurus, and the reasons for the differences we find. Both authors criticize the position of the Atomists from the standpoint of the theory they themselves subscribe to. Theophrastus' I am grateful to members of the Symposium for comments and criticisms at the time particularly to David Sedley, Andre Laks, Mario Mignucci, Geoffrey Lloyd, James Hankinson, and Eyolfur Emilsson. Malcolm Schofield gave me generous help in revising the paper for publication. I have learnt much from these comments; the remaining faults are, of course, mine. 1 Contrast Sextus Math. VII. 135-6 on Democritus with Math, vii.203 and vm.63 on Epicurus. 72
Democritus and Epicurus on sensible qualities
73
stance in general is clear from a remark included in his comments on Plato: On the natures of theflavours,he does not say what each of them is, if in fact the total differences are four, but explains the affects that arise from them. He says that the astringent contracts the pores and the salty purges - the affect in ourselves - and so on with the rest. But we seek ratherfor their objective reality and the cause of their action, since we actually observe the affects, (de Sens. 89) 2
A passage of Plutarch shows something of the assumptions that guided him in his criticism of the Atomists. In his own voice, he says to Colotes that there was after all no need to adopt the atomic theory from Democritus, but once he has done so, he must take the consequences, however awkward, such as the following: Take for example the quality called hot: where have you got it from, and how has it supervened upon the atoms, which neither came with heat in them nor became hot when they joined together? The first alternative implies the possession of quality, the second the natural capacity to be affected, and you say that neither may properly belong to the atoms because of their indestructibility. So what? [Golotes replies.] Didn't Plato too, and Aristotle, and Xenocrates find themselves generating gold from not-gold and stone from not-stone and everything else from four simple and primary (elements)? Of course. But in their theory, at the moment when the first principles join together to generate each thing they bring along with them generous contributions of their own in the qualities inherent in them; and when they join together and combine, wet with dry, cold with hot, hard with soft, being bodies that change when affected by each other and alter through and through and through, their joint product varies with their various blends. (Plutarch, Against Colotes 8-9, 111 I.DE) 3 2
TCOV 8e x^Acbv TOCS uev <J>OCJEIS OU Aeyei TIS eKaoros, et ocpcc Teaaapes ai Trdaai 8iaopai, TOC 8e TraOri TOC av|j|3aivovTa car' OCUTOOV SrjAoT. avvdyetv y a p TOUS Tropous TO OTpv<J>v6v f\ CTTUTTTIKOV KCU KOc6aip6lV TO d A | J U p 6 v , OTTEp TTCCOOS &7T1V f ) P £ T 6 p O V 6|iOlCOS &*£ KCU TOC OcAAa.
3
£nToOuev 8e OOCTICCV udAAov KCU 8id TI TOOTCC Spcoaiv, eTrei TOC ye Trd6r| 6ecopo0uev. I n my translation the n o u n 'affect' is chosen, reluctantly, to represent the Greek TTOCOOS. oiov euOOs, TO KaAoOiievov 0epp6v uulv iroOev <5C4>IKTOCI KOU TTCOS emyeyove TOCIS CTTOIJIOIS, at lifjT* fjX8ov ix o u < 7 0 c l tepuoTriTa uf)T' iysvovTO Oepuai CTuveA6o0aai; TO pev y a p IXOVTOS Troi6TT|Ta TO 8e TrdCTxeiv TT6<1>VK6TOS, o08eTepov Se TaTs OTOMOIS uTrdpxeiv <|>aTe Trpoof]Kov elvai 8id TT\V dc|>8apaiav. c Ti ouv; oux'i xai TTAOCTCOVI owepaive Kai 'ApiaTOTeXei Kat EevoKpocTei XP UCJOV ^K MT) Xpv/aoO Kai XiOov ex \xr\ AiOou Kai TdAAa yevvocv ex Teaaapcov OCTTACOV Kai TTpcbTcov
; TTOCVU jiev oOv. dAA* eKeivois uev euOOs TB CTUviaaiv ai dpxoci Trpos Tf|V eKaaTou yevsaiv coorrep CTUMPOAOCS peydAas <|>epouaai TOS ev ocuTals Troi6TT|Tas, Kai OTOCV o~uveA6coaiv eis TO OCUTO Kai auuTreacocn §r)pois u y p d Kai y u x p d Oepuols Kai OTeped piaAOocKoTs, CTcb|iocTa KivoOueva Troc6r|TiKcbs Cnr* dAAr|Acov Kai peTapdAAovTa 8i' oAcov, eTepocv d<J>' eTepas Kpdaews auvocTroTiKTei yeveciv.
74
DAVID FURLEY
Theophrastus, we saw, seeks the real nature (ousia or phusis) of the objects of sense perception: that is to say, the sensible qualities that according to Aristotelian theory are inherent in substances. The position favoured by Plutarch attributes qualities to the basic elements, and those qualities, by mixing or some other process, can generate other qualities. But all the Atomists deny that the elements have such qualities. It is true that there is a considerable problem about Plato's position: Plutarch happily puts Plato and Aristotle together, whereas Theophrastus points out that Plato often comes close to Democritus on this subject. But we can bypass that complication for the moment. It is clear enough from these passages and others that our two authors quarrel with the Atomists because both of the latter deny that sensible qualities exist at the level of primary elements. The fact that Theophrastus and Plutarch declare their own position and do not pretend to give us an objective history of Atomism helps us to understand and make allowances for the emphases they give to their account of the Atomic theory. It is of course deplorable that we have no reliable touchstone on which we can test the genuineness of reports of Democritus, and indeed only sketchy fragments of Epicurean theory in the original. But so it is, and we have to make the best of it. We have some sort of control in the case of Theophrastus, de Sens., because the sixth book of his de Causis Plantarum goes over some of the same ground, on the subject of flavours and smells, and because he reports and criticises portions of the theory put forward in the Timaeus. He wrote with the text beside him: that is clear from a comparison (to take an example) between de Causis Plantarum vi. 1.4-5 a n d Timaeus 65C7-66C7.4 There are omissions - forgivably, Theophrastus cuts out several lines from the impossibly complex sentence in which Plato explains the quality called oxu - but the order, sense, vocabulary, and emphasis of Plato's statement are preserved. In de Sens. 84 the passage is still more abbreviated, but nothing is added or distorted by Theophrastus. Admittedly, it is possible to find places in both de Sens, and de Causis where Theophrastus makes small but significant distortions in his reports of Plato, and it can be shown that they are made with a view to strengthening his objection that Plato fails to give a proper 4
Through the kindness of Professor W. Fortenbaugh, I have been able to consult the Loeb edition in advance of publication.
Democritus and Epicurus on sensible qualities
75
account of the real nature of the flavours.5 But in spite of this, a fair idea of Plato's general strategy in the Timaeus could be retrieved from the text of Theophrastus. He may have treated Democritus with less respect: we can hardly know. But especially when he applies peculiarly technical words to Democritean atomic shapes and their actions on the physiology of the observer, his treatment of Plato gives some reason for expecting him to be fairly accurate in this case too. Theophrastus attributes a causal theory of perception to Democritus; he gives details of the sizes and shapes of atoms that according to Democritus regularly produce this or that affect {pathos) in the perceiver. There is no reason to doubt that he knew the theory of Democritus, or that he had an accurate understanding of its nature. The question to be discussed is: what are the differences between Democritus and Epicurus on the objects of sense perception, according to these two accounts? We may begin with what they certainly have in common: the elements of the material world have no qualities other than size, shape, the capacity to resist, and perhaps weight. Although none of these qualities is in itself imperceptible, they are of course imperceptible as properties of atoms, since no atom is in itself perceptible. Differences arise between Democritus and Epicurus concerning the status of those properties of bodies that are normally listed as the special objects of the five senses: colour, sound, flavour, smell, and feel (i.e. hard/soft, rough/smooth, etc. - strange that there is no proper word for it). I want to make an effort to identify these differences as precisely as possible. These differences have sometimes been expressed in what I take to be misleading language. An example from a recent article on Epicureanism may help to explain the point: The world now [sc. on Democritus' theory] turns out to be in reality utterly different from the way it is perceived to be: all the colours, states of mind, etc., turn out to be not real (etee), but just arbitrary [my italics] ... constructions placed by the experiencing subject on atomic aggregates which in the last analysis are quite devoid of such properties.6 It may be helpful to consider what according to Democritus' theory may truly be called arbitrary, and what may not. I concede that the actual sounds or letters used to express the properties of 5
See Thompson (1941), pp.112-18.
6
Sedley (1988), p. 298.
76
DAVID FURLEY
perceived objects are arbitrary: Democritus did not take the position taken by Socrates in Cratylus 424b ff., that syllables and letters imitate the essence of things, as a painter's colours imitate the appearance of things. But given that 'red' is the right name for the top light of a traffic signal, we cannot arbitrarily choose whether to say 'the light is red' or 'the light is green'. We must say, if we intend to speak the truth, how it looks to us. Moreover, how it looks to us is not an arbitrary matter. In normal conditions, it looks red to us, although in abnormal conditions it may look yellow - that is, it may look to us the same colour as the second light. But it is not an arbitrary matter whether we now say 'it looks yellow to me' or something else. If we know what we mean, and intend to speak the truth, 'yellow' is what we must say. To generalise, I want to claim that there is nothing arbitrary, according to Democritus' theory, about the colours we see, or about our other sense experiences. But sometimes we say 'the light is red' rather than 'the light looks red to me'. As we shall see, this is strictly improper in Democritus' theory, but it is still not arbitrary. The theory holds that the appearance of red is explained by particular dispositions of the atoms in the object, which produce just this appearance when the sense organs of the perceiver are adapted to receive it. The statement 'this is red' therefore makes a claim about the composition of the object itself; if it did not contain atoms of such and such a kind disposed in such and such a way, it would not produce the appearance of being red. Let me confront the obvious objection at once. What did Democritus mean by his most famous remark about sensible qualities, mentioned for the first time by Plutarch: He [sc. Colotes] says that Democritus' claim that by convention there is colour and by convention sweet, and by convention a compound and the rest, in truth, the void and atoms, is said in opposition to the senses, and anyone who remains consistent with this saying and puts it into practice couldn't even conceive that he himself is a man and alive. (Plutarch, Against Colotes VIII. 11 IO.EF = DK.68B9)7 7
TO y d p v6|ico XpoiT)v £ * v a i K «i v6|jcp yAuxO KOCI v6|jcp auyKpicnv, (KCCI TOC aAAa, ETef) 8E TO KEVOV Kcci) T&S onrouous, dn/Tsiprmevov <|>rjo"iv OTTO AriiJOKpiTou TOLS aicr6f)a6ai, KCCI TOV EUUEVOVTOC TW Aoyco TOOTCO Kai xpwpevov o08' dv CCOTOV cos (dv0pcoTr6s) EOTIV f| £rj 5iavor|6fivai. T h e r e are m a n y textual problems in these lines. T h e word aOyKpicxiv does not occur in either of the two later quotations of this saying; TriKpov occurs in its place in both: in Galen ( = D K 68.B. I 25 'v6|ico XPOI11> voucp yAvKU, vopcp TriKpov', eiTrcbv, 'ITETJ 6' onroua KCCI KEVOV') and Sextus ( = D K 68.B.9 'voucp' y d p (j>r|<Ti 'yAuKU, [KOU] vouco TTiKpov, vouco Oepuov, vouco yuxpov, vouco XPOI11» £T6 tl S £ cxTOjia Kai KSVOV') . I n line 2 of
Democritus and Epicurus on sensible qualities
77
The true sense ofnomoi in this fragment, as Westman (1955) saw, is suggested by the well known fragment DK B 9 of Empedocles about generation and destruction, also quoted by Plutarch in this same essay: When they [the elements?], mingled, come into the air in the form of a man, or of the kind of wild beasts, or of bushes, or of birds, then they say it was a birth; and when they are separated, then again they name baneful death. They are (?) right; I myself follow the convention too. (Plutarch,
Against Colotes xi. 1113.AB)8
Whatever the reading in the last line, where the MS text does not scan and editors commonly insert a negative, Empedocles' meaning is on the whole clear enough. There is no real birth, but only rearrangement of the elements in new patterns. But no harm will come from using the language of birth and death, so long as we realise what each of them really means: that is to say, how each of them is to be analysed. I think we can understand Democritus along similar lines, although the situation is more complicated in this case. The compound called honey is a mass of atoms and void; its components have no taste. The sensation of taste is an affect of the taster's sense organs produced by the shapes and distribution of the component atoms. By convention, we call honey sweet, but it is really of course just atoms and void. The convention does not consist in the creation of a the above quotation from Plutarch, Sandbach (1941) suggested that iriKpov should be read, for the MS auyxpiaiv: 'It is absurd to say,' wrote Sandbach, 'there is anything "conventional" about combination (of atoms, presumably) . . . Perhaps the error was due to a marginal note KOCTOC ovyKpiaiv (sc. sori XP°l
oi 6' 6T£ pev KOTOC <J>WTOC uiyevT' eis cciOep' I(KCOVTCCI) (?)
f\ KOCTOC 8-npwv dypoTEpcov yivos f\ Konrd ©auvcov f)6 KOCT' oicovcov, TOTS UEV TO (Aeyoucn) ysv£cr6ai, eOTg 5' 6rrroKpiv6cb(Ti, TO 6' au 5v<j8aiuova TTOTUOV f) ? OEUIS ? KaAeouat, vouco 5' eirtyriui KCCI OCUTOS.
78
DAVID FURLEY
name for something that is merely imagined: it consists in taking this name to be the name of some property of a collection of atoms - a property like shape, size, or weight. 'By convention there is colour [in the object] . . . ' 9 Neither Democritus nor Epicurus doubted that terms denoting sensible qualities have a use, and that whatever it may be that they denote can be identified and reidentified. We all know what we mean by black and white; we can in general agree with each other where to find in our shared visual field objects that produce the same appearances in each of us. Although this is true in general, at least of human beings, it is not always true: people in good health have different pathe from sick people, in circumstances that are otherwise the same. Animals may certainly be thought to e x p e r i e n c e d ^ that are different from those of humans. This is a principal reason why Democritus held that perceptible qualities are not elementary. But what the theory reported by Theophrastus clearly aims to explain is how it comes about that there is a regular and normal association between certain external objects and certain pathe. That the pathe of perception are determinate according to Democritus is explicitly testified by Theophrastus: And yet, even if all don't experience the sweet and the bitter from the same things, still the nature of the sweet and the bitter appear the same to all.
This even he himself would appear to attest. For how could what is bitter to us be
sweet or astringent to others, unless these [qualities] had some determinate nature? (Theophrastus de Sens. 70)10
Democritus claimed that the honey is not sweet: the sweetness is an epiphenomenon, which occurs in some way when it touches the sense organs. But the epiphenomenon in this case is rightly called 'sweetness', and it would be wrong and perverse to call it 'bitterness'. If we take Democritus' saying along these lines, we can understand why he offered the elaborate analysis of sensations reported by Theophrastus in On the senses. There is no point in analysing some9
10
M. Mignucci, in a written note to me, suggests that there is a difference between Empedocles and Democritus in the use of vouco. 'Birth' is used metaphorically according to Empedocles. Sweet is predicated truly of honey according to Democritus, but the use is conventional in the sense that 'sweet' has no denotation. KOUTOI 6i |if) KOU Sia TCOV auTcbv yiveTai TTOCCTI TO yAuxu KCCI TO mKpov, aAA' f\ ye 4>0<JIS TOO
TTiKpoO Koci TOO yAuxeos T) auTf) <j>aiv£Tai Traaiv. OTrep KOU OCUTOS av 86£ei6v STnuapTUpeTv. TTCOS yap dv TO TJIJITV mKpov aAAois r\v yAuKU KOCI orpu<|>v6v, el \ir\ TIS r\v cbpta|i£vr| <|)0ais auTcbv;
Democritus and Epicurus on sensible qualities
79
thing that is merely arbitrary (in the sense of 'capricious') except perhaps as an exercise in psychoanalysis. In fact, Theophrastus reports Democritus' theory in great detail and obviously from first hand knowledge of sight (50-4) and hearing (55-6), and then switches to the question of the objects of sense-perception (ta aistheta): tangible qualities (61-2), tastes (65-70), and colour (73-82). In all these he relates particular sensible qualities with particular shapes and arrangements of the elements. Theophrastus contrasts Democritus with Plato on the question of ta aistheta:
But Democritus and Plato went most deeply into the subject, in that they define each individually, the latter, however, without taking away the nature of sensible objects, Democritus making them all affects of the sense. (deSens. 60) n
Heavy and light, hard and soft (he continues) are in a special category in the atomic theory, because these are properties of atoms. But of the other sensible objects he says that none has a phusis but all are pathe of the sense as it is altered, from which arises the phantasia ... A sign that they are not in nature is that they do not appear the same to all animals: what appears sweet to us appears bitter to some, pungent to others again, or piquant, or sour, and so on. Furthermore, [people] themselves change in their temper according to their affections and their age. Hence it is clear that the disposition causes the appearance. (63-4)12 What one perceives is not something that has & phusis of its own: it is a pathos of the sense. But Theophrastus goes on to report a dogmatic causal theory of each pathos, based on the shapes, sizes, and distribution of the atoms composing the object from which the sensation originates. He cannot offer an explanation of how the disposition of the perceiver changes the perception; indeed he seems to complain that Democritus leaves this question wide open. After mentioning some other criticisms he continues: ATIMOKPITOS 6e Koci TTAaTcov em TTAEICTTOV eiaiv fjiinevoi, KOC0' eKoccrrov y a p d(J>opi£ouai, TTATJV 6 |iev OOK orTTOorepcov TCOV aicr6r|Tcbv Tr\v (J>UCTIV, ArmoKptTos 5e iravTa TTOOTI THS aicr0f)CTecos TTOICOV.
TTEpi [xkv < ouv > papeos KOCI KOU<|>OU KCCI <7KAr|poO xai liaAaxou ev TOOTOIS aopi£ei. TCOV 6S ccAAcov aia6r|Tcbv ouSevos eivai 4>0aiv, aAAa TravTa Tr&8r| Tfjs aicj6f|avTaaiav... armsTov 8' cos OUK eicri uorei TO [xr\ TOUTOC TT&CTI <j>aivecr6ai TOIS £COOIS, dAV 6 f|nfv yAuKu, TOOT' aAAois iriKpov KOU eTepois 6£0 KOU dAAois SpijiO TOIS 6E OTpu<|>vdv KOU Ta aAAa 6' cbaauTcos. 2TI 6' auTous |i€Tap6AAgiv TT\ Kpaaei KOTO TOC TTOOTI Kai Tas f|AiKias- ij) Kai <j>av6pov, cos T) 6id6£ais alTia Tf^s (J>avTaaias.
80
DAVID FURLEY
It is absurd also to claim that the same thing appears to everyone perceiving the same things, and to question the truth of these things, having also said previously that things are unlike to unlike dispositions and again that one person attains the truth no more than another. You'd expect the better [to attain the truth] more than the worse, and the healthy more than the sick, [the former being] more in accord with nature. (69-70)13 But it is worth noting that in describing aistheta of which it is particularly true that they appear differently to different people he says firmly (for instance) that 'The sharp [taste] is, in shape, angular and twisted and small and delicate . . . The sweet is composed of round shapes that are not too small' (65). So far as I have noticed, there is no hint in de Sens, that these same shapes produce different pathe in different people. This is important. In so far as he accounts for the differences at all, he does so by means of the thesis that each perceptible compound contains a mixture of shapes. Of all the shapes none is unblended and unmixed with the rest, but in each of them there are many, and the same [taste?] has something of the smooth and rough and round and sharp and the rest. Whatever there is most of in it, that has most strength with regard to sensation and power. Furthermore, the kind of disposition that it enters makes no small difference. For that reason the same thing produces opposite affects and opposites produce the same affect sometimes. (67)14 The theory must involve the notion of the fitting of different shapes to different dispositions of the sense organs - a theory that is amply attested for Empedocles in the earlier sections of Theophrastus' de Sens. If the shapes that produce a sweet tastefitthe structure of the sense organ, the taste is sweet; but if those shapes do notfitand some others in the mixture do, then even if the former are preponderant the taste is different.15 13
14
OTOTTOV 5E Koci TO TTaaiv d£iouv TOO/TO <J>aiv£o6ai TCOV OCUTCOV aiadavouEvois, Kai TOOTCOV TT^V aAf)6eiav IXeyxsiv, Kai TCCC/TOC eipriKOTa TrpoTEpov TO TOIS dvouoicos SIOKEIUEVOIS dvouota <J>aiv6a6ai Kai TrdAiv TO urjOlv udAAov ETEpov ETEpou TuyxdvEiv Tfjs dAr|0Eias. etKos y a p TO (3EATIOV TOO x £ ip o v °S K a l T o uyiatvov TOO KauvovTOS" KOTO (|>UCTIV y a p lidAAov. diravTGov 6e TCOV axTiuaTCOv ou5ev dxepaiov eivai Kai duiyES TOIS dAAois, dAA' EV EKOCCTTCO TTOAAOC Eivai Kai TOV OCUTOV EXEIV Aeiou Kai Tpaxeos Kai Trepi<|>EpoOs Kai 6£eos Kai TCOV AOITTCOV. oO 5' dv £vri TTAEICTTOV, TOOTO ndAiaTa gviaxOeiv irpos TE TTJV a!a0Tiaiv Kai T-nv Suvajjiv. ETI 8E EIS OTToiav i£iv dv EIQEAOTI, 5ia<|>£pEiv [ y a p ] OUK oAiyov. Kai 5id TOOTO TO auTO TdvavTia Kai TdvavTia TO OOTO TTOOOS TTOIEIV EVIOTE.
15
In de Causis 11.2-3, Theophrastus complains that Democritus does not, as he should, explain the effects of the sense-organs on the nature of the sensation, and he takes it for granted that the same atomic configuration can produce different affects (TO yE TOO X U ^O^ crxfiiJia TOCUTO, Kai EV dAAcp EOIKEV OU TOUTO 6Ovaa0ai TravTaxoO TTOIEIV). This is, however,
Democritus and Epicurus on sensible qualities
81
If this is right, then Democritus is entitled to claim from the evidence of the senses, not that the perceived external object is sweet or red or violet-scented, but that it contains atoms of such shapes and dispositions as cause these appearances. He can claim that the senses allow inferences from the pathe to the atomic composition of external objects. It is indeed by inference from sense perception that we get some (not all) of the various theses of the atomic theory. So Galen reports with approval Democritus' famous retort of the senses, after the mind has pointed out that sensible qualities have no existence at the 'real' elementary level: 'Poor mind, do you take your convictions from us, then throw us down? That throw is your fall' (Galen, Med. Exp. 15.7 = Democritus B.125).16 It is time to return to Plutarch and Colotes, and consider the Epicurean Colotes5 charges against Democritus. The first objection he [Colotes] brings against Democritus is that when he says that every single thing is no more such than such, he has thrown our life into chaos. (1 IO8.F)17 It should be clear from our reading of Theophrastus that if Democritus did indeed say this, he meant that the ordinary objects of the sensible world [pragmata) do not in themselves have sensible qualities: they are no more hot than they are cold, no more black than they are white, and so on. They differ from each other in the shapes, sizes, and distribution of their component atoms, and these cause the appearance of different qualities. We know that the appearance may change according to the disposition of the perceiver, because different dispositions of the perceiver's soul register or respond to different atomic configurations. Plutarch continues the passage by claiming, wrongly, in my opinion, that Democritus cannot have held the view that nothing is more such than such, because he attacked Protagoras for asserting precisely this. Plutarch claims that Colotes simply misread someTheophrastus' own notion of how Democritus must explain conflicting sensations. In de Causis vi.7.2 he suggests that a mix of atomic shapes in an object must be the cause of its changing its flavour: the cause cannot be a change in the atomic shapes, hence it must be a resorting of the atoms of different shapes. These are Theophrastus' speculations. They seem to indicate that Democritus' text was itself unclear about these questions. 'TdAaiva (j>pf)v, Trap' fjuecov Aa|3o0cra T&S TTICTTEIS rmsas KcrrapaAAEis; Tncoua TOI TO
16
17>
EyKaA€T 8' OCUTCO irpcoTov, OTI TCOV Trpayu&Tcov EKCCOTOV eimbv ou uaAAov TOIOV
elvoa ovyKEXVKe TOV piov.
82
DAVID FURLEY
thing that Democritus wrote: namely, that 'the den is no more than t h e mederC.
There is no reason to doubt that Democritus wrote this latter phrase, meaning that what really and fully is, namely the atom, is no more than what we call 'nothing', namely the void. The actual phrase quoted by Plutarch is not confirmed by any other source, but Galen (Hipp. Elem. 1,2 = DK 68.A.49) confirms that Democritus used the odd word den to refer to atoms, and meden to refer to void. But there is no reason to think that Colotes twisted this sentence into TCOV Trpayu&Tcov eKaaTov . . . ou uaAAov TOIOV f| TOIOV eivca, since Theophrastus shows that Democritus must have made some such statement. 18 The qualities that enable us to call a thing toion e toion do not belong to things in themselves, but grow from the interaction of the atoms composing things and the sense organs of the perceiver. Admittedly, there is some remaining obscurity about the status of those properties that do belong to compounds of atoms per se, namely shape, size, and weight. But for the rest of the range of qualities, ou mallon toion e toion is a pretty accurate statement of Democritus' view. But Plutarch's objection, that this makes Democritus' position the same as that of Protagoras, whom he attacked, is invalid. If Plato was interpreting Protagoras correctly in Theaetetus I5i.e— 152.a, then Protagoras inferred from the different perceptions of one and the same thing by two different people that the object is what both of them take it to be. The wind is both hot and cold, no more one than the other, according to Protagoras; according to Democritus, it is neither. Democritus would think that Protagoras was quite wrong about the nature of sensible qualities. But Plutarch now claims that the charge levelled by Colotes against Democritus actually applies to Epicurus: he too was committed to the view, ou mallon toion e toion, and so he also threw our lives into chaos, if Democritus did. There is, of course, no statement of Epicurus that says this directly; Plutarch infers it from Epicurus' notorious claim that all sensations are true, together with the admission that people may receive different sensations from the same thing. 18
There is confirmation that Democritus used the notion of ou uaAAov TOIOV f| TOIOV in fr. 8 of Theophrastus Phys. Opin. = DK 68.A.38; but there it refers to the shapes of atoms. See further De Lacy (1958b).
Democritus and Epicurus on sensible qualities
83
If one of two persons says the wine is dry and the other says it is sweet, and neither is wrong in his sensation, how is the wine more dry than sweet? . . . If someone says that one and the same phainomenon is this to this person and that to the other, then he has unwittingly said that it is both. (1 iog.B-c) 19
This is a very genuine problem that Plutarch has raised, and it seems to me that the Epicureans were very hard pressed to find an adequate answer. Epicurus took over from Democritus the thesis that sensation is brought about by an encounter between eidola flowing from the surface of external objects and the organ of sense: External things would not imprint their own nature in respect of colour and shape ... so [well] as when certain 'prints' enter into us, from the external objects, similar in colour and shape, according to the appropriate size either into our sight or into our mind. (Ep.Hdt. 49)20 This text begins to show both the similarity between Democritus and Epicurus, and a difference between them. Both adopt the theory of eidola; but since Epicurus says that the tupoi (presumably another term for eidola) are 'similar in colour and shape' to the external objects, he is giving to external objects and to eidola at least one property denied them by Democritus, namely colour. He stays with Democritus, of course, in denying colour to atoms. Lucretius confirms that colour is a property of the external object itself (de Rerum Natura iv.72-89), and in the following lines he implies that the same is true of smell, the tangible quality of heat, and taste. We certainly see that [objects] not only from deep inside, as we said before, throw off and distribute many [bodies] but also often from their surfaces their colour itself, (iv.72-4)21 19
20
EI y a p 8uoTv AeyovTcov TOO JJEV auor-npov eivoa TOV oivov TOO SE yAi/Kuv p yeOSEToa TTJ orio"6f)<7Ei, TI [idiKKov 6 oivos a u o r r i p d s "n yAi/Kus EOTI; . . . EI y a p TO aCrrd 4>aiv6|JEVov ETEpov ETEpco <|>daK£i TIS, a|j<j>6TEpa Eivai XEycov XEATIOEV. TO auTO <(>aiv6|jEvov is t h e r e a d i n g of W y t t e n b a c h (i 7 9 5 - 1 8 3 0 ) . T h e MSS h a v e a u T O cj>. E i n a r s o n a n d D e L a c y (1967) r e a d auTO T O $ . , a n d t r a n s l a t e ' t h e object itself is w h a t a p p e a r s o n e t h i n g to o n e person e t c ' . W e s t m a n (1955) inserts (d|j(j)OTEpois TOIOOTOV Eivai) after (f>acrKEi TIS. ou y a p a v EvaTroCT<|>payiaaiTO TCJ E£CO TT)V EauTcov <|>0CTIV TOO TE xPc*DJJia"TOS K 0 " T"ns HOp^ilS . . . OUTCOS COS TUTTCOV TIVCOV ETTEiaiOVTGOV f | | i l v
21
OTTO TCOV TTpOyMaTCOV OJJIOXPOGOV TE KCCl
6|ioio|i6p<j>cov KaTa TO EvapiaoTTov |jiEyE8os E!S TT^V oyiv f\ TT\V Sidvoiav. nam certe iacere ac largiri multa videmus, non solum ex alto penitusque, ut diximus ante verum de summis ipsum quoque saepe colorem.
84
DAVID FURLEY
Furthermore, the reason why smell, smoke, heat and the like come streaming out of objects in shapeless clouds is that they originate in their inmost depths, (iv.90-2)22 From certain objects there also flows a perpetual stream of odour, as coolnessflowsfrom rivers, heat from the sun, and from the ocean waves a spray that eats away walls round the sea-shore. Sounds of every sort are surging constantly through the air. When we walk by the sea-shore, a salty tang of brine enters our mouth; when we watch a draught of wormwood being mixed in our presence, a bitter effluence touches it [sc. our mouth], (iv. 218-24)23 These texts are sufficient to show that the Epicureans parted company from Democritus in allowing that perceptible qualities are properties of external objects, not merely pathe of the senses. And this is plainly an important difference. But before we return to Plutarch and the challenge he posed, it may be worth digressing to note that Lucretius' language seems to suggest some hesitation with regard to at least one of the senses, the sense of taste, when he comes to give a fuller account of it later in bookiv: When the trickling particles of the juice are smooth, they affect the palate sweetly and sweetly tickle all the moist regions of the tongue in their circuitous flow. Others, in proportion as their shape is rougher, tend to prick and tear the organs of sense by their entry. The pleasure derived from the juice does not extend beyond the palate, (iv.622-7)24 This may well suggest that the sweetness or sourness is generated only in the palate. 22
praeterea omnis odor, fumus, vapor atque aliae res consimiles ideo diffusae (e) rebus abundant, ex alto quia dum veniunt extrinsecus ortae.
23
perpetuoque fluunt certis a b rebus odores; frigus ut a fluviis, calor a b sole, aestus a b undis aequoris exesor moerorum litora circum. nee variae cessant voces volitare per auras, denique in os salsi venit umor saepe saporis, cum mare versamur propter, dilutaque contra cum tuimur misceri absinthia, tangit amarum.
24
hoc ubi levia sunt manantis corpora suci, suaviter attingunt et suaviter omnia tractant umida linguai circum sudantia templa. at contra pungunt sensum lacerantque coorta, quanto quaeque magis sunt asperitate repleta. deinde voluptas est e suco fine palati.
Democritus and Epicurus on sensible qualities
85
But not everyone or everything derives pleasure from the same substances, and that needs explanation. First, there are 'atoms mingled together in things in many ways' (semina multimodis in rebus mixta teneri, 644). This is confirmed by other passages of Lucretius (i.8i4ff, 895-6; 11.333-80). Second, there are differences in the atomic configurations of the sense organs (differre necessest I intervalla viasque, foramina quae perhibemus, I omnibus in membris et in ore ipsoque palato, 649-51). When something sweet to some is bitter to others, it must be because its smoothest particles palpably penetrate the palate of the former, whereas those to whom the same thing is bitter inside them are evidently invaded in the gullet by particles that are rough and jagged, (iv.658-62) 25
When someone has a fever, his body is upset, and the position of his component atoms are changed, and so different particles are now filtered along the channels in his body. What used to taste sweet now tastes bitter. The exact words used by Lucretius are interesting: It happens that the bodies that previously fitted his sense now do not fit, and others are better suited, which can penetrate and generate a bitter sense; for both kinds [of bodies] are mingled in the flavour of honey, as we have often made clear to you above, (iv.668-72)26 He has not actually said before that honey has both smooth and rough atoms, but only that things in general have a mixture of atoms in them. What is interesting is that the atoms generate [progignere) the taste in the sense organ. In itself, it would seem, the honey is neither sweet nor bitter, nor both. These passages about taste might well suggest that Epicureans agreed with Democritus about the ontological status of tastes. They suggest that, strictly speaking, honey is not sweet (in spite of the famous simile in which he compares his own poetry to honey on the lip of the cup of bitter medicine, 1.936-41). Sweetness is an adverbial property of the experience of tasting something, and it depends both 25
hoc ubi q u o d suave est aliis aliis fit a m a r u m , illi, cui suave est, levissima corpora debent
contractabiliter caulas intrare palati, at contra quibus est eadem res intus acerba, aspera nimirum penetrant hamataque fauces.
26
fit prius a d sensum (ut) q u a e corpora conveniebant n u n c non conveniant, et cetera sint magis apta, q u a e penetrata q u e u n t sensum progignere acerbum. u t r a q u e enim sunt in mellis commixta sapore; id quod i a m supera tibi saepe ostendimus ante.
86
DAVID FURLEY
on the shape of the atoms composing the food or drink and on the state of the sense organs. Such was the conclusion of Giussani and Bailey in their notes on the passages.27 But in view of the earlier passage quoted above, where tastes are treated in exactly the same way as other sensible qualities, it seems more likely that Lucretius means that there is both sweetness (consisting of smooth particles) and bitterness (rough particles) in the honey. If honey were explained in a systematically different way from the other senses, one might expect some emphasis on that fact, especially in the programmatic lines iv. 218-24, quoted above. Lucretius goes on to talk about smells, and seems to be more confident about their ontology than he is about tastes. Taste is a matter of contact on particular occasions, and it cannot easily be given a durable identity, except as experienced. But smells are explained by Lucretius in much the same way as vision, with the difference that the emanation in the case of smell is said to come from deep inside the object, rather than from its surface (DRJV iv.694 ex alto ... ex re). Without explicitly making the point, Lucretius appears to make smells, like colours, properties of the object. The impact of a smell touches the nostrils (naris adiectus odoris I tangat, iv.693) • A 'flow of
smells' (fluctus odorum) come from objects. Bees are drawn by the smell of honey. A smell is said to be 'made of larger particles than sound' (iv.699), because it can't penetrate walls as sound can. On the other hand, whether a smell is attractive or unpleasant is explained as in the case of taste: it is a matter of the 'fit' between the particles constituting the smell and the structure of the sense organ (iv.677). Epicurus said much the same as Lucretius, in fewer words (Ep.Hdt. 53). 2 8
Plutarch presses on with his criticism in the next chapter (5). There is a well-known Epicurean theory, he says, of summetriai and harmoniai of the ducts associated with the sense organs, and there are polumixiai 27
28
I n the presentation I m a d e to the Symposium, I followed the line taken by Giussani (1896-8) a n d Bailey (1947). I have been persuaded b y David Sedley's comments o n that occasion to look at the whole m a t t e r again, a n d I have concluded t h a t Giussani and Bailey are probably wrong, a n d Lucretius does not intend to make a difference between taste a n d the other senses. Lucretius concludes this section with a thoroughly confusing passage (706-21). I t m a y possibly be out of place: L a c h m a n n (1860—6) bracketed it, Giussani (1896—8) moved it u p twenty lines, Bailey (1947) calls it a digression. T h e passage picks u p the idea of tastes a n d smells being good for one species a n d b a d for another, a n d applies it to visible forms a n d
Democritus and Epicurus on sensible qualities
87
of the seeds dispersed in all the flavours and smells and colours; together, these account for the different appearance of things to different people. He challenges Colotes: 'Don't these theories compel objects in [your] view to be no more this than that?' He produces a more dangerous piece of evidence against the Epicurean theory in a later chapter. He first quotes from Epicurus' Symposium. Epicurus holds a dialogue with Polyaenus about the heating quality (to thermantikon) of wine (ch. 6, 1109.F ff), in which Epicurus concedes that there is a mixture oiphuseis in wine such that a certain amount of it may affect one person in one way and another in another. Some people, he continues, deceived by this, conclude that wine is universally cooling, others that it is universally heating. But that concession, Plutarch claims, should have made him grant that everything is no more this than that. Well, one who says that the majority have been deceived when they suppose that what is heating is heating and what is cooling is cooling has been deceived himself if he denies that 'everything is no more this than that' follows from what he has said. (11 IO.A-B) 29 Plutarch is wrong. It does not follow, because wine has different heating effects on some people from others, that wine does not have other properties that are intrinsic to it. But Plutarch finds better ammunition in chapter 7, in Epicurus' book Reply to Theophrastus, and it has to be quoted. Unfortunately it is not agreed among commentators where the quotation from Epicurus ends. This has to be discussed, so for convenience I print the agreed quotation with single quotes and the dubious part with double quotes. Now Epicurus himself, in the second book oi Reply to Theophrastus, says that colours are not inherent in the nature of bodies but are generated according to certain arrangements and position in relation to the sight; and with this statement he concedes that a body is no more uncoloured than coloured. And he has already written earlier on (and I quote): 'But also, quite apart from this section, I don't know how one should say that these things have colour when they are in the dark.'
29
colours. He produces the weird example of lions that are said to be turned offby the sight of roosters, because there are certain seeds in roosters that slash the pupils of lions when they penetrate them (Bailey quotes someone who tried to confirm this by experiment; but the lion ate the chicken). 6 Sf] Aeycov I^TiTrorrnaOai TOUS TTOAAOUS TO OepuaTvov OepiJiavTiKov f^ TO vyOxov yuKTixov UTToAau|3avovTas, si |ifi voui£oi TO *jjir| IJCXAAOV eivai TOTOV f\ TOIOV iKccorov' &KoAou6eTv ois 6ipr|K6v,
88
DAVID FURLEY
"Yet often, when the surrounding air is of the same degree of darkness, some perceive a distinction of colour and others don't, because of weakness of sight; moreover, on entering a dark house we see no sight of colour, but after a short interval we do." So each of these bodies will be said not to have colour any more than it has. But if colour is relative, white will be relative, and dark-blue, and if these, then also sweet and bitter, so that in truth 'it no more is than it isn't' will be predicated in respect of every quality. For to those so affected it will be such, to those not so affected, not. (i i IO.C-D) 30 Plutarch concludes that by attacking Democritus for the thesis 'no more this than that' Colotes has shot himself in his own Epicurean foot. Westman31 argued that the sentence in double quotes belongs to Epicurus, and he is followed by Einarson and De Lacy. Westman's argument is that without the text within double quotes Plutarch has no basis for concluding that Epicurus is committed to 'no more this than that', and he could not understand why Plutarch should introduce the example of entering a dark room. We have confirmation that the Epicurean theory included the thesis that objects have no colour in the dark from Philodemus On Methods of Inference 25 and Aetius 1.15.9, but most of all from Lucretius DRN11.795-809. Lucretius' aim is to show that atoms have no colour; he does not directly discuss whether or not compound objects have colour, which is Plutarch's concern. But in passing I think he makes it clear what Plutarch's argument must be. Lucretius first outlines the argument: colours cannot exist without light; but atoms individually never emerge into the light; so there cannot be colours in atoms. Then he returns to the first premiss and produces arguments for it. Things change colour according to the 30 OOTOS y a p oOv 6 'EiTiKoupos ev TCO 8euT6pcp TCOV Trpos 0e6<|>pacTTov OUK eivai Aeycov TCC Xpcbiiorra auiJiun TOIS acbpacnv, aAAa y6vvdcj8ai KCCTCC TTOIOCS TIVOCS TCC^EIS xai GECTEIS Trpos TT)V ovyiv, o u paXAov 4>r|ai Korrd TOUTOV TOV Aoyov a x p c o i i d n a T o v acona eivai f\ x p & p a £Xov. avcoTEpco 8E KOTOC M ^ I V TCCC/TOC yEypa<|>£v 'dXXd Koct x ^ p f c TOUTOU TOO pepous OOK O ! 6 ' OTTGOS 8et TOC IV TCO CTKOTSI TOUT' OVTOC <j>fi<7ai Xpcbporra EXEIV.' "KaiToi TTOXAOCKIS dtEpos ojioicos OKOTCO8OUS 7T£piKExv/|iEvou oi \xkv ai<j8avovTai XP 00 " H&TCOV 8ia<|>opas oi 6 ' OUK a i a 6 a v o v T a i 61' an|3ACnT)Ta TTJS oyEcos- ETI 5 ' EICTEAOOVTES ets CTKOTEIVOV OIKOV o08£|jiav o y i v xpwiiorros 6poo|JEv dvaiiEivavTES 6E piKpov 6pa>|jiEv." ou liaAAov ouv EXEIV f\ \xr\ syew xP&Ma j^TidfjaeTai TCOVCTCOMOTCOVEKacrrov. £i 8E TO x p " ^ a TTpOS T l , KOCt TO XEUKOV E Q T a i TTpOS T l KOI TO KUOCVOUV, fit 8 E TOCUTa, KOI TO y X U K U KOI TO
TTiKpov. COOTE KOTOC TraoTjS TTOI6TT|TOS 6:XT|6COS TO [xi] ^aX^ov Eivai f\ \xf\ ETVCCI Korrr|yopETCT6aiTOIS y a p OUTGO 7T6CCTXOUCTIV l o r a i TOIOUTOV, OUK i a T a i 8E TOTS (if) Tra 31
Westman (1955) p p . 141-3-
Democritus and Epicurus on sensible qualities
89
direction from which the light falls on them - for example, a dove's neck-feathers, or a peacock's tail. And since these colours come into being because of a certain blow of light, plainly we must hold that they cannot come to be without it. (11.808-09)32 The language suggests that the colour is created in the object by the 'blow of light'. It has already been explained that colour is generated from uncoloured atoms by the shape and arrangement of the atoms on the surface. The theory of light seems to be that it is itself an atomic compound which strikes compound bodies in such a way that a rearrangement takes place in their surface atoms. Nothing in the Lucretian passage shows that colour is generated as an interaction between the external object and the seeing eye. To return to Plutarch. What is said by Epicurus in the passage in single quotes is that bodies have no colour in the dark. The sentence in double quotes adds that in the same light conditions perception of colour differs. The object in a dark interior is seen as coloured by one person, uncoloured by another. It cannot be an effect of the light, since the light is the same. So the object itself, Plutarch concludes, must be no more uncoloured than it is coloured: the perception of colour must be in the mind or sense-organ of the beholder. This would be a severe blow to Epicurean theory, if that conclusion could be sustained. Colour is the proper object of vision. That is made clear in the very interesting papyrus fragment by an anonymous Epicurean, brought to our attention by Long and Sedley: We hold that vision perceives visibles and touch tangibles, that the one is of colour, the other of body, and that the one never interferes with the other's sphere of discrimination ... If, then, visible shape is nothing but the external positioning of the colours, and visible size is nothing but the positioning of the majority of the colours in relation to what lies outside, it is perhaps possible for that whose function it is to register colours themselves to perceive the external positioning of the colours. (PHerc 19/ 698 = Long & Sedley 16c, their translation)33 32 33
qui quoniam quodam gignuntur luminis ictu, scire licet, sine eo fieri non posse putandum est. TT)V \xk[v] y[6c]p [6]vfiv 6pccT& KOTa[A]a|i|3|a]v6iv fjyoOiaeSa, TTJV 8E <5c<j>f]v dorra, KCC[I] TT|V nsv xp^MocTolils, TT)V 51 (TCOMCXTOS, Koci [T]T)V Prepay TOO TT\S trkpas [xp]i|ja[Tos] |ir|8£v Tr[oA]uTrpayM[o]veTv- . . . [et]-TT6p oOv oO0ev §Tep6[v] icrri TO axfj^a TO 6pa[Tdv] f| T « V Xpco|j&Tco[v] ^l [£5GO]TOTCO 06CTIS, oG8e TO MJsJyeOos TO opaTov f| TW[V] TrAeiovcov XP°°lionxo[v] f] KOTOC TO §£[co] 06or[is, TCJJGOS SUVCCTOV TT)V CC[0T]CC>V xpcd\xcrTG3V OUCTOCV [&]VTI-
Ar|TmKf)v [TT)V] Ibarra™ Oiaiv TQ[V] XP 0 0 ^ 0 7 1 0 ^ KOTaXa|j[pd]v6iv.
90
DAVID FURLEY
If, then, Plutarch could show that colour is not a real property of the object of vision, but an interactive experience in the mind or sense-organ of the beholder, he would have gone a long way towards making his case, and saddling the Epicureans with the fatal proposition, ou mallon toion e toion. His case would then be the following: Epicurus concedes that objects have no visible colour in the dark; but some people see colours in bad light, when others do not; hence colours must be nothing but subjective experiences. But as a matter of fact, Epicurus would not need to concede this conclusion: weak vision can account for failure to see objectively existing colour, just as poor hearing can account for failure to hear real noises. Epicurus could defend himself against Plutarch's charge on these lines: in total darkness there is no colour; a little light generates some colour in the illuminated objects by its effect on their surfaces, but this is sometimes not perceived by those with weak vision, or perceived only gradually as their eyes adjust to the poor light. It appears, then, that Epicurean theory gave to the sensible properties corresponding to each of the senses the status of real properties of objects in the external world. And they claimed that all sense impressions are aletheis. But it is notorious that objects appear differently to different people or animals, even without special conditions such as the darkness postulated in the passage just considered. So the problem must be faced: do objects in the external world have contradictory properties simultaneously? Or, to phrase the question as Plutarch does, was Epicurus committed, as Democritus was, to conceding that the object is ou mallon toion e toion? Three methods of saving Epicurus have commonly been tried. The first is to read the word alethes as meaning 'real' rather than 'true'. The claim that all sense impressions are aletheis then has the force of denying that they can be imaginary: they are events that happen in fact. This reading gets its main support from passages like the following Epicurean summary by Diogenes Laertius: Our seeing and hearing are facts, just as having a pain is ... The figments of madmen and dreaming are true. For they cause movement, whereas the non-existent does not move anything, (x.32, trans. Long & Sedley)34 34
0eaTT|K6 S E T O TE 6 p a v fjiaocs KOCI OKOVEIV w o r r E p T O d A y e l v . . . T & TE TGOV JJOUVOHEVGOV <j>avTdCTpocTa KCCI (TOC) KOTT' o v a p a A r | 8 f i , KIVE! y a p - T O 8 E \xr\ 6 v o u KIVET.
Democritus and Epicurus on sensible qualities
91
Several moderns have adopted this interpretation. 35 But this interpretation takes Epicurus out of the main stream of the epistemological debate of his period, and is inconsistent with the sense given to his theory by his supporters and opponents in antiquity. He was generally taken to be asserting that sense impressions report something true - i.e. some true propositions about the external world.36 There are still (at least) two lines of interpretation designed to rescue Epicurus from an apparently impossible position. One is to suppose that the aistheta are not (or not in all cases) the external objects themselves, but the eidola that flow from external objects to the sense organs and cause sense impressions.37 The main support for this is Sextus Empiricus [Math, vii.206-10). He explains, on behalf of Epicurus, that although there is a colour that belongs to the solid body itself, there is also a colour that is 'outside the solid body and adjacent to it'. 38 The large square tower in the distance appears small and round because the eidola have been changed by travelling through the intermediate space. And it is the peculiar function of sensation to grasp only what is present to it and moves it, such as colour, but not to distinguish that the object here is one thing and the object there another. Hence for this reason all sense impressions are true. [Math, vii.210)39 The outstanding difficulty in this interpretation seems to be that it makes it impossible that we should have any access to the sensible properties of external objects themselves without inference, except in the case of touch and taste. In these cases, there is no opportunity for eidola to be altered: they must preserve the features of the object itself. But in all cases of sight, hearing, and smell, we would have to make an inference from what we perceive - the eidola - to the object 35 36 37
38
Myself w h e n y o u n g e r (Furley, 1971). Also R i s t (1972). See Striker (1977) a n d T a y l o r (1980) for a r g u m e n t s t o this effect. Since writing a n d revising this p a p e r I h a v e r e a d S t e p h e n Everson's defence of the thesis t h a t t h e objects of p e r c e p t i o n i n E p i c u r e a n t h e o r y a r e EiScoAoc (Everson 1990). Sometimes T a y l o r (1980) a n d L o n g & Sedley (1987) seem to take this view. I still think the thesis is refuted b y the objection set o u t below. TOO 8E xpwuocTos TO UEV EOTIV E*TT' OCOTOO TOO aTEpEuviou, KaOairgp e m TCOV aOveyyi/s KOCI EK [TOO] lAETpiOU 6iaOTf||JO:TOS (3AETTOUEVG0V, TO 8 ' EKTOS TOO CTTEpEUVlOV K&V TOIS i ^ E ^ S TOTTO1S UTTOKeifJEVOV.
39
aia6f|CT6cos 5E TSIOV uirnpxE ( T O ) TOO TrapovTos |i6vov Kcci KIVOOVTOS CCUTTJV &VTIAOCU|3&VECT8O:I, olov xpcbuorros, ouxi 8E TO SiaKpivEiv OTI aXXo UEV EOTI TO EVOOCSE OCAAO 5E TO £V©&5E C/TTOKEIUEVOV. SioiTEp a ! UEV <|>avTaaiai 6 i d TOCOTOC T r a a a i EICTIV <5CATI0ETS.
92
DAVID FURLEY
itself. And this is just what Epicurus meant to deny. 40 It cannot have been his own thesis that we perceive only eidola: so that thesis cannot have been the basis for his claim that all aistheta are true. Lucretius exclaims that touch is the sense of the body 'either when something from without finds its way in, or when a thing which is born in the body hurts us, or gives pleasure' (11.434-6). The theory of eidola makes all the senses like touch: through the eidola we have contact with external objects. As Epicurus himself says, We must also suppose that it is when something enters from external objects that we see and grasp their forms; for external objects would not impress their own nature, both of colour and of form, [in any other way] so well as they do from the entrance into us of certain 'casts' (tupoi) from the objects, of the same colour and form ... And the sense-impression we get when we focus the mind or the senses, whether of shape or of properties, that is the shape of the solid body. (Ep.Hdt. 49-50)41 The point to be emphasized is that it is the actual colour and form of the external object itself that is impressed upon us by means of the eidola.
The third mode of interpretation is to accept that Epicurus asserted both that sense impressions are true, and that they directly convey the truth about external objects themselves, but that the content of sense impressions needs to be specified very exactly. Error arises from leaving out qualifications. Thus when we see the familiar square tower in the distance, our sense impression is not correctly described as 'a small round tower' but perhaps as 'a distant tower with angles indeterminable at this distance', or something like that. 42 This, Epicurus might claim, is indeed a real property of the external object (i.e. the tower together with the intervening distance), and it is the direct content of the sense impression (distance is part of the content of a sense impression). So, presumably, the wine that tastes sweet to one person but dry to another should be said to have the properties 'sweet-tasting to one with such-and-such a 40
Sensation is ocAoyos (Diog. Laer. x.31).
41
AeT 8e KCU voui£siv
42
£TT61CTI6VTOS TIVOS cVrrd TCOV §C;CO8EV TOCS |iop<|>&s o p a v
f)|aas KCCI
6iavos!(70ai- ou y a p a v 6vcnToa<|>payi<jaiTO TOC IC;CO Tfiv ECCVTCOV <|>0(Tiv TOO TE xpcoiiccTos KCCI TT\S |iop<|>f\s . • • OUTCOS COS TOTTCOV TIVCOV £7T6iai6vTcov fjuiv &TT6 TCOV TrpayuaTcov 6noxpocov T6 Koci 6uoiou6pcov . . . Koci f^v a v Aa|3cou8v <J>avTacriav 6TTI(3ATTTIKCOS TTJ Siavoia f\ TOTS aia6r|Tnpiois SITE p i o p ^ s eft-e av/|jp6|3r|K6Tcov, uopf| ecmv OOTT) TOO orspeMviou. This m a y be w h a t Asmis (1984) means w h e n she writes (p. 114) ' T h u s the tower seen from afar really is small a n d roundish, even t h o u g h t h e tower seen from n e a r b y is large a n d s q u a r e ' . But it should n o t b e p u t like that. If the tower is square, it is n o t true t h a t it is roundish. W h a t the distant view truly reports is 'tower-at-a-distance'.
Democritus and Epicurus on sensible qualities
93
palate' and 'dry-tasting to one with a different palate'. These are then different properties, but not contradictory ones. This line of interpretation is admittedly strained to the limit by Epicurus' reported insistence that the figments of madmen and dreaming are true (Diog. Laer. x.32). His line of defence would have to be that there really are eidola causing the impressions, the eidola originate (somewhere, some time) in real external objects, and these external objects taken together with their context of hallucination or dream really have the properties they appear to have. The difference between Democritus and Epicurus is a matter of ontology. The objects of perception (ta aistheta) are not to be found among the realities of the external world according to Democritus; they are real properties both of eidola and of the parent solid objects according to Epicurus. There is nothing in the surviving Epicurean texts to tell us exactly why Epicurus made this change, but one may hazard some guesses. The reason why Democritus denied the objectivity of sensible properties other than shape, size and weight is to be found in his inheritance from Eleatic philosophy - particularly perhaps the argument of Melissus fr. 8. If we believe, Melissus argued, that there are real distinctions between hard and soft, black and white, it is because our senses tell us so. But the senses also tell us that what is white becomes black, and what is soft becomes hard. 'But if there is a change, then what is has perished and what is not has come to be' (fr. 8.6). Hence, on the Eleatic principle that what is not cannot enter into a rational discussion, change must be eliminated, and therefore also differences of quality. 43 Between Democritus and Epicurus comes the discussion of being and not-being in Plato's Sophist and Aristotle's distinction of substratum and quality in Physics 1. Epicurus, although he still wished to base his physics on unchanging elements, saw no reason to deny the reality of change at the level of compounds. If Democritus' ontological position is comparatively clear, his epistemology is not. If Theophrastus' evidence has any value, as I believe it has, he put forward a theory that linked atomic shapes and dispositions with particular perceptions: he did not, in other words, hesitate to make inferences from perceptions to imperceptible 43
See further Furley (1987), pp. 117 fif.
94
DAVID FURLEY
objects. Yet he notoriously made sceptical comments about the worthlessness of human perceptual knowledge. We might link Democritus with Parmenides in this respect, and ask of both of them how it is that they claim to distinguish acceptable from non-acceptable statements about perceptual experience, while at the same time holding that such statements are 'deceitful5 (Parmenides fr. 8.52) or 'bastard' (Democritus fr. 11).
PART THREE
The Passions
CHAPTER 5
Poetry and the passions: two Stoic views Martha C. Nussbaum
The spectator of the dramatic theatre says: 'Yes. I have felt the same. I am just like this. This is only natural. It will always be like this . . . I am weeping with those who weep on the stage, laughing with those who laugh.' The spectator of the epic theatre says: 'I should never have thought so. That is not the way to do it. This is most surprising, hardly credible. This will have to stop . . . I am laughing about those who weep on the stage, weeping about those who laugh.' Bertolt Brecht Stop wanting your husband, and there is not one of the things you want that will fail to happen. Stop wanting to remain in Corinth. And in general stop wanting anything else but what the god wants. And who will prevent you? Who will compel you? No one, any more than anyone prevents or compels Zeus. Epictetus, addressing Medea There is surely no principle of fictitious composition so true as this, - that an author's paramount charge is the cure of souls. Henry James, 'Miss Prescott's AzariarC (1865) Listening to poetry, wrote Plutarch, is like eating fish-heads: absolutely delicious, but it can give you bad dreams [How the young person should listen to poetry i5bc). Believe this, as all the major Stoic thinkers do, and what follows? A lover offish-heads would prefer, clearly, to discover a way to go on eating them in good health, without suffering the disturbing consequences. And this is, on the whole, the Stoic response where poetry is concerned: not sweeping censorship or hostile denunciation, but a stern yet affectionate programme of reform, aimed at preserving and enhancing the health of the soul. But to design such a programme, one clearly needs to know a I am grateful to Jacques Brunschwig, Miriam Griffin, Stephen Halliwell, Gisela Striker, and Richard Sorabji for comments that have helped me in my revision of this piece. I am conscious that I have by no means answered all the questions they have raised.
97
98
MARTHA G. NUSSBAUM
great deal about the systems that are affected. To solve the problem of fish-heads, a doctor will need to know a lot about the digestive system, and how fish-heads interact with it. Two scientists with radically different accounts of digestion will be likely to propose correspondingly different accounts of proper fish-head eating. So too for poetry. Before the doctor of the soul can productively ask how education can retain poetry's delights while avoiding its dangers, he or she must first have well-articulated views about the relevant questions in psychology. Above all, she must have an account of the passions, and of the operation of poetry on the passions. For it is generally agreed, among Hellenistic thinkers of several schools who disagree on much else, that poetry makes its impact on the soul above all by altering its passions. And it is this aspect of poetry's causal role that is the basis of the most serious attacks upon it. As in the digestive case, two thinkers who have radically different views about the nature of the passions and of poetry's causal interaction with them are likely to differ, as well, in their accounts of correct poetic education. Both Stoics and Epicureans grapple with this problem, connecting their discourse about poetry closely with their analysis of the psychology of desire and passion. All major thinkers of both schools seem to agree that poetry has a powerful and in some respects dangerous effect on the passions of the soul; all ask what these effects are, and whether a reform of poetry can retain and cultivate whatever benefits poetry offers, while avoiding its dangers. I have elsewhere discussed some aspects of the Epicurean treatment of these issues, focussing on Lucretius.1 Here I want to investigate some complexities of the Stoic positions.2 THE PARADOX OF STOIC POETRY
As soon as one embarks on the study of Stoic views of poetry, one encounters a paradox. For, on the one hand, the Stoics clearly took a very extreme position concerning the passions, holding that they should be not just moderated, but completely extirpated from 1 2
Nussbaum (1989). Anyone who works on this topic owes a great debt to P. De Lacy's fine article, 'Stoic views of poetry' - De Lacy (1958a). It seems to me a major shortcoming of De Lacy's discussion that he did not clearly separate the different Stoic views of the passions, as I try to do here. But it is a fine, path-breaking study.
Poetry and the passions: two Stoic views
99
human life.3 No other school showed the passions such singleminded and obsessive hostility. On the other hand, no other ancient school is more sympathetic to the poets, those notorious feeders of passion. We see this enthusiasm in many ways. First, we notice the striking number of Stoic treatises devoted to the topic of poetry. Zeno wrote a Peri poietikes akroaseos {On listening to poetry) and may also have
discussed poetry in his Peri lexeon {On diction) (Diog. Laer. vii.4). Cleanthes wrote a Peri ton poietou {On the poet) (vn.173) and, of course, poetry - on which we shall comment later. Timon, noting his slowness in scientific reasoning, mocks him as a 'slow-witted lover of verses' (vn.170). Chrysippus wrote a Peripoiematon {Onpoetry) in one book, a Peri tou pos dei poiematon akouein {On how one should listen to
poetry) in two books (vii.200), and several other works that may have dealt with poetic matters. Diogenes of Babylon made important contributions to the debate in Peri phones {On voice) and Peri mousikes (see below). Posidonius gave the topic a new direction, as we shall see; and poetry is discussed often in Epictetus and Seneca, as well as in the Stoic geographer Strabo. But treatises may be condemnations; surely the number of Stoic treatises on the passions is no index of support. More significant, then, is the fact that the major Stoic thinkers were especially fond of citing poetry at moments of importance, apparently with approval. Zeno allegedly died with lines from the JViobe of Timotheus on his lips (vii.28). Cleanthes is depicted as quoting casually from Homer and Euripides (vn.172). Whether true or not, such stories indicate the prevailing habits and sentiments of the school. And about Chrysippus we know a great deal more. We know that he took the practice of poetic quoting to an extreme, drawing frequently and copiously on the words of his favourite authors: above all Homer, Euripides, and Menander, but also Hesiod, Stesichorus, Empedocles, Tyrtaeus, Orpheus, and other dramatic writers - a thaumaste aperantologia, comments Galen. We have, through Galen, many examples of such quotation, and indications of many more. Diogenes Laertius reports that Chrysippus was said to have copied nearly the entirety of Euripides' Medea in one of his treatises (vii.180); and while that report may be hostile, the interest in the play that is evident in the fragments and reports of the Peri pathon, 3
See Nussbaum (1987) for an account of these arguments.
IOO
MARTHA G. NUSSBAUM
especially in Galen's de Placitis Hippocratis et Platonis (hereafter PHP) make the story seem plausible enough. 4 Another report states that if you remove the poetic quotations from Chrysippus' writings, the pages would be bare ( V I I . I 8 I ) . While this is clearly malicious, Galen's numerous complaints along the same lines make one believe that the bulk and variety of poetic quotation in Chrysippus' work must have been striking. And the use of the poets is plainly above all positive: he turns to the poets as genuine sources of insight. Finally, we find explicitly favourable Stoic discussion of the effects of poetry, and explicit defence of its role in the education of the young - in authors from Cleanthes through Posidonius to Seneca and Epictetus. These establish conclusively that the Stoics did not write about and cite the poets only to repudiate them. Since these arguments will be a central theme of this paper, I shall say no more about them now. The Stoics were aware of Plato's arguments, in the Republic, in favour of the censorship of most existing poetry. They share, it would seem, many of Plato's motivations for censorship. For while the Republic eliminates fear, grief, and pity from the lives of the guardians, but leaves in place some military anger, the Stoics wish to reject passion in an even more sweeping way. Why, then, do they not reject the artists who 'feed fat the emotion of pity' {Rep. 6o6b) and the other dangerous emotions? And how, given that they retain the bulk of conventional poetry, do they propose to prevent it from causing bad dreams? I shall argue that there are two very different Stoic answers to this question, related to two different Stoic views of the passions. One view, developed above all by Chrysippus, but probably also by Zeno in some form, and continued by Seneca and Epictetus, I shall call the cognitive view. It holds that the passions are judgments, assents to appearances; they are therefore modifications of the rational faculty of the soul. They are educated by an education of this faculty. The other view I shall call the non-cognitive view. It is found in its clearest form in Posidonius, who defends (what he takes to be) Plato's tripartite account of the soul; but I shall argue that it is also found in a preliminary form (under cover of an allegiance to the cognitive view) in Diogenes of Babylon. This view holds that the passions are 4
See the excellent account of Chrysippus' interpretation of the Medea in Gill (1983).
Poetry and the passions: two Stoic views
ioi
movements of a separate irrational part of the soul. This part cannot be modified by a modification ofjudgments; it must be 'harmonized' and balanced through non-rational means. Obviously these two views are likely to yield very different accounts of the impact of poetry on the soul's passions, and consequently different accounts of the correct poetic education of the young. I shall argue that they do, and try to describe the differences. I shall begin, briefly, with Plato; for I believe that it is significant that not just one but (pace Posidonius) both of these views could find their archetype in Plato's Republic, through certain ambiguities and indeterminacies in the Republic's account of the passions. I shall then briefly mention an Aristotelian contribution to the debate that adds a factor of central importance. Then, turning first to the chronologically later of the Stoic views, the non-cognitive view, as the more direct descendant of Plato's primary line of argument, I shall examine Posidonius' account of the passions and their education through poetry/music. I shall argue that the essential features of this account are already present in Diogenes of Babylon, although he to some extent presents himself as an orthodox follower of Chrysippus. Then I shall turn to the very difficult task of reconstructing the view of poetry that was connected to the cognitive view of passion, piecing together the evidence about Cleanthes and Chrysippus, and relying, also, on Epictetus and Seneca - but turning also to some less mainstream sources - above all Plutarch's Pos dei ton neon poiematon akouein {How the young person should listen to poetry), which, used with
proper caution, contributes valuable information about Chrysippan views. I shall then ask why partisans of the cognitive view, according to which all passions are false judgments, still wish to retain poetry as a valuable part of education. And I shall above all ask how these Stoics solve the problem of the fish-heads - arguing that their devices for improving poetic digestion are multiple and ingenious, but perhaps not altogether foolproof.5
5
Among the many works that offer potential insight into these questions, the reader will notice that I have not included discussion of Philodemus' On the Poets, which does discuss some Stoic positions. Since the work is occupied with questions about what makes a poem formally good, and since in many cases these questions are kept separate from questions about poetry's effect on character and on the passions, it is not directly pertinent to this discussion. And in fact the Stoic material it contains is far less interesting in any case, I think, than the material from Diogenes of Babylon discussed by Philodemus in the Peri mousikes.
IO2
MARTHA G. NUSSBAUM NOTE ON TERMINOLOGY
In the debate I shall describe, there is a difference of opinion not only about what poetry does to the passions and how, but also about which poems are of greatest interest and value in the educational process. Differences about the structure of the passions lead to differences about the causal interaction between poetry and the passions; and these in turn lead to differences about what the most educationally useful material is. All partisans to the debate deal with texts in verse that have - at least in part - musical accompaniment. All, then, are speaking about mousike in the old sense of the word, in which it included poetry as well as instrumental and vocal musical performance, and in which text and accompaniment were usually regarded as intimately linked, complementary parts of a single whole. But one position (the non-cognitive position) tends to focus on the musical element, treating the text as a part of the complex auditory experience in which the most important elements are melos, rhuthmos, and harmonia. The cognitive position tends, by contrast, to emphasize the text — and above all its narrative and dramatic structure - though not ignoring what the other elements contribute to the presentation and communication of that structure. The noncognitive group tend to use the word mousike, the cognitive group the word poietike - though this is by no means always the case. One might therefore wonder: am I comparing apples with oranges, A's position about music with B's position about poetry? I think that this is not the case, that the two positions are really rivals. And I shall go on defending this judgment as I proceed. But let me begin with some clarifying observations. First, throughout I am beginning from a direct conflict between the two groups: the conflict about the structure of the passions. Both groups agree that the resolution of this issue has major consequences for one's views about education, and the place of poetry in that education. They portray themselves, then, as rivals about a single goal: namely, the proper education of the young, where the passions are concerned. And what I am attempting to do is to map out that debate, and to look at its implications for the role of poetry in education. If one group focusses on rhythm, the other on judgment, this is not because they are pursuing different inquiries: it is because they have come to different conclusions about the soul. Thus beginning, as the Stoic thinkers themselves begin, from the passions, helps
Poetry and the passions: two Stoic views
103
us to see an underlying unity to what might at first glance look like different inquiries. But there are some other points of clarification that should be made here, in order that we should understand clearly just how close mousike and poietike are, in any case, throughout the period with which we are dealing. First of all, then, the Stoics all considered poetry to be an auditory event. Although reading is mentioned once in the evidence I have scrutinized (and this by Plutarch, not by a Stoic), listening is always uppermost. (And of course reading itself, in this period, is above all an auditory experience in which metrical elements are always present.) Zeno's treatise on poetry and Chrysippus' work on the audience speak in the titles of listening and hearing; Plutarch's title follows Chrysippus'. The study of poetry was included by the Stoics under the branch of dialectic concerned with voice, phone (Diog. Laer. vii.44); and the Stoic definition of poiema ranges it under lexis - itself a species ofphone- describing it as 'lexis in metre or rhythm, going outside of prose in its structure' (VII.6O: the definition is ascribed to Posidonius). Poiesis is defined as 'significant poiema, including representation of gods and human beings' (60). Thus poetry is already pulled towards music by the emphasis on its auditory character; and all the poetry considered by the writers on both sides of the debate, including epic and drama, would have had prominent musical elements. On the other side, music is still considered by the Stoic thinkers we shall consider to be inseparable from a text. Music, too, is for the Stoics a branch of the study of phone, and thus included under dialectic (vn.44). And Diogenes of Babylon gave a widely received definition ofphone that makes it apparent that human, as opposed to animal, phone is all 'articulate and sent forth in a deliberate way' (D.L. VII. 55). If we combine this view with the view that mousike is a part of the study of phone, it seems to follow that human mousike is itself articulate: not just textless vocalizing, but the singing of a text. (And notice that bare instrumental music is apparently neglected altogether in this Stoic classification, since that cannot possibly be classed as phone, except insofar as someone sings along with it.) So mousike is drawn towards poietike. And even though Diogenes of Babylon focuses, as we shall see, on the elements of lyric performance that we might tend to regard as purely musical, it is clear that he did not so regard them. For (in a passage that we shall study later) Philodemus represents Diogenes' supporters as charging
104
MARTHA G. NUSSBAUM
Philodemus with being agroikos, because he uses the term 'mousike' too narrowly, excluding the poetic text. It seems to these opponents to be a reductio ad absurdum of Philodemus5 position that Pindar and Simonides will not count, for him, as mousikoi; so clearly, though these poets would have been famous as the authors of texts, they are central cases of mousike on the non-cognitive view. And in general Diogenes never mentions instrumental music without text: he focusses on lyric poetry, hymns, laments, and other vocal/textual performances. As we shall soon see, there are good and evident reasons why partisans of the two different views of the pathe should focus on different parts of the complex event that is the performance of a poetic text with musical accompaniment. But I believe that we are entitled to regard these differences as just that, differences of selection and emphasis, stemming from the different answers these people give to the question, 'How does poetry move and influence the passions of the soul?' A N T E C E D E N T S : PLATO AND ARISTOTLE
The seminal treatment of poetry's influence on emotion, for all Stoic thinkers, is in Plato's Republic. (The Laws has some importance too, as we shall see.) Plato's texts, both complex and susceptible of multiple interpretations, serve as a starting point, officially or unofficially, for both of the competing views. Defenders of the noncognitive view, such as Posidonius and Galen, explicitly appeal to Plato's tripartite view of the soul. Defenders of the cognitive view, while rejecting the tripartite soul, still appear to follow certain Platonic lines of argument - especially the arguments of Republic n—in, in which poetry is criticized for its influence on beliefs about the self-sufficiency of the good person, and, through these, on the emotions of fear and grief. It would be an important part of a full account of this history to trace these connections in detail; limits of space prevent this. But a very cursory summary of the main points is necessary in order to set the stage for what follows.6 Posidonius will hold that emotions or passions7 such as fear, grief, pity, and anger are neither judgments nor dependent on judgment. 6 7
Many of these points are dealt with at greater length in Nussbaum (1992). I am using these words with no significant distinction - see my comments in Nussbaum (1987) footnote 2.
Poetry and the passions: two Stoic views
105
The soul has three parts: the reasoning part, the epithumetic part (concerned apparently with hunger, thirst, and sexual appetite) and the emotional or thumotic part. Like the epithumetic part, the thumotic part appears in animals and young children, as well as in human adults; like the epithumetic, it can be trained and harmonized, but not through cognitive changes. This view, explicitly modelled on Platonic sources, can legitimately claim to derive from at least some of Plato's statements about emotion and poetry's influence on it. The general account of the soul's tripartite structure is clearly based upon Republic iv (though one might well feel that Posidonius makes the thumotic part more independent of belief and reasoning than Plato did there). It can also claim support from book iv's account of the education of thumos, in which it is to be 'calm(ed) by means of harmony and rhythm' (441c), and from the assault on the poets in book x, where they are criticized for 'feeding' the soul's 'hunger' for grieving, strengthening that irrational appetite and thus making it more difficult to control in actual life (6o6a-d). 8 Here again, passions are treated, apparently, as irrational urges to be developed or undermined by non-cognitive strategies. Above all, however, Posidonius can (and does) claim support from the Laws, where the young human is said to be full of natural non-rational movements certain cries and jumps (672cd) - that must be ordered by rhythms and melodies; these are taken to operate apart from reasoning and judgment, imposing a structure on something fundamentally noncognitive. Laws VII follows the same line, focussing on the causal properties of melody and rhythm, and giving an account of the ways in which pregnant mothers can educate the emotions of their foetuses, through a kind of non-cognitive habituation. Chrysippus, on the other hand, holds that episodes of passion are identical with evaluative judgments - that is to say, assents to certain (propositional) appearances; and not just any judgments, but judgments with a particular content. All involve the ascription of a very high value to 'external goods', that is to say, items in the world that the agent does not control. 9 The basic idea seems to be that if we understand the reasoning part as both dynamic and involved in evaluation (not just calculation), there is every reason to 8
9
Another source of difficulty for the interpreter of Plato's position is the fact that Republic x appears to recognize only two parts of the soul, and appears to classify emotions such as grief and pity along with bodily appetites that demand 'replenishment'. See the longer account in Nussbaum (1987).
IO6
MARTHA NUSSBAUM
think that our loves and fears are activities of that 'part', and consist in the assent of that part to the appearance that some external item (that is at hand, or lost, or whatever) has (or had) enormous worth. Passions are not animal urges or stirrings, non-rational motions, but choices about how to view the world. They can, therefore, be modified and educated only by an education of reason. Thus, when Achilles mourns for Patroclus, that grief is an acknowledgement of the importance he accords to the person he loves. It is an acceptance of the judgment, 'A person who is extremely important to me has died'. If such an attitude is to be modified, then one will have to change Achilles' view about what has importance. And this, notoriously and extensively, the major extant Stoic accounts of passion's therapy - in the fragments of Chrysippus, in Cicero and Seneca and Epictetus - all undertake to do. This view can also claim to derive from Platonic sources. First of all, it can claim to be offering an alternative and perhaps a superior reading of the passionate part of the soul, as Plato presents it in Republic iv. For Plato insists on that part's responsiveness to belief and judgment, calling it an 'ally of the reasoning part' (441a) and a 'partner of judgment' (440b); and there are texts that also suggest that this part has its own internal beliefs and judgments. Chrysippus, of course, rejects tripartition. But he could plausibly point out that Plato's own characterization of the cognitive resources of passion makes the reasons for its division from reason less than compelling; and he could claim to be following what is deepest in Plato's account when he accepts the characterization of passion and refuses the division. More important for our purposes, he can clearly claim to be following Plato's account of the reform of poetic education in Republic 11—in. For there Plato reforms the emotions of fear and grief by reforming the beliefs (doxai) taught by poetic texts. Socrates focuses his attention above all on the portraits of the gods and heroes in epic and dramatic poetry - figures who are assumed to be the objects of a certain sort of admiring identification on the part of the audience. He then points out that these figures are repeatedly shown as attaching importance to events that actually have no importance, and thus as finding occasions for the emotions of fear and grief where a really good and self-sufficient person would not find them. For example, the grief of Achilles for Patroclus must be removed: for it shows a hero thinking it terrible to be deprived of a loved one,
Poetry and the passions: two Stoic views
107
whereas the really good and self-sufficient person would not think this (387~388a). In short, poetry contains false doxai {STjbj); and these doxai are then 'taken into the soul' of the young hearer, and give the developing soul a certain shape (377ab). Thus in books 11 and ill's assessment of poetry the focus is persistently on the evaluations poetry contains and promotes; and this cognitive content is understood to be closely linked to the poem's narrative structure, which sets up relations of identification and emulation (these, too, based on judgments, e.g. that so-and-so is a good person), which are then the vehicle through which the characters' passions enter the listener's soul. And Socrates also suggests — significantly for the development of the Chrysippan account — that a change in the relation of spectator to characters, a change that disrupts these admiring doxai, will often be sufficient to disrupt the transmission of the passions. If we put Achilles' speech into the mouth of a character who is inferior, or female (387eg-388ai), or in some way the object of mockery, we can prevent the malleable soul from being formed in accordance with the views expressed. It is not the purpose of this paper to ask which Stoic position can claim to have read Plato more correctly, or indeed, even to ask whether there is a single consistent position on the emotions and their education in Plato, both in the Republic itself and in Plato's work as a whole. What is evident is that Plato's complex arguments provided more than one starting point for Stoic investigations. Aristotle's actual influence on this complex history is unclear although the wide influence of the Poetics makes a causal relation to Stoic positions more likely than it is in some other areas. But in any case, Aristotle adds to the debate several ingredients that any good account of these matters needs to incorporate; so mentioning them will guide us in asking questions about the Stoic accounts. First, he introduces a far more explicit account of the emotions aroused by poetry, especially fear and pity, articulating the beliefs on which they rest and showing their common basis in the idea that events outside the agent's own control have importance for the agent's pursuit of the good. Second, he develops more explicitly than Plato an account of the process of identification with the tragic hero through which tragedy is able to show the audience 'things such as might happen' (Poetics ch. 9) in a human life. Thus he shows how
IO8
MARTHA G. NUSSBAUM
one's emotional response to a drama can be connected with thought about one's own life.10 But Aristotle's most original contribution to this debate is his recognition that the very literary form of a tragic drama embodies a commitment to the beliefs that ground the emotions. The form most characteristic of tragedy involves a significant reversal in fortune that is taken to be important enough to be the occasion for pity and fear. Without that plot structure, there is no tragedy - or at least no good tragedy. But that plot structure is not ethically innocent. By depicting the gifts or damages of fortune as if they have real importance for the lives of people who are 'good' and even 'better than us' - people with whom the spectator is encouraged to identify - they present already a certain view about the world and the importance of external happenings in it - a view that Plato and the Stoics will reject, and that Aristotle will (with some qualifications)11 accept. So Aristotle shows that the reform of poetry (if one is a Platonist or Stoic) cannot involve anything so simple as changing this or that line, this or that passage: it must involve the entire formal structure of the work, the shape of the dramatic action. One cannot purify the content without reforming the form. Plato did anticipate this point when he suggested that we might reform tragedy by giving the offending sentiments to a character the audience is encouraged to disdain. But he never reflected on the alterations this would bring about in the literary form of drama, in its choice of plots above all. Aristotle's contribution seems to me to be essentially new, and extremely important. We shall have to ask, later, to what extent Stoic proposals for literary reform really take account of it. 10
11
This material is discussed much more fully in Nussbaum (1986a) Interlude 2 and in Nussbaum (1992). See also the admirable discussion in Halliwell (1986), especially persuasive on the issue of identification; the discussion is continued in Halliwell (1992), with an extensive account of the concept of tragic recognition. The qualifications I have in mind have to do above all with the question whether misfortune can affect character. In Nussbaum (1986a) ch. 11 I argue that Aristotle admits damage to eudaimonia through impediments to the expression of character in action, but that, at least in Nicomachean Ethics 1, he is convinced that character itself remains firm. Material in the books on friendship and in the Rhetoric may, however, suggest a more complicated position.
Poetry and the passions: two Stoic views
109
THE NON-COGNITIVE VIEW! POSIDONIUS ON THE RHYTHMS OF THE SOUL
As we approach the Stoic views of poetry's effect, it will be useful to bear in mind a schematic contrast between two views. Non-cognitive view
tripartite soul emotions are non-rational movements emotions in humans and (most) animals poetry educates by imposing structure on the non-rational movements emphasis on rhythm, harmony, melody reform by appropriate selection of rhythm, harmony, etc. greatest interest in lyric poetry with musical accompaniment; hymns, dirges, etc. favourite famous poets Pindar, Simonides, etc.
Cognitive view
one-part soul emotions are evaluative judgments emotions in humans only poetry educates by forming or changing judgments emphasis on cognitive structure: narrative, identification, character reform by disruption of identification, textual rewriting, allegorical interpretation, etc. greatest interest in drama and epic favourite famous poets Homer, Euripides, Menander, Publilius Syrus
I begin with the non-cognitive view because, though later in origin, it is also more explicitly traced to Platonic sources. It is also relatively easy to see what the view is, and what scheme of poetic education is proposed in connection with it. Posidonius claims to follow Plato. In fact, he calls Plato 'divine', on account of his writing about the passions (Galen, PHP p. 286 De Lacy (1978/89) - hereafter D). He refers to Republic iv, the Timaeus, the Phaedrus, and the Laws, and was evidently fond of copying out passages from these works to support his case (see below). His view is that the soul has three parts: the logistikon, the thumoeides, and the epithumetikon. All the elements that older Stoics call pathe are located
110
MARTHA G. NUSSBAUM
outside of the logistikon, somewhere in the two non-rational parts. Obligingly explicit, at least here, he tells us that all four of the generic Stoic categories of passion - fear, distress, longing, and delight - are so located (PHP iv.5, 266 D); and his examples include not only the familiar Platonic case of anger, but also fear and grief (e.g. Galen, PHP, iv.5, 268 D). Concerning the nature and status of the passions, Galen informs us, Posidonius 'completely departed' (telos apechorisen), both from Chrysippus' view that they are identical with judgments and from Zeno's view that they supervene on and are necessarily produced by judgments: 12 Tor he does not regard the passions either as judgments or as supervening upon judgments, but as coming about through the thumoeidetic and epithumetic power, following in every respect the ancient account' (PHP iv.3, 248D). Later, Galen repeats the claim: 'He both praises and accepts Plato's view and disputes the view of the followers of Chrysippus, proving that the passions are neither judgments nor supervenient upon judgments, but certain movements of other non-rational powers, which Plato called epithumetikon and thumoeides' (PHP v.i, 292 D = Posidonius Edelstein & Kidd (1989) fr. 152). It is one of the frustrations of Galen's account that he never tells us how Posidonius divided these two non-rational elements, or to which he assigned the various pathe and on what grounds. The natural assumption would be that everything the Stoics call pathos fear, anger, grief, pity, etc. - is in the thumoeides, the bodily appetites in the epithumetikon. But the two Galen passages I have quoted do not support this, suggesting, instead, that the pathe were divided between the two non-rational parts. Perhaps this is just Galen's confusion; but it prevents us from confidently asserting the obvious view. In another significant passage (EK fr. 31 = PHP v.6, 332 D), Galen represents Posidonius as comparing passions to the two horses in the Phaedrus\ but again, this might just be careless. We can see, however, that the distinction between epithumetikon and thumoeides becomes relatively unimportant, on the account Posidonius has produced. For Plato in Republic iv, the importance of the separation was that the thumoeides had a responsiveness to judgment and reflection about the good that the appetitive part was explicitly said to lack. This had important educational implications, developed 12
Galen says that there is a difference of this sort between Chrysippus and Zeno; but we should be sceptical, since he also claims, implausibly, that Cleanthes supports Posidonius'
Poetry and the passions: two Stoic views
111
in books n—in. But in other texts (for example Republic x, Laws n and Laws m), Plato seems to take a less cognitive view of the passions and their education. Posidonius clearly follows that lead. Insisting on his independence even from Zeno (though he claims to find support for his view in Cleanthes — PHP, 332D) he denies passions are produced by judgments. They are simply 'non-rational movements', found, Posidonius tells us, in animals as well as humans, and in children from birth. And all of them are explicitly said to be unresponsive to judgments about the good. There is, Posidonius says, a natural oikeiosis towards pleasure through the epithumetikon, to victory through the thumoeides, to the kalon through the logistikon, and this alone (EK 160 = PHP v.5, 318 D). Although this view gives some vague basis for a distinction between thumos and epithumia (pretty vague indeed, since grief and pity, which have to be classified somehow, have no obvious connection with either pleasure or victory), more important to Posidonius clearly is the fact that it breaks the link between passion and judgment about the good. The non-rational parts become far more like one another than either is like reason. One other passage that distinguishes the two non-rational parts is equally vague: for it holds that some animals have epithumia without thumos, namely, 'all animals that are not easily moved and are attached like plants to rocks or something like that' (EK 33 = PHP v.6, 334 D); other animals 'all have both'. Once again, this is remarkably uninformative as to what the distinction actually is: for a friend of Chrysippus sees remarkably little difference between a sponge and a rat, where grief and love and anger and pity are concerned. And it certainly would not be safe to assume without further evidence that Posidonius could not have believed that sponges have grief and love. Still, once again, the primary message is clear. Grief and pity and love, even if not found in sponges, are certainly in rats and rabbits and worms and mosquitoes, and are to be seen as movements that go on in the absence of reasoning, without any natural orientation to reasoning and judgment. They are simply what Posidonius repeatedly calls them (Galen says it was a favourite term): pathetikai kineseis (PHP v.5, 322 D). Posidonius offers a related account of the origins of vice: these origins are innate, in the movements of the natural disposition, not learned from outside as a part of the creature's acquisition of judgments and beliefs. Here he believes that he has the twofold
112
MARTHA G. NUSSBAUM
advantage of following Plato and at the same time solving a problem that Chrysippus' account (as he sees it) was unable to solve (EK 35 = Galen, Quod animi Mores 819-20; cf. also PHP v.5, 318-20 D). His own explanation has both a general and a particular aspect. All human beings are born with non-rational motions that need balancing and will cause trouble if they do not get it. To that extent, all stand in need of appropriate education. But individual constitutions also vary: and here Posidonius turns both to physiognomy and to climatology for 'explanations'. Animals and humans that are warmer and more broad-chested are more prone to anger; those that have wide hips and are colder are more timid. Climate affects the pathetikai kineseis too, producing differences in emotional character. For this reason, some people can be easily balanced, while others must be 'blunted' with considerable difficulty ( ? / / ? v.5, 320-2 D). Posidonius insisted on the great benefits of his account of the passions for the educational theorist. He wrote: 'When the explanation of the passions was grasped, it removed the absurdity [apparently that of Chrysippus' view], and showed the origins of distortion in what is desired and avoided, and distinguished the methods of training' (EK 161 = PHPv.6, 328 D; cf. also EK 168). In particular, he insisted on the crucial importance for education of recognizing that the goal, where emotion is concerned, is not learning or judging, but the production of a balanced movement whose sharp edges have been 'blunted by good habits' and that is weak and gentle enough to receive the rule of the rational part. 'For knowledge does not arise in the non-rational powers of soul any more than it does in horses, but these get their appropriate excellence through a kind of non-rational habituation [ethismou tinos alogou], the charioteer through rational teaching' (EK 31 = PHP v.5, 324 D). Posidonian education must begin, he holds, with a programme for pregnancy; and Plato is praised for his thoughts about the movements of the foetus and the exercise of the pregnant mother (PHP 324 D). In the first book of his Peri pathon (On the passions), Posidonius made a collection of Plato's various statements about the child (presumably drawing above all on Laws vn), and commended Plato's prescriptions for the harmonization of the child's passions through movements. 'For this,' he writes, 'is the best education of children: the preparation of the passionate part of the soul, so it will be as amenable as possible to the rule of the rational part' (EK 31 = PHP 423 D). Following Laws vn, he gives mousikea, substantial role in this
Poetry and the passions: two Stoic views
113
early education. But, as in the Laws, the focus is on the ways in which tunes and melodies of various types will habituate these non-reasoning elements, changing people's mood and disposition in a non-cognitive way. Plato showed little interest, in those passages, in dramatic or narrative poetry, focussing on hymns and other songs and lyrics. Posidonius seems to follow this lead. In keeping with his physiognomic and climatological interests, Posidonius' regime is more highly individualized than Plato's, however. Galen summarizes: We should bid some people to spend their time in some rhythms and harmonies and exercises, others in others, as Plato taught us - raising the dull and heavy and spiritless in high-pitched rhythms and harmonies that forcefully move the soul and in exercises of the same kind, and those who are too spirited and who madly dart about in the opposite sort. Why was it, by the gods - I ask this also to the followers of Chrysippus - that when Damon the mousikos came upon a flute-girl playing in the Phrygian mode to some boys who were drunk and acting wild, he told her to play in the Dorian mode, and the boys immediately stopped their wild behaviour? For obviously they are not taught anything by the music of the flute that changes the opinions of the rational faculty. But, since the passionate part of the soul is non-rational, they are aroused and calmed by means of non-rational motions. For the non-rational is helped and harmed by non-rational things, the rational through knowledge and ignorance. (EK i68 = Pi/Pv.6, 330 D) (The passage is anomalous for its inclusion of apparently textless music as an example; but since the example is probably Galen's, not from Posidonius, it gives no indication of the range of Posidonius' own material.) Since a question has been asked, the friend of Chrysippus should answer it, in a contentious style that emulates Chrysippus' own (see below). 'You are talking here about bodily feelings, or objectless moods, not emotions. And perhaps the non-cognitive aspect of hymns and songs, their bare rhythm and melody, can influence moods and bodily feelings. Perhaps it could indeed calm the heat of drunkenness, which nobody, not even you, would take to be an emotion. But no real emotion will respond to such treatment. If I am angry because someone has murdered my child, whom I deeply love, you could play Mozart until the year 2000 without altering my state. If I am grieving because my lover has died, Beethoven's Seventh will do me no good. If I fear a nuclear holocaust, marches by John Philip Sousa have nothing to do with my condition. If we
114
MARTHA C. NUSSBAUM
want children to learn to grieve and fear and love appropriately [which, for Chrysippus, means not at all - but one can be a friend of his analysis and take issue with the normative beliefs that lead to that conclusion] what we must do is to get them to understand what matters and what does not, what is worth caring about and what is not. Rhythm and pitch without understanding - as Plato himself saw, at least in Republic ii-m - has nothing to do with the formation of the passions of the soul, a task in which poetic mousike, as we can agree, plays a central role.' Now we can see the non-cognitive proposal. It is obvious that I have little sympathy with it. But a defender of Posidonius might now charge that I am, after all, comparing two different things. For Posidonius, she might say (unconvinced by our remarks on pages 102-4) *s talking about music, not poetry. Perhaps he has another and more compelling view about the interaction between poetry itself and the passions, one that does not neglect its cognitive content. This defence will not do. It is not only that there is no sign of such an account. We can see that, given his view of the pathe, Posidonius cannot have such an account. The only way to approach emotion is non-cognitively, through a modification of the soul's non-rational movements. Posidonius leaves no doubt at all about this: he prides himself on this view and repeats it. This being so, it is perfectly natural that tunes and rhythms spring to the fore in his account of education, while the text (still around, apparently, in the hymns and songs he praises from Plato's Laws) is eclipsed, and narrative and drama disappear more or less entirely. That is not a change of subject matter; that is a change in selection of poetry that results directly and appropriately from Posidonius' position on the emotions. Poetic texts may of course give information of various kinds to the pupil's reason, in much the way a treatise can. But the only way they can have any impact on the passions - on pity or fear or grief or anger or love - is through the non-rational habituation of psychological motions, through a process in which a dog could also participate. Posidonius and Chrysippus share a goal: to put the passions of the soul in the condition that reason would approve. They differ about what the passions are, and thus about the causal processes that will produce the best state. Therefore they differ, as well, in their selections of causally beneficial poetry.
Poetry and the passions: two Stoic views
115
DIOGENES OF BABYLON! A PREDECESSOR OF THE NON-COGNITIVE VIEW
Posidonius' view of the passions and the parts of the soul has usually been regarded as new within the history of Stoicism, and as containing a degree of unorthodoxy that earlier (and many later) Stoics would be reluctant to allow. And, strictly speaking, this view is more or less correct. But it is interesting to observe that most of the salient points in Posidonius3 treatment of music/poetry are already present in a considerably earlier text: in the Peri mousikes of Diogenes of Babylon, Chrysippus' pupil and follower. Diogenes still mentions the Chrysippan view of the soul; and his account of poetic education adheres officially to the cognitive view. On the other hand, his real interests and emphasis lead very much in the Posidonian direction. So it may be that reflection about poetic education motivated the development of the new view of the soul. Rather than flowing simply from an independently supported philosophy of mind, thought about poetry seems to shape philosophy of mind. Posidonius says with emphasis that we really need to get clear about the causes of the passions before we can properly describe 'the methods of training'. Perhaps this is a gentle rebuke to Diogenes, who let his keen interest in paideia carry him past an essential stage in the argument. At any rate, I think we will see that Diogenes' statements about poetry are very close indeed to Posidonius; this makes one suspect that Posidonius knew and used them, furnishing them with the necessary psychological background. We do not have a great deal of information about Diogenes' character as a philosopher. (For the fragments, see SVF m, pp. 210-46; henceforth I shall refer to these by fragment number, not page.) He seems to have been generally orthodox in his views, but not afraid to propose new arguments and even new definitions. Galen reports that he offered new argument in favour of Chrysippus' view that reason is in the heart, basing his claims on the fact that voice issues from the chest (Galen P//P11.5 = SVFm (Diogenes) 229-30). On the other hand, Galen mocks him for inconsistency, saying that, having defined soul as an exhalation, he forgets he has done so and later says that it is blood, 'agreeing with Empedocles and Critias, not Cleanthes and Chrysippus and Zeno' (in. 30). If this is not mere carelessness (or Galen's malice) it is unorthodoxy of a major kind.
Il6
MARTHA G. NUSSBAUM
In a revealing passage of the de Officiis, Cicero reports that Diogenes had a difference of opinion with Antipater about the moral obligations of the seller. Antipater held that the seller was morally obliged to disclose any fault in his merchandise. Diogenes held that he needed to do so only up to the point required by law, no further: 'It is one thing to conceal, another to be silent . . . What indeed could be more stupid than to tell the buyer the flaws in what he is buying?' (m.50 = SVF m (Diogenes) 49). I suggest that this principle guides, in effect, the exposition of doctrine in the Peri mousikes. Diogenes presents a view ofpaideia that is hard to reconcile with the cognitive view of the passions, and strongly suggests something like the Posidonian tripartite view. He does not say this to the Stoic buyer. He is silent about the novel properties of his wares. So it is up to the buyers to examine them closely. This we shall now attempt to do. Diogenes was known for his keen interest in poetry and music. He wrote a treatise Peri phones that gave definitions of phone, lexis, and logos (see page 103); this was the major attempt in that area until Posidonius; to judge from the brief references in Diogenes, it contained discussions of phonology, of diction, of grammar, of rhetoric. He wrote a work on rhetoric, fragments of which are also preserved (cf. SVF). And he apparently wrote a work Peri mousikes, concerning the effects of poetry/music on the character of the young. The difficulties of reconstructing this work are considerable, since it is preserved only via hostile paraphrase in the Peri mousikes of Philodemus. The textual and philological problems of this fragmentary work are great (though progress has certainly been made by Neubecker's new edition (1986) of book iv, the most important book for our purposes). In addition to these problems one also has to face the problems caused by Philodemus himself, a stridently hostile and not terribly acute reporter. 13 The subject of Diogenes' treatise is mousike. Philodemus usually treats this as if it means 'music' as opposed to 'text' (semasia). He reports, however, that certain unnamed critics (tines) call him agroikos because he takes the term this narrow way; they charge that his whole critique of Diogenes fails on that account. Appealing to Plato 13
So as not to encumber the text unduly with references to the several editions of the work, I shall give citations from book iv using the chapters and page numbers of Neubecker's edition - Philodemus (1986) - and from other books using the SVF numbers. For one passage from book in that is not in SVF, I shall cite van Krevelen's edition - Philodemus (1939)-
Poetry and the passions: two Stoic views
117
as their authority, they say that one would have to be pretty apaideutos to forget that melody and rhythm are wedded to text and have their effect in connection with text. On Philodemus' view, they say, Pindar and Simonides will not count as mousikoi; and they seem to treat this as a reductio ad absurdum of Philodemus' position. Plainly, then, for the defenders of Diogenes these are central cases o£ mousike (cf. above pages 103-4). Philodemus replies that the opponents do in fact neglect, or at least subordinate, the text; in doing the same himself, he is simply following their lead. And besides, he adds, Aristoxenus uses the term mousike in the narrower way (iv chs. 17-18, 73-8 N). As for the lyric poet, according to Philodemus he is a mousikos when he is producing asemanta, a poet when producing logos. But it is clear that this account of mousike belongs to Philodemus; Diogenes had a more inclusive understanding. It appears, then, that Diogenes followed Plato - probably with explicit reliance on Plato's authority - in considering mousike to include text and its musical accompaniment taken together; the lyric poets are central cases. On the other hand, it will become clear that Philodemus is in a way correct: Diogenes does neglect the text as a cognitive entity, focussing on the effect of auditory stimuli on non-rational movements. It is interesting to observe that, here already, the authority of Plato is apparently linked with a departure from Chrysippus. Diogenes' general thesis is that mousike is valuable for anyone who is not perfect (SVF 54), for people of all ages, Greek and barbarian (58). Its chief value is that it 'makes the disposition [hexis] very harmonious and very rhythmical' (SVF 56). Note here the noncognitive language, which points directly ahead to Posidonius. He cites the musician Damon as authority for the fact that music contributes to the virtues (SVF 56). (Recall that Damon was cited by Galen in defence of Posidonius, possibly with reference to some use of his authority by Posidonius himself.) In fact, claims Diogenes, mousike produces characters that are noble and serious (IV.I, 37 N), drawing the hearer to good dispositions (sunephelkomene, 37N). Philodemus reports that these changes in character come about, for Diogenes, 'according to non-rational perception' (kata ten alogon epaisthesin, 39 N).
Most of Diogenes' arguments seem to have consisted of asserting a series of specific cases of this general thesis, and then supporting these by examples. He discusses the usefulness of music in connection
Il8
MARTHA G. NUSSBAUM
with each of the virtues, taken one by one; the claim (as in Posidonius) seems to be that mousike contributes to virtue by balancing and harmonizing the character: in other words by developing the noncognitive side of virtue. Philodemus is especially amused by the idea that tunes and rhythms can make people more just: a claim that Diogenes apparently supported with reference to Plato (iv.15, 70 N). At least in Philodemus' report, what Diogenes said was that vocal sounds that are kinetikai akoes alogou (that 'move irrational hearing') can be important to the disposition of the soul where justice is concerned. Philodemus scoffs at this, saying that no disposition concerned with choice and avoidance in connection with one's fellow citizens could possibly be produced or enhanced in such a way. But above all Diogenes seems to have focussed more directly on the passions, holding that music can console grief, assuage the torments of love, moderate the feelings displayed at drinking parties, contribute to friendship by producing friendly feeling (philophrosune) (iv, chs. 8-10). Again and again, Philodemus shows that he, at least, finds Diogenes' position to be a non-cognitive one, one that relies on the ways in which music affects non-rational movements and perceptions, rather than judgments. For again and again he argues against Diogenes that real moderation, real consolation, real erotic balance - these require correct belief about what is worth pursuing and what not. He ascribes to Diogenes, with ridicule, the view that music contains something that is 'stirring by nature' (kinetikon phusei) in much the way that fire contains something kaustikon phusei (iv ch. 5). To Diogenes' story that Thaletas and Terpander stopped a civil war by playing music, Philodemus asks, how on earth can aloga mele really do anything for a logike diaphora (iv. 11, 63 N)? This story is so like Posidonius' story of the flute girl that we have another reason to suspect continuity between the two thinkers. Still a further similarity is to be found in the fact that Diogenes, like Posidonius, holds that mousike must be chosen bearing in mind the particular disposition of the hearer: 'mousike can awaken an unmoving and sluggish soul and lead it to the disposition that the appropriate melody naturally awakens in it. For all are not moved in the same fashion by the same thing. Or the opposite: from a darting and rushing disposition it can calm it and make it serene' (SVF 62). The similarity of this passage to Posidonius is obvious. We meet once again the same two extreme characters - the sluggish one
Poetry and the passions: two Stoic views
119
and the darting one; and the passage even contains direct verbal resemblances. It seems to imply a Posidonian prescription for the reform of poetic education, one that would consist in the judicious selection of mousike for each occasion and each individual. The focus of Diogenes' attention is on various types of lyric verse set to music. We hear mention ofenkomia, marriage hymns, laments, love songs (iv ch. 3) - all lyric forms. Elsewhere Diogenes mentions military music, singing, dancing, and lyre-playing at sumposia (SVF 79). Tragedy and in general drama are mentioned in one passage (SVF 67) - but according to Philodemus the emphasis, even here, is on the lyric element: for he denounces Diogenes' focus on choral dancing, which according to Philodemus, contributes nothing valuable to the dramatic performance (iv ch. 4). There are passages in which Diogenes seems to ally himself with Chrysippus' view of pathe: but I think we will find that they do not imply a very deep commitment to that view. In an early passage, he says that a poem with music moves the logike dianoia more than one without (SVF 71): but his examples are of the general impression of weight and gravity produced by the musical element, not of any really cognitive effect. In general, his talk of the motion-imparting properties of music is Posidonian rather than Chrysippan, focussing on balance and harmony in psychological motions. So even if he uses the words logike dianoia (cf. also book in, van Krevelen (1939) p. 50), the use seems more cosmetic than functional. Much the same is true of his claim that the hearer's auditory perception of mousikeis a special kind of epistemonike ('knowledge-related') perception. For it is clear that this is a way of hooking up musical motions with the reasoning element in the soul; but nothing seems to be done with the device. It certainly does not seem to have the consequence that mousikeis seen as having an impact upon judgment (especially book iv ch. 1). And when Diogenes holds that music represents likenesses of character traits (SVF 63 and iv, 40-4 N) - even then, he does not seem to be thinking of the way in which an epic or dramatic poem represents a hero of a certain sort. For he says that these 'likenesses' are crafted and presented to the ear 'by harmoniai\l* which give an impression of 'splendour and humility and manliness and weakness and orderliness and boldness' (SVF 63; IV.I =40-1 N). Both the alleged mechanism and the choice of traits - not full-blown virtues, 14
There is actually considerable unclarity about the text at this point. I translate von Arnim's version; but Neubecker, more conservative, does not restore the word 'harmonies'.
120
MARTHA G. NUSSBAUM
but what we might call manners or styles - suggest that we are dealing with a position closer to that of the contemporary aesthetic theorist Suzanne Langer, who is herself close to Posidonius, than to that of Chrysippus.15 The tunes themselves, with their rhythms, contain likenesses of human feelings and dispositions, in that they are kinetically isomorphic to the movement pattern of those traits. Thus a sluggish melody 'represents' sluggishness, a darting one a 'darting' character, and so forth. Some elements of Diogenes' view could be imagined as parts of a cognitive view, and are at least compatible with it. The idea, for example, that music arouses the sense of hearing (kinetikai akoes alogou) might be associated with the cognitive view's idea that passional change is accompanied by a change in the way things appear to the agent, that is to say in the group of (usually propositional) appearances to which the agent will assent. The cognitive view could conceivably also have used music to 'change the topic', so to speak, getting the agent to focus for the time being on something other than the object of the passion, and thus preparing her for reasoning. (In this function it would not necessarily be privileged above other distractions, such as doing logical exercises or going in for sports.) It is always possible that Philodemus' report has altogether distorted the point and emphasis of Diogenes' account, and that he really had a cognitive view of the passions and their therapy that Philodemus simply fails to mention, in relation to which the account of irrational musical influences is an adjunct in one or both of the ways just mentioned. Possible, but, I think, unlikely: for there is no sign of these functions for music in the fragments of Diogenes' position. There is no indication that music works by altering propositional appearances or assent to them. (That music 'arouses hearing' surely implies no more than that this is the sense that takes it in.) And there is no sign that music is, for Diogenes, just one among many equally beneficial distractions. (Nor is there any evidence I know of that Chrysippus gave music either of these two roles.) On the other hand, there is much evidence that Diogenes gives mousike a special place precisely because of the power it has to move in non-cognitive ways. So there appears to be no serious obstacle to understanding 15
S. Langer (1953, 1957) has defended the view that music contains motions that are isomorphic with the movements of human emotions; she develops a very interesting account of musical expression along these lines.
Poetry and the passions: two Stoic views
121
Diogenes as a Posidonius in the making - a Posidonius, that is, without a developed Posidonian philosophy of mind. It seems clear that Posidonius was much influenced by Diogenes, and follows his general line of argument. I suspect that some reports that are now traced to Posidonius - for example Galen's flute-girl story - may actually have their roots in Diogenes. And both thinkers consider Plato their mentor. But Posidonius seems right to say that there is a major task (not performed, apparently, by Diogenes) that needs to be performed before the topic of paideia can be given a really convincing treatment: namely, the overthrow of Chrysippus' account of the pathe. If one is heading in the direction they both are, where poetry is concerned, then that is what one must do. And Plato, guide to Diogenes already, is there to lend a hand. It is somewhat disturbing to find oneself in agreement with Philodemus on any matter of importance, since his way of arguing is so crude and boringly polemical, his way of expressing his points so repetitious. But I believe that he is in essence right to say that this is a weak position, one that betrays a serious misunderstanding of the passions and the contribution poetry can make to their education. Poems are not just sound, tune, rhythm - they are idea, statement, action, confrontation. And by neglecting these elements, Diogenes and Posidonius have saddled themselves with a very thin account of mourning and consolation, of love, of anger and reconciliation. To suppose that hearing jolly or mournful tunes really alters a passion that is based on a cognitive commitment seems superficial. As Chrysippus has argued, passions are commitments made with the core of the personality, commitments about what is really important, assents to certain value-laden ways of seeing the world. The non-cognitive view yields an impoverished view of education, reducing the formation of a child's emotions to a kind of dog-training. THE COGNITIVE VIEW
The non-cognitive view did not altogether displace the cognitive view. Chrysippus' general account of the passions prevailed, on the whole, in later Stoicism; it was adopted by Epictetus and Seneca, and it also heavily influenced non-Stoics such as Cicero and Plutarch, who think of it as the central Stoic view. And yet, no extensive whole discussion of poetic paideia by a major Stoic thinker in this
122
MARTHA G. NUSSBAUM
tradition survives. Chrysippus' On how one should listen to poetry is the work we would like to reconstruct. (And the thought of Cleanthes plays a part here too, as we shall see.) But we must approach this task tentatively, and using a variety of sources. We can draw, first of all, on numerous statements of Chrysippus himself about the poets, and on his practice of poetic interpretation. We also find helpful material in Seneca and Epictetus. Strabo, who explicitly portrays himself as a Stoic reporting Stoic views, contributes some helpful remarks. But many points on which we would like to be informed are not covered in these sources. We need, therefore, to turn for elaboration to another work of surprising interest: Plutarch's How the young person should listen to poetry. Philip De Lacy has argued that
this work relies heavily on Chrysippus' work of similar title, and can be used as a source for Chrysippus' view. Obviously one must proceed with caution here, but I believe that De Lacy is basically right. Certainly Plutarch here refers to Chrysippus and to Stoicism with approval, and describes (with crucial citations) parts of the contents of Chrysippus' work. He describes the purpose of his work as one of arguing against an Epicurean rejection of poetry; and throughout he shows himself in general sympathetic to Stoic positions. (He is critical of allegorical reading; but, as we shall see, that device is not as central in Stoic theory as is sometimes made out.) Where we do find parallels between Plutarch's argument and material in more orthodox Stoic sources, agreement is close. So I believe that we may use the work cautiously, to flesh out a Stoic picture of cognitive poetic education. The general character of the cognitive view of the passions should by now be fairly clear. The associated theory of poetry holds - to turn here to Strabo - that the aim of poetry is didaskalia, not a non-teaching sort oipsuchagogia. The ability of poetry to improve ethe and pathe and praxeis derives from its cognitive content. In fact, poetry has a close kinship with philosophy (1.2.3). This view need not neglect poetic form: for we have seen that, in Republic ii-m and especially in Aristotle's Poetics, such a view prompted a searching inquiry into elements of poetic form - especially narrative, character, and structures of audience identification. But it will be interested in form as expressive of certain commitments as to the way the world is, certain patterns of salience. It will treat form as showing us something about how we might see ourselves, our relationships, our commitments, our vulnerabilities.
Poetry and the passions: two Stoic views
123
THE DEFENCE OF POETRY: DANGERS AND BENEFITS
We have, then, a view according to which the passions are (false) recognitions of great value in external and uncontrolled things and people. According to this view, poetry affects the hearer's emotional disposition through the judgments it leads the hearer to form about what is important and what not. The strand in Plato's argument that focussed on the acquisition of belief or judgment concluded that most existing poetry would have to be banned, since it seemed so obvious that most of it shows alleged heroes as fearing and grieving for things beyond their control. Aristotle compounded the problem, arguing that the very structure of the tragic genre, its characteristic ways of engaging its audience, relies upon fellow-feeling with a hero who suffers an undeserved and significant reversal, and on the emotions of pity for the hero, fear for oneself, that the witnessing of such an event evokes. If this is so, we might expect a Chrysippan Stoic to conclude that tragic poetry - and no doubt most of epic as well - is false and pernicious in its very structure, indissolubly wedded to values that must be rejected from Stoic education. And we might expect Stoics, like Plato, to permit only praises of the goodness of good gods and heroes. This, of course, is not what happened. The partisans of the cognitive view are also zealous defenders of poetry. Even more clearly than the non-cognitive theorists, they rely heavily on existing poetic texts in many ways - Homer and Euripides being the favourites. We must now ask to what extent they perceived the dangers of poetry for their ethical position, why they felt it was so important that it should be retained nonetheless, and how they proposed to guard against the dangers. Dangers
Evidence that this group of Stoic thinkers saw the dangers of poetry for their pupils is not copious, but it is telling. Cicero's account of the Stoic position on the origins of vice names the poets as among the environmental culprits (along with parents, teachers, and nurses): they take the 'soft and unformed souls' and 'bend them as they wish' (de Legibus 1.47 = SVF in (Chrysippus) 229b). This is probably a reference to Chrysippus' much-discussed account of the origins of vice. Seneca presents the dangers at greater length; and similar
124
MARTHA G. NUSSBAUM
views are developed in an interesting way by Plutarch, who may tentatively be taken here to be reporting, roughly speaking, Chrysippan views. Seneca's primary complaint is that poets are indifferent to the (ethical) truth. They do not think the search for truth is important. {de Beneficiis 1.3.10, 1.4.5); a n d they t e ^ many stories that are false and potentially pernicious. These include false stories about the underworld {Consolatio ad Marciam 19.4), false stories that the gods are involved in vice - for example, stories of the love affairs of Jupiter {de Brevitate Vitae 16.5, de Vita Beata 26.6), false praises of wealth (Epistulae 115.12 ff). What interests Seneca about these poetic falsehoods is their effect on the passions of the hearer, through a formation of judgment. (He seems to be interested primarily in dramatic and epic poetry - poetry with a marked narrative content; he once expresses disdain for the trivialities of lyric - Ep. 49.5.) The tales of the lower world 'stir us up with empty fears' because they make us believe that there are bad things waiting for us after death. Stories ofJupiter's lust 'feed human error' by making people believe that it is a good thing to desire and act that way (Brev. Vit. 16.5); their shame is removed {Vit. Beat. 26.6) when they see that the gods themselves act this way. Tales of the glories of wealth, and especially of the material luxury of the lives of the gods, 'set a torch to our passions' by making us believe that wealth is an extremely important thing, the best thing the gods can have or give {Ep. 115, 12). Thus, as in Republic ii-m, it is not just the presence of the value judgment in a poetic work, but its connection with figures who are displayed as admirable or divine, that produces, through a complex of judgments, the bad effect. Plutarch makes these same points and adds two others of considerable interest. First, he notes that the mechanism through which passion is aroused frequently includes identification with a character or characters who are suffering or feeling fear. This was implicit in Plato and Seneca, but he makes it explicit and develops the point at length. He quotes several agonized speeches of characters who fear either their own death or the death of a loved one (drawn from Homer and from Euripides' Iphigenia at Aulis) and comments: 'These are the voices of people who are suffering [pepontkotdn] and who have been snared beforehand by opinion and delusion. For this reason they seize hold of us and thoroughly shake us up, as we become
Poetry and the passions: two Stoic views
125
infected by their passion and by the weakness from which the speeches are said' (i7cd). This passage, whose terminology of weakness and passion seems to mark it as close to a Stoic source, indicates, plausibly, that it is our sympathy with the characters, our feeling of closeness, that permits the transmission of passion, as we come to share their deluded opinions about what is happening to them. The same speech said by a tough or evil or unsympathetic character, or without all context of character, would not move us as the speech of Iphigenia does. Furthermore, Plutarch adds a further very important point: Poetry as an art is thoroughly committed to certain ideas that are rejected by philosophers. Among these ideas are: the idea that the good and the bad are closely intermingled throughout human life in complicated ways - an idea, Plutarch remarks, 'that very definitely says goodbye to the Stoics [polla panu tois Stoikois chairein phrazousa],
who will have it that nothing bad attaches to virtue and nothing good to vice' (25c); the idea that sudden highly significant changes take place in life, that unexpected happenings have great importance; and, above all, the general idea that life is a complicated business - Tor the simple is without emotion [apathes] and boringly flat [amousonY (23d). In short, poets hold their audience by committing themselves to a non-Stoic view of the world. Our interest in narrative is closely linked to a certain false sort of interest we have in our lives and their events. Even the gods, in order to figure in a good poem, must, Plutarch continues, be falsified; they cannot be represented as 'free from emotion or error' 'in order that the exciting and striking element in the poetry should not be absent because of an absence of risk and struggle' (23d). It would be nice to believe that these wonderful comments, which Plutarch connects closely with Stoic ethics, have Chrysippan roots; they are extremely perceptive, and show a greater sensitivity to the connection between literary form and philosophical content than we usually find in ancient literary discourse. But whatever their source, they raise a most difficult challenge for the Stoic defender of poetry. For they tell her that poetic forms themselves, and the springs of the hearer's interest in poetic forms, are thoroughly wedded to a morality that Stoics must repudiate. The choice appears clear: either dull flat poetry, or goodbye to Stoicism.
126
MARTHA G. NUSSBAUM
Benefits The Stoics chose neither horn of this dilemma. They remained fascinated with the poetry that was already around and already loved. They offer several arguments for this continued engagement, claiming that poetry has such importance in human life that it should be retained despite its risks. First of all, poetry simply presents a great deal of information that the student, and even the philosopher, should know and use. Strabo indicates that he is following Stoic doctrine in regarding the poets as teachers about all sorts of things: in his case, geography above all. And indeed, he does appear to be following Chrysippus5 example. For it is clear that Chrysippus regarded the poets as a major source of evidence about psychology and physiology. He irritates Galen no end by appealing to Homer and other poets in support of his views about the location and structure of the soul. Seneca makes this same point about the ethical sphere, holding that poetic maxims are a major source of ethical information: 'How many things poets say that the philosophers either have said or ought to say' {Ep. 8.8). But such information, once discovered by the philosopher, could from then on be transmitted in a prose form. So this argument, while it gives the philosopher a reason to read the poets himself, gives him no reason to make use of them in education. The Stoics, however, have further arguments. In a famous comparison, Cleanthes insisted that poetic form sharpens and condenses meanings, making truths clearer to the audience: 'Just as our breath gives out a clearer sound when it is drawn through the long narrow passage of the trumpet and pours out from the broader opening at the end, so the narrow necessity of poetry [carminis] makes our meanings clearer' (Sen. Ep. 108. 10). Seneca, discussing this view, insists that moral maxims enclosed in poetry are more easily grasped and digested by the young; furthermore, their pithy quality inspires self-scrutiny and self-recognition in the audience at any age. He reports that when a clever verse about avarice is recited in a mime of Publilius Syrus, 'even the most mediocre person applauds and delights in accusing his own vice' (108, 8-9). The usual 'uproar' in the theatre when such truths are uttered (108, 12) shows that they lead to 'confession of the truth'. When such things are said in prose, 'we listen to them more carelessly and are less struck; but when metre is added and the excellent meaning is put in line by a definite
Poetry and the passions: two Stoic views
127
rhythm, the same maxim is as if twisted more tightly by the arm' (108,
10).
Such a defence does little, as yet, for existing works of poetry. For all that seems to be said here, we could replace the mimes people love with a collection of maxims in verse, thus avoiding the dangers attaching to drama in general. This was actually done in Seneca's time, to avoid the immorality of mime performances. Seneca the Elder was fond of such collecting; and collections of this sort played an important role in later education - St Jerome studied from one. But I suspect that Seneca's view of poetry here is more complex. The effect that he describes as happening in the theatre is surely not independent of the whole experience of going to a mime of Publilius. The relaxed atmosphere, the audience's identification with characters who live and to some extent speak like them and whose lives are full of engaging and amusing incident — all this appears essential to the effect of self-recognition and 'confession' that Seneca describes, although he is not explicit on this point. The fact that this passage is written later than most of Seneca's remarks about the dangers of poetry - and yet he still does not advocate replacing mime by a collection of maxims — indicates that he thinks of the beneficial effect as belonging to the theatre and the dramatic experience. A more extensive Stoic defence of dramatic and epic poetry is made in Strabo and in Epictetus. Strabo, referring to the authority ofhoipalaioi and to the thought of'our people' (i.e. Stoics), says that poetry is 'a kind of first philosophy' (1.2.3), which contributes to virtue in the hearer by luring him into a moral lesson with a novel and engaging story. Children, he says, love stories and the marvels and novelties stories present. This is why stories can serve us as 'bait' to get them involved in moral inquiry that would otherwise bore them. Furthermore, this moral instruction does not simply consist in maxims inserted in an otherwise morally neutral story: the story itself inspires virtue by encouraging the imitation of excellent heroes and by frightening the hearer away from vice. A philosophical argument or a simple exhortation on these points would have little effect; but the dramatic structure promotes involvement (1.2.8). Here we find, once again, stress on identification as essential to poetic experience; this time it is seen as morally beneficial. Strabo adds that poetry in this way is important not just for children, but actually for women as well, or indeed for any 'random
128
MARTHA G. NUSSBAUM
mob' - in short, for anyone lacking the patience and the intellectual training to understand a philosophical argument. And at any time this is likely to mean most people. Philosophy is 'for the few, while poetry is more friendly to the masses and can fill up the theatre' (1.2.8).
We now have two arguments in favour of the poets; both defend not only the setting of maxims in verse but also the narrative/ dramatic structure of conventional genres. Strabo's argument defends the value of poetry only for non-philosophers; the Cleanthes/Seneca argument, based on clarity and self-recognition, seems to defend its value for everyone, suggesting that the forcefulness of poetic language, and probably also (in Seneca) the power of identification with dramatic situations, plays an important role in self-scrutiny and 'confession' for anyone who is trying to be good. One final argument pushes this general defence further. It is found in both Epictetus and Plutarch; but Chrysippus' famous treatment of the Medea is likely to be a source for Epictetus in some passages, and what Galen preserves of Chrysippus' comments on Euripides suggests that he argued along similar lines. The argument is that hearing poetry - in this case, tragic drama above all - prepares one for the various misfortunes that can take place in a human life. At the same time it shows how morally ruinous it is to come to misfortune with a mistaken set of values. The importance of being prepared for misfortune is, of course, a constant theme of Stoic thought about the 'therapy' of the passions; no other school makes such a point of this. Plutarch, I think, shows the Stoic credentials of his poetic thought clearly when he emphasizes this feature — saying that it is a consequence of having experience of poetry that 'we ourselves when we encounter misfortunes will not be struck down or disturbed, but will bear calmly with ridicule and reviling and laughter' (35 D). Epictetus takes the same thought one step further. By seeing in tragic drama how many misfortunes can befall people, especially people of wealth and high degree, you will learn, he tells the pupil, not only to be prepared for misfortune, but also not to become excessively attached to the things that can be altered by fortune, and not to envy people who have a lot of such things. Having seen a tragedy such as the Oedipus, you will, he says, have a new way of looking at people who have a great many external advantages.
Poetry and the passions: two Stoic views
129
. . . remembering that tragedies take place among rich men and kings and tyrants, but no poor man has a role in a tragedy, except as a member of the chorus. Kings begin well: 'Hang the hall with garlands.' But by the third or fourth part, we hear: 'O Kithairon, why did you receive me ?' Slave, where are your crowns, where your diadem? Aren't your guards going to help you? When you see a person like that, remember all this: you are approaching a tragic character - not the actor, but Oedipus himself. (1.24.16-18) And, almost certainly following the example of Chrysippus, Epictetus adds a further elaboration. When we look at a tragedy, we do not only see what misfortunes life can bring; we also see that these things are devastating misfortunes only to people who have the wrong scheme of values. When we hear the tragic hero's laments, when we see the wreckage of that life - and see that these piteous laments are being uttered by someone who was supposed to be a hero, we see that real heroic dignity is incompatible with the attachment to externals that brings these people low. Setting all this moaning over against an example of real heroic dignity - for example, the life of Socrates - we learn to disdain externals and to reject the values embodied in the tragic hero. If we had not witnessed his downfall, we might not have seen how ridiculous, how lacking in dignity, such a person actually is. In attending to tragedies, then, a person's goal must be: to study how to remove from his own life mournings and lamentations, and such expressions as 'oimo? and Halas ego\ and misfortune and ill fortune, and to learn the meaning of death, exile, prison, hemlock - so he can say in prison, 'O dear Crito, if this is what pleases the gods, so be it' - and not that other exclamation, 'Poor me [talas ego], an old man, it is for this that I have kept my gray hairs.' Who says this? Do you think I am talking about some insignificant lowly person? Doesn't Priam say these very things? Doesn't Oedipus? In fact, don't all kings talk like this? For what else are tragedies but the sufferings of people who have been wonderstruck by external things, displayed in the usual metre? (1.4.23-30) And in making chance Medea,
a later passage he gives us a further 'definition' of tragedy, the same point: 'Look how tragedy comes about: when events befall fools' (2.26.31). His elaborate reading of the very close to Chrysippus' own surviving treatment (see
130
MARTHA C. NUSSBAUM
below, pages 142-3), makes a similar point. We are to see that what is wrong with this woman is that she has valued the wrong things; without the initial moral failing the mischance she encounters would have seemed no disaster and would not have given rise to murderous rage. Thus, when we see the terrible consequences of depending too much on the external, we have powerful motives, connected with our sense of our own dignity and our moral health, for leading a Stoic life. Thus tragedy is not simply a source of true statements: the tragic plot itself, and its central action, serve the purposes of Stoicism. I shall later return to this reading of the Medea; for I think it takes us to the heart of the fascinating and radical proposal the Stoics make for the reform of poetic spectatorship. But by now we can already see a difficulty emerging for their account, one that we will not be able to resolve until we understand this proposal. The difficulty is that the attack on poetry and the vindication of poetry do not seem to go together. The attacks assume that the listener or spectator is experiencing the work as an involved participant, in more or less the usual way: not just absorbing its truths, but caring about the characters, involved with their fate, identifying with them, 'infected' with their emotions. And Plutarch's attack suggested that things must be this way: poetry, without that kind of involvement in an interesting and surprising story, would be flat and uninteresting. Some of the defensive arguments do appear to be compatible with these observations. Strabo thinks of the nonphilosopher as caught up in the plot as 'bait', fascinated by strange adventures. Seneca, too, seems in Letter 108 to assume at least a certain degree of relaxed participation and identification. Epictetus, on the other hand, seems to be thinking of a very different spectator: a watchful, critical, actively assertive spectator, who dissociates him or herself strongly from characters like Oedipus and Medea, and refuses to participate in their fate. I shall be arguing that this new conception of the spectator is crucial to the Stoic rehabilitation of poetry; it explains why Stoics can at one and the same time love and revere the poets and also teach that the Stoic wise man is the best or only poet. But before we can get to these conclusions, we need to survey the whole arsenal of devices the Stoics have at their disposal for taking the danger out of poetry.
Poetry and the passions: two Stoic views
131
MAKING POETRY HARMLESS! 'SOME U P R I G H T STANDARD OF REASON'
The Stoics, then, have a problem on their hands. On the one hand, they agree with most of Plato's claims (in Republic n-ni) about the dangers of poetry, where the formation of value-judgments and the related passions are concerned. Indeed, they appear to go even further, holding that those dangers of passion are intrinsic to poetic genres. On the other hand, they see some compelling reasons for retaining poetry in the curriculum; and these reasons require the retention not simply of verse maxims, or praises of the goodness of good men, but of interesting narratives, and, above all, of tragic dramas. Whether based on a Stoic source or not, Plutarch's vivid metaphor gives an accurate summary of the Stoic position and its difference from the harsher position of Epicurus: Shall we, then, stopping up young people's ears with a hard and unyielding wax, as the ears of the Ithacans were stopped, force them to put to sea in the Epicurean boat, and to avoid poetry and steer their course clear of it? Or shall we instead, standing them up against some upright standard of reason, and binding them there securely, straighten and watch over their judgment, so that it will not be carried away by pleasure toward that which will harm them? The Stoics choose, with Plutarch, the latter course. Where do they find the upright standard of reason? And how do they apply it to the poetry they wish to preserve? I shall argue that the Stoic rationalizing arsenal has at least four weapons - the last of which is, I believe, by far the most important and the most interesting, though others have received more attention. These are: censorship (?); writing new poetry; allegorical interpretation; and what I shall call the art of critical spectatorship. Censorship
For Stoic censorship there is little evidence. But since there is said to be some it must be mentioned. In Chrysippus' account of his ideal republic, according to Plutarch, he forbade citizens certain 'things delightful to hearing and sight' {De Stoicorum Repugnantiis iO44d). Philip De Lacy (1958a) has interpreted this - not altogether implausibly - to refer to some sort of censorship of the arts. Perhaps it does.
132
MARTHA G. NUSSBAUM
But on the other hand the bulk of what we know about Chrysippus indicates that he was determined to find something uplifting about even the most objectionable art works (see pages 133—4 below). I conclude, therefore, that the evidence is too vague and thin to tell us anything. No other Stoic of the cognitive view seems to provide evidence leading in this direction. Writing new poetry
The Stoics famously held that the wise man was best at all the arts, including the art of poetry (SVFm. 654-6). And several of them put this into practice, Cleanthes, obviously, above all. His non-mythological and yet (to many) stirring verses show that it is possible to retain certain elements of poetic tradition while purifying tradition of harmful theological and psychological content. Seneca's tragedies and Lucan's epic extend the experiment to the two central poetic genres. But how and to what extent they do this with fidelity to Stoicism is a very complex issue that we cannot begin to discuss here. More local and piecemeal attempts were also made. Stoics were fond of rewriting existing poetic works so that they would yield a more acceptable sense. Diogenes Laertius tells us that Zeno rewrote two lines of Hesiod so that following a good example would be praised above discovering something oneself - rather than, as in Hesiod, the other way round (vii.25). He also rewrote Sophocles' lines: Whoever comes to do business with a king Is his slave, even if he arrives a free man. so that the second line read: Is not a slave, if he arrives a free man. Cleanthes rewrote Euripides Electra 428: Give to your friends, and when your body falls ill Save it by spending, to read: Give to your whores, and when your body falls ill, Wear it out by spending. Epictetus produces a satirical Romanized version of Euripides' Phoenissae 368, where Polyneices longs for 'The gymnasia in which I was reared and the water of Dirce', mocking an analogous Roman
Poetry and the passions: two Stoic views
133
youth's longing for 'Nero's baths and the waters of the Marcian aqueduct' (2.26.31). Plutarch suggests a long series of such changes. Interpolations in dramatic texts were also common practice. In such ways the Stoic thinkers maintained their control over the text, and showed the pupil that in the poetic experience philosophy was always in charge. Chrysippus is not mentioned as part of this tradition of composition. On the other hand, there is some evidence that he did, on occasion, write philosophy dramatically. Most accounts of his style are very negative; he is blamed for verbosity, bad diction, and flatness. And it must be said that the surviving fragments appear to have little literary distinction. But according to Fronto, his style was vividly dramatic: cHe . . . asks questions, describes, invents characters, puts his sentiments into the mouth of another' (SVF 111.27). Plutarch, too, speaks of the vivid and dramatic way in which Chrysippus presents opposing positions (How the Young Person 32). We may possibly, then, have here some attempt to emulate the dramatists, and to show how a philosopher can perform that function. Allegorical interpretation
But the Stoics were, on the whole, too fond of the existing poets to displace them in favour of their own work. And the primary task they undertook was to keep the usefulness and delight of poetry while preventing its potentially harmful elements from harming. One famous Chrysippan device to this end, which altered the course of literary history for centuries, was the device of allegorical reading. To the persistent charge that the artists are liars, Chrysippus replied by showing that any work, however apparently false and even morally pernicious, can be a source of truth about the universe, if only one follows its suggestions (huponoiai) and does not stop with the first apparent meaning. The most notorious and, in many ways, most revealing example of this technique concerns not a literary text but a painting. To the considerable disgust and amusement of later thinkers, Chrysippus produced a very serious explication of a painting that apparently showed Hera fellating Zeus. Obviously this painting is a paradigm of what would have been banned in Plato's ideal city; and it doesn't even meet with the approval of the various people who narrate parts of the story (SVF 11.1071-4). Chrysippus not only refuses to condemn it, but actually draws attention to it as a
134
MARTHA C. NUSSBAUM
source of insight. What he apparently said is that the painting represents the matter of the universe receiving the spermatikoi logoi. (It is impossible to tell, here or elsewhere, whether Chrysippus had a sense of humour.) Diogenes Laertius exclaims: 'He says what nobody would soil his mouth with on a bad day . . . Even if he praises it as philosophy of nature, the language used is more appropriate to whores than to gods' (SVF 11.1071 = Diog. Laer. vii.187). The attempt here is presumably representative. If I can make something respectable and even illuminating out of this work of art, Chrysippus seems to say, I can do it anywhere, and so can you. The choice of an otherwise not distinguished work seems best explained by some generalizing intention. And the example of oral sex - regarded as profoundly disgusting and degrading in the Greek world - makes it a 'worst-case scenario' for the rationalizing interpreter. A work of art that might have inspired unacceptable erotic desires, if seen with uneducated eyes, is now seen in a new way, under the guidance of correct teaching and an 'upright standard of reason'. This teaching tells the pupil to regard the painting not with any anthropomorphic identification or sensory memory and anticipation, not with any human fellow-feeling with the depicted figures, or any infection with their involvement in the body and its sexual activities as things of great interest and importance - but, instead, as a set of signs to what is really important - reason, and reason's plan for the whole universe. By a change in the relation of spectator to work, the work is rendered harmless, and even possibly helpful. And one can imagine a further bonus. Someone who becomes accustomed to looking at artistic depictions of sexual activity as signs of Zeus's rational plan might begin, after a while, to take up the same attitude to him or herself and his or her own body. He might begin, that is, to think that what is important about it and its activities is not that it feels this sensation or that, or has intercourse with this or that particular person - but, rather, that it plays its part in Zeus's scheme of things. I have focussed on this case from visual art since it shows the important points especially clearly. But it is plain that Chrysippus was fond of doing the same thing for literature as well - as is clear from Galen's account of his arguments for the placement of the hegemonikon in the region of the heart. Here Chrysippus uses an extensive allegorical reading of the Hesiod story of Athena's birth from Zeus's head, in order to show that it can be made to support and not undermine his theory. The idea is that Athena is conceived
Poetry and the passions: two Stoic views
135
somewhere in Zeus's middle, but comes out through the head - i.e., says Chrysippus, the mouth - just as reason is in the heart, but comes out as voice through the mouth. The myth is called a sumbolon of these physical facts. (Here it is not a question so much of defending the literary work as of defending his theory against the apparently contrary authority of the literary work.) Chrysippus has a lot of trouble with some details of the story — especially with Hesiod's claim that she came out from the 'top of the head' (koruphe); Galen delights in his discomfiture and the absurdities it produces. But it appears that where literature was concerned, allegory was not Chrysippus' first avenue of interpretation. He turns to it here only when challenged to defend his view; and most of his many readings of literary works are much more straightforward. It is fortunate, I believe, that this was so. For the laboured ingenuity of these allegories would hardly have been easily available to the young, to whom poetry is above all commended by the Stoics. This point, interestingly enough, was already made in Republic 11, where Socrates renounces this approach to the rationalization of literature: 'For a young person cannot judge what is a huponoia and what is not, but whatever opinions they take in at that age are likely to be hard to wash out, and unchanging' (378de). Even where sophisticated readers are concerned, ad hoc manoeuvres such as those by Chrysippus did not and do not inspire confidence. Galen is more than usually convincing when he tells Chrysippus to give up that approach. He refers to Plato's Phaedrus, where Socrates criticizes rationalizing explanations for the myth of Boreas: For my part, in the words of Plato himself, I 'regard5 all 'such' interpretations of myth 'as otherwise delightful, but as the work of a man who is excessively clever and hard-working and not altogether fortunate, for no other reason than that after this he must rectify the species of the Hippocentaurs, and then that of the Chimera; and a whole throng of such figures comes flooding in, gorgons and Pegasuses and an absurd crowd of other impossible and fantastic natures. And if someone, doubting them, is going to reduce them all to the probable, using a rustic kind of cleverness, he will need a lot of leisure time.' Chrysippus should have read this passage and then abandoned myths, and should not have wasted his time explaining their huponoiai. For if a person once gets involved in this, a countless number of mythical narratives 'comes flooding in,' so that anyone who examines them all will spend his whole life at it. [PHP 111.8, 230-2 D)
I36
MARTHA G. NUSSBAUM
Allegorical reading has great interest as one mode of a more general strategy: to break the links of poetic identification, producing a detached and critical spectator. It is not the most reliable such device, especially where the young are concerned. Critical spectatorship
But the Stoics have other devices that are as sophisticated as allegory without being as sophistical, devices that keep the reader involved with the literary work, and connected to the reasons for which it originally seemed important, while still protecting the soul from passion. Allegory wards off passion through a very radical shift in the position of the spectator. He is not only not to identify with the characters represented, he is not even to think of them as human beings at all. Instead, he sees them as symbols, usually of some non-anthropomorphic aspect of nature or reason. Such a shift is difficult to achieve and to sustain, especially in the young, given the vividness of the anthropomorphic depiction for which poetry is valued. And it might also forfeit some of poetry's benefits, since some, at least, seem to depend upon regarding the events as human events happening to people who are in many ways like oneself. The Chrysippan reading of the painting will not promote recognition and 'confession' in any person but one who already thinks of his body as a piece of the rational order of the universe. Such a person is already a Stoic philosopher. Nor, furthermore, will such a reading help anyone prepare for the many changes of fortune that human life brings. Accordingly, in the bulk of what the Stoics of the cognitive view write about spectatorship, we find another conception of detachment. From Epictetus' general remarks and concrete interpretations, from Chrysippus' handling of Euripides' Medea, from Plutarch - in this case sometimes explicitly reporting Chrysippus - we can reconstruct a complex and coherent picture of 'how one ought to listen to poetry' - the title of Chrysippus' treatise, and Plutarch's essay. The mutual reinforcement among these different sources of evidence indicates that we have here, at any rate, if not Chrysippus' view, at least a genuine Stoic view of the audience, and one that is consistent with what we know of Chrysippus' critical practice. We have seen that the passion-arousing effects of tragic poetry (and also, to some extent, of epic) depend upon certain formal
Poetry and the passions: two Stoic views
137
features of tragic spectatorship. First, it is crucial that the spectator identify with the hero or heroine, believing that the tragic action shows possibilities for him or herself as well. Second, the spectator must believe that the sufferings of the characters are really important. If one thought that the hero's loss was trivial, or even a good thing, deep emotions of pity and fear would not be possible. The hero takes his loss extremely seriously; and the unreformed audience does so too, entering into the way he views the world. Finally, the spectator must believe that the characters are in some sense powerless to prevent the damages - and that the spectator herself, being like the characters, is powerless to prevent similar misfortunes in her own life. Losses of loved ones, defeats in war, are objects of pity and fear because they come to the tragic characters from outside, in a way they cannot help. One will not view in the same way the arrest of a drug-dealer, let us say, or the punishment of a wife-beater. These events inspire no pity, since we believe that the bad outcome is the result of the person's own bad character and bad choices. Accordingly, insofar as we are trying to be good, we dissociate ourselves from that outcome and those possibilities. We therefore feel no fear, since we can easily prevent ourselves from committing similar crimes. Clearly, from the point of view of Stoicism, these three beliefs, all constitutive of tragic spectatorship, are false and pernicious. The sufferings shown in tragedy are important only to one who has the wrong view about what is important. And they are beyond our control only when we do not take control of our own lives, extirpating the attachments to externals on which these sufferings are based. We cannot control the events; we can control ourselves and the ways in which events matter to us. Accordingly, the Stoic spectator must take up a new attitude to the tragic characters. The essence of this attitude is a concerned but critical detachment. By supplementing the works of the poets with the continual guidance of philosophical commentary (which provides, as Plutarch puts it, a correct paidagogia - 15c), the Stoics hope to form a spectator who is vigilant rather than impressionable, actively judging rather than immersed, critical rather than trustful. The first step in this re-formation, and one that is insistently emphasized, is to get the spectator to realize that poetry can lie. What we have before us is not the way the world is, but only someone's appearance {phantasia, phainomenon). And we have to ask
I38
MARTHA G. NUSSBAUM
ourselves, as we do about any impression that comes before us, whether we are going to assent to it or not. In this process, the spectator has to realize that she cannot rely on the poet: as elsewhere, she has to rely on herself and her own judgment. Plutarch says that, therefore, the young must be armed from the first with the saying that poetry cares little for the truth (iyde). Epictetus and Seneca give similar advice. From the first, the main aim must be to produce a spectator who is vigilant and probing, active rather than receptive. This vigilance is directed above all at the relationship formed between spectator and characters. For the aim of critical spectatorship is to break the bonds of tragic fellow-feeling and identification, substituting a different bond, one that does not presuppose a shared scheme of values, but is based upon what we might call reforming zeal and compassion. Again and again, Epictetus warns the spectator not to yield to the view of the world held by one or another literary character, but to realize that this is just someone's appearance - and, as with any other appearance, we are the ones who must choose. Furthermore, when we see what happens to these characters who are attached to externals and who do not regard these appearances critically, we will observe that the impressions lead them both to suffering and to wrongdoing. So the spectator, once having suspended initial identification, will derive from her study of the plot a new motive to persist in that critical posture: Am I stronger than Agamemnon and Achilles, - that they, through following what appeared to them to be the case, both do and suffer such evils, and yet I am satisfied with what appears? And what tragedy has any other starting point? What is the Atreus of Euripides? Appearance. What is Sophocles' Oedipus? Appearance. Phoenix? Appearance. Hippolytus? Appearance. What sort of person, then, do you think it is who pays no attention to this problem? What do we call people who follow every appearance? Madmen. - Are we, then, acting any differently? (1.28.31-3) The appearances in question are the appearances of the characters within tragedy or epic - as the initial examples of Agamemnon and Achilles show. They are appearances not, clearly, in the sense of sensory illusions, but in the sense of impressions about matters of importance, impressions as to what is worth pursuing and what is not. Epictetus agrees with Republic n—111 that such impressions are starting-points of tragedy. But he then claims that tragedy can be, for this very reason, highly beneficial. The spectator is to
Poetry and the passions: two Stoic views have sufficient fellow-feeling with the characters to recognize that, as human beings, they share a common problem: what to choose, what to value, how to exercise control over appearances. Seeing what happens to people who solve this problem in the tragic hero's way - loving things external - she acquires again and again, as often as she sees a tragedy (and the importance of repetition and a variety of related works is stressed), a powerful motive to choose the Stoic solution. And the more she moves toward the Stoic solution, being watchful over herself and her passions, scrutinizing each action and each appearance, the more she will also be a good Stoic spectator of tragedy. The two practices reinforce one another. Above all, then, Stoicism undoes tragic identification by attacking the hero's (and thus the spectator's) belief that what we see here is a serious misfortune that could not have been prevented by intelligence. The misfortune, Epictetus insists, comes not from the events, but from the characters' scheme of values, which gives them a certain way of viewing the events, the way that makes for passion. And one's way of seeing the world is always in one's control. Therefore there is nothing grand or inevitable about tragedy - and it is a record of deplorable, if extremely common, foolishness. Tragic heroes are not the grand things that the tradition takes them to be; we should call them 'people who have been wonderstruck by things external', and their tragedy nothing more solemn than 'what happens when chance events befall fools'. Several techniques contribute to the formation of this detached and critical attitude. Most centrally and pervasively, our Stoic texts use and recommend the technique of philosophical commentary. The spectator is encouraged - first by example, gradually in her own efforts - to provide a running commentary on the action of the work she hears. In doing this, to begin with she should look for the poet's own commentary upon his characters - for when she does so, as both Seneca and Plutarch insist, she will frequently discover a criticism that approximates to the criticism philosophy itself would offer. This source of insight, Plutarch holds, is frequently lost through premature allegorizing, and is to be sought by staying closer to the literal sense of the text. His examples are taken from Homer (who seems to be critical of excesses of love and anger), Euripides and Menander. Seneca develops the same point with an example from Euripides: he imagines the poet himself reproving the audience for hasty reading; 'Euripides' urges them to wait to see what becomes of a character
139
140
MARTHA G. NUSSBAUM
who has a defective ethical view (Plutarch 19-20, Sen. Ep. 115, 15). Plutarch makes many sensitive and subtle points about the way in which a poet may undercut our enthusiasm for an initially attractive character, both through explicit statements and through the course of the dramatic action. This is, I think, an important part of the Stoic programme. Finding such judgments in the plays and poems themselves is important for them, if they wish to defend existing works of literature, retaining the 'classics' as a positive force in education. I think it is significant that their attention is focused on authors - especially Menander and Euripides - who could plausibly be said to be at times almost Stoically critical of ordinary human passions and the values that underlie them. Perhaps the emphasis simply reflects these authors' popularity; but I suspect there is more to it. Sophocles' scheme of ends would be far harder to accommodate. But the Stoic spectator is critical even when the poet himself does not oblige. She is encouraged to foster detachment further, through two further devices: generalization, and humour. Chrysippus, according to Plutarch, advised the spectator to apply poetic statements and situations to other similar situations - making characteristic use of a medical analogy. When doctors see the effectiveness of a drug in one disease, they apply it to other relatively similar diseases. So too, says Chrysippus, when a poetic statement seems valuable, we should 'not allow it to be linked to one matter only, but move it over to all similar cases, and accustom the young people to see the common link and to make intelligent transfers of what is pertinent, through many examples getting practice and training in sharpness of discernment. For example, when they listen to Odysseus' rebuke to Achilles, as he sits idly amusing himself with the maidens in Scyrus, they are to think how this rebuke applies to other profligate and wasteful people' (34b ff.). Such generalizing is as old as Plato; and Aristotle already made it a reason for calling the poetic experience philosophical. Chrysippus now uses it to explain how poetry helps the spectator interpret his or her own life. Notice, however, that generalizing does not work the way it does in Aristotle, through approval of and identification with the tragic hero. Instead, the spectator is detached from the types he observes. He may, as in Seneca's example, recognize himself in what he sees on stage; on the other hand, he may simply reprove and condemn. In neither case does he acquiesce in the literary flow and
Poetry and the passions: two Stoic views
141
let it take him along. He relies on the active guidance of his very own judgment. And notice that the active guidance of judgment forges a new sort of bond between characters and spectator. The spectator now views himself as a kind of doctor-in-training of the soul, seeking for explanations of human diseases and weaknesses. (See Plutarch 28b for emphasis on the importance of always seeking explanations.) He now sees poetry as a representation of human disease and weakness: Let the young person ... think that poetry is a representation of character and life, and of human beings who are not perfect or pure or entirely above criticism, but all mixed up with passions and false opinions and ignorance, but who on account of the goodness of their nature often change for the better. (Plutarch 26a) In what looks very much like a Stoic's reply to Aristotle's Poetics, Plutarch appears to endorse Aristotle's account of poetic representation, but changes it in a decisive way. The fundamental attitude of the spectator is now not admiring fellow-feeling, but a peculiarly Stoic blend of criticism and compassion. He doesn't suffer with the characters, he discerns their diseases, and wants to heal them. It is significant that Chrysippus urges the listener to take up a partnership with Odysseus, whose role is to offer criticism and wise advice. Both of them, like doctors, will be concerned about the diseases of the hero. Finally, as a further weapon against excessive involvement with tragic characters, the Stoic uses humour and satire. The topic of Stoic humour sounds rather unpromising. But I believe it is not, and that one could do interesting research along those lines. Epictetus listens to the lament of Polyneices in the Phoinissae of Euripides, we recall, and quickly produces a satirical Roman version, mocking the hero's attachments and inducing a mocking self-recognition in the young spectator. As often in Epictetus' treatment of the interlocutor, the grief and pathetic sufferings of tragic characters are, here and elsewhere, treated with a brusque and mocking vigour. In this way the spectator is urged to find (his own or another's) excessive involvement in trivial things foolish, to laugh at the weepers. Such laughter cements the distance between hero and audience. It says, in effect, T am not like you. I see things differently. You are a silly creature. You could have done otherwise. This did not have to happen to you. What you call tragedy is simply your own foolishness.' We can see why Menander is important for Stoic theatre when
142
MARTHA G. NUSSBAUM
we reflect that this structure is frequently present in New Comedy itself, in its mockery of obsessive types of various kinds. Plutarch concludes with several further pieces of advice that may well also be taken from Chrysippus - at any rate, they are adjacent to the explicit reference we have just discussed. They flesh out the general programme that we have described. This advice is: to seek the reasons for each thing that happened; to argue with poetic statements; to look for useful lessons, especially lessons in moderation; to comment and amplify with related philosophical remarks on the same point. We can get an idea of how these pieces fit together if we look at the approach of Chrysippus and Epictetus to Euripides' Medea. Much of Chrysippus' treatment is preserved by Galen. And Epictetus devotes more time to Medea than to any other fictional character. First, we see clearly that both find in the play evidence that the passions are false judgments and that the conflict between passion and reason is a debate about how (in evaluative terms) to see the world. This is especially plain in the evidence about Chrysippus, which I have discussed elsewhere; and his interpretation of the play has recently been convincingly defended by Christopher Gill (explicitly), and by Bernard Knox (implicitly).16 Second, it is equally plain that both philosophers are anxious to break, in the case of the reader or spectator, the ties of identification that would cause similar judgments, similar passions, to arise in them. Chrysippus insists that Medea is an example of a diseased person, who thinks life is not worth living without an external undependable item: all her suffering and her murderous rage stem from that error, an error that the audience is encouraged to repudiate. The general context uses a number of the devices I have mentioned to promote distancing: grotesquely ugly description of anger, satire and ridicule, philosophical generalization from her case to other related cases. Epictetus' treatment is perhaps even more interesting, since he gives us, in his lively address to Medea, an example of Stoic spectatorship at work, a spectatorship that is vigorously concerned, yet detached, that points out error, reproves, says things do not have to be the way they are: Medea, since she could not endure [not getting what she wanted], ended up killing her children. At least she acted with great spirit. For she had the 16
See Gill (1983), Knox (1977).
Poetry and the passions: two Stoic views
143
proper impression of what it means not to get what one wants. 'Well then, I shall avenge myself on the man who has wronged and insulted me. But what good do I get out of his being in a bad state? How can that be accomplished? I kill my children. But I shall get revenge on myself also. But what do I care?' This is the outburst of a great-sinewed soul. For she did not know where the ability lies to do what we wish, that it must be sought not from without, not by changing and rearranging things. Stop wanting your husband, and there is not one of the things you want that will fail to happen. Stop wanting him to live with you at any cost. Stop wanting to remain in Corinth. And in general, stop wanting anything else but what the god wants. And who will prevent you? Who will compel you? No one, any more than anyone prevents or compels Zeus. (2.17.19-22; cf. also 1.28.7-9)
In what precedes and follows, he generalizes this to the situation of the pupil. In short, the spectator is permitted to admire Medea and to be intensely concerned with her - but not with 'infection' or fellow-feeling with her love or her anger. The love is treated drily, distantly, viewed critically as a regrettable error. The ordinary spectator of Medea's tragedy would find something deeply painful in the way in which great and loyal love, betrayed, leads on to disaster; for they would think of such love as a fine thing, and it would seem horrible that the interaction between love and the world would produce such a morally disturbing result. She would fear for her own life, seeing that she has, in all likelihood, attached to love a similarly high value. The Stoic spectator feels none of this emotion. She is concerned about Medea as a doctor is concerned: with active energetic commitment to making her better, and to using her case to improve the health of others. What we have here, then, is a retention of the literary work, accompanied by a radical transformation of the spectator's relation to it, through the omnipresence of vigilant philosophical commentary. Does this reformed Stoic theatre escape Plutarch's original charge that a literature without surprise and passion on the part of the audience would be dull? Does it escape Aristotle's claim that the bonds that hold us to the tragic poem must be bonds of pity and fear, forged by fellow-feeling? To a surprising extent, I believe that it does. In answering these questions, the Stoics will follow Chrysippus' literary use of the medical analogy. The good doctor, they will insist, is not bored with his patients. He does not have to become infected with their diseases to be strongly concerned for their well-being. So too, the good Stoic
144
MARTHA C. NUSSBAUM
spectator is concerned with the characters - with a friendly and humane concern that is perfectly compatible with detachment from their sicknesses of passion. He views them as fellow imperfect human beings, and regrets that they are not sufficiently concerned with their progress. His attitude to them is one of rational wish (boulesis) for their good, and friendly warning (eulabeia) about the consequences of their sickness. If he could, he would talk to them and give them help. As for his own self-interest: in seeing their illnesses, he will frequently, as Seneca saw, recognize himself. Such recognition appears to require only the Stoic sort of concern and not any stronger sort. The possibility of self-recognition and 'confession' will generate in him still further interest in the literary work, and increase his motivation for engaging himself seriously with literature. The Stoic will claim that here is a way in which theatre can be truly interesting without being corrupting, truly exhilarating rather than producing people who passively wallow in pity. There is no true compassion in wallowing; in advice there is real fellow-feeling and love of humanity. It is, I believe, both useful and striking to compare to these attitudes the criticisms made by Bertolt Brecht of the 'dramatic theatre', and his famous proposal to replace it with what he called the 'epic theatre'. The problem with the usual dramatic theatre, as Brecht saw it, was that the spectator wallowed passively in human suffering, acquiescing in bad states of affairs as if they could in no way be helped. Thus the spectator surrendered control and ceased even to ask the question, 'What is under my control and what is not?' Brecht's response, notoriously, was to create a theatre in which dramatic illusion was regularly broken by philosophical commentary - in which the spectator would become aware of herself as an active, critical, reflective being, a being who asks, 'How did this come about? How can this be changed? How can these people live better?' There is fellow-feeling with the characters in the epic theatre, an awareness of common humanity and a concern with well-being. But this fellow-feeling issues not so much in pity and fear as in practical reflection and the giving of advice. And it expresses itself structurally in a Stoic kind of alternation between representation and commentary, making significant use of mocking detachment and humour to construct the new spectatorial relationship. There is, of course, an absolutely fundamental difference between epic theatre and Stoic spectatorship. For in the former the things
Poetry and the passions: two Stoic views
145
that need changing, the things to which the practical attention of the audience is directed, are material and institutional conditions, whereas in the latter they are internal diseases of thought and choice. Brecht, as a Marxist, was profoundly opposed to the idea that changes in thinking were the most essential to the improvement of human life. And Marx's own rejection of that idea was closely connected with his turn away from Hellenistic philosophy as a source of insight. There is disagreement, as well, about the nature of the spectator's detachment. For while Brecht's spectator is made to adopt a detached and mocking attitude to many of the evaluations characteristic of ordinary life - to the obsession with money and power, even to the obsessions and jealousies of erotic love - some areas of agreement with ordinary belief remain undisturbed, where a Stoic would certainly disturb them. I think above all of the approval of an all-encompassing love of a mother for a child, which is the mainspring of audience interest in The Caucasian Chalk Circle; a Stoic would surely have qualms here. On the other hand, in some equally fundamental ways Chrysippus and Brecht can agree: above all, in rejecting the Aristotelian idea that true insight is gained through and in emotional experience itself; in the view that the theatre will be practical only if it creates a critical and unemotional spectator, one who is actively in control of his or her own judgment. The nearness of the parallel can be appreciated if one considers the passage I have quoted as the epigraph to this paper, comparing it with Plutarch's prescriptions and Epictetus' practice.17 If we allow for the differences in the positive direction of thought proposed by the two thinkers - political change in the case of Brecht, change of thought in the case of Chrysippus - we might well have here - even in its muscular and dramatic style - a piece of Chrysippan writing. And we begin to see, through considering the parallel, how poetic drama can be seen with critical detachment and still be engaging. For the experience of seeing a play of Brecht is engaging, in a peculiarly exhilarating way - for the combination of insight, reflection, humour, and human compassion it evokes. Euripides seen through Chrysippus' eyes would yield, it is claimed, a similar experience. 17
Brecht's remark, made in 1936, is quoted in Tynan (1964) pp. 143-4.
146
MARTHA G. NUSSBAUM POETRY REFORMED?
We have, then, a plan for the reform and yet the retention of poetry: a plan that justifies the Stoics' proud boast that the wise man is the only true poet. For the plan puts the Stoic philosopher firmly in charge of the poetic experience, forging, in effect, a new work compounded from philosophical commentary and the poetic original. The claim is that such a partnership, and the new experience that issues from it, will retain the great advantages of poetry for human life - in particular its advantages as an explorer of 'foolishness' and passion - without incurring the risks run in the unmediated poetic experience. This proposal seems to me exciting. In contrast to the proposals of the non-cognitive school, it really grapples with the intelligence of great poetry, with its cognitive commitments, and with the identifications and evaluations usually built into the poetic experience. It does not treat a poetic text as simply a machine that imparts sub-rational animal motions. It has not only (I believe) a far richer and more adequate view of the passions, but also a richer view of the poetic text itself, one that does justice to its complex cognitive and narrative structure, and to the importance of our relations with poetic characters. And it treats the spectator like an intelligent human being - and perhaps even to a greater extent than does the Aristotelian view, in that it continually asks him to take charge of his own judgment. The proposal also seems to me superior to the allegorical strategy, in that it allows poetry to stay close to the daily lives and passions of human beings, where, the Stoics argue, its insights have much to contribute, generating recognitions of many kinds. In so doing, it seems to me to develop what is best and deepest in the original Platonic picture, bypassing what is weaker and in tension with that. And yet it avoids Plato's stern rejection of poetry by reinventing its audience. Will this work? First of all, we have to say that the institutional side of things remains, so far as we can tell, altogether undeveloped. Brecht did not just make remarks about the spectator: he actively forged a theatre in which the spectator would be what he imagined. Chrysippus, Plutarch, Epictetus apparently tolerated the retention of current theatrical practices, and merely added their comments and training to them, for any young people who fell under their influence. This seems to me a rather weak approach; for surely one's
Poetry and the passions: two Stoic views
147
experience of spectatorship is influenced far more by what actually goes on in the theatre - the reactions of one's fellow audience members, the style of acting, the mode of presentation - than by instruction one gets before or after concerning how to watch. The limits of the approach are even more apparent if one reflects that many of the spectators with whom the Stoics are most concerned are young and inexperienced in philosophy, not ready to hear Chrysippan lectures. These will form bad habits of spectatorship before even encountering philosophy - as would not happen, for example, to an audience brought up on Brechtian dramas, in which drama and commentary are an inseparable whole and the thought is woven into the form. A related problem is that Stoics seem vastly to underrate the extent to which their own evaluative commitments are at odds with those of the poets they praise. Perhaps one reason why they are willing to allow the young to hear unreformed Euripides or Homer, even before encountering philosophical instruction, is that they are convinced that there is an alliance between Stoicism and these poets on the most important matters. Even though their works portray characters who are passionate, the poets themselves, they usually suggest, subtly urge balance and apatheia. This seems to me to be plainly false, at least of many central cases. The works of the great tragedians do, as Aristotle believed, have built into their very structure the bases of fear, pity, and grief. Someone brought up on them will learn that it is really dreadful to lose a child or a husband, to be made a slave, to die. Euripides' Medea is the object of pity and sympathy, not simply of moral disapproval: and this is written into the play itself. Achilles is not simply a person who foolishly fails to listen to good advice: he is a hero who claims our love, and whose risks we allow to stand, in certain ways, for our own; again, this is written into the poem, and cannot so easily be removed. To some extent Euripides is a better ally for the Stoics than other major Greek poets: for he shows the ugly consequences of deep love and other external commitments so clearly that one might well read him as calling for the extirpation of passion - though I believe that it would be incorrect to do so, and that plays such as the Hippolytus and the Bacchae could not possibly bear such an interpretation. In short, it seems difficult to reform the poetic experience without reforming the poetry. A certain conception of the spectator's response is built into the structure of a tragic poem: in this Aristotle
148
MARTHA G. NUSSBAUM
was correct. One cannot so easily keep the very same poem and change the experience. It seems to me, then, that the Stoics should have gone further in the direction taken by Brecht, with his thorough-going reconstitution of the theatrical text and its mode of presentation. They could have left in far more of poetry's delights than Plato did - and Brecht is a valuable example of how this can be done. But they need to intervene more in both dramatic writing and dramatic production. The only Stoic who seems to have reflected profoundly about these issues is Seneca. For of course Seneca did produce poetic dramas that are in some way related to his Stoicism. And he did, apparently, produce them as recitation-drama, not as staged drama — a mode of production that might be thought suitable for the more didactic and dialectical approach to theatre that Stoicism requires. It is not possible to launch into the large question of Senecan drama and its relation to his philosophical work here, at the end of what is already much too long a paper. I have discussed some of the relevant issues elsewhere.18 But two points can be briefly sketched. First, Senecan drama presents Stoic psychology of passion and passional conflict with greater explicitness and clarity than any non-Stoic poetic text - even one, such as Euripides' Medea, that might be invoked in Stoicism's support. In this way, it furthers Seneca's didactic purpose. And, second, the dramatic structure of Senecan drama actively impedes sympathetic identification, promoting critical spectatorship and critical reflection about the passions. For the central characters repel the spectator, making it very difficult to be 'infected' by their passions, difficult to view them as anything but diseased. And the Chorus, frequently a guide for the spectator's response, is moralizing and orthodox to a degree unknown in Greek tragedy; it usually lacks sympathy with the principal character. In these ways, Seneca promotes Stoic spectatorship - although the complexities of his dramas make it clear that the tragic genre, even in such careful and sophisticated hands, is not an altogether reliable tool for Stoic moralizing. But my greatest objection to the Stoic revision is a deeper one. It is that, while apparently treating the poets as wise men and sources of insight, Stoic thinkers never really admit the possibility that poetry might actually have something to teach them - not just about 18
In Nussbaum (1991).
Poetry and the passions: two Stoic views
149
diseases, but about full health, not just about aberration but about the complete human life. Chrysippus makes a great show of learning from the poets; and no doubt he is convinced that he has done so. He even makes it fundamental to his philosophical method in psychology that he should turn to the poets as to reliable authorities. But where the most important matters are concerned, he does not do this at all. He does not really ask Euripides or Sophocles or Homer what view of the worthwhile and the good shows itself in the totality of their work. Nor does he open himself up to the avenues of learning through emotional response, above all - that their poetry makes available to the spectator who follows its lead. His proposal for defensive spectatorship (if it is his) is arrived at out of an experience that is itself thoroughly defensive and dogmatic. In all the hundreds of Chrysippan references to and discussions of poetic scenes and speeches, one never has the sense of a person who is puzzled, or troubled, or in any way changed by the poetic text - who is ever at a loss for what to do, what to say. But, as even Plutarch saw, being bewildered by the complexity of life, being surprised and troubled, is a great part of what poetry has morally to offer us. And if one steels oneself so adamantly against all this, what is the point of having poetry in one's life at all? Plutarch's treatise ends with a telling image. The young person, he says, needs a good helmsman in matters of poetry, so that, 'not prejudiced but educated, in a kindly and friendly and appropriate way, he may be sent forth from poetry into philosophy' (36b). The Stoic boat knows where it is going, and exactly who is steering it. 'Our people say that the only real poet is a wise man' (Strabo 1.2.2). Not with explicit Platonic opposition, but in a kindly and friendly and appropriate way, the poets have been banished - not from the city, but from control over their own meanings, their own truths.
CHAPTER 6
Seneca and psychological dualism Brad Inwood
SENECA AND ORTHODOXY
Seneca's 'philosophy of mind' presents us with many puzzles, but in this it is no more than a faithful mirror of his philosophy as a whole. In this paper I am interested in how his views on the structure and operations of the human soul relate to those of so-called orthodox Stoics, and how one goes about assessing such an issue. That Seneca is not a slavish or unimaginative representative of Chrysippean Stoicism seems clear from the most casual reading of his work and the most cursory glance through the abundant secondary literature. But it is not clear just how his views on the soul differ from those of the early Stoics. Seneca is sometimes described by the (traditionally pejorative) term 'eclectic'. But as Pierluigi Donini and others have shown, 1 we can no longer take for granted the usefulness of that simple description in the study of later ancient philosophy; indeed, it is not clear that we can readily agree about its meaning. I want at least to set aside the negative associations of the word: even if Seneca is in some sense an 'eclectic' it should not be assumed that he is for that reason a derivative, less powerful or less interesting thinker. 2 Further, I want to argue that in one important area of the philosophy of mind, the theory of the passions, we should not be calling Seneca an eclectic at all. His approach is open, but not eclectic. He is prepared to propose changes in traditional Stoicism, but those changes never Donini (1988a). I follow Donini (1982) ch. 4 in his general estimation of Seneca's 'eclectic' character: his proneness to adopt and integrate facets of other schools' thought into his own basically Stoic framework is a mark of his ability to appreciate the larger significance of philosophical theories and tailor them to his own insights and needs, a sign, that is, of his freedom from the restrictions of unreflective school loyalty. 150
Seneca and psychological dualism
151
threaten the core of Stoic ideas - as a change from psychological monism to dualism (let alone tripartition) would do. 3 Seneca's orthodoxy is inevitably judged by comparison with the early Stoic scholarchs. What it means to be a Stoic is determined in large measure by a thinker's relation to Chrysippus, Cleanthes and Zeno; in the normal way of treating the development of Stoic thought, these three philosophers constitute a benchmark to use in measuring the 'Stoicness' of other thinkers. Before proceeding too much further we should pause to reflect on one of Seneca's own statements on the question of school allegiance and orthodoxy. Hoc Stoicis quoque placere ostendam, non quia mihi legem dixerim nihil contra dictum Zenonis Chrysippive committere, sed quia res ipsa patitur me ire in illorum sententiam, quoniam si quis semper unius sequitur, non in curia sed factione est.4 For Seneca, being a Stoic does not mean following an inflexible rule to abide by the views of Zeno and Chrysippus; but it does involve following their doctrine when the facts themselves permit him to do so; the initial expectation is that Stoic views will be accepted, but Seneca assumes no obligation to do so. When, in his own view, their doctrine conflicts with the facts, then the facts prevail. Of course, this leaves out the difficult questions of how Seneca goes about determining what the Tacts' are and how strict a notion of conflict he employs; but I will not try to explore those complexities here. So how in practice can we assess Seneca's orthodoxy? Let us concede, first of all, that orthodox consistency in early Stoicism is not something to which we have direct access; rather, it is a methodological construct. I do not mean that orthodoxy itself is a myth, nor do I doubt that it remains proper to regard the first three scholarchs as a kind of criterion of what it meant to be a Stoic. After all, even Epictetus used to lecture to his aspiring philosophers on 3
4
The sense I assign to the terms 'psychological monism' and 'dualism' is only that described in Inwood (1985). See in particular chapters 2 and 5. To quote from page 28: one must distinguish two questions: 'does the soul in its entirety or just the hegemonikon have a multiplicity of powers? and the more important question, does the mind . . . have in it a power which can oppose reason and impair its functioning in the control of a man's actions and life? The latter is the key psychological question which bears on Stoic ethics.' And it is the key diagnostic question to ask when assessing a theory for monistic or dualistic character. This central point seems to have been missed by Ioppolo (1987). de Otio 3.1. This text and several others are discussed by Rist (1989). See esp. pp. 1999-2003 for his discussion of the subjects which I will discuss at length below.
152
BRAD INWOOD
Chrysippus.5 But the picture which modern scholars have made for themselves of an orthodox Stoicism teaching internally consistent doctrine, grounded on clear general principles - this picture seems to me to be an artefact of our reconstructive methodology. We generate such a construct by means of tacit assumptions which most scholars make when attempting a reconstruction of early Stoic doctrines: we assume (1) that school doctrine was on most points fairly unified, (2) that the high-level philosophical motivations for these doctrines were fairly consistent, and (3) that each of the early scholarchs held determinate and internally coherent views on most issues, and held them in more or less one form over his entire career. And yet, these are all questionable assumptions. The chances that any of them was strictly true in the real world are small. We have evidence of divergences among the three scholarchs on important theories in physics, logic, and ethics; but when there is little or no convincing evidence of divergence we tend to postulate doctrinal harmony, taking comfort from the ancient testimony about the internal consistency of the school in its early phase. Moreover, within the work of one of the early scholarchs we assume a fairly high degree of internal cohesion, referring often to the Chrysippean theory on this or that topic. But consider the length of Chrysippus' working life, the size of his corpus: compare this to the working life of Plato or Aristotle and to their respective corpora. Do we speak with comfort of the Platonic theory of forms? or the Aristotelian doctrine of substance? Of course not; or at any rate we should not. Why then do we speak of the Chrysippean theory of the pathe or of the early Stoic conception of the soul? On many, if not most, topics we must reconstruct the doctrines from scattered bits of primary and secondary evidence of quite various degrees of reliability. But although most of the applicable parallels militate against a harmony postulate strong enough to support a definite reconstruction, without something of the sort we cannot effectively proceed. And we have good reason for doing so: anything else would make the reconstruction of philosophically interesting theories virtually impossible and reduce the historian of thought to an activity little better than doxography. I do, however, mean to doubt that agreeing in all significant points with that 'holy trinity' was often regarded as a criterion for considering oneself a Stoic, or being considered a Stoic
Seneca and psychological dualism
153
The limitations of our reconstructive method matter little for the early Stoa considered on its own, since we can do nothing about the loss of direct evidence from the period. But things are different when we come to an author like Seneca, for whom we have a large part of his total philosophical output. We tend to judge him by comparison with early and orthodox Stoics in two ways: we ask how much he diverges from them, and we ask whether he was as rigorous as they. On both counts a more sober appraisal of our knowledge about the early Stoa will work to Seneca's advantage: it is harder than we like to think to show that a given doctrine diverges from orthodoxy, and it is highly unlikely that Chrysippus was ever as internally consistent as we might wish to present him as being. His doctrine of the passions leaves us with some uncomfortably loose ends to tie up for ourselves. Perhaps things are not all that different with Seneca; perhaps the looseness sometimes detected in his theory was not all that different from what we might find in the work of Chrysippus if we recovered it tomorrow. What I will be doing in this discussion is to take a very familiar text, de Ira 11.1—4, for closer examination and use it to try to show how Seneca's openness operates in one particular case, and how it differs from derivative eclecticism. But before doing so I will need to comment briefly on the recent work ofJanine Fillion-Lahille (1984) on the de Ira (for my approach to Seneca is quite different from hers) and to reflect on three preliminary issues: the significance of literary form, the importance of distinguishing the different kinds of dualism which we may find in a philosophical text, and apparent examples of explicitly dualistic statements. OLD PROBLEMS AND NEW APPROACHES
My method here represents a long-standing dissatisfaction with the procedures of Quellenforschung. In the detailed discussion below I take issue with Holler (1934) in particular, since it is his work which most clearly reveals the limitations of that method in advancing our understanding of ancient texts. Moreover, Holler is in a way the classical representative of the view that Seneca was a confused eclectic. In Holler's view Posidonius is most to blame for the dualistic muddle to be found in Seneca's work. by others. Seneca bears comparison with the even more difficult case of Marcus Aurelius; see Rist (1982).
154
BRAD INWOOD
More recently there has been another and better study of Posidonian and other influences on the de Ira, by Janine Fillion-Lahille (1984). For the particular text which interests me (i.e. the early chapters of book 11), she too finds significant Posidonian influence. The foundation of her analysis of the de Ira is, without a doubt, traditional Quellenforschung. A detailed analysis of Galen's de Placitis Hippocratis et Platonis and other sources (such as Cicero's Tusculan Disputations) is used to reconstruct in great detail the plan and doctrine of Chrysippus' and Posidonius' books on the passions. The quality of the analysis is higher than is found in Holler's work and the results are presented with a much more realistic caution. As far as the de Ira is concerned, her conclusion is that book 1 is basically of Chrysippean inspiration and book 11 of significantly Posidonian inspiration. Her further arguments about influences in the latter part of book 11 and in book 111 are new, but not relevant to my immediate concerns. More interesting is her fundamental re-evaluation of the relation between Chrysippean and Posidonian philosophies of mind. The traditional view, which I still hold despite Fillion-Lahille's argument, is that Posidonius' analysis of mind and the passions is plainly inconsistent with Chrysippus'. By contrast, Fillion-Lahille regards them as being complementary and argues at some length that we should regard Posidonius as being essentially orthodox,6 blaming Galen for the common misperception of Posidonius as a radical innovator. The overall effect of her study, put bluntly, is that she retains the old-fashioned Quellenforscher's theory of Posidonian influence in book 11 of the de Ira, but strips that theory of what is to me one of its most unacceptable corollaries, viz. that Seneca is a muddled eclectic. If one were concerned to do no more than rehabilitate the philosophical reputation of Seneca, this would be a tempting theory to support. But my concerns about the traditional view are also founded on a deep scepticism about the possibility of reconstructing Posidonian or other lost views by traditional methods. In that regard, Fillion-Lahille has not advanced much. Her Quellenforschung is better, but we are still shuffling back and forth the familiar counters: Cicero, Galen, Arius Didymus, Plutarch, Diogenes Laertius; and we are still resting hypotheses on unanswerable questions 6
Esp. pp. i54ff. She outlines three key theses which define orthodoxy. But these are insufficient to define orthodoxy; and the evidence cited seems insufficient to support the claims made.
Seneca and psychological dualism
155
such as: is Galen more misleading when giving us evidence about Chrysippus or about Posidonius? (I tend to think he is most misleading about Chrysippus, Fillion-Lahille stresses how misleading his Platonizing version of Posidonius is.)7 What I am trying to do in this paper is to break free from such sterile analysis and to ask afresh whether we have adequate grounds for even suspecting 'foreign' influence on the early chapters of de Ira 11. Fillion-Lahille's analysis of the Posidonian elements in 11. 1-4 (pp. 163 ff.) leaves Seneca committed to a form of dualism incompatible with views of Chrysippus, whereas I do not think that dualism is to be found there any more than it is to be found in Chrysippus himself. Once again the appeal of Posidonius has encouraged a Quellenforscher to postulate inconsistency in Seneca's text. Of course I still rely, for better or worse, on my own understanding of the early Stoic doctrine of the passions. Fillion-Lahille's analysis of it is noticeably different, and different in ways which make it easier for her to argue that Posidonius is not fundamentally at odds with the early Stoics, as Galen claims. For what it is worth, my analysis of the early Stoic theory allows Posidonius to play the role that Galen gives him.8 In other respects too Fillion-Lahille differs from me on the earlier Stoic theory. In particular (see pp. 128-30), she does not see that the key issue for orthodoxy is not the relation of irrationality in the soul to the passions, but whether or not there is any irrational part. By tackling the theory of passions on its own, outside the context of the entire analysis of mind and action, she enables herself to present Posidonius' addition of irrational powers as a minor complement to Chrysippean theory which helps with the causal explanation of the pathe.9 And perhaps from that narrow point of view it is no more 7
8
9
I have benefitted from reading an unpublished paper on Posidonius' doctrine of the passions by John Cooper, 'Stoic Theories of the Emotions'. He is critical of Galen's portrayal of Posidonius. Of the texts cited to justify rejecting Galen's view (pp. 123 ff.), only the Plutarchan Lib. Aegr. 6 (in fr. 154 EK) is of any weight at all; and even this does not say what Fillion-Lahille hopes it says. It does not say that 'desire, fear, and anger are nothing but judgments'. What it does is to give a description to distinguish purely psychic pathe from others: in contrast to mixed and bodily pathe, these involve judgments and suppositions. Fillion-Lahille says that this formally contradicts the assembled evidence of Galen (p. 124). Well, it does not. Cooper (op. cit. n. 15) has a more plausible view of the passage. The clearest index of how different our views of Chrysippean monism are comes from the fact that she argues (pp. 160 ff.) that it is a mark of compatibility with Chrysippean philosophy of mind that Posidonius locates the irrational powers in the heart, rather than in
156
BRAD INWOOD
than that. But as Galen saw, such an addition is in fact of much greater philosophical significance.10 Perhaps the simplest description of my position on the early chapters of de Ira 11 is this. Posidonius is not overtly invoked in the chapters in question. Seneca claims to be following the views of his school, and there is good reason to doubt that Posidonius represents that view. Consequently I limit myself to direct examination of this text in comparison with earlier Stoic doctrine. Fillion-Lahille has no difficulty in holding that Seneca's philosophical stance is significantly different in books 11 and 1, though she tries to play down the difference by bending Chrysippus and Posidonius towards each other (in my view unsuccessfully); I, however, prefer to assume that the two books are in harmony until forced by evidence to postulate internal contradiction. I do not think that book 11, properly interpreted, conflicts with anything in book 1. Let us turn to the other preliminaries. THE IMPORTANCE OF LITERARY FORM
With few exceptions we have very little detailed information about the literary form and technique of early Stoic treatises.11 But that many works of Chrysippus had a literary form is certain;12 Cleanthes' poetry and Zeno's book titles also point in that direction. We may be confident that the doxographical writers and later authors who discuss the early Stoics have stripped off what they understood to be the irrelevant literary apparatus which might mislead us, and in so doing they may have got things right or wrong. But with Seneca, the problem of controlling the literary apparatus is much more immediate and acute. We possess Seneca's works, and
10
11 12
a Platonically inspired lower part of the body. But the issue is whether the soul contains any irrational powers, not where they are located. In addition, I would want to disagree with Fillion-Lahille about the importance of assent, the nature of the excess which constitutes a passion, the proper understanding of Posidonius' pathetike holke (156ff.), and about the relation between Zeno and Ghrysippus. FillionLahille holds that Posidonius revived the dualism of Zeno (see esp. p. 163) and reasserted it against Chrysippus; since the case for Zeno's 'dualism' is weak (Inwood (1985) pp. 29 ff.), Posidonian dualism, if such it be, would be a more serious departure from orthodoxy, and Seneca's acceptance of it would be a much more serious departure from his normal attitude to the early scholarchs than Fillion-Lahille believes. See Stroux (1966) and Rolke (1975). The extensive amount of quotation from Euripides' Medea in his On Passions is the clearest sign. In the same work he makes crucial use of an illustrative comparison with a runner (SVF m.462), and this will be relevant to our appraisal of Seneca.
Seneca and psychological dualism
15 7
therefore must decide for ourselves how to handle the question of literary form. We cannot forget, as we might be tempted to do when dealing with Chrysippus, that Seneca, like Plato, writes deliberately in a literary genre, or genres, and the apparatus which naturally accompanies such work tends to exert an influence on our interpretation of his philosophical doctrine. It is not possible to provide a general theory of the interaction between philosophical content and literary form; the observations which follow are best regarded as an exhortation to caution in approaching the text of a literary philosopher: I do not claim that there is a single right way of interpreting metaphors and similes in philosophical texts, merely that a close and sceptical look will never be a waste of time. 13 I believe that one reason for imputing to Seneca an 'unorthodox' philosophy of mind flows from his use of images to portray more vividly psychological relationships and actions. Consider the theory of the passions, the central problem area for later Stoic psychology and the one in which suspicions of unorthodoxy and eclecticism are most likely to arise. The 'orthodox' Stoic view is that all passions are causally dependent on rational decisions of the mind, and that a conflict between one's rational decision-making powers and the passions of the soul, of the sort which Plato attributes to Leontius in Republic iv, is simply not possible. Yet in many passages Seneca writes as though the passions had a kind of life and force of their own, which enables them to take up a position in opposition to reason; and this has encouraged scholars, most noticeably Holler, to impute an eclectic form of psychological dualism to Seneca. A good example of this can be found in de Tranquillitate Animi 11.8, where the absence of inner harmony is described in terms which suggest dualism: 13
One might sketch out a typology of the relationships between metaphor and philosophical content. (i) The metaphor may be purely ornamental and evidently so, such that one feels no temptation to impute doctrinal significance to it; (2) at the other extreme, the metaphor may be an essential component of the philosopher's conception - his doctrine may have been developed by thinking through the metaphor itself and so the two are inextricably intertwined (the work of Lakoff & Johnson (1980) is interesting in this connection); (3) the metaphor may be used to enliven or reinforce the impact of a theory without being meant to determine the content of that theory, and in so doing it (i) may or (ii) may not introduce misleading elements or ideas; (4) there may be selfconscious analogies drawn to familiar experiences, analogies whose implications are meant to help determine the sense of a doctrine, and these analogies (i) may be essential to the justification of the theory or (ii) may be purely illustrative. There are surely many other possible categories. The kind of
I58
BRAD INWOOD
Tune illos et paenitentia coepti tenet et incipiendi timor subrepitque ilia animi iactatio non invenientis exitum, quia nee imperare cupiditatibus suis nee obsequi possunt, et cunctatio vitae parum se explicantis et inter destituta vota torpentis animi situs.
Most of the psychic instability here appears to be a matter of temporal vacillation, a perfectly orthodox analysis: 14 at one moment the agent feels or wishes one thing and shortly afterwards something else. But the phrase quia nee imperare cupiditatibus suis nee obsequi possunt
hardly admits of that interpretation: to speak of giving orders to or obeying the orders of one's desires would suggest the sort of quasiindependence for the desires which orthodox Stoic theory does not permit. Yet from the larger context of this work it is clear that Seneca is not turning to Platonism or to some sort of Posidonian amalgam of Platonic and Stoic psychology. The figure of speech which Seneca uses here would, if pressed for its literal descriptive significance, entail a non-monistic view of the soul. The question interpreters must ask is whether a given passage should be so pressed. Another text, this one not from Seneca, shows how difficult it is to determine whether to take such language literally. In Stobaeus' report of Arius Didymus' doxography, 15 there occurs a well-known borrowing from Plato's Phaedrus. The Stoic idea of excessiveness in the passions (sphrodotes) is illustrated by the Platonic image of a horse which is disobedient to its master; this is supported by a quotation from Euripides which puts cognitive powers (gnome) into conflict with the nature (phusis) of the agent. If one takes seriously the Platonic allusion, then this text reveals an openness to a synthesis of Platonic and Stoic psychologies. If one attends more to the Euripidean quotation, then things are less clear - for to find someone's 'nature' to be in conflict with his powers of rational decision is much less clearly inconsistent with what we understand as Chrysippean psychology. The 'nature' in question could be understood as bodily. As has been recently argued (though perhaps not as sharply as one might wish),16 there is a kind of dualism in Chrysippean Stoicism, a
14 15 16
cases I am interested in fall into types (3) and (4); (3-i) is the type most relevant to the parts of Seneca under consideration. Better seen in Ep.Mor. 52. Cf. Inwood (1985) p. 138. Eel. 11.89.4 ff. See Inwood (1985) p. 142 and notes. Couloubaritsis (1986). We should also distinguish metaphysical dualism, which deals with the principles underlying Stoic physics, from psychological dualism. Thus the Stoic division between god and matter can be seen as a form of dualism, and so can the
Seneca and psychological dualism
159
body/soul dualism which is compatible with monism as applied within the soul.17 If, in the passage of Stobaeus now before us, we focussed on the Euripidean text, nothing would determine for us which kind of dualism we should attribute to the author; it is only by giving primacy to the Phaedrus allusion that we feel confident in concluding that the psychology is unorthodox. The delicacy of this interpretation is enhanced by reflection on the evolution of Plato's own psychology. The dualism between reason and desire is, at its first appearance in the Phaedo, a body/soul dualism; great weight is placed on the internal unity of the soul. A reader of the Phaedo must take the phenomena associated in the Republic or Phaedrus with a lower, irrational part of the soul as being mere bodily disturbances. If we put ourselves in the position of someone who believes that the differences in psychological doctrine between the Phaedo and the later dialogues are superficial, it becomes more difficult to decide whether a Platonically flavoured dualism is or is not compatible with Stoic views. David Sedley's paper in this volume is acutely and correctly sensitive to the relevance of the Phaedo, read in the context of the rest of Plato's corpus, to the interpretation of Stoic thought. Because it is so easy to blur the issue of Platonic dualism, it becomes more and more difficult to decide whether a given problematic discussion by a Platonically prone Stoic is or is not 'orthodox'. It is at this point that the nature of the text becomes crucial. If we are considering a doxographical report, like that in Arius Didymus,18 it seems reasonable to press the significance of the allusions and quotations quite hard. It is hard to argue that Arius or his source is using this deliberately borrowed Platonic allusion in a
17
18
distinction between god (or providence) and the cosmos. These metaphysical dualisms appeal to Seneca, and he sometimes finds in them (correctly, I would say) similarities to important Platonic ideas - see esp. Ep.Mor. 65 and the Quaestiones Maturates, esp. the proem to book 1. I owe this observation to P. Donini. It may also be true that Seneca, in Stoic fashion, understands the importance of the parallel between the microcosm of man and the macrocosm of nature, according to which the god/cosmos or god/matter contrast is analogous to the mind/body opposition in man. But even so, this sort of dualism is not relevant to my analysis of Seneca's position on psychological dualism, which turns exclusively on the relations within the soul of man. I cannot discuss fully here the question of how Stoic materialism as applied to the soul is compatible with dualism. Suffice it to say that there is no a priori need for the two entities opposed in a dualism to be of different ontological orders - as body and soul are in a Platonic dualism. The corporeal status of soul and its perfect blending with body are compatible, as far as I can see, with an ethically relevant dualism between body and soul. For Arius' character see Kahn (1983).
l6o
BRAD INWOOD
casual or merely literary way and that he does not realize that his readers will assume that he is endorsing a full Phaedrus-style dualism. With the Senecan text we have considered above, it is more difficult to be sure, but I suggest that the general nature of the text and immediate context should dispose us to strip off the figures of speech and leave Seneca with a reasonably orthodox if not technically precise theory, at least in this place. Seneca's own explicit views on the use offiguresof speech are also worth considering. In Epistulae Morales 59, which is distinguished for its self-conscious attention to the technical accuracy of style, his own and Lucilius', Seneca offers some remarks on metaphor and simile (translationes, imagines) in philosophical writing. Despite the accuracy of Lucilius' style, Seneca does find some examples of these figures of speech (59.6): he groups them as 'comparisons' [parabolae], and adds: quas existimo necessarias, non ex eadem causa qua poetis, sed ut imbecillitatis nostrae adminicula sint, ut et dicentem et audientem in rem praesentem adducant. The effect of a good imago or translatio is simple, to compensate for our weaknesses as readers and writers and bring us face-to-face with the reality of what we are trying to say. Sextius, whose influence on Seneca is visible in a number of places, is offered as an example of this kind of philosophical writing; Seneca then translates and paraphrases an extended comparison between a man's relationship to his virtues and a general's relationship to his troops when on the march in hostile territory. The central point of comparison is made explicit (59.7-9)* Quod in exercitibus iis quos imperatores magni ordinant fieri videmus, ut imperium ducis simul omnes copiae sentiant, sic dispositae ut signum ab uno datum peditem simul equitemque percurrat, hoc aliquanto magis necessarium esse nobis ait. UK enim saepe hostem timuere sine causa, tutissimumque illis iter quod suspectissimum fuit: nihil stultitia pacatum habet; tarn superne 1111 metus est quam infra; utrumque trepidat latus; sequuntur pericula et occurrunt; ad omnia pavet, imparata est et ipsis terretur periculis. Sapiens autem, ad omnem incursum munitus, intentus, non si paupertas, non si luctus, non si ignominia, non si dolor impetum faciat, pedem referet: interritus et contra ilia ibit et inter ilia. Nos multa alligant, multa debilitant. Diu in istis vitiis iacuimus, elui difficile est; non enim inquinati sumus sed infecti. Another splendid illustrative imago is used at Ep. Mor. 72.8, this one being the work of Attalus. The Tool' is like a hungry dog being
Seneca and psychological dualism
161
tossed tidbits by his master; whatever he gets he swallows whole and then gapes for the next morsel. The point of the parallel is once again made explicit: Fortune is like the master, and fools who devote themselves to anticipation of good things wind up not enjoying them. The use of this imago does not commit Attalus to a belief in a personal Fortuna, any more than the image of the general and his troops commits Sextius to a belief that virtue is a distinct component of the mind which issues orders to other parts of one's personality, orders which may travel quickly or slowly.19 The point of such comparisons is limited and much must be discarded as theoretically unimportant. The same procedure will be in order when the imagines Seneca uses are ones he creates himself. Possibly the best and most important example of this would be at Ep. Mor. 37.4—5, where the highly coloured personification might well be taken as a sign of a dualistic attitude to the relationship between the passions and reason: but the dissection or analysis of the soul undertaken here is much more plausibly regarded as a mark of deliberately vivid presentation. Humilis res est stultitia, abiecta, sordida, servilis, multis adfectibus et saevissimis subiecta. Hos tarn graves dominos, interdum alternis imperantes, interdum pariter, dimittit a te sapientia, quae sola libertas est. Una ad hanc fert via, et quidem recta; non aberrabis; vade certo gradu. Si vis omnia tibi subicere, te subice rationi; multos reges, si ratio te rexerit. Ab ilia disces quid et quemadmodum adgredi debeas; non incides rebus. Neminem mihi dabis qui sciat quomodo quod vult coeperit velle: non consilio adductus illo sed impetu inpactus est. Non minus saepe fortuna in nos incurrit quam nos in illam. Turpe est non ire sed ferri, et subito in medio turbine rerum stupentem quaerere, 'hue ego quemadmodum veni?' EPISTULAE MORALES 92
Let us move on to another traditional reason for a dualistic reading of Seneca. It is usually held that Ep. Mor. 92.1-2 commits Seneca to a Platonic dualism in regard to the soul.20 And if Seneca is here 19 20
O n e might compare the externalization of the pathos love by Panaetius, as reported by Seneca in Ep.Mor. 116.5-6. See Zeller (1923) p . 737; Holler (1934) p p . n ff.; Pohlenz Die Stoa (1955) vol. 2 p . 112; H a d o t (1969) p . 91; Voelke (1973) p . 165; Donini (1982) p . 203. T h e most detailed discussion is that in Holler, and it depends almost completely on a by now outdated method for reconstructing middle Stoic, esp. Posidonian, doctrine. (One indication of this is the central role he assigns to Seneca's use of Posidonius' definition of anger (e.g. p . 23), which according to our source Lactantius (de Ira Dei 17.13 = de Ira 1.2) was only one of m a n y cited by Seneca.)
l62
BRAD INWOOD
clearly committed to that position - usually attributed to Posidonian influence - then in other places it becomes quite possible or even likely that he slides into psychological dualism. But if he does not commit himself to such Posidonian or Platonic views here, then it is less reasonable to suspect dualism in other cases. For there is no other text which appears so categorically to commit Seneca to this stance. But in fact this letter does not commit Seneca to such a position. The dualism is offered only as an agreed basis for discussion, a shared starting point for conversation which Seneca wishes to use in order to show that on this basis too the Stoically appropriate conclusions can be drawn. In other words, it is a merely dialectical or didactic move to agree with the middle Platonic psychology here canvassed. To support this claim, let us look at the letter more closely. Seneca begins (92.1) tentatively enough: 'I think,' he says, 'you and I will agree that external things are acquired for the body, that the body is taken care of for the sake of the soul, and that in the soul there are subservient parts (by which we are moved and nourished) given to us because of the leading part of the soul.' In what follows Seneca slides out of indirect discourse and into direct statement; but it is evident that it is effectively governed by the 'agreement': 'in this leading part there is something irrational and also something rational; the former serves the latter, and this latter is the one thing which is not referred to something else but which refers everything else to itself. For divine reason is in charge of everything and is itself subordinate to nothing, while this reason of ours, which derives from it, is the same.' There is nothing in the word 'agree' (convenire) which commits Seneca to belief in the propositions governed by the agreement. It is quite possible, even likely, that the agreement proposed here (note the use of the future conveniet) is merely conventional or dialectical. In a debate we agree to accept certain premises, but it does not follow that we actually commit ourselves to the truth of those propositions. I may well say to my Christian friend, 'OK, we will agree that God sent his Son to save the world from sin, but even so it does not follow that the route to salvation is solely through Christ'. Obviously I am not committed to any Christian beliefs; I am not even expressing sympathy with or openness to such beliefs. Seneca may well be making a dialectical proposal of this type. In 92.2 he
Seneca and psychological dualism
163
goes on: 'if there is agreement between us on this, it follows that we also agree that a happy life depends exclusively on having the reason within us perfected.' In the paragraphs which follow the key to human happiness is 'greatness of soul', the ability of man to rise above misfortunes and trials. For Seneca the key to this is the place of human reason in the hierarchy of nature, i.e. at the same level as divine reason. Consequently, it is most important to stress the subordination of other values to those of reason. Hence Seneca returns to the structure of the soul for further dialectical purposes in sections 5-10. When an Epicurean conception of the telos is rejected, this rejection is carried out by associating pleasure with the lower parts of the soul and then adding, in section 8, that the previously mentioned irrational part of the soul has the same subdivisions as Plato recognized in book iv of the Republic, the spirited and the appetitive or pleasure-loving part. In rejecting an Epicurean goal, then, Seneca will draw on the conceded Platonic views about the soul: in 9 he claims that the Epicureans have reversed the natural hierarchy of the soul by making pleasure the goal. The shared conclusion of Stoicism and Platonism (that the highest human happiness is to be found in the perfect development of reason, which we share with the gods) is supported more readily on the basis of the conceded Platonic premises than it would be if one worked from the technical standpoint of Stoic psychology. What we learn from this passage, then, is not that Seneca has come to share Platonic premises in his psychology, but that he is writing in an environment influenced by Platonic ideas and that he finds himself prepared to grant them for the sake of making clear progress on the substantial points of ethics he is most concerned with.21 And in this letter what is at issue ethically is something on which Seneca and the Platonists agree. His concession of an agreed starting point, then, should be put in perspective: Seneca has not turned dualist in letter 92, he has merely agreed to entertain certain Platonic theses which yield shared conclusions. It is these conclusions he cares about in this letter, and they are just as much a part of orthodox Stoicism as they are of middle Platonism. 21
I thank P. Donini for his comments at this point. He directed my attention to the relevance oiEp.Mor. 58 and 65. The significance of Platonism for Seneca is indicated very clearly in Ep.Mor. 58.25 ff.
164
BRAD INWOOD
So I would claim that we do not find any clear evidence of Platonic dualism in Seneca's own psychological theory, merely an openness to the invocation of it when it suits his purpose. This openness is surely encouraged by the respectability of body-soul dualism in Stoicism and by the ease with which psychological dualism and body-soul dualism can be made to yield the same conclusions (I am thinking of the development from the Platonic views of the Phaedo to those of the Republic). The way we interpret other apparently dualistic texts is affected by this reading of Ep. Mor. 92. When Seneca says (Ep. Mor. 71.27) that the sage himself is composed of two parts, one irrational (and this can be 'bitten, 22 burned and feel pain') and the other rational (and possessed of unshakable opinions), we should not conclude that the dualism invoked here is psychological - the rest of the letter hardly encourages this reading anyway. We should take special note of 71.29, which makes it clear that problematic psychological phenomena, such as the trembling, pain and pallor of the sage, are mere bodily affects: hi enim omnes corporis sensus sunt. We must conclude,
then, that the irrational thing which is subject to such sufferings is the body.23 We have, then, some clarifications which I hope will be useful in evaluating Seneca. We may distinguish body/soul dualism from psychological dualism; and we may distinguish primarily technical or doxographical contexts from more literary ones. And since there are no texts which clearly commit Seneca to dualism - Ep. Mor. 92 being explained differently - a much more straightforward view of Seneca can be entertained as the initial hypothesis for our consideration of de Ira. THE PSYCHOLOGY OF DE IRA III PROPATHEIAI
REVISITED
Holler (1934), pp. 16-24, argued, in keeping with his general account of Senecan psychology, that the doctrine of de Ira 11.1-4 owes a very great deal to Posidonius' adaptation of Stoic psychology, and thus that Seneca is committed by this theory to some 22 23
See I n w o o d (1985) p . 178. Also Rist (1989) p p . 2 0 0 1 - 2 . Pain is not in itself a problematic item here, since it is the term which is standardly used of what is merely bodily, in distinction from the pathos pain, which is a mental affect. But the tremors and pallor here, like the non-rational reactions of de Ira 11, are harder to dismiss to the level of the merely bodily.
Seneca and psychological dualism
165
clearly unorthodox psychological theses.24 In contrast, the effect of my present approach will be to make the issue of Seneca's alleged eclecticism and dependence on Posidonius secondary, and his use as a source for Posidonian doctrine quite limited. That, however, is a small loss compared with the gain of being allowed to see Seneca's development of earlier Stoic themes in their own right. In the first four chapters of de Ira 11 Seneca examines the genesis of anger, and asks whether it comes about by judgment or by impulse. The causal analysis of anger is of great general significance because anger is a passion. The theory of passions was one of the most vigorously attacked features of early Stoicism, and Posidonius' revised theory shows clear signs of Platonic influence. Moreover, the theory of the passions is a subset of the theory of action;25 consequently serious revisions to this theory entail a correspondingly significant change in the larger field of the Stoic philosophy of mind. And the theory of anger here has often been suspected of involving such a major revision, indeed of being a sign of eclectic psychological dualism, betraying the direct influence of Posidonius. I think that this suspicion is wrong, though it is one I have myself entertained. 26 The development of the Stoic doctrine of the passions which occurs here is a characteristically Senecan development. Without any commitment to theses borrowed from other schools, Seneca modifies in a sensitive and open way the established Stoic doctrine; the result is a theory more closely corresponding to 24
25 26
I d o not think that he is right about this, though it does not seem worth the space to present a detailed refutation here. I n the discussion below a few particular points c a n be brought out. But the essential problem with Holler's a p p r o a c h is that he treats Seneca as though he were a doxographical source for middle Stoic ideas, thus subordinating the understanding of Seneca to his plan to find Posidonian fragments in the text of Seneca. Since he antecedently believes Posidonius (and Panaetius too) to be synthesizers of early Stoic a n d Academic/Peripatetic ideas, he is actively seeking signs of such synthesis in Seneca, since that provides h i m with 'evidence' for Posidonius' views. T h e general method he employs for reconstructing middle Stoicism is not well founded. Fillion-Lahille, despite the general superiority of h e r Quellenforschung, succumbs (notably on p p . 163 ff.; also p . 45) to the tendency inherent in the method to find superficial a n d uncompelling similarities between the well-known author (Seneca) a n d the hypothetical source (Posidonius) a n d to use such alleged similarities to support a more sweeping thesis about historical influence. Given the state of o u r sources, this form of a r g u m e n t cannot be a n y more convincing than one antecedently wants it to be. A point which sounds less paradoxical when the Latin adfectus is used rather than the Greek pathos. Inwood (1982) p . 330, a n d Inwood (1985) ch. 5, esp. 175-81, m y first treatment of propatheiai a n d of de Ira 11.1-4; for references to earlier literature, see the discussion there; more recently, note the remarks of Long & Sedley (1987) vol. 2. p p . 417-18.
l66
BRAD INWOOD
Seneca's own experience and insight but which still entails orthodox conclusions, especially in ethics. The most important feature of the passions for Seneca and for earlier Stoics as well is that they should be 'rational5 in a broad sense: not in the sense that they embody or reflect the correct use of reason in the guidance of human action and reaction, but in the sense that they are products of human reason and so subject to control by that reason. The challenge for the Stoic theory is to account for what seem to be ineliminable affective aspects of human experience;27 such affective reactions form the focus of Seneca's discussion in the early chapters of book n, since Seneca wishes to distinguish them from affective reactions, like anger, which are rational in the broad sense, but misguided. The ineliminable reactions are in a sense exempt from moral judgment, since even the sage cannot control them. 28 It is clearly important to understand how such reactions relate to reason, within the context of Stoic moral psychology. How can affective reactions stand in relation to reason? They can be rational in several senses. (i) They can be the product of assent in a mature rational animal. This reflects the early Stoic theory of action and the passions.29 (2) In a stronger sense, they can be rational in that they are produced by human reason and are correct when judged against the standards of divine reason. Such would be the affective reactions of the sage. (3) They could be rational in the sense that the reaction can be described adequately by the use of 27 28 29
See I n w o o d (1985) 175-81. See de Ira 1.16.7, Ep.Mor. 57.3-6 a n d I n w o o d (1985) p p . 177-8. I cannot accept the argument of Ioppolo (1987) that Zeno and Cleanthes held a different theory, according to which the impulse (which would include affective reactions) precedes and is confirmed by the assent. Her argument rests on three texts, Cicero de Fato 40, Plutarch Adversus Colotem 26 and Seneca Ep.Mor. 113.18. Of these, only the Plutarch text can be said to deal with the pre-Chrysippean Stoa; Cicero's text does not impute Stoic authorship to the doctrine in question, though it does distinguish it from Chrysippus' views; Seneca makes no mention of early Stoic authorship for the doctrine in question, even though he is careful to distinguish Cleanthes' from Chrysippus' views elsewhere in the same letter (section 23). It is only if the Plutarch text clearly attributes to Zeno (against whom Arcesilaus must be reacting) the theory of impulse before assent that we have any reason to assign the theory to the pre-Chrysippean Stoa. But despite Ioppolo's repeated assertions, this text tells us nothing about the causal or chronological order of impulse and assent; it merely explores the polemical suggestion of Arcesilaus that impulse could occur without assent - a doctrine which Arcesilaus' Stoic opponents naturally reject. It is only Ioppolo's suggestion that the claim that assent is not necessary for impulse to occur entails the claim that assent follows impulse. We should recall that where the problem of ineliminable affective reactions is discussed (as at Gellius Nodes Atticae 19.1), Zeno and Chrysippus are treated together as having the same theory.
Seneca and psychological dualism
167
intentional language, by spelling out the upshot of the feelings or reactions in lekta. In this minimal sense, even the actions and affective reactions of non-rational animals are 'rational', as are the desires and urgings of the epithumetic and thumoeidetic parts of the Platonic soul. (4) They could be rational in the sense that they are the result of conscious deliberation (of an Aristotelian sort) such that the agent can always say that he feels as he does for stateable reasons. (5) They could be rational in the sense that good reasons for acting or feeling as one has done can be given after the fact, regardless of whether conscious reasoning of that sort actually caused the action or feeling. No doubt other senses of rationality could be distinguished, but this provisional list should do for now. For my purposes the central question is whether these ineliminable affective reactions are rational!. Sense 4 is clearly inapplicable; sense 5 is possible, but like sense 2 it is disqualified by the fact that Seneca is concerned with affective reactions which are not ultimately justified, reactions which if confirmed would produce the pathos anger. Sense 3, though, is more important. An affective reaction which is non-rational in this sense would be one which just cannot be got into words, which even in principle cannot be spelled out in a form appropriate for analysis and approval by reason. Any reaction which is irrational 3 is truly irrelevant from a moral point of view, even if such reactions can be made morally relevant by their acceptance. Such acceptance would, I suggest, have to take the form of adding propositional content to the reactions so that they can be approved of and so cause a genuine affective response, one which is rational : . To illustrate: a person may have ice-cold water doused on his back while sun-bathing. The shock of such an event inevitably produces an intangible and inexpressible feeling; this feeling is irrational 3 and there is nothing anyone can do about it. But if one supposes that the feeling, or some aspects of it, are or become rational 3 , expressible in lekta, then it is possible to scrutinize the reaction and accept or reject it. Whatever reaction is thereby produced is going to be rational r. In brief, I would hold that strictly orthodox Stoic psychology requires that anything which is rational 3 in an adult human is also rational r - i.e. that people are responsible for all that happens in them which is formulable in lekta. This is, of course, a very strict theory and no ancient evidence
l68
BRAD INWOOD
unmistakably compels us to attribute it to the early Stoics. But it takes proper account of their determination to advance a theory which justifies strict accountability for all of one's morally relevant actions and reactions. The affective reactions discussed in Ep.Mor. 57.3-6 are plausibly held to be irrational 3 , and so even the sage is allowed his ineliminable affective reactions. Those discussed by Gellius at Nodes Atticae 19.1 are similar; in these texts the emphasis is on the merely physical nature of the preliminary affective reaction: 30 Seneca uses the verb sentire and refers to the reaction as a movement; Gellius uses similar language and explicitly denies that any opinion is involved; yet the physical effects on the psychic pneuma have the ability to influence the mind; this is what one would expect in a psychophysical theory - Chrysippus even held that a sage could lose his virtue by the effects of alcohol (Diog. Laer. vii.127). The suggestion that irrational3 reactions are converted into rational 3 reactions (and thus into rational j reactions) is derived from the word prosepidoxazei in Gellius. His account suggests that the propositional content has to be added to the preliminary and irrational 3 experience before acceptance or rejection by reason is possible.31 Now, it is clear that one way in which such irrational 3 feelings can be caused is by hormetic presentations. That presentations, even rational presentations, should have aspects which are irrational 3 seems very plausible - it would be silly to suppose that every aspect of a presentation can be captured in verbal form and that only the verbalizable aspects of such an experience can affect us. The orthodox theory of the passions does not hold that only rational r events can occur in a rational animal, but merely that all rational 3 events are rational^ 32 And it is this orthodox theory which Seneca is 30 31
32
Cf. Ep.Mor. 71.29. Long & Sedley (1987) vol. 2 p . 418 suggest that prosepidoxazei is equivalent to 'assent'; I think it is more likely to be t h e necessary preliminary to assent, viz. t h e addition to a n irrational 3 experience of explicit propositional content. M . Frede claims (1986) that each presentation comes with more t h a n propositional content and that in assenting to the presentation one commits oneself to the acceptance of the non-propositional 'way' of receiving the presentation as well. In discussion at Syam, Frede suggested that this broader and looser 'way' of receiving the presentation was an important part of the hormetic presentation and that this hormetic force of the presentation was responsible for the ineliminable reactions Seneca discusses in de Ira 11.1-4 under the description of primus pulsus etc. But that cannot be quite right, since at 11.3.5 the prima agitatio animi is distinguished from the presentation. Moreover, it seems that Seneca thinks of the primus pulsus or prima agitatio as a kind of horme and not as a component of the presentation. More important for the suggestion I am making here is the apparent fact that every example of the 'way' in which a certain propositional content is thought could also be cashed out in the form of further propositions. This is the crucial fact about the early Stoic
Seneca and psychological dualism
169
working from, extending it and adapting it, though never pushing it over the line into dualism. In book 1 of the de Ira there are several passages of considerable interest as background. Let us look at them to see what philosophical baggage Seneca carries into the passage which concerns us. In 1.3.3-8 an essential Stoic thesis is maintained: the sharp gap between brute beasts and rational animals is made a part of the explanation of human psychology. In understanding anger, Seneca holds that one must remember that only humans can have this passion; any tendency to say that anger might be a useful feeling in small doses, or that it is an ineradicable part of our animal nature and so not subject to strict moral evaluation, should be rejected. What animals have is not anger; anger can only be understood as an expression of reason. This is a crucial theoretical position for Seneca to take, for it ties anger to the activities of reason; just as any phenomenon which occurs in a non-rational animal cannot be anger (on these theoretical grounds), so too any phenomenon which is independent of reason, even in a rational animal, cannot really be anger. When Seneca further specifies criteria for determining what is subject to reason in a person, he will have a basis for distinguishing between anger and pseudo-anger. This he will do in the early chapters of book 11. For now what is most relevant is that the stand he takes here is in every respect an orthodox one. In chapter 7 Seneca's orthodoxy is equally strong: the comparison between the person falling into a passion, such as anger, and the person losing physical control of his motions because of inertia is drawn directly from Chrysippus.33 In 1.7.4 this comparison is used just as it was in Chrysippus, and there is no suggestion of dualism: quarundam rerum initia in nostra potestate sunt, ulteriora nos vi sua rapiunt nee regressum relinquunt. ut in praeceps datis corporibus nullum sui arbitrium est nee resistere morarive deiecta potuerunt, sed consilium omne et paenitentiam inrevocabilis praecipitatio abscidit et non licet eo non pervenire quo non ire licuisset, ita animus, si in iram amorem aliosque se proiecit adfectus, non permittitur reprimere impetum; rapiat ilium oportet et ad imum agat pondus suum et vitiorum natura proclivis.
33
philosophy of mind, that everything which one assents to and so is responsible for is something formidable in lekta, and that nothing not so formidable is of moral significance. Hence I hold that the early Stoic theory is that assent is properly speaking only given to axiomata (Ed. 11.88) and that references to assent being given to presentations are best understood as loose expressions of the theory (see Inwood (1985) pp. 56—9). SyFiii.462. See Inwood (1985) pp. 155 ff. The term 'inertia' is taken from Galen, PHP iv.2.30 (p. 373 K): rhope. See De Lacy's note at p. 643 of his edition of PHP.
170
BRAD INWOOD
But in 1.7.2-3 things may look different:34 we see here very clear examples of the externalization and objectivization of the pathe. But the vivid expression should not lead us to suppose that the reason and the passions are actually distinct; such near-personification need not commit Seneca to any particular doctrinal stance. The external and distinct things which reason opposes are not psychic forces, but, as becomes clear in 1.8.1, the first inritamentum and the semina of anger - i.e. the presentations which tend to move the mind to anger.35 These presentations are experienced as pathos-producing, and so it is easy to treat them as though they were actual pathe, independent psychological states which one might give in to. Note in what follows how Seneca combines a perfectly clear monistic analysis of passion (complete with the voluntary act of reason yielding to the external stimulus) with the apparent reification of the pathos as an external thing admitted into the citadel of the soul: nam si coepit ferre transversos, difficilis ad salutem recursus est, quoniam nihil rationis ubi semel adfectus inductus est iusque illi aliquod voluntate nostra datum est: faciet de cetero quantum volet, non quantum permiseris. Seneca's view is that our reason yields control and authority to the external enemy when it voluntarily accepts the stimulus to passion. When he says that there is no reason left once passion has been introduced into the mind, he does not mean that one thing is displaced by another: rather, the reason we have is transformed into a corrupt form, which is a passion. Our minds are still rational r minds and still function through reason. What has left us is the correct use of reason (rationality2): passion drives out right reason when our unitary minds cease to use reason properly and when they voluntarily (voluntate nostra) give up their authority. Thus 'reason' here means right reason in control, as it often does in Chrysippus too;36 pathos means reason which has turned its back on right reason. But we get the appearance of dualism from the fact that personification and externalization are used. These are effective techniques for vivid presentation, and here we see them unmistakably linked to orthodox monistic psychology. In the next section (1.8.2) this technique is used even more dramatically to hammer home the moral, which is a simple and 34 35 36
Note especially 1.7.3: ratio ipsa, cuifreni traduntur, tarn diu potens est quam diu diducta est ab adfectibus; si miscuit se Mis et inquinavit, non potest continere quos summovere potuisset. The reactions they do cause, and cannot help but cause, are an important subject of 11.2-4. I n w o o d (1985) p . 156.
Seneca and psychological dualism
171
orthodox one: one must nip the proneness to a passion in the bud, so that one does not lose control of one's own mental processes: in primis, inquam, finibus hostis arcendus est; nam cum intravit et portis se intulit, modum a captivis non accipit. neque enim sepositus est animus et extrinsecus speculatur adfectus, ut illos non patiatur ultra quam oportet procedere, sed in adfectum ipse mutatur ideoque non potest utilem illam vim et salutarem proditam iam infirmatamque revocare. non enim, ut dixi, separatas ista sedes suas diductasque habent, sed adfectus et ratio in melius peiusque mutatio animi est. Note the effective use of a military metaphor which externalizes and reifies passion as the enemy. But it is also clear just how much of a distortion the metaphor is, and Seneca says as much with his comments on the structure of the soul: what he here explicitly rejects is the kind of Platonizing psychological dualism which one is so often tempted to lay at the door of Seneca the eclectic.37 Moreover, he goes on to deal with Peripatetic objectors, and it becomes clear in what follows that the externalizing language does more than just make vivid the psychological struggle Seneca is describing. It also gives him a verbal common ground with his opponents and so makes possible the debate. Nevertheless, Seneca is careful to distance himself from the implications of the language. In the sections which follow (through to the end of 1.11) military metaphors play an important polemical role, as well as providing a unifying literary grace. In 1.9.1 the catapult is used as an illustration of the properly controlled use of the soul's proneness to action (impetus): the catapult is effectively used just because the artilleryman in charge of it makes the decision about how much force is to be used in launching the missile.38 The motive power of the soul which is under rational control is not anger, a passion which is by definition disobedient to reason. In section 4 Seneca alters the military imago slightly, concluding that the useful soldier is he who scit parere consilio; adfectus quidem tarn mali ministri quam duces sunt. By
comparing the passion anger to a disobedient soldier Seneca buttresses his position against the Peripatetics; but again, doing so entails a certain reification of the passions and also of the motive power of the soul (impetus). The best sign that this reification should 37 38
For the importance of the theme of psychic transformation, see Inwood (1985) p . 138. The parallel to the runner example is quite interesting, though unexploited here. Just as inertia puts the runner's legs outside his control at a certain point, so too the missile is not recoverable once launched, nor is its trajectory alterable.
172
BRAD INWOOD
not be taken to have strict doctrinal implications is perhaps to be found in 1.11.8, where even virtue itself is reified: like the good soldier, ilia certissima est virtus quae se diu multumque circumspexit et rexit et
ex lento ac destinato provexit. The general is portrayed as being distinct from both his good and bad soldiers. We cannot suppose that virtue itself is to be considered as separate from the active reason of the good man; 39 so it would be unreasonable to find dualistic implications in the metaphorical way in which anger is handled. The general picture which emerges from book 1 is straightforward enough, in my view. The psychological and ethical theses which Seneca wishes to defend seem to be perfectly orthodox and traditional, but the argument against the Peripatetic position, that the passion of anger is good provided that it is moderate, is developed with the aid of an elaborate military comparison which if taken literally would entail some very dualistic-looking views. On balance, though, it seems wrong to attribute to Seneca any proneness to dualism just on the basis of such rhetorically effective devices. 40 Book 11 begins with a warning: up to this point the discourse has been fairly easy, but more difficult topics must now be dealt with. The first of these, and the only one I want to deal with, is described thus: quaerimus enim ira utrum iudicio an impetu incipiat, id est utrum sua sponte moveatur an quemadmodumpleraque quae intra nos <non>*1 insciis
nobis oriuntur ( I I . I . I ) ; and it extends to the end of chapter 4. The question seems a simple one: does anger begin with a judgment (i.e. a rational, and therefore controllable decision) or with mere impulse (horme, for which Seneca's Latin term is the familiar impetus)? What makes the question more difficult is the explication of it introduced by id est: I take it that the arrangement of what follows is chiastic, with sua spontefleshingout mere impulse and the rest of 39
40
41
I n 1.17.2-3 even reason is externalized in this way: nil aliis instrumentis opus est, satis nos instruxit ratione natura. hoc dedit telum,firmum perpetuum obsequens, nee anceps nee quod in dominum remitti posset, non ad providendum tantum, sed ad res gerendas satis est per se ipsa ratio . . . quid quod actiones quoque, in quibus solis opera iracundiae videtur necessaria, multoper se ratiofortior est? nam cum iudicavit aliquidfaciendum, in eo perseverat. Consider too the stance Seneca takes in Ep. Mor. 113 o n the early Stoic thesis that the virtues are animals. Note too how Galen tries to get dualistic conclusions out of the r u n n e r example used by Chrysippus: PHP iv.2.28-38 (pp. 372-375 K ) also iv.5.12 (p. 394 K ) a n d see D e Lacy's commentary on p . 643 of his edition. If even Chrysippus c a n be smeared with the t a r of dualism because of his use of a vivid illustrative image, how much easier is it to treat Seneca this way. This supplement, which clearly makes a major difference to the rest of the discussion, seems inevitable in view of the a r g u m e n t in the rest of 11.1-4. But Holler (1934) p . 16 n. 40 prefers to delete sua a n d avoid the chiasmus.
Seneca and psychological dualism
173
the sentence giving further description of mental events which occur by rational decision. If this is so, then the distinguishing mark of merely impulsive mental events is their causal autonomy; they happen 'of themselves'; and the mark of 'decisions' is the fact that they happen 'in the manner of most mental events which occur non insciis nobis\ accompanied by 'our' awareness. The 'F must be reason, which is at the core of personal identity for any Stoic. We do not, then, have a tidily exclusive set of properties to describe the main contrast between impulsive and decided action. The most I would want to conclude from this introductory sentence is this: to preserve consistency with the general thrust of book 1, Seneca should clearly opt for the second choice; the spontaneous and possibly unselfconscious character of impulsive mental events runs counter to the whole Stoic account of anger in book 1. In 11.1.3 Seneca moves on to a statement of the uncontroversial: the first stimulus of anger is a perception of unfairness, a presentation {species) oiiniuria. Hence the incipiat of I I . I . I is not referring to the first stimulus, but to the first decisive internal component of the complex event known as anger. The contrast of impulsive and decided action is redescribed again: does anger follow immediately on the presentation once it is received, without the consent of the mind (non accedente animo), or is it set in motion by the mind's assent, i.e. is it rational^ Seneca follows the Stoic view (nobis placet: 11.1.4), that anger does not 'dare' to do anything on its own but needs the approval of the animus*2 The reasons he gives for accepting the orthodox view are interesting - and it is important to see that at this point Seneca is giving reasons for staying with the orthodox view, not simply stating that view. There is, in Seneca's opinion, an impulse which is set going without our voluntas, but it is not able to do the kind of processing and quasi-reasoning which is characteristic of a genuine anger-response: for in anger we receive the presentation that we 42
It seems that Mam must take ira as its antecedent, rather than species; so once again we may note the reification of a mental state. Contrast the comment of Holler (1934) pp. 21-2; for
him, this treatment of anger is ieine der wenigen Stellen in de ira, aus denen klar hervorgeht, dafi Seneca ein besonderes thumoeides . . . neben der animus . . . annimmt\ This, he says, is something new: 'halb altstoisch, halb peripatetisch-akademisch.'1 And he asks, 'Sollte diese Synthese Senecas eigenes Werk sein? Oder ist sie die unbekannte mittelstoische aus dem Suntagma peri orges des
PoseidoniosV Once one acknowledges the importance of Seneca's use of literary techniques, this sort of fancy becomes impossible. Time and again Holler loses his grip on good judgment because of his simplistically doxographical pursuit of Posidonian phantoms.
174
BRAD INWOOD
have been treated unfairly and then long for vengeance and then link together the two views: that we ought not to have been harmed and that we ought to fight back. This is a complicated impulse indeed, whereas the other sort, the sort which can occur without our voluntas, is simple.43 The complexity of cognitive structures which the anger-response requires, including understanding, resentment, adverse moral judgment, revenge, simply cannot occur without the assent of the mind to the experiences and thoughts by which it is affected. The contrast is pursued in the next section (11.2). After clarifying what is at stake in this theoretical debate, namely that if anger can occur without our 'will' it will be immune to rational control (i.e. will be irrational^, Seneca gives a wealth of examples of each kind of mental event. The general principle at issue is clearly stated: omnes enim motus qui non voluntate nostrajiunt invicti et inevitabiles sunt.^ He will
allow that there are reactions which are not voluntary, but not that any such reaction could count as genuine anger. To do otherwise would conflict with the stance taken by the early Stoics and reaffirmed in book 1. These 'inevitable' reactions are interesting. Shivering when splashed with cold water and a creepy feeling when touching certain objects seem to me to be strictly bodily reactions. But the bristling of our hair when frightened by bad news and blushing at improper language - these reactions clearly involve considerable cognitive and social sophistication. Feeling dizzy when looking down from a height seems to fall between the two extremes. A physiological fear reaction to hearing (and understanding) bad news and blushing seem to be rational phenomena; recall Mark Twain's observation that man is the only animal which blushes, or needs to. 45 Clearly Seneca's criterion for 'rationality' z in the reaction to potentially passionate stimuli is not that reconstructed for the earlier Stoics, i.e. expressibility in or reliance on lekta (rationality 3 ). 46 His 43
44
45
46
T h e simple impetus is n o t impulse proper; note that in 11.3.4 impetus without qualification is used of a reaction which is always accompanied by assent. This reflects early Stoic use of the term. This is compatible with t h e view attributed to Zeno in 1.16.7, that even a sage will have some faint traces of passions - these being inevitable a n d i m m u n e to t h e perfect rational control of a sage - b u t no passions. These examples from de Ira 11 a r e worth c o m p a r i n g to those mentioned at Ep.Mor. 57.4; even the sage has certain reactions in his soul which a r e n o t pathe: non est hoc timor, sed naturalis adfectio inexpugnabilis rationi. Cf. de Ira 1.16.7. In wood (1985) p p . 66 ff.
Seneca and psychological dualism
175
own criterion is much narrower and, to our surprise, leaves many clearly cognitive and verbally expressible phenomena on the other side of the line, as irrational or pre-rational reactions. For Seneca the criterion appears to be susceptibility to change by a conscious rational decision. And one might reasonably conclude that Seneca thinks that the complexity he is describing here gives a better foothold for conscious control than does mere expressibility in lekta. And conscious control is evidently quite important to Seneca. It follows that 'being in our power' is used in a different sense from what we are used to in earlier Stoic discussions. Quorum quia nihil in nostrapotestate est, nulla quominusfiant ratiopersuadet (11.2.1). W h e n a n early Stoic talks of things being ephy hemin he does not have such a narrow notion in mind. 47 There is no reason to believe that orthodox Stoics ever propounded the sort of distinction Seneca is outlining here between two kinds of impulse (simple and complex). 48 Of course, they may have done so, or some of them may have at some time; but the relatively tidy reconstruction which makes the most of our scattered sources does not suggest it. Hence my suggestion here is that Seneca himself is offering the distinction in an attempt to justify his retention of the essential element of the early Stoic analysis of anger, i.e. the claim that the passion is rational!. But does this innovation commit Seneca to any sort of psychological dualism? If we look closely at what Seneca says in these chapters, 47 48
Inwood (1985) p p . 8 9 - 9 1 , Appendix 4; Sorabji (1980) p p . 86, 245, 252. However, the idea of there being a n impulse before assent, the assent being a confirmation for the preliminary impulse, does occur in two places in the Latin tradition: Cicero de Fato 40 and Ep.Mor. 113. I n the de Fato the theory of an adpetitus preceding assent is attributed to anonymous philosophers, neither Stoics n o r Antiochean veteres, in the context of a polemic about choice a n d necessity. I n Ep.Mor. 113 Seneca is discussing the Stoic thesis that the virtues are animals, expounding arguments pro a n d con, b u t essentially dismissing the idea as ridiculous. I n 113.18, part of a n a r g u m e n t against the thesis, he presents a n analysis of rational action involving impulse preceding assent a n d confirmed b y it - t h e point being that any rational animal must be capable of this sort of action a n d that the virtues are not. Earlier in t h e letter (113.2) Seneca uses the inapplicability of action to the virtues for t h e same purpose. Here Seneca accepts a n analysis of action not attested for the earlier Stoics in order to refute a n early Stoic theory - a n i m p o r t a n t indication of his independence of mind. A n d yet the general result of his theory is still very m u c h in line with Stoic doctrine. W h e n we shift o u r minds to the de Ira, it is clear that the theory of Ep.Mor. 113.18 is compatible with it, though the de Ira text is much less clear-cut and neat. It is h a r d to tell whether the de Fato text presents the same theory, or whether it is meant to present a theory at all. T h e r e may be some hint of the origin of Seneca's theory of impulse before assent in Cicero's text, but I d o u b t it; certainly nothing in Cicero suggests that we tie it to Posidonius or to a n y form of psychological dualism. For Adversus Colotem 1122D-C, see Inwood (1985) p p . 8 7 - 8 and above n. 29.
176
BRAD INWOOD
we do not find him indirectly positing a distinct and non-rational part of the soul. The soul he describes is unitary; the complexity which Seneca introduces centres not on parts or powers, but rather on a temporally dynamic process. Chrysippus (SFFm.462) described a passion by means of a comparison with the body in motion, so over-committed to its trajectory by its own inertia that it cannot regain self-control and balance; and there are other hints in early Stoicism of what we might call psychological inertia, all readily explicable within the framework of monistic psychology. 49 In book 1 Seneca accepted that comparison and made it his own. In book 11 Seneca goes beyond the Chrysippean theory, but he does so by the enrichment of this essentially Chrysippean insight and by a reasonable reappraisal of the concept of assent. Assent now involves conscious rational control: implicit assent is not good enough for moral improvement - it is probably significant that Seneca shows little interest in the problem of determinism and responsibility which provides the most congenial focus for implicit assent in earlier Stoicism. And the idea of a preliminary, almost mechanical response which is unstoppable and so not subject to the strict moral appraisal which attaches to assented actions grows smoothly out of the idea of psychological inertia already found in early Stoicism. The criterion which Seneca uses to draw a distinction between a passion and a non-passion is rationality in a new sense, not rationality!, but voluntariness in his own sense. What he asks is: can this mental event be stopped or altered by our rational powers if we consciously suspend the process and apply critical scrutiny to the presentation which sets it in motion? In his view, anything not subject to this procedure is in an importantly different moral category. Thus in 11.2.2 he says that the 'voluntary' vices of the soul are those which can be eliminated by the application of moral precepts. Those which come with the human condition and which no one can avoid are not subject to any conscious control and so are, in the relevant sense, involuntary. 50 Involuntary vices, though they are not discussed here, would seem to be less grave and certainly less amenable to immediate action by the agent. Character flaws which do not yield to the assertion of will would, on this theory, be exempt from a significant kind of moral blame. 49 50
See Inwood (1985) p p . 146 ff. on the 'fresh opinion' a n d the fading of passion with time. This latter class also afflicts the sage; see de Ira 1.16.7, a n d discussion above.
Seneca and psychological dualism
177
That is why the most important thing to Seneca here is not what I think Chrysippus would have called the mere rationalityr of the presentation and our response to it, but whether it is consciously controllable. Hence the primus ictus animi51 which affects us after we get an opinionem iniuriae can be put in the class of involuntary things, even if it is only possible through what I would have to describe as rational mental processes, since they require lekta. The linkage between rationality3 and rationality! is being weakened, but the linkage between rationality r and morally significant actions is not.52 For the psychology of action the most important non-passions in 11.2 are the preliminary passions, the emotional responses which if left unchecked will actually lead to a full blown passion. But from our point of view the most interesting application of this distinction comes in what follows (11.2.3-6). Here Seneca exempts our emotional responses (for so they must be called) to literary and dramatic experience from voluntariness and so from moral judgment, apparently (and justifiably) on the grounds that the artificiality of the experience causes us to suspend our critical faculties. If we read about moving historical events or see them on the stage, or if the musical accompaniment stirs passions in our souls, or even if a painting with grim subject matter upsets us, still our reactions are not judged in the way that reactions to events in real life are judged. This sort of detachment is, it seems to Seneca, what makes possible that imaginative sympathy with the feelings of portrayed characters which is such an important part of artistic experience (11.2.5). In his view these emotional reactions are no more passions than is the emotion of the actor in expressing himself or the frisson of fear which a reader feels in reading of the impending Hannibalic attack. From a moral point of view they have more in common with shivers and involuntary blushes. quae non sunt irae, non magis quam tristitia est quae ad conspectum mimici naufragii contrahit frontem, non magis quam timor qui Hannibale post Cannas moenia circumsidente lectorum percurrit animos, sed omnia ista motus sunt animorum moveri nolentium, nee adfectus sed principia proludentia adfectibus. sic enim militaris viri in media pace iam togati 51 52
By this point Seneca is becoming reluctant to describe this non-rational emotional response as a n impulse. By 11.3.4 he will have a b a n d o n e d that description. Similarly the reactions of the sage described in de Ira 1.16.7 seem to involve rational 3 mental states b u t not rational , assent. This suggests that the doctrine in both books is meant to be the same a n d that its orthodoxy focusses on the connection between rationality, a n d actions or passions, rather t h a n on the connection between rationality 3 a n d rationality,.
178
BRAD INWOOD
aures tuba suscitat equosque castrenses erigit crepitus armorum. Alexandrum aiunt Xenophanto canente manum ad arma misisse (11. 2.5-6). Clearly the theory of aesthetic emotional response which Seneca is invoking here is of considerable interest and might bear comparison with Aristotelian views on tragedy (especially Poetics i455a3O-~4) and with Horace's claims in the Ars Poetica (101-5). But what is curious and crucial for us is the originality of Seneca's psychological position. He has grouped together a striking range of phenomena under one description: what they all have in common and what justifies grouping them together53 is that they do not involve what Seneca would consider to be voluntary commitment: they are not the sort of emotional responses which it is reasonable to expect will be brought before the court of reason for approval or rejection.54 The range of these phenomena is remarkable: mere bodily shivers, engaging only the body or the vegetative level of the soul; unavoidable responses to sudden bad news or improper behaviour; vertigo; the emotional response one tends to feel in performing or viewing drama, or in reading history or fiction, or in viewing powerful graphic art, or in our emotional responses to music and sound effects; he is even willing to include the involuntary response to martial stimuli of soldiers, even in peacetime. These phenomena are all 'involuntary' in a reasonable sense of the word. But it is striking to find a Stoic willing to take them out of the ambit of the passions. At least some of these phenomena express character very clearly; and as expressions of character they are highly relevant to passions and to the moral evaluation of that character. In earlier Stoicism as normally understood we would not 53
54
N o t all of these should be called preliminary passions, since so m a n y of these responses a r e not exempted o n the grounds that they are too early in their development to merit description as passions; many are privileged by their context a n d would never develop into passions. Some might well be preliminaries to passions, others are merely analogues to them, a n d others are merely bodily. See Holler (1934) p p . 18-19 f ° r m s summary of what these phenomena have in common. H e seems to find their grouping less remarkable than I do. W h a t they have in common, according to Holler, is (1) that even the sage is exposed to them, (2) that they are stimulated directly by the phantasia, (3) that they are passive not active, a n d (4) that they are only treatable by habituation and, in short, belong to the irrational part of man. His fourth point is the most problematic. For habituation will have little or no effect on the strictly bodily responses. Moreover, there is no attribution of these phenomena by Seneca to a n irrational part of man; that is a contribution from Holler's own theory. N o r does Holler distinguish here between the irrational body a n d the irrational part of the soul, a fuzziness which helps him in his justification of Seneca's grouping of non-rational responses.
Seneca and psychological dualism
179
find all these reactions grouped with morally irrelevant bodily shivers; there, the criterion of moral relevance lies in the relation of action and thought to character. Moreover, in early Stoicism we find no insistence that the assent given to a stimulus must be voluntary in our and Seneca's sense - that is, a consciously chosen response55 to a stimulus. For earlier Stoics it is enough if it is a response to a rational 3 stimulus in a rational animal. 56 In 11.3 the demarcation between passion and action is revisited: the stimuli which do not entail a passion are described as happening fortuito, i.e. non-voluntarily. When subject to these, the soul does not act - and that is what is required for a passion to occur. The emotional response which no one can help is afortuitus motus, and simply having that is not very important. What matters morally is yielding to that and confirming it by a voluntary judgment and assent. Again, the examples of involuntary responses are strikingly different from what one would expect of the old Stoa: paleness, tears, sexual arousal, deep sighs, excited gaze (11.3.2), involuntary manifestations of fear or nerves even in a brave soldier, general or accomplished orator (11.3.3). Some of these Seneca is surely right to describe as purely bodily, and so even an early Stoic would be happy dismissing them as non-passionate. But others strike us too much as character-related and cognitively significant to write off as merely bodily. Yet that is just what Seneca does in 11.3.2: fallitur nee intellegit corporis hos esse pulsus.57
Seneca's position is that anger is different from all of these because it is a deliberate acceptance and confirmation of a starting point which is of this character; it is not just a response (moveri) but an excessive one, a runaway response (excurrere) - with this Seneca is once again invoking the Chrysippean idea of'psychological inertia'. That is because anger is a genuine impetus (Jiorme) and so is dependent on assent - as the first or primitive emotional/bodily response (now called an agitatio, not an impulse) is not. And since it needs assent, it must come under conscious mental control (11.3.4). For the soul would be aware of the sort of activity needed for anger: the sort of quasi-inference we met above. Here Seneca is integrating his new criterion of conscious voluntariness with the earlier Stoic idea of assent. At 11.3.4-5 anger is presented as a response of the reason to 55 56
T h o u g h we start to see this in the phrase tes sunkatatheseos kath' hormen ouses, which occurs in some of our sources for early Stoicism (see I n w o o d (1985) A p p e n d i x 4 ) . 57 I n w o o d (1985) ch. 3. Cf. Ep.Mor. 71.29.
l8o
BRAD INWOOD
the soul's prior response to a stimulus, and that prior response is no more anger than the stimulus itself [species) is:58 ira non moveri tantum debet sed excurrere; est enim impetus; numquam autem impetus sine adsensu mentis est, neque enim fieri potest ut de ultione et poena agatur animo nesciente. putavit se aliquis laesum, voluit ulcisci, dissuadente aliqua causa statim resedit: hanc iram non voco, motum animi rationi parentem: ilia est ira quae rationem transiliit, quae secum rapit. ergo prima ilia agitatio animi quam species iniuriae incussit non magis ira est quam ipsa iniuriae species; ille sequens impetus, qui speciem iniuriae non tantum accepit sed adprobavit, ira est, concitatio animi ad ultionem voluntate et iudicio pergentis. numquam dubium est quin timor fugam habeat, ira impetum;59 vide ergo an putes aliquid sine adsensu mentis aut peti posse aut caveri. Anger occurs by assent, by which Seneca means a consciously voluntary concession to a pre-existing emotional or at any rate non-responsible reaction. In ii.4 Seneca summarizes his causal account of anger, and distinguishes three distinct motus or psychological reactions. First is the primus motus non voluntarius, the proto-passion which Seneca has been carefully isolating. Second is the conscious judgment that one ought to act, which he calls assent. It is cast in Stoic form as a hormetic judgment (the alter motus) mentioned in 11.4.1). The uncontrolled form of this (tertius motus is the passion of anger. This passage also gives a justification for regarding the primus motus as non-rational and involuntary, a justification which can be put into the form of an inference. (1) The first response, Seneca says, cannot be changed by judgment. (2) But what is produced by judgment can be cured by it (11.4.1—2): tertius motus est iam impotens, qui non si oportet ulcisci vult sed utique, qui rationem evicit. primum ilium animi ictum effugere ratione non possumus, sicut ne ilia quidem quae diximus accidere corporibus, ne nos oscitatio aliena sollicitet, ne oculi ad intentationem subitam digitorum comprimantur: ista non potest ratio vincere, consuetudo fortasse et adsidua observatio extenuat. alter ille motus qui iudicio nascitur iudicio tollitur. 58
59
See above on the idea of a n impulse before assent, which is confirmed or rejected by it. While Seneca has no hesitation about calling this preliminary response an impetus in Ep. Mor. 113, in this text he shows a more proper a n d orthodox caution: he at first calls this response an impulse, a n d then withdraws the label in favour of something quite vague. I n de Fato 40 the impulse before assent is called a n appetitus. At this point impetum is not a translation of 6puf| b u t means 'attack' (compare/ttgam above); I thank D. Konstan for this observation.
Seneca and psychological dualism
181
(3) Therefore the first response must be classified as non-rational. And this will hold good just as long as Seneca maintains (1), i.e. that the emotional responses put into the class of simplex impetus or primus motus are totally immune to change by rational judgment, and so resemble bodily events more closely than they do rational responses. And it seems that Seneca will do that just as long as he maintains that 'will' rather than expressibility in lekta is the crucial mark of rationality. This is the most spare and clear elaboration of the theory Seneca has been discussing. The three motus are distinct: the relationship between the first and third cannot entail dualism because reason has nothing at all to do with the first. And while the third does oppose reason [rationem evicit), it is itself a product of judgment and the reason it defeats is not the agent's own, but the right reason which is abandoned by anyone who yields to a passion (i.e. it is rational but irrational 2 ). There is novelty in Seneca's theory. But as far as I can see there is nothing which commits him to psychological dualism or which need owe anything to Platonic or Posidonian influence. The goal of the discussion is the maintenance of Stoic theses in ethics, and the changes have considerable independent interest and merit. We might ourselves want to pursue Seneca's obvious interest in widening the class of emotional experience which is exempt from the strictest moral relevance; that is something anyone concerned about the moralistic constraints of earlier Stoicism would want to do. We may also want to consider the merits of his decision to alter the concept of assent; he does so by linking it to his own notion of voluntas, which is a fruitful and original contribution to the development of the Stoic tradition. 60 But this is not the place for a full consideration of Seneca's contributions. What we have to consider in closing is one last doubt: is there not one respect in which we might detect the cloven hoof of eclecticism? If one thing, reason, is in a position to sit in judgment on something else, which has a kind of life of its own in the soul, do we not find ourselves on the doorstep of dualism? 60
See Kahn (1988), esp. 251-5 on Epictetus and Seneca. Kahn offers a larger view of the evolution of the notion of'will' than I could hope to do, and he puts particular stress on the importance of the shift from early to late Stoicism (typified by Epictetus and Seneca). But he does not provide an account of the philosophical motivations Seneca may have had for the shift towards 'the intense preoccupation with the inner life' (p. 253); this description is,
l82
BRAD INWOOD
The answer I give is obviously no, and for familiar reasons. The tendency to speak of reason as distinct from what it is judging is part of the literary apparatus Seneca uses throughout: he is reifying the psychological forces which are the building blocks of his theory. He is no more committed to dualism by this device here than he is anywhere else. The temporally dynamic process which Seneca invokes owes much to Chrysippus' comparison with the runner, which is used effectively in the orthodox theory. When appraising Seneca's orthodoxy, we should pause to reflect that one of the implications of Seneca's runner example - as an illustration - is, as Galen points out with some satisfaction,61 that there is a difference between the self which tries to check the inertial drift of the bodily movement and the body itself; it suggests a distinction between reason and the rest of the soul; indeed, Galen insists that the example proves that Chrysippus is committed to accepting not just a separate, but even an irrational part of the soul. One suggestion which I make is that part of Seneca's innovation is rooted in Chrysippean ideas of psychological inertia, ideas which Chrysippus would have wanted to claim were compatible with his fully rationalistic monism. Seneca, I think, develops some of these ideas in his own way, and certainly drops the commitment to a fully rationalistic account of such experiences. But he does not thereby drop his commitment to monism. What follows for Seneca is that his theory is no more dualistic than Chrysippus'.62 If, as is so often alleged, we are on the doorstep of dualism with Seneca, then we were already there with Chrysippus. It is quite possible that we should be re-examining the early Stoic theory for its adequacy; it may, as Galen alleged, contain internal inconsistencies which undermine its integrity (though I personally
61 62
I think, meant to refer to what I see as Seneca's interest in consciously controllable assent and his reduced interest in overall character. See above n. 40. T h e same might be said about the use of medical parallels in de Ira m. 10; for these are well documented in the early Stoa, a n d it is h a r d to see that Seneca's use of them is unorthodox. Holler, however (1934) p p . 23-4, takes the language of 111.10, with its emphatic use of personification and reification, as being paradigmatically dualistic, so m u c h so that he uses it to cast d o u b t on the monistic character of 1.8, with its clear enunciation of monism (1.8.3, see above). As usual, the main effect of Holler's technique is twofold: Seneca becomes self-contradictory a n d his internal contradictions become crucial evidence in the quest for Posidonian doctrine. Seneca's integrity as a philosopher, even as a writer, is readily sacrificed on the altar of Quellenforschung.
Seneca and psychological dualism
183
do not think that it does). But if it does, then Seneca will naturally reflect that situation; and the resulting theory will then not properly be described as eclectic or Platonizing: we will simply have discovered more about the Platonic roots of Stoic psychology.
CHAPTER 7
Actions and passions: affection, emotion, and moral self-management in Galen's philosophical psychology James Hankinson
Galen is remembered mainly as a doctor; and it is on his fame as a medical theorist and practitioner that his reputation principally rests. Yet he was no mere physician; he thought of himself as equally a philosopher,1 and if the tradition has been less kind to his strictly philosophical works than to his medical writings (here as elsewhere interest has to a large extent dictated survival), none the less enough This paper has benefited greatly from the helpful comments supplied by friends and colleagues both at the Symposium and elsewhere; I should like to thank Julia Annas, Michael Frede, David Gill, David Sedley, and Dolores Iorizzo for useful remarks; my colleague Bob Kane for a helpful discussion of the intricacies of freedom and responsibility; and in particular Jonathan Barnes for his prompt and incisive editorial remarks. Note on citations. Referring to Galen is problematic. The only edition of the greater part of his work done in modern times is that of Kiihn (1821-33, re-issued 1965). Kiihn's work is invaluable; but it was done in a hurry, with the needs of doctors rather than scholars in mind and is consequently riddled with errors of both a trivial and a substantial kind. In some cases later and better texts of Galen exist, particularly in the Teubner series (notably the three-volume Galeni Scripta Minora [SM], edited by Marquardt, Helmreich and Miiller, Leipzig, 1884-93), a n d the Corpus Medicorum Graecorum (CMG: Berlin): the latter was begun with high hopes and comprehensive aims at the end of the last century, and work proceeded relatively rapidly until the First World War; but since then publication has slowed to a trickle (although the stream shows recent signs of revival). However, it has seemed best for brevity's sake to refer principally to Kiihn even where later and better editions are available, since they all cross-refer to it. At the first mention of any Galenic text, I give its full title, followed by an abbreviation which is standard for the rest of the article (there is as yet, unfortunately, no general orthodoxy in these matters), followed by a Roman numeral indicating Kiihn volume, and an Arabic numeral for the page; and, also on its first appearance, I refer to the later, better editions of the work, if any exist. 1 In the autobiographical parts of his work, he emphasizes the catholicity of his philosophical education. He imbibed Platonism under a pupil of Gaius: see On the Diagnosis and Cure of the Affections oj the Soul v.41 (AffDig., = SM 1 1-44 Marquardt [1884], = CMG v.4.1 1 De Boer [1937]; Aff.Dig. is translated into English in Harkins & Riese, [1963]). Galen studied at Smyrna under another important Platonist, Albinus: de Libris Propriis (Lib.Prop., = SM 2 91-124 Helmreich [1891]) xix.i6fT. In Rome he hobnobbed with Peripatetics: Eudemus, upon whom he performed a famous cure early in his Roman career: see de Praenotione (Praen., = CMG v.8.1, Nutton [1977]) xiv.605-19; and Alexander of Damascus, with whom he had a not untypically violent quarrel (ibid. 627-9): see de Anatomicis Administrationibus (AA) 11.217-18. As a boy, he learned Stoic logic from a Chrysippean: de Ordine Librorum Propiorum (Ord.Lib.Prop., = SM 2 80-90 Helmreich [1891]) xix.43.
184
Actions and passions
185
of his philosophical temper shines through his medical oeuvreforus to be able to judge his claims to that title. For Galen, fortunately (although at times infuriatingly), was not a dry and merely technical writer. On the contrary, his massive corpus abounds with digressions, asides, interludes; irrelevant hares are started, pointless chimerae chased. If Galen insists (as he does to the point of tedium) that his wordiness is the direct result of the idiocy of those whom he has to refute, he himself cannot be acquitted of the charge of prolixity and verbosity; even when hymning the philosophical and literary advantages of brevity, he does so at extravagant length. Galen loved the sound of his own voice. And it is lucky for us that he did. For it is in those asides and irrelevances that we find a pot-pourri of scattered remarks and fragments of enormous philosophical importance. It is to Galen's exuberant largesse that we owe much of our knowledge of the schools of later Greek philosophy; and that debt has long been acknowledged. But it is less frequently recognized just how important and incisive a thinker Galen was in his own right; and it is to the originality of his contributions to the field of philosophical and moral psychology that this paper is principally devoted. The range of Galen's philosophical interests was wide; he wrote on logic2 and epistemology,3 on proof,4 on the nature of science;5 he elaborated a powerful teleological conception of the structure of the natural world;6 he developed a sophisticated account of causation and explanation; 7 and his work in the philosophy of science led him to sketch a tantalizing theory of individuation, natural kinds, properties and reference.8 However, it is with his work in ethics and the philosophy of mind that we shall be mainly concerned. Elsewhere I have analyzed Galen's philosophical and physiological account of the structure and functions of the soul; 9 its principal purpose was to show how Galen manages to develop, by 2 3 5
6 7 8 9
See Barnes (forthcoming). Barnes' paper deals with all aspects of Galen's philosophical view of science; and see also Barnes (1985b) and Hankinson (forthcoming 3). 4 Frede (1981). See Barnes (forthcoming). See Barnes (forthcoming). See also my translation and commentary of the first two books of Galen's De Methodo Medendi (MM: x 1-1024), Hankinson (1991a); Hankinson (forthcoming 1) and Hankinson (1991c). See Hankinson (1988a) and Hankinson (1989). See Hankinson (forthcoming 4) and (forthcoming 5) and Hankinson (1988b). See Barnes (forthcoming) and Hankinson (1991a, 1991c, forthcoming 1,3). Hankinson (1991b).
l86
JAMES HANKINSON
judicious selection from a variety of sources,10 a powerful analysis of the physiological bases for psychological activity. He marries a methodology drawn ultimately from Aristotle (one starts with a description of the functional phenomena, and then proceeds to investigate their causal bases), 11 but given an injection of empirical legitimacy by Galen's own deep conviction of the necessity for persistent anatomical inquiry. It is anatomy that will uncover and unlock the mysteries of the transmission of neural influence; and will serve to show the locations of the various psychic powers (rational in the brain, spirited in the heart, desiderative in the liver: the account is, substantially, Plato's of the Timaeus supported by detailed empirical investigation). I further suggested that Galen's wedding of philosophy and physiology, of conceptual analysis and neuroscience, was not simply a marriage of convenience, but rather the outcome of a sophisticated and self-conscious reflection on the proper nature of any scientifically and phenomenologically satisfying account of the mind; that Galen, in effect, was engaged in the enterprise that Fodor has called 'speculative psychology'.12 He attempts, then, to chart a via media between a purely empirical approach to psychology (which runs the risk of conceptual sterility), and an excessively aprioristic and rationalistic philosophy of the mind carried through with no regard for empirical adequacy. Naturally, in order to discover a Galen of this sophisticated sort, one needs both to show him engaged in neuro-science (or a recognisable antecedent of it) at a relatively high level, and to give an account of the way in which he tries to integrate this with his picture of the phenomenology of mental life, and with an overall analysis of the nature of psychological causation and of its relation with the physical. The other paper tried to accomplish the first of these tasks, discerning a broadly (but cautiously) materialist Galen, committed 10
11
12
His principal acknowledged sources are Hippocrates in medicine and Plato in philosophy: indeed, he wrote On the Doctrines of Hippocrates and Plato (PHP = CMG v.4.1.2, De Lacy [1978-84]) to demonstrate their agreement on all major issues, pre-eminently the nature of the soul; but he also owes a debt to the work of others, notably Aristotle and the eclectic Stoic Posidonius. PHP was also edited for CMG by von Miiller (Leipzig, 1874); von Miiller's text is a huge advance on Kiihn, but is no substitute for the magnificent new CMG edition of De Lacy (3 vols., including English translation, commentary and indices) 1978-84; 'De Lacy, 1984' refers to the third volume of commentary and indices. For Aristotle, see PA 1.1, 639b7ff.; 640a i4ff.; for Galen, see On the Natural Faculties (Nat.Fac.) 11.9-10; On the Substance of the Natural Faculties (Sub st.Nat.Fac.) iv.760. This methodology perhaps found its way into the medical tradition via Herophilus, the great third-century BC Alexandrian physician: see von Staden (1989) pp. 1 i8ff. See Fodor (1975), p. vii.
Actions and passions
18 7
to a functionalist account of psychology coupled with the claim that psychological events were (in some sense left perhaps deliberately vague) 13 dependent upon physiological ones. In this study, I try to undertake the second of those projects, and to show how his views on psycho-physical relations, and on the nature of causation, can be reconciled with his account of human mental life, human responsibility, and our status as agents. The starting-point will be Galen's contribution to the protracted Hellenistic and Later Greek debate on the status and management of the emotions, and of the proper attitude to be taken towards them. I THE STATUS OF THE EMOTIONS
Emotions, for the Greeks, arepathe, affections: things that happen to us. Their very name suggests that they are somehow beyond our control - and the literary history of eros, the most powerful emotion of them all, tells strongly in favour of this cultural donne: it is an alien force, anikate machan, according to Sophocles (Antigone 78iff.); a compulsive, uncontrollable force, in the plays of Euripides (notably Medea: see e.g. 530-1; but see also Hippolytus 443);14 a sort of madness according to Plato (Phaedrus 244a ff.). In Gorgias' Defence of Helen (sections 15-19), it is something so outside an agent's control that it can absolve them of all responsibility for their actions;15 and 13 14
15
See Lloyd (1988) esp. pp. 33-9; see also Hankinson 1991b. Euripides' Medea became in Hellenistic times perhaps the principal post-Homeric literary quarry from which philosophers mined their examples: Chrysippus was so prone to this that one work of his was referred to ironically as 'the Medea of Chrysippus'; and Herillus of Carthage even wrote a Medea (Diog.Laer. vii.166, = SVF 111.409), although we know nothing of its contents. At PHP v.403-24, Galen discusses a sequence of quotations made by Chrysippus in his On Passions from various plays of Euripides: Alcestis 1079-80 (413); 1085 (419), Andromache (629-30: 405), and Electra (15-6: 423) as well as Medea (1078-9: 408, also quoted at 307 and mentioned at 372), in addition to citations from the Dictys, Oeneus, Phrixus, and Stheneboea, which have not survived; the common thread to the discussion is that the passages chosen by Chrysippus are as a matter of fact inconsistent with his account of a unified soul, and indicate clearly that the affections are separate and distinct from the reason. In particular, Galen notes that the Chrysippean doctrine of passion being identical with judgment is difficult to make consistent with the evident fact that the judgment in accordance with which the passion was originally aroused can remain quite unaltered long after the passion itself has subsided (the anesis tes lupes: 419, = SVF in.466) - Chrysippus' account of passion as a 'fresh opinion' (doxa prosphatos: 416—7, = SVF HI.481) simply will not do in Galen's view: 416-24; for a discussion of this issue, and a defence of the Chrysippean position, see Nussbaum (1987) pp. 144-58; and see also Gill (1983). On the concept of distress (lupe), see n. 80. Fr. 82.B.11 DK: in spite of the fact that Gorgias himself styles the Helen a paignion, or rhetorical plaything, it seems reasonable to conclude that nevertheless it embodies elements of a serious contemporary debate on the issue.
l88
JAMES HANKINSON
Aristotle feels constrained to argue against the view that 'acts due to appetite or anger' 16 can be removed from the sphere of the voluntary, acutely aware that any such concession to the mitigating force of the passions will threaten the coherence of the entire structure of his account of voluntary action and responsibility (Nicomachean Ethics in. i, 1111 a22fif.; cf. 111 obgff). Galen spends much time attacking the Stoic (Chrysippean) conception of the structure of the soul; and he hopes to demonstrate thereby that the Stoic notion that the emotions are simply excessive alterations of the reasoning part of the soul is false (PHP v.26gff.: = SVF n.887ff., 2g4ff.). No amount of poetic witness will serve to establish Chrysippus' claim; but more importantly it is simply false to hold, as Chrysippus does, that we can feel thought taking place in the chest. We are perhaps aware of anger and fear as involving localized strong sensations in that region (indeed, much of the evidence of the poets goes to suggest that) - and that perhaps justifies the belief that anger and the other emotions have a cardiac location (270-1); but no special motion of the heart is evident in the act of thinking, learning and teaching, even though such a motion is clearly apparent in the case of the emotions [pathe].11 (Pi/Pv.271) And Galen proceeds to demonstrate the truth of his contentions at great length, forced to do so, he tells us, by his opponents' prolixity and obtuseness, and by their habit of invoking arguments that are ... completely worthless ... introducing the testimony of women, non-experts, etymologies, motions of the hands, upward or downward movement of the head, or poets. (PHP v.327 = SVF 11.884)18 16 17 18
For anger as a source of actions outside the agent's control, see Medea 1078-9, and the passages in which Galen discusses Chrysippus' discussion of it cited in n. 14. De Lacy translates 'pathos' neutrally as 'affection' - but the denotation is clear enough here. Women, according to Chrysippus, point to their chests and say 'that doesn't sit well here' when they disagree with something {PHP v.323: = SVF 11.892): this presumably is the reference to 'the motions of the hands' (cf. 224-5); on the futility of etymologies as a guide to facts, see 213-18 ( = SVF 11.883, 895), anc * the case of 'ego*: when you pronounce it, your jawbone points to your chest, indicating, according to Chrysippus, the cardiac location of the personality (cf. 224-5, 227-8, 327-9: = SVF n.896). On non-experts and poets, see 214, 227-8, and in general book m. See also 315-16: what could be more anile, more tedious, more like a schoolmaster [cf. 339], or further removed from the sort of demonstration that a philosopher ought to use, than to mention right at the outset the Platonic view, then to disregard it and reject it completely, not mentioning how it was supported by his followers, and to neglect also to answer the arguments set forth by these people, or present a refutation, in which he should have
Actions and passions
189
In fact, almost everything that Chrysippus says is consistent with anger and fear arising in the heart, but desires and intellectual activities having different locations. It is only by begging the question at issue ('is the soul unitary or divided?') that Chrysippus can move from the claim that the heart is the evident seat of the emotions to the assertion that it is also the location of the rational and controlling part of the soul, the hegemonikon. Galen also accuses Chrysippus of equivocation, at times denying the existence of the spirited and appetitive parts, at times 'assigning them to a single locality in the body' (364: = SVF111.461). And perhaps most importantly, he accuses Chrysippus of being misled by an equivocation in the term 'irrational' (alogon). Something can be alogon (a) if it makes no sense to predicate rationality of it, like a stone; or (b) if it is of such a class as to possess reason, but fails in this instance to do so. Emotions, according to Galen are aloga (a); but Chrysippus treats them as being aloga (b), and hence locates them in the reasoning part of the soul (378-89: = SVF 111.441, 462, 464, 476). I shall not assess the force of Galen's objections here, nor discuss the extent to which they are fairly raised against Chrysippus.19 What we need to do now is to examine Galen's concept of a pathos, and to see what sort of work it can do within a general theory of the rational and the irrational in human conduct. Galen distinguishes rigorously between errors (hamartemata) and affections (pathe); indeed, he devotes separate books to the treatment of their diagnosis and cure (AJf.Dig. v. 1-58; On the Diagnosis and Cure of the Errors of the Soul [Pecc.Dig.] v.59-103). 2 0 Sometimes in fact Chrysippus does likewise:
he very properly distinguishes affections from errors. For errors are faulty judgments and reason that has been mistaken about the truth has erred. But affection, on the other hand, is not something that has erred, or makes a mistake in reasoning; it is a motion of the soul disobedient to reason. (PHP v.371-2) confirmed his own doctrine with scientific demonstrations, b u t instead to quote poets a n d to call in a multitude of non-experts as witnesses, a n d to record the expressions used by women, a n d to have no success even with t h e witnesses he invokes? 19
20
A n d cf. 358-9. Although I think that Galen is both fairer a n d less disingenuous in his treatment of the Stoic position than Inwood, for instance, allows in his recent account: Inwood (1985), esp. ch. 2; for more on the alleged Stoic self-contradiction here, see n. 54. O n this, see Donini (1988b) a n d Vegetti (1984).
190
JAMES HANKINSON
Galen offers an example: A man who scorns his children's life in order to save his country ... acts as he does because it seems to him noble . . . and he engages in a certain reasoning. Medea on the other hand was not persuaded by reasoning to kill her children. Quite the reverse ... she says she understands how evil are the acts she is about to perform, but her anger is stronger than her deliberations ... her affection ... does not obey reason ... but departs and disobeys the command, implying that it is the work or result of some power other than the rational. For how could anything disobey itself or reject itself or fail to follow itself? (372) The last point is of great importance: Galen is asserting that certain two-place relations are self-evidently irreflexive; and that, in conjunction with empirical data establishing the fact of such psychic conflict, will for Galen no less than Plato logically entail the separability (at least in analysis) of the psychic faculties. Chrysippus' own example of the downhill runner carried away by his own momentum and unable to stop at will {PHP v.369-70: = SF.F111.462 [part]) is, Galen thinks, grist to his own mill; people in such a condition are clearly moved by two different causes, their will qua human being, but their momentum qua physical body - and the latter is quite distinct from, and hence potentially in conflict with, the former (Galen dissects the downhill runner case at length: PHP v.372-7). Applying the model which Chrysippus has so helpfully, if unwittingly, supplied, we get the following story. If the affections are not kept in check, they take over and dominate the reason. What happens then is not a failure of reasoning or an error, but something altogether different - the normal channels which link deliberation to action are somehow short-circuited, and for that to happen the factor which renders them impotent must be logically and causally separate from them. Chrysippus' model, in fact, simply fails to account for the basic facts of psychic conflict. Of course more can be said in Chrysippus' defence. He may be right to think that psychic conflicts of this sort never actually occur: rather the soul simply oscillates rapidly between two incompatible sorts of 'hormetic presentation', 21 and it is a general tendency of those who oppose the Stoics on this question that they are too easily inclined to think that obvious facts of experience are on their side. 22 21 22
O n this concept, see I n w o o d (1985) p p . 56, 59, 76-7, 78-9; Gosling (1987) esp. p p . 181-6; on oscillation, see N u s s b a u m (1987) esp. p p . 155-8. See Gosling (1987) p p . i86ff.
Actions and passions
191
But whatever one thinks of that, Galen has a point when he castigates the Stoics for dithering as to just what the appropriate description of the soul in a state of affection really is. They can't hold reason responsible (for then there would be no such thing as an irrational act); but if reason isn't responsible, something else must be, or they will be committed to uncaused events (389-91: = SVF m.476):23 They ebb andflowlike the Euripus.24 After making reason and opinion the cause of the affections, they change to the position that when the soul moves in a certain way without a cause it falls into the affections; and again from this random and uncaused motion they return to the view that the movements of the affections are rational. For the refutation of either answer is at hand, and it allows them to persist in neither: when they claim the movements of the affections are produced by reason, they cannot distinguish error from affection; and when they remove reason as the cause, they are driven to the view that some motion is uncaused. (391) Galen's language of akrasia, akolasia, and sophrosune is strikingly Aristotelian,25 and he clearly sees himself as upholding the Peripatetic view of the nature of psychic conflict (one for which, given his general syncretist tendencies, he will discern a Platonic antecedent) against the psychologically incredible account of Chrysippus and his followers,26 one broadly in line with Posidonius' revisionist account. However, there are some points of contact between Galen and his Chrysippean opponents, notably in the notion that the passions of the soul are like diseases of the body, both structurally and in their requiring treatment (Nussbaum (1987) details the history and deployment of this account). 27 Still, it remains to be seen precisely 23
This is a n o t h e r passage in which Galen evinces his a d m i r a t i o n for Posidonius (see n. 10): - 8 3 , 99, 59 E K ; see further Galen's approval of Posidonius' rejection of the Chrysippean account of a. pathos as a type ofj u d g m e n t : PHP v. 4 2 4 - 6 , 429-30: T.95 E K , F.152, 157, T.98, 62 E K ; a n d also Galen's extended discussion of the n a t u r e of psychological illness a t PHP v . 4 3 2 - 5 4 , esp. 4 3 2 - 5 : F.I 59, 163, T.64 E K . A favourite image of Galen's: cfMM x.35, On the Function of the Parts (de Usu Partium [UP]) 111.454; it is owed ultimately to Plato: Phd. 90c. UP was edited by Helmreich (1907-9) for the Teubner series. />///> v . 3 7 5 - 6 . F o r a recent more sympathetic account of the Chrysippean position, see Gill (1983). A n d see H u t c h i n s o n (1988) for the early history of this connection; the idea of philosophy as the t h e r a p y for t h e ills of the soul is of course paradigmatically Platonic: see e.g. Phd. 82C-83C; Rep. x . 6 o 8 e - i ia; b u t it is not confined to the Platonic tradition: cf. e.g. Aristotle, EN 11.4, iiO5bi2ff; the connection is later a commonplace: cf. Nussbaum (1986b). It should be stressed that Galen takes issue with Chrysippus' view that it is the disposition to anger or grief or whatever that should be considered the disease; Galen makes it clear that these dispositions are akin to the bodily weakness which predisposes individuals to physical sicknesses, and hence are not themselves identical with the sickness: PHP v.439-42; cf.
T
24
25 26 27
192
JAMES HANKINSON
what the substantial differences between the two views amount to. I propose to approach that issue from two different angles. First of all, I shall consider Galen's philosophical analysis of the distinction between errors and passions (or more generally affections: the ambit is clearly at least potentially broader than that of the emotions here, as well as being directed at a particular sub-class of affections in general, namely the detrimental ones). Having done that, I shall turn to Galen's recipe for the proper treatment of each condition, and contrast it with the Stoic views on the same issue. II ERRORS AND PASSIONS
We have already seen what Galen's differentiation between hamartemata and pathe amounts to. And, of course, given the Platonic tripartition of the soul that he claims to have established, plus the assimilation of the rational to the brain and the source of the motor functions, and of the emotional to the heart and the arterial functions,28 that Galen should consider the two psychological affects to be more than merely accidentally distinct is hardly surprising (in a sense of course it is clear that they are distinct no matter what view of the soul you choose to adopt). At the beginning of Aff.Dig. (v.2), Galen notes (in connection with the work of one 'Antonius the Epicurean' On Controlling One's Own Passions; ibid. 1) that passions and errors are easily confounded, and that both are 'commonly called errors in a more general sense'.29 However, he insists that a proper analysis should keep them distinct, in the manner of PHP v.371: an error arises from a false belief, while a passion results from an irrational power within us which refuses to obey reason. {Aff.Dig. v.2-3; cf. 7)30 The Stoics, pre-eminently Chrysippus, dispute this, holding that properly analyzed passions really are errors, immoderate beliefs. The Stoic view has one apparent advantage over traditional accounts of psychic conflict, in that the identity of the agent is never in doubt 28 29 30
448-51, 451-3; compare Seneca's careful distinction between anger (ira) and irascibility (iracundia): de Ira 1.2.1-3. See Hankinson (1991b). H e sometimes speaks of errors when he means passions (Aff.Dig. v.3—6, 34); on this, see Donini (1988b) p p . 66f. Here I disagree with the suggestion of L. G. Ballester (1988) in his learned a n d thoughtprovoking article, p p . 140-1, that Galen makes all errors the result of passions (which would reverse the Stoic order of dependence): see further n. 49.
Actions and passions
193
(one is tempted, in the case of tripartitional accounts, to ask just who it is within whom the warring factions contend); 31 but it buys that advantage (if it is one) at the expense of what Galen takes to be an intolerable inadequacy to the phenomenology of the passions themselves. Nevertheless, the problems with the identity of the subject involved in any decompositional account of the structure of the soul, such as that adopted by Galen, do occasion a certain sensitivity in the use of the language of affection. After all, in its original sense a pathos is simply something that happens to something,32 and the traditional Greek conception of love is consistent with this. Indeed, we need not confine ourselves to love: consider Achilles' response to the embassy in book ix of the Iliad (644-55), where Achilles having listened to the ambassadors' speeches replies that his reason agrees with what they have said, but his thumos compels him otherwise.33 But if that is right, how can we do justice to the internal sense of our own involvement that a pathos invariably carries along with it? The phenomenology of passion does not seem to be the affective equivalent of the phenomenology of sickness, for example: it is we who are angry, while our bodies (or parts of them) are sick; or perhaps even if we want to say that we too are ill, it generally makes sense in the case of physical illness, as it apparently does not in the case of psychological affect, to ask what part of us feels ill.34 Furthermore, if we emphasize the external nature of passion, if we assimilate pathos to the pathological, how can we reasonably hold the agent in 31
32
33
34
A n d in general a n y account that decomposes t h e psychological functions into a variety of c o m p o n e n t parts must explain both how the different bits can function as a coherent whole, a n d how it can seem to the individuals that they are individuals, a n d not collections. This is just as true for c o n t e m p o r a r y neurophysiological accounts as it was for Aristotle, p e r h a p s the first writer in the western tradition to b e a w a r e of the problem: see de Anima in. 1-2. I t is cognate with the general verb of causal patiency paschein; G a l e n discusses the origin a n d development of the sense of the word a t PHP v . 5 0 5 - 1 3 ; On the Differences of Symptoms (Symp.Diff) vii.44-7; MM x.89—91. I n these passages, he generally wants to reserve t h e term pathos for some u n n a t u r a l alteration to the body's n o r m a l condition, although h e is well aware that this was not by any means its only original use (cf. Tim. 6id-66c, where sense perception is described as a kind of pathos); in Symp.Diff. vn.44-7, Galen holds (invoking Plato), that properly speaking, a pathos is still in a state of development, contrasting it with a disease (nosos) which is established - this connotation is not directly relevant in our contexts, but Galen will stress, in connection with affections, that a pathos is a kinesis - something actually has to be going on. thumos is a difficult concept here - it is something akin to w o u n d e d pride, a sense of d a m a g e d amour-propre; it seems, at a n y rate in H o m e r , to be self-regarding (or at least its a m b i t is confined to one's immediate philoi). See usefully M a n u l i (1988). She remarks, not without justification, that the word thumos is 'quasi intraducibiW (pp. 185, 188). This m a y b e controversial: Achilles, in t h e passage just referred to, does speak as though anger affects a part, the phrenes; a n d cf. Aristotle de An. 1.
194
JAMES HANKINSON
the grip of passion responsible at all? Passion, if it really is the uncontrollable alien force depicted in tragedy and epic, not merely diminishes responsibility - it removes it altogether.35 Here again, as Nussbaum (1987) has emphasized, the Stoics have something on their side: allow passion that degree of independence from the agent, and you run the risk of making praise, blame, and the ascription of responsibility, at least in certain cases, at best vacuous or metaphorical, at worst positively false and unjust, as indeed Aristotle clearly saw. Consider E.N. m.1-5: our very practices of praise and blame indicate that individuals are as a matter of fact responsible for their actions (except in extremis); and hence any acceptable moral theory needs to take that fact into account.36 The theory of responsibility must show how it is that we are, as our evaluative practices indicate we are, responsible for what we do. 37 So the existence of powerful pathe threatens to undermine the coherence of our ordinary notions of responsibility, provided that the following conditions, or something very much like them, hold: (1) it is beyond the agent's control whether or not the agent suffers from pathos P at time t\ and (2) if an agent suffers from P at /, then his actions are beyond his control. Those conditions are strong (and (2) certainly needs refining); and any weakening of them will require that they be supplemented from other sources if the threat to responsibility is to be a serious one. How might such difficulties might be overcome? Broadly speaking, it seems that such a project can adopt one of four strategies. Either (Si) you can hold that the will is unconstrained, that in some sense to be explicated our actions flow from original events internal to us which are themselves uncaused, or self-caused (that is, at least if David Sedley is right, the Epicurean 35 36
37
Gorgias' jeu d'esprit the Helen involves perhaps the earliest surviving explicit recognition of this fact, a n d its i m p o r t a n c e for forensic matters; see Barnes (1979) vol. 2, p p . 221-8. A n d Aristotle is quite clear that if you allow actions done as a result of appetite a n d emotion to fall outside the scope of the voluntary (and hence for h i m outside the scope of responsibility as well), the resulting restrictions o n the class of actions for which h u m a n beings can be held responsible will be so great as to fail conspicuously to preserve anything like our ordinary moral intuitions on the issue of legitimate praise and blame: EN in. 1. 1111 a24ff. For a clear and philosophically rich analysis of Aristotle's theory of responsibility, see Irwin (1980). These remarks do not do justice to the subtlety of Aristotle's discussion, a discussion which set the agenda for all subsequent inquiry into human responsibility: but they will do for now. I have more to say about the Aristotelian position at the end of this paper; and see n- 39-
Actions and passions
195
view),38 even if at times this internal freedom is over-ridden by the passions (such a position runs the risk of allowing, however, that in the latter cases we are indeed acting under duress, and hence are not responsible for what we do). Alternatively (S2) you can try to show that in some suitably strong sense we are causally in control of those of our actions and of our dispositions for action which are ultimately evaluatively relevant (including our susceptibility to the passions), even though our actual acts may themselves be directly and ineluctably caused by those dispositions (in addition of course to suitable sense-impressions: this is essentially Aristotle's strategy). 39 Or (S3) you can attempt to show how, appearances notwithstanding, the passions are deeply grounded within the individual, and hence part of the individual's personality, and justify holding the individual responsible in that way: although I am acting under the influence of anger, it is none the less my anger. This is, I think, the view of the Stoics40 and of Hume; and perhaps it is partially at least Aristotle's 38
39
40
See Sedley (1989); it was also, as I u n d e r s t a n d it, the view of K a n t : Critique of Pure Reason A-534-
Developed in ZsjVm.1-5; Aristotle's subtle, b u t ultimately unsatisfactory, account of how our actions can still be free even if they are now ineluctably determined b y the combination of a d a m a n t i n e disposition a n d external circumstances just in case at some time in the past the development of our dispositions was u p to us (see in particular his discussion of the
p r o p e r analysis of addiction at EN v. n i 4 a 3 f f . ; cf. 11.1-3), has been adopted by some modern writers on the subject, notably P. van I n w a g e n - b u t it is still unsatisfactory (crudely, because the idea that our past development of dispositions was u n d e r our control requires that it be shown how the dispositions can have arisen as a result of genuinely open decisions on the p a r t of the agent which were not themselves either merely r a n d o m , or else constrained by desires, information-states, a n d so on, whose own existence requires antecedent causal explanation). For the necessity of having the right type of impressions (phantasiai: evaluatively-laden sensory impressions), see 111.5, 1114a32ff.; and see H a n k i n son (1990a). I should add here that it is by no means clear that Aristotle's theory is exclusively of the (S2) type (as I remark below, the strategies are neither rigidly d e m a r cated, nor in all cases necessarily exclusive). Irwin (1980) indeed elaborates a n account for Aristotle which is a good deal closer to the (S3) c a m p , such that I a m responsible for x just in case I a m capable of deciding rationally a b o u t *, and I d o # voluntarily (i.e. without external coercion a n d not as a result of mitigating ignorance); the account ties the theory of responsible action closely to the idea that one is responsible for w h a t one does w h e n one's reasons for action are p a r t of the causes of one's action (for this type of analysis, see Davidson (1980) p p . 6 3 - 8 2 ) . Irwin rightly notes (against some recent commentators) that this sort of account is quite indifferent as to the t r u t h or falsity of determinism. However, it seems clear that even if this type of (S3) m a n o e u v r e can explain how we can act responsibly even w h e n there is now no genuine causal possibility of our acting otherwise, it will not serve to show that our dispositions themselves are suitably free from external influence - and to that extent, Aristotle is pretty clearly an (S2) theorist. I a m grateful to J u l i a Annas for discussion of these points. For the Stoic account of responsibility, see Stough (1978), a n d particularly Ioppolo's recent penetrating study (1988) on the internal development of the Stoic view (which was shared by Spinoza) from Zeno through Cleanthes to Chrysippus.
196
JAMES HANKINSON
too: see note 39. Effectively to adopt (S3) is to deny that anything other than the immediate causes of the action need be evaluatively relevant (so the question of how my disposition arose need not matter). Finally (S4) you can try to explain how it is that, even if the passions are in some clear sense beyond the individual's control, and perhaps even outside the individual's ambit altogether, none the less individuals must be held responsible for their outcomes: that is, one can simply deny that causal issues have anything necessarily to do with the case. This is Strawson's position41 and, I think, it is Galen's too. All four strategies pose notorious problems; and the lines between them are not sharply drawn (at the limits they tend to shade into one another). I shall return to them in the final sections of this essay. But for the time being, I want to turn briefly to Galen's account of the relation between passion and activity, pathos and energeia, to show how sensitive he was to the difficulty of even getting one's descriptive language right in this area. At the beginning of book vi of PHP, Galen writes: it is essential that the meaning of every term be accurately defined. Indeed, some of the questions that we earlier deferred are solved by this means ... [e.g.] whether desire, anger and the like are to be called energeiai or pathe. (PHP v.506) and he continues by remarking that an energeia is an active motion, i.e. one that comes from the moving object itself, while a pathos is a motion in one thing that comes from something else. Thus [the two] can often be combined in one and the same basic situation, differing not in reality but in definition. (PHP v.506-7)42 Galen's example is that of cutting: the dividing by the cutter and the being separated in the object cut are one and the same thing, but it is an energeia of the cutter, a pathos of the thing cut. In the same way anger is an energeia of the thumoeides, but apathema43 of the other two parts, and indeed of our whole body, when our body is forcibly driven to act by anger. (PHP v.507)44 41 42
43 44
Strawson (1962). I a m grateful to Bob K a n e for pointing out this m o d e r n parallel. T h e final r e m a r k owes something of course to Aristotle; for examples of things that are on his account the same in n u m b e r but different en tot einai, see de Insomn. 1. 459a 15-22 (which refers back to de An. 11.3. 429a iff.): a n d particularly in this type of causal context, Phys. 111.3. 2O2a22ff. N o t h i n g turns o n the use oipathema in place of pathos here: see Symp.Diff. vii.45. Note in this context Galen's de Causis Contentivis 9 ( = CMG S u p p . O r . 11, p . 140.23-37) where Galen urges care lest in causal contexts a mere re-description in active terms of the
Actions and passions
19 7
Let us label those senses energeial and pathosY. Galen then distinguishes two different senses (energeia2 andpathos2), where one denotes a motion according to, the other one contrary to nature; thus the heart's beating is an energeia^ while any palpitation is a pathos2 of it. 45 But even given that, certain types of pulse are not energeiai2 since they are preternatural, although they are energeiail (nothing else is moving the heart). Thus, given this second distinction between activity and passion, it is not surprising that a single thing may happen to be called both a pathos and an energeia (e.g. a pulse which is weaker than normal), but not with the same meaning. It is an energeiax for the movement in this pulsation too is active; but it is not an energeia^ but a pathos^ since the movement is not in accordance with nature. The same applies to anger and the other pathe: they are all energeiai^ of the affective part of the soul; but insofar as they are runaway and immoderate (ekphoroi te kai ametroi) motions and not according to nature, they would be said to be not energeiai2 butpathe2. (PHP v.509) Galen develops this line of thought, borrowing imagery from Plato, 46 in order to illustrate the variety of different possible combinations of energeiai and pathe. But what matters for us is the care Galen devotes to his choice of language here: he is aware that our analysis can go wrong at the very outset if we are unclear about the precise senses in which we are applying our descriptive terms. We have, then, two related claims about the terms we use to describe affects of this type: first of all we should distinguish pathe from hamartemata; and secondly we need to take care as to precisely which sense ofpathos we are employing when we describe affects like anger, grief, and fear.
45
46
effect be mistaken for the cause of that effect: see Hankinson (1987); a n d cf. Nat.Fac. 11.70: 'they have not given an answer to the question, b u t have simply reported what happens, thinking that by so doing they have assigned a n explanation to it.' For Galen's distinction between heart-beat a n d palpitation, which he thinks Hippocrates confused, see in Hippocratis Prognosticum (Hipp.Prog., = CMGv.9.1 [Heeg] 1915) XVIIIB 87-9; and On the Differences of Pulses (Diff.Puls.) vm.497; see also de Crisibus (Cris. [Alexanderson] 1967) ix.760. Galen was extremely proud of his sphygmatology, on which he wrote four major works: Diff.Puls. vm.493-765; Dig.Puls. vm. 766-961; de Causis Pulsuum (Caus.Puls.) x. 1-204; de Praesagitione ex Pulsibus (Praes.Puls.) ix.205-430; and two abstracts (the second of doubtful authenticity) de Pulsibus ad Tirones (Puls.) vn.453-92; and Synopsis de Pulsibus (Syn.Puls.) ix.431-549; s e e Hankinson (forthcoming 2), section 7. For the connection between pulsation a n d the faculties of the soul, see Praes.Puls. ix.421-30 (given Galen's account of the location of the parts of the soul, a n d of the mechanisms for the transmission of psychic power, such a connection should not come as a surprise). T h e charioteer and his horses, Phdr. 253c—64a: ibid. 510, 513; the m a n with the lion a n d the m a n y headed beast, Rep. ix.588c-d: ibid. 513.
I98
JAMES HANKINSON III THE THERAPY OF THE PASSIONS
The Galen who has emerged so far is a figure with a deep sense of intellectual history, of the tradition of ideas and of his own place in it; he is deliberately backward-looking, seeking the basis for his syncretist view of the soul in his distant Hippocratic, Platonic, and Aristotelian forebears. He has also set himself up in implacable opposition to the Chrysippean picture of the soul's structure, and of the proper analysis of its workings. Of the Stoics, only Posidonius, who on Galen's account at least was a syncretist of his own temper, bent on reviving the Platonic picture of the soul (and whom Strabo described as 'Aristotelizing': 11.3.8 = T85 EK), comes through with credit as being scientifically-minded, and not committed to the defence of the doctrines of his school at all costs (PHP v.390; That the Powers of the Soul Depend upon the Temperament of the Body
(QAM)47
iv.819-20).
However, Galen's position is not entirely incompatible with that of the Stoics - and the extent to which he owes them a debt 48 depends at least in part how far he believes that the passions must be extirpated altogether rather than simply reduced and domesticated. 49 The treatise Aff.Dig. outlines his programme for psychotherapeutic control of the strong and destructive emotions. The first stage in developing self-control involves us realizing that we are prone to error (in the broad, generic sense of the word); everyone minimizes their own faults, and is blinded by self-love (Aff.Dig. v.3-6). 50 Moreover, it is better to get rid of the passions first, 'since it is probably as a result of these as well that we form false opinions' (ibid. 7).51 Given the difficulty of discovering our own faults, we 47 48
49
50
51
Quod Animi Mores Corporis Temperamenta Sequuntur iv. 767-822: it appears in SM 2 32-79, Muller (1891). A n d perhaps to the Epicureans as well, although he himself would be loth to acknowledge such a debt: cf. e.g. DRN in.48-93, although of course for the Epicureans the ultimate cause of all psychological ills is the groundless fear of death. Nussbaum (1987) p p . i6iff., emphasizes the differences between the Stoic a n d Peripatetic outlooks - for the Stoics merely moderating the passions is not enough: they have to be eradicated root a n d branch. Galen refers to Aesop's fable of the two sacks suspended from our necks: 'the one in front is filled with the faults of others, the one behind with our own' (Halm (1889) P- 359) 5 an( ^ t o Plato's remark at Laws v . 7 3 i e that the lover is blind with regard to the object of his own love. H e returns to this theme later in Aff.Dig. v.35-6. I a m grateful to P. Donini for pointing out to me the force of the kai ('as well') here, which is a further reason for rejecting Ballester's view: see n. 30.
Actions and passions
199
must rely on someone whose judgment we trust to point them out for us (ibid. 8-14).52 Galen emphasizes the difficulty and lengthiness of the task ('each of us needs almost a lifetime of training to become a perfect man' (ibid. 14)); and the limitations imposed by our individual natures: we can't all be like Hercules - nor can we all achieve perfect mental calm; but almost all of us are amenable to some extent to training, in mind as well as body (ibid. 15-16). The emphasis on training forms the very heart of Galen's programme. We need first of all to become aware that there is something wrong with us (which may prove difficult for readily understandable psychological reasons); then we need to develop in consequence of that realization a strong desire to become better people, either from a sense of shame, or for some other motive (ibid. 16), a desire reinforced by constant verbal repetition of the goal in mind, both by ourselves (ibid. 30), and others (7-14, 30). 53 All of these ideas have clear modern psychotherapeutic resonances,54 and are paralleled in later Stoicism. Galen writes: when I was still a youth and pursuing this training, I watched a man trying desperately to open a door. When things did not turn out to his satisfaction, I saw him bite the key, kick the door, blaspheme, glare wildly like a madman, and all but foam at the mouth like a wild boar [cf. 22]. When I saw this, I conceived such a hatred for anger that I was never thereafter seen behaving in an unseemly manner because of it. At present this will be enough to keep you from blaspheming, from kicking and biting stone and wood, from looking wild looks; this will be sufficient motive for you to hide your anger and keep it within you. A man cannot free himself from the habit of anger as soon as he resolves to do so, but he can keep in check the more unseemly manifestations of his passion. If he does this frequently, he will discover that he is less prone to anger than he formerly was. Things which are unimportant will not rouse his wrath; and even if he does become angry over supremely important matters, his anger will be slight, (ibid. 16-17, trans. Harkins, with alterations) 52
53
54
Galen outlines the qualities such counsellors must have, a n d w h a t o u r attitude towards them should be: someone who frequents the rich a n d powerful is n o good; equally the rich and powerful have themselves little chance of finding a Diogenes w h o will point out their faults for them (Aff.Dig. v. 13-14). O n the importance of verbal repetition to the therapy of the soul, see Ballester (1988) p . 145; Ballester draws attention to the parallel between Galen's prescription a n d that of Plato Charm. 157a. F o r a very brief general history of psychotherapy, see Riese (1951).
2OO
JAMES HANKINSON
That the irrational acts provoked by passion are absurd and grotesque is an ancient commonplace, particularly in the Stoic55 and Epicurean 56 tradition; as is the view that they can on occasion become more than merely ludicrous and undignified. Galen recounts the story of a friend, whose ungovernable temper caused him to injure a slave in a moment of passion by bringing a sword (albeit in a scabbard) blade-down on his head (ibid. 17-21). He concludes as follows: that friend of mine, then, took thought for himself, and in a year had become a much better man. Even if you should not become much better, be satisfied if in the first year you have advanced and shown some small measure of improvement. If you continue to withstand your passion and to soften your anger, you will show a more remarkable improvement during the second year; then . . . you will notice a great increase in the dignity of your life in the third, fourth, and fifth years, (ibid. 20-1; cf. 53-4, 55-6)
The dominating motifs of these passages are (1) that self-control requires long practice; (2) that it is easier first of all to control the manifestations of the emotion than the emotion itself; (3) that if the manifestations are controlled, gradually the emotions too begin to wither; (4) that the gains in self-control increase exponentially. Moreover (5) one should attempt always to wait until the initial burst of passion has subsided before attempting to act in accordance with it, and that one should try as far as possible (at least in the case of slaves) to get someone else to exact the appropriate punish55
Seneca, de Ira 1.3-7; Chrysippus (SVF 111.478, ap. PHP v.414-15) indeed describes t h e effects of passions using some of the same examples as Galen does (a fact which perhaps suggests that Galen's purportedly historical story is in fact merely a literary device): we take such leave of ourselves a n d get so far outside ourselves a n d a r e so completely blinded in our frustrations that sometimes if we have a sponge or wool in our hands we lift it u p and throw it, as if we would thereby accomplish anything. If we h a d h a p p e n e d to have a knife or some other object, we should have used it in the same way . . . Often in this kind of blindness we bite the keys a n d beat against the doors when they a r e not quickly opened, and if we stumble on a stone we take punitive measures, breaking it a n d throwing it somewhere, a n d all the while we use the strangest language . . . F r o m such actions a person would get a notion of the irrationality in the affections, a n d would perceive how we are blinded o n such occasions, as though we were n o longer the same persons w h o h a d earlier engaged in philosophical conversations.
56
Galen takes these remarks as further instances of Chrysippus' tendency towards self-contradiction in his account of the effects of the passions: see p p . 188-90. Note, however, t h e tantalizing hint in the last sentence of an awareness on Chrysippus' part of the implications of the passions for questions of personal identity, although t h e precise nature of his views here seems impossible to determine. Philodemus, de Ira (Wilke 1914), p p . xlii—xliii.
Actions and passions
2o 1
ments. 57 Finally (6) Galen insists on the importance of frequent repetition of the appropriate moral maxims and exhortations (see note 53). It becomes apparent that the relation for Galen between passion and judgment is a complex one. Not merely is the first frequently responsible for errors in the second, 58 the second, properly applied, can control the first. One of the evil effects of passion is precisely that it dulls the judgment, and promotes false beliefs as to the appropriateness of certain courses of action (particularly in the case of punishments: this theme is taken up several times in Aff.Dig.; e.g. v.21-2); and false evaluations of the importance of certain external goods and of things that affect us. 59 At 37-52, Galen tells a long story about a young acquaintance who, in spite of advantages of wealth and birth, was malcontented; Galen, having hymned the virtues of frugality (which includes among other things being satisfied with one pair of shoes, instead of fifteen: 47), remarks that 'you have twice as much as I, and are in possession of your citizen's rights - so I don't see what could be the cause of your grief except insatiate desire and greed' (44). This is a type of mis-evaluation, a form of erroneous judgment: but it is important that Galen does not identify the passion (in this case greed - which he views as a sort of Ur-passion: ibid., 48-93 51? 53) with a judgment (cf. Nussbaum (1987) pp. 142-4, for an analysis of the relations between passion and belief); rather, the passion causes a mistaken judgment, which in its turn fuels the passion. So Galen quite clearly exhibits the relation between passion and judgment as involving a type of feedback; the more a passion is indulged, the greater it waxes (and conversely, the more it is starved, the more manageable it becomes); but equally, the more a passion is indulged, the stronger become the associated judgments and the stronger they become, the harder it is to achieve the state of reflective self-awareness appropriate to beginning the therapy designed to reduce the passions (cf. ibid. 28, 32). Hence the importance of involving others in the process, both as monitors of one's behaviour and as executors of one's wishes. And there is a further 57
58 59
Galen praises his father's example in both these regards: Aff.Dig. v.17, 4 0 - 4 1 ; a n d h e tells the story of the emperor H a d r i a n w h o blinded a slave in a fit of passion. F o r anger as a cause of excessive behaviour a n d cruelty in punishment, see Seneca, de Ira 1.2.5, 3- I - 2 >
5.2-3, 6.1-5, 9.4, 18.3-6, 19.2-4.
See n n . 30, 51 above. For the various types of distinction that Galen makes, see Donini (1988b) p p . 68-72.
202
JAMES HANKINSON
refinement: we can distinguish between the person who has managed to prevent the passion spilling over into action, and the person who has extirpated the passion altogether: 60 do not consider him wise who only stands acquitted of the action, namely kicking, biting and stabbing bystanders. Such a man is, it is true, no longer a wild beast, but he is not yet indeed wise - he is somewhere between the two. (ibid. 23) This is of interest for three reasons. Firstly, Galen thinks that there is a way to short-circuit the normal routes that connect passion to anger, presumably for empirical reasons - we know that we can have the anger, but control it, bottle it up with an effort of the will. Here again Galen is attempting to be faithful to the phenomenology of affect. Secondly, he considers the short-circuiting to be part of the means of attaining genuine freedom from the passions (although it should not be mistaken for the attainment itself). Even if it has been short-circuited, we still have the passion, considered as the disposition to behave in a certain sort of way; but the very cutting off of the route to its expression contributes, Galen thinks, to the process of causing it to atrophy altogether. Thirdly, this is a case of someone who is able to give an appearance of being of a certain type, while not really being so, like Plato's unjust man who gives the impression of justice. 61 And like Plato he holds that anyone satisfied with this condition is in serious error: for they do not understand the nature of mental health: anger is 'a sickness of the soul',62 and no one who understands health and sickness could genuinely prefer the appearance of health to the genuine article. 60 61 62
This distinction mirrors Aristotle's between the enkrates a n d the sophron: EN 1.13. iiO2b25ff.; VII.I. 1145a35ff.; vn.2. 1 i46agrT.; vii.9. 11510321! First considered at Rep. 1.345a; worked out in detail by Glaucon at the beginning of book 11:
358e-6id.
Aff.Dig. v.24; cf. 35, 37, 52. Ballester (1988) p p . 142-6, distinguishes between 'diseases of the soul' a n d 'mental diseases', where the latter are to be described (and hence treated) in purely physiological terms. For diseases of the type Ballester has in mind, see On the Affected Parts (Loc.Aff.) viii.i3iff., 156, 160—6, 173, 178, i95ff., 226-7, 327; see also Galen's Commentary on Hippocrates' 'Prorrhetics' [Hipp .Prorr.) xvi.517-18, on phrenitis; see also in this context the article of Pigeaud (1988) esp. pp. 157-66 (and Pigeaud (1981) pp. 55-70). While there no doubt is such a distinction to be made, and while in the case of 'diseases of the soul' the 'therapy of the word' (Ballester (1988) pp. i4off.: see Lain Entralgo (1957)) was particularly important, it is essential not to lose sight of the fact that they, too, are dependent on physiological states: cf. Pigeaud (1981) p. 59: len sommepour Galien, les maladies de Vame sont des maladies du corps'.
Actions and passions
203
The reference to the 'wild beast' in the last quotation is not mere rhetoric either - Galen views the desiderative soul as essentially untameable, 'not like dogs or horses, but befitting the wild boar and goat and any of the wild beasts which cannot be domesticated' (ibid. 28); the best you can do with it is beat it down: the chastisement of the desiderative power consists in not furnishing it with the enjoyment of the things it desires. If it does attain this enjoyment, it waxes great and strong; if disciplined and corrected, it grows small and weak. (ibid. 28) By contrast, the passionate soul can be domesticated - here Galen leans away from the Stoics and towards Plato and Aristotle. The thumoeides is not seen as a radically destructive tumour on the soul to be excised as completely as possible, but rather as a source of force and energy that can be directed into the right channels: I described at length in my work On Characters63 how someone might make his soul excellent, and how it was not necessary to dissipate its strength, any more than in the case of the dogs and horses that we use, but rather that we need to cultivate obedience in it, just as in them. The same work showed no less how you can employ the power of the thumoeides against the other part, called the epithumetikon by the philosophers of old, which carries us unreasoning towards bodily pleasures, (ibid. 27) But Galen's account is not as lucid as it might be - and indeed it seems in danger of breaking down under the combined weight of the separate phenomena (and traditions) that he is trying to integrate. On the one hand, strong emotions can be ungovernable, and are destructive; and they are closely connected with (at times apparently identical with) out-of-control desires (for wealth, for power, for sensual enjoyment); on the other hand the emotional part can be reason's ally. However, Galen soft-pedals the latter Platonic aspect — and the tamed thumoeides seems, for him, to have a merely decorative capacity, rather like the tame lion which it is supposed to resemble (on domesticating the passions, see Aff.Dig. v.26-7). Finally, did Galen side with Plato and Aristotle, thinking that the passions should be moderated, or with the Stoics, believing that only their total eradication could ensure mental health? It seems to me that there is no clear answer to be found in our texts, possibly for the 63
Peri ethon (not to be confused with On Habituations {Peri ethori)) does not survive in Greek (Galen refers to it at Lib.Prop, xix.45), although we have an Arabic epitome of it - Kraus (1937); an English version was published by Mattock in Stern et al. (1972). On this see the useful article of Walzer (1949).
204
JAMES HANKINSON
excellent reason that Galen himself had no clear answer. But one might tentatively suggest something like the following to be his account, based upon the distinction between the two senses of pathos.^ An uncontrollable anger will be a pathos^ of the affective part of the soul, in that it will be contrary to nature. In that sense, it must be entirely extirpated. But that does not entail that the smooth functioning (energeia^ of the affective part is to be eradicated as well - rather it is a proper function, and as such needs maintaining and fostering.65 Of course, that smooth functioning itself can be viewed as a pathosl of the individual taken as a whole (or perhaps of the reasoning part); but that will not in itself show that it is something pathological (for that would be to treat it as though it were a pathos^). So, what needs eradicating (on the Stoic model) is the excessive, preternatural pathos2; but that eradication takes the form of moderating the impact and influence of what is, under other circumstances, a perfectly natural and healthy part of the psyche, of bringing the restless horse under control, and allowing it to take up its proper role in the overall well-ordering of the soul. Galen differs from the Stoics precisely in that what will be left for him in the well-ordered soul will still be passions - controlled, channelled, constructive, but passions nonetheless, psychological drives capable of supplying a bloodless disengaged reason with the fire and commitment of motivation. These will not be the pale, desiccated husks of the Stoics' eupatheiai.66
IV PASSIONS, DESIRES, AND THE STRUCTURE OF LIFE
But, however successful we think Galen has been in describing the various types of mental affect, it is clear that he has an account of a type of practical psychotherapy which is more than negligible. And it is clearly one that depends upon there being such a thing as weakness of the will. For, in order for Galen's account to be success64 65
66
I a m indebted in w h a t follows to discussions with Dolores Velkley. There is more to be said regarding the notion of a proper function, a n d of the connection of that conception with Galen's strongly teleological ontogeny (see Hankinson 1988a a n d 1989); b u t crudely, if you think that nature does nothing in vain (at t h e level of organogenesis), a n d you think also that you have identified a distinctly functioning part (in this case the affective part), your general teleology will prevent you from thinking that it could be entirely useless, indeed positively harmful, for the proper functioning of the organism of which it is a part. F o r the eupatheiai, see Diog.Laer. vii.115 ( = SVF 111.431; cf. 432, 438-9); see Nussbaum (1987) a n d Inwood (1985) ch. 5, esp. p p . 173-5.
Actions and passions
205
ful, there has to be a genuine distinction between first- and secondorder desires,67 and one which is such that the object of the secondorder desires may be a desire-set different from that actually constituted by the first-order desires. One needs to be able to reflect upon one's own life, and be able to decide rationally that it would be better if in certain respects the structure of one's dispositional psychological relationship with the world were different. Hence, one will be in a position where one wishes not to be angry, say (i.e., one wishes not to be the type of person who angers easily), thinks it better that one be not angry, and yet one is as a matter of fact enraged. Clearly Galen's own version of faculty psychology is wellequipped to deal with this case; it is less clear, however, that the Stoic unitary account of the soul simply could not do so. Indeed, this sort of weakness of the will looks as though it can be handled perfectly well on the unitary model; to have meta-desires whose objects do not correspond to one's actual desires may be unfortunate, distressing, damaging to one's sense of psychological integration and so on - but it does not entail that one need have quite distinct concurrent incompatible views. If 'x desires f (for some suitably non-behavioural notion of desire) entails 'x believes t h a t j is desirable', then being in the condition of wanting something but wishing one didn't (because it's harmful, or married to someone else, or whatever) will turn out to have the following structure: (1) x believes thaty is desirable; and (2) x believes that it is desirable that x does not believe that y is desirable; and (1) and (2) are perfectly consistent. But psychological peace is not to be won, for Galen any more than Aristotle, simply by ensuring that one's first- and second-order desire sets are somehow compatible, and hence securing a measure of psychological integration, of accord between one's view of oneself, and one's actual actions - for the akolastic meets that basic criterion,68 and the akolastic is not someone whose psychological condition is emulable: indeed, as Galen drily notes any man is glad to make progress in that which he loves; so it is that drunkards are glad when they outstrip their companions in drinking; gluttons are happy to surpass those who delight in the abundance of foods; 67 68
F o r the development of this concept see Frankfurt (1971). As Aristotle himself points out: EJVvu.y, 1 i5oai5fF; 8, 11500291!.
206
JAMES HANKINSON
the voluptuary in eating rejoices when he outdoes those who find their joy in cakes ... And I have known some men who were conceited about the great number of their sexual exploits. (Aff.Dig. v.32) What Galen needs is a way of grounding the claim that a life in which the contribution of the passions is reduced to the role of supporting and providing commitment to the deliverances of reason on the Platonic (Rep. iv.439c—42c) model, while the desiderative element is starved and forced to wither away, is somehow obviously rationally preferable to that of the akolastic. The key notion here is self-mastery. Again, this idea has a long philosophical pedigree; it is common to the Platonic, Aristotelian, Stoic and Epicurean traditions, and is consequently one of the few genuine commonplaces in ancient moral thought. Galen repeatedly adopts the language of slavery to describe a man in the grip of passions;69 and equally, he describes anger as 'a sickness of the soul' (Aff.Dig. v.24, 35, 37, 52), albeit one that is not always recognized as such.70 The combination of these two styles of description yields an answer to the question of how Galen can commend the life of someone who has eradicated their passions as being objectively the better life for human beings. As we have seen, Galen thinks it unthinkable that anyone who recognized the truly pathological nature of anger and the passions should want merely to give the impression of having dealt with them - that would be like painting over the spots, treating the symptoms. It might be objected that there's nothing inconsistent in holding both that it's wrong to act on one's pathe (and hence that they should be resisted), and that it is none the less a good thing to have pathe (because you might hold that to struggle successfully against temptation was virtuous - or perhaps even, bizarrely, that it was fun). That is, I think, the standard Christian view, in which moral heroism turns out to be something desirable. It is, however, profoundly un-Greek (witness Aristotle's suggestion that sophrosune is to be preferred to enkrateia); and in any case if Galen really can show that pathe in this sense are properly to be regarded as diseases, he is on firmer ground — it is surely a mark of a certain psychological aberrance to think that struggle against 69
70
Aff.Dig. v . 2 3 , 24; QAM iv.778; cf. ibid. 782, 787; a n d That the Best Doctor Be Also a Philosopher (Opt.Med.) 1.59; the metaphor is of course venerable: see Nussbaum (1987) pp. 162-71, esp. p . 165, n. 71. For the difference between Galen's view of the nature of the sickness, a n d that of Chrysippus, see PHP v.432-54: = .SKF111.471-3; on the concept of a 'sickness of the soul', see n. 62.
Actions and passions
207
disease, however successful, is to be preferred to health. But to justify the claim that anger and its congeners really are diseases of the soul, Galen needs to show how they fit into his own general pathological scheme. For Galen disease consists in damage to or destruction of one of the essential activities (energeiai) of the body;71 hence for anger and the like to be diseases of the soul, they must involve or consist in damage to some essential psychic function. And there is an obvious candidate for such a function, namely reason. Galen agrees with Aristotle that reason is the supremely human quality: Man alone as compared with other things has the special gift of reason; if he casts this gift aside and indulges in anger, he is living and acting like a wild animal rather than a man. {Aff.Dig. v.22-3) Abrogate reason, and you destroy your own humanity. Here a problem surfaces: Galen's language suggests that we are strongly responsible (i.e. both causally and morally) for our submission to passion. Yet that picture does not seem to square with the idea of passion as a disease, or as something external to us. None the less, it is reasonably clear how Galen wants to conceptualize the pathological nature of passion: passions interfere with, and render impotent, our reasoning abilities, causing us to do things that we would never have done on cool reflection, such as inflicting incurable injuries72 and berating inanimate objects like doors, stones and keys (ibid. 22).73 The first step in combatting this infection is to prevent the passion from issuing in action, as it does in the case of those fully in its grip. But that is only the first stage — following that, we have to eradicate it. Galen assigns, traditionally enough, the power of voluntary action to the rational part of the soul;74 what happens in the grip of passion is precisely that that normal function gets diverted or usurped altogether. What is needed, then, is the re-establishment of the proper lines of communication between the rational soul, the brain, and action. Part of the physiological theory he needs to ground this can be found in his account of the separate location of the passions in the 71 72 73 74
See e.g. MM x. 78-81; and see Hankinson (1991a) ad loc; this is part of his reason for rejecting Chrysippus' dispositional account: see n. 27. W h i c h G a l e n considers ' t h e act of a m a n insane o r of a wild a n i m a l ' : ibid. 22; cf. Seneca, de Ira 1.1.3—7, 19.1-4. See nn. 55, 57. PHP v.600; Symp.Diff. vii.55-6; MM x.636; see Hankinson (1991b); see also Pigeaud (1988) pp. 170-1, and the references collected there.
2O8
JAMES HANKINSON
heart, and of their mediation via the arteries. Of course, that in itself will not fully explain the causal mechanisms involved in the usurpation of the role of reason by the passions, and indeed such an account is difficult to square with his neurophysiology (if one can call it that). More precisely, Galen can show how the passions might interfere with the dictates of reason, causing a type of paralysis; 75 it is much harder to give an account of how the passions might not merely prevent rational action, but actually promote conflicting types of action; and although such an account can be sketched on Galen's behalf, he does not himself attempt one. 76 We nevertheless now have a reasonably clear account of how control of the passions relates to Galen's notion of the good human life, construed on the broadly Aristotelian model. For Galen, just as for Aristotle, the good man is autarkes (E.N. x.7. 1177a! iff.); and autarkeia is defined by Galen as being the condition in which one's actions are under one's rational control, where one genuinely acts as opposed to simply being affected; thus, in the case of the good man, Galen can adopt a type of (S3) justification, in which one is responsible because one's reasons for action are part of the action's causes (see note 39). Galen will sometimes use the phrase erga kai pathe as though nothing turned on the distinction between them; 77 but 75 76
77
Although he never in fact does so, all the necessary material is contained in his remarks on the difficulty of making precise neural sections in the spinal column: PHP v.235ff. We could begin to generate an account for Galen if we allowed that the boiling blood in the arteries penetrated the brain and caused alterations of its normal functioning, which in turn caused non-normal neural signals to be sent out from it. This would be to adapt and adopt a theory Galen considers and then rejects on the Stoics' behalf, namely one in which the brain is the source of the nerves, but only mediately the source of psychic power, which itself originates in the heart and proceeds via the brain: PHP\. 255-6, 261-4; see Hankinson (1991b) for a full discussion. Note that Galen's reasons for rejecting the Stoic theory explicitly do so only as an explanation of normal functioning (the body could not be so constructed with regard to its normal functions, for teleological reasons: see Hankinson (1988b) and (1989) and n. 65); but that, of course, makes it quite coherent for him to view this as a possible type of abnormal functioning. But it must be stressed that, tempting as this is, it is all speculative. Thus he describes the 'unseemliness of behaviour due to erotic desire, gluttony, drunkenness, and luxurious eating' as lerga and pathe of the desiderative power of the soul': Aff.Dig. v.27-8. But perhaps Galen does not mean to suggest that erga and pathe have the same sense here - rather an analysis akin to that given for pathos and energeia is in order: gluttony and the like are indeed actions of the desiderative soul; but they are emphatically not (or at least are only incidentally and not in any genuine sense) the actions of the whole man, but are at best pathe of that. The association of erga and pathe in discussions of the soul goes back at least to Plato (Phdr. 245c); and cf. Aristotle, de An. 1.1, 403b 12: 'the natural scientist is concerned with all the erga and pathe of this body, and of this type of matter'. The distinction is between what the soul can do and how it can be affected, in a quite general way; that is, with its causal powers broadly construed.
Actions and passions
209
elsewhere he puts them in pointed contrast - thus at Aff.Dig. v.41, he writes: when I compared the noble deeds [erga] of my father with my mother's disgraceful passions [pathe], I decided to embrace and love his deeds and to flee and hate her passions. The father78 is described as performing actions, whereas the mother is not - and the reason for that distinction should now be obvious. Galen, then, leans towards the notion of a disinterested, hegemonic principle of rationality quite distinct from desires and interests that was a central part of the philosophical tradition from Plato through Descartes until Hume. But reason for Galen has that privileged position precisely because in its untrammelled condition it is able to take the long view, and to work out the correct course of action from the point of view of a life taken as a whole in which each element is accorded its proper value. And for Galen, unlike the Stoics, that does not require literally valuing at naught the ordinary comforts of life, treating them at best as preferred indiflferents - rather it involves evaluating them accurately, for the overall contribution they make to a life well-lived. Now, such positions are notoriously prone to charges of circularity — surely one can only make such strategic determinations about life plans and the like when one has already sketched a rough conception of what the good life for oneself is to consist in: but that is entirely up for grabs. Such charges can be levelled against Aristotle too, but in both cases they miss the mark; for Galen, like Aristotle, thinks that there is a determinate, non-subjective way of determining the best life available to a human being, a method founded upon biological analysis. Broadly, there are two ways in which one might seek to justify the claim that a certain way of life is objectively better for human beings, irrespective of their incidental beliefs about it. The first (Strategy A) consists in showing that a life lived in a particular manner, articulated by a particular set of desires and goals, will as a matter of fact issue in a greater level of satisfaction for the liver than any of the alternatives; Strategy A does not seek to show that a particular set of articulating principles, or guiding second-order desires if you prefer, 78
Whom Galen lauds as being 'mild-mannered, just and kind': Aff.Dig. v.40, in sharp contrast with his mother, whom he represents as a Xanthippe prone to biting her housemaids.
210
JAMES HANKINSON
is better independently of the effects they may have on one's life, but rather that they are to be preferred precisely because a life lived in accordance with them is better from the inside than one that isn't. Such accounts raise interesting but not insuperable difficulties centred around the concept of identity - but this is not the place to follow them out. The second strategy, B, is quite different from A, although because of certain structural similarities it bears to A (principally in that both strategies make use of the concept of'a better life' which is in both cases to be explicated in partially non-subjective terms),79 the two are sometimes confused. Strategy B involves attempting to ground the view that a certain type of life is superior, and superior for everybody, without direct reference to what it is like to live that or any other type of life at all. Proponents of Strategy B, Aristotle among them, regularly claim that as a matter of fact, the superior life will be a better one subjectively for the agent to live: but they do not have to do so. In the ancient world, the Epicureans are the principal representatives of strategy A;80 nowadays most people are, I think, A-strategists. Galen was clearly in the B-strategy camp. Of course he does hold that the life in accordance with reason will 79
I should perhaps clarify what I mean by 'partially non-subjective' here - and the difference between Strategies A a n d B is principally to be located in the different ways in which they are non-subjective. Strategy A is 'partially non-subjective' (or 'weakly objective', if you prefer), just because it is taken that the fact that one (Lj) out of a sequence of possible lives < L n L 2 , . . . L n > is as a matter of objective, absolute a n d non-relative fact better for /,, where ' I , ' picks out some individual; b u t that is of course compatible with some other life from the set Lj being better for I 2 . Strategy B is 'totally non-subjective', (or strongly objective) just because it allows the individual indices to be dropped: or alternatively because it holds that, if it is true that some life L; is as a matter of fact better for some individual I, then it is better for any individual you care to mention. A, then, entails that (1) (x) (3/,) (Z, is best for*); B entails the stronger
80
(2) (3Z,)(*)(Lisbestforx); and the quantifier marks the difference between partial and total non-subjectivity. (i) and (2) are not, however, intended to be full analyses of the difference between A and B. Actually, this claim is controversial: whether or not it is strictly true depends, I think, on whether for the Epicureans it is a merely contingent fact that certain types of life are actually more pleasant than others - it is certainly true that the Epicureans are naturalists about h u m a n nature. I suppose the answer to that question depends on whether they would reject the counterfactual claim that a life consisting in the satisfaction of only natural and necessary desires would be better even if as a matter of fact it was not the (only) way to secure a m a x i m u m of pleasure over pain; I assume they would, hence my enrollment of them as A-strategists; but the case is unclear. I a m grateful to David Gill for useful discussions on this topic.
Actions and passions
211
be subjectively a better life - it will be less turbulent, and not prone to chronic and unalleviable dissatisfactions (Aff.Dig. v.45-56), dissatisfactions which stem from the very nature of the structure of desires that such a life involves (51-2); in these passages there is almost an Epicurean ring to his moral exhortation (cf. Lucretius de Rerum Natura m.48-93; note 48): the man who hates insatiate desire and greed loves to be self-sufficient and independent. If, then, being free from grief lies in this alone, and this lies in our power, we can now be entirely free from distress (lupe)81 by keeping the doctrines on greed and independence ready at hand. (Aff.Dig. v.52) But crucially, this fact in itself is not the reason for avoiding one life and choosing the other - that choice can be justified simply by appeal to the greater nobility of the life of reason, a nobility that is the direct expression of its greater intrinsic worth as a life for a human being, and these goods (principal among them being freedom from distress: alupia)82 are derivative goods.83 Galen stresses, as does Aristotle, that pleasure is not something specifiable independently of a description of character and way of life, and hence cannot serve as an independent point of reference and justification for a choice of life.84 Rather, we find pleasant those types of things that are as a matter of fact in accordance with the sort of life to which we are committed; if our second-order desires to have a desire-set with a certain structure are realised, then our pleasures will consist precisely in the satisfaction of those desires. 85 Galen 81
82
83
84 85
lupe is translated by Harkins (1963) as 'grief, which is a general Greek meaning of the term - here, however, it is being used in its semi-technical sense of'any unpleasant mental disturbance' - hence 'distress' seems the appropriate translation. The Stoics have a great deal to say about lupe, defining it as 'an opinion that some recent {prosphatori) evil is present' {PHP v.416, = SFFi.2i2, 2.463; cf. /WPv.332, = SVF i.2io, 3.481; SFF1.211); Galen, though, will dispute the identification of distress directly with beliefs (on the importance of the qualification that the beliefs be 'recent' or fresh, see n. 14). T h e Stoics classed alupia as one of the 'final goods {telika agatha)\ i.e. things that were good in themselves: Diog. Laer. vii.96 ( = SVF111.107; c^- 112)> along with confidence, thought, freedom, joy, good-disposition a n d all actions in accordance with virtue; a n d given the close connection between alupia a n d ataraxia, it is clear that both the Epicureans a n d the Sceptics would agree with them here: cf PH 1. 25ff., M 11. 1 ioff., 1416°. At this point it is worth noting that, although Galen stresses the aesthetic advantages of such a life over its competitors (passions a n d desires make one ridiculous a n d ugly), for Galen, aesthetics itself should be functionally-based - something is beautiful just in case it fulfills its function well (see UP m.897-9; an<^ s e e Hankinson (1989)). Although there can still be 'natural' a n d ' u n n a t u r a l ' pleasures: EN vn.5. 1148b 156°. Aff.Dig. v . 3 2 - 3 , 44ft. Galen's asceticism is most in evidence when he discusses what types of things are appropriate objects for pursuit - a n d here again his tone becomes almost Epicurean. Avoid costly garments, jewelry, a n d things which are thought to be beautiful (but in fact are not, because they fulfil no function: see n. 83) a n d pursue only those things
212
JAMES HANKINSON
agrees with Aristotle, then, that the good life will be pleasant - but it will be pleasant because it is good, and not good because it is pleasant. V REASON AND RESPONSIBILITY
The last passage I quoted may serve as a convenient bridge to the final set of issues I want to discuss. Galen describes the psychotherapeutic process in causally determinate, if not necessarily deterministic, terms; you bring it about that your desires have a certain structure; but this process is supposed to make you free. But a familiar problem suggests itself: how can you cause yourself to be free? That partly depends on what freedom means here. If our concept of freedom is simply negative then the problem vanishes we can certainly free ourselves of things. And in a sense, Galen's notion of freedom is negative, since to be free for him is to be free of the passions. Nevertheless, it is also clear that for Galen being free is a matter of being rational, or rather of being such that one's reason controls one's behaviour (this is not equivalent to a Davidsonian analysis note 39 — in which reasons may simply be desires); and yet, given that one can become free, he has to allow that the actions of an unfree individual can be partially causally responsible for that individual's freedom. It is frequently taken as axiomatic that 1i) it is a necessary condition of any individual Fs actions being free that they are not caused by anything or anyone other than /; let us then suppose that / is an individual who has successfully completed Galen's course in psychotherapy, and hence that (2) Fs actions are free; if, then it turns out that (3) something other than / i s a cause (or at least a partial cause) of Fs actions, we have a patent inconsistency. If we think that / is actually substantially different as a result of the psychotherapy, then we will have reason for asserting (3); but even if we reject that fragmentary view of personal identity, there will be many other reasons for asserting (3), at least in its weaker formulation - contingent facts which contribute to bodily health: Aff.Dig. v.46-9; see also the discussion of the relations between sex and pleasure at Loc.Aff. vm.417-20.
Actions and passions
213
about circumstances and environment surely have some effect on our actions. I do not have the space here to follow these issues out in detail and in any case, as we shall see, they are not directly relevant to Galen's position. But, briefly, anyone impressed with (1) and (2) might reject (3) on the grounds that while external circumstances were causally relevant to one's free decisions, they were no part of the causes of them (such a view need not merely be a desperate stipulation). Hence, on this account (call it [Ai]), given that (3) may be false, (1) and (2) can be safely maintained. However, I find no trace of this strategy in the ancient world. Alternatively (or [A2]), one might deny (1), and hold that one's actions can be caused by things outside one's (present) control, and yet still be the actions of a free agent: my actions are freely mine either (a) simply because they issue from my desire-set plus my information states; or (b) because (i) at some time in the past I acted in such a way as to bring it about that I would behave in a certain way, and (ii) my action then was free; both of these (which are versions of (S3) and (S2) respectively, and which are not incompatible: see note 39) amount to saying that my actions are free just because, as Aristotle would put it, the motive cause is within me.86 The trouble with (a) is that it invites the question 'so what?' Why should it matter that my desires and states are mine in that sense? As well hold the knife responsible for the stabbing because its sharpness was instrumental in the parting of the flesh.87 The problem with (b), the (S2) part of Aristotle's solution, centres on clause (ii): how, and in what sense, was the earlier action free in a 86
87
EN m.i, 11 ioa2ff., ai4ff., biff., 111 ia22ff; cf. 5, 1113b3off. Compare the Stoic notion of an action's being ' u p to us (eph'heminy: see Alexander of Aphrodisias de Fato 13 (p. i8if. Bruns, = Sharpies (1983), p p . 190-1), = SVF 11.979: 'they [i.e. the Stoics] say that what occurs in accordance with impulse a n d through our own agency is u p to us'; cf. ibid. 33 (p. 205 Bruns, = Sharpies, p . 207). Furthermore, on the Chrysippean view, assent (sunkatathesis) is u p to us: Cicero Acad.Pr. 11.37-8, = SVFn. 115 (cf. Origen de Or. 6, = SVF 11.989) on the proper interpretation of these doctrines, see especially Inwood (1985) ch. 3; a n d Ioppolo (1988). I should emphasize that neither (a) nor (b) is intended to provide a complete account of free action (although (a) might conceivably d o so), b u t only of how actions outside a n individual's current control m a y still be free actions of that individual. T o give a complete picture, (b) needs to be supplemented by a n account of how the original action referred to in (i) a n d (ii) was then free. But I take it that that presents no particular difficulties. This sort of objection was of course canvassed in antiquity: see classically Antiphon's Tetralogies. Of course, I d o w h a t I d o because I want to, whereas the knife has no desires of any kind - b u t the question still remains whether I can be held responsible for wants outside m y control.
214
JAMES HANKINSON
way that subsequent ones were not? Clearly, given the nature of the case described it was free from the control of the dispositions - but that clearly does not entail that it was free from absolutely everything else. Indeed, given Aristotle's account of how we become good, with its emphasis on education, habituation, and learning by moral example, it is hard to resist the conclusion that the actions by which we develop our future dispositions to act virtuously not merely are, but indeed ought to be constrained by factors beyond our control. And if that is true, the question immediately arises: how then can (ii) be true in a sense strong enough to ensure that we are indeed responsible for the acts with which we shape our dispositions, and hence transitively for their subsequent outcomes? And supposing we accept (3) (at least in the weaker sense obtained by stressing the parenthesis 'or at least a partial cause'), as I believe both Aristotle and the Stoics would have to do - then we seem to be committed to denying (1), at least for a strong sense of 'cause'. We are left with the apparently paradoxical claim that one's actions can be free even though they are causally conditioned (at least in some sense) by things outside one's control. I am inclined to think, though, that the extent to which either strategy is plausible (at least as an account of freedom) depends entirely on how close the alleged causal link is between the external occasioning forces, and the supposedly free action - the tighter the link, the less tempting it is to claim that, simply because I am (i.e. my desire-set is) instrumental in bringing about the action, the action is freely mine - hence [Ai] and [A2] tend to collapse: (1) is plausible only for a strong reading of'cause'; while (3) seems true only for a much weaker interpretation; hence (1) and (3) need not conflict. That seems to me to be at least a sketch for a workable compatibilist defence of freedom and responsibility; although one might legitimately wonder just how compatible with determinism it really is. But in a sense, for Galen at least, all of this argument is irrelevant for his notion of freedom does not even run the risk of entailing (1). For Galen (as indeed for the Stoics),88 an act is free just in case it proceeds from a unified psyche in which reason is the controlling element; freedom, then, amounts to proper functioning, in exactly the same way as we say of a car that it is running freely when nothing is fouling up the works. Hence an action's being free does 88
via their own concept of freedom, or eleutheria: see SFF111.355 ( = Diog. Laer. vii.121), 356.
Actions and passions
215
not entail that it be completely causally independent of anything external to the agent - only that it now be wholly explicable (in the circumstances which as a matter of fact obtain) on the basis of the agent's internal states of character, belief, and desire. Whether this concept is capable of satisfying our ordinary notion of human freedom, I do not know: but I am inclined to think that the 'ordinary concept' is both incoherent and ultimately unnecessary, as perhaps Galen himself saw. Aristotle worried himself with problems in the causal transitivity of responsibility: how can we be responsible for actions we are unaware of performing?89 More importantly, if our actions are the result of what appears to us to be good, how can we be responsible for them if we are not responsible for the contents of our phantasiai?90 Galen simply brushes these difficulties aside. We are, for him, straightforwardly the products of our inherent natures and our upbringing. In the ancient analogue to the modern debate between the claims of inheritance and the environment to be the chief factors in determining character, Galen comes down strongly on the side of heredity. He emphasizes, against the Stoics, the innate distinctions of character observable in very small children: some are always smiling and even-tempered,91 others sullen and difficult; some are naturally quick, others naturally dim-witted; some naturally truthful, others natural liars (although rarely do we come across cases of children either wholly good or irremediably evil). We can only educate them to the extent that they possess suitably malleable temperaments;92 and that extent is severely circumscribed. Character is largely a matter of genetics. This is Galen's version of the 'cradle argument'; 93 he rejects the Stoic view that all children are naturally good, but are corrupted by parents and teachers 89 90 91 92
93
EN in. 1, 11 iobi8.ff; 5, 1113b22ff.; here at least Aristotle has a more plausible account than Locke: Essay 2 2 7 20: cf. Leibniz, Nouveaux Essais 2 2 7 20. isjVin.5, 1114a3~bi5; see n. 39. QAMiv.768-9, 816-78 Aff.Dig. v.37-40; cf. Plutarch, Cons. 6o8d, on the natural goodness of his daughter. This is a common refrain throughout Galen's work: coming to understanding requires both natural ability, a n d hard work (this is stressed particularly in Opt.Med. 1.53-63) - that is why he despises the 'easy road' of the Methodical doctors to knowledge (AfAf x.5, 781, 927; de Sectis (SI) 1.83) as much as he abominates that of the Cynics in ethics: Pecc.Dig. v.71; PHP v.233-4; note Galen's pointedly amusing anecdote concerning the relations between Theagenes the Cynic a n d Attalus the Methodist doctor at MM x.909-14. See Brunschwig (1986) for a fine discussion of the forms of this argument in later antiquity. T h e account of the inheritance of character was no d o u b t developed at much greater lengths in the four books of the lost Peri ethon: see n. 63.
2l6
JAMES HANKINSON
(QAM iv.818-19), since we cannot then explain the original fall from grace (who could have corrupted the teachers?). Equally (and for symmetrical reasons) they cannot all be naturally attracted only by base pleasure and away from the good. But insofar as they are educable, we can, given a suitable course of paideia, instill in them the appropriate sorts of beliefs and make of them properly social animals.94 Surely if this is true individuals cannot be causally responsible for their states and actions? Possibly. Let us consider a second triad, closely analogous to (i)-(3): (4) we are responsible for our actions; (5) we are only (morally) responsible for things for which we are causally responsible; (6) we are not (ultimately) causally responsible for our actions. Again (4)-(6) appear inconsistent; again, similar moves can be made: weakening the sense of'cause' for (5); strengthening it (or putting all the weight on the 'ultimately') in (6), in order to defuse the inconsistency. Now, it is generally held that (7) we can only be (legitimately) praised/blamed/punished/ rewarded for actions for which we are responsible; hence if (5) and (6) force us to reject (4), we cannot (justly at any rate) be punished; and crucial to this line of thinking is the notion that the sense of 'responsible' in which (7) is true is a strong one, otherwise it becomes almost a truism. Mere instrumental or mediate responsibility won't do: hence (7) on this reading amounts to (7*) we can only be praised/blamed/punished/rewarded for actions which we genuinely caused. 95 For Galen, by contrast, if responsibility matters at all, it does so only insofar as we can be held responsible for actions, hence judged, condemned, praised, blamed, and so on: there need be no antecedent way of determining responsibility based on causal considerations which may alter our views as to who is or is not to be 94
95
QAM iv.812-14 stresses the Platonic antecedents (in the Laws: V.747CUT.) for this view; cf. ibid. 808: our regimen is directly causally linked to the type of character we develop although as Lloyd (1988) p p . 37fF. notes, the exact nature of this dependence is left unclear by Galen. T h e importance of social a n d moral education was not merely a Platonic concern: see Aristotle EN 11.1-4; Pol. v m . 1 - 3 ; it is, in fact, a Greek educational commonplace. There are indeterminacies even with this; a n d for it to be plausible as a n account of responsibility, it needs either to be strengthened to include cases of negligence, or alternatively those cases must be subsumed by a sense of'cause' such that we can be said to have caused those things which we failed to prevent.
Actions and passions
217
punished. The justification for punishment (and for that matter approbation) does not depend on any set of conditions of the type of (4)-(7) being satisfied. For Galen rejects (7), at least for any strong sense of 'responsible' expressible by way of (7*): mediate responsibility will be enough to secure praiseworthiness and punishability. Thus Galen adopts strategy (S4). In a moment we shall see why he does so. But let us map out the territory a little first. Generally, there seem to be three more or less plausible lines one can adopt if one rejects (7) in its full strength; crudely, one can justify punishment even if (7) fails on grounds of (a) public safety; (b) deterrence; or (c) rehabilitation, (a), (b) and (c) can all bypass the issue of whether or not we can attribute strong responsibility to agents, since they derive their justification from their alleged causal consequences: in the case of (a) and (b) the creation of a better society, in the case of (c) by causing the individual agent to be such as to integrate better into that society. Both (a) and (b) are in a sense agent-neutral - they neither depend nor have any direct bearing upon the agent's state of mind, or character, or his deserts in any sense; (c) is not agent-neutral, but nor is it desert based either (and of course one's theory can involve combinations of (a), (b), and (c)). Galen's view is striking, in that he subscribes at least partially to all three views. We shall examine his versions of (a) and (b) in a moment; and in a sense his commitment to some form of (c) is obvious: after all, passions (and by extension bad states of character) are diseases - and diseases require treatment, hence insofar as treatment is still a possibility, Galen will recommend it. 96 And that treatment will frequently involve both discipline and punishment, as we have seen (e.g. at Aff.Dig. v. 19-20). But, just as for physical ailments, so too with diseases of the soul the case may simply be a hopeless one, admitting of no therapy (ibid. 29, 52). 97 In that case, the only available treatment is surgical excision. 98 Galen himself asks how responsibility might be possible at QAM iv. 784. Let us consider his answer: 96
97 98
Although as we have seen t h e treatment cannot simply be externally applied - it requires a n effort of commitment on the part of the patient, in a n interestingly pro to-Freudian way - the agent must still have at one level the desire to better himself. T h e notion t h a t diseases frequently progress to a stage beyond therapy is a n ancient commonplace: see the Hippocratic treatises Prognosis i, a n d On the Art 3-5. Galen sometimes remarks that we would prefer to a m p u t a t e a useless organ rather than simply carry it around with us uselessly: UP iv.347.
2l8
JAMES HANKINSON
Nor is this line of reasoning [sc. that our characters are dependent upon the nature of our bodies] destructive of the fruits of philosophy, but rather it is contributory towards them, and a teacher of them, even though some philosophers have failed to realize this, at least to some extent. For those who hold that all men are receptive of virtue, and those who hold that no one properly gets hold of justice, have each seen a half of human nature. For it is neither the case that everyone is naturally born friendly to justice, nor that they are all hostile to it, since everyone becomes what they are on account of their bodily temperaments [kraseis]. How then, they say, can someone be justly praised or blamed, hated or loved, when they have become either wicked or virtuous not as a result of themselves, but on account of their temperament, which is clearly the result of other [external] causes? Because, we reply, it belongs to all of us to rejoice in the good, and to pursue and love it, and to turn away from the bad, and detest it and flee from it, paying no regard to whether it is generated or not." Thus we destroy scorpions, poisonous spiders, and vipers100 even though they have become what they are by nature, and not as a result of themselves. Consequently we can with justification hate the wicked among men without further calculating that they are the active cause of their becoming thus. Furthermore, we put the irremediably wicked to death for three good reasons: (i) so that they cannot do us wrong while they are alive; (ii) so that those like them will be afraid lest they too be punished by those they have wronged; and (iii) because it is better for these people themselves to die given that their souls are in such a state of corruption that not even the Muses themselves could educate them, nor could they be made better by Socrates or Pythagoras.101
(i) corresponds to (a); (ii) to (b); and, as we are dealing here with the irremediably reprobate, (c) does not in this case apply, although of course as we have seen it does elsewhere. It is (iii), however, and 99
100
101
Here I disagree with the emphasis at least of Lloyd (1988) pp. 36-7; referring to this passage, he claims that it shows that 'Galen is concerned to disassociate himself from a straightforward physical determinist position' on the grounds that '[he] rebuts the charge that on his view no-one may rightly be praised or blamed'. On the contrary, his position is compatible with the hardest determinism you care to espouse, and his motivation in this passage is precisely to show that whatever view one takes of causal determination, you can still rescue a notion of responsibility that will do the work you require of it in forensic and moral contexts. Similarly Pigeaud (1981) (having quoted with approval Daremberg (1865) p. 85) writes: 'cette contradiction [sc. between an intrinsic and unalterable moral nature on the one hand, and a determinist materialism on the other] existe. On doit bien la constater' but this seems to me to be quite wrong. Cf. Seneca de Ira 1.16.5 (on the proper attitude one should adopt to punishment): et cum cervicem noxio imperabo praecidi et cum parricidas insuam cuello et cum mittam in supplicium militare et cum Tarpeio proditorem hostemve publicum imponam, sine ira eo vultu animoque ero, quo serpentes et animalia venenata percutio. The reference to Pythagoras is not mere rhetoric: Galen himself makes use of the repetition of Pythagorean maxims as part of the process of moral self-improvement: Aff.Dig. v.30; note also the maxim 'choose the life which is best - living with it will make it pleasant',
Actions and passions
219
the line of reasoning that has led up to it that is of particular interest. In effect, Galen denies that our rationale for punishment, indeed for execution, need have anything to do with determining desert on the basis of a notion of responsibility strongly dependent upon free actions. The wicked are just like poisonous animals - no one would hold them responsible in the strong sense for what they are, but we do not destroy them because they are responsible for what they are, but simply because they are what they are. Their natures alone, whatever their causal provenance, make them fit subjects for eradication. In the case of wicked human beings, the case is given a powerful twist in the tail by the contention that such a life is objectively not worth living. Here as elsewhere individuals' subjective states can simply be ruled out of account: given that the evaluation of lives is strongly objective for a B-strategist like Galen, whether a particular form of life is pleasant from the inside is quite irrelevant - moral sickness, like any other form of disease, can be imperceptible to its sufferer.102 quoted approvingly by Galen at ibid. 33, attributed to Pythagoras by Stobaeus, Anth. 102
11.1.29.
I should stress that here too Galen's position has powerful Platonic antecedents: in Gorgias 476a~78a, Plato argues that punishment is in the interest of the punished; and he distinguishes between curable and incurable states of the soul: 478eff., 525bff.; however in the latter passage he seems to hold that punishment of the irremediably reprobate is to be justified on deterrence grounds, and not because it is in the interest of the individual punished. At Rep. 111.4x^-410a, Plato hints at the analogy between mental and physical health, but makes this type of mental health (i.e. justice) rather the preserve of the lawmakers and judiciary - none the less he does hold (410a) that those who are physically weak should be allowed to die, while those who are 'incurably evil in their souls' should be executed. And at iv.445a-b it is assumed that if we can show that life with a corrupted body is not worth living, then it will follow a fortiori and without further examination that life with a corrupted soul is equally worthless (cf. rx.59ib: all these amount to a justification of type (iii)); for an examination of Plato's position in the Republic, see Kenny (1973) pp. 1-27. However, the most obvious source for Galen's views is surely Laws v, which, as we have seen (n. 30), he is familiar with and quotes at length. The distinction between remediable and irremediable ills is made at 73id, where those suffering from the former are to be pitied, while those in the latter state should be the objects of anger; earlier, at 728D-C, Plato details the increasing wretchedness of the man who falls in with evil company, and gradually becomes like them, describing this as the 'greatest penalty' (megistedike); and he distinguishes between the man who is cured of this affliction, and the one who is not, who is 'destroyed in order that many others might be saved' (i.e. justification of type (a), or (i)). However, a theme common to the Republic (ni.4O9aff) and the Laws is the analogy with stock-breeding and the emphasis on the need to improve the blood-stock of the state by weeding and breeding out the undesirable elements (Laws v.735bff.): this type ofjustification is, so far as I know, absent from Galen's treatments of these subjects. As Mackenzie (1981) writes, p. 190: 'Plato's priority [sc. in the Laws] seems to have changed from the early period. He considers the problem of the incurable criminal now from the point of view of the state as a whole: if the person is a danger, that is sufficient reason for his removal.' In Book IX.853D—c, Plato returns to the topic of the sanctions of the
220
JAMES HANKINSON
I do not endorse that position: but it is at the very least consistent, in ways in which naive accounts of freedom and responsibility simply are not; and it avoids the manifold, perhaps insuperable, difficulties which assail any attempt to make the actions of human rational agents somehow causally free of the world around them. Galen's notions of desert and punishability may seem to us rebarbative and inhumane (although if they do, they do so more because of the type of punishment envisaged and its severity, which is of course in a sense incidental to the account, than because of their rationale) - but they are not negligible. vi
CONCLUSIONS: THE DIRECTION OF PSYCHOLOGICAL CAUSATION
Finally, I want to suggest some answers to a difficulty which has been lurking uncomfortably in the background of the discussion for some little time. The whole thrust of QAM is to show how psychological states supervene upon, and are caused by, physiological ones. criminal law, and introduces the concept of deterrence (type (b) or (ii) justification) before embarking on the question of whether the lawmaker should educate or coerce his subjects (857c-63b), a passage which recalls the earlier comparison of empiricist with rational medicine (iv.72oa-e): and the upshot is that, in the case of the better natures (those which are not damaged beyond repair, and which are receptive to such things) education (i.e. reform, type (c) justification) is appropriate, while for the others (which are not irremediably damaged, but are beyond the call of reason) one is reduced to coercion, of a type that will involve considerations of deterrence. In the limiting case, of course, here as elsewhere if the case is hopeless, then there is nothing for it but death and banishment. Thus Plato's account, particularly the mature, synthetic picture of the Laws, serves as a powerful model for Galen - but he is not slavishly tied to it; and it is noticeable that he invariably places more emphasis on the notion of medical therapy, which frequently for Plato (the Timaeus passages, quoted above n. 30, are an exception) functions as nothing more than a metaphor. Mackenzie (1981) (I am greatly indebted to her clear and philosophically sensitive presentation of the issues, particularly in chs. 10-12), holds that ultimately Plato finds it difficult, perhaps impossible, coherently to reconcile inherently conflicting views of the nature of punishment (essentially the conflict is between considerations of type (c) and retributivism). It is worth stressing that Galen never adopts a pure retributivist theory, of the type which one might think at best justified by a principle as strong as (7*), and as such is not an obvious prey to such an incoherence. Furthermore, however, if one justifies retributivism on the grounds that it is better for the objects of punishment that they undergo the punishment in question (and hence the retributivism is no longer pure, but derived), then that alleged incoherence starts to dissolve - and the dissolution is complete if one insists as Galen does (and perhaps Plato does not) that the medical model is more than merely an analogy or metaphor. Plato would perhaps have done better to concentrate on the Timaeus (86b-87b, 89b—c) treatment of criminal behaviour as a disease; and as Saunders (1983) points out in his review of Mackenzie's book, perhaps Mackenzie should have taken it more seriously on Plato's behalf. I am grateful to D. Gill for stressing the importance of the Laws passages in this connection.
Actions and passions
2 21
The analysis is rigorously carried through: not merely adventitious and evanescent things like moods and feelings, but the basic facts of our character and dispositions are dependent upon the body, its structure^ and its environment Galen can sidestep the difficulties this view might create for a strictly Aristotelian account of responsibility; but two questions suggest themselves. First, how can that general picture of the direction of psychological causation be made compatible with the notion that our rational selves initiate actions, cause things to happen, are responsible (in more than merely a mediate sense) for our behaviour? And second, how can purely mental events, like the decision to initiate a course of moral self-improvement, cash out in what must be, on the QAM view, a physical alteration, namely a change in the physical substances upon which the structures of our desires and passions depend? First of all, it is worth stressing that for Galen at least, given the materialism of his psychology in QAM and elsewhere, 103 there is no generalized, Cartesian problem of interaction, no need to resort to pineal glands. Thinking and deciding are, like everything else of causal significance, material processes. Hence the second of those questions is improperly formed. The appropriate answer must be (although it is never worked through in detail by Galen, and is indeed enormously hard to work through) that the decisions themselves cause alterations in the physiological states, either directly or mediately (by deciding to refrain from the consumption of certain types of food, let us say). 104 So what of the first question? What follows is a sketch only. But, I suggest, the following general account is not incoherent. The sort of person one is directly depends upon one's physiological structure; hence one's dispositions, including one's dispositions to have certain types of thought, are at least a partial function of that structure. 103
I discuss Galen's attitude to the nature of the soul in detail in Hankinson (1991b); M. Frede objected that Galen's caution about the soul's real nature (see PHP v. 791-5; QAM iv.772—3; On the Formation of the Foetus iv.700-2; Subst.Nat.Fac. iv.759, 763-4) makes
104
it inappropriate to describe him as a materialist. Galen is cautious: but he holds that at the very least there is some connection between body and soul (see Lloyd (1988) p. 37ff.); and avows that he finds it difficult for a variety of reasons to think that an immaterial (and hence divine) soul could be under the thumb of a material body in this way (QAM iv.788; cf. 776-7). Consequently, the evidence seems strong enough to suggest that Galen is a materialist, in spite of the fact that he is quite unclear about what sort of material the soul could be made of. Here again there are intriguing modern parallels: recent research on the immune system suggests that its effectiveness may be dependent upon individuals' states of mind; this type of interactive effect can be explained on the model I offer for Galen below.
222
JAMES HANKINSON
Thus I could not (for instance) form the second-order desire to give up drinking and smoking unless my body were composed in some determinate way; of course, its physical structure also accounts for my rampant first-order desires for brandy and cigars. But, if suitable conditions are realized (and these for Galen are not merely internal, but involve among other things having the right type of friend and associate around to supply positive and negative reinforcement),105 then our decisions in accord with our second-order desires become part of a causal process leading to the alteration of our first-order desires. Everything in that causal process is material, and antecedently caused (that poses no problem for Galen's theory); part of the process takes place at the conscious level, but it need be no less material for that. We have, effectively, a model in which the overall state S of some system generates outputs 0 which causally contribute to the creation of a new state S'. The system, as we would say, exhibits positive feedback. Galen clearly viewed the human organism as such a system; equally he thought that some of its states were clearly better objectively than others. In some systems, there will be a natural tendency to progress towards the better states;106 in others, the system will have become too corrupt to be self-ameliorating, and indeed even beyond external repair (just as diseases can progress to the stage where they are incurable). Our conscious states are part of (i.e. causally connected with) the system as a whole, hence our decisions feed both into and out of the system's states. Whether Galen ever really espoused that picture in its full generality is unclear. That he is committed to it seems obvious enough on the basis of the evidence I have presented. That it is both coherent, and a sophisticated and explanatorily powerful model of the role of the mind in our physiological functioning, also seems clear enough. I submit, independently, that it, or something very like it, is likely to be true. Whether Galen realised that - and whether he was better able to purge himself of the deeply rooted fallacy that the phenomenology of free decision-making simply rules out of account any such instrumental account of deliberation and action than some of his modern descendants - I do not know. 105 106
Aff.Dig.v.y-i^ 30-1. All of this is consistent with Galen's strong general teleology: see Hankinson (1988a) and (1989)-
PART IV
Stoic Psychological Concepts
CHAPTER 8
De la 'nature phantastique' des animaux chez les Stoiciens Jean-Louis Labarriere
'Si nous voulons apprendre ce qu'est rhomme, nous devrons savoir d'abord ce qu'est 1'animal.' Sextus Empiricus, AM vm. 87
Quiconque s'en va sejourner quelque temps du cote des Stoiciens apprend bien vite que, selon eux, les hommes, dans leur quete du bien, se sont tout d'abord, comme les animaux, laisses conduire par leur inclination {horme) a rechercher le premier approprie {to proton oikeion), a savoir ce qui est susceptible de preserver la constitution (sustasis) a laquelle ils sont naturellement attaches par un lien tout autant affectif que providentiel. Bien vite egalement il apprendra qu'a la difference des animaux prives de logos, les hommes, dont 'la raison est comme un artisan qui s'ajoute a l'inclination' (Diog. Laer. VII. 86), n'en resterent pas la, ce qu'expliquent aussi bien la celebre metaphore de la recommandation que l'acquisition de la notion de bien chez Ciceron (de Fin. m. 20-3, 33-4). Or, si Ton s'est souvent interesse au role que pouvaient jouer les animaux, voire les enfants, dans la theorie de Yoikeiosis — leur fonction est de montrer que le premier approprie n'est pas le plaisir, mais aussi qu'il existe un sentiment naturel d'attachement a sa progeniture - on s'est en general moins interesse pour elles-memes aux fonctions ou facultes cognitives mises en oeuvre par les animaux dans cette recherche de l'approprie. Ce fut pourtant l'objet d'une belle polemique entre Stoiciens et Academiciens.1 La question, lourde de sens pour la cohesion du Je tiens a remercier les participants au Symposium pour leurs critiques, conseils et encouragements. Je pense plus particulierement aux discussions avec B. Inwood, A. Long et R. Sorabji, mais rien de tout cela n'aurait ete sans la fidele amitie de J. Brunschwig. 1
Cf. Philon d'Alexandrie, Alexander ou de Animalibus, de Providentia; Plutarque, de Sollertia Animalium, Bruta Animalia Ratione Uti sive Gryllus, de Esu Carnium Orationes 11; Sextus
Empiricus, Hypotyposes Pyrrhoniennes 1. 40-78; Porphyre, de Abstinentia in. Ces textes, d'une
225
226
JEAN-LOUIS LABARRIERE
systeme stoicien, etait en effet de savoir si les actes qu'accomplissent les animaux temoignent ou non en faveur de leur possession du logos, ce qui ne saurait etre sans consequence sur le probleme de la justice, puisque, si les animaux avaient part au logos, alors il y aurait lieu d'entretenir des rapports de justice avec eux, contrairement a la doctrine professee par Chrysippe.2 Autant dire que c'etait alors la doctrine de la Providence qui se trouvait menacee ainsi qu'en temoigne ce fait que Philon d'Alexandrie ecrivit les deux livres de son de Providentia pour reduire au silence ses detracteurs, a commencer par les Neo-Academiciens ici representes par Alexandre, son neveu-apostat. Epictete (Diss. 1.16; 11. 8) et Origene (Contra Celsum iv. 79-82) soutiendront egalement qu'accorder le logos aux animaux revient a en faire nos egaux et done a nier l'oeuvre de la Providence. Rien d'etonnant des lors, tant sont lies les divers pans du systeme, a ce que cette menace etende ailleurs son empire, ainsi que semble s'en etre avise Sextus Empiricus lorsque, dans sa discussion de la conception stoicienne du signe, il remet en question que son existence suive de l'idee de consequence (ennoian akolouthias) que l'homme seul possederait, la Providence l'ayant distingue de toutes les autres creatures, lesquelles sont au mieux douees de la 'presentation simple' (phantasia haple), tandis que l'homme est equipe de la 'presentation transferentielle et synthetisante' (metabatike kai sunthetike, AM vm. 275-6 et 285-6). 3 Dans le meme sens, au sein d'un developpement destine a brocarder les Dogmatiques et donne 'par surcroit' (eh periousias, HP 1. 62-78), fameux parce que Ton y trouve une des mentions selon lesquelles Chrysippe aurait soutenu que le chien fait usage du cinquieme indemontrable (69), Sextus renforce ses critiques a l'egard du critere stoicien en rappelant que le premier mode de la suspension du jugement dont l'expose proprement dit couvre les §§40-61, revient a conclure a la non-superiorite des phantasiai humaines sur celles des animaux reputes 'irrationnels', en raison de extraordinaire richesse, meriteraient une etude approfondie. Ne pouvant m'y livrer ici, je me bornerai a y introduire en renvoyant a un proche avenir leur analyse detaillee. Cf. Plutarque, Soil. 964A-F, 970 B; ESU 999A-B; Sextus, AM ix. 131. On se reportera egalement sur ce point a la belle anecdote montrant Cleanthe assister a l'echange que des fourmis, etrangeres les unes aux autres, firent entre un cadavre et une larve. Cf. Elien, NA vi. 50 et Plutarque, Soil. 967E ( = 6TF1.515). Cf. infra p. 235 et n. 19.
De la enature phantastique3 des animaux chez les Stoiciens
227
la diversite des phantasiai a propos des memes objets, diversite due a celle-la meme des animaux. 4 La nature meme de ces enjeux, leur horizon polemique, m'ameneront done a m'interroger ici sur la liaison chez les animaux entre leurs capacites cognitives et leurs capacites pratiques. Soit cette question: une fois admis que phantasia et horme distinguent canoniquement les animaux des plantes — 1'unite des premiers est 'tendue' par une psuche tandis que celle des secondes ne Test que par une phusis - comment comprendre la nature des phantasiai chez des etres prives de logos? Comment produisent-elles leurs hormai? Quelles connaissances, et quel degre de reflexivite, si reflexivite il y a, leur apportent-elles? Conjuguees avec les inclinations naturelles, ces phantasiai suffisent-elles a rendre compte des actes animaux, ou bien faut-il supposer qu'il existe dans le quasi-hegemonikon animal quelque chose comme un assentiment 'rudimentaire'? L ' A M P L E U R DU PROBLEME
La mise en parallele de deux passages celebres, Tun d'Origene (de Princ. in. 1-5), l'autre d'Epictete (Diss. 1. 6), 5 nous permettra de mesurer la question dans ses dimensions a la fois physiques, epistemologiques et pratiques, puisque nos deux auteurs y rapportent les actes des animaux a la nature de leur constitution et de leurs presentations en marquant nettement leurs differences tant avec les humains qu'avec les plantes. Ainsi Origene, fidele a la distinction stoi'cienne entre hexis, phusis et psuche — et au sein de cette derniere entre ce que nous pourrions appeler la psuche aisthetike et la psuche logike- rapporte une doctrine que Ton peut representer par un arbre ou les numeros (1) (2) (3) (4) indiquent les quatre sortes d'etres auquels nous avons affaire: 4
5
On remarquera le soin avec lequel Sextus parle d'animaux 'dits non-rationnels' (HP i. 61, 62, 74, 75). La meme chose vaudrait de Porphyre, mais il leur arrive d'employer tout simplement l'expression traditionnelle ta aloga (//Pi. 60, 62, 77, 78; de Abst. 11. 2, 3 = N. 133). On se reportera sur ces points a Annas & Barnes (1985), pp. 41-8. Pour Origene, cf. Inwood (1985), pp. 21-7 et 78-9. Pour Epictete, cf. Long (1974) pp. 174-5.
228
JEAN-LOUIS LABARRIERE
etres mus i du dehors, exothen (les transportables doues de hexis, soit les mineraux) 2 d'eux-memes, ex heauton = ta apsucha, les plantes
du dedans, en heautois (phusislpsuche) par eux-memes, aph'heauton = ta empsucha 3 alogoi, les
animaux
4 logikoi, les
humains6
Animaux et humains different done des plantes en ce que, bien que tous aient la cause (aitia) de leur mouvement 'en eux-memes', les etres inanimes, e'est-a-dire les plantes, sont mus 'd'eux-memes', tandis que les etres animes sont mus 'par eux-memes'. Reste maintenant a distinguer parmi ces derniers entre les animaux et les humains, mais aussi a resorber l'objection de ceux qui voudraient doter certains animaux d'une intelligence, en raison de leurs comportements, voire de l'art qu'ils y deploient. Tel est le tres explicite propos d'Origene: Et les etres animes sont mus par eux-memes, car une representation se produit en eux qui provoque l'impulsion [phantasias egginomenes hormen prokaloumenes]. Mieux, chez certains animaux, des representations se produisent provoquant l'impulsion, la nature imaginative [phuseos phantastikes] actionnant l'impulsion de facon ordonnee [tetagmenos kinouses ten hormen]: ainsi chez l'araignee la representation de tisser se produit et l'impulsion a tisser suit, la nature imaginative la provoquant a cela de fa^on ordonnee, car l'animal n'a rien recu d'autre que cette nature imaginative. La meme chose se produit chez l'abeille pour fa^onner la cire. Mais l'animal raisonnable [logikon zoion], outre la nature imaginative, possede la raison [logon] qui juge les representations [ton krinonta tasphantasias], refuse les unes et accepte les autres, pour que le vivant se conduise selon elles. (de Princ. in. 1. 2~3 = S V F I I , 988)7 Le logos discriminateur vient done marquer la difference humaine et, en consequence, meme les animaux qui paraissent avoir des 6
7
Dans le de Oratione vi. 1 ( = SVFn. 989), Origene precise que le mouvement des humains est 'di'heauton*. Long & Sedley (1987) font a juste titre remarquer que Origene dit 'je pense' a propos de cette denomination, et qu'en consequence elle lui est peut-etre propre, (vol. 11, 53A, p. 310). Traduction de H. Crouzel et M. Simonetti, Editions du Cerf, Paris 1980.
De la 'nature phantastique' des animaux chez les Stoiciens
229
conduites reflechies sont bornes a leur seule 'nature phantastique'. 8 En effet, bien que les phantasiai soient, chez certains animaux, endogenes et non pas seulement exogenes, il n'en reste pas moins qu'ils les suivent sans jamais les mettre a distance et les passer au crible d'une quelconque 'faculte de juger'. Or, la est le point principal puisqu'il s'agit de soutenir que l'ordre et la regularite qui president aux actions de certains animaux, sont incommensurables avec ce logos, qui, puissance d'affirmation et de negation, nous rend responsables de nos actions, quelles que soient la nature et l'origine de nos phantasiai (ibid. 3-5). II faut done, ainsi que le dit le Contra Celsum iv. 81, a propos des abeilles et des fourmis, bien comprendre 'la superiorite des actions accomplies par raison et reflexion [apo logou kai logismou] sur celles qui proviennent d'une nature sans raison [alogouphuseos] et de sa simple constitution [kataskeues psiles]'. Aussi,
si Ton peut trouver chez certains 'une nature sans raison qui, meme sans raisonner, assiste les etres qui n'ont pas merite d'avoir la raison' (ibid.), il n'en reste pas moins que ces animaux restent des animaux et que leur 'nature phantastique' n'excede en rien leur 'nature animale', laquelle ne contient rien de plus que la sensation, la phantasia et l'impulsion. C'est pourquoi, par dela la difference relative a l'origine des phantasiai chez certains animaux, la notion de 'nature phantastique' me semble pouvoir etre etendue sans grand dommage a l'ensemble des animaux et servir a les distinguer de la 'nature simple' des plantes tout autant que de la 'nature rationnelle' des humains, seuls ces derniers devant ce qu'ils sont a autre chose qu'a leur seule nature. En effet, tel sera le theme constant de la reponse des Stoiciens a leurs opposants ainsi qu'outre Origene, en temoignent aussi bien Plutarque et Porphyre que Philon dans sa reponse a Alexandre: les animaux agissent en tout grace a la nature qui est en eux, ils n'inventent ni n'apprennent rien et il est en consequence inutile - et impie - de leur supposer un quelconque logos, puisque, de ce point de vue, certaines plantes agissent comme eux, et pourtant elles ne sont pas douees d'ame! 9 Les choses sont d'ailleurs telles qu'on pourrait se demander si, pour les Stoiciens, un peu a la maniere de 8
9
Ce sont la, bien entendu, des exemples traditionnels, mais ils avaient du etre retravailles par les Stoiciens et les Neo-Academiciens dans leurs polemiques puisqu'on les trouve developpes chez Philon, cf. Prov. i. 25, 52; 11, 40; Anim. 17-21, 42, 77. Cf. Philon, Anim. 77-81, 92-5; Porphyre, de Abst. in. 6. 4 ( = N. 194); 10. 1-3 ( = N. 199—200); 19. 1-2 (= N. 208-9); Plutarque, Soil. 968c, 973E; Bruta 991F-992A; Origene, de Princ. in. 4-5.
23O
JEAN-LOUIS LABARRIERE
ce qui sera chez Descartes, l'art et l'extreme regularite que certains animaux manifestent dans leurs actions, temoignent non qu'ils ont moins de raison que nous, mais 'plutot qu'ils n'en ont point, et que c'est la nature qui agit en eux selon la disposition des organes' (Discours de la Methode v). 1 0
Tout le probleme, nous le verrons par la suite, est alors de parvenir a penser un mouvement 'par lui-meme' qui doit tout a la nature et rien a une quelconque conscience ou reflexivite, puisque aussi bien Origene (de Oratore vi. 1 = SVFu. 989) qu' Epictete (Diss. 1. 6. 13), si on doit les tenir pour de fideles temoins de l'orthodoxie, rapportent la parakolouthesis, la 'conscience reflechie' ou 'intelligence', au logos propre a l'homme. Rappelant que l'usage de la pensee 'par inference' est plus propre a l'homme que la simple reception des empreintes dans la rencontre des objets sensibles (ibid. 10) et qu'honorer la Providence consiste a faire usage d'une telle pensee, Epictete, peut-etre plus radical encore qu'Origene, assied son argumentation sur la distinction entre usage (chresis) et intelligence ou conscience reflechie des phantasiai. Aux animaux le premier, aux hommes la seconde. Ainsi en a voulu Zeus qui a donne aux hommes un 'equipement' (paraskeue, 37) susceptible de le distinguer des autres animaux, lesquels n'ont done pas conscience de ce qui leur arrive (13). Quant a leurs fonctions, qu'il s'agisse de leurs fonctions naturelles (14) ou de leur destinee (18: servir de nourriture, de bete de somme, a faire du fromage . . . ) , elles ne necessitent rien d'autre que l'usage des phantasiai, puisque la constitution des animaux est adaptee au seul usage (17). Maintenant, si Ton se tourne vers cet usage lui-meme, s'il renvoie sans doute chez Epictete au simple fait d'etre impressionne lors de la rencontre des choses sensibles, voire, si Ton veut pousser le parallele entre Origene et Epictete, egalement a l'usage de phantasiai endogenes, alors force est de constater que cet usage est defini bien negativement, puisque Epictete soutient que pour accomplir les fonctions que les animaux ont a accomplir, il n'est nul besoin de pouvoir comprendre (parakolouthein) etdiscerner (diakrinein) les phantasiai (18). Tout comme chez Origene, les animaux sont done depourvus de toute espece de faculte critique vis-a-vis de leurs phantasiai, ce qui les condamne a les suivre de fafon mecanique, aveugle, puisqu'ils n'ont pas besoin, a la difference des humains, d'en com10
Ce n'est pourtant pas une raison pour, tel Inwood (1985) p. 78, rendre a mon sens un peu durement tetagmenos par automatically.
De la (naturephantastique3 des animaux chez les Stoiciens
231
prendre l'usage (13). Voila qui revient, sans nul doute, a les priver d'assentiment, ce qui semble orthodoxe des lors qu'on se souvient que c'est a des axiomata que l'homme donne son assentiment, or cela ne saurait etre sans qu'il soit doue de logos.11 Si done Ton admet qu'en raison de la doctrine de Voikeiosis, il doit exister chez les animaux quelque chose comme un quasi-hegemonikon, point focal ou sont ramenees les sensations afin que 1'animal soit a meme de satisfaire les exigences de son attachement a sa constitution, alors ce centre ne semble devoir comprendre, ainsi qu'en temoignent la majorite de nos sources, que deux instances, l'inclination et laphantasia (ou la sensation). 12 Des lors, quelque sens que Ton donne a la conscience qu'ont les animaux de leur constitution, voire de leur attachement a cette constitution, et ce, nous le verrons, quel que soit le terme que Chrysippe ait reellement employe, suneidesis ou sunaisthesis, cette conscience, a suivre Epictete et Origene, ne devrait dependre que de ce que Ton peut appeler leur 'nature phantastique'. Comprenons qu'elle ne saurait s'interpreter comme un pouvoir de mise a distance des phantasiai, comme une faculte de triage, puisqu'un tel pouvoir excede leur nature. De la qu'a priori on doive s'attendre a trouver dans la phantasia (ou dans la sensation) de quoi donner sens a cette conscience. Ce qu'il reste a verifier. UNE TENTATIVE DE RECONSTRUCTION DE LA PHANTASIA DES ANIMAUX
Une premiere constatation s'impose: nos sources sont excessivement minces sur ce point et rien n'assure que cette minceur soit principalement due a I'etat de nos sources. En temoigne a mon sens de fafon significative le fait qu'au livre second des Premiers Academiques, Ciceron, a une notable exception pres (37-9), ne fasse jamais appel a la phantasia des animaux, alors que c'est la un des themes chers au premier mode de la suspension du jugement chez les Sceptiques (HP 11. 26). En ce sens, peut-etre pourrions-nous supposer de la part des premiers Stoiciens une certaine indifference a ce sujet, laquelle, outre un ininteret pur et simple, pourrait tres bien se fonder sur une 11 12
Cf. Stobee, Eel. n. 7, 9b (W. p. 88). Pour Vhegemonikon cf. Arius Didyme apud Eusebe, Prep. Ev. xv. 20-5 ( = SVF 11. 821); Ciceron, ND 11, 29; Philon, Anim. 29; Sextus, AM ix. 102, 119. Pour la phantasia et Vaisthesis, cf. Philon, Leg. 1. 30; 11. 23 (= SVFn. 844 et 458); Ps-Galien, Intr. xiv. 726. 7-11 (= SVFn. 716); Ciceron, ND 11. 34, 122, 133; Off. 11. 11; Calcidius, in Tim. 220 ( = SVFu. 879). On se reportera egalement a Inwood (1985) p. 25 avec les notes.
232
JEAN-LOUIS LABARRIERE
assertion de ce genre: 'Des animaux, que pouvons-nous dire, sinon ce que nous en apprend la Providence?' Rien n'interdirait alors de penser que, par dela le chien de Chrysippe ou les fourmis de Cleanthe, les Sto'iciens auraient ete contraints de rompre avec cette indifference initiale afin de repondre aux critiques que leur adressaient les Academiciens. De cela, temoigne sans doute le fait qu'Antipater ecrivit un livre sur les animaux. 13 Quoi qu'il en soit, une des mentions les plus claires de laphantasia des animaux marque une certaine indecision, ou du moins un certain vague: En outre, parmi les presentations, les unes sont rationnelles [logikai], les autres irrationnelles [alogoi]; rationnelles, celles des animaux rationnels [ton logikon zoion], irrationnelles, celles des irrationnels [ton alogon]; les rationnelles sont des pensees [noeseis], tandis que les irrationnelles n'ont pas re^u de nom [ou tetuchemasin onomatos]. (Diog. Laer. vn, 51)
De ce passage, tout ce que nous pouvons tirer avec certitude relativement a notre objet, c'est que, si les animaux possedent bien des phantasiai, elles ne sont ici determinees que negativement, par privation. 'Irrationnelles', ce ne sont pas des 'pensees', mais on ne sait trop ce qu'elles sont puisqu'elles n'ont pas d'onoma pour les specifier. En se souvenant de Diog. Laer. vn. 49, ou il est dit que la pensee est 'causeuse' (ehlaletike), nous pourrions sans doute soutenir que les phantasiai des animaux sont 'muettes' et non pas 'causeuses' puisqu'elles sont le fait d'etres alogoi. N'etant done pas des 'pensees', elles ne sauraient etre autres que 'sensibles', e'est-a-dire, d'apres Diog. Laer. vn. 51, 'saisies [lambanomenai] a travers un ou plusieurs organes corporels', puisque, toujours selon le meme paragraphe, les phantasiai non sensibles sont celles qui sont saisies 'a travers l'intellect [dia tes dianoias], comme celles des incorporels et des autres choses saisies par le raisonnement [logoi lambanomenonY. Mais, a la difference de ce qui se passe chez les humains, les phantasiai sensibles des animaux restent irrationnelles puisqu'ils sont eux-memes irrationnels, tandis 13
Cf. Plutarque, So//. 962F (= SVF111. 47); Aetia Physica 38. Tout ce que Ton sait de ce livre c'est qu'Antipater y faisait remarquer que les anes et les moutons n'etaient pas particulierement propres. Sans doute etait-ce afin de montrer qu'il ne fallait pas idealiser les animaux. Mais peut-etre pourrait-on se demander si Antipater de Tarse n'aurait pas applique le principe selon lequel la comparaison est inegale entre la masse des animaux et l'homme (Sextus, HP 1. 62-3). Certains Stoiciens auraient alors oppose la diversite a un argument 'massif, et Ton pourrait alors penser que le debat se serait poursuivi par la recherche, de part et d'autre, de contre-exemples, ce dont nous aurions une trace dans le de Soil, de Plutarque puisque son vrai titre est: 'Lesquels des animaux sont intelligents, les terrestres ou les aquatiques'.
De la 'nature phantastique3 des animaux chez les Stoi'ciens
233
que celles des humains sont deja rationnelles puisqu'ils sont euxmemes rationnels. 14 Autrement dit, etant donne que les animaux restent radicalement etrangers aux lekta, meme leur phantasia sensible differe de celle des humains puisque, chez ces derniers, la phantasia sensible se 'presente' necessairement, si Ton ose ainsi s'exprimer, sous la forme d'une proposition (Sextus, AM vm. 70). De la que, si humains et animaux sont chacun pour leur part doues d'une phantasia sensible, ils n'y accedent pas sous la meme modalite, le logos faisant ici son oeuvre et la non. On pourrait alors se demander si, non denommee, la phantasia animale n'aurait pas de phantasia que le nom. En effet, bien que de jure ce qui est dit d'un genre (ici la phantasia) doive valoir pour toutes ses especes, tout se passe de facto comme si en regie generale ce qui etait dit de laphantasia-genrt ne valait que pour une de ses 'especes', la phantasia logike, a vrai dire 'genre' elle-meme puisqu'elle comprend les diverses 'especes' de phantasiai humaines. On ne saurait done transposer ce qui est dit d'un genre, en fait genre autant qu'espece, a la simple phantasia alogos. Comprenons alors qu'il en irait pour la phantasia comme du terme 'psuche* dont Sextus nous dit que de subtils Stoi'ciens faisaient remarquer qu'il pouvait etre utilise en deux sens: soit pour designer l'espece de tension unifiante du compose, soit pour designer V hegemonikon (AM VII. 234). Or, il y a la quelque chose comme un double sous-entendu: (1) il va de soi (dans le contexte de la discussion) que psuche signifie hegemonikon; (2) rien de ce qui est dit du second sens ne vaut pour le premier. Voila qui, mutatis mutandis, vaut en regie generale des discussions sur la phantasia, ainsi qu'en temoigne AM vn. 227-62, puisque, etant donne que e'est en regie generale aux humains que Ton s'interesse, e'est alors de la phantasia logike que Ton parle quand Ton parle de la phantasia tout court. Toute transposition devient par la tres dangereuse car les Stoi'ciens se mefient comme de la peste du 'comme si'. En effet, si Ton en croit Plutarque, Philon et Porphyre, le 'comme si' ne devait aux yeux des Stoi'ciens surtout pas etre entendu comme invitant a rechercher chez les animaux une quelconque faculte susceptible d'en rendre compte, mais bien plutot comme signifiant qu'il n'y a rien de tel chez eux, comparaison ne valant pas raison. 15 14
15
Cf. Imbert (1978) p. 227. Long ne semblait pas initialement partager ce point de vue (cf. Long (1971) p. 83) mais il semble qu'il ait modifie son jugement, cf. Long & Sedley (1987) vol. 1, p. 240. Cf. Philon, Anim. 40 et 83; Plutarque, Soil. 961E-F; Porphyre, de Abst., in, 22, 5-6 ( = N. 215); SVFi. 65. II y a la, me semble-t-il, un argument susceptible de contrer Sorabji
234
JEAN-LOUIS LABARRIERE
La tache de l'interprete s'en trouve alors singulierement compliquee. J'en prendrai un exemple chez Sextus Empiricus. Presentant le point de vue des Stoiciens sur la phantasia aisthetike, il ecrit: Pour que se produise une presentation sensible selon eux, telle celle de la vue, il faut que cinq choses concourent, l'organe sensoriel, Pobjet sensible, le lieu, le comment, l'intellect [dianoian], puisque si Tun seulement est absent tandis que les autres sont presents, par exemple si l'intellect est dans un etat contre-nature, alors, disent-ils, 1'apprehension ne sera pas sauvee [ou sothesetai he antilepsis]. (AMvu.
424)
Puisqu'il s'agit ici de la phantasia aisthetike, que cette expression soit celle de Sextus ou celle des Stoiciens eux-memes, 16 la tentation est grande de vouloir transposer cela aux animaux. Mais comment faire, puisqu'ils sont par definition prives de dianoia? Faut-il, en tenant radicalement compte de cette privation, soutenir qu'il ne se produit pas chez eux de presentation sensible, ce qui semble difficile a admettre, ou bien faut-il penser qu'il ne s'en produit que de non susceptibles de 'sauver' l'apprehension? D'un certain point de vue, la seconde solution est sans doute preferable puisque, d'apres le contexte de la discussion, il n'est point de phantasia kataleptike sans apprehension sauve. Mais quel sens cela pourrait-il bien avoir aux yeux des Stoi'ciens, jeunes ou vieux, puisque, selon eux, les animaux sont en dehors du domaine de la verite? De ce point de vue, nous pourrions alors peut-etre soutenir qu'a leurs yeux, les animaux n'ont part aux presentations sensibles qu'autant que Ton insiste sur leur cote sensible et non sur leur cote presentationnel. 17 Rapportee a celle des humains, la presentation sensible des animaux pourrait a la limite etre tenue pour une phantasia akataleptos en ce sens que, si elle provient bien d'un objet reel, elle n'est 'ni nette ni bien gravee' (Diog. Laer. vn. 46). 18 De la que, nous l'avons vu, quand bien meme les comportements de certains animaux pourraient laisser croire a d'aucuns qu'ils se conduisent avec intelligence, leurs presentations n'en resteraient pas moins tres inferieures a celles des humains et done a 'dedaigner' (Philon, Anim. 95).
16
17
(1990) qui soutient que les phantasiai des animaux sont propositionnelles parce qu'il suffit qu'elles soient verbalisables et conceptualisables par nous. Lafindu §424 invite a rapprocher ce passage de ce que dit ailleurs Sextus de la clause ajoutee par 'les Jeunes Stoi'ciens': 'pourvu qu'il n'y ait pas d'obstacle' (AM vn. 253). Ce point est ici de peu d'importance puisque la presente discussion ne porte pas sur ce qui est propre aux Jeunes Stoiciens. 18 Cf. Inwood (1985) pp. 75-7. Cf. Frede (1983) pp. 65-93.
De la 'nature phantastique' des animaux chez les Stoiciens
235
Peut-on toutefois s'avancer un peu plus dans la determination de la presentation sensible des animaux? Nous savons par Diogene Laerce que 'la saisie du noir et du blanc, du rugueux et du lisse, se produit par la sensation [ginetai aisthesei\ (vn. 52) et ce sont les memes exemples que nous retrouvons quand Aetius definit la phantasia et lephantaston [Plac. iv. 12. 1-4 = SVF n. 54). Que laphantasia
soit canoniquement pensee en reference a la phantasia sensible et qu'elle revele elle-meme et sa cause, nous importera moins pour le moment que la relation qu'elle entretient avec la sensation. Soit la question: chez les animaux, le senti et le presente sensible se confondent-ils? Ainsi que le souligne fortement Ciceron, chez les Stoiciens, toute phantasia, a fortiori sensible, s'enracine dans la sensation, mais il n'en reste pas moins que la phantasia humaine deborde le cadre strict de la sensation, de la que 'cet objet est blanc, cela est saisi', dit-il, 'non pas directement par les sens, mais cependant en quelque fa^on par les sens', c'est-a-dire par 'un acte de comprehension de Fame (et) non par les sens' (Acad. Pr. 11. 20-2; 1. 40-42). Or, chez les animaux, cet acte de Fame ne saurait etre celui du logos. Des lors, ce que dit Sextus de la saisie des choses sensibles ne devrait pas valoir tel quel pour les animaux. S'attachant, il est vrai, a ce qui est connu de faf on evidente par les sens 'selon la rencontre' [kata periptosin, AM in. 40) ou 'selon l'approche' (kat'empelasin, AM
ix. 393), il precise: selon l'approche des choses evidentes, sont complies [noeitai] des choses comme le blanc et le noir, le doux et Tamer; car si ces choses sont bien sensibles, dies sont neanmoins con^ues. (AM ix. 394) Etant donne que les animaux sont depourvus de noesis, on voit mal comment ils pourraient concevoir quoi que ce soit. Reste done a s'efforcer de comprendre comment ils pourraient user de ce qui est connu par les sens 'selon la rencontre', sans toutefois concevoir. Tournons-nous pour ce faire vers ce qui peut etre considere comme la seconde grande mention de la phantasia des animaux, soit vers les paragraphes 275-6 de AM vm, la ou Sextus, relan^ant sa polemique contre la conception stoicienne du signe, rapporte que les Dogmatiques, c'est-a-dire ici les Stoiciens, distinguent entre une phantasia 'simple' (haple), qu'humains et animaux semblent partager, et une phantasia propre aux humains, laquelle est 'transferentielle et synthetisante' (metabatike kai sunthetike): Mais les Dogmatiques restent sourds a chacune de ces objections [ = celles au sujet de l'existence du signe], et afin d'etablir le contraire ils disent que
236
JEAN-LOUIS LABARRIERE
rhomme ne differe pas par le discours profere [toi prophorikoi logoi] des animaux irrationnels (car les corbeaux, les perroquets et les pies proferent des sons articules) mais par le discours interieur [toi endiathetoi], pas plus qu'il n'en differe par la presentation simple seulement [tei haplei motion phantasiai] (car ceux-ci ont aussi la presentation), mais par la transferentielle et synthetisante [alia tei metabatikei kai sunthetikei]. Cela revient a sa possession de l'idee de consequence [ennoian akolouthias] par laquelle il se saisit immediatement du concept de signe (semeiou noesin). Car le signe est lui-meme de la forme: 'si ceci, alors cela'. Done l'existence [huparchein] du signe suit de la nature et de l'equipement [kataskeuei] de l'homme. En laissant ici comme hors de propos ce qui concerne le signe et la distinction entre les deux types de logoi, remarquons que le sens obvie de la distinction pretee aux Stoiciens est de distinguer les phantasiai produites 'par inference' de celles qui ne le sont pas et sont done par le fait 'simples' puisqu'elles ne sont pas 'complexes'. A la suite d'A. Long et de M. Burnyeat qui tous deux le soulignent, quoique avec quelque difference, il faut bien entendu rapprocher metabatike kai sunthetikedt ce que nous savons par ailleurs des notions. 19 J u s q u ' a nouvel ordre, l'une des mises au point les plus claires de cette question me semble avoir ete celle d'une note de Rodier. 2 0 J e mettrai done, sans plus les rediscuter, ses conclusions sous forme de tableau:
Ennoiai
ennoiai adventices, formees naturellement (phusikos, anepitechnetds) = prolepseis
immediatement par la sensation dans la rencontre (periptosis)
par la pluralite et la repetition des experiences (empeiria) ennoiai factices resultant d'une inference operee sur les sensations claires (metabasis)
19 20
par ressemblance
par synthese (kata
(kath'homoioteta)
sunthesin)
par analogie f
kat'analogiari)
par augmentation^
par diminution
(auxetikos)
(meiotikos)
Cf. Long (1971) p . 87 (avec les notes); Burnyeat (1982) p p . 206-8; Long & Sedley (1987) vol. 11, 53T, p. 319. Cf. Rodier (1926). II s'agit de la note 4 p p . 276-8.
De la 'nature phantastique' des animaux chez les Stoiciens
237
Si Ton reporte cela a la distinction des phantasiai, alors il s'agit bien d'opposer ce qui est produit mediatement, 'par inference', a ce qui se produit naturellement, soit, pour parler comme Aetius, ce qui est produit 'par notre instruction et attention [di'hemeteras didaskalias kai epimeleiasY (Plac. iv. 11. 3) a ce qui se produit 'naturellement et sans travail' (ibid.). Or, ce qui se produit ainsi tout seul et sans effort, c'est, toujours selon Aetius, ce qui se produit par la sensation, la memoire et l'experience. Que les animaux possedent ou non effectivement la memoire, voire l'experience, chez les Stoiciens,21 importe assez peu ici puisque, de toute fa^on, la phantasia simple qu'ils possedent seulement ne peut etre que celle qui procede de la nature et ne demande aucun effort, aucun travail, et non pas celle qui demande instruction et attention. De la, ainsi qu'il a deja ete signale, le theme constant de la replique des Stoiciens a leurs contradicteurs: c'est la nature et elle seule qui agit dans les animaux car ils n'apprennent rien. Si done Ton se demande comment les animaux peuvent user de ce qui est connu par les sens sans toutefois concevoir, la solution la plus sage me semble de se reporter a l'habituelle comparaison avec les enfants, comparaison d'ailleurs utilisee par Philon pour soutenir que les animaux sont denues de toute forme de pensee (Anim. 96). Nous savons que les enfants commencent, grace a l'acquisition desprolepseis, a acquerir quelque chose du logos des l'age de sept ans et qu'ils en ont la possession complete a quatorze ans grace a l'acquisition des notions de bien et denial, age dont Diogene de Babylone nous dit qu'il est celui de l'achevement de la voix humaine. 22 Au depart, s'il faut croire ce que dit Varron a propos de Chrysippe, les enfants n'ont qu'un quasi-logos (ut loqui) somme toute comparable au logos prophorikos des oiseaux (de Lingua Latina vi. 56 = SVF 11. 143),
autrement dit, faute de dire des choses en manipulant des lekta, leurs phantasiai ne sont que 'confuses et inarticulees' (Seneque, de Ira 1. 3. 7). Ces etres non stabilises et inacheves sont done comparables avec les animaux, a ceci pres qu'ils acquerront le logos. Maintenant, si Ton retourne la comparaison, en prenant bien sur soin de ne pas faire des animaux de futurs humains adultes, alors, 21
22
Nos sources sont assez divergentes sur ces points. Seneque (Ep. 121. 19-23) et Hierocles (Eth. 1. 51-3) accordent, par exemple, l'experience aux animaux, tandis que Plutarque (Soil. 961c), Porphyre (de Abst. in. 22. 5 = N. 215) et Calcidius (in Tim. 220) rapportent que les Stoiciens les privaient de memoire. Cf. D . L. VII. 55; Aetius, Plac. iv. 11, 4; Seneque, Ep. 121. 15; Giceron, de Fin., in. 21, 33-4.
238
JEAN-LOUIS LABARRIERE
puisque ce sont des etres animes, ils doivent posseder la connaissance sensible de ce qui est necessaire a la preservation de leur constitution. Soit des quasi-prolepseis formees 'naturellement et sans travail', contrairement a ce qui s'acquiert 'grace a l'attention et a l'instruction', puisque cela ne peut etre qu'apres un debut d'acquisition du logos. Mais ces quasi-prolepseis, pour que ce soient bien celles d'animaux et non celles de futurs humains adultes, ne doivent point etre congues a l'instar de celles qui pourront servir de base a l'acquisition du logos. De la, que ce ne soient que des qu&si-prolepseis restant definitivement enfermees dans l'ordre de ce qui se forme naturellement. De la encore, que si telle est bien la nature de la connaissance sensible des animaux, elle reste, malgre sa remarquable efficacite dans certains cas, 'obscure' car 'inexprimable' (Philon, Anim. 99). COMMENT ET DE QUOI LES ANIMAUX ONT-ILS CONSCIENCE?
Reste maintenant a determiner comment les animaux font usage de ces phantasiai, lesquelles ont done, par rapport a celles des humains, des capacites cognitives tres reduites, quand bien meme elles sont censees etre suffisantes pour produire des impulsions les menant a accomplir des actes appropries a leur constitution. Soit deux questions: (1) La definition de la phantasia donnee par Aetius (Plac. iv. 12. 1-3) et Sextus (AM vn. 162), ou elle est dite 'revelatrice d'elle-meme et de quelque chose d'autre [ = ce qui l'a causee]', est-elle applicable telle quelle a la phantasia des animaux? En effet, s'il semble aller de soi que des objets exterieurs se revelent aux animaux, en est-il de meme de leurs phantasiai? Autrement dit, la celebre comparaison avec la lumiere (phos), remarquablement 'eclairee' par A. Long et D. Sedley, vaut-elle de la meme fafon pour les humains et pour les animaux?23 Ce qui revient a ceci: a travers l'attention qu'ils portent aux objets sensibles, les animaux sont-ils egalement attentifs a leurs phantasiai et, si oui, sous quelles modalites? (2) Etant donne que Chrysippe semble avoir dit que 'le premier approprie a tout animal est sa constitution et la conscience qu'il en a' (Diog. Laer. vn. 85), quelle est done, qu'il ait ou non employe le terme de suneidesis, la nature de cette conscience? Signifierait-elle, 23
Cf. Long & Sedley (1987) vol. 1, p. 239.
De la 'nature phantastique' des animaux chez les Stoiciens
239
ainsi que le soutiendra fortement Hierocles (Eth. 1. 39-51), que Yaisthesis, puisqu'il n'emploie peut-etre pas a dessein le terme de phantasia, ne saurait etre bornee a la seule perception des objets exterieurs? En quoi cette conscience, suneidesis ou sunaisthesis, puisqu'elle semble devoir deborder le cadre de la stricte perception des objets exterieurs pour se porter egalement vers la constitution que Ton a en propre, est-elle tributaire de Fattention portee aux phantasiai, si telle attention il y a? J'aborderai ces questions, apparemment d'ordre plus epistemologique que pratique, a travers le probleme de Faction animale. En effet, puisque ici aussi 'la nature ne fait rien en vain', il faut se demander a quelle necessite peut bien repondre, chez les animaux, une telle attention et une telle conscience. La reponse ne saurait etre autre que celle-ci: a preserver leur constitution (Seneque, Ep. 121. 6—15). La question: 'de quoi les animaux ont-ils besoin pour ce faire?', entraine done que les questions d'ordre epistemologique et d'ordre pratique soient intimement liees puisqu'il s'agit, au sein d'une scala naturae, de marquer la place des animaux (dotes d'une psuche) entre les plantes (dotees seulement d'une phusis) et les humains (dotes d'une psuche logike). Autrement dit, aisthesis, phantasia et horme suffisent-elles a bien marquer cette place et a faire droit aux animaux les plus performants? Mes questions (1) et (2) s'entremelent done en ce que tout depend du contenu donne a Yaisthesis et/ou a la phantasia.24 En effet, si Ton y inclut quelque chose comme une conscience, alors cela signifie qu'une aisthesis et/ou une phantasia bornee a la seule perception ou revelation des objets exterieurs ne saurait suffire, meme conjuguee a une impulsion, a rendre compte de faf on satisfaisante de Faction des animaux. L'automaticite aveugle presidant a leurs actions risquerait en effet fort de les 'vegetaliser' en ramenant leurs mouvements 'par eux-memes' a de simples mouvements 'd'eux-memes'. De ce point de vue, si certes il faut se garder de trop accorder aux animaux, car ce serait alors la specificite humaine qui se trouverait menacee, les critiques que Hierocles adresse a ceux qui n'accordent a la sensation que la fonction de percevoir des objets exterieurs, peuvent fort bien se comprendre comme le refus de ne voir dans Yoikeiosis des animaux qu'une reponse automatique n'induisant aucune faculte cognitive.25 24 25
J'userai de la double coordination 'et/ou' afin de pourvoir inclure aussi bien les temoignages faisant etat de Yaisthesis que ceux faisant etat d ' u n e phantasia. Cf. supra n. 11. Cf. I n w o o d (1984) p p . 151-84.
24O
JEAN-LOUIS LABARRIERE
De la que Hierocles soutienne qu'en raison de l'etroite intrication (krasis) du corps et de Tame, la sensation de soi (aisthanesthai heautou) implique une attention (antilepsis) a toutes les parties de son corps et de son ame (Eth. 4. 38-53). J'en reviens ainsi a mes questions (1) et (2). Si en effet les phantasiai animales ne se revelent pas elles-memes a leurs recepteurs, revelation qui devrait entrainer chez ces derniers quelque forme de conscience, ne serait-ce que sous la forme d'un assentiment, lui, automatique, alors sont-ce bien encore des phantasiai puisqu'elles n'honoreraient qu'un cote, et encore le mineur, de la definition donnee en (1)? A l'inverse, si elles se revelent elles-memes a leurs recepteurs, alors quelle est la nature de cette 'revelation'? Impliquet-elle quelque chose comme une 'pause', sans que pour autant celle-ci soit une reelle mise a distance des phantasiai impliquant qu'on puisse n'y pas assentir? On sait que le propos prete a Chrysippe en (2) a suscite suffisamment de perplexite chez certains pour qu'ils proposent de lire sunaisthesis en lieu et place de suneidesis afin de ne pas doter les animaux d'un assentiment par trop humain. 26 Mais J. Rist a su attirer l'attention sur ce passage de Plutarque soulignant que dans Yoikeiosis il y a sensation et apprehension (aisthesis kai antilepsis, St. Rep. 1038 B-C), ce qui doit impliquer que, quelle que soit l'expression que Chrysippe a reellement utilisee, elle devait renvoyer non seulement a la simple sensation, mais encore a cet acte mental d'apprehension de ce qui est propre. De la que, selon J. Rist, il n'y ait pas lieu d'amender cet extrait. 27 On le voit done, tout le probleme est bien celui du contenu de Yaisthesis et/ou phantasia: soit on s'en forge une conception large, et alors on y inclut quelque chose comme une conscience; soit on s'en fait une conception plus restrictive, et alors vient se poser le probleme de savoir comment Ton rendra compte de Faction animale, surtout si, de plus, on prive les animaux de toute forme d'assentiment. De la, me semble-t-il, une alternative: soit Ton a une conception etendue de Vaisthesis et/ou phantasia, et alors il n'est pas necessairement besoin de supposer un quelconque assentiment chez les animaux, puisque quelque chose comme une conscience s'y trouve deja; soit Ton en a une conception plus restreinte, et alors, nouvelle alternative, soit Ton accorde une certaine forme d'assentiment aux animaux, ce qui pourrait sembler heterodoxe, mais 26 27
Cf. Pohlenz (1970) vol. 11, p . 6 5 . Cf. Rist (1969) p p . 4 3 - 5 versus Striker (1983) p . 145 n. 2.
De la 'nature phantastique' des animaux chez les Stoiciens
241
permettrait de rendre aisement compte de 1'action des animaux, soit a Pinverse, et semble-t-il de fagon orthodoxe, Ton refuse l'assentiment aux animaux et il devient alors difficile d'expliquer de fafon satisfaisante 1'action des animaux. 28 Remarquons que, quelque position que Ton adopte, cette position, sauf a sombrer dans d'inextricables difficultes, implique que Ton ait recours a une instance, incluse ou non dans Yaisthesis et/ou phantasia, qui ne soit pas bornee a la seule perception ou revelation des choses exterieures, mais qui soit aussi une antilepsis ou une sunaisthesis.
Or, si les Stoiciens, dans leur volonte de desanimaliser l'humain, ont sans doute tout fait pour preserver sa specificite, rien n'assure qu'ils aient tous adopte la meme solution dans le traitement du probleme animal. De la que certains leur aient peut-etre accorde une forme d'assentiment, tout se passant comme si, aux yeux de ceux-la, horme, aisthesis et/ou phantasia, ces dernieres ayant alors une extension moins large que chez Hierocles, ne suffisaient pas a rendre compte de 1'activite des animaux. Ainsi de Ciceron qui, dans un passage sans doute influence par Antiochus, et pour cette raison conteste,29 fait dire a Lucullus: Puisque la principale difference entre l'etre inanime et ranimal est que ranimal est actif . . . ou bien il faut lui enlever la sensation ou bien reconnaitre que la faculte d'assentiment est en son pouvoir. Or vouloir que les animaux ne puissent ni sentir ni donner leur assentiment, c'est en quelque sorte leur enlever Tame . . . II est absolument necessaire que, avant toute action, il y ait une representation et un assentiment a cette representation. Aussi nier la representation ou l'assentiment, c'est retirer toute activite a la vie. (Acad. Pr. 11. 37-9) 30
Quand bien meme ce passage devrait etre tenu pour infidele a l'orthodoxie en ce qu'il concede peut-etre trop aux critiques neoacademiciennes, et pourrait etre rapproche, en raison du caractere automatique de l'assentiment, de ces propos de jeunes Stoiciens soutenant que, pourvu qu'il n'y ait pas d'obstacle, la phantasia kataleptike 'nous saisit par les cheveux' (Sextus, AM vn. 257), il n'en resterait pas moins que, fidele ou non sur ce point a l'orthodoxie, ce passage conserverait neanmoins le merite de bien souligner le probleme relatif aux animaux: que serait leur ame, et done leur activite 28
29 30
O n voit par-la comment u n theme a p p a r e m m e n t mineur a p u d o n n e r lieu a u n e belle polemique q u i , q u a n d elle sera recuperee p a r Sextus, en viendra a porter sur u n theme majeur, a savoir celui d u critere d e la verite. Cf. S a n d b a c h (1971) p . 15; Inwood (1985) p . 76. T r a d u c t i o n de Brehier, La Pleiade, Paris 1962.
242
JEAN-LOUIS LABARRIERE
vitale, s'ils etaient prives d'assentiment? Autrement dit, la sensation ne peut se penser sans l'assentiment et 'elle est suivie d'action parce qu'elle a pour consequence une inclination' (Ciceron, Acad. Pr. 11. Telle est d'ailleurs, dans un passage trop peu sollicite des lors que Ton s'interroge sur les facultes animales, voire sur le probleme de 1'assentiment en general, la position que Porphyre pretera aux Stoiciens: Les Stoiciens au contraire placent non seulement la sensation dans la presentation [ouk en tei phantasiai histanton morion], mais en rapportent encore l'essence a l'assentiment (alia ten ousian anartonton apo tes sunkatatheseos). Car
la presentation sensible est l'assentiment ou la sensation de 1'assentiment selon l'impulsion [e aisthesis tes sunkatheseos kijith'hormen ouses]. (de Fac. Anim.
apud Stobee, Egl. 11. 25, W. 349, 23-27, = SVFn. 74)31
Maintenant, est la chose et plus evidente chez Ciceron que chez Porphyre, cet assentiment est automatique, puisque sentir, c'est immediatement assentir. Qu'une telle position soit sans doute peu orthodoxe et reflete peut-etre une concession aux Academiciens, n'importera moins ici que la 'concession' de l'assentiment aux animaux. En effet, cette 'concession', si e'en est bien une, pourrait fort bien avoir pour fin de souligner que Faction des animaux, certes immediate et non reflechie, n'est pas pour autant aveugle. En d'autres termes, les animaux ne seraient pas seulement 'agis', comme il semble que le soient les plantes privees d'ame, mais seraient encore le 'sujet' de leurs actions en ce sens que, s'ils n'ont certes pas la possibility de ne pas assentir, ils sont neanmoins 'conscients' de ce qu'ils font, sans etre toutefois capables de discerner entre leurs phantasiai, lesquelles ils suivent immediatement. Aussi, si telle n'etait peut-etre pas la position initiale des Stoiciens, alors nous aurions peut-etre moins affaire, relativement aux animaux, a une position infldele qu'a une position affinee grace a des critiques de Carneade qui, rapporte Sextus, marquerent Antiochus: 31
J e conserve la version manuscrite et ne suit pas la correction proposee par Von Arnim: aisthetikei gar phantasiai. Si Ton acceptait cette correction, le sens de la deuxieme phrase serait alors celui-ci: 'car la sensation est un assentiment a la presentation (ou impression) sensible'. La position des Stoiciens est presentee par Porphyre comme s'opposant a celle de Numenius (fr. 45 Des Places) qui soutenait que la presentation [to phantastikon] est 'un accompagnement [sumptoma] (de l'assentiment), non pourtant son acte ou son effet, mais son satellite \parakolouthemd\\
De la 'nature phantastique' des animaux chez les Stoiciens
243
Puisque 1'animal [£<JzW] differe des etres inanimes par la faculte sensible [aisthetikei dunamei], il apprehendera [antileptikongenesetai] grace a elle tout a
la fois [pantos] lui-meme et la chose exterieure. {AM vn. 160)
Remarquons done, pour conclure ce point, que si action animale il doit y avoir, et que si cette derniere ne doit pas se confondre avec les mouvements pretes aux plantes, alors il doit bien y avoir chez les animaux quelque chose comme cette antilepsis non seulement de la chose exterieure, mais encore de soi. Or, si Ton veut etre consequent, cet acte d'apprehension de ce qui est propre, implique, meme de fafon confuse, que les animaux ne se contentent pas d'avoir des 'presentations d'objet', des 'presentations presentees', mais encore des 'presentations se presentant', des 'presentations presentantes'. L'attention qu'ils portent a leur constitution, la conscience qu'ils en ont, implique que les animaux, en meme temps qu'ils perfoivent quelque chose, se perfoivent comme le percevant et soient, pour ainsi dire, avertis d'un certain soi qui perjoit, afin que l'impulsion elle-meme puisse devenir effective sous l'effet de leur phantasia. La doctrine de Voikeiosis semble done devoir conduire les Stoiciens a poser une certaine forme de conscience chez les animaux. Mais cette position risque de heurter les gardiens d'une stricte orthodoxie car elle implique que le discernement de Futile et du nuisible soit ramene a un quelconque centre de la perception, le quasi-A^monikon animal. Or la difficulte tient moins a la position de ce quasihegemonikon qu'a la necessite d'y inclure, sous une forme ou sous une autre, une instance cognitive debordant la stricte impulsion et la stricte sensation-presentation, tout en en dependant fortement puisque cette instance ne saurait dependre de la pensee ou du raisonnement reflexif. UNE APPROGHE DE L ASSENTIMENT ANIMAL
Toute l'ambiguite de la position stoicienne, tient, me semble-t-il, a ceci qu'irrationnels, parce que prives de logos, les animaux et leurs comportements sont susceptibles d'etre analyses selon un double point de vue: compares aux humains, ils doivent etre reduits a la portion congrue en ce qui concerne leurs capacites cognitives et pratiques, tandis qu'en raison de la doctrine de Yoikeiosis, ils se voient reconnaitre de plus grandes capacites sur les deux versants de cette doctrine, celui de la recherche du propre et celui de la 'justice'. II en va d'ailleurs de meme lorsque Ton rapporte le statut des
244
JEAN-LOUIS LABARRIERE
animaux a la doctrine de la Providence: leurs capacites peuvent se voir minimisees lorsqu'on les rapporte a celles des humains, mais elles peuvent aussi etre magnifiees, si Ton ose dire, lorsqu'il s'agit de louer les oeuvres de la Providence. De la cette delicate question de l'assentiment animal: son enjeu n'est pas seulement celui de la difference entre l'homme et l'animal du point de vue des facultes mises en jeu dans Faction, mais aussi celui de la coherence d'une doctrine relativement au statut des facultes animales mises en jeu dans le mouvement des animaux. C'est pourquoi, malgre les denegations fondees sur d'autres sources tenues pour plus sures par B. Inwood,32 j'avoue pour ma part tenir comme assez fondees les assertions de Ciceron, d'Alexandre d'Aphrodise, de Nemesius, voire de Porphyre, gratiflant les animaux d'assentiment.33 Encore une fois, on voit mal comment les Stoiciens pourraient eviter d'introduire dans le quasi-hegemonikon des animaux cette troisieme instance, en plus de l'impulsion et de la phantasia. Ainsi, quand bien meme on devrait accepter la distinction proposee par B. Inwood entre eixis et sunkatathesis, et ne doter les animaux que de Veixis car seuls les humains assentissent vraiment, l'assentiment portant sur un axioma et etant done solidaire du logos, toujours resterait-il que cette 'action de ceder', que Ton pourrait peut-etre rendre par 'emprise' afin d'indiquer que les animaux ne sont pas libres de ceder ou non a leurs phantasiai, ne devrait pas induire une automaticite telle qu'elle reduise a neant toute conscience de leur propre constitution et de ce qui est utile a leur preservation. Si tel etait en effet le cas, qu'apporteraient de plus la phantasia et Yhorme par rapport a la phusis des vegetaux? Nous tiendrions sans doute la le principe d'une difference organique, 'physico-physiologique', mais certainement pas celui d'une reelle difference 'psychologique', laquelle doit bien etre, a moins de supposer que la psuche non logike ne soit qu'un nom designant les seuls phenomenes vitaux, les Stoiciens 'anticipant' alors tant sur la theorie du corps-machine que sur l'hypothese de 1'animal-machine chez Descartes. De la sans doute les reticences d'A. Long et de D. Sedley a ceder aux instances de B. Inwood puisque, pour leur part, ils tiennent que le quasi32 33
Cf. Inwood (1985) pp. 66-91 (tout particulierement p. 75 pour l'eventuelle distinction entre eixis et sunkatathesis). Cf. Nemesius, Mat. Horn. 291. 1-6 (SVF 11. 991). Les passages d u de Fato d ' A l e x a n d r e d'Aphrodise seront discutes infra.
De la 'nature phantastique3 des animaux chez les Stoi'ciens
245
hegemonikon animal cornporte, sous une forme 'rudimentaire', trois des quatre pouvoirs de Yhegemonikon humain, seul le logos leur faisant radicalement defaut. 34 Si le logos est bien, comme chez Aristote, la pierre de touche de la difference entre les animaux et les humains, on ne peut pour autant, malgre la rationalite incluse dans nombre de definitions de l'assentiment, conclure de l'irrationalite des presentations et inclinations animates, a la pure et simple inexistence de l'assentiment chez les animaux, mais seulement a l'inexistence chez ces etres de tout assentiment rationnel. Or, etre prive d'assentiment rationnel, ne revient nullement 'a n'avoir point d'assentiment du tout'. On peut d'ailleurs tenir pour un indice de cette difficulte le fait que, meme aux yeux de B. Inwood, il faille quelque chose comme un troisieme pouvoir, ce que serait precisement Veixis. Si l'on-veut traiter ce point et eclairer ce que Ton peut entendre par 'rudimentaire', il faut tout d'abord se garder des ecueils dus a, nos sources. J'en prendrai pour exemple l'apparente contradiction entre le de Fato d'Alexandre d'Aphrodise et le de Principiis d'Origene, le premier accordant l'assentiment aux animaux tandis que le second le leur denie, ce qui a entraine, surtout si Ton tient compte des intentions polemiques d'Alexandre et de sa mauvaise foi a l'egard d'adversaires qu'il ne nomme d'ailleurs jamais, que Ton a mis en doute son temoignage. 35 Or, reste a savoir si la question de la fidelite est la meilleure question a poser au de Fato. Ne pourrionsnous nous efforcer de le comprendre comme temoignant d'une des difficultes du Stoicisme et, par la, recevoir son temoignage comme etant assez fonde? Remarquons done tout d'abord qu'Alexandre sait lui aussi etre fidele. Ainsi, il distingue bien le mouvement naturel des etres inanimes, entierement du a des causes exterieures (exothen), de celui des etres animes, du a l'impulsion et par consequent au 'pouvoir' de ces derniers (xm. 182. 10-16; xiv. 184. 24). De meme il insiste a juste titre sur la difference stoicienne entre Faction 'directe' du destin par le biais des causes exterieures, biais qui n'a rien d'un biai, et son action 'indirecte' a travers Pinclination qui en rend done maitres ceux qui en sont les sujets (di'hormes, xm. 182. 30). Tout le probleme tient done a ce que cette 'maitrise' soit renvoyee a Pinclination et a l'assentiment sans distinguer entre les animaux et les humains. Ici 34 35
Cf. L o n g & Sedley (1987) vol. 1, p . 332 et 11, 530, p . 317; Long (1971) p p . 170-3. Cf. Sharpies (1983) p p . 141-4 et la controverse qui oppose I n w o o d (1985) a Stough (1978)
pp. 203-31.
246
JEAN-LOUIS LABARRIERE
interviennent sans nul doute les intentions polemiques et la mauvaise foi d'Alexandre, mais, s'il est injuste envers les Stoiciens en faisant comme si ce qu'ils soutenaient revenait a rabattre la liberte des humains sur le comportement 'spontane' des animaux (xiv. 183. 22-184, 4), et en fin de compte les mouvements des etres animes sur ceux des etres inanimes, leur est-il tout autant infidele en ce qui concerne les animaux en faisant comme si la possession de l'inclination allait avec celle de l'assentiment? Voyons pour ce faire ce qu'Alexandre entend par assentiment puisque selon lui, a la difference d'Origene, la doctrine stoicienne de l'assentiment, liee a leur theorie du destin et de la Providence, ne laisse aucune place au 'libre-arbitre', qu'il definit comme la possibilite de choisir entre deux possibilites (XII). En effet, bien qu'Alexandre rapporte des propos d'origine stoicienne selon lesquels il faut distinguer, au sein des animaux mus selon l'inclination, entre ceux qui 'agissent seulement' (energesei monon) et ceux qui 'accomplissent des actions raisonnables' (praxei ta logika, xxxiv. 205. 30-5), il ne credite pas les Stoiciens de la distinction entre l'assentiment sensible des etres irrationnels et l'assentiment rationnel des etres humains, assentiment qu'il assimile alors a la deliberation et au choix (xxxm. 205. 14; Quaestiom. 13). L'assentiment qu'il prete aux animaux n'est done jamais que l'acte de ceder a une phantasia (eixein, xiv. 183. 22-184, 4), ce a quoi il oppose l'acte du logos deliberateur, dont les Stoiciens, a ses yeux, n'auraient pas vu qu'il constitue le propre de l'humain. Cet assentiment n'est done pas l'assentiment rationnel. En consequence, l'assentiment dont, selon Alexandre, jouiraient animaux et humains d'apres les Stoiciens, n'est pas l'assentiment 'standard', a l'oeuvre par exemple chez Origene. De la qu'a mon sens ces deux temoignages, aux visees bien differentes, ne se contredisent pas reellement puisqu'ils ne se fondent pas sur la meme notion d'assentiment. Ici, nous n'avons affaire qu'a l'assentiment 'sensible' qui, cedant aux 'seules presentations' (tais phantasiais monais, 184. 1), est sous leur 'emprise'; la, nous avons affaire a l'assentiment 'rationnel' dont le maitre d'oeuvre est un logos discriminateur. D'un cote, Alexandre rapporte done fidelement les distinctions qui sont bien celles des Stoiciens, tandis que d'un autre, pour les besoins de sa polemique, il les ecrase. On pourrait en ce sens se demander s'il ne negligerait pas l'etape representee par les vegetaux doues de phusis afin d'aller plus vite dans sa polemique: au bout du
De la 'nature phantastique' des animaux chez les Stoiciens
247
compte, le mouvement selon l'inclination, meme renforce par l'assentiment, ne differe guere de celui des pierres ou du feu puisque le determinisme y pese tout autant, et qu'il ne laisse guere qu'une seule possibilite a tout se qui se meut. Tout se passerait ainsi comme si, dans sa polemique, l'Exegete avait eu en tete ce chapitre de la Metaphysique (© 5), ou Aristote distingue entre les puissances irrationnelles productrices d'un seul effet, celles des inanimes et ce qui dans les animes ne se meut pas en tant qu'anime, et les puissances rationnelles productrices des contraires, celles des animes, chez lesquels il est done necessaire qu'a la necessite qui preside a l'actualisation des puissances irrationnelles, vienne s'ajouter le desir (orexis) ou le choix preferentiel {prohairesis). Le reproche principal qu'il adresserait alors aux Stoiciens serait d'avoir tout confondu et d'avoir traite tous les etres comme s'ils n'etaient dotes que de puissances irrationnelles. On comprend alors que dans le feu de sa polemique, Alexandre puisse crediter d'assentiment aussi bien les animaux que les humains, l'assentiment ne changeant rien a ses yeux puisqu'il l'interprete comme si les Stoiciens n'y avaient vu que l'acte de ceder a une phantasia^ ce qui, a ses yeux, est la marque des animaux et non des humains, lesquels n'assentissent pas aux seules pkantasiai: Car tout ce qui apparait n'est pas representation [ou gar phantasia to phainomenon pan]. La representation en effet, simple et irrationnelle [haple te kai choris logon], se produit dans la rencontre des choses exterieures [hupo ton exothen prospiptonton ginetai], elle est modelee sur les activites sensorielles, et e'est pourquoi elle a aussi sa plus grande force chez les animaux sans raison [dio kai ten ischun en tois alogois zoois echei malista], tandis que certaines
donnees tiennent evidemment la cause de leur apparition de la raison et d'un raisonnement [phainetai de tina kai dia logon te kai para sullogismou ten
aitian tonphainesthai lambanonta], mais personne ne dirait d'elles que ce sont des representations: celui en effet qui, par le raisonnement effectue par lui-meme dans la deliberation, donne son assentiment a quelque chose est lui-meme, pour lui-meme, agent causal de son assentiment. (xv 186. 5-!
2)36
L'extraordinaire est que cette distinction que l'Aphrodisien s'attribue et qui, n'etait la notion de deliberation, sonne de fafon tres stoicienne, pourrait donner une assez bonne exegese des positions stoiciennes: les animaux assentissent, comprenons: cedent a ce qui leur apparait a travers leurs sensations, parce que seule la phantasia 36
Traduction de P. Thillet, legerement modifiee. Les Belles Lettres, Paris 1984.
248
JEAN-LOUIS LABARRIERE
simple leur appartient; tandis que les humains assentissent non a ce qui leur apparait sous l'effet d'une phantasia simple, mais a ce qui leur 'apparait' a Tissue d'un raisonnement critique. Par-dela son caractere nettement polemique, le temoignage d'Alexandre est done precieux en ce que, mieux sans doute qu'Origene, plus fidele d'un autre point de vue, il insiste sur l'importance de la difference introduite par l'inclination dans la scala naturae, et done sur la portee fondamentale du mouvement kath} hormen.37 En effet, si Ton veut que le mouvement qui en procede soit bien different de celui qui n'en procede pas, alors cela implique une certaine maitrise du mouvement selon l'inclination chez ceux qui le possedent, car sinon on ne comprendrait ni quelle serait la difference introduite par ce type de mouvement, ni comment les animaux se mettraient en mouvement (184. 23-5). Or, a cet egard, l'assentiment, meme rudimentaire et automatique, vient bien mieux satisfaire cette exigence qu''une phantasia simple. Sans doute chez les animaux cet assentiment se produit-il de fafon automatique et repetitive, et sans doute cette regularite sans faille peut-elle etre interpretee comme le signe que 'les betes n'ont point de raison du tout', mais cette automaticite, surtout quand elle est le fait d'animaux qui se conduisent de fafon ordonnee en suivant des phantasiai produites de fafon endogene, ne saurait se confondre avec la necessite brute des causes exterieures. Or pour qu'une telle confusion n'ait pas lieu, il faut sans doute qu'a l'inclination et a la presentation vienne s'ajouter quelque chose comme un assentiment arm que l'animal qui se meut kath'hormen ne soit pas seulement, si Ton ose dire, le lieu de ce mouvement, mais soit aussi en quelque fafon son 'sujet'. Comment comprendre sinon la suneidesis ou sunaisthesis? Cette 'conscience de soi (percevant)', comme 'attention a soi' et 'connaissance de soi', n'est certes en rien rationnelle, mais ce n'est pas une raison suffisante pour en faire une absence totale de 37
J e n'ignore pas bien entendu le remarquable appendice 4 de Inwood, (1985) pp. 250-7, mais ce qui y est dit ne me semble pas s'appliquer a ce que je m'efforce de montrer. D'autre part, bien que je sois globalement d'accord avec le theme de cet appendice montrant comment l'expression kath'hormen a perdu chez les commentateurs platoniciens ou aristoteliciens, le sens technique qu'elle avait chez les Stoi'ciens, pour finir par signifier tout simplement le 'mouvement volontaire', il ne me semble pas que tel soit toujours le cas, ainsi de Porphyre et d'Alexandre d'Aphrodise. Enfin, je ne comprends pas bien ce que peut avoir d'aristotelicien la synonymie entre les expressions kath'hormen kinesis et kata prohairesin kinesis, puisque Aristote, et Alexandre a sa suite, prennent precisement grand soin de distinguer le mouvement hekousion, 'volontaire', du mouvement selon laprohairesis, et ce afin de pouvoir distinguer entre hommes et animaux. Cf. EJV in. 4, 111 ib6-i8; Eth. Eud. 11. 8.
De la 'nature phantastique' des animaux chez les Stoiciens
249
conscience, sous pretexte qu'il ne saurait y avoir d'assentiment que rationnel. EN GUISE DE CONCLUSION
En raison des multiples problemes que souleve 1'analyse des facultes cognitives et pratiques des animaux, il semble difficile de brosser de fafon ferme un tableau des positions stoiciennes puisque, des que Ton s'efforce d'expliciter tel ou tel point, alors, bien souvent en fonction de telle ou telle perspective, la doctrine, a supposer qu'il y en ait bien une sur cette question, accorde plus ou moins aux animaux. De fafon generale, il me semble que le probleme est tout autant methodologique que doctrinal: etant donne que les novations stoi'ciennes sont enoncees dans une conceptualite, et non seulement dans un langage qui ne vaut que pour les humains et les dieux, comment penser la 'rudimentarite' des facultes cognitives et pratiques des animaux, toute transposition, meme affecte du signe negatif, etant de jure proscrite? Ainsi, malgre la notion de scala naturae, la coupure introduite par le logos entre les animaux et les humains semble bien plus tranchee que chez Aristote, et bien entendu, pour ce que nous en savons, chez Theophraste et Straton. Sans doute Poriginalite de la conception stoicienne de la phantasia kataleptike et du critere de la verite les conduisit-elle a marquer les animaux non du signe moins, mais de 1'alpha privatif, en interdisant par la tout recours a l'analogie ou a la difference selon le plus et le moins. Mais on comprend alors le prix a payer pour cette innovation dogmatique: celui des polemiques avec ceux qui ne pouvaient se satisfaire de positions aussi tranchees et a leurs yeux incertaines. Or, comme je le disais en introduisant cette etude, s'attaquer par ce biais au systeme stoicien, c'etait, en s'attaquant a un de ses points obscurs, tendre a le remettre tout entier en question. Et sans doute, faute d'avoir clairement perfu par avance la pertinence de la question animale, les Stoiciens s'exposaient-ils a ce genre de critiques. En effet, s'il est vrai que la position consistant a soutenir: des animaux, que ipouvons-nous dire? A une grande force, cette force ne se trouve-t-elle pas contrebalancee par l'usage que les Stoiciens font des animaux pour asseoir certains pans de leur doctrine, a commencer par celles de Voikeiosis et de la Providence? Comment en effet faire un usage doctrinal de ce dont on ne peut pas dire grand-chose?
CHAPTER 9
Le concept de doxa des Stoi'ciens a Philon d'Alexandrie: essai d'etude diachronique Carlos Levy
Le but de ce travail est d'esquisser une etude des variations dans la psychologie de la doxa depuis Zenon jusqu'a Philon d'Alexandrie. L'approche que nous avons choisie peut paraitre surprenante a double titre: elle est diachronique, alors que les etudes actuelles ont quelque peu delaisse ce type de demarche; par ailleurs, elle se refere a un personnage qui, non seulement n'appartient pas a la periode hellenistique, mais dont la qualite de philosophe est fortement contestee, et le plus souvent meme niec II convient done de preciser ici notre propos et la methode choisie. Les etudes philoniennes ont toujours eu un statut special par rapport aux recherches sur l'histoire de la philosophic ancienne. Le 'philonien' est quelqu'un qui etudie un corpus ou tout est exprime en fonction de concepts philosophiques, mais auquel on a denie, pour des raisons parfois contradictoires la qualite de texte philosophique. 1 Or ces dernieres annees ont vu la remise en question de cette attitude, et une certaine rehabilitation de l'apport de Philon a la philosophic 2 II n'est pas question d'entrer ici dans le probleme de fond; nous souhaitons montrer a propos d'un probleme particulier, celui de l'opinion, comment la reflexion de Philon ne peut se comprendre qu'a la lumiere des debats anterieurs, et apporte a ceux-ci une contribution que Ton sera en droit de juger peu satisfaisante, mais dont nous pensons qu'elle a sa place dans une etude d'histoire de la philosophic Le choix de Philon comme terme de notre etude n'a pas pour origine notre seul interet pour cet auteur. II se trouve que travaillant depuis longtemps sur la Nouvelle Academie, nous avons pu constater 1
2
On trouvera une analyse critique des etudes philoniennes des annees 1937-82 in Radice (1983). La critique la plus coherente des lectures philosophiques de Philon est celle de Nikiprowetzky (1977), qui a elabore la these de la philosophic comme instrument au service de l'exegese. Cf. Runia (1986) et Mansfeld (1988).
250
Le concept de doxa des Stoiciens a Philon d'Alexandrie
251
tater que, par la mediation de cette entite, conceptuellement interessante mais historiquement tres problematique qu'est 'le sceptique', 3 cette ecole est etudiee le plus souvent dans la perspective du neopyrrhonisme.4 Notre demarche sera differente: tout en rejetant categoriquement la fable du dogmatisme esoterique,5 nous essaierons de montrer a travers l'exemple de l'opinion comment les Neoacademiciens ont prepare, en quelque sorte malgre eux, l'avenement d'un platonisme dogmatique, d'ou le fait que des gens comme Ciceron ou Plutarque ne voient aucune contradiction a unir dans une meme admiration Arcesilas et le Platon de la theorie des Formes. La position de Philon est encore plus complexe: il n'aime pas les Academiciens, pas plus que les Sceptiques,6 mais son oeuvre temoigne puissamment de leur influence et, volens nolens, il est, du moins nous semble-t-il, Tun des beneficiaires de Tune des formes du devenir de leur pensee. LE CONCEPT DE DOXA DANS LA PHILOSOPHIE STO1CIENNE
Le concept de doxa dans la philosophie du Portique nous parait presenter deux caracteristiques, contradictoires pour quelqu'un qui regarde le systeme de l'exterieur, mais evidemment coherentes pour peu que Ton accepte de se situer dans la logique, au sens le plus large, de cette doctrine: - la doxa est, dans un sens que nous aurons a preciser, un concept mineur, nous dirons meme sciemment minore, de la psychologie stoicienne; - c'est dans ce concept, dont il serait exagere de dire qu'il est introuvable, mais qui assurement fait probleme, que le stoi'cisme a subsume des son origine une realite humaine considerable, a laquelle ses predecesseurs avaient attribue une nature et un fonctionnement autonomes, a savoir les passions. Pour des raisons qui tiennent essentiellement a la conception 3
4
5
II s'agit la d'un probleme considerable, qui est parfois bien percu par ceux qui parlent du 'sceptique', cf. ce qu'ecrit a ce sujet Frede (1987), p. 215. La position exactement inverse est celle de Kramer (1971), p. 52, qui refuse le terme de 'scepticisme' a propos de la Nouvelle Academie et prefere celui d'aporetisme. La principale exception a cette tendance nettement majoritaire est constituee par Tarrant (1985), qui a interprete la philosophie de Philon de Larissa comme un pont entre 1'Academie sceptique et le moyen platonisme. 6 Gf. Levy (1978). Cf. Levy (1986).
252
CARLOS LEVY
generale de ce travail, nous avons choisi de n'aborder ici que le premier aspect du probleme. La doxa est un concept mineur de la psychologie stoicienne en ceci qu'aucune place specifique ne lui est reservee ni dans la description de la structure de Tame ni dans celle des mecanismes de son fonctionnement. Cela constitue une innovation tres considerable par rapport a Platon et a Aristote. II est inutile d'insister ici sur Pimportance de la dunamis doxastike dans la description platonicienne du fonctionnement de Tame.7 De meme, nous nous contenterons de remarquer brievement que le doxastikon est facilement reperable dans la division aristotelicienne de Tame. 8 En revanche, le stoi'cisme a elimine toute reference a la doxa de ses divisions psychologiques, pour la simple raison qu'il n'y a pas chez lui de faculte specifiquement productrice de doxa, et il n'y a pas puissance doxastique de Tame parce que rien dans le monde n'est destine a etre necessairement objet de doxa. L'univers stoicien, de par la rationalite immanente aux choses, n'a rien en lui qui rende impossible une science parfaite. La plupart des representations apportent une image correcte des choses et, lorsque tel n'est pas le cas, le sage pourra fort bien rever, avoir des hallucinations, ou etre confronte a des questions insolubles, il lui suffira de suspendre son assentiment, ce qui n'alterera en rien la science qu'il a de l'ordre du monde. Aucune des huit parties de Pame n'est naturellement source d'opinion, mais les choses ne sont guere plus faciles a discerner lorsqu'on cherche a determiner a quoi correspond celle-ci dans le fonctionnement de Phegemonique. II est a cet egard tres revelateur que l'opinion soit absente de la celebre metaphore zenonienne de la connaissance;9 dans la succession: main ouverte (representation), main contractee (assentiment), main fermee (comprehension), poing maintenu ferme par l'autre main (science), ou est la doxa? Meme si elle correspondait exactement a l'un des maillons de cette succession, Pabsence du terme lui-meme ne saurait etre consideree comme insignifiante et devrait deja etre interpretee comme le signe d'une volonte de rompre avec la problematique precedente de Platon et 7 8 9
Cf. Platon, Rep. 477b7-c5 et le commentaire qu'en font Annas (1981), p. 201, et Y. Lafrance (1986), pp. 124-5. Cf. Aristote EN 1140D26-30. Cf. Ciceron, Luc. 145. Ce passage a fait l'objet d'une interessante etude de Gorier (1977) qui s'est efforce de montrer, contre la tradition generalement admise, que ce passage n'exprime pas une succession chronologique, mais une hierarchie dans l'ordre du savoir.
Le concept de doxa des Stoiciens a Philon d'Alexandrie
253
d'Aristote. Mais la question est d'autant plus complexe que l'opinion ne peut etre assimilee sans difficulte a la main legerement ouverte et que sa definition par rapport a la katalepsis souleve d'epineuses questions sur lesquelles nous allons avoir a revenir. Notons egalement que la doxa ne figure pas non plus dans le texte de Jamblique 10 ou les quatre qualites specifiques de l'esprit humain sont: la representation, l'assentiment, la hormett le logos. La encore il y a au moins matiere a discussion. En effet, si Ton estime, avec A. Long,11 que le logos n'est pas ici le quatrieme terme de l'enumeration et qu'il designe ici ce qui est present dans les trois autres qualites, il n'y a pas d'ambiguite: la doxa est une variete particuliere d'assentiment, et il reste a determiner laquelle. En revanche, si on rejette cette interpretation, intellectuellement satisfaisante, mais a laquelle on peut objecter la forme meme de l'expression, et si on estime, comme l'a fait A. J. Voelke12 que logos designe dans ce texte un aspect particulier de l'activite de l'hegemonique, la localisation de la doxa est encore plus delicate. Ces difficultes n'ont rien de fortuit. Lorsque Zenon a elabore ses concepts psychologiques fondamentaux, il ne visait pas a rendre compte le plus rigoureusement possible des erreurs humaines, mais a demontrer qu'en depit de ces aberrations Tame est rationnelle et que l'aboutissement naturel, bien qu'exceptionnel, de son developpement est la sagesse. Comme le montre si bien la metaphore de la main, sa psychologie, qui a pour point de depart cette perfection immediate que constitue l'activite des sens, est tout entiere orientee vers la sagesse, perfection conquise et definitive. Cela ne signifie pas qu'il se soit desinteresse de la doxa, mais il l'a reduite a n'etre qu'un processus de dysfonctionnement d'un systeme confu pour etre parfait. Pour etudier ce dysfonctionnement, nous avons choisi de commencer notre analyse par un texte exprimant un etat de la doctrine probablement posterieur a Zenon, mais qui permet de mieux poser les problemes. Nous n'entrerons pas ici dans la discussion sur la source de ce passage de Stobee, vraisemblablement Chrysippe, mais on ne peut pas aller dans ce domaine au dela d'une conjecture raisonnable.13 En 10 11 12 13
Jamblique, de Anima, ap. Stobee, SVF 11.831Long (1982b), p. 51: The hoyos of the human soul is not one faculty among others, but the mode of the whole soul's operation. Voelke (1973), p. 24. Stobee, SVF 111.548. Pour Ioppolo (1986), p. 100, n. 33, la source de Stobee serait Chrysippe. Cette these a ete recemment critiquee par Maconi (1988), p. 240, qui voit dans le texte de Stobee 'a farrago of different views'.
254
CARLOS LEVY
tout cas, il s'agit d'un texte precis, qui a pour vocation de demontrer l'heterogeneite absolue de la sagesse et de la doxa. De cette derniere il nous est dit qu'elle revet deux formes: la doxa est, d'une part, l'assentiment a ce qui n'a pas ete perfu (akataleptoi), d'autre part, V hupolepsis asthenes.l4r La premiere definition, qui se refere au cas ou Ton a assenti soit a une representation vaine (phantasma), soit a une representation ne correspondant pas a son objet, soit encore a une representation vraie, mais obscure, ne fait pas veritablement probleme;15 il reste a preciser le sens de la deuxieme et le lien qui existe entre les deux doxai. Le terme hupolepsis est rare dans les temoignages stoiciens et il est un de ceux qui ont le plus souffert de la disparition des textes de cette ecole, puisqu'il ne nous reste rien du Peri Hupolepseos de Chrysippe.16 L'analyse des differentes occurrences n'apporte guere de lumiere supplementaire, a une exception pres, nous semble-t-il. II s'agit d'un texte ou Plutarque, paraphrasant Chrysippe, ecrit que pour celui-ci la representation ne saurait etre cause de l'assentiment parce que, si cela etait le cas, le sage, qui est parfois oblige de mentir aux phauloi, devrait etre considere comme responsable de leur erreur et de la tromperie. La traduction que nous proposons de ce passage est celle-ci:17 souvent les sages font usage du mensonge a Pintention des sots et leur suggerent une representation persuasive. Celle-ci n'est cependant pas cause de leur assentiment car, si elle l'etait, elle serait egalement responsable de leur conviction [tes hupolepseos] erronee et de la tromperie.
Ce texte montre clairement que hupolepsis n'est pas un simple doublet de sunkatathesis. La difference, a en juger en tout cas par ce texte, parait etre celle-ci: la sunkatathesis represente l'acte d'assentiment envisage independamment de la proposition a laquelle on assentit, alors que la hupolepsis est inseparable de celle-ci. L'analyse de ce texte confirme que l'invention par Zenon de la sunkatathesis comme concept philosophique ne fut pas une simple innovation terminologique; elle correspondait a la necessite d'exprimer une experience pour laquelle le terme de hupolepsis etait inadequat, celle 14 15 16 17
Ibid.: AITT&S yap eTvai 86£as, TTJV pev aKaTaAf)7TTco cjuyKaT&Oeaiv, Cf. Diog. Laer. vii.46 = SVF 11.53, e t Al. de Anima p. 71, 10 Bruns = SVF 11.70. Le traite de Plutarque ffepi UTroArjyEQos est mentionne par Diogene Laerce, vii.201. Plutarque, Sto. rep. 47.1055 f-i056a = SVF 11.994: TTOAA&KIS yocp 01 ao9oi iTpos TOUS 9aOAous Kai 9avT0ccriav Trapicrrdai TTi0avf|v ou \XT\V ai-riav TX\% ai Kai TX\S dTr
Le concept de doxa des Stoiciens a Philon d'Alexandrie
255
du face a face du sujet avec ses representations. Si nous tentons de generaliser cette conclusion nous nous trouvons confronte aux questions suivantes: (a) Le terme de hupolepsis est-il reserve dans le stoi'cisme a des croyances resultant d'assentiments faibles et/ou errones? Le fait meme que Chrysippe-Plutarque utilise l'expression tes hupolepseos tes pseudous montre qu'une telle interpretation est inexacte. On doit done rejeter une correction de Wachsmuth, acceptee par von Arnim et neanmoins tres contestable, qui substitue katalepsin a hupolepsin dans la definition de la pistis, laquelle est presentee comme une caracteristique du sage.18 Le fait d'avoir des hupolepseis est commun a tous les hommes et seule la qualite de celle-ci differencie le sage du sot. On retrouve ainsi chez les Stoiciens, et dans un contexte evidemment different, les variations d'Aristote, qui tantot associe etroitement doxa et hupolepsis, les considerant comme susceptibles d'erreur et les differencial des hexeis de verite, tantot subsume dans hupolepsis a la fois la science, l'opinion et la prudence; 19 (b) Lorsque la source stoicienne de Stobee dit que Tune des deux varietes de doxa est la hupolepsis asthenes, que faut-il entendre par la? S'agit-il necessairement d'une croyance faible, expression d'une ame elle-meme depourvue de cette tension qui caracterise la sagesse, ou bien peut-on imaginer que cette hupolepsis faible resulte de Pintegration d'un assentiment non-faible (e'est a dire d'une katalepsis) dans une ame qui n'a pas la force de la sagesse? Les chercheurs ont des opinions divergentes sur cette question20 et il faut reconnaitre que ce texte lui-meme ne permet pas de trancher. Pour essayer de voir plus clair dans ce probleme, nous allons revenir a la source, e'est a dire a Zenon, et done au fameux passage qui se trouve a la fin de l'expose de Varron: 21 18
19
20
21
Stobee, SF.F111.548: TT)V 5e TTICTTIV dorelov Crrrapxeiv, elvai yap KaxaAriyiv iaxupav pspa CTOCV TO UTToAaiiPavopievov. L'edition Meineke de Stobee conserve, en revanche, le texte des manuscrits. Aristote, EN 113^17; de An. 111.427b24.-27. II est a remarquer que dans les Topiques V.i3obi6-i7, Aristote donne une definition de la science qui pourrait etre tout a fait stoicienne: UTTOATI^IV duETaTTEioTOV OTTO Aoyou. Long (1974), p. 129: "'Weak assent" describes the cognitive state of someone who has "grasped" the object or what is really the case' L'interpretation de Long a ete critiquee par Gorier (1977) pp. 88-9, qui definit l'aa$evf|S UTroAriyis comme independante de la verite et de l'erreur des representations et se caracterisant par la precipitation avec laquelle a ete donne l'assentiment. Ioppolo (1986) p. 22, n. 21, critique elle aussi Long et affirme que la marque distinctive de la KonrdAriyis n'est pas la faiblesse mais la mutevolezza. Ciceron, Ac. post. 1.41-2.
256
CARLOS LEVY
Quod autem sensu comprehensum, id ipsum sensum appellabat, et si ita erat comprehensum ut convelli ratione non posset, scientiam; sin aliter inscientiam nominabat, ex qua exsisteret etiam opinio, quae esset imbecilla et cum falso incognitoque communis. Sed inter scientiam et inscientiam comprehensionem illam, quam dixi, collocabat, eamque neque in rectis neque in pravis numerabat, sed soli credendum esse dicebat. Nous ferons a propos de ce temoignage les remarques suivantes: - II s'agit ici d'un texte de caractere gnoseologique, qui montre que l'approche zenonienne de la doxa n'etait pas purement ethique et qu'il pouvait envisager la question de l'opinion autrement que comme une faite morale. 22 Bien sur il n'y a pas dans le stoi'cisme de veritable separation entre logique et ethique (la remarque a propos de la comprehensio le confirme si besoin en etait), mais il y existe neanmoins une certaine autonomie des points de vue, et l'approche choisie par Zenon est bien celle de l'analyse du processus de la connaissance. Par ailleurs, l'expression cum falso incognitoque, maintes fois citee par les chercheurs, ne nous semble pas avoir ete suffisamment analysee. D'une part, cette construction de communis avec uniquement la preposition cum est un 'hapax' grammatical, 23 qui en rend Interpretation malaisee, meme si Ton peut admettre que notre expression est equivalente kfalsi incognitique communis. D'autre part, la comparaison avec d'autres passages des Academica montre que falsum et incognitum ne sont nullement employes indifferemment par Ciceron:24 si le premier est la traduction de to pseudes, le second rend to akatalepton. Nous pouvons done affirmer avec quelque vraisemblance que Zenon avait lui-meme fortement theorise la doxa en l'evaluant selon deux approches, l'une cognitive, l'autre psychologique; - la premiere phrase de notre citation indique de maniere tres claire que toutes les comprehensions autres que celles du sage sont, une fois incorporees, de Yinscientia. La katalepsis du sot sera de l'ignorance, non parce qu'elle ne perfoit pas bien l'objet x, mais parce que cette perception sera integree dans un reseau conceptuel instable qui ne permet pas l'intellection de ce qu'est x dans le 22 23
24
Cf. l'opinion contraire d e Ioppolo (1986) p . 23. L a construction la plus frequente d e communis est celle qui comporte, d ' u n e part, le datif, et, d ' a u t r e part, la preposition cum, cf. Tusc. v.5, 14, a propos d e la vie heureuse: negatque ei cum dolore et cruciatu quicquam esse commune. L a construction avec le genitif se trouve, p a r exemple, dans le Lucullus 33: visio verifalsique communis. Gf. Ac. post. 1.45: cum aut falsa aut incognita res approbaretur; Luc. 29: constitutionem verifalsi, cogniti incogniti; 59: numquam adsentiri rei velfalsae vel incognitae; 68: adsentiri quicquam aut falsum aut incognitum.
Le concept de doxa des Stoiciens a Philon d'Alexandrie
257
monde. Cette katalepsis peut-elle pour autant etre qualifiee de doxa? Pour tenter de repondre a cette question il faut analyser mot par mot la proposition inscientiam . . . ex qua exsisteret etiam opinio. L'analyse du
verbe exsistere montre que celui-ci a toujours chez Ciceron un sens fort.25 II exprime une dynamique qui se developpe a partir d'un point precis, ce qui implique que Zenon n'identifie pas la doxa a Vagnoia, il la definit precisement comme une sorte d'eruption de celleci. Mais que signifie alors le etiam? A priori trois sens sont possibles: (1) ou bien on donne a ce terme le sens, qui est le plus frequent en latin, de 'aussi, egalement', et cela implique que de la katalepsisagnoia, sortent a la fois l'opinion et autre chose qu'elle, qui ne pourrait etre qu'une autre katalepsis; (2) ou bien on le prend comme un synonyme de rursus et Ton affirme que, comme dans un circuit ferme, la katalepsis du sot est deja une opinion qui donne naissance a d'autres opinions, ce qui rejoindrait Interpretation proposee par A. A. Long; (3) ou bien, enfin, on l'interprete comme un simple jalon chronologique, et Ton comprend que de la comprehensio-inscientia, qui n'est pas elle-meme une opinion, nait l'opinion. La premiere lecture nous semble devoir etre ecartee parce que trop contraire a la structure meme du texte ciceronien. 26 Le choix entre les deux autres peut se resumer a la question suivante: qu'est-ce qui chez le sot peut differencier la katalepsis de la doxa, compte tenu du fait que, si la premiere est chez lui une agnoia,27 elle se definit comme un assentiment faible et inconstant, ce qui correspond aussi a la definition de la doxa? La reponse se trouve dans le texte de Ciceron lui-meme, qui n'accorde aucune valeur positive a l'opinion, alors que la comprehensio est la seule chose a laquelle on puisse faire confiance (soli credendum esse). Cela signifie done que, meme si du point de vue de l'hegemonique du sot, il n'y a pas de difference de nature entre la katalepsis et la doxa, il existe un autre point de vue qui differencie ces concepts et qui est celui de la gnoseologie. Autrement 25
Cf. la phrase tres significative q u e Ton trouve d a n s Rep. 1.69: ut existat ex rege dominus, ex optimatibus factio, ex populo turba et confusio. Le verbe existere, generalement mal compris, est,
26
27
en revanche, bien traduit {arisefrom) p a r A r t h u r (1983) p . 77. O n p e u t c e p e n d a n t regretter qu'il n'ait pas percu les difficultes issues d u etiam. T o u t ce passage a, en effet, une forte structure chronologique: visum ... I cum acceptum iam et approbatum esset I ... Quod autem erat sensu comprehensum I ... scientiam ... inscientiam ... ex qua exsisteret etiam opinio. L'ocyvoioc est ainsi definie p a r Stobee, SVF 1.6S: TT\V ayvoiocv lieTaTTTcoTiKfjv eTvou cruyKaTa8Ecriv Koci &a8evfj; cf. egalement Sextus, Adu. math, vii.432: Trotaa 9oOAou KOTT' OCUTOUS CrnroAriyis d y v o i a eori.
258
CARLOS LEVY
dit, meme chez le sot la perception sensorielle a une valeur gnoseologique, alors que celle-ci est deniee a la doxa, ce qui fait assurement probleme, pour nous, mais temoigne bien de ce qu'etait l'inspiration premiere de la logique stoi'cienne. A partir de ces remarques, nous proposons Interpretation suivante: - quod autem erat sensu comprehension, id ipsum sensum appellabat: la comprehension est ici envisagee comme neutre (ni faible ni forte), non pas qu'elle soit incorporelle, mais parce que le point de vue choisi est celui de la prehension sensorielle de l'objet; l'existence de ce point de vue nous parait confirmee par le fait que la comprehensio est definie comme moralement neutre (neque in rectis neque in pravis), ce qui ne serait pas le cas si elle n'etait envisageable que comme etat d'un hegemonique particulier; — si ita erat comprehensum . . . nominabat: la realite est exactement la meme, mais envisagee du point de vue de la qualite de l'hegemonique; - ex qua exsisteret etiam opinio: des comprehensions qui resultent d'une perception sensorielle correcte, mais appartiennent a des ames faibles, ne sont des opinions que potentiellement, mais donnent naissance a des doxai, lesquelles expriment precisement a la fois l'etat de Fame du sujet et son incapacite a comprendre la signification veritable de l'objet. Si Ton admet cette interpretation, il faut comprendre la doxa zenonienne comme le devenir des katalepseis quand elles ne sont pas le fait de Tame du sage; — quam esset imbecilla et cumfalso incognitoque communis: cette expression pose le probleme de la relation entre le texte de Ciceron et celui de Stobee. Y a-t-il ici, comme chez Stobee, l'indication qu'il existe deux types d'opinion, ou bien Zenon confondait-il ce qui devait etre dissocie par ses successeurs? Jusqu'ici Ciceron a suivi le devenir de la katalepsis dans l'ame du sot et imbecilla se refere a l'opinion issue de cette comprehension. Mais precisement, l'expression cum falso incognitoque communis ne peut concerner directement une doxa qui nait d'une katalepsis^ c'est-a-dire de l'apprehension d'une realite. II faut done admettre que Ciceron greffe a cet endroit une autre definition de l'opinion, celle de l'assentiment a ce qui n'existe pas ou que Ton ne perfoit pas, et que sur le fond son temoignage ne differe done pas de celui de Stobee. Cependant la maladresse de l'articulation entre les deux formes d'opinion dans le texte ciceronien met paradoxalement en evidence la coherence de la pensee: que les assentiments du
Le concept de doxa des Stoiciens a Philon d'Alexandrie
259
sot soient donnes a des representations comprehensives ou a des fantasmes, ils n'aboutissent en aucun cas a un savoir veritable. Cette lecture du passage ciceronien, concorde-t-elle avec Sextus, Adv. math, vn, 150 ss., qui ne mentionne pas explicitement Zenon mais dont on peut penser qu'il se refere a lui par Arcesilas interpose? II ne nous semble pas que les deux temoignages soient incompatibles, compte tenu du fait que Sextus, a la difference de Ciceron, n'expose pas la pensee zenonienne dans sa purete, si Ton peut dire, mais dans la perspective du debat academico-stoicien. De ce fait, il ignore une nuance aussi importante que celle du exsisteret etiam ciceronien et il ne mentionne nulle part Yinscientia. Mais nous nous demandons s'il faut voir la autre chose qu'une contraction de ce qui ne paraissait pas indispensable pour la comprehension du conflit Arcesilas-Zenon. En revanche, la neutralite de la katalepsis est affirmee en des termes qui ne sont pas essentiellement differents de ceux de Ciceron et la position de la comprehension comme critere de la verite est exprimee sans ambiguite. En realite, ce qui complique les choses, c'est qu'Arcesilas s'est refuse a admettre dans sa complexite cette volonte de prendre en compte a la fois les donnees objectives de la connaissance et la qualite du sujet, qui semble avoir caracterise la pensee de Zenon. En objectant au Stoicien que la meme realite est opinion chez le sot et science chez le sage, Arcesilas feignait d'ignorer l'originalite profonde du stoicisme et reduisait celui-ci a un relativisme sans prise sur la realite des choses. En raisonnant comme si Zenon ne s'etait preoccupe que du sujet dans sa definition de la doxa, Arcesilas allait, volontairement ou non, retrouver une attitude platonicienne, celle consistant a definir l'opinion par rapport au statut de son objet. 28 Nous illustrerons cette difference entre Academiciens et Stoiciens en utilisant une metaphore qui se trouve chez Epictete. 29 Celui-ci compare Tame a une cuvette pleine d'eau, la representation a un rayon et il dit que 'quand l'eau est agitee, il semble que le rayon aussi soit agite; or il ne Test pas'. Mais precisement, pourquoi cette eau est-elle agitee? Un Stoicien comme Posidonius estimera devoir expliquer, en des termes qui ne sont plus ceux du stoi'cisme orthodoxe, pourquoi cette cuvette dont l'etat naturel devrait etre le repos se trouve en perpetuelle agitation et il elaborera la theorie des dunameis irrationnelles. La 28 29
L'importance d u probleme de l'objet dans la refutation academicienne d u stoicisme a ete justement per$ue p a r Ioppolo (1986) p p . 59 ss. Epictete, Diss. 111.3.20.
260
CARLOS LEVY
critique academicienne est beaucoup plus radicale, elle ne nie certes pas que la representation existe, mais elle conteste qu'elle soit ce rayon de lumiere qui nous revelerait la verite des choses. LA NOUVELLE ACADEMIE: UNE ECOLE SANS PSYGHOLOGIE?
Quiconque s'en tiendrait strictement aux temoignages relatifs a Arcesilas et a Carneade aboutirait a la conclusion que pendant leurs nombreuses annees de scholarquat ces Academiciens ne se sont interroges ni sur la nature de Tame ni sur le probleme du fonctionnement psychique autrement qu'en termes stoi'ciens. Ce sont des Stoiciens (Pane this, Posidonius), et non des gens institutionnellement lies a la tradition platonicienne, qui ont critique le monisme psychologique du Portique dans un langage rappelant le Phedre ou la Republique. Et si les Tusculanes montrent que Carneade a critique la methode therapeutique preconisee par Chrysippe a l'egard de la passion,30 on y chercherait en vain un quelconque indice des attaques que Ton eut ete en droit d'attendre de la part d'un tel dialecticien contre l'assimilation du pathos a un jugement. Comme ce n'est pas l'epistemologie de la doxa que nous etudions ici, mais la psychologie de celle-ci, deux points vont nous interesser plus precisement: le probleme general de l'attitude de la Nouvelle Academie a l'egard de la question de Tame et celui d'un possible arriere-plan platonicien dans sa maniere d'utiliser certains concepts psychologiques stoi'ciens lies a la doxa. Le texte le plus important pour definir l'attitude de la Nouvelle Academie a l'egard de la psychologie se trouve a lafindu Lucullus.31 Voici la traduction que nous en proposons: Mais je reviens a Tame et au corps. Savons-nous suffisamment quelle est la nature des nerfs et des veines? Connaissons-nous la nature de Tame, sa localisation? Savons-nous meme si elle existe, ou si, comme le croit Dicearque, elle n'existe meme pas? Si elle existe, a-t-elle trois parties, comme le 30 31
Cf. Tusc. in.59-60, et Particle d e Ioppolo (1980), ou il est montre q u e C a r n e a d e assume r&Trpoo"56Kr|Tov comme cause des passions, mais a titre dialectique. Ciceron, Luc. 124: Sed redeo ad animum et corpus. Satisne tandem ea nota sunt nobis, quae nervorum natura sit, quae venarum? Tenemusne quid sit animus, ubi sit, denique sitne an, ut Dicaearcho visum est, ne sit quidem ullus? Si est, trisne partes habeat, ut Platoni placuit, rationis, irae, cupiditatis, an simplex unusque sit: si simplex, utrum sit ignis an anima, an sanguis, an, ut Xenocrates, numerus nullo corpore, quod intellegi quale sit vix potest: et, quidquid est, mortale sit an aeternum? Nam utramque in partem multa dicuntur. Horum aliquid vestro sapienti certum videtur, nostro ne quidmaxime quidem probabile sit occurrit: ita sunt in plerisque contrariarum rationum paria momenta. Nous avons modifie sur plusieurs points la traduction Brehier-Goldschmidt.
Le concept de doxa des Stoi'ciens a Philon d'Alexandrie
261
pense Platon, la raison, la colere et le desir, ou bien est-elle simple et une? Si elle est simple, est-elle du feu, un souffle, du sang, ou, comme le pense Xenocrate, un nombre incorporel, ce qui est a peine inconcevable? Et, quelle que soit sa nature, est-elle mortelle ou eternelle? En effet, il existe bien des arguments dans les deux sens. Votre sage tiendra pour certaine Tune de ces theses, le notre ne voit meme pas quelle est la plus probable: tant il est vrai que sur la plupart des questions les arguments contradictoires ont un poids egal.
Ce texte ne presente pour ainsi dire pas de difficulte d'interpretation et il apporte une reponse que Ton est en droit de juger satisfaisante aux problemes que nous avons souleves: nous n'avons pas de document psychologique en ce qui concerne la Nouvelle Academie, parce que ses scholarques avaient estime que cette question etait particulierement obscure et qu'aucune assertion sur ce point n'etait pas meme probable. Remarquons en outre que cette position semble avoir ete conservee meme par ceux qui avaient estime devoir proceder a une attenuation de la position d'Arcesilas. En effet, alors qu'Arcesilas proclamait que tout est recouvert de tenebres et que toutes les affirmations sont equipollentes, dans le passage du Lucullus consacre au dissensus la metaphore des tenebres est reservee aux seuls problemes de la physique, et il y a une restriction importante en ce qui concerne le principe de l'isosthenie, puisque celui-ci n'est plus universel, mais concerne seulement le plus grand nombre de questions (in plerisque). Comment des hommes qui etaient institutionnellement les successeurs de Platon en sont-ils arrives a une telle position? II est a remarquer que c'est Ciceron lui-meme qui semble nous donner lui-meme la reponse, lorsque, quelques lignes avant le passage que nous avons cite, il oppose aux dogmatiques l'exemple de Socrate et d'Ariston de Chios qui, eux, affirmaient ne pouvoir rien connaitre sur les questions de la physique.32 Lorsque Arcesilas ou Carneade tournaient en derision les pretentions des physiciens, ils pouvaient se referer a l'exemple de Socrate qui, parce qu'il considerait que la solution de tels problemes etait hors de la portee d'un esprit humain, comparait les physiciens dans leur relation reciproque a des fous.33 Les limites d'une telle demarche sont cependant evidentes, puisque l'aversion de Socrate pour le dogmatisme des physiciens ne l'empechait pas d'avoir lui-meme des croyances en ce qui concerne la 32 33
Ibid. 123: Liber igitur a tali irrisione Socrates, liber Aristo Chius, qui nihil istorum sciri putat posse. X e n o p h o n , Memorables 1.1.13.
262
CARLOS LEVY
nature de Tame. Mais qu'en etait-il de Platon dans une telle perspective aporetique? Dans ce texte le fondateur de l'Academie apparait bel et bien comme quelqu'un qui s'est differencie de Socrate sur le probleme de la physique, et qui a soutenu a propos de Tame une position precise, dont il ne nous est certes pas dit qu'elle etait dogmatique. Or la Nouvelle Academie n'a jamais pretendu rompre avec Platon, elle Pa au contraire presente comme Pune des references de sa pensee et de sa methode. II est done a priori peu probable qu'elle ait fait sienne Popposition entre un Socrate uniquement preoccupe de morale et un Platon pythagoricien, la presence de ce theme chez Ciceron devant plutot etre consideree comme une consequence de l'enseignement d'Antiochus d'Ascalon.34 En realite, pour comprendre la relation entre la Nouvelle Academie et Platon sur cette question de Fame, il faut se reporter a ce que dit Ciceron lorsqu'il explique pourquoi il n'y a pas eu de veritable rupture dans l'histoire de l'Academie:35 Platonem ... cuius in libris nihil adfirmatur et in utramque partem multa disseruntur, de omnibus quaeritur, nihil certi dicitur. On remarquera la conjonction d'une proposition generate (rien n'est affirme dans ses livres) et d'une proposition restrictive (beaucoup de choses sont discutees de maniere contradictoire). Autrement dit, meme lorsque Platon ne s'en tient pas a la disputatio in utramque partem, il etait presente par Arcesilas et par Carneade comme leur precurseur, parce que ne procedant pas selon eux de maniere dogmatique. La Nouvelle Academie ne pouvait nier que Platon avait eu des convictions en ce qui concerne Fame, mais elle devait valoriser toutes les variations ou les formules dubitatives des textes psychologiques platoniciens, de maniere a montrer qu'il n'y avait pas la une veritable doctrine. On notera, en outre, que le texte ciceronien ne dit pas que 1'attitude du sage doit etre seulement celle du constat d'isosthenie. L'expression ne quid maxime quidem probabile sit occurrit signifie, en raison meme du sens du verbe occurro,36 que le sage academicien, place devant toutes les doxai sur Fame ne voit pas, a cause precisement de l'isosthenie des discours, quelle est la theorie la plus vraisemblable. Mais cela n'implique pas qu'un tel domaine doive etre a priori exclu de sa recherche. Ce qui est certain, en 34 36
35 Cf., par exemple, Ciceron, de Rep. 1.16. Ciceron, Ac. post. 1.46. Le verbe occurro indique l'immediatete, cf. Luc. 35: Nam si quod cuique occurrit et primo quasi adspectu probabile videtur, id conjirmatur, quid eo levius?
Le concept de doxa des Stoiciens a Philon d'Alexandrie
263
revanche, c'est que ce sage n'abordera pas le probleme de Tame de la meme maniere que les pkysici ou les Stoiciens. II faut done admettre que la doxographie du Lucullus ne doit pas etre lue comme si toutes les doxai evoquees avaient le meme statut. Platon etait aux yeux d'un Neoacademicien beaucoup moins coupable que les autres philosophes cites, puisqu'il n'avait pas dogmatise sur un tel sujet, comme d'ailleurs sur aucun autre. II reste neanmoins une question difficile et importante: la Nouvelle Academie en est-elle restee a cette dissociation du contenu de la psychologie platonicienne - ni plus ni moins defendable que n'importe quelle theorie psychologique - et de sa forme, consideree par elle comme non dogmatique? A-t-elle, au contraire, marque, sous une forme ou sous une autre sa preference pour le contenu meme de la pensee de Platon dans ce domaine? Remarquons d'abord la situation particuliere de cette doxographie de Tame du Lucullus. A partir du § 116, et pratiquement jusqu'a la fin de l'ouvrage, Ciceron utilise des doxographies pour demontrer la complexite des problemes dans chacune des trois parties de la philosophic et sa tendance generate dans cette partie du discours est a l'accentuation de la difficulte des problemes, de fa9on a mettre en evidence qu'ils ne sauraient etre traites avec lapropeteia des dogmatiques. Mais ces doxographies ne sont pas un aboutissement, elles constituent le point de depart d'un programme de recherches qui est exprime de la maniere la plus claire dans les dernieres phrases du dialogue.37 Or le probleme de la nature de Fame est repris par Ciceron, non pas dans une oeuvre de physique, mais dans un texte appartenant a l'ethique, la premiere Tusculane,38 le passage d'un domaine a l'autre etant deja en lui-meme tres significatif. Les ressemblances entre les deux textes sont tres frappantes et la doxographie du Lucullus apparait tres precisement comme la forme abregee de celle de la Tusculane.39 II est remarquable, en outre, que cette Tusculane traite du caractere perissable ou immortel de Fame, e'est-a-dire precisement de la question sur laquelle s'achevait la doxographie du Lucullus. On remarquera enfin que la conclusion que tire Ciceron est sensiblement la meme dans les deux cas: 37
38 39
Gf. Ciceron, Luc. 147: Posthac tamen, cum haec quaeremus, potius de dissensionibus tantis summorum virorum disseramus, de obscuritate naturae deque errore tot philosophorum, qui de bonis contrariisque rebus tanto opere discrepant. Uobscuritas naturae sera traitee dans le de Natura Deorum, le de Divinatione et le de Fato; l'ethique dans l'ensemble de Finibus-Tusculanes. Ciceron, Tusc. 1.18-23. Cf. notamment, au § 20, la presence du groupe Xenocrate-Platon-Dicearque.
264
CARLOS LEVY
n'exprime-t-il pas aussi a la fin de la doxographie de la Tusculane l'idee que la connaissance de la nature de Fame est au-dessus de ce dont est capable un esprit humain et qu'il n'est meme pas possible de discerner immediatement la plus grande vraisemblance?40 Or, a partir done de ces memes conclusions, Ciceron va neanmoins essayer de demontrer dans la Tusculane que la these de l'immortalite de Fame est superieure a celle de sa disparition apres la mort, parce qu'elle apporte un espoir qui est moralement utile.41 Les Tusculanes sont une oeuvre complexe qui doit etre envisagee dans son ensemble et non livre par livre. Dans l'etude que nous en avons faite ailleurs,42 nous avons essaye de montrer que cet ensemble se rattache non pas a des courants heterodoxes du stoi'cisme, mais a l'Academie de Philon de Larissa. Le probleme qui nous interesse ici peut neanmoins etre detache de la question generale des sources des Tusculanes et il doit etre formule ainsi: a qui attribuer cette articulation entre la suspension du jugement en ce qui concerne la nature de Fame et l'acceptation, sur le mode de l'espoir, de la these de son immortalite?43 Nous sommes la assurement dans le domaine de la conjecture et, plutot que d'enumerer les differentes hypotheses possibles, il convient de poser directement la question suivante qui pourra paraitre hasardeuse mais qui nous parait inevitable si Ton veut approfondir le probleme de la psychologie de la Nouvelle Academie: Arcesilas et Carneade - nous choisissons volontairement les cas les plus difficiles - pouvaient-ils articuler eux aussi la disputatio in utramque partem et une psychologie 'transcendantale'? Precisons la question: il ne s'agit pas de dire si logiquement les deux sont 40
41
42 43
Ibid. 23: Harum sententiarum quae vera sit, deus aliqui viderit; quae veri simillima, magna quaestio est. La premiere partie de cette phrase semble apporter u n element nouveau p a r rapport au Lucullus, a savoir l'idee que Dieu serait detenteur de la verite. Dans cette distinction entre savoir h u m a i n et savoir divin, on reconnaitra u n theme socratique important, cf. Vlastos (1985). Y a-t-il dans cette reference a u n savoir divin quelque chose qui introduise u n e difference fondamentale entre les deux textes? Sans sous-estimer l'importance d u probleme, on ne peut exclure q u e le deus aliqui viderit soit ici u n e b o u t a d e de Ciceron lui-meme. Cf. le § 71, ou Ciceron d e m a n d e a son interlocuteur, qui a ete totalement convaincu p a r la demonstration de l'immortalite de Fame, d e ne pas considerer celle-ci comme u n dogme. Le probleme pose dans cette Tusculane est d'ordre moral: comment faire disparaitre l'angoisse de la mort? L a solution de l'immortalite n'est pas preferee pour des raisons d'ordre physique - de ce point de vue-la, elle n'est ni plus ni moins defendable q u ' u n e autre - mais parce qu'elle apporte a Fame u n plus grand espoir et une plus grande resistance a la peur de la mort. Levy (en cours d'impression). Ni Panetius, ni Posidonius, ni Antiochus, qui sont cites p a r Gigon (1955), p . 59, comme sources possibles de cette Tusculane ne paraissent a priori susceptibles d ' u n e telle demarche philosophique.
Le concept de doxa des Stoiciens a Philon d'Alexandrie
265
compatibles, mais plus modestement de chercher si nous avons des traces de ce qui apparaitrait alors comme la racine de la position ciceronienne dans la Tusculane. Sans pouvoir apporter une reponse ferme, nous donnerons quelques precisions susceptibles de montrer que le probleme est plus complexe encore qu'il n'y parait: - en ce qui concerne Arcesilas lui-meme, trois fragments peu souvent cites montrent qu'il ne s'etait pas contente de pratiquer le contra omnia, mais qu'il avait fait etat de croyances positives, temoignant d'un souci de perfection morale et de la presence en lui de themes ayant une certaine resonance socratique: il avait fait l'eloge de la pauvrete, la comparant a une bonne nourrice et la definissant comme un exercice de vertu; 44 par ailleurs, il disait que la mort est, parmi ceux que les hommes considerent comme des maux, le seul qui n'ait jamais fait souffrir par sa presence mais par son attente; 45 enfin il reprochait aux hommes de s'interesser exclusivement aux biens d'autrui et de negliger ceux qu'il portent en eux. 46 La simple prudence implique de ne pas surexploiter ces textes malheureusement trop courts et formules de maniere topique. On peut cependant se demander si un homme qui semble avoir ete fascine par le personnage de Socrate 47 a reussi a se detacher totalement de la these de Pimmortalite de Pame, ou s'il Pa conservee en lui et dans son enseignement comme espoir, auquel cas la premiere Tusculane exprimerait, sous une forme 'modernisee', quelque chose qui existait des Porigine de la Nouvelle Academie; — le second texte, qui se trouve dans les Prolegomena in Platonis philosophiam, est encore plus interessant, car il fait, lui, explicitement reference au Phedon.*8 II mentionne aussi explicitement la Nouvelle Academie, mais il faut reconnaitre que cette allusion est formulee de maniere quelque peu ambigue. 49 Admettons, ce qui parait malgre tout vraisemblable, que les arguments cites en faveur d'une lecture sceptique de Platon aient ete reellement soutenus dans PAcademie. 44 45 46 47 48 49
Plutarque, frag. 152 S a n d b a c h = Arcesilas, frag. 11 Mette. Plutarque, Consolation a Apollonios, 15 = frag. 10 Mette. Plutarque, irepl sudvfjfas 9 = frag. 9 Mette. Cf. de Fin. 11.1, a propos de la continuite entre Socrate et Arcesilas dans la dialectique, a quoi il faut ajouter le refus arcesilien de l'ecrit. Westerink (1962) p p . 2 1 - 2 . Ibid. 2 1 , 1.1-6. L ' A n o n y m e commence p a r dire q u e Platon est superieur a la Nouvelle Academie, puis il ajoute q u e 'certains' o n t affirme q u e Platon avait professe l'acatalepsie. Le texte ne permet pas d'affirmer avec certitude qu'il s'agit de Neoacademiciens. Mais il n'y a rien d'invraisemblable a ce que les Tives, s'ils n'etaient pas eux-memes des Neoacademiciens, aient utilise u n e argumentation elaboree dans l'Academie.
266
CARLOS LEVY
Que dit le quatrieme de ces arguments? Que Platon est de maniere evidente quelqu'un qui pense que rien ne peut etre per^u avec certitude, puisqu'il se defie des sens et qu' il affirme que Tame n'apprehende rien par l'intelligence, 'entrelacee qu'elle est a ce mal qu'est le corps'. 50 II y a la une interpretation evidemment abusive du Phedon, et il est possible que les Neoacademiciens, a la recherche d'une legitimite platonicienne, ne se soient interesses dans ce passage du Phedon qu'a la these socratique de la necessaire incompletude du savoir ici-bas. Trop heureux de trouver une telle justification, ils se seraient alors empresses de radicaliser cette affirmation et de l'etendre a l'ensemble de l'oeuvre platonicienne, sans pour autant assumer de quelque maniere que ce soit l'enracinement metaphysique d'un tel scepticisme. Mais c'est la une hypothese seulement, a laquelle on peut opposer une autre hypothese, a savoir que lorsque le Phedon etait etudie dans l'Academie d'Arcesilas et de Carneade la frontiere entre interpretation aporetique et croyance metaphysique etait moins tranchee que nous n'aurions tendance a le penser, a partir notamment de schemas neopyrrhoniens. On aura beau jeu de nous repliquer qu'il n'est pas utile d'entrer dans de pareilles considerations, puisque, a supposer meme que les scholarques de la Nouvelle Academie aient garde comme une croyance la psychologie platonicienne, cela ne les a pas empeches de se referer exclusivement, en tout cas a juger par nos temoignages, aux concepts psychologiques du Portique. Effectivement, dans la mesure ou la Nouvelle Academie ne s'est exprimee qu'avec des concepts stoiciens, on pourrait penser que toute reference de la part de l'historien a la psychologie et a la metaphysique platoniciennes n'aboutit qu'a ajouter de la confusion a un debat deja passablement complique. Nous voulons essayer, au contraire, de montrer que ce sont les textes eux-memes qui peuvent nous inviter a faire intervenir ces speculations et a nous poser le probleme de la lecture de Platon dans la Nouvelle Academie. Nous nous appuierons pour cela sur deux temoignages, tres etroitement lies au probleme de la doxa. Le premier texte est le fameux passage du Contre Colotes relatif a la horme et il est generalement attribue a Arcesilas, mais sans qu'on sache exactement s'il faut lui attribuer tout ou partie de cet ensemble qui va de 1122A, a 1124B.51 Or ce probleme de la delimitation est 50 51
Ibid. 1.25-31, cf; Platon, Phedon 66b et 79c. Long & Sedley (1987), frag. 69a, arretent leur citation a 1122, en supprimant les passages qui, a l'interieur de cette delimitation, font explicitement reference a u x Epicuriens. U n tel
Le concept de doxa des Stoiciens a Philon d'Alexandrie
267
essentiel, car selon que Ton choisit l'une ou l'autre attitude, l'image de la Nouvelle Academie sur cette question apparaitra differente. Par ailleurs, on ne semble pas s'etre interesse au probleme des etapes intermediates entre Arcesilas et Plutarque, etapes pendant lesquelles la pensee du scholarque a pu etre a la fois conservee dans sa lettre, si Ton peut dire, et incluse dans un reseau d'arguments qui etaient etrangers a son esprit. Nous ferons done a ce sujet les remarques suivantes: - on ne saurait contester a Arcesilas la paternite de la these - nous employons ce terme en fonction de son aptitude a exprimer soit une proposition a laquelle on adhere, soit une proposition que Ton defend dialectiquement - selon laquelle la presence d'une image mentale oikeia suffit a mettre en branle la horme sans qu'il y ait intervention de l'assentiment, et done sans opinion. Mais cela ne nous renseigne pas sur le probleme essentiel, a savoir le statut de cette these, et, par ailleurs, nous savons grace a Plutarque que d'autres Neoacademiciens avaient defendu la meme proposition. 52 Or, si Ton regarde de pres le texte du Contre Colotes, que constate-t-on? Plutarque mentionne expressement Arcesilas en ce qui concerne la volonte de se rattacher a la tradition philosophique la plus reculee, 53 mais lorsqu'il aborde la question de Vepoche, il n'evoque plus ce scholarque;54 en revanche, il se reftre, en des termes qui ne sont pas tres clairs, a des adversaires de Yepoche qui n'etaient pas Stoiciens et qui avaient neanmoins emprunte au stoicisme 1'argument de Yapraxia dirige contre l'Academie. 55 Bien plus, il fait etat d'agones nomimoi engages par les Academiciens avec ces gens-la, et il oppose la qualite de leurs reponses aux sottises de Colotes. Pourquoi done cette longue allusion qui rompt le face a face Arcesilas-Colotes tel qu'il avait ete presente en 1121F? N'est-ce pas parce que Plutarque va utiliser pour defendre Yepoche'une argumentation qui s'est forgee dans ces nomimoi agones, et qui done est tres posterieure a Arcesilas?
52 53 54 55
choix fait probleme: pourquoi les paragraphes posterieurs a 1122f ne figurent-ils pas dans la citation, alors qu'ils sont construits exactement c o m m e celui-ci, a savoir comme des reponses a des critiques d ' u n meme objecteur? P a r ailleurs, Striker (1980), p . 65 n.29, et Ioppolo (1986), p . 139, accordent u n e importance certaine pour la comprehension de la pensee d'Arcesilas a la citation de Platon {Rep. 458d), qui figure en 1122e, mais ne precisent pas ou s'arrete selon elles la citation d'Arcesilas. Cf. Plutarque, Sto. rep. 1057a. Plutarque, Adv. Col. 1 1 2 i e - i 122a, j u s q u ' a eis 'ApxeaiAaov d m ^ a i v o v n . Ibid. ii22a-b. Ges gens ne peuvent etre que des Epicuriens posterieurs a Colotes; l'hypothese de D e Lacy (1956), p . 74, q u i voit la u n fragment d'Antiochus, ne nous semble pas fondee.
268
CARLOS LEVY
- c'est la precisement le coeur du probleme. Quelles raisons precises avons-nous de penser que, contrairement a Interpretation qui a ete donnee de ce texte, 56 le naturalisme que Ton trouve dans ce passage est le fait de Plutarque ou de sa source, non d'Arcesilas? Pour Plutarque, au moins dans ce texte, la nature est effectivement, comme dire Montaigne dans un autre con texte, 'un doux guide', qui presente des phantasiai oikeiai par lesquelles il suffit de se laisser guider pour pouvoir agir sans courir le risque de se fourvoyer dans l'opinion. Deux details nous laissent cependant penser que Plutarque interprete Arcesilas en fonction de ses propres convictions. Tout d'abord il fait grief aux Epicuriens de ne pas croire a la divination et, en cela, il est fidele a lui-meme, puisque nous savons par le Catalogue de Lamprias qu'il avait ecrit un ouvrage pour demontrer que l'Academie ne rejetait pas la divination. 57 Etait-ce la la position de la Nouvelle Academie? II suffit de se referer au de Diuinatione ciceronien pour montrer que Carneade avait refute avec force la divination stoi'cienne. Plus revelateur encore: Plutarque reproche aux Epicuriens de ne pas accepter cette evidence absolue qu'est l'amour des parents pour les enfants. Or il s'agit la d'un des aspects de Yoikeiosis stoi'cienne, et nous savons par une lettre de Ciceron, ecrite a Atticus a l'occasion de la naissance de la fille de celui-ci, que Carneade contestait fortement l'existence de ce lien, mais beaucoup plus dans un esprit de provocation dialectique que dans l'optique d'une veritable doctrine de l'egoi'sme naturel. 58 Devant un tel texte plusieurs demarches sont possibles, compte tenu du fait qu'il nous parait totalement impossible d'y voir la transcription fidele de propos d'Arcesilas. II se peut que les themes stoiciens que nous y trouvons aient ete defendus par la source neoacademicienne de Plutarque disserendi causa, a la seule fin de contredire les Epicuriens. II se peut aussi que ce melange de scepticisme et de confiance en la nature exprime reellement la pensee de gens qui se reclamaient de la Nouvelle Academie, mais qui etaient tres posterieurs a ses grands scholarques. Or la pensee de ces derniers nous semble pouvoir se resumer en deux propositions: (a) la nature permet Faction, meme si Ton suspend son assentiment; (b) le bien agir exige l'intervention de la raison, qui verifie autant que faire se peut les donnees des sens, et qui rend possible la justification d'une conduite. II y a une delimitation des domaines respectifs de l'instinct 56 58
Cf. Ioppolo, (1986) pp. 139-40. Ibid., cf. Ciceron, Att. vii.2.4.
57
Plutarque, Adv. Col. 1123a, cf. Catalogue 128.
Le concept de doxa des Stoi'ciens a Philon d'Alexandrie et du rationnel, qui se differencie profondement de l'esprit du stoicisme, et a laquelle il n'est pas impossible de trouver une racine platonicienne, comme le montre la comparaison avec un passage de la Republique cite par Plutarque lui-meme, mais a contre-sens.59 Platon y evoque une action instinctive, qui n'est done pas guidee par des 'contraintes geometriques', et la presence chez Plutarque de cette citation a pu etre invoquee a l'appui d'une interpretation naturaliste de la philosophic d'Arcesilas. II nous semble, au contraire, que Platon n'y exprime pas sa confiance, mais sa defiance a l'egard de l'instinct.60 Rappelons ce qu'il en est. A cet endroit de la Republique, Platon parle de la reforme de la famille, et plus precisement du mariage des gardiens, mais en laissant la part la plus petite possible a ce qui ne releve pas de la 'necessite geometrique'. Que dit-il, en effet? Qu'il faut, de meme que Ton a choisi les hommes, choisir les femmes et assortir les natures semblables, ce qui constitue deja une premiere entrave au libre jeu de l'instinct. II faut ensuite donner a ces hommes et a ces femmes 'une education commune' et cette promiscuite fera qu'ils formeront des unions, du fait de la 'necessite naturelle'. Mais Platon, comme s'il tenait precisement a ecarter toute interpretation de sa pensee qui ferait la part trop belle a l'optimisme naturaliste, ajoute immediatement que former des unions au hasard serait une impiete dans une cite heureuse et qu'il s'agira done de faire les manages le plus saints possibles. Ainsi done la necessite naturelle se trouve doublement controlee: par la selection prealable et par le necessaire assentiment des chefs de la cite. Contrairement a l'epicurisme et au stoi'cisme, la Nouvelle Academie n'a pas cru trouver dans les donnees de la nature ni les premices du bonheur ni celles de la science. Arcesilas n'a pas commis l'incoherence qui eut ete de pretendre que le monde - qu'il decrivait comme recouvert de tenebres, et done comme une caverne sans sortie - etait phusikos guide de Faction. II est vrai que les Neoacademiciens n'ont pas epargne leurs critiques au logos, et cependant, e'est lui, et lui seul, qui exprime pour eux 1'aspiration au vrai et le souci de bien agir. Dans ce refus de confondre la raison et la nature, e'est l'un des 59 60
Platon, Rep. 458a. Sext. E m p . , Adv. math, vii.158. Sur la signification ad hominem, cf. Striker, (1980) p p . 65-6, et Burnyeat, Carneades was no probabilist, dont l'auteur a bien voulu nous faire parvenir le manuscrit. A l'inverse, Ioppolo (1986) p p . 121-30, a d o n n e une signification fortement positive a 1'EuAoyov arcesilien.
269
270
CARLOS LEVY
aspects importants de la pensee platonicienne qui nous parait se perpetuer, sous une forme que Ton est en droit de juger surprenante. Le deuxieme texte a deja fait l'objet lui aussi de nombreux commentaires et il a ete notamment interprets par M. Frede 61 comme l'acte de naissance d'un scepticisme negatif, posant dogmatiquement l'impossibilite de percevoir avec certitude quoi que ce soit. Voici la traduction que nous en proposons.62 Alors Catulus dit: 'Moi, je me reporte a l'opinion de mon pere, celle qu'en tout cas il attribuait a Carneade: je pense que rien ne peut etre percu et que cependant le sage donnera son assentiment a ce qu'il ne percoit pas, c'est-adire qu'il opinera. Mais je crois qu'il le fera en comprenant qu'il opine et en sachant qu'il n'y a rien qui puisse etre apprehende ou percu. C'est pourquoi, tout en approuvant que Ton suspende son assentiment en toute occasion, j'assentis avec force a cette proposition: il n'y a rien qui puisse etre percu.
Ce passage est presente generalement comme Pexpression de la pensee de Philon de Larissa, le seul probleme etant pour les tenants de cette these de determiner s'il s'agit du Philon d'avant ou d'apres les livres romains. 63 Nous avions deja combattu cette these il y a quelques annees et elle a ete egalement refutee par J . Glucker.64 Rappelons done que Catulus n'est pas dans le dialogue ciceronien le defenseur des innovations philoniennes, mais bien au contraire celui qui les a combattues avec une certaine brutalite, si Ton en juge par la phrase suivante: 65 Philo autem, dum nova quaedam commovet... aperte mentitur, ut est reprehensus a patre Catulo. S'il y a done une certitude
dans cette affaire assurement tres embrouillee, c'est que la position de Catulus ne peut etre assimilee a celle de Philon de Larissa, ou alors il faut taxer Ciceron d'incoherence et nous devons rejeter cette 61 62
63
64 65
Frede (1987) p . 212. Ciceron, Luc. 148: Turn Catulus: 'Egone?' inquit: 'ad patris revolvor sententiam, q u a m quidem ille C a r n e a d i a m esse dicebat, ut percipi nihil putem posse, adsensurum non percepto, id est, o p i n a t u r u m sapientem existimem, sed ita ut intellegat se opinari sciatque nihil esse quod comprehendi percipique possit. Q u a re ETTOXTIV illam omnium rerum comprobans, illi alteri sententiae, nihil esse quod percipi possit, vehementer adsentior.' Le qua re est une correction de Manutius, les manuscrits d o n n a n t per. Pour Long & Sedley (1987) t. 2, 451 = frag. 69 k, Catulus representerait ici le Philon d'avant les innovations romaines. O n pourrait a la rigueur admettre - encore q u ' a u c u n texte n'en fasse mention - q u e , deja a Athenes, Philon avait pris quelques libertes p a r rapport a la tradition de Clitomaque. Mais il nous semble impossible d'affirmer q u e la position de Philon etait alors celle de Metrodore. C'est a R o m e seulement que le scholarque adopta les theses de cet Academicien marginal, d'ou precisement le scandale. Cf. Levy (1980) p p . 30-46, ou nous avions sous-estime l'importance de Platon pour la nouvelle Academie, et Glucker (1978) p p . 396—7. Ciceron, Luc. 18.
Le concept de doxa des Stoiciens a Philon d'Alexandrie
271
solution de facilite. Mais, par ailleurs, il est au moins risque de faire le commentaire philosophique de ce texte sans savoir qui est le locuteur reel. Que disaient Metrodore et Philon de Larissa? Leur originalite par rapport a Interpretation 'orthodoxe', c'est a dire clitomaquienne, de la pensee de Carneade se manifeste sur deux points: - ils affirmaient que l'incomprehensibilite du monde n'est pas absolue, mais que les choses sont assurement acataleptiques si Ton s'en tient au critere stoi'cien;66 - ils affirmaient egalement que le sage donnera parfois son assentiment a l'opinion. Comment comprendre cette proposition? En 1'absence de toute autre precision, et en tenant compte de cette restriction temporelle, dont J. Glucker a souligne l'importance, 67 mais d'une maniere que nous n'approuvons pas, il convient d'interpreter cette these comme une mise en valeur de l'humanite du sage, capable comme tout un chacun de se tromper. Ainsi done Philon et Metrodore se seraient rattaches a l'inspiration presentee par Ciceron lui-meme comme platonicienne et peripateticienne. 68 Remarquons a cet effet que dans le passage du Pro Murena69 ou Ciceron refute le stoi'cisme de Caton, la proposition ipsum sapientem saepe opinari quod nesciat est 66 67
Cf. Sext. E m p . , Hyp. Pyr. 1.235 e t Augustin, Contra Academicos 111.41, a propos de Metrodore. Pour Glucker (1978) le texte d u Lucullus comporte trois interpretations d e la position d e C a r n e a d e sur l'assentiment d u sage: - l'exegese clitomaquienne: C a r n e a d e n'avait defendu q u e dialectiquement (disputatum) la these d u sage d o n n a n t son assentiment a l'opinion (78); - l'exegese de Philon et de Metrodore: C a r n e a d e avait approuve (probatum) cette meme these (ibid.); - la middle of the road interpretation: celle qui procede de maniere attenuee, en ajoutant des adverbes comme non numquam, interdum ou aliquando (59, 67, 112).
68 69
En realite, la distinction entre la deuxieme et la troisieme est tres contestable. E n effet, s'il est vrai q u e Ciceron n ' a pas exprime d'attenuation a u § 78, la lecture des § 59, 67, 112, montre qu'ils font allusion aux innovations philoniennes. Par ailleurs, on voit mal comment Glucker pourrait identifier le § 148 a sa middle of the road interpretation, alors q u e les attenuations dont il a lui-meme souligne l'importance, n ' y figurent pas. Cf. Luc. 112. Ciceron, Pro Murena 63: 'Nostri a u t e m - fatebor enim Cato me q u o q u e in adulescentia diffisum ingenio m e o quaesisse a d i u m e n t a doctrinae - nostri, i n q u a m , illi a Platone et Aristotele, moderati homines et temperati, aiunt a p u d sapientem valere aliquando gratiam: viri boni esse misereri: distincta genera esse delictorum et dispares poenas: esse a p u d hominem constantem ignoscendi locum: ipsum sapientem saepe aliquid opinari quod nesciat, irasci non n u m q u a m , exorari e u n d e m et placari.' Le Pro Murena n'est pas u n traite philosophique, mais cela n'empeche pas Ciceron d e s'exprimer avec precision. R e m a r q u o n s le pluriel (moderati homines et temperati), dans lequel on peut penser q u e Ciceron englobe a la fois Philon et Antiochus, faisant ainsi u n resume de ce qui lui apparait comme l'essentiel d e leur enseignement.
272
CARLOS LEVY
incluse dans un ensemble ou tout est destine a montrer que le sage n'est pas radicalement different du reste de l'humanite: il se met parfois en colere, il n'est ni inexorable ni implacable, il change parfois d'avis, etc. A la difference de Philon et de Metrodore, Catulus affirme sans restriction d'aucune sorte que tout est acataleptique, et sa conception de l'assentiment du sage a l'opinion est radicalement differente de celle de ces deux philosophes, puisqu'il s'agit non pas d'une faute qui rend le sage semblable a tous les autres hommes, mais d'un acte mental reflechi dans lequel s'exprimera sa sagesse. A partir de la un certain nombre de constatations nous paraissent s'imposer: - en ce qui concerne la forme meme de notre texte, il faut remarquer que Ciceron termine a chaque fois son etude d'une partie de la philosophic par une proposition dans laquelle les interlocuteurs-adversaires pourront au moins partiellement se retrouver. 70 Dans cette fin du Lucullus, Ciceron cherche de toute evidence a trouver une conciliation entre ceux qui preconisent Yepoche peri panton et ceux qui considerent que le sage n'a que des assentiments. Le vehementer adsentior par lequel Catulus, defenseur de Yepoche, commente la sententia Carneadia que lui avait transmise son pere, est de toute evidence destinee a suggerer cette conciliation finale; — la position de Catulus le pere se rattache non pas au courant faillibiliste philonien, mais au courant clitomaquien, dont il represente une variante interessante. Le fait d'affirmer que rien ne peut etre per^u ne constitue pas sur le fond une deviation par rapport a la pensee carneadienne, puisque Carneade lui-meme considerait que cette affirmation, etant autodestructrice, ne devait etre aucunement consideree comme une marque de dogmatisme. 71 II reste done cette idee que la specificite du sage est de donner son opinion, mais avec la conscience lucide qu'aucune connaissance certaine n'est possible. Quelle est la difference entre cette formule et le couple carneadien 'classique' epoche-pithanon? On peut evidemment discuter pour savoir si les deux attitudes (suivre le probable en suspendant son jugement/ donner son assentiment, mais en sachant que rien n'est cataleptique) sont deux attitudes psychologiquement equivalentes. L'essentiel 70
71
Ciceron, Tusc. v . 8 3 , ou Ciceron dit q u e C a r n e a d e avait l ' h a b i t u d e d e soutenir contra Stoicos q u e tous les philosophes, malgre leurs divergences, identifient la vertu a la vie heureuse; 120, o u C a r n e a d e est presente c o m m e u n honorarius arbiter, q u i renvoie dos a dos les Peripateticiens et les Stoiciens; cf. egalement la conciliation entre libre-arbitre et determinisme dans de Fato 44. Cf. Ciceron, Luc. 28.
Le concept de doxa des Stoiciens a Philon d'Alexandrie
273
nous semble etre cependant dans le changement de references. Le couple epoche-pithanon est constitue a partir de 1'utilisation au moins partiellement dialectique de concepts stoiciens. L'idee que le sage se caracterise par la conscience du caractere conjectural de ses affirmations et par la perception de l'universelle acatalepsie, renvoie au sunoida emautoi de Socrate, un Socrate revu et corrige par Arcesilas. Si Ton accepte l'idee que Interpretation de Catulus constitue l'expression en termes socratiques de la pensee de Carneade, on peut formuler la conclusion que ceux-la memes qui defendaient l'interpretation aporetique de la philosophic neoacademicienne avaient conscience de la continuite existant entre leur epoche et la demarche reflexive de Socrate. A cela il faut ajouter un element important, mais que la nature meme du sujet choisi pour ce colloque ne permet pas de traiter a fond: il s'agit du statut de l'objet de la doxa. Le stoicisme avait profondement innove par rapport a Platon en defendant cette idee que rien dans un monde parfaitement organise par la Providence n'est naturellement objet de doxa et que l'origine de l'opinion est a chercher dans le sujet rationnel lui-meme, qui est incapable de percevoir immediatement l'ordre des choses dans sa totalite et son excellence. En utilisant les termes de la psychologie stoicienne pour parler de l'opinion, la Nouvelle Academie montrait, au contraire, que le probleme de la doxa pouvait fort bien etre aborde independamment de toute conception personnelle de l'ame, et elle pressait son adversaire de revenir a la mise en question du monde. De fait, ses scholarques ont toujours rattache leur approche de l'opinion a une definition du statut des choses qui, perdue dans la duree, apparait comme liee au mythe de la caverne. Les choses sont entierement entourees de tenebres pour Arcesilas, et cette interpretation pessimiste du mythe s'explique au moins partiellement par la reaction a ce monde lumineux (cf. le rapprochement phoslphantasia) qu'est l'univers stoicien. Les choses ne sont plus qu'acataleptiques avec Carneade. 72 Elles deviennent comprehensibles en puissance avec Philon. Au fur et a mesure done que la confrontation avec le stoicisme perdait son caractere d'urgence absolue, 1'Academie a done retrouve une perception plus conforme a Platon de ce qui dans le monde est ombre et lumiere. Dans le meme temps, sa perception de la doxa a evolue. Au depart, le consensus academico-stoicien sur 72
Cf. Ioppolo (1986) pp. 66-70; 197.
274
CARLOS LEVY
la proposition 'le sage n'opinera pas' reposait sur un antagonisme radical: pour Zenon, le sage est sans opinion parce qu'il a compris la perfection du monde; pour Arcesilas cette absence d'opinion exprime le refus de s'engager dans un monde ou rien ne peut etre connu. Dans Interpretation orthodoxe-catulienne de Carneade, Yepoche peri panton se traduit en assentiment a l'opinion, mais avec la conscience que tout est acataleptique, ce qui subvertit profondement le sens que les Stoiciens avaient donne a l'assentiment, celui d'une adhesion totale. Avec Metrodore et Philon, la doxa apparait comme un risque que le sage, redevenu en fait philo-sophe, doit affronter, de par sa presence dans un monde ou, malgre la fragilite des sens, la connaissance des choses est possible. Nous avons essaye de montrer comment, a l'interieur meme de la Nouvelle Academie, le probleme de la doxa ne peut etre entierement coupe d'un arriere-plan platonicien. C'est un vaste probleme que de savoir si le devenir d'une pensee revele necessairement des elements qui etaient des l'origine inherents a celle-ci, et la relation entre le moyen platonisme et la Nouvelle Academie constitue assurement un probleme difficile. Mais, par ailleurs, le recours aux concepts de 'source' et de 'RezeptiorC occulte trop souvent cette realite qu'est revolution des idees elles-memes. On pourrait sans doute imaginer dans l'abstrait un Neoacademicien du premier siecle apres J. C. raisonnant exactement comme Arcesilas, ou un Stoicien de la meme epoque qui ne serait qu'un Chrysippe redivivus, mais l'experience des textes suffit a montrer la vanite d'un tel deni de l'histoire. Philon d'Alexandrie n'est pas une conscience intemporelle qui puiserait au gre de son libre-arbitre dans les systemes anterieurs, il n'est pas seulement quelqu'un qui lit la philosophie en pensant a la Bible, il represente egalement - comme les autres Medioplatoniciens — une attitude et une culture philosophique qui ont leurs racines, diverses et complexes, dans la confrontation entre l'Academie et le Portique. Cela n'annule pas la liberte du penseur, mais cela permet aussi d'examiner sa position sur la psychologie de l'opinion dans le prolongement du debat que nous avons decrit.
Le concept de doxa des Stoiciens a Philon d'Alexandrie
275
LA PSYGHOLOGIE DE PHILON D'ALEXANDRIE. JOSEPH OU LA DOXA 'MIXTE'
Nous essaierons a l'interieur de cette partie de montrer: (a) comment la psychologie de Philon amalgame des elements a la fois stoiciens et platoniciens, et comment cet eclectisme trouve une certaine coherence philosophique dans 1'utilisation par l'Alexandrin de la thematique neoacademicienne sur l'impossibilite de connaitre Tame avec certitude; (b) comment avec Interpretation du personnage de Joseph apparait une nouvelle figure de l'opinion, qui est a mettre en relation avec la notion d'ame moyenne, presente dans d'autres textes medioplatoniciens. La critique neoacademicienne avait mis en evidence les difficultes auxquelles se heurtait la conception stoicienne, purement subjective, de la doxa, et elle avait renoue a sa maniere avec la tradition platonicienne du fondement objectif de l'opinion. Le probleme ontologique se trouvait done pose et le moyen platonisme devait affirmer explicitement ce qui etait deja implicitement suggere par les Neoacademiciens, a savoir que la question de l'opinion ne prend tout son sens que par rapport a un etre transcendant au monde. (a) II n'est pour ainsi dire pas de page de l'important corpus philonien dans laquelle on ne trouve au moins une notation psychologique, et souvent ce sont meme de longs developpements qui sont eonsaeres a la problematique de l'ame. Si cet abondant materiau n'a suscite qu'un interet relatif chez les historiens de la philosophic, e'est qu'une fois circonscrits les quelques passages dans lesquels on reconnait telle quelle une doxa philosophique, par exemple ceux qui figurent dans les SVF, le reste semble destine a lasser la patience du chercheur le mieux intentionne, tant Philon semble s'etre ingenie a rendre impossible dans ce domaine aussi toute systematisation de sa pensee. E. Brehier, qui a eu le merite de chercher a mettre un peu de clarte dans cette confusion, a dresse la liste des passages ou Ton trouve les differentes divisions antiques de lapsuche:73 platonicienne, veteroacademicienne, peripateticienne et stoicienne. Son travail a ete approfondi per D. Runia, qui a etudie l'anthropologie platonicienne de Philon74 et en a conclu que la doctrine psychologique de celui-ci serait sur le fond celle de l'Ancienne Academie, relayee par 73
Brehier (1925), p. 138.
74
Runia, (1986) pp. 296-314.
276
CARLOS LEVY
un certain nombre de philosophes, parmi lesquels Posidonius. C'est cette idee que nous voudrions nuancer en etudiant les passages ou Philon n'utilise plus la division platonicienne, mais celle des Stoiciens. L'examen des textes stoicisants semble confirmer l'interpretation dualiste, et cela a deux niveaux: - meme lorsqu'il emploie la division de Fame en huit parties, il arrive a Philon de la ramener a une conception dualiste, en isolant l'hegemonique rationnel des autres parties irrationnelles. Ainsi, nous trouvons dans Quaest. Gen. 1.75, la remarque suivante: 'Quum octo partibus anima nostra constet, ex rationali individua et irrationali in septem partes distingui solita, in quinque nempe sensus, et in vocis instru-
mentum et geniturae' Est-il necessaire de preciser que dans aucun texte stoicien la division de Tame n'est ainsi interpretee dans un sens dualiste? — la transformation dans ces textes du monisme stoi'cien en un dualisme d'inspiration platonicienne apparait clairement aussi dans le choix des metaphores.75 Pour suggerer la nature de la relation existant entre l'hegemonique et les sept autres parties, les Stoi'ciens avaient utilise des metaphores variees (le poulpe, l'araignee, le roi et les messagers), mais ayant en commun de ne jamais suggerer une double nature de Tame. Or on s'aperfoit que Philon substitue parfois a ces metaphores stoiciennes des images qui, tout en conservant la division en huit parties de Tame, subvertissent profondement le sens de cette psychologic Ainsi dans Opif. 117, l'hegemonique est compare a un montreur de marionnettes qui, au moyen de fils, commande les sept autres parties, 'lesquelles tantot restent en repos, tantot se meuvent, chacune selon les figures et les mouvements qui lui sont appropries'. En se referant ainsi a l'image platonicienne du montreur de marionnettes,76 Philon assimile les parties de Fame humaine autres que l'hegemonique a des objets par eux-memes inertes, comme le sont les bras et les jambes d'une marionnette. Cette division, exprimee par la metaphore, entre, d'une part, l'hegemonique vivant et autonome, et, d'autre part, le reste de l'ame compare a des objets-instruments, est etrangere a 75
76
Sur le probleme des metaphores dans le stoicisme, on se referera a l'ouvrage de Rolke (1975), qui ignore cependant l'originalite de Philon dans le traitement des metaphores stoiciennes. Platon, Lois 644d. Cf. egalement Sacrif. 104, ou Philon utilise p o u r exprimer la relation d e l'hegemonique aux autres parties de l'ame, la metaphore platonicienne du Politique, celle du berger et du troupeau.
Le concept de doxa des Stoiciens a Philon d'Alexandrie
277
l'hylozoisme stoicien et va dans le sens d'une platonisation de la psychologie du Portique. Rien n'est cependant simple chez l'Alexandrin et si cette transformation en profondeur des metaphores stoiciennes montre qu'il y a eu de la part de Platoniciens (Eudore?), ou de Stoiciens heterodoxes - mais il faut reconnaitre que Ton ne trouve rien de semblable dans les fragments de Posidonius - une attitude qui consistait a parler le langage de la psychologie du Portique, pour exprimer quelque chose d'etranger a celle-ci, on trouve aussi dans le corpus philonien des metaphores stoi'ciennes non modifiees. Ainsi, dans Leg. 1.28, l'image de la source et des courants est reprise dans l'esprit d'un stoicisme orthodoxe:77 l'intelligence est tendue vers le sens et amene celui-ci a la saisie de l'objet. Plus interessante encore est la metaphore de la souche que Ton trouve dans Agr. 30: Philon dit que Fame fait naitre 'comme d'une seule souche' un double rejet: l'un insecable, qui est l'esprit, Tautre six fois divise correspondant aux cinq sens et aux deux autres organes. Runia a eu raison de signaler a propos d'autres textes que la metaphore de la racine a sa lointaine origine dans le Timee.18 Neanmoins, telle qu'elle est utilisee dans ce passage, elle correspond a l'une des expressions du monisme stoicien, comme le montre la comparaison avec un passage de Calcidius.79 Par ailleurs, Philon affirme a plusieurs reprises que la distinction entre rationnel et irrationnel n'implique pas a ses yeux que Fame ne soit pas une. 80 Y a-t-il une logique a tout cela, et est-il possible de definir une psychologie de Philon qui ne se resume pas a la simple application de lieux communs, choisis selon leur adequation a tel ou tel aspect du texte biblique? II nous semble que les variations philoniennes apparaissent comme moins incoherentes qu'il n'y parait, pour peu que Ton se refere a un certain nombre de textes dans lesquels apparait la dette de Philon a l'egard de la Nouvelle Academie, probablement par l'intermediaire des Neopyrrhoniens. Ainsi, il arrive a Philon d'evoquer la division tripartite de Fame, mais en ajoutant que pour certains il s'agit de veritables parties, tandis que pour d'autres il n'y a que des puissances, et lui-meme se garde bien de prendre parti. 81 Ailleurs sont evoquees les controverses sur la 77 78 80 81
Cf. egalement Fug. 182. Sur cette metaphore, cf. SVF 11.879, e t Rolke (1975) pp. 453-4. 79 R u n i a , (1986) p p . 3 2 4 - 5 , cf. Timee 90a. SVF 11.879. Cf., p a r exemple, Leg. 11.8. Cf. Leg. m . 115.
278
CARLOS LEVY
localisation de Fame, toujours dans le meme esprit de Vepoche.82 Mais le texte le plus important est sans conteste Mut. 10. En des termes qui ne sont pas sans rappeler ceux utilises par Ciceron dans la premiere Tusculane, Philon dit: 'Qu'y a-t-il d'etonnant que l'Etre soit imperceptible aux hommes, quand l'esprit de chacun de nous est inconnaissable? Qui a vu, en effet, l'essence de Tame? Son manque de clarte a suscite des milliers de disputes chez les Sophistes qui ont apporte des avis contraires.' L'allusion aux disputes est evidemment ici une reference implicite a la doxographie que Philon n'a pas juge bon d'exposer dans le detail. Mais le plus important est de constater que, comme le Ciceron du Lucullus et de la Premiere Tusculane, Philon considere que la nature de Fame fait partie des problemes sur lesquels l'homme ne peut aboutir a une connaissance certaine. Cela ne signifie certes pas qu'il s'en tienne a l'isosthenie de toutes les theses, mais cela reduit sa preference pour le dualisme a n'etre que le choix d'un pithanon. (b) Sur le probleme de Fame Philon arrive done, par le biais de la Nouvelle Academie, a une position qui n'est pas sans rapport avec celle de Platon lui-meme: point de certitude parfaite, mais des croyances. II y a done la une certaine rehabilitation de l'opinion, a condition qu'elle soit sous-tendue par Vepoche. Mais le plus souvent les hommes adherent a la doxa, et l'oeuvre de Philon est pleine de developpements sur l'aveuglement de ceux qui prennent leurs opinions pour un savoir veritable. Entre ces deux types d'opinion, Joseph represente un type d'ame et de doxa particulierement interessant. Pour l'etudier, nous nous bornerons ici a l'etude d'un texte tire du de Somniis.83 La traduction que nous en proposons, et qui differe sensiblement de celle de l'edition Sources Chretiennes, est celle-ci: Garde-toi bien d'isoler un element pour l'attribuer a Joseph, mais sache qu'il est l'image d'une opinion heteroclite et melee. Car en lui est visible cette manifestation de la raison qu'est la maitrise de soi, cette marque de la masculinite que lui a imprimee son pere Jacob. Se manifeste aussi Pirrationnel de la sensation, herite de sa mere et qui le fait ressembler a 82 83
Cf. Somn. 1.31-2; Leg. 1.91. Philon, Somn. 11.15: |iTi8ev ouv 6V d-TTOKpivas TTOTE Trpos *** TW 'IGOOT)s £9' a p | i a 81a T O K0O90V dvapaivEi ( G e n . 4 1 , 4 3 ) , 9uacbpEVos xai |i£TECOpOV alcOpCOV EaUTOV ETTl KaSaipEQEl taOTT|TOS.
Le concept de doxa des Stoiciens a Philon d'Alexandrie
279
Rachel. Se manifeste aussi la semence du plaisir physique que lui ont imprimee les frequentations des grands echansons, des premiers officiers de bouche et des maitres de cuisine. Se manifeste aussi le souci de la gloriole sur lequel il monte comme sur un char {Gen. 41, 43) se gonflant et se projetant dans les airs apres s'etre debarrasse de l'egalite. Le texte lui-meme presente au moins deux difficultes ponctuelles d'interpretation. II contient une lacune evaluee a dix ou onze lettres que Ton a propose de combler par prosklerou monon ou par prosklerou toutoi, hypothese qui nous parait vraisemblable.84 Par ailleurs, l'expression eikon polumigous kai kekramenes doxes est ambigue: elle a ete interpretee par les traducteurs comme si toute doxa etait chez Philon le resultat d'un ensemble de plusieurs composantes ('ce cocktail bariole qu'est l'opinion'), 85 ce qui fait de polumigous et de kekramenes des predicats. Cette interpretation doit etre selon nous ecartee, car le propre de la doxa representee par Joseph est de se situer a mi-chemin de deux doxai qui ne peuvent pas etre qualifiers de mixtes: la doxa kene, la gloriole, qui est Tune des parties integrantes de la personnalite de Joseph, et la doxa alethes, qui est l'equivalent de la science.86 Par la conjonction gar Philon exprime qu'il va enumerer les divers ingredients qui ne sont pas constitutifs de toute opinion, mais dont resulte la doxa mixte que represente Joseph. A cela s'ajoute un probleme plus global: est-il possible de faire une lecture philosophique de ce texte et que peut-il apporter a l'histoire du concept de doxa a l'epoque hellenistique? II convient a cet egard d'apporter une precision, qui permettra de mieux cerner le probleme. Le personnage de Joseph est Tune des enigmes de l'oeuvre philonienne. Ce personnage, que la tradition juive venere, fait l'objet dans le corpus philonien d'un traitement surprenant. Disons, pour resumer une question fort complexe, que Ton oppose le de Iosepho, ou Joseph est presente sous un jour favorable, aux traites constitutifs du Commentaire allegorique, dans lesquels il fait l'objet de critiques 84 85
86
Cf. l'edition Cohn & Wendland, t. 3, Berlin, 1898, 261, 1.25. C'est la traduction de Savinel dans Sources Chretiennes. Elle correspond a la traduction de Colson dans Loeb: ' H e represents Opinion with its vast medley of ingredients.' En revanche, la traduction correcte est celle de Heinemann: 'er das Abbild einer vielfacher und gemischten Lebensauffassung ist.' Le concept de KEVT) 86£CC est tres frequent chez Philon, qui y rattache tous les themes philosophiques lies a la condamnation du TO90S, cf. sur ce concept Decleva Caizzi (1980). En Fug. 19, Philon ecrit: 66£oc &Ar|8f|S £7ncnT)|Jir|S &8ia<popo0cja, ce qui est une adaptation d u Menon 98b. Nikiprowetzky (1977) p p . 212-13, remarque tres justement q u e pour Philon i ' o p i n i o n vraie c o m m a n d e tout le domaine de la theologie'. Aux exemples qu'il donne on peut ajouter Leg. 11.7, ou il est dit q u e 'chez le mechant, l'opinion vraie sur Dieu s'est obscurcie et s'est cachee'.
28O
CARLOS LEVY
severes, le de Somniis etant sans doute de tous le moins favorable a ce patriarche. V. Nikiprowetzky, a qui nous devons l'etude la plus penetrante de cette question, 87 a montre que cette opposition s'explique par la nature meme de Joseph: il est le Politique, il possede l'art royal, mais il a choisi de l'exercer dans l'Egypte, la cite profane, symbole du corps. II est quelque peu comparable au philosophe dans la caverne, avec cette difference que dans son cas 'son stage dans la province du corps procede d'un choix volontaire, d'une vocation qui l'oriente contradictoirement vers le monde exterieur et vers Dieu, qu'il finira par preferer'. 88 Le Joseph de Philon, fils de la vertu active et de la vertu de l'irrationnel - dualite que T. Mann reprendra a sa maniere 89 - est une realite complexe, qui doit beaucoup plus a la pensee grecque qu'a la tradition juive. Cet apport de la philosophic est decelable dans la construction meme du passage que nous etudions et il est etrange que personne, a notre connaissance, n'ait per9u dans les articulations de ce texte la division de Tame en huit parties, chere aux Stoiciens et a Philon lui-meme. Voici, en effet, comment se fait la correspondance: Phegemonique: to logikon enkrateias eidos les sens: to alogon aistheseos la reproduction: to tes somatikes hedones sperma le langage: to tes kenes doxes
Si cette structure a echappe aux commentateurs, c'est parce que les differentes parties ne figurent pas telles quelles, mais modulees en fonction de l'idee selon laquelle Joseph represente une ame 'melee', aux aspirations contradictoires. Examinons brievement ce qu'il en est pour chacun de ces elements: - le concept d'enkrateia est un de ceux dont le sens a le plus varie en fonction des ecoles. Contrairement a Aristote, pour qui elle n'est qu'un etat intermediate, 90 les Stoiciens la considerent comme la capacite de Fame a ne pas transgresser les regies de la droite raison et comme une vertu qui nous permet de nous abstenir des choses dont il semble difficile de s'abstenir. 91 II faut remarquer que l'exemple donne par Sextus pour illustrer cette definition est precisement celui 87 89
90
88 Nikiprowetzky (1977) p p . 218-19. Ibid. 219. Cf. T . M a n n , Joseph et sesfreres, trad. L. Vic. col. 'L'imaginaire', Paris, 1988, p . 8: 'Joseph etait repute pour son charme, herite de sa mere, qui avait ete jolie et belle comme la pleine lune et la planete Ishtar, q u a n d elle flotte doucement dans le p u r ether. De son pere, il tenait ses facultes spirituelles et le depassait sous certains rapports.' 91 Aristote, Eth. Eud. 1127b 12-16. Cf. Sext. Emp., Adv. math, ix.153.
Le concept de doxa des Stoiciens a Philon d'Alexandrie
281
de Phomme qui a la possibilite de jouir des charmes 'de Lais, de Phryne ou de quelque autre creature de ce genre', et qui y renonce. La similitude avec Pepisode de la femme de Putiphar est frappante. Philon a certainement lu ce passage de PEcriture avec en tete les exemples scolaires destines a illustrer Venkrateia, mais cela ne signifie pas que sa conception de celle-ci soit la meme que celle de PAncien Portique. En revanche, il n'est pas impossible que la place considerable que tient chez lui cette vertu soit a mettre en relation avec la place qu'elle occupe dans la psychologie de Posidonius; 92 - la sensation appartient au genre de Valogon, mais cet alogon est lui-meme une creation de Dieu et, a ce titre, il ne peut etre entierement mauvais. II existe une Forme de Yaisthesis, et cette Forme est une pensee de Dieu. 93 La sensation est done partie integrante de Pordre naturel voulu par Dieu, et sa fonction naturelle est de fournir a Pintelligence le materiau sans lequel elle ne pourrait pas travailler. 94 Lorsque la sensation aide Pintelligence a comprendre le monde, 95 elle constitue une etape dans le cheminement au terme duquel se produira Pemigration, e'est a dire la perception par Pame que rien ne lui appartient en propre. 96 Mais les sens sont attires par le monde et ils veulent exister par eux-memes, d'ou Petat de revoke et les guerres internes a Pame; - par un jeu subtil sur les signifiants, qui eut ravi J. Lacan, Philon passe de to spermatikon a sperma tes hedones. L'une des huit
parties de Pame se trouve ainsi devoyee, mais plus precisement encore, e'est le plaisir lui-meme qui est detourne de sa finalite naturelle. En effet, si les invectives contre le plaisir sont monnaie courante dans le corpus philonien, il arrive aussi a Philon d'utiliser des themes epicuriens pour souligner Putilite naturelle du plaisir, lequel a dans sa pensee une fonction precise, celle de permettre la fecondation et la procreation. 97 Chez Joseph, la relation naturelle hedone-sperma est subvertie et le jeu de mots revele comment ce qui 92 93
94 95 96 97
Galien, Hipp. Plat. dog. v.5.32-6. A. M . Ioppolo nous a signale que le concept d'eyKporraa etait deja important chez Cleanthe, cf. SVF 1.563. Philon, Leg. 1.21, a propos de l'idee archetype d e la sensation. Sur les idees comme pensees de Dieu chez Philon, cf. R u n i a (1986) p p . 164-5. L a sensation est qualifiee de neutre dans Leg. 111.67. Cf. Leg. 1.25; 11.5. Le tort de Rachel, vertu de la partie sensible, est de croire q u e tout vient d e ^intelligence h u m a i n e , alors que tout vient de Dieu, cf. Leg. 11.46 et Post. 135. Sur le theme d e Pemigration de l'intelligence hors d'elle-meme, cf. Deter. 1 ss. Gf. Opif. 161-2, ou Philon reprend, dans les termes memes d e l'epicurisme, le theme d e la recherche d u plaisir des la naissance.
282
CARLOS LEVY
ne devait etre qu'un moyen tend a se transformer chez lui en une fin; - la kene doxa est une des formes du theme cynique par excellence, celui du tuphos;98 et il faut reconnaitre que sa relation a la fonction psychologique du langage n'est pas immediatement perceptible. Rappelons cependant que, dans Mut. 91, Philon dit, en se referant a la Genese, 41—5, que le surnom egyptien de Joseph est Psontampenech, qu'il traduit par 'la bouche qui juge en reponse', et il ajoute a ce propos: 'tout insense se figure qu'il est capable de repondre aux questions de qui l'interroge'. Mais surtout, le logos prophorikos, lorsqu'il n'est pas au service de l'intelligence, est identifie par Philon au bavardage inconsistant. Ainsi, dans Abr. 29, evoquant les differentes parties de Fame, il caracterise le logosprophorikos comme 'bavardant, bouche sans frein sur des milliers de sujets qu'il fallait passer sous silence', et dans Ebr. 70, il est dit de la parole proferee qu"elle depose en nous des opinions fausses pour la ruine de notre bien le plus precieux, la verite'. En mentionnant la doxa kene plutot que le logos prophorikos, Philon nous semble rester sur le fond fidele a sa methode, qui est de mettre en evidence dans cette description des parties non rationnelles de Fame de Joseph, l'utilisation negative de la fonction plutot que la fonction elle-meme. Joseph nous parait etre l'allegorie d'un type psychologique auquel le moyen platonisme s'est beaucoup interesse, celui de Fame 'moyenne', partagee entre des aspirations contradictoires. Albinus parle d'une metaxu diathesis, et Apulee ecrit a propos de cet homme moyen: neque sinceras esse virtutes nee vitia tamen mera et intemperata, sed
hinc atque indepermixta esse." L'originalite de Philon aura ete d'exprimer ce caractere mixte a travers la structure de la doxa de Fame, et de mieux en preciser la signification philosophique. Joseph est l'etre qui, tout en ayant confinement conscience de l'evanescence du monde du Devenir, est incapable de s'en detacher. Sa tunique bigarree represente les pseude, euloga, pithana, eikota, dont la multiplicite contraste avec le caractere unique de la verite.100 II est, au mieux, l'homme du probable, mais d'un probable auquel il manque cette epoche pratiquee au contraire par Abraham, qui s'en remet totalement a Dieu. Commentant l'episode du sacrifice d'Isaac, Philon explique101 en effet que le belier qui est substitue a l'enfant represente Vepoche 'parce que la meilleure victime e'est 1'immobilite 98 100 101
Cf. n. 86. " Cf. Albinus, Epit. 30.2 et Apulee, Plat. dogm. 19.247. Cf. Somn. 1.240 et Ios. 32. OS. Fug. 136.
Le concept de doxa des Stoiciens a Philon d'Alexandrie
283
et la suspension du jugement sur les points ou les preuves font totalement defaut' et il ajoute: 'On peut declarer seulement ceci: "Dieu verra. L'univers lui est connu, il l'eclaire d'une lumiere tres eclatante, a savoir lui-meme. Le monde cree ne peut declarer rien d'autre: sur lui sont repandues d'epaisses tenebres, et dans les tenebres, il est prudent de rester tranquille."' Le concept d'epoche, la metaphore des tenebres, renvoient a la culture de la Nouvelle Academie, mais Philon, en representant du moyen platonisme, y ajoute cette lumiere dont Arcesilas avait prive la caverne neoacademicienne. Nous resumerons ainsi les conclusions auxquelles nous a conduit cette etude: — la grande nouveaute du stoicisme par rapport a Platon et a Aristote, en ce qui concerne la doxa, fut d'affirmer que rien dans le monde ne constitue un obstacle a Vepisteme du sage. Le fait que celui-ci doive suspendre son assentiment sur certains points ne l'empeche pas d'avoir une science sans faille de l'ordre des choses. A l'inverse, les sots ne doivent s'en prendre qu'a eux-memes, et ils ne peuvent trouver une justification a leur doxa en invoquant la presence d'objets qui seraient naturellement doxastiques; - contre cette conception, Arcesilas revient en un certain sens a la tradition platonicienne, puisqu'il redonne un fondement objectif a l'opinion. S'il innove effectivement par rapport a Platon, c'est parce qu'il affirme, contrairement a lui, que tout dans un univers recouvert de tenebres est objet de doxa. Pour le stoicisme, le sot opine parce que son logos n'est pas parvenu a un degre suffisant de tension, et qu'il ne pergoit pas la rationalite du monde. Pour Arcesilas et Carneade, ce que nous percevons du monde, par la raison ou par les sens, a toujours un caractere trompeur. Le monde tout entier n'est qu'une immense invitation a la doxa, le propre du sage etant de s'en tenir a l'ecart par Yepoche. C'est sans doute dans la tension tradition/ innovation a l'interieur de l'ecole platonicienne qu'il faut chercher l'explication de la double posterite, sceptique et medioplatonicienne, de la Nouvelle Academie. A partir du moment ou le scholarque Arcesilas decrivait le monde comme entoure de tenebres, la force de la tradition platonicienne poussait, si Ton nous permet cette expression, a retablir la lumiere, et c'est ce qui s'est fait progressivement. Mais simultanement la dynamique propre a la these nouvelle du 'tout est incomprehensible' conduisait a Enesideme et au
284
CARLOS LEVY
neopyrrhonisme. Parce que la recherche actuelle nous parait privilegier trop exclusivement le second aspect, nous avons essaye de montrer la possibility de la permanence dans la psychologie neoacademicienne de themes traditionnels: la mefiance platonicienne a l'egard d'un naturalisme naif, et la conception socratique de la sagesse comme conscience du non-savoir; - Panalyse de la psychologie philonienne et du personnage de Joseph nous a montre la permanence chez PAlexandrin des grands themes de la confrontation academico-stoi'cienne. Cependant, en affirmant, a la fois comme croyant et comme philosophe, que Dieu est transcendant au monde, Philon sort de cette alternative qui avait oppose la Nouvelle Academie et le Portique: la nature donne-t-elle a Petre humain toutes les conditions de la science et du bonheur, ou faut-il considerer que la seule attitude sage est de pratiquer Vepoche par rapport a un monde recouvert de tenebres? Le Joseph de Philon, allegorie de Popinion mixte, c'est-a-dire de Fame qui est tentee a la fois par le Bien et par les plaisirs terrestres, represente done un apport interessant a Phistoire de la doxa.
CHAPTER
10
Seneca on reason, rules, and moral development Phillip Mitsis
Diogenes Laertius reports that Zeno was the first to introduce the word 'kathekon'1 and to write a treatise on the topic. He also suggests2 that, in connection with these reflections on kathekonta, Zeno reversed Hesiod's well-known tag about the value of following good advice relative to that of knowing things for ourselves.3 According to Zeno, The best man of all is the one who can accept another's sound advice, Good also the man who knows all things for himself. Although few things are so characteristic of the Stoics as their I am much indebted to the discussions at Syam for helpful criticism and especially to Julia Annas, Brad Inwood, David Konstan, and Martha Nussbaum for their many concerted attempts to rescue me from error. An earlier version of this paper was read at the Ancient Philosophy Workshop in Chicago and at Duke University; I am grateful for criticism on both those occasions and for subsequent comments from Elizabeth Asmis and Paul Vander Waerdt. Joe DeFilippo, Terry Irwin, Chris Shields, and Jennifer Whiting get special thanks for extremely valuable written comments and discussion. Finally, I am indebted to the Howard Foundation and the ACLS for their very generous support. 1 Diog. Laer. vii.25; cf. vii.4. Although it is likely that Zeno included katorthomata among kathekonta, as argued by Ioppolo (1980b) p. 99, a more restricted use is possible in this particular context. 2 phasi de kai proton kathekon onomakenai kai logon peri autou pepoiekenai. tous th'Hesiodou stichous metagraphein houto. Since Diogenes sometimes lists bits of information with no apparent connection, it is not certain that Zeno's revision of Hesiod should be read against the background of his interests in kathekonta. Diogenes seems to suggest a connection, however, by linking them as items in the same report {phasi de . . . tous th')\ in a similar context, moreover, Seneca argues that one who can follow advice is better equipped for moral action than one who is doctus (see below p. 291). 3 Cf. Proclus ad Hesiodi Opera et Dies (ad loc), Themistius Orationes VIII.IO8C, xm.iyid = SVF 1.235. Cf. SVF i.2i§ for a similar literary revision of some importance.
285
286
PHILLIP MITSIS
penchant for dispensing advice,4 Zeno's remark is initially rather puzzling. Surely, we might object, Hesiod is the one who has it right: merely following another's advice, whether in the form of specific individual instructions or more general rules, should hardly take precedence over knowing things for ourselves. Zeno, however, apparently defended his own ranking by postulating a critical link between an ability to accept advice and action. A crucial prerequisite for praxis, he explained, is being capable of accepting and following the sound advice of others; by itself, knowing things on one's own holds no special guarantees with respect to one's actions.5 Yet, this added bit of explanation seems hardly less puzzling, especially given Zeno's overall endorsement of the Socratic claim that knowledge is sufficient for virtuous action. Thus the question remains, how can Zeno, or any other Stoic for that matter, think that someone who is able to follow advice is better equipped for action than one who 'knows all things for himself? To get our bearings on Zeno's claims, it might be helpful to juxtapose them against another set of philosophical remarks on these same verses of Hesiod, namely those of Aristotle at the end of Nicomachean Ethics 1.4. Although it would be rash to suppose that Zeno is specifically responding to Aristotle's discussion,6 his somewhat cryptic assertions about the requirements of praxis begin to take on more point when read against this Aristotelian background. Unlike Zeno, Aristotle endorses Hesiod's order of preference. 7 He 4
5
6
7
Diog. Laer. vii.84 = SVF m.i for 'encouragements and discouragements' as a recognized branch of ethical philosophy. Epictetus (Dissertationes in. 19-20) claims that in the same way that Socrates' speciality was refutation and Diogenes' was rebuking, Zeno had a special calling for both instruction (didaskaliken) and laying down doctrines (dogmatiken). kreittona gar einai ton akousai kalos dunamenon to legomenon kai chresthai autoi ton di'hautou to pan sunnoesantos- toi men gar einai monon to suneinai, toi a" eu peisthenti proseinai kai ten praxin. (Diog. Laer. vii.26) Since these verses of Hesiod (Opera et Dies 293 ff) achieved almost proverbial status in antiquity, it may be nothing more than a coincidence that they were invoked by Aristotle and Zeno in similar ethical contexts. Nonetheless, it is noteworthy that these lines elicited such opposing reactions. By the same token, although any similarities may be coincidental, it is intriguing to compare Zeno's comments about praxis with EN iO95a6~7 (epeide to telos estin ou gnosis alia praxis). Cf. EN 1143b! 1 ff for Aristotle's general view on the value of advice. Most commentators have taken Aristotle to be citing these verses to illustrate the primacy of those who already have a belief that something is true (i.e. what Aristotle calls to hoti, 'the that') to those who can acquire that belief. Although this is the most natural reading of the passage, one cannot rule out the possibility that Aristotle is distinguishing instead those who understand the reason (to dioti) for their belief from those who merely have the belief (to hoti) for a defense of such a reading, see Burnyeat (1980) pp. 71, 88. On the latter interpretation, however, Aristotle's endorsement of Hesiod leaves open (though it does not require) the
Seneca on reason, rules and moral development does so, moreover, just as he is beginning to signal his wider disagreements with intellectualist accounts of moral psychology and moral development. Since Aristotle's warnings about the dangers of Socrates' intellectualism bear on Zeno's own version of intellectualism as well, a brief recapitulation of those criticisms might help bring into focus not only their different reactions to Hesiod but also several larger issues lurking in the background. As is well known, the primary difficulty Aristotle finds with intellectualism is that it ignores the central role of the pathe and their habituation in our moral lives. A strictly cognitive grasp of moral principles is insufficient,8 he insists, both for making moral judgments and for our continuing moral development. It is only by the additional habituation of our affective natures to virtuous conduct that we can hope to develop properly as ethical beings and learn to make the kinds of individual moral judgments appropriate to specific circumstances. Sometimes attributed to Aristotle in this context as well is a further objection to intellectualism based on worries about the nature and efficacy of general moral rules - an objection therefore especially relevant to Stoic intellectualism. Myles Burnyeat, for example, suggests that at the heart of Aristotle's critique of intellectualism one also finds deep misgivings about the adequacy of general rules for guiding our moral lives: since 'the noble and the just do not, in Aristotle's view, admit of neat formulation in rules or traditional precepts', it is necessary for moral agents to develop an 'ability to internalize from a scattered range of particular cases a general evaluative attitude which is not reducible to rules or precepts'. 9 Linking Aristotle's rejection of intellectualism to questions about rule-following in this manner is hardly unproblematic, however, and some relevant historical and philosophical considerations should be kept in mind. For instance, one readily can agree with Burnyeat that Aristotle is concerned that moral agents develop an appropriate sensitivity to particular moral situations, a sensitivity
8
9
possibility that one can have 'the because' without already having or acquiring 'the that' - a state of affairs that Zeno deems unpromising for praxis. A Stoic might have even stronger reasons for disagreeing with and correcting these Aristotelian remarks, if EN 1.4. is taken as Burnyeat suggests. This formulation is not quite precise, since it leaves open the question of whether Aristotle would agree in principle that a purely cognitive grasp of general moral rules is possible. (See note 13.) Burnyeat (1980) p. 72.
287
288
PHILLIP MITSIS
manifested as a special evaluative attitude or mode of perception. Much less clear, though, is Aristotle's conception of the role actually played by rules and generalizations in that development. According to Burnyeat, Aristotle's emphasis on the particularity of moral situations is intimately tied to a corresponding belief that general rules are insufficiently fine-grained to capture the specific details of texture and context relevant to individual moral situations. Thus, he argues, the Nicomachean Ethics seeks to question the view that adhering to moral rules can adequately develop an agent's capacity for making particular moral judgments. 10 For historical and philosophical reasons, however, it would be a mistake merely to assume, as Burnyeat appears to, that Aristotle sees or should see any straightforward connection between intellectualism and rule-following. In the first place, such an assumption is troubling from an historical perspective since it flirts with anachronism. Although Burnyeat claims quite plausibly that the target of Aristotle's attack on intellectualism is Socrates, he nowhere shows that Socratic intellectualism relies (or that Aristotle thinks that it relies) on any particular conception of moral rules or rule-following. Such an historical link needs to be established, however. Otherwise, Burnyeat's view gives us an Aristotle deeply concerned to counter a form of intellectualism that none of his philosophical predecessors had yet dreamed of holding. Of course, one might, on Burnyeat's view, give Aristotle credit for formulating and then dismissing this novel and more complex version of intellectualism. Indeed, if the early Stoa were particularly well-informed about Aristotle's ethical writings, one might even conclude that Stoicism found the inspiration for its own theory in Aristotle's discovery of the possibility of bolstering intellectualism with a theory of general rules. From an historical point of view, however, none of these options is very attractive. 11 10
11
If Burnyeat were right about this, Zeno's endorsement and elevation of rule-following would conflict very directly with Aristotelian priorities. Among other things, this might allow for a very neat account of their different reactions to Hesiod. It is not clear that Aristotle's and Zeno's attitude towards the value of rules differs so markedly, however. Striker's argument (1987) that the Stoics were the first to formulate a conception of universal, exceptionless moral principles offers a more compelling historical framework for understanding these problems. We might reasonably expect Aristotle's remarks about moral rules to be neither very explicit (see note 13) nor obviously connected to his worries about intellectualism if his target is Socrates and if questions concerning the connections between intellectualism and rule-following come into the foreground only with the advent of Stoicism.
Seneca on reason, rules and moral development Second, we need to be careful about what we take intellectualism to entail for Aristotle. One obviously can reject intellectualism and at the same time endorse a suitably non-particularist account of moral judgment based on a belief in the general applicability of moral rules. Or, by the same token, one can deny the applicability of general moral rules, while maintaining that the relevant perception of moral particulars is strictly cognitive. By itself, Aristotle's rejection of intellectualism carries with it no special presumptions in favour of the sort of particularism ascribed to him by Burnyeat: in rejecting intellectualism one is forced neither to espouse a particularist account of moral judgment nor to shun rules as being strictly in the domain of reason. 12 Accordingly, one needs an independent textual warrant over and above his denial of intellectualism for claiming that Aristotle either affirms or denies the relevance of general moral principles to particular decisions. Yet, extracting any clear-cut or explicit message about rules from Aristotle's text is no easy task13 - a feature of his argument that perhaps gives us further grounds for doubting that his rejection of intellectualism is connected to any particular suspicions about moral rules. Where Aristotle clearly differs from both Socrates and the Stoics, 12 13
I am indebted to T. Irwin and C. Shields for clarification on several of these issues. Burnyeat (1980) takes the discussion of exceptions to laws (EN v.10) and the difficulty of formulating generalizations (EN ix.2) as evidence for Aristotle's denial of the existence of general moral principles (p. 72) and, presumably, as important elements in Aristotle's attack on intellectualism. Yet, although Aristotle obviously would disagree with intellectualists such as the Stoics about what it takes to grasp and follow rules, it is far from clear that he would deny the existence and value of general moral principles. Aristotle's fullest discussion of general rules occurs at EN 1137b 13-32. He claims that for some things (cases?) (peri enion, 1137b 14, 1137b28), we cannot lay down a general rule and we must therefore correct our judgments by deciding what the lawgiver himself would have decided if present. Aristotle does not here question the existence, validity, or usefulness of all general laws; indeed, he argues that even if an individual law fails to apply to a specific case, the law itself is still correct (orthos), remaining sensitive to the error involved (EN 1137b 16 ff). Thus, he in no way suggests that the practice of adhering to legal rules dulls our sensitivity to the demands of particular cases. Nor does he suggest that every new case presents an occasion for improvising from scratch. (For a representative range of opposing views of this passage, see Nussbaum (1986a) p. 301 and Sherman (1989) pp. 13-22.) At EN ix.2, Aristotle himself seeks to broaden generalizations to cover some casuistic exceptions and concludes with a suitably wider generalization (EN 1165a 14 ff). By the same token, Aristotle nowhere explicitly claims that there always may be potential exceptions to general moral rules such as 'Do what is just and fine' or 'Be virtuous' (see Irwin (1988) p. 600), and it is expressions of this type that the Stoics were to elevate to the status of moral principles. Nor for that matter is it even clear that Aristotle believes that there must be exceptions to some kinds of lower level rules such as 'Don't run away from battle to save your life', see Putnam (1983) p. 195.
289
29O
PHILLIP MITSIS
therefore, is in his denial that moral development is solely the province of reason. Detailed exploration of the status of moral rules and their connection to intellectualism comes into the foreground of philosophical discussion only with the beginnings of Stoicism. Consequently, while it might be philosophically instructive to speculate about how Aristotle's views of moral judgment and development fare against rule-based intellectualist accounts, 14 we have little reason to be confident that Aristotle himself ever thought he was confronting such theories. The Stoics, on the other hand, are convinced that moral development depends solely on a deepening cognitive grasp of both universal and more determinate moral principles; they hold, moreover, that moral rules can structure our understanding of a particular moral situation in ways that guarantee our sensitivity to its specific demands. For the Stoic, developing the requisite sensitivity to moral particulars is strictly a matter of cognitively grasping the applications of such rules.15 By the same token, what enables agents to make the critical transition from general moral principles to moral particulars is the ability that Zeno thinks both crucial for praxis and even more valuable than knowing things for ourselves - an ability to follow advice.16 Thus, in a certain sense, our opening anecdote about Zeno encapsulates the Stoics' programme for refurbishing Socratic intellectualism, an intellectualism that now takes morality to depend crucially on knowing how to grasp and apply moral rules.17 In a passage that both corroborates and further explains 14
15
16
17
In sections n and in of this paper, I try to confront Aristotelian positions with Stoic arguments to see how much of what Aristotle values in moral development and judgment can be captured by a cognitive and rule-based theory. I do not suppose, however, that such a procedure represents any actual historical debate. For an opposing view of the Stoics, see Inwood (1986) who argues that the sage's moral perception in particular circumstances may depend on an ability to break or transcend universal rules and to follow nature and the craft of life in 'a higher and more flexible sense'. I doubt, however, that a sage is ever called upon to break the constitutive rules of morality embodied in dogmata or decreta; see Mitsis (1986) and Striker (1987) for the claim that, for the Stoics, all moral choices that reflect Right Reason are structured by such constitutive, and therefore exceptionless, rules. In this context it is worth remembering that, for the Stoics, reason alone is the source of all motivation, desire, actions, etc.; reason is not, therefore, as on a Humean conception, inert and purely instrumental. For the Stoic, then, cognitively grasping a particular rule carries with it the implication that one will be able to act on it; action requires no supporting non-rational source of motivation, since knowledge is sufficient for action. I leave open the question of whether the Stoics might have found anything in Socratic thought corresponding to exceptionless moral rules. Even if they did, however, an independent argument would be needed to show that Aristotle viewed Socratic intellectualism in the same light. See note 11.
Seneca on reason, rules and moral development
291
Zeno's claims, Seneca argues that one may have a grasp of general moral principles (decreta) and yet still be unable to discover what a particular moral situation demands. It is the practice of receiving advice, admonitio, in the form of more determinate moral precepts (praecepta), that teaches agents to recognize what individual cases require and therefore enables them to act: 'Si quis' inquit 'recta habet et honesta decreta, hie ex supervacuo monetur.5 Minime; nam hie quoque doctus quidem est facere quae debet, sed haec non satis perspicit. Non enim tantum adfectibus inpedimur, quominus probanda faciamus, sed inperitia inveniendi quid quaeque res exigat. Habemus interdum compositum animum, sed residem et inexercitatum ad inveniendam officiorum viam, quam admonitio demonstrat.18 (Ad Lucilium Epistulae Morales xciv.32 (Reynolds))
For the Stoics, therefore, moral judgment and development are structured at every level by rules; and these rules are grasped by reason alone. As is well known, rule-based conceptions of morality have become the target of a wide variety of criticism. 19 It is argued, for instance, that such theories employ moral rules merely as mechanical 'gridlines' for administering systematically uniform and impersonal judgments. 20 Consequently, the judgments they deliver are inevitably prone to narrowness, inflexibility, and remoteness from individual and collective moral experience. A further charge is that these theories achieve ethical consistency only by defusing the force of particular, and potentially conflicting, moral obligations. What drops out of such systems of moral rules, it is claimed, is precisely an Aristotelian concern for the individuality of moral experience. At the same time, such criticisms are often coupled with a general attack on the very notion of rules and rule following.21 If we are worried that rules merely pick out repeatable features of moral situations, thereby failing to capture what is unique about them, sceptics about rules offer even more unsettling prospects. They argue that the realist conception of rules underlying moral systems like that of the Stoics is incoherent (as are its attendant notions of meaning, mental representation, intention, etc.). By denying that 18 19 20 21
Seneca repeats this claim at Ep. Mor. xcv.12. See Nussbaum (1986a) p . 31ff.for discussion a n d p. 426ff.for references to a wide range of contemporary literature on this issue. H e r m a n (1985) attempts to meet such charges within a K a n t i a n perspective. I a m much indebted throughout this paper to her discussion. See McDowell (1979) with the response of Blackburn (1981) p p . 163ff;Dancy (1983).
292
PHILLIP MITSIS
rules are objective patterns that extend, without limit, 'independently of where we ourselves take them', 22 they attempt to undermine the claim that there are any objective, repeatable features of moral situations that can guide us in making moral judgments or forming moral intentions. By the same token, such sceptics reject the view that our actions depend on an ability to understand and make determinate projections based on fixed, external patterns. Thus, they challenge as well the coherence of accounts of the stability of semantic meanings or of the attribution of determinate mental content that rely on a realist conception of rules. 23 In what follows, I primarily look at Stoic resources for meeting criticisms of the first type, i.e. that moral rules fail to capture the particularity of moral experience. Several common objections to rule-governed theories, I argue, fail to touch their account. In fact, a surprising feature of the Stoics' conception of moral rules is how well it accommodates insights that are often taken to be alien to rulebased theories. Nonetheless, their account remains in jeopardy until it can be shown to meet more general sceptical attacks on rules 24 and much the same holds for their corresponding accounts of action and mental representation. Discussions of the Stoics' theory of action and philosophy of mind have for the most part neglected such worries, concentrating almost exclusively on sorting out the logical and causal relations of individual episodes of presentation, impulse, and assent. It also is necessary, however, to understand the wider connections between the Stoics' psychology and their conception of moral principles. We need to ask, for instance, how individual acts of assent can be guided and linked by moral rules, or how individual actions and mental states can be said to give expression to objective moral principles. Until such issues are sorted out, our picture of the Stoics' theory of action, mind, and morality will remain at best incomplete, if not distorted. Thus, as a preliminary step towards addressing some of these larger problems, it may prove helpful to 22 23
24
McDowell (1979). This rides roughshod over several difficult issues since it fails to factor in internalist defences of rules based on linguistic practices a n d so-called communal 'forms of life'; nor does it take into account the extent to which non-externalist theories can claim to b e realist or objective, etc. I take it, however, that for the internalist, the Stoic view that moral norms exist independently of us must come to grief as must their belief that grasping a rule guarantees future behaviour (as opposed to the internalist claim that rules go where we take them, i.e. it is because we behave in certain ways we can be said to have grasped a rule). For a particularly clear statement of the relevant distinctions, see Lear (1983). I n section iv, I briefly discuss one such Stoic argument combatting scepticism about rules.
Seneca on reason, rules and moral development
293
come to a more precise understanding of the Stoics' conception of moral rules, and the ways that these rules enter into our moral life and guide our moral judgment. 11
Two connected letters of Seneca, Epistulae Morales xciv and xcv, provide the most detailed surviving account of the Stoics' conception of moral rules. They defend the importance of both specific rules {praecepta) and universal moral principles (decreta) for our moral lives and explain their joint role in human action. Since Seneca's didactic and literary aims make little sense from the vantage point of a theory that, like Aristo's, denies any value to more concrete and individualized moral recommendations {praecepta), it is clear that he has a very personal stake in these arguments. 25 It would be a mistake, however, to dismiss his account as being merely idiosyncratic. Orthodox Stoics regularly defend the practice of giving instruction by precepts, viewing it as part of a general natural impulse to benefit others.26 Moreover, Seneca structures his discussion around what has all the hallmarks of an authentic controversy emanating from the early Stoa.27 I will therefore proceed on the assumption that his account, at least in its major outlines, is representative of an ongoing consensus among orthodox Stoics about the nature of rules and their place in morality.28 25
26
27
28
See Cancik (1967) p p . 42 fffor discussion of the general role that decreta and praecepta play in structuring the form a n d content of Seneca's Epistulae Morales. Cf. note 4 for Zeno's concern with both doctrine a n d instruction. de Finibus 111.65-6: impellimur autem natura ut prodesse velimus quam plurimis in primisque docendo rationibusque prudentiae tradendis. As Cicero notices, the Stoics claim that this impulse can be seen working at a more elementary level in the care that animals take rearing their young (111,62). See Bellincioni (1978) pp. 73 fffor a detailed discussion of Stoic didactic methods and aims in Ep. Mor. xciv and xcv. Seneca's use and explanation of technical Stoic vocabulary (both Latin a n d Greek) a n d his attempt to clarify technical distinctions throughout these two letters suggest, though of course d o not prove, that he is taking care to recount the terms of an earlier debate. This claim is controversial a n d has been doubted on various counts by Watson (1971), Inwood (1986) (on which see note 15 above), and in more detail by V a n d e r W a e r d t (1989) w h o argues that, whereas the early Stoa viewed moral conduct a n d natural law strictly in terms of katorthomata (which he claims cannot be subsumed u n d e r general rules or principles), post-Antiochean accounts (especially through Cicero's influence in de Legibus) hold t h a t the content of natural laws c a n be summarized completely in kathekonta without appealing to katorthomata. I find these claims unconvincing a n d agree with Kidd's arguments (1971) against Pohlenz (1970), Reesor (1951), a n d v a n Straaten (1946) that Stoicism never wavers on the fundamental distinction between katorthomata a n d kathekonta, though particular Stoics might concentrate on one or the other for purposes of exposition
294
PHILLIP MITSIS
I begin with the discussion of praecepta in Ep. Mor. xciv. Seneca's account of the dispute between orthodox Stoics and Aristo over the usefulness of praecepta has generally been taken to turn on the question of whether praecepta can furnish non-trivial, exceptionless rules for action.29 On this view, both sides in the dispute agree that rules function as a kind of external evaluative grid that we map on to our experience to pick out relevant factual instances. Difficulties occur when there are problems of fit; rules may have exceptions or an individual case may fall under too many rules.30 Depending on our purposes, of course, we might be able to tolerate some indeterminacy in the rules we use. However, both Aristo and orthodox Stoics are standardly taken to agree that, to be of any use, rules embodied in praecepta must not only be unequivocally informative in specific situations but also entirely immune to exception.31 Since orthodox Stoics commit themselves to defending the usefulness of praecepta^ they become faced, on this view, with the dubious task of satisfying these two prima facie conflicting demands. When the dispute is framed in this way, however, it can hardly come as much of a surprise that Seneca's examples of praecepta get low marks both for their immunity to exception32 and for their ability to give any very
29 30
31 32
(pp. 150-72). Seneca in no way diverges from the early Stoa's emphasis on katorthomata (see Ep. Mor. xcv.43; also xcv.57); nor does he ever claim that morality can be summarized completely in praecepta (or officia). Moreover, he gives a plausible account of how katorthomata reflect universal constitutive moral principles (for a discussion of this problem in early Stoicism see Striker, (1987)). Thus, I find nothing in Seneca's overall treatment of moral rules that conflicts with the general plan a n d intention of early Stoic thinking on natural law (nor, for that matter, do I see any striking innovations in Seneca's discussion, pace Watson a n d V a n d e r Waerdt; cf. Leg. 1.18-35, e s P- J9> 3 3 5 I L 8-13; Resp. in.33). This latter evidence casts doubt as well on Dihle's claim that Seneca's account depends crucially on Posidonian innovations (1973) p . 50 ff. A view defended by Rist (1969) p p . 63 ff; Kidd (1978) p p . 252-4; a n d White (1985). T o anticipate a bit, the problem of too many (or conflicting) rules does not explicitly feature in this dispute. Since this would be such a n obvious move for someone wishing to deny the usefulness of rules it seems to me that this m a y furnish further evidence against viewing the question in quite these terms. White (1985). W h i t e (1985), citing Ep. Mor. x c i v . 3 5 , claims that Seneca 'just flatly denies' t h a t there are no exceptionless praecepta a n d thinks t h a t 'we can include e n o u g h qualification in o u r rules to make t h e m i m m u n e to exceptions' p . 301. (Cf. de Officiis 1.33, for Cicero's general suspicions a b o u t such procedures.) T h i s offers t h e Stoics t h e very u n h a p p y prospect of m a i n t a i n i n g a c o m m i t m e n t to exceptionless rules only by continually qualifying t h e m to fit exceptions. A n d it is h a r d to see w h a t the point of calling such rules exceptionless is supposed to be o r h o w they can h o p e to continue qualifying a rule without m a k i n g it look trivial. Moreover, even if such saws as 'Greed is never satisfied' or 'Expect from others w h a t you d o to t h e m ' c a n escape the charge of triviality, it does not take too m u c h ingenuity to think u p exceptions.
Seneca on reason, rules and moral development
295
precise moral guidance. Indeed, it is difficult not to view the entire dispute as being somehow misguided. Characterizing Stoic conceptions oipraecepta in this way, however, misconceives both the terms and goals of their overall argument. Part of the problem is that it can be fairly tempting to assimilate the Stoics' account to a rigorist model of rules33 in which exceptions take on paramount importance. If we view praecepta as rules specifying and enjoining obligations to perform certain actions, it becomes necessary, as in the Kantian tradition, to handle exceptions by arranging rules into hierarchies and articulating further principles of priority. 34 Although it is fairly clear that Seneca is committed to classifying praecepta?5 and showing how they are connected to basic features of human agency and rationality, 36 he nowhere implies that they are ranked in hierarchies or ordered along some fixed grid of priorities. Moreover, he appears to be relatively untroubled about problems of exceptions37 or conflict. From the perspective of a rigorist all this would be very puzzling; so it is perhaps worth 33
34 35 36
37
By 'rigorism' I m e a n the view, often attributed to K a n t , that morality is structured by rules that admit of no exceptions. See D o n a g a n (1977) for a recent defence of rigorism on the basis that all possible exceptions are implicitly implied in a rule when it is properly understood. I see no evidence for such a move a m o n g the Stoics, however. O r , as in Rawls, precepts might be lexically ordered with the rider that no precept c a n be followed if it violates a precept higher in the series (1971) p . 4 3 . H e does so at Ep. Mor. xcv.47 ff, classifying praecepta with respect to gods (47-50), m e n (51-3), a n d things (54). I t is easy to get the impression from recent discussions that praecepta are mostly limited to expressions such as 'Nothing in excess' or 'Greed is never satisfied'; this is very misleading, however. Seneca argues, for instance, that there a r e basic precepts concerning attitudes towards others which revolve a r o u n d certain natural conditions of h u m a n agency, i.e. that agents have a common share in reason, are particles of god, share natural kinship, etc. (Ep. Mor. xcv.51 ff; cf. Off 1.7). At Ep. Mor. x c i v . 3 5 - 6 Seneca attempts to meet Aristo's objection that praecepta are infinita, thus useless in practice: 'Inftnita' inquit 'praecepta sunt.} Falsum est; nam de maximis ac necessariis rebus non sunt infinita; tenues autem differentias habent quas exigunt tempora, loca, personae, sed his quoque dantur praecepta generalia. This might suggest a rigorist model of rules that handles exceptions either by (a) subsuming them under various sub-rules until a complete and detailed code of rules is explicitly laid out, or by (b) developing a practical science for qualifying and interpreting rules. The latter alternative suggests a casuistic model in which agents are trained to make moral judgments by arguing about hard cases. Again, Seneca's emphasis is different from either of these alternatives. Praecepta can help an agent pick out morally relevant features of a situation; but in doing so they do not need to encompass a whole battery of possible exceptions or qualifications since they are not rules for making judgments. Judgment proceeds from decreta and it is when the question ofjustification arises that exceptions and qualifications need to be weighed. (On a similar issue, cf. Herman (1985) p. 420.) Seneca's point is that praecepta can lead agents to see that a particular situation needs moral attention and judgment, and they do so in particular situations without being endlessly qualified or becoming 'infinita'.
296
PHILLIP MITSIS
questioning whether such a model is adequate for understanding the Stoic conception of praecepta. We can situate problems about praecepta somewhat differently, I believe, and perhaps account for what might look to a rigorist like odd omissions by looking more closely at Seneca's explanation of the manner in which rules enter moral experience and how they function in moral development. His most basic disagreement with Aristo is over the way that individuals come to acquire an awareness of the requirements of particular moral situations. He maintains, against Aristo, that even someone acquainted with the fundamental principles of morality {deereta) needs the help of praecepta in order to learn to recognize what particular moral situations demand. 38 Without developing this kind of moral awareness, agents will not be able to act in ways appropriate to particular circumstances.39 Praecepta, he claims, teach us to recognize morally salient features of individual situations; moreover, by enabling us to recognize that particular situations have moral characteristics, they give us a starting point for moral development.40 To be sure, praecepta cannot provide us with the reasons or justifications for our actions. But they lead us to recognize what should be done and how it should be done. 41 In 38
Seneca's claim that praecepta are necessary for discovering 'quid quaeque res exigaf (Ep. Mor. xciv.32) is ambiguous between types a n d tokens, b u t the contrast d r a w n a t Ep. Mor. x c v . 12 more strongly suggests action tokens. A n d in a n y case, although the content of a praeceptum is not a particular, adoptions a n d assertions of a praeceptum are. Cf. Nell w h o suggests comparing how the fact that a sentence, e.g. ' T h e world is round', is not a particular in no w a y prevents m y belief that the world is round from being one. (1975)
39
Ep. Mor. x c i v . 3 2 - 3 ; cf. x c v . 12 where this claim is repeated about someone who has ratio. I a m using 'starting points' for the sake of contrast with Aristotle b u t also to capture what Seneca calls 'prima fundamental {Ep. Mor. xcv.35), i.e. the initial cognitive starting points or foundations of virtue. T h e Stoics talk about aphormai (cf. Diog. Laer. vii.89; SVF in.264 = £c/.ii.6o.g; SFF1.566 = Ed. 11.65.7 W) m t n e sense of natural inclinations to virtue, and these might in some sense be characterized as starting points. Presumably, aphormai are built-in propensities to encounter experience in certain ways; however, without the right cognitive input a n d instruction, they are insufficient to provide a secure starting point for the acquisition of virtue. For a general discussion of Stoic claims about innateness, see Scott (.988). Cf. Ep. Mor. xciv.23 • • • auid ei quemadmodum debeamus facere, discendum est (i.e. by means of praecepta). Cf. Burnyeat (1980) pp. 71 ff for a discussion of'starting points' at EN iO95b213 and Aristotle's view of the role that non-cognitive elements have in their acquisition. Aristotle claims that someone who has acquired the requisite starting points already loves virtue (cf. Burnyeat, p. 75 on EN 1179b4~31). Seneca makes a similar claim at xcv.35: . . . ita in his quos velis ad beatam vitam perducere prima fundamenta iacienda sunt et insinuanda virtus. Huius quadam superstitione teneantur, hanc ament; cum hac vivere velint, sine hac nolint. I hope it becomes clear in the ensuing discussion in what sense this process can be viewed as being strictly cognitive. White argues (1985, p. 303) that praecepta pick out purely descriptive, factual conditions for evaluative claims. But this is too schematic and relies on the notion
40
41
P- 1 3 - •
Seneca on reason, rules and moral development
297
addition, Seneca maintains, unlike Aristotle, that this kind of moral sensitivity can be acquired and developed strictly through cognitive means. Seneca provides a richly detailed catalogue of the cognitive methods by which praecepta structure and develop our moral awareness, but for present purposes it will perhaps be sufficient to touch on only a few of the most significant. He argues that praecepta are necessary because our beliefs have become corrupted and we have consequently become blind to the moral aspects of our actions. In an account reminiscent of Plato's in the Republic, he traces the sources of corruption to the bad advice and example of those who glorify power, wealth, and pleasure.42 What is needed, he claims, are therapeutic praecepta that can counteract such influences and restructure our beliefs by removing the obstacles to our moral vision.43 A significant attraction of praecepta, Seneca claims, is that they are not merely coercive;44 unlike laws, they sometimes can persuade solely by reminding us or by calling our attention to things that we have forgotten.45 At the same time, they can help us bring coherence to the moral views that we already have and also make
42
43
44
45
that rules act as a n externally imposed, normative grid that can be mapped on to o u r experience in order to discover the relevant factual instances. T h e picture that emerges from Seneca is quite different. Praecepta alert us to the moral characteristics of a situation not by showing us that it answers to an appropriate factual description, b u t by calling our attention to its morally salient features. For corresponding reasons, I think it is misleading to claim, as White does (pp. 304-5), that the Stoics are attempting to respond to openquestion arguments a n d thus to difficulties about moving from factual premisses to evaluative conclusions. Nulla ad aures nostras vox impune perfertur; nocent qui optant, nocent qui execrantur. Nam, et horum inprecatio falsos nobis metus inserit et illorum amor male docet bene optando; . . . Non licet, inquam, ire recta via. t r a h u n t in p r a v u m parentes, t r a h u n t servi. (Ep. Mor. x d v . 5 3 - 4 ) Cf. Diog. Laer. v n . 89 for Ghrysippus' claim that rational beings are led astray by the instruction or communication of companions (ten katechesin ton sunonton). Cf. SVF 111.519; also Sedley, p p . 313-31 in this volume. Multis contra praeceptis eget animus ut videat quid agendum sit in vita. (Ep. Mor. xciv.19) T h r o u g h o u t these two letters Seneca sustains a n elaborate medical analogy between saving a n d restoring one's sight a n d saving a n d restoring one's ability to recognize moral requirements. This is not to say that Seneca thinks, as some particularists have argued, that moral j u d g m e n t is actually a kind of'vision'; but he shows a similar concern with capturing the individuality of moral experience. 'Leges' inquit ' u t faciamus quod oportet n o n efficiunt, et quid aliud sunt q u a m minis mixta praecepta?' P r i m u m o m n i u m o b hoc illae n o n persuadent quia minantur, at haec non cogunt sed exorant; deinde leges a scelere deterrent, praecepta in officium a d h o r t a n tur. (Ep. Mor. xciv.37) I t a q u e subinde ad m e m o r i a m reducendus es; n o n enim reposita ilia esse oportet sed in
promptu. Quaecumque salutaria sunt saepe agitari debent, saepe versari, ut non tantum nota sint nobis sed etiam parata. Adice nunc quod aperta quoque apertiora fieri solent. (Ep. Mor. xciv.26)
298
PHILLIP MITSIS
them more explicit.46 Similarly, they enable us to look at things in new ways, correct corrupt beliefs,47 and banish disturbing passions.48 In short, through these and similar cognitive techniques, praecepta help us recognize that particular situations have morally salient characteristics. They do so by altering the structure of our beliefs and developing a new framework for our moral awareness.49 We still need to ask, of course, how plausible it is for the Stoics to maintain that we can develop this sort of moral awareness through cognitive methods alone. But at this point, we can at least defend their view from some common complaints against intellectualist and rule-governed theories of morality.50 As Seneca makes clear, rules do not enter into our moral experience as sources of abstract and inflexible authoritarian guidance (Ep. Mor. xciv.37). Praecepta call our attention to the particular, changing circumstances of our moral experience; therefore, moral recognition is built up both from within particular situations and in response to their specific features. The moral awareness that we acquire through Stoic praecepta is not, therefore, like an abstract or inflexible moral viewpoint imposed from without. Nor is it in any way remote from individual and collective experience. Stoic praecepta offer a concrete and flexible strategy for acquiring the starting points necessary for further moral development. They do so, however, without offering a completely 46
47 48
49
50
Praeterea q u a e d a m sunt quidem in animo, sed p a r u m prompta, q u a e incipiunt in expedito esse c u m dicta sunt; q u a e d a m diversis locis iacent sparsa, quae contrahere inexercita mens non potest. Itaque in u n u m conferenda sunt et iungenda, ut plus valeant a n i m u m q u e magis adlevent. (Ep. Mor. xciv.29) Cf. Inwood (1985) p p . 83 ff. o n the notion of agents making explicit to themselves propositions to which they have assented. Ingenii vis praeceptis alitur et crescit novasque persuasiones adicit innatis et depravata corrigit. (Ep. Mor. xciv.30) Q u o d si est, non t a n t u m scita sapientiae prosunt sed etiam praecepta, quae adfectus nostros velut edicto coercent et ablegant. (Ep. Mor. xciv.47) In other words, praecepta help us to restrain a n d remove emotions, though they themselves are not coercive (cf. note 44 above). I t is not that praecepta help us to acquire patterns of relevant factual information in order to guide us when we are faced with a particular moral situation. T h e y help to create an overall framework for moral j u d g m e n t a n d action. See H e r m a n (1985) p . 419 w h o provides an account of the ways that a 'practical framework' for action is created on a K a n t i a n model of rules. I a m much indebted to her general account of the mechanisms of 'moral salience'. It should be clear that their theory escapes from the following general defect in intellectualist theories described by Burnyeat (1980) p . 70: 'Intellectualism, a one-sided preoccupation with reason a n d reasoning, is a perennial failing in moral philosophy. T h e very subject of moral philosophy is sometimes denned or delimited as the study of moral reasoning, thereby excluding the greater part of what is important in the initial - and, I think, continuing - moral development of a person.'
Seneca on reason, rules and moral development
299
codified set of moral priorities or by requiring practice in the purely theoretical complexities of casuistry.51 We might reasonably wonder, of course, about the efficacy and reliability of the cognitive techniques described by Seneca and why he is convinced that sound moral perception can develop in the face of so many sources of corruption.52 It might be objected, for instance, that whatever his account gains in flexibility and attention to moral detail it threatens to lose in structural clarity and objectivity. That is, if agents at this stage are not guided by some external grid of priorities, what is there to guarantee that their moral perception will not develop in ways that are completely idiosyncratic? For instance, at Ep. Mor. xcrv.26, Seneca describes the case of someone who adheres to a sexual double-standard. 53 Praecepta presumably can bring to one's attention the asymmetry of treatment that such conduct involves together with its moral dissonances. Yet, even if we concede to the Stoics their denial of akrasia, incorrigible desires, and so forth, what prevents individuals from revising their beliefs about such cases in the wrong way and, hence, developing defective moral attitudes? That is, we need an account not only of the formation of our attitudes, but also some way of further justifying the acquisition and the content of those attitudes. Seneca is aware of this potential problem with praecepta and through a series of images and arguments attempts to explain why decreta are necessary for a solution. But what are decreta and how do they function?54 It sometimes is claimed that decreta are merely general and more inclusive kinds of praecepta. And in fact, Seneca 51
52
53 54
This is not to deny that a kind of casuistry ever m a d e a n a p p e a r a n c e in the Stoa (cf. Cicero's report about the disputes between Diogenes of Babylon a n d Antipater {Off. in. 51 ff; 91) a n d Hecato's q u a n d a r y about whether in a storm one should toss overboard a n expensive horse or a cheap slave {Off 111.89)). But Cicero does not take a n y of these to be h a r d moral cases, since they all involve a conflict between utilitas a n d moral good a n d are thus easily resolvable in favour of the latter. Nulli nos vitio natura conciliat: ilia integros ac liberos genuit. {Ep. Mor. xciv.56) Part of the answer is that there is a natural presumption in favour of developing the right kind of moral awareness (cf. D L vii.53 a n d Long & Sedley (1987) 1 p . 375 for discussion). Scis i n p r o b u m esse q u i a b uxore pudicitiam exigit, ipse alienarum corruptor uxorum; scis u t illi nil c u m adultero, sic tibi nil esse debere cum paelice, et n o n facis. Seneca uses 'decreta', iscita\ and iplacita> to translate 'dogmata' {Ep. Mor. xcv.io), on which see Barnes (1982) and Caizzi (1986) p. 154. Presumably, dogmata correspond to (or more precisely, are justified by) what Chrysippus calls koina theoremata {SVF 111.295; 2I4> 2 7^)See Ep. Mor. xcv.63-4 for general connections between decreta and ratio and xcv.io for connections between decreta and the governing principles (presumably connected to theoremata) in contemplative and practical crafts. On theoremata in general, see Schofield (1984) p. 91. Cf. Galen /W> v.324.
3OO
P H I L L I P MITSIS
himself provides an explicit textual basis for such a view.55 However, he also characterizes decreta in enough other ways to suggest that this is at best only part of the story. At Ep. Mor. xcv.59, for instance, he offers an illustration of the organic connections he sees between decreta and praecepta.56 He emphasizes their mutual dependence by likening decreta to branches and praecepta to leaves of a tree (an image he completes at xcv.64, by comparing reason to its sustaining roots). This suggests an overall relationship that fits rather awkwardly with the claim that decreta are just generic praecepta. At Ep. Mor. xcv.64, moreover, he likens praecepta to hands, and decreta to the hidden, directing powers of the mind. If we take this analogy at all seriously, it suggests that decreta somehow embody active, dynamic principles of reason and judgment. 57 Again, however, it is hard to see how this claim plausibly can be reduced to a question of their respective generality. Elsewhere, as well, Seneca makes several other claims on behalf of decreta: they provide reasons and justification for moral actions {Ep. Mor. xcv.5); 58 they lend both stability and coherence to our lives (xcv.44—6); they enable one to attain a fixed and unchanging understanding of life and nature as a whole.59 All of these claims suggest differences between decreta and praecepta that are not solely a matter of degrees of inclusiveness. We still need to inquire further into the actual mechanisms of decreta, though, in order to see how 55
56
57 58
59
Quid enim interest inter decreta philosophiae et praecepta nisi quod ilia generalia praecepta sunt, haec specialia? U t r a q u e res praecipit, sed altera in totum, particulatim altera. (Ep. Mor. xciv.32) O n the basis of this passage, White (1985) argues that ' T h e Stoics think that the contrast [between praecepta a n d decreta] is merely between general a n d more specific injunctions, that is, that decreta are really just praecepta of a general sort . . . ' p . 303. I agree with Kidd, however, that 'this is a puzzling line for them to take . . . ' (1978) p . 254 and try to offer some ways of reading this distinction based o n evidence in Ep. Mor. x c v as well, which, after all is specifically devoted to discussing decreta. Note, however, pace White, that this passage does not in a n y case commit the Stoics to a n ascending grid of ranked priorities; it only draws a general contrast between the universality of decreta a n d the particularity of praecepta. Cf. Ep. Mor. x c v . 12 for the comparison of decreta to membra and praecepta to elementa which suggests n o t so m u c h a contrast between general a n d particular b u t a similar kind of organic (sine radice, 12.22) and causal connection (ilia et horum causae sunt et omnium, 12.27). T o take just o n e example of a claim that h e repeats in several places: Iustum autem honestumque decretorum nostrorum continet ratio (Ep. Mor. xcv.63). Cf. Leg. 1.18. I n this respect they correspond to katorthomata. T h e relationship between decreta and praecepta differs from that of katorthomata a n d kathekonta, however, since not every decretum is also a praeceptum. At Ep. Mor. xciv.33-4, Seneca discusses the relation between praecepta and officia: officia praeceptis disponuntur . . . U t r a q u e enim inter se consentiunt: nee ilia possunt praecedere ut non haec sequantur, et haec ordinem sequuntur suum; unde apparet ilia praecedere. Decreta sunt quae muniant, quae securitatem nostram tranquillitatemque tueantur, quae totam vitam totamque rerum n a t u r a m simul contineant. (Ep. Mor. xcv. 12)
Seneca on reason, rules and moral development
301
they enter into our moral development and shape our moral judgment. Unfortunately, Seneca does not provide a list of such decreta, but if what we expect are very high-level evaluative principles, a decretum will presumably be something on the order of'One always should be virtuous' or, perhaps, 'One who is virtuous always acts in accordance with nature'. 60 If we think of decreta only in terms of their generality and their immunity to exception, we may, as in the case of praecepta, be somewhat disappointed. White, for instance, thinks that the Stoics are faced with the dubious prospect of mapping these kinds of very general evaluative principles onto particular factual instances. In so far as he finds such general principles uninformative and empty, he questions whether such a procedure would be of any use to agents faced with making particular moral judgments. 61 We need to be careful, however, about what kinds of tests we are asking these principles to pass. First of all, it is not clear on what grounds such general principles are held to be uninformative. We might reasonably claim, for instance, that 'One always should be virtuous' informs us exactly what to do in a particular moral situation: it informs us that we should be virtuous.62 The problem seems not to be so much one of information, but of our needing further help in the form of more determinate principles to recognize particular actions that are in fact virtuous. But this need for further determination hardly demonstrates that general principles are uninformative, unless we assume that a principle provides information only if we apply it directly to an individual situation without the help of more determinate principles. It is not at all clear, however, what makes this latter requirement a reasonable test for principles. In the same way, general principles illuminate the connections between our actions and set significant restraints on them. Take the following example. Suppose that a friend of mine is depressed and that I advise him that hard work is the best cure for depression. This 60 61
62
I t is not clear whether decreta need to be in the form of imperatives. See Ep. Mor. xcv.63. Cf. D L VII. 109 = SVF in.496 for corresponding examples of aei kathekonta. White (1985). I see no evidence that the Stoic account of decreta is responding to problems about moving from factual to evaluative claims (cf. note 41). Inwood (1986) a n d V a n d e r W a e r d t (1989) make use of a similar claim about the triviality of general principles to support their claim that the early Stoa believed that katorthomata cannot be captured in general rules. T h e y merely assume, however, that general principles are trivial. For further defence of these claims about general principles see Nell (1975) p p . 59—93.1 a m indebted to T . Irwin and J . Whiting for further clarification about these issues.
3O2
PHILLIP MITSIS
general recommendation is informative and makes a whole network of associations intelligible in a way markedly different from my telling him that to cure his depression he should get up in the morning and shovel snow, paint his garage, refinish his piano, etc. Offering a list of instances in this case is less informative in important respects than giving a general recommendation. By the same token, Seneca plausibly maintains that decreta are to be augmented by praecepta not because decreta are uninformative, but because praecepta help to convey the special information provided by decreta to the requisite moral contexts. Nonetheless, it is still difficult to see, at least at first glance, what justifies Seneca's claim that decreta are in any sense dynamic principles of reason and judgment. If we take Seneca's descriptions of the benefits of decreta as well as his supporting imagery seriously, however, we can begin to see why Stoics think that decreta have such a central, dynamic role in individual moral development. 63 At this point, it will be helpful to return to the difficulties with praecepta that Seneca thinks can be solved by appealing to decreta. A central worry is that, by themselves, praecepta may not be able to prevent us from developing merely idiosyncratic modes of moral perception. To meet such objections, Seneca gives an account of the governing conditions used for administering praecepta and discusses methods for properly guiding and testing them. For instance, he develops an elaborate contrast between the moral prospects of those living in a city, surrounded by vice, and those who are leading solitary, selfsufficient lives in the country. 64 He argues that only the right kinds of praecepta will begin to strip away the layers of worldly distraction that prevent us from viewing ourselves as utterly self-sufficient and independent moral agents. 65 As Seneca repeats again and again, vice needs a stage on which to act, and we can regain our health only by seeking retirement and retreating into ourselves. 66 A limiting condition on praecepta, then, is that they must help us to regard ourselves as approximating an ideal. Seneca's controlling idea here 63
64
65 66
Limiting the discussion to questions of individual development is somewhat artificial. I t is clear that decreta also m a y have a role to play as social principles, as instruments of social change, etc. Ep. Mor. XGIV.59-74. Nussbaum (1987) p . 135 for other i m p o r t a n t dimensions of Seneca's tactic of getting us to view ourselves as radically detached from t h e world a n d entirely self-sufficient. Cf. Ganss (1952) a n d Trillitzsch (1962) for the general background. O r as Seneca says robigo animorum effricanda est (Ep. Mor. xcv.37). Cf. Ep. Mor. xciv.71—2: Ambitio et luxuria et inpotentia scaenam desiderant: sanabis ista si absconderis.
Seneca on reason, rules and moral development
303
is in some ways an interesting anticipation of one of Kant's formulations of the categorical imperative.67 Kant argues that one categorical limitation on the permissibility of our maxims is that we view ourselves as ideal moral agents, regardless of our moral defects or those of others. Seneca, too, thinks that one of the requirements governing praecepta is that they lead us to a correspondingly ideal vision of ourselves. Of course, Seneca is not explicitly laying down a categorical restraint here precisely in the manner of Kant. 68 Yet, he clearly is concerned to elicit the limiting conditions and rational requirements that underwrite effective uses of praecepta; thus, a comparison with Kant, when properly qualified,69 can be instructive. A similar Kantian parallel70 occurs in Seneca's discussion of the general principles guiding our dealings with others. He argues that the praecepta concerned with various helping actions are governed by a decretum enjoining that we respect other persons as mutually related parts of God and nature. 71 Thus, Seneca thinks that by means of the underlying rational constraints provided by decreta, this class of praecepta can be prevented from becoming merely a haphazard collection or heap and our moral perception of individual helping actions can be similarly protected from developing in ways that are merely idiosyncratic. Seen in this light, it becomes easier to understand why Seneca insists that decreta and praecepta are mutually 67
68 69
70
71
Foundations of the Metaphysics of Morals (Beck, p . 54 ff). T h e r e are some questions about the extent to which this counts as a separate formulation of the categorical imperative; I follow A u n e (1979). Both K i d d (1978) p . 253 a n d Kerferd (1978) p p . 134-5 a r e rightly cautious about this. K i d d says that decreta ' a r e more like categorical imperatives' (thanpraecepta are). T o forestall possible misunderstandings, it should be noted that the way K a n t i a n agents are sometimes thought to test the permissibility of their maxims against the categorical imperative is very different from the m u t u a l functioning of Stoic praecepta a n d decreta. Decreta elicit rational requirements from within a set of moral characteristics that have been perceived by agents using praecepta. At least on some readings of K a n t , agents approach the categorical imperative with no prior independent moral awareness. (For further discussion, cf. Nell (1975) ch. 3.; for the possibility of K a n t i a n agents having prior recognition of moral features see H e r m a n (1985) p . 416.) I.e. K a n t ' s second formulation in FMM. Long points to a similar parallel between K a n t ' s first formulation of the categorical imperative a n d the Stoic's conception of right reason, (1974) p . 208. T h e r e are hints of this as well in Seneca's account at Ep. Mor. xcv.48 ff. N a t u r a nos cognatos edidit, c u m ex isdem et in eadem gigneret; haec nobis a m o r e m indidit m u t u u m et sociabiles fecit. Ilia a e q u u m iustumque composuit; ex illius constitutione miserius est nocere q u a m laedi; ex illius imperio p a r a t a e sint iuvandis manus. {Ep. Mor. xcv.52-3) Cf. xciv.67 ff. ( I n this section ofEp. Mor. x c v , Seneca is discussing the decreta that guide our relations to gods a n d men.)
304
PHILLIP MITSIS
supporting and necessary for moral development. Praecepta enable us to grasp morally salient features of individual situations. They are therefore able to give us our initial purchase on the world of concrete moral experience. Decreta, in turn, provide us with the rational constraints necessary for further articulating and guiding these moral perceptions. But without our initial grasp of the particular textures of moral experience, decreta would be like branches without leaves or a mind powerless to make contact with the world. Moreover, as suggested in Seneca's images, praecepta and decreta are mutually strengthening. As we develop, our cognitive grasp of both begins to deepen and their mutual connections are gradually reinforced.72 Accordingly, induction from particulars and deduction from principles have mutually supportive and overlapping roles to play in our ongoing development as well. 73 What emerges from Seneca's discussion of moral rules is something very different, then, from the static, grid-like conception of morality that is often ascribed to rule-based moral theories. The Stoics envision a process that is both flexible and tied to particulars, while giving agents a way of progressively articulating and assessing their moral experience both from within their own individual circumstances and with the help of rationally grounded principles. No doubt, difficult questions remain about the adequacy of this account, even as a plausible story about cognitive development; but, at the very least, we can perhaps begin to acquit the Stoics of some of the more flagrant charges commonly levelled at rule-based theories. in
It is difficult, nonetheless, not to notice how much of what we intuitively might find important has dropped out of Seneca's account. We might reasonably wonder, for instance, whether our 72
73
This developmental picture perhaps begins to provide part of an answer to Long's worry that, since only sages can truly understand Nature's rules, 'Nature provides a destination which is approachable by no known road.' (1970) p. 102. By emphasizing the role of praecepta in helping agents to recognize morally salient features, I do not mean to imply that they are formed strictly inductively from responses to individual situations. It is not that decreta serve only to provide rational constraints for praecepta once we have already acquired praecepta non-deductively. There are certainly other deductive relations between decreta and praecepta from the very beginning, since grasping the Tight, praecepta will require the right Stoic conception of human nature, nature as a whole, etc. At the same time, however, moral development and judgment are not just a matter of mechanically deducing the relative instances from general axioms. It might be that, as argued by Striker (1987), moral justification is strictly deductive. But the account
Seneca on reason, rules and moral development
305
grasp of particular moral situations through praecepta might not be helped rather than merely undermined by various kinds of feelings or settled habits. Or we might be worried about the possibility that, in a particular situation, morally salient features might come into conflict. These two problems - the role of non-cognitive elements in morality and the possibility of moral conflict - obviously raise a wide variety of serious challenges to the Stoics' account of rules and moral development. To give a fully satisfactory defence of the Stoic position, one would have to show, for example, why the Stoics are justified in thinking that moral values are uniform and commensurable, hence not prey to irreducible conflicts. Similarly, it would be necessary to defend the Stoics' moral psychology on a wide variety of fronts. Since I cannot pursue these larger questions here, what I propose to do is to concentrate on just one argument, albeit one that unites elements of both of these basic challenges to the Stoics in a particularly telling way. Obviously, such a procedure will leave important questions about rules untouched; nevertheless, it may at least suggest further contours of the Stoics' general position and give some indication of its potential strengths and weaknesses. In the course of discussing whether it is ever right to break common moral rules such as rules that we should keep promises, tell the truth, and so forth, W. D. Ross makes a claim that has attracted much attention: When we think ourselves justified in breaking, and indeed morally obliged to break, a promise in order to relieve someone's distress, we do not for the moment cease to recognize a prima facie duty to keep our promise, and this leads us to feel, not indeed shame or repentance, but certainly compunction, for behaving as we do; we recognize, further, that it is our duty to make up somehow to the promisee for the breaking of the promise.74 Many have objected that there is an element of contradiction in Ross's overall account. 75 On the one hand, Ross thinks that in conflict situations, we can reach clear-cut and completely justified solutions. In this passage, however, he maintains that overridden duties still retain some residual emotional influence and moral force. 74
75
given by Seneca of moral development relies on the continuous interplay of inductive and deductive elements. Ross (1930) p . 28. F o r a representative range of criticism, see Phillips & M o u n c e (1970) p p . 79 ff; Williams (1973) p p . 175 ff; N u s s b a u m (1986a) p p . 47 flf; G o w a n s (1987); Santurri (1987) p p . 41 ff. Santurri (1987) pp. 47 ff sets out the range of contemporary responses to this issue in great detail; I am indebted to his discussion in what follows.
306
PHILLIP MITSIS
Yet, if an overridden duty still holds any emotional or moral claim on us, so it is claimed, it remains both a genuine and persisting obligation. A potential contradiction thus arises for Ross,76 since we might ask how a decision can be described as unequivocally justified (i.e. morally obliged), if it has violated what is still viewed as a genuine and persisting obligation. To avoid contradiction, one might claim that reaching an unequivocal solution just means that, all things considered, there is no residual moral claim; hence agents should feel no compunction or remorse over the obligations they set aside.77 For many, however, the intuitive costs of such a move have seemed too high and they insist that it is an inescapable feature of taking 'seriously the fact that we are human persons, and not (even potentially) some kind of gods' 78 that we have such feelings as regret or compunction in situations of conflict and that we regard these feelings as important components of human value. If we take such feelings to reflect a continuing obligation, however, we will be forced to deny that conflicts admit of unambiguous solutions. In a corresponding discussion of whether it is ever permissible to break a promise (de Officiis ni.92-5), 79 Cicero, reflecting a Stoic position, gives the following example: What of the case of Agamemnon, when he had vowed to Diana the most beautiful creature born that year in his realm and was brought to sacrifice Iphigenia, for in that year nothing was born more beautiful than she? He 76
77
78 79
As Santurri points out (1987), Ross may have the means, from within his own theory, to escape contradiction. He might claim that we can feel regret for the loss of a moral good represented in the action set aside, without thinking there is any loss of a moral right. My action overall is right, though I can still feel regret at the loss of a moral good; what I feel in this case is not compunction, however, since compunction and remorse are reactions to a moral wrong. Thus, it might be claimed that in this passage Ross misdescribes the emotional phenomenology of conflict situations in terms of his own theory. It makes a difference here whether we view conflicts in terms of beliefs or desires. Those who argue that there is a moral remainder when one of two conflicting obligations is put aside often liken moral conflicts to conflicts of desire rather than to conflicts of belief. 'When we act on one of two conflicting desires . . . the rejected desire is not thereby abandoned; this can show itself in the regret we feel for what is missed. On the other hand, when we accept but one of two conflicting beliefs, the rejected beliefs abandoned; there is no regret in losing what we now regard as false.' Gowans (1987) p. 14. The Stoics' moral psychology constrains them to see conflicting obligations in terms of conflicts of beliefs. C o o p e r (1989) p . 36. It is unclear, however, exactly how precise the correspondence is between Ross's and Cicero's views, since Ross's view of prima facie obligations commits him to a strongly particularist moral epistemology, whereas Cicero's view in this passage seems compatible with either a particularist or non-particularist account.
Seneca on reason, rules and moral development
307
ought to have broken his vow rather than commit so horrible a crime. [Offic. in.95) (trans. Miller, adaptation)80 Cicero, like Ross, has no doubt that there is a correct solution to Agamemnon's perplexity. At the same time, it is clear that Stoics can ascribe no value to any accompanying emotions of regret in such situations. Throughout his account of conflicts in de Officiis in, Cicero repeatedly emphasizes the Stoic claim that no matter what kind of sacrifice is involved in making the right choice, there are no reasons for regret. Indeed, to see the extremity of the Stoics' attitude on this issue, we need only look at a brief passage from Seneca. In describing the many goods bestowed on us by philosophy, he claims that perhaps its chief benefit is that it enables us never to regret our own conduct.81 For the Stoics, Agamemnon's choice and his attitude towards his choice should be unambiguous. Admittedly, Cicero's account is not exactly the stuff of tragedy. However, before dismissing it as yet another instance of Stoic insensitivity to fairly obvious features of our ordinary emotional and moral experience, we need to see how well it fares against competing accounts. First, it is important to notice that, by itself, the Stoics' emphatic rejection of the value of regret does not mean that they could not allow for residual moral claims in cases of conflict. If a particular moral conflict were in agreement with the rational order of the universe,82 a Stoic might accept and value its competing 80
81
82
This is one of the very few instances of conflict in de Officiis that might qualify even as a n apparent moral dilemma, since the dilemmas Cicero discusses regularly involve conflicts between morality a n d utility. Cicero believes that morality consists of a uniform set of principles, though he allows that in following moral rules one m a y face conflicts among nonmoral values; thus, while practical dilemmas may occur, there can be no genuine moral dilemmas since any a p p a r e n t conflict between morality a n d utility can be resolved in favour of morality. By severely limiting the scope of what is morally relevant, Cicero a n d the Stoics are far less inclined to see a n y possibilities of conflict. See Gowans (1987) p p . 16 ff for general discussion of this issue. Itaque hoc tibi praestabit, quo equidem nihil maius existimo: numquam te paenitebit tui. Ep. Mor. cxv. 18. Cf. Andronicus peri Pathon 2 = SVF 111.414 for metameleia: Metameleia de lupe epi hamartemasi pepragmenois hos di hautou gegonosin. Also, SVF HI.563; Seneca de Benef. 4.34 (describing the wise man) ideo numquam ilium poenitentia subit, quia nihil melius Mo temporefieri potuit, quam quodfactum est, nihil melius constitui, quod quam constitutum est. Cf. EN ix.4 for a corresponding view in Aristotle. Notice as well that whereas non-Stoics might locate their regret in the fact that they are faced with a particular decision (and not in their actual decision), Stoics can feel no regrets of this sort either. I doubt, however, that the Stoics can concede the possibility of genuine moral dilemmas (as opposed to merely a p p a r e n t ones), given their view of N a t u r e as providential, rational, a n d morally coherent. T h e possibility of genuine a n d irresolvable moral dilemmas would involve, in a certain sense, the repugnant spectacle of God a n d N a t u r e deconstructing themselves. At the same time, given the connections that Stoics see between eudaimonia a n d
308
PHILLIP MITSIS
claims without feeling the emotional pangs of remorse. Within their own theory, the Stoics could attribute to agents caught in such conflicts a corresponding 'reasonable emotion' based on a recognition of the value of the actions that are set aside.83 That the Stoics ever made such a move is no doubt entirely unlikely; but even its theoretical possibility should alert us to how they can resist arguments that make any straightforward inferences to moral realities from feelings or emotions. For the Stoic, our emotional states depend on (or are) our rational judgments about value; they are not separate or competing sources of motivation, action, or knowledge. Accordingly, the Stoics think that, taken by themselves, emotions have no independent explanatory force, nor can they give us any reliable indications about the status of moral realities. In order to show that the emotions of agents who feel caught in a dilemma reflect anything about the actual metaphysical status of dilemmas,84 critics of the Stoic position need to explain how emotions of regret, remorse, and so forth, can be described without merely presupposing a prior belief in the reality of moral dilemmas. Similarly, they need to demonstrate that the affective elements in these emotions give us reliable or determinate indications about the status of moral realities. It is not clear how either of these claims can get off the ground, however, without merely begging the question in favour of the existence of dilemmas. One might attempt, for instance, to specify certain kinds of observable behaviours or dispositional characteristics indicative of remorse. But then several well-known problems quickly arise. Sartre's budding young Resistance fighter, for instance, might be groaning aloud, wringing his hands, and cursing the horrible choice that he must make between helping his country and tending to his sick mother. But, then again, he might really only be play-acting; perhaps he is greatly relieved to be finally getting away from her apron-strings. What can be observed, in any straightforward sense of this term, is behaviour, not remorse. Moreover even if, like many behaviourists, we were to
83
84
morality, a genuine moral dilemma might put into j e o p a r d y the Stoic claim that happiness consists in following N a t u r e , since a genuine moral dilemma would make it impossible to know h o w one c a n rationally follow N a t u r e . For some further connections between eudaimonist a n d theological considerations, see Long (1989) p p . 77 fF. Admittedly, this is not a very lively option, since I take it that, unlike for fear, pleasure, a n d desire, there is n o evidence for a n y rationally reconstructed forms oflupe. O n eupatheiai, cf. Frede (1986) p . 9 3 . Here I pass over problems associated with the arguable claim that such emotions accomp a n y conflict situations as a matter of course.
Seneca on reason, rules and moral development
309
extend behavioural criteria to include non-observable mental states and dispositions, any determinate account of our hero's disposition to feel remorse still must include some reference to his belief that he is faced with a moral dilemma. 85 By the same token, it is unclear how we can specify behavioural or affective criteria for distinguishing remorse from regret, for instance, without making reference to agents' beliefs (or disposition to believe) that they are still under some unfulfilled obligation. Yet, any reference to agents' beliefs that they are in a dilemmatic situation merely begs questions about moral dilemmas. Thus, no matter how elaborately or carefully we attempt to catalogue the emotional phenomenology of agents who take themselves to be caught in conflicts, the Stoic can point out that in order to be in the required emotional state of remorse, agents must believe that they are caught in a genuine dilemma. And it is precisely such a belief in the existence of genuine moral dilemmas that is at issue. Thus, we are brought back full circle to the Stoics' claim that there are no separate affective routes to moral realities. By the same token, it might be objected that, whatever the difficulties in explaining the remorse of others, we have direct introspective access to our own remorse. We just know what it feels like to be caught between clashing obligations. Here the Stoics can agree that our remorse perhaps may feel to us like some sort of gnawing distress. But it is open to them to claim that our distress is irrational or pathological unless we really are caught in a genuine dilemma. It may be, for example, that I have an entirely mistaken belief about my obligations or a mistaken belief that they are in conflict. Or perhaps my remorse indicates deeper pathological problems. Suppose that I believe that I am obliged both to keep promises and to kill cats. In what I take to be an appropriately dilemmatic circumstance, I am forced to make a choice between these conflicting obligations and choose to keep a promise. I also, however, feel great remorse in setting aside my perceived obligation to kill cats. Clearly, my remorse is pathological and my perceived dilemma spurious. In terms of its phenomenological feel as a kind of gnawing distress, however, it is not at all clear how my remorse in this instance differs in any way from other, less pathological, episodes of remorse. But then it also is unclear how affective elements in my 85
Cf. McDowell (1979) pp. 346-7 for some parallel problems in specifying what a disposition is a disposition to. Also, Santurri (1987).
310
PHILLIP MITSIS
remorse can, on their own, indicate anything about the status of moral realities. None of these considerations demonstrates, of course, whether genuine moral dilemmas exist or whether remorse is ultimately a proper human response to moral conflict. Such larger issues about the nature of moral reality are better answered at the appropriate metaphysical level. But they do show how difficult it is to construct a convincing argument defending the possibility of separate non-rational paths to such moral realities. One still might object, nevertheless, that even if non-rational elements fail to offer an independent route for approaching these metaphysical issues, the emotions offer important support both for correct beliefs and for our ongoing development. The Stoics, though, can use a similar strategy to challenge the usefulness of non-cognitive elements even as supports for moral judgment and development.86 For instance, if I am urged to develop and make habitual my feelings of remorse, the Stoic can rightly ask for more precision about what this habituation consists in and what value it has. If my remorse indicates nothing about moral reality or if it provides me merely with an occasion for wallowing in self-reproach and for admiring the depths of my own sensibilities, I might be better off not making remorse a habit. If, on the contrary, the purpose of habituating me in remorse is to provide me with a deeper awareness of moral demands and to furnish me with moral motives for my future actions, this arguably begins to sound more like a cognitive story, and it is unclear why the cognitive regimen prescribed by the Stoics is not a more reliable method for developing the requisite kind of awareness and for acquiring the right sorts of motives. 86
Cf. Nussbaum (1987) p. 165 for a discussion of emotional motives. It is worth making the following distinction, however, with respect to emotional motives. In talking about the human value of regret, remorse, etc. we can distinguish: (a) valuing an emotion for itself or as an indication of personal emotional sensitivity, sensibility, etc. (b) valuing an emotion as a means to moral action or as an indication of moral awareness. Whereas (a) is self-referential and may turn out to be self-indulgent, (b) is likely to be other-directed, more helpful, etc. (Cf. Herman (1986) pp. 424 ff for a Kantian defence of (b).) Kant's suspicion that (a) does not provide reliable motives for virtue in many ways resembles Stoic worries. Moreover, if what we value is (b), the Stoics might plausibly argue that what we are isolating is a cognitive element, and hence not a separate non-cognitive element at all. Moreover, this distinction might help to absolve the Stoic account from the charge of circularity (cf. Nussbaum, p. 165); in the case of an emotion like regret, for instance, it may be that Stoics can get their opponents to agree that what they are valuing in seemingly affective responses is really (b).
Seneca on reason, rules and moral development
311
IV
I want to conclude by looking at an argument in Seneca apparently meant to combat scepticism about rules. In discussing the Stoics' view of morality, a number of scholars have been attracted to an analogy between moral rules and game rules.87 In some ways this analogy has been helpful, but it also can be seriously misleading. It is helpful because games offer a relatively uncontroversial model for activities in which our immediate strategic aims are not identical or strictly instrumental to our ultimate goal in playing games. Thus, it can be argued that games offer a suitable model for capturing the Stoic distinction between the skopos and the telos of moral actions. At the same time, however, this analogy is potentially misleading because it fails to capture both the realist and normative elements that rules have for the Stoics. Game rules are conventional and can be changed to some extent whenever players agree to do so. When analogies are made between games and social mores, linguistic practices, and so forth, it is usually for the purposes of showing the conventional nature of these practices and the purely descriptive nature of their rules. Clearly, the Stoics need and want something more from moral rules than mere communal agreement. In many ways, the difficulties they face with rules correspond to their problems in defending the objectivity of their general moral notions. It is fairly clear that the so-called consensus omnium arguments used by Stoics to generate moral preconceptions can at best demonstrate that such preconceptions have several admirable uses88 either in moral inquiry or in the moral lives of individuals and communities. On their own, however, such arguments are insufficient to establish the objective and normative validity of moral preconceptions without relying on some questionable inferences to nature and to divine providence. Yet, without some kind of objective and normative backing, Stoic moral claims become much more vulnerable to attack. So the question arises, can the Stoics give any defence of the objectivity and normativity of their moral rules? As is well known, their answer here relies in part on a series of parallel inferences to 87 88
Cf. K i d d , (1971) p p . 157, 170 a n d (1978) p . 253 a n d in more detail by Striker (1986) p p . 185 ff. K i d d draws attention to Epictetus Diss. 11.5, 15 for a possible ancient analogue. Cf. Long & Sedley (1987) 1 253 on Adversus Mathematicos v i n . 3 3 i a - 3 3 2 a n d on Epictetus' defence at Diss. 1.22 1-3, 9-10; also Schofield (1980) p p . 283 ff.
312
PHILLIP MITSIS
Nature and to divine providence. I want to look at an argument, though, that suggests some further possibilities. At Ep. Mor. xcv. 60-1, Seneca sets out the following argument: Praeterea non intellegunt hi qui decreta tollunt eo ipso confirmari ilia quo tolluntur. Quid enim dicunt? praeceptis vitam satis explicari, supervacua esse decreta sapientiae [id est dogmata]. Atqui hoc ipsum quod dicunt decretum est tarn mehercules quam si nunc ego dicerem recedendum a praeceptis velut supervacuis, utendum esse decretis, in haec sola studium conferendum; hoc ipso, quo negarem curanda esse praecepta, praeciperem. Seneca maintains that both decreta and praecepta are inescapable features of our linguistic and conceptual framework. If we attempt to deny that there are decreta, we will violate the pragmatic and logical requirements of any argument for that denial. Corresponding attempts to do away with praecepta are similarly self-refuting. Although it is fairly clear that his argument is aimed primarily at those who think they can dispense with either decreta or praecepta, it can be pressed into service against a more general kind of scepticism about rules. Seneca claims that denials of moral rules somehow must invoke further moral rules. Consequently, they are not just a matter of conventional agreement, since they are an inevitable structural feature of any thought or talk about morality. 89 Even if we grant Seneca the force of this argument, 90 however, it would only show that some rules are inevitable; it fails to secure for the Stoics a justification of their own particular brand of moral rules. Nonetheless, this argument potentially contains a stronger normative and realist defence against sceptical challenges than that of consensus omnium arguments. If moral rules are shown to be part of the intrinsic framework of all thought and talk about morality, they might have a normative and objective status that goes beyond mere agreement. For the Stoics, of course, any talk about the intrinsic structure of logos will still involve inferences to Nature and God. But compelling arguments about the objective and normative credentials of logos and its structuring rules may not rest only in the lap of God. 89 90
See Winch (1958) p p . 57 ff for a similar attempt t o show how denials of rules are self-refuting. See Williamson (1989) for an extremely dismissive attack on such claims. H e argues that the use o f ' r u l e ' to describe what the sceptic relies on in denying rules is a 'mere pun'. I n any case, the sceptic about rules can perhaps avoid self-refutation in much the same way as the sceptic about knowledge, i.e. by refraining from following a rule on each particular occasion without dogmatically rejecting rules. Nevertheless, there remains the question of whether such a non-dogmatic sceptic is still acting in accordance with a rule.
CHAPTER I I
Chrysippus on psychophysical causality David Sedley
INTRODUCTION
This chapter will in effect be an extended commentary on a single passage, Cicero, de Fato 7—9, which indirectly attributes to Chrysippus a strong thesis of psychophysical causality. I shall hope to show what that thesis amounts to, and how it fits into the teleological structure of the Stoic causal nexus. At the same time, I shall be offering reasons for seeing the Stoic theory in question as largely a development of Platonic psychology. This last point needs some qualification. Chrysippus himself was openly opposed1 to at least one feature of Plato's psychology, namely the tripartition of the soul, as expounded in the Republic, Phaedrus and Timaeus. But equally, he was sympathetic to, and almost certainly deeply influenced by, the more 'Socratic' psychology to be found elsewhere in Plato's dialogues - the monistic theory presupposed in the Protagoras, developed in the Phaedo,2 and arguably still observed in the Theaetetus - according to which the soul is in itself a purely intellectual faculty. Given the further fact that Socrates, unlike Plato, was from the start revered as an absolute authority by the Stoics,3 it is probably most correct to locate the background to Stoic psychology not in Platonic but in Socratic psychology.4 At the In revising this paper I have had the benefit of comments from many sources, including extensive criticisms and suggestions from Julia Annas, Jonathan Barnes, Susanne Bobzien, Nicholas Denyer, Jaap Mansfeld, the late Paola Manuli, Richard Sorabji, Robert Sharpies, and Robert Wardy. I am most grateful for all the help received, although I am of course entirely to blame for whatever uses I have put it to (or failed to put it to). 1 See Galen, PHP IV.I.6, where, as M. Schofield has pointed out to me, eiircbv cos KTA, probably means 'having said that Plato believed that . . . ', not that Chrysippus 'takes the Platonic position that . . . ' (De Lacy); cf. in.i. 14—15. (The contradiction which Galen claims to find in Chrysippus is made explicit not in this sentence but first at iv.1.14.) 2 3 Cf. B. Inwood's remarks, pp. 158-9. See Long (1988). 4 Even Posidonius' reinstatement of the tripartite thesis, although inspired by Plato, was not presented as replacing Socrates' authority with Plato's, but as reverting to the more ancient
313
314
DAVID SEDLEY
same time, it would be misleading to imply that any ancient reader of Plato operated with an entirely clearcut distinction between historically Socratic texts on the one hand and Platonic texts on the other. Even if, as seems likely, the Stoics' preference for a monistic psychology was influenced in part by the belief that it was more authentically Socratic, it would be unwise to assume that they never used the less obviously Socratic dialogues, 5 wherever these latter were useful for supplementing the ideas contained in the former. BIOGRAPHICAL TELEOLOGY
In the lacuna between chapters 4 and 5 of Cicero, de Fato much is lost, including an argument of Chrysippus' whose only echo comes in the following words at 7-9: Let us return to Chrysippus' snares, and reply to him first about the world's interconnexion itself [de ipsa rerum cognatione],6 then pursue the rest later. We see how great the differences are between the natures of places. Some are healthy, some disease-ridden. In some the people are phlegmatic to the point of overflowing, in others they are utterly dried out. And there are many other immense differences between places. At Athens the atmosphere is rarefied, resulting in the Attics' reputedly sharp wits; while at Thebes it is heavy, so that the Thebans are stout and tough. Yet neither will that rarefied atmosphere bring it about whether someone attends Zeno's lectures or those of Arcesilaus or of Theophrastus, nor will the heavy atmosphere bring it about that someone competes at the Nemean rather than the Isthmian games. Carry the distinction further. What can the nature of the place do to make us walk in Pompey's colonnade rather than on the Campus? Or to walk with you rather than with someone else? Or on the Ides rather than the Kalends? Just as the nature of the place has a bearing on some things but not on others, so too let the influence of the stars [qffectio astrorum] have power over some things, if you like: it certainly will not have power over everything. But, he will say, given that men's natures differ, so that some love sweets while others love savouries, some are passionate while others are irascible
5
6
authority of Pythagoras (Posidonius F151, 165.168-72 Edelstein/Kidd (1989)), whose follower Plato was taken to be in this matter. Thus Plato was for the Stoics never the ultimate authority, but always the spokesman or interpreter of some higher authority. Even the Timaeus need not have been regarded as entirely un-Socratic: Socrates is at least a willing listener to Timaeus' discourse. The lack of early Plato commentaries makes it hard to make a confident guess at Hellenistic views on this matter, but note, at least, that Socrates' similar role in the Sophist was seen by Proclus (In Platonis Parmenidem 622.24, 629.37-630.1) as sufficient reason for attributing to him the ideas of the dialogue's main speaker. The reading of some MSS reported by Turnebus, cognatione, is well defended against the more favoured contagione, in both this passage and ib. 5, by Luck (1978).
Chrysippus on psychophysical causality
315
or cruel or arrogant, and others shrink from such vices - given, he says, such gulfs between different natures, why should it occasion surprise that these dissimilarities are the products of different causes? In arguing this, he does not see what is at issue, and on what the cause rests.7 For it does not follow, if some people are more inclined to some things because of natural antecedent causes, that there are natural antecedent causes even of our volitions and impulses. For if that were so, nothing would be in our power. But as it is, while we admit that it is not up to us whether we are clever or stupid, strong or weak, anyone who thinks it necessarily follows that even whether to sit or walk is not a matter for our own volition does not see what follows from what. For granted that both gifted people and dim-witted people are born that way through antecedent causes, and likewise the strong and the weak, it still does not follow that even sitting and walking and performing some given action are marked out and established by principal causes [principalibus causis]. Apparently, Chrysippus has appealed to the causal influences of native environment and of the stars to establish the pervasive power of 'natural' causes in human psychology. There are points in this text to which we will need to return, but the first issue I want to raise is how this natural causation is supposed to square with Stoic providentialism. To allow the automatic determination of character by non-rational causes may sound like an incautious concession to a mechanistic world-view. The nature and magnitude of this problem can best be grasped by turning directly to chapter 30, where Stoic causality can be viewed from the opposite pole. Chrysippus is responding to the Lazy Argument, according to which decisionmaking is superfluous if just the same outcome is fated to ensue regardless: Some events in the world are simple, he says, others are complex. 'Socrates will die on such and such a day' is simple: his day of dying is fixed, regardless of what he may do or not do. But if what is fated is 'Oedipus will be born to Laius', it will not be possible to add 'regardless of whether or not Laius has intercourse with a woman'. For the event is complex and 'co-fated'. 7
No doubt the primary sense of 'in quo causa consistaf is 'the point with which the argument is dealing' (Rackham [1942]), *ce quifait lefond du debaf (Yon [1933]), or 'worin der Streitgegenstand bestehf (Bayer [1963]) - cf. Cicero, Pro Quinctio 32, 79 for the expression. But in context I find it hard to avoid seeing a punning second sense: 'what a cause consists in'. For Cicero's ensuing objection is precisely to what he sees as a muddle about what constitutes a cause; cf. his own (Carneadean) more rigid notion of cause, developed below (especially Fat. 34, 36), as that which produces its effect efjicienter or sua vi. Hence a looser but more apposite translation might be 'he is fighting a lost cause'.
316
DAVID SEDLEY
While the doctrine of'co-fated' events has attracted a good deal of discussion, little attention seems to have been paid to Chrysippus' example of 'simple' fate, the death of Socrates. Given Chrysippus' well advertised belief that no event escapes the causal nexus of fate, it is at first sight hard to make sense of his distinction between simple and complex. Surely the route to Socrates' death is just as fated as the event itself. It is not enough to respond that the distinction is one between outcomes fated via an individual's actions and those fated independently of them. It might have some superficial plausibility that someone with a terminal illness, or living in an earthquake zone, will die at such and such a time regardless of anything he does or fails to do. But it will hardly do for Socrates' death. Was it not famously the case that Socrates need not have died on that day, if he had yielded to Crito's entreaties to escape? If the problem has gone unnoticed hitherto, it is no doubt because 'Socrates' has been assumed to play his role, familiar in Aristotle and elsewhere, as a mere specimen individual - the role usually assigned to Dion in Stoic texts, although occasionally also to Socrates.8 I find this hard to believe. The examples used in the de Fato debate are far from haphazard. They constantly recur to carefully chosen and exploited stories, such as the death of Scipio, the behaviour of philosophers, and the Oedipus story, which is here directly juxtaposed with the Socrates example. To throw in, casually and without awareness of its significance, that uniquely important event in the history of philosophy, the death of Socrates, would be an astonishing departure from the pattern. 9 The answer, I suggest, lies rather in Plato, Crito 44a—b (already cited by Cicero's Stoic spokesman at de Divinatione 1.52, the preceding work in the trilogy), where we hear of Socrates' prophetic dream while in prison that he would die on the third day. To a Stoic, it will have seemed that Socrates' decision to stay and die was to a large extent guided by the revelation that he was bound in any case to die on that day - a divine indication that it was morally preferable for him to do so willingly. (This is the kind of consideration which underlies the Stoic analogy of man as a dog tied to a cart, who, willing or unwilling, must follow regardless,10 and Chrysippus' 8 9
10
E.g. by Diogenes of Babylon at Diog.Laer. vii.58. That the examples used in this passage are Chrysippus' own is demonstrated by Barnes (1985a). SVF 11.975 = LS 62A (LS = Long & Sedley (1987)).
Chrysippus on psychophysical causality
317
assertion that 'If I actually knew that I was fated now to be ill, I would even have an impulse to be ill.') 11 The moral rightness of Socrates' willingly drinking the hemlock was intimately bound up with his understanding of fate: he had had, but declined, the opportunity to die z/wwillingly on that day. Hence, in Stoic eyes, his death on that day had to be unconditionally fated. Thus Socrates' death brings out a feature of Stoic causality which can easily go unnoticed, and which makes it look as distant from mechanistic determinism as it could conceivably be. There is not just a single-stranded causal chain leading from Socrates' birth to his death as a result of drinking hemlock, but one which repeatedly branches out into hypothetical alternative strands, such as Socrates' escaping from prison but thereafter being re-arrested, or falling ill in flight, or whatever, and thus dying on that very same day on which he in actuality drank the hemlock. For both logical and causal reasons, only one of these chains of events was ever going to be realized. But although it was true all along that things would in fact turn out thus, it was not necessary that they should,12 and the very meaning of Socrates' moral achievement depended on an intelligently planned network of diverging but ultimately re-converging alternative opportunities, 13 only one of which represented a willing acceptance of what fate had ordained as the final outcome. If this is right, the Stoic causal nexus, far from being mechanical, exhibits to a quite astonishing degree the meticulous workings of an intelligent teleology. The choice of Socrates' death as an example has already suggested that the Crito and Phaedo were crucial texts in the Stoic hagiography. This is confirmed by the manifest importance of the Phaedo to Stoics as a guide to the proper attitude to suicide. 14 Socrates' declaration at Phd. 62C6-8 that one should not kill oneself until god sends some necessity evidently influenced, among others, Zeno's own suicide 11
12
13
14
SVFHI . I 9 I = L S 5 8 J .
For the importance to Chrysippus of severing future truth from necessity by establishing the existence of counterfactual possibilities, cf. Fat. 13 = LS 38E 3.1 would take these to be, in essence, opportunities (LS 1.235). It may seem odd to use simplicia, of all words, for events which are determined by a complex web of interlocking causal chains. But the term merely represents the designation of such events as 'absolutely' (frrrAcos) fated (SVF 11.998, p. 292,31). (On 2 July 1990, 1,426 pilgrims died in a stampede at Mecca. King Fahd blamed them for disobeying traffic instructions, but added, 'It was fate. Had they not died there, they would have died elsewhere, and at the same predetermined moment.') Well documented by Rist (1969), ch. 13, and Griffin (1986).
318
DAVID SEDLEY
story,15 in which he was said to have recognized a divine sign that the proper time for his departure had arrived. But the Stoics' debt to the Phaedo runs deeper than that. Even the teleological causal network within which Chrysippus located Socrates' decision to drink the hemlock is likely to have arisen through reflection on the causal theories of the Phaedo.16 Socrates maintains at Phd. 96-9 that no cosmic state of affairs is adequately causally explained until one has isolated the good end served and thus related it properly to the controlling intelligence. To make his point, he appeals to the analogy of the cause of Socrates' sitting there in prison (98b-99b). No amount of talk about the position of his bones and sinews will reveal the true cause of his sitting there. That will only emerge when attention is drawn to certain decisions by the Athenians and Socrates himself as to what it is best to do. The chosen example is, at least on the surface, only an analogy for the operations of cosmic intelligence. But you hardly have to be a Stoic to be left wondering whether this series of events surrounding Socrates' death is not to be understood as itself due ultimately to the teleological workings of the cosmic intelligence. 17 That the operations of the cosmic intelligence do indeed include subordination of the details of individual lives to the overall good is affirmed with sufficient emphasis in the Laws18 to make this a quite natural reading of the Phaedo.19 We are thus left with the likelihood that the Chrysippean worldview, where the explanation even of individual human decisionmaking requires a teleological causal framework which extends vastly beyond the individual's internal psychology to a kind of 15
16 17
18 19
Diog.Laer. vii.28 (cf. Cleanthes' death, ib. 176); Rist (1969), pp. 242-3. Seneca, Ep. 104.21 confirms that Zeno's death was compared to Socrates', but adds a subtle distinction: 'Socrates will teach you how to die if you must, Zeno how to die before you must.' The point, I take it, is that while Socrates knew he was bound to die on that very day, Zeno recognized no more than that the time for his death was approaching. Frede (1980) notes how the Stoic distinction between amov and ocma takes up the usage found in the Phaedo. I have in fact been asked this very question by a (non-Stoic) member of a lecture audience. It gets a little backing from Phd. 62D-C, where Socrates links his own conduct concerning his death with the gods' superintendence of his life. Laws 9O3b-d; cf. Long (1974), p. 151 for its relevance to Stoicism. Stoicism differs from Aristotelianism in not calling final causes causes. Plato is less clear on this point, but it is certainly easy enough to read both the Phaedo and the Timaeus as making Intelligence the true cause, and explanations of the form 'because it is best' no more than a shorthand for the causal role of Intelligence. Read in this way, Plato is in close agreement with the Stoics, who regard the cosmic principle god as in a way the ultimate cause (S.E. M ix. 75-6 = LS 44c, Sen. Ep. 65. 2 = LS 55E).
Chrysippus on psychophysical causality
319
cosmic super-psychology, is a direct descendant of the Phaedo. I shall shortly be trying to trace the Phaedo's influence even further. We are now ready to return to de Fato 7—11. The example of 'natural' causation discussed there looked surprisingly mechanistic in tone. If you are born at Thebes, the heavy atmosphere makes you intellectually dense but physically tough. If you are a native Athenian, the rarefied atmosphere makes you sharp-witted. It may well all sound like a matter of blind luck, and quite foreign to the intensely rational causality of the Stoic world.20 But it is actually quite straightforward to incorporate this environmental causation into the rational causal nexus. To a Stoic, after all, it will hardly be an accident that Socrates was born at Athens and not at Thebes. I shall amplify this observation shortly. The atmospheric differences between Athens and Thebes reflect on a small scale a crucial feature of the overall arrangement of the cosmos. Cicero's Balbus in the de Natura Deorum puts the point as follows (ND 11.17), m a n argument possibly deriving from Chrysippus:21 Again, don't we even understand that all higher things are better, whereas the earth is the lowest, surrounded by the densest air? Hence it is through this very cause that what we see happen in certain zones and cities namely that people's intellects are duller because of the fuller nature of the atmosphere - this very same thing has befallen the human race, because they are located on the earth, which is the world's densest zone? In context, this is an argument for the Stoic conclusion that the world exists not just for us but for higher and better beings too, the gods. Once again, its basic idea is drawn from the Phaedo — this time from the myth, in which our own wretched habitations in coarse air, although better than the even denser atmosphere of water in which fish are condemned to live, are vastly inferior to the aetherial realm up above, in which purified souls can dwell in wisdom and happiness, close to the gods. It is important here to remember that the Stoics themselves likewise held that at least virtuous souls enjoy after 20
21
O f course, the tenet that physical environment influences intellect a n d character was in itself a commonplace - cf. the Hippocratic Airs, waters and places 12, 24; Plato, Tim. 24.c4.-cl3, Laws v.747d2-e2; Aristotle, Politics vn.6. W h a t occasions surprise is the degree of influence attached to it by Chrysippus, in view of the absolute centrality Stoic ethics assigns to rationality in moral development. Chrysippus is named as author of the argument quoted in 11.16, a n d it is left unclear precisely where the citation from h i m ends.
320
DAVID SEDLEY
death a discarnate survival in the upper atmosphere22 - yet another sign of the Phaedo's decisive importance for them. 23 As in the Phaedo, so too in Stoicism, atmospheric variations form part of a teleological structure in which souls are assigned to regions with a degree of atmospheric density or rarefaction appropriate to the moral or intellectual character they are to possess.24 If we now return to de Fato 30 (translated above), we may observe that it can be no accident that the first two examples of fated events chosen by Chrysippus are the death of Socrates and the birth of Oedipus. 25 Socrates was a native Athenian, who, as we have seen, had the intellectual understanding to respond in the right way to the promptings of fate with regard to his death. Laius, born in the dense atmosphere of Thebes, was intellectually disadvantaged. It is therefore only to be expected that he should have responded as he did, in the wrong way, to the promptings of fate concerning his own death, by disregarding the oracle which warned him that if Jocasta bore him a son the son would kill him, and also perhaps by subsequently exposing the infant Oedipus in a vain attempt to thwart the oracular warning. 26 (That this oracle featured in the debate is proved by Carneades' use of it in his retort at Fat. 33.) 22 23
24 25
26
Cf. S F F 11.774, 809-22; H o v e n (1971). T h e Phaedo m y t h itself is, strictly speaking, a b o u t reincarnation, n o t discarnate survival. But it emerges a t the e n d (114.C2-6) t h a t fully purified souls d o achieve discarnate existence, a n d it is a n easy inference t h a t the almost discarnate life of the aether-dwellers is primarily a didactic myth pointing forward to the soul's total escape from the body. T h e evidence of Cicero's Tusculan Disputations makes it clear that Hellenistic interest in Plato's views on the afterlife played d o w n reincarnation in favour of discarnate survival. I argue this, with regard to the Phaedo, in Sedley (1989). For the Stoics, see Hoven (1971), pp. 66ff. Against m y attempt to pair off the examples at Fat. 30 with those at Fat. 7—11, it has been urged by one critic that Fat. 30 adds a third example, Milo's wrestling at Olympia, which has a m u c h less exact match with the earlier passage. But (as N . Denyer has helpfully pointed o u t to me) Milo came from Croton, a town proverbial for its healthy climate, which was considered to account for the numerous successful athletes it produced (Strabo vi. 1.12); hence it is entirely plausible that Chrysippus h a d mentioned Croton along with Athens a n d Thebes in the lost passage between Fat. 4 a n d 5. I n the debate between the Stoics a n d their opponents reported at Alexander, de Fato 202, the focus is on Laius' mistake of getting d r u n k and sleeping with Jocasta, in disregard of the oracle's warning (see Euripides, Phoenissae 13-31 for the version of the story followed; on the Stoic treatment of conditional oracles, see Sharpies (1983), p p . 166-7); b u t a t SVF 11.939 Chrysippus is interested in Laius' attempt to kill the infant Oedipus. (We need not be bothered by the fact that a Stoic might not consider Laius' d e a t h at the hands of Oedipus an 'evil': within the framework of the myth his decision was, uncontroversially, a mistaken one.)
Chrysippus on psychophysical causality
321
PRINCIPAL CAUSES
All this puts us in a better position to understand Chrysippus' argument alluded to at de Fato 7-9, to which I now return. It is an argument for cosmic sumpatheia (represented by Cicero's 'rerum cognatione'),27 the total interaction of all parts of the world, which Chrysippus adduced in favour of his fate doctrine. 28 In the preceding chapters (5-6), we have had the tail end of Cicero's retort to the arguments of Posidonius for fate, all of them apparently appealing to the evidence of amazing coincidences. Cicero is there inclined to accept those cases which can be attributed to natural sumpatheia (naturae cognatio, 5) without any added idea of rational destiny, i.e. cases where a purely natural explanation is available, but to reject as mere flukes those in which rational destiny would otherwise have to be involved, such as the uncanny fulfilment of prophecies. Thus sumpatheia has already been marked off by Cicero as a thesis which restricts itself to the natural causal interaction of the world's parts, and need not include any element of rational destiny. So when in chapter 7 he turns to Chrysippus' arguments on the same issue of sumpatheia, we should expect the cases to emphasize mechanical rather than purposive causation, as indeed they do. They are, specifically, the influence of atmosphere on intelligence and physique, the influence of the stars,29 and the congenital origin of character traits. Cicero's response is not outright rejection, but an insistence that there is no proper further inference from a general causal influence of such factors on character to the conclusion that they determine even agents' individual decisions. This might give the impression of substantial common ground between Chrysippus 27
28 29
Translated 'coniunctio naturae\ 'concentus' and 'consensus' at Div. 11.34, an(^ ' convenientia et coniunctio naturae" at ib. 11.124. But there is no reason to doubt that 'cognatio rerumlnaturae' in Fat. is another translation, introduced in the lost part. For the reading, see note 6 above. SFF11.912. Discussions of the role of astrology in Stoic arguments (especially those a t Fat. 11-16) sometimes fail to take account of the fact that stellar conjunctions at the moment of birth are for the Stoics causes of subsequent events in a person's life, a n d not, as other omens are (see e.g. Div. 1.118), merely god-given signs. For Posidonius, see Augustine, Civitas dei v.2 = Posidonius F H I Edelstein-Kidd (1989) (deriving from the lost portion of de Fato), 'vim ... siderum quaefuerat quo tempore concepti natique sunC\ for Chrysippus, 'affectio astroruirC at Fat. 8; cf. also, for the general point, Div. 11.89, Manilius iv.105. This very feature that astrological signs are causally untypical compared with dreams, oracles etc. seems sufficient explanation of the fact (if it is a fact: pro, see Long (1982a), contra, see Ioppolo (1984)) that Chrysippean Stoicism gave astrology a low profile in its treatment of divination.
322
DAVID SEDLEY
and Cicero on sumpatheia. But what we have now learnt about Chrysippus' world view shows how misleading that impression would be. The material influences on human psychology are exploited by Chrysippus as integral features of the purposive working of rational destiny in individual lives. If Chrysippus did, as Cicero could be thought to suggest, conclude that such material influences were individually sufficient to determine the outcomes of individual decision-making processes, that was indeed unjustified. But if my earlier reconstruction is correct, that is not Chrysippus' point. The rarefied atmosphere of Attica was, taken by itself, far from being a sufficient condition of Socrates' choosing to drink the hemlock. After all, Crito was another native of Attica, but he would not have made the same decision if caught in the same predicament. Even so, Chrysippus evidently held that, as a major contributory factor to Socrates' intellectual powers, the atmosphere of Attica was in some sense a salient cause of his action. Chrysippus' claim does, in fact, seem to have been along these lines. What Cicero denies, and hence presumably what Chrysippus asserted, is that these congenital influences on character are causae principales of individual decisions {Fat. 9, translated above). 30 What are causae principales? The answer can, I hope, be squeezed out of de Fato 41-3. In apportioning the responsibility for human actions, Chrysippus distinguished two classes of cause: (a) perfectae et principales, and (b) adiuvantes et proximae. When a cylinder is rolled, the initial push falls into class (b), its roundness or 'rollability' (volubilitas) into class (a). And when a human agent responds to an impression by acting, the impression belongs to class (b), the agent's psychological character to class (a). Chrysippus' object is to argue that, while the main responsibility for any event lies in causes belonging to class (a), the causal chains that constitute fate belong to class (b). This much is familiar, and widely discussed.31 But what can we 30
31
This passage has not generally been recognized as alluding to a specifically Chrysippean usage of principales causae. It is, I suppose, conceivable that Cicero has himself supplied the term in his response to Chrysippus, but I doubt it. I t is no part of Cicero's own causal terminology, in Fat. or elsewhere, a n d it is after all explicitly attributed to Chrysippus at Fat. 41. I prefer not to tackle here the various controversies surrounding this text. I shall simply affirm my belief that it is an authentic report of one particularly interesting Chrysippean treatment of fate. That it seriously distorts Chrysippean doctrine is impressively argued by Donini (1975). But I do not, at least, feel prepared to endorse his complaint that it is
Chrysippus on psychophysical causality
323
infer from it about causae principales? First, it is crucial to realise that the two terms within each class are non-synonymous. That in class (b) the 'auxiliary' and 'proximate' causes are non-synonymous is relatively easy to argue. The proximate causes seem to be 'procatarctic', 32 i.e. initiating or triggering causes, while auxiliary causes (sunerga aitia) are known to be those which intensify or facilitate effects.33 It would be natural to group the two kinds together in a single class, not only because both constitute secondary causal factors, but also because one and the same item will often serve in both guises: one push starts the cylinder rolling, a second push makes it roll faster.34 Again, in class (a), causaeperfectae must translate the Greek aitia autotele and these are known to be sufficient conditions:35 the effect lasts precisely so long as the 'complete' cause is present. Now it should be obvious36 that Chrysippus' examples of the cylinder rolling and the agent's response to an impression cannot be invoking either the cylinder's roundness or the agent's character as causae perfectae, since evidently the cylinder does not roll for the entire time that it is round, nor does the agent perform the same action for the entire time that he has his present character (otherwise he would not need the stimulus of an initial impression). Rather, these must be taken to be salient causes, which determine the object's or agent's behaviour only when combined with a set of further necessary or appropriate conditions. And since, on the model of class (b), we by
32
33 34 35 36
unStoically indeterministic in tendency: its principal novelty, compared with other Stoic accounts, seems to m e to lie in its restriction of the n a m e 'fate' to one part of the (fully deterministic) causal nexus. This seems the natural inference to d r a w from the fact that in Fat. Chrysippus identifies fate with proximate causes, while at Plutarch, de Stoicorum repugnantiis 1056B = LS 55R it is maintained by his apologists that h e identifies fate with procatarctic causes. Cf. Frede (1980) p . 241, for the same conclusion. I agree with those who take it that at Cicero, Topics 59 ''alia autem praecursionem quandam adhibent ad efficiendum\ which corresponds to the role of the 'proximate' causes in Fat., accurately conveys the meaning of 'procatarctic'. I have been unconvinced by the attempt of Gorier (1987) to identify 'procatarctic' causes with Cicero's principales, largely because I find myself in disagreement with h i m as to w h a t a procatarctic cause is. But I owe to Gorler's article the realization that perfectae a n d principales must be distinct species of cause, for which he rightly cites Fat. 42, 'non . . . perfectae neque principals' (comparing also Academica 11.99 f °r Cicero's use of genera). I can see few merits in the suggestion of D u h o t (1989), p p . 170-2, that it may be Cicero, not Chrysippus, w h o is responsible for adding principales a n d adiuvantes. Cf. L S 551, a n d commentary; S.E. PH in. 15. Cf. L S 551 5, a n d commentary, 1.342. L S 551 2, echoing Zeno's original very strict notion of cause (LS 55A 1-3). Although it seems not to have been to most scholars writing about this text. W h a t follows is a partial retraction, or modification, of remarks at L S 1.341-2.
324
DAVID SEDLEY
now expect principales in class (a) to differ fromperfectae, it is hard to resist the view that causa principales are salient causes — those which bear a major share of the responsibility for the effect without however being sufficient to bring it about unaided. 37 (Indeed, unless something fills that role, there will be a massive gap in the Stoic classification of causes.)38 Returning to de Fato 9, we can now interpret Chrysippus' thesis that material influences on human psychology are causae principales. The atmospheric conditions which fostered Socrates' intellectual powers did not by themselves determine that he should willingly drink the hemlock. Nor however were they mere 'procatarctic' or 'auxiliary' causes. They played a leading causal role. This, however, should not be taken to mean that such a causal factor, even taken individually, is causa principalis of your every action. If your native environment were the salient cause of your present behaviour, that would exclude such other crucial factors as your moral education from a comparably important role. It is hard to imagine any Stoic countenancing that. Rather, we had better suppose Chrysippus' position to be that it is the entire set of these external or antecedent causal influences, including congenital character, stars, environment and upbringing, that jointly constitute the causae principales oi your every action. Most discussions of the two classes of cause at Fat. 41-3 equate the perfectae et principales with 'internal' causes, although this is not explicit in the text. It must now be clear that such a characterization is misleading, since atmospheric influences on character, although undeniably external, still count among the causae principales of action 37
38
I t is h a r d e r to decide w h a t Greek term principales translates. Michael F r e d e (1980) suggests e.g. KupicbTccTa. D o r o t h e a F r e d e h a s urged o n m e t h e merits of TrporiyoCueva, which certainly seems to fit the use of the term in Alexander, Fat. 173.14 (chance events lack a TTporiyounevn ocmcc: see further, Sharpies (1983), p p . 132-3), a n d which is a p p a r e n t l y a t h o m e in C h r y s i p p e a n causal terminology (SVF11.912). But if so, two other uses of the same term must be distinguished - 'antecedent cause', common in Galen, a n d 'antecedent internal cause', attributed to the Pneumatic doctor Athenaeus: see Hankinson (1987). T h e equation of principales with irporiyoOuEva actually goes back at least to Y o n (1944), xxix n. 1, although h e takes a different view of w h a t a irporiyoOuevov a r n o v is. Principalis certainly makes a very natural translation of irporiyoOuevos, combining its senses 'primary' and 'leading'; cf. Seneca's translation of fjyeuoviKOV as principale. See also Sharpies (1975), P- 49If even the agent's entire moral character falls short of being the 'complete' cause of a n action, how c a n there be a n y role for complete causes in Stoic theory of action sufficient to justify Chrysippus' inclusion of them in his classification? A plausible answer (for which I thank A n n a M a r i a Ioppolo) is that assent, itself the inevitable product of character plus impression, is in turn the complete cause of the ensuing action.
Chrysippus on psychophysical causality
325
for Chrysippus. By seeing this, we can avoid treating the Stoics as if they isolated internal character as an autonomous entity and simply overlooked the fact that on their theories character will itself be causally predetermined by factors external or antecedent to itself. Rather, the Chrysippean theory takes it as obvious from the start, in true Stoic spirit, that the self which is causa principalis of your every action is not a hermetically sealed intellect pitted against the external world, but you in the broadest sense, incorporating your entire genetic and environmental background. PSYGHOPHYSIGS
The de Fato has helped us to tie down the teleological role of material factors in Stoic psychology. But that leaves us in the dark as to just how material and mental factors interact causally. It is to this that we must now turn. To start with the historical background, it might be thought that here at least we have an issue on which Stoicism radically rejected the tenets of Platonic psychology. After all, Platonic minds are incorporeal - leaving it desperately unclear how they interact with bodies - while Stoic minds are emphatically corporeal. But I doubt, at least, whether the Stoics saw it that way. The first manifestation of their corporealism was in the doctrine of the twin physical principles, 'matter' and 'god': god is a corporeal cause, present throughout matter and causally acting upon it.39 And far from being an outright rejection of Platonism, this was almost certainly seen as a development or interpretation of it. By Zeno's day, there was a current reading of the Timaeus, accepted by Theophrastus, 40 according to which Plato posits two principles, (a) 'matter', and (b) 'a moving cause which he connects with the power of god'. Although the origins of this interpretation remain obscure (at least to me), 41 it seems clear that the Stoic corporealization of the active cause was, relatively speaking, a minor further development of it, rather than a rejection. The Stoics' own grounds for identifying all existing things with bodies turn largely on the supposed dependence of causal properties on corporeality. And even here the prin39 41
40 LS 44-5. Physic, opin. fr. 9, Diels, Dox. pp. 484-5. I assume that the identification of god with an immanent causal principle is inferred at least in part from 77m. 34D3-9.
326
DAVID SEDLEY
ciple has a well recognized origin in Plato's debate between the materialists and the idealists in the Sophist.42 The corporealization of the soul is founded on this same thesis about causality.43 Soul is a specially attuned portion of pneuma44 pervading the passive elements of the body (earth and water). This identification has no obvious Platonic background, but reflects one strand of current medical thought in Zeno's own day. Significantly, it seems to have had the backing of Zeno's contemporary the doctor Praxagoras,45 whose canonical standing for the Stoics is attested by Chrysippus' later appeal to his support regarding the location of the soul's commanding-faculty in the heart. 46 Thus we are at liberty to read the Stoic adoption of the pneuma theory not as a rejection of Platonic psychology, but as an attempt to update it in the light of the latest science. This is borne out by the fact that the Stoics' pneumatic soul is capable of very much the same discarnate survival as Plato had defended in the Phaedo. Indeed, the inherent cohesiveness of pneuma in Stoic physical theory provides them with a new theoretical underpinning for Socrates' confidence, in the Phaedo, that the soul can survive death intact and not disperse like a puff of wind.47 Leaving aside many issues associated with this doctrine, I shall turn directly to its causal implications. As pneuma, the soul stands as a special kind of 'quality' to the material 'substrate' (hupokeimenon) or 'substance' (ousia)48 This gives rise to a temptation to look for an Aristotelian-type distinction between the roles of mind and matter. For example, at de Anima 1.1 Aristotle observes that both material and formal analyses may be available for one and the same psychological event, e.g. anger either as a boiling of the blood around the heart, or as a desire for revenge.49 And one can then go on to reconstruct Aristotle's position on the precise explanatory relationship that obtains between the material and the formal aspect - e.g. 42 43 44
45 46 47 48
Esp. Soph. 246a-b; see further, L S 45, a n d Brunschwig (1988), p p . 64ff. Esp. Cleanthes, SFF1.518 = LS 45c. I a m persuaded by Sorabji (1987), p p . 8 5 - 9 that Stoic pneuma need not be a n a m a l g a m of fire and air b u t can be either one of them individually. But I remain convinced that psychic pneuma, at least, is always a n a m a l g a m of the two, both w a r m t h and breath being integral to animate life (cf. SKF1.135,11.786—7; Cic. Tusc. 1.42). See Harris (1983), ch. 3; S a n d b a c h (1985), p . 47; for similar ideas in Hippocratic texts, see Flat. 15, Nat.puer. 17. Galen, PHP i.y.i. S.E. M i x . 7 1 - 2 . Cf. Plato, Phd. 70a, 77d-8oc, a n d below p . 328 on the soul's 'strength'. 49 LS 28F. Aristotle, de Anima
Chrysippus on psychophysical causality
327
that the material change does not in any way necessitate the formal change, but is hypothetically necessary for it. But any attempt to perform a similar dissection between quality and matter in Stoic psychology seems to get nowhere. It is perhaps only with regard to the problem of individual identity over time that the Stoics exploited such a distinction. (Chrysippus insisted that identity is preserved despite changes in an individual's constituent matter, so long as there is the right kind of qualitative persistence, located in the lifelong 'peculiar quality'.) 50 When it comes to such topics as behavioural explanation, the pneuma carries the role of causal agent with respect to a purely passive substrate,51 the latter being simply that in which all the changes are effected. Even a primitive physical change, such as heating or cooling, is just as much the action of the qualitative pneuma on or in the substrate as is a fully 'psychological' change like impulse or assent.52 Thus in Stoic psychology we should not be looking for any kind of dualism between material and intentionalist description. All psychological changes alike will be pneumatic changes, and none material. Since pneuma is in its nature an intelligent causal agent, there is in principle no puzzle about where the psychic pneuma's distinctively mental properties come from. If anything, we are owed an explanation of how pneuma can come to lack those properties when it characterizes plants (as phusis) and stones (as hexis).53 There is, of course, a physical difference between psychic pneuma and the other two pneumatic states, one regularly described in terms of its greater tension, warmth, dryness etc. But this carries with it a further difference. Mental states are analysable partly in terms of the lekta which are correlated with them. A 'rational' impression, for example, although itself a physical modification of the pneuma (see below), is rendered significant by its propositional content.54 A desire, likewise, must be directed at something; and that something is the predicate which one wishes to become true of oneself. These 50 52
53 54
51 L S 28; amplified in Sedley (1982). L S 47, 55F-G. T o w h a t extent bodily changes are regarded as the work of £§is a n d <|>Ocris, a n d to w h a t extent the functions of these a r e subsumed u n d e r yvxT), is debatable (see Long (1982b)). But the answer should not affect the general point I a m making here a b o u t causality. LS 47. I leave aside the difficult question whether the same applies in some indirect w a y to the impressions of non-rational animals. Cf. the p a p e r of Labarriere in this volume, a n d Sorabji (1990).
328
DAVID SEDLEY
propositions, predicates etc. are lekta, which thus have a crucial role in marking the difference between a mere physical state on the one hand and a significant thought on the other. Is this the ontological basis for some kind of psychological dualism? Against any such possibility, it should be stressed that, unlike mental states, these lekta are incorporeal and hence causally inert. Although the structure and content of thoughts are analysed by mapping them onto lekta, the lekta themselves can have no active role in the structuring of the thoughts,55 let alone be thoughts. And such conformity with lekta is not even an exclusive hallmark of mental entities: it is equally a property of sentences.56 The lekton is as important a component of Stoic psychology as it is of Stoic theory of language, but it does not introduce any kind of dualism of the mental and the physical. This non-dualism is, I think, further borne out by much of the terminology of Stoic psychology. To start with a simple example, a soul's 'strength5 is its moral health; but that very same strength can be cited as the basis of its durability, which enables it to survive outside the body.57 This kind of usage defies any attempt to drive any sort of wedge between psychological and physical analysis. It is instructive to see Galen 58 trying hard to do just that. He quotes Chrysippus as regularly attributing moral failures of resolve to the soul's 'weakness' or 'slackness', defects which make it too readily influenced by external threats and temptations. These conditions, Galen argues, must surely be irrational ones, separate from judgment. But it is striking that, despite extensive verbatim quotation from Chrysippus in a bid to prove the point, he comes up with no such assertion of dualism. And our evidence overwhelmingly suggests that Chrysippus would have equated 'weakness' and 'slackness' with imperfect intellectual states. For example, the list of the nontechnical Stoic virtues juxtaposes the definition of the soul's 'health' as 'a good blending of the doctrines in the soul', with that of the soul's 'strength' as 'sufficient tension in judging and in acting and refraining from acting'. 'Tension' and 'blending' are physical properties of the pneuma, but the definitions are framed with the expectation that no 55 57 58
56 L S 27E, 45. Cf. also t h e role of lekta in t h e analysis of causation, L S 55. SVF 1.563,11.810. PHP iv.6. On the background to this insensitivity, on Galen's part, to the niceties of Stoic linguistic usage, see Manuli (1986).
Chrysippus on psychophysical causality
329
category mistake will be involved in applying these terms directly to the soul's intellectual properties themselves.59 The same applies, in Stoic discussions of the passions, to such key terms as eparsis (swelling), sustole (contraction), tapeinosis (lowering), diachusis (melting), ptoia (fluttering), epitasis (intensification), and dexis (biting). Although some of these nouns, or at least the corresponding verbs (e.g. epairesthai, diacheisthai), had an established metaphorical use as terms for psychological states, the Stoics' predilection for them undoubtedly reflects their physical analysis of pneuma as conditioned by its varying 'tension'. To Galen's ear, at least, they clearly designated irrational physical events.60 But for the Stoics themselves, it is not that each is a purely physical term, matched by some purely psychological counterpart. They are themselves integral to the language in which psychological analysis of passions is carried out, and sit comfortably enough among unambiguously intentionalist terms, without apparent discrimination. For example:61 Pleasure is an unreasonable swelling [eparsis] at something that seems to be choiceworthy. Under it are ranked enchantment [kelesis], gloating [epichairekakia], delight [terpsis], and melting [diachusis].
Here the two emphasized terms, both of which embody the notion of pneumatic tension, are used to designate the genus and one of the four species, without any suggestion that we have crossed a boundary between two alternative modes of analysis. Another area of Stoic psychology in which it is hard, and probably mistaken, to separate physical from intentionalist language is the analysis of phantasia. Zeno's definition of this as a tuposis led Cleanthes, the most physicalist of the Stoics, to interpret it as a literal printing, as in wax62 - which does, if nothing else, confirm the term's origin in Plato's celebrated image at Theaetetus i()iaff. of the soul as a wax tablet suitable for receiving imprints. 63 But Chrysippus 59
60 62 63
Stobaeus n.62.20-6 = SVF 11.278, cf. 279. See Galen, PHP v.2.2off. for the Chrysippean origin of this material. Chrysippus, as cited by Galen, speaks of a n 'analogy' between bodily a n d mental states, b u t without a n y indication that 'tension' etc. a r e only used metaphorically in the case of the mental. 61 Galen, P//P1V.3.2, v.1.4 = SVF 1.209. Diog. Laer. VII.I 14. Gf. Cic. Tusc. iv.66-7. S.E. Afvn.227ff. T h e complex influence of the Theaetetus on the d e b a t e between Zeno a n d Arcesilaus about the cognitive impression is the subject of Ioppolo (1990), which I h a d the good fortune to
read in draft.
33O
DAVID SEDLEY
rejected this in favour of a metaphorical 64 reading of tuposis as an 'alteration' (heteroiosis) of the soul, a term which again eludes any straightforward differentiation between the physical and the mental. 65 A heteroiosis is just a shift from one quality to another, and 'quality', as we have seen, covers all pneumatic states. The language of tuposis survived this debate and remained integral to Stoic psychology, for example in the standard definition of the cognitive impression as 'moulded and impressed' in accordance with its object. One likely instance of it as a concept which importantly spans, or rather unites, the physical and the mental is in the production of speech from the hegemonikon. Galen, in book n of On the doctrines of Hippocrates and Plato (PHP), has a long tirade
against Zeno's notorious syllogism about the location of the hegemonikon. Zeno argued as follows: Voice comes [chorei] through the windpipe. If it were coming from the brain, it would not come through the windpipe. Where speech [logos] comes from is where voice also comes from. But speech comes from the mind. Therefore the mind is not in the brain. (PHP 11.5.8) Galen's critique of this turns largely on the charge of equivocation: 'comes from' has to mean 'comes out of5 in the second premise, 'is produced by' in the fourth. But he recounts how he once met a Stoic who thought he could rescue Zeno by denying that his term 'comes' (chorei) describes locomotion at all (ibid. 11.5.22-3). Clearly this Stoic understood the transmission of speech from the mind in a way which erased any possible distinction between its causal and its local origin. Galen claims that the Stoic, asked what this special sense of chorein could be, was unable to answer him. But Galen's own text contains the materials for reconstructing a suitable answer, when a few pages earlier he dismisses the unattributed view that the voice is produced 'when pneuma in the lungs is imprinted [tupoumenou] in a certain way by the pneuma in the heart, and then imprints in accordance with itself [suntupountos] the pneuma in the windpipe' (PHP 11.4.40). This sounds like a Stoic elaboration, in terms of tuposis, of Zeno's definition66 of voice as 'pneuma stretching from the hegemonikon to the windpipe, the tongue, and the appropriate 64 65
66
Cf. Diog. Laer. vii.46, Chrysippus' objection to Cleanthes' reading (S.E. loc. cit.; Diog. Laer. vii.50), that several imprints cannot coexist in the same wax, is perhaps best interpreted as exposing the excessive crudity of so narrowly physicalist a n analysis of a mental event. SVF 1.150.
Chrysippus on psychophysical causality
331
organs'. Nothing actually travels from the hegemonikon to the outer air, but an 'imprint' is passed along from one portion ofpneuma to the next. So far the description seems purely physical. But there is one link in the chain still to supply. Diogenes of Babylon, amplifying Zeno's syllogism with the help of Plato's Theaetetus,67 suggested that speech itself is 'impressed [ensesemasmenon — or 'made significant'?] and, as it were, imprinted [ektetupomenon], by the conceptions [ennoiai] in the mind' [PHP 11.5.12). There is nothing absurd about this: although conceptions are the very stuff of rationality,68 they are also themselves a kind of phantasia, and hence a kind oftuposis.69 But it seems clear that the smooth chain of transmission from mental impression, perhaps via internal speech,70 to an eventual vocal imprint on the external air requires us to obey the spirit of Stoic psychophysics and to refuse to impose any division between physical and mental imprinting. From the first stage to the last, the tuposis is a rational articulation of the pneuma. And we can glimpse here a theoretical foundation of the extraordinarily close link forged by the Stoics between rationality and speech.71 Returning one last time to de Fato 7-9, in view of what we have seen to be an uninterrupted transmission of imprints from the rational mind to the outer air, it should by now be quite clear that Chrysippus would not feel any special difficulty about what is in effect the reverse process - how such physical factors as the local atmosphere can causally act upon a person's psychological condition. A psychological condition is a pneumatic state. And what could have a closer bearing on your pneumatic state than the air you breathe? 72 67
68 70 71 72
Diogenes' words closely echo the definition of speech (Aoyos) at Tht. 2o6d as 'making one's own thought manifest through the voice, with descriptions a n d names, imprinting (EKTU7ToO|iEVov) one's belief into the stream that flows through the m o u t h as if into a mirror or water'. 69 LS 39E, 53V. Plutarch, de Communibus Notitiis 1084F. See S.E. M v m . 2 7 5 = L S 53T - a n o t h e r legacy of the Theaetetus (189c). See L o n g (1982b), p . 5 1 . F o r t h e c o n t r i b u t i o n of b r e a t h i n g to t h e psychic pneuma, cf. SVF 11.782 (for whose Stoic p r o v e n a n c e see L S vol. 2, note o n 53E), 783, 792 (p. 219, 43), 879 a n d 885 (p. 238, 3 3 ) . Native atmosphere is an especially important determinant of character, because, as J. Mansfeld has pointed out to me, it is the pneuma's 'cooling' (yO^is, from which Stoic etymology derives yuxr)) at the moment of birth that makes it a soul (SVF 11.804-8, Philo, Somn. 1.32, Iamblichus, de Anima ap. Stob. 1.366.15-16).
Bibliography
We are grateful to Steven Rutledge for his help in preparing this bibliography ANCIENT SOURCES Alexanderson, B. (1967). Galeni de Crisibus, Studia Graeca et Latina Gothoburgensia 23 Annas, J. & Barnes, J. (1985). The Modes of Scepticism: Ancient Texts and Modern Interpretations, Cambridge Arnim, J. von (ed.), (1964). Stoicorum Veterum Fragmenta, 4 vols., Leipzig, 1903-24, reprinted Stuttgart Arrighetti, G. (1973). Epicuro, Opere, 2nd edn., Turin Babbitt, F. C. (ed.), (1927). Plutarch's Moralia, vol. 1, Cambridge, MA Bailey, C. (1947). Titi Lucreti Cari De rerum natura libri sex, Oxford Bayer, K. (1963). Cicero, de Fato, Munich Bollack, J. (1975). La pensee du plaisir. Epicure: Textes Moraux, Commentaires, Paris Bruns, I. (ed.), (1892). Alexander of Aphrodisias De Fato, Supplementum Aristotelicum 2.2. Berlin De Lacy, P. H. (ed.), (1978/84). Galen De Placitis Hippocratis et Platonis, 3 vols., Berlin Diano, C. (1946). Epicuri Ethica, Florence Diels, H. (1915). Philodemus: Ueber die Goetter 1. Buch. Abh. Preuss. Akad. Wiss., phil.-hist. Kl. 7 Edelstein, L. & Kidd, I. G. (eds.), (1989). Posidonius / The Fragments, 2nd edn., Cambridge Einarson, B. & De Lacy, P. (eds.), (1967). Plutarch: Reply to Colotes, Loeb Classical Library, Cambridge, MA Giannantoni, G. (1983/5). Socraticorum Reliquiae. Collegit, disposuit, apparatibus notisque instruxit G. Giannantoni, 4 vols., Naples Giussani, C. (1896/8). Lucrezio: De rerum natura, Turin Giusta, M. (1964/7). I Dossografi di Etica, 2 vols., Turin Halm, C. (1889). Aesop, Fables, Leipzig Helmreich, G. (1893). Galeni Scripta Minora, vol. 3. Leipzig (1907/9). Galeni de Usu Partium, 2 vols., Leipzig Jones, H. L. (ed.), (1968/83). Strabo Geography, 8 vols., Cambridge, MA Krevelen, D. A. van (ed.), (1939). Philodemus - De Muziek, Hilversum 332
Bibliography
333
Kuhn, C. G. (ed.), (1821/33). Galeni Opera Omnia. Leipzig, 20 vols. in 22, reissued in facsimile, Hildesheim 1965 Lachmann, C. (1860/66). Lucretius, de Rerum Natura, 3rd edn Long, A. A. & Sedley, D. N. (1987). The Hellenistic Philosophers, 2 vols. Cambridge Long, H. S. (ed.), (1964). Diogenes Laertius Lives, 2 vols., Oxford Longo Auricchio, F. (ed.), (1988). Ermarco, Frammenti, Naples Mannebach, E. (1961). Aristippi et Cyrenaicorum Fragmenta, Leyden, Cologne Marquardt, J. (1884). Galeni Scripta Minora, vol. 1, Leipzig Muller, I. (1894). Galeni de Placitis Hippocratis et Platonis, Leipzig (1881). Galeni Scripta Minora, vol. 2, Leipzig Neubecker, J. (ed.), (1986). Philodemus Vber die Musik iv. Buch. Naples Nutton, V. (1979). Galeni de Praenotione (CMG v 8.1). Berlin Oldfather, W. A. (ed.), (1925). Epictetus The Discourses as Reported by Arrian, The Manual, and Fragments, 2 vols., Cambridge, MA Rackham, H. (1942). Cicero, De Oratore, book m, de Fato, London, Cambridge MA (Loeb Classical Library) Reynolds, L. D. (ed.), (1965). Seneca AdLucilium Epistulae Morales, 2 vols., Oxford (ed.), (1977). Seneca Dialogi, Oxford Sharpies, R. W. (1983). Alexander of Aphrodisias on Fate, London van Straaten, M. (1946). Panetius, sa vie, ses ecrits, et sa doctrine, Amsterdam Westerink, L. G. (ed.), (1962). Anonymous Prolegomena to Platonic Philosophy, Amsterdam Wilke, K. (1914). Philodemus, de Ira, Leipzig Wyttenbach, D. A. (1785/1830). Plutarch, Moralia Yon, A. (1944). Ciceron, Traite du destin, Paris SECONDARY SOURCES Annas, J. (1981). An Introduction to Plato's Republic, Oxford (1992). Hellenistic Philosophy of Mind, California Antoniades, E. (1916). Aristipp und die Kyrenaiker, Diss. Gottingen Arthur, E. P. (1983). 'The Stoic analysis of mind's reactions to presentations', Hermes 111, pp. 69-78 Asmis, E. (1984). Epicurus3 Scientific Method, Ithaca Aune, B. (1970). Kant's Theory of Morals, Princeton, NJ Ballester, L. G. (1988). 'Soul and body, disease of the soul and disease of the body in Galen's medical thought', in Le Opere Psicologiche de Galeno ed. P. Manuli & M. Vegetti, Naples Barnes, J. (1979). The Presocratic Philosophers, 2 vols., London (1982). 'The beliefs of a Pyrrhonist', Proceedings of the Cambridge Philological Society N.s. 28, pp. 1-29 (1985a). 'Cicero's Defato and a Greek source', in Histoire et structure: a la memoire de Victor Goldschmidt, ed. J. Brunschwig et al., Paris
334
BIBLIOGRAPHY
(1985b). 'Uma terceira especie de silogismo: Galeno e la logica das relacoes', Andlise 2 pp. 35-61 (forthcoming). 'Galen on logic and therapy', Proceedings of the ig82 Kiel Galen Conference
Barnes, J., Brunschwig, J., Burnyeat, M. & Schofield, M. (eds.), (1982). Science and Speculation, Cambridge Barnes, J. & Mignucci, M. (eds.), (1988). Matter and Metaphysics, Naples Bellincioni, M. (1978). Educazione alia Sapientia in Seneca, Brescia Blackburn, S. (1981) 'Rule-following and moral realism', (reply to McDowell 1979) in Wittgenstein: to Follow a Rule, ed. S. Holtzmann & C. M. Leich. London Brehier, E. (1925). Les idees philosophiques et religieuses de Philon d'Alexandrie, Paris Brunschwig, J. (ed.), (1978). Les Stoiciens et leur logique, Paris (1986). 'The cradle argument in Epicureanism and Stoicism', in The Norms of Nature, ed. M. Schofield & G. Striker, Cambridge (1988). 'La theorie stoicienne du genre supreme', in Matter and Metaphysics, ed. J. Barnes & M. Mignucci. Naples Burnyeat, M. F. (1980). 'Aristotle on learning to be good', in Essays on Aristotle's Ethics, ed. A. Rorty. Berkeley (1982). 'The origins of non-deductive inference', in Science and Speculation, ed. J. Barnes, J. Brunschwig, M. Burnyeat, & M. Schofield, Cambridge Caizzi, F. D. (1980). 'Tuphos, Contributo alia storia di un concetto', Sandalion 3 pp. 53-66 (1986). 'Pirroniani ed Accademici nel iii secolo a.c' in Aspects de la Philosophie Hellenistique, ed. H. Flashar & O. Gigon, Vandoeuvres-Geneva Cancik, H. (1967). Untersuchungen zu Senecas Epistulae Morales, Hildesheim Classen, C. J. (1963). 'Compte rendu de E. Mannebach (1961)', Gnomon 35, PP- I35-4I Cooper, J. (1989). 'Some remarks on Aristotle's moral psychology', in Aristotle's Ethics (Spindel Conference 1988) ed. T. Roche. The Southern Journal of Philosophy 28, pp. 25-42 Couloubaritsis, L. (1986). 'La psychologie chez Chrysippe', in Aspects de la philosophie Hellenistique, Entretiens Hardt vol. 32 Dancy, J. (1983). 'Ethical particularism and morally relevant properties', Mind 92 pp. 530-47 Daremberg, C. (1865). La Medecine. Histoire et Doctrines, Paris Davidson, D. (1980). 'Freedom to act', in Essays on Action and Events, ed. D. Davidson. Oxford De Lacy, P. H. (1956). 'A new "Fragment" of Antiochus', AJP 77, p. 74 (1958a). 'Stoic views of poetry', AJP 69, pp. 241-71 (1958b). 'ou jidXAov and the antecedents of ancient scepticism', Phronesis 3>PP- 59-71 Dennett, D. C. (1973). 'Mechanism and responsibility', in Essays on
Bibliography
335
Freedom of Action, ed. T. Honderich, London; reprinted in Free Will, ed. G. Watson, Oxford, 1982 Dierauer, U. (1977). Tier und Mensch im Denken der Antike, Amsterdam Dihle, A. (1973). 'Posidonius' system of moral philosophy', Journal of Hellenic Studies 93, pp. 50-7 Dillon, J. M. & A. A. Long (eds.), (1988). The Question of 'Eclecticism3: Studies in Later Greek Philosophy, Berkeley, Los Angeles, London Donagan, A. (1977). The Theory of Morality, Chicago Donini, P. L. (1975). 'Fato e volunta umana in Crisippo', Atti dell'Accademia delle Scienze di Torino, 109, pp. 187-230 (1982). Le scuole, Vanima, Vimpero: Lafilosofiaantica da Antioco a Plotino, Turin (1988a). 'The history of the concept of eclecticism' in The Question of c Eclecticism3: Studies in Later Greek Philosophy, ed. J. M. Dillon & A. A. Long. Berkeley, Los Angeles, London (1988b). 'La tipologia degli errori e loro correzione secondo Galeno', in Le Opere Psicologiche di Galeno, ed. P. Manuli & M. Vegetti, Naples Doring, K. (1988). Der Sokratesschiiler Aristipp und die Kyrenaiker, WiesbadenStuttgart ( = Abh. D. Akad. d. Wiss. und d. Lit. Mainz, 1988/1) Duhot, J. J. (1989). La conception stoicienne de la causalite, Paris Everson, S. (1990). 'Epicurus on the truth of the senses', in Epistemology, ed. S. Everson. Cambridge (ed.), (forthcoming). Philosophy of Language, Cambridge Fillion-Lahille, J. (1984). Le De ira de Seneque et la philosophic stoicienne des passions, Paris Fodor, J. (1975). The Language of Thought, Harvard Frankfurt, H. (1971). 'Freedom of the will and the concept of a person', Journal of Philosophy 67 pp. 5-20 Frede, M. (1980). 'The original notion of cause', in Doubt and Dogmatism. ed. M. Schofield, M. Burnyeat & J. Barnes. Oxford (1981). 'Galen's epistemology', in Galen: Problems and Prospects ed. V. Nutton, London (1983). 'Stoics and Skeptics on clear and distinct impressions', in The Skeptical Tradition, ed. M. Burnyeat, Berkeley (1986). 'The Stoic doctrine of the affections of the soul', in The Norms of Nature, ed. M. Schofield & G. Striker, Cambridge (1987). 'The Skeptic's two kinds of assent and the question of the possibility of knowledge', in Essays in Ancient Philosophy, Minneapolis Fritz, K. von (1934). 'Theodorus' (32), Realenzyklopddie, VA2, col. 1825-31 Furley, D. (1971). 'Knowledge of atoms and void in Epicureanism', in Essays in Ancient Greek Philosophy, ed. J. P. Anton & G. L. Kustas. New York (1987). The Greek Cosmologists. Vol. 1: The Formation of the Atomic Theory and its Earliest Critics, Princeton Ganss, W. (1952). Das Bild des Weisen bei Seneca, Fribourg
336
BIBLIOGRAPHY
Giannantoni, G. (1958). / Cirenaici. Raccolta delle fonti antiche, traduzione e studio introduttivo, Florence Gigon, O. (1955). 'Die Erneuerung der Philosophic in der Zeit Ciceros', Ent. sur VAnt. class. 3, pp. 25-79 Gill, C. (1983). 'Did Chrysippus understand Medea?' Phronesis 28 1, pp. 136-49 Giusta, M. (1962/3). 'Passi dossografici di morale epicurea nel x libro di Diogene Laerzio', Atti dell1 Accademia delle Scienze di Torino 97, pp. 120-74 Glucker, J. (1978). Antiochus and the Late Academy, Gottingen Goldschmidt, V. (1977). La Doctrine d3Epicure et le Droit. Paris Gorier, W. (i977)."Acr6evf)s luyKOCT&OEcris Zur stoischen Erkenntnistheorie", WJA N.S. 3, pp. 83-92 (1987). ' "Hauptursachen" bei Chrysipp und Cicero', Rk.M. 130, pp. 254-74 Gosling, J. C. B. (1987). 'The Stoics and akrasia\ Apeiron 20, pp. 179-202 Gosling, J. C. B. & Taylor, C. C. W. (1982). The Greeks on Pleasure, Oxford Gowans, C. (1987). Moral Dilemmas, Oxford Griffin, M. (1986). 'Philosophy, Cato and Roman suicide', G&R 33, pp. 64-77, 192-202 Grilli, A. (1959). 'Compte rendu de G. Giannantoni (1958)', Rivista Critica di Storia della Filosofia 14, pp. 343-51 Hadot, I. (1969). Seneca und die griechisch-romische Tradition der Seelenleitung, Berlin Halliwell, S. (1986). Aristotle's Poetics, London (1992). 'Pleasure, emotion, and understanding in Aristotle's Poetics', in Essays on Aristotle's Poetics, ed. A. Rorty, Princeton Hankinson, R. J. (1987). 'Evidence, externality and antecedence: inquiries into later Greek causal concepts', Phronesis 32, pp. 80-100 (1988a). 'Galen explains the elephant' in Philosophy and Biology, ed. M. Matthen & B. Linsky, Canadian Journal of Philosophy Supp. Vol. 14 (1988b). 'Galien: la medecine et la philosophic antisceptique', Revue de Philosophie Ancienne 2 pp. 227-67 (1989). 'Galen and the best of all possible worlds', Classical Quarterly 39 1 (1989) pp. 206-27 (1990a). 'Perception and evaluation: Aristotle on the moral imagination', Dialogue 19, pp. 18-39 (1990b). 'Saying the phenomena', Phronesis 35 3 (1991a). Galen on the Therapeutic Method, Books 1 and 2, Oxford (1991b). 'Galen's anatomy of the soul', Phronesis 36 1 (1991c). 'Galen's account of scientific knowledge', in Galeno: Obra, Pensamientoy Influencia, ed. J. A. Lopez Ferez, Madrid (forthcoming 1). 'Usage and abusage: Galen on language', in Ancient Philosophy of Language: A Philosophical Introduction, ed. S. Everson, Cambridge
Bibliography
337
(forthcoming 2). 'Galen's concept of scientific progress', Aufstieg und Niedergang der Romischen Welt, 11, 37 1 (forthcoming 3). 'Galen and the logic of relations', in Aristotle and Later Antiquity, ed. L. Schrenk (forthcoming 4). Galen on Antecedent Causation, Cambridge (forthcoming 5). 'Galen's theory of causation', Aufstieg und Niedergang der Romischen Welt, 11, 37 2 (ed.) (1988). Method, Medicine, and Metaphysics, Alberta Harkins, P. W. & Riese, W. (1963). Galen on the Passions and Errors of the Soul, Columbus Harris, C. R. S. (1983). The Heart and Vascular System in Ancient Greek Medicine from Alcmaeon to Galen, Oxford Herman, B. (1985). 'The practice of moral judgment', Journal of Philosophy 82, pp. 414-36 Holler, E. (1934). Seneca und die Seelenteilungslehre und Affektpsychologie der Mittelstoa, Kallmiinz Hossenfelder, M. (1986). 'Epicurus - hedonist malgre lui', in The Norms of Nature, ed. M. Schofield & G. Striker. Cambridge Hoven, R. (1971). Stoicisme et stoiciensface auprobleme de Vau-dela, Paris Huby, P. (1969). 'The Epicureans, animals and freewill', Apeiron 3 Hutchinson, D. S. (1988). 'Doctrines of the mean in fourth-century medicine, rhetoric and ethics' in Method, Medicine and Metaphysics ed. R. J. Hankinson, Alberta Imbert, C. (1978). 'Theorie de la representation et doctrine logique' in Les Stoiciens et leur logique ed. J. Brunschwig, Paris Indelli, G. (ed.) (1978). Polistrato: Sul Disprezzo Irrazionale delle Opinioni Popolari, Naples Inwood, B. (1982). 'A note on desire in Stoic theory', Dialogue 21, pp. 329-31 (1984). 'Hierocles: theory and argument in the 2nd century AD', in Oxford Studies in Ancient Philosophy, vol. 2 (1985). Ethics and Human Action in Early Stoicism, Oxford (1986). 'Goal and target in Stoicism', Journal of Philosophy 83, pp. 547-56 Ioppolo, A. M. (1980a). 'Carneade e il terzo libro delle Tusculani', Elenchos 1, pp. 76-91 (1980b). Aristone di Chio e lo stoicismo antico, Naples (1984). 'L'astrologia nello stoicismo antico', in La scienza ellenistica, ed. G. Giannantoni & M. Vegetti. Naples (1986). Opinione e scienza, Naples (1987). 'II monismo psicologico degli Stoici antici', Elenchos 8, pp. 449-66 (1988). 'Le cause antecedenti in Cic. "de Fato" 40', in Matter and Metaphysics, ed. J. Barnes & M. Mignucci. Naples (1990). 'Presentation and assent: a physical and cognitive problem in Stoicism', CQ40
338
BIBLIOGRAPHY
Irwin, T. H. (1980). 'Reason and responsibility in Aristotle', in Essays on Aristotle's Ethics•, ed. A. Rorty. California (1988). Aristotle's First Principles, Oxford Kahn, C. H. (1983). 'Arius as a doxographer' in On Stoic and Peripatetic Ethics: The Work of Arius Didymus, ed. W. W. Fortenbaugh. New Brunswick, NJ (1988). 'Discovering the will' in The Question of 'Eclecticism': Studies in Later Greek Philosophy, ed. J. M. Dillon & A. A. Long. Berkeley, Los Angeles, London Kenny, A. J. P. (1973). 'Mental health in Plato's Republic', in Kenny The Anatomy of the Soul, Oxford
Kerferd, G. (1978). 'What does the wise man know?', in The Stoics, ed. J. M. Rist. Berkeley Kidd, I. G. (1971). 'Stoic intermediates and the end for man', in Problems in Stoicism ed. A. A. Long, London, first published in CQN.S. 5(1955), pp. 181-94 (1978). 'Moral actions and rules in Stoic ethics', in The Stoics, ed. J. M. Rist. Berkeley Knox, B. (1977). 'The Medea of Euripides', Tale Classical Studies 25 pp. 193-225 Kramer, H. J . (1971). Platonismus und hellenistische Philosophie, Berlin, New
York Kraus, P. (1937). Kitab al-ahlaq li-Galinus, Bulletin of the Faculty of Arts of the University of Egypt 5, Sectio arabica, Cairo Lafrance, Y. (1986). La theorie platonicienne de la doxa, Montreal, London Lain Entralgo, P. (1957). 'La curacion por la palabra en la Antigiiedad clasica', Archivos Iberoamericanos de Historia de la Medicina 9, pp. 199—241
Lakoff, G. & Johnson, M. (1980). Metaphors We Live By, Chicago, London Langer, S. K. (1953). Feeling and Form, New York (1957). Philosophy in a New Key: a Study in the Symbolism of Reason, Rite, and
Art, 3rd edn., Cambridge, MA Laursen, S. (1987). 'Epicurus, On Nature Book xxv', Cronache Ercolanesi 17, pp. 77-8 (1988). 'Epicurus On Nature xxv (Long-Sedley 20 B,G and j)', Cronache Ercolanesi 18, pp. 7-18 Lear, J. (1983). 'Ethics, mathematics and relativism', Mind 92, pp. 38-60 Levy, C. (1978). 'Scepticisme et dogmatisme dans l'Academie: "l'esoterisme" d'Arcesilas', REL 56, pp. 335-48 (1980). 'Opinion et certitude dans la philosophie de Carneade', RBPh 58, pp. 30-46 (1986). '"Le scepticisme" de Philon d'Alexandrie: une influence de la Nouvelle Academie?' in Hellenica et Judaica ed. A. Caquot, M. HadasLebel & J. Riaud, Leuven-Paris (forthcoming). Ciceron et la Nouvelle Academie, Collection de l'Ecole Francaise de Rome
Bibliography
339
Lloyd, G. E. R. (1988). 'Scholarship, authority and argument in Galen's Quod animi mores', in Le Opere Psicologiche di Galeno ed. P. Manuli & M. Vegetti. Naples Long, A. A. (1970). 'The logical basis of Stoic ethics', Proceedings of the Aristotelian Society 71, pp. 85-104 (1971). 'Language and Thought', in Problems in Stoicism ed. A. A. Long, London (1974). Hellenistic Philosophy, 2nd edn. 1986, London (1982a). 'Astrology: arguments pro and contra' in Science and Speculation ed. J. Barnes, J. Brunschwig, M. Burnyeat & M. Schofield. Cambridge (1982b). 'Soul and body in Stoicism', Phronesis 27, pp. 34-57 (1988). 'Socrates in Hellenistic philosophy', CQ38, PP- I 5°~7I (1989). 'Stoic eudaimonism', in Proceedings of the Boston Area Colloquium in Ancient Philosophy 4, ed. J. J. Gleary & D. C. Shartin, pp. 77-101 Lopez Ferez,J. A. (ed.) (forthcoming). Galeno: Obras Pensamientoy Influencia, Madrid Luck, G. (1978). 'On Cicero De fato 5 and related passages', AJP 99, PP- 155-8 McDowell, J. (1979). 'Virtue and reason', Monist 62, pp. 331-50. Mackenzie, M. M. (1981). Plato's Theory of Punishment, California McKim, R. (1984). 'Democritus against scepticism: all sense-impressions are true', in Proceedings of the First International Congress on Democritus, ed. L. G. Benakis, Xanthi Maconi, H. (1988). 'Nova non philosophandi philosophia', OSAPH 6, PP- 230-53 Mansfeld, J. (1988). 'Philosophy in the service of the Scripture', in The Question of'Eclecticism3: Studies in Later Greek Philosophy, ed. J. M. Dillon & A. A. Long, Berkeley, Los Angeles, London Manuli, P. (1986). 'Traducibilita e molteplicita dei linguaggi nel Deplacitis di Galeno', in Storiogrqfia e dossogrqfia nellafilosojiaantica ed. G. Cambiano, Turin (1988). 'La passione nel De Placitis\ in Le Opere Psicologiche de Galeno ed. P. Manuli & M. Vegetti, Naples & Vegetti, M. (eds.), (1988) Le Opere Psicologiche de Galeno, Naples Matthen, M. & Linsky, B. (eds.), (1988). Philosophy and Biology, Canadian Journal of Philosophy supp. vol. 14 Mattock, J. N. (1972). 'A translation of the Arabic epitome of Galen's book Peri ethon\ in Islamic Philosophy and the Classical Tradition eds. S. M. Stern, A. Hourani & V. Brown, Oxford Menghi, M., & Vegetti, M. (1984). Le Passioni e gli Errori del Anima, Venice Mitsis, P. (1986). 'Moral rules and the aims of Stoic ethics: reply to Inwood', Journal of Philosophy 83, pp. 556-7 (1988). Epicurus3 Ethical Theory, Cornell Studies in Classical Philology XLVIII
34°
BIBLIOGRAPHY
Natorp, P. (1894). 'Annikeris', Realenzyclopddie, 12 cols. 2259-61 Nell, O . (1975). Acting on Principle: An Essay on Kantian Ethics, New York Nikiprowetzky, V. (1977). Le commentaire de VEcriture chez Philon d'Alexandrie,
Leiden Nussbaum, M. (1986a). The Fragility of Goodness, Cambridge (1986b). 'Therapeutic arguments: Epicurus and Aristotle', in The Norms of Mature, ed. M. Schofield & G. Striker. Cambridge (1987). 'The Stoics on the extirpation of the passions', Apeiron 20 2 pp. 129-58 (1989). 'Beyond obsession and disgust: Lucretius' genealogy of love', Apeiron 22, pp. 1-59 (1991). 'Serpents in the soul: a reading of Seneca's Medea', in Essays in Honor of Stanley Cavell, ed. T. Cohen, P. Guyer, & H. Putnam, Lubbock, Texas (1992). 'Tragedy and self-sufficiency: Plato and Aristotle on fear and pity', in Essays on Aristotle's Poetics, ed. A. Rorty, Princeton Phillips, D. Z. & Mounce, H. O. (1970). Moral Practices, London Pigeaud, J. (1981). La Maladie de VAme, Paris (1988). 'La psychopathologie de Galien', in Le Opere Psicologiche di Galeno ed. P. Manuli & M. Vegetti. Naples Pohlenz, M. (1970). Die Stoa, 4th edn. 2 vols., Gottingen Putnam, H. (1983). 'Taking rules seriously - a response to Martha Nussbaum', New Literary History 15, pp. 193-200 Radice, R. (1983). Filone di Alessandria, Naples Rawls, J. (1971). A Theory of Justice, Cambridge Reesor, M. E. (1951). 'The "Indifferents" in the Old and Middle Stoa', Transactions and Proceedings of the American Philological Association, 82, pp. 102-10
Riese, W. (1951). 'An outline of a history of ideas in psychotherapy', Bulletin of the History of Medicine 25, pp. 442-56
Rist,J. M. (1969). Stoic Philosophy, Cambridge (1972). Epicurus: An Introduction, Cambridge (ed.), (1978). The Stoics, California (1982). 'Are you a Stoic? The case of Marcus Aurelius', in Self Definition in the Greco-Roman World., ed. B. F. Meyers & E. P. Sanders, Philadelphia (1989). 'Seneca and Stoic orthodoxy', ANRW36.3, pp. 1993-2012 Rodier, G. (1926). 'La coherence de la morale stoicienne', Anneephilosophique repr. in Etudes de philosophie grecque. Paris Rolke, K.-H. (1971). Bildhafte Vergleiche bei den Stoikern, Hildesheim, New York (1975). Die bildhaften Vergleiche in der Fragmente der Stoiker von £enon bis
Panaitios, Hildesheim Rorty, A. O. (ed.), (1980). Essays on Aristotle's Ethics, California Ross, W. D. (1930). The Right and the Good, Oxford Runia, D. (1986). Philo of Alexandria and the Timaeus of Plato, Leiden
Bibliography
341
Sandbach, F. H. (1941). 'Some textual notes on Plutarch's Moralia' pp. 110-18
(1971). 'Phantasia kataleptike' in Problems in Stoicism ed. A. A. Long. London (1985). Aristotle and the Stoics, Cambridge Santurri, E. (1987). Perplexity in the Moral Life: Philosophical and Theological Considerations, Gharlottesville Saunders, T. J. (1983). 'Review of M. M. Mackenzie's Plato's Theory of Punishment', Journal of Hellenic Studies, pp. 173-4 Schofield, M. (1980). 'Preconception, argument and god', in Doubt and Dogmatism: Studies in Hellenistic Epistemology, ed. M. Schofield, M. Burnyeat & J. Barnes. Oxford (1984). 'Ariston of Chios and the unity of virtue', Ancient Philosophy, 4, pp. 83-96 & Striker, G. (eds.) (1986). The Norms of Nature, Cambridge Schrenk, L. (ed.) (forthcoming). Aristotle and Later Antiquity Schwartz, E. (1903). 'Diogenes (Laertios)(4o)', Realenzyclopddie vol. 1, cols. 738-63 Scott, D. (1988). 'Innatism and the Stoa', Proceedings of the Cambridge Philological Society 214, pp. 123-53 Sedley, D. (1973). 'Epicurus, On Nature Book xxvm', Cronache Ercolanesi 3, pp. 14-17, 27-34 (1982). 'The Stoic criterion of identity', Phronesis 27, pp. 255-75; reprinted as 'Le critere d'identite chez les Stoiciens', Revue de Metaphysique et de Morale 4 (1989), pp. 513-33 (1983). 'Epicurus' refutation of determinism', Suzetesis, Studi offerti a Marcello Gigante, pp. 11—51, Naples (1988). 'Epicurean Anti-Reductionism', in Matter and Metaphysics, ed. J. Barnes & M. Mignucci. Naples (1989). 'Teleology and myth in the Phaedo', Proceedings of the Boston Area Colloquium in Ancient Philosophy, 5, pp. 359-83 Sharpies, R. W. (1975). 'Responsibility, chance and not-being (Alexander of Aphrodisias, mantissa 169-72)', BICS 22, pp. 37-64 Sherman, N. (1989). The Fabric of Character, Oxford Sifakis, G. M. (1986). 'Learning from art and pleasure in learning: an interpretation of Aristotle Poetics 4 I448b8-i9', in Studies in Honour of T. B. L. Webster vol. 1, ed. J. H. Betts, J. T. Hooker & J. K. Green. Bristol Sorabji, R. (1980). Necessity, Cause and Blame, London (1987). Matter, Space and Motion, London (1990). 'Perceptual content in the Stoics', Phronesis 35 Staden, H. von (1989). Herophilus: the Art of Medicine in Early Alexandria, Cambridge Stern, S. M., Hourani, A. & Brown, V. (eds.), Islamic Philosophy and the Classical Tradition, Oxford
342
BIBLIOGRAPHY
Stough, C. (1978). 'Stoic determinism and moral responsibility', in The Stoics, ed. J. Rist, Berkeley Stratton, G. M. (1917). Theophrastus and the Greek Physiological Psychology before Aristotle, New York
Strawson, P. F. (1962). 'Freedom and resentment', Proceedings of the British Academy 48, reprinted in Free Will ed. G. Watson, Oxford, 1982 Striker, G. (1977). 'Epicurus on the truth of sense-impressions', Archiv fur Geschichte der Philosophie 59, pp. 125-42
(1980). 'Sceptical strategies', in Doubt and Dogmatism, ed. M. Burnyeat, J. Barnes & M. Schofield, Oxford (1983). 'The role of oikeiosis in Stoic ethics', in Oxford Studies in Ancient Philosophy 1
(1986). 'Antipater on the art of living' in The Norms of Nature, ed. M. Schofield & G. Striker, Cambridge (1987). 'Origins of the concept of natural law', in Proceedings of the Boston Area Colloquium in Ancient Philosophy, 3, ed. J. J. Cleary, Lanham Stroux, L. (1966). 'Vergleich und Metapher in der Lehre des Zenon von Kition', Diss. Heidelberg Tarrant, H. (1985). Scepticism or Platonism? The Philosophy of the Fourth
Academy, Cambridge Taylor, C. C. W. (1980). 'All sensations are true', in Doubt and Dogmatism, ed. M. Schofield, M. Burnyeat, & J. Barnes. Oxford Thompson, G. R. (1941). 'Theophrastus on Plant Flavors and Odors: Studies on the Philosophical and Scientific Significance of De Causis Plantarum vi', Diss. Princeton Trillitzsch, W. (1962). Senecas Beweisfuhrung, Berlin Tynan, K. (1964). 'Bertolt Brecht', in Essays in the Modern Drama, ed. M. Freedman, Boston Usener, H. (1977). Glossarium Epicureum, ed. M. Gigante & W. Schmid, Rome Vander Waerdt, P. A. (1988). 'Hermarchus and the Epicurean Genealogy of Morals', Transactions of the American Philological Association, 118
pp. 87-106 (1989). 'The Stoic Theory of Natural Law', Diss. Princeton Vegetti, M. (1984). 'La terapia dell'anima. Patologia e disciplina del soggetto in Galeno', in Le Passioni e gli Errori del Anima ed. M. Menghi & M. Vegetti, Venice Vlastos, G. (1985). 'Socrates' disavowal of knowledge', PhQ 35, 139 pp.1-31 Voelke, A.-J. (1973). Videe de volonte dans le stoicisme, Paris Walzer, R. (1949). 'New light on Galen's moral philosophy' CQ 43, 1, pp. 82-96 Watson, G. (1971). 'The natural law and Stoicism', in Problems in Stoicism, ed. A. A. Long, London (ed.) (1982). Free Will, Oxford
Bibliography
343
Westman, R. (1955). Plutarch gegen Kolotes. Ada Filosojica Fennica 7, Helsinki White, N. (1985). 'Nature and regularity in Stoic ethics: a discussion of Anna Maria Ioppolo, Aristone de Chio e lo stoicismo antico\ Oxford Studies in Ancient Philosophy, 5, pp. 289-306 Williams, B. (1973). Problems of the Self Cambridge Williamson, C. (1989). 'Following a rule', Philosophy 64 pp. 487-504 Winch, P. (1958). The Idea of a Social Science and its Relation to Philosophy, London Zeller, E. (1923). Die Philosophie der Griechen, 5th edn. Leipzig
Subject index
Abderitians, 47 Academicians, 72, 225-6, 232, 242, 259, see also New Academy; Old Academy action, Stoic theory of, 165-6, see also agency actions, and passions in Galen, 196-222 advice, good, 285-6, 290-1 aesthetic emotional response, Senecan theory of, 177-82 aesthetic pleasure, 37-9 affection, emotion and moral self-management (Galen), 184-222 agency, Epicurean concept, 53-71 allegorical interpretation, Stoic, 133-6 altruism, 29, 35-6, 48-9 anger, 164-83, 196, 202, 206-7 animals: assent in, 240-2, 243-9; consciousness/conscience of, 238-43; reason of 65-9; Stoic idea of fantastic nature, 225-49; w ^ 61-2 Annicerians, 28-30, 34, 35-6, 39-49 anti-Epicureanism, 18-49 anticipation, 27 aponia, 15-16 arbitrary qualities, 75-9 assent, 166, 170, 173-6, 179-81, 292; in animals, 240-2, 243—9 ataraxia, 15-16, 66, 67 atmospheric influences on character, 319-20, 321-5 atomism, 51-94; criticisms of, 72-93 atoms, 55, 57, 58-9, 66; properties of 79—84 behaviourism, 308-9 being and not being (Plato), 93 belief, see doxa beliefs: and agency, 70; corrupt, 297-8, 309 biographical teleology, 314-20 body-mind relationship, see mind-body relationship categorical imperative (Kant), 303
causality: Epicurean 54, 62-3; psychophysical (Chrysippus), 313-331 causation: direction of psychological (Galen), 186-7, 220-2; natural 315-19, 321-5 causes, principal, 315, 321-5 censorship, 100, 131-2 children, compared with animals, 225, 237 choice, and constraints, 64-5 cognitive view of passions (Stoic), 100-4, 121-2
colour, perception of, 76, 83, 87-90 constitution, 55, 58, 60-3; of animals, 225, 231, 238-9 convention, 76—8 corporealism, Stoic, 325-6 Cyrenaics, 3, 4, 5, 9, 17, 18-49, see a^so Annicerians, Hegesians; evolution of beliefs, 30-6; orthodox and dissident, 24-30; versus Epicureans, 18-49 decreta, 291, 293, 299-304, 312 desires: necessary and unnecessary, 10; passions and the structure of life, 204-12 detachment, Stoic critical, 136-45 determinism, and freedom, 64-5, 315-17, 319, 321-5 developments, 55-7, 60-3 Dogmatists (Stoic), 226, 235 doxa: concept in Stoic philosophy, 251-60; mixed, 279—84 dualism, Platonic, 159—60, 161-4, 276; see also psychological dualism education: Posidonian, 109-14; role of poetry in, 100, 102-4, 106-7 egoism, 34-5 eidola theory, 83-4, 91-3 Eleatic philosophy, 93 emotions, status of for Galen, 187-92 Epicurean hedonism, 3-17 Epicurean psychology, 51-94, 268
344
Subject index epiphenomena, 78 errors, 254-5; an<^ passions (Galen), 189-97 ethics, and psychology of hedonism, 1-49 fate, doctrine see Providence, doctrine of forms, theory of, 251, see also literary form freedom, Galen's notion, 212-15 friendship, 35, 36, 43, 48 good, concept of the, 8, 31 good life, desires, passions and the, 204-12 happiness: as goal of our actions (Aristotle), 30-1; hedonist conception, 3, 8-9, 17; and pleasure, Cyrenaic beliefs, 25, 27, 30-6; and reason, 163 hedonism: Epicurean, 3-17; psychology of, J -49 Hegesians, 23, 24, 28-30, 34-6, 48, 49 humour, Stoic, 141-2 idealism, 325—6 imagination, see mental pleasures individuality of moral experience, 291-2 inertia, psychological, 169-70, 176, 179, 182 inheritance, and the environment in character development, 215-16, 321 intellectualism, dangers of Socratic, 287-90, 298 Jewish tradition, 279 judgment, and passion, 105-7,
IO
9
katalepsis, 253-9 katastematic pleasure, 3-4, 5, 14-17, 20-1, 25,27 kinetic pleasure, 3-4, 5, 14-17, 20-1, 27, 32 language, 69, 282, 328-31 'Lazy Argument', 315-16 lekta, 327-8 literary form, 108, 156-61 materialism, 325-6 Medio-Platonists, 274, 275, 282, 283 memory, 27, 60, 68 mental pleasures, 18-49: doctrine of, 36-9 mimetic pleasure, 38-9 mind: Chrysippean philosophy of, 154-9, 168—70, 176-7, 179, 182, 313-31; Posidonian philosophy of, 154-6, 164—5; Seneca's philosophy of, 150-83 mind—body relationship, 325—31, see also pleasures, bodily and mental monism, 159, 182, 260, 276-7, 313, 314 moral conflict, 305—10 moral development, reason and rules, 285-312
345
moral precepts, see praecepta moral principles, see decreta moral self-management, and emotion (Galen), 194-222 moral situations, particularity of, 288—9 morality, role of non-cognitive elements in, 305-10 mousike, 102-4, IJ4> 115~i9> 120 movement, 25, 26, 30, 43 nature: and divine providence, 311-12; Epicurean concept, 57-8; and sense perception, 7-8 Neo-Pyrrhonism, 251, 277, 284 New Academy, 250-1, 260-74, 277, 283 New Comedy, 142 non-cognitive view of passions (Stoic), 100-4, 109-21 objectivity of moral values, 311-12 Old Academy, 275-6 passions, 95-222; and actions in Galen, 196-222; desires and the structure of life, 204-12; and errors (Galen), 189—97; anc ^ poetry in Stoicism, 97-149; Seneca's theory of the, 150-83; therapy of the 198-204, 260 patriotism, 36, 37, 38, 43 perception, see sense perception Peripatetics, 5, 72, 171-2, 191 pessimism, 35-6, 49 philosophy, kinship with poetry for Stoics, 107-8, 122, 125, 126-30, 146-9 pleasure: as absence of pain, 3, 4, 6, 9—10, 11-13, 21, 25; bodily and mental, 25-6, 27-8, 36-9, 40-1, 45; as the criterion of choice and avoidance, 6-10; denial of possibility of, 35, 48; as goal of actions, 20, 22, see also telos; and happiness, Cyrenaic beliefs, 25, 27, 30-6, nature of, 11—17; see also katastematic; kinetic pleasure; and reproduction, 281; as state of someone asleep or dead, 3, 25, 41-2,
43-5
poetry: Aristotle on, 107-8; benefits for Stoics, 126-30; dangers for Stoics, 123-5; and passions in Stoicism, 97-149; role in education, 126-30; Stoic reform of spectatorship, 108, 130, 131-45, 146—9; Stoic rewriting and new, 132—3 poietike, 102-4 praecepta, 291, 293-9, 300, 304, 312 pregnancy, 112-13 Providence, doctrine of, 226, 230, 232, 244, 246, 249, 315-19, 321-5
Name index
346
psychological causation, direction of (Galen), 186-7, 220-2 psychological concepts, Stoic, 223-331 psychological dualism, 150-83, 328 psychology: Galen's philosophical, 184-222; Platonic, 313-14, 316-20, 325-6, 331 psychology of hedonism, ethics and, 1-49 psychophysics, 325-31 punishment, 217-20 Quellenforschung, 153-4, i82n rationality, human, 67-71 reason: of animals, 67-9; in the heart (Chrysippus), 115; and the passions, 166-83; a n d responsibility (Galen), 194-6, 212-20; rules and moral development, 285-312 representation, 25, 38—9 reproduction, and pleasure, 281 responsibility, 54; and causation, 322-4; and reason (Galen), 194-6, 212-20 rhythms of the soul, 109-14 rules: hierarchy of, 295-6; moral and game, 311; reason and moral development, 285-3* 2 scepticism, 231, 251, 270-4, 283, 291-2; about rules, 311-12 self-control, 194-222, 280-1 Senecan drama, 148 sensations, truth of, 82-4, 90-3
sense perception: nature and, 7-8; value difference between Epicurus and Democritus, 72-94 smell, 86 soul: dunamis doxastike of, 252-60; and joy, 37-8; nature of the (New Academy), 260-74; pneumatic, 326-31; poetry and health of, 97-8; rhythms of the 109-14; Seneca on the, 150-83; tripartite (Plato), 104-7, I 0 9~ r l-> 3*3 spectatorship, critical Stoic, 130, 136-49 speech, and rationality, 330-1 Stoic dogma, according to Philo of Alexandria, 250-84 Stoic poetry, paradox of, 98-101 Stoic psychological concepts, 223-331 Stoicism, Seneca and orthodox, 150-3 Stoics, 4-5, 8; cognitive and non-cognitive views of the passions, 100-22; poetry and the passions 97-149 substratum, and quality (Aristotle), 93 suicide, 316-20 taste, ontological status of, 77, 78, 80, 84-6, 9^92-3 teleology, biographical, 314-20 telos, 5-6, 9-10, 29, 31, 47-8 theatre, dramatic and epic, 144-5 time, and happiness, 32-3 touch, 91-3 tragedy, 108, 128-30, 136-7
Name index (Names of philosophical schools can be found in the subject index) Aelianus, Claudius, 226n Aenesidemus, 283 Aesop, ig8n Aetius, 88, 235, 237, 237n, 238 Aithiops, 22n, 23n Albinus, i84n, 282 Alexander of Aphrodisias, 7n, 2i3n, 226, 229, 244, 244n, 245-8, 248n Alexander of Damascus, 184n Anaxagoras, 47 Andronicus, 3O7n Annas, J., I95n, 227n, Anniceris, 18-49 Antiochus, 241, 242, 262 Antipater, 22n, 23n, 116, 232,
Antiphon, Antisthenes, 47n Antoniades, E., 19, 23n 'Antonius the Epicurean', 192 Apollodorus of Cyzicus, 47 Apuleius, Lucius, 282 Arcesilaus, i66n, 251, 259, 260, 261, 262, 264, 265, 266-7, 268, 269, 273, 274, 283, Arete, ign, 22n, 23n Aristippus of Cyrene, the Elder, 4, 18, 19, 20, 22, 23, 24, 30, 32n, 35, 40, 48 Aristippus the Younger, 18, 19, 2on, 22n, 32,40 Aristo, 5, 261, 293, 294-6
Name index
346
psychological causation, direction of (Galen), 186-7, 220-2 psychological concepts, Stoic, 223-331 psychological dualism, 150-83, 328 psychology: Galen's philosophical, 184-222; Platonic, 313-14, 316-20, 325-6, 331 psychology of hedonism, ethics and, 1-49 psychophysics, 325-31 punishment, 217-20 Quellenforschung, 153-4, i82n rationality, human, 67-71 reason: of animals, 67-9; in the heart (Chrysippus), 115; and the passions, 166-83; a n d responsibility (Galen), 194-6, 212-20; rules and moral development, 285-312 representation, 25, 38—9 reproduction, and pleasure, 281 responsibility, 54; and causation, 322-4; and reason (Galen), 194-6, 212-20 rhythms of the soul, 109-14 rules: hierarchy of, 295-6; moral and game, 311; reason and moral development, 285-3* 2 scepticism, 231, 251, 270-4, 283, 291-2; about rules, 311-12 self-control, 194-222, 280-1 Senecan drama, 148 sensations, truth of, 82-4, 90-3
sense perception: nature and, 7-8; value difference between Epicurus and Democritus, 72-94 smell, 86 soul: dunamis doxastike of, 252-60; and joy, 37-8; nature of the (New Academy), 260-74; pneumatic, 326-31; poetry and health of, 97-8; rhythms of the 109-14; Seneca on the, 150-83; tripartite (Plato), 104-7, I 0 9~ r l-> 3*3 spectatorship, critical Stoic, 130, 136-49 speech, and rationality, 330-1 Stoic dogma, according to Philo of Alexandria, 250-84 Stoic poetry, paradox of, 98-101 Stoic psychological concepts, 223-331 Stoicism, Seneca and orthodox, 150-3 Stoics, 4-5, 8; cognitive and non-cognitive views of the passions, 100-22; poetry and the passions 97-149 substratum, and quality (Aristotle), 93 suicide, 316-20 taste, ontological status of, 77, 78, 80, 84-6, 9^92-3 teleology, biographical, 314-20 telos, 5-6, 9-10, 29, 31, 47-8 theatre, dramatic and epic, 144-5 time, and happiness, 32-3 touch, 91-3 tragedy, 108, 128-30, 136-7
Name index (Names of philosophical schools can be found in the subject index) Aelianus, Claudius, 226n Aenesidemus, 283 Aesop, ig8n Aetius, 88, 235, 237, 237n, 238 Aithiops, 22n, 23n Albinus, i84n, 282 Alexander of Aphrodisias, 7n, 2i3n, 226, 229, 244, 244n, 245-8, 248n Alexander of Damascus, 184n Anaxagoras, 47 Andronicus, 3O7n Annas, J., I95n, 227n, Anniceris, 18-49 Antiochus, 241, 242, 262 Antipater, 22n, 23n, 116, 232,
Antiphon, Antisthenes, 47n Antoniades, E., 19, 23n 'Antonius the Epicurean', 192 Apollodorus of Cyzicus, 47 Apuleius, Lucius, 282 Arcesilaus, i66n, 251, 259, 260, 261, 262, 264, 265, 266-7, 268, 269, 273, 274, 283, Arete, ign, 22n, 23n Aristippus of Cyrene, the Elder, 4, 18, 19, 20, 22, 23, 24, 30, 32n, 35, 40, 48 Aristippus the Younger, 18, 19, 2on, 22n, 32,40 Aristo, 5, 261, 293, 294-6
Name index Aristocles, 2911 Aristotle, i2n, 1711, 30-1, 38, 3911, 64, 73, 93, 101, 104-8, io8n, 122, 123, 140, 141, 143, 145, 147-8, 152, 178, 186, 188, 194, 195, 20211, 203, 205, 206, 207, 208, 209, 210, 211-12, 213-15, 2i6n, 245, 247, 24811, 249, 252-3, 280, 283, 286-90, 297, 316, 31911, 326-7 Aristotle of Cyrene, 2311 Aristoxenus, 117 Arius Didymus, 154, 158-9, 23 m Arnim, J. von, 11911, 24211, 255 Arrighetti, G., 5411, 5511, 5611, 5711, 5811, 5911, 6on, 6in, 64n Arthur, E. P., Asmis, E., 6 Athenaeus, Attalus, 160-1, Augustine, Saint, 27m, 32 m Aune, B., Bailey, C , 86, 86~7n Ballester, L. G., ig2n, i98n, iggn, 2O2n Barnes, J., 57n, i85n, i94n, 227n, 299n, 3i6n Bayer, K., Beck, 3O Bellincioni, M., Blackburn, S., 29 m Bollack, J., 22n, 23n, 24n, 34n, 4111, 46n Brecht, B., 97, 144-5, l4^» J 4^ Brehier, E., 24 m, 275 Brunschwig, J., 7n, 2i5n, 326n Burnyeat, M., 236, 269n, 286~7n, 287, 288, 289, 296n, 2g8n Calcidius, 23m, 237n, 277 Cancik, H., 293n Carneades, 33n, 242, 260, 261, 262, 264, 266, 268, 271, 272, 273, 274, 283, 320 Cato, 271 Catullus, 272-3 Cebes, 23n Chrysippus, 99, 100, 101, 103, 105-6, n o , i n , 112, 113, 114, 115, 119, 120, 121, 122, 123, 124, 126, 128, 129, 131-2, I33-5. i36> 140-1, 142-3, 145, 146-7, 149, 151, 152, 153, 154, 155, 156, i66n, 168, 169, 170, i72n, 176, 177, 182, i87n, 188-9, I 9 ° J I9I> X92> 2oon, 2o6n,226, 231, 232, 237, 238, 240, 253-4, 260, 297n, 299n, 313-31 Cicero, 3, 4—6, 8, 9, ion, 11—17, 21, 33, 35n, 106, 116, 121, 123, 154, i66n, i75n, 231, 235> 237n, 241, 242, 244, 251, 256, 257, 258, 259, 260-1, 262, 263-5, 268, 270-1,
347
272, 278, 293n, 294n, 299n, 306-7, 313, 314-16, 319-20, 321-5 Classen, C. J., 29n Cleanthes, 99, 100, 101, i n , 115, 122, 126, 128, 132, 151, 156, i66n, 226n,232, 3i8n, 326n,329, 33on Clement of Alexandria, 3, 20, 24, 3on, 32n, 37n, 38n, 39, 41-8 Colotes, 72, 73, 81-3, 87, 267 Cooper, J., i55n, 3o6n Couloubaritsis, L., i58n Critias, 115 Crito, 23n, 322 Crouzel, H., 228n Damon, 117 Dancy, J., 29m Daremberg, C , 2i8n Davidson, D., i95n, 212 Decleva Caizzi, F., 279n, 299n De Lacy, P., 77n, 82n, 83n, 88, 98n, 109, 122, 131, i6gn, i72n, i86n, i88n, 267^ 3i3n Democritus, 47, 72-94 Dennett, D. C , 62 Denyer, N., 32on Denys the Dialectician, 23n Descartes, R., 63n, 209, 230, 244 Diano, C , 54n Diels, H., 69n, 325n Dierauer, U., 6gn Dihle, A., 294n Diogenes of Babylon, 99, 100-1, 103-4, 115-21, 237, 3i6n, 331 Diogenes Laertius, 3, 7, 8, 9, 18, 20, 21, 22-6, 29-30, 34, 36, 37-8, 39, 40-1, 42, 46, 48-9, 90, 93, 99, 103, 134, 154, 225, 232, 234, 235, 285, 286n, 296n, 297n, 3i8n
Diotimus, 47 Donagan, A., 295n Donini, P., 150, i59n, 16in, i63n, i89n, ig2n, i98n, 20m, 322n Doring, K., 19, 20-1, 23n, 24, 27n, 3on, 3m, 32n, 37n, 39, 40-1, 43n, 45-6
Duhot, J. J.,
Edelstein, L., 32 m Einarson, G., 77n, 83n, 88 Empedocles, 77, 78n, 80, 99, 115 Epictetus, 99, 100, 101, 106, 121, 122, 127, 128-30, 132-3, 136, 138-9, 141, 142-3, 145, 146, 151-2, 18m, 226, 227, 230, 231, 259, 286n, 31 in Epicurus, 3-17, 20, 21, 25, 27-8, 3m, 34n, 37, 41, 44-5, 46, 48, 49, 53-71, 72-94,
348
Name index
Epiphanes, 3211, 3511 Epitimides, 2211, 2311 Euclid, 2311 Eudemus, 18411 Euripides, 99, 109, 123, 124, 128, 132, 136, 139, 140, 141, 142-3, 145, 147, 148, 15611, 158, 187, 32011 Eusebius, 19, 2211, 2611, 32, 4311, 23111 Everson, S., 91 n Fillion-Lahille, J., 153-6, 16511
Fodor,J., 186 .
Frankfurt, H., 20511 Frede, D., 32411 Frede, M., i68n, 18511, 21 in, 234n, 25m, 270, 3o8n, 3i8n, 323^ 324n Fritz, K. von, 19, 2on, 2in, 23n, 41 Fronto, Marcus Cornelius, 133 Gaius, i84n Galen, 76n, 81, 82, 99, 100, 104, 109, n o , i n , 112, 113, 115, 117, 121, 126, 128, 134-5, 142, 154, 155, 156, i69n, i72n, 182, 184-222, 28m, 299n, 3i3n, 324n, 326n, 328, 329, 330 Ganss, W., 3O2n
Gellius, Aulus, i66n, 168 Giannantoni, G., i8n, 19, 2on, 2in, 22n,
Herophilus, i86n Hesiod, 99, 132, 134-5, 2 ^ 5 , 286, 288n Hierocles, 237n, 239-40, 241 Hieronymus of Rhodes, 5 Hippobotos, 25 Hippocrates, i86n, ig7n, 2i7n, 3i9n Holler, E., 153, 157, 16m, 164-5, I 7 2n » i73n, i78n, i82n Homer, 99, 109, 123, 124, 126, 139, 147, 193 Horace, 178 Hossenfelder, M., 2in Hoven, R., 32on Huby, P., 67n Hume, D., 195, 209 Hutchinson, D. S., 19m Iamblichus, 253, 33 m Imbert, C , 233n Indelli, G., 68n Inwagen, P. van, i95n Inwood, B., i8gn, igon, 2O4n, 2i3n, 227n, 23on, 23 m, 234n, 239^ 24m, 244, 245, 248n, 29on, 293n, 298n, 30 m, 3i3n Ioppolo, A. M., 33n, 15m, i66n, i95n, 2i3n, 253n, 255n, 256n, 259n, 26on, 267^ 26911, 273n, 28m, 285^ 32m, Irwin, T., i95n, 289^ 30 m
Gigon, O., Gill, Christopher, ioon, 142, i87n, 191 n Gill, David, 2ion, 22on Giussani, C , 86 Giusta, M., 2on, 22n, 24n, 2gn Glaucon, 23n Glucker, J., 270, 271 Goldschmidt, V., 7on Gorgias, 187, i94n Gorier, W., 252n, 255n, 323n Gosling, J. C. B., i4n, 15, i5-i6n, igon Gowans, C , 3O5n, 3o6n, 3O7n Griffin, M., 3i7n Grilli, A., 43-4, 47n Hadot, I., 16m Halliwell, S., io8n Harkins, P. W., i84n, 21m Harris, C. R. S., 326n Hecataeus of Miletus, 47 Hegesias, 2on, 22, 23, 24, 28, 2gn, 34-5, 37n, 42, 48, 49 Helmreich, G., i84n, 19m Heraclitus, 47 Herillus of Carthage, i87n Herman, B., 29m, 295n, 298n, 3O3n, 3ion Hermarchus, 67-8, 7on
James, Henry, 97 Jerome, St, 127 Johnson, M., I57n Joseph,275-84 Kahn, C. H., i59n, 18m Kane, R., i96n Kant, L, i95n, 295, 303, 3ion Kenny, A. J. P., 2i9n Kerferd, G., 3O3n Kidd, I. G., 293n, 294n, 3oon, 3O3n, 31m, 32m Knox, B., 142 Konstan, D., i8on Kramer, H. J., 25 m Kraus, P., 2O3n Krevelen, D. A. van, 1 i6n, 119 Kuhn, C. G., i84n Labarriere, J.-L., 327n Lacan, J., 281 Lachmann, C , 86n Lactantius, 16m Lafrance, Y., 252n Lain Entralgo, P., 2O2n Lakoff, G., i57n
Name index Laks, A., 711, 1311, 5611 Lamprias, 268 Langer, S. K., 120, 12011 Laursen, S., 53-4, 5611, 5711, 5911, 6on, 6in, 6511 Lear, J., 29211 Leibniz, G. W., 2i5n Lloyd, G. E. R., i87n, 2i6n, 2i8n, 22m Locke, J., 2i5n Long, A., 15, 53n, 54n, 55n, 56n, 57n, 58n, 59n, 6in, 89, 90, 9 m , i65n, i68n, 227n, 228n, 233n, 236, 238, 244, 245n, 253, 255n, 257, 266n, 27on, 29gn, 3O4n, 3o8n, 31m, 3i3n, 3i6n, 3i8n, 32m, 327n, 3 3 m Longo Auricchio, F., 67n Lucan, 132
Lucilius, Gaius, 160 Luck, G., 3i4n Lucretius, 59n, 66-7, 69, 69n, 83-6, 88-9, 92, 98, 211 McDowell, J., 29m, Mackenzie, M. M., 2ign, 22on Maconi, H., 253n Manilius, 32 m Mann, T., 280 Mannebach, E., i8n, 19, 2on, 2in, 29n, 38n Mansfeld, J., 25on, 33m Manuli, P., i93n, 328n Marcus Aurelius, Marx, Karl, 145 Mattock, J. N., Melissus, 93 Menander, 99, 109, 139, 140, 141-2 Metrodorus, 271, 274 Mignucci, M., 78n Mitsis, P., i4n Montaigne, M. E. de, 268 Mounce, H. O., Miiller, I., i86n,
Natorp, P., 45n Nausiphanus, 47 Nell, O., 296n, 30 m, Nemesius, 244 Neubecker, J., 116, Nikiprowetzky, V., 25on, 279n, 280 Numenius, 242n Nussbaum, M. C , 33n, i87n, i9on, 191, 194, I98n, 201, 2O4n, 2o6n, 289n, 29 m, 3O2n, 3O5n, 3ion Nutton, V., Origen, 23m, 226, 227, 228-9, 230, 231, 245-6, 248
349
Orpheus, 99 Panaetius, 25, 40, 16m, i65n, 260 Paraibates, 22n, 23n Parmenides, 94 Phillips, D. Z., 3O5n Philo of Alexandria (Judaeus), 225n, 226, 229, 23m, 233, 234, 237, 238, 250-84, 33in Philo of Larissa, 264, 270, 271, 273, 274 Philodemus, 6gn, 88, 10m, 116-17, IJ8> 119, 120, 121, 2oon Pigeaud, J., 2O2n, 2O7n, 2i8n Pindar, 104, 109, 117 Plato, 10, i2n, 14, i5n, 23n, 36n, 41, 73, 74, 79,82,93, 100, 101, 104-8, 109, I I O - I I , 112—13, 114, 116—17, 118, 121, 122, 123, 124, 131, 133, 135, 146, 152, 157-9, 163-4, 186, 187, 197, 202, 203, 206, 2o8n, 209, 2i9n, 22on, 251, 252-3, 261, 262, 263, 265-6, 269, 273, 278, 283, 297, 313, 316, 317-20, 325-6, 329, 331 Plutarch, 3, 29n, 38n, 39, 72, 73, 74, 76-7, 77n, 81-3, 86—90, 97, 101, 103, 121, 122, 124-5, I2&5 130, 131, 133? 136, 137-8, 139, 140, 141, 142, 143, 145, 146, 149, 154, i66n, 2i5n, 225n, 226n,229, 232n, 233, 237n, 240, 251, 254, 267, 268, 269, 323n, 33m Pohlenz, M., 16m, 24on, 293n Polystratus, 68, 6gn, 70 Porphyry, 225n, 227n, 229, 233, 237n, 242, 244, 248n Posidonius, 99, 100, 101, 103, 104-5, 109-14, 115, 116, 117, 118, 120, 121, 153, ! 54» !55> J56> 164-5, I91* J98> 259> 260, 276, 277, 281, 3i3-i4n, 321 Praxagoras, 326 Proclus, 285n, 3i4n Protagoras, 81, 82 Publilius Syrus, 109, 126-7 Putnam H., 28gn Pythagoras, 47, 218, Rackham, H., Radice, R., 25on Rawls, J., 295n Reesor, M. E., Reynolds, L. D., 291 Riese, W., i84n, i99n Rist, J. M., 9m, 15m, i53n, i64n, 240, 294n, 31711, 3i8n Rodier, G., 236 Rolke, K.-H., i56n, 276n, 277n Ross, W. D., 305-7 Runia, D., 25on, 275, 277, 28m
350
Name index
Sandbach, F. H., 7711, 24m, 26511, 32611 Santurri, E., 30511, 30611, 30911
Sartre, J.-P., 308
Saunders, T. J., 22011 Schofield, M., 29911, 31 in, 31311 Schwartz, E., 19, 2011, 3911, 41 Sedley, D., 15, 53, 5411, 5511, 5611, 5711, 58, 58-911, 6in, 64n, 67n, 68-9n, 75n, 77n, 86n, 89, 90, 9 m, 159, i65n, i68n, 194-5, 228n, 233n, 236n, 238, 244, 245n, 266n, 27on, 297n, 299n, 31 in Seneca, 99, 100, 101, 106, 121, 122, 123-4, 126, 127, 128, 130, 132, 138, 139, 140, 144, 148, 150-83, 2oon, 20in, 2O7n,
2i8n, 237, 239, 285-312, 3i8n, 324n Sextus Empiricus, 26n, 2gn, 72n, 91, 160, 161, 225, 226, 227n, 23m, 232n, 233, 234, 235, 238, 241, 242-3, 257n, 259, 26gn, 27m, 280-1 Sharpies, R. W., 245n, 32on, 324n Sherman, N., 289n Shields, C , 289n Sifakis, G. M., 39n Simmias, 23n Simon, 23n Simonetti, M., 228n Simonides of Ceos, 104, 109, 117 Socrates, ign, 22, 76, 106-7, 129, 135, 218, 261, 273, 286n, 287-90, 313-14, 316-18, 320,322,326 Solon, 3 m Sophocles, 132, 140, 187 Sorabji, R., i75n, 233-4^ 326n, 327n Speusippus, 23n Spinoza, B., i95n Staden, H. von, i86n Stern, S. M., 2O3n Stesichorus, 99 Stilpon, 23n Stobaeus, 3m, 158-9, 2i9n, 23m, 242, 253, 255> 257n, 258, 32gn, 33m Stough, C , i95n, 245n Strabo, 21, 22, 4on, 48, 4gn, 99, 122, 126, 127-8, 130, 149, 198, 249, 32on Strawson, P. F., 196 Striker, G., 3m, 34n, 47n, 49n, 9m, 24on, 267n, 26gn, 288n, 29on, 294n, 304-5^ 31m Stroux, L., i56n
Tarrant, H., 25 m Taylor, C. C. W., i4n, 15, i5-i6n, 9In Theagenes the Cynic, 215n Themistius, 285n Theodorus 2on, 22n, 23, 24 Theophrastus, 23n, 72-3, 74-5, 78-80, 80-1 n, 82, 93, 249, 325 Thillet, P., 247n Thompson, G. R., 75n Timon, 99 Timotheus, 99 Trillitzsch, W., 3O2n Turnebus, Tynan, K., Tyrtaeus, 99 Usener, H., 56n, 5 van Straaten, M., Vander Waerdt, P. A., 68n, 293n, 294n, 30m Varro, 237, 255-6 Vegetti, M., i8gn Velkley, D., Vlastos, G., Voelke, A. J., 16m, 253 Wachsmuth, 255 Walzer, R., Watson, G., Westerink, L. G., Westman, R., 77, 83n, 88 White, N., 294n, 296-7^ 3oon, 301 Whiting, J., 30m Williams, B., 3O5n Williamson, C., 3i2n Winch, P., 3i2n Xenocrates, 73, 261 Xenophon, 26 m Yon, A., Zeller, E., 18, 19, 16in Zeno Stoicus, 99, 100, 103, n o , i n , 115, 132, 151, 156, i66n, i74n, 250, 253, 254, 255-8, 274, 285-6, 290-1, 317-18, 325? 326, 329, 330-1
Index ofpassages cited
Aelian NA vi.50
Aristocles fr.3 Heiland
22611.2
Aetius
Aristotle
Placita I-I5-9 iv. 11.3
4 iv. 12.1-3 1-4
de Anima
88 237
19311.34, 32611.49
403b12 42 9a iff. 425012
23711.22 238 235
427D24-27
in. 1-2
Albinus Alexander of Aphrodisias
EE 1127012-16 11.8
<& Anima
EN
Epit. 30.2
p.71, 10 Bruns
in Top. 9, Us. Fr. 404
28211.99
1095a19-20 109701 1097b1-6
25411.15 246 21311.86 245 245 246 246 245 248 247
1098a18-20 IIO2b25ff. I1 ioa2ff. I11 oa 14ff. 111 ob 1 ff. 111 obgff. iuobi8ff. nna22ff. 1111 a24ff. 1111b6-18 I I I3b22ff. 1113b3ofT.
32011.26
246 246 246
Andronicus Peri Pathon 2
30711.81
Apuleius Plat. dogm. 19.247
28211.99
Arcesilaus fr. 9 Mette fr. 10 Mette fr. 11 Mette
2650.46 2650.45 26511.44
19611.42 1211.7
ii37bi 4 1 i37bi6f ii37b28 1139b17
19311.31 28011.90
2480.37 3011.46, 3311.61
31 28611.6 2960.41 3111.50 2O2n.6o 1910.27 21311.86 21311.86 21311.86 188
21511.89 188, 2i3n.86 19411. 36 24811.37 21511.89 21311.86 i95 n -39 2150.90 195n-39 2890.13 2890.13 2890.13 2890.13 2520.8
351
Index of passages cited
352 Aristotle cont. 1143b11
1145a35ff. 1146agff. 1148b 15ff. 1150a 15ff. ii5ob2gff. ii52b2-6
1i53bg-i4 1165a 14ff. 1177a1 iff 117gb4~31 1
4 11.1—3
1-4
m. 1-5 V.IO IX.2
4
28611.6 2O2n.6o 2O2n.6o 21 in.84 20511.68 20511.68 2O2n.6o 15n.11 1711.13
58J
1.235 SVF 11.912 939 in. 191
229b
III.3.2O2a22ff. Poet.
1448b15-17 I 455 a 3°-4 4 9 Pol. VII. 6 VIII. 1-3
Top. v.i3obi6—17
Athenaeus XII, 544a
HI.41
Chrysippus LS28 38E3
123
2g6n.4i io8n.i 1
481
286, 28711.7 ig5n. 3g 2i6n.g4 194. !95 n - 39 28gn.i3 28gn. 13 3O7n.8i ig6n.42
Cicero
1.41-2 45 46
39n-87
231
1.40-2
235 235
11.20-2
37-8 37-9
99
26811.58
118
32m.2g 32m.27 32m.2g 32m.27
38
11.34
89 124
255n.ig
de Fato
5 5-6 7 7-9 7-11
23m.12 32n.55 32m.2g 27m.66
327 n -5° 3i7n.i2
316
3i4n.6 321
32on.25 313, 314-15, 321, 331
3r9
11-16
32in.2g 322,324 32on.25 32m.2g
13
317^12
8 91 1
3° 33 34 36
40
23m.12, 237n.2i
242
VII. 2.4 de Divinatione 1.52
107
2i6n.g4
241
323^32
Att.
178
3ign.2O
255n.2i, 256 25611.24 262n.35
Acad. Pr.
111
93 ig6n.42
i87n.i4 2oon.55
Acad. Post.
i86n. 11 186n.11
Calcidius
in Tim. 220
317n.11
15611.12
208
Augustine Civitas Dei v.2 Contra Academicos
32m.28, 324^37 32on.26
463 466 478
28gn.i3
Arius Didymus ap Eusebius Praep. Ev xv.20-5
31711.11 31711.12
i6gn.33, 176
Physics 1
323n-33 323n-35 323n.34
462
de Insomn.
1.459a 15-22 PA 63gb7ff 640a14ff.
551 55I 2 55I 5
42
44
320 320
3:5n-7
i66n.2g, i75n.48 322, 324 272^70
de Fin. 1
4
Index of passages cited 29
29-3° 29-42 32-3
34-6 37
37-8 39
41 42
55 41-2 11 1
5 7.20 9
5-J7
6-7
5-i7 6-7
5-!7 5-*7
6-7, 11
26511.47
7n-3
2111.7
4 16
13 16
i3n.8 3,4 5 4 5 i3n.8
9 31-2 35 4i
in.21 20-3
23711.22
33-4
225, 23711.22 29311.26 29311.26
62
65-6 67
1.47
123
72 de Legibus
deOff. 33 11.11
m. 50 5iff89 9i
92-5
Lucullus 18 28
122
133 pro Quinctio 32
79
Rep.
29411.32 23111.12 116 29911.51
299 n -5 I
29911.51
306-7
116
263
124
26011.31
23111.12 23111.12 23111.12 23111.12
3!5n-7 3!5n-7
1.16 69
26211.34 25711.25
59
323 n -32
Topics Tusc.
1.18-23 20
23
3483 42 7i
in. 13.28-31 52
75f. 59-6o iv.66-7 14
83
2950.36
112
67 68 78
34
v-5
27011.65 2720.71 25611.24 25611.23 26211.36 25611.24, 2 7 m . 6 7 27111.67 25611.24 27111.67 2711111.67, 68
29 33 35 59
31911.21
3*9
225
4 5 5
IV
27111.69
29
1311.8
6-7 3,4
25211.9 26311.37 27011.62, 27111.67
63 ND 11.16
5-J7 5-i7 6-7n.i, 8
10
l
145 147 148 pro Murena
353
120
Cleanthes SVF1.518
26311.38 26311.39 26411.40
35n-69
32611.44 26411.41 3311.60 3311.60 3311.60 26011.30 32911.61 25611.23 25611.23 2720.70 27211.70 32611.43
Clemens Alexandrinus Stromateis
11.21 21.127, 2 21.130, 7-8 xxi. 127.2 128.1
3
32n.55 2on .7,41-8 4811.120 4811.120
Us. Fr. 451 Democritus
3
B.i 25
81
fr.n
y4
Diogenes of Babylon SVF54r
56 58
62
117
,
J I
7
117 118
Index of passages cited
354 ogenes 01 miDyion cont.
109 114
119
63 67
JI
"9 "9
71
119
79
in. 30
"5
logenes Laiertius 22 n.65-85a 66-83a 22 23n.22, 47n.i 17, 49^127 85 22 85b-86a 18 86-93 86b-go 24n.23 860-91a 24-30 86b~93a 23 86b-104
87-8 89 91
9 3 24n.23, 33^62, 38n.84,
96* 96b—97a
97D-103
98 105 Ir
3
125
VII. 4 25 26 28
44 46 49 50 51 52
53 55 58 84 85 89 96
22n.i9
28^38, 35n.68, 36n.75 23
34n.66, 35n.7o 340.67
3 n -22 3 n -2i 23n.22 99, 285n. 1 132, 285n.i
286n.5 99,31811.15 103
234, 254n.i5, 33on.64 232
330^65 232
235 299n.52 103, 237n.22 3i6n.8 286n.4 238 296n.4o, 297n.42 21 in.82
99
181
100
187
134 99
200 201
xi.82 x.31 32
9O.93 8
33 34
136
Empedocles DKB9
15,
77
Epictetus
1.4.23-30 16-18 6 16 22.1-3 9-10
42n.i02 23 23
2
n.i5
180
28^38, 29, 35n.73, 36^76,
2
99 99 99
173 176
23n.22
42n.iO2
9i-3 92 93 93°~9^a 94 95
170 172
116
229-30
2i4n.88 168
166
"5
49
5
121 127
30111.60
28.7-9 3J~3 8 17.19-22 26.31 in.3.20 19-20
129 129 227, 230-1 226 31111.88
3iin.88 H3
138 31 in.87 226 142-3 129,132-3
Epicurus Ep. Her. 40
48 62
67 68
57 n - 2 3 57n.23 59
71
Ep. Pyth. 92
97 "3 Hdt. 49 53 63
56n. 12
55n-9
57n.2i 57n.2i
83 86 ion.5
Index of passages cited ion.5 ion.5
72
37-8 49-50 KD 3 7 16 18 l
^^n.Q
69*1.55 58n.24 58n.24 6 711.47, 7on.6o
9
25
32 Men. 128 129
6n.i, 3m.52, 45n.io8 6n.i, 7n.3, 9-10, 16 14
131 132
133 on Nature
[24] [48], 7, 17 [24] [49] 4, 8, 27 [37] [35] " - 1 5 PD 3 22
Sent. Vat. 25 81 fragments Arrighetti (1973) [11]
[15H20] [17] 5-6
[20] 151T. 6—14 [21] 11-12 16
[2i]-[22] [23] 22-5 [24] 3-i5
[25] 5-9 19-34
69*1.53
34*1.66, 5 6n.i2, 58^24
57 n -23 57*i-23 58n.24 14*1.9 ion.6
ion.6, 58n.24 55*i-9
53n.2, 5511.8 55*in 5411.6 54*i.6 53n.2, 61
20C (1)
(4) (8)
92
37n.8o ^8n.24 O9n.54
355
20J
Sedley (1983) pi 9 1-8 15-19 PP19-20 p20 38-41 P36 5-7 8 PP36-8 Usener (1977) 76 84
55*i. 11 58*1.29 54*i-6 58n.26 57*1.20
56^13, 64-5 57*i-2i 57n-2i c^on.24
423 471
^on.24
Epiphanius Adv. Haer.
m.2.9
32n.55, 35n.69
Euripides Alcestis
1079-80 1085
i87n.i4 i87n.i4
Andromache
629-30
i87n.i 4
Electra
57*1-22 59-60, 62 64^40 65*1.44 60 Ron.26 57*1.20
56n. 13, 59^31,64-5
60, 64 61
61
r r~-\ [26] I-I5 [26H3O] 58*1.29 [27] i-9 [28] 19-27 54*i-6 [32] 22-5 58n.28, 64n.4O 58n.28 [33] 3-4 [34] [33] i-7 54*i-5 [36], [16] 6 55 n -9 Laursen (1988) pp 7-9 56*1.13, 64-5 Long& Sedley (1987) 20B 53*1.2, 56n.i3, 59^31,64-5 58n.26 (2) 57*1-20 (3)
,87,I4
15-16 Hippolytus 443
l87
Medea
53O-I 1078-9
187 i87n.i4, i88n.i6
Phoenissae I
3~3 I
32on.26
Eusebius Praeparatio Evangelica
iv.B.5 G xiv.18.31-2 xiv. 18.32 xiv. 19.1 xv. 20-5
1911.3, 2onn.9, 11, 22n.i8, 26n i9n.3, 2onn.9, 11, 22n.i8, 26n 31, 32n.57, 43^105 23m.12
Fronto J
SFF111.27
33
Galen
All references between pp 184 1 and 222 are to
Ktihn (1821/33). AA 11.217-18
i84n.i
356
Index of passages cited
Galen cont.
Aff. Dig. V.I
192
1-58
189
2
192 192
2-3
3-6 7 7-14 8-14 13-14 14
15-16 16-17
n
19211.29, 198 192, 198 199, 22211.105 199
i99 n -5 2
199 199 199 20111.57
17-21 19-20 20-1 21-2 22-3
200 217 200 201 207
23 24
202, 20611.69 20211.62, 206, 20611.69
27
203 203
26-7
27-8 28 30
30-1 32
32-3 33 34 35
35-6 37 37-40 37-52 40
40-41 41
44
44ff.
45-56 46-9 47 48-9
20811.77 201, 203 199, 21811.101 222n. 105 201, 205-6 2HII.85 2 1911.101 19211.29 20211.62, 206 I981I.5O 20211.62, 206 215^91
201
2IIII.85 21 I 21211.87 201 201 201
51
211
52
2O211.62, 206
53
201 200
53-4 55-6
x. 1-204
200
I961I.44 I
97 n -45
197*45 i97n.45 i97 n -45 i97n.45 22in.iO3 82
1.2
Hipp. Prog. XVIIIB 87-9
Hipp. Prorr.
v.5.32-6
xvi.517-18 Lib. Prop. xix.i6fT. 45 Loc. Aff VIII. 13 iff. 156
160-6 73
J
178
i95ff. 226-7 327
417-20 Med. Exp. J 5-7 de Methodo Medendi
x.35 19m. 78-81 89-91 636
201
2O9II.78 2OIII.57 18411.1, 209
51-2
Caus. Cont. 9 Caus. Puls.
Cris. ix. 760 [Alexanderson] Diff. Puls. vm.493-765 497 Dig. Puls. viii. 766-961 On the Formation of the Foetus iv. 700-12 Hipp. Elem.
28in-92 2O2n.62 i84n. 1 2O3n.63 2O2n.62 2O2n.62 2O2n.62 2O2n.62 2O2n.62 2O2n.62 2O2n.62 2O2n.62 2i2n.87 81
18511.5 24
207^71
927
i93 n -3 2 207^74 2i5n.92 215^92 215^92
70
186n.11 i97n.44
78i
909-14 Nat. Fac. 11.9-10 Opt. Med. I
i97n.45
-53-63
59 Ord. Lib. Prop. xix.43 Pea. Dig. v. 59-103
215^92 2o6n.69 i84n.i 189
7i
2i5n.92
1.7.1 11.4.40
326^46
PHP
5
8 12
330 115
330 33i
Index of passages cited 22~3 in. 1.14-15 8.230-2 D IV. 1.6 14
2.28-38 (pp.372-5 K) 30 (p.373 K) 3-2
248 D 5.12 (p.394 K) 266 D 268 D 6 v.i.4 2.20ff. 213-18 214
224-5
227-8 233-4 234
235ff. 255-6 261-4 269^ 271
286 D 292 D 307
315- 1 6 318D 318-20 D 320-2 D
322 D 323
324 D 327
327-9 328 D 330 D 332 332 D 334 D 358-9 364 369-70 371 371-2 372
372-7 375-6 378-89 389-91 390
403-24
330
414-15
i87n.i4 i87n.i4 2oon.55
416
21 in.81
405 413
3I3n-l
135 31311.1 3i3n.i i72n.4O
416-17 419
i6gn-33 32gn.6o no i72n.4o no no 328n.58 329n.6o 329^59 i88n.i8 i88n.i8 i88n.i8 i88n.i8
423
2i5n.g2
600
299n-54 2o8n.75 2o8n.76 2o8n.76 188 188 109
110
i87n.i4 i88-9n.i8 111 112 112 111
i88n.i8
423 D 424-6 429-30 432-54 439-42 448-51 45 I "3 505-13 506-7 509 510
791-5
vn.453-92 QAM iv. 768-9 772-3 776-7 778 782
784 787 788 812-14
"3
814-16 816-78 818-19
190 192
189-90 i87n. 14, 190 190
19m. 25 189 191
198
i87n.i4
19m.27 192^27 192^27 196 197
22in.iO3
Praes. Puls. ix. 205-430 421-30 Puls.
112
189
igin.23,
xiv.605-19 627-9
i88n.i8
i89n.i8
19m. 23 igin.23
Praen.
112 188
110, 21 in.81 111 111
357
819-20 de Sectis
22in.iO3 22in.i03 2o6n.6g 2o6n.6g
217 2o6n.6g
22in.iO3 218 215 112, 198
1.83 Subst. Nat. Fac IV
-759 760
763-4 Symp. Diff.
vii.44-7 45 55-6
Syn. Puls.
ix.431-549
22in.iO3 i86n. 11 22in.iO3
Index of passages cited
358 Galen cont. UP 111.454 19m. 897-9 iv-347
55 24 21111.83 21711.98
Fragments, DK
68.A.49 68.B.125
82 7611.7
Galen, PseudoIntr. xiv. 726.7-11
23111.12
Gellius, Aulus Nodes Atticae 19.1
16611.29, 168
Gorgias Defence of Helen 15-19
187 18711.15
Fr. 82.B.11 DK Hermarchus 12.5-6
6711.47 6711.46
Hesiod Opera et Dies 293ff.
285, 28611.6
IV.38-53 Hippocratic Corpus Airs, Waters and Places, 1224 On the Arts 3-5 Flat. 15 Nat.puer. 17 Prognosis
239 23711.21 240 31911.20 217*1.97 32611.45 32611.45 21711.97 193
Horace Ars Poetica 101-5
178
Iamblichus de Anima
ap. Stobaeus 1.366.15-16 ap. Stobaeus 11.831
33 in -72 25311.10
Lactantius de Ira Dei 17.13
Long & Sedley (1987) 27E 28 39E 44-5 44c 45 47
Lucretius 1-15 163 8i4ff. 895-6 936-41 11.161-76 170 25^93 265 268 333-8o 343 434-6 795-8O9 808-9
921 1081-3
in. 48-93 288-320
Hierocles Eth.
Homer Iliad 644-55
69A 69K
16111.20 32811.55 32611.48 33111.68 31811.19 32611.42, 32811.55 32711.51, 32711.53 33111.68
299
320-22 iv. 72-4 72-89 90-2 218-24 622-7 644 649-51 658-62 668-72 677 693 694 699 706-21 986-1010 1192-1207 v.i056-90 Manilius iv. 105 Mannebach (1961) fr. 186 fr. 190 Melissus fr. 8 fr. 8.6 Nemesius Nat. Horn. 291.1-6
32811.56 32711.51 26611.51 27011.63 6911.56 6911.56
85 85 85 59n-30 67 67 67 67 85 6911.56 92 88-9 89 6911.56 6911.56 811.48, 211
66 67 66
83 83 84 84 84 85 85 85 85 86 86 86 86 8611.28 69*1.59 69*1.58 69*1.57 321*1.29 29*1.44 3811.86
93 93 244*1.33
Index of passages cited Numenius fr. 45 Des Places Origen Contra Celsum iv. 79-82 iv.81 deOr. VI 1
de Princ. in. 1.2-3 3-5 4-5
24211.31
226 229 230 21311.86 22811.6 228 227 229 22911.9
Parmenides fr. 8.52 PHerc 19/698
94 89
29 Agr.
282
30 Anim. 17-21 29 40 42
277
83 92-5 95 96 99 Dter. iff Ebr. 70 178-80 Fug. 19 136 182 Jos. 32 Leg. 1.21 25 28 30 9i "•5 7
27711.80 23111.12 28111.95 27711.81 28111.93 282
Opif 117 161-2 Post. 135 Prov. 52 11.40 Quaest. Gen.
276 28m.97 28m.95 22911.8 22911.8 22911.8 276
Philo Abr.
77 77-81
23 46 in. 15 67 Mut.
359
22911.8 23111.12
233 n I 5
22911.8 22911.8 22911.9 22911.9 234 237 238 28111.96 282 5611.12 27911.86 28211.101 27711.77
Sacrif 104 Somn. 1.31-2 32 240 11.15
27611.76 27811.82 33111.72 28211.100 27811.83
Philodemus de Dis 1.12-15 6911.55 de Ira (Wilke 1914) 20011.56 On Methods of Inference 25 pp. xlii-xliii 88 Peri Mousikes Book iv (Neubecker 1986) 1 119 4 119 5 118 8-10 118 11,63 118 15, 70 118 17-18 117 37 117 39 117 40-1 119 40-4 119 73-8 117
Plato Charmides Cratylus
28211.100 Crito 28111.93 28111.94 277 23111.12 27811.82 28111.94 27911.86
44a-b
Gorgias 476a~7 478cff. 525D1T. Laws 644c!
76 316 21911.102 21911.102 21911.102 27611.76
36°
Index of passages cited 6o8e-6na
Plato cont. 72oa-c
728D-C 73id 73ie 735bff. 747dff. 747d2-e2 VII
105 22on. 102 2I9-2On.IO2 220n. 102 2i9n.io2 2i9n.io2 1980.50 2190.102 2i6n.94 3190.20 3i8n.i8 105
Magna Moralia 15n.11 Meno 98b Phaedo 62b-c 62C6-8 66b 70a 77d-8oc 82C-83C 90c 114c2-6 Phaedrus
26611.50 32611.47 r
_
I9in.27 I9III.24
4In-99 3611.77 3611.77 2O2n.6i 2O2n.6i 107 107
378d-e
135 107
439C-42C 440b 441a 441c 445a-b
206 106 106 105
64a-65b
6507-6607
5 88c-d
3611.77 1970.46
59^ 6o6a-d 606b
105 100
82 3611.77
33In-7° 329
1930.32 360.77 74 22On.IO2 22On.IO2
277^78
Plutarch Adv. Colot. 26
1108.F 1109.B-C I I09.Fff I I IO.A-B IIIO.C-D IIIO.EF I I I I.DE I I I3.AB II2OC-II2IC II2IC-II22A I I2IF II22A-II24B II22A-B II22B-C II22C I I23A Aetia Physica 38 Bruta 991F-992A Catalogue 128
26711.51, 269^59
93
34O3-9 oia-ooc
90a
3i8n. 17
I97n.46
Republic 345^ 358e-6id 377a-b
189c 191 aff. 2o6d Timaeus
86b-87b
245c
33d 43bi-5
Theaetetus I5i.c-i52.a
89D-C
.87
Philebus 2ib-c 33C5S
Sophist
de Communibus Notitiis 1084F Cons. Apoll.
i66n.29 81
83 87 87
88-9 76 73 77 2911.39 2670.53 267 266 2670.54 1750.48 2670.51 2680.57 2320.13 2290.9 2680.57 3310.69
6o8d 2150.91 Esu. 999A-B 2260.2 How the Young Person Should Listen to Poetry 101 15DC 97 15c 137
Index of passages cited i7cd i7de
19-20 25c 2
5d
26a 28b 32
34b ff. qcH
JJ U
36b 0
Lib. Aegr.
6
138
151
139-40 125 125
152 160 161
141 141
165 168
133
168-72
139-40 14.Q
i55n.8
Peri euthumias
265n.46
9 Quaest. conv. Soil.
961c
961E-F
962F 964A-F 967E 968c 970B 973E St. Rep.
1038 B-C iO44d 1055^ 1056a 1056a 1057a fr. 152 Sandbach
38n.86 237N.21 233NI5 232N.13 226N.2 226N.2 229N.9 226N.2
22gn.9 240 131
254^17 32311.32 267^52 265n.44
vii.6-8 de Abstinentia
1.7-12, 26, 4 in
6.4 10.1-3 19.1-2 22.5 22.5-6 de Fac. Anim.
ap. Stobaeus, Eel. 11.25, W.349, 23-7 Posidonius Edelstein and Kidd (1989) 31 33 35
229^9 229^9 229^9 237n.2i 233 n i 5
3i4n-5
Prolegomena in Platonis Philosophiam (anon) 1.1-6
265^49
25-31
26(^1.50
Seneca
de Beneficiis
1.3.10 4-5 4-34 de Brevitate Vitae 16.5 Consolatio ad Marcian 19.4 Epistulae
8.8 65-2 104.21 108
374-5 52
i n 112
1-2
no, 112
285^3
622.24, 629.37-630.1
57.3-6 4 58.25ff. 59 6 7-9 65 71.27 29 72.8 92.1
242
3i4 n 4
ad 293ff.
Epistula Morales
Porphyry
3J4n4
112, 113
In Platonis Parmenidem
225n.i
68
111 112
ad Hesiodi Opera et Dies
67n.46
On Irrational Contempt for Popular Opinions
32m. 29 3!4 n 4 no
Proclus
108.8-9 108.10 108.12 II5.I2ff. 115.15 121.6-15 121.15 121.19-23
Polystratus
361
124 124
3O7n.8i 124 124 126
3i8n.i9 3i8n.i5 130 126
126-7 126 124 140
239
237n.22 237n.2i 161
I58n.i4 i66n .28, 168 i74 n 45 i63n.2i
160 160 160
I59n.i6, i63n.2i 164
164, i68n.3o, 17911.57 160-1 162
161-4
362
Index of passages cited
Seneca cont. 2
5-10 94 l 9 23
26 29 30
32 32-3 33-4 35 35-6 37 47 53-4 56 59-74 67ff. 71-2 95
294 2971-43 29611.41 297^.45. 299 29811.46 29811.47 291, 29611.38, 30011.55 29611.39 30011.58 29411.32 295n-37 29711.44, 298 29811.48 29711.42 29911.52 30211.64 30311.71 30211.66
299n-54 12 29111 18, 2961111.38, 39, 3001111.56, 35 37 48ff. 5iff. 59 63 64 44-6 52-3 60-1 63-4 47-54 43 57 113 2
18 23 115.18 116.5-6 de Ira 1.1.3-7 2 i-3
5 3.1-2 3-8 7 3-7 5-2-3
59 2961111.40, 41 30211.65 30311.70 29511.36 300
30011.57, 30111.60 300 300
30311.71 312
299n-54 295n-35 29411.28 29411.28 17211.39, 17511.48 1751-48 16611.29, 17511.48 16611.29 30711.81 16111.19 20711.72 16111.20 19211.27 20111.57 20111.57 169 237 20011.55 20111.57
169 18211.62
1
170
2
170-1 18211.62 171 171
9-i
20111.57 4 l7l> 172 11.8 21811.100 16.5 7 16611.28, 1741111.44,45, 17611.50, I771-52 17211.39 17.2-3 20111.57 18.3-6 19.1-4 20711.72 2-4 20111.57 II.I.I—1.4 172-3 i-4 153, 16811.32, 164-5 2
!74> 177 175 176 177 177-8
1
300
10
170
3 3-11
293 5
20111.57
6.1-5 7-2-3 4 8
162-3 163
2
3-6 5-6 2-4 3 4 4-5 5 4
i7oi-35 179 17411. 43, I771-51* 179 179-80 16811.32
1-2
in. 10 Leg. 1.18 i8-35 11.8-13 de Otio 3-i
Quaestiones Naturales 1. proem Resp. 11133
de Tranquillitate Animi 11.8 de Vita Beata 26.6 Sextus Empiricus de Abst. 11.2, 3 Adv. M. in. 40 m.33ia-332 31111.88 vn.i5off. 158
180 180
18211.62 30011.57 29411.28 29411.28 15114 15911.16 29411.28 157-8 124
22711.4 235 259 26911.60
Index ofpassages cited i6o 162 191-8 199 227-62 234 253 257 424 432 VIII. 70
87 275-6 285-6 IX. 102 131 153 199 393 394 Math. 11.11 off. i4iff. VII. 135-6 203 206-10 VIII. 63
275 ix.71-2
75-6
243 238 2911.39 233 233 241
234 25711-27 233 225 226, 235-6 226 23m.12 226n.2 23m.12 235 235 21 in.82 21 in.82 72n.i 72n. 1 9i 72n.i 33m.70 3i8n.i9
PH I.25ff. 40 4O-61 60 6l 62 62-3 62-78
69 74 75 77 78 100-17 235 11.26 in. 15 fragment, DK 68.B.9
21 in.82 226
226 226
56n.i2 27m.66 231 323n-33
Sophocles
Antigone 78 iff
187
Stobaeus
1.366.15-16
331^72
363
11.62.20-6 11.77.16f.
329n-59 3m. 52
Anth. 11.1.29
2i9n. 101
Ed. 11.7 9b 2 5> w - 349> 23-7 60.9 65.7 W 88 89.4 ff. SVF 1.68 11.548 831 in. 548
23m.11 23m.11 242 296^40 296^40 169^32 15*^1.15
Strabo 1.2.2 3 8 11.3.8 vi. 1.12 XVII.3.22
122, 127 127-8 io8 2in. 14, 4on.95, 48n. 121, 49nn.i26, 127
SVF 1135 150 211 219
563 566 II.228ff. 774 782 783 786-7 792 804-8 809-22 810
879 885 975 998 1071-4 in. 112 214 264 278 295 355 356 43iff 5*9 563
33on.66 21 in.81 28m.92, 328n.57 2in.i6 32on.22 33m.72 33111.72 33m.72 331^72 32On.22 277nn.77, 79, 33111.72 33in-72 3i6n.io 3i7n.i3 133 21 in.82 296^40
2i4n.88 2i4n.88
37n.8o 297^42 307^81
364 SVF cont. 654-6
Index of passages cited 132
see also under named authors
Themistius Orationes VIII. 108c
xiii. 17 id
28511.3 28511.3
Theophrastus de Causis
11.2-3 vi 1 -4r5 7.2 Dox. 484-5
70
73-82 84 89
Physic, opin.
80-in. 15 74 74 8111.15
de Sens.
50-4 55-6 60 61-2
03-4 65 65-70 67 69-70
79 79 79 79
fr. 8 Diels fr. 9 Diels Varro de Lingua Latina vi.56
79
80
79 80 80
78 79 74 73 8211.18
3250.40 237
Xenophon Memorabilia 1.1.13
26111.33
Zeno LS 55A 1-3
323n.35
