GISELA STRIKER The following rather lengthy piece is a revised version of the Nellie Wallace Lectures I gave at Oxford University in the spring of 1984. My aim in this series of six talks was to put together an outline of Stoic ethics that would permit an audience of non-specialists to see some of the connections between the notorious bits of Stoic doctrine with which-or so I assumed-most of us are familiar. For example, most philosophers or classicists will have heard that the Stoics believed the universe was governed by a divine reason, identified with nature; that they defended the view that virtue is the only good, and that the virtuous person would be free from all emotion. But I thought that it was not so clear, given our fragmentary sources, how these doctrines hang together, and so I tried to offer a more or less historical sketch of the development of Stoic ethics as one way in which the pieces of the puzzle could be put together. The first five sections can I think be read as a continuous account of Stoic theories about the goal oflife and of morality; the last one deals with 'freedom from emotion', pick- ing up what is perhaps the most striking feature of Stoicism. I am aware ofthe fact that a lot has been written on all these topics since 1984, and that it is now much easier, thanks to the sourcebook of A. A. Long and D. N. Sedley (The Hellenistic Philosophers [Cambridge, 1987]), to find one's way through the bewildering array of sources for Stoic doctrine. I have not attempted to deal with all these contributions; partly for lack of time, but also because I hope that a concise sketch might still serve a purpose distinct from a more detailed scholarly treatment: either as a kind ofintroduction, orto provoke discussion. The first version of the lectures was written while I held a research scholar- ship from the Deutsche Forschungsgemeinschaft, whose support I gratefully acknowledge. Since the typescript has been around for some time, I have collected so many helpful comments from colleagues and students that it has become impossible to remember them all. However, I should like to register my special debts to Julia Annas and Tony Long for encouraging me to publish the lectures; to Michael Hardimon, Donald Morrison, and Jennifer Whiting for discussing most of the material with me in the autumn of 1985, and to Mary Mothersill and Heda Segvic for a new set of comments in the summer of 1990 that helped me through the revision. Gisela Striker 199! 2 Gisela Striker 1. Why is it good to follow nature? THE moral theories of the Hellenistic schools are presented to us in the doxographic sources as being based, like Aristotle's ethics, on an account of the goal oflife. The thesis that there is such a goal, argued at some length by Aristotle in Nicomachean Ethics 1. r, and that it is what everybody calls eudaimonia, happiness or the good life for man, seems to be taken for granted. l But an echo of discussions about this point can perhaps be discerned in the Stoic definitions of the term telos. The Stoics defined the end as 'what all actions in life are appro- priately referred to, while it itself is not referred to anything else' (Stob. 2. 46. 5-7), and also as 'the ultimate aim of desire, to which everything else is referred' (ibid. 76. 22-3). The word 'appropriately' (kathekontos) in the first formulation brings out what is probably implied by the second as well (since orexis, in the strict Stoic termino- logy, is directed only at the good): namely, that the goal oflife is seen as something that everyone should or ought to pursue, rather than as an aim that everyone does in fact pursue. We do not have an official Epicurean definition of the term felos, but it is probable that Epicurus would have defined the end rather as 'that for the sake of which we do everything in life' (cf. D.L. ro. r28), that is, what we do aim at rather than what we ought to pursue. So the Stoics presumably argued that nature has set a goal for human beings that they ought to pursue, while the Epicureans held that the existence of a single final aim in all our actions can be established by observation of the natural behaviour of human beings. The difference would show up, for example, in differ- ent explanations of why people are not happy. While the Stoics would say that most people are miserable because they do not desire the right sort of things, and hence would be unhappy even if they got what they wanted, the Epicureans would say that unhappy people do not under- stand what it is that they really desire, and hence do not get what they want. Apart from this difference, the reports about the Hellenistic debate begin with Aristotle's second question: what constitutes happiness? It is assumed that the rest of the theory depends to a large extent upon the answer to this question, as is not surprising, for if ethics is the theory of how we should act and what kind of person we should try to 1 For some discussion of this thesis and its implications see my 'Greek Ethics and Moral Theory', Tanner Lectures on Human Values, 9 (1988),185-8. Following Nature 3 be, it is clearly most important to know that to which every action should be (or is) referred. As Aristotle put it, 'knowledge [of the end] will have a great influence on our lives, since, like archers who have a mark to aim at, we will be more likely to hit upon what we should' (NE 1. 2, 194"22-4). It may be that the impression of uniformity is partly due to systemat- izing tendencies in the doxographers; but it is at least fairly clear that all Hellenistic schools-including even the Pyrrhonists, but with the possible exception of the sceptical Academy-offered some answer to the question about the end, and there is good reason to think that they took this to be the foundation of their subsequent theories and recom- mendations. The Stoic doctrine about the goal of life is given in a rather bewildering variety of versions. The best-known of these is presum- ably the thesis that virtue is the goal; but we also get things like perfect rationality, peace of mind (ataraxia), or even freedom from emotion (apatheia). There exists, however, one formula that must have fulfilled the role of Aristotle's famous definition of the human good in NE 1. 7 (1098"16), namely homologoumeniis tei phusei zen, 'living in agreement with nature' or perhaps 'following nature' (akolouthiis lei phusei zen, D.L. 7. 88). (In the following I shall sometimes use the shorter 'follow- ing nature', for the sake of convenience, as an abbreviation of the longer official formula.) It is this definition, I think, to which one would have to appeal for an explanation as to why certain things should or should not be done, and even why one should lead a life of virtue. If the Stoics had an argument for their conception of the human good, it must have led to the conclusion that the good life for human beings consists in living in agreement with nature. It requires another argument to show that the good life is a life of virtue. This second claim should be derived from the thesis that the end is a life in agree- ment with nature, by establishing that such a life must be identical with a life of virtue. As Aristotle's ethics is an investigation of virtue and what belongs to it because the end is supposed to be a life in accordance with complete virtue, so Stoic ethics, I believe, is an investigation of what it is to live in agreement with nature. The first question to ask about Stoic ethics would therefore seem to be: why is it good to follow nature? A complication must be mentioned right at the start. Stobaeus tells us (2.75. 11-76. 3) that Zeno, the founder of the Stoic school, origin- ally defined the end simply as 'living in agreement' or 'living 4 Gisela Striker consistently' (homologoumenos zen). By this he meant, according to Stobaeus, 'living in accordance with a single harmonious principle',2 since (as he explained) those who live in conflict (makhomenos) are unhappy. Stobaeus goes on to say that Zeno's successors put in clarifications (prosdiarthrountes), holding that Zeno's first formulation was 'too short' (elatton kategorema), and thus Cleanthes changed the official formula to read homologoumenos tei phusei zen. This story about the incompleteness of the first version, though a nice invention, is unlikely to be correct. The words homologoumenos zen can stand without a complement, although the phrase may not be either very colloquial or very elegant Greek, and they do not mean the same as homologoumenos tei phusei zen. What they do mean is brought out by Stobaeus' paraphrase-leading a consistent and coherent life, one in which no conflicts occur and that is unified by adherence to a single principle. One does not need a reference to nature to make this intelligible. In fact, neither of the two formulae seems logically to imply the other. It is evident neither that following nature will lead to a harmonious life, nor that consistency and harmony could be achieved only by following nature. On the other hand, the way in which the two definitions are connected in Stobaeus' story indicates that they were intended to express the same doctrine, and that there was no serious disagreement between Zeno and his pupils. After all, Diogenes Laertius reports (7.87) that Zeno himself was the first to use the phrase 'living in agree- ment with nature' as a definition of the goal oflife. We should there- fore try to understand why the Stoics might have thought that consistency or harmony in one's life is the same as following nature, and see if their arguments about the final good can shed some light on this. Before I turn to the arguments that have been offered by ancient authors or modern commentators as the theoretical basis of Stoic ethics, I should say a few words about what might be meant by the phrase 'living in agreement with nature'. It does not mean, as one might at first be inclined to think, 'living naturally' as opposed to unnaturally, leading the kind of life that is natural for human beings. That contrast would be expressed in Stoic terminology by the phrase 2 I use 'principle' here to translate the notoriously untranslatable Greek word '\6Y0\,-'reason', 'proposition', 'argument', 'principle', 'language', 'speech', etc. I suppose that a native speaker of ancient Greek, if confronted with such a choice of different possible meanings, would often have been at a loss. Following Nature 5 kata phusin zen (living according to nature), as opposed to para phusin (contrary to nature).3 It was apparently taken for granted at the time that a good human life would have to be natural rather than unnatural in this sense, so that the question about the goal of life could also be put more precisely as 'What is the goal of a natural human life?' What the Stoics meant by agreement with nature is expressed most explicitly in another definition of the goal, ascribed to Chrysippus: 'living according to one's experience of what happens by nature' (kat' empeirian ton phusei sumbainonton zen, D.L. 7. 87). 'Nature' in this case refers to universal nature, as Chrysippus also pointed out (ibid. 88). So the Stoics are telling us that the goal of life, happiness, consists in a conscious observation and following of nature's wil1. 4 Why should that be happiness for man? I think one ought to realize that it is a rather strange suggestion, far removed from the traditional competitors virtue, pleasure, or fame-things that were standardly seen as valuable or desirable. Of course, the Stoics eventually ended up arguing for a life of virtue; but if they gave such prominence to their strange official definition, it must have had a fundamental theoretical role. What, then, was the Stoic argument for this doctrine? There seem to be only two sources that offer an argument precisely for the thesis that the end is a life in agreement with nature, as distinct from arguments that purport to demonstrate that the human good is perfect rationality, or virtue-namely Cicero (Fin. 3. I5-21) and Diogenes Laertius (7. 86--7)5 Like most of the evidence we have, the two passages in Cicero and Diogenes are sketchy and incomplete. Their interpretation has J To keep the distinction clear, I shall use the phrases 'according to nature' or 'natural(ly)' as translations of KUT<i q,VaLv. For the terminology see e.g. D.L. 7. 105; Alex. Aphr. De anima 2. 167. 18; Stob. 2. 81. 3; Cic. Fin. 4.14-15. 4 What is the relation between the 'natural life' and the life 'in agreement with nature'? If it is natural for human beings to reach agreement with nature, then agree- ment must be included in the natural life, though not identical with it. One way of seeing the point of the distinction is to consider it in the context of human development. A child may lead a natural life in the sense of behaving in the way nature has planned for her, but until she comes to realize that she is acting in accordance with nature's plan and accepts this as the best way of organizing her life, she will not be living in agreement with nature. For more on this distinction see sect. 5. There is also an illuminating discussion of the difference between the natural life and the life in agreement with nature in 1\. P. White, 'The Role of Physics in Stoic Ethics', Southern Journal of Philosophy, supp!. 23 (1985), 57-74. 5 One might perhaps add a passage in Epictetus (Diss. I. 6. 16-20) that is not really meant to state an argument, but mentions 'conduct in harmony with nature' (uv",<pwvov 8'E!;aywy>]v Til <pVUH) alongside contemplation (8Ewpia) and 'following up the con- sequences' (7TapaKoAov8YJaLs) as one of the functions of reason. 6 Gisela Striker however been made more complicated by commentators who tend not to distinguish arguments by their explicit conclusions, and take for granted some of the Stoic identity-theses like 'living in agreement with nature is the same as living virtuously', or 'virtue is the same as perfect rationality'. Perhaps it helps to keep in mind that we should concen- trate on the concept of agreement with nature and not on any of those other descriptions of the end. I begin with a brief analysis of the Cicero passage because it seems to be the source of the most influential standard account of Stoic ethics. 6 Cicero ostensibly bases his argument about the goal oflife on the famous Stoic theory of natural concern (oikeiosis),7 according to which all animals, including humans, are endowed by nature with an instinct for self-preservation. That is, more precisely, they are born with some not very articulate conception of their own selves, or their 'constitution' (sustasis), and they are inclined to be concerned about it. This instinct enables them to recognize and pursue things needed for self-preservation, and to recognize and avoid things that might lead to their destruction. Thus their lives can be seen to be regulated by the overall goal of keeping themselves alive and healthy and, at least in the case of humans, by the inclination to use and develop their various capacities. As a human being matures, she comes to recognize more and more things as belonging to her (oikeia) or being in accordance with her nature (kala phusin), and as she follows the rule oftaking what accords with her nature, rejecting what goes against it, her conduct eventually comes to exhibit a regular pattern that is, according to Cicero, consentaneumnaturae (Fin. 3. 20), in agreement with nature. The details of this development are obscure, but this need not concern us at the moment. The next step is crucial: when a person has reached " The classical statement of this view is in M. Pohlenz, 'Die Oikeiosis', in id., Grund- fragen der stoischen Philosophie (Giittingen, 1940), 1-47. The most recent versions of this interpretation, quite different in detail but assuming the same strategy of argument, are B. Inwood, Ethics and Human Action in Early Stoicism (Oxford, 1985), ch.6, and T. Engberg-Pedersen, 'Discovering the Good', in M. Schofield and G. Striker (cds.), The Norms r1Nature (Cambridge, 1986), 145-84. I have discussed the alleged 'argument from oikeiosis' in more detail in 'The Role of Oikeiosis in Stoic Ethics', Oxford Studies in Ancient Philosophy, I (1983), 145-67. The present section is an attempt to arrive at a clearer version of the actual Stoic argument for the goal of agreement with nature, which was very inadequately sketched in the earlier paper. 7 The Greek term seems impossible to translate adequately, as one commentator after the other has pointed out. My rendering provides, I hope, an expression that is neither all too clumsy nor all too misleading. For a detailed discussion of the meaning of oiKEiwat> see S. G. Pembroke, 'Oikeiosis', in A. A. Long (ed.), Problems in Stoicism (London, 1971), "4-49. Following Nature 7 this stage, she comes for the first time in her life to understand 'what really deserves to be called good' ('quod vere bonum possit dici', Fin. 3.21). She sees what Cicero describes as 'the order and, so to speak, harmony of conduct' and comes to value such order and harmony above anything she had valued before. Thus she arrives 'by insight and reasoning' (cognitione et ratione) at the conclusion that the human good, and the reference-point of all action (that is, the goal oflife), is what the Stoics call homologia, consistency and coherence (Cicero's term is convenientia). Cicero's description ofthis process is vague, but we may assume, I think, (1) that 'seeing the order .. .' means seeing that there is a certain order and harmony in one's conduct, and (2) that the term homologia refers to this very order. So we can perhaps say that the agent realizes that there is some admirable order in her way of acting, and then decides that this is the highest good. This is as far as Cicero goes. Apparently he thinks he has shown what the highest good is, and that it is living in agreement with nature (ibid. 26). Since the regular pattern of action has just been described as agreeing with nature, we are probably meant to understand that 'order and harmony' and 'agreement with nature' refer to the same thing, though one can hardly say that Cicero explains why this should be true. But he offers no further argument, and proceeds to take the Stoic thesis as established. Now I think it is fairly clear that Cicero has produced no such thing as an argument to show that the end is living in agreement with nature. The story he tells presupposes that the highest good is order and harmony, but it says nothing about why that should be so. For to say that someone realizes that X is the highest good is indeed to imply that X is the highest good, but it is not an argument to show that what the person allegedly realizes is in fact the case. So Cicero's argument is at best incomplete; at worst, it is lj. confusion, since he presents as an argument for the thesis that 'order and harmony' are the human good a story that is based upon this very thesis. I think Cicero's account is a confusion; but let me first turn to the standard interpretation of the whole passage. Since Cicero explicitly tells us that the theory of oikeiiisis is the proper starting-point of ethics (Fin. 3. 16), but does not explain what the insight or reasoning is that leads one to conclude that order and harmony are the highest good, commentators have looked at other accounts of oikeiiisis to supplement Cicero. The most influential text seems to have been a passage from Seneca (p. 121. 14-18), in which 8 Gisela Striker Seneca sets out to show that 'man is concerned about himself not qua animal, but qua rational'. He argues that concern for self-preservation takes on different forms according to the stages of human develop- ment, and that its final stage is concern for a rational being. The 'insight' that Cicero mentions also comes at the final stage of human development, and so commentators have inferred that one comes to realize that one's true selfis reason, and thereafter ceases to care about everything except one's rationality. Now even if it were granted that man's true self is reason, it would be strange to see self-love turn into an exclusive concern for rational- ity, since presumably one does not cease to be an animal with various other needs that must be looked after for the sake of self-preservation. (This is in fact the main objection Cicero raises against the Stoic doctrine in Fin. 4.) Secondly, and more importantly, the ostensible demonstrandum of Cicero's argument is not that one comes to value rationality, but that one comes to value order and harmony above all else. Why should love for rationality--one's own rationality-be the same as love for the order and harmony of nature? Pohlenz (Die Stoa [Gottingen I959], II7) suggests that these are the same because, according to the Stoics, reason is the same in the universe as in man, and thus concern for one's own rationality is eo ipso concern for nature's rationality; and that can be described as a desire to live in agreement with nature. But this seems to me to be a very peculiar argument. First, we should at least expect to be told that the develop- ing human being comes to believe, at some stage, that his reason is identical with, or part of, universal reason. Second, the identity between human and cosmic reason can hardly be more than identity in kind, or possibly, since the Stoics supposed reason to be pneuma, some sort of airy stuff, an identity in material, as between a glass of water and water in general. But concern for my glass of water is not concern for yours, or for the whole mass of water in the universe. Why, then, can we equate concern for one's own rationality with concern for the rationality of nature as a whole? So long as there is no explicit testimony for it, I should hesitate to saddle the Stoics with this feeble reasoning. It might seem simpler to argue that concern for one's rationality implies the desire for agreement with nature because it is rational to follow nature. But to do so would open up another gap in the argu- ment: why is it rational to follow nature? Besides, it seems then super- fluous to say that the desire for agreement with nature is a form of Following Nature 9 self-love. We could simply argue that man desires agreement with nature because he is rational, not because he wants to be. And in fact, while Cicero says nothing about recognizing one's true self, he does indicate that it is reason that leads one to value order and harmony above all else. Nor does he say that the concern for self-preservation takes on a different form at the moment of the alleged insight; he seems rather to suggest that self-preservation is replaced as a primary goal by the desire for order and harmony, once one has achieved the crucial insight about the good. And so we are brought back to the question the commentators have tried to answer through the theory of concern for one's self. What is the reasoning behind this insight? It would seem that recognizing oneself as a rational being is not enough to explain the exclusive desire for order and harmony, or agreement with nature, though reason is apparently needed to arrive at the postulated insight. The connection between reason and agreement with nature as the goal oflife recurs in the argument reported by Diogenes Laertius, and I think this might help us to see more clearly why the Stoics might have thought that it is rational to follow nature. Diogenes Laertius' argument starts 8 from a consideration of how the lives of different kinds ofliving things are organized. Nature her- self regulates (oikonomei) the life of plants and some parts of animal life. Animals besides humans have impulse (hormi) to direct them towards what belongs to them; so for these it is natural to have their lives organized (dioikeisthai) by impulse. Rational animals have been given reason 'in accordance with a more perfect form of management' (kata teleioteran prostasian); hence it is natural for them to live by the guidance of reason (kata logon), which supervenes as a 'craftsman of impulse'. From this statement Diogenes Laertius jumps with a rather abrupt 'therefore' (dioti) to the thesis, 'first proposed by Zeno in his book On the Nature o/Man', that the end for man is living in agreement with nature. Here we find an explicit transition from being rational to following nature. To understand the 'therefore', and probably also to understand the insight postulated by Cicero, we need an argument to show why agreement with nature must be the ultimate aim of a life guided by reason. This is the place, I think, where Stoic ethics draws upon Stoic 'physics', that is to say, in this case, Stoic theology. According to 8 I assume that the actual argument begins at 86, not 85, as most commentators have assumed. For a defence of this view see 'The Role of Oikeiosis'. IO Gisela Striker Plutarch (Stoic. repugn. I035c-D), Chrysippus would never begin any treatise on ethics without a preface about 'Zeus, Destiny, and Provid- ence'; and he used to say that 'there is no other or more suitable way of approaching the theory of good and evil, or the virtues, or happiness (than) from universal Nature and from the dispensation (dioikesis) of the universe' (trans. Cherniss). We no longer have any of the prefaces cited by Plutarch, but it is not difficult to show, for example from Cicero's account of Stoic theology in the second book of his De natura deorum, that the Stoics were committed to the thesis that the order of the universe is the best possible rational order. Hence if it is natural for human beings to be guided by reason, as Diogenes Laertius says, then following nature might be the best way to provide rational order in one's life-and this would explain the transition from rationality to agreement with nature. Let me first set aside an argument suggested by Stoic determinism and graphically illustrated by the famous comparison of man's rela- tion to nature to that of a dog tied to a cart (Hippolytus, Philosophou- mena 21. 2). The dog will be better off if it follows willingly; if not, it will be dragged along. This suggests that it is rational to follow nature because one cannot avoid it, and it might be better to accept the unavoidable than to rebel against it. But it seems all too obvious that the dog would be still better off if it were not tied to the cart-fatalism is not usually a form of optimism. Following willingly can only be the best if one is convinced that it is good to be led in this way. And this is indeed, I think, what the Stoics would want to argue. The key term in the argument reported by Diogenes Laertius, which recurs in the description of all three types oflives, is some expression for organiza- tion. And the reason why a rational creature will want to organize its life by living in agreement with nature should be that nature's organ- ization provides the best rational order. The Stoics argued on a variety of grounds that universal nature must be rational. One prominent argument was, for example, that the universe is superior to all its parts, including man, but man is superior to the rest in virtue of possessing reason; so the universe must possess reason too (ND 2. I8ff.). This cosmic reason is what organizes and governs the world, as can be seen from the order, regularity, and coherence of all its parts; and reason organizes the world in the best way possible (ND 2. 80-1, 86-7). Reason, then, is presented as occupying the highest place in the hierarchy of things, and its good- ness or superiority is explained in terms of order and harmony. Hence Following Nature II I would suggest that the understanding of the good that a mature human person reaches, according to Cicero, might consist in realiz- ing, not that human beings are essentially rational, but that nature's order is the best, or indeed 'what really deserves to be called good'. Such an interpretation would seem to be confirmed by the frag- mentary remains of the Stoic theory of the good, mentioned only very perfunctorily by Cicero (Fin. 3. 33), but set out in slightly more detail in some of Seneca's letters. Cicero mentions with approval a definition of the good as 'what is perfect according to nature' ('quod esset natura absolutum', presumably a translation of to teleion kata phusin). I take it that this is a definition of the word 'good', which says that 'good' means 'perfect', so that a good X is a perfect X, one that is complete in its own nature. Seneca uses the same definition in Ep. 76. S-11. But as Seneca points out later, in Ep. 124. 13, the definition is actually too generous, since it allows us to speak of goodness also in respect of plants or irrational animals. These, however, are called good only as it were by courtesy (precario ). Real perfection appears only in beings of a higher order-those that possess reason, that is, humans and gods. 'These other beings are perfect only in their own nature, not truly perfect, since they lack reason. Only that is (really) perfect which is perfect in accordance with universal nature. But universal nature is rational' (Ep. 124- 14, cf. also Epict. Diss. 2. S. 2-3 for the point that goodness in the strict sense presupposes rationality). It seems, then, that 'good' ought to be defined more narrowly as 'perfection of rational nature' -and this is indeed a definition independently reported by Diogenes Laertius (7. 94: TO TEAELOV KunJ. q,VULV AOytKOV cOs AOYtKOV). The definition does not tell us, of course, what it is that perfects rationality. But there is good reason to think that the good- making features are the order and harmony recognized, according to Cicero, as what is truly to be called good. Seneca remarks (Ep. 124. IS) that what is good is never disorderly or confused (inordinatum aut turbidum). And it is precisely the orderliness and coherence of all parts of the universe that makes the world so admirable and leads the Stoics to conclude that it must be governed by reason. This seems also to be the notion of the good we grasp, again according to Cicero (Fin. 3 33) and Seneca (Ep. 120. 5), by analogy, ascending from things that are easy to observe to those that must be grasped with some effort. As there is health in the body, so there is health in the soul, and both are kinds of ' good mixture' (cf. Stob. 2. 62. 20-63. 5). Again, what makes us recognize perfect virtue is its 'order, beauty, constancy, and the 12 Gisela Striker harmony of all actions among themselves' (Sen. Ep. 120. II). 'What really deserves to be called good' seems to be the order and harmony produced by rational organization, and this is exhibited to the highest degree by the organization of the universe. One might still ask why it would follow that the rational order of the universe is also the best order for a human individual. Could one not try to imitate or reproduce nature's admirable order and harmony in other ways than by adapting oneself to what one takes to be nature's will? It does not seem self-evident that what is good for nature as a whole must also be good for each of its parts. But the Stoics, of course, denied that there could be a discrepancy, let alone conflict, between what is good for the whole and what is good for a part. They claimed that nature is not only rational, but also provident and benevolent, so that the best order of the universe must eo ipso be the best for each part of the whole (for some most emphatic statements of this point see Marc. Aur. Med. 2. 3, 2. II). We can now try to put together the reasoning that should lead a fully developed person to the conclusion that the human good must be a life in agreement with nature. It would go roughly as follows. First one discovers the order and harmony of one's own natural conduct. Then one realizes that this must be a result of rational planning by nature, since real goodness can be found only where there is rationality. (This is probably what Cicero indicates by saying that one 'grasps the notion' -presumably of goodness.) As a rational being one then comes to think that the best or only way to bring order and harmony to one's own life consists in following nature. But the final good is the best human life-therefore it must be a life 'in agreement with nature'. The argument does not say anything about impulse or desire. But the end was also defined as the ultimate aim of desire; and so one might well ask why it is that a creature made to care about its own preservation should eventually come to abandon this concern and be interested only in agreement with the rational order of nature. I think that the account of human development given by Cicero was probably intended to answer precisely this question. It might be useful to remember at this point that the Stoics held the end to be what one should desire, not what every one of us does desire. The Stoic concep- tion of the end does not arise as a natural continuation of one's concern for self-preservation, but rather as the result of one's reflec- tion upon the way nature has arranged human behaviour in the con- text of an admirable cosmic order. Whether a person will come to Following Nature 13 have the right conception of her final good, and hence the right sort of desire, depends upon her ability to reach the stage of reflection that will make her realize what really deserves to be called good. This is what she ought to desire, according to the Stoics, not only because nature has provided her with reason, but also because it is both necessary and sufficient for a happy life. It remains true, as Seneca says (Ep. I21. 17), that she will desire the good for herself, but her desire for agreement with nature will not simply be an en- lightened form of self-love. If this reconstruction of the Stoic argument is correct, then the foundations of Stoic ethics are to be sought, as Chrysippus said, in cosmology or theology, and not in human psychology. Perhaps we can now explain the connection between Zeno's first definition of the goal of life, 'living consistently', and the later official formula, 'living in agreement with nature'. Both can be taken as ways of saying what it is to lead a life guided by reason. But while the first mentions only the consistency or harmony that should characterize a good life, the second introduces the standard that a rational creature would look to in order to achieve a harmonious life. To describe the life guided by reason merely as consistent or harmonious would be rather vague, and invite the question, to pick up Stobaeus' explana- tion of the first definition, of what the unifying principle might be that produces order and harmony. Given the Stoic theory of nature, the answer would be that it is agreement with nature's will, or observation of the natural order, that provides consistency. Thus one can quite well imagine that Zeno used both versions of the definition, not indeed to say the same thing more or less completely, but to explain how order and harmony can be attained. And I would also say that agreement with nature proved to be the theoretically more fruitful and important notion for, by contrast with the bare appeal to consistency, it offers a basis for an account of how one should live by the guidance of reason-namely, as Chrysippus expressed it most accurately, by adhering to what experience shows to be the natural course of events. If Stoic ethics itself had some kind of order and coherence, it was provided, as Chrysippus seems to have thought, by the starting-point, universal nature and its organ- ization. Gisela Striker 2. Chrysippus vs. Aristo: establishing orthodoxy When the polymath and geographer Eratosthenes came to Athens around the middle of the third century BC, he thought he had hit upon a fortunate moment, since two great philosophers, Arcesilaus of the Academy and Aristo the Stoic, were flourishing in the same place at the same time (cf. Stoicorum VeterumFragmenta [SVF], ed. I. von Arnim [Berlin I903-6], i. 338). Strabo, who reports this, remarks upon Eratosthenes' lack of judgement for thinking of these two as the out- standing philosophers of the period-both men who no longer had any followers in his own, Strabo's, time-while not even mentioning Zeno and his successors, that is, Cleanthes and Chrysippus. But of course it is most likely that Eratosthenes' impression reflected the actual reputation of the philosophers at his time, and this seems to indicate that Zeno's successors were not considered to be very prominent. Chrysippus became head of the Stoa only in 232 BC, though he must have been an important member of the school long before; but the establishment of Zeno's school as one of the leading, if not the most important, seems to have been mainly due to him. He is credited with defending Stoic doctrine both against criticism from the outside and against heresies within the school, and also with providing more systematic foundations for Zeno's views. It was Chrysippus who defended the Stoa against the objections of the sceptical Academy (Plut. Comm. not. I059B), and he appparently once and for all eclipsed Aristo, who had no more followers after Chrysippus. Now Aristo was himself a pupil ofZeno, and considered a Stoic; so the debate between him and Chrysippus was an internal affair of the Stoic school, con- cerning, so it seems, the foundations of Stoic ethics. Later critics of the Stoa like Carneades, Antiochus, Plutarch, or Galen sometimes appealed to Aristo's views in order to argue that Chrysippus' innovations, if such they were, did not represent Zenonian orthodoxy, or that Aristo was more consistent than the official Stoa. Thus Cicero, following Antiochus, argues throughout book 4 of the De finibus that the Stoics, to be consistent, would have either to adopt the view of Aristo, and thus keep their dogma that virtue is the only good, or admit that their doctrine did not differ at all from that of the Peripatetics. I think that an investigation of the dispute between Chrysippus and Aristo over the definition of the goal oflife can show why the Stoics would not have wanted to follow Aristo, Following Nature IS and might serve at the same time to bring out the point of the concep- tion of the human good encapsulated in the formula 'living in agree- ment with nature'. Aristo was one of those Hellenistic philosophers who, like the Cynics, rejected all parts of philosophy except ethics. 9 He claimed that physics is above us, logic is no use, since it does not make us lead better lives; the only topic that concerns us is ethics. For this view he could appeal to the authority of Socrates (cf. SVF i. 351-7). Even ethics he wished to restrict to a very few fundamental principles, leav- ing aside detailed advice for various types of everyday situations as being more appropriate for nurses and nannies (Sen. Ep. 94. 2). It is not surprising that such a man should not accept an account of the goal of life that involved a reference to nature and hence depended upon what was then called physics. Aristo declared that the goal oflife was indif- ference (adiaphoria), or, more precisely, indifference to what is between virtue and vice, or what is neither good nor bad. We do not know how he argued for this thesis. Cicero, who tends to group him with Pyrrho the Sceptic, suggests in one place that he might have appealed to ataraxia, the kind of peace of mind that allegedly arises from not caring about anything except moral character and hence being free from sorrow, desire, and fear (Fin. 4. 69: 'vives ... magnifice atque praeclare, numquam angere, numquam cupies, numquam timebis'). However that may be, at the moment I wish to look at what appears to have been Chrysippus' main objection to Aristo's doctrine of indifference as the goal oflife. This objection is stated most clearly, I believe, in a passage in Plutarch's De communibus notitiis (1071 F-1072A): What would you say is more at odds with the common conception than that people who have neither grasped nor got a conception of good desire the good and pursue it? Because you see this is rather the perplexity to which Chrysip- pus also reduces Aristo, on the ground that things (do not provide) for getting the notion of indifference to what is neither good nor evil if there has not been a prior notion of the good and the evil, for thus the state of indifference would obviously have subsistence prior to itself, if a conception of it cannot be had without prior conception of the good, but only itself and nothing else is the good. (Trans. Cherniss [Loeb Classical Library, 1976], with modifications) 9 Aristo's philosophical views and his position among the first generation of Zeno's students are discussed in detail in A. M. loppolo, Aristone di Chio e 10 stoicismo antico (Naples, 1980). The fragments are collected in SVF i, pp. 73-90. Gisela Striker The introductory sentence seems to be Plutarch's, and it is not quite clear why he thinks it describes the same difficulty as the sub- sequent argument. It is not even clear that what he describes is para- doxical at all, for why should it be absurd to say that people desire and pursue something of which they have no conception? Freud was certainly not the first to say that some people are ignorant of what they really want. To see what Plutarch had in mind, we must look at the argument he describes. It appears that Aristo, according to Chrysippus, was telling his hearers what the good was in terms of the good itself. The telos was said to be indifference, which in turn was explained in terms of indif- ference. This seems at least to be the point of the argument explicitly attributed to Chrysippus by Plutarch. The argument is somewhat elliptically stated, but it is fairly clear that it went as follows: (1) One cannot get a conception of indifference to what is neither good nor evil unless one already has a conception of good and evil.-This means, I suppose, that one cannot understand the expression 'indifference to what is neither good nor evil' unless one already understands the words 'good' and 'evil' which occur in it; and that seems uncontroversial. From (1) it is inferred, though not explicitly stated, that (2) Good and evil must subsist prior to indifference to what is neither.-If this does not go beyond the first premiss, it should say that the notions of good and evil must be grasped before the notion of indifference to what is neither good nor evil can be grasped. lo But, according to Aristo, (3) Indifference and nothing else is itself the good; hence it seems to follow that (4) Indifference must subsist prior to itself; which is absurd. The schema of the argument is: B is grasped prior to A; A is identical with B; hence A is grasped prior to A, which is absurd. What is the force of this argument? Whether it is valid depends, I think, on how we are to take the identity-premiss (3), 'Indifference and 10 That the Greek term used here, vcpiaTaa8aL, means 'be grasped (or graspable) by thought' is shown by a later passage in Plutarch (108IF; cf. I08Ie, 10661'), in which Chrysippus is reported to have said that the past and the future only subsist, but do not exist (v7TliPXELv), while only the present exists. Also, Sextus Empiricus uses the word 7TpoerrLvoEia8aL (to be grasped before) as a synonym of 7Tpo';cpEaTaVaL in an argument of the same type; cf. M. 11. 186. Following Nature nothing else is itself the good'. If this is taken merely to assert that indifference (to what is neither good nor evil) and the good are one and the same thing, which might be a contingent fact, then the argu- ment is invalid. For the first premiss is about conceptions of things, the second about the things we have conceptions of; and from the premiss that A and B are identical it does not follow that grasping the notion of A is also grasping the notion of B. To use an example suggested by Sextus, writing and exercising the skill of writing may be one and the same thing, and yet the notion of writing is prior to the notion of exercising the skill of writing. Obviously, the notion of exer- cising the skill of writing is not therefore prior to itself. But perhaps the identity-premiss should be given a stronger inter- pretation, namely that the notion of indifference is identical with the notion of the good. This could not be due to 'indifference (etc.), and 'good' having the same meaning, of course, but it might be said to hold because, within some theory, the term 'good' is to be explained or defined by means of 'indifference'. In other words, the identity- premiss might be an explanatory definition or a statement of essence, so that to understand, not just the word 'good', but what the good is, is to see that it is indifference. This is a stronger sense of 'grasping the notion of' something than was suggested by the initial example of an expression contained in another expression, but it could be applied there too. With this interpretation the argument becomes valid. As an objec- tion to Aristo's theory, its point would be to show that the theory is circular and uninformative, since it defines the good, indifference, in terms of itself. I think that this was indeed Chrysippus' point-and furthermore, that he was right, though of course it is not likely that Aristo would ever have said anything as blatantly circular as 'the goal of life is indifference to what is neither indifference nor its opposite'. From the few things we hear about Aristo's doctrine we can, however, reconstruct the reasoning that would have led Chrysippus to his accusation of circularity. According to the sources, Aristo held the following theses: (1) The goal of life is indifference to what is between virtue and vice; and (2) The only good is virtue, the only evil is vice. Now we can assume without explicit testimony for Aristo that the goal is the human good, since this was a presupposition in arguments about I8 Gisela Striker the goal oflife. But if, according to (2), the only good is virtue, then the human good, indifference, must be identical with virtue. That this was indeed Aristo's view is confirmed by Cicero, who lists Aristo among those who held that 'living virtuously' (honeste vivere) is the goal (Fin. 4. 43).11 Supposing that substitutions based on identity are permissible, we can already infer the rather circular-sounding proposition that the goal is indifference to what is between indifference and its opposite, or, with Plutarch, that the good is indifference to what is neither good nor evil. But this might be as harmless as saying that medicine is the endeavour to bring about the object of medicine, which, for all its obvious uninformativeness, neither is false nor implies that we cannot say what medicine is without previously saying what the object of medicine (so described) should be. The thesis that indifference is an attitude directed towards what is neither itself nor its opposite is reached via the premiss that indifference is identical with virtue. Now if there were some way of explaining what virtue is that did not bring in indifference-if virtue were not defined as indifference-then it would be wrong to say that in order to understand what virtue is, we must understand what indifference is, and vice versa. So we should look at Aristo's account of virtue. According to Galen (De placitis Hippocratis et Platonis [PHP] 7.2.2; SVF i. 374) Aristo held-with Zeno-that virtue was knowledge of goods and evils, and that the four cardinal virtues justice, wisdom, temperance, and courage were this same knowledge applied in differ- ent kinds of situations. With 'goods and evils' we seem to get danger- ously close to virtue and vice again, given the thesis (2) that virtue and vice are the only good and evil, respectively. However, it still seems possible that Aristo might have defined virtue, or the four specific virtues, in some other way, e.g. by saying what just, courageous, temperate or wise action consists in, given certain types of situations like dangers, temptations, distributions, and the like. It appears, however, that Aristo quite explicitly and deliberately refused to do this. His most famous divergence from what came to be Stoic orthodoxy was his rejection of the theory of 'preferred' (proegmena) and 'dispreferred' (apoproegmena) things, the theory that the Stoics used to explain what is appropriate action (kathikon). As we have seen before, the other Stoics held, just like Aristo, that virtue- 11 Strictly speaking, virtue and living virtuously are not the same, but I take it that 'virtue is the goal' is an abbreviated way of saying that living virtuously is the goal, since the goal-happiness-had to be some sort oflife. Following Nature understood as agreement with nature-was the only good; but they nevertheless admitted that other things could have more or less value depending on whether they were more or less in accordance with human nature (kala phusin or para phusin). In order to distinguish this kind of value from the absolute value of virtue, they had introduced a terminological distinction, saying that natural things are 'preferred', though not good, and unnatural things 'dispreferred', though not bad. Appropriate action would consist in taking or doing what is in accord- ance with one's nature, hence 'preferred', and avoiding what goes against one's nature and hence is 'dispreferred'. Virtuous action could then be described as doing the appropriate action because it is in agreement with nature to do this. If Aristo had accepted this doctrine, he could have used it, perhaps, to explain how the wise person will act in the kinds of situation relevant for the different virtues. But Aristo argued, on the contrary, that nothing except virtue could be said to be naturally preferable or more valuable than other things (SVF i. 36I = Sextus M. I1. 64, cf. PH 3. I92). SO, for example, health, which was held to be naturally preferred, would not be preferable to illness in a situation where some tyrant had decided to press into his service, and thereby ultimately to kill, all and only those in good health, so that only the sick would avoid his service and thereby also avoid destruc- tion. In general, Aristo said, what is to be preferred depends entirely upon circumstances (peristasis), and just as in writing we determine which letter comes first by what word we want to write and not by any intrinsic order of preference among the letters we use, so what is to be preferred is a question of the given situation, not of any natural order of value among the things we opt for or against on each occasion. Aristo's example-and we have no other-is unconvincing, for it seems pretty obvious that the wise man who opts for illness rather than health in such an exceptional situation is not preferring illness to health, but rather survival with illness to destruction with health; that is, he thinks, quite plausibly, that survival is more important than health. Which does nothing to show that health is not in general preferable to illness when there is no disadvantage connected with health. So this example is not likely to convince anyone that there is no natural order of preference among external things or bodily states. In fact, the other Stoics apparently took care of Aristo's example by saying that some things are appropriate aneu peristaseos, when there are no exceptional circumstances-for example, looking after one's health-others, such as maiming oneself or throwing away one's 20 Gisela Striker fortune, only under exceptional circumstances (D.L. 7. 109)' The simile of the letters has no force by itself and might indeed be rejected as misleading, since there is usually only one way of spelling a given word, while there may be many different ways of reaching a certain result, even in special situations, and we might well prefer one means of action to another apart from the circumstances. Aristo, however, insisted upon the total indifference of things 'between virtue and vice', on the ground that calling some of them 'preferred' was equivalent to calling them goods, and he like the other Stoics wished to make it absolutely clear that nothing but virtue could be a good. Like many later critics, he evidently rejected the kind of double evaluation system the Stoics had introduced by saying that the 'preferred' things were indifferent indeed, of no value with respect to happiness, but did in spite of this have different values with respect to impulse, or to their contribution to a life in accordance with nature (cf. Stob. 2. 80. 9-13; Cic. Fin. 4 47). On the other side, Chrysippus argued that declaring all bodily and external things to be totally indifferent and valueless would leave the wise man with no method for making a selection among them. To which Aristo apparently replied that the wise man would do 'whatever came to his mind' or 'whatever occurred to him'. This mysterious doctrine, barely mentioned by Cicero (Fin. 4. 43, 69), is unfortunately not explained in any other source. But whether we take it, with Pohlenz (Die Stoa, i. 123, ii. 70), as an appeal to appearances (phaino- mena) in the manner ofPyrrho, or, with Ioppolo (Aristonedi Chio, 183), as referring to some 'rational judgement' of the sage, who simply sees what is the virtuous thing to do given the circumstances, it seems clear that we are not given an account of correct action. In fact, we might seem to get into another circle here if virtuous action is to be defined as what the wise man would do, for the wise man will of course be defined as the one who knows what is good or evil. This does not mean that Aristo's theory was completely tauto- logous. After all, 'indifference' is not the same word as 'virtue', and Aristo could assume that his audience would have some conception, albeit not a very precise one, of what was meant by such words as 'justice', 'wisdom', etc. So they could understand Aristo's doctrine about the goal of life as telling them that the good and happy person would be concerned about nothing but virtue, and thereby perhaps would lead an undisturbed life. Chrysippus' charge of circularity would be relevant only if one demanded, as Chrysippus probably did, Following Nature 21 that virtue should not be left unexplained or defined by a list of specific virtues, but should be characterized, for example, through some general account of correct action. Chrysippus could plausibly argue that a philosophical theory is supposed to explain, rather than assume, a conception of what constitutes just or wise or courageous conduct. Aristo seems to have thought that there could be no such general account, and hence refused to attempt one. Besides, it may never have occurred to him to apply to his teaching the standards of theoretical coherence Chrysippus invoked in his criticism. But if Aristo's doctrine were subjected to those standards, then Chrysippus' objection would seem to be justified. We cannot explain what indiffer- ence to things between virtue and vice should be unless we know what virtue and vice are, and as far as Aristo's theory goes, the only available answers to the question of what constitutes virtue and vice are-either indifference again, or knowledge of good and evil, that is, once again, knowledge of virtue and vice. Chrysippus could justifiably argue that Aristo's theory offers no advice for action and hence does not really explain what the goal of life consists in or how it is to be attained. Aristo's mistake was not, as Cicero seems to think, that he held such paradoxical views as that it makes no difference whether we are ill or in good health, for in a certain sense the orthodox Stoics believed the same-health or illness are of no value or disvalue with regard to happiness. What is missing in Aristo's scheme is some rule for correct action that will tell us in each case what will be the appropriate thing to do. Now according to Chrysippus that rule was indeed 'select the naturally preferred things', so that we should, barring exceptional circumstances, generally prefer health to illness. But Aristo could in principle have decided to offer as a rule 'do whatever will lead to the greatest pleasure', which does not imply a general ranking of things to be selected, but would have saved him from the charge of circularity, though the Stoics would of course have argued that such a rule did not agree with the standard presuppositions about virtuous conduct (cf. Cic. Fin. 4. 46). Since Aristo to all appearance refused to produce an account of virtue or virtuous action in terms other than good and evil, Chrysippus was right, I think, to say that his theory was ultimately uninformative. But now Plutarch claims, in the paragraphs following the passage just discussed, that the same argument could with equal justice be applied to the orthodox Stoics themselves. For, Plutarch says, if you ask them 'What is good?', they will say 'Nothing but wisdom' 22 Gisela Striker (phronesis); if you then ask 'What is wisdom?', they will answer 'Nothing but knowledge of goods' (Comm. not. I072A-B). Both these propositions are well attested as Stoic doctrine (cf. Cherniss's note ad loco for parallels), and so it seems that Chrysippus landed in his own trap-wisdom is apparently defined in terms of the good, and the good in terms of wisdom. But in this case it is fairly easy to see that the appearance of circular- ity is superficial. Both propositions are orthodox, but the first is not a definition, and so it is not true that the Stoics explained the good in terms of wisdom and vice versa. Rather, as we saw before, goodness should be defined in terms of the rational order of nature, and the human good as agreement with that order. The Stoics could then argue, without circularity, that a human being will have attained the goal of life when he knows 'goods and evils' (that is, I take it, both that the human good is agreement with nature, and what things are good as exhibiting agreement with nature); that a person who has this know- ledge thereby lives in agreement with nature, and hence that a good human life-happiness-is a life of wisdom. The proposition that 'the good is wisdom' is then not a definition of the good in terms of wisdom, but a theorem derived, inter alia, from the more fundamental premiss that the human good is a life in agree- ment with nature. A complete derivation of this theorem from the 'axioms' of Stoic ethics would be quite complicated, and I shall only try to give an outline here, as a warning for commentators who would treat such identity-statements as more or less evident assumptions in dealing with Stoic theory. The derivation, then, could go roughly as follows: (r) Goodness consists in rational order as exhibited by nature. (2) Wisdom is knowledge of good and evil, comprising both know- ledge of what goodness is, and of which things or actions are good or bad. (3) Knowledge of what is good entails acting in a good way.12 (4) The person who has wisdom will act in agreement with nature. (Assuming that 'acting in agreement with nature' amounts to 'exhibiting rational order in one's conduct', this can be derived from (r), (2), and (3)') 12 The Stoics, like Socrates, denied the possibility of weakness of will (uKpaa[a). For their arguments on this point see Inwood, Ethics and Human Action, ch. 5. Following Nature (5) The human good is a life in agreement with nature. (6) The human good is a life of wisdom (from (4) and (5. The last line is a more exact statement of what Plutarch compresses into the somewhat misleading formulation 'the good is nothing but wisdom' 13 -or if this is not so, then one might want to argue that the thesis is not orthodox after all. The possibility of such a derivation shows, I think, that Chrysippus' theory could provide what he said was lacking in Aristo's-a non- circular account of the object of that knowledge which, according to both Aristo and Chrysippus, constitutes virtue. Cherniss (in his note (b) on I072B, p. 76r) points out that the ancestor of Plutarch's argu- ment against the Stoics is to be found in Plato's Republic (505 B 6-c 5), where Plato derides philosophers who say that the good is wisdom or knowledge (phronesis), but who, when asked to specifY what know- ledge, have no better answer than 'knowledge of the good' (phronesis agathou). We know that Chrysippus had read Plato, since he wrote a number of treatises against him, and so it is not unlikely that he had Plato's argument in mind when attacking Aristo, but also that he saw his own metaphysical foundation of ethics as a way of answering Plato's objection. After Chrysippus, Cicero tells us, Aristo ceased to be taken seriously as a philosopher. This is no doubt a slight exaggeration, since Seneca (Ep. 94) shows that Aristo's writings continued to be read, and of course critics of Chrysippus like Antiochus, Plutarch, and Galen kept referring to him, though mainly to point out that Chrysip- pus' doctrines were no improvement, or no closer to Zeno's original version of Stoicism than Aristo's. But it is perhaps no great com- pliment to Aristo that he was most often cited by people who wanted to claim that Chrysippus was just as bad. I suspect, in any case, that Aristo's fame among his contemporaries was due more to his person- ality and his apparent talent for witty remarks than to his philo- sophical acumen. As for Chrysippus, I think we can learn two things from his criticism of Aristo: first, that he was indeed concerned about the systematic coherence of Stoic doctrine, more so perhaps than Zeno or Cleanthes, trying to provide something like an axiomatic structure; and second, 13 Alternatively, one might understand Plutarch as saying that goodness (for human beings) is the same as wisdom. In this case, (6) would have to read 'A good life is a life of wisdom', where 'a good life' is equivalent to 'the human good'. Gisela Striker that for him at least the doctrine of the goal as living in agreement with nature was a conscious attempt to offer an account, in non-circular terms, of what constitutes virtue. 3. Virtue as a craft From the third century BC on we find Greek philosophers not only presenting themselves as members of schools, but also concerned about the philosophical genealogy of their teaching. Both the Stoics and the sceptical Academy claimed Socrates as one of their ancestors. But while the Academics emphasized the aporetic character of Socrates' philosophical method, the Stoics were more interested in the positive doctrines that sometimes seem to lie beneath the surface of the early Platonic dialogues, and were presented more outspokenly in the writings of other followers of Socrates. One of the most striking Socratic inheritances of the Stoic school is the revival of the doctrine that virtue is a craft. The Stoics must have thought that they could overcome the dif- ficulties that had led Plato and Aristotle to abandon or criticize this view; and they certainly thought that they could say more clearly than Socrates what kind of knowledge would be required for the craft-not knowledge of any Platonic Form, but knowledge of nature and its order. So they said that virtue was a craft or art ofliving that produces happiness (techni peri ton bion eudaimoniaspoiitiki, Alex. Aphr. De anima 2.159 34). The evidence for the views of the early Stoics is very fragmentary, and we do not possess a continuous treatise in which their theory of virtue is introduced and defended. Most of what we hear about this topic refers to a debate that took place in the second century BC, when the Stoics came under attack from Cameades, the most important of the Academic sceptics. 14 Cameades' counterpart at the time as head of the Stoa was Antipater of Tarsos, whom most doxographers find inferior to his very famous critic. I shall argue that Antipater was not as weak as people like Plutarch would make us believe. But before I describe the controversy between Antipater and Cameades, I must try 14 I have discussed the debate between Carneades and the Stoics of his time in greater detail in an earlier paper, 'Antipater, or the Art of Living', in Schofield and Striker, The Norms of Nature , r85-204, with which this section overlaps. Hence my treat- ment of the sources here will be somewhat dogmatic. Following Nature 25 to give a sketch of the specific version of Stoic doctrine that Carneades was attacking. Although we do not have a precise record of their arguments, it does not seem difficult to explain why the Stoics thought that virtue was a craft. While the premisses of their argument for this claim are perhaps problematic, the argument itself is not very hard to find. It would begin presumably with the fundamental premiss that the goal oflife is living in agreement with nature. The next step would be a clarification of this formula: living in agreement with nature consists in following the pattern of behaviour that nature seems to have provided not only for humans, but for other animals as well, namely trying to preserve and perfect oneself by selecting the things that accord with one's own nature, rejecting and avoiding things that go against it. 'Natural' or 'preferred' things (ta kata phusin) are those that are needed for self- preservation, like health, strength, intelligence, and the means to keep or acquire these, like wealth and reputation; also, the company and well-being of others, that is, one's family first of all, but also the citizens of one's state, and human beings in general. To be n:jected would be things like illness, weakness, ugliness, poverty, and lack of company. The leading Stoics of the second century, Diogenes of Babylon and Antipater of Tarsos, actually defined the goal of life as 'reasonable selection of things in accordance with nature, and rejec- tion of things that are against nature' (cf. SVF iii. 44-6, p. 2I9, for Diogenes, iii. 57-8, p. 253, for Antipater). (Note that what is said to be the goal is the reasonable selection, not the things selected.) This way of acting was claimed to be the same as living virtuously, and hence we are invited to conclude that a human being will be living in agreement with nature precisely if she acts virtuously. This step is clearly very questionable, but I shall for the moment assume it without asking how it might be justified. It turns out, then, that virtue can be described as a rational disposition in the selection of things natural or unnatural. According to Stoic psychology, the 'governing part' of the human soul (higemonikon), the part that determines how one acts, is reason, and reason alone. Now reason is a capacity for discovering the truth, and ifit functions as it should, it will guide us to pursue whatever it has correctly recognized as being good. There are no irrational sources of motivation in the human soul, contrary to what Plato and Aristotle had maintained, and so once reason has discovered that the good is to be found in the order of nature, it will lead us to act in agreement with 26 Gisela Striker nature. Knowledge of nature's rules will therefore be sufficient for virtue, since that consists in following nature. But knowing the rules and acting accordingly seem to be exactly what it means to exercise a craft, and so virtue could also be defined as the craft concerned with the selection of things according to nature (Alex. Aphr. De anima 2. r61. 5-6). I do not want to claim that this was precisely the argument the Stoics put forward. Certainly the premisses could be introduced in a different order, and with elaborate justification. Also, there were no doubt auxiliary arguments in support of the theory, e.g. in Chrysippus' treatise 'On expertise and laymanship' (Peri technes kai atechnias, D.L. 7.202), cited in a list of works on the virtues. For example, one feature that would lend support to the classification of virtue as a craft was probably the regularity and unfailing correctness that characterize the performance of a craftsman who knows his skill-and also, according to the Stoics, the workings of universal nature herself. Furthermore, such a conception could be used to clarify the relation between virtue and the things to be selected or rejected: those were, according to the Stoics, the materials (hufe) that the virtuous person uses in exercising her craft, and thus were prerequisites, but not constituents, of the virtuous life. But the core of the Stoic theory was, I believe, the idea that knowledge of the rules of nature is necessary and sufficient for attaining the goal oflife, agreement with nature. Carneades' attack on this theory was focused on the two proposi- tions (r) that the goal of life is the reasonable selection of things in accordance with nature, and (2) that virtue is the craft of selecting things in accordance with nature. According to Plutarch (Comm. not. 26, I070F-I07r B), Carneades confronted the Stoics with a dilemma: on the Stoic theory, either there will have to be two goals oflife, namely the acquisition of the 'natural' things on the one hand, and the selection itself on the other; or the goal will be distinct from the reference-point of all action. Both alternatives were unacceptable for the Stoics. They had defined the goal as 'that to which all actions in life are appropriately referred, while it itself is not referred to anything else', hence if the reference- point of all action was not the goal, we would seem to get the contra- diction that this both is-by definition-and is not-by hypothesis- the goal of life. The charge that there were two goals could of course have been avoided by saying that the combination of the two was the ultimate end, as some philosophers were, for example, prepared to Following Nature combine virtue and pleasure (Cic. Fin. 2. 34-5). But the Stoics wanted to hold that success in getting the natural things is irrelevant to happi- ness-that is precisely the point of saying that the reasonable selection is the goal, not the things selected. So they had to insist on a single goal that can at the same time be seen as the reference-point of all action. The argument that leads to the dilemma is set out, I think, by Plutarch (Comm. not. 27, I072C-D) and Alexander of Aphrodisias (De anima 2. I67. I3 ff.), and proceeds as follows: A selection can be called reasonable only if it aims at some goal that is either identical with or contributes to the goal of life. But according to the Stoic theory, the acquisition or possession of the natural things is not the goal of life, nor does it contribute to happiness. It follows that there will be two goals-one for the selection, namely attaining the natural things, the other, as the Stoics say, the reasonable selection itself. For if attainment of things natural is not a goal, then the selection will not be reason- able-or else we shall get the other horn of the dilemma: the reference-point of all action will be distinct from the goal of life. The crucial move in this argument is the assumption that a selec- tion can only be reasonable ifit aims at something that has value for, or is identical with, the goal oflife. Since the Stoics certainly did not want to make attainment of things in accordance with nature part of the goal, they could refute this argument only by showing that it is not unreasonable to select things that do not contribute to the final good. And this is what Antipater set out to do. As a reply to the 'two goals' objection Cicero introduces (Fin. 3. 22) a simile that seems to have been used by Antipater. Imagine an archer who is preparing to shoot at a target. According to the Stoics, the primary goal of his action is not-as one would expect-hitting the target, but 'doing everything in his power to hit the target'. The word- ing shows that the simile was meant to illustrate a new definition of the goal that Antipater had introduced: 'doing everything in one's power, constantly and without wavering, in order to attain the things in accordance with nature' (Stob. 2. 76. I3-IS). The point of the archer simile does not seem hard to discover. One can easily imagine a situation in which the proposed description of the archer's action would be quite appropriate. For example, the archer might just be shooting in order to practise his skill, or he might be 28 Gisela Striker complying with an order, fulfilling a task that was not his own concern. If under such circumstances the wind were to carry away his arrow, he would not find that he had failed to do what he had set out to do, for he would have practised none the less, or he would have fulfilled his task, and that was what he primarily intended to do. Hence one might say that his proper goal in this action was fulfilling a task; hitting the target would be a subordinate goal, one that he tried to attain only in order to reach his real, primary goal. And actually hitting the target would not even be a necessary prerequisite for that. What holds for the archer in such a case is supposed to hold quite generally for the virtuous person who tries to attain the natural things: the primary goal of all her actions is following the rules of nature, and she is trying to get these things only because nature has ordered her to do so. Possession of things in accordance with nature is at most a sub- ordinate goal-one that determines, indeed, what steps she takes, but not the end for the sake of which she is acting. Understood in this way, the archer simile might seem to offer quite a good example to demon- strate that it need not be irrational to attempt to achieve something that has little or no value in itself, and that the value of an action may occasionally lie in the correct performance rather than the result. Indeed, a number of commentators have thought that the archer simile provides a brilliant refutation of Carneades' objections, which turn out to be rather superficial and simply due to lack of under- standing. And so it might seem-until one realizes that those objections concern, not the aims of particular actions in particular circum- stances, but rather the goal of an entire craft. Carneades argued (Cic. Fin. 5. 16-20) that if there was a craft of selecting things in accordance with nature, then the attainment of these things, and not its own exercise, must be the goal of the craft. In other words, the question he raised was not, as the archer simile seemed to suggest, whether it could be reasonable on a particular occasion to aim at something that is not the primary goal of one's action, but whether there could be a craft that consisted in trying to achieve a result that was not itself the goal of the craft. In order to maintain this, one would have to say that the craft was normally exercised, not for the sake of the intended result, but for some other end; and presumably also that the craft had been invented for the sake of this higher goal. Seen in this perspective, the archer simile loses its plausibility. For who would want to say that archers normally shoot, say, for the sake of Following Nature 29 practising their skill, or that archery was invented for the sake of shooting correctly? The fact that a different goal may be pursued under exceptional circumstances does nothing to show that hitting targets is not in general the goal of bowman ship. However, it seems clear that Antipater maintained there were entire crafts the goal of which was not the result the craftsman would normally try to achieve. He apparently appealed to the so-called 'stochastic' crafts, like medicine, navigation, or rhetoric, to cite the standard examples. What is characteristic of these crafts is that the result of the craftsman's action depends not only upon his expertise and correct performance, but also upon a number of incalculable external factors, such as the nature of the patient or the disease, or the weather, or the mood of the audience, that need to be favourable in order for the craftsman's action to lead to success. Antipater seems to have argued that the goal of these crafts is not the intended result-the cure in the case of the doctor, a safe journey in the case of the navigator, etc.-but rather the complete performance of all that belongs to the craft. The arguments for this claim are reported (anonymously) by Alex- ander of Aphrodisias (Quaestiones 2. 61 ff., cf. SVF iii. 19). The differ- ence between stochastic crafts and others like shoemaking or weaving does not lie in a different relationship between the craftsman's performance and the goal to be achieved, but rather in a difference between kinds of goal. In the case of the non-stochastic crafts, correct performance generally leads to the intended result, and if the result is not forthcoming, we conclude that the performance was incorrect. For example, if someone tries to write a certain word, and then produces a string ofletters that do not represent the word she set out to write, we conclude that she does not know how to spell. In the case of the stochastic crafts, however, the technical performance may be perfectly correct and yet the intended result not be attained. We do not neces- sarily conclude that the doctor made a mistake if her patient is not cured, and we do not blame the navigator ifhis ship sinks in a storm in spite of all his efforts. What we should infer from such examples, according to Antipater, is that the goal of the stochastic crafts is not the projected result, but rather the correct performance of all that belongs to the craft, as Alexander puts it, or, to use Antipater's formula, 'doing everything in one's power to attain the intended result'. It seems to me that this argument has at least some prima-facie 3 0 Gisela Striker plausibility. It is based on the implicit assumption that the goal of a craft should be whatever makes us decide that the craftsman has performed well. While producers of artefacts will be judged in terms of the excellence of their products, doctors and navigators are judged in terms of their technically correct or incorrect performance. More- over, it is easy to see why one might wish to compare virtue with this kind of craft, for it seems characteristic of moral evaluation that it does not refer to the success of an action, but rather-to emphasize the analogy-concerns the point whether the agent acted in accordance with moral rules or not. Still, it seems more plausible to follow Alexander, who at this point introduces a distinction between the goal of a craft and the task or function (ergon) of the craftsman, and says that while the goal, even in the stochastic crafts, is the intended result, the task of the performer consists in doing everything in his power to achieve the result. IS In the stochastic crafts, one may complete one's task and yet not achieve one's goal, whereas in other cases completion of the task and attain- ment of the goal will coincide. Even in the stochastic crafts, though, the intended result remains 'that for the sake of which' everything is done, and the reference-point of all intermediate action. For even though a doctor may on occasion justify his prescription by saying that this is the kind of treatment prescribed by the book, it is clear that the recommendation of a specific drug for a specific kind of disease can only be justified in terms of its contributing to the restoration of the patient's health. Were we to follow Antipater, we would have to say that medicine is practised for the sake of correct therapy. This is not only an uncom- fortable thought for patients. Rather, it seems perfectly plain that the craft of medicine was invented because people are interested in the conservation and, if need be, restoration of health, and that every particular activity of the doctor will have to be justified and explained " Alexander is obviously developing a point made by Aristotle with respect to dialectic (Top. I, IOIbS-IO) and rhetoric (Rhet. I, 135sbro-I4), as is shown by a similar discussion of the differences between stochastic and 'productive' crafts in his comment- ary on Topics 1. 3 (pp. 34-6). Although Aristotle does not make an explicit termino- logical distinction between i!pyov and 7('''0" he does say, in the passage from the Rhetoric, that 'the function (i!pyov) of rhetoric is not to persuade, but to find the avail- able means of persuasion about each given subject'. This does not mean, of course, that the orator will not aim at persuasion; it merely accounts for the fact that this aim may be impossible to achieve if; for example, the orator's case is very weak, or the audience hostile and recalcitrant. Nevertheless the orator may have performed well-as well as possible-and so fulfilled his task. Following Nature by reference to an expected cure. So medicine and navigation do not seem to provide the required support for Antipater's position. But, one might say, there is still a class of activities that does offer the appropriate model, namely sports and competitive games. These are arguably stochastic, in that success depends as much on the others' performance and abilities as upon one's own, and they in- dubitably provide an example of a kind of activity in which the result aimed at is not the ultimate goal. One may assume, I think, that such things as running races were not invented for the sake of setting records, or reaching the line before all other competitors, but rather, say, for the sake of strengthening the body, or perhaps for entertain- ment. This is why the result pursued by the performers, winning the race, is subordinate, even to the extent that the primary goal can be achieved independently: the runner who loses strengthens himself just as well as the winner. This sort of relation between projected results and overall goals is of course particularly evident in the case of games. What the players do is try to win; yet one would not want to say, for example, that the goal of chess-playing is checkmating one's opponent, or the point of soccer getting the ball between two posts. Rather, these activities are pursued for other purposes, such as intellectual exercise, perhaps, in the case of chess, amusement or whatever in the case of soccer. It may not always be clear what the superordinate goal might be, and in fact there seem to be several in most cases, yet it seems beyond doubt that the intended result-winning-is not what motivates and justifies the whole activity. It remains true that all actions within the game aim at the intended result, namely winning, in the sense that individual steps in the procedures have to be explained by reference to this; but what these examples show is that it is a mistake to assume that this 'reference- point of all action' must coincide with the primary goal, that for the sake of which one engages in the game in the first place. Now this seems to be precisely the point Antipater wanted to make, and whether he thought of these examples or not (this is doubtful, but Epictetus did: Diss. 2. 5. r-23), it is clear that the distinction between intended result and overall goal can be used to disarm Carneades' dilemma by showing that there may be a 'reference-point of all action' that is distinct from the primary goal and not a necessary means to it, but clearly subordinate. One should admit, however, that the archer simile is not a very good Gisela Striker illustration. Archers are normally judged by their success in hitting targets, not by whatever elegance they may display in bending their bow. One would not want to say that archery is a stochastic craft in the sense in which medicine or navigation is. But given the etymo- logy of the word stokhastikos, which means 'taking aim', it would be hard for a speaker of Greek to maintain that. I think that the example is more likely to come from Carneades than from Antipater, who, since he could not well refuse to admit it, probably tried to make the best of it. Thus far, I think, we have no reason to agree with those ancient sources that found Antipater incapable of resisting Carneades' attacks. The elaboration of the concept of the 'art of living' as a stochastic craft was no doubt provoked by Carneades' questions, but what Antipater offered in reply is at least sufficient to show that the Stoic doctrine was not incoherent. We now have a fairly precise notion of the craft that virtue is supposed to be. Its characteristics are: (I) It is stochastic in the sense that the result of the technical performance depends in part upon external factors not con- trolled by the craft. Therefore, individual performances are not to be judged by their success, but by their correctness. This seems to capture an important point about moral evaluation. (2) Its primary goal is not identical with the result pursued through the exercise of the craft. Though every step will be referable to the intended result, that result is not the ultimate end for the sake of which the activity is performed. The ultimate aim of virtuous action lies in the correct performance itself as being our way ofliving in agreement with nature. We have seen that such a craft is not inconceivable. But this does not, of course, settle the further question whether it is plausible to say that moral virtue is a craft of this kind. Now although the analogy between the attitude of players to winning a game and the attitude of the morally good person to the success of her projects is attractive, I think that this only disguises the fact that the comparison of virtue with the skill or craft of a player is still misleading. Briefly, one could say that the analogy between rules of games and moral rules may be instructive, but not because virtue is a craft. The point comes out if one tries to see which rules of games arc supposed to be analogous to the rules of morality. Roughly, we can perhaps distinguish three types of rules that belong to all or most competitive games, namely: Following Nature 33 (r) constitutive rules, i.e. those that define the game, like the rules for the moves of chess figures; (2) restrictive rules: rules that set limits to the means players may use in trying to win. These rules are usually meant to ensure that all players have roughly the same initial chances, so that winning or losing depends upon skill rather than, say, strength, cunning, or the use of brute force. Examples would be the rules about fouls in soccer games, rules determining what distance a player must be from the goal, the rule that the ball is not to be thrown by hand, and the like. Not all games require such rules- chess does not-but in some they are clearly important; (3) strategic rules: rules that are not properly speaking rules of the game, yet quite important in games that require either co- operation or complicated planning, like soccer and chess, respectively. (1) and (2) differ from (3) in that constitutive or restrictive rules may not be broken without disqualifYing the player. An incorrect move at chess is void, and the use of brute force on the soccer field leads at least to a penalty. By contrast, rules of strategy are not binding. If a player neglects them, she may be foolish or just inexperienced, but all she loses is her prospects of winning. Now the good or skilful player is obviously the one who has the best strategy, not the one who violates neither constitutive nor restrictive rules, for that is expected of every participant and does not require any special skill. If there is a craft involved at all, it consists in developing good strategies, and the performance of players is evaluated-just as in non-stochastic crafts-in terms of their success. The best player is the one who wins most often, even though she may occasionally lose, or she may not be able to make up by herself for the weaknesses of other members of a team. But this shows that the analogy between virtue and the skill of a good player breaks down, since moral evaluation, as was emphasized before, is not based upon success. The plausibility of the game analogy is probably not due to the supposed fact that follow- ing moral rules can be considered as a craft, but that we inadvertently tend to compare moral rules to constitutive or restrictive ones. Thus a breach of moral rules might be said to disqualify a person as a member of the community, or make her subject to penalties, just as cheating disqualifies a player. And neglect of moral rules, unlike neglect of strategic rules, is not usually treated as a mere matter of foolishness or 34 Gisela Striker lack of experience. We need, perhaps, to introduce yet another distinction among goals: between the point of an activity as a whole and the specific aim of the craft involved. Even if winning is not the point of playing games, it remains the case that the aim of any craft or skill that goes with the game will be success, though the success may only be sought for the sake of entertainment. Characteristically, we tend to speak of'good players' in two sorts of ways. On the one hand, the good player is the one who does not get angry when she loses, does not try to cheat, etc.; on the other hand, the good player is she who wins most of the time. And if there is a skill involved, it belongs on the side of effi- ciency and success, not on the side offairness and friendliness. It seems that games do not after all provide an example of skills that aim at some- thing distinct from the result they are supposed to bring about. With respect to the Stoic proposal for a craft ofliving, then, I would conclude that Carneades was right. Even for stochastic crafts, the goal must be the intended result, whether they are normally exercised for the sake of this result, as with medicine, or for some other purpose, as in the cases of games. Moreover, the distinction between the point of a skill and the point of using the skill can be applied to non-stochastic crafts just as well: one might, for example, learn to knit sweaters not because one needs them, but because one enjoys the activity. So the fact that moral evaluation does not depend upon success in action will not show that virtue can be considered as a stochastic craft. Nevertheless, we can use the distinction between the goal of a craft (in the sense of the intended result) and the point of practising it to understand what Antipater wanted to say about virtue. Chess-players, to return to this example, might develop their skills not in order to win more often, but because the game is more interesting when played skilfully. In this case again, success or failure becomes less important than performing well. So one might well prefer to play against a dis- tinguished chess master and lose, rather than to win an easy victory against a novice. Similarly, according to the Stoics, one might acquire the skill of pursuing what belongs to a natural life for human beings, not because one is interested in possessing as many of the preferred things as possible, but because this way of acting is what fits best into the order of nature. And surely a wise person should be more content to see everybody being as skilful as herself than to outwit all the fools. Morality, ifit can be compared to a game at all, should presumably not be compared to a competitive game in which one player's success implies another's failure. Following Nature 35 Carneades' objection, then, is not fatal for Antipater's theory of virtue. For Antipater could have simply used the analogy with skills to show that even a goal-directed activity may have a superordinate purpose that need not be identified with the result one is trying to achieve. So it is not nonsense to say that we should try to secure the things in accordance with nature only for the sake of living in agree- ment with nature. This, however, is still not enough to show that the Stoics were right about virtue, for now the crucial question becomes whether they were right to say that virtue consists in adapting oneself to the order of nature. 4. Laws of nature? One of the central doctrines of Stoic ethics is the claim that 'nature leads us to virtue' (D.L. 7. 87), or that following nature means living virtuously. It has also become one of the most influential doctrines, under the label of Natural Law Theory. For the Stoics maintained that the reason which governs the universe can be described as a universal lawgiver-it prescribes what ought to be done, and prohibits what must not be done. In Cicero's solemn words, 'law is the highest reason, implanted in nature, which commands what ought to be done and forbids the opposite' (Leg. 1. 18, trans. Keyes [Loeb Classical Library]). This law, as Cicero goes on to say, should also be the basis of correct human legislation. It would be an exaggeration to say that the idea of a natural law was invented by the Stoics. The expression (nomos tis phuseos) is first used in Plato's Gorgias (483E), as an intentional paradox; but the fourth century Be is full of attempts to show that justice is natural in some sense-this is the continuation of the old controversy over nature (Phusis) and convention (nomos) that started in the fifth century with the sophists. However, the Stoics were the first to introduce the idea of nature as a kind of personal lawgiver,16 which probably made the theory attractive to later Christian authors, who could simply promote the decalogue to the status of natural law. Now the Stoics certainly knew that to speak of universal nature as 16 For a discussion of the difference beteen the Stoics and their predecessors see my 'Origins of the Concept of Natural Law', Proceedings of the Boston Area Colloquium in Ancient Philosophy, 2 (1987), 79-9+ Gisela Striker prescribing or forbidding certain ways of acting is to speak metaphor- ically, since nature, though arguably rational, does not have a language in which to speak to us, nor can her laws be read in any book. If they were not just trying to provide an impressive cosmological background to their preaching of more or less standard morality, the Stoics must have offered some way of finding out what nature's laws might be. We may take it, I think, that the thesis that nature prescribes virtuous conduct was at first an optimistic prediction. Having tried to establish that our lives will be most successfully organized if we try to live in agreement with the order of the universe, the Stoics expected that it would turn out that virtue as commonly understood, that is, courage, temperance, justice, and wisdom, was in conformity with the natural order-they were not setting out to overthrow our ideas about virtue by investigating what natural behaviour would consist in, and then declaring that to be virtue, as Callicles had suggested. 17 How did they try to demonstrate their thesis, and to what extent did they succeed? A first indication about how we discover nature's laws is contained in the definition of the goal of life ascribed to Chrysippus: living in accordance with one's experience of what happens by nature (D.L. 7. 87).18 So we should be guided by our expertise concerning what happens by nature. In particular, if we want to find out how we humans should live, we should look at the way nature has created us. This was at least Chrysippus' view. Cleanthes apparently disagreed, saying we should follow only universal nature (D.L. 7. 89). How he 17 This does not mean that the Stoics were trying to be conformists or defenders of conventional morality. On the contrary, Zeno's RejJUblic apparently shocked its readers by its thoroughly unconventional descriptions of the 'natural life'. None the less, the Stoics accepted the traditional canon of moral virtues, beginning with the four cardinal ones. This shows that they were prepared to revise ordinary conceptions of virtue, but not to reject them altogether. 18 'Experience' ("7Tpia) sees to be a weaker term than 'knowledge' (E7nar/W'I), but we should probably not be misled here by Aristotelian associations. E("7rELpia does not seem to have been a technical term of Stoic epistemology. It apparently occurs only once in an epistemological context, in the famous Aetius passage on the origin of common and other notions (4- II = SVP ii. 83), but there the description is so close to Aristotle's Metaphysics A I that I would not want to place any weight on it. On the other hand, EfL7rELpia and especially the adjective EV7rELPOS' is frequently associated with technical expertise by Epictetus, and that connection seems to be supported by similar passages in D.I .. (7. 48) and Stobaeus (2. 99. 9). There is not much of a distinction between TEX""! and <maTiJv"! in Stoic sources-both are systems of cognitions (KaTa- /..'i1"(5). I Ience Cicero may be right in simply translating the word E("7TELpiu as scienlia, Fin. 3. 31. Chrysippus might have used it to indicate the 'art ofiiving'. Following Nature 37 went on from there to build a moral theory we do not know-the theory that has come down to us as Stoic seems to be, as so often, Chrysippus'. What one learns by experience about human nature is set out in the theory of natural concern (oikeiosis), which provides the theoret- ical background for the Stoic teaching about appropriate action (kathikonta). We are told that nature has endowed every animal with some awareness of its own self and with an impulse to preserve itself or its own constitution, that is, its normal, healthy state. This is argued on teleological grounds, from the consideration that it would not be reasonable for nature to create an animal and then not to endow it with the capacity to keep itself alive. Furthermore, the fact that animals have such a capacity is discovered by observation: they all seem to be able to recognize as belonging to them things that contribute to their self-preservation and to avoid things that might lead to their destruction. As the human animal grows up, more and more things come to be recognized as belonging to it and being in accordance with its nature. Thus children learn to walk on their feet rather than to crawl; after a while, they become interested in finding out about the world and develop a natural inclination for truth and aversion to falsehood; they also naturally try to acquire technical skills. The most important development, however, is that they come to recognize other people as belonging to them. It is not clear exactly at what stage and how this is supposed to happen. The standard argument Chrysippus seems to have used to establish the existence of this social form of concern refers to the love of parents-both animal and human-for their young, arguing teleologically, as before, that it would not be reasonable for nature to create animals so as to produce offspring and then not to care for them. But \t is obviously implausible to say that our social instincts should begin to develop only so late in life, and one's concern for others is not supposed to be limited to one's children-on the contrary, it should eventually embrace all mankind. It is also not clear how the concern for others is supposed to be related to the initial concern for self- preservation. Should we assume that there are two distinct forms of natural concern, egocentric and social, and that they develop in some temporal sequence, or should we take it that there is only one basic form, namely self-concern, which eventually comes to comprise concern for others as a way of expressing one's own human nature? This is the way Epictetus sees it (Diss. I. I9. II-IS), and he may well Gisela Striker be right. 19 However that may be, the general lesson we are to learn from a survey of the way nature has provided for animals and humans is that we are made to take what accords with our nature, and to avoid what goes against it. Following nature is thus seen to consist in select- ing what is in accordance with, and rejecting what goes against, one's human nature. Since it is to be our primary goal to obey the laws of nature, what is good about this way of acting is the activity of selecting itself, not the things selected; and the Stoics tried to keep this clear by insisting that the things selected or rejected are just 'in accordance with nature' (kata phusin) or 'against nature' (para phusin), not 'good' or 'bad', and that they are 'selected', not 'chosen'. Choice (hairesis) has to do with real goods and evils, while selection (eklogi) deals with things that are indifferent as far as the final good or happiness is concerned. Now it is obvious that we do not always succeed in getting or doing what is natural or avoiding what is against our nature; yet according to the Stoics everything that happens is planned by providence or universal reason, and so also in accordance with nature. When we call health or wealth 'in accordance with nature', illness and poverty 'con- trary to nature', we are considering the human being as it were in isolation (Epict. Diss. 2. 5. 24). When we consider her as part of the universe, however, it may turn out that illness and poverty are assigned to her, and hence must be taken to be in accordance with nature in a different sense. Since we want to live in agreement with universal nature, we will have to accept such things, admitting that our plans turned out to be mistaken; but as we cannot foresee the future, our actions will still have to be guided by what we perceive to be nature's general tendency. It is not certain, but reasonable to think that nature wants us to stay alive and well, given the way she has made us. So we will be acting in agreement with nature, not by always correctly anticipating what she intends, for this is impossible, but by always acting in such a way that what we have done will admit of a reasonable justification. 'As long as it is unclear to me what will come next, I always cleave to what is better adapted to obtaining the things that are in accordance with nature,' said Chrysippus, 'for God himself 19 For a discussion of these questions see H. Gorgemanns, 'Oikeiiisis in Arius Didymus' (with comments by B. Inwood), in W. W. Fortenbaugh (ed.), On Stoic and Peripatetic Ethics (Rutgers University Studies in Classical Humanities, I; New Brun- swick and London, 1983), 165-201, and]. Annas, 'The Hellenistic Version of Aristotle's Ethics', Monist, 73 (1990), 80-96. If the theory was mainly intended as a support for claims about appropriate action, it is perhaps not so surprising that the Stoics did not bother to elaborate the psychological details. Following Nature 39 has created me with the capacity to select these' (Epict. Diss. 2. 6. 9-10). 'What, when done, admits of a reasonable justification' is the definition of appropriate action (Cic. Off. 1. 8; D.L. 4- I07; Stob. 2. 85. 14), and the 'art of living' will consist in a reasonable selection of things in accordance with nature. 20 An action may be objectively appropriate without being done for the right sort of reason. For example, people will eat and drink, or avoid precipices, without considering whether such practices conform to nature's general rule to keep oneself alive and well. Here the action will be merely appropriate, but not perfectly right, not what the Stoics called a katorthiima, since what makes an action perfect or good in the strict sense is that it be done with the intention of following nature, and from a stable disposition so to act. Only thus will one act in agree- ment with nature, as opposed to merely leading a natural life (kala phusin bios). Virtue, according to the Stoics, is achieved when one's every action is both appropriate and done for the right reason. Since such a stable disposition is only reached at the stage of complete wisdom, perfectly right actions are reserved to the wise. But why should we believe that selecting things in accordance with nature in this way actually amounts to virtuous conduct? Most descriptions of the development of natural concern seem to offer two rather general tendencies-towards self-preservation and towards some concern for others. We might agree that both these tendencies are observably natural, and still not find that we have been given an 20 To do what is reasonable without success will not count as acting against nature's will-as opposed to doing what is unreasonable, which is 'against nature'. In fact, nature might wish precisely that people should try even if they fail. The point of such an assumption can be brought out by an example which, though not Stoic, may capture their thought. Small fish have an instinct to avoid predators and thereby to preserve their life. By trying to keep alive and healthy, reproduce themselves, etc., they follow a natural pattern. Their natural aim in life would seem to be to survive to the end of their life-span. But those small fish notoriously also serve as food for the big fish, and hence it is also part of nature's plan that some of them should get eaten. So success in leading a natural life is not necessarily part of nature's plan for a small fish, but it remains the case that the balance of nature is maintained through each species following its innate impulses. Indeed, if the small fish did not have the instinct to escape, they would soon be finished off, and the big fish would starve as well. This shows that nature's order depends upon each creature's pursuing its natural aim, but not upon each individual's attaining it. One might protest at this point that it does not follow that trying correctly, but unsuc- cessfully, is also good for the small fish, or for that matter for a human being-all that is shown, if the story is true, is that nature has not provided for every creature's happiness. I would agree with this objection, but it only serves to point out that the Stoic position is implausible, not that it is incoherent. Gisela Striker argument to show that following these tendencies will be a guarantee for morally praiseworthy behaviour. The Stoics asserted that the social form of oikeiiisis is the foundation of justice, but to establish the existence of some form of natural altru- ism does not seem sufficient to show that human beings will naturally come to act in the way we consider just. We have to look at the details of the Stoic account of natural development in order to see how virtuous dispositions were supposed to arise from the primary impulses. (For simplicity's sake I shall limit the discussion to the four cardinal virtues.) Unfortunately, we do not have any of the older Stoic treatises on appropriate action (of which there were many). What we do have is Cicero's De officiis, in which Cicero himself professes to be following a work on appropriate action (Peri tou kathekontos) by Panaetius, head of the Stoa in the second half of the second century Be. The first book of De officiis contains the kind of argument one would expect, namely an attempt to show how virtue arises out of natural tendencies (Off I. I 1-14). But Panaetius' argument seems peculiar in several ways. First, the standard cardinal virtues of courage and temperance are replaced by magnanimity and 'propriety' (decorum, Tel 71pE7rov, cf. Off 1. 93); second, in addition to the usual two primary impulses we find two other alleged natural tendencies, at the root of magnanimity and propriety, respectively: a 'certain desire to dominate' (appetilioquaedam principatus) and a sense of order and proportion which pertains not only to aesthetic features, but even more to 'beauty, constancy, and order in thought and action'. Panaetius is said to have defined the final good as 'living according to the starting-points provided to us by nature.'21 I suspect that he might have tried to simplify the argument by postulating an innate tendency for each of the virtues. Hence I shall rely on the De officiis passage only to the extent that it is parallel to other accounts of natural concern. The least problematic among the virtues is obviously wisdom, because it can be seen as the perfection of human reason (cf. Off 1. IS). Cicero argues that a desire for knowledge is apparent already in young children, who will be pleased if they have figured out something for themselves, even if it is of no immediate use to them. At a later stage, 21 Clem. Slro111. 2. 21. I29. I-5 TO I;ijv KUTa Tal OEOOI""vo., Y;j.ilV EK q,uawJ> uq,Opj.ia,. Panaetius' aq,opj.iui should probably not be identified with the impulses (opj.iar) men- tioned in the theory of OiKEiwUL', but rather compared to the 'seeds' of virtue that Cicero mentions in several places, c.g. Fill. 5. 18. Following Nature technical expertise appears as desirable from the same motive, as consisting of pieces ofknowledge methodically organized (Fin. 3 17- IS). It is not surprising that logic ('dialectic') and 'physics' should count as virtues also: they are needed to avoid error and to compre- hend the order of the universe so that we can live in agreement with nature (Fin. 3. 72-3). In a similar vein, Seneca argues that human beings will be concerned about themselves qua rational (p. 121. I4- IS), and Epictetus says that nature has provided humans with reason so that they can not only make use of their impressions-as other animals do-but observe and reflect upon them. 'God has brought man into the world to be a spectator of himself and his works, and not merely a spectator, but also an interpreter' (Diss. 1. 6. 19-20, trans. Oldfather [Loeb Classical Library]). But the acquisition of knowledge does not appear to be necessarily linked to moral virtue, and so we still do not see why a natural human development should lead to a morally virtuous life. The weight of the argument for this claim evidently falls upon the natural impulse to care about others which was declared to be the foundation of justice. The existence of such an impulse, as we saw before, was established by observation and teleological argument, beginning from the love of parents for their children. Furthermore, the gregariousness of human beings, which goes beyond mere family ties, is undeniable and indeed so strong that no one would be willing to live a solitary life, even if she could enjoy all sorts of pleasures. Human beings are also said to have a natural inclination to benefit as many others as they can by teaching them whatever they have found out, and to use exceptional strength or talent to protect or help others, as was done by such mythical heroes as Heracles (Cic. Fin. 3. 65-.6).22 Now it seems clear that concern for others should take the form of helping them to preserve and develop themselves and protecting them against physical harm and deprivation. If we grant the assumption that concern for others will eventually extend to all human beings simply as human beings, we can see how it might lead one to act according to 22 In De officiis Cicero, i.e. Panaetius, twice (r. 12 in passing, I. 59 with emphasis) appeals to language as a bond that exists only between humans and shows that they were made for social living. This is so close to Aristotle's Politics (I. 2, 1353'9-18) that I am inclined to think it was introduced by the 'Aristotelizer' Panaetius himself. The Stoics did indeed believe in a natural community of all rational beings, including not only humans but also the gods, but this is not a matter of a common language, and could hardly be adduced as evidence for a natural impulse to sociability. Gisela Striker the rule 'to each his due',Z3 understood as 'to each according to his needs' (cf. Off. I. 59), and this might seem to encapsulate the essence of justice. The virtues of courage and temperance are not directly traceable to the natural impulses but can be shown to be necessary if one wishes to act in accordance with wisdom and justice. The Stoics held, like Socrates, that the virtues were inseparable and implied one another; and Chrysippus explained this by saying that they have their theorems in common (D.L. 7. 125). The knowledge that underlies all virtues concerns of course 'what ought to be done' (ibid. 126), and if the things to be done were determined by the basic tendencies towards self- preservation and social living, it would be quite plausible to claim that the actions thereby prescribed often require courage or temperance. 24 For example, maintaining one's health might require enduring painful medical treatment and abstaining from excessive eating and drinking; care for one's children might require the willingness to protect them against dangerous enemies, if necessary at one's own risk; and respect for other members of a community would obviously imply that one should not deprive them of the means of sustaining themselves. That this was indeed the Stoic view seems to be borne out by the definitions of specific virtues as forms of knowledge: temperance is knowledge of what ought to be chosen or avoided (or neither), courage is knowledge of what is terrible or not terrible (or neither), etc. (cf. SVF iii. 262ff.). These definitions all refer back to wisdom (phronesis )25 as knowledge 23 The standard definition of justice was <t" arrov<I-'7]TtKr, TOU KaT' atiav EKaaT'I' (cf. SVF iii. 262, 263, 266, 280), literally 'a disposition to assign to each his share according to worth'. The phrase KaT' Miav in this definition was explained by Diogenes of Babylon (ap. Stob. 2. 8+ 13-17, SVF iii. 125) as meaning 'what belongs or falls to someone': xpiJa8at 0' EvioTE TcjJ OV()f-LaTl rr),; attu.;; aVTi TOU Eruj3ci/V\OVTOS' we; EV n{J TryS cHKuwauv1jS opc.p 7TapEiI\Yj'TTTUL, oTav l\iYTJTUL Elvat EgLS d1TOVEJ.1-YJ'TLWij TOU KaT' attnv EKaOTlfJ' EaTL yap olov TOU E1n{36./.,l\ovTos EKciunp. Cicero's usual translation is suum cuique: cf. Off I. IS, Fin. 3. 68, Leg. I. 19; he sometimes interprets this as 'to each accord- ing to need' (Off I. 49, 59), sometimes as 'according to desert' (Off I. 42, 46), sometimes as 'according to closeness of relationship' (Off I. 50). So I think 'to each his due', vague as it is, might be the best translation. 24 An example of this kind of argument can be found before the Stoics in Plato's Gorgias, where Socrates convinces Callicles that even his unjust 'strong man' will not be able to do without these two virtues: cf. 49' B-C, 506c-507 A. li In the lists of virtues, <pp6v7]a" (wisdom) appears only as knowledge of things to be done or not done, or knowledge of goods and evils, rather than knowledge of truth and falsehood, or of the order of the universe. This is understandable if we assume that these definitions are meant to indicate that part of universal wisdom which underlies virtuous action. But it seems clear that wisdom as perfection of human rationality included more. Although the Stoics do not seem to have taken over Aristotle's distinc- tion between theoretical and practical reasoning, the difference between theoretical and Following Nature 43 of goods and evils, or knowledge of what ought to be done, and the only indication about the content of this knowledge seems to come from the definition of justice: giving to each his due. If this outline of the theory is correct, it follows that virtuous conduct in the Stoic sense comprises not only morally praiseworthy action, but also actions aimed at self-preservation and the acquisition of knowledge. Hence we can understand why the Stoics claimed that every single action of the wise person would count as 'right' (katorthoma) and virtuous: the decisive factor is not whether an action is altruistic, say, or socially useful, but whether it is in accordance with human nature and done from the intention of agreeing with universal nature. However, the Stoic account of the origin of justice shows that morally virtuous conduct as ordinarily understood will be included in a full development of human nature as guided by the primary impulses, and so the claim that 'nature leads us to virtue' can be understood in both the larger (Stoic) and the narrower sense of virtue. What all these more or less well-established observations amount to is not, of course, a Code of Natural Law; but one might accept them as evidence for the weak, but in this context significant, claim that the cardinal virtues arise out of certain natural tendencies of human nature. One might perhaps use this to derive a few very general principles of conduct that could then serve to justify or criticize more specific rules of action. For example, one could appeal to natural benevolence to prohibit murder and fraud, and to enjoin mutual assistance in cases of need. However, the standard definition of justice-'to each his due' -goes beyond this. It seems likely that here the Stoics would have appealed to the common notions or preconcep- tions (prolepseis) that arise naturally in human souls when they come to the age of reason. These could serve to guide their understanding of what it means to develop one's human capacities and to lead a civilized social life. Nature's way of leading us to virtue would then seem to be twofold-giving us the right sort of fundamental tendencies, and letting us grasp, as rational animals, the preconcep- tions that should regulate our attempts to live by those tendencies. But one has to admit, I think, that there is only a rather tenuous connection between these highly general observations about human practical knowledge seems still to be implicit in the different definitions given of oo(a and p6V'YjOlS: ooia was defined as 'knowledge of things divine and human' (SVF ii. 35, 36), covering all three parts of philosophy (logic, ethics, physics). The Stoics must have held that ooia includes p6v'YjotS. See Cic. Off. 1. 153 on sapientia and prudentia. 44 Gisela Striker nature and the often very specific rules of conduct we find in Cicero's De officiis and the books of many later Stoics. The differences between the teachings of different generations of Stoics themselves illustrate the vagueness of the label 'in accordance with nature'.26 At the begin- ning, it seems, Stoic ethics contained some rather revolutionary ideas-not only Zeno's Republic, which was so strongly influenced by Cynicism as to be an embarrassment to later generations, but also the books of Chrysippus. For example, the Stoics argued, in contrast to Plato and Aristotle, that all rational beings were by nature equally equipped to attain virtue, and hence in principle deserved equal respect. The earlier Stoics also apparently scandalized others by pointing out that certain practices abhorred by the Greeks, like incest, were not demonstrably against nature and hence, if not commendable, at least permissible. But the moral advice given by the later Stoics in Roman times by and large closely resembles the conventional morality of their respective times and social classes. Apart from vagueness, the main problems with the Stoic account of natural impulses as a foundation of morality seem to be, first, that the optimistic assumption that our natural instincts are all for the good makes it hard, if not impossible, for the Stoics to explain why most people in fact turn out to be bad rather than virtuous; and second, that the theory introduces two potentially conflicting tendencies without at the same time providing a method for deciding which one is to be given precedence in cases of actual conflict. This second point is perhaps the more important, since it indicates that even if one could establish that human beings have just the respectable instincts accepted by the Stoics, it would not follow that they would naturally come to lead a virtuous life. The crucial case, as one might expect, is the virtue of justice, allegedly arising out of the natural tendency towards social living, or general benevolence. What happens if this tendency conflicts with the equally natural tendency to protect one's own life? In an argument preserved in an anonymous (middle-Platonic) commentary on Plato's Theaetetus,27 the Academic sceptics argue that justice will lose out against egoism (cols. 5 24-7 14) '6 For an interesting example of conflicting interpretations within the Stoic school see l. Hadot, 'Tradition stoi"cienne et idees politiques au temps des Gracqucs', Revue des etudes latines, 48 (I970), 133-79. 27 Anonymer Kommentar zu Platolls Theaetet, cd. II. Diels and W. Schubart (Berliner Klassikertexte, 2; Berlin, 1905). For the date and contents of this commentary see II. Tarrant, 'The Date of Anon. III Theaetetum', Classical Quarterly, 33 (I983), 161-87. Following Nature 45 They claim that oikeiiisis can only serve as a foundation for justice if it can be shown that our concern for every other human being is equal to our concern for ourselves. Why this should be necessary is not explained right away, but comes out at the end of the argument. First it iis argued that we love ourselves more than we love others. Evidence for this is, for example, that we feel alienated from others, but not from ourselves, for their bad deeds and evil character (poneria). Besides, it is dear that natural concern admits of degrees in the case of parts of our own bodies, some of which are obviously less important to us than others, as hair and nails are less important than, say, hands and feet. Now the Stoics might admit that natural concern is not equally strong in all cases, and thus save their general benevolence (philan- thriipia) to the extent that it only requires some degree of respect for all members of the human race. But this will not suffice to guarantee justice in exceptional circumstances (peristasis), as when only one of two persons can be saved. As Diels notes in his edition of the text, this is probably an allusion to the kind of case described by Cicero (Off. 3. 90), in which two men are left after a shipwreck with a plank that will carry only one of them. The question is, will it be correct for the stronger to push off the weaker man and thus save his life? The point of the argument, which is not fully stated in the text, should be that if the Stoics admit, as they apparently must, that self-love is naturally stronger than benevolence towards a total stranger, then they will have to admit that it is natural, and hence appropriate, to save one's life at the expense of another's. But this is clearly not what one would expect from a virtuous person, and so it seems that general benevolence will not suffice as a foundation for justice. The critic admits (col. 6. 25) that such extreme situations need not arise in everybody's life, but whether they do or not is not to the point, for what was to be shown is that we cannot appeal to our natural instincts to justify our view of what would be the appropriate course of action in this sort of situation. In fact, as the Academics say, the example shows that the argument that the Stoics used to reject the Epicurean account of justice can be turned against them. The Stoics used to claim that for Epicurus justice was not based on general benevolence, but on the desire to avoid the fear of discovery and punishment in case one offended against the contract not to harm others that is the foundation of civilized human society.28 This could " Cf. e.g. Cic. Fin. 2. 53; Sen. Ep.97. IS. It does not matter for the present argument that this is not a fair account of Epicurus' theory of justice. Gisela Striker not be enough for perfect justice, however, since the Epicurean would, following his own principles, have to break the rules of justice if this was to his advantage and he could be certain not to be found out, as for example in the case of a promise given to a dying man in the absence of witnesses (Cic. Fin. 2. 55). In such circumstances virtuous conduct could be explained only by appeal to general benevolence. So the Stoics argued against the Epicureans from exceptional situations. Now the shipwreck case points to a similar difficulty for the Stoic theory by showing that under extreme conditions, when one has to decide between the instinct for self-preservation and general benevolence, the Stoics should, on their own principles, admit that egoism must outweigh concern for others. But, as the Academics rather maliciously conclude, just as a single error suffices to show that a craftsman's expertise is not complete, so a single exception will suffice to show that virtue is not perfect. The Academic argument brings to light what seems to have been a weakness in Stoic treatments of 'selection'. By concentrating on the simple case of choosing what accords with nature over what is contrary to it, the Stoics had apparently paid less attention to the equally important possibility of deciding between two things on one side of the divide. Obviously, one ought to opt for what is 'more' in accordance with nature, or what is 'less' against it, but how does one determine which is which? It is probably no accident that Cicero repeatedly complains about the absence of rules for such cases in Panaetius' book (cf. Off. 1. 10, I52; 2. 88). Ifhis predecessors had dealt extensively with these problems, it is not likely that Panaetius would have omitted them. Now the argument about the shipwreck case suggests that a plausible 'natural' criterion for deciding whom to help would be the greater or lesser degree of natural concern for a person. Indeed, such a criterion is quite acceptable in other types of situations (cf. Cic. Off. 1. 50, 53-4), but in this case it leads to unacceptable consequences. This kind of argument explains, I think, why we suddenly find, in the second half of the second century Be, the prominent Stoic Panaetius proposing to write about 'conflicts between apparent expediency and virtue' (Cic. Off. 1. 9, 3 7-I3, 33-4). Cicero is careful to stress the word 'apparent' (quod videtur), for as he points out, according to official Stoic doctrine there can be no real conflict between expediency and virtue, since the noble or morally good (honestum, Ka'\6v) and the expedient or useful (utile, aVfLEpoV) are Following Nature 47 coextensive, and there can be no doubt that Panaetius was orthodox on this point (Off. 3. 34). Panaetius never wrote his book on the problem, which was to be the last of his work on appropriate action, although it is clear that he intended to write it, and Posidonius said that he lived for another thirty years after he had finished the first three books (Off. 3. 8). Is is altogether implausible to suggest that he did not write it because he never found a satisfactory solution? What the Stoics needed was some argument to show that by nature altruism limits egoism to the extent that we never pursue our own advantage to the detriment of some other person. Cicero sets out to fill the gap left by Panaetius, but the solution he offers seems to highlight the problem rather than to solve it. He proposes a 'formula' by which to deal with those apparent conflicts, namely: 'To take away some- thing from another or for a man to promote his own advantage through the disadvantage of some other man, is more against nature than death, or poverty, or pain, or anything else that might happen either to the body or to external things' (Off. 3. 2I). SO we are told that harming another human being is more against nature than other things that are admittedly also against nature; so much so that harm to others must be avoided even at the cost of accepting those other things. But how does nature indicate to us that this is so? It cannot be, as the Academic argument pointed out, because altruism is naturally stronger than egoism, for it is not. Cicero argues that society could not exist if everybody were to rob or injure every other for his own profit. This is certainly correct, but it is not enough, since we can normally indulge both our instinct for self- preservation and our natural benevolence towards others without having to give up the one for the sake of the other. It may sometimes be difficult to decide how far we can go in the pursuit of our own material advantage, but I suppose the Stoics could have appealed to the idea that external things like money are valuable only as means for securing advantages that are intrinsically in accordance with our nature (D.L. 7. I07), and one should think that our fellow humans are directly appreciated as belonging to us (though even that is not clear, cf. Cic. Off. 3 50-7, 89-9I ). But if it comes to actual loss oflife, as in the ship- wreck example, why should benevolence be stronger than self-love? One argument that recurs in this context (cf. Off. 3. 22, Fin. 3 64, Hierocles ap. Stob. 3. 732. I-I3) compares human society to an organic body to show that the advantage of the parts or members coincides with that of the whole. It would be unreasonable, says Gisela Striker Hierocles, to prefer one finger to the five (of a hand) rather than the other way around, for if the five perish, so does the one. But the analogy fails, of course, since by contrast with the members of a body, a member of a group can survive the group, not to mention a single member of the group. Hierocles, who notes this, argues that one can- not survive as a citizen the destruction of one's city, any more than a finger can survive, qua part of hand, the destruction of the hand. This is true, but it does not hdp, for we needed to be shown why being a citizen is more important than, say, being alive and healthy. Besides, there is no cogent reason to think that an occasional act of injustice will actually destroy the community, and the argument was not that it is not against nature at all to harm others, bUi, as Cicero quite correctly puts it (Off. 3. 26), poverty or pain may seem worse than inflicting harm upon one other person . . ,So Cicero's 'more against nature' does not seem to be well founded, and though he has clearly seen what is needed, one cannot say, I think, that he succeeds in providing it. At any rate, I would not want to agree with him when he hopes (Off. 3. 33) that Panaetius would have treated the problem in the same way. It appears that later Stoics did not make serious attempts to solve the problem, with the exception perhaps of Posidonius, who wrote, according to Cicero (Off. 3. 8), that there was no more urgent topic in the whole of philosophy. Cicero complains that in spite of this Posidonius 'only briefly touches upon it in some commentaries'. But Cicero may have looked in the wrong place. In his Tusculan Disputa- tions, when expounding the Stoic theory of the emotions, he shows no awareness of Posidonius' criticism of Chrysippus' psychology, on which Galen reports so extensively in his De placitis Hippocratis et Platonis. According to Galen, Posidonius thought that his own account of the emotions would also be needed to arrive at the right view about 'happiness and consistency' (eudaimonia kai homofogia, PHP 5. 6. 5, fro r87 Edelstein-Kidd). He rejected the Chrysippean version of the theory of oikeiosis, apparently because it claims that our natural impulses wi!! lead us only towards what is morally good (kalon, cf. PHP 5. 5. 8, fr. I69 EK). It seems clear from Galen's report that Posidonius was primarily concerned with the problem of the origin of evil. But if he rejected Chrysippus' account of oikeiosis, saying that only reason, the better part of the human soul, is naturally drawn towards the morally good, then presumably he must have based his rules for appropriate action on reason alone. Now it may be worth Following Nature 49 notIcmg in this connection that purported literal quotations in Galen's summary of Posidonius' ethics (fr. 187 EK) always have the expressions homologia (consistency) or homologoumenos zen (living consistently) as terms for the goal of life, whereas the usual formula, homologoumenos tei phusei zen, is used only by Galen. This does not point to any lesser reverence for the order of nature on the part of Posidonius-in fact, the 'definition' of the final good ascribed to him by Clement of Alexandria (Strom. 2. 21. I29. 1-5, fro 186 EK) indicates just the opposite. But it might be an indication that Posidonius emphasized order and consistency as being those features that account for the goodness of nature's rational order. As consistency should also characterize the conduct of the wise man, Posidonius might conceivably have argued that it would be inconsistent for a rational being that had accepted the notion of justice as a good, and hence adhered to the principle 'to each his due', to harm another person for the sake of his own advantage. Posidonius would have been right, I think, in holding that such an argument appeals exclusively to reason and its desire for consistency, and not to any natural instincts that might in fact lead in the direction suggested by the Academic objection we discussed before. It would also, incidentally, support the kind of solution to the shipwreck case we find in Cicero (Off. 3 90; taken from Hecato, a contemporary of Po sidon ius): the man whose life is more valuable for himself or his country should be allowed to survive; or if there is no such difference of value, the lot should decide. Hecato was not a 'heretic' like Posidonius, and so perhaps consistency (with moral preconceptions) was also invoked by other Stoics. One hardly needs to point out nowadays that consistency will not be suffi- cient to solve all priority problems, but it might have helped with the rather crude cases of egoism vs. altruism that Cicero lists in Off. 2. 50-7 and 89-91. What the Stoics evidently did not do was to apped to our intuitions about morality and introduce a distinction between 'prudential' and 'moral' values in the field of natural things in order to guarantee the right sort of priorities. Such a move would indeed have made nonsense of the project of explicating appropriate action in terms of the order of nature. The Stoic theory of things preferred or in accordance with nature also does not seem to offer an equivalent of such a distinction-caring for one's health is just as natural as looking after one's family or fighting for one's country. Rather than deplore the absence of a vital distinction, however, I would claim that this basic assumption is quite plausible. What seems to be lacking is a 50 Gisela Striker ranking of values within the field of things preferred or dispreferred such that acting in conformity with the priorities so determined results in morally good conduct. But while the two tendencies that are said to underlie preferences and rejections are rather uncontroversially natural, it seems far from clear whether nature can also provide the appropriate rules of priority. However, rather than criticizing the Stoics for their failure to find a solution, I would suggest that we should admire them for having attempted the task at all. In any case, the Stoic conception of the Natural Law evidently survived the sceptical attack on its foundations, thus proving to be more attractive than the allegedly dangerous and pessimistic Epi- curean idea that justice might be founded on human needs and mutual agreements alone. 5. Carneades on moral theory In the last two sections I have tried to describe some of the details, and some of the difficulties, of the system of ethics based upon the idea of following nature. For the difficulties, I have mostly referred to Carneades, or at any rate Academic critics. Typically, Carneades pointed out difficulties in the Stoic doctrine by way of arguing that the premisses the Stoics accepted would lead to most unwelcome conclu- sions. This was of course Carneades' general strategy as a sceptic, which has been more fully described in the field of epistemology. By drawing out alleged consequences of Stoic assumptions, Carneades constructed as it were a 'corrected' version of the Stoic theory, which he then apparently defended as his own alternative for polemical purposes. From his critique of the Stoic conception ofvirtue as a craft, and of general benevolence as a foundation for justice, we can recon- struct the outlines of what one could, in quotation marks, call Carneades' Moral Theory-or rather, one of his moral theories, for it is well attested that he defended different views of the goal of life, and hence different theories, depending upon the occasion. It would for example be most interesting to have the arguments he used to support the thesis that 'whatever the conflicting views of the philosophers were about the goal of life, still virtue would offer sufficient support for a happy life' (Cic. Tuse. 5. 83)' But as it happens, Carneades was better remembered as a critic, especially of Stoic theories, and so the only theory we have some information about is the one he used, as Cicero Following Nature emphasizes, to attack the Stoics (Fin. 2. 42,5.20). It is in fact a caricature of Stoic ethics, but an instructive caricature, and one that, like Carneades' theory of 'plausible impressions' in epistemology, may have been taken more seriously by some later philosophers than by its author. In this section I shall try to pull together the strands of this anti- Stoic theory, and contrast it with a pair of very different reactions-on the one hand, that of the orthodox Stoic Antipater, on the other hand, that of Antiochus of Ascalon, the man who tried to put an end to the sceptical era in the Academy and to reinstate what he saw as the original doctrines of its founders, Plato and the Peripatetics. Carneades' theory can I think be reconstructed from two main sources: the so-called Carneadea divisio in Cic. Fin. 5. 16-20, and the remains of his speech 'Against Justice' in Cicero's De re publica 3. It starts from the conception, generously but anachronistically attrib- uted to all philosophers, of wisdom or prudence (prudentia) as an art of living (Fin. 5. 16). Carneades began, it seems, with a bow in the direc- tion of Chrysippus: he said that every craft must have an object distinct from itself. In am not mistaken, this is an allusion to Chrysip- pus' argument against Aristo. Carneades is pointing out that one cannot explain what a specific craft is by reference to the exercise of the craft itself, e.g. by saying that dancing is the craft of exercising the skill of dancing. The object or goal of the craft must be describable independently, and this holds regardless of whether the object is an activity like dancing or a product or result of the craftsman's activity, as in the case of medicine or navigation. Perhaps Carneades was also replying to those unfortunate Stoics who had rejected Antipater's analogy between virtue and the stochastic crafts in favour of an analogy with dancing and acting, and who had ended up claiming that wisdom is the only craft that is concerned entirely with itself ('in se tota conversa est', Fin. 3. 24; cf. Fin. 5. 16 'null am artem ipsam in se versari'). Wisdom, then, must also have an object other than itself. Now Carneades claimed that it was 'agreed on all sides' that the object or goal of this craft must be in accordance with nature and such as to attract by itself the 'appetite of the soul', what the Greeks called horme. Disagreement among philosophers begins with the question what it is that so 'attracts nature from the very beginning oflife' (Fin. 5. 17). This is a fundamental question for, again according to Carneades, the answer to it will be decisive when it comes to determining the ultimate end of desire, that is, the goal oflife or the highest good. 52 Gisela Striker There were-according to Carneades--exactly three possible answers (ibid. 18). The first object of desire must be either pleasure, or absence of pain, or what the Stoics called 'the first things in accord- ance with nature', that is, integrity of the body and its parts, health, strength, beauty, etc. 29 Needless to say, since Carneades wanted to attack the Stoics, he adopted the third answer. These first steps look very much like orthodox Stoic doctrine, but it is not clear that they are. What is absent is of course any argument for the thesis that the goal is living in agreement with nature. Instead, we find the plausible-sounding but dangerous assumption that the ultimate goal of desire must somehow be derivable from the first objects of desire. I have already argued (Section r) that this is not what the Stoics said-in fact, I think, the appropriateness of our concern for the things in accordance with nature is derived from the thesis that we should live in agreement with nature, and not vice versa (cf. Gc. Fin. 4. 48). So there is reason for suspicion. But let me continue with Carneades'theory. Given that the art ofliving must be concerned with the first objects of desire, we are now told (Fin. 5. 19) that the theory of right and noble action must agree with this, in the sense that virtuous action should consist either in trying to obtain the objects of desire, even if one does not succeed, or in actually obtaining them. The distinction between trying to obtain regardless of success and actually obtaining is of course an allusion to Antipater's distinction between the goals of stochastic and other crafts, so that, according to Stoic doctrine, virtuous action should consist in trying to obtain the things in accord- ance with nature even if one does not obtain them (ibid. 20). At this point Carneades argued against the Stoics that the goal of their art of living ought to be obtaining the things in accordance with nature rather than trying regardless of success. And since the goal of the art of living was assumed to be identical with the goal of life, Carneades maintained that the highest good must be obtaining and enjoying the things in accordance with nature (ibid.). This is, I believe, the theoretical background against which we should set Carneades' most famous piece of anti-Stoic polemic, the 29 I think that the 'sparks and seeds of the virtues' that appear on the list in Fill. 5.18 are out of place here, being an addition from Antiochus, who is Cicero's source (cf. 5. 43)' Lists of 7Tpwm KaTd q,VUlV tend to end with an 'etc.'-e.g. Stob. 2. 47 20-48. s-but they never contain the virtues, and this is no accident. In fact Cicero says explicitly that Carneades' conception of the end did not include virtue: cf. Fin. 5. 22, 2. 35. Following Nature 53 speech he gave on the occasion of his embassy to Rome (ISS Be). That this was indeed an anti-Stoic argument has long been recognized, and appears clearly enough from the fact that Cicero introduced it as an attack on what is evidently a Stoic theory of justice as the foundation of the best state. All the evidence we have about Carneades' speech comes from the fragments of Cicero's De re publica 3 and a few paraphrases and summaries in later authors-mainly Lactantius, Tertullian, and Augustine-all of whom depend on Cicero. Since Cicero no doubt adapted whatever report he may have used to the purposes of his context, and also surely intended the speech to remind readers of the views defended by Thrasymachus in Plato's Republic, we cannot assume that the words he gives to L. Furius Philus accurately repro- duce Carneades. 30 So I would not want to claim that we can recon- struct the actual order of arguments in Carneades' speech, but I think that the outline of his critique of Stoic doctrine is recognizable. The theory that Carneades criticized was set out in Cicero's De legibus (written right after De re publica) and probably in the speech of Laelius that followed that of Philus-Carneades. Cicero argued, with the Stoics, that the foundation of the best state must be the law of nature, which should serve as a standard by which to judge the laws proposed by actual lawgivers (Leg. I. 19)' This natural law, based upon the social instincts of human beings and their mutual benevolence, was said to hold for all persons and all nations, and 'whoever offends against this law thereby forsakes his own human nature and thus suffers the severest punishment, even ifhe escapes the sanctions of the state' (Laelius speaking, Rep. 3. 22. 33)' Knowing and obeying this law will of course be what constitutes wisdom or virtue, as the Stoics held (Leg. I. 19, 58-66). Against this Carneades, it seems, argued in the following way. 1. There is no such thing as a universal natural law. This appears from the fact that different states have all sorts of different and conflicting laws, and that even the laws of a single state can be seen to change as time goes on (Rep. 3. 8. I3-IO. 17)' If there were a universal law, it would have to be the same everywhere (8. 13)' It follows that there is also no natural justice, as a virtue of character. For virtue 30 No doubt Cicero was also drawing directly upon Plato, as for example in the rather pointless rhetorical use he makes of the contrast between the virtuous man who is despised and treated as a criminal and the immortal man held in the highest esteem, taken from Glaucon's speech in Republic 2. 54 Gisela Striker cannot consist in obeying the law that happens to be valid in a specific place and time, since (as the Stoics say) it is constant and unchanging, and nature admits of no such variability (11. IS). Thus far Carneades has obviously argued on the assumption that justice is embodied in the law, and that the virtue of justice must consist in following the law. 2. Now one might say (and the Stoics certainly did say) that the diversity of the laws of actual states is due to general corruption (cf. Cic. Leg. 1. 47; 42). The good man, however, will by nature follow real, not putative, laws, and thus his justice will consist in giving to every- one his due (11. IS). The Stoics would undoubtedly have added that this natural justice is based upon the nature of human beings as social animals. Could one not say that even the worst constitution contains certain laws that are needed as a basis of social life? 3. It seems that at this point Carneades appealed to the contract theory of justice, sketched in Plato's Republic 2 and revived in Hellen- istic times by Epicurus, in order to argue that general benevolence is not needed to explain the existence of some rudimentary form oflaw and justice within human communities. 'The mother of justice is neither nature nor choice, but weakness', which makes humans agree to a contract 'neither to commit injustice nor to suffer it' (13. 23).31 Individuals will obey the law, not spontaneously or out of love for others, but because they fear the sanctions of the law, and could not lead an undisturbed life even if their offences remained secret, since they would always have to live with the fear of discovery-as Epicurus had said (16. 26). So justice, as far as it can be said to exist, will be limited to comply- ing with the existent laws of a given state. 4. The differences between the laws of different states are not due to corruption of the rulers or ignorance of the natural law. They can in fact be explained once one sees that each nation adopts those laws that serve its own advantage (I2. 20-1; Lactantius). And this is a sign of wisdom, not of ignorance; for if states were to adopt the rules of Stoic natural justice-that is, not to infringe upon the rights of others; to care for the welfare of mankind in general; to give to each his due; not to touch the public sanctuaries of other nations-then all the great and powerful nations, first among them the Romans, would have to return 31 One should probably notice that Cicero has taken over Plato's biased wording for the contract, 'neither to do nor to suffer injustice', rather than Epicurus' 'neither to do nor to suffer harm'. Following Nature 55 to their ancestral huts and live in poverty and misery. But humans and other animals, under the guidance of nature (12. 21), pursue their own advantage, and thus act wisely. It is wisdom that tells nations to increase their power and wealth, to extend their empires, to rule over as many others as they can; and so great generals are praised in public monuments by the phrase fines imperii propagavit ('he extended the boundaries of the empire'), which of course could not have happened unless the land had been taken from others. Thus the laws of states can be seen to agree with wisdom or prudence, but they have nothing to do with what is claimed to be natural justice; on the contrary, 'natural' justice turns out to be the greatest stupidity. 5. The same contrast between wisdom and justice can be exempli- fied in the private dealings of individuals. If a man were to forgo his own advantage in order to help or not to harm another, he would be considered just, but not wise; if he pursued his advantage at the expense of others, he would be wise, but not just (19. 29-31; Lactantius). Here Carneades introduced the rather shocking kind of examples known also from Off. 3 (50-6; 89-91), of which I shall cite only one. If a man wanted to sell a house which he alone knew had serious defects, should he reveal the defects to the potential buyer, or keep them to himself? If he were to be honest and reveal the defects, he would be a just man, but a fool; ifhe did not reveal them, he would be wise in looking after his own advantage, but he would not be just. A long list of similar examples is given by Cicero in De officiis, and they seem to have become a standard topic for debate in the Stoic school. While it is evident that Carneades used these examples to demon- strate that wisdom was incompatible with justice, Cicero tells us that the Stoic Diogenes of Babylon would argue that the egoistic actions cited by Carneades as wise, but unjust, were actually compatible with virtue. Diogenes was one of Carneades' teachers, and he was also a member of the embassy to Rome during which Carneades gave his famous speeches. I find it hard to decide whether Carneades was reacting to Diogenes, pointing out that the narrow legalism Diogenes advocated was incompatible with what the Stoics themselves conceived as justice, or whether Diogenes himself was reacting to Carneades in trying to argue that a remarkable degree of cunning and selfishness could still be reconciled with the wise man's virtuous disposition, so that wisdom and justice could go together after all. As regards the examples, the second hypothesis is probably more 56 Gisela Striker plausible, since those stories are so hair-raising that they could hardly have been introduced by Diogenes himself. In any case, Carneades' speech 'Against Justice' led to the conclu- sion that' either there is no such thing as natural justice, or if there is, it is the greatest stupidity' (12. 21; Lactantius). Now of course this attempt to demonstrate that justice and wisdom are irreconcilable, which sounds, and is meant to sound, very much like Thrasymachus, could be taken as no more than an appeal to vulgar conceptions of cleverness, and of success as constitutive of happiness. But this could hardly have impressed anyone who, like the Stoics, believed that most people were bad and stupid anyway, and did not see where their proper good was to be sought. The point of Carneades' argument lies in the assumption that the wisdom that is said to guide imperialists in their conquests and individuals in cheat- ing their neighbours is wisdom in the Stoic sense, that is, observation of nature's rules. This time the argument goes further than the one outlined in the Theaetetus commentary (above, Section 4), in that Carneades seems to be denying universal benevolence altogether, and insisting that a human being's only natural impulse is towards the material advantage of herself or her social groUp.32 If the art of living aims at obtaining the things that accord with one's human nature-as the Stoics held-and happiness consists in obtaining and enjoying these-as Carneades had no doubt argued often before-then wisdom will be exactly what Carneades claims it is, and the Stoic sage will turn out to be, not a paragon of moral virtue, but an intelligent criminal. The argument is also limited to the Stoic theory, I think, since it leaves open the possibility of saying, for example, that the common core of positive laws, those that forbid murder and fraud, are 'natural' in the sense of being necessary for the stability of human commun- ities. 33 As Alexander of Aphrodisias points out (De anima 2. 157. 32 It seems possible that the argument of tbe Anonymus (111 Theaet.) also occurred in the speech, presumably as a rejoinder to the suggestion that human beings do have a natural affinity to other human beings. It could be used to show that even if the sordid examples of cheating were thereby ruled out, g'eneral benevolence would not suffice to produce perfect justice, since there would be at least one type of case in which justice would conflict with wisdom. In any case, Lactantius sets examples of this type off from the others by saying 'transcendebat ergo ad maiora', which suggests perhaps that some objection had intervened, 33 For this line of argument see Alexander ofAphrodisias' essay OT( ,pUUE< TO OiKawv ('That what is just is natural'), De anima 2. 158. 13ft: In the absence of Carneades' speech 'In Defence of Justice', Alexander's little treatise could perhaps provide some idea of what it might have been like-unless we should believe that Carneades' second speech demolished exactly the position he had defended the day before. Following Nature 57 19-2I), even those who hold that justice derives from a contract think that it is natural for men to arrive at such conventions. Besides, the content of those contracts is not arbitrary (ibid. 158.24-7). But this is not an appeal to general benevolence, but to human weakness, and the Stoics had argued against Epicurus that that was not enough. What Carneades' argument purported to show, then, was that the derivation of justice from natural tendencies of human nature alone would not work. I do not think that this can tell us anything about Carneades' own views on moral questions. 34 Nobody even in antiquity believed that he seriously meant to advocate unrestrained egoism and injustice. What his anti-moral theory does is to expose weak points in the Stoic doctrine. There seem to be three main problems, connected with three steps in Carneades' 'theory'. 1. Natural concern. Carneades argued that the ultimate object of desire must be identical with the first. This raises the question how the Stoic theory of the good can go together with the psychological doctrine of oikeiiisis . Ifhuman beings were made by nature to seek self- preservation and self-realization, how can they eventually come to accept agreement with nature as their only good? If agreement with nature is not the same as self-realization, how can one get from the first object of desire to the last? Or if it is the same, how can one preserve and develop one's human nature while totally neglecting bodily and external advantages? 2. Virtue as a craft. According to Carneades, the goal of the art of living ought to be obtaining the things in accordance with nature, not just trying to obtain them. If Carneades were right on this point, it would follow that happiness consisted in precisely those things that the Stoics had taken great care to declare indifferent, not real goods, though objects of rational selection. It would also follow that happi- ness was a matter of chance and good luck, not something that is in our power. 3. Appropriate action. It would seem, on Carneades' account of human nature, that following nature results, not in virtue, but on the contrary in a reckless pursuit of one's own (or one's group's) advantage, at the expense of one's neighbours. This was of course an 34 Contrary to Jeanne Croissant, 'La morale de Carneade', Revue internationale de philosophie, 1 (1939), 545-70, who was herself the first to emphasize the connection between Carneades' speech and Stoic doctrine. There is no reason to think that Carneades preferred Epicurus' theory of justice to the Stoic. 58 Gisela Striker attack on the thesis that nature leads us to virtue, challenging the Stoics to show that Carneades was wrong about human nature. All three problems were taken up, I think, by Antipater, and also later by Antiochus, but in opposite ways. For while Antipater tried to refute the fundamental premisses ofCarneades' objections, Antiochus accepted the first two objections, and then claimed that the 'old doctrine' of the Academy and the Peripatos could accommodate them, and would have no difficulties with the third point. I have already dealt with Antipater's defence of the doctrine of virtue as a craft (Section 3), and so I shall only briefly describe what I take to have been his reaction to the other points. In response to the question about the connection between the first object or objects of desire and the ultimate goal it was probably Anti- pater who introduced the complicated account of psychological devel- opment we find in Cicero (Fin. 3. 20-1), ostensibly as an argument for the Stoic conception of agreement with nature as the goal oflife. I have tried to show that this psychological story cannot very well be taken as an argument in the way Cicero wants to take it, since it would pre- suppose what it is allegedly meant to prove. 35 The crucial step in the psychological development described by Cicero seems to be the reversal brought about by the insight that the only thing that really deserves to be called good is the order and harmony of the universe. The postulated change of attitude is meant to explain, I think, how a human being can come to give up self-preservation or self-realization as a primary goal in favour of agreement with nature, so that the objects of her initial desires are then pursued only because nature apparently wants her to pursue them. Since nature obviously has not provided for unfailing success in the pursuit of the natural things, obtaining these will be indifferent with respect to our agreement with nature's order. Hence the attainment of things in accordance with nature will remain the reference-point of our actions, but it will not be the ultimate goal. This corresponds, of course, to Antipater's subtle account of the stochastic craft that is supposed to be virtue. It is no accident that Cicero introduces the archer simile in the context of his argument for the Stoic conception of the goal oflife. As regards the third point, appropriate action understood as follow- 35 Cicero's error would be understandable if he had-under the influence of Antiochus?-accepted the Carneadean scheme, in which the ultimate end must be derived Ii'om the first object of impulse. Then Antipater's account would seem to be the most plausible candidate for such an argument. Following Nature 59 ing natural human impulses, Antipater's reaction can to some extent be inferred from Cicero's report about his disagreements with his predecessor Diogenes of Babylon (cf. Off. 3.50-6; 91-2). It is clear that Antipater rejected Diogenes' attempts to show that a rather alarming degree of selfishness and dishonesty is compatible with justice. He insisted, it seems, that natural attachment to human society and fellow-feeling for other human beings would rule this out-that is, he must have thought that the desire to be of assistance to others excludes the possibility of seeking one's own advantage at their expense. This shows that Antipater adhered to the standard doctrine of sociability and general benevolence as foundations of justice; and here a Stoic would probably have relied upon anti-Epicurean arguments to show that egoism could not serve as the basis of human social life. After all, the Stoics did not have to accept the Epicurean part of Carneades' argument. It is not clear whether Antipater also tried to refute the more insidious argument according to which even general benevol- ence would not be sufficient to guarantee virtue in extreme situations like the shipwreck case. So much for Antipater. Perhaps the most striking example of Carneades' influence can be found, I believe, in the moral theory of Antiochus of Ascalon, the philosopher who started out as an Academic, but then turned into a dogmatist hardly distinguishable, as Cicero remarks (A cad. pro 132), from a genuine Stoic. Antiochus' version of what he saw as the common doctrine of the old Academy and the Peripatos is set out at length by Cicero in Definibus 5, and used to criticize the Stoics in De finibus 4. Now it is surely significant that the Carneadeadivisio is intro- duced with approval at the beginning of De finibus 5; and in fact Antiochus' theory looks very much like an attempt to fit the 'old Peri- patetic' doctrines into the mould suggested by Carneades. Antiochus apparently accepted without hesitation the idea that the goal oflife must be derived from the first object of desire. 36 He claimed that the older philosophers agreed with the Stoics that every animal's first impulse is for self-preservation-in fact, Antiochus said, the Stoics had simply taken over this doctrine from their Academic and Peripatetic teachers (Fin. 5. 22-3). But then it turns out that on Antiochus' view the class of things pursued for the sake of self- realization includes far more than the Stoics had envisaged-namely 36 I lis Peripatetic contemporaries did the same: see Alexander of Aphrodisias' review of opinions on the question of' The first object of natural concern according to Aristotle', De anima 2. ISO. 20ff. 60 Gisela Striker the virtues qua perfections of human rational capacities. Since the mind is superior to the body, Antiochus argued, its perfection would also be far more important than the healthy state of the body; but as a human being has both body and soul, perfection of the body would have to be a part, however negligible, of the human good (Fin. 5. 34-S). According to Antiochus' theory, then, self-realization is not only the first, but also the ultimate object of desire. Assuming that the Stoics held the same view-and thereby ignoring the distinction between the merely natural life (kata phusin bios) and the life in agreement with nature-he could then criticize them for neglecting the bodily side of self-preservation and treating the human being as though she were nothing but her mind (Fin. 4. 2S; 32-4; 4I). This objection is beside the mark, however, for in so far as self-realization was the reference-point of appropriate action, the Stoics did not neglect the physical aspects of self-preservation; but since the ultimate goal was agreement with nature, not self-realization, they could also hold that all that pertains to self-preservation is merely in accordance with nature, not a real good. Having included the virtues among the objects of the primary impulses, Antiochus could then proceed to accept Carneades' argu- ment about the art ofliving, according to which its goal should consist in obtaining the things in accordance with nature, and to ascribe this view also to the 'older philosophers' (Fin. 4. I5, 25-6; cf. Cic. Acad. post. 22). At the same time, the Stoic distinction between 'preferred' and 'dispreferred' things on the one hand, 'goods' and 'evils' on the other, could be dismissed as a mere verbal manreuvre (Fin. 4. 20-3, 46-S). Moreover, once the virtues were included, problems of priority among things valued could of course be solved by the sweeping rule 'virtue first', which does not tell us much about what virtuous conduct is, but effectively sets aside the complicated casuistry of the Stoic attempts to find out what is more or less in accordance with nature. Antiochus' project of keeping the best of two worlds failed with regard to the self-sufficiency of virtue for happiness. He introduced a distinction between the happy life (vita beata) and the happiest life (vita beatissima) and maintained that virtue was sufficient for happi- ness, but not for the greatest happiness (Fin. 5. SI; Acad. post. 22). But, as the Stoics were quick to point out, this will not work. Once the dis- tinction between preferred things and real goods is abandoned, it will no longer be possible to maintain that the virtuous person who lacks external or bodily goods lacks nothing that could make her life any Following Nature 61 better, and this had been a condition for happiness ever since Aristotle (cf. Cic. Fin. 5. 8r-6; Sen. Ep. 85. 19-23). Above all, by including the virtues among the things desired for the sake of self-realization, Antiochus entirely missed the point of Chrysippus' theory. For if virtue is among the objects of the primary impulses, the theory of natural concern will no longer provide a non- circular account of morally correct action. What Antipater saw, and Antiochus missed, is that the old doctrine of virtue as perfection of the human soul needed to be supported by some argument to show that virtue is in fact the best state of a rational animal, or the completion of a natural development. Antiochus' theory, I submit, was not a good idea. By contrast, I hope that it may serve to underline the merits of the brave-if unsuccessful-attempt of the Stoic philosophers to spell out what it is to act virtuously in terms of following the natural tendencies of human nature. 6. Why no emotion? 'That to which no emotion whatever attaches is better than that to which emotion is congenital, and the law has no emotion, whereas every human soul must have it', said Aristotle (Pol. 3, 1286 a q, trans. Robinson), speaking about the best ruler for a city. Aristotle said this, of course, because he believed that there is an irrational element in the human soul which is the origin of the pathe-passions or emotions. The Stoics, as is well known, did not think that the human soul neces- sarily included such an element of emotionality. They maintained, against Plato and Aristotle, but possibly agreeing with Socrates, that the 'governing part' (hegemonikon) of the human soul is only one thing, reason. If we sometimes behave in ways contrary to reason, and even contrary to our own better insight, this is not due to a part of the soul independent of and not controlled by reason, but to a weakness of reason itself. And if the emotions are 'irrational', they are so in the sense of going beyond or against right reason. The Stoics thought that it was not only possible, but also natural and desirable, that one should be free from all emotion. The wise person, they said, would be entirely without it, apathes. How did they think this was possible, and why would it be right? To see how it would be possible, we must look at the Stoic theory of impulse (horme), since the emotions were considered by the Stoics as 62 Gisela Striker being a kind of impulseY Human beings are not born as rational animals. Before they arrive at the age of reason, their behaviour will be guided, like that of other animals, by their capacity to react, either favourably or unfavourably, to certain kinds of impressions (phan- tasiai). Their natural instinct for self-preservation enables them to recognize certain things or activities as being either in accordance with or against their nature; and if something presents itself as being in accordance with nature or against it, this impression will lead to an impulse either to pursue or to avoid the object so presented. This is the kind of impulse that nature has given to animals to make them 'proceed towards what belongs to them' (D.L. 7. 86). With the advent of reason, however, human impulses take on a different form. The impressions that rational creatures receive have a rational content, that is, they are expressible in language, and for the impulse to action to follow the impression, it is no longer sufficient that something be presented as in accordance with nature or contrary to it, attractive or unattractive. The rational animal will have an impulse to action only if reason has assented to the impression, and if reason functions properly, it will assent only to impressions that are not only true, but also 'cognitive' (kataleptikai), that is, so clear and distinct that what is presented could not possibly be otherwise. It is because of the intervention of assent that adult humans can be held responsible for their actions and beliefs, for assent can be either given or withheld, whereas animals and small children merely respond automatically to the way things present themselves to them. It is also at the age of reason, presumably, that human beings acquire the concepts of good and evil, and it seems that their actions will then largely be guided by these since, as Cicero tells us most explicitly (TusC.4. 12), the human mind has a natural and irresistible inclination towards what it sees as good, and a corresponding aversion from what it sees as bad. All human impulses, including the emotions, are then to be considered as acts of assent to a certain kind of impression, which the Stoics called 'impulse-arousing' (hormitike, Stob. 2. 86. 18). They distinguished four most general classes of emotions, depending on the different types of impressions that provoke them: appetite (epithumia) 37 For recent discussions of impulse and emotion in Stoic theory see e.g. J. Gosling and C. C. W. Taylor, The Greeks on Pleasure (Oxford, I982), ch. 2I; Inwood, Fthics and Human Aaion, ch. 5; M. Frede, 'The Stoic Doctrine of the Affections of the Soul', in Schofield and Striker, The Norms of Nature, 93-IIO. I shall here mention only a few points relating to my specific questions. Following Nature and pleasure (hidoni) are assent to the impression of an impending or present good, fear (Phobos) and distress (lupi) are assent to an impend- ing or present evil, respectively (Cic. Tusc.4- II; Fin. 2. 35; D.L. 7. III; Stob. 2. 88. 14-15). Thus it can be seen that, far from being independ- ent of reason, the emotions are actually one of the ways in which reason directs our behaviour. If other philosophers had tended to think that the emotions were independent of reason, this must be ascribed to the fact that they had overlooked the crucial role of assent. The Stoics would argue that whether we feel pleasure or distress, fear or longing, depends entirely upon whether we accept or do not accept the way things present them- selves to us. This could be shown by comparing the different ways in which people may react to the same impression. The prospect of being killed may appear as a terrifying evil, and if we assent to this, we will be overcome by fear. But if we keep in mind, as a good Stoic would, that loss of life is just one of the things that go against our nature, but are not real evils, we shall resist the impression, and assent at most to the thought that being killed is one of the things to be avoided. So we may indeed try to escape from the danger, but we shall not be afraid, for fear results only if we assent to something's being an intolerable evil. But in people whose reason is not strong enough always to be on guard against unclear or deceptive impressions, assent to such a terrifying impression may come almost automatically, so that they are not even aware of their assent. And thus they will claim to have been overwhelmed by an affection caused by a frightening object, when in fact it was just the weakness of their reason that produced the emotion. Similarly, if something appears as being extremely attractive, as for example the prospect of winning the Nobel Prize, one may simply give in to the impression without thinking and thus be carried away by ambition, whereas a more cautious person would have withheld assent to make sure that the impression that appealed to her was unmistak- ably true; that is, that trying to win the Nobel Prize could be given a reasonable justification. Up to this point we can see, I think, why the Stoics held that the emotions were judgements (kriseis, D.L. 7. III; cf. Cic. Tusc. 4. 14, Stob. 2. 88. 22-89. 2). In fact, it is not quite clear whether one should say that the emotions were simply judgements, or rather movements of the soul caused by such judgements. It may be safest to assume that both the judgement, that is, the assent, and the consequent movement of the soul were covered by the term pathos, but that the accent should Gisela Striker be placed on the judgement, since a movement of the soul would be described as appetite or fear, pleasure or distress, depending upon the belief that caused it (cf. Cic. TusC.4. IS). Now from the fact that all emotions are judgements or caused by judgements we can indeed infer that they can always be avoided, since assent can be either given or withheld; but it is not yet clear why the Stoics thought that emotions are always unreasonable and hence ought to be avoided. But this is what they maintained. Zeno is said to have defined emotion as 'an irrational and unnatural movement of the soul, or an excessive impulse' (D.L. 7. lIO).38 Why should that be correct? Might it not be the case that we sometimes have good reason to be afraid, for example, so that only some, but not all, emotions should be avoided? Given the Stoic distinction between things that have some value, but are neither good nor bad, and real goods and evils, they presum- ably declared all emotions to be unreasonable because they result from our taking to be good or bad things that are actually at best in accordance with or contrary to our nature, and hence strictly speaking indifferent. Thinking of illness as not only to be avoided, but truly bad and relevant to our happiness might lead to an excessive reaction, like tears and lamentations, no longer justified by reason. Now this is probably correct (cf. Sen. Ep. 7S. I I ff.), but it might sound like a mere terminological trick. For if the Stoics agree that we should avoid such things as illness, why did they not identify what others saw as accept- able emotions with our assent to the impression of something as being preferred or dispreferred, in accordance with or contrary to nature? Why, indeed-so Lactantius says-insist on purely verbal distinctions, and speak of caution as being entirely different from fear when we might just as well say that it is a reasonable degree of fear? (SVF iii. 437; cf. Pluto Virt. mor. 449A-B.) Plutarch and Lactantius introduce here the terms for the so-called eupatheiai, states of being well affected, that the Stoics reserved to the wise. The fact that they are 3K The Greek is "AOyo, Kai 7Tapa <pvatv .pvxij' KivYjul> 7j oPfLij 7TAova(ovua; cf. Stob. 2.88. 8. The 'or' in this formula is curious, though not without parallel in Stoic sources. It is not clear whether we are offered two alternative definitions, or whether the 7j is explicative, so that emotion is said to be irrational and unnatural because it is an excessive impulse (or vice versa). Chrysippus apparently tried to show that the two expressions are equivalent; cf. Galen, PHP 4. 2. T T, SVF iii. 462. So perhaps Zeno used both and did not explain the connection. Each is sometimes quoted independently-cf. Cic. '[usc. 4. II, 47; Stob. 2. 39. 5; but mostly thcy go together. I shall assume that Chrysippus was right. Following Nature 65 reserved to the wise shows that these terms cannot be used as descrip- tions of ordinary avoidance and preferring. It is indeed a mistake to think, as Plutarch (Virt. mar. 444 B) and Lactantius do, that all actions must originate in some form of emotion. The Stoics must have thought that most action, whether based on true or false belief, can go on without emotion, so that 'natural' impulses are not weak cases of the same sort of movement that occurs in emotional behaviour. They evidently thought that what they called 'excess' was an essen- tial characteristic of emotion, and indeed it seems to explain both the 'unnatural' and the 'irrational' in Zeno's definition. The falsity of the judgement underlying the impulse is not the only factor that accounts for its irrationality. So Chrysippus said 'it is not in view of the judge- ment that each of these things is good or bad that we call these [sc. the emotions] illnesses, but in view of the fact that people go out of their way for these things more than is in accordance with nature' (Galen, PHP 4. 5. 2I; SVF iii. 480). What happens in emotional states is, rather, that the natural commensurateness (summetria, PHP 4. 2. IS; SVF iii. 462) of impulse and reason is disturbed. This is what makes emotion 'unnatural'. And the emotions are irrational not only in the sense of being due to an erroneous judgement, but primarily in the sense of going beyond or against what one rationally thinks one should do. This special sense of alogos, which Galen claims he cannot understand (PIlP 4. 4. 14-15; SVF iii. 476), is illustrated by the famous simile of the runner whose legs have acquired such momentum that he is unable to stop himself, as we might say, 'at will'. He is running voluntarily, but his running makes it impossible for him to stand still when and where he wants to. So his legs move on, as Chrysippus says, against his impulse (para tin hormin). By analogy, the person who has lapsed into an emotion will not be able to stop herself doing things that, upon reflection, she would not find it right to do-her impulse goes against her reasoning (para ton logon, PHP 4. 2. 15-18; SVF iii. 4 62 ). Galen's difficulties in understanding the sense of 'irrational' Chrysippus wanted to use in describing the emotions may be due to the fact that it includes the more normal sense 'erroneous'. It is neces- sary, but not sufficient, for an emotion to occur that the corresponding judgement be mistaken. If we merely act upon an erroneous judge- ment, for example, that we should eat something which in fact we should not eat, we may still be acting exactly as we think we should, and this will become evident if we refrain from eating when our error is 66 Gisela Striker pointed out to us (cf. PHP 4. 4. 25-7; SVF iii. 476). There will be an emotion only if the impulse is so strong as to overcome the control of reason, so that it can no longer be changed 'at will'. And this will typically happen when things that are in fact indifferent are seen as real goods or evils. If emotion essentially involves excess, then mere preferring or avoiding does not involve any emotion at all. This is indeed not implausible. For example, ifI go out in the rain, I will take an umbrella if! happen to have one, but I will not be either upset or distressed if! get wet, nor will I necessarily snatch my neighbour's umbrella because I ardently desire one. The Stoics are suggesting that our normal attitude towards preferred or dispreferred things should be just like this-to take them if we can, but not to go out of our way to obtain them, or to avoid them if we can, but not to be upset if we cannot. 39 By banishing all emotion, then, the Stoics did not deprive ordinary human beings of all springs of impulse; they only meant to exclude those impulses that are apt to interfere, by their excessive strength, with our capacity to live according to our best insight. So they could say that the wise person will not be emotionless or unaffected (apathis) in the sense of being insensitive, 'harsh and relentless' (D.L. 7. 117, cf. Sen. Ep. 9. 2-3)-she will have all the normal inclinations and aversions, but no excessive ones. As I said before, emotionless impulses of the ordinary person should not be confused with the so-called eupatheiai of the wise. These are all directed at real goods or evils, things that it is reasonable to rejoice about, or to desire, or to beware of. The Stoics admitted only three kinds of eupatheiai, defining them as counterparts to pleasure, appetite, and fear, namely joy (khara), defined as reasonable elation, wish (boulesis)-reasonable desire-and caution (eulabeia), reasonable aversion (D.L. 7. II6, cf. Cic. Tuse.4. 12-14). There is no counterpart "J It is possible that some Stoics also recognized that there might be harmless, non- emotional cases of satisfaction or dissatisfaction, just as there seem to be non- emotional impulses to do or avoid something. This might explain their wavering in the case of pleasure (cf. Gosling and Taylor, The Greeks Oil Pleasure, 416-17)' According to their general theory of the emotions, pleasure ought to have been rejected; yet it seems that some Stoics-including even Chrysippus-accepted it as something natural, though not of course a good. I would think that the term ",oovTJ might have been used both for the excessive impulse described as 'irrational elation' and for the satisfaction that could arise out of the judg-ement that one has attained one of the natural things. The latter would be acceptable and natural, the former not. It may be relevant that Cicero tends to translate ",8ovTJ, in the context of his treatment of the emotions, as laetilia gesticils. Following Nature to distress (lupe), presumably not only because the wise person never slides back into vice, and hence will never be confronted with an evil in her own person, but also because the real evils she sees around her- self, in the foolishness of the rest of mankind, must be considered to be part of the order of nature, and hence providing no reason to be upset. Like right actions (katorthOmata), the eupatheiai are reserved to the sage, because she is the only person who truly recognizes real goods and evils and therefore desires and shuns the right objects, and who does so without wavering. The desire of a fool for wisdom is presumably reasonable, but since it is not unwavering, it apparently cannot count as a case of eupatheia. By introducing these counterparts to ordinary appetites, pleasures, and fears the Stoics seem to make room for the old Platonic and Aristotelian requirement that the good life should also be the most pleasant. They also emphasize that some things are seriously to be desired, not just preferred, or seriously to be avoided even at great cost. The counterparts or 'contraries' (D.L. 7. r r6) of the emotions can of course never become excessive, since they only reinforce the wise person's firm agreement with the order of nature. So the Stoic sage, who has come to virtue by learning to value the order and harmony of nature as the only true good, and to regard as indifferent the preferred and dispreferred things she is trying to attain or to avoid in everyday life, will lead a life of undisturbed joy in the contemplation of the natural order of things and her own agreement with it, and she will pursue the good she has seen, and avoid evil, with unwavering deter- mination. Thus far, I think, one might admit that the psychological theory the Stoics are offering is not unintelligible, and perhaps even rather plausible in its diagnosis of the irrational character of certain emotions. If we still find that there is something inhumane about the portrait of the wise man who lives in constant joy about his own and his friends' virtuous conduct, who is never afraid, but just cautiously avoids any action that might not be entirely justifiable by reason, it is not, I believe, because we think that it is even theoretically impossible to be thus free from all disturbing emotion. That it would be extremely difficult to reach this state of complete serenity and detachment from worldly affairs the Stoics would be the first to admit. What is discon- certing about the portrait of the Stoic sage is rather that it is presented to us as a portrait of perfect virtue. It is disconcerting to be told that the wise person will indeed love 68 Gisela Striker her friends, if they are virtuous-true love is not an emotion, but another privilege of the sage-but not to the extent of being distressed if one of them dies, or longing for them when they are absent, or being pleased upon seeing them again (cf. Cic. Tusc. 4. 72, Sen. Epp. 9 passim, 59. 1-4) It is no comfort either to think that this virtuous person will risk her life to save a drowning child, but that she will not be sad or disappointed if she fails, being content with the reassuring knowledge that what she did, and also what happened, was in accordance with nature. What has gone wrong here is not, I think, the suggestion that we could be without emotion, but that we should try to be. And the reason for this lies not in the Stoics' theory of emotion, but in their theory of what is good or bad, and consequently what should make us feel elated, what we should seriously wish for, and what we should make every effort to avoid. It is obvious that the distinction between the wise person's 'good states' and the emotions of the fool corresponds to the distinction between real goods and evils on the one hand and indifferent, though possibly preferred or dispreferred, things on the other. The emotions are described as judgements in terms of good and evil. Joy, wish, and caution are reasonable because directed at real goods and evils; ordinary emotion is never justified because directed at things that are indifferent, but taken to be good or evil. The Stoics claimed that acceptance of their system of values was a prerequisite of virtue. They said that virtue would 'have no foundation' ('non posse constitui', Cic. Fin. 4. 40) if anything besides virtue and vice were to count as good or bad. By this they probably meant that if anything besides virtue were considered as a good, we could never exclude the possibility that other goods outweighed virtue, if not individually, then collectively, so that one might come to act immorally for the sake of some non-moral goods-for example, trying to save one's life by some act of injustice. But in fact the Stoics offered no cogent argument to show that perfectly virtuous conduct could not be achieved if one adhered to the postulated natural order of preferences, thereby coming to believe, as Cicero suggests, that harming another person is worse than anything else that might befall one, and yet regarded the 'preferred' and 'dis- preferred' things as good and bad, respectively. One could accept the Stoic distinction between the desire to follow the laws of nature and the desire to obtain what belongs to a natural life, even say that the natural life should be pursued only because this is nature's will, and Following Nature 69 still consider the natural advantages or disadvantages as goods or evils, though different in kind and less important than agreement with nature. If one was convinced that going against nature was worse than any other evil, why would one be tempted to disobey? Why can there be degrees of value in the field of natural advantages, but not among goods and evils? Since joy, wish, and caution were said to be reasonable attitudes with respect to goods and evils, one could then suggest that those would also be appropriate responses to natural or unnatural things, provided that they were proportionate to the value assigned to the good or bad things on the natural scale. This would in effect be what the Peripatetics called metriopatheia, the disposition to feel the right degree of various kinds of emotion, with the difference that where the Stoics would speak of the contrast between eupatheia and pathos, the Peripatetics would speak of a correct or excessive amount of one and the same thing, emotion (pathos). But the Stoics vehemently rejected such a suggestion, saying that to speak of an adequate amount of pathos is as absurd as to speak of an adequate amount of vice or illness (Cic. Tuse. 4. 39-42). The Stoics' refusal to group eupatheiai and pathe together might look at first sight like sheer obstinacy, given that the eupatheiai do seem to be what we would call emotional responses, and were described by the Stoics themselves in the same way as the pathe, namely as elation (eparsis) or desire (orexis) (cf. D.L. 7. II6). It has certainly led to some confusion in the interpretation of their thesis that the sage will be apathes. However, it was part of the philosophical tradition to see the emotions as irrational, albeit in the sense of non- rational, and hence the expression 'rational emotion' might have sounded like a contradiction in terms. Since the good affections of the wise person were emphatically not supposed to be irrational, the Stoic terminology has a point. Now the Stoics evidently thought that there could be no appropri- ate attitude or response to the natural or unnatural things that goes beyond mere taking or avoiding. By declaring everything except virtue and vice to be indifferent, they implied that no emotional attachment or aversion to anything else could be reasonable, and hence appropri- ate for the virtuous person. The oddity of this theory of value as a foundation for virtue can perhaps best be illustrated by the examples of courage and self- control or temperance. According to the Stoics, the courage of the wise man consists, as one might expect, in the total absence of fear, Gisela Striker based upon the firm conviction that nothing that happens to a human being can do him any harm. Chrysippus is said to have defined courage as 'knowledge of what is to be endured, or disposition of the mind in suffering and enduring obedient to the supreme law, without fear' (Cic. Tusc.4- 53). There might be a sense in which one could say that courage consisted in not being afraid-thus we speak of courageous explorers or mountaineers, people who apparently do not fear things most others would find rather terrifying. But on the other hand, one might ask what courage would be needed for if there was nothing to be feared? So Socrates reminds us in the Laches (I93A 3-I8) that the knowledge that there is no real danger makes for confidence, not courage. If we admire the courage of Socrates himself on his last day, it is not because he did not mind being poisoned, but because he was willing to give up his life for the sake of his moral convictions. The same goes for temperance: what is going to be so admirable about self-control when there is nothing to be sacrificed in giving up certain advantages? To use an example introduced by Chrysippus himself, and a favourite of Plutarch's (cf. Stoic. repugn. I038E-F, Comm. not. I06oF-I06IA), we do not admire the man who 'temperately abstains from an old woman with one foot in the grave'. We might find it admirable, perhaps, if someone gave up her long-deserved vacation to look after a sick friend, but our admiration would not be increased by being told that she did not care for the vacation anyway. And we would, I submit, be rather disgusted if it turned out that she was actually pleased to find such an excellent opportunity to display her virtue. But this is exactly what Seneca, this great admirer of heroic virtue, suggests when he admits (Ep. 66. 52) that one should wish for great calamities in one's life because only thus could one really prove one's virtue. Not only was Socrates happy in goal, it seems-he is actually to be envied for his fate. (It should be said, to Epictetus' credit, that he thinks this is nonsense; cf. Diss. I. 6. 35-6.) Seneca is trying to account for the fact that the courage and endur- ance of a martyr or a hero seem more admirable than even the firmest adherence to the principles of virtue in a person who is never made to suffer for her convictions. By maintaining that all external things are indifferent, the Stoics had left themselves no plausible way of making such a distinction of degree. On their theory, it might indeed seem most plausible to follow Seneca in saying that what is most admirable is also most desirable. But in this case that seems to be rather perverse. If virtue is not only hard to attain, but occasionally hard to maintain, Following Nature this is so because it may, though it need not, require the sacrifice of real goods, or the acceptance of real evils. And as it is not reasonable to desire evils, so it is also not reasonable to desire the fate of a martyr. One might perhaps reasonably wish to be or become like Socrates, but not to suffer Socrates' fate. The Stoics would in fact have agreed with this, since it is in accordance with nature to avoid pain and suffering, but they could not say that a life without serious hardships is happier than a life of suffering and torture, and as Seneca's example shows, if they had admitted degrees of goodness in the field of'real' goods, they might have been led to say that the happiest life is that of the wise man on the rack, since it also seems to display the greatest virtue. The Stoics were no doubt right in saying that we evaluate others and ourselves as good or bad people according to moral standards and regardless of success or bodily attractions (Sen. Ep. 76. 11- IZ), but this is not because moral virtue is the only true good, but presumably because virtuous conduct contributes or tends to contribute to the well-being of all members of society, and we evaluate people as good or bad qua members of society. The moral value of a person is indeed not a matter of her success, beauty, or wealth, but it does not follow that the value of virtue in general has nothing to do with the material welfare of human beings in general. Hence the contentment of the Stoic sage who is elated by her grandiose display of virtue in spite of her failure to achieve what she set out to do seems out of place, for in congratulating herself she shows contempt for the very things that made, not indeed this particular action, but this way of acting, good and admirable in the first place. To come back to apatheia, then: in so far as the Stoics wanted to say that we ought not to be upset or excited about things that have no real value, we might perfectly well agree with them, but we should reject their 'freedom from emotion' on the ground that it makes us indiffer- ent to things we ought to appreciate. Far from being a necessary condition of virtue, Stoic apatheia actually seems to be incompatible with it. But by rejecting emotionlessness, we would also have to give up a claim that was clearly extremely important, not only to the Stoics, but to most Hellenistic schools (with the honorific exception of the Peri- patetics), namely that happiness is entirely in our power, depends upon nothing but ourselves. If such things as the well-being of friends and family, or even-horribiledictu-our own health were to count as goods, and their opposites as evils, then obviously virtue would not be Gisela Striker sufficient to guarantee a life in which no good is lacking (cf. Cic. Tusc. 5. 83-5; Epict. Diss. I. 22. 13; Sen. Ep. 95. 14ff). In fact, our own attitudes or decisions (proaireseis), as Epictetus frequently insists, are the only thing that really is in our power (e.g. Diss. 2. 5 4-5). Epictetus goes on to suggest that good and evil must therefore also lie in our decisions or, as he most often puts it, in the right use of impressions (khresis hoia dei phantasion, e.g. Diss. 1. 20. 15-16). Now if one could attain the wholeheartt:d acceptance of the order of nature the Stoics describe as virtue, then one would have achieved what Zeno called the 'easy flow of life' (euroia biou), and also what later Stoics like Seneca and Epictetus call tranquillity or peace of mind (euthumia and ataraxia), the inner state of the happy person. Stoic freedom from emotion, that is, indifference towards everything except conformity with the order of nature, is of course just the state of mind that would guarantee tranquillity. Note that it is not the same: if one cared about nothing but agreement with nature, but could not be certain to achieve this, one would have reason to be disturbed. But as it happens, the same process by which we learn that all external things, and even our own bodies, are valueless compared with the rational order of the universe also leads us to see what this order is. And since reason is irresistibly drawn towards what it has clearly seen to be good, this wisdom, once achieved, cannot be lost again. So the wise man knows that all the good he desires is within his reach, and all evil can be avoided. Hence his mind will be ever serene, 'as clear as the sky above the moon', to use Seneca's impressive image (Ep. 59. r6). This has proved to be an attractive ideal for many generations of moralists. But is it a plausible conception of the human good? Tranquillity was only the psychological side of happiness,4o and it seems obvious that the Stoics were primarily interested in showing that the human good is a life of virtue. The claim that 'nature leads us to virtue' might have been defended, I think, without the strong thesis that agreement with nature is the only good. If, as I have argued, virtue not only does not require, but actually excludes indifference to the welfare of other human beings, then one will have to abandon the claim that it brings absolute equanimity and peace of mind. Which should lead one to wonder, not indeed whether virtue is a necessary part of the best human life, but whether tranquillity 40 For this point, and the notion of tranquillity in Hellenistic ethics more generally, see my 'ATARAXIA: Happiness as Tranquillity', Monist, 73 (1990),97-1 IO. Following Nature 73 as conceived by the Stoics is necessary for happiness, or even desirable. If the Stoics misdescribed virtue in order to guarantee its un- wavering stability, they also, I think, misdescribed happiness in order to make it depend upon nothing but ourselves. IIaroard University