Irwin Review Of Annas

REVIEW ESSAY

Happine ss,

Vi rtue,

and

Mo ral ity* T. H. Irwin i Though Greek ethics has long been one of the most intensively studied areas of Greek philosophy, Annas's book is both innovative and welcome. I cannot think of anyone who has even tried to do what she has done with conspicuous success; a careful reading of her book will be both compulsory and enjoyable for everyone with any sort of interest in Greek ethics. It is full of instruction for the specialist, since it does not shirk detailed discussion of controversial issues. But it is also accessible to the less specialized reader. Annas does not presuppose familiarity with the philosophical and interpretative issues she discusses, but provides an admirably lucid and informative introduction before going on to present her own position. She refers frequently to primary and secondary sources and so introduces readers to what they need to read to pursue the issues further. A large mass of source material, and a series of complex and interconnected issues, are skillfully organized into a book that offers the reader both a readable narrative and a sustained and persuasive philosophical argument. Annas tells us, "The book, as I emphasize, is not a comprehensive study of ancient ethics" (p. v).1 It is a study of some leading themes in ancient ethical theories and, more specifically, of the themes that define the distinctive character of ancient theories in contrast to modern ethical theories. The point of Annas's constant comparisons, parallels, and contrasts between ancient and modern theories is not to make the ancient theorists seem appealing to the modern tour-

* A review of Julia Annas, The "Morality of Happiness (Oxford: Oxford University Press, 1993), pp. x + 502. 1. Unfortunately, this point escaped the writer of the jacket blurb, which claims that this book is "the only comprehensive treatment of ancient ethical theory." The blurb also claims that Annas's books include eleven volumes of Oxford Studies in Ancient Philosophy.

Ethics 105 (October 1994): 153-177 © 1994 by The University of Chicago. All rights reserved. 0014-1704/95/0501-0007J01.00

1

2

Ethics

October 1994

ist—either to the adventurous tourist with a taste for exotic dishes or to the reluctant tourist who prefers fish and chips—but to see what is distinctive and important about their concerns and arguments. Her discussion of "ancients" is deliberately restricted. Socrates, Plato, and the Cynics are mentioned only occasionally,2 and Annas concentrates on Aristotle, the Stoics (usually the earlier Stoics), Epicureans, Skeptics, Cyrenaics, and the later Peripatetic defenses of Aristotle (see pp. 17-20). This restricted focus is quite reasonable in the light of Annas's general aim of examining and correcting some modern assumptions about Greek ethics. In the seventeenth and eighteenth centuries some writers tended to regard the Stoics as the only Greek philosophers who formed a genuine conception of morality; and so Plato and Aristotle were often left out of the discussion. In the last two centuries or so, however, the revival in the philosophical study of Plato and Aristotle has sometimes tended to overshadow the study of Hellenistic ethics. Annas's treatment avoids any one-sided focus either on Aristotle or on the Hellenistic writers. Many of the earlier critics who have supposed that a clear conception of morality is absent from Aristotle have assumed that something like the principles of natural law prescribe the basic elements of morality and that a philosopher who overlooks these principles cannot have a clear conception of morality. This is why Jean Barbeyrac believes that Aristotle misses the central elements of genuine morality: "Thus this vast genius of nature, this philosopher, for whom such numbers have so great a veneration; proves to be grossly ignorant of, and without any scruple treads under foot, one of the most evident principles of the law of nature."3 To illustrate his charge, Barbeyrac espe-

32. There are good reasons for writing a book confined largely to Aristotle and his Hellenistic

33.

successors. But the reason that Annas gives for leaving out Socrates and Plato is not a good one. She mentions the difficulties that (she believes) we face in deciding how many of the views assigned to Socrates in the dialogues can confidently be attributed to the historical Socrates or to Plato (p. 19). This is rather surprising, (a) Given the difficulties that students of Stoicism face in reconstructing the details of the Stoic position, largely from reports by other people, it is strange to complain about difficulties in interpreting Plato, {b) Even if we don't know that they are Plato's views, they are interesting enough to deserve discussion. Nor ought we to ignore their historical influence. Chrysippus, e.g., apparently discussed the ethical argument of the Republic in some detail. See Plutarch Stoic. Rep. 1040d. (c) In some cases attention to Plato allows us to see significant points in later discussions that would not be so clear if we looked at the later discussions without thinking of Plato. Annas notices such cases on p. 48, n. 3, and p. 428, n. 1. In the rest of this review I will mention a few other places where it is useful to keep Plato in mind. Jean Barbeyrac, An Historical and Critical Account of the Science of Morality, trans. Carew, which is printed as an appendix to Pufendorf on the Law of Nature and Nations, 4th ed., trans. B. Kennett (London, 1729), p. 66. Barbeyrac is discussed by R. Tuck, Natural Rights Theories (Cambridge: Cambridge University Press, 1979), appendix.

cially condemns Aristotle's willingness to permit abortion in some circumstances.4 The example is not entirely fortunate, since it may not strike everyone as a gross violation of natural law. Among recent writers, Alan Donagan is sympathetic to the conception of morality as natural law, and he also argues that Aristotle has no clear conception of morality: "The Stoics, rather than Aristotle or Plato, are to be credited with forming the first reasonably clear conception of morality. . . . Obviously there is a strong moral element in Aristotle's theory of ethical virtue; but he did not succeed in distinguishing moral virtue as such, the virtue of a man as a man, from political virtue, the virtue of a citizen of a good city."5 This specific charge is difficult to defend as it stands, 6 but it is easy to suppose that something like Donagan's charge identifies a significant shortcoming in Aristotle's position. Annas does not explicitly discuss the claims of Barbeyrac and Donagan, but they provide a useful starting point for reflection on her discussion, since she discusses several of the issues that they raise. By examining both Aristotle and the Hellenistic moralists, she allows the reader to see not only the elements of continuity in ethical discussion, but also the Hellenistic developments that give some support to Barbeyrac and Donagan. Hellenistic moral philosophy is difficult to study in detail because the Greek and Latin sources are somewhat scattered and not always easily accessible. Annas assembles the relevant evidence and presents it clearly, so that we can understand the philosophical point and interest of the issues discussed by Aristotle and his successors. Annas brings out very clearly both the importance of Aristotle's position for understanding Hellenistic debates and the important advances made by Hellenistic moralists; in particular she does justice to the later Peripatetics who saw that Aristotle's position needed defense and, in Annas's view, succeeded to a considerable degree in modifying it to make it more defensible. "Modern" ethical theories are less precisely characterized. For purposes of comparison with ancient theories Annas has in mind "one or more of the range of types of ethical theory which enter into debate and dialectic among moral philosophers at the present day" (p. 16). These normally include "theories of a consequentialist, deontological

Irwin

Happiness, Virtue, and Morality

3

4. Aristotle Politics 1335bl9-26^

34.Alan Donagan, The Theory of Morality (Chicago: University of Chicago Press, 1977), p. 4. 35.Donagan's claim seems to conflict with Aristotle Politics 3.4, where Aristotle's discussion would be unintelligible unless he drew the distinction that Donagan says he did not draw. Since Aristotle thinks it is worthwhile to argue that a good person must be a citizen of a good city, he must be able to distinguish the two concepts. Perhaps Donagan means to object to Aristode's view that the virtue of a man qua man must include political virtue (for the reason suggested in Nkhomachean Ethics [EN] 1142a9-10); but that objection presupposes that Aristotle has a conception of moral virtue, not that he lacks one.

and virtue ethics type." In practice, however, she tends to treat deonto-logical theories as Kantian, and consequentialist theories as utilitarian, so that she refers freely both to Kant and to the utilitarian tradition, as embodied especially in Sidgwick. Annas acknowledges quite candidly that her view about which questions, contrasts, and generalizations are important might well need to be modified in the light of a fuller examination of the history of modern ethics. I will simply raise one question prompted by her observation: Might fuller historical understanding not cast doubt on the usefulness of the division between "ancient" and "modern" altogether? If we suppose that "ancient" goes up to Marcus Aurelius, and that "modern" goes back to Hobbes or even to Machiavelli, that still leaves thirteen hundred years or so out of consideration. If we take account of medieval ethical theory, ought we to treat Annas's contrasts as contrasts between "ancient-cum-medieval" and "modern" views, or as contrasts between "ancient" and "medieval-cum-modern" views? Or do these divisions miss the points of greatest historical and philosophical significance? Once we even notice the complexity of these questions, we can see why Annas is wise to put them aside. Her decision to limit the scope of her comments on "modern" ethical theories is quite understandable, considering the amount of material that she already has to organize. It is less clear that she is wise to generalize as much as she does about modern ethical theories. In attributing views to the ancients, Annas cites chapter and verse; it would have been useful if she had always done the same in comments on modern theories. The book begins and ends (like the Nicomachean Ethics) with happiness.7 Part 1 gives an outline of happiness, virtue, and morality in ancient ethics. Parts 2 and 3 articulate this outline by discussing the role of nature as a foundation for ethical theory, and by examining the relation between the agent's good and the good of others. Part 4 sums up this discussion by considering the place of virtue in happiness. Annas believes, quite reasonably, that we cannot reach an informed view on the eudaemonistic character of Greek ethical theories until we have seen what these theories have to say about questions that

7. Annas does not point out that the structure of her book has this ancient precedent, because she takes a highly skeptical view about the unity of the Nicomachean Ethics (p. 216n.) and so does not even think it is worth asking the question whether book 1 and book 10 have conflicting views about the nature of happiness (though she answers this question at p. 261 n.). Her skepticism seems to me to go too far. Nichomachean Ethics 10.6-7 refers back several times to previous discussions that seem very similar to those in the previous books of the EN, and it seems quite legitimate to consider how far the following claims about happiness are consistent with books 1—9. Annas endorses the doubts of A. J. P. Kenny (The Aristotelian Ethics [Oxford: Clarendon Press, 1978], pp. 29-36) about the unity of the EN; these doubts are convincingly answered in J. M. Cooper's review of Kenny, Nous 15 (1981): 381-92, p. 389.

both ancients and moderns recognize as central for ethical theory. This arrangement of the book involves some repetition and dispersion; but the plot is easy to follow, and when she returns to the same questions, Annas has prepared us to see them in a new light. Annas's detailed exposition of different theories is full of interest and deserves detailed discussion by specialists. Here I will confine myself to some issues that bear especially on contrasts between ancient and modern theories—both the contrasts that Annas rejects and those she seeks to display. II Annas presents the aims and structure of ancient ethical theories by contrast with the ambitions of modern theories. She believes that "a great deal of discussion of ancient ethical theory is flawed . . . because the authors have carried over to their study of the texts, unquestioned, their own expectations as to what an ethical theory is and should do" (p. 3). What are these misleading expectations? It is a widespread modern assumption, in Annas's view, that an ethical theory must be (1) "hierarchical," in the sense that "some set of notions is taken as basic, and the other elements in the theory are derived from these basic notions" (p. 7), and (2) "complete," in the sense that "the theory claims to account for everything in its area in terms of the basic concepts and of other concepts insofar as they are derived from the basic ones" (p. 8). Ancient theories do not satisfy these demands. They include notions of the agent's final end, of happiness and of the virtues, but these are "primary" without being "basic": "These are the notions that we start from; they set up the framework of the theory, and we introduce and understand the other notions in terms of them. They are thus primary for understanding; they establish what the theory is a theory of, and define the place to be given to other ethical notions, such as right action.

4

Ethics

October 1994

However, they are not basic in the modern sense: other concepts are not derived from them, still less reduced to them" (p. 9). In these ways the aspirations and aims of ancient theories are different from those of modern theories. Are these correct or useful contrasts? I don't see that Annas's description of modern theories matches very many of them. Admittedly, it matches a certain kind of teleological theory that begins with some concept of the good, understood independently of the morally right, and then defines the right as what produces the good. I do not see, however, that deontological theories standardly claim to derive the good from the right; characteristically, they take the right to be independent of the good without claiming that the good is derivative from the right. One might defend Annas by arguing that I have understood 'ba-sicness' too restrictively, and that a deontological theory counts as hierarchical because it makes the right morally prior to the good, even though it does not strictly reduce the good to the right. Annas quotes with approval Mackie's account of basicness: "But what would make a theory X-based in the most important sense is that it should be such a system not merely formally but in its purpose, that the basic statements about Xs should be seen as capturing what gives point to the whole moral theory" (quoted, pp. 7-8). Mackie's criterion includes some claim about derivability, but the type of basicness that Annas has in mind might best be captured by omitting the 'merely' from the sentence just quoted. It might plausibly be argued that many modern theories, while not hierarchical in the stricter sense, are none the less hierarchical in this broader sense. Perhaps Annas means to claim that ancient theories are not hierarchical in the broader sense. At any rate, this claim deserves discussion. In most ancient theories statements about happiness seem to capture "what gives point to the whole moral theory." The point of having an ethical theory is not primarily to find out which actions or states of character are virtuous, but to find out what happiness is and how to be happy; moreover, we have sufficient reason for being virtuous and for doing the actions required by virtue, insofar as this is the way to secure our happiness. Such a view certainly does not imply that happiness is definitionally or epistemologically prior to the virtues or to right action, but it implies some priority in justification. This is one way to express what Aristotle has in mind when he claims that an account of happiness provides a "principle" (arche; EN 1098a33-bl 1), and that practical reasoning begins from the principle of the form, "since this is the character of the end and the best."8 I am not sure whether Annas agrees or disagrees with what I have just said. She discusses the role of the concept of happiness at the "entry point" for moral reflection (p. 27), but she does not say much about the role of an appeal to happiness in providing justifying reasons for our choice of actions or of lives. This justifying role is most easily seen in book 2 of Plato's Republic. Glaucon and Adeimantus take it for granted that they will have found a good reason to be just if and only if they see how justice promotes their happiness; this is what makes the question about the relation between justice and happiness an urgent question for people who are wondering about whether they have good reason to be just. I don't know whether Annas believes that this justificatory role for happiness is peculiar to Plato's ethical theory and is discarded by later theorists. If she does believe this, it would be interesting to know why she believes it; I do not find it easy to believe.

8. Aristotle EN 1144a31-36.

If Annas agrees that an appeal to happiness has this justificatory role, then ancient theories are hierarchical in the broad sense. Since it would be difficult to show that most modern theories are hierarchical in any more restrictive sense, I doubt whether Annas's appeal to hierarchy marks out a difference between the ancients and the moderns. Annas also seeks to contrast ancient with modern theories on the relation of moral theory to moral practice. She claims, "It is a common modern assumption that a moral theory should help us to decide what it is right for us to do, and in particular that it should help us to resolve moral dilemmas and difficult moral cases" (p. 6). Annas is well aware that ancient theories are often thought to be relevant to practice, but she sees a difference from modern theories: Ancient theories assume that the moral agent internalizes and applies the moral theory to produce the correct answers to hard cases; but the answers themselves are not part of the theory. Nor are they produced by the theory in the sense that applying the theory to a simple description of a hard case will automatically generate a correct answer. Thus for ancient theorists it is true that there is not much to be said in general about hard cases. Modern theorists often see it as a demand that they be able to generate answers to hard cases in a comparatively simple way; and to this extent ancient ethics fails to meet modern demands on casuistry. [Pp. 6-7; cf. p. 443] Annas's picture of a "modern" theorist might fit some utilitarians setting out to teach courses in applied ethics, if they claim that utilitarianism provides an effective method for deciding who should be allowed to use kidney machines, what we owe to future

Irwin

Happiness, Virtue, and Morality

5

generations, and so on. These specific practical ambitions, however, do not seem typical of modern theories. Sidgwick, for instance, is often cited by Annas as a paradigmatic modern theorist, but he firmly insists that his aim is the reverse of Aristotle's because it is theory rather than practice.9 The conclusions of Sidgwick's argument bear out his initial claim; for he argues that if we are convinced of the truth of utilitarianism, we may well find it advisable not to apply it openly to our moral practice. 10 Nor does it seem characteristic or typical of Kantian or intuitionist theories to promise relatively direct solutions to moral problems or dilemmas. On the other side, ancient theories seem to respond to definite practical concerns. In the Crito Socrates tries to justify his decision to stay in prison; he argues from rather high-level ethical principles

9. H. Sidgwick, The Methods of Ethics, 7 th ed. (London: Macmillan, 1907), p. vi. 10. Ibid., p. 490.

affirming that justice is sufficient for happiness and that justice prohibits returning evil for evil. In the Gorgias Socrates and Callicles argue about whether Callicles is or is not misguided in his pursuit of power and status. In the Laches the discussion is about how to educate the young men Aristeides and Thucydides. In the Euthyphro it is about whether Euthyphro is right to prosecute his father. Aristotle clearly does not shrink from casuistical questions. On the contrary, he regards it as his business to discuss them (EN 1103b34-1104all, 1164b27-1165a4). Annas might argue that answers to casuistical questions are not strictly "part of the theory." But I am not sure how the distinction between a "part of the theory" and an "application of the theory" can be drawn in a way that will mark an important difference between ancients and moderns. I wonder whether Annas's contrast, and in particular her reference to "hard cases," might result from attention to the fact that modern theorists often focus on specific examples in order to distinguish one theory from another, or to show how all moral theories lead to awkward conclusions. Utilitarians focus on cases in which it seems absurd (for instance) not to tell a lie in order to save a life, and antiutili-tarians focus on cases in which the pursuit of utilitarian reasoning to its logical conclusion produces repellent results. This appeal to examples, however, does not mean that modern theories are primarily devised to tell us when it is or is not right to lie to a murderer, or when we ought to cut up one person to perform lifesaving operations on ten people, or what we ought to do about runaway trolley cars. Attention to these examples (in which the nonmoral facts are usually stipulated with unrealistic precision)11 helps us to understand different views about what the morally relevant considerations really are. For this purpose, counterfactual examples may often be as good as, or better than, actual examples. This theoretical use of actual or counterfactual examples does not seem to distinguish modern from ancient theories. Book 2 of the Republic appeals to mythical and counterfactual examples (in Gyges' Ring and in the comparison of lives) to explain what Plato thinks is the crucial theoretical issue about the value of justice. Cicero seems to understand this use of examples quite well when he compares Plato's with Epicurus's view on justice.12 Once again, then, I am not convinced that Annas has picked out a genuine contrast between ancient and modern theories.

11. This appropriateness of such stipulation may itself be a point of theoretical dispute. See P. T. Geach, The Virtues (Cambridge: Cambridge University Press, 1977), pp. 106-8, 115-16, 143-44. 12. Cicero De Off. 3.39. Annas herself discusses cases of this sort at p. 273.

Though Annas believes that the character and aims of ancient moral theories differ from those of modern theories on the points I have mentioned, she believes that the actual content of ancient theories is not as alien as has sometimes been maintained. She believes that once we examine some of the details, we can see the error in some common modern criticisms. Among the stock criticisms she rejects are these: (1) The ancients have no conception of genuinely moral virtues. (2) Their eudaemonism makes their views "egoistic" in some morally controversial way. (3) They attempt to justify their views by appeal to some conception of "nature" that has been superseded by modern science. Annas describes some of these criticisms as "drearily familiar" (p. 440). One of her aims is to show that the eudaemonist and naturalist elements in Greek ethical theories are misunderstood by critics who suppose that the theories are open to these criticisms. In rejecting the criticisms, Annas rejects some spurious (as she believes) contrasts between ancient and modern theories and tries to replace these spurious contrasts with a clearer sense of what is really distinctive and important about Greek approaches to ethical questions. On the points about egoism, morality, and nature, Annas believes that ancient and modern views are not as different as the "drearily familiar" criticisms take them to be. I believe there is some truth in the familiar criticisms, and so I do not find them

6

Ethics

October 1994

dreary. Some modern moralists have certainly supposed that some aspects of Greek eudaemonism are controversial. In Hobbes's view, "there is no such finis ultimus (utmost aim) nor summum bonum (greatest good) as is spoken of in the books of the old moral philosophers."13 Others have argued that even if there is such a thing, it cannot have the practical supremacy that Greek moralists attribute to it; this is why Sidgwick believes Butler marks the decisive break from ancient ethics in his discovery of the dualism of practical reason.14 What does Annas think of Hobbes's and Sidgwick's description and assessment of ancient views? First of all, is Hobbes right to believe that there is some room for controversy about whether there is a final good? Or does his belief rest on a misunderstanding of what the old moral philosophers were talking about? Annas recognizes that Hobbes has some company among ancient moralists, since the Cyrenaics deny that we have any

36.Hobbes, Leviathan, chap. 11. 37. Sidgwick, Outlines of the History of Ethics, 5th ed. (London: Macmillan, 1902), pp. 197-98. Sidgwick's claim is discussed by William Frankena, "Sidgwick and the History of Ethical Dualism," and by T. H. Irwin, "Eminent Victorians and Greek Ethics," both in Essays on Henry Sidgwick, ed. B. Schultz (Cambridge: Cambridge University Press, 1992).

reason for concern with happiness, understood as a supreme end for my life as a whole. In her view, this is because the Cyrenaics "rejected the importance of one's life as a whole for one's ethical perspective" (p. 290). Annas does not mention any other reason one might have for rejecting the notion of a final good; and so we might infer (though I am not sure whether she actually does) that if one recognizes the importance of one's life as a whole for one's ethical perspective, then one must also agree that one has sufficient reason for concern with a final good, understood as happiness. If this is right, then Hobbes must either (1) fail to recognize the importance of considering one's life as a whole or (2) misunderstood what the old philosophers mean when they speak of a final good. Is either of these a fair description of the basis of Hobbes's objection to the old philosophers? The first does not seem plausible, since it is difficult to understand a person's concern to "secure for ever the way of his future desire"15 unless we suppose he has some concern for his life as a whole. The second description is more plausible, but I am not sure it is right. Hobbes seems to suggest that the only thing one can sensibly do in planning for one's life as a whole is to "secure for ever the way of his future desire." If this is what he means, then he apparently objects to the predominant ancient (and medieval) view that if we are concerned for our lives as a whole, we have good reason to aim at a good with some relatively definite structure, attaching intrinsic value to specific states and activities. If this is his objection, then it rests on a correct understanding of the ancient view,16 and a defender of the ancient view needs some answer to Hobbes. To see where Annas stands on this issue, we may consider her claim that "the final end is not unlike the modern notion of a life-plan, the idea that all my activities make sense and are ordered within some overall plan for my entire life" (p. 38). Does Hobbes accept the relevant notion of a life plan? Suppose, first, that following Hobbes's advice to secure forever the way of our future desire counts as a minimal sort of life plan; in that case, acceptance of a final good seems to involve significantly more than having a life plan. Alternatively, suppose that following Hobbes's advice does not count as having a life plan; in that case, acceptance of concern for one's life as a whole does not, or does not obviously, commit us to acceptance of a life plan. On either supposition about life plans, the appeal to a final good is more controversial than Annas makes it appear. What might decide the dispute between Hobbes and the old philosophers? Hobbes's advice might seem reasonable if we consider a

38.Hobbes, chap. 11. 39. Epicurus is—in my view, but not in Annas's view—a partial exception to this generalization. ten-year-old child who wants to be a pilot. It would be foolish for the child to form, or for us to form on the child's behalf, a life plan that is centered on becoming a pilot; for by the time she can do anything to put the plan into effect, she may have lost her enthusiasm for being a pilot and want to be a rock star or a stockbroker. In this case the best advice would perhaps be to secure the way of her future desire, not to tie herself down to one specific plan of life. Hobbes's advice to us might be reasonable if all our ends had a status relevantly similar to that of a child's and passing fancies. Hobbes may assume, then, that the different things that are good for us are good because we happen to desire them, or we desire ends to which they are means; he may deny that anything is good for us independently of whether we happen to desire it (or an end to which it is instrumental). Some such assumption about goods is widespread (though not universal) in modern moral theories. It is sharply opposed to Aristotelian and Stoic eudaemonism; and the opposition is closely connected to Aristotelian and

Irwin

Happiness, Virtue, and Morality

7

Stoic claims about human nature (which I will discuss below). I don't know whether Annas agrees that this is an important difference between ancient and modern theories. I believe we can see its possible importance once we consider what is implied in the acceptance of a final good. IV If this is what we say about Hobbes's objection, what are we to say about Sidgwick's objection? A rough way to state the point that Sidgwick derives from Butler would be to say that conscience is a source of justifying reasons independent of the justifying reasons provided by self-love; the fact that an action is required by conscience is a sufficient reason for doing it, whether or not it is also required by self-love. 17 In saying this we do not say that self-love and conscience must conflict; we simply reject the view that they must agree because they appeal to the same end. Sidgwick believes that ancient moralists, in contrast to Butler, are monists about practical reason, because they recognize one's own happiness as the only ultimate source of justifying reasons. Since Sidgwick's belief in the independence of conscience (morality) and self-love (prudence) is widely shared, he seems to have focused on an important difference between ancient and modern ethics—if he is right to claim that ancient ethical theorists are monists about practical reason.

17. I am not saying that this is exactly what Butler means, or that Butler is in fact the first one to state Sidgwick's division. One might consider Duns Scotus; see Duns Scotus on the Will and Morality, trans. A. B. Wolter (Washington, D.C.: Catholic University Press, 1986), p. 178.

Annas emphasizes, quite correctly, that in most ancient theories the good of others matters in its own right, not simply because it promotes the agent's happiness (pp. 127, 329). She also emphasizes, equally correctly, that ancient theories do not regard one's initial conception of happiness as a conception of such a determinate and self-confined end that a conflict between prudence and morality is bound to arise (p. 323). She argues that "eudaimonistic theories do not permit this kind of split [namely, between prudential and moral reasoning] to develop" (p. 323), and so I infer that she agrees with Sidgwick's view that ancient ethical theories are monistic. If ancient theorists are monists about practical reason, is their position controversial? Sidgwick believes it is both controversial and false, even though he does not attribute an excessively determinate initial conception of eudaimonia to ancient theorists. Some of Annas's remarks, however, suggest that the type of monism we find in ancient theories is not really controversial. For she may believe that the initial conception of happiness is so indeterminate that anything that matters to me when I consider my life as a whole automatically becomes part of my revised conception of happiness. If that is all it takes to be counted as part of my happiness, my belief that morality promotes my happiness is simply a trivial result of my belief that morality matters. If this is all that monism involves, then we can see why Annas does not follow Sidgwick in suggesting that it is a major point of difference between ancient and modern theories, and why she thinks it is not an especially controversial part of ancient theories. She seems to suggest that monistic theories need not deny the possibility of conflicts between the demands of morality and other aspects of my happiness; for the claim that morality promotes my happiness will simply be another way of saying that morality matters to me, however it may be related to the other things that matter to me. If this is right, then the claim that morality promotes my happiness does not give a reason for being moral beyond the reasons that make morality matter to me in the first place; since the fact that morality contributes to my happiness is simply a trivial consequence of the fact that morality matters to me, the further fact seems to have no force as a justifying reason. I am not entirely sure, however, that this is what Annas means to say about the reason-giving force of appeals to happiness. For she does not always suggest that X's being part of my conception of my happiness is simply a trivial logical consequence of X's mattering to me for its own sake when I consider my life as a whole. She rejects the view that the final good is simply "the collection of all the goods we aim at, taken together." The predominant ancient view is that "we seek a final good by seeking all our other goods in an organized way" (p. 36). In forming a more determinate conception of happiness, "I will both change, adjust and harmonize my ends and will end up with a more coherent set of values" (p. 330). If this is right, then the claim that what matters to me contributes to my happiness is not a trivial truth; however we explain coherence and harmony, they seem to imply more than the bare claim that something matters to me or is a reasonable object of concern. The eudaemonistic outlook requires us to claim that all our reasonable concerns can be harmonized and made into a coherent set of values, and if we can show that a course of action promotes our happiness we must be able to show that it fits into this relatively harmonious and coherent set of values. If this is what we say, then the appeal to happiness has some reason-giving force that is not contained in the claim that something matters to me.

8

Ethics

October 1994

If we understand appeals to happiness in this way, then Sidgwick is right to claim that ancient monism about practical reason is a substantive and controversial claim. Once we see this, we can ask some questions. (1) Do ancient theories really include this apparently substantive claim, that all our rationally justified concerns must be fitted into some harmonious and coherent set of values? (2) If ancient theorists claim this, have they any good reason for their claim? For the reasons I have given, I believe that Annas answers yes to the first question, but I am not sure how she answers the second question. In emphasizing the intuitive aspects of the appeal to happiness, she gives less emphasis to its more controversial elements. I have tried to suggest why the more controversial elements of eudaemonism seem to mark an important difference from the trend in modern ethical theory that is expressed in Sidgwick's claims about the dualism of practical reason. V I have suggested that Annas's exposition and defense of ancient eudaemonism blunts its controversial edge. I believe something similar is true of her discussion of appeals to nature. While she argues effectively against some objections that rest on misunderstandings of ancient views, I believe her defense makes the ancient theorists' claims appear less controversial than they really are. Ancient ethical theorists come under suspicion because (in the view of their critics) they appeal to facts about nature to provide a basis for ethical argument and theory. This appeal may be rejected for two quite different reasons: (1) the ancients make claims about nature that we now have good-reason (derived, perhaps, from Newton, or from Darwin, or from Freud) to believe are false; (2) the sort of argument that the ancients attempt is misguided in principle, whether we begin from ancient or from modern beliefs about nature. The first of these objections concerns (for instance) Aristotle's "metaphysical biology" (to use Maclntyre's phrase);18 as Annas points out very effec18. A. C. Maclntyre, After Virtue (London: Duckworth, 1981), p. 139.

tively (p. 139), the objection often rests on a vague and inaccurate idea of Aristotle's actual views about nature and teleology. The second objection appeals especially to those who believe in a sharp and important division between facts and values, or between descriptive and normative judgments; these critics accuse ancient theorists of arguing in ways that ignore these important divisions. I will mainly discuss Annas's reply to the second objection, though some of the issues are also relevant to the first. What are we to count as evidence for or against these objections? In assessing Aristotle's appeals to nature, Annas considers only passages that include the word phusis or its cognates. She denies, for instance, that Aristotle's argument from the human function counts as evidence of his views about nature, on the ground that he does not use phusis to express his main claims (p. 144). This exclusive focus on the word phusis is open to at least three objections. (1) Aristotle's remarks about phusis show that he regards the nature of a living organism as (in one way of speaking of nature) its form and essence. Moreover, he identifies the soul with the form and essence of a living body. He authorizes us, then, to take remarks about the essence of a human being (or what a human being is) and about the human soul as appeals to human nature. (2) Aristotle's critics have no reason to focus exclusively on his use o f phusis. If their charge turned out to be false about his use of this word, but true about his uses of 'form', 'essence', and 'soul', then the main point of the objections would be vindicated. (3) Annas makes the important observation that Arius Didy-mus and Alexander use 'phusis' in contexts in which Aristotle does not use it (pp. 147-49). But to decide whether this is a development of doctrine (as Annas supposes), or mainly a change in terminology, we need to consider not only whether Aristotle uses phusis, but whether he uses the other terms that indicate that he is talking about human nature. In replying to the objection that appeals to nature by Aristotle (and other ancient theorists) involve an attempt to base values on facts, Annas argues that the claims about nature underlying ancient ethical theories are characteristically normative claims, so that ancient theorists do not try to derive normative from nonnormative claims. If we agree with her this far, we might say (i) that claims about nature provide a normative but nonethical basis, or (ii) that they rely on ethical judgments that do not themselves depend on an ethical theory, or (iii) that they are part of an ethical theory. Annas accepts the third answer: "The appeal to nature is part of an ethical theory; it supports the other parts, but is not itself an appeal outside the theory altogether" (p. 138). The first two answers, however, deserve consideration. When Aristotle appeals to facts about what a human being essentially is,19 he 19. See, e.g., 1097b22-1098a20, 1102a5-32, Illla33-b3, 1139al5-17, 1144a6-ll, 1168b28-1169a6, 1170bl0-14, 1178a2-8, 1178b3-7.

uses these appeals as premises of his ethical argument, but he does not seek to establish them by the specific ethical argument that he develops in the ethical treatises. They are affirmed and defended, however, in Aristotle's works on nature and on first

Irwin

Happiness, Virtue, and Morality

9

philosophy, especially in the De Anima. To this extent Aristotle's appeal to them seems to be an appeal outside his ethical theory. It would be difficult to claim that the relevant claims in Aristotle's nonethical works are not normative in any sense; it may be difficult, for instance, to say what the form and essence of an organism is without some conception of what is good for the organism. But the natural philosopher's conception of what is good for a plant or a sea anemone does not seem to be a specifically ethical judgment. If judgments about the human form and essence are also normative, but nonethical, they seem to provide some nonethical basis for ethical theory. And even if they are ethical judgments (when applied to human beings), it does not follow that they are simply parts of an ethical theory.20 Annas does not give us sufficient reason, therefore, to reject the view of modern critics that Aristotle seeks to rest ethical theory on an appeal to facts about nature that are not affirmed simply as parts of ethical theory. I conjecture that she disagrees with modern critics because she assumes that if their view of Aristotle were correct, then their objections would be unanswerable. I do not find this assumption plausible; if it is Annas's assumption, I would like to see it defended. Perhaps a little more history would help to clarify the philosophical issues here. An appeal to human nature is basic in Butler's account of self-love and morality. Butler claims that his first three sermons "were intended to explain what is meant by the nature of man, when it is said that virtue consists in following, and vice in deviating from it; and by explaining to show that the assertion is true." He takes himself to be explaining the views of ancient (especially Stoic) moralists: "That the ancient moralists had some inward feeling or other, which they chose to express in this manner, that man is born to virtue, that it consists in following nature, and that vice is more contrary to this nature than tortures or death, their works in our hands are instances."21 It seems to me that Butler understands the ancient moralists correctly, that his conception of nature is not simply a part of an ethical theory, and that neither scientific nor philosophical argument

20. I am not denying that ethical theory may also improve our theory about human nature, and so I am not claiming that the justificatory relation between ethical theory and claims about nature always goes in the same direction (hence I agree with much of what Annas says on p. 217). I take it that the main issue is whether Aristotle's ethical theory relies on some claims about human nature that are justified by appeal to something other than ethical theory. 21. Butler Sermons 1.13.

has shown his claims about nature to be false or misguided. It is quite true that Butler's claims have been widely controverted; if we compared the issues raised by Butler's claims with the issues raised by Aristotelian and Stoic claims about nature, the ancients and the moderns might throw some light on each other. Annas argues that no ethically relevant conception of nature is neutral common ground between different ancient ethical theories. She is certainly right in claiming that, for instance, Epicurus holds sharply different views about nature from those of Aristotle and the Stoics. This, however, does not show that different views about nature are simply aspects of different ethical theories; for (as Annas of course recognizes) Epicurus differs sharply from these other moralists on nonethical issues in epistemology, metaphysics, and natural philosophy that are relevant to claims about nature. When we set Epicurus aside, I wonder whether Annas somewhat exaggerates the extent to which different moralists "end up with conceptions of nature which cannot help to judge between the theories but rather stand or fall together with each theory" (p. 218). A dispute between Antiochus and the Stoics may illustrate this point. Antiochus criticizes the Stoics because their conception of the good fails to take account of human nature. The human good must be the good for the sort of creature that a human being is, and since a human being is not simply pure practical reason, the good for a human being cannot consist simply in virtue, which is the perfection of practical reason.22 The Stoics do not seem to suppose they are free to answer this criticism by devising their own ethically based conception of human nature and arguing that according to them human nature consists exclusively in pure practical reason. Their response to the objection is to argue that their theory does not in fact ignore the complexity of human nature, because it enjoins rational concern for the preferred indifferents. This is why they describe the good as doing all we can to secure the preferred indifferents (p. 397). Admittedly, the appeal to a shared conception of nature does not settle the dispute between the Stoics and their critics, but this does not mean it plays no important role in ethical argument. On the contrary, it imposes an important condition of adequacy that the Stoics are careful to respect; I will return to the importance of this condition of adequacy when I comment on Annas's discussion of virtue and happiness. VI One aspect of Annas's comparison of ancient with modern theories concerns

10

Ethics

October 1994

impartiality. She takes this to involve "the requirement that, 22. Cicero De Fin. 4.19, 25, 27-28.

from the moral point of view, the agent (a) not weight his own interests merely because they are his own and (b) not weight his own particular attachments and commitments merely because they are his own" (p. 267). This requirement itself may be interpreted in different ways, depending on how we understand the 'merely because' clause. Two interpretations would be the following. (1) The fact that something is in the agent's own interest or is a commitment of the agent's cannot have the sort of moral significance that would justify the agent's giving priority to this interest or commitment. (2) This fact cannot carry moral weight unless it can be shown to have the right sort of moral significance. In both of these cases the facts about the agent's own interests or commitments do not "by themselves" carry any moral weight, but the second interpretation of impartiality (we may call it "weak" impartiality) does not exclude the possibility that they could be shown to embody some moral property that carries the appropriate weight. The first interpretation ("strong impartiality") asserts that they cannot embody the relevant moral property. Weak impartiality imposes a nontrivial requirement, since it requires us to explain how the agent's relation to his own actions could be a morally relevant feature of his actions; still, it does not imply that the agent's relation to his actions could not be morally relevant. This division between weak and strong impartiality is still rather imprecise, but it may help us to raise one or two questions about Annas's argument. As far as I can see, she takes the requirement of impartiality to imply strong impartiality,23 and she tries to show that some ancient theories accept it. In particular, she believes that the Stoics accept it, and she appeals to a passage in Hierocles that describes an agent's relation to other people by assuming several "circles" that depend "on their different and unequal relations to one another" (quoted, p. 267). The first of these circles includes one's body; others include one's immediate and extended family, one's fellow citizens, and eventually the whole human race. Hierocles suggests that we belong to these circles whether we are virtuous or vicious, but that if we are virtuous we draw people in the more remote circles into closer circles. Annas believes that Hierocles intends this process to result in our drawing everyone into the first circle, so that eventually we have the same degree and type of concern for everyone; "the process is supposed to end up with my having the same attitude to all humans" (p. 268), which embodies strong impartiality. Annas's view depends on supposing that ( a ) we draw the members of the second circle into the first circle; ( b ) we draw the members of 23. But perhaps she refers to weak impartiality on p. 290 (end of penultimate paragraph).

the third circle into the second circle, and since the second circle has already been drawn into the first circle, "drawing in" the third circle involves drawing them into the first circle; ( c ) for the same reason "drawing in" the fourth circle involves drawing them into the first circle, and so on for the other circles. A different picture, however, would be this: ( a ) we draw the members of the second circle into the first, by treating them in the way we previously treated members of the first circle; ( b ) we draw the members of the third circle into the second by treating them in the way we previously treated members of the second circle; ( c ) and in general we "draw in" members of circle n by treating them in the way we previously treated members of circle n-1. If this is what Hierocles has in mind, his advice to "draw in" the outer circles does not commit him to collapsing all the circles into one and so does not commit him to strong impartiality. Annas claims that Hierocles tries and fails to describe a process that ends with strong impartiality. But she does not explain why we ought to interpret "drawing in the circles" as an attempt to argue for strong impartiality by reducing all the circles to one. While any attempt to make the picture of "drawing in the circles" precise will probably run into difficulties, the difficulties raised by Annas do not clearly affect the way he actually uses the picture; for nothing he says conflicts with the second picture I suggested, which does not involve reduction of all the circles to one. The second picture seems to fit better with Cicero's account of the different obligations to people in different "circles."24 No commitment to strong impartiality is evident in the other Stoic remarks that Annas cites. When, for instance, Stoics claim that virtuous people benefit each other equally, they do not claim that other relations, putting people in one or another of Hierocles' circles, are always morally irrelevant. To say that some moral relations prohibit discrimination in favor of the "inner circles" is not to say that all moral relations prohibit it. We might argue that a judge is obliged not to favor her own children when they face a criminal charge without arguing that she is obliged never to aim at their welfare rather than the welfare of strangers.25 The strongest evidence for a Stoic belief in strong impartiality is the criticism that Annas quotes from Anonymous in in Theaetetum. Anonymous seems to say that the Stoics can vindicate the requirements of justice only if they can show that we can

Irwin

Happiness, Virtue, and Morality

11

form equal concern for ourselves, our close friends, and someone completely unknown to us and unconnected with us; he objects that we do not in fact form this

40.Cicero does not actually use the metaphor of circles (De Off. 1.50-60). 41.A similar case is considered by Cicero (ibid., 3.40). sort of equal concern. Anonymous seems to assume that the Stoics can defend the requirements of justice only if they can defend strong impartiality. This objection is appropriate if Anonymous supposes that the Stoics advocate strong impartiality. But it is also appropriate if he believes that they accept the second picture of Hierocles' circles. In that case, Anonymous's objection will be that the Stoic claims about extending circles do not lead to the strong impartiality that would be needed to vindicate the requirements of justice. And so even if we assume that Anonymous understands the Stoic view correctly, we do not have sufficient reason for attributing a belief in strong impartiality to the Stoics. Once we see this, we ought also to be doubtful about Annas's case for attributing a belief in strong impartiality to other ancient theorists. Antiochus certainly claims that our concern, properly trained and developed, extends to every other human being, but the passage that Annas cites seems to say nothing about strong impartiality. 26 Similarly, though Annas attributes a belief in strong impartiality to Arius, all I can find there is an argument to show that one's concern ought to be extended to all human beings (pp. 282-83). I see no demand for equal concern or equal treatment applied to anyone. Annas recognizes that Arius also allows for differential concern and suggests that he faces a difficulty in fitting this into a theory that has already accepted the Stoic demand for equal concern, understood as strong impartiality (p. 290). I am not convinced that there is any such Stoic demand or that, if there is one, Arius accepts it; and so I do not see that this particular difficulty arises for Arius. Annas is right to draw our attention to Arius's interesting attempt to combine some of Aristotle's views on self-love and friendship with Stoic views on oikeidsis; but I am not convinced that he departs as far from Aristotle as he would if he accepted a demand for strong impartiality. Annas's arguments, then, do not show that ancient theorists accept strong impartiality. If a commitment to strong impartiality is a hallmark of modern moral theory, the absence of this commitment in ancient theories seems to mark a significant difference between ancients and moderns. Annas's discussion should prompt us to ask at least two further questions. (1) Is strong impartiality really characteristic of modern ethical theories, or are they usually committed only to weak impartiality? (2) Do ancient theorists accept weak impartiality? Treatment of these issues would require a more careful discussion of impartiality than I have offered so far, and so I will not pursue them any further here. The fact that Annas's discussion raises these questions for us confirms her claim that when we study the treatment of these

26. See Cicero De Fin. 5.65-66.

issues in ancient theories, we can improve our understanding of the issues themselves. VII Annas suggests that there is a considerable degree of consensus in ancient theories about impartiality. Another area of consensus concerns the relation of virtue to happiness; once again, she argues that a better understanding of ancient views allows us to reject some modern criticisms. She agrees that the Cyrenaics regard virtue as purely instrumental to the agent's pleasure, and not to be chosen for its own sake. On this point the Cyrenaics reject the demand that is stated by Glaucon and Adeimantus in book 2 of the Republic and accepted by Aristotle and the Stoics. Annas, however, believes that Epicurus agrees with the Platonic, Aristotelian, and Stoic view; she believes that Epicurus, like Aristotle, regards virtue as a part of happiness. If this is true, then the major ancient theories cannot be criticized for failure to advocate virtue for its own sake. This claim about Epicurus seems to conflict, as Annas sees, with Epicurean hedonism; for virtue and virtuous activity are states and activities that can be sources of pleasure but, apparently, cannot themselves be parts of it. To answer this objection, Annas explains Epicu-rus's position by comparing it with Mill's (p. 339). Mill first identifies happiness with pleasure, but he goes on to claim that different activities can themselves be parts of happiness; Annas suggests that this is also what Epicurus has in mind. Does this appeal to Mill show that Epicurus's position, as Annas understands it, is coherent? Mill's different claims about happiness have often been taken to conflict. How, we may ask, could happiness both be a feeling of pleasure and a compound of states and activities that are sources of pleasure? The suspicion that Mill equivocates on 'happiness' is strengthened by an apparent equivocation on 'pleasure'. When he claims that an activity such as music counts as a pleasure, his claim is true only if'a

12

Ethics

October 1994

pleasure' means "a source of pleasure" rather than "being pleased"; but "being pleased" is apparently what Mill means when he speaks of happiness as pleasure and the absence of pain. Annas does not say whether she thinks Mill can reasonably be understood in a way that answers the objection I have mentioned. If Mill is open to this sort of objection, then an appeal to his conception of happiness does not show that Epicurus can consistently hold the position that Annas attributes to him.27 27. A sympathetic account of Mill's position is presented by F. R. Berger, Happiness, Justice, and Freedom (Berkeley: University of California Press, 1984), pp. 30—38. While Berger shows convincingly that some elements in Mill's position cannot be captured by a simpleminded hedonist interpretation, he is less successful in showing that Mill's

In Annas's view, however, this sort of objection can be seen not to damage Epicurus's position, once we understand his conception of happiness. She appeals to the distinction between kinetic and static pleasures and points out that the kind of static pleasure that Epicurus identifies with happiness is ataraxia, which she describes as "the condition of normal functioning unimpeded by pains or worries" (p. 338). According to this view, "we achieve our final end. . . by doing whatever we are doing in a way which is not hindered or made miserable by pains and 'troubles'" (pp. 33738). Let us suppose that this is Epicurus's conception of happiness; does it show how virtue and virtuous action could reasonably be taken to be choiceworthy for their own sakes, and therefore (given the completeness of happiness) a part of happiness? I do not see how the difficulty apparently raised by hedonism is removed by this description of happiness. Epicurus seems to believe that your happiness is the untroubled state that you achieve by being virtuous and acting virtuously, not the virtue and the virtuous action themselves. Epicurus might intelligibly try to show that virtuous action results in less trouble and disturbance than results from any other course of action; but he would surely be confused if he inferred from this that he had shown how virtue is to be chosen for its own sake, and not simply as instrumental to happiness. Annas's arguments, then, do not show that Epicurus's conception of happiness is consistent with the view that virtue is a part of happiness. We are entitled to ask, therefore, for very clear evidence to show that Epicurus really regards virtue as a part of happiness. The claim that virtue is necessary and sufficient for happiness will not provide such evidence, since this claim is consistent with a strictly instrumental status for virtue. Moreover, Annas points out that Epicurus sometimes affirms that virtue is valuable only insofar as it results in pleasure. Her best evidence to show that he counts virtue as a part of happiness is his remark that "the virtues have grown together with (sumpephukasi) living pleasantly, and living pleasantly is inseparable from them."1 Our view on Epicurus's doctrine depends on the translation of sumpephukasi here. Annas translates it by 'grown to be a part of, which incorporates the claim that virtue is a part of happiness; she remarks that this choice of translation is influenced by Mill's claim about parts of happiness. Are we entitled to translate this passage so that it implies that virtue is a part of happiness? We might have a case for such a translation if we had clear evidence elsewhere that Epicurus maintains the relevant claim about pleasure and virtue, and position as a whole is consistent. Indeed (see p. 37, and p. 305, n. 11), he refrains from committing himself on the question of Mill's consistency. And so even this sympathetic account of Mill does not show how Epicurus's position could be both plausible and consistent.

that he uses sumpephu-kenai to make this claim. But we have no such evidence; indeed I have suggested that if Epicurus maintained the relevant claim, he would raise difficulties for the rest of his position. Moreover, an interpreter who attributes Mill's position to Epicurus may not make his position more defensible; for in allowing him to respect the commonsense view that virtues are to be chosen for their own sake, we introduce an apparent incoherence into his position (if what I said above is right). For these reasons I would prefer a rendering of sumpephukasi that does not import the claim that virtue is a part of happiness.29 In this case, then, Annas seems to exaggerate the degree of agreement among ancient moralists about the relation of virtue to happiness. If she were right, then Epicurus would agree with Plato, Aristotle, and the Stoics in believing that virtue is to be chosen for its own sake. In fact, however, Epicurus appears to have been unwilling to adapt his hedonism to fit this belief about virtue. Indeed, he may have supposed that this belief about virtue is simply an unreflective assumption exploited by unenlightened moralists such as Aristotle.30 In this case (and in other cases that Annas discusses) he may be less "conservative" (p. 340) in his attitude to common sense and to Aristotelian assumptions than he appears in Annas's picture. VIII The first three parts of the book lead up to Annas's long and important discussion, in

1Diogenes Laertius 10.132; quoted on pp. 238-39, 340.

Irwin

Happiness, Virtue, and Morality

13

part 4, of different views about the relative place of virtue and of external goods in happiness. Readers who are initially inclined (as I am) to suppose that Aristotle must be right to insist that virtue is insufficient for happiness, and to suppose that the Stoics are clearly misguided to identify virtue with happiness, will especially benefit from considering Annas's careful defense of the Stoic position. While she argues, as we have seen, that Epicurus's position is more similar to the Aristotelian and Stoic position in attaching noninstru-mental value to virtue, she believes that there is a sharp conflict between the Aristotelian and Stoic position on external goods.

42.A. A. Long and D. N. Sedley, The Hellenistic Philosophers (Cambridge: Cambridge University

43.

Press, 1987), vol. 1, p. 114, translate, "the virtues are naturally linked with living pleasantly." G. Arrighetti, Epicuro: Opere, 2d ed. (Turin: Einaudi, 1973), p. 114, translates, "le virtu sono infatti connaturate alia vita felice." B. Inwood and L. P. Gerson, Hellenistic Philosophy (Indianapolis: Hackett, 1988), p. 125, translate, "the virtues are natural adjuncts of the pleasant life." For some pertinent ancient comments on Epicurus's claim see P. Mitsis, Epicurus' Ethical Theory (Ithaca, N.Y.: Cornell University Press, 1988), p. 61, n. 5 (Seneca Epp. 85.18 is especially relevant). Plutarch Non Posse 1091b. It has been suggested (see Arrighetti, p. 563) that peripate(i) refers to Aristotle.

Annas uses this conflict to suggest a parallel between ancient and modern ethical debates. She argues that a fair examination of the arguments on each side of the issue should lead us to sympathize with Cicero's inability to make up his mind about which side is right. In her view, "even to someone familiar with all the arguments and the different demands which the different theories make on our intuitions, it can seem unclear where the truth lies" (p. 433). She takes this result to be parallel to the predicament of modern students of ethical theory: "It is clear that morality is important to us, but what we urgently need to know is, just how important. But on this, it appears, reflection can never be decisive" (p. 435). Annas rejects one reconciling suggestion that was already put forward by ancient critics of Stoicism. According to this suggestion, it appears difficult to decide between the Stoic and the Aristotelian positions simply because the two positions—once each is properly explained so as to accommodate the truth in the other—are substantively so close together (p. 393). I agree with Annas in rejecting the reconciling suggestion as it stands; for the differences between the two positions are not merely verbal. I believe, however, that a different reconciling suggestion—that the two views agree on the most important ethical questions—deserves more consideration. It is easy to see why Aristotle (following Plato) rejects the Socratic and Cynic view that virtue is sufficient for happiness. Since this view is eudaemonist, assuming that all rational action aims at happiness, it apparently cannot explain why we have any reason to aim at anything that does not promote happiness; but if virtue is sufficient for happiness, then we apparently have no reason to aim at external goods that are not necessary for virtue. Aristotle insists on a specific condition of adequacy: an account of happiness and virtue must explain why we have good reason to prefer, ceteris paribus, more external goods rather than fewer. He insists on this condition of adequacy not simply because it is intuitively plausible, but also because it seems to him to follow from the requirement that happiness must be complete (p. 366). Annas perhaps lays insufficient stress on the fact that the Stoics do not reject the Aristotelian condition of adequacy but try to show how it can be satisfied within a theory that identifies happiness with virtue. They agree that we have good reason to prefer, ceteris paribus, more external goods31 rather than fewer. External goods are objects of rational concern because they are preferred indifferents, and the rational person prefers virtue together with preferred indifferents over virtue alone (p. 395). The Stoics differ from Aristotle in believing that

31. In using this phrase to describe the Stoic position I am taking advantage of the permission given by Chrysippus (Plutarch Stoic. Rep. 1048a).

the achievement of this result is not happiness, but the life in accordance with nature.32 In identifying virtue with happiness, they identify happiness with only one component of the life in accordance with nature; hence, from Aristotle's point of view, they identify happiness with only one of a rational person's objects of rational concern. Once we give the Stoic doctrine of preferred indifferents its proper weight, we may wonder why they contract happiness so that it does not embrace every object of rational concern. In saying this, I do not mean that the Stoics' claims about virtue and happiness are trivial or unimportant. I simply want to emphasize the extent to which their theory takes it for granted that Aristotle is right about the importance of external goods in the rationally preferable life. It is easy to get the impression that the Stoics want to give virtue a more important place than Aristotle wants to give it in a rationally planned life, and want to give external goods a less important place; but this impression is misleading. Annas suggests that the Stoics have the better of the argument on some points, because of "unstable" elements in the Aristotelian position. One source of instability is

14

Ethics

October 1994

Aristotle's claim that while happiness requires some adequate level of external goods, the addition of further external goods can make the happy person happier. The difficulty consists in specifying the "adequate" level needed for "simple" happiness that still allows us to become happier (p. 382). This difficulty is not peculiar to Aristotle and so does not show that his position is worse off than its rivals. The Stoics must presumably face an analogous problem; for they must agree that it is possible to live the life in accordance with nature, not simply a virtuous life, with some adequate level of external goods below the maximum possible, while still allowing that some increase above the adequate level would be desirable. Annas is also concerned by Aristotle's claim that the happy person who loses external goods but remains virtuous does not become "miserable" (athlios), whereas the vicious person is miserable even if he is well supplied with external goods. 33 Annas asks, "What is the point of holding . . . that virtue in rags is not happy, if we have to hold that the wicked person in luxury, apparently flourishing and unrepentant, is really unhappy? Why is vice so much more powerful in its effects than virtue?" (p. 418). I doubt whether this is a severe difficulty. Since Aristotle believes that virtue is the dominant component of happiness, so that it contributes more than any other good contributes to happi-

44.Diogenes Laertius 7.105. 45. Annas attributes this "rather startlingly Platonic position" (p. 418) to Arius, not to Aristotle. But Aristotle commits himself to it at 1166b26-28.

Irwin

Happiness, Virtue, and Morality

15

ness, it is reasonable to mark this fact by saying that the vicious person is "miserable" or "unhappy" because he lacks the dominant and crucial component of happiness. We might say that a house is merely damaged as long as its basic structure survives, but that it is ruined once its basic structure has collapsed; this seems to be the point that Aristotle wants to make about the role of virtue and vice in a worthwhile life. Aristotle's claims do not seem to make his position "unstable"; on the contrary, they seem to make an important point about the role of virtue in happiness. For these reasons I am doubtful about Annas's assessment of the ancient debate about virtue and happiness. While she believes that both the Aristotelian theory and the Stoic theory face such grave objections that it is difficult to be satisfied with either theory, perhaps we ought to conclude instead that the success of each theory in answering objections makes it difficult to be dissatisfied with either theory. The degree of convergence that we find in these ancient theories is partly the result of mutually beneficial controversy. If that is so, then we may draw a more optimistic conclusion than Annas draws from this debate about the capacity of ethical debate to resolve some apparently radical disagreements; for while some disagreements are left in this ancient debate, it is not clear that they are radical. In this case we may say that neither side scores a knockout blow. But if we avoid evaluating philosophical disputes in metaphors drawn exclusively from war and competitive sports, and if we avoid relying on a picture of scientific progress that requires one theory to "win" over its "rivals," then we can recognize that genuine progress has been made even if not all disagreements have been resolved. Annas quite rightly warns against the misleading influence of sporting analogies (p. 402) and of hasty comparisons between ethical and scientific theories (p. 7); this is perhaps a further case in which we should be guided by her warnings.