Jump to ContentJump to Main Navigation
The Morality of Happiness$

Julia Annas

Print publication date: 1995

Print ISBN-13: 9780195096521

Published to Oxford Scholarship Online: November 2003

DOI: 10.1093/0195096525.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (oxford.universitypressscholarship.com). (c) Copyright Oxford University Press, 2021. All Rights Reserved. An individual user may print out a PDF of a single chapter of a monograph in OSO for personal use. Subscriber: null; date: 10 May 2021



(p.3) Introduction
The Morality of Happiness

Julia Annas (Contributor Webpage)

Oxford University Press

Abstract and Keywords

The book's methodology is set out: we must be critically aware of the theoretical assumptions we bring to the study of ancient ethics, or we risk importing anachronism. The limits of the ancient evidence should also be respected. We must also be aware of the structures of modern ethical theories and prepared to find that ancient theories differ. The ancient traditions and their major sources are listed: Aristotle, Stoics, Sceptics, Cyrenaics, Epicurus and hybrid theories.

Keywords:   ancient ethical theories, Aristotle, Cyrenaics, Epicurus, methodology, modern ethical theories, Sceptics, Stoics

Ancient Ethics

This book is a contribution to the history of ethics. But it is not a ‘history of ancient ethics’. It is a book about the form and structure of ancient ethical theory. Hence it is restricted in scope and selective in its interests. This is not the comprehensive account of ancient ethics which we badly need, but I hope that it will clear some preliminary ground and make the production of such a work more feasible.1

The primary aim of this book is historical: to study the ancient theories from a systematic and thematic point of view. Historical study of ancient ethics, however, is more elusive and difficult than historical study of ancient physics, or logic; for the perspectives of our own which we inevitably bring to study of ancient accounts of nature and reasoning are not as disputed in their very nature as our ethical perspectives are. Thus when we study ancient ethical theories, we cannot avoid reflecting on our own expectations as to what an ethical theory is. To go directly to the texts without any such reflection might look like a more strictly historical way of proceeding. But in fact, if we do not reflect on our own ethical perspective we shall just carry it with us unnoticed. A great deal of discussion of ancient ethical theory is flawed, in my view, 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.2 But this hinders understanding, and tends to produce an attitude that is either over‐ or under‐critical. Before we rush either to praise or to blame ancient theories for not, for example, providing detailed solutions to hard cases, we should ask ourselves whether the ancients shared the view that this is the kind of thing which an ethical theory ought to be doing. And we can hardly produce an adequate answer to this if we have not reflected on the centrality or otherwise of our own assumption that this is something which an ethical theory ought to provide.

Thus although this is a historical study, it is guided and formed by systematic and thematic concerns which arise from reflection on modern as well as on ancient ethics. I have done this not in order to make the ancient theories seem fashionably interesting. (If they are not interesting in their own right, any such attempt would only falsify them, in any case.) My purpose has been the opposite: to have the best chance of finding out the intellectual structure of ancient ethics, rather than imposing on the subject our own conceptions of what the appropriate structure is.

(p.4) Ancient and Modern

Because of these considerations, this book may also be of interest to those who have a serious interest in moral philosophy even apart from its history. In recent years, there has been a growing sense that there is something deeply inadequate about the view that when we systematize theories about our ethical views we are faced with the traditional option, a simple choice between consequentalist and deontological ways of thinking. If this is our option, then we must choose between calculating consequences to discover the right way to act, or rely on moral rules to guide us positively and negatively. But this is to take a modern journey with a mediaeval map—an artificially neat map combining detail in places with large unsatisfactory areas, unexplored and assigned only to monsters. In the last decade much attention has been paid to an option not even marked on previous maps—that of so‐called virtue ethics. Philosophers have begun to take seriously the idea that morality might be importantly concerned with the agent's view of her life, with happiness and with virtue.3 And, since these notions are the controlling ones in ancient ethics, there has been a growth of serious interest in ancient ethical theories, particularly that of Aristotle.4

There are many reasons why ancient ethics should appear as a source of fresh insights in areas neglected by most of twentieth‐century academic moral philosophy. Ancient ethical theories are concerned with the agent's life as a whole, and with his character. Concern with character and choice, with practical reasoning and the role of the emotions, is central rather than marginal. This has seemed to many to be a useful corrective to modern theories which operate with a narrow and abstract notion of what is relevant to morality, and which are frequently criticized for producing theories which are seriously at odds with our conceptions of what matters in our lives.

Ancient ethics, further, is not based on the idea that morality is essentially punitive or corrective, ‘the notion that morality is a life harassed and persecuted everywhere by ‘imperatives’ and disagreeable duties, and that without these you have not got morality.’5 Its leading notions are not those of obligation, duty and rule‐following; instead of these ‘imperative’ notions it uses ‘attractive’ notions like those of goodness and worth.6 Ancient ethical theories do not assume that morality is essentially demanding, the only interesting question being, how much does it demand; rather, the moral point of view is seen as one that the agent will naturally come to accept in the course of a normal unrepressed development.

(p.5) Stated thus generally, these advantages of ancient ethics are genuine, and it is clearly sensible to look to ancient ethics in the hope of remedying some of the gaps and failures of understanding that are felt in modern ethical theories. Nonetheless, there are reasons to be cautious, rather than enthusiastic about the idea of approaching ancient ethics from the primarily therapeutic viewpoint of remedying what are seen as our own deficiencies.7

To be any good, a study of ancient ethics (like this one) has to take its subject matter seriously, whether it aligns with our concerns or not. If we foist onto Aristotle or Epicurus ways of thinking that we find currently fashionable, we shall merely do bad history. And if what we find is not relevant to modern concerns, this is just something we must accept. It is entirely possible that ancient ethics is like ancient physics, which we study without being able to accept the theories as ours.

We might consider this warning needless, since when we read ancient ethics we do not find concepts as remote from ours as form and matter. Ancient ethics centres on the notions of happiness, of virtue and of the agent's deliberations about his life as a whole.8 These are notions which we recognize from our everyday ethical deliberations without any difficulty. We appeal all the time to the place of certain projects and values in a person's life taken as a whole, to the notion of a life which is satisfactory, and to qualities or dispositions of people which are admirable or deplorable. We probably do not use the word ‘virtue’ often, if at all,9 but of course it does not follow from this that we do not recognize virtues and vices; few use the word ‘deontology’ either, but that does not mean that we cannot recognize duties and obligations. When we read Aristotle on the virtues, or Epicurus on pleasure, it certainly seems as though we understand what is said unforcedly, that we know how to use, and extend the use of, these very concepts.

Does it follow from this that when we read Aristotle we are rediscovering our own tradition, finding our own roots? Some have thought so, and have concluded that we have been thinking all along in terms of virtue and happiness, and that it is only bad moral philosophy that has obscured this fact from us. If this is so, then the study of ancient ethics will have a very direct and exciting relevance to our own ethical thought: it will be the way we think ethically, and although there will be some aspects of ancient thought that we cannot carry over to the modern world, the central insights of the ancients will be our insights.

Here, unfortunately, the unexciting intellectual virtue of caution is required. Ancient ethics is indeed relevant to modern moral philosophy. But if we rush to establish similarities too soon, they will peter out. Only if we take due account of the sources of deep difference between ancient and modern ethics will we also do justice to the deep similarities. There are, I think, three major sources of differences between ancient and modern ethics, all of which we must bear in mind.

(p.6) First, modern ethical thinking is the product of several ethical traditions. One is that of the Judaeo‐Christian religious framework of ethics: the idea that morality is in some way guaranteed by God. Most prominent, however, are the two types of ideas which when made systematic by theory we call deontology and consequentialism. The former embodies the idea that the basic questions in ethics are those concerning what one ought to do and what one's duties are; the latter embodies the idea that the fundamental ethical questions are rather those as to how one should produce the best consequences.

Thus, even if study of ancient ethics were to redirect us towards concern with virtue and happiness in our own moral thought, this would not provide us with a complete alternative framework; for we have an ethical framework which is concerned not only with bravery and cowardice but also with duty and obligation, and with producing the best outcomes. Rather, studying ancient ethics may help us in a more indirect way. We shall see in this book that surprisingly many of the features characterizing aspects of modern morality—rule‐following, the notion of duty, appeal to what is beneficial or useful—do have a place within the ancient theories. But, because they have a place within the ethics of virtue, their role and status are very different from those of their modern analogues. Studying ancient ethics, then, will not lead us to reject, in favour of virtue, notions like rule‐following and appeal to benefit; but it may make us rethink the role they play in modern theories, and the ways we relate them to virtue and goodness. And this is a realistic goal, unlike the idea that we could jettison or replace notions which have a settled place in our moral theories.10

Second, we find it natural to make a number of demands on a moral theory which ancient theories do not make. 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.11

It might be replied to this that no theory can, on its own, settle hard cases for us; all theories, whether deontological, consequentialist, or virtue‐based, direct us towards the principles we need to apply to produce right answers, but this does not mean that they give us the answers; we apply them to give the answers.12

There is a difference here, however, between ancient and 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 (p.7) a correct answer. Thus for ancient theories it is true that there is not much to be said in general about hard cases. Modern theories 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. The source of this difference is easy to locate: it is the demand, explicit since Sidgwick, that we identify, systematize and formalize out of our moral thinking certain ‘methods’ or procedures for coming to ethical conclusions. This demand in turn rests transparently on the demand that ethics become more like the physical sciences; just as they enable us to make particular predictions as to what will happen,13 so a ‘scientific’ ethical theory should enable us to make particular decisions very directly, without the intervention of further deliberation on our part. This often goes with a general attitude that ethics, as it stands, is a mess, and needs to be sanitized by scientific methods.14

Ancient ethics accepts no such demands; we shall see that the intellectual model it finds appropriate for ethical understanding is quite different. So, to the extent that this demand is prevalent in modern moral philosophy, ancient ethics will again fail to meet all our demands on an ethical theory. It may well be that the demand itself is a mistaken one, and that the methods of ethics are utterly different from the methods of science; if so, the difference between ancient and modern will be to the credit of the former. The reader will probably have formed a view on this by the end of the book. But, whether mistaken or not, the demand and the consequent need, if one rejects it, to argue against it have become a settled part of modern moral philosophy; they do not depend simply on philosophical argument, and they cannot be wished away.

The Structure of Ethical Theory

There is a third point of contrast, the most important. It is a widespread modern assumption that an ethical theory must have a structure of which two things must be true: it must be hierarchical and it must be complete. By hierarchical I mean that some set of notions is taken as basic, and the other elements in the theory are derived from these basic notions. Consequentialism, for example, takes as basic the notion of a good state of affairs, and then defines the notions of duty, and of the admirable kind of person, in terms of this.

What is meant here by basicness? John Mackie expresses a commonly shared modern assumption here when he says,

a moral theory is X‐based if it forms a system in which some statements about Xs are taken as basic and the other statements in the theory are derived from them, perhaps with the help of non‐moral, purely factual, premises. But what would make a theory X‐based in the most important sense is that it should be such a (p.8) 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.15

It need not be, of course, that the basicness of the basic elements is obvious to most people (consequentialism is notoriously counter‐intuitive in this respect). Rather, they are basic in that when the whole theory is laid out these are the elements which explain and justify the other elements, but not vice versa. In consequentialism it is the production of good states of affairs which explains what the right thing to do is, and not the other way round. This kind of basicness is quite distinct from other kinds—for example, epistemological basicness. (Even full‐fledged consequentialists need not think that their understanding of duty and character is based on a prior and better understanding of good states of affairs.)16

The ways in which theories are hierarchical can differ. Some theories are reductionist, claiming that the derived elements can be reduced to the basic ones. The strongest form of this would be conceptual reductionism, the claim that the derived ethical concepts can be reduced to the basic ones. But weaker versions are obviously possible; the derived elements are in some way to be explained and/or justified in terms of the basic ones.

By complete I mean 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. Consequentialism, for example, accounts for judgements we make about someone's being cowardly, or hostile, in terms of good states of affairs, and in terms of character insofar as that is seen as a producer or inhibiter of the production of good states of affairs. Unfavourable judgements that we make about cowardice have no independent standing; if our dislike of a cowardly act is justified, it must be because that act in fact produced worse consequences than the alternative.17

Until recently, it has often been claimed that modern virtue ethics is like other modern theories in being hierarchical and complete.18 As a critic of modern virtue ethics puts it:

Just as its utilitarian and deontological competitors begin with primitive concepts of the good state of affairs and the intrinsically right action respectively and then derive secondary concepts out of their starting points, so virtue ethics, beginning with a root conception of the morally good person, proceeds to introduce a different set of secondary concepts which are defined in terms of their relationship to the primitive element. Though the ordering of primitive and derivatives differs in each case, the overall strategy remains the same. Viewed from this perspective, virtue ethics is not unique at all. It has adopted the (p.9) traditional mononomic strategy of normative ethics. What sets it apart from other approaches, again, is its strong agent orientation.19

If this is so, then virtue ethics will differ from other kinds of theory simply in its choice of virtue and character as basic elements in the theory. It will share with other theories the ambition to explain and justify its derived elements in terms of its basic ones, and to account for everything that an ethical theory should account for in terms only of the basic notions, and of others only insofar as they are derived from the basic notions.

But as a claim about all modern virtue ethics, this is not true. A great deal of recent work which centres on virtue has tended to query the demand for hierarchy and completeness. By now it is accepted that virtue theories may be more or less ‘radical’, and many have found attractive the idea of a ‘moderate’ virtue ethics, by which is meant precisely a virtue theory that renounces claims to hierarchy and/or completeness.20 So far, however, there has been little by way of precise and extended study of the form such a theory would take.

Ancient virtue theories, at any rate, do not aspire to be hierarchical and complete. In them, the notions of the agent's final end, of happiness and of the virtues are what can be called primary, as opposed to 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. However, they are not basic in the modern sense: other concepts are not derived from them, still less reduced to them. For example, all ancient theories understand a virtue to be, at least, a disposition to do the morally right thing; but the notion of the morally right thing to do is not defined or justified in terms of (still less reduced to) the disposition to do what will produce or sustain the virtue. We need to grasp in its own right what is the morally right thing to do. Indeed, if we do not do this, we will not have understood what makes this disposition a virtue, rather than some disposition which does not involve morality. Similarly the good of others is introduced in ways which make it formally part of the agent's own good; but we fail to grasp its place in ancient theories if we think of it as derived from or justified in terms of the agent's own good—for if that were the case, we would be misconceiving what the good of others is.

Ancient ethics has a structure—the notions of happiness and virtue are primary in it—but it is not a hierarchical structure. Lack of hierarchy leads directly to lack of completeness; the non‐primary elements are not derived from the primary ones, so a fortiori the theory cannot account for everything it aims to account for, in terms only of the primary elements and the others insofar as derived from the primary ones.

It is important not to confuse this with a modern suggestion that the ethics of virtue is autonomous and independent of other areas of morality.21 The modern (p.10) suggestion is that judgements about what people should do have no direct connection to judgements about what people should be like, what characters they should have. This certainly avoids the problems that accrue to the assumption that an ethical theory must be hierarchical and complete. If we accept that, then virtue must come in either as a derived element in a theory making duty or good states of affairs basic (which trivializes it) or as a basic element in an ethics of virtue in which duty and other elements are derived from it (which poses numerous problems).22 But the price of making virtue independent of other elements of morality is a high one. We normally say that both people and actions are brave, just, and so on. If we must separate judgements of character from judgements of acts, large areas of our moral discourse turn out to be systematically ambiguous. Further, to leave judgements about acts and judgements about agents unrelated is surely to give up too soon in terms of systematizing our moral judgements in a theory. Ancient theories do not have the strong structure which we find in many modern theories, especially those which are consciously based on a scientific conception of theory. But they do have some structure. They are not like some modern forms of virtue ethics which are consciously a‐theoretical, and do not even attempt to bring their basic concepts together in a single structure.23

Ancient ethical theories are theories : the notions of final good, virtue, nature, happiness and so on are systematically connected. A difficulty in one of these areas is likely to show up in others and to weaken the entire structure. Hence I have felt free to use the word ‘theory’ throughout to describe what I am talking about. Ancient theories, however, are unlike modern ones in not being hierarchical and complete; and one of the main results of the book will be, I hope, to make this clear, and also to make clear how important this is for our understanding of ancient ethics.


What then do we get from a study of ancient ethics? First and foremost, we get what we get from any history, if it is honestly done: a sensitivity to the options that we now take increased by understanding those that we now no longer do, and a grasp of our own intellectual situation deepened by seeing what its sources are.

We renew our emphasis first on the notions of happiness, final end and virtue in our own thinking, and put them more firmly in an ethical context. For, although we respond to talk of cowardice and generosity, of the good life and happiness, the trend of moral theorizing in the twentieth century has been such that it has been hard to take these thoughts seriously as part of moral thinking. They have remained important in everyday thinking, but most moral theories have not found room for them, and reflection on them has tended to migrate into popular psychology.24 (p.11) Studying ancient theories in which happiness and virtue are not only respectable but central concepts may encourage us to give these notions more respectful and serious attention when we reflect on our own use of them. Increased attention to ancient ethics could make modern ethics more realistic, readier to take seriously thoughts that have never left the lives of most people but which most ethical theories brusquely banish or downgrade.

Second, studying theories that are not hierarchical and complete may open our minds to the possibility of kinds of theories that are different in form from the ones we are used to. This may encourage a more critical attitude to the theories that we do have, and an awareness of their limitations. In particular this may encourage a more critical attitude to the common assumption that the model for an ethical theory must be that of a scientific one, with basic and derived concepts, and with reduction and theoretical simplicity seen as major aims. We may come to find alternatives other than accepting this picture or rejecting theory altogether as an aim in ethics.

Third, modern virtue theorists may benefit from seeing the formal basis of and limitations on ancient ethical theories, which are theories of virtue. It is the aim of this book to show that ancient theories share a formal framework (which one may miss if one concentrates on just one theory, like Aristotle's) which, despite the deep differences, may be fruitful to consider for modern writers exploring the general structure of a virtue theory.

And finally, modern virtue ethics can benefit not only from the broad patterns of ancient theories but also from their detail. One disadvantage that has been felt by modern writers on virtue is the lack of agreed context for discussions of happiness and the various virtues. When it comes to distinguishing among kinds of virtues, or their connections, or the connection of happiness with the good of others, it sometimes seems as though the debate lacks rules, that the writers are making their way through a medium with too little resistance. Useful insights are often made, but sometimes it is not clear quite what we are to do with them, what is the exact context of debate in which they make a contribution. Ancient theories are well worked‐out; often debates developed between different schools, giving a context for the development of important positions. Again, we cannot take over the debates; but studying them may give us a better idea of what is at stake in our own.

The Plan of the Book

The first part of the book examines the notions of a final end and of virtue, the core notions in ancient ethics. I begin by tracing what for ancient ethics is the entry point for ethical reflection: it is the agent's reflection on her life as a whole, and the relative importance of her various ends. This contrasts strongly with modern theories, for which hard cases and ethical conflicts which are often taken to be the spur to ethical thinking. The notion of a final end emerges from considerations about the ways in which the agent's ends and priorities fit together, and from the dissatisfaction which it is assumed that most thoughtful and reflective people feel about their lives and their priorities. Ancient ethics takes its start from what is taken to be the fact that people have, implicitly, a notion of a final end, an overall goal (p.12) which enables them to unify and clarify their immediate goals. Ethical theory is designed to enable us to reflect on this implicit overall goal and to make it determinate. For, while there is consensus that our final end is happiness (eudaimonia), this is trivial, for substantial disagreement remains as to what happiness consists in. (Because of this verbal agreement, ancient theories may conveniently be referred to as eudaimonistic.)

The other major concept in ancient ethics is that of virtue. Because the ancient notion is so different from what we generally understand by virtue, I discuss virtue from several angles: the dispositional aspect of virtue, the way in which developing a virtue affects one's emotional and affective side, the intellectual development which virtue demands, and the demands which an ethics of virtue does and does not make on the structure of ethical reasoning. After some more minor points the part closes by discussing two issues. One is the general assumptions which ancient theories make as to the relationship of virtue to our final end. The other is the extent to which virtue, in ancient theories, is the locus of what we call specifically moral value and the moral point of view. I argue, against the modern orthodoxy, that in ancient theories virtue does occupy the conceptual area which we assign to the moral, despite obvious differences. The result of this is that ancient ethics is concerned with what we would characterize as the place of morality in an agent's life. The other three parts explore three aspects of this point.

Part II is concerned with the role in ancient theories of the appeal to nature to ground or justify the theory in some way. I begin by distinguishing ancient concerns here from modern concerns with ‘naturalism’. I then look at the role of nature in various ancient theories: Epicurus, the Sceptics, the Stoics, Aristotle and later Aristotelian theories. One result of this is to show that nature figures in these theories in several ways; ‘the appeal to nature’ does not refer to one monolithic strategy. Nonetheless, analogies can be discerned, and nature figures, roughly, in two roles. One is to give the constraints which all theories must respect; some things are inevitable for human beings, and a theory which is to be livable by human beings must take account of them. Other than the Sceptics, however, all ancient theories hold that within these constraints there is a great deal which humans can do to achieve a better, rather than a worse, life. All appeal to human rationality as what enables us to do this; some are more optimistic than others as to what human reasoning can achieve. Thus nature appears also as a goal which ethical theory can help us to reach. Nature in this sense is not an ethically neutral, ‘purely factual’ notion. I conclude the Part by discussing the kind of justification for an ethical theory which such a conception of nature can provide, and by contrasting it with what modern theories characteristically demand by way of justification.

Part III examines the role in ancient theories of the interests of others. It is very often assumed that ancient theories must be, at some level, egoistic, just because their starting point is the agent's reflections on his life, and because they demand that the agent have concern for his acquisition of virtue. I begin by discussing those ancient theories that do arguably have a problem with allowing the interests of others a non‐instrumental role in the agent's overall goal, namely those theories that hold that our final end is pleasure or tranquillity. Other theories, however, assume that the interests of others will have intrinsic value for the agent. Their differences come in the scope that other‐concern has. Aristotelian kinds of theory assume that (p.13) other‐concern will extend only as far as ‘friendship’ or commitment to particular other people. Stoic kinds of theory, on the other hand, take it that other‐concern will extend beyond the range of personal acquaintance and commitment, and eventually result in the virtuous agent having concern, from the moral point of view, for any rational being. In this part I also consider justice, which has been found a problem for eudaimonistic theories. It is often assumed that in such theories justice can figure only as a virtue of character along with the other virtues, and that this leaves ancient theories with nothing to say on the subject which is most prominent in modern discussions of justice, namely the justice of institutions. I show, by considering various theories, that this is mistaken, and also that justice fits into ancient theories better than is often thought.

The final part raises again the question, how eudaimonistic theories can really be theories of happiness. The more we take into the account the point that ancient theories of virtue are theories of what we characterize as morality, the harder it is for us to take them seriously as theories of happiness. I look at two kinds of theory: those, like Epicurus' and the Sceptics', which lay more weight on the idea that the happy life must be a life which the agent enjoys and is positive about, and consequently have problems in giving virtue a non‐instrumental role, and also the theories which begin by giving virtue a large role, and consequently have to be extremely revisionary about our conception of happiness and its intuitive suggestions. The major obstacle to our achieving an adequate understanding of ancient ethical theories is the difficulty we are bound to find in giving due weight both to the point that these are theories of happiness, however redefined, and the point that they are theories of morality.

In the conclusion I bring together some of the results of the different parts, and draw some conclusions about the structure of ancient ethical theories.

It is plain, even from this skimpy summary, that the method followed in this book is to let the structure of ancient ethical theory emerge from the material. In each of the parts I bring together various different theories, by comparison, or through the debates between them, in order to let the similarities and differences in their treatments of nature, or the interests of others, emerge. In every case the result is that we can discern different content in the same form, different ways of answering the same questions.

Some readers may find this way of proceeding unsatisfactory. It would certainly be neater and more satisfying if I first laid out clearly what I take the main theses of ancient ethical theory to be, and then presented the material as illustrative of these theses. A book written in this way would first lay out the bare abstract structure of ancient ethical theory, making clear what the alternative possibilities are, and then fill in the details from the various ancient texts. I have consciously rejected this way of proceeding, for two reasons. One is that I am too aware of the difficulties of establishing the form of ancient ethics to feel confident in presenting any such bare structure. Such a project is worthwhile, but will have to build upon more cautious studies, including the present one. Where ancient ethics is concerned, we have all too many confident pronouncements about what is and what is not essential to it‐often pronouncements which are backed by knowledge of a very limited number of the ancient theories. A project like the present one, which tries to extract the structure from comparative treatment of several theories, has a better chance of not (p.14) prejudging important issues at the start, and of letting the wide variety of ancient ethical options register more effectively. The second reason is that such a finished theoretical grid laid over the material from the start would inevitably have a coercive effect on the reader; if the material is presented merely as illustrative of an existing definite plan, it will be read more one‐sidedly. My way of proceeding leaves it more open to the reader to agree or disagree in an informed manner with the conclusions I draw. Of course the material is not presented in a wholly neutral way—there can be no such thing; but I hope that for all that, the reader can see whether or not the structure I discern is really there. I have obviously not been able to follow through systematically every significant feature of every theory. Rather, I have tried to lay down the main lines of what is needed to get a grasp of the shape of ancient ethical theory.

Methodological Issues

There are (at least) three kinds of difficulty the reader may feel, which I shall try to isolate and (to some extent) meet. First, presenting the material thematically—discussing virtue in ancient ethics, rather than Aristotle on virtue—necessitates frequent cross‐referencing, some repetition and some passages where an important point has to be taken for granted and explored only later on. These expository problems come from the rich and dense nature of ancient ethical theories; explaining what a virtue is, for example, requires following up in detail several kinds of consideration, some of them complex. No simple linear explanation is possible which would go from what is clear and obvious to modern thought to what is less so; rather, several strands in the notion—the emotional side, the intellectual side, the structure of reasoning‐have to be introduced separately and brought together at the end. I do not think that this procedure in fact creates major difficulties for a modern reader, and I have kept the cross‐referencing as clear as I could. The problems a modern reader has with ancient ethics are less those of grasping abstruse or unfamiliar material than those of coping with an unfamiliar standpoint on familiar things.

Second, confusion may arise over the words ‘ethical’ and ‘moral’. One of the major theses that this book hopes to establish is that ancient ethical thought is a recognizable form of moral thought. That is, the ancient theories are theories of what modern moral theories are theories of; they are not a different, lost form of thought that we can study but have no practical interest in. I have emphasized one failing I hope to avoid: that of seeing ancient ethics as too readily available for our use. But I hope also to avoid the opposite failing: that of seeing it as so fundamentally different from modern moral thought in content and method as to form a disjoint alternative. This is a fairly influential view, and I shall examine the arguments for it in detail in the final section of the first Part. I shall thus talk of ancient ethics as a form of morality, even though at several places it is necessary to contrast some aspects of it with aspects of modern morality.25

(p.15) Finally, the project of discussing the issues of ancient ethics in a way that brings the different ancient theories together is vulnerable to two kinds of complaints, especially from philosophers whose field this is. One is that it is misleading to compare pieces of ancient theories—says Aristotle on virtue with the Stoics on virtue. We can understand the relevant piece of each theory, it is claimed, only if we make the effort to understand the whole theory: how virtue fits into Stoic ethics, and possibly how Stoic ethics fits into Stoic philosophy as a whole. This objection points to something: parts of a theory lack support if extracted from the whole theory, and are open to misinterpretation. But I do not think this danger serious enough to undermine the project. Sustained discussion and comparison of the different schools' ethical theories will, I claim, show this to be just as legitimate as the comparisons of modern moral theories with which we are more familiar.

A more robust form of this complaint is that ancient ethical theories are not in fact sufficiently independent of the rest of the philosophical theory in question for us to be able legitimately to extract and compare them. It has been claimed, for example,26 that Aristotle's main ethical claims depend in a substantial way on prior Aristotelian metaphysical principles. And it is sometimes assumed that Stoic and Epicurean ethics depend in a strong way on their physicalism. In this book I do not have the scope to meet these claims directly, by examining all the metaphysical theories in question. I hope that the argument of this book will establish that ancient ethical theories can be legitimately studied in a relatively autonomous way. For the book finds a structure common to philosophers whose metaphysical principles are mutually conflicting; what is thus shared cannot be dependent on the metaphysical principles (though of course I would not go so far as to claim that nothing in ancient ethics depends on such principles).

The second kind of complaint takes the form of a claim that there is not enough in common between, say, Aristotle, the Stoics and Epicurus on virtue for a study of virtue in ancient ethics to establish anything except what is too general to be of interest as a position in ethical theory. This kind of complaint can only be met by the book as a whole; what it tries to establish, for a number of issues, is that we can in fact find in substantial outline the forms and limits of ancient ethical theory. Such a case can only be made cumulatively, and the book should be judged as a whole. Any philosophical reader is likely to disagree with my reading of the evidence on some points, on philosophical grounds; indeed one thing I hope the book will do is to stimulate discussion on several problems in ancient ethics which become more salient when different theories are compared. But it is begging the question to assume that ancient ethical theories are so different that we can not find clear and specific positions which show us the formal and limiting framework of ancient ethics, and display the different theories as taking up different options within it.

The book does not pretend to be a guided tour to the ancient systems of ethics; I have not tried, on each issue, to list what each theory has to say on it, but have discussed only those theories that define the terms of the debate or significantly add to it. Thus the book does not form a comprehensive account of ancient ethics.27 The (p.16) selections and omissions are guided by the overall aim of producing a clear account of the intellectual structure of ancient ethics. I have been aware throughout that this is a difficult task, and risks inadequacies on both the historical and the philosophical side. I have persevered because it seems to me one of the most important tasks in studying ancient ethics.

Who Are the Moderns?

Throughout the book, when I refer to modern ethical theories, what I have in mind is one or more of the range of types of ethical theory which enter into live debate and dialectic among moral philosophers at the present day. For these are the theories which actually shape our ongoing understanding of what ethical theory is, and provide the presuppositions we need to be aware of in trying to recover historical understanding of very different kinds of theories. Hence I have had in mind for the most part theories of a consequentialist, deontological and virtue ethics type; I have tried to make it clear when I am referring to a feature of one of these types of theories, and when to assumptions which are more generally shared. I have occasionally referred to theories of a different type,28 but have tried to be careful in making clear when a theory is one which is not currently regarded as a type of viable ethical theory, one which forms part of the current dialectic.

One major objection to this proceeding is extremely obvious. Is this not a somewhat parochial way of proceeding? Perhaps it is just a matter of temporary fashion that these are the theories which are at the moment the ones which are taken seriously. Would it not be better, especially in a work which takes seriously the task of accurately delineating the options in ancient ethics, to try equally seriously to give a historical account of the major intellectual trends in modern ethics, dealing not just with the options which face us now but with the historical traditions which underlie this situation, and which shape our options for us?

I agree entirely that when one wants to contrast ancient with modern ethical theory, it would be most desirable to have a full historical understanding of the forms of modern as well as ancient ethical theories. But at the moment it does not seem that we have either, and it does not seem sensible to try to work at both ends of the contrast, ancient and modern, at the same time; any such attempt is bound to be thoroughly confusing, and probably confused. In this book I have concentrated on exploring the ancient texts, from the basis of what I admit to be a less than full historical understanding of modern theories. To achieve a proper understanding of the kind of contrast that exists between ancient and modern ethical theories we need, in addition to studies like this one, studies which clarify the modern options in ethics, on the basis of historically informed understanding of them.29 Only then will we be in the position needed to give a satisfactory answer to questions about the ways in which ancient and modern ethical theories are similar and different. In the meantime, I have stayed aware of my deficiencies in understanding modern ethical (p.17) theories, and hope that my claims about the ancient theories do not actually rely on a parochial understanding of particular contrasts with modern theories.30

Who Are the Ancients?

The book has one important limitation: I start the enquiry with Aristotle, and focus on Aristotle's own writings, the early (and sometimes middle) Stoics, the Sceptics, the Cyrenaics, Epicurus and later writers of Aristotle's school, called Peripatetics. I occasionally refer to later Stoics of the Roman period, and to other later writers, but my main concern is with ethics from Aristotle to Cicero—for Cicero, though not himself an original philosopher, is our source for some major debates.

Why start with Aristotle? Aristotle is the first thinker whose works we possess at any length who writes a work called Ethics, dealing systematically with ethical theory. It is Aristotle who first lays out for us the framework of Greek ethics (and hence Aristotle who figures largest in the first part of the book), and subsequent schools walk in his footprints; they share his view that ethics is a distinct subject matter with certain agreed starting points—the importance of the agent's final good, of an account of virtue and the place in it of the emotions, of the role played in the good life by commitments to others‐ and certain settled areas of contention: all agree that our final good is happiness, but disagree as to how that is to be informatively specified.

Thus this book is not a ‘history of Greek ethics’ of the kind which starts from texts with recognizably ethical material, such as Homer, and goes through the Sophists, Socrates, Plato and Aristotle. For while all these figures are of interest for ethics, they are not of equal interest when one's concern is with explicit ethical theory. From this point of view it is Aristotle who opens the debate, and the Hellenistic schools who further it; and hence the book concentrates on these.

Still, a few words are in order about the omission of systematic discussion of the Sophists and Plato. The Sophists can indeed reasonably be seen as the real originators of ethical and political theory. They discussed nomos, law or convention, and the ways in which it contrasts with nature or phusis. They raised and discussed questions like the source of authority of the laws and the nature of society. Why not then begin with them? The reason is simply that the remains we have of their ethical ideas are too fragmentary to come to conclusions firm enough to use in a systematic study of Greek ethical theory. Of the three major figures relevant to ethics, Protagoras,31 Antiphon32 and Democritus, the first two still give rise to continuing (p.18) controversy as to what their ideas were, and fundamentally differing interpretations are current;33 and the relation of Democritus' ethics to his atomism is still disputed.34 This situation is especially frustrating in the case of Democritus, as he came to be regarded, in the Hellenistic period, as something of an ethical pioneer, and was regarded as having adumbrated an ethical theory which Hellenistic thinkers took to be primitive, but recognizably like their own in form. Unfortunately we possess no fragments relevant to this highly interesting claim,35 and scholars disagree as to whether Democritus really did pioneer a eudaimonistic form of ethical thinking, or whether he merely put forward some suggestive claims which later ethical theorists interpreted in an anachronistic way, in line with their own developed theories. The latter option is generally seen as the more plausible.36 So, although I shall refer to these earlier thinkers for specific arguments where this is relevant, I shall not begin with them, to avoid starting on a speculative and uncertain note.

The exclusion of systematic discussion of Plato may surprise and dismay more people; and of course I am not claiming that Plato can be left out of histories of Greek ethics. But in a book which examines explicit ethical theory Plato is problematic, because of his deliberate use of the dialogue form and his consequent rejection of systematic discussion of ethical theory. There is no Platonic dialogue which is ‘about ethics’ in the way that the Nicomachean Ethics is about ethics. Each dialogue has its own theme, or group of themes, which often do not answer neatly to our subject divisions. Even the dialogues which are most obviously relevant to ethics, like the Republic or the Philebus, are also, and equally, about metaphysical and epistemological issues. And even in the dialogues which are most clearly about ethics, Plato in using the dialogue form deliberately refrains from identifying and laying out his own position. It would be extremely naive simply to identify Plato's position with that which is allotted to Socrates in a given dialogue. And even if we do, we find immensely different positions occupied by Socrates in different dialogues; the structure of Platonic ethical theory is greatly underdetermined by what we find in the dialogues.37

Thus if we study ‘Platonic ethics’ we face a problem of extraction: we have ourselves to select and extract the subject matter of ethics, and this is our imposition of a later framework on a writer who deliberately writes in a very different way. Of course this is a perfectly legitimate procedure. But it points up the large differences (p.19) between Plato and a writer like Epicurus. Whatever our problems with Epicurean texts, we are in no doubt that Epicurus wrote about a subject called ethics, and that we are looking for his systematic thoughts about that. Plato, however, chose to write dialogues, not treatises; so to write about ‘Platonic ethics’ is already to have made certain important decisions about interpreting Plato and to have taken a stand on very contentious matters. In a project designed to let the structure of ancient ethics emerge from the material examined, clearly the less reliance on contentious premises the better.

In the case of Plato there are additional problems. One is the perennial ‘Socratic problem’. Within Plato's dialogues can we distinguish a distinctive ‘Socratic’ position from a later ‘Platonic’ one? And if so, what relation does the former bear to other accounts of Socrates, such as Xenophon's? This problem is especially pressing in ethics, where we seem to see some clear changes between dialogues. In recent years the old, confident assumption that we can distinguish a ‘Socratic’ from a ‘Platonic’ ethics has crumbled somewhat; it has increasingly been seen to rely on naive ways of reading the dialogues, and the whole idea of Platonic ‘chronology’ has been increasingly questioned. Another problem is the status of Plato's arguments. Are the arguments put forward by Socrates simply Plato's own arguments? Or is Plato more interested in exploring arguments and issues than in building up a system of ethics in the first place? This would certainly explain why he chooses the form of writing that he does. The more seriously we take this possibility, the more problematic it becomes to see particular arguments as simply part of a single system of ideas. This is another area where old views are under fire and old certainties are crumbling.38

Aristotle probably develops his ethical position by way of reaction to some Platonic theses as well as by reflecting on common opinion. And some of the Hellenistic schools, while rejecting Plato, take Socrates as an exemplary ethical figure.39 But, during the period of debate with which this book is mostly concerned, Plato's ethics did not figure as a contender; Plato's own school held no substantive ethical views of its own, but limited itself to criticizing those of others.40

It is, of course, perfectly legitimate to try to extract a ‘Platonic ethics’ from the dialogues. But we should be aware that this involves both strong methodological assumptions and a very large amount of system‐building and priority‐choosing on our part. And equally, once we have extracted the framework of ancient ethical theory from later writers, it is legitimate to go back and apply it to Plato's works, if we are both cautious and clear about our methodological assumptions. We need, in fact, a work or rather several works on Platonic ethics to help us to understand the following: the distinguishing features of ‘Socratic’ ethics and how they differ from ‘Platonic’ ethics; the theory of the ‘middle’ and ‘later’ dialogues, and the relation of (p.20) this both to the ethical theories of the Old Academy and to the later position of Middle Platonists such as Plutarch. It is clearly of great interest to examine the extent to which modern readers of the dialogues agree, or disagree, with later writers in the ancient world who read Plato in the light of more developed and explicit expectations about ethical theory. However, I hope it is by now clear why a book with the aim of this one has to start with Aristotle.


All translations are my own. A ‘Cast of Characters’ at the end of the book gives brief information, relevant to the concerns of this book, about the major figures discussed.


Aristotle's ethics have come down to us in three versions: the Nicomachean Ethics (NE), which has the best text and has in the twentieth century been the main focus of study; the Eudemian Ethics (EE), which shares three books with the Nicomachean version and covers roughly the same issues, but which sometimes differs and has a much more problematic text; and the Magna Moralia (MM), a compendium or textbook of Aristotelian ethics.41 Since the nineteenth century there has been intense scholarly debate over the chronology and authenticity of these works.42 While the Nicomachean Ethics remains the main text studied, there has been an increasing tendency to use the Eudemian Ethics also. Scholarly debate over whether the Eudemian Ethics is earlier or later than the Nicomachean Ethics has settled down inconclusively, and the most widely shared assumption is that both are by Aristotle, the Eudemian Ethics probably earlier. The Magna Moralia remains more controversial, and I go along with the majority view that it is a handbook produced in Aristotle's school not long after his death. The view that it is a juvenile production by Aristotle, predating both Nicomachean Ethics and Eudemian Ethics, has also been argued.43 For the purposes of this book the point is not of enormous importance, since I am concerned with the general form of Aristotle's views and the (p.21) ways in which they can be developed; the Magna Moralia is good evidence for this even if it is not by Aristotle himself. Also important for Aristotle's ethical theory are the Politics and the Rhetoric, the latter being a useful source for commonsense views on ethical matters.

The Stoics

No complete work by any early Stoic has come down to us; but for Stoic ethics we have the benefit of three continuous accounts by later writers, which contain material much of which goes back fairly directly to early Stoic writers.

  1. (1) Cicero in de Finibus (On Final Ends), Book III, presents an account of Stoic ethics put forward by a Stoic spokesman.44

  2. (2) A collection of extracts by the fourth‐century writer John Stobaeus contains a long extract on ethics, generally attributed to Arius Didymus, court philosopher to the Emperor Augustus.45 The extract falls into three parts: (a) an introduction, in which Arius draws from two other writers of the first century B.C., Philo of Larissa and Eudorus of Alexandria (this introduction is of great interest for considering the general assumptions and formal framework of ancient ethics); (b) an account of Stoic ethics and (c) an account of Peripatetic ethics. The account of Stoic ethics is a handbook resume. There is at present no English translation available, but I am working on a translation with notes for the Clarendon Library of Later Greek Philosophy.

  3. (3) Diogenes Laertius, a writer of a ten‐volume Lives of the Philosophers, of uncertain date but probably in the second century A.D., puts into his Life of Zeno an account of Stoic philosophy as a whole. The section on ethics clearly derives at many parts from the same sources as the Arius Didymus account (2), and the two usefully complement each other. The only available English translation is that in Volume 2 of Diogenes in the Loeb Classical Library, but I shall use my own translation, which was done together with that of the Arius passage, so that the two are consistent.46

The Cyrenaics

The Cyrenaic school, traditionally founded by Socrates' follower Aristippus, seems to depend extensively on Aristippus' grandson, Aristippus the Younger, for its theory and arguments. Our main sources for this are the second book of Diogenes Laertius, (p.22) and passages in later authors, principally Sextus Empiricus.47 There were later figures in the Cyrenaic school, some of whom made considerable alterations in its theories. The fragments of all the Cyrenaics can be found in three collections,48 but English translations are available only of the major sources. Much about the Cyrenaics remains puzzling, but there is an increasing amount of accessible discussion.49


Epicurus' own major work on ethics is lost, but we possess a short exhortatory work, the Letter to Menoeceus (Ep Men). This and a short account of Epicurean ethics are in Diogenes Laertius' Lives Book X. Apart from this we have to rely on fragments of or about the ethical works preserved in later writers; on fragments of later Epicurean writers like Philodemus preserved on papyri from Herculaneum; and on passages in the long Epicurean poem On the Nature of Things written in Latin by the first century B.C. poet Lucretius. We also have long speeches for and against Epicurean ethics in books I and II of Cicero's de Finibus.50 Reconstruction of the form and much of the content of Epicurean ethics is more hazardous and uncertain than it is with other ancient theories. English translations are available of Epicurus apart from the papyrus fragments, Cicero and Lucretius. For papyrus fragments of Epicurus and other Epicurean writers we are dependent on modern editions, only some of which are in English.

Hybrid Theories

Aristotle's School, the Peripatetics

Aristotle's school developed his ethical ideas in ways that were heavily influenced by later schools, especially the Stoics. Our best source for this later version of Aristotelian ethics, which I shall refer to as Peripatetic to distinguish it from Aristotle's own works, is the account of Peripatetic ethics in Arius Didymus.51

(p.23) Antiochus

Another figure is relevant here: the first‐century philosopher Antiochus of Ascalon.52 Antiochus was a member of the sceptical Platonic Academy who broke away to set up his own school in opposition to prevailing sceptical trends. He turned back to positive or ‘dogmatic’ teaching. And he thought up the idea that what he called the ‘Old Academy’ of Plato's successors was the source of a common tradition shared by all the schools of his day except the Epicureans. In ethics he maintained that Plato and his successors, and Aristotle and his school, shared a fundamentally similar ethical framework, and that this was also the position of the Stoics, whose innovations were merely tiresome verbal ones. While nonsense as history, Antiochus' thesis enabled him to build up an ethical theory which is basically Aristotelian, but with heavy Stoic influences. In Book V of de Finibus Cicero has a supporter of Antiochus' theory present it at length; and in Book IV he uses Antiochus' arguments against the Stoic theory.53 Antiochus' ethical theory has not received much serious attention, partly because it is supposedly based on history which is blatantly false, partly because it is an avowed mixture of Aristotelian and Stoic elements, an enterprise which most scholars have seen as doomed. But Antiochus' use of Stoic elements to recast fundamentally Aristotelian ideas is actually of great interest in itself as a contribution to debates between Stoics and Aristotelians on important issues. And Antiochus' project brings him in many ways near to the Peripatetic, Stoic‐influenced version of Aristotelian ethics in Arius Didymus.54 We shall see that there are many important areas of disagreement between Aristotelian and Stoic ethics, and that these hybrid theories have interesting contributions to make to them. Until recently, these theories have tended to receive scholarly abuse or condescension as being ‘eclectic’. But familiarity with ethical debate enables one to see that ‘eclecticism’ is not necessarily an unintelligent combination of diverse elements; it may be a sophisticated attempt to avoid problems with theories which are untenable in their unqualified form.55


There are two major sceptical schools in the ancient world:

  1. (1) The Academics. In the third century B.C. Plato's Academy turned to scepticism under its head, Arcesilaus. The tradition was carried on, notably by the brilliant Carneades, until the Academy petered out in the first century B.C. 56 Arguing in the tradition of Socrates, the Academics left no writings of their own, but many of (p.24) their arguments have been preserved, along with evidence about their attitude to our final end and the good life.

  2. (2) The Pyrrhonists. This was a breakaway radical movement from the Academy in the first century B.C. They took their inspiration from the legendary figure of Pyrrho, an earlier sceptic who left no writings. Our major source for Pyrrhonism is the extensive work of Sextus Empiricus, a doctor probably of the third century A.D., who wrote a shorter and a longer account of Pyrrhonism with a collection of Pyrrhonist arguments in all areas of philosophy. For ethics the relevant sections are the account of the sceptic's life and aims (which we have only in Book I of the shorter work, Outlines of Pyrrhonism) and the sections of ethical argument in both the Outlines and the longer work, Against the Professors.57

General Sources

Some authors serve as sources for more than one school. Plutarch, a second‐century Platonist philosopher, wrote many works from which we get useful information about Academics, Stoics and Epicureans, though its value is lessened by Plutarch's hostile and polemical attitude. Alexander of Aphrodisias, a second‐third century commentator on Aristotle, is our source for much valuable information about Stoic‐Peripatetic debate, and for what Alexander takes Aristotle's own position on the issues to be. Galen, a third‐century doctor, who wrote voluminously on many philosophical topics, is our best source for some issues, for example, the Stoic theory of the emotions.

Long and Sedley (1987) is a collection of extracts from Sceptics, Epicureans and Stoics, arranged topically. Volume 1 contains English translations, and volume 2 the original texts. Inwood and Gerson (1988) is a collection of passages from Hellenistic philosophy translated into English, with notes. Gerson and Inwood translate entire passages rather than collecting extracts by topic. Both these collections are valuable; I have not referred to passages in them, since their selections of ethical material differ considerably from mine.


(1) In Michael Frede's terminology, this is a work of doxography; I am concerned to give an account of past theories, from a present viewpoint which is inevitably selective both as to which theories are still philosophically live and as to which terminology will be best understood. See Frede (forthcoming). However, the term ‘doxography’ is currently understood in such a pejorative sense that I hesitate to use it to label the book.

(2) For interesting discussion of some twentieth‐century examples of this, with regard to ancient philosophy more generally, see Cambiano (1988) and N. White's forthcoming book.

(3) See Foot (1978), Wallace (1978), Dent (1975) and (1984). There is a survey of recent work in Pence (1984). See also Louden (1984), reprinted in a collection containing other useful papers, Kruschwitz and Roberts (1987); also Baron (1985). Recent collections are French, Uehling and Wettstein (1988); Philosophia 20.1–2 (1990), special issue on virtue ethics; Flanagan and Rorty (1990). For a clear and vigorous discussion of common misunderstandings of virtue ethics, see Hursthouse (1991).

(4) See MacIntyre (1981) and (1988); Williams (1981a) and (1985).

(5) Bradley (1876), Essay VI, p. 215. Bradley's own brilliant account of ethics owes much to the leading ideas of ancient ethics; but it has been uninfluential in the mainstream of analytical moral philosophy.

(6) The distinction is drawn in these terms by Sidgwick (1907), p. 106; it governs the way he reads ancient ethics in his history of ethics (1931). It has been influential in the way constrasts have been drawn between ancient and modern ethical theories (not always to the benefit of the former). See Larmore (1990), N. White (forthcoming b).

(7) A therapeutic viewpoint can be seen even in the title of Taylor's (1988). For a good discussion of this issue see Striker (1988).

(8) Ancient ethical thinkers are all sexist in that the agent they discuss in ethics is always male. Women are sometimes discussed, but always as a special case. I have not followed them in this; I have tried roughly to alternate genders when talking of ethical agents in both modern and ancient ethics. This sometimes produces an odd effect with ancient texts, but the oddity is superficial.

(9) Williams (1985), p. 9: ‘The word ‘virtue’ has for the most part acquired comic or otherwise undesirable associations, and few now use it except philosophers’.

(10) Some modern writers have been undeniably incautious here. See Scheffler (1987), p. 416: if our ethical concepts are in many ways opposed to those of the ancient Greeks, ‘and if this opposition is not only correlated with but also partly responsible for the dramatic differences between ancient Greek society and our own, and if moreover any set of ethical ideas can flourish only in appropriate social and historical circumstances, then it seems hard to see how the ideas of the ancient Greeks could have more to offer those living in our society than do the ideas that flourish in our society itself’.

(11) Cf. Louden (1984), p. 229: ‘[P]eople have always expected ethical theory to tell them something about what they ought to do, and it seems to me that virtue ethics is structurally unable to say much of anything about this issue. . . . [O]ne consequence of this is that a virtue‐based ethics will be particularly weak in the areas of casuistry and applied ethics.’ Louden's objections are to an ethic that makes virtue primary; more will be said on this below. See also Solomon (1988).

(12) Solomon (1988) presses this objection.

(13) At least, some of them do; enthusiasts for the analogy generally ignore sciences like biology, where prediction is not so important.

(14) This attitude, very clear in Sidgwick, is still present in modern ethical works which lay great weight on technical and quantitative methods.

(15) Mackie (1978), p. 358.

(16) For perhaps the whole theory develops together in a coherentist way, none of the elements having epistemological primacy. This is a modern concession; Sidgwick feels the need to claim that the principles of consequentialism are ‘self‐evident’ in a way that principles about duty are not.

(17) See Slote (1988).

(18) Or sometimes simply assumed without argument. Cf. Frankena (1973), p. 63: ‘What would an ethics of virtue be like? It would, of course, not take deontic judgements or principles as basic in morality . . . instead, it would take as basic aretaic judgements like ‘That was a courageous deed,’ ‘His action was virtuous,’ or ‘Courage is a virtue,’ and it would insist that deontic judgements are either derivative from such aretaic ones or can be dispensed with entirely.’

(19) Louden (1984), p. 228. (But for second thoughts see Louden [1990].)

(20) See, for example, Baier (1988), Louden (1984) and (1990); also Hudson (1986), esp. chapters 1–5, 10, on non‐reductive theories of virtue.

(21) See Trianosky (1986).

(22) I do not go into these problems here; see Louden (1984), Solomon (1988). The most prominent one is that an ethics of virtue (so understood) cannot produce an account of right action that is not trivial.

(23) This is recognized in Louden (1990). This issue is further discussed in Part V.

(24) In our society we have to turn to popular self‐help manuals to find extensive discussion of questions of the best life, self‐fulfillment, the proper role of the emotions, personal friendships and commitments, topics which in the ancient world were always treated in a more in tellectual way as part of ethics.

(25) Apart from these sections of the book, I have discussed this issue in Annas (1992c).

(26) Irwin (1988b), chs. 15–21.

(27) At present there is no such account which is extensive enough to deal with the arguments of the different theories in much detail. See, however, Vegetti (1989) and Prior (1991).

(28) For example, Bradley, as a representative of the Hegelian school of ethics.

(29) Work of this kind is presently being done by Schneewind. See his (1990a), (1990b), and (1991) as well as his (1993) and forthcoming paper. The forthcoming book and papers by N. White are also valuable contributions to this project.

(30) Where they do, I rely on critics to point this out. It is worth noting, however, that (apart from mentioning the methodology of Rawls' A Theory of Justice) I do not engage with modern contractarianism. Although this has some forebears in ancient political theory, it is quite alien to ancient ethical theory, and so is unhelpful from the interpretative point of view. (I would also claim, though of course it cannot be argued here, that the contractarian model cannot capture morality as that is understood in the eudaimonistic, deontological and utilitarian traditions, and represents a radical departure from the notion of morality which this book aims to understand.)

(31) Protagoras' views come down to us mainly through their very hostile presentation in Plato; see Woodruff (forthcoming). See also Nill (1985); Barnes (1979), vol. 2, ch. 10 (b)—(d).

(32) On Antiphon see Saunders (1977–78); Furley (1981); Barnes (1979) ch. 9 (a); Nill (1985). On the papyrus fragment, which has recently been re‐read and re‐edited, with striking changes, see Decleva Caizzi (1985), (1986a) and (1986b); also Barnes (1987).

(33) For Protagoras, Antiphon and other figures like Gorgias (but unfortunately not Democritus), see Classen (1985).

(34) On Democritus see Vlastos (1945–6); Taylor (1967), 6–27; Nill (1985); Barnes (1979) ch. 9 (d).

(35) For the claims, see Cicero de Finibus V 23, 87–88; Arius Didymus ap. Stobaeus, Eclogae II 52.13–53.1. The latter passage also contains a long comparison of Democritus with Plato in the Laws on reason and pleasure (53.1–20). I discuss the Hellenistic revival of interest in Democritus as an ethical thinker (in striking contrast to Aristotle, who ignores his ethics while paying much attention to his physical theories) in my (1993d).

(36) See Kahn (1985) and Striker (1990). For the view that Democritus was the first to produce a systematic ethical theory see Gosling and Taylor (1982). See Farrar (1988) for a discussion of Democritus which stresses the political significance of his ideas.

(37) This situation is most striking in the case of the hedonism apparently defended by Socrates in the Protagoras. There is no consensus as to whether the hedonism is Socrates' or Plato's, and, if Plato's, what its relation is to positions defended by Socrates in other dialogues.

(38) Some traditional preconceptions as to how to read the ‘Socratic dialogues’ are attacked in Kahn (1981a). Kahn's position in this and other articles is usefully discussed by McPherran (1990a). Recent collections of articles which explore the methodological problems are Griswold (1988) and Klagge and Smith (1992).

(39) See Long (1988a). Among these are the Cynics, who will not figure much in this book, just because, while they are important in some respects (notably of helping to populari ze Socrates as an ethical exemplar in the Hellenistic period) they did not contribute to the development of explicit ethical theory, which is the concern of this book.

(40) See Annas (1992b).

(41) For the NE I use the Oxford Classical Text (Bywater), and for the EE and MM the Teubner (Susemihl) as well as the recent Oxford Classical Text of the EE by Walzer and Mingay. There are translations with commentaries on all three by F. Dirlmeier: Eudemische Ethik, Berlin 1962, Nikomachische Ethik, Berlin 1966, Magna Moralia, Berlin 1968. There are many commentaries on the Nicomachean Ethics in English, notably J. Burnet (London 1900) and H. Joachim (Oxford 1955), and a full and useful one in French—L'éthique à Nicomaque, R. A. Gauthier and J. Y. Jolif, Louvain 1970. The best translations in English are W. D. Ross' Oxford translation, revised by J. O. Urmson in the Revised Oxford Translation, ed. J. Barnes (Princeton 1984), and the translation with notes by T. Irwin (Hackett 1985a). The MM has not been so well served. For the EE see the partial translation with notes by M. Woods (Oxford 1982).

(42) See von Arnim (1924); the works by Dirlmeier cited in the previous note; Kenny (1978). Kenny defends the position that the EE, rather than the NE, is Aristotle's mature ethical treatise.

(43) Cooper (1973).

(44) I follow Madvig's text of the de Finibus (1876; reprinted Hildesheim 1965). Translations are available by H. Rackham in the Loeb Classical Library (Cambridge, MA 1914, repr. 1971), and by Wright (1992).

(45) I follow the text of C. Wachsmuth (1884). The attribution to Arius of the Stoic section is not certain, but it is plausible. On this and other issues see Fortenbaugh, ed. (1983).

(46) I use the Oxford Classical Text, though this is not very satisfactory. A new text by M. Marcovich is forthcoming. The only complete English translation is by R. D. Hicks in the Loeb Classical Library (Cambridge, MA 1925, repr. 1972). On Diogenes' treatment of the Stoic and other schools see Giannantoni, ed. (1986).

(47) For Sextus see below under Sceptics.

(48) Giannantoni (1958); Mannebach (1961); Giannantoni (1990).

(49) For bibliography see the discussion of the Cyrenaics in Part III.

(50) The standard text of Epicurus is Epicuro: Opere, ed. G. Arrighetti (Turin 1960). H. Usener, Epicurea (Leipzig, 1887, repr. Stuttgart 1966) collects ancient evidence and testimonia on Epicurus. Because Diogenes Laertius transcribes three letters and many sayings of Epicurus in his Life, much of the evidence can be found in Diogenes (see n. 46). Apart from the Loeb Diogenes, an English translation of Epicurus can be found in the edition by C. Bailey (Oxford 1926), which however omits papyrus fragments. We know the work of later Epicureans principally through the Herculaneum papyri; the writer most relevant to this book is Philodemus, and references to those of his works referred to in the book can be found where these are discussed. The standard edition with translation of the Roman Epicurean Lucretius is by C. Bailey (Oxford 1947, repr. 1986). A variety of English translations of Lucretius are available.

(51) On Arius see n. 45. At present no English translation is available, but one is forthcoming in the Clarendon Library of Later Ancient Philosophy, by S. White, in his translation with notes of Cicero Fin V (the theory of Antiochus).

(52) See Glucker (1978) and Barnes (1989).

(53) For the de Finibus, see n. 44.

(54) Many scholars have thought that this passage is based on, or is actually a version of, Antiochus' ethics. However, the passages are sufficiently different, even if they do have a common source, that it is better to keep them distinct rather than risk confusion by conflating them. They provide quite different solutions to some Stoic difficulties with Aristotle's ethics, for example (see chapter 20).

(55) Cf. the collection of studies in Dillon and Long (1988). I make some remarks on the interest of these hybrid and eclectic theories in my (1990a).

(56) Glucker (1978); also my (1992b).

(57) Sextus' arguments ‘against the ethicists’ are found in the last part of book III of the Outlines of Pyrrhonism (PH) and in book XI of Against the Professors (Adversus Mathematicos [M]). For Sextus I use the Teubner text (Mutschmann, revised Mau). The only complete English translation is that by Bury in the 4‐vol. Loeb edition (Cambridge, MA 1933–47, repr. 1967–71). The Sceptical ‘Modes’ or argument‐forms are translated, with commentary and an introduction to ancient scepticism, in Annas and Barnes (1985).