Jump to ContentJump to Main Navigation
Women, Culture, and DevelopmentA Study of Human Capabilities$

Martha C. Nussbaum and Jonathan Glover

Print publication date: 1995

Print ISBN-13: 9780198289647

Published to Oxford Scholarship Online: November 2003

DOI: 10.1093/0198289642.001.0001

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. date: 14 June 2021

Emotions and Women's Capabilities

Emotions and Women's Capabilities

(p.360) Emotions and Women's Capabilities
Women, Culture, and Development

Martha C. Nussbaum (Contributor Webpage)

Oxford University Press

Abstract and Keywords

Nussbaum takes on the Western and non‐Western claim that women are too emotional to be fully rational, arguing that when an adequate conception of emotion is developed, the opposition between emotion and reason will be recognized as incoherent and emotions will be seen as Adam Smith saw them—as essential ingredients in rational ethical judgement. Nussbaum contends that emotions are best understood as forms of recognition of neediness and dependency regarding the most important things in life, and are therefore as appropriate and rational as are those recognitions and the beliefs that support them.

Keywords:   dependency, emotions, gender justice, impartiality, neediness, rationality, recognition, Adam Smith

Dazed with fear, Okonkwo drew his matchet and cut him down. He was afraid of being thought weak . . .

When did you become a shivering old women?' Okonkwo asked himself. ‘You are known in all the nine villages for your valour in war. How can a man who has killed five men in battle fall to pieces because he has added a boy to their number? Okonkwo, you have become a woman indeed.’

Chinua Achebe, Things Fall Apart

So, for fear of ‘female’ emotion and its weakness, Okonkwo kills his foster son lkemefuna, whom he loves. So, for loathing of his love and the ‘womanish’ passivity into which it casts him, shaming him before others, he kills a piece of his own upbringing and his own character.1

Women are emotional, emotions female This view, familiar in Western and non‐Western traditions alike, has for thousands of years been used in various ways to exclude women from full membership in the human community and to shape—often in ways detrimental to their own flourishing—the moral education of men.2 If we are to make a convincing (p.361) case for women's equality and to develop a single account of human functioning and capability for both women and men, we evidently need to come to grips with this issue and take some stand on the role of emotion in good human functioning.

1 Emotions and Rationality

This project is urgent, since emotions are condemned as enemies of reason not only by many traditions but by the view of rationality that dominates our public life, the view of economic Utilitarianism. Charles Dickens already brilliantly made the point, when he imagined the following exchange between the Utilitarian pupil Bitzer and Mr Gradgrind, economist and grief‐stricken father:3

‘Bitzer,’ said Mr Gradgrind, broken down, and miserably submissive to him, ‘have you a heart?’

‘The circulation, sir,’ returned Bitzer, smiling at the oddity of the question, ‘couldn't be carried on without one. No man, sir, acquainted with the facts established by Harvey relating to the circulation of the blood, can doubt that I have a heart.’

‘Is it accessible,’ cried Mr Gradgrind, ‘to any compassionate influence?’

‘It is accessible to Reason, sir,’ returned the excellent young man. ‘And to nothing else.’

Bitzer understands Reason to be a calculative faculty altogether distinct from the emotions, and sufficient without them to accomplish an adequate judgement. Indeed, so well has he learned his lessons in economic rationality that he does not grasp what an emotional response is. Like Okonkwo, Bitzer won't hear of softness. Unlike Okonkwo, he doesn't even seem to fear it—perhaps because he has a ‘scientific’ argument for his position, derived from Mr Gradgrind himself. For Utilitarianism prefers reasoning that is detached, cool, and calculative, concerned with quantitative measurement. And it has a way of erecting this preference into a norm, defining Reason and the Rational in terms of it, branding everything else as (p.362) mere Irrationality. Such thinking has had and is having a profound effect on our public culture, not only in economics but also in the law.

One typical contemporary example will indicate the situation. In his 1981 book The Economics of Justice, Richard Posner (the leading thinker of the law‐and‐economics movement) begins by announcing that he will begin from the assumption ‘that people are rational maximizers of satisfactions’; the ‘principles of economics’, after all, are ‘deductions from this assumption.’4 Posner now goes to propose an extension of economic analysis to all areas of human life. He justifies this extension by appeal to a norm of rationality that is defined by contrast to emotion:

Is it plausible to suppose that people are rational only or mainly when they are transacting in markets, and not when they are engaged in other activities of life, such as marriage and litigation and crime and discrimination and concealment of personal information? Or that only the inhabitants of modern Western (or Westernized) societies are rational? If rationality is not confined to explicit market transactions but is a general and dominant characteristic of social behavior, then the conceptual apparatus constructed by generations of economists to explain market behavior can be used to explain nonmarket behavior as well . . . I happen to find implausible and counterintuitive the view that the individual's decisional processes are so rigidly compartmentalized that he will act rationally in making some trivial purchase but irrationally when deciding whether to go to law school or get married or evade income taxes or have three children rather than two or prosecute a law suit. But many readers will, I am sure, intuitively regard these choices . . . as lying within the area where decisions are emotional rather than rational.5

Posner's official purpose is to show respect for members of non‐Western as well as Western societies (and, we later see, for women as well as men) by showing that their decision‐making processes are rational. But without argument of any kind, he simply excludes emotion‐based reasoning (and much else besides) from his account of the rational, which he takes over uncritically from the narrowest of economic conceptions. This entails that any decision‐making process that later turns out not to conform to this narrow model must be regarded by Posner as ‘irrational’, therefore substandard. If women and non‐Western peoples do not, after all, conform to the economic–utilitarian model—then so much the worse for them. And the book makes this consequence evident. For though often Posner manages to construe in economic terms behaviour that may seem quite nonutilitarian,6 (p.363) by the end of the book he is forced to recognize many examples of decision‐making that do not conform to this model—above all, the US Supreme Court's privacy jurisprudence, especially in the areas of abortion and contraception. He draws the predictable conclusion: the Supreme Court's reasoning is a ‘topsy‐turvy world’, for which the ‘rational’ reader should have no respect at all.7 The same conclusion would be drawn, evidently, about the reasoning of a woman, or a member of a non‐Western society, were it to become clear to Posner that this reasoning relied on the emotions or on other non‐economic processes. Elsewhere Posner refers to a mother's grief at the death of a child as an example of a ‘nonrational’ process.8

Is Posner Okonkwo? If so, then the matchet of utilitarian rationality is ready for the softness in us all: for Posner's view, repudiating emotion‐based judgement as irrational and (so to speak)9 womanish, is entirely typical in public life. Sometimes it is claimed that women, on account of their emotional ‘nature’, are incapable of full deliberative rationality, and should not perform various social roles in which rationality is required. To take just one example: in 1872 Myra Bradwell was denied a licence to practise law in the state of Illinois. The US Supreme Court, upholding the (p.364) denial, wrote that ‘the natural and proper timidity and delicacy which belongs to the female sex evidently unfits it for many of the occupations of civil life.’10 This Okonkwo‐like line of thought11 does not require Utilitarianism to get going, clearly; indeed I regret to say that in the Western philosophical tradition it can be traced at least as far back as Aristotle himself.12

Sometimes the appeal to women's emotional nature takes an apparently more positive form—though with consequences no less pernicious for women's equality. Here it is claimed that, on account of their emotional ‘nature’, women are in fact well equipped to perform certain valuable social functions: for example, raising children, caring for the needs of a husband. These functions are indeed important; and women's fitness for them gives them social importance. But they must be exercised in the home; and, properly exercised, they will confine women to the home for much of her life. Again, this is a very old argument in both non‐Western and Western traditions of thought. It is prominent in both Hindu and Islamic traditions, and in Chinese thought as well.13 It is at least as old, in Western philosophy, as Xenophon's Oeconomicus, which faithfully reflects Athenian popular thinking of the time on this issue. In modern times it is again extremely common. It is central to Rousseau's arguments for denying women the sort of education in judgement that fits males for public life.14 It has recently played a very influential role in the social sciences in the United States—in Erik Erikson's assertions that women's ‘inner space’ defines for her a special field of activity, connected with nurturing and the home; in Talcott Parsons's very influential account of the family, in which the woman plays an ‘expressive’ (emotional, nurturing) role, the man an ‘instrumental’ (rational, practical, calculating) role, and in which the woman's role prevents her, in all but the most unusual cases, from holding a full‐time job.15 Advocates of women's equality have often uncritically accepted this contrast between emotion and civic rationality, arguing that women who wish to prove their fitness for citizenship must follow the Stoics (p.365) and get rid of (or at least very much curtail) the emotions, in order to be fully rational.16

Women in developing countries are particularly vulnerable to such ‘arguments’: for they are the intersection point of the traditional gender‐based denigration of emotion with a colonial form of ‘argument’ that holds the people of developing countries to be, in general, excessively emotional and unfit for self‐government. Stereotypes of the people of India, or of Africa, as ‘intuitive’, ‘irrational’, ‘emotional’ are too common to require illustration—and one can add that this simple portrait is all too often produced by alleged admirers of these societies as well as by detractors.17 The intersection between the ethnic and the gender‐based denigrations is sometimes made explicit, as in Havelock Ellis's striking observation that among women, ‘as among children, savages, and nervous subjects’, emotions are dominant over reason.18

Evidently if we wish to defend a universal norm of human functioning that includes women as fully equal participants, we must do what we can to answer these claims. And we must begin by scrutinizing the opposition between emotion and reason that is unreflectively assumed in them. In this chapter, then, I shall argue that this contrast, as it is frequently drawn, rests on an inadequate philosophical understanding of the emotions and their relation to belief and judgement; that a more adequate understanding shows that emotions pose no obstacle to the elaboration of a universal norm in which women are fully equal subjects.

I shall argue (1) that the strongest philosophical account of the emotions shows that they are not brutish irrational forces, but intelligent and discriminating elements of the personality, closely related to perception and judgment; (2) that other common objections that have led to the conclusion that emotions are (in a normative sense) irrational can also be effectively answered; (3) that there is no good reason to think that women, as such, are more emotional than men: such differences as do appear are best explained by appeal not to ‘nature’ but to socialization and ways of life; and (4) that, all this being the case, it is up to us to decide what forms of emotional life are fruitful and appropriate, within an overall picture of human functioning, and to transmit the capability for these emotions to all human beings, without regard to gender.

(p.366) 2 Objections to Emotion

In order to answer the charge that emotions are irrational, and thus inappropriate in good deliberation, I must first make them more precise. For a number of very different charges have been made against the emotions; and all of these can be, and have been, expressed using the convenient umbrella term ‘irrationality’. And yet there really are several very different arguments here, and in some cases built upon incompatible views of what emotions are. So any defence must begin by disentangling them. I do not claim that this list is exhaustive; but the objections I introduce here are the ones that seem to me the most germane to a debate about women's social role.

First, then, there is the common objection that the emotions are blind forces that have nothing (or nothing much) to do with reasoning. Like gusts of wind or the swelling currents of the sea, they push the agent around, surd unthinking energies. They do not themselves embody reflection or judgement, and they are not very responsive to the judgements of reason. (This picture of emotion is sometimes expressed by describing emotions as ‘animal’, and elements of an animal, not fully human, nature in us. That idea is itself often closely connected to the idea that emotions are somehow ‘female’ and reason ‘male’—presumably because the female is taken to be closer to the animal and the instinctual.) It is easy to see how this view of emotions would lead to their dismissal from the life of the citizen and the good judge. For forces of the sort described do seem to be a threat to good judgement; and their dominance in an individual would indeed seem to call into question the fitness of that individual for the functions of citizenship. It is this picture, presumably, that underlies the pejorative judgements of writers like Ellis and other anti‐emotion psychologists; and indeed the view was at one time dominant in psychology, though never, I think, in philosophy.19

A very different argument is made in the chief anti‐emotion works of the Western philosophical tradition. Variants of it can be found in Plato, Epicurus and Lucretius, the Greek and Roman Stoics, and Spinoza.20 It appears to be common in the Chinese and the Indian philosophical traditions as well.21 These philosophers all hold a view of emotions incompatible with the view that underlies the first objection. They hold, namely, that emotions are very closely related to (or in some cases identical with) (p.367) judgements. So lack of judgement is not at all their problem. The problem, however, is that the judgements are false. They are false because they ascribe a very high value to external persons and events that are not fully controlled by the person's virtue or rational will. They are acknowledgements, then, of the person's own incompleteness and vulnerability. Fear involves the thought that there are important bad things that could happen in the future and that one is not fully capable of preventing them. Grief involves the thought that someone or something extremely important has been taken from one; anger the thought that another has seriously damaged something to which one attaches great worth; pity the thought that another is suffering in a non‐trivial way, through no fault of his or her own; hope involves the thought that one's own future good is in important respects not under one's own control.

In all of these cases, the emotions picture human life as something needy and incomplete, something that has hostages to fortune. Ties to children, parents, loved ones, fellow citizens, country—these are the material on which emotions work; and these ties, given the power of chance to disrupt them, make human life a vulnerable business, in which complete rational control is neither possible nor, given the value of these attachments to the attached, even (from their point of view) desirable. Thus emotions are sources of softness, holes, so to speak, in the walls of the self.22 It is this view, it seems, that underlies Okonkwo's reaction against emotion, his view that a real man would have a hard and not a gentle self.

But according to these philosophers—and according to ordinary people who inspire and/or follow them—that picture of the world is in fact false. As Socrates already said, ‘A good person cannot be harmed.’ One's own virtue and thought and will are the only things of real worth; and one's virtue and thought cannot be damaged by fortune. Another way of expressing this is to say that the good person is completely self‐sufficient.

This argument is sometimes connected with a relative of the first argument, through the idea of stability. A good judge, these philosophers insist, is someone stable, someone who cannot be swayed by the currents of fortune or fashion. But people in the grip of the emotions, because they place important elements of their good outside themselves, will change with the gusts of fortune, and are just as little to be relied upon as the world itself is. Now hopeful, now in tears, now serene, now plunged into violent grief, they lack the stability and solidity of the wise person, who takes a constant and calm delight in the unswerving course of his own virtue. Thus this second picture can lead to some of the same characterizations as the first, and even to some of the same associations of emotion with the female—since women, they observe, often have commitments to children and loved ones that are (for contingent social reasons, on this view) more (p.368) intense than those of males. But it is important to notice how different, in the two cases, the reasons for the conclusions are. On the first view, emotions are innate; on the second, they are taught with the teaching of evaluative beliefs. On the first, they can be neither educated nor entirely removed; on the second, both are possible. On the first, emotions are unstable because of their unthinking internal structure; on the second, because they are thoughts that attach importance to unstable external things.

As will become clear in what follows, I very much prefer the second objector's view to the first, in the sense that I think it is based on a far more profound and better argued view of the relationship between emotion and belief or judgement. On the other hand, it should already be plain that one might accept this analysis of the emotions and yet refuse to accept the Stoic conclusion that the emotions are (in the normative sense) irrational and wholly to be avoided when we seek to deliberate rationally. For, as one can see, that conclusion is based on a substantive and highly controversial ethical view, according to which the good person should be totally self‐sufficient, and ties to loved ones, country, and other undependable items outside the self are without true worth. Okonkwo, like many males in many cultures, probably holds some such view. But one might dispute it. And then one would wish to retain the evaluative judgements contained in (or, as the Stoics think, identical with) the emotions, and to draw on them in practical reasoning, as acknowledgements of what is in truth valuable and good in a finite human life.

A third objection, while compatible with some respect for the emotions in parts of private life, assails their role in public deliberation. (It is compatible with the second objector's analysis of emotions as closely linked to judgements about the worth of external objects, and probably not compatible with the first objector's claim that they are altogether without thought.) Emotions, this objector charges, focus on the person's actual ties or attachments, especially to concrete objects or people close to the self. They consider the object not abstractly, as one among many, but as special—and special, in part at least, on account of its prominence in the agent's own life. They always stay close to home and contain, so to speak, a first‐person reference. Thus, love ascribes great worth to a person who is in an intimate relationship with the agent; and its intensity usually depends on the existence of an actual connection of some sort between agent and object. Grief, again, is usually for a loss that is felt as cutting at the roots of one's own life. Fear is usually either completely self‐centred or felt on behalf of friends, family, loved ones. Anger is aroused by slights or damages to something that is one's own. In all these cases, emotions seem to bind the moral imagination to items that lie close to the self and stand in a relationship to the self. They do not look at human worth or even human suffering in an even‐handed way. They do not get worked up about distant lives, unseen (p.369) sufferings. This, from the point of view of many moral theories—including Utilitarianism, Kantianism, and their relatives in non‐Western traditions23—would be a good reason to reject them from a public norm of rationality, even though they might still have some value in the home. I believe that this objection, although rarely clearly distinguished from the first two, underlies quite a few arguments that connect women's emotionality with restriction to a domestic role, suggesting that exactly the abilities that make them good in that role undermine their status as citizens.

A closely related further objection is that emotions seem to be too much concerned with particulars in general, and not sufficiently with larger social units, such as classes. And this is a point that has seemed to many Marxists, and to other political thinkers as well, to make emotion‐based reasoning altogether unsuitable in political reflection, and even subversive of it. (Or, in some versions, it is in a sense political, but so committed to bourgeois individualism that it is unsuited for critical political reflection.) Love, it is claimed, directs the attention to an individual, and deflects it from impartial reasoning about collective action; grief, too, is a form of personal self‐indulgence. In Doris Lessing's The Golden Notebook24 the heroine, a Marxist novelist, faces this objection from her political allies in the Communist Party: they claim that her emotional way of perceiving events (which they connect both with her femaleness and with her literary profession) betrays a residual attachment to the bourgeois world, and is inconsistent with her politics. On some versions of this objection, emotions may be useful enough in a private ethical domain, so long as they do not overstep their bounds. On the orthodox Marxist version, which does not recognize an ethical domain apart from the political, they are altogether without worth.

A fifth objection, finally, focuses not on the emotions as a class, but on erotic and romantic love in particular. For even some moral views that ascribe a positive role to other emotions, such as fear, pity, hope, sympathetic brotherly love, and even anger, holding that they would figure in the deliberations of an ideally rational and sympathetic judge, view romantic love with a special scepticism. Whereas other emotions need to be balanced and moderated, but will still be retained and even cultivated within social morality, this one, and the erotic desire to which it is closely linked, are seen as subversive of the public domain. A powerful version of this objection is, I believe, in Adam Smith's The Theory of Moral Sentiments, a work that on the whole defends emotions as essential ingredients in deliberative rationality.

Smith's objection25 is that love seems to have about it something hidden (p.370) and mysterious, something that both repels social scrutiny and impedes the lover from participation in the social world. Love is based to a great extent on a strong and not fully governable response to morally irrelevant particularities; to that extent, it cannot be explained or justified to another, cannot become a part of the giving and receiving of reasons that is the essence of public rationality. Adam Smith observes that a person who has been wronged can explain to a friend the nature of the wrong and can reasonably expect the friend both to comprehend and to share his anger. But a person who is in love cannot completely explain why it has happened in the way it has, for this person at this time. And the sentiment is one that he both cannot and would not ask a friend to share. At the same time, love asks lovers to avert their eyes from the wider world of social concern. It demands an exclusivity of vision that does not balance the claims of the loved one against other claims, and often prevents the due consideration of those other claims. In short: the world of love is a world of mystery and privacy, in which lovers neither see nor are seen with the vision characteristic of general social and political concern. It thus proves not only irrelevant in social reasoning, but powerfully subversive of it. (Once again, this objection is independent of the non‐cognitive view of emotion put forward in the first objection: for Smith's view of the passions is in fact a strongly cognitive view.)

One might well wonder why this objection should be linked, as it often has been, with a denunciation of women's alleged irrationality: for surely not only women fall in love, not only women feel erotic desire. And yet, the objection is in fact in its history strongly gendered, in both non‐Western and Western traditions, males being portrayed as reasonable beings who ought to be above all that, and usually are; females as dangerous seductresses, natural inhabitants of that world of passion and passivity, who lure unfortunate males to their destruction—or that of their rationality.26

And finally, even if we should manage to give a successful answer to all these objections, and thus to defend the claim that emotions are a valuable part of the ‘capabilities’ of the good person and the good citizen, we will (p.371) face one further very shrewd objection, having to do with the need for emotion in public life. Suppose we grant, then, that the most perfect social thought, both private and public, uses emotions. Nonetheless, given that the production of emotionally balanced people is likely to be a taxing job, requiring a radical reordering of human development for men as well as women from infancy on up, what priority should politics give to this task? It would seem a recipe for political disaster to tell it that it must accomplish this goal before it reshapes social institutions. Politics cannot wait around until people perfect themselves inside: it must design institutions so as to bring about intelligent results, even in the absence of good judgement and response on the part of individuals. This is what economic analysis enables us to do, and this is why it is so important. So to place the accent on emotion in a project in development economics is to engage in idle Utopian thinking instead of addressing the urgent issues.27

Each of these objections is a profound one. To answer them all definitively would require me to elaborate and defend a full theory of the emotions. This obviously cannot be done here—all the more since that is a project I have undertaken elsewhere.28 What I propose to do here is to sketch out plausible answers to the six objections, drawing, often, on ancient Greek and Roman philosophical traditions in which a cognitive view of emotions finds, I believe, an especially valuable defence. I am now convinced that these theories of emotion are closely related to both classical Indian and classical Chinese theories, and indeed to very many non‐theoretical intuitive accounts of emotion from very many parts of the world. This is no surprise, if the view is true.29 I hope, therefore, that I shall be forgiven for focusing on the philosophical material I know best, and simply pointing to some of the non‐Western parallels.

3 Answering the Objections

3.1. Emotions as Blind Animal Forces

The first objector insists that emotions are irrational in the normative sense, that is, bad guides to choice, because they do not partake in reason in even the broadest descriptive sense—are just blind impulses that neither contain a perception of their object nor rest on beliefs. This position is in one sense, I feel, hardly worth spending time on, since it has never been strongly (p.372) supported by major philosophers who have done a great deal of their most serious work on the emotions—including many who for other reasons dislike the emotions intensely. And by now it has been widely discredited even where it once was popular: in cognitive psychology, for example, and in anthropology.30 But it still has a hold on much informal thinking and talking about emotions, where the correct observation that emotions are not included in Mr Gradgrind's sense of ‘reason’ is all too often taken to licence the conclusion that they do not partake of any sort of reasoning or cognition.31 Therefore it seems important to say something about what has led to the widespread conclusion that it is not a tenable view.

Western philosophers as diverse as Plato, Aristotle, the Greek and Roman Stoics and Epicureans, Spinoza, and Adam Smith32 have agreed that it is very important to distinguish emotions such as grief, love, fear, pity, anger, and hope from bodily impulses and drives such as hunger and thirst, and this in two ways. First, emotions contain within themselves a directedness toward an object; and within the emotion the object is viewed under an intentional description. That is to say, it figures in the emotion as it appears to, is perceived by, the person who experiences the emotion. My anger is not simply an impulse, a boiling of the blood: it is directed at someone, namely, a person who is seen as having wronged me. The way I see the person is itself intrinsic to the nature of my emotion, and to its role in action. (Why would a boiling in my blood move me this way rather than that?)33 Gratitude contains an opposed view of another person's relation to my good; distinguishing anger from gratitude requires giving an account of these opposed perceptions. Love is not, in the relevant sense, blind: it perceives its object as endowed with a special wonder and importance. Once again, this way of perceiving the object is essential to the character of the emotion. Hatred differs from love in nothing so much as the opposed character of its perceptions. Emotions do oppose one another, but not (p.373) blindly: they differ with one another in a way to which perception is central—fear struggling with hope, anger with gratitude, love against panic.34 Emotions, in short, whatever else they are, are at least in part, ways of perceiving. And if one does not bring in this intentional content, it is impossible to understand how they can play the role in action that they clearly do play.

Furthermore, my second point, emotions are also intimately connected with certain beliefs about their object. The philosophical tradition I have mentioned is not unanimous about the precise relation between emotion and belief: some hold that the relevant beliefs are necessary conditions for the emotion, some that they are both necessary and sufficient, some that they are constituent parts of what the emotion is, some that the emotion just is a certain sort of belief or judgement. Let us begin, therefore, with the weakest view, on which most agree: the view that emotions are so responsive to beliefs of certain sorts that they cannot come into being without them. What leads these philosophers to accept that view? Consider, again, a case of anger.35 Anger, they argue, seems to require the belief that I, or something or someone important to me, have or has been harmed by another person's intentional action. If any significant aspect of that complex belief should cease to seem true to me—if I change my view about who has done the harm, or about whether it was intentional, or about whether what happened was in fact a harm, my anger can be expected to abate or to change its course accordingly. If, for example, I discover that B, and not A (as I had believed) is the murderer of my child, I will transfer my anger from A to B. If I believe that C has been insulting me behind my back, and then discover that no such insults in fact took place, my anger can be expected to fade. Even if some residual feeling of excitation and irritation should remain, it would not seem right to call it anger. If I am angry at D because she has forgotten my name (an example actually given by Aristotle in the Rhetoric), and should then become (deeply and fully) convinced that forgetting a name is not an offence at all, then, once again, I can expect my anger to abate. Anger, then, requires certain beliefs for its existence; and it can be modified, or even removed, by the modification of belief.

Much the same is argued for other major emotions, such as grief, fear, pity or compassion, love, and gratitude. Fear requires the belief that important damages may happen to me in the future, and that I am powerless to prevent them. I do not fear what I believe I control: thus, I do not fear that I will suddenly start telling lies or smashing up the room, if I believe that I am not the sort of person who is out of control with respect (p.374) to these actions. Nor do I fear what I believe to be trivial: I do not go around fearing the loss of a coffee cup.36 Again, if I become convinced that a damage is impossible, I will cease to fear it. If some residual upheaval remains, it will seem right to call it excitation or agitation, but not fear. Pity requires the belief that another person is suffering big and important things, through no fault of his or her own; it is in that way the other‐directed analogue to fear.37 And so forth. The beliefs in question may be very deeply rooted in the person's psychology; so getting rid of them cannot be expected to be, always, the job merely of a one‐shot argument. Some of them may even be held beneath the level of consciousness.38 But beliefs they are, none the less, as we can tell when we manage to bring them to the surface for inspection. And without these beliefs, no emotion can take root.

This, as I have said, is the weakest position. Most of the ancient Greek and Roman thinkers go further. The next plausible step to take is to hold that the beliefs in question are also constituent parts of the emotion, part of what identifies it, and sets it apart from other emotions. For it seems highly implausible that we can individuate and identify complex emotions such as anger, fear, and pity, simply by reference to the way they feel. To tell whether a certain pain is fear or grief, we have to inspect the beliefs that are bound up with the pain. To tell whether a certain happy, pleased feeling should be called love or gratitude—once again, we must inspect not just the tonality of the feeling, but also the beliefs that go with it. Thus most ancient Greek and Roman definitions of emotion (and many modern Western definitions as well—Spinoza is, of course, the most conspicuous here) mention beliefs as well as feelings, and mention them not just as causes, but as constituents of the complex passion.39

It appears to many thinkers, furthermore, that the beliefs we have mentioned are—usually, at any rate—sufficient for the emotion. That is, if I really succeed in making you believe that B has been insulting you behind your back, that will suffice to make you angry with B; I do not need, as well, to light a fire under your heart. Whatever fire there is, is fire about the insult, and is produced by awareness of the insult. If I want to make the (p.375) people of Athens furious with the Spartans, once again, all I need to do is to convince them that they have been insulted, offended, mocked, or damaged in some significant way by deliberately undertaken Spartan actions. Much of the ancient science of rhetoric rests on this observation; and modern political speech is no stranger to it either. When George Bush wanted to make the American public fear the prospect of a Dukakis presidency, he did not need to doctor their insides or inject ice water into their veins. All he needed to do was to make them believe that a Dukakis presidency would mean big and significant dangers for them that they would be powerless to control: Willie Hortons running free in the streets of every city, ready to prey on their innocent families. Okonkwo's belief that holding back from killing Ikemefuna would be evidence of womanish softness was sufficient to fill him with fear; his feeling that his own standing in the community is injured by his father's gentleness is sufficient to inspire him with rage; and so forth. This position is compatible with the view that emotions have further, non‐cognitive components (such as feelings or bodily alternations): but it insists that the relevant beliefs are sufficient causes of these further components.

The greatest Stoic thinker, Chrysippus, went one step further, holding that emotions were simply identical with beliefs or judgements. His position is not a simple one to grasp, since his notion of judgement is far more dynamic than most modern notions. His idea is that to judge that p is the case, is a two‐step process: one first entertains the ‘appearance’ of p, puts it before one's eyes, so to speak. Then one can do one of three things: one can reject it, saying that's not the way things are really. One can suspend judgement about it, neither accepting nor rejecting it. Or, finally, one can accept or assent to it, embrace it and take it into oneself as the way things are. It is his view that this activity of assenting to a certain sort of claim about the world is what having an emotional experience is. For example, grieving just is embracing the fact that (say) the most important person in my life is dead and will never return. Such activities of assenting or accepting are not calm or inert: reason itself moves (reluctantly or with alacrity, cautiously or imprudently) to embrace the relevant appearance. Nor are they without their affective side: accepting the fact that my loved one has died is itself a wrenching of my entire existence—like putting the world's knife into my own insides. Only an analysis of this sort, Chrysippus believes, can do justice to grief's discriminating power and its object‐directed intentionality. Emotions are not irrational pushes and pulls, they are ways of viewing the world. They reside in the core of one's being, the part of it with which one makes sense of the world.

This is, I believe, a powerful analysis; and I have defended it at greater length elsewhere. (One might argue that this is the position that best makes sense of Okonkwo's fear: for it is portrayed as a way of seeing seated ‘deep within himself’, at the heart of his interpretations of the world.) One does (p.376) not need to accept this view, in order to question the simple opposition between emotion and reason; any of the weaker versions of the cognitive position will do for that purpose. But it is probably the position in this family with the strongest philosophical arguments in its favour; and it certainly has had the most distinguished philosophical history, in the Western tradition, being endorsed by Chrysippus, Seneca, and, much later, Spinoza.40

Notice that this family of cognitive views still makes ample room for saying of some (or even of all) emotions that they are, in the normative sense, irrational. For emotions must now be assessed by assessing the relevant beliefs or judgements. These may be either true or false; and they may be either rational or irrational. (These are two independent dimensions of assessment: a belief might be false but rational, if I formed it on the basis of good evidence, and yet it happened to be wrong; it may also—and more often—be true but irrational, if I formed it hastily and uncritically, but it just happens to be the case.) But in no case will emotions be irrational in the sense of being totally cut off from cognition and judgement; and thus there is no more reason to think them unsuited for deliberation—just because they can go wrong—than there is reason to dismiss all beliefs from deliberation, just because they can go wrong.

3.2. Emotions as Acknowledgements of Neediness

We must now turn to the second objector. For in answering the first objector I have endorsed that objector's account of what emotions are, finding in them both object‐directed intentionality and a close relation to beliefs. According to this objector, the emotions are linked to beliefs with a particular subject matter: beliefs that ascribe high worth or importance to things and persons outside the self. To make these judgements of worth is to acknowledge one's own neediness and lack of self‐sufficiency. We can now locate the cognitive dimension of the emotions more precisely: they enable the agent to perceive a certain sort of worth or value. And (for those to whom such things do have worth) emotions are thus necessary for a full ethical vision.

But are these acknowledgements of neediness and incompleteness good? The objector states that beliefs that one has these deep needs from the world are always false: the only resources one really needs come from within oneself and one's own virtue. These false beliefs are, moreover, (p.377) socially damaging, sapping confidence and robbing action of its force and its stability. One can get rid of them: and a life without them is a more satisfactory and stable ethical life than the one in which they occur.

This means, for the Greek and Roman Stoics, radically rewriting the vision of the world and what has importance in it that their young pupils would be likely to have absorbed from their culture. For as they saw (here following Plato), most influential literature treats the events that befall finite and vulnerable people as deeply significant, and involves the audience in their good or ill fortune. It shows a hero like Achilles grieving for the death of Patroclus, rolling in the dirt and crying out, rather than recognizing that such things have no true importance.41 What we should have, instead, are paradigms of exemplary self‐sufficiency and detachment: for example, the story told by Cicero of a father who, informed of his son's death, calmly remarked, ‘I was already aware that I had begotten a mortal.’42 Socrates, in this tradition, is the exemplum par excellence, the great antidote to tragedy. Thus Epictetus writes that a person's goal should be:

To study how to remove from his own life mournings and lamentations, and such expressions as ‘Alas’ and ‘Wretch that I am’, and misfortune and ill fortune, and to learn the meaning of death, exile, prison, hemlock—so he can say in prison, ‘O dear Crito, if this is what pleases the gods, so be it’—and not that other exclamation, ‘Wretch that I am, an old man, it is for this that I have kept my gray hairs.’ Who says this? Do you think I am talking about some insignificant lowly person? Doesn't Priam say these very things? Doesn't Oedipus? In fact, don't all kings talk like this? For what else are tragedies but the sufferings of people who have been wonderstruck by external things, displayed in the usual metre? (Disc. 1.4.23–30)

This is a profound vision of the ethical life: profound, first of all, because it is based on a deep conception of what emotions are, one that I believe to be more or less correct, and one that has deep roots in the experiences of people in many different cultures. Profound, as well, because it raises deep questions about what a good human life should be, what sorts of vulnerabilities are compatible with the constancy that the ethical and political life require. These questions can hardly be resolved by a hasty dismissal of the tradition, or by charging that it is simply, here, the institutionalization of unreflective male prejudice against emotion. (Indeed, reading the Roman Stoics makes it perfectly clear that these people were concerned about the dignity of each and every human being: about the ability of females to achieve full human freedom, about the urgency of getting males to refrain from the obsessive pursuit of worldly wealth and honour that made the (p.378) politics of Nero's time irrational by anyone's normative standards.) And the view is profound, finally, because, like all the most searching and incisive philosophical thought, it shows its own argumentative structure to its reader, and thus shows, as well, how and where one might take issue with it. In particular, it shows both friends and opponents of the emotions that the radical anti‐emotion conclusion rests on normative premises about self‐sufficiency and detachment that are highly controversial. Let us at least begin to question those premises.

In my earlier chapter in this volume, I presented a sketch of the central elements in human life, claiming—plausibly, I think—that it was derived from intuitions that are both widely and deeply held across many differences of place and time. The list included both capabilities and needs, and showed how, in complicated ways, the two shape one another. What the Stoics want to do, we can now see, is to make human life anew, focusing entirely on capability and marginalizing need. In effect, they want to give the human being what, in my analysis, looks like the life of a god, rather than the life of a human being. They claim that making this change does not remove social cohesiveness and ethical virtue. But one might have some doubts about this.

Consider the emotion of compassion (pity).43 Compassion requires, as Aristotle long ago argued, the belief that another person is suffering through no fault of his or her own. The suffering must be thought to be of serious importance. And, finally, the person who has compassion must believe that his or her own possibilities are similar to those of the person who suffers. This acknowledgement of one's own vulnerability is traditionally thought, and not implausibly, to be linked with beneficence and a generous vision of the enemy; the refusal of pity, as in Dickens's Bitzer, as in Okonkwo preparing to wield his matchet, is connected with a hard and arrogant disposition of mind.

The foundation of compassion, and its social role, is the belief that many common forms of bad luck—losses of children and other loved ones, the hardships of war, the loss of political rights, bodily illness and deficiency, the prospect of one's own death—are in fact of serious importance. In order to remove compassion from human life (and the other closely related emotions with it) the Stoics have to remove that fundamental belief. But then we must ask what reasons they give their good person to care profoundly about the bad things that happen to others; what reasons they give her to get involved, to take risks, for the sake of social justice and beneficence.

It has always been difficult for philosophies based on the idea of the self‐sufficiency of virtue to explain why beneficence matters. No major thinker (p.379) of this sort is willing to say that it does not matter: and yet for Socrates, for the Stoics, for Spinoza, for Kant, it seems difficult to motivate consistently, given the alleged moral irrelevance of external goods, the self‐sufficiency of the virtuous will. Repudiating pity, as they all do, leaves very few motives for the acts usually prompted by pity; and if they are performed out of very different motives, say, on account of pious obedience to Zeus's will, it is not clear that their moral character is the same. This problem is especially deep for the Roman Stoics. For in order to convince any member of their society that theirs is even a plausible view, they must show that it makes room for the defence of family and country, for risk‐taking acts of other‐regarding virtue, and for beneficence in the distribution of material goods. Seneca repeatedly endeavours to satisfy these demands—with significantly mixed results.

The moral vision of the emotions, by contrast, beginning as it does from the idea of the vulnerability of human life and the importance of certain ‘external goods’, has no such problem. Here the natural response to the sight of the need of another similar to oneself is to come to the aid of that need; for one believes that the need is very important, and that the social world is inhabited by weak creatures who can survive and flourish only if they come to one another's aid. One believes that one is oneself such a creature. Rousseau shrewdly argues that the absence of this belief is what leads to much social callousness and cruelty:

Why are kings without pity for their subjects? It is because they count on never being human beings. Why are the rich so harsh to the poor? It is because they do not have fear of becoming poor. Why does a noble have such contempt for a peasant? It is because he never will be a peasant . . . It is the weakness of the human being that makes it sociable, it is our common sufferings that carry our hearts to humanity; we would owe it nothing if we were not humans. Every attachment is a sign of insufficiency . . . Thus from our weakness itself, our fragile happiness is born. (Emile, Bk IV, 263, 259)

Utilitarianism, we know, begins from the fact of common suffering, and is, at its best, motivated by a wish to relieve pain. This it shares, for example, with Buddhist and Hindu traditions, in which compassion for suffering is appropriately made a central source of social motivation. So it is a very serious internal criticism of utilitarianism if it can be shown that the ways of reasoning that it designates as ‘rational’ undermine the full acknowledgement of suffering, the growth of compassionate fellow‐feeling.

In short, if we reject the Stoic tradition on the matter of self‐sufficiency, we must, to be consistent, reject its arguments for the dismissal of emotion—they stand or fall together. The argument with the Stoics is not finished; and yet we have some good reasons to feel dissatisfied with their position as a basis for social morality. If, however, we should reject their claims about the worth of external goods, then we have no good reason to (p.380) reject the emotions: and we must grant that they are, in many central cases, at least, both true and rational, worthy of guiding our deliberation.

And something more seems to follow. If one agrees with most of this tradition in holding that certain sorts of beliefs about the importance of externals are not only necessary, but also sufficient for emotion—and this seemed to be a very plausible position to take—then one will have to grant that if emotion is not there, that judgement is not (fully) there. This means that those who accept the judgements about the worth of externals that the Aristotle/Rousseau tradition puts forward (against the Stoics) must, to be consistent, admit emotions as essential elements in good reasoning about these matters. And this has consequences for economic reasoning. Economic accounts of human motivation as based on ‘rational self‐interest’ have recently been criticized a good deal, both in philosophy and in economics itself, on the grounds that such accounts fail to do justice to the way in which good reasoning ascribes intrinsic worth to persons, and to the ties of sympathy and commitment that bind people to one another in defiance of self‐interested calculation. A leading example of such criticism is Amartya Sen's famous paper, ‘Rational Fools’, which defends sympathy and commitment in a manner closely related to my argument concerning pity or compassion.44 What reflection about the structure of the emotions suggests, however, is that one cannot fully articulate Sen's more complex normative theory of reasoning without including prominently the emotions in which parts of that reasoning are embodied. It does not seem to be a consistent position to accept an anti‐Stoic account of what has worth, and yet to reject the guidance of love, grief, pity, and gratitude—since such judgements of worth seem to be sufficient for those emotions.45

The world from which emotions are banished is an impoverished world, in which no ultimate ends exist but states of oneself. This norm is not only a very dubious basis for an account of good reasoning, but also an empty universe that cannot sustain an agent's interests long or fulfil her search for meaning. Consider Dickens's Bitzer, with whom we began: we hardly understand how Bitzer goes on living, or what he lives for. With his ‘mechanical laugh’, he strikes the reader as a machine propelled from behind, rather than a human being summoned to action by some beauty or wonder before her.46 It cannot be rational to make ourselves into such beings—even when we write economics!

(p.381) 3.3. Emotions and Impartiality

The calculating intellect claims to be impartial and capable of strict numerical justice, while emotions, it alleges, are prejudiced, unduly partial to the close at hand. Each human being should count as one, and none as more than one, it insists. But in the emotions, the loves of family and close friends seem all‐encompassing, blotting out the fair claims of the distant many. Thus the emotional reasoner, accustomed to cherishing particulars rather than to thinking of the whole world, reasons in a way that is subversive of justice.

This, however, we may doubt. The abstract vision of the calculating intellect is a blunt and relatively obtuse instrument, unless aided by the vivid and empathetic imagining of what it is really like to live a certain sort of life. The person who approaches a distant life with compassion, experiencing what it is like to live such a life, is enabled by that imaginative work to see further and deeper than a person for whom that life was just a number in a mathematical function. Here I want to turn to Dickens once more: for, in his account of the political economy lesson of the circus girl Sissy Jupe, he shows us vividly the rationality and reach of emotion‐based vision, the relative coarseness of mathematical calculation.

Sissy is told by her utilitarian teacher that in ‘an immense town’ of a million inhabitants, only twenty‐five are starved to death in the streets. The teacher M'Choakumchild asks her what she thinks about this—plainly expecting an answer expressing satisfaction that the numbers are so low. Sissy's response, however, is that ‘it must be just as hard upon those who were starved, whether the others were a million, or a million million.’ Again, told that in a given period of time 100,000 people took sea voyages and only 500 drowned, Sissy remarks that this low percentage is ‘nothing to the relations and friends of the people who were killed.’ In both of these cases, the numerical analysis comforts and distances: what a fine low percentage, says M'Choakumchild, and no action, clearly, need be taken about that. Intellect without emotions is, we might say, value‐blind: it lacks the sense of the meaning and worth of persons that the judgements internal to emotions would have supplied. Sissy's emotional response, by contrast, invests the dead with the worth of humanity. Feeling what starvation is for the starving, loss for the grief‐stricken, she says, quite rightly, that the low numbers don't buy off those deaths, that complacency just on account of the low number is not the right response. No such number is low enough. Because she is always aware that there is no replacing a dead human being—the central fact, we might say, around which the emotions weave their fabric—she pushes for more relief, more protection, less complacency. Dealing with numbers one can always say, ‘This figure is all right’—for none of them has any intrinsic meaning. (And really, notice that 500 deaths out of 100,000 is incredibly high for ocean crossings, whether by sea or air.) (p.382) Dealing with imagined and felt human lives, one will accept no figures of starvation as all right, no statistics of passenger safety as low enough. Judge which approach would lead to a better public response to a famine at a distance, to the situation of the homeless, to product testing and safety standards.

We may add to this general argument a genetic thesis. It is, that intimate bonds of love and gratitude between a child and its parents, formed in infancy and nourished in childhood, seem to be the indispensable starting‐point for an adult's ability to do good in the wider world of adult social concern. Such initial cares need further education, to be sure; but they must be there if anything good is to come of education. This point is at least as old, in the Western tradition, as Aristotle's criticism of Plato in Politics II, where he insists that removing the family—rather than ensuring impartial and equal concern for all citizens—will ensure that nobody cares strongly about anything. It is brilliantly developed in Dickens's chilling account of the education of the young Gradgrind children, who are taught to calculate but never encouraged to love.47 Their minds and hearts become thoroughly listless, lacking in any motivational energy for good, and in understanding of what human good there is to be done. We see the same point in Achebe's tragic portrait of Okonkwo's son Nwoye. Deprived of love not only by his father's refusal of feeling but also by the murder of his only companion, he becomes sluggish and listless—until the other‐worldly promises of Christianity give him a source of meaning beyond the life he has previously known. It is clearly the promise of Christ's inextinguishable love that effects his conversion.48

3.4. Emotions and Classes

As for the related objection that emotions are too much concerned with the individual, and too little with larger social units such as classes, we must grant that in fact the emotions do tend to see the world in a variegated and particularized way. In this sense, the vision of community embodied in the emotions is a liberal vision, in which each individual is seen as valuable in her own right, and as having a distinctive story of her own to tell.49 While emotions emphasize the mutual interdependence of persons, showing the world as one in which we are all implicated in one another's good and ill, (p.383) they also insist on respecting the separate life of each person, and on seeing the person as a separate centre of experience—including emotional experience.

It is in this sense no accident that mass movements frequently fare badly in an emotionally rich form such as the novel: as the Communist friends of Doris Lessing's literary heroine are quick to notice, as Lukács was quick to notice when he condemned as ‘petit bourgeois’ the political vision of Tagore's The Home and the World.50 To the extent that they neglect the separate agency of their members, their personal experiences and their qualitative differences, mass movements will be seen as obtuse, and perhaps also as repressive. Whether we are dealing with Dickens's Coketown or Tagore's West Bengal, the violence done by abstract politics to individual lives is likely to grab our sympathy.

This political attitude has its dangers; and sometimes the suspiciousness of any form of collective action that is natural to emotion‐based reasoning can lead to error—as when, in Hard Times, Dickens seems to assume that trade union movements must be repressive and obtuse toward their members. Far more often, however, this attitude is compatible with, and actually motivates, radical socio‐economic criticism—as when, in Sissy Jupe's lesson, the emotions themselves showed the real meaning of hunger and misery, directing the calculative intellect to interpret the numbers in a new and more critical spirit; as when, in Tagore's mordant portrayal of early Indian nationalism, we find Sandip and Bimala neglecting, in their abstract nationalistic zeal, the real economic misery of the poor traders who cannot earn a living unless they sell the cheaper foreign wares, whereas the compassionate particular vision of Nikhil51 makes each human life really count for one.52 And indeed it seems appropriate for any form of collective action to bear in mind, as an ideal, the full accountability to the needs and particular circumstances of the individual that an emotion‐based form like the novel recommends, in its form as well as its content. This, after all, is what Sen's capability approach has all along recommended for development economics: the recognition that each individual has a separate life story, and that good planning should aim to understand those stories, so (p.384) that it can make each distinct individual capable of fully human functioning. An account of class actions, without the stories of individuals, would not show us the point and meaning of class actions, which is always the amelioration of individual lives. Raymond Williams put this point very well, defending traditional realist narrative and its emotional entanglements against this political criticism:

Moreover we should not, as socialists, make the extraordinary error of believing that most people only become interesting when they begin to engage with political and industrial actions of a previously recognized kind. That error deserved Sartre's jibe that for many Marxists people are born only when they first enter capitalist employment. For if we are serious about even political life we have to enter that world in which people live as they can as themselves, and then necessarily live within a whole complex of work and love and illness and natural beauty. If we are serious socialists, we shall then often find within and cutting across this real substance—always, in its details, so surprising and often vivid—the profound social and historical conditions and movements which enable us to speak, with some fullness of voice, of a human history.53

So too, I think, in Lessing's and Tagore's novels: their heroine and hero, on account of the rationality of their emotions, are able to grasp the human significance of political movements and events in a way that their more dogmatic associates are not.54 So too, as well, with Sen's capability approach: precisely because it is determined to come to grips with each persons' concrete life story, precisely for this reason it can pursue radical goals in a truly productive way.

3.5. Love and Blindness to the General Good

This objection is closely related to our third one: but it takes a narrower focus. Adam Smith's analysis concludes that most of the major emotions, including anger, grief, pity, fear, and some forms of love, are all appropriate in the good citizen and judge, and capable of taking cognizance of the good of distant others in morally appropriate ways. To this extent he accepts the argument I made: that emotion‐based reasoning can see very well into the good of people far away, just on account of the special vividness and worth with which the emotions endow the objects of their imagining. Smith holds, however, that romantic and erotic love are special exceptions to this rule, always inappropriate when we seek to deliberate rationally and powerful obstacles to good citizenship and judgement. He leaves little doubt in his reader's mind that women are a primary cause of this problem: seductresses of male rationality and far gone in irrationality themselves.

(p.385) This is a difficult objection to answer briefly, since what we need in order to answer it properly is a comprehensive theory of the development of moral motivation in the child and the links between this motivation and erotic desires and attachments. I shall therefore confine myself to suggesting, tentatively, an answer that emerged from my analysis of Dickens's David Copperfield, which I read as implicitly responding to and criticizing the sort of position Smith defends.55

David Copperfield presents the reader with a fascinating picture of the complex interweavings between erotic desire and novel‐reading, and between both and moral motivation. For David, when he retreats from the gloomy ‘firmness’ of the Murdstones to the room where his father has left a collection of novels, spends hours passionately ‘reading for life’—and not just reading, but also enacting his favourite novel plots, some of which (the plot of Tom Jones for example) have a marked erotic content. He casts himself always, he tells us, in the hero's role, Murdstone in that of the villain; and it is perfectly clear, though he does not name his mother, that she plays the heroine's role. In this way, the activity of novel‐reading nourishes and gives expression to his already passionate emotions for his mother, which include a perception of her as vulnerable and in need of help, a desire to come to her aid, a fierce jealousy of Murdstone. The romantic plots of the novels strengthen his fantasy life, giving new intensity to the emotions he feels for the real people, whom he now sees as the characters. Love fuels imagination, and imagination gives new strength to love.56

What the novel makes clear, however, is that this erotic susceptibility in David—although not altogether morally unproblematic—is not the simple obstacle to good judgement that Smith believes it to be. For it is an essential part of what has made him the generous, exuberant, merciful, gentle moral being he is. The suggestion is that people who are capable, at times, of becoming completely wrapped up in another particular individual in a way that is not based altogether on judicious moral assessment are likely, too, in social life to prove more generous, less retentive and judgemental, than people who always see things from the moral point of view. Their passionate involvement in life, furthermore, gives them more energy and exuberance in social beneficence; and since David's erotic imagination focuses from the first on images of protecting that which is weak and needy, those images, too, enter into his conduct as a social agent.

This romantic conception of citizenship needs further development and argument.57 But I hope this suffices to indicate in what direction I believe (p.386) we ought to move in answering this objector. Such an answer would show that a stance toward life for which women have frequently been criticized might actually be an ingredient in the best conception of citizenship.

3.6. Emotions and Institutions

But why focus so much on emotions? And why urge development economics, which has so many other urgent tasks, to take emotions so seriously? For even if emotions are essential for human flourishing and even for ideal citizenship, what we should be doing is to design things so that they will produce good results even when we don't have ideal people. This is a profound objection, and I shall not be able to give it a comprehensive answer. But four points, I believe, can be made in defence of this emphasis.

First, in any well‐ordered society we do in fact need ideals of the citizen, and of rationality, in order to guide moral and civic education and norms of both family and political life. Every society does employ some such ideal. Even when we are aware that we will never succeed in producing a nation of ideal people, even when we are also determined to design institutions to protect people against the non‐ideal behaviour of their fellow citizens, still, we do bring people up in one way rather than another, in accordance with a certain public conception or conceptions. If those conceptions are not adequate, they will be inadequate. For example, if we do not teach the equality of all citizens, we will by default be giving support to racism and other forms of prejudice. In this case, the ideal of rationality that I recommend, if adopted, would reshape the education of children and the structures of family relations in many ways; and I am arguing that this will have an impact on gender conceptions and gender relations in our society, even if not perfectly realized. One may note that Adam Smith did not fail to investigate the emotions just because he believed that we could not produce emotionally perfect people.58 In fact, he devoted half of his philosophical output to this topic, probably a larger ratio than any other major philosopher in the history of the subject.

Secondly, we should recall that the overall aim of the capability approach in development economics is to secure to each and every person the necessary conditions of fully human functioning in all the major areas of human flourishing. This means that we need to understand what human flourishing requires, in order to support it more adequately. An ideal of rationality that includes emotion is a component part of an ideal of human flourishing. And so my conclusions give direct instructions to policymakers concerning what requires support. This can be expected to have an impact on the design of institutions (especially in the areas of education and (p.387) of family policy) in many ways. In other words, when we depart from classical liberalism by taking a greater interest in the good, and in self‐realization, we need to do some hard close thinking about what the good requires, precisely in order to design institutions that will give people support on the way to being at least capable of flourishing.

Thirdly, although it is correct that in many areas of public life we can indeed design things institutionally in such a way as to yield the results that an ‘ideal person’ would have chosen, even when we don't have such ‘ideal people’, there are always going to be some areas of public life in which we really need the individual people to be fully rational in something like the complete way an ideal would recommend. In US political life, for example, one of these areas is the judiciary. The design of judicial institutions leaves latitude for flexibility and individual interpretive and normative reasoning. There are good reasons for leaving this latitude: for no document can contain instructions so precise and so unambiguous that it will settle every problem in advance, and even to try for this would no doubt conduce to a baneful rigidity in the law.59 But once the latitude is there, we need judges who exemplify rationality; and so we need an adequate conception of rationality to give them guidance. In the area in question, it is especially clear that the absence of an adequate account of the relationship between emotion and reason has led to some very bad judicial reasoning in the criminal law, and in constitutional law as well. What I am saying in effect is that we can't produce an entire society of ‘perfect people’, but we had better do the best we can with our judges, by trying to make the conceptions of rationality involved in their education more philosophically adequate.60

Finally, there is a point that should already be obvious from my discussion of Adam Smith. It is some ways the deepest and most important point in favour of ideal thinking. It is that the design of public institutions itself embodies an ideal conception of rationality. Adam Smith's account of the ‘judicious spectator’ is not just a nice account of the way nice people think. It is a device that is central to his thinking about how social and economic institutions are well designed. The idea of what a judicious spectator will feel, think, and approve is an artificial construction by which we gain insight concerning justice, and the just structure of institutions. This point is even clearer if we consider John Rawls's A Theory of Justice,61 (p.388) since Smith tends to believe that things are in good order as they are, whereas Rawls actually puts his construction of the person to work in the design of principles and policies. Rawls makes it explicit that a conception of the person, which includes a comprehensive normative account of rationality, is basic to his argument for his conception of justice. Only when we understand what these ‘ideal people’ are like, what information they do and do not use, and how they reason, will we have a purchase on the proper design of institutions and the just distribution of resources. I believe that Smith and Rawls are correct: we need an ideal of reason in order to design institutions justly. But if this is so, and if, as I have argued, many conceptions of reason (including Rawls's)62 err by their neglect of the emotions, then, in order to have just social institutions, we need to think more about emotions.

I have argued that the emotions are best understood not as blind and brutish, but as intelligent forms of evaluative perception, either identical with or very closely linked to judgements; that, while Spinoza and the Stoics are right to link emotions with beliefs about our lack of self‐sufficiency, there are reasons to think such beliefs both true and politically valuable; that while emotions do begin at home, so to speak, the love and gratitude of the home is a necessary prerequisite for an adequate awareness of the needs of those at a distance; that emotions do focus on particulars rather than on classes, but that in so doing they show the point and ultimate goal of class action; that even the connection of emotions with extra‐moral sexual energies can make a valuable contribution to both private and public life; that even when we know we will be dealing with imperfect people in an institutional setting, there are strong reasons for giving these features of our rationality our most serious consideration.

4 Emotions and Social Construction, Emotions and Gender

It will by now be clear that my view of the emotions is closely connected to recent views (defended in anthropology above all) according to which the emotions are ‘socially constructed’. What is meant by this is that there is no emotional experience without belief; beliefs, however, are learned, and learned in society. Therefore both the fact that a person has emotions (rather than living a Stoic life) and the particular repertory of emotions she has will be best explained by examining the society in which she was raised, and asking what it has taught her to think.

My view is, as I have said, closely related to these views: and yet it is in (p.389) certain ways distinct from them also.63 For I insist on stressing the connection between emotional experience and judgements about the importance of ‘external goods’—people and things outside the person that the person him or herself does not fully control. Such judgements are likely to be found in some form in any society given the general structure of human life, its mortality, its susceptibility to illness and fatigue, its dependence on goods that are in short supply. Society may give emotions a particular focus and colouring: Roman beliefs about what the object of love is like, about what is worth getting angry about, about what is really to be feared, differ in subtle ways from Greek beliefs on similar topics—and so forth. But it would be surprising indeed if there were not a great deal of overlap among societies in the general repertory of emotions they teach. And if we add to this the picture of early infant experience that I sketched in my previous chapter, it seems likely that emotions of fear and love and gratitude and envy and anger have their roots in infancy itself, long before society has had a chance to shape the infant very much in its own image. The infant's interactions with its parents are, of course, still shaped in certain ways by societal teachings that vary from one culture to another; but I think it is still possible to develop a general account of infant experience, where emotions are concerned, that will convincingly bring together the common experience of many different cultures. This means not only that it is highly unlikely that we could ever develop a fully Stoic society, even if we wanted to do so (even the Stoics did not think this a real possibility); it means, too, that the emotional repertories of different societies will be likely to have a great deal of overlap, in such a way that it is possible for us to see ourselves in their accounts of what they fear, and cherish, and get angry about, as they themselves in ours.

What now of gender? For it is almost universally, if unclearly,64 observed—even among philosophers who hold a cognitive view of the emotions—that women are frequently more emotional than men. How should someone with a cognitive account of emotion explain these phenomena, or purported phenomena?

My view holds that the emotions do not arise from ‘nature’ at all: then, a fortiori, they do not arise from some female as distinct from male nature. In fact, on the account of the emotions I have offered, there is no intrinsic reason why the emotions should be distributed along gender lines at all. (p.390) And of course they are not always so distributed. In some societies, extremes of fear, anger, and grief are judged normal and socially appropriate for males as well as females.65 In still others, emotions are judged inappropriate for both.66 In some societies, emotions are judged appropriate for both, but different emotions in each case.67 But to the extent that women are commonly more associated with the emotional than are men, there is within my view an explanation ready to hand.

Emotions, I have said, are acknowledgements of neediness and dependence, acknowledgements of the importance of things outside the self that the person does not control. To some extent all human beings have such attachments to undependable external things. (Even the Stoics, after the time of Zeno and Cleanthes, stopped claiming that any fully ‘wise’ (i.e., completely unemotional) person was really alive.) But one can see that contingent social facts about control and dependency can shape the course of these attachments. Two areas of social difference between males and females seem to be especially significant. First, in many societies, from Okonkwo's Umuofia to the contemporary US, women are more likely to be brought up to think that it is right to have strong attachments to others, and to think of their own good as a strongly relational matter; males are more likely to be urged in the direction of self‐sufficiency and the curtailment of deep need.68 This is not a universal cultural fact; and the ancient Greco‐Roman world was, it seems to me, far less gender‐divided than modern American society is on this question. There are emotions, furthermore, for which it is less likely to hold, even for Americans: angers, fears, and anxieties connected with status, money, and reputation are very likely to be expecially linked with the socialization of males, in US society as much as in ancient Rome and in Umuofia. But the diagnosis has some explanatory force.

A second difference, of a very different sort, takes us further. The emotions, we recall, are connected with judgements about the importance (p.391) of external things that one does not control. Now obviously there are some important features of human life that nobody ever fully controls. One cannot make oneself immortal, one cannot will that one's children should be healthy and happy, one cannot will oneself happiness in love. But nonetheless, differences in degree of social autonomy and control between males and females do affect the extent to which the sense of powerless governs the course of one's life. The heroines of Seneca's tragedies, driven mad by emotion, are usually wives who have no avenue to happiness but through their marriage. They have neither full citizenship nor full education; they have no avenue to productive self‐expression outside the home. Thus, for them, everything hangs on the fate of this one attachment; and when it collapses they are, accordingly, in a state of complete disorder, as though they had no self at all. Roman males never experienced exactly that: for they had many things to cherish, not all of them dependent on the whim of a single other person (who is not encouraged by socialization to be especially dependable).

This situation is not just history or fiction: it is a daily fact of life for women in most parts of the world. We can see it at work in the life of Saleha Begum, for whom a husband's disability—given social laws against women's work—meant hunger for herself and her family; and even more clearly in the story of Metha Bai, whose only hold on life is her father's fragile health. Such women are far more dependent on fortune than are most males in their society, in all sorts of ways: dependent even, frequently, for their daily food and medical care. What wonder, then, if they should experience more fear, more grief, and also, frequently, more anger?

These two social facts point, it seems (not incompatibly) in opposite directions. The first suggests that one might need to reform the moral education of men such as Okonkwo, in order to make them value more properly the attachments to others that are so large a part of women's lives. The second, however, suggests that there is a degree of neediness, dependency, and powerlessness that is not connected with any intrinsic good in anyone's life—and that it would be best to move women, and indeed everyone, out of this condition of extreme dependency. In both cases, what we need, clearly, in order to make further progress, is an account of the good human life and good human functioning that will tell us which attachments to externals are truly valuable and which are not. I have tried, in my first chapter, to sketch a candidate for such an account: but much more remains to be done if it is to be concrete enough to offer any real guidance, especially where the structure of family attachments is at issue.

‘But women in every culture love and support their children.’ Both anti‐feminists and certain (alleged) feminists have pointed to the universal fact of mothering as evidence that there is, after all, a separate female nature, intuitive and emotional rather than calculative and intellectual. Appeals to mothering play a large role in many types of anti‐feminist and ‘feminist’ (p.392) argument, ‘explaining’ why logic is not a female domain, why women's thinking is bound to be intuitive and associative rather than logical, why women are better suited for some jobs than others, etc. So any analysis that claims, as mine does, that emotions are gender‐linked only for contingent social reasons needs to consider this case and take a stand on it.

I have already shown the direction in which I would search for the response, by saying that my approach combines an interest in social teaching with an interest in much earlier and less explicit learning. The emotional repertory is very largely shaped in infancy. And yet infancy is not a time in which unshaped ‘nature’ simply expresses itself; it is a time of complex interaction and highly charged experience. I have said that it seems plausible that this experience, though differently shaped by different societies, has certain common features, since all societies have in common certain pervasive human problems. I can now add that it seems very likely that most societies have shared certain gender‐based patterns in this crucial early phase of experience, in such a way that ‘mothering’ is reproduced in females from generation to generation. The work of Nancy Chodorow has, I believe, convincingly established this conclusion.

Chodorow's work is convincing because she realizes that the phenomena cannot be explained simply by pointing to behavioural patterns that a child acquires during its recognizable socialization: they lie deeper, in such a way as to be part of the structure of the personality itself:

The capacities and orientations I describe must be built into personality; they are not behavioral acquisitions. Women's capacities for mothering and abilities to get gratification from it are strongly internalized and psychologically enforced, and are built developmentally into the feminine psychic structure. Women are prepared psychologically for mothering through the developmental situation in which they grow up, and in which women have mothered them. (See above, n. 68)

But to say that these orientations are built into personality is not to say that they derive from an innate ‘nature’: for the structure of the personality is itself formed, though at an earlier date and at a level of depth that is not readily accessible to conscious analysis.69 By focusing on these earliest interactions and the structures of personality they transmit, Chodorow is able to explain gender‐differentiation as an outgrowth of social tradition—here operating at a very deep level both to shape the personality of the mother and her care of female and male infants and then, in turn, to reproduce that gender‐differentiated structure in the lives of infants who will themselves grow up to take part in gender‐differentiated societies.

(p.393) Chodorow's argument is highly complex, and I hesitate to attempt, here, a summary that is bound to be but a caricature of her more subtle and finely demarcated analysis. But her central conclusion is that, very widely if not universally, mothers' interactions with female infants reproduce the maternal personality, forming people who conceive of their own maturity and adulthood as involving close ties of a nurturing sort. The mothering of male infants, by contrast, tends to produce people who conceive of maturity and independence in terms of the denial of need and dependence. Thus Chodorow agrees with the social diagnosis of the ancient Stoics, though she extends the causal analysis of these differences back into infancy in a way that they did not (though they would very likely have sympathized with the project). But her normative evaluation agrees more with my argument here than with the Stoic argument: for she argues that the degree of self‐sufficiency males come to expect of themselves is in various ways crippling to them, both in their emotional lives generally and above all in their relationships with women. And women's confinement to the mothering role has, itself, adverse social consequences, affecting their self‐definition and what they can bring to work, to citizenship, even to relationships with others.

The diagnosis ends with a proposal. These things will not be easy to alter, given the depth and pervasiveness of the causes, Chodorow holds. If they are to be altered, it will only be through a reformation of parenting itself. But this, she believes, is possible:

Children could be dependent from the outset on people of both genders and establish an individuated sense of self in relation to both. In this way, masculinity would not become tied to denial of dependence and devaluation of women. Feminine personality would be less preoccupied with individuation, and children would not develop fears of maternal omnipotence and expectations of women's unique self‐sacrificing qualities. This would reduce men's needs to guard their masculinity and their control of social and cultural spheres which treat and define women as secondary and powerless, and would help women to develop the autonomy which too much embeddedness in relationship has often taken from them.

We live in a period when the demands of the roles defined by the sex‐gender system have created widespread discomfort and resistance. Aspects of this system are in crisis internally and conflict with economic tendencies. Change will certainly occur, but the outcome is far from certain. The elimination of the present organization of parenting in favor of a system of parenting in which both men and women are responsible would be a tremendous social advance. This outcome is historically possible, but far from inevitable. Such advances do not occur simply because they are better for ‘society’, and certainly not simply because they are better for some (usually less powerful) people. They depend on the conscious organization and activity of all women and men who recognize that their interests lie in transforming the social organization of gender and eliminating sexual inequality. (See above, n. 68)

(p.394) I wholeheartedly agree with this conclusion. I would only add that when people of this sort ask themselves where their interests lie, they should not, if they are truly rational, use a narrow economic conception of ‘interests’ and think of themselves simply as maximizers of their own feelings of satisfaction. The question should, instead, be considered in terms of the interest a really rational being has in leading a rich and flourishing life that would combine into a more or less coherent whole all those forms of human functioning that have intrinsic value.

5 Emotions and Gender Justice

The question of the emotions thus becomes part and parcel of the general question of human functioning and human capability. In asking what emotions it is good to have, what emotions form part of truly rational judgement, we are asking, in effect, what forms of dependence and affiliation with undependable items outside ourselves it is good for a human life to have. And this involves asking how much uncertainty and dependency a person can live with while retaining integrity and practical reason; how far trust in others is a good thing and at what point it becomes naïvety; whether, indeed, it is best to live a life of what Thrasymachus (speaking of justice) called ‘a very noble kind of naïveté’, or a life that consists in prudently maximizing one's own satisfactions; whether it is good to love anyone at all, given the depths of the pain that love can inflict; whether it is good to build a society on the basis of these needs, loves, and attachments, or rather on the basis of respect for the self‐sufficiency of reason. In short, it is the question of the human good.

It is not surprising that there should be deep division—among human beings and indeed perhaps within each human being—over these questions. For vulnerability is indeed painful, and trust difficult to sustain in an inhospitable and uncertain universe. The anti‐emotion position will keep coming back, in philosophy and in life itself, as long as there are people who cannot bear to live in a world of deep attachments, who seek another purer world. And it will remain, therefore, a topic of dispute whether the emotions, seen as acknowledgements of incompleteness, are good or not, whether their associated beliefs are true and rational or false and pernicious. It seems right to end this paper not with a dogmatic answer to that question—though my argument will have made clear where my own sympathies lie—but with an acknowledgement of its complexity. These lines of Wallace Stevens seem to me a wonderful metaphor of the emotional life, in all of its mystery, wonder, and pain:70 (p.395)

  • The body walks forth naked in the sun
  • And, out of tenderness or grief, the sun
  • Gives comfort, so that other bodies come,
  • Twinning our phantasy and our device,
  • And apt in versatile motion, touch and sound
  • To make the body covetous in desire
  • Of the still finer, more implacable chords.
  • So be it. Yet the spaciousness and light
  • In which the body walks and is deceived,
  • Falls from that fatal and that barer sky,
  • And this the spirit sees and is aggrieved.


I am extremely grateful to all participants in the WIDER conference for their valuable comments, and especially to Cathy Lutz, whose work I have long admired, for the quality of her commentary. Others whose comments have played a role in the rewriting of the paper are Martha Chen, Nancy Chodorow, Jonathan Glover, Christine Korsgaard, Onora O'Neill, Richard Posner, John Roemer, Cass Sunstein, and Susan Wolf.

(1) Achebe, Things Fall Apart (London: Heinemann, 1958, repr. Picador 1988), 25: ‘His whole life was dominated by fear, the fear of failure and weakness. It was deeper and more intimate than the fear of evil and capricious gods and of magic, the fear of the forest, and the forces of nature, malevolent, red in tooth and claw. Okonkwo's fear was greater than these. It was not external but lay deep within himself. It was the fear of himself, lest he should be found to resemble his father. Even as a little boy he had resented his father's failure and weakness, and even now he still remembered how he had suffered when a playmate had told him that his father was agbala. That was how Okonkwo first came to know that agbala was not only another name for a woman, it could also mean a man who had taken no title. And so Okonkwo was ruled by one passion—to hate everything that his father Unoka had loved. One of those things was gentleness and another was idleness.’

It is a significant point in the novel that anger is perceived as manly, not womanish, and that anger frequently can be used to keep down the gentler emotions. Thus Okonkwo's cruelty and anger toward his wives and children is analysed as a direct outgrowth of his fear of weakness. I shall have more to say about this in what follows.

(2) For examples from India, and an argument that both male and female flourishing are damaged by this situation, see Roop Rekha Verma's paper in this volume, and also her ‘Development and Gender‐Essentialism’, in Glover, Nussbaum, and Sunstein (eds.), Women, Equality, and Reproduction (forthcoming); for connections between Indian emotion categories and female receptivity, see Marya Simon, The Weavers (a cycle of poems) (forthcoming). For Chinese traditions, which depict the best man as passionless, see Lee Yearley, Mencius and Aquinas (Albany, NY: State University of NY Press, 1990), and for a penetrating analysis of traditional Chinese emotion categories, which links them closely to the analysis of emotion in ancient Greek Stoicism, see Lothar von Falkenhausen (forthcoming). Other valuable examples are cited in Catherine Lutz, Unnatural Emotions: Everyday Sentiments on a Micronesian Atoll and Their Challenge to Western Theory (Chicago, Ill.: University of Chicago Press, 1988). The Utku Eskimos studied in Jean Briggs's Never in Anger (Cambridge, Mass.: Harvard University Press, 1981) urge the elimination of most of the major emotions, and especially anger, for both males and females—see further below.

(3) I discuss the novel's account of imagination and emotion more fully in Poetic Justice: The Literary Imagination in Public Life (Boston: Beacon Press, forthcoming 1995). Ch. 2 has appeared in an earlier version in New Literary History, 22 (1991), 879 ff. For further discussion of Hard Times in connection with development economics, see M. Nussbaum and A. Sen, ‘Introduction’, in Nussbaum and Sen (eds.), The Quality of Life (Oxford: Clarendon Press, 1993). The edition of the novel I use is the Penguin edition, ed. David Craig (Harmondsworth, 1969).

(4) Richard A. Posner, The Economics of Justice (Cambridge, Mass.: Harvard University Press, 1981), 1. I have discussed Posner's utilitarianism in two reviews of his book Sex and Reason (Cambridge, Mass.: Harvard University Press, 1992), in which he proposes to extend economic analysis to the sphere of human sexual relations. A short review is in The New Republic 20 April 1992, and a longer one, more concerned with the issues of this paper, in The University of Chicago Law Review, fall 1992. Posner has now significantly modified his views about the adequacy of utilitarianism as a normative theory of public reasoning, since he recognizes that it does not suffice to generate the Millean libertarian principles that he would favour.

(5) Posner, Economics of Justice, 1–2.

(6) Here Posner follows the lead of Gary Becker, The Economic Approach to Human Behavior (Chicago, Ill.: University of Chicago Press, 1976), A Treatise on the Family (Cambridge, Mass: Harvard University Press, 1981). For criticism of Becker's models, see Sen's paper in this volume, and my discussion of it in the Introduction.

(7) Posner, Economics of Justice, 345. Posner's view of the privacy issue has now significantly changed. In Sex and Reason he argues that it is legitimate to fill gaps in the explicit provisions of the Constitution by finding a general right to privacy, and very important for courts to do so if they are to prevent arbitrary and wilful impositions of legislative power. He strongly upholds the Supreme Court's contraception rulings, supports the outcome, though not all the reasoning, in Roe v. Wade, and argues that Bowers v. Hardwick (in which the Supreme Court refused to recognize a right to privacy where consensual adult homosexual relations were concerned) was both wrongly decided and extremely ill argued. His own ‘libertarian’ theory of sexual legislation holds that the law may intervene in sexual matters only in order to protect the ‘liberty or property’ of others.

(8) Posner, The Problems of Jurisprudence (Cambridge, Mass.: Harvard University Press, 1990) 453, n. 40: ‘And recall the nonrational component in Aristotelian deliberation. See Martha C. Nussbaum, The Fragility of Goodness: Luck and Ethics in Greek Tragedy and Philosophy, 307–10 (1986).’ Posner's reference is to a section in my discussion of what I call ‘non‐scientific deliberation’ in Aristotle where I argue that the emotions are, for Aristotle, an important part of rational deliberation: ‘Far from seeing them as obstacles to good reasoning, he makes proper passivity and passional responsiveness an important and necessary part of good deliberation’ (307). The central example is one of grief for a child's death, taken from Euripides' The Trojan Woman. Posner is so thoroughly convinced that emotions cannot be rational that he can read this discussion and summarize it as a discussion of ‘non‐rational’ components. In more recent work, however, Posner significantly modifies this view: see especially the essay on rhetoric in Overcoming Law (Cambridge, Mass.: Harvard University Press, 1995).

(9) This qualification is necessary where Posner is concerned, since he has explicitly said that he thinks the association of emotion with women is misleading: females are just as rational, in the utilitarian sense, as males. But many holders of such a view do not so stipulate. And Posner himself, in The Problems of Jurisprudence, discusses compassion and related emotions in a chapter entitled ‘Literary and Feminist Perspectives’.

(10) Bradwell v. the State, 83 US (16 Wall.) 130 (1872). For a discussion of the case I am indebted to Susan Okin, Women in Western Political Thought (Princeton, NJ: Princeton University Press, 1979), 252–6. See also Cass Sunstein's paper in this volume.

(11) Sometimes we find a related position about women's anger that would not have been endorsed by Okonkwo, who thinks of anger as ‘male’: thus Seneca's Medea claims, in a way typical of the Greek tradition, that women's jealous rage is stronger than any ‘violence of fire or swelling gale’. Rage, like timidity, can be thought a barrier to full civic equality—especially when it is the rage of a woman that is in question.

(12) Politics 1260a 12–13: women have the deliberative faculty, but it is ‘lacking in authority’—apparently, vis à vis their ‘irrational’ emotions.

(13) On Islam, see Valentine Moghadam, Modernizing Women: Gender and Social Change in the Middle East (Boulder, Colo., and London: Lynne Rienner, 1993). On Hindu traditions, see Verma's paper in this volume; on China, see Li.

(14) On these arguments, with references to some other analyses, see my paper ‘Human Capabilities’ in this volume.

(15) For an excellent account of these views, see Okin, Women, ch. 10.

(16) Mary Wollstonecraft is a striking example of this line of argument; like most ethical thought in her century, hers is profoundly influenced by Stoicism. See Jane Roland Martin, Reclaiming a Conversation (New Haven, Conn.: Yale University Press, 1985).

(17) For criticism of some of this, where India is concerned, see M. Nussbaum and A. Sen, ‘Internal Criticism and Indian Rationalist Traditions’, in M. Krausz (ed.), Relativism (Notre Dame, Ind.: Notre Dame University Press, 1989), 299–325. And Sen, ‘India and the West’, The New Republic, June 1993. On Africa, see the excellent rebuttal of the simple view in Anthony Appiah, In My Father's House: Africa in the Philosophy of Culture (New York: Oxford University Press, 1992).

(18) Havelock Ellis, Man and Woman (Boston: Houghton Mifflin, 1929), cited in Catherine Lutz, Unnatural Emotions, 74.

(19) Lutz, Unnatural Emotions, discusses these associations well; but she is wrong, I think, to claim that this objection to emotion is the predominant one in the history of Western philosophy, and that criticizing it entails a fundamental criticism of ‘Western theory’.

(20) I discuss these views historically in The Therapy of Desire: Theory and Practice in Hellenistic Ethics (Princeton, NJ: Princeton University Press, 1994), and defend a related position in my own way in Upheavals of Thought: A Theory of the Emotions, The Gifford Lectures for 1993, delivered April–May 1993 at the University of Edinburgh and forthcoming with Cambridge University Press.

(21) On the Chinese analysis, see Yearley and von Falkenhausen; for Indian cognitive views of emotion, see Simon.

(22) For an analysis of Seneca's imagery for emotion, with many such examples, see Nussbaum, Therapy, ch. 12.

(23) On China, see Yearley, where I believe that this argument can be found alongside the second.

(24) Lessing, The Golden Notebook (London, New York and Toronto: Bantam, 1981).

(25) I discuss Smith and the argument in detail in ‘Steerforth's Arm: Love and the Moral Point of View’, in Love's Knowledge: Essays on Philosophy and Literature (New York: Oxford University Press, 1990).

(26) Thus Smith speaks of a ‘charm’ from women that temporarily eclipses men's rational powers. Nor are women treated as moral subjects: the argument depicts them as if they lived entirely in the emotions' enclosed and irrational realm. For the American analogue, consider the passage from Thomas Jefferson that forms the epigraph to my paper ‘Human Capabilities’ in this volume. For Indian, Latin American, and Chinese versions, see the papers by Verma, Valdés, and Li. Rabindranath Tagore's novel The Home and the World (London: Penguin, 1985, first published in India, 1915–16, first English publn. Macmillan, 1919) offers a fascinating contrast on this issue. The traditionalist and hyper‐patriot Sandip takes the view that women embody ‘the living image of Kali, the shameless, pitiless goddess’; they are prevented from destroying men, insofar as they are, only by artificial social ‘fetters’. Nikhil, the critical humanist, denies this: the fault lies in society generally, for its excessive exaltation of erotic desire: ‘Man has so fanned the flame of the loves of men and women, as to make it overpass its rightful domain, and now, even in the name of humanity itself, he cannot bring it back under control. Man's worship has idolized his passion. But there must be no more human sacrifices at its shrine’.

(27) I owe this objection to John Roemer.

(28) In Upheavals of Thought (forthcoming, 1996). Especially pertinent to this project are Lectures 5 and 6, dealing with compassion and its role in public life.

(29) To say that the view is true for all human beings (and in fact for higher animals as well, as I argue in Upheavals of Thought) is not to neglect the evident differences among societies in the particular repertory of emotions they construct and recognize. Recognition of the element of ‘social construction’ in emotion is, in fact, a prominent part of the theory itself.

(30) Lecture 1 of Upheavals of Thought is devoted to stating and answering this objection; in Lecture 2 I examine a wide range of materials in cognitive psychology and anthropology that support my reply; in Lecture 3, turning to a developmental account of emotion, I link my cognitive view with recent work in object‐relations psychoanalysis. I refer the reader to this material for references to the literature. Lecture 1 will appear before the book's publication, in a volume in memory of Bimal Matilal, ed., J. N. Mohanty. For related arguments, see also my Therapy of Desire, ch. 10.

(31) This is strikingly true in the law, where there is a growing literature on empathy and emotion in which the defenders of emotion seem to accept the sharp distinction between emotion and reason foisted on them by emotion's attackers. See, for example, Toni Massaro, ‘Empathy, Legal Storytelling, and the Rule of Law: New Words, Old Wounds’, Michigan Law Review, 87 (1989), 2099–127; Lynne Henderson, ‘Legality and Empathy’, Michigan Law Review, 85 (1987), 1574–653. The same thing happens in actual judicial opinions: even the defenders of compassion and mercy speak of them as ‘irrational’. For examples, see my ‘Equity and Mercy’, Philosophy and Public Affairs (Spring 1993). A welcome exception to the trend is M. Minow and E. Spelman, ‘Passion for Justice’, Cardozo Law Review, 10 (1988), 37–76.

(32) Since these later philosophers are deeply indebted to Stoicism, I shall in what follows be focusing on the original ancient Stoic positions.

(33) I owe this way of putting the point to Christine Korsgaard.

(34) I am grateful to Cass Sunstein for valuable observations on this point.

(35) I discuss ancient Greek and Roman views of anger in Therapy, chs. 3, 7, 11, and 12. The most important texts are Aristotle, Rhetoric II. 2–3, Nicomachean Ethics IV. 5; the fragments of Chrysippus' Peri Pathon, collected in Stoicorum Veterum Fragmenta, ed. J. von Arnim (Stuttgart: Teubner, 1903, repr. 1979) Vol. III; Seneca, On Anger (De Ira) and On Mercy (De Clementia); Cicero, Tusculan Disputations III–IV.

(36) The Stoic philosopher Epictetus urged the too‐fearful person to do exercises in which he would begin by considering the loss of a cup, and progress, by gradual steps, to the acceptance of the possible loss of other things—his household pets, his children, his wife, his own life! In each case, he grows to regard them as trivial and dispensable.

(37) On ancient Greek analysis of pity, see my The Fragility of Goodness: Luck and Ethics in Greek Tragedy and Philosophy (Cambridge: Cambridge University Press, 1986) Interlude 2, and ‘Tragedy and Self‐Sufficiency: Plato and Aristotle on Fear and Pity’, Oxford Studies in Ancient Philosophy, 10 (1992) and, in a shorter version, in A. Rorty (ed.), Essays on Aristotle's Poetics (Princeton, NJ: Princeton University Press, 1992). The later history of the debate about its role in public judgement is traced, and adjudicated, in Gifford Lectures 5 and 6.

(38) The Epicurean tradition suggests that this is especially likely to be true of beliefs about the badness of one's own death.

(39) Von Falkenhausen argues convincingly that much the same is true of the Chinese tradition; and I believe that the Indian tradition has a similar position, though perhaps intentional perception is stressed more than belief.

(40) In the Gifford Lectures I argue that the view needs to be modified in two ways in order to be philosophically adequate: we must get rid of the claim that emotions always involve linguistically formulable propositions, making room both for non‐linguistic emotional experiences (Lecture 4 is devoted to emotion and music) and also for experiences that are more indefinite and inchoate than the Stoic propositional analysis would allow. Even in these cases, however, emotions involve a view of the world that contains ideas of value or salience; and I argue that this is true even of the emotions of small infants and non‐human animals.

(41) See Plato, Republic, 387–8; on this critique of literature, see ‘Tragedy and Self‐Sufficiency’, and also ‘Poetry and the Passions: Two Stoic Views’, in J. Brunschwig and M. Nussbaum (eds.), Passions & Perceptions (Cambridge: Cambridge University Press, 1993).

(42) Cicero, Tusculan Disputations III. 30. Yearley gives closely related examples from the Mencian portrait of the sage. And Indian traditions contain similar pictures of detachment. Okonkwo seems to aim at a similar self‐sufficiency.

(43) I use both words because ‘pity’, in recent times, has acquired connotations of condescension to the sufferer that it did not have earlier, and still does not have when used as a translation for Greek eleos, Latin misericordia, or Rousseau's pitié.

(44) A. Sen, ‘Rational Fools’, Philosophy and Public Affairs, 6 (1977), 317–44, repr. in Sen, Choice, Welfare, and Measurement (Oxford: Blackwell, 1982), 84–106.

(45) In Gifford Lecture 6, I use this point to establish that Posner has compassion, through a reading of two of his recent judicial opinions in which the evaluations that (on this view) suffice for emotion are amply present.

(46) Okonkwo is different, because, like Mr Gradgrind, he is not consistent. The love for Ikemefuna that turns him into a ‘shivering old woman’ is a powerful source of motivation throughout the novel, as is his intense attachment to his own status and reputation. In fact it is the tension between these two attachments that produces the tragedy.

(47) Their father refers to love as ‘the misplaced expression’. For a more detailed account of these aspects of the novel, see chapters 2 and 3 of Poetic Justice (Boston: Beacon Press, 1995.

(48) And it is really too late, where the issues of this argument are concerned. For the next novel in Achebe's trilogy, No Longer At Ease, shows that Nwoye, now christened as Isaac, proves an inept and unperceptive father, whose rigidity causes his son Obi's disaster. Thus the tragedy of non‐emotion transmits itself from generation to generation.

(49) See also Lionel Trilling, The Liberal Imagination: Essays on Literature and Society (New York: Charles Scribner's Sons, 1950).

(50) He writes that the novel is ‘a petit bourgeois yarn of the shoddiest kind’, and that Tagore is ‘a wholly insignificant figure . . . whose creative powers do not even stretch to a decent pamphlet.’ Quoted in Anita Desai's Introduction to the Penguin edition of the novel (1985), p. 7. I have discussed the related case of Henry James's The Princess Casamassima in ‘Perception and Revolution’ in Love's Knowledge.

(51) This is seen also in his refusal to dismiss the English governess: ‘I cannot . . . look upon Miss Gilby through a mist of abstraction, just because she is English. Cannot you get over the barrier of her name after such a long acquaintance? Cannot you realize that she loves you?’ This is a self‐referential moment: for it is evident that the vision of the novel as a whole is this sort of particularized vision.

(52) This is complicated, since Sandip and Bimala think of themselves as very emotional creatures indeed, and in a certain sense are so. But Nikhil seems correct when he diagnoses their state as one of a self‐generated overheated excitement, not an acknowledgement of a real tie to any other person or entity.

(53) Raymond Williams, The Politics of Modernism (New York: Verso, 1989), 116.

(54) As the impoverished trader Panchu says to Nikhil, ‘I am afraid, sir . . . while you big folk are doing the fighting, the police and the law vultures will merrily gather round, and the crowd will enjoy the fun, but when it comes to getting killed, it will be the turn of only poor me!’

(55) I discuss the novel in ‘Steerforth's Arm’ in Love's Knowledge, in ‘Equity and Mercy’, and also in Gifford Lecture 3, my developmental account of emotion.

(56) Later on, he links this receptivity with images of gender change as, telling stories to Steerforth in the moonlight, he becomes himself a heroine, ‘Daisy’ and ‘The Sultana Scheherezade’.

(57) I develop it, with reference to Joyce and Whitman, in Gifford Lecture 10.

(58) For an excellent account of his view of the limitations, see R. Coase, ‘Adam Smith's view of Man’, Journal of Law and Economics 19 (1976), 529–46.

(59) Furthermore, there are general results in the theory of social choice that show systems of general rules to be insufficient to determine a rational political outcome. In particular, Allan Gibbard has shown that it is impossible to design a voting system that cannot be manipulated by strategic voting. This result has general implications for the rationality of other systems of rules. This gives us still broader reasons to hold that we need to be concerned with the full personal rationality of individual social actors, and not merely with systematic and institutional considerations.

(60) On emotion in the criminal law, see ‘Equity and Mercy’, and Gifford Lecture 6.

(61) Cambridge, Mass.: Harvard University Press, 1971.

(62) For an account, and criticism, of Rawls's view of the emotions, see my ‘Perceptive Equilibrium’ in Love's Knowledge.

(63) For a more extended account of this, see ‘Constructing Love, Desire, and Care’, forth‐coming in Laws and Nature: Shaping Sex, Preference, and Family (New York: Oxford University Press, 1995).

(64) As Christine Korsgaard observed in discussion, this claim is so vague as to have very little content. Does it mean that women have a larger number of types of emotion than men? A larger number of episodes of the same types? Longer, or more intense episodes? Episodes directed at a larger number of objects? That they spend a larger proportion of their time in an emotional frame of mind? That they make more of their decisions by appeal to emotion? And so forth.

(65) Ancient Athens and Rome—prior to philosophical intervention—seem to be such societies.

(66) See especially Jean Briggs's account of an Eskimo tribe, the Utku, who are very like Stoics in their theory and practice of emotional extirpation. Some Indian norms of detachment also apply equally to serious people of both genders.

(67) Modern America, like Okonkwo's Africa, seems to be such a case, anger being judged more appropriate for males than for females, fear and compassion and love more appropriate for females. Anne Fausto Sterling assembles an impressive amount of data showing that emotions of babies are differently labelled according to the sex the observer believes the babies to have. In one American experiment, for example, alleged girls were labelled frightened when they cried, while alleged boys (actually the very same individual infants) were labelled angry. Such labelling is a part of the process of norm‐transmission that culminates in actual emotional differences.

(68) See, here, Carol Gilligan, In a Different Voice (Cambridge, Mass.: Harvard University Press, 1982), and Nancy Chodorow, The Reproduction of Mothering: Psychoanalysis and the Sociology of Gender (Berkeley, Calif.: University of California Press, 1978) on which see further below.

(69) Both contemporary cognitive psychology and object‐relations psychoanalysis now hold that even very young infants have very rich and complex attitudes to external objects. For one gripping summary of recent research, see Daniel Stern, Diary of a Baby (New York: Basic Books, 1992).

(70) Wallace Stevens, ‘Anatomy of Monotony’, in Collected Poems (London: Faber and Faber, 1923), 107.