The Problem

Consider a lawyer. He is in his mid-forties, with a lucrative law practice in a small town. He took the practice over from his father who is now retired. He is secure and comfortable, has a good enough family life, and a respected place in his community. But he sees the future ahead of him and finds it uninspiring. His family life is affectionate but not passionate. The humdrum cases he competently handles no longer challenge or engage him. He is largely going through the motions, taking diminishing satisfaction in his life. He asks himself whether this is all he could expect until he, like his father, retires and plays golf with similarly dispirited partners. He finds himself dissatisfied with both his personal existence and the political conditions of his society.

He always had an interest in politics and he was quite active in it during his college days. Indeed, there was a time when he was undecided between pursuing a career in law or in politics. He picked law largely because his father expected it of him and because he wanted for himself the comfort, security, respect, and status his father obviously enjoyed. He also saw the uncertainties, risks, and compromises entailed by a political career. He opted for law, and got the satisfactions of the conventional life he sought. But now he finds that its satisfactions are meager. In this unquiet state he thinks back to his choice of career, and, although he does not quite regret the choice he had made, he is questioning it. He asks himself whether he should not enter politics now and give up his law practice. He has not changed his political ideals, he thinks that his country needs people like him and ideals like his, and that there are good reasons for exchanging his insufficiently satisfying life for the active pursuit of his ideals.

He is nothing if not prudent, and after careful reflection he sees that there are good reasons also for continuing his law practice. He believes that the continued well-being of his family rests on his income, and that his wife and children count on the affection he genuinely feels for them and shows, day in and day out, by the attention he pays to them. He also feels loyalty to the many clients who, throughout the years, have entrusted their affairs to him. The sense of rectitude he was taught to have has become his second nature, and he thinks that it would be wrong to abandon his freely chosen obligations to his family and clients. He realizes that if he entered politics he would have to go against the second nature to which he was true for many years and cultivate the unformed self of his heady college days that he had left behind when he chose law over politics. Yet he is also conscious of his growing dissatisfactions. He wants to do what is best for his country, family, clients, and himself, but he is deeply ambivalent about what that is.

The source of his ambivalence is his divided self. His present lawyerly self prompts him to act one way and his once and future political self, which would have to be more deliberately formed were he to have it, prompts him to act in another way. He is ambivalent, because he does not know whether he should be guided by his present self or by the future self he might develop. He wants to do what is best, but cannot decide whether it is better to remain as he is or better to try to become what he might be. He sees that his present self is good enough, but he also sees that the good enough is not enough. I now leave him temporarily in mid-thought.

The lawyer is meant to illustrate a predicament that is absent only from very few lives. All of us have a present self and a conception of a future self that we think of as better. Virtually everyone wants to have a better self, because most of us are aware of some of our deficiencies and are at least sometimes dissatisfied with how we are. Our commitments may be those of a present self or of a better future self. We do not know whether the key to overcoming our present dissatisfactions is to live and act in closer conformity to the commitments of our present self or according to the commitments that would lead to a better future self. Thus we are ambivalent. However we try to resolve our ambivalence, we will have to violate our own commitments and feel badly about ourselves. And the stronger the violated commitments are, the worse we feel. This is a painful emotion that we all want to avoid. So we are ambivalent, like the lawyer.

This kind of ambivalence is widespread but not universal. Saints and monsters are not ambivalent, because they are living and acting as they think they should, and their present self is as good as they can make it. Nor are people living in miserable circumstances ambivalent. They are concerned with staying alive, coping with adverse conditions, and have neither time nor energy to reflect on whether they should transform their present self into a better future one. Ambivalence is also absent from the lives of those who live in highly structured hierarchical societies, born into particular positions and relationships, and how they should live has been settled at birth. Few of us, living in contemporary Western societies, are saintly or monstrous; our conditions are rarely miserable; and our societies are pluralistic, democratic, and characterized by much mobility. For us, ambivalence is a typical state. The problem is how we deal with it.

The Kantian Ideal of Autonomy

One of the shibboleths of the present age is that we should overcome our ambivalence by becoming more autonomous. According to the influential Kantian interpretation, autonomy is a requirement of both practical reason and morality. To be autonomous is to act as the categorical imperative dictates.Footnote ¹ In my opinion, the Kantian ideal is mistaken, autonomy cannot overcome ambivalence, and the categorical imperative to pursue it fails to appreciate the depth of the predicament created by our ambivalence. I will proceed by considering an explicit defense of the Kantian ideal by one of the most respected contemporary Kantians, Christine Korsgaard.Footnote ² I acknowledge that there are non-Kantian interpretations of autonomy, but I will not discuss them.

Korsgaard thinks that ‘there is only one principle of practical reason, the categorical imperative viewed as the law of autonomy, but there are different ways to fall away from autonomy and the different principles of practical reason really instruct us not to fall away from our autonomy in these different ways’ (71–2).Footnote ³ ‘My view is,’ she writes, ‘that action is self-constitution … [and] the essential characteristic of an agent are efficacy and autonomy. … An agent is autonomous when her movements are in some clear sense self-determined. … These two properties, efficacy and autonomy, correspond to Kant's two imperatives, hypothetical and categorical’ (82–3). Following Kant, Korsgaard thinks that only autonomous actions are practically rational, morally good, and self-constitutive. Non-autonomous actions fail on all three of these counts.

Korsgaard would say that the lawyer's trouble was that he did not act autonomously. Contrary to appearances, he did not have reasons for his actions and he failed to do what is morally right. He was not even an agent, because ‘it is essential to the concept of agency that an agent be unified … to regard some movement … as my action, I must see it as an expression of my self as a whole … it must result from my entire nature working as an integrated whole’ (18–19). Korsgaard says to someone like the ambivalent lawyer: ‘stop dithering and bring your deliberation to an end: Make up your mind, or even better, Pull yourself together’ (126). It is a mistake, in her view, to think that the lawyer was reasonable and morally right when he was affectionate and attentive to his family, faithfully served the interests of his clients, governed his actions by unfailing rectitude, and agonized over whether he ought to enter politics. He was divided between his present fortuitous and possible future deliberate selves, so he could not act autonomously, could not be guided by reasons, and could not act morally.

What, then, is it to act autonomously? Korsgaard's answer is that it is to act on the basis of deliberation. ‘To deliberate is to formulate a maxim, stating the complete package of considerations that together favor the performance of a certain action. … Your maxim, once formulated, embodies your proposed reason. You then test it by the categorical imperative, that is, you ask whether you can will it to be a universal law, in order to see whether it really is a reason. Universalizability is a condition on the form of reason, and if a consideration doesn't meet this condition, then it is not merely outweighed – rather, it is not a reason at all,’ (51) Furthermore, ‘the work of practical deliberation is reunification, reconstitution: and the function of the principles that govern deliberation – the principles of practical reason – is the unification of the self … the function of practical reason is to unify us into agents who can be the authors of our actions.’ (126)

I have quoted these passages at length, because readers may find it incredible to suppose that Korsgaard really says that without a unified self and autonomy we cannot be human agents, cannot have reasons for what we do, and cannot do what is right. But that is what she says. I do not see how the conclusion could be avoided that what she says is utterly unrealistic.

Lack of Realism

To begin with, it follows from Korsgaard's Kantian ideal that we can be said to act reasonably and do what is morally right when we keep an appointment, apologize for a tactless remark, pay the bill for what we bought, or signal before we turn the car we drive only if we have met the following conditions. We have deliberated, formed a maxim, stated to ourselves the complete package of relevant considerations, and tested the maxim by the categorical imperative, asking whether we would wish the maxim to become a universal law. If we do not meet these conditions, says Korsgaard, we fail to follow rational principles because ‘rational principles [are] the principles that govern self-conscious self-determination’ (106); we are self-determining beings only if ‘we choose our own maxims, the content of our principles. And the categorical and hypothetical imperatives are rules for doing this – rules for the construction of maxims’ (108). Moreover, we also fail to act as human beings, because ‘a human being acts when she self-consciously determines herself to be the cause of a change in the world’ (127). Human action ‘requires autonomy and autonomy requires universalizability’ (180). Korsgaard's claim is that unless we self-consciously go through this mind-bogglingly complicated process, we can have no reason for action, we cannot do what is morally right, and, in fact, our actions cannot even qualify as human, since they are mere movements of the kinds animals perform.

Speaking for myself, I act from time to time, have reasons for my actions, and at least sometimes do what is morally right. Yet I have never in my life gone through the process Korsgaard claims is necessary for action, for having reasons, and for doing what is right. Nor do I think that I am unique or deceiving myself. I think that what is true of me in this respect is true of the overwhelming majority of people, including the lawyer, who do not cease to be human beings on account of having fallen short of Korsgaard's requirements. In fact, I can think of no one from my personal experience, or from reading literature, history, and sociology who meets the requirements Korsgaard says are necessary for being human, having reasons, and doing what is right.

The lack of realism in these requirements cannot be avoided by supposing that Korsgaard must mean that for reasonable and morally committed people meeting the requirements she insists on has become habitual, so they do not have to go through every time they act the necessary steps Korsgaard says are required. This would be a less unreasonable view than Korsgaard's, but it is not hers. She repeatedly emphasizes that autonomy consists in self-conscious deliberation every time we act autonomously. If I act out of habit and without deliberation, I am not acting autonomously. If I act autonomously ‘I must see it as an expression of my self as a whole’ (18), autonomy involves acting according ‘principles that govern self-conscious self-determination’ (106), ‘to be autonomous or self-determined is to choose the principles that are definitive of your will’ (108), and ‘a human being acts when she self-consciously determines herself to be the cause of change in the world’ (127).

The lack of realism in this view is even more glaring if we realize that autonomy as a moral ideal has emerged gradually, beginning in sixteenth and seventeenth century Europe. It was ‘only from about the early eighteenth century that the effort to create a theory of morality as self-governance became conscious.’Footnote ⁴ Before this time in Europe and in the rest of the world, the dominant moral ideal was obedience to divine or secular authority. It was taken for granted that people acted rightly or wrongly depending on whether they obeyed the reigning authority. It follows from Korsgaard's position that only in the eighteenth century and only in Europe have proto-humans become fully-fledged human beings who replaced traditional authorities with their own and thereby began acting on reasons and doing what was morally right. Since, in Korsgaard's view, living and acting with self-conscious deliberation, autonomously, and according to the categorical imperative is a universal requirement of practical reason and morality, no one before the emergence of autonomy was either reasonable or moral. Only someone blinded by a theory can believe this.

Let us suppose, however, for the sake of argument that the lawyer resolves to conform to Korsgaard's unrealistic requirements. He states to himself ‘the complete package of considerations’ he needs to deliberate about in order ‘to formulate a maxim’ that ‘embodies [his] proposed reason’, then ‘test[s] it by the categorical imperative … in order to see whether it really is a reason’, accepting that it will be a reason only if he can ‘will it to be a universal law’. If he cannot will it, then his supposed reason ‘is not a reason at all’ (51). There are two, to my mind decisive, reasons why he cannot do what Korsgaard says he should.

First, being prudent and intelligent, the lawyer has already thought a great deal about the complete package of considerations. His problem is that his thinking has yielded only uncertainty about which of two incompatible courses of action he should follow. He still does not know whether he should continue with law or abandon it for politics. Both courses of action are open to him and the complete package of considerations allow him to opt for either. That is why he is ambivalent. Korsgaard would tell him that ‘the function of the principles that govern deliberation – the principles of practical reason – is the unification of the self’ (126). Now the lawyer would certainly like to unify his self and reach a decision, but he does not know how to do that, because his self is divided and he is ambivalent. If he knew how to change that, he would not have a problem. He would resolve to live either according to his fortuitous or deliberate self. His self would already be unified, he would not be ambivalent, and he would not need to struggle with formulating a maxim, and testing it by the categorical imperative. Korsgaard's advice that he should do what he has again and again tried and failed to do is as useless as advising the sick to be healthy, the weak to be strong, or the stupid to be intelligent. And if the lawyer, contrary to the facts, had a unified self and was not ambivalent, then he would not need to do what Korsgaard advises him to do. Consequently the whole rigmarole Korsgaard insists on going through as the key to following the unrealistic requirements of reason and morality is either useless or unnecessary.

Secondly, suppose the lawyer unifies his self, overcomes his ambivalence, and decides to continue with law. Does this mean that if his decision was based on deliberation and followed the requirements of practical reason and morality, then he had to have given himself a universalizable maxim? What would this maxim be? It cannot just be that whenever anyone faces a choice between continuing to practice law and entering politics, he should choose the former. For if a lawyer was rich and had no need of earning a living, or had no family, or found the legal system utterly corrupt, or was married to a wife who earned a great deal of money, or was unexpectedly appointed to replace a dead legislator, or was given six months to live, or was psychoanalyzed and found that his ambivalence was the product of an overactive superego, or inherited a fortune from a distant relative who died intestate, and so on and on endlessly, then he may well have made a different decision.

The universalizable maxim cannot be formulated, because any maxim that might be proposed would have to be qualified again and again to rule out all conceivable exceptions, otherwise it would not be universalizable. The maxim would sink under the weight of a thousand qualifications. But if it did not sink, if it could actually be precise enough to resolve the lawyer's ambivalence, while barring exceptions, then it would have to be so precise as to apply only to the lawyer in his particular circumstances. It could, then, be said that reason and morality require anyone in those precise circumstances to act on that maxim, but if that precise maxim was as precise as it would have to be to rule out exceptions, it would obligate only the lawyer's actions, because the precise circumstances of others would be different from the lawyer's. It is self-refuting to claim that a maxim that applies only to one person is universalizable.

I conclude that Korsgaard's defense of autonomy as a requirement of practical reason and morality fails. My objections to her defense were not meant to deny that autonomy is good to seek or to have. Rather, they were meant to show that something has gone very wrong with the extravagant claims Korsgaard makes for what she understands as autonomy. Autonomy, in her sense, is not the key to coping with the routinely ambivalent commitments prompted by our present and future selves.

Groundless Optimism

I think that one main problem with Korsgaard's approach in particular and the general approach of which hers is an example is that they are pervaded by an unrealistic optimism. A telltale sign of it is that they attempt to cope with our ambivalence by way of constructing an ideal theory, which, they optimistically assume, will inspire people to follow it.

Korsgaard's account of self-constitution aims to be such an ideal theory. She says that in formulating it, she is using ‘Rawls' theory for a model’ and that the ‘ideal theory is worked out under certain assumptions’. Some of them are that ‘everyone will act justly … that historical, economic, and natural conditions are such that realization of the ideal is feasible,’ and she quotes Rawls in assuming ‘that everyone is “rational and able to manage their own affairs”.’ She realizes, of course, that actual conditions are not ideal, but the point of formulating an ideal theory, as Rawls says of his version, is ‘to work out our ideal view of justice on the assumption that people, nature, and history will behave themselves so that the ideal can be realized, and then to determine – in the light of that ideal – what is to be in actual circumstances.’Footnote ⁵

This raises two obvious questions: why suppose that if we had an ideal theory, we would follow it? and if our present conditions are recognized as less than ideal, then why start with an ideal theory? I will discuss Rawls' answer to the first question now and his answer to the second in the next section.

In answer to the first question, Rawls says that ‘the ideal theory is fundamental,’Footnote ⁶ because it ‘presents a conception of a just society that we are to achieve if we can.’Footnote ⁷ And we can, because ‘the sense of justice is continuous with the love of mankind’ and ‘men's propensity to injustice is not a permanent aspect of community life; it is greater or less depending in large part on social institutions.’Footnote ⁸

Underlying the claim that ‘men's propensity to injustice is not a permanent aspect of community life’ is the assumption, widely shared by optimists, that human beings are basically good. Hume writes of ‘moral distinctions deriv'd from so noble a source’ as ‘the generosity and capacity of human nature. It requires but very little knowledge of human affairs to perceive, that a sense of morals is a principle inherent in the soul, and one of the most powerful that enters into the composition. But this sense must certainly acquire new force, when reflecting on itself, it approves of those principles, from whence it is deriv'd, and finds nothing but what is great and good in its rise and origin. … So that nothing is presented on any side, but what is laudable and good.’Footnote ⁹ According to Rousseau: ‘man is naturally good; I believe I have demonstrated it’ and ‘the fundamental principle of all morality, about which I have reasoned in all my works … is that man is a naturally good creature, who loves justice and order; that there is no original perversity in the human heart.’Footnote ¹⁰ Kant says virtually the same: man is ‘not basically corrupt (even as regards his original predisposition to good), but rather … still capable of improvement’ and ‘man (even the most wicked) does not, under any maxim whatsoever, repudiate the moral law. … The law, rather, forces itself upon him irresistibly by virtue of his moral predisposition.’Footnote ¹¹ John Stuart Mill speaks of the ‘leading department of our nature … this powerful natural sentiment … the social feeling of mankind – the desire to be in unity with our fellow creatures, which is already a powerful principle in human nature, and happily one of those which tend to become stronger, even without inculcation.’Footnote ¹² Friedrich Hayek writes: ‘It is … no argument against individual freedom that it is frequently abused. Freedom necessarily means that many things will be done which we do not like. Our faith in freedom does not rest on the foreseeable results in particular circumstances but on the belief that it will, on balance, release more forces for the good than for the bad.’Footnote ¹³ Philippa Foot treats ‘moral evil as “a kind of natural defect”,’ claims that ‘acting morally is part of practical rationality’, and that ‘no one can act with full practical rationality in pursuit of a bad end’.Footnote ¹⁴ And this is also the assumption that underlies Rawls' and Korsgaard's commitment to an ideal theory.

I begin with two finely expressed dissenting views – views I share. Stuart Hampshire writes of ‘the innocence in moral philosophy’ that proceeds ‘as if it was sufficient to establish some truths about the great goods for mankind, and then to deduce from these truths the necessary human virtues and vices and the necessary social policies. It is not sufficient’,Footnote ¹⁵ because, as he writes elsewhere, ‘the unchanged horrors of human life, the savage and obvious evils … scarcely vary from culture to culture or from age to age: massacre, starvation, imprisonment, torture, death and mutilation in war, tyranny and humiliation – in fact, the evening and the morning news. Whatever divergences in conceptions of the good, these primary evils stay constant and undeniable as evils to be at all costs averted, or almost all costs.’Footnote ¹⁶

The other view is Bernard Williams’ who writes that ‘there are areas of philosophy which might be supposed to have a special commitment to not forgetting or lying about the horrors, among them moral philosophy. No one with sense asks it to think about them all the time, but, in addressing what it claims to be our most serious concerns, it would do better if it did not make them disappear. Yet this is what in almost all its modern forms moral philosophy effectively does.’Footnote ¹⁷

I will not dwell on the horrors of which Hampshire and Williams rightly remind us, but merely mention the mass murders of falsely persecuted innocent men, women, and children that have been committed in Stalin's Russia, Hitler's Germany, Mao's China, as well as in Afghanistan, Argentina, Bosnia, Cambodia, Chile, the Congo, India, Iran, Iraq, Rwanda, Sudan, Turkey, and Uganda. As historians write, ‘the hundred years after 1900 were without question the bloodiest in modern history, far more violent in relative as well as absolute terms than any previous era. … By any measure, the Second World War was the greatest man-made catastrophe of all time.’Footnote ¹⁸ ‘There is no question that the twentieth century witnessed a mind-bogglingly large number of deaths by organized violence. Estimates for the total number … range from 167 million to 188 million.’Footnote ¹⁹ ‘Archeological, anthropological, as well as all surviving documentary evidence indicates that war, armed conflict between organized political groups, has been the universal norm in human history.’Footnote ²⁰

Moved by an amazingly unrealistic optimism, Rawls and Korsgaard's proceed in the face of these horrors to suppose that the ideal theory they propose will motivate people to live and act according to it. How could people as thoughtful and well-informed as Rawls and Korsgaard obviously are suppose that if Nazi and Communist mass murderers, Mao's rampaging Red Guards, various terrorists, and fanatical perpetrators of tribal, religious, and ethnic atrocities were given access to an ideal theory, then they would desist from what they are doing and live and act morally just as the ideal theory says they should? Even the lawyer, who certainly was not an evil person, would not be helped by trying to do what Korsgaard's ideal theory says he should, so it is by many multiples more unrealistic to suppose that mass murderers would be turned into decent people if they were given the ideal theory. Yet it is only by making that assumption that Rawls and Korsgaard could reasonably claim that the ideal theory will motivate people to live and act according to it.

It may be thought that Rawls and Korsgaard could combine the acknowledgment of the horrors with a denial of their relevance to ideal theories. The horrors, they might say, are among the facts that make prevailing conditions non-ideal. But it is not irrelevant to ideal theories why conditions are non-ideal. Rawls says that they are non-ideal, because, although, ‘men's propensity to injustice is not a permanent aspect of community life; it is greater or less depending in large part on social institutions, and in particular on whether these are just or unjust.’Footnote ²¹ And Korsgaard says that ‘bad actions, actions that are contrary to justice and the categorical imperative, then, are defective actions’ and ‘badness is a kind of heteronomy. The bad person is determined from outside, for he is a conduit of forces working in him and through him’ (161). But neither Rawls nor Korsgaard asks the question that must be asked and to which they owe an answer: why are social institutions and conditions non-ideal?

Defective social institutions and conditions are made and maintained by us. If they are defective, it is because we, who make and maintain them, make mistakes. It may be that we make mistakes, because we have been corrupted by defective institutions and conditions, but sooner or later this explanation comes to an end. It must eventually be recognized that we are the causes, and the institutions and the conditions we make and maintain are the effects. To try to explain our mistakes by the defects of institutions and conditions is to try to explain causes by their effects. The conclusion is inescapable that conditions are non-ideal, because we are non-ideal. And if we are non-ideal, then what prevents us from accepting an ideal theory are not non-ideal conditions, but our non-ideal selves. The optimistic illusion of Rawls, Korsgaard, and other ideal theorists is that our non-ideal selves will follow an ideal theory, unless non-ideal conditions corrupt us. But we make non-ideal conditions non-ideal, and that is the unpleasant fact that optimists refuse to face.

It may be thought that Rawls is not committed to an unrealistic optimism about human motivation, because his ideal theory is about what the basic structure of a just society ought to be, not about human motivation. It is hard to see, however, what could be the point of an ideal theory about society if it was not assumed that human beings would be motivated to live and act according to it. But I will not pursue this point, because I do not want to get bogged down in interpreting Rawls. Whatever the truth may be about Rawls's ideal theory resting on on an unrealistic optimism, there can be no doubt that Korsgaard's ideal theory rests on such optimism. For her ideal theory is about how human beings ought to be motivated, and her theory would be pointless unless she assumed that most people who understood her ideal theory would be motivated to live and act according to it. She is explicit in making this assumption: ‘The moral law is the law of self-constitution, and as such, it is a constitutive principle of human life itself’ (xiii) and ‘the normativity of obligation is a form of necessity … its manifestation … [is] a psychological force’ (3).

This is not to say, of course, that Korsgaard's untenable optimism should be replaced by pessimism that supposes that we are basically bad, rather than good. Pessimism would be just as untenable as optimism is. The reasonable view is that we are ambivalent. We are forever divided between our present and future selves, just as the lawyer was. There is no guarantee that a future self would be more rational or more moral than our present self. The human condition excludes such guarantees. The optimists suppose that if only we thought harder and deliberated better, we could transform our faulty present self into a much improved future self. But they do not see that the very same ambivalence that they advocate overcoming would permeate and handicap our efforts to think harder and deliberate better. They are trying to formulate ideal theories in non-ideal conditions and the resulting theories could not escape the non-ideal conditions that they are meant to correct. Their proposed cure is unavoidably infected by the disease they intend to cure.

The key to becoming more rational and moral cannot be to become more autonomous, because, as Korsgaard clearly recognizes, although Kant does not, ‘self-conception is, in short, the root of both moral good and evil’.Footnote ²² The key to personal understanding, I conclude, is to understand how a person struggles with the particular form of the ubiquitously human ambivalence created by the conflict between his fortuitous and deliberate selves.

Misguided Theorizing

The second question about ideal theories I posed earlier is: why should we begin with an ideal theory since our conditions are non-ideal? Why should we think that having a theory about what the best conditions are would help us to cope with the problems that we have in the prevailing conditions that fall far short of being the best? Rawls' answer is that ‘the reason for beginning with ideal theory is that it provides, I believe, the only basis for the systematic grasp of … more pressing problems’.Footnote ²³

Now, this is an extraordinary belief to hold, since many approaches to how we should live and act begin with the ‘more pressing problems’ that arise in everyday life, not with ideal theories. This is where Aristotle, Pyrrhonian skeptics, Epicureans, Machiavelli, Montaigne, Locke, Hume, and, closer to our own time, Ross, Oakeshott, Berlin, Hampshire, and Williams, among many others, begin. Rawls was familiar with the works of these people, so it needs to be explained why he would claim that the only basis for coping with pressing problems is an ideal theory. Why do ideal theorists consider only their own and perhaps competing ideal theories, but not radically different approaches? They do so, I believe, because they assume that a way of coping with pressing problems can be reasonable only if it is universalizable. Ideal theorists tend to discuss only the contractarian, consequentialist, and Kantian theories that proceed on this assumption, because only they merit serious consideration. Since approaches that do not accept the assumption cannot, in their view, be reasonable, they are ignored. As Korsgaard puts it, ‘universalizability is a condition on the form of a reason, and if a consideration doesn't meet this condition … it is not a reason at all’ (51). (I note in passing that Korsgaard does consider Aristotle in some detail, but only because, ‘flying in the face of received opinion’ (14), she interprets him as a precocious Kantian.)

There are, I think, two ways of contesting the claim that we must begin with an ideal theory in order to cope with pressing problems. One is to show that the assumption on which it rests, namely, that reasons must be universalizable, is mistaken. I will say nothing about it here. The other is to show that it is possible to cope reasonably with pressing problems without beginning with or in any way relying on an ideal theory. And this is what I will now do by considering how the lawyer could cope reasonably with the pressing problem of his ambivalence.

Korsgaard is right in this: the lawyer should deliberate further. But she is wrong to suppose that his deliberation ought to abstract more and more from his particular circumstances and bring him ever closer to formulating a universalizable maxim prescribing what anyone in his situation ought to do. The more he abstracts, the more he tries to formulate a universalizable maxim, the less relevant his deliberation becomes to himself. Contrary to Korsgaard, Kantians, and ideal theorists, his deliberation ought to become more and more personal. It ought to lead him to consider more critically than he has hitherto done what has led to his ambivalence. He should redirect his deliberation from considering which of his alternatives he should opt for to asking about the source of his ambivalence. Perhaps if he examined the conflicting beliefs, emotions, and motives that have led to his ambivalence, he would find that some of them are unreasonable, and then he could cope with his ambivalence by dissolving it.

Perhaps he should talk to his wife and children to find out whether his continued practice of law is really as important for them as he is supposing. Perhaps his wife is anxious to leave housewifery behind and make a career for herself. Perhaps his children are aching to leave home and become independent. Perhaps they love him deeply enough to accept less comfort and security if it enables him to do what he really wants. Perhaps he has exaggerated the extent to which their well-being depends on him. Perhaps a career in politics would merely replace the drudgery of his law practice with that of a low-ranking junior politician. Perhaps his lifelong rectitude would be outraged by the inevitable compromises, give-and-take, and self-serving negotiations that are unavoidably involved in political life. Perhaps what he imagines to be his boredom with his law practice is merely a familiar manifestation of a banal mid-life crisis that assails some men, much as menopause assails some women. Perhaps what is happening to him is that he is unwilling to accept that his adolescent dreams were no more than dreams, and he should leave childish things behind.

Deliberating in this way is a matter of testing the realism of his beliefs, emotions, and motives. It is to entertain seriously the possibility that they reflect illusions that sustain a false picture of himself. The aim of this kind of deliberation is not to formulate a universalizable maxim, but to find out whether the beliefs, emotions, and motives that focus on how we are and want to be are realistic, or whether both or either of our fortuitous and deliberate selves are wholly or partly formed of illusions that prevent us from seeing the facts as they are.

In general, deliberations about coping with the unavoidable ambivalence created by the conflict between our present and future selves can proceed in two directions. One is toward the future, and it is about considering which course of action would lead to wherever it is we want to go. The other is toward the past, and it involves considering what has led to our present ambivalence. And we may find that the sources of our ambivalence are unrealistic beliefs, emotions, and motives that we hold, because we are reluctant to question the illusions we nurture about ourselves. I have presented the lawyer so far as deliberating in the first way, looking toward the future, and I am suggesting now that he could make progress toward coping with his ambivalence if he reflected also in the second way, looking toward the past.

It may be said, of course, that it is a universalizable maxim that we should all deliberate in both ways about the realism of the beliefs, emotions, and motives that concern how we are and want to be. But this would be no better than sophistry, because it disguises the fact that if we follow the maxim, then we are doing something that no one else can do. It is an unavoidably personal deliberation about the illusions we might hold, about the realism of our evaluations of the facts about ourselves, and about whether we respond ‘to the right person, to the right extent, at the right time, with the right aim, and in the right way’. As Aristotle saw, ‘that is not for everyone, nor is it easy’.Footnote ²⁴ I would merely add that it is not merely not for everyone, but that is only for the person who is deliberating. The direction of this deliberation is deeper into ourselves, but its aim is not the solipsistic digging ever deeper in our soul, but the prudent and practical effort to eliminate illusions about ourselves that prevent us from responding to our circumstances realistically. The deliberation involved in getting from our present self to a future one, therefore, is to a considerable extent a matter of overcoming illusions. An essential part of it is to understand how successful a person is in overcoming the inevitable ambivalence created by the often incompatible and ill-formed attitudes prompted by his present and future selves.

The alternative, then, to beginning with an ideal theory that, according to Rawls, is ‘the only basis for the systematic grasp of … more pressing problems’Footnote ²⁵ is the rather obvious one of beginning with the more pressing problems themselves and trying to find a way of coping with them. Coping consists in making the problems less pressing, not on acting according to a theory that is made ideal by being universally applicable regardless of differences among persons, circumstances, illusions, and problems. Many and perhaps the most pressing problems cannot be made less pressing by an ideal theory that aims to cope with them by formulating a universal maxim, because the problems are personal and particular, and they are often caused by the illusions of the person whose problems they are.

Implications

Explaining actions would be much simpler if all it required was to identify what we want and explain our actions as aiming at satisfying it. Complications enter because we are often unclear about what we want, we want incompatible things, and may be wrong in supposing that what we want would be satisfying. I suggested that one source of many of these complications are the conflicts between our present and future selves that make us ambivalent. This much is uncontroversial, I think. Basic disagreements emerge, however, between those who share the prevailing consensus that the key to coping with our conflicting and ambivalent selves is to pursue the ideal of autonomy, and those who think that the consensus as mistaken.

According to defenders of the ideal of autonomy, the key to overcoming the conflicts and resolving the ambivalence depends on abstracting from personal and particular considerations and evaluating lives and actions, our own or other people's, on the basis of their conformity to a law that reason and morality require everyone to follow regardless of individual differences. They suppose that following this law is a universal requirement, and its violation is contrary to both reason and morality. I have rejected the ideal of autonomy on the grounds that it cannot be reasonable to require conformity to a universal law, because we differ greatly in genetic inheritance, upbringing, education, and experiences, in the possibilities we have, the commitments we make to realizing them, in our conceptions of well-being, and in the circumstances in which we live. And even if we could do what is surely impossible, namely to free ourselves, from all personal and particular considerations and all the influences that have formed us, and live and act according to a putatively universal law, there is still no guarantee that we would do so in conformity to morality. If we were autonomous, in complete control of our lives and action, we may or may not act reasonably and morally depending on what our present and/or future selves prompt us to do. And if we were non-autonomous and lived in unquestioned obedience to some religious or secular authority, we may still live and act reasonably and morally if the authorities are reasonable and moral. Autonomy is neither necessary nor sufficient for living and acting reasonably and morally.

This is not to say, of course, that autonomy is bound to be unreasonable and immoral. It may or may not be. But, however it turns out, autonomy is merely a local conception of well-being held by many people – I might as well admit, myself included – who live in contemporary Western societies, in tolerably civilized circumstances, in adequate comfort and security, having sufficient leisure, education, and inclination for habitual deliberation about how to live and act. It gives a bad name to reason and morality to demand in their name that everyone must either live in this way or fail to be reasonable and moral. What I have rejected are extravagant claims made on behalf of autonomy. It is merely one fallible conception of well-being among many others. Its pursuit is as liable to go wrong because of our conflicts and ambivalence as the pursuit of other conceptions of well-being.

But if the universality of autonomy is rejected, if it is acknowledged that its pursuit may lead us astray because of our conflicting present and future selves and the resulting ambivalence, then how can we guard against going wrong? We can do so by doing what we can to overcome the unrealistic beliefs, emotions, and motives by which we nurture comforting illusions about ourselves. We cannot be certain that we will succeed. But we can be certain that if we do not try, we will suffer the debilitating consequences of our conflicted and ambivalent selves.Footnote ²⁶