INTRODUCTION IN this book I offer answers to three questions. What is the best way to understand "well-being'? Can it be measured? Where does it fit in moral and political thought? The third question is huge, and my answer is incomplete. I concentrate on issues that make up the subject of justice (a glance at the Table of Contents will show which I take them to be). Together they also make up much of moral and political philosophy, but not all of it. What I shall say about them has implications for the rest of moral and political philosphy, but to work out those implications, to add to them, and to generalize enough to answer my third question fully would take vastly more work than I have done here. My three questions correspond to the three parts of the book. None of the questions has priority. We cannot just ask, What is the best account of well-being?, as if "best' could mean "most accurate'. Our job is not to describe an idea already in existence independently of our search. Before we can properly explain well-being, we have to know the context in which it is to appear and the work it needs to do there. It may be that different notions of well-being are needed in different theore- tical contexts. Nor can we first fix on the best account of "well-being' and independently ask about its measurement. One proper ground for choosing between conceptions of well-being would be that one lends itself to the deliberation that we must do and another does not. I should say something about how I see argument in normative ethics proceeding. To justify my views on that subject would take another book, but I should at least explain them. The most common method in normative ethics is piecemeal appeal to intuition. "It follows from what you say that it would be all right to do such and such, but that's counter-intuitive, so you're wrong.' Piecemeal appeal to intuition has been severely attacked, and the attacks seem to me fatal. Intuitions as a class have no probative force. Still, there are one or two intuitions that are solider than anything 2 INTRODUCTION else that moral thought is likely to come up with---say, that battering babies is wrong. However, the few moral beliefs of this solidity do not get us far; they weed out only the wildest of moral theories. There is also the problem of distinguishing the solid from the only apparently solid. The role that intuitions can play in moral philosophy is the role that we are content to let them play in other departments of thought (it is only in moral philosophy that they have risen so far above their epistemological station). In mathemat- ics, the natural sciences, and other branches of philosophy, finding a conclusion intuitively repugnant does not close an argument; it is a reason to start looking for a good argument. I think that there are two tests for good argument in normative ethics. One is the test of completeness. We all wish to provide answers to roughly the same rather varied questions about what to do, and our answers will have to have a fair degree of complexity and we shall have to know how the various parts of what we say work together to produce the answers. In this minimal sense every moral philosopher has to develop a theory. It may be thought that introducing the word "theory' already begins begging questions. It may look as if it suggests one particular, contentious conception of what it is that we need to test and also, within certain limits, what suitable tests for it will be: namely, that morality consists of a system of highly abstract principles and that it is rather like a natural science. But I am using the word "theory' far more broadly than that---broadly enough to be acceptable to some, though perhaps not quite all, philosophers who think that morality does not constitute a "theory', on some narrower understanding of the word. Having to produce a moral theory in my broad sense helps to guard against at least some errors. It forces an individual to find answers to moral questions that are both coherent and complete. For instance, contractualists, as I shall be arguing later, respond to the counter-intuitive permissiveness of utilitarianism by adopting principles that, seen against that permissiveness, look very attractive indeed, but then, when shifted on to different ground, begin to look counter-intuitively strict. Indeed, I suspect that, one way or another, quite a few moral theories that we now regard as 3 INTRODUCTION plausible contenders will fail the test of completeness. It is a test of no small power. Still, it is not powerful enough. All that coherence and completeness in the end require is that we come up with an answer to any moral question that presents itself and that we do not encounter conflicts of belief in the course of doing so. They do not require that the beliefs be appropriate or correct or reliable or true or whatever the standard is that one wants to appeal to in order to mark the difference---if it can be marked in the moral domain---between what is merely believed and what ought to be believed. What we need, if we are to get beyond completeness, is to find an exit from the circle of beliefs. If we find one, it would supply the second and more important test of good argument, the test of correctness. There are various possible exits. We might think that a form of foundationalism is possible: we might, for instance, think that substantive moral beliefs could be derived from formal features of rationality or from the logic of key moral terms, although neither of these might be enough on its own to constitute an exit. We might become convinced by a form of naturalism: for example, we might decide to clarify the not entirely clear questions of prudence and morality by turning them into questions answerable solely by appeal to empirical facts and logic. We might accept constructivism: we might accept that some notion of what rational persons can be brought to agree upon constitutes what we mean, and all that we can mean, by "correctness' in normative ethics. Or we might accept realism: we might decide that, when we speak about prudential and perhaps also moral values, we are speaking, without reduction to non-value terms, about something in the world. The way to submit a moral theory to the test of completeness is to spread the theory as widely as possible, especially into areas where the chances are best of its running into trouble. The way to submit it to the test of correctness is harder to decide. We should have to decide which exit, if any, is to be taken. But even without settling that question---which is the question that needs a book to itself---we can do something. A good place to start on the search for standards of correctness in prudential and moral judgement is with 4 INTRODUCTION developing as rich a substantive account of prudence and morality as one can. (I am using "prudence' here in the philosopher's especially broad sense, in which it has to do not just with a due concern for one's future but with everything that bears on one's self-interest.) One cannot hope to find a possible route to standards of correctness until substantive prudential and moral theories take on far more detail than they have yet had. For instance, how can we decide on the role of reason and desire, how can we decide whether values are valuable because desired or desired because valuable, until we see in great detail what can be advanced in the way of argument for the existence of values? Some philosophers have in the past tried to derive substantive moral beliefs from the nature of rationality, or from the semantics of key moral terms, or from the need for moral restrictions to fit the human psyche. But if one believes, as I do, that none of this is quite enough to do the job, then one has to find another way to justify substantive moral beliefs. The most promising way, I believe, is to come to understand the nature of prudential and moral deliberation much more fully than we now do, and then to look in that fuller understanding, perhaps especially in an understanding of the relationship between prudence and morality, for the pieces out of which the standard of correct- ness, the exit from the circle of mere beliefs, is to be constructed. We ought not, I think, to treat meta-ethics as something that can be pursued for long independently of normative ethics. We cannot derive conclusions about the status of values from highly speculative and abstract meta- physical considerations that anyway need a decision about the nature of value to help prop up the abstract argument. What, to my mind, we need is a period in which philosophers build wide-ranging, detailed, substantive prudential and moral theories. That, I think, would help not only normative ethics but meta-ethics too. PART ONE MEANING PART TWO MEASUREMENT PART THREE MORAL IMPORTANCE I UTILITARIAN ACCOUNTS: STATE OF MIND OR STATE OF THE WORLD? HOW are we to understand "well-being'? As "utility', say the utilitarians, aware that this technical term itself needs explain- ing.1 What is "utility'? "Pleasure and the absence of pain', the classical utilitarians said,2 not realizing how much the words "pleasure' and "pain', especially in the stretched sense they attached to them, needed explaining. Two main traditions about "utility' have grown up. One sees it as a state of mind, the other as a state of the world. Is "utility' mental states (e.g. pleasure, pain) or states of the world which fulfil desires (e.g. economists' "preference')? If mental states, is it only one sort, or many? If many, what links them? If fulfilment of desires, desires as they happen to be, or in some way improved? If improved, how? We can forget morality for the moment. Utilitarians use our rough, everyday notion of "well-being', our notion of what it is for a single life to go well, in which morality may have a place but not the dominant one. This does not mean that our job is merely to describe the everyday use. It is too shadowy and incomplete for that; we still have to be ready for stipulation. 1.Mental state accounts When some utilitarians have spoken of mental states such as pleasure and pain, they have meant these terms so widely that their accounts get very near desire accounts of "utility'. So we cannot always take this verbal difference as marking any real difference. Still, the difference is often real enough. Bentham and Mill are, with ample reason, taken to be offering a psychological 8 MEANING account of "utility'. Pleasure or happiness is presented as a "state of feeling', and pain or unhappiness as a feeling on the same scale as, and the opposite of, pleasure or happiness. And the utilities of all our experiences are supposed to be determin- able by measuring the amount of this homogeneous mental state that they contain. The trouble with thinking of utility as one kind of mental state is that we cannot find any one state in all that we regard as having utility---eating, reading, working, creating, helping. What one mental state runs through them all in virtue of which we rank them as we do? Think of the following case. At the very end of his life, Freud, ill and in pain, refused drugs except aspirin. "I prefer', he said, "to think in torment than not to be able to think clearly'.3 But can we find a single feeling or mental state present in both of Freud's options in virtue of which he ranked them as he did? The truth seems, rather, that often we just rank options, period. Some preferences---Freud's seems to be one---are basic. That is, preferences do not always rest upon other judgments about the quantity of some homogeneous mental state found in, or produced by, each option. When, in these cases, one speaks of one thing's yielding greater satisfaction than another, this seems best understood as saying that having the first is the fulfilment of a greater desire than having the second would be. One wants the first more than the second. But these desires are not ranked by independ- ent quantities of satisfaction. So, if the mental state account takes this simple form, the objections to it are insurmountable. And if we do not want to go over to a desire account, there are two ways we might now move. We might accept that utility is not one mental state but many, and then look for an explanation of how they are linked. Or we might, on the other hand, decide that utility is neither a matter of mental states nor of desire-fulfilment but of something, in a way, in between; we might say that it is a matter of finding enjoyment in various things, where "enjoy- ment' is what we might call an attitude, or a conscious state, or a state of a person. I want to leave the second move until later; it is not easy to grasp, and it will be easier after we have looked at both the mental state and the desire accounts. So let us go back to the first move. 9 STATE OF MIND OR STATE OF THE WORLD? 2.Sidgwick's compromise Suppose we said that utility consisted of several different mental states. What then would make them into a set? The obvious candidate would be desire; we could say, following Henry Sidgwick in borrowing something from each of the competing accounts, that utility combines a psychological element and a preference element. "Utility', we could say, is "desirable consciousness', meaning by "desirable' either con- sciousness that we actually desire or consciousness that we would desire if we knew what it would be like to have it.4 The trouble with this eclectic account is that we do seem to desire things other than states of mind, even independently of the states of mind they produce. This is the point that Robert Nozick has forcefully made with some science fiction.5 Imagine an experience machine programmed to give you any experience you want; it will stimulate your brain so that you think you are living the most ideal life, while all the while you float in a tank with electrodes in your brain. Would you plug in? "What else can matter to us', Nozick asks, "other than how our lives feel from the inside?' And he replies, surely rightly, that we also want to do certain things, to be certain things, and to be receptive to what there is in life beyond what humans make. The point does not need science fiction; there are plenty of examples from ordinary life. I certainly want control over my own fate. Even if you convince me that, as my personal despot, you would produce more desirable consciousness for me than I do myself, I shall want to go on being my own master, at least so long as your record would not be much better than mine. I prefer, in important areas of my life, bitter truth to comfortable delusion. Even if I were surrounded by consummate actors able to give me sweet simulacra of love and affection, I should prefer the relatively bitter diet of their authentic reactions. And I should prefer it not because it would be morally better, or aesthetically better, or more noble, but because it would make for a better life for me to live. Perhaps some such preferences, looked at with a cold eye, will turn out to be of dubious rationality, but not all will. This fact presents a serious challenge to the eclectic account of utility. If not all desirable things are mental states, yet they 10 MEANING matter to our well-being, the eclectic account is fissile. Which part of it should we retain: desire or mental states? It is hard to retain mental states, for if we did, we should then have, puzzlingly, to accept that when, with eyes wide open, I prefer something not a mental state to a mental state and so seem to value the former more than the latter, I get greater utility from what I value less.6 Of the two, it is better to retain desire.7 Of course, "mental state' is a vague expression. Perhaps Sidgwick and others use it broadly enough to include, say, knowledge. However, that does not seem to be Sidgwick's intention, and in any case it would still not be broad enough. I also want to be my own master, and it would take more broadening to include that. It seems more promising to abandon "mental state' altogether and to try defining "utility' solely in terms of desire: utility consists, we might try saying, in the fulfilment of desire. 3.The actual-desire account The simplest form of desire account says that utility is the fulfilment of actual desires. It is an influential account. Economists have been drawn to it because actual desires are often revealed in choices and "revealed preferences' are observable and hence a respectable subject for empirical science.8 Also the same account of utility can then do service in both moral theory and theory of action; explanation of action has to appeal to what we in fact want rather than to such ideal notions as what we ought to want or would want if well-informed. And both philosophers and social scientists have been powerfully drawn to it because it leaves no room for paternalism; if actual desires determine distributions, con- sumers are sovereign and agents autonomous. Yet, notoriously, we mistake our own interests. It is depressingly common that when even some of our strongest and most central desires are fulfilled, we are no better, even worse, off. Since the notion we are after is the ordinary notion of "well-being', what must matter for utility will have to be, not persons' actual desires, but their desires in some way improved. The objection to the actual-desire account is overwhelming. 11 STATE OF MIND OR STATE OF THE WORLD? In any case, considerations of autonomy are, on reflection, no recommendation of it. Well-being and autonomy, no doubt, both matter morally. It is even likely that living autonomously would be part of any enlightened person's conception of a good life. But it just confuses two quite different ideas to adopt the actual-desire account of well-being just because it makes autonomy prominent. One consideration to keep in mind is that the question, "What is the best account of {"utility{'?', should be kept distinct from the question, "What is the account of {"utility{'---perhaps highly artificial and ad hoc---that yields a one principle, utility maximizing, moral theory that comes closest to adequacy?' It is wrong to try to build into the notion of "utility' all the restrictions that morality needs, if they fit more naturally elsewhere in the theory. 4.The informed-desire account At this point, an obvious move is to say that desires count towards utility only if "rational' or "informed'. "Utility', we might try saying, is the fulfilment of desires that persons would have if they appreciated the true nature of their objects. But we shall have to tone this definition down a bit. Although "utility' cannot be equated with actual desires, it will not do, either, simply to equate it with informed desires. It is doubtless true that if I fully appreciated the nature of all possible objects of desire, I should change much of what I wanted. But if I do not go through that daunting improvement, yet the objects of my potentially perfected desires are given to me, I might well not be glad to have them; the education, after all, may be necessary for my getting anything out of them. That is true, for instance, of acquired tastes; you would do me no favour by giving me caviar now, unless it is part of some well-conceived training for my palate. Utility must, it seems, be tied at least to desires that are actual when satisfied. (Even then we should have to stretch meanings here a bit: I might get something I find that I like but did not want before because I did not know about it, nor in a sense want now simply because I already have it; or I might, through being upset or confused, go on resisting something that, in some deep sense, I really want.) It 12 MEANING is hard to get the balance between actual and informed desires quite right. But, to be at all plausible, the informed-desire account has to be taken to hold them in a balance something like the one I have just sketched. The move to "informed-desires' marks the first important break with the classical utilitarian tradition (we shall see several more in the course of the discussion). Bentham and Mill used "utility' both to explain action and to set a moral standard; they used its empirical role in arguing for its moral role. But now "utility' has taken on a shape to fit it for a normative role (it need not be only in moral theory; it could also be in an account of one person's well-being or an account of practical reason), and it is of doubtful relevance to a purely empirical account of motivation. So this account of "utility' should no longer be seen to be attached, except historically, to certain theories of action. It is not committed to the view that action is the result purely of a vector of desire-forces. It is not committed to any Hume-like account of the role of reason and desire. We can no longer use historical connections as a guide to theoretical connections. The informed-desire account starts with the recognition that actual desires can be faulty. What sort of faults matter? Obviously, for one, lack of information. Some of our strongest desires rest on mistakes of fact. I make my fortune, say, only to discover I am no better off because I was after people's respect all along and mistakenly thought that making a fortune would command respect. Or I want an operation to restore me to health, not realizing that some pill will do just as well. What matters is the ultimate, not the immediate, object of my desire, and factual mistakes creep into matching the one to the other.9 Or I develop one set of material desires not realizing that they are the sort that, once satisfied, are replaced by another set that are just as clamorous and I am no better off. The consumer-desires at the centre of the economists' stage can be like that.10 Then another relevant fault is logical mistake. A lot of practical reasoning is about adapting means to ends and, like any reasoning, it can be confused, irrelevant, or question-begging. Then there are subtler faults.11 Some- times desires are defective because we have not got enough, or the right, concepts. Theories need building which will 13 STATE OF MIND OR STATE OF THE WORLD? supply new or better concepts, including value concepts. For instance, it is easy to concentrate on desires to possess this or that object, at the cost of the more elusive, difficult- to-formulate, desires to live a certain sort of life. And it is almost impossible to strike the right balance between the two main components of happiness---on the one hand, the dis- content that leads to better and, on the other, contentment with one's lot.12 One needs more than facts and logic to sort those problems out: one needs insight and subtle, perspicuous concepts. And with information, more is not always better. It might cripple me to know what someone thinks of me, and I might sensibly prefer to remain in ignorance.13 What seems most important to the informed-desire account is that desires have a structure; they are not all on one level. We have local desires (say, for a drink) but also higher order desires (say, to distance oneself from consumers' material desires) and global desires (say, to live one's life autonomously). The structure of desires provides the criterion for "informed' desire: information is what advances plans of life; information is full when more, even when there is more, will not advance them further. So there is only one way to avoid all the faults that matter to "utility': namely, by understanding completely what makes life go well.14 This brings out another break with classical utilitarian tradition. Bentham, Mill, and Sidgwick all saw utility as having to enter our experience. But we desire things other than states of mind; I might sometimes prefer, say, bitter truth to comforting delusion. The informed-desire account has the advantage of being able to accommodate such desires. But the desire account does this by severing the link between "fulfil- ment of desire' and the requirement that the person in some way experience its fulfilment, dropping what we might call the Experience Requirement.15 If the delusion is complete, one believes that one has the truth; the mental states involved in believing something that really is true and believing a successful deception are the same. Or if a father wants his children to be happy, what he wants, what is valuable to him, is a state of the world, not a state of his mind; merely to delude him into thinking that his children flourish, therefore, does not give him what he values. That is the important point; the 14 MEANING informed-desire account does not require that fulfilment of desire translates itself in every case into the experience of the person who has the desire, and that is what gives the account its breadth and attraction as a theory of what makes life valuable. This seems to me the way that the informed-desire account has to develop. The definition itself is short: "utility' is the fulfilment of informed desires, the stronger the desires, the greater the utility. The way that the account develops, however, shows that all of those key terms are to a fairly large degree technical. (a) "Desire'. In the present technical sense, desires clearly do not have to have felt intensities; they need not be linked exclusively with appetitive states (some are, but others are aims we adopt as a result of understanding and judgement); they need not have existed before fulfilment. Rather, desiring something is, in the right circumstances, going for it, or not avoiding or being indifferent to getting it. (b) "Informed'. In its technical sense, "informed'16 is the absence of all the faults that I listed just a moment ago. There is a historically important account of practical reason that goes roughly like this: reason alone can never determine action. The end of action must be something fixed on, in its own reasonless way, by desire; we reason, but deliberation is only of means.17 It is hard to see what is at issue between those who say, with Hume, that reason alone cannot supply a motive and those who say, with Kant, that it can. But those of the latter pursuasion are right to this extent: in deciding how to act, we must try to understand what properties things and states of affairs have, and we must put our desires through a lot of criticism and refinement to reach this understanding.18 In this sense, deliberation may be of ends, and important deliberation often is. So an "informed' desire is one formed by appreciation of the nature of its object, and it includes anything necessary to achieve it.19 (c) "Fulfilment'. Being "fulfilled' cannot be understood in a psychological way, or we should be back with mental state accounts. A desire is "fulfilled' in the sense in which a clause in a contract is fulfilled: namely, what was agreed (desired) comes about. (d) "Strength'. "Strength of desire' has several senses, 15 STATE OF MIND OR STATE OF THE WORLD? appropriate to different theoretical settings. The "strongest' desire can be the winner, or it can be the most intensely felt. But strength of desire, in its technical sense here, has to be understood in connection with the structure that informed desires have. One does not most satisfy someone's desires simply by satisfying as many as possible, or as large a proportion. One must assess their strength, not in the sense of felt intensity, but in a sense supplied by the natural structure of desire. The desires I feel most intensely could be satisfied by your constantly imperilling my life and saving me only at the last moment,20 whereas I should clearly prefer peace to peril; anyway, felt intensity is too often a mark of such relatively superficial matters as convention or training to be a reliable sign of anything as deep as well-being. That I prefer peace to peril suggests that global desires provide, in large part, the relevant notion of strength of desire: I desire the one form of life more than the other. True, sometimes we form global desires only on the basis of having summed local desires (for example, the global desire for a way of life based on a reckoning that day-to-day pleasures will be maximized that way). But even then we must rank that way of life against others that it excludes, and our preference between them will, it seems, be basic---that is, a global judgment not based on any other quantitative judgments. This means that the relevant notion of aggregation cannot be simply that of summing up small utilities from local satisfactions; the structure of desires already incorporates, constitutes, aggregation. It means also that the relevant sense of "strength' is not simply the desire that wins out in motivation. If my doctor tells me that I shall die if I do not lay off drink, I shall want to lay off it. But I may later crack and go on a binge, and at that point my desire to drink will, in a perfectly clear sense, be strongest. If strength were interpreted as motivational force, then "utility' would lose its links with well-being; what would be good for me would then be satisfaction not of my informed desires but of what I "ought to desire' or "have reason to desire'. So to retain the links with well-being, the relevant sense of "strength' has to be, not motivational force, but rank in a cool preference ordering, an ordering that reflects appreciation of the nature of the objects of desire.21 16 MEANING 5.Troubles with the informed-desire account There are strong objections to such an account. Is it even intelligible?22 If our desires never changed with time, then each of us would have a single preference order, by reference to which what most fulfilled his desires over the course of his life could be calculated. However, life is not so simple; preferences change, and not always in a way that allows us totally to discount earlier ones. Suppose that for much of his life a person wanted his friends to keep him from vegetating when he retired but, now that he is retired, wants to be left to vegetate. Is there any intelligible programme for weighing desires that change with time and hence for maximizing fulfilment? If not, we may be driven back to a happiness or mental state account. Yet all the problems that we have just seen with mental state accounts remain; defects in one account do not obligingly disappear with the appearance of defects in another. How do we determine how happy a person is? Is happiness a single mental state? If many, how are they linked? Mental state accounts are hardly a refuge from troubles. Moreover, there may be an acceptable programme for handling cases where preferences change with time. The notion of an informed desire needs still further development and may eventually be able to supply the weighting of desires that we need in these troublesome cases. Has our retired friend simply forgotten the satisfactions of a busy life? If so, his later desire has much less weight. Is it just a change in taste, on the model of no longer liking ice cream? If so, his earlier desire has much less weight. We shall have to come back to these problems when we discuss measurement, but for now I have to be content with suggest- ing that the prospects of making the informed-desire account work are certainly not less rosy than those of making a mental state account work.23 The other troubles are much more worrying. The breadth of the account, which is its attraction, is also its great flaw. The account drops the Experience Requirement, as we called it. It allows my utility to be determined not only by things that I am not aware of (that seems right: if you cheat me out of an inheritance that I never expected, I might not know but 17 STATE OF MIND OR STATE OF THE WORLD? still be worse off for it), but also by things that do not affect my life in any way at all. The trouble is that one's desires spread themselves so widely over the world that their objects extend far outside the bound of what, with any plausibility, one could take as touching one's own well-being. The restriction to informed desire is no help here. I might meet a stranger on a train and, listening to his ambitions, form a strong, informed desire that he succeed, but never hear of him again. And any moderately decent person wants people living in the twenty- second century to be happy and prosperous. And we know that Leonardo had an informed desire that humans fly, which the Wright brothers fulfilled centuries later.24 Indeed, with- out the Experience Requirement, why would utility not include the desires of the dead? And would that not mean the account had gone badly awry? And if we exclude these desires that extend beyond the bounds of what affects well-being, would we not, in order to avoid arbitrariness, have to reintroduce the Experience Requirement, thereby losing the breadth that makes the informed-desire account attractive? The difficulty goes deep in the theory. In fact, it goes deep, one way or other, in any account of well-being. Another attraction of the account is that desires have to be shaped by appreciation of the nature of their objects. Without that restriction, the account is not even a starter. But with it, do desires even matter any longer? It may be somewhat too simple to say that things are desired because valuable, not valuable because desired. Yet the informed-desire account concedes much of the case for saying so. What makes us desire the things we desire, when informed, is something about them---their features or properties. But why bother then with informed desire, when we can go directly to what it is about objects that shape informed desires in the first place? If what really matter are certain sorts of reason for action, to be found outside desires in qualities of their objects, why not explain well-being directly in terms of them? It does not seem that it is fulfilled desire that is the basis of well-being, but certain of its objects. And that points us, depending on what we decide those objects are, either back towards mental states or beyond utility altogether.25 18 MEANING 6.Is there something between mental state and desire accounts? Mental state accounts are too narrow, desire accounts too broad. We ought to look in between. I said a while ago that one way of correcting the flaws in the simple mental state account would be by making utility neither a mental state nor desire-fulfilment. I then postponed taking up that suggestion until after we had looked at desire accounts, and it is time to return to it now. There is a cluster of terms which even in their everyday use seem to fall conveniently between mental states and fulfilment of desires: namely, enjoying or liking things, finding them pleasing or satisfying or fulfilling, being pleased or happy with them. Let us use the term enjoyment to cover them all. Enjoyment, in its ordinary use, is not anything so narrow as experiencing a single mental state or one of a range of states, but let us explicitly make it part of the present use of enjoyment that it is not. In similar spirit, let us specify that it is nothing so broad as merely having desires (even informed ones) fulfilled. Also, let us allow that people enjoy things other than states of mind; in fact, to treat the account sympathetically we should make the range of objects of enjoyment wide---wide enough to include, say, helping others or advancing knowledge. But let us put a limit to the range by requiring that all the objects of enjoyment fall within our experience. So we both push the boundaries out beyond its ordinary use and, at the same time, limit the expansion by reinstating the Experience Require- ment.26 The enjoyment account will need a lot more work to make it clear. Is it just a desire account with the Experience Requirement tacked on? Does the notion of enjoyment do any work of its own? Enjoying and liking, in their ordinary uses, are closely connected with desire: the acid test for whether I enjoy or like something is whether, other things being equal, I go for it, or do not avoid it, or am not indifferent to getting it. Enjoyment in its technical sense is even more closely connected with desire. Many everyday uses of like and enjoy suggest a certain psychological tone, but that disappears in the technical sense. However, if enjoying is just having a favourable attitude, the revival of the Experience Requirement runs the risk of 19 STATE OF MIND OR STATE OF THE WORLD? being ad hoc. I have favourable attitudes towards many things that do not enter my experience. Why do they not count too? Still, the enjoyment account is clear enough for it to face one serious trouble. It is attractive because, with the revival of the Experience Requirement, it restrains over-wide desire accounts. The intention is right, but the particular restraint applied seems not. It seems in the end simply too drastic. It bans things that our ordinary notion of well-being cannot, without damage, do without. It is common that, as many persons' values mature, such things as accomplishment and close authentic personal relationships come more and more to fill the centre of their lives. If the Experience Requirement excludes these values from "utility', then "utility' will have less and less to do with what these persons see as making their own lives good. And those values do seem excluded. Suppose that someone is duped into thinking that those close to him are behaving authentically. What enters his experience is the same whether he has the real thing or a successful deceit. But it is only the real thing, he thinks, that makes his life better. According to the enjoyment account, what affects well-being can only be what enters experience, and the trouble is that some of the things that persons value greatly do not. My truly having close and authentic personal relations is not the kind of thing that can enter my experience; all that can enter is what is common to both my truly having such relations and my merely believing that I do. And this seems to distort the nature of these values. If I want to accomplish something with my life, it is not that I want to have a sense of accomplishment. That is also desirable, but it is different from, and less important than, the first desire. And if I want to accomplish something, it is not necessary that I want my accomplishment to enter my experience---say that I know about it. That too is desirable, but it is still not the first desire. If either I could accomplish something with my life but not know it, or believe that I had but not really have, I should prefer the first. That would be, for me, the more valuable life. "Valuable life', of course, is full of ambiguity. It can mean a life that is valuable because of its value to other persons. It can mean a morally valuable life, or an aesthetically valuable one, or one valuable in terms of some code, such as a code of chivalry. But my 20 MEANING ground for preferring the first sort of life would not be any of these; I should prefer it because it would be, considered on its own, considered simply as a life I must lead, a more fulfilling one. So it is a value that has to be found a place within the bounds of "utility'. The enjoyment account, too, has its serious troubles. To my mind, the best prospect for a utilitarian account of well-being is to hold on to the over-wide desire account and look for good reasons to rein it in.27 It is harder to correct the over-narrow accounts. However, developing the desire account is a large job, which will be the subject of the next chapter. II UTILITARIAN ACCOUNTS: THE DESIRE ACCOUNT DEVELOPED 1.How may we restrict the desire account? THE informed-desire account will have to be abandoned unless we can find a way to restrict the desires that count. But we cannot do it with the Experience Requirement. That is where the last chapter left us. The trouble, you will recall, comes from examples like these: I want the sympathetic stranger I meet on the train to succeed; I want people in the twenty-second century to prosper; Leonardo wanted humans to fly. All of them informed desires, but (the trouble is) their fulfilment not part of well-being. The notion we are after is not the notion of value in general, but the narrower notion of a life's being valuable solely to the person who lives it. And this must itself impose restrictions on which desires count. As these examples show, the desires that count have to enter our lives in a way beyond just being our desires. So what we need to do is to make clear the sense in which only certain informed desires enter our lives in this further way. Think of the difference between my desire that the stranger succeed and my desire that my children prosper. I want both, but they enter my life in different ways. The first desire does not become one of my aims. The second desire, on the other hand, is one of my central ends, on the achievement of which the success of my life will turn. It is not that, deep down, what I really want is my own achievement, and that I want my children's prosperity only as a means to it. What I want is their prosperity, and it distorts the value I attach to it to make it only a means to such a purely personal end as my own achievement. It is just that their prosperity also becomes part of my life's being successful in a way that the prosperity of the stranger on the train does not. 22 MEANING But that can be only part of the story. It is not that informed desires count only if they become the sort of aims or goals or aspirations on which the success of a life turns. Good things can just happen; manna from heaven counts too. So we should try saying, to introduce more breadth, that what count are what we aim at and what we would not avoid or be indifferent to getting. What counts for me, therefore, is what enters my life with no doing from me, what I bring into my life, and what I do with my life. The range of that list is not so great as to include things that I cannot (e.g. the prosperity of our twenty-second-century successors) or do not (e.g. the sympathetic stranger's success) take into my life as an aim or goal. And Leonardo's wanting humans to fly would not count either; to the extent it became an aim of his life it was unsuccessful, and to the extent it was merely a wish it does not count.1 In a way the account is now circular. I appeal to our rough notion of well-being in deciding which informed desires to exclude from this account of well-being. But that, I think, does not matter. If what we were doing were taking a totally empty term, "well-being', and stipulating a sense for it, then we could not, in the middle of the job, appeal to "well-being'. But our job is not that. The notion of "well-being' we want to account for is not empty to start with; utilitarians use our everyday notion, and our job is to make it clearer. So we are free to move back and forth between our judgments about which cases fall inside the boundary and our descriptions of the boundary. Every account of this type will do the same. There is the same sort of undamaging circularity in mental state and enjoyment accounts, because they need to get beyond the ordinary senses of "pleasure' and "enjoyment', and they would have to go about fixing a new boundary in just the same way. This narrowing of the desire account still does not get rid of the great embarrassment of the desires of the dead. Of course, a lot of the desires of the dead do count morally, but that is because they affect the living. There is a good case for honouring wishes expressed in wills. Inheritance satisfies the desires of the living to provide for their offspring and encourages saving that benefits society generally. There is a good case, too, for granting rights to the dead---say, to 23 THE DESIRE ACCOUNT DEVELOPED determine whether their bodies are used for medical purposes.2 But that, again, does not require appeal beyond the well-being of the living. And, anyway, that a desire of a dead person counts morally does not show that it counts towards his well-being. The real trouble is our counting the fulfilment of aims even if (as it seems we must) we do not require that the fulfilment enter experience. Some of our aims are not fulfilled until we are dead; some, indeed, being desires for then, could not be. But is this so embarrassing, after all?3 You might have a desire---it could be an informed one, I think---to have your achievements recognized and acknowledged. An enemy of yours might go around slandering you behind your back, successfully persuading everyone that you stole all your ideas, and they, to avoid unpleasantness, pretend in your presence to believe you. If that could make your life less good, then why could it not be made less good by his slandering you with the extra distance behind your back that death brings? You might well be willing to exert yourself, at risk of your life, to prevent these slanders being disseminated after your death. You might, with eyes full open, prefer that course to longer life with a ruined reputation after it. There seems nothing irrational in attaching this value to posthumous reputation. And the value being attached to it does not seem to be moral or aesthetic or any kind other than the value to be attached to the life as a life to be lived.4 Here is another example. It would not have been at all absurd for Bertrand Russell to have thought that if his work for nuclear disarmament had, after his death, actually reduced the risk of nuclear war, his last years would have been more worthwhile, and his life altogether more valuable, than if it all proved futile. True, if Russell had indeed succeeded, his life clearly would have been more valuable to others. But Russell could also have considered it more valuable from the point of view of his own self-interest. For instance, it would not have been absurd for Russell to think the same about devoting his last years to some purely intellectual project without effects on others' well-being, such as patching up the holes in the Theory of Descriptions. A lot of desires of the dead would be ruled out on the grounds we have already mentioned, but it seems right for some still to count.5 24 MEANING 2.Why we should resist restricting it more Excluding some desires raises the general question of whether the best account of "utility' will not exclude desires of several further sorts. Should not other-regarding desires be excluded?6 Those who not only want their own welfare but also, luckily for them, have others wanting it too count more heavily than those who do not; for instance, orphans count less than children with loving parents. But that yields Bentham out of Orwell: each to count for at least one but some for more than one. Should not irrational desires be excluded?7 The principle of utility is a normative principle and ought perhaps, therefore, to grant weight only to what are, by its own standards, good reasons, such as benefit and freedom from harm, and to grant weight only to desires justifiable in terms of these reasons. Should not, for obvious reasons, immoral desires be excluded?8 Indeed, should not desires of any sort of moral character be excluded?9 If the concern of the principle of utility is with what ought to be done, then the desire for something because it is what ought to be done appears when the principle delivers its result and seems improper as a ground for the result. Do we not, in general, need a Theory of Types in utility theory to exclude certain desires from the argument place in utility functions? I think not. First of all, it is impossible to separate self-regarding and other-regarding desires. Each of us wants certain pure states of himself (e.g. to be free from pain); but we also want our lives to have some point, and this desired state can be hard to separate from states of others. Also, if we accepted the restriction to self-regarding desires, we should sever the connection between utility and happiness (and happiness is a large part of utility even on the informed-desire account). A father's happiness can be at stake in his child's happiness---two persons' welfare riding on one person's fate. Allowing that is no violation of everybody's counting for one; it merely allows the father, like everyone else, also to count for one. We have to swallow a little harder when we shift from involvements such as love to involvements such as hate, envy, spite, prejudice, and intolerance. If these sorts of desires are going to count too, what awful distortions will creep into 25 THE DESIRE ACCOUNT DEVELOPED political decision? But if a lover's happiness counts, so does a hater's schadenfreude. It is an ugly sort of pleasure, and as pleasures go slight and troublesomely mixed, but still a pleasure. If it ought not to have weight in moral or political calculation, then we had better find a way to keep it out. But out of moral and political calculation, where it probably does not belong, not out of "utility', where in some small way it probably does. Also letting other-regarding desires enter moral calculation seems to distort the notion of a moral reason. If one of my aims is to convince Britain that it ought to go over to comprehensive schools, why should you think that my desire constitutes yet another moral reason for going over to comprehensive schools? It would be absurd to introduce comprehensive education because it satisfied the desires of its advocates. What has overwhelming weight here, of course, is the good of the children and of society at large. But all that one really has to swallow is that the happiness of the advocates of comprehensives may in some small way turn on what happens, and that, at least, seems right.10 Simply to rule out irrational desires would also go too far. A compulsive hand-washer's desire is irrational, but its fulfilment affects his utility. So since irrational desires cannot be excluded wholesale, why not let them in, and if their fulfilment is sometimes morally intolerable, look to other moral matters besides utility to block it. True, the fulfilment of other sorts of irrational desires is more worrying. A misogynist might be put off his food by a woman's sitting next to him in the Senior Common Room. Consistency would seem to require that his desire not to have women around counts too. Well, why not? The suggestion earlier was that desires that are irrational on utilitarian grounds should not be given weight, because no utilitarian value is at stake. But if someone is upset or distressed, then there is a utilitarian value at stake. The theoretical oddity would come, not in giving weight to such desires, but in giving them none. What if desires are not only irrational but downright immoral? Should we count, for instance, sadistic desires? This has seemed more of a challenge than it really is, simply because people still tend to think of "utility' in rather narrow hedonistic terms. Anyone with much understanding would 26 MEANING regard his own sadistic desires---even purely from the point of view of how good his life is for him to lead---as making virtually no claim upon being fulfilled. He would have formed second-order desires not to encourage or indulge them; he would know that, in his case at least, their gratification is mixed and brings no lasting or deep enjoyment; he would know that their opportunity cost is enormous. In fact, it is hard to think of any fairly normal sort who would not be better off, from his own point of view, frustrating his sadistic desires and trying for something better. Still, it would be a mistake simply to rule out sadistic desires. Not everyone is fairly normal. Perhaps there is someone for whom sadistic kicks are all he has, who is incapable of better. It might even be right, if he were also an inept sadist who aims to shock and upset but succeeds only in boring, to play along with him. The same holds for a desire to do something because it would be morally right. There are people for whom living morally is so much at the centre of their lives that their success there is a large part of their lives' being successful. The ideal develop- ment of human nature is for "ethical push' (self-interest) and "ethical pull' (obligation) to get progressively closer to each other.11 That too complicates the notion of "well-being', but desirably so.12 All these cases for further restrictions focus not on "well-being' itself but on how it fits into moral theory. So this is another time when we have to remind ourselves that the question, "What is the best account of {"utility{'?', is quite distinct from the question, "What is the best account, as ad hoc as you like, that yields the most adequate one principle, utility maximizing, moral theory?'13 3.How value and desire are related The danger is that desire accounts get plausible only by, in effect, ceasing to be desire accounts. We had to qualify desire with informed, and that gave prominence to the features or qualities of the objects of desire, and not to the mere existence of desire. Then, to prevent informed desires from spreading too widely, we had to give prominence to only a certain range of features or qualities. Does this not confirm the suspicion 27 THE DESIRE ACCOUNT DEVELOPED that desire is no longer playing any real part, and remains only as a token of piety to a utilitarian tradition that has now effectively been abandoned? The issue widens. Could desire be a ground of value, or is it at best only a mark of it? Are things valuable because desired, or desired because valuable? And widest of all, what place do reason (cognition, perception, judgment) and desire (will, appetite, conation) have in explaining value?14 In a way the order of explanation must be from value to desire.15 We see that an object has certain features, such as that it is pleasant or healthy or that it gives security, or that it would be an accomplishment. And therefore we desire it. We have always to be able to cite some features that makes the desirability of the object intelligible; otherwise the notion of "value' loses hold. And that feature has to be generally intelligible as one that makes things desirable. No one can just make something valuable by adopting it as his own personal aim.16 Of course, people can disagree in their values. I might find mountain-climbing exciting and value it highly; you may find it simply terrifying and not value it at all. But we do not disagree here in our values in any deep or interesting way. Virtually everyone values excitement and does not value pure terror, though people differ in what they find exciting and terrifying. We all have to be able to connect what we value to some generally intelligible desirability feature. What is more, we sometimes discover values. You may be happy-go-lucky and not even think about accomplishing anything with your life, but then come upon someone whose accomplishment makes his life seem to you exhilarating and fulfilled. And with time you may come to discover more and more what this desirable accomplishment really is; you see how to separate it from mere achievement and its value from merely gaining praise. When you see what accomplishment is, you form a desire. And there need not be any pre-existent background desire (except those of vacuous generality) of which your new desire is merely another instance. But what is interesting is how little any of this shows. It still leaves a strong case for saying that the order of explanation is quite the reverse: from desire to value. True, objects are valuable because of their features. But how do we explain these 28 MEANING various desirability features? How do we separate desirability features from the rest? Here we have to guard against taking one or two examples as paradigms, and missing the variety of cases. So consider the following ones. Case 1: I have tasted both apples and pears. I like both but prefer pears. How do we explain my attaching more value to having a pear? The only relevant desirability feature is that they taste good. However, it is not a plausible explanation of tasting better that I perceive that pears possess this desirability feature to a greater degree than apples. We need to explain my liking pears more in terms of my wanting them more. That is true whether different persons' tastes coincide or not. But another important feature of tastes is that often they do not. We have no reason to expect, with many tastes, that differences in valuing shows that there is any lack of perception or understanding. My preference for pears is not open to criticism (though others of my tastes are---for lack of discrimin- ation, experience, attention). There is a tradition, especially strong in the social sciences, that sees all preference on the model of the simplest tastes: a pre-existent motivation, not subject to criticism, unaffected by understanding; the explana- tion running from desire to value. But this is only one kind of case. Case 2: A recluse may see what he is missing and come to prefer good company. Here perception plays a large part; it may even be a case of discovering a value. But why is good company itself seen as desirable? The explanation cannot be just in terms of perception; there is an important pre-existent motivation. The motivation is not a taste, which typically can vary from person to person; it is more a feature of human nature. We are social creatures; we want and, other things being equal, go for company.17 Case 3: Freud, in his last days, preferred thinking clearly to drugged comfort. Here there is a large element of perception but no obvious, at least simple, pre-existent motivation. We get no plausible explanation of this case unless we bring in both understanding what it is to think clearly and wanting it more. Explaining the state of thinking clearly as a desirability feature needs both perception and desire, without priority to either. To see this feature as desirable and to desire it on seeing 29 THE DESIRE ACCOUNT DEVELOPED it are the same. There is no plausible explanation of the one in terms of the other.18 Case 4: A person who in the past has frittered his life away comes to value accomplishing something with it. Here under- standing plays an enormous role, and desire may seem to disappear altogether. Accomplishment, in the general sense I have in mind (making one's life valuable and not just frittering it away), is valuable for everyone; anyone who fails to recognize it as valuable lacks understanding. It is true that there will be odd types for whom, all things considered, it will be better not to try for it. Perhaps someone for whom any ambition sets up intense anxieties had better not. But that is a case of conflict of values, in which accomplishment is still a value. Does the priority now run the other way? Do we see a life of accomplishment as valuable and then, on the basis of that, form a desire for it? Clearly, this is not a case of first perceiving facts neutrally and then desire's entering and blindly fixing on one object and not another. The way in which we talk about the objects we value is far from neutral; we call it "accomplishment' and explain it in terms of "giving life weight and substance' or "not wasting life'. The language we use in reporting our perceptions already organizes our experience and selects what we see as important; it is designed to show how we view certain things in a favourable light. Desire here does not blindly fix on an object; it is obviously pointed in certain directions by what we perceive favourably. But all of this, though true, explains too little. We also have to explain what goes on in our perceiving things favourably. And here desire comes back at a deeper level, as part of this explanation. Hume was wrong to see desire and understanding (appetite and cognition) as distinct existences. He was wrong to make desire blind. But it is a variety of the same mistake to think that one can explain our fixing on desirability features purely in terms of understanding. It is a mistake not only to keep understanding out of all desire but also to keep desire out of all understanding. Some understanding---the sort that involves fixing on certain features and seeing them in a favourable light---is also a kind of movement.19 It requires a will to go for what has those features. There is no adequate explanation of their being desirability features without appeal 30 MEANING to this kind of movement. So we cannot, even in the case of a desirability feature such as accomplishment, separate understand- ing and desire. Once we see something as "accomplishment', as "giving weight and substance to our lives', as "avoiding wasting our lives', there is no space left for desire to follow along in a secondary, subordinate position. Desire is not blind. Under- standing is not bloodless. Neither is the slave of the other. There is no priority. It may still seem that a value such as accomplishment has to have some priority to desire. It may seem that such a value cannot even depend upon what, if informed, persons would desire. One thing that would show this is our deciding that accomplishment is of absolute value. But few of us believe that; it is much more plausible to think there could be the very rare case in which trying to accomplish something was so painful that it was not worth doing. Another thing that would show it is our deciding that the value of accomplishment, while not absolute, is not given by its place in informed desires either, in other words that reflectively wanting accomplish- ment and recognizing its value to one can differ. But what would the difference be? Where, for instance, would trade-offs sanctioned by these different conceptions of the value of accomplishment diverge? So desire is more than merely a mark of value. It is a ground, in the following sense: it is part of the full explanation of prudential value. But this does not give desire priority. Nor does the appearance of desiring in valuing mean that we are free to make an existential choice of values. The desires that count are not brute and unconstrained; they are informed. It is the strength of the notion of "informed desire' that it straddles ---that is, does not accept any sharp form of---the divide between reason and desire. The advantages of the informed desire account, therefore, seem to me to be these. It provides the materials needed to encompass the complexity of prudential value. It has the advantages of scope and flexibility over explanations of "well-being' in terms of desirability features. It has scope, because all prudential values, from objects of simple varying tastes to objects of universal informed agreement, register somewhere in informed preferences. It has flexibility, because 31 THE DESIRE ACCOUNT DEVELOPED not everyone's well-being is affected in the same way by a certain desirability feature, and we want a notion sensitive to these individual differences. We want to know not only that something is valuable, but how valuable it is, and how valuable to different persons. 4.A formal account An old and potent objection to the utilitarian way of thinking is that it assumes that we value only one kind of thing, whereas we value many irreducibly different kinds of things. It seems to me undeniable that we do value irreducibly different kinds of things. But that point counts against certain mental state accounts, not against the informed-desire account. On the desire account one can allow that when I fully understand what is involved, I may end up valuing many things and valuing them for themselves. The desire account is compatible with a strong form of pluralism about values. However, the desire account may purchase its pluralism at the price of emptiness.20 If I advise you to maximize the fulfilment of your desires, I have not helped you much. I have not supplied you with the dominant end of human action by appeal to which you can resolve conflict between your subordinate ends. Nor have I given you a principle of choice of ends. It is no use to you to be told that you should decide what to go for by seeing what gives you most of what you decide to go for. Maximizing the fulfilment of one's desires does not yield, but presupposes, a hierarchy of goals. In contrast to this, the old notion of utility as a pleasurable mental state was both of these things---a dominant end and a principle of choice. And this contrast can easily give the impression that the desire account makes "utility' almost empty. But the charge that the new notion of "utility' is empty is, I think, partly the charge that it does not do what the old notion would do, and that is certainly correct. "Utility', on the old monistic interpretation, was the super, over-arching, substantive value. But now, "utility', on the desire account, is not to be seen as the single over-arching value, in fact not as a substantive value at all, but instead as a formal analysis of what it is for something to be prudentially valuable to some 32 MEANING person.21 Therefore, utility will be related to substantive values such as pleasure or accomplishment or autonomy, not by being the dominant value that subsumes them, but by providing a way of understanding the notion "(prudentially) valuable' and hence the notions "more valuable' and "less valuable'.22 So when, for whatever purposes, we shift from everyday talk of pursuing various different ends to theoretical talk of maximizing a single quantity, "utility', this quantity should not be understood as an end of the same kind, only grander.23 There is simply no case for reducing these various ends to a single end in this sense. The most that can be said is that a person's ends are unified only in being his ends, things he values. When our various values conflict, we may attempt to resolve the conflict by trying to realize as much "value' as possible, but the only substantive values present remain the various values that originally appear in our system of ends. We are still able to go for the most "value', to step far enough back from all of our various particular ends and sacrifice the lesser for the greater, even in the absence of a single substantive end as mediator. We mediate, but without such a mediating value.24 Is this, then, a "neutral' account of utility, in the sense in which accounts in recent economics and decision theory are? Yes and no. Yes, because this account, unlike hedonism or ideal utilitarianism, mentions no substantive values. But no, if the account is taken more widely to include the arguments for it, because then substantive values have to appear. Is this account "objective' or "subjective'? By "subjective', I mean an account that makes well-being depend upon an individual's own desires, and by "objective' one that makes well-being independent of desires.25 It may look as if an informed-desire account could not be anything but subjective, since it makes "desire' part of the explanation of prudential value. But that entirely depends upon how we take the phrase "an individual's own desires'. Values do not rest upon one person's desire. Values cannot be entirely personal, the result simply of someone's wanting the thing. That would not even be intelligible; persons generally have to be able to see a prudential value as something to go for if it is to be a prudential value at all. But the informed-desire account does 33 THE DESIRE ACCOUNT DEVELOPED make well-being depend upon variant, individual desires in this sense: it gives a place to both actual and ideal desires. Lafite may be worth much more than Coke, and might be to anyone at all if he appreciated all the flavours they contain, but is worth less to me with my untrained palate. And the account is certainly incompatible with some versions of an objective-list approach to well-being. An objective-list ap- proach says that a person's well-being can be affected by the presence of certain values (which it lists) even if they are not what he wants. The informed-desire account can allow that the values on the list (enjoyment, accomplishment, autonomy, etc.) are values for everyone, but it also allows that there may be very special persons for whom any value on the list (say, accomplishment), though valuable for them as for everybody, conflicts enough with another value (say, freedom from anxiety) for it not, all things considered, to be valuable for them to have. If a certain objective-list approach denies this, then it is different from the informed-desire approach. If it does not deny it, and even plausibly includes enjoyment on its list, and furthermore accepts the complex view about the relation of value and desire that I set out in the last section, then it gets very hard to distinguish from the informed-desire approach.26 Some philosophers treat the distinction between objective and subjective as if it marked a crucial distinction between accounts of well-being.27 They do, because they attach great importance to whether or not well-being is made to depend upon an individual's desires, tastes, feelings, or attitudes. But, as we just saw in the last section, the depend- ence of prudential value on desire is much less simple, less a matter of all or nothing, than they assume. The best account of "utility' makes it depend on some desires and not on others. So the distinction between objective and subjective, defined in the common way that I have defined it, does not mark an especially crucial distinction. It would be better if these terms (at least in this sense) were put into retirement.28 But if they are not, if the question "Subjective or objective?' is pressed, then the answer has to be "Both'. That answer shows how far what seems to me the best account of "utility' has to move away from its classical beginnings. It has to move from mental state accounts to a 34 MEANING desire account; it has to move from an actual-desire to an informed-desire account; and it has to set the standards for a desire's being "informed' in a place not too distant from an objective-list account. This is a stiffer standard for "informed' or "rational' desire than other writers have wished to adopt,29 so much so that it might seem that I should use a different label. But this label, it seems to me, has merits. It records the fact that this account is a development of one utilitarian tradition, that there is no plausible stopping point for the notion of "utility' short of it, but that this point is still short of objective-list accounts. And it records the role both of understanding and of desire. But we need a fuller explanation of how prudential values are identified and what sort of argument about them is possible. I shall come back to that subject in chapter IV, section 3. 5.Maximization and the unity of life Maximization, in some sense, is our prudential policy: we want to have the most valuable life we can. But is maximiza- tion, in this uncontroversial sense, a matter of aggregation? To answer that, we have to be clear about the sort of unity life has. On desire accounts, aggregation is explained in terms of the structure of desires. If desires are fairly well informed, the structures are plans of life. This talk of plans of life does not mean that valuable lives must be highly planned.30 One could have a life plan to take each day as it comes. Or one could have a minimal plan to live for short-term pleasures that, in aggregate, reach the greatest total. But for us to be able to evaluate any approach to life, even these, we have to see them in the fairly long term: as ways of living, which exclude other ways of living, all of which have---in those terms---to be ranked. We can never reach final assessment of ways of life by totting up lots of small, short-term utilities. We can do this totting up of short-term quantities, it is true, with certain prudential values in certain areas of life. We can say of pleasure, for instance, that this moment is pleasanter eating a pear than an apple, and the next moment will be pleasanter washing it down with Lafite than with Coke. But that cannot be the model for our final, authoritative calculation of utility. It 35 THE DESIRE ACCOUNT DEVELOPED has to take a global form: this way of living, all in all, is better than that.31 Let us return now to the question whether the (prudentially) most valuable life is the one with the greatest sum of utility. It is not, some persons think, because aggregation is insensitive to distribution, and distribution can matter in prudence as well as in morality.32 I might prefer a life with a lower sum to one in which the bad periods got very bad---the sort of "minimum acceptable level' requirement often imposed be- tween lives now applied inside a life. But how could I know that it had a lower sum? That calculation seems to need the sort of totting up that I have just suggested could not, in the end, be utilitarian calculation. We might try various ways of introducing a metric---say, by comparing each of the items to some entirely different item about whose value we are clear. In this spirit, I might calculate: I should be willing to sit under the dentist's drill for a thousand hours to avoid the awful year, for two thousand hours to get the glorious year, and only four hundred hours for each of the two middling years; so there is more good in the good year than bad in the bad year, and more net good in that package than in the package of the two middling years. But then we have to go on to say: still I prefer, all things considered, the second package to the first. The difficulty is to know what is going on in the first set of estimates. The second, the final preference, ranks the packages as stretches of life to be lived. So that seems to be the only judgment on which talk of utilities can be based. The first set of estimates, if they can be given a sense, must be doing something else.33 The desire account, therefore, detaches aggregation from a totting-up conception. So, to understand aggregation, we have to understand how desires are given structure. Philosophy has not yet provided a fully satisfactory explanation of how we bring system to our various ends.34 Pleasure, accomplishment, autonomy, loving relationships are all valuable. A life with only one or two of them, even in large quantities, would not be the best life. We cannot introduce system to our ends by calculating amounts of this value determined in isolation and amounts of that one, also determined in isolation. To some degree, estimating the values of these various ends is like 36 MEANING determining the value of ingredients in cooking. We can measure the quantities of wine and beef and onions separately, but we can only measure their value to the dish by considering them in various combinations. With our different ends, too, the important estimate of how valuable they are is their contribution to a whole life. And although we can talk about the amount of some of our ends (e.g. pleasure) in isolation, with others it is more difficult (e.g. autonomy), and in any case once we talk about amount of utility we have to see these items as part of a life. So the amount of value cannot be decided by attaching a value to each separately and then adding. There is a notion of the amount they contribute to a life that is independent of anything we can say about each end on its own. On that overall judgment we are in a clearer position with cooking: there is a further value that we appeal to in assessing various combinations of ingredients---how good they taste. But when we assess various combinations of our ends, there is no further substantive value, no supreme or dominant end, to rank them. There is not even an inclusive end, some larger value of which they are a part---at least, if that value is thought of as substantive.35 All that we can appeal to, in the last analysis, is our notion of "a valuable life'. The form of the judgment is, this combination makes a more valuable life than that. And the notion "a valuable life' used here does not serve to summarize further, more substantive notions; judgments about more or less "valuable life', in those terms, are basic. So if by "aggregation' one understands addition on the totting-up model, then the desire account abandons aggrega- tion. But I think that there is a strong sense in which it retains it. What are central to utilitarian calculation, on the desire account, are trade-offs. We make judgments of the form: this loss of value is less than that gain. And it is part of the desire account that the best prudential policy is going for net gains. We began this section wondering whether maximization, in the uncontentious sense, is different from aggregation, in the sense that the desire account attaches to it. It seems to me that it is not. The desire account may be too all-encompassing to be the account of "well-being' that suits moral or political theory. But it seems unavoidable in prudential theory. In that 37 THE DESIRE ACCOUNT DEVELOPED sphere, I want next to suggest, it comes close to being what we mean both by "good' and by "value'. 6.A (restricted) interest theory of value The desire account has all along been designed to keep "utility' close to "what is in a person's interest'. It is not that the expressions "fulfilling one's informed desire' and "being in one's interest' come out meaning the same. They certainly do not mean the same in the usual sense of "desire'. Even in the technical sense of "desire' that I adopt, they do not mean quite the same. For one thing, "informed desire' is a formula devised with humans in mind, whereas "interest' can be applied without strain to animals---even to plants for that matter. "Interest', in that respect, is wider in application than "in- formed desire', while in other respects it is narrower. For instance, I may desire turkish delight simply through having a weakness for it, but to say that it is in my "interest' applies a heavy word to an entirely light subject. None the less, these are all pretty minor differences, and the application of the two expressions is very close. The link between "interests' and "good', however, is one of meaning. The word "good' has associated with it "the condition of answering certain interests, which interests are in question being indicated either by the element modifying or the element modified by {"good{' or by certain features of the context of utterance'.36 The use of "good' that we are especially concerned with is "good for such-and-such a person'. The relevant interests, therefore, are that person's interests, and for something to be good is for it to satisfy those interests of his. And since utility is linked to satisfying a person's interests, it must in the same way be linked to a person's good. Now let us introduce value. A good starting point is the simple Interest Theory of Value, which says: for a thing to be of value is for it to be the object of, or the satisfaction of, desire or interest.37 But that must be too simple. At best the Interest Theory fits the case of prudential value, and the word "value' is used of far more than just prudential values. Morality itself may be said to be valuable, but on no plausible moral theory is its value exhausted by its status as an object of desire, even 38 MEANING informed desire. So the only form of the Interest Theory that is plausible is one restricted to prudential value. There are two weighty doubts about it. It does not make value prior to desire, and there are cases where value seems to have that priority. And it does not allow the value we attach to our various ends ever to diverge from their place in the hierarchy of our informed desires, and there seem to be cases of such divergence. But these are doubts we have already tried to allay.38 So this restricted version seems to me right. There are, none the less, differences in tone between "valued' and "desired'. "Valued' and "valuable' are, like "good' and "interest', rather weighty words and do not sit altogether comfortably in cases where what one wants lacks importance. But these are minor troubles, the sort of linguistic strains that most theories involve. The links between utility, good, and value show how broad a notion utility, on the desire account, is and how strong its claims for the central place in prudential theory really are. 7.Is this account still utilitarian? The desire account may seem to have few links left with the utilitarian tradition. But that is not quite right. The desire account is a natural development of classical utilitarianism. Bentham often sounds like a thorough-going hedonist. But even in The Principles, once having made his declaration of allegiance to hedonism, in the next breath he talks of what is "good' for people, or to their "advantage' or "benefit'.39 Mill, we know, felt forced to introduce his famous distinction between quantity and quality of pleasure. He explains quality in terms of informed desire: the pleasure of higher quality is the "one to which all or almost all who have experience of both give a decided preference'.40 Mill intro- duced "quality' as a necessary supplement to talk about "quantity', but the supplement then swallows up the rest of the theory. The criterion for differences in quality is the only one with the scope necessary for the theory, and the only intelligible talk of quantity is what can be defined on preference orderings. In fact, the desire account may be seen 39 THE DESIRE ACCOUNT DEVELOPED as a development of one side of Mill's theory to a point where the rest can be accommodated in subordinate places.41 Sidgwick knew well how much explanation the words "pleasure' and "happiness' needed in order to carry the theoretical burden utilitarians put on them. He adopts an informed-desire account (although one where "consciousness' is a necessary part of our ultimate good). For instance, he explains a person's future good as "what he would now desire and seek on the whole if all the consequences ... were accurately forseen and adequately realised in imagination ...'42 The trouble with Sidgwick's account is that its two elements, consciousness and desire, come apart, and it seems best, for reason that we have seen, to follow where the desire element leads. Again, when Sidgwick's account is developed far enough, the notion of informed desire dominates, and consciousness assumes a subordinate place. It does because that arrangement yields the best account of "well-being'. Is it the best? That brings us to the most serious doubt of all about the desire account. If "well-being' is defined to include fulfilment of desires that are trivial, abnormal, cheap, disgust- ing, and immoral, perhaps it is too wide. Doubtless it forms one notion of well-being, the broadest one, and one that we should sometimes want to use. Perhaps even it is the notion to use in prudential value theory. But do we not need a much narrower, more sharply focused, more discriminating notion for moral theory? III OBJECTIVE ACCOUNTS 1.Two concepts of well-being ONCE we start using "well-being' in moral theory, we come under strong pressures to make it much narrower than utilitarians do. Why should I accept that your mere desires make a claim on me? You may, when informed, want champagne, but if you do not need it, why should my obligations to you be at all engaged? On the face of it, the notion of "need' suits morality far better than "desire' does. Your informed desires are bound to range more widely than those interests of yours that the rest of us ought to weigh in deciding how to act. And consulting desires distorts moral calculation. If you are rich or bright, you will rationally form a conception of the good life adjusted to these resources and capacities of yours. But why should informed desires shaped, as they are, by resources that may have been unfairly distributed in the first place and genetic capacities that in any case have no moral standing then become the basis for determining what is just? It would be a crazy standard of justice. Clearly we have to get behind desires and expectations to the deeper considerations that show which desires and expectations have moral force. Fulfilling any informed desire may increase utility, but what morality seems to need is not the gross notion of utility, but a notion more finely focused on the vital interests, the basic needs, the central human concerns, that create obligations. We have two concepts of well-being, one broad and the other narrow. Both, no doubt, have their uses. But only the second, it seems, suits morality.1 The nature of our job, therefore, has now changed. Until now we have wanted to explain the rough, broad, everyday notion of well-being that is used outside moral theory. Our question was, What account describes and clarifies that notion best? Now we want to find the possibly much narrower notion 41 OBJECTIVE ACCOUNTS that suits moral theory best. So instead of starting with a notion that we might later fit into morality, we start with morality and ask which notion fits it. In the end, the only conclusive way to answer the question is to wait and see what happens when we build an adequate moral theory. But quite a lot can be done in a preliminary way. Of course, it is too simple to speak of just two notions of well-being. There is merit in a notion somewhere between the broad and narrow ones, or in an eclectic concept that borrows from each. One quite attractive position, for instance, is that "well-being' includes both basic needs and mere desires, but that needs have priority over mere desires.2 These are all possibilities that we should now consider. 2.The need account Needs are not a sub-class of desires. They are not, say, strong or widespread or central desires. While "desire' is, "need' is not an intentional verb; I can only need a thing if I need anything identical with it. So, while "desire' is, "need' is not tied to a subject's perception of the object; if I need a thing because it will cure my headache, it really will cure it.3 Desires have to do with how a subject of experience looks out on the world; needs have to do with whether one thing is in fact a necessary condition of another. Needs do not even have to be attached to subjects of experience. More than just living things have needs. An element needs a free electron to conduct electricity.4 Statements of need are of the form: x needs a in order to u.5 And even judgments about human needs are made from the same detached standpoint and take the same form. Humans need vitamin C to avoid scurvy. Terrorists need cool nerves to plant bombs. But let us focus just on human needs, in particular on the needs that persons have because of their aims or ends. And here there are two sorts. Some of them are instrumental: they are needs we have because of the ends we happen to choose. Some, however, are basic: they are needs we all have just by being human. We need food to survive; though we might in special circumstances choose not to survive, in general survival is part of what human existence aims at and is not in any 42 MEANING ordinary sense an object of choice for us.6 We usually speak of basic needs as if they were not only basic but also absolute: humans need food and rest and health---not for anything; they just do. But these claims are still of the form: x needs a in order to u. Humans need food to survive, rest to go on functioning, health to do much of anything.7 The question whether statements of basic needs are elliptical should not be confused with the question whether they, like hypothetical imperatives, depend upon the adoption of some purpose. If they do not fill the last place ("in order to u') they are elliptical; but they do not depend upon ends that we just happen to adopt. The ends that fill that last place are the normal ends of human life. The rough explanation of basic needs is clear: they are what we need to survive, to be healthy, to avoid harm, to function properly. It is obvious that instrumental needs as such have no moral weight; some have and some have not. But basic needs, on the other hand, seem to have a quite special moral importance simply in virtue of being basic. For one thing, the presence of the notions of health, harm, and proper function make statements of basic need normative---in the proper, strict sense of the term that is not the contrary of "descriptive'. They all involve a norm falling below which brings malfunction, harm, or ailment. And that explains why basic needs have an especially strong link with obligation: my ailment makes a claim on others that my whims, hankerings, pleasures, and even happiness cannot. And these claims on others are of a strong sort; they depend not upon this or that person's particular wish or purpose, but upon something deeper and objective---human nature. These last thoughts lead naturally to the following proposal: Well-being, at least that conception of it to be used as the interpersonal measure for moral judgment, is the level to which basic needs are met.8 3.Can we give a tolerably clear sense to "basic need'? The key notions of "ailment', "harm', and "malfunction' are too indeterminate as they stand to do the work expected of them by the need account. It is no good repeating the obvious example of food: certainly without food we shall ail. But it is 43 OBJECTIVE ACCOUNTS not enough that a few needs fall clearly one side or the other of the boundary of "basic needs' if most sit right on the line. Is interesting work a basic need? Well, without it, alienation, a kind of social pathology, results. Is education a basic need? Without it, one's intellect will atrophy. And how much education? There are many ways of introducing more determinateness. We could try developing the idea of a minimum provision. We need a certain amount (not just of resources, but also of liberty and leisure and education) to be able to make something valuable of our lives; what we have beyond that minimum amount may make our lives pleasanter or easier, but it is not necessary to the basic task of living a worthwhile life. If one does not have that amount, then one's life is maimed. I do not want to gainsay any of this. This notion of the minimum ingredients of a worthwhile human life is indispensable to moral theory; for one thing, we could give no adequate account of "natural' or "human' rights without it. Since the notion is indispensable, it has to be made determinate. But it is certainly not determinate enough yet. How much education would the basic kit contain? Literacy? Enough to ponder the meaning of life? And what level of health? When life expect- ancy is thirty, or fifty, or seventy? One does not make a notion such as "harm' determinate merely by linking it (correctly) to notions such as "minimum provision' or "human rights', both of which suffer from the same indeterminateness. We need a principled way to make all the members of this interconnected set more determinate. In any case, this way of making "basic need' more determinate may be making it too basic, too ungenerous, for it to perform its function in moral theory. If we want a notion of well-being to introduce considerations beyond the sort of minimum provision guaranteed by human rights, we had better not link it too closely to human rights. We might, instead, try saying that basic needs are needs for the all-purpose means to whatever ends different individuals may choose.9 This would be to develop one part of our explanation of "basic needs', namely, that they connect with ends that we do not choose but which are characteristic of human existence. Each of us chooses his own life plan; none of us chooses the sorts of things that are necessary for living out 44 MEANING any life plan. But though this is of help with one sort of indeterminateness (namely, with the content of basic needs; they are, it says, all-purpose means), it is of no help with another. How much of these all-purpose means are needed? If we answer "Any quantity', we have abandoned the conception of basic need that linked with obligation, because that conception has to retain some sort of reference to a minimum. Of course, many supporters of a need account readily admit its indeterminacy. They admit that basic needs have a large "conventional' element.10 But they say that, although the line between basic and non-basic needs may change as society changes, there is still a rationale for fixing the line wherever we do. For instance, we can banish the indeter- minateness by defining a standard on people's natural expecta- tions in that society; expectations adjust to possibilities, and a standard of minimum acceptable level of life, admittedly very rough, will naturally emerge.11 Although that is true, it is not clear why we should give this standard any moral weight. If we did, we should merely detach well-being from objective features of human nature and connect it instead to accidental social changes that have no obvious moral significance. A person's natural expectations are formed by much more than the real possibilities that face him; they are formed too by how good his education happens to be or by how much hope traditions or forms of government have left him. Since this approach is unpromising, we might instead try defining the minimum as what most persons have.12 A few persons may be relatively well off, but what virtually everyone has can be taken as the lowest level society now finds acceptable. But this is not really any better. In a primitive country which has just struck oil, most persons may be at grim subsistence level and a few in utter luxury, but the mere largeness of the number at subsistence level does not make it morally acceptable. It is not that there are no reasons, given the level of resources in a particular society, to fix a minimum at any particular place---and to revise it as resources change. There are, to my mind, perfectly good ones. The question is, rather, how much these reasons are supplied by the notion of basic needs. There is usually a large gap between the least and the most that a society can provide. And as the society gets richer 45 OBJECTIVE ACCOUNTS and richer, how do we decide where the minimum stops? Does the amount of education that is a "basic need' just increase in pace? The most sympathetic interpretation of the suggestion that basic needs have a large "conventional' element is that the key terms---"health', "harm', "proper function'---have to be given fresh interpretations in each social setting. But it is not just a matter of applying them to new conditions; their own conceptual resources are not enough to enable us to apply them at all. We cannot just interpret; we have to stipulate. As things stand, the notion of a "basic need' is playing very little role.13 4.The link between need and obligation This indeterminacy loosens the link between need and obliga- tion. If, over a wide area, what is harm, ailment, and malfunction is a matter of arbitrary or conventional stipulation, then over that area what constitutes an obligation is also arbitrary or conventional, which would be odd. Or if, in order to make "basic need' determinate, we must largely forget the notions "harm', "ailment', and "malfunction' and bring in others, then it is the others that do the important work, not "basic need'. Another consideration that loosens the link is this: whatever in the end goes on the list of basic needs, there are likely to be persons who want things off the list more than they want things on it. A group of scholars may, with full understanding, prefer an extension to their library to exercise equipment for their health. And part of what makes us think that basic needs, such as health, are more closely linked to obligation than are desires is that basic needs seem the "bread' of life and desires mere "jam'. But an extension to the scholars' library may not seem like "jam' to them. On the contrary, if the scholars' preference is sufficiently informed then the library is of greater value to them. But then to maintain that needs create obligations where mere desires do not, or that they create stronger obligations, is to say that we have an obliga- tion, or a stronger one, to the scholars to give them what they themselves value less, which would be odd. Much of the attraction of the need account comes from our deep intuition 46 MEANING that mere desires are morally lightweight. But the need account, though intended to exploit just this feature of desires, does not really do so. It leaves in the class of mere desires many heavyweight values---as values go, far weightier in many cases than needs, and weightier in a way that morality cannot ignore. Just as we have to work our way behind desires to considerations with greater moral significance, it seems that we shall have in some way to get behind needs too. There are plenty of reasons to resist this conclusion, but they all trade, I think, on a confusion. We ought to distinguish the question, What are the moral standards for distribution? from the question, What are the standards appropriate to distribution on a large social scale by a state? A social worker, confronted with a cripple who asks for a record player of equal value to the wheelchair that the state offers him, will turn him down. Similarly perhaps it is quite right for a state to go in for jogging tracks but not for manuscripts for the scholars' library. Still, what we are after, in the first instance, are basic moral principles---the principles of individual morality as much as of political morality. For many reasons, the principles of political and of individual morality are likely to be somewhat different. If the cripple is my son, and he says that he would much prefer an education in philosophy to lifts and wheelchairs, and I am satisfied that he knows what is involved, I should use my resources as he wants. His would certainly not be everybody's scale of values, but he himself might find ignorance more crippling than bad legs.14 The conventional notion of handi- cap does not, from a moral point of view, go very deep. We are nowhere near tolerably firm moral bedrock.15 So the link between needs and obligation, at closer look, is more tenuous than it first seems. Some harms are trivial. This is not just again the point that the notion of "harm' is indeterminate. Even if we take the notion at its tightest and least disputable---for instance, we are certainly harmed if our health is damaged, particularly if the span of our life is reduced---some harms are still minor, some reductions in span still insignificant, compared to other things, not harms accord- ing to need accounts, that can blight life. Why should harm to health matter morally and help to learning not? Health on its own is not valuable; it is necessary for life, out of which each of 47 OBJECTIVE ACCOUNTS us in his own case can make something valuable. But then what moral status has a necessary condition of a good life, in a case where achieving it will not allow one, and may prevent one, from having a good life? It is an odd conception of well-being that gives moral weight to trivial effects on what is only a necessary condition and no, or less, weight to what can be central to the state for which it is necessary.16 Of course, needs generally trump desires. What is at stake with basic needs generally matters much more to our lives than what is at stake with mere desires, and governments should concern themselves with basic needs and not at all, or at least not much, with mere desires. But it is a great mistake to move from that truth to the conclusion that needs as such count morally and desires as such do not, or anyway count less. Not all basic needs are morally important; some mere desires are. What we need are deeper categories. We have to get behind talk about needs and desires to their deeper significance in our lives. This, then, becomes a serious threat to need accounts, for if prudential value turns out to be that deeper category we shall be back with the informed-desire account. 5.Avoiding distortions to moral thought We do not, as a matter of fact, form our plans of life as if they were, in effect, choices from a Good Fairy's List---"whatever you want, just say the word'. Our desires are shaped by our expectations, which are shaped by our circumstances. Any injustice in the last infects the first. There is no denying that some accounts of well-being will, therefore, distort moral thought in this way. Actual-desire accounts will. A moral theory should not use as its base persons' actual expectations. It has to get behind them to what are in some sense legitimate expectations. The informed-desire account and the need account agree on this. They both appeal not to persons' actual desires but to what really would increase or decrease the quality of their lives. It does not matter that some persons have modest expectations; their informed desires include what they would want if they raised their sights, and their needs are anyway 48 MEANING independent of what they happen to desire. To that extent the two accounts agree. After that, though, they diverge. The case for the need account is that the class of "legitimate' expecta- tions is smaller than any desire account will make it. We spoke a moment ago about one reason---in the end not a very strong reason---for thinking so; some informed desires seem to create no moral obligations. However, another reason, which we should turn to now, is that "legitimate' seems already to incorporate elements of fairness, and that unless we use this tighter sense moral thought still runs the risk of being distorted. Of course, someone might develop expensive tastes just to skew allocation in his direction, but that cannot in the end be the important point.17 What is important is that many persons, when informed, develop expensive desires for perfectly ordinary, even admirable, reasons---simply because they are necessary to what they see as living a good life. Someone dedicated to advancing knowledge, for instance, will come to want expensive libraries and laboratories. Why should these desires be excluded from "legitimate' expectations? One answer is that we should otherwise give too much weight to chance, which has no moral standing.18 It is true that if we do not exclude them, whether we maximize well-being or make it equal, then someone with expensive desires will get more of society's resources than someone without them. Certain abil- ities, such as great intelligence, are in part the result of the genetic lottery, and they lead naturally to certain aims in life, such as advancing knowledge, that happen to be expensive. But this case for excluding them runs into trouble over the indeterminateness of "basic need'. Some of the ends of life adopted by an especially intelligent person are as much needs as desires. If giving more to him is unfair, the unfairness is likely to appear on the need account. And in any case, it is not true that these expensive desires are largely a matter of chance. Some persons have expensive desires because of ways of life that they have chosen after reflection. It is no good advising them to develop less expensive desires, although it is true that they could, because these desires are essential to what they see as making life valuable. Giving these desires weight is not, contrary to this argument, giving weight to chance. And 49 OBJECTIVE ACCOUNTS treating these desires as matters of chance is seeing them too much as if they were tastes. Applied to simple tastes, the argument has some force; one can expect a person to look after the consequences of his expensive tastes. But applied to central prudential values---non-idiosyncratic, well-grounded, essential to human flourishing, not however basic needs---it has much less force.19 Is the point, rather, that unless we narrow "legitimate' expectations, one person's vital interests can never be made secure from the swollen demands of others?20 Is the narrowed sense of "legitimate' the dike that morality must erect around a vital interest to protect it against inundation by aggregations of large demands of non-vital interests? If so, the need account is no answer; basic needs---health, for instance---can be less vital that non-needs. In any case, whether the protection is of each person's title to an equal share of all-purpose means or, more modestly, to a minimum acceptable level of life, it seems far too strong. Some societies do consciously satisfy the desires of minorities for libraries and universities, at the expense of, say, the level of road safety. Nearly everyone thinks it right to do so, even though it means that persons will lose their lives through no fault of their own who would not have done if non-needs had not been favoured over an all-purpose means or over a constituent of the minimum acceptable level. We cannot really assess this case until we determine exactly how strong morality should make this protection. That takes a well-worked-out, well-tested moral theory. The case often comes down to the claim that, to put it roughly, since an individual has to have strong protection, moral theory has to use the narrowed sense of "legitimate' provided by the need account. But once we think for a moment, it is not at all obvious that the protection should be as strong as the one provided by the need account. Or is the explanation this, that moral theory should concern itself with only all-purpose means in order to stay neutral between different persons' conceptions of the good life?21 Equality of opportunity is often interpreted, especially in modern political life, in a way that makes it morally unattractive. It is thought of as being satisfied just by bringing to some starting-post persons who are unequal in education 50 MEANING and influence and, unless their genes have been tampered with, also unequal in intelligence and health, and then letting competition largely settle who gets what. It is unattractive, because it totally ignores the demands of a deeper form of equality. We are all equal as moral persons; our fates matter equally; our fates should not be left to the mercy of these sorts of chances. Perhaps the same defence should be made of all-purpose means. Our fates matter equally; that you have a chance to make something valuable out of your life matters just as much as that I do. It does not matter that I think, perhaps rightly, that your conception of the good life is less defensible than mine; it does not matter that your life will in fact turn out less valuable than mine. In a deep sense, it is still just as important as mine. It is this deep sense of equality that the narrowed sense of "legitimate' can be seen as designed to capture. If well-being is confined to basic needs or universal human concerns or all-purpose means, then morality will remain neutral between conceptions of the good. However, this neutrality, though attractive at first glance, is pretty meagre. Can any moral theory get by with such a thin notion of well-being? Perhaps a government has good reason to remain largely neutral between conceptions of a good life. None the less, a father in distributing resources between his children, one of whom will simply waste them, ought to take this fact into account. And even a government cannot do its job if it ignores all values except basic needs. There comes a point when threat to some basic need, such as health, is so slight compared to what else is at stake, as not really to matter. But how can a government decide when it does not matter without deciding how valuable the competing value is? How can a government decide when to stop improving road safety and divert resources to art or education or libraries? It can make no sensible judgment about when that point is reached without making judgments about how valuable art, education, and libraries are. And it cannot make those judgments without implicitly making judgments about the relative value of at least some individual, non-neutral conceptions of the good. Throughout these various explanations of a restriction to "legitimate' expectations, questions of scale continually crop 51 OBJECTIVE ACCOUNTS up. Some principles are especially plausible on the large social scale. It makes sense to concentrate on moral principles, especially principles of justice, that apply on the social scale, if, as many think, the features that stand out clearly on the large social scale turn out to be essential, though less obviously so, on the small interpersonal scale as well. For instance, govern- ments lack knowledge of each citizen's conception of the good life; to get it would require intolerable intrusions on privacy; and in any case the information would usually be too unreliable and too little agreed upon to serve as a basis for interpersonal comparisons for social decisions that will com- mand assent. All of this clearly points to governments' being in large measure neutral between different conceptions of the good. Then we may find different and deeper reasons---for instance, the deeper notion of equal treatment that we have just looked at---for neutrality in individual morality as well. But once these deeper reasons begin to look dubious, we should look again at the whole question of scale. Perhaps much of what is appropriate on the social level is not essential to a principle's being a moral principle and so will not appear on the small interpersonal scale. We are trying to establish moral principles, principles that are relevant---perhaps in different forms---on any scale. Certainly, what matters on the social scale (chance factors such as degree of knowledge and agreement) is often irrelevant on the small scale.22 So we have to try to get behind these differences of scale as well. 6.A more flexible need account Most need accounts make the mistake of being too rigid; they rank basic needs above mere desires, regardless of the levels or amounts of satisfaction of each that are in question, which is implausible. That was the moral of the story of the scholars' choice between a library extension and exercise equipment. But this mistake can be corrected. We can introduce flexibility into need accounts by introducing a new notion: how important or vital or urgent a basic need is in a particular case.23 A basic need, we can say, is generally less important the more it is already met, and at some level of satisfaction it will cease to be important at all. We can make how important 52 MEANING it is also depend upon the level of satisfaction of other basic needs; if my needs for liberty and minimum material provision are not met, and are so unlikely ever to be met that life is not worth living, then health must matter less too. And if the amount to which a basic need will be met is only slightly affected by some change, we can regard the change as of little importance. We can then amend the need account of well- being: well-being is the level to which basic needs are met so long as they retain importance. We have not yet got a general account of importance (we should, for instance, need to define an interpersonal use), but even as far as we have gone problems start to emerge. Important to what? The answer must in some form or other be: important to how good one's life is. But if we leave the explanation as wide as that, the need account is at risk of being absorbed by an account of prudential value. Yet it is hard to see how to make the explanation narrower and still have a notion of importance that does the work required of it. There must be a point, we all agree, at which the moral demands of health are fully met and beyond which we are not obliged to go. There must also be a point at which the demands of health, although not fully met, are so little affected that they matter less than some mere desire. But how do we decide where these points are? With eyes wide open, the scholars prefer a few more books to a few extra weeks' longevity. Is this basic need still important enough to trump this mere desire? It seems impossible to form any estimate of how important the need is without appeal to the same standard that gives us the value of the mere desire---namely, how each affects the overall quality of life. And there seems to be no criterion by which to decide whether the demands of health are fully met, no matter how minimal we think these demands are, without seeing what else, including mere desires, people value and how greatly they value them---in short, by seeing how all the competing options affect the overall quality of life. It is unlikely that the conception of importance can be, as the account requires it to be, independent of our general conception of prudential value.24 The introduction of flexibility into the need account really takes us behind the importance of need to deeper values, to 53 OBJECTIVE ACCOUNTS the broader base of prudential values of which basic needs are just a part. All basic needs will have their place, and their importance marked, in the hierarchy of prudential values. The notion of a basic need is attractive because it goes deeper than the notion of desires. It goes behind the actual desires to the concerns that give them importance; needs connect more securely to the real stuff of well-being than actual desires do. But informed desires, as they figure in prudential values, go deeper than basic needs, as it is clear that we must. 7.Neutrality, objectivity, and moral depth Suppose that we drop need accounts. There are other, and to my mind more plausible, objective accounts. All need accounts rest on a distinction between, on the one hand, things that we aim at simply as normal human beings rather than as the particular human beings we are, things that are both necessary to and sufficient for a recognizably human existence, and, on the other hand, things that, as the individuals we are, we choose to go for. Keeping to the senses we used before,25 "subjective' standards are those that depend upon tastes, attitudes, or interests, while "objective' standards are those that do not. So the argument for regarding the need account as objective is fairly easy and straightforward. Its standard rests on aims flowing from human nature and not on any flowing from a person's particular tastes, attitudes, or interests.26 On this view, what is objective is also universal and neutral between differing conceptions of the good life.27 But there is no good reason why what is objective has also to be either universal or neutral. Why not think that behind our varying commitments to particular ways of life lie objective standards by appeal to which we can, among other things, settle the value of these commitments? Admittedly, the need account is not incompatible with this wider form of objectivism. But some of the appeal of the need account comes from the idea that these neutral, universal values are the only ones that could be objective. The root idea, I think, is this. Once we get outside the realm of basic needs, we are in the domain of mere desires: in this domain aims rest on the tastes, attitudes, and interests that we happen to have; 54 MEANING value is contingent upon our having certain desires.28 But this thought gets its plausibility from treating variant, personal values as if they came from tastes or attitudes that we just happen to have, or from commitments that we blindly make. But most of our non-universal, non-neutral values are not at all like that. They are not capricious or accidental or arbitrary. Most of them are not a matter of taste. And it is odd to think even that we choose them; generally they choose us, by being the sorts of values that we have only to perceive clearly to adopt as goals.29 A person might have as a central, animating goal of his life achieving some understanding of life itself. Certainly not everyone has this goal; it is a paradigm of a variant, personal goal. None the less, everyone may be worse off for not achieving this understanding. And it may be a goal that the rest of us should foster, even at some cost to health and survival. And the value does not depend upon desires in a way that makes it "subjective': it, like many other values, owes is status as a value to its nature, the proper perception of which would lead anyone to adopt it as his goal. So the most plausible form of objectivism allows that many values---many more than the ones linked to basic needs---are objective. The thought behind forming them into a standard of well-being would be this: when they appear in a person's life, then whatever his tastes, attitudes, or interests, his life is better. However, if that standard is taken to mean that we can measure changes in a person's well-being just by the amount that he realizes objective values, it must be wrong. As we have seen,30 there may be an unusual person for whom autonomy sets up such great anxieties for it not to be worth it to him. This is not to deny autonomy its objectivity; it is not to deny that it is a universal value; it still allows that autonomy would, other things being equal, make his life better. It is only that an objective-list account has to made more flexible, more sensitive to individual differences. It has also to be made more concrete; one objective value is enjoyment but what any particular person enjoys is very much a matter of the particular person he is. In just the way that we can make a fetish of goods---by using them, and not their effect on our lives, as our index of well-being---we can make a fetish of values: even objective universal values matter only by making individual lives better. 55 OBJECTIVE ACCOUNTS We have to find some mode of deliberation about values that sees them as they fit into particular lives. The manifestation of these objective values in particular lives is the deepest measure of value. And that again takes objective accounts very close to prudential value theory. Objective accounts have obvious strengths. They get close to what really matters morally, certainly closer than the flux of actual desires gets us. But the need account is not rich enough. The simple objective-list account is not flexible or sensitive enough. Informed desires include both basic needs and the items on the objective list; they give them structure as prudential values. There remains, however, the persistent embarrassment of the desire account's over-extension; not all of one person's desires make claims on other persons. Another strength of objective accounts is their concern with the special demands on a moral notion of well-being. But we should not approach moral theory in too narrow a way. The structure of moral theory may, in the end, provide good grounds for counting some informed desires and discounting others. If so, then what moral theory anyway independently provides does not have also to be built into our notion of well-being. IV PERFECTIONISM AND THE ENDS OF LIFE 1.Prudential perfectionism WHEN we move on to perfectionist accounts of well-being, we change the nature of our exercise once again. Desire accounts aim at capturing everything that comes into human well- being. Objective accounts aim at explaining a narrower conception of well-being, one thought (wrongly, I have argued) to be needed by moral theory. Perfectionist accounts aim at explaining a notion not clearly moral in a modern sense, closer at points to prudential value theory, but more concerned with what a good life is for humanity in general. Desire accounts focus on a person in all his individuality; objective accounts focus on an index of goods that are good to everyone, regardless of the differences between them; perfec- tionist accounts focus on a species ideal. At least, those are the differences that strike us at first sight. But just as the more adequate we make desire accounts and objective accounts the less clear the boundaries between them become, the more plausible perfectionist accounts get the less sharp are the differences.1 This is because there are many different "perfectionist' beliefs, some a lot more plausible than others. The historically most important one says nothing in the first instance about how we ought morally to act but tells us rather what makes an individual life good---a form of "prudential perfectionism', in contrast to "moral perfectionism'.2 It says that there is an ideal form for human life to take, a form in which human nature flourishes and reaches perfection. And it offers an account of well-being strikingly different from either desire or need accounts: the level of well-being for any person is in direct proportion to how near that person's life gets to this ideal. 57 PERFECTIONISM AND THE ENDS OF LIFE But is there a single ideal? Well, let us start by saying so. We can say that this single ideal may manifest itself in lives that outwardly look quite different (butcher, baker, philosopher, etc.); all that we have to insist on is that, to the extent that the ideal is really manifested in them, they will all be, on a higher level of generality, the same sort of life. In the most famous version of this argument, Aristotle is confident that there is a single ideal because he believes that there is a single human nature. And it is hard to see what other reason one could have. Aristotle's own teleological conception of the single human nature, however, is a good deal less obvious. He believes that everything has its own natural function or task (ergon), the performance of which is its final end.3 So he sets off on a search for the task or end of human beings. It is not simply life, he says, because plants have that too; nor is it sentience, because animals have that. It is, rather, rational activity.4 But, as Aristotle is well aware, the expression "rational activity' covers different things. It covers pure rationality, divorced from everyday life, the contempla- tion of truths; and it covers practical rationality, weighing the various demands of everyday life and striking a reasonable balance between them. These two forms of rationality leave Aristotle rather in two minds, so he ends up giving a two-part answer.5 A human's supreme function, he says, is pure contemplation; that is the peak of human existence.6 But the secondary human function, he adds, is living in conformity with practical reason. Not everybody, he allows, can attain the supreme end or, if they are able, sustain it for long; so the secondary end is more relevant to most persons' aspirations, a lower but more realistic peak of human existence.7 Neither of Aristotle's answers is persuasive. The first is clearly dubious. It pictures human flourishing as a God-like review of eternal truths as they march in orderly formation before the mind. This passive, narrow, austere, even rather boring activity would not go far towards making life valu- able or giving it substance. There are many other activities that are also valuable in themselves: enjoying oneself, accomplishing things, deep personal relations. And it is not that one of them is the single peak either; no one of them on its own would do. Only a quite relentless, unsubtle 58 MEANING application of a teleological conception of existence could yield such an unlikely result. So we ought to shift to the far more plausible notion of rational activity not as pure contemplation, but as the practical weighing and balancing of the elements of a good life. Anyway, Aristotle's suggestion that this practical rational- ity is a matter of second-best does not carry conviction, because even if we accept the assumption that human life has a single supreme function to which all others are subordinate, it is as likely to be a blend of contemplation and action as contemplation alone. And "rationality', in the practical sense in which we are now to understand it, is not the content of the ideal life (or else a version of the objection to the first conception of rationality would apply), but the capacity to determine what it should be. This shift to a formal conception of practical rationality, though more plausible as an account of the role of rationality in an ideal life, does not really help to advance an account of that life. The ideal life is one containing the right balance of virtues. Practical rationality is the procedure to decide what that is. But the ideal is not exercising our capacity to strike the right balance; the ideal is the balance. When we look for the ideal human life, we are looking for its content. Practical rationality---for example knowledge, autonomous choice---is part of its content, but only along with, for instance, enjoyment, accomplishment, and deep personal relations. So the shift to a formal concep- tion of practical rationality is no improvement, because we do not want a formal account of the ideal. The first (material) Aristotelian answer is the right sort of account, but it is implausible. The second (formal) Aristotelian answer is plaus- ible, but it is the wrong sort of account. Nor is there a single right balance. The right balance is very likely to vary from person to person. Autonomy is valuable to everyone, but, as we saw,8 a particular person might be made so anxious by being autonomous that it would be best for him not to be. That, admittedly, is an extreme case, but most of us, in some way or other, to some degree or other, fall short of the ideal. Certain of our capacities and skills are more limited than normal, and certain others more highly developed, and the ideal life for one of us would not be ideal for another. Even 59 PERFECTIONISM AND THE ENDS OF LIFE if we ignored the damage that the accidents of life can inflict on human capacities, even if we confined attention to the natural range of normal human capacities, there would still be no single ideal life. The natural and normal range would include an artist creating beauty, a researcher discovering truth, a politician realizing the good, a citizen tending his own garden. There is no one balance of prudential values in all of these lives, unless the description of the balance is so general ("the balance that realizes most value') as to be vacuous. And when we add the reality that most of us, as a result of the accidents of upbringing, have developed our native capacities to different degrees, then the perfectionist account of well- being seems not at all plausible. Your well-being and mine are not in direct proportion to how close our lives get to a single ideal life. There is no such ideal. And even if there were, it would be too insensitive to variations between persons to be the basis of a measure of each individual's well-being. Are there, then, several ideals for human life? After all, one does not have to be a monist to be a perfectionist. One can accept a plurality of values, not just because for one person there are irreducibly many different values but also because for different persons there are irreducibly different ideal forms of life. Different forms of life (the artist, the researcher, the politician, the private citizen tending his garden), to the extent that they are good, might instantiate one or other of these ideals. A pluralist form of perfectionism would have to say more than merely that there is an ideal life for each person, because there is nothing in that claim to make it especially "perfection- ist'. Anyone with a conception of prudential value would say as much: the ideal life for each person is the one that is most prudentially valuable. To turn this sort of pluralism into a form of "perfectionism' we should have to limit the ideal lives in some way, most plausibly by making them depend upon a certain number of kinds of human nature. But that is a tall order. People have different capacities and skills, which they possess in many different combinations and degrees. It would not be too tall an order to draw up a list of prudential values: accomplishment, enjoyment, deep personal relations, etc. We might even be able to complete the list. We can say that 60 MEANING persons with different combinations of capacities, or capacities to different degrees, would be better off with different propor- tions of these prudential values. And we can perhaps go so far as to give an account in the case of a particular person of what balance of those values would give him the best life. But there seems nothing in between the highly general list of prudential values and the very particular judgment about an individual life. For instance, there seems to be no place for an account of a certain number of different ideal forms of human life, made ideal by being the flourishing of a certain number of different kinds of human nature. What we lack is a rationale for the division of human nature into two or three (or fifteen or twenty) different kinds. 2.Moral perfectionism Perfectionism grows naturally from a prudential claim about ideal human lives into a moral claim about what we ought to do to promote them. One form of moral perfectionism might be seen as a development of Aristotle's first conception of human function, namely rationality as pure contemplation. Aristotle's own view seems to have been that very few persons can manage pure contemplation, but that this sort of life of the mind is the very peak of human existence. It is an easy step from that---though not one Aristotle took---to saying that therefore that sort of life ought to be enhanced above all others. It is the Superman version of perfectionism: make the peaks as high as possible, even if only one or two persons will reach them; maximize height.9 It has an obnoxiously e$0litist sound to modern ears, but one does not have to go back at all far in history to find it regarded as natural. Nietzsche's view that we should all live for the good of "the rarest and most valuable specimens', like Socrates and Goethe, might be thought a bizarre exception.10 But the young Bertrand Russell could write without a blush that an equitable distribution of wealth "might, if things didn't improve extraordinarily, cut off all the flowers of civilisation. The Shelleys and Darwins and so on ... And surely one Darwin is more important than thirty million working men and women.'11 61 PERFECTIONISM AND THE ENDS OF LIFE We all find the Superman version unacceptable.12 And that is not mere modern dogma. The Superman version is arid in spirit, immoral in recommendation, and, in the end, confused in formulation. Imagine an obsessional museum director who wants, say, only the one supreme Rembrandt and will pass up all the rest of the Rembrandts in the world to get it. For him, the single supreme Rembrandt would be better than any number of other Rembrandts, whatever their quality or variety of achievement, no matter how close many of them came to that single best, no matter even whether in aggregate they constituted something more important and more valuable. Even as an approach to art, it lacks any breadth. But writing off less than supreme works of art is a far cry from writing off less than supreme human lives. The museum director's policy would be merely narrow, but the comparable policy with human lives would be monstrous. There is agreement between moral theories that no theory is moral unless it grants some form of equal respect and unless it justifies trade-off of one person's good against another's by appeal to a certain range of human values. What is interesting about the Superman version is that it detaches the notion of perfection from prudential value, moral value, even aesthetic value. It makes the notion, in the end, unintelligible. Perfection- ism urges us to go not for quantity of value, but for the height of its realization. But how are we to measure "height'? We might have had a coherent account of it, if Aristotle's programme of finding the supreme function succeeded and if, in addition, he showed that "supreme' in his sense was also a reason for action. But the programme certainly fails. For instance, if Aristotle's argument is that pure contemplation is our supreme function because it is the purest form of rationality, then it succeeds only if "pure' is taken to mean something like "most detached from effort to reach practical goals', in which case there is no reason to accept a move from "purest' to any conclusion about how we ought to behave. What human achievement has greatest "height'? The con- templation of truth? Why not the discovery of truth? Why not the creation of beauty? Why not a full, varied, and satisfying life? We can, of course, appeal to prudential value theory for answers to these questions, but that would be to admit, as I 62 MEANING think we must, that "height' is not a coherent independent notion.13 The Superman version ruins a good point by drastic overstatement. It is entirely plausible to think that there are values that represent human existence at its very best, and rationality and creativity are obvious candidates. So it is plausible also to think that we ought, either absolutely or relatively, to favour lives of learning and art in which they are embodied. Perhaps the good point is that we should go, not for the highest peak, but instead for either as many lives of this sort as possible or the highest aggregate achievement in art and learning. This is still e$0litism, but of a welcoming, proselytizing, non-exclusive sort. This is the Spread of Civilization version: work for a multitude of faithful rather than one or two saints. It is certainly not obnoxious; there are even elements of truth to it. But it is not yet the good point buried in perfectionism. Its serious defect is that it picks out certain (few) features that make life good---for instance, knowledge and beauty---and elevates them above the others, and the collapse of prudential forms of perfectionism puts this whole procedure in question. Why are these few features to be preferred to the rest? They do not constitute ideal forms of life, because there are no such things. There are, it is true, features valuable to any life, but their value can be explained only within a theory of prudential value, where they are neither the only nor the supreme values. So to promote one part of what makes life valuable over the rest makes no sense as a moral policy. There is no form of prudential perfectionism that generates an acceptable standard of well-being. And once the link to well-being is cut, there is no longer any reason to move from prudential perfectionism to moral perfectionism. Still, there is one more version of moral perfectionism---the most plausible of all, I think---and it escapes this criticism. It does so by not resting its case on a shaky substantive account of ideal forms of life. Aristotle wavered not only between pure contemplation and practical rationality as the supreme func- tion of human life, but also between material and formal accounts of practical rationality. It was his teleological metaphysics that pulled him towards the first option in each pair, but the most sympathetic interpretation would in each 63 PERFECTIONISM AND THE ENDS OF LIFE case choose the second. So we should regard practical rationality, and the virtues in which it displays itself, not as a substantive account of the ideal life, but as a formal account of the modes of approach that will fix on the best life. Moral perfectionism, thus, does not rest on the description of what the ideal life consists in; rather both are generated by the complex ways of functioning that make up human excellences ---by display of wisdom, courage, temperance, industry, humility, hope, charity, justice, etc. Or, rather, it is not that practical rationality fixes separately the content both of prudential good and of moral obligation, but that it treats living a good life as a whole, not separating prudence and morality. Indeed, what many persons find attractive in this version is that it treats good as penetrated thoroughly by right and that it makes prudence part of a whole in which it is subordinated to morality. It sees acting rightly as acting in accordance with human excellences, the virtues; and acting that way, it says, is also the flourishing of human life. Another attraction for many persons is that it offers no formulas, no general principles, for moral decision; it sees moral decision not as the fitting of rules or laws onto experience but as the working out of human excellences as each particular case requires. This Virtue version of moral perfectionism sees perfectionism as an account of the perfect way for a human to function---the spirit in which to approach life, meet its trials, decide on goals and action. But for all its plausibility, it is still not the good point buried in perfectionism. It tries to make an account of the virtues do more work than it can. It proposes an occasion-by-occasion approach, but we have never been given, and cannot easily find, an explanation of what this approach is that gets beyond the vaguely suggestive.14 No account of the virtues seems rich enough to provide answers to all the questions that constitute the core of moral concern. For that we need a moral theory with more, and a different kind of, content. For instance, the virtues include justice, but justice is not merely approaching other persons in a certain spirit, respecting their interests. We have to know in what such respect consists. What restrictions, if any, are there on trade-offs of one person's interests against another's? What are 64 MEANING our obligations to the Third World or to future generations? How partial to family and friends can one be and remain moral? When are abortion and euthanasia permitted? Is the policy of nuclear deterrence immoral? To fill out the account of virtues for it to answer these questions would be to abandon it purely as an account of practical rationality. Filling it out would require determining the sort of content for moral "principles' that the formal account of "perfect human function- ing' was meant to do without. The virtues are "corrective', serving us in such ways as bolstering our defences at likely points of temptation or increasing our spirit when it is likely to let us down.15 Getting our motivation in good order is a large part of the moral battle. But it is only part. We also need answers to certain central questions. 3.The ends of life The place to start putting content into our theories is with prudential values. Up till now, we have talked mostly of their formal features. But we certainly have also to know what substantive prudential values there are. For one thing, moral theory cannot get on without that knowledge. We cannot begin to work out a substantive theory of rights---what rights there actually are---without a substantive theory of goods. We need to know what is at the centre of a valuable human life and so requires special protection. And we cannot decide anything about the permissibility or extent of trade-offs between well-being and rights, or between one right and another, without knowing what makes them rights. What sort of deliberation goes on in deciding the ends of life? How can claims about them be criticized and assessed? From time to time, I have used accomplishment as an example of one of the ends of life, and let me take it now as my chief example. The rough, intuitive case for it is plain. We all want to do something with our lives, to act in a way that gives them some point and substance. That much is clear. But it is harder to see the sort of accomplishment that would have this status. It would have to be kept distinct from simply wanting one's life to be valuable. It cannot be part of what makes life valuable if it is identified with the whole. And the idea behind 65 PERFECTIONISM AND THE ENDS OF LIFE accomplishment is that doing something with one's life is part, but only part, of what makes it valuable. It would also have to be kept distinct from bare achievement. The compulsive achiever may reach the goals he sets himself, but that need not be the sort of accomplishment that we are after. The achiever may simply love competition and derive great joy from success in the contest, but that, though valuable as enjoyment, is not quite accomplishment. In any case, bare achievements, though difficult, can themselves lack point or substance; walking on one's hands from Oxford to London is a remarkable deed, of the Guinness Book of Records sort, but itself lacks the worth that is part of accomplishment. This shows too that the development and exercise of skills, though usually satisfying, is not accomplishment either. Nor can winning respect and admiration be taken as a sign of accomplishment. It is very easy to confuse what others, even what recognized authorities, admire with what is a real accomplishment, and then to confuse having accomplish- ment as one's goal with merely wanting respect and admiration. And it is terribly easy to confuse respect for one's accomplishment with respect and affection for one's person, and to want to accomplish something for the personal affection it is thought, usually wrongly, to win. Accomplish- ment, as a value, is valuable independently of its conse- quences, and it is not accomplishment that a person values if what he is after is one or other of what he supposes its consequences to be. Accomplishment need not be of something lasting. Developing a treatment for a major disease would be a great accomplishment, even if it were quickly succeeded by a better one. Nor does it have to be of something of wide or public importance. A person who lives a rich, rewarding personal life could make out of it, say, a private work of the art of life, and he could rightly regard it as no mean accomplishment. This is one kind of reasoning that claims about ends are clearly subject to, the definitional. It whittles away what only looks like, or is confused with, the end itself. After whittling away, there might be nothing left; or there might be no new value, only an old one now separated from confusing appearances. 66 MEANING But why, once a candidate value is tolerably well defined, accept it as a value? Well, all that one can do to show that it is not a value is to keep whittling away---to show that what makes a candidate value look attractive is something else with which it is confused, or is something meretricious, which when isolated and seen plain is no longer attractive. But once all the whittling is done, one must just make up one's mind: is accomplishment, now that it is separated off and seen plain, worth going for? It is not that this is the end to all possible argument. It is just that, once the whittling is done, the argument has to become more radical. Accomplishment, like all prudential values, connects with many other values: what is accomplished has, by definition, to be worthwhile: it has to be the sort of thing that gives life point or substance. But then, to ask a radical question, does anything give life point? When we step far enough back from our everyday concerns, when we see life sub specie aeternitatis, is anything worthwhile? We do not even need the view sub specie aeternitatis; the view de lecto mortis is enough. But from death-beds a few things also look more important than ever. The arguments that nothing is worthwhile, though certainly important, seem all to rest on one confusion or other. Why is it all worthless? Because sooner or later we shall die and be forgotten, was Tolstoy's answer before returning to Christianity; therefore nothing can come of life, nothing is worth doing, there is no true accomplishment.16 Because the constant increase of entropy ensures "the vast death of the solar system', Russell added.17 However, if merely my death makes my life meaning- less, eternal life would restore its meaning. But if it would, then my life itself is not the problem. Why should more time allow something to come of life? If something can come with more time, it can come with less. Mere time cannot make the difference. And if it is not eternal life that answers but only eternal bliss, temporal bliss ought to answer a bit.18 These steps back provide no new perspective. Radical argument about values is certainly possible,19 but one can wonder whether all of the steps back that it might come up with are not also, like this sample of them, steps into confusion, steps off the edge of coherent talk about values.20 67 PERFECTIONISM AND THE ENDS OF LIFE The final results of deliberation of these sorts---the proposals that emerge from the definitional sort, and the lack of effective challenge from the radical sort---are, I think, a list of prudential values something like this. (a)Accomplishment. (b)The components of human existence. Choosing one's own course through life, making something out of it according to one's own lights, is at the heart of what it is to lead a human existence. And we value what makes life human, over and above what makes it happy. What makes life "human', in the distinctly normative sense the word has here, is not a simple thing. The systematic way to understand its complexities is to understand the complexities of "agency'. One component of agency is deciding for oneself. Even if I constantly made a mess of my life, even if you could do better if you took charge, I would not let you do it. Autonomy has a value of its own. Another component is having the basic capabilities that enable one to act: limbs and senses that work, the minimum material goods to keep body and soul together, freedom from great pain and anxiety. Another component is liberty: the freedom to read, to listen to others; the absence of obstacles to action in those areas of our life that are the essential manifestations of our humanity---our speech, worship, and associations. To value our humanity is to value all of these components. (c)Understanding. I have already argued that practical rationality is not one of the things that constitutes a good life, but that it is, rather, the capacity that allows us to identify what makes it good. Still, even putting practical rationality aside, there is something important left. Simply knowing about oneself and one's world is part of a good life. We value, not as an instrument but for itself, being in touch with reality, being free from muddle, ignorance, and mistake. (d)Enjoyment. We value pleasures, the perception of beauty, absorption in and appreciation of nature---the enjoyment of the day-to-day textures of life. (e)Deep personal relations. Personal relations can make life worse, or not affect its quality at all, but when they become deep, authentic, reciprocal relations of friendship and love, then they have a value apart from the pleasure and benefit 68 MEANING they give. They fit Aristotle's model of the human ergon better even than his own candidate, rationality; they in themselves go a long way towards filling and completing life.21 This sort of deliberation about ultimate values, about ends, is consistent with any major moral theory. It is a misunder- standing to think, for instance, that utilitarianism limits deliberation to means-end. The conception of a self free from the inexorable working of appetite and taste, capable of assessing ends and choosing those that, by their nature, deserve to be chosen, is available to all major moral theories. Conceptions of the self do not serve to distinguish the good theories among them from the bad.22 4.How morality fits into prudence Do ethical push and ethical pull eventually meet? Generally, the more mature one's prudential values are, the more important among them is living morally. Perhaps the end towards which this maturation approaches (not asymptot- ically) is the intersection of ethical pull and ethical push. Perhaps the very finest life for a person to live, from his own personal point of view, however testing his circumstances, is a fully moral life. Certainly the most secure and stable condition for morality is when a person's conception of the right is rooted, not in inhibitions and self-denial, but in flourishing and self-fulfilment. And a sense of moral right belongs to any sensible person's good: peace with oneself and others, stability and richness in personal relations, fulfilment, even agreement with nature---certainly with human nature, possibly even with a moral reality. But I doubt that ethical push and ethical pull will ever meet. Sometimes the idea that they do is just a confusion. Moral reasons do trump purely prudential reasons; morality does eventually take over. But that does not mean that push and pull meet or that moral reasons are generated out of prudential ones. It shows only that once our various reasons for action---prudence, morality, etiquette, and the rest---are brought together in a single hierarchy of reasons for action, moral reasons come out on top. But we want to know something different. When we organize the various things that 69 PERFECTIONISM AND THE ENDS OF LIFE make an individual life better into a hierarchy of prudential values, does morality then come out on top? Does it, in cases of conflict, trump, perhaps even extinguish all other prudential values?23 So the strong argument we need, which indeed many will make, runs something like this. There is no coherent concept of good in conflict with right; right is prior to good; good has sense only within limits laid down by right. Therefore, it would make no sense to say that morality required a sacrifice of good. Suppose I come upon several children trapped in a blazing building. Suppose, too, I ought to try and save them (for whatever reasons are strong enough to create an obligation: it is one life against many; I am risking fewer years than they have ahead of them; I am responsible for them). I might reason: I could well die in the attempt; my life---if I funk doing what I ought---would still be a lot better than no life at all; by funking it I would, of course, suffer terrible guilt and remorse, but I would eventually face up to them and go on living what altogether would be a good life. But this strong claim means that I cannot intelligibly reason like this. The moral failure would make it impossible for it to be a good life; it could not be a better life than one without the moral failure. It could not, no matter how small the moral failure, and how great the disaster to me.24 It is an extravagant claim, and one that again, I think, rests on a confusion. We need to split the notion of "a good life' into two. There is a sense in which moral failure, being a failure to act for the best reasons, is a falling off from an ideal---and not just in the trivial, circular sense that it is not the most moral or most rational life. It is not the finest life: the life one would hope to lead. But there is another conception of a good life, a life one would hope to lead. It is the sense that appears in judgments such as that it is better to be moral and alive than to be moral and thereby lose one's life, or that it is sometimes better to fail morally and stay alive than not to fail and thereby lose one's life. And it is this second conception that should be the base for judgments of well-being in moral theory. This is not at all paradoxical. We want as the base for at least some moral judgment considerations of quality of life that are restricted to the prudential values on my list. And 70 MEANING being moral enters that list only in a limited way: only by being part of what it is to be at peace with one's neighbour and with oneself. This sort of peace is a prudential value, and when morality enters consideration under that heading it takes on prudential weight. But as a purely moral considera- tion, not subsumable under one of these headings, it has no prudential weight. We do not want to lose this prudential notion of a good life by merging it altogether with morality. And we do not for reasons of moral theory; it needs the notion. 5.What is the good point buried in perfectionism? The good, almost unavoidable, point in perfection is this. There are prudential values that are valuable in any life. There are not enough of them, nor is a specific balance between them prescribable universally enough, to constitute a form of life. They are the values on the list of the ends of life. As we have seen, our skills and capacities differ too much for all of us to want to realize them in quite the same balance. Indeed sometimes they should not be realized by a particular person at all (e.g. autonomy, when it sets up great anxieties in a particular person). Persons differ not so much in basic values as in their capacity to realize them. Some basic values (e.g. enjoyment) depend in part on persons' individual, varying tastes, but not many depend on taste. The values on the list of the ends of life provide us with an important standard for judging most (ordinary) human lives. They let us say, roughly, though only roughly, how good the life is and how it could be better. They do, because they rest on general features of human nature; for example, autonomy is central to living a human existence, and we all value our humanity.25 Finally, these values are not restricted to all-purpose means to a good life. They include accomplishment, deep personal relations, and the enjoyment of beauty, none of which is a means to the realization of one's own particular conception of a good life, but is, or should be, a central part of that conception. This is the unavoidable form of prudential perfectionism. It has a derivative moral perfectionism: if certain prudential values are valuable for anyone, why ever not promote them? We ought to go out and, in missionary spirit, lead persons to 71 71 PERFECTIONISM AND THE ENDS OF LIFE them. Certain values---understanding and enjoyment of beauty, say---enrich lives, and in promoting them a society would lead its members, if not in this generation then in later ones, to better lives. Is this paternalism? That all depends on how far one's missionary zeal carries one. One can be non-neutral between different conceptions of the good life (e.g. putting forward the case for education and art, making access to them easy) without being paternalistic. Paternalism is not the central issue about neutrality. Anyway, this form of moral perfection- ism has built-in checks on paternalism. A particularly stringent one is that one heavyweight prudential value is autonomy; any promotion of other prudential values must respect this (non-absolute but still important) one. We are not in a position yet to carry the discussion of these issues much further. It would take a fairly fully developed moral theory to say precisely how far we can go with this form of moral perfectionism. But even the very modest degree that I am broaching now seems to offend the neutrality of a lot of modern "liberal' political theory.26 Society, that theory says, should stay neutral between values, except for a few, special, all-purpose means. That too raises issues that need the context of a moral theory to settle finally. But I suspect that sometimes neutrality looks attractive because of certain beliefs simply about the nature of values. For instance, neutrality is suggested by the need account of well-being. But, as we have seen,27 "basic needs' do not go deep enough to yield a satisfactory account of well-being, and when we go deeper, we encounter values that no longer support neutrality. Neutrality is also suggested by a wide- spread belief about the difference between all-purpose means on the one hand and all the rest of the prudential values on the other. The belief is that once outside the realm of objective all-purpose means we are on the shifting grounds of mere preferences---tastes and subjective attitudes. Sometimes poli- tical theorists write as if the diversity of conceptions of the good life is ineliminable,28 as if due to the diversity of tastes, which can itself be eliminated only in New Men, genetically engineered and regularly brain-washed in a Huxley "Central London Hatchery and Conditioning Centre'. But this greatly 72 MEANING misjudges the rest of the prudential values. They are not, in any simple sense, subjective. They are not a matter of taste. They too are, in an important way, objective goods.29 6.The primacy of prudential value theory To get an account of well-being that would be of use in moral theory we have to move beyond forms of perfectionism (even my modest one) and on to what is valuable to the particular persons affected in each case we judge. For special reasons about Jessica, autonomy might not be the thing for her to go for. For Nicholas, totally absorbed in his work, with no taste for day-to-day pleasures, enjoyment may count for little and accomplishment for a lot. For Edward, risk stimulates; for Sarah, it is off-putting. We cannot make any useful judgments about how to act so as to promote well-being without having a notion of well-being that captures all of these variations. We found with need accounts that, to give them the required flexibility, we had to go beyond the notion of a "basic need' to the notion of its "importance', which led us to a form of judgment that could take place only in a full prudential value theory. Now we find with perfectionism that it gives us no useful account of well-being unless it is incorporated in a full prudential value theory. If we are interested, as we ultimately are, in the conception of well-being needed by moral theory, it seems that it must be the one supplied by the prudential value theory. Of course, the pressures of building a moral theory may put things in a different light. But the pressures would have to be very great to get us to abandon that conception. The end of the argument comes only when we have built an adequate moral theory with one conception or other of well-being. Short of that, all conclusions are tentative. But in this spirit we can draw a conclusion. Well-being is the same in prudence and morality. In any case, the conception of well-being that has emerged in this Part of the book has important consequences both for measuring well-being and for the place it might occupy in moral thought, to which I now turn.