‘It’s pretty hard to tell what does bring happiness: poverty and wealth have both failed,’ says Kin Hubbard’s creation Abe Martin. Since the pursuit of ‘the greatest happiness of the greatest number’ has proved so perplexingly difficult in practice, discussion of the philosophical foundations of utilitarianism can easily appear to be otiose. Sometimes it seems obvious that the pursuit of general welfare is what morality and public policy must be all about now that the theological underpinnings of traditional morality have crumbled, and a waste of time to argue over that when the technical obstacles are so pressing. At other times the very elusiveness of welfare, happiness and related notions can make it seem equally obvious that the whole enterprise is shot through with hubris.
And yet utilitarianism continues to provide an inexhaustible talking-point for philosophers, as well as those with professionally grubbier knees, such as economists, whose policy prescriptions frequently rest on theories that bear resemblance – in varying degrees – to its currently popular versions. Even Bernard Williams, who ten years ago incautiously expressed the hope that ‘the day cannot be too far off when we hear no more of it,’ has bowed to the inevitable and edited, with Amartya Sen, this substantial collection of essays by philosophers and economists. And although the title and the balance of the contributors’ sympathies might lead one to suspect a last psychotic attempt to kill the theory off for good, its overall effect is to leave one with the impression that many of the important issues raised by utilitarianism are only just beginning to be explored.
In spite of the Science Fiction resonances of the title, the contributors to Utilitarianism and Beyond with one or two exceptions concentrate on re-surveying territory previously charted (sometimes by themselves), rather than on setting off for the unknown. But the territory now looks much broader – and more full of pitfalls – than it did even ten years ago, and is more or less unrecognisable as the terrain explored by Bentham and Mill. The view that utility consists in pleasure and the absence of pain, or even (G.E. Moore’s view) in the presence of mental states of intrinsic worth, has very much fallen from prominence. Utility as now conceived is grounded firmly in the fulfilment of desires, and the dispute among utilitarians is chiefly about how that is to be understood. Sophisticated utilitarianism, like sophisticated architecture, incorporates split-level design; complex theories of multi-level desires have been developed. The links between utilitarianism and theories of rights and justice have been explored. And the gleaming technology of social choice theory, decision theory and the economics of information has been put to work (though the formal apparatus is kept discreetly in the background for most of this book).
The fact that utilitarianism can now appear in a variety of sophisticated guises makes it hard to know exactly what it consists in. How far are we justified in refining it before it ceases to be properly called utilitarianism at all? This is not a terminological quibble, but a question about whether utilitarianism represents a distinctive way of thinking about morality. All the contributors implicitly agree that it does, but many of them take the nature of its distinctiveness for granted, which can be confusing since they do not seem to agree on what it is. The main exception is the editors’ own introduction, which tackles this crucial question directly, and also provides a valuable brief survey of most of the main areas of contention in the subject as a whole. ‘Utilitarianism,’ Sen and Williams remind us, ‘can be regarded as the intersection between two different kinds of theories.’ Consequentialism is the theory that ‘actions are to be chosen on the basis of states of affairs which are their consequences’; welfarism is the theory that states of affairs derive value from ‘welfare, satisfaction, or people getting what they prefer ... Utilitarianism is thus a species of welfarist consequentialism – that particular form of it which requires simply adding up individual welfares or utilities.’
So far so familiar. But how elastic can this definition be made? Clearly much hangs on the way welfare is conceived. The editors discuss three devices they claim to be characteristic, to varying degrees, of utilitarianism. These are reduction: ‘the device of regarding all interests, ideals, aspirations and desires as on the same level, and all representable as preferences’; idealisation: the use, not of individuals’ actual preferences, but of cleaned-up versions (those corrected for faulty information, for instance, or purged, as John Harsanyi requires, of ‘sadism, envy, resentment or malice’). Finally there is abstraction: the tendency to ignore as of secondary importance the practical application of the theory, particularly by asuming that the preferences on which utilitarianism operates are in some way both given and known independently of the process of utilitarian action itself.
All three of these devices are examined and criticised in relation to utilitarianism conceived both as a theory of personal morality and as a theory of public choice. The first two I shall discuss shortly; the third, abstraction, comes under heaviest fire in Jon Elster’s essay, in which he points out that the opportunities open to individuals can act to change their preferences. The most important forms of this are adaptive preference change (the ‘sour grapes’ effect) and counter-adaptive change (the ‘grass is always greener’ effect). If preferences are thus contingently dependent upon the actual state of affairs, the satisfaction of these preferences seems a much less obvious candidate for the goal of morality: and a utilitarian might find that changing preferences to match possibilities rather than vice versa was a great deal easier;
This of course raises the old spectre of Brave New World. But Elster’s careful treatment of the problem leads him to suggest a mechanism for getting at the true ‘autonomous’ preferences by weeding out the dubious ones. I must admit to being less sanguine than he is about the possibility of providing a formal distinction of this kind, particularly since the more successful his mechanism is at weeding out adaptive preference changes, the more likely it seems to be to admit counter-adaptive ones. But even if that is so, it still makes the issues he raises of fundamental importance.
The discussions of abstraction make the question of the distinctiveness of utilitarianism inescapable. Several writers point out that in realistic applications the pursuit of utilitarian consequences may well require the abandonment of any attempt to apply utilitarianism very directly. It is not quite clear what the import of this is supposed to be: the editors clearly think it an embarrassment to the theory, while others, like Hare, find it a positive virtue. Still others never resolve the uncertainty, which can leave the thrust of their essays a little obscure: both Frank Hahn and Partha Dasgupta, for instance, argue that utilitarianism is compatible with respect for individual rights, Hahn because of the utility which people derive from the fulfilment of their rights and Dasgupta (more persuasively, perhaps) because of constraints on the information available about individual abilities. But in both cases it remains unclear whether these considerations are supposed to make theories of intrinsic rights appear redundant, or whether the authors are making the (important but probably uncontroversial) claim that utilitarianism in practice will be less manipulative than it appears in grim lack of detail.
Reduction and idealisation can be thought of as representing, respectively, the aggressive and defensive strategies of utilitarianism: reduction to preferences is the foundation on which utilitarianism is constructed, and the idealisation of those preferences is the way in which several immediate and awkward objections are countered. Both processes are best exemplified in the three essays that give the theory their full support: those of Richard Hare, John Harsanyi and James Mirrlees. The first two are the only reprinted articles in this collection, and represent two by now well-known and influential approaches to justifying the theory. Hare’s essay expounds the derivation of utilitarianism from his theory of universal prescriptivism, according to which moral judgments are characterised by universalisability, which is really another way of talking about impartiality. Impartiality requires giving equal weight to the interests of all those affected by an action, and this, it is claimed, is tantamount to utilitarianism.
Harsanyi’s article attempts to derive utilitarianism from the Bayesian approach to rational decision-making. His arguments amount to considering what an individual would decide to do if he were ignorant of the position he would occupy among those affected by the action concerned, and assigned himself an equal probability of occupying each such position. This heuristic device is similar to the ‘veil of ignorance’ used by Rawls in A Theory of Justice (though Rawls uses it to reject utilitarianism).
Powerful though the device is, it is far from obviously the only compelling way to think about morality, and an expanded defence of this would greatly strengthen Harsanyi’s case. In general, the essays of both Hare and Harsanyi have a striking elegance and power, but both take on board assumptions that may be far less innocuous than the authors allow. Both, for instance, make consequentialism more or less true by definition. Both assume that the distinction between personal and moral preferences is unproblematic. And both effectively rule out the possibility that decision-makers might set some store by equality of utilities (as opposed to incomes). It is also worth remarking that Hare’s reasons for thinking that utilitarianism will tend to prescribe equality of incomes appeal to the fact that ‘inequalities tend to produce ... envy, hatred and malice’, which are precisely what Harsanyi would exclude from relevance to the utilitarian calculation, as being incompatible with the ‘general goodwill and human sympathy’ which is ‘the fundamental basis of all our moral commitments to other people’. This suggests that some of the defences which utilitarians adopt, even if individually compelling, may not always be compatible with one another: energies devoted to a face-lift may leave other parts of the corpus in awkward shape.
The doctoring of preferences is dealt with at length in Mirrlees’s essay, which is very clear and explicit about the assumptions involved in a certain kind of utilitarian approach. He argues the desirability of grounding the suspiciously unobservable notion of utility in an account of choice. What people prefer is defined as what they would choose under certain ideal circumstances. In normal cases, what people choose will differ widely from what will most serve their own welfare, because they have false beliefs, because they are sometimes irrational and because they are sometimes committed to other things than their own interests. So Mirrlees proposes that we use ‘what people would do if they could succeed in conforming to the simple rational-choice model ... as a standard for judging what is best for them individually’. But as he himself points out, once we start doctoring preferences it is hard to know where to stop. ‘It must be possible to allow for convictions about what is good for one that, though unshakable, are nevertheless mistaken.’ This more ‘objective’ notion of welfare is kept in tenuous contact with actual choices by requiring that ‘with full understanding’ the individual ‘would come to accept the rightness of the altered utility function or rather of the underlying preferences’. The problem here is not just the rather mystical role played by the notion of ‘full understanding’ (how can we recognise it, for instance, other than by chauvinistically requiring that individuals share our own idea of what is good for them): it is also that the explanatory value of defining utility via choice seems much reduced if the appropriate notion of choice must itself be established by appeal to an ‘underlying’ utility function: to conform to the simple rational-choice model is to maximise the expected value of one’s utility. The process is not entirely circular, for the ‘full understanding’ clause is clearly intended to rule out preferences which an individual would not assent to under any circumstances (otherwise a Nazi might claim that the Holocaust had been tailored to the ‘true’ underlying preferences of Jews for self-destruction). But the sense in which utility is made observable by this method is a rather weak one. This could be seen as the result of unnecessarily restricting (to choice behaviour) the range of evidence that is relevant to establishing what people’s desires are.
Also I suspect that opponents of utilitarianism (like Charles Taylor and T.M. Scanlon in this volume) would reply that this account of utility is so elastic as to make us all utilitarians now. Few people would seriously dispute that moral goals worth pursuing must be acceptable to us at some underlying level of preference, or that welfare in its traditional sense is of some importance in morality. But this raises a deeper point. Once we take seriously the rich texture of human desires, as utilitarians are surely right to do, then we must accept that such desires are interrelated in very complex ways. Some are desires about what desires we should have, some are about what kind of society we should live in, or about what conceptions of morality we should develop. Many are fundamentally incompatible with one another. Not only is the relation between these desires and our own welfare a very perplexing one, but the project of adding them up as utilitarianism requires is formidably difficult. It is fair to say that non-utilitarians have been no clearer than utilitarians about how a successful moral theory might copy with these complexities.
Some hints can perhaps be gleaned from Stuart Hampshire’s essay ‘Morality and Convention’. He takes utilitarianism to task for failing to respect the distinction between moral requirements that are ‘to be recognised by reason, and as founded in the nature of things, as not essentially diverse, and as not contingent upon any specific type of social order’, and those that are local, distinctive, dependent upon a particular culture or society, maintained by convention, and essentially ‘non-converging’. This suggests that some individual desires and preferences may derive their source from attachment to a whole culture or way of life (which must be evaluated in its entirety) rather than the other way round. This opens up many possibilities, but Hampshire is vague about perhaps the most interesting question of all: namely, what relation these two forms of judgment bear to one another. Are not judgments of convention subject to some kind of scrutiny by the claims of reason, particularly in the hard cases where the requirements of conformity begin to hurt individuals or minority groups? Unless a fuller account is given, it is hard to distinguish convention from repressive snobbery at one extreme, and from mere convenience at the other – like having lunch at the same time each day. Neither is the least threat to a utilitarian. However, Hampshire’s theme seems closely related to the objection that Williams has raised to utilitarianism in the past – that it cannot make sense of individual integrity. This suggestion I have found intriguing but rather obscure. Hampshire’s appeal to the integrity of a way of life may help to elucidate it.
Bernard Williams’s previous criticisms of utilitarianism have taken the form of a sustained bombardment on several fronts, challenging not only its specific prescriptions but its very approach to moral thinking: Amartya Sen by contrast has tended to adopt what one might call the Exocet tactic – he has launched single explosive counter-examples that have to be avoided if the theory is to survive. One such is his objection to the claim that utilitarianism will tend to prescribe equality of incomes: what of the handicapped, he says, who require greater quantities of resources to attain any given utility level? Under utilitarianism, such people would receive even lower levels of resources than others – a conclusion which seems strongly offensive to our intuitions about justice. The only defender of utilitarianism here who even picks this missile up on the radar is Mirrlees; he takes vigorous evasive action, even sending up a decoy in the argument that such instances are very difficult to verify. He concludes that in certain hard cases the admission that we do not know the answer may be appropriate.
One thing which becomes clear at a number of points is that traditional utilitarianism has defined preferences, not for particular objects or events, but for fully specified alternative states of the world. These, of course, are just the same as the modal logician’s ‘possible worlds’, which are not a dish to suit every philosophical palate, and may not, after all, be the best context in which to discuss preferences. Frederic Schick urges that preferences are defined, not for states of the world, tout court, but for such states under particular descriptions. Thus someone may derive utility from the coming of sunny weather, and disutility from the beginning of the hay-fever season, and can do so even when these are the same event. Schick’s essay raises deep questions without, perhaps, exploring them as thoroughly as one might wish. I have a similar feeling about Rawls’s essay, which develops his theme that the proper province of social justice lies in attending, not to the desires as such of its citizens, but to the primary goods (including the ‘social bases of self-respect’) that provide the capability to participate in a just society. ‘The fact that we have a compelling desire does not argue for the propriety of its satisfaction any more than the strength of a conviction argues for its truth.’ This as it stands is a suggestive analogy, holding out the hope that the move from a private towards a more public perspective may yield valuable insights into political philosophy, as it undoubtedly has done in epistemology and the philosophy of language. Equally interesting are his arguments that a theory of justice (and especially of capabilities) relies importantly upon a particular conception of the person. But in both cases the argument is very lightly sketched.
Undoubtedly the finest and most original essay in this collection is T.M. Scanlon’s ‘Contractualism and Utilitarianism’. He argues that the main reason why utilitarianism ‘occupies a central place in the moral philosophy of our time’ is that it rests on a well-articulated view about the subject-matter of morality: ‘namely the thesis that the only fundamental moral facts are facts about individual well-being’ – a view he labels ‘philosophical utilitarianism’. A successful alternative to utilitarianism must provide ‘a clear account of the foundations of non-utilitarian moral reasoning’; and this he proceeds to do. ‘An act is wrong,’ he proposes, ‘if its performance under the circumstances would be disallowed by any system of rules for the general regulation of behaviour which no one could reasonably reject as a basis for informed, unforced general agreement.’ I wish I had space to show how powerful a tool this becomes in Scanlon’s hands. A full development of this idea, however, will undoubtedly have to cope with some difficult problems. For example, the exact nature of the moral reasoning underpinning the contract requires elucidation: just what are ‘reasonable grounds for rejection’? Secondly, the theory will have to try and explain the scope of morality: whom is it proper to bring within the compass of our moral concern? Creatures that can feel pain all fall within it, according to Scanlon, because ‘if a being can feel pain, then it constitutes a centre of consciousness to which justification can be addressed’: yet surely the ability to feel brute pain is more primitive than the reflective and self-conscious capacity to ask the question ‘why?’
What I miss most in this collection is an account of what a full, subtle and specific theory of human welfare would look like – a psychology of welfarism. The editors write that the use of utilities alone in making moral judgments ‘amounts to taking ... a remarkably narrow view of being a person’. If the utilities they refer to are those defined in the way characteristic of economics, I could hardly agree more – but that may just be the result of taking a remarkably narrow view of utility. It is ironic that the word ‘happiness’ should now be so unfashionable in utilitarianism, largely – I suspect – because of its associations with the crudity of the hedonistic and mental state versions of the theory. Yet to understand happiness, and what it is for human beings to flourish, will require a much more sophisticated insight into the human personality than that afforded by the simple definitions that utilitarians have come up with so far.