2. UTILITARIANISM AND CONTRACTUALISM
In this chapter I shall examine two theories (or classes of theory) that stand some chance of proving acceptable, both in respect of the explanations provided by their governing conceptions of the source of moral notions and moral motivation, and in respect of their basic normative output. The theories are, namely, utilitarianism and contractualism.
Utilitarianism and its governing conception
Although utilitarianism is named after its basic normative principle (the principle of utility), and although not all utilitarian thinkers have felt the need to provide their theory with a governing conception, I shall begin my discussion by outlining what I take that conception to be. The main idea is that morality may be viewed as the set of decisions that would be made by an impartial benevolent observer - an observer who is aware of all the conflicting interests in a given situation, and of the consequences that different policies would have for those interests, and who is equally sympathetic towards all of the parties involved. The governing conception of utilitarianism is thus an imaginary construction (as is the governing conception of contractualism, as we shall see later). The moral point of view is a sort of God’s eye view, but independent of any belief in an actual God. It is the point of view that we would take if we could be fully aware of all the consequences of our actions, and could be equally sympathetic towards all those affected.
Plausibly, what an impartial benevolent observer would always choose, is the option that would have as a consequence the greatest happiness, or desire satisfaction, or pleasure. (These formulations are not exactly equivalent, and lead to rather different versions of utilitarianism. But the differences will not concern us, for the most part. Throughout most of this book I shall use ‘utility’ in a consciously ambiguous way, to equivocate between the different possible ultimate values of utilitarians, only being more precise when something of importance turns on the issue.) So the most basic normative principle of utilitarianism is that one should maximise overall utility (whether in total or on average - again there are differences here that need not concern us). For observers who are truly impartial will be indifferent as to whose utility is in question, and being truly benevolent, they will wish to do as much good as they can.
Most of the problems with utilitarianism arise from its basic normative principle, as we shall see shortly. In contrast, the greatest attraction of the theory lies in its governing conception. This can provide us with a satisfying explanation, not only of the origin of moral notions, but also of the source of moral motivation. Both explanations start from an hypothesis of natural (though limited) benevolence. This involves, like the contractualist accounts we shall consider later, a claim about an innately given aspect of human nature. The idea is that human beings will naturally feel an impulse of sympathy for the unhappiness, frustrated desires, or sufferings of other humans. Such an impulse will only actually occur, of course, when other things are equal - for example, in the absence of prior hatred for the individual in question. It will also only occur, in the first instance, in response to more or less direct confrontation with the suffering of another - as when one sees someone crying with grief over the dead body of a loved one, or observes them moaning from the pain of a broken leg. At any rate, it is surely very plausible that a normal person will feel sympathy in such a situation.
The next step in the explanation, is that the natural impulse towards benevolence should become rationalised - developed and extended through the impact of rational considerations. For there is, plainly, no rational difference between the suffering of someone whom one can see, and the suffering of someone whom one does not. Nor is there any rational difference between the suffering of someone whom one knows and cares for, and the suffering of a stranger. Each of these types of suffering is equally real, and may be of equal severity. Reason then demands that one should respond equally. Hence the picture of the impartial observer - one should be equally sympathetic towards the sufferings of all those who suffer equally.
According to utilitarianism, then, morality arises in the first place when the natural impulse towards benevolence is universalised through the impact of reason. And then the source of moral motivation is that we cannot, by our very nature, avoid the sympathetic impulse; nor can we, in so far as we are rational, avoid its universalisation. These are very plausible explanations, avoiding the excesses of strong objectivism and intuitionism.
Problems with utilitarianism
While the main strength of utilitarianism lies in its governing conception, its main weakness lies in its basic normative principle, the injunction to maximise utility. This gives rise to a number of notorious problems. The first is that it entails very counter-intuitive solutions to questions of distributive justice. Since all that utilitarianism regards as mattering, in the end, is total (or perhaps average) utility, the intense sufferings of a few can in principle be justified in terms of the marginal benefits of many. For example, one of the standard objections to utilitarianism is that it might be forced to legitimate a system of slavery, if the total number of slaves were few and the benefits to the slave-owners were sufficiently great. Yet we would surely want to insist against this, that it is only the worst off individuals who should be taken into account. (This intuition will be captured by the so-called difference principle of contractualism, as we shall see later.) Unless it can be shown that the slaves themselves must inevitably be even more degraded and unhappy under any realistic alternative to slavery, the system will stand morally condemned.
Consider a less dramatic example to make the same point. Suppose that Daniel is a doctor who possesses only a limited quantity of a very expensive drug. Unusually, this drug has a dual use - in very small quantities it can permanently cure acne, and in very large quantities it can cure some life-threatening disease. Daniel may therefore face a choice between removing a major source of unhappiness for many hundreds - perhaps thousands - of adolescents, and saving the life of a single individual. Here even the simplest of utilitarian calculations may show that the drug ought to be used as a cure for acne, allowing the individual in question to die. But this is hugely counter-intuitive. I think we should want to insist that the sufferings of the adolescents, although perfectly real, simply do not count when set against a person’s life. (This does not imply that doctors ought never to prescribe asprin for a headache, or to set a broken finger. In real life it is rare for them to be faced with choices as stark and clear as Daniel’s.)
Utilitarians do have various avenues of reply to the problem of distributive justice. The most plausible of these is to appeal to some sort of two-tier system of morality, two rather different versions of which will be outlined shortly. But first let me introduce another traditional objection, which is that utilitarianism, in contrast with contractualism, cannot provide adequate protection for individuals. Apparently anything can be done to a person, provided that it produces more utility (either total or average) than any alternative course of action. One famous version of this objection is that a utilitarian may be forced to condone the judicial punishment of the innocent.
Consider the following example. Some years ago there was a series of murders of young black men in Atlanta, evidently performed by the same hand. The killings were thought to have a racial motive, and there were persistent rumours of police connivance in failing to catch the killer. These rumours sparked a series of race-riots, in which many people were killed and injured. Now suppose that Prunella, the state prosecutor, has come to believe that the killer was in fact white, but is now dead. Her evidence is not of such a nature to convince a court of law, however, or to bring an end to the riots. Now by some chance she stumbles across evidence that could easily be used to frame a particular black man. (She also has conclusive evidence that he was not in fact the killer, that she could easily suppress.) What should Prunella do? As a utilitarian, it seems she should frame the man, resulting in one execution or life sentence, thus putting an end to the riots and saving many lives. But this seems intuitively abhorrent.
Notice that two of the considerations to which utilitarians often appeal, in trying to explain why judicial punishment of the innocent is wrong, have no application in this example. For since Prunella is the only one who will know that there has been a miscarriage of justice (aside from the person she frames, of course), there need be no long-term weakening of respect for the rule of law resulting from her action. Nor need the weakening of her own disposition towards the impartial application of the law have any bad consequences in the future. For we can imagine that she is on the point of retirement, and will henceforward have no role as an officer of the law.
Some utilitarians have tried to respond to these difficulties by retreating from the act-utilitarianism we have been considering so far, to a version of theory sometimes known as rule-utilitarianism. The idea is that, instead of actions being judged directly in terms of the utility of their consequences, actions should be judged by their conformity with a set of rules, and it is the rules that are justified by appeal to their general utility. These utilitarians have thus claimed that the rules outlawing slavery and forbidding the judicial punishment of the innocent are generally good ones, compliance with them producing the most utility in almost all cases. Then the actions of instituting a system of slavery, and of arranging for the punishment of an innocent man, are wrong because they conflict with rules justified by appeal to utility, even though in the particular case the actions in question may produce more utility than any alternative.
There can be no reflective equilibrium here, however. For rule-utilitarianism is an unstable position. There are two reasons for this. The first is that there is a problem about selecting the right rules (that is, the rules that accord with ordinary belief). For in place of the rule forbidding judicial punishment of the innocent, why should we not adopt a rule that forbids such punishment except where the utility of doing so is very great and except where it can be known that there will be no weakening of respect for the law as a result? It certainly looks as though this rule would be productive of more utility in the long run, which lands us back with the original problem.
The second reason why the position is unstable is even more devastating. It is that rule-utilitarianism, as a basic normative principle, appears inconsistent with the governing conception of utilitarianism. For why would rationalised sympathy lead one, in the first instance, to justify a system of rules? It is the sufferings of individuals that matter, surely. Indeed, rule-utilitarianism appears vulnerable to the charge of unmotivated rule-worship. For in a case where I can see clearly that breach of a rule would produce more utility than would compliance, what possible utilitarian justification can there be for insisting on compliance? If utilitarianism is to find a position of reflective equilibrium with ordinary belief, it needs to pursue a different strategy.
The way forward for utilitarians is to claim that the primary object of moral assessment is neither actions nor rules, but qualities of character. This is in fact the view of John Stuart Mill in Utilitarianism (1863), though Mill is often mistaken for a rule-utilitarian. It is also the view developed by Richard Hare in Moral Thinking. The idea is that our primary duty is to develop qualities of character - dispositions of thought and feeling - whose possession is likely to produce the greatest overall utility. Such an account would be supposed to be grounded in a more realistic view of human rationality, and of the springs of human action, than is presupposed by act-utilitarianism.
It is often remarked that act-utilitarians seem to regard human beings as mere calculating machines. Their picture of moral agency requires that people should be constantly calculating the likely consequences of the various actions open to them on any given occasion. But this is unrealistic. Besides powers of rational calculation, we also have emotions and desires, and these are likely to exert a profound influence on our judgements of right and wrong. We also frequently find ourselves in situations where we are required to act quickly, without time for careful assessment of consequences, or estimates of resulting utility. The very least that can be said is that appropriate dispositions of thought and feeling are likely to make a very great difference to our effectiveness as utilitarian agents.
More importantly, since moral judgements are so susceptible to influence by emotion and desire, we may argue, on utilitarian grounds, that moral agents should develop dispositions of thought and feeling that exclude certain options from consideration. For example, it may be important that our law officers should never even think of breaking the rules, since they may otherwise be tempted into breaking the law in cases where they should not. Then in the case of the Atlanta murders discussed earlier, a utilitarian may condemn Prunella for framing the man, because of the weakness in her moral character thereby displayed. This apparently accounts for our common-sense judgement, while allowing that the action did, in the circumstances, turn out for the best.
Compare, here, the utilitarian attitude to adultery. A utilitarian may be expected to argue that in principle some acts of adultery are permissible (because productive of more utility than would otherwise occur), but that in general adultery is wrong because of its likely harmful consequences. A quality-of-character-utilitarian will argue further, that it will be best to develop an attitude in oneself such that adultery is never a serious option. For the circumstances in which the question of adultery may arise are likely to be ones in which we are in no fit condition to weigh up the likely consequences, because blinded by passion. A utilitarian may then condemn an act of adultery, even though it was in fact productive of utility, on the grounds that the agent has thereby displayed a faulty character - adultery should never have been considered.
Quality-of-character-utilitarianism may be successful in meeting at least some of the problems utilitarians face in connection with justice. It is by no means obvious, however, that it adequately accounts for our intuitions in the case of Prunella, the prosecutor. For her character failings will not, in fact, lead to bad consequences in the future (recall that she is on the point of retirement). So while utilitarians can condemn her for having failed to become the sort of person she should have become, they still have no basis for condemning her action in this case. On the contrary, they still seem committed to the unpalatable conclusion that her action is, in the circumstances, the right one. Moreover, it is, in any case, doubtful whether there will always be appropriate qualities of character to hand, to reconcile utilitarianism with common-sense judgement. For example, in the case of Daniel, the doctor, discussed earlier, what weakness of character would be displayed if he were to use the drug to cure many cases of acne rather than to save a life? The only plausible candidate is his very preparedness to weigh up minor benefits to many people against major harm to an individual. But then it seems we cannot, in utilitarian terms, explain why this would be a bad attitude to have - on the contrary, it is just what utilitarians think we ought to do.
The demands of morality
The move to quality-of-character-utilitarianism is also powerless in the face of the other main charge made against utilitarianism - that it is too demanding. Notice, to begin with, that an act-utilitarian holds that there is only one duty, that is binding at all times, namely to maximise utility. But it seems, intuitively, to be too much to ask that I should run around doing good all the time. There must surely be some private space left to me, in which I can be free to pursue my own projects, and attend to my own interests and the interests of those whom I love.
There are, in fact, two rather different objections here, that are not often distinguished from one another, one theoretical and one practical. The theoretical objection is that utilitarianism is committed to a two-fold classification of actions into those that are duties (morally required of us) and those that are against duty (morally forbidden) - at least if we ignore those rare cases where there may be a tie between two incompatible actions in our assessments of overall utility. (In such cases our obligation is only to perform one or other of these actions, rather than either in particular.) Our ordinary moral thought, in contrast, classifies actions three-fold, into those that are duties, those that are against duty, and those that are neither. We tend to think that there is a non-moral space left to us (perhaps an extensive one), within which we are free to do what we wish, provided that we do nothing wrong. When I relax in front of the TV in the evening, for example, I am (often) neither performing a duty nor doing anything wrong. This non-moral space is lost to us under utilitarianism.
(Note that the standard utilitarian response to the practical objection, to be considered in a moment - that the greatest utility will in fact be produced if people generally attend to their own happiness and the happiness of those close to them - is no response at all to the current theoretical objection. For it does not deliver the desired conclusion that I am often morally free to do what I want, but only that it is often my moral duty to do what I want, in so far as doing what I want will in fact be productive of the greatest utility.)
There is a related theoretical point to this one. It is that utilitarianism, of whatever variety, is incapable of drawing a distinction between obligation and saintliness (or between duty and what is beyond duty). Since we are obliged to act for the best (or to follow rules, or develop qualities of character that will be productive of the best) it is impossible, on utilitarian grounds, to do more than one ought. Yet our common-sense morality certainly includes the idea that there are actions and qualities of character that are supererogatory - above and beyond the call of duty.
One possible response to this difficulty would be to shift to a version of utilitarianism that would require us only to seek satisfactory consequences, rather than the best. But this would then lose touch with the governing conception of the theory. It is impossible to see why an impartial benevolent observer should accept a satisfactory outcome, rather than the best. For selecting a situation that is sub-optimal would generally mean that at least some of the claims and interests of some people would have been ignored. Moreover, even this version of utilitarianism will not reinstate our three-fold moral classification of actions. It will only mean that there are many more cases where I can fulfil my duty by performing any one of a range of actions, each of which would have satisfactory consequences. This is still not the conclusion we want, that there are many circumstances in which I have no duties at all.
The practical objection to utilitarianism is that, even setting aside the above theoretical considerations, utilitarians are committed to claiming that I must be constantly at work in the service of others, which seems, from a common-sense perspective, to be too much to ask. The reply to this difficulty has generally been to argue that the greatest utility will in fact be produced if people mostly attend to their own happiness, and the happiness of those who are close to them, whose desires and needs they know best. For it is easy to get the desires of strangers badly wrong, and one may readily make mistakes in promoting remote consequences. This is the suggestion that one should develop in oneself an attitude of restricted attention, and is thus, in effect, a move towards quality-of-character-utilitarianism. The idea is that in general one should pay attention to the happiness of those close by, only considering the utility of strangers when the situation demands it. Now this reply might have considerable plausibility in a world where all are roughly equally well off. But in a very unequal world, such as ours, any given quantity of money or time spent in pursuit of utility will be of much more value to those who have least - a crust of bread may mean nothing to me, but may mean everything to one who is starving. Moreover, in the case of extreme suffering there seems to be no special problem in knowing what people want - they want food, shelter, warmth, and peace. It looks, then, as if in our world the difficulty remains - a utilitarian must claim that I should give up all (or almost all) for the sake of beneficence.
It is hard to estimate the precise force of the current - practical - objection to utilitarianism. For utilitarians can claim that the old morality is the morality of complacency and selfishness. Indeed, the considerations above suggest that the only option for utilitarians is to concede that on this matter they are out to reform common-sense morality, rather than to accommodate it. My own view is that the resulting position is highly implausible, especially when it can be contrasted with a form of contractualism that would require me to develop dispositions towards beneficence sufficient to account for my fair share of alleviating suffering (but no more), and which also allows for a private space that is outside the moral sphere altogether. But I recognize that this is not a matter that admits of decisive proof.
I conclude that while utilitarianism has its advantages, it also has very considerable draw-backs. We might be well advised to seek some alternative theoretical framework if we can. With this in mind, I shall now turn to consider contractualist approaches to morality.
Varieties of contractualism
Contractualism is named after its governing conception, which views morality as the result of an imaginary contract between rational agents, who are agreeing upon rules to govern their subsequent behaviour. It needs to be stressed that the contract in question is hypothetical, not actual. The idea is not that moral rules have resulted from some explicit contract entered into by human beings in an earlier historical era, a claim that is almost certainly false. (John Locke seems to have held a view of this sort.) Nor is the idea that we are, now, implicitly committed to a contract of the ‘I won’t hit you if you don’t hit me’ variety, which implausibly reduces moral motivation to self-interest. (This was the position of Thomas Hobbes.) While there have been versions of contractualist moral theory of these sorts, this is not the way in which I propose to use the term in this book. A contractualist moral theory, as I shall understand it, is an attempt to justify a system of moral principles by showing that they would be agreed upon by rational agents in certain ideal circumstances. It is an attempt to exhibit the rationality of moral rules, not an attempt to legitimate those rules by appeal to past agreement or present self-interest.
What all the different varieties of contractualism have in common is that they deem morality to be a human construction, created by human beings in order to govern their relationships with one another in society. Contractualism, too, can thus avoid all the difficulties that attend strong objectivism and intuitionism as theories of ethics. Moral values are not regarded as being ‘out there’ in the world, any more than the difference between spices and other foodstuffs is, except in the innocuous sense that human beings have created a conceptual system that puts them there. All the same, most contractualists have thought that moral rules are imposed upon us, in one way or another, by our very rationality. So moral rules may be held to be objective, not just in the weak objectivist sense that, given a system of moral concepts, there are then objective truths about what one may or may not do, but also in the stronger sense that we cannot, in so far as we are rational, select an alternative system of concepts. But since it goes without saying that human beings do not always behave or choose rationally, all versions of contractualism employ some variant on the idea of an imaginary contract, so as to idealise away from our limited rationality, and to represent what we would choose if we were perfectly rational.
The main historical exponent of contractualism, in my sense, was Immanuel Kant. He regarded morality as arising out of a process of rational construction, moral rules being those that rational agents could not rationally wish to be universally ignored. On this account, the rule ‘Do not kill for gain’ is a moral one because you cannot rationally want a society in which all are prepared to kill for gain. While there is no explicit mention of a contract here, still the distinctive feature of Kantian moral construction is that agents should be seeking rules that all can rationally agree to. (Kant would say that their very rationality requires them to seek such rules, but that is another matter.) Kant’s own formulations give rise to many technical and interpretative difficulties, however, and are embedded within theories of human nature and the origin of knowledge that are highly controversial. Accordingly, I shall confine my discussion of contractualism to some more recent theoretical work that can nevertheless be described as ‘Kantian’, in that it continues to see morality as arising out of a process of rational construction by rational agents.
Undoubtedly the most famous and influential contemporary version of contractualism is that presented and defended by John Rawls in A Theory of Justice. Since Rawls is mainly interested in political philosophy, his application of the contract idea is primarily to determine the basic institutions and structures of a just society. But I propose that we may deploy his version of contractualism for the wider purpose of constructing a general theory of morality. The basic idea, then, is that we are to think of morality as the rules that would be selected by rational agents choosing from behind what Rawls calls a veil of ignorance. While these agents may be supposed to have knowledge of all general truths of psychology, sociology, economics, and so on, they are to be ignorant of their own particular qualities (their intelligence, physical strength, qualities of character, projects, and desires), as well as the position they will occupy in the society that results from their choice of rules.
No real person could ever be in such a position, of course, nor is Rawls committed to claiming that they could. No one could ever be in ignorance of such basic facts about themselves as their desires, their rough age, or their sex. Yet agents under the veil of ignorance would know none of these things. The point of the restrictions is to eliminate bias and special pleading in the selection of moral principles. Rawls thinks that, since the position behind the veil of ignorance is a fair one (all are equally rational agents, none has any more knowledge than any of the others), the resulting system of morality must also be fair. Hence his proposal is, in fact, that moral rules are those that we would rationally agree to if we were choosing from a position of complete fairness.
Most importantly, the agents behind the veil of ignorance must not be supposed to have, as yet, any moral beliefs. For part of the point of the theory is to explain how moral beliefs can arise. Moreover, since the idea is that we may seek to resolve moral disputes by subjecting the beliefs in question to the test of their derivability from a particular moral theory, we cannot allow moral beliefs themselves to figure within the theory, or all the original disputes will simply replicate themselves at that level also. The choice of moral principles is rather to be made in the light of broadly self-interested desires (such as those for happiness, freedom, and power) that the agents know they will possess whatever particular desires and interests they subsequently come to have. The agents will know that, whatever they want, they will want their most important desires to be satisfied, they will want the freedom to satisfy them, and they will want the power to do so.
Rawls’s theory is by no means the only form of contemporary contractualism, although it is certainly the best known, and probably the most thoroughly worked out. Since one of our goals in later chapters will be to see what consequences contractualism as such has for the question of the moral standing of animals, it is important that we should consider at least one other formulation of the doctrine. Otherwise we may easily be misled in thinking that we are drawing out the implications of a contractualist approach to ethics, when in reality we may only be tracing out the consequences of Rawls’s particular version of it.
I propose, in addition, to consider the version of contractualism that has been developed more recently by Thomas Scanlon. Scanlon acknowledges a debt to Rawls, but thinks his formulation is able to avoid many of the difficulties that have been raised against A Theory of Justice. His account of morality is roughly this: that moral rules are those that no one could reasonably reject as a basis for free, unforced, general agreement amongst those who share the aim of reaching such an agreement.
The basic picture of morality here is essentially similar to that of Rawls, and would arguably justify many of the same normative principles, as we shall see in the next section. It is much less artificial, however, and it contains many fewer idealisations. For here the agents concerned are supposed to be real ones, with knowledge of their own idiosyncratic desires and interests, and of their position within the current structure of society. The only idealisations are that choices and objections are always rational (whereas real agents are only sometimes rational, often making mistakes), and that all concerned will share the aim of reaching free and unforced agreement (whereas many real agents may seem indifferent to such considerations). Yet, arguably, these idealisations can achieve the same effect as Rawls’s veil of ignorance (that is, the elimination of bias and special pleading). For the contractors will know that there is no point in rejecting a proposed rule on grounds special to themselves, since others would then have equal reason to reject any proposed rule.
It may be worth noticing how Scanlon’s version of contractualism is reminiscent of one of Kant’s formulations of the basic moral perspective, namely that it is a ‘kingdom of ends’. Kant’s picture is of an association of rational agents, each of whom is legislating for all - that is, each of whom is trying to devise principles that could be freely and rationally acceptable to all. This is, in effect, Scanlon’s formulation minus the qualification that the agents share the aim of reaching free agreement. Kant would think that the qualification is not needed, because he holds that the basic principles of morality can be derived from reason alone. The only idealisation necessary, for Kant, is that the agents should always decide rationally.
Notice that if Kant’s version of contractualism were acceptable, we should have an immediate explanation of the source of moral motivation. For if moral principles can be derived from pure reason, in such a way that immoral action is always at the same time irrational action, then the answer to the question ‘Why care about morality?’ is plain. It is, that concern with morality is forced upon us, in so far as we are rational. Along with most contemporary thinkers, however, I believe that this aspect of Kant’s programme is hopeless. Since there is no prospect of deriving morality from reason alone, Scanlon’s qualification - that the agents must share the aim of reaching free agreement - is a necessary one. I shall return in a later section to the question whether contractualism, thus construed, is capable of explaining the source of moral motivation.
The normative output of contractualism
A collection of rational agents choosing moral rules from behind a veil of ignorance would plausibly agree, most fundamentally, to respect one another’s autonomy. They would agree not to interfere with one another’s projects, except where this is necessary to prevent a similar interference with their own. Most importantly, they would agree not to attack or kill one another, except in self-defence. It seems likely, indeed, that the most fundamental moral principle under any version of contractualism will be a principle of respect for autonomy - that agents should be granted as much freedom to act and pursue projects as is compatible with a similar degree of freedom being granted to all the others. For given that you do not know, or cannot legitimately appeal to, your particular strengths, weaknesses, or life-plans, or your position under the system of rules to be selected, you will wish to preserve as much space for yourself as possible. Under these conditions, rational agents may be expected to value their own rational agency above all else.
Undoubtedly the main attraction of contractualism, for many thinkers, lies in its basic normative principle. There is a powerful intuitive appeal in a rule requiring that people should not, so far as possible, interfere with one another’s plans and projects. The principle of respect for autonomy is attractive in providing adequate defence for individuals against interference from other individuals or groups, or, indeed, from the state itself. It provides the basis for an acceptable set of rules regulating people’s interactions with one another, while at the same time leaving a substantial domain within which they remain free to get on with their own lives, and to develop their own concerns and interests, without direction from morality. Here, then, are already two significant contrasts with utilitarianism - namely, that the lives and interests of individuals cannot be interfered with merely to subserve greater overall utility, and also that our common-sense belief in non-moral space is preserved.
Rawls himself also argues that contractualism entails a particular solution to the problem of distributive justice, which he calls the difference principle. Since agents in the original position are to be ignorant of their place within the structure of the society that will result from their choice of basic rules, there is an initial presumption in favour of equality of distribution of goods and duties. For none will wish to be disadvantaged. They may rationally accept deviations from this basic equality, however, provided that the resulting increase in efficiency leaves the worst off person under the new system better off. This, then, is the difference principle: differences in wealth etc. are only acceptable if those who are worst off under the system are better off than the worst off people would have been under any alternative system. This normative implication of contractualism, too, strikes many as having a powerful intuitive appeal - although what its implementation would mean in practice, of course, will depend upon highly contentious claims of psychology, economics, and political theory.
In fact, Rawls himself is only able to derive the difference principle, within his system, by disallowing gambling from behind the veil of ignorance. Otherwise the contractors might rationally opt for a highly unequal distribution, provided that their chances of being in the worst off group were slim. But that gambling should be disallowed has seemed to many to have been a theoretically arbitrary restriction. This is one of those places where Scanlon’s version of contractualism may have the advantage. For it seems likely that some version of difference principle will also be a consequence of Scanlon’s contractualism, if those who are worst off under a proposed system of distribution could reasonably reject that system provided that there is an alternative which would not leave anyone worse off than themselves.
It is worth noticing, also, that although contractualism is supposed to legitimate a system of rules, it has no difficulty with the sorts of case - such as the judicial punishment of the innocent - that cause such trouble for utilitarianism. For it must be assumed that the moral rules agreed under the contract will be publicly known - it is part of the very ethos of contractualism that rules and practices should be publicly justifiable. (In contrast, utilitarianism need not assume that everyone will be a utilitarian. On the contrary, it should not assume what is false.) Since contractualism pictures morality as arising out of agreement between rational agents, the resulting rules must be supposed to be known to all. Indeed, close to the heart of the contractualist approach lies an ideal of publicity - namely, that moral rules and principles should be publicly negotiable and defensible in open debate. And then rational agents selecting rules to govern their interaction would, of course, choose principles of justice and punishment that are impartial and exceptionless. For it would obviously be intolerable for it to be commonly known that there is a general rule allowing the punishment of the innocent in the service of expediency, since such a rule would undermine confidence in the whole system.
One difficulty that many people see in the normative output of contractualism is that the resulting rules are too minimal and negative. There will, it seems, be rules enjoining us not to interfere, but no rules enjoining us to assist. There will be rules forbidding us from taking lives, but no rules requiring us to save lives. In short, the accusation is that contractualism places all of its emphasis on justice, to the exclusion of beneficence. This is certainly a serious charge. For there are, intuitively, a great many sorts of situation in which one ought to help others. If contractualism must entail that there are no such situations, then it will surely have been shown to be inadequate under reflective equilibrium.
To see how powerful is the case for an obligation of beneficence, consider a dramatic example of Peter Singer’s. While Carl is walking to work one day he passes a shallow pond in which he can see that a child is drowning. No one else is around. Surely he ought to wade in to save the child. The only cost to himself is that he will be slightly late for work and that his clothes will become wet and muddy, whereas the gain will be the life of the child, with all that means both to itself and its parents. If Carl chooses to ignore the child’s plight, indeed, then this would surely be an extremely serious moral failing - not quite tantamount to murder, perhaps, but almost as bad in its complete callousness. Yet no principle of non-interference would have been breached. By failing to save the child he could not, by any stretch of the imagination, be said to have interfered with its plans or projects. His failure to save, in these circumstances, is in conflict, not with justice, but with beneficence. So what is a contractualist to say?
I do not believe that there is any special problem for contractualism here, in fact. While it is true that contractualists have tended to devote most of their time to developing principles of justice, and have talked largely about rights to non-interference rather than obligations to render assistance, this seems to me to have been an accident. It just so happens that this is the area of morality in which contractualists have been most interested. Partly, no doubt, this is because the differences between contractualism and utilitarianism are at their greatest at this point, as we have just seen. For what could be more natural than to place most emphasis upon those aspects of your theory that differentiate it from its main rival? I propose to postpone to Chapter 7 the task of showing in detail how contractualism can accommodate principles of beneficence. But the basic idea is simply that contracting agents should, if they are rational, agree to develop in themselves an attachment to the welfare of others, of sufficient strength to ensure that they do their fair share in alleviating suffering. So callous Carl may be severely criticised for failing in this regard, even though he infringes no one’s rights.
The governing conception of contractualism
If it is true that contractualism can account for duties of assistance as well as those of non-interference, then at this stage in our investigation it begins to seem likely that contractualism can meet one of the main constraints placed upon the adequacy of a moral theory, namely that it should have a normative output that is, by and large, intuitively acceptable. But what of the other main constraint, that an acceptable moral theory should provide a plausible account of the source of moral notions, and of the basis of moral motivation? Let us take these in turn.
Contractualism certainly suggests a plausible theory of how morality can come to exist at all. As we saw earlier, moral notions are presented as human constructions, that arise in order to facilitate human co-operation and the life of community. In any actual society, of course, many different forces and pressures will have shaped the structure of its morality. Contractualism presents us with a way of seeing what our morality should be, if the only constraints on its content are rational ones.
Now consider what contractualism can say about the source of moral motivation. This seems much more problematic. For why should we be interested in what rational agents would choose from behind a veil of ignorance, for example? Why would this be something worth dying for, in the way that many have lain down their lives in the service of justice? Indeed, why should we ourselves feel constrained by what rational agents would agree from behind the veil? Why do I have any reason to accept the rules that they would accept? Rawls provides a partial answer to these questions in terms of the notion of fairness - since the position behind the veil is fair, it is guaranteed that rational choice from such a position will be fair. Although helpful, in the end this just pushes the problem further back. For why are we interested in fairness? What is the source of our motivation towards arrangements that are fair? Until these questions are answered, contractualism has not really justified its governing conception.
One approach, implicit in some of Rawls’s later writings, might be to maintain that we care about morality, as contractualism pictures it, because without it there can be no lasting and non-violent solutions to areas of social conflict, given the basic nature of modern society. Since we can no longer appeal to theological authority to resolve moral disputes, and since no body of traditional belief can now hope to secure universal assent, the only way in which we can have a chance of achieving moral consensus is through reasoned agreement. This explains the source of moral motivation for us, living as we do in pluralistic and relatively free societies. But even those traditional societies that still remain might have reason to accept the need for rationally agreed rules and principles, since the nature of our modern world is such that all communities must necessarily interpenetrate one another to some degree.
Another approach, defended by Scanlon, maintains that the solution to the motivation problem is simply to postulate that human beings have a basic need to justify their actions to one another in terms that others may freely and rationally accept. That human beings do have such a need is certainly very plausible. Even scoundrels will characteristically attempt some such justification for what they do. As Scanlon remarks, the notorious difficulty of motivating people to do the right thing need not result from any deficiency in their underlying moral motive, but rather from the ease with which it is deflected by self-interest and self-deception.
Scanlon supposes that the requisite desire to justify oneself is produced and nurtured by moral education. I think it might more plausibly be maintained that it is innate (inborn), in such a way as to emerge gradually at a given stage in maturational development. The case for believing that a good deal of human cognition is innate, including knowledge of the basic principles of human psychology, is a powerful one. What more natural, then, than that the basic springs of characteristic human motivations (including moral motivation) should be innate also? Certainly one would expect such an innate feature of human beings to have considerable survival value, given that we depend crucially for our survival upon co-operative modes of living.
If the desire to be able to justify oneself to others in terms that they may freely and rationally accept were innate, however, one would certainly expect it to be universal. Yet, it may be objected, there have been many communities in the course of human history that have not conceived of their morality in anything like contractualist terms. We may note in reply that what someone may rationally accept will depend, in part, upon their background beliefs. If you believe that the world is ruled by a benevolent and just God, for example, who watches over us as his children and who wishes us to arrange our lives along the hierarchical lines of feudal societies, then you might freely and rationally accept the rules that assign you your role as a serf in such a society.
Once again an analogy with other areas of our cognition may prove helpful. There is a powerful case for saying that the basic principles we employ in justifying our beliefs about the world - particularly inference to the best explanation of a given phenomenon - are innate. Yet what counts as the best explanation of something is, in part, a function of the other beliefs that you hold, since one feature of a good explanation is that it should cohere well with surrounding beliefs and theories. This enables us to discern an underlying unity beneath the very different sorts of explanations that may be preferred within different human communities. I am suggesting something similar in connection with the manifest diversity in human moralities - that they may reflect the same (innate) underlying conception of moral justification, put to work in the context of different social practices and metaphysical beliefs. My hypothesis is thus that the contractualist concept may form the core of all the different moralities that human beings have endorsed. At any rate, this is sufficiently plausible to make contractualism, when taken together with the appeal of its normative output, a very strong contender overall for the title of most acceptable moral theory. But we have yet to consider some powerful criticisms that have been levelled against the theory.
Replies to criticisms
One criticism of contractualism is that it is parochial - that it merely expresses the values of liberal capitalist democracies. Now one aspect of this charge is easy to rebut, namely the commitment to capitalism. Whether or not contractualists should be capitalists will depend upon the facts of economics and psychology necessary to apply the difference principle. If it were true, as socialists generally claim, that socialist economies serve to unleash the forces of production, leading to better living standards for all as well as to greater control by individuals over the course of their own lives, then it would be socialism that should be selected. It is just that many contractualists doubt whether these claims are true.
The commitment to liberal democracy, however, seems to go deeper. Rawls, indeed, is happy to accept this commitment, and the associated charge of parochialism, claiming it as a positive virtue. He regards the apparatus of the veil of ignorance as modelling the values implicit in liberal democracies, rather than as providing an account of universal morality as such. It is merely a device to enable us to gain a clearer self-image, and to work out the consequences of our most fundamental values. If the suggestions I made in the last section were on the right lines, however, then Rawls may have been too pessimistic about the prospects for contractualism to serve as the universal morality. If the contractualist conception is innate to human beings, then liberal democracy may merely be what you get once morality is stripped of its connections with religious belief, and of its association with such beliefs as that ill-educated people, women, and members of other races are not properly rational agents, needing in important respects to be treated as children.
While some have charged contractualism with parochialism, others have levelled the opposite - communitarian - charge against it. They have claimed that in aiming to be universal it has become too abstract, breaking the connection between moral values and the communities and practices in which they are embedded. But this is to miss the point made above, that what rules it may be rational to accept or reject will depend, in part, upon the other beliefs and values current in the community. We should therefore expect a good deal of cultural variation in the way that contractualism gets embodied in concrete institutions and practices. Moreover, as Scanlon points out, there may be a number of different sets of principles that pass the contractualist test of non-rejectability, and yet any moral community will need to settle upon one from amongst the options available. (Compare: while there are no theoretical considerations that determine whether we should drive our cars on the right or the left, we plainly need to settle upon one or other of these in particular.) Contractualism would then entail that many obligations and duties are to be established by convention, and may therefore vary from society to society.
The above objection can be restated in a stronger form, however. It can be claimed that the apparatus of the veil of ignorance entails that rational agents can exist prior to possessing any particular set of desires and values, having an identity that is independent of them. Whereas, it may be said, such agents are always, and essentially, members of some moral community. Indeed, the very identities of those agents may be tied to the values and practices that form a necessary part of the communities that have nurtured and shaped them. But in my view contractualism carries no such presupposition about the identities of persons. All that it presupposes is something much weaker and more plausible, namely that it is possible for rational agents, in some of their reasonings, to bracket off even those values and concerns that contribute to their identities, subjecting those values to examination and possible criticism from a more abstract perspective. Even if my sense of loyalty as a feudal serf forms a necessary part of who I am, as an individual, it is surely possible, temporarily setting those loyalties to one side, that I should raise the question whether feudal institutions might reasonably be rejected by those seeking to agree on moral rules. It is also possible that I might arrive at a negative verdict, and that I should, in consequence, come to change my moral commitments, and hence my own identity as a person.
My conclusion is that the communitarian challenge to contractualism fails. So we are left with a theory whose governing conception is at least as plausible as that of utilitarianism, but whose normative output is considerably more attractive.
I have developed versions of both utilitarianism and contractualism that can claim not only to present a satisfying explanation of the origins of morality and of moral motivation, but also to accommodate a good deal, at least, of common-sense moral judgement. My own view is that contractualism is, under reflective equilibrium, by far the more plausible moral theory. But both are sufficiently powerful that we shall need to consider the consequences of each of them for our treatment of animals. It will be a further test of the adequacy of our two candidate theories that they should entail acceptable consequences on such matters.
 A charge pressed forcefully by Jack Smart in the book jointly written with Bernard Williams, Utilitarianism: for and against (Cambridge University Press, 1973).
 Oxford University Press, 1981. See also his discussion of the utilitarian attitude towards slavery in his contribution to Peter Singer ed., Applied Ethics (Oxford University Press, 1986).
 See Michael Slote, ‘Satisficing Consequentialism' in his Common-Sense Morality and Consequentialism (Routledge, 1985).
 Such a view has been defended by Singer, Practical Ethics, ch. 8.
 See his Two Treatises on Government (1690).
 See his Leviathan (1651).
 See especially his Groundwork to a Metaphysic of Morals (1785).
 See his ‘Contractualism and Utilitarianism' in A.Sen and B.Williams eds., Utilitarianism and Beyond (Cambridge University Press, 1982).
 See his Groundwork to a Metaphysic of Morals, ch. 2.
 But for a contrary view, see Onora O'Neill, Constructions of Reason (Cambridge University Press, 1990).
 Taken from his Practical Ethics (Cambridge University Press, 1979), ch. 8.
 See his papers ‘Justice as Fairness: Political not Metaphysical' and ‘The Priority of Right and Ideas of the Good' in Philosophy and Public Affairs 14 (1985) and 17 (1988) respectively.
 See ‘Contractualism and Utilitarianism', pp. 116-17.
 See my Human Knowledge and Human Nature, chs. 6-8.
 See my Human Knowledge and Human Nature, ch. 6.
 This is my preferred mode of defending the coherentist conception of knowledge, mentioned briefly in Chapter 1 above. I claim that it is an innate constitutive part of human reason that the justification for a belief consists, in part, in its coherence with surrounding beliefs. See my Human Knowledge and Human Nature, ch. 12.
 See Michael Sandel, Liberalism and the Limits of Justice (Cambridge University Press, 1982).
 ‘Contractualism and Utilitarianism', p. 112.