The appearance​ of a new ethical theory is a rare event, but it happened in 1982 with the publication of T.M. Scanlon’s essay ‘Contractualism and Utilitarianism’. Although he called his theory contractualism, Scanlon does not postulate an actual social contract between the members of a society, like Hobbes, or a hypothetical contract under imaginary conditions, like Rawls. Instead, he finds the foundation of morality in principles that reasonable people would find mutually acceptable. He holds that an action or policy is wrong if any principle that permitted it could be reasonably rejected by someone affected adversely by that principle. Whether such a rejection is reasonable depends roughly on a comparison between the burdens that the principle imposes on the rejector and the benefits it provides to other individuals, taken one at a time. (I’ll say more about what this means below.)

Scanlon’s theory was fully developed in his book What We Owe to Each Other (1998) and extensive commentary and criticism have grown up around the work. The book also incorporated ideas about the problem of freedom and moral responsibility first presented in Scanlon’s 1988 Tanner Lectures on ‘The Significance of Choice’. He has never stopped thinking about these topics, responding to criticisms, modifying his views and offering additional arguments. Morality and Responsibility collects twelve essays – four on the general theory of morality and eight on responsibility – that reflect on his theories and the way they compare with the alternatives.

Let me begin with the topic of freedom and responsibility, on which Scanlon holds distinctive and complex views. The traditional problem is this: we blame and punish people for the bad things they have done, which seems to presuppose that they have free will. But even if those actions were a result of their bad character, that character is ultimately caused, if we go back far enough, by biological, psychological and social factors over which they have no control. Whether the causal order of nature is deterministic or partly random, people do not create themselves. So how can it make sense to blame them for what they do? Punishment makes sense only as a deterrent, not as retribution, and feelings of condemnation and resentment towards a criminal are as out of place as they would be towards a tiger.

Scanlon’s response is complicated. First, he doesn’t see how people can have free will, since this implies that ‘our thoughts and actions are not all caused, ultimately, by factors over which we have no control.’ Second, he believes that whether or not we have free will in this sense, retributive punishment is never justified, because it is never a good thing for someone to suffer, no matter what he has done. But, third, he holds that many forms of adverse treatment and negative attitudes can be warranted by the bad things a person has done and by what he is like, even though he is not ultimately responsible for his character. Most of our moral reactions and judgments of responsibility do not require free will in order to make sense.

Instead of searching for a single general condition of moral responsibility, Scanlon argues, we should identify the kinds of change in treatment or attitude towards an offender that seem called for in each particular case, and determine whether they are appropriate in light of the breach in our relations that is revealed by the offender’s conduct. Like anyone treating this topic, Scanlon is indebted to Peter Strawson’s essay ‘Freedom and Resentment’ (1962), which drew attention to the unavoidability of ‘reactive attitudes’ like indignation and resentment towards people who have violated the terms of our human relations with them – relations we could not give up without ceasing to live a recognisably human life. We could not have friends, for example, without being vulnerable to feelings of blame and resentment towards a friend who betrayed us. But Scanlon emphasises other responses as well:

1. Withdrawal of trust

2. Decreased readiness to enter into special relations such as friendship with the person

3. Decreased willingness to help the person with his projects

4. Decreased tendency to take pleasure in things going well for the person and to feel sad or regretful when they do not

These responses are appropriate whether the person has caused you personal harm or just violated the standards that apply to him as a member of the moral community – to which Scanlon believes all human beings belong. All these attitudes and exclusions impose costs on the offender, but they are morally justified not because it is good that he should suffer, but because what he has done, and what it shows about his indifference to the interests of others, makes it unreasonable or risky to treat him with the goodwill we adopt as a default towards decent people. Even if the relation is not ended (the universal moral relation, for example, cannot be), it will have to be modified.

The fourth reaction listed above shows that Scanlon is not a saint, but he insists that it is distinct from a belief in retribution: ‘Not being sad or regretful when something bad happens to a person [who has done wrong] is not the same thing as thinking it to be a good thing, or justified, that the person should suffer this fate.’ In other words, we may not think, ‘He got what was coming to him!’ but only: ‘I’m not sorry.’

This is a fine line, but in Scanlon’s view it has practical consequences. He holds that permissible alterations of feeling do not justify imposing suffering on the wrongdoer or acquiescing in it – only not caring about it. But I wonder. If you wouldn’t regret a villain suffering as a result of his crimes, wouldn’t you be inclined to acquiesce in his suffering, and decline to prevent it, even if you could? Scanlon seems genuinely immune to the retributive sentiment of wanting the wicked to suffer (except perhaps in the case of wicked fictional characters, as he suggests at one point), but it is a feeling that I suspect is natural to most people, and as Scanlon says in another context, desire always involves there seeming to be something good about the thing desired. It is not clear how we can settle the stark conflict between such feelings and his claim that ‘it is never a good thing, morally speaking, for anyone to suffer, no matter what they have done.’

However, if he is right about this, it changes the problem of the conditions of moral responsibility. Scanlon points out correctly that the main motive for seeking a condition of free will is that retribution makes no sense if a person’s character is not under his own control. But the question of ultimate control seems irrelevant to other appropriate reactions to bad conduct and bad character because the person’s conduct and character themselves have an effect on our relationship to him that finds morally justifiable expression in those reactions. Refusing help to a friend who has betrayed us does not entail an assumption that he is ultimately responsible for his character. It is just a natural consequence of his destruction of our friendship. Scanlon concludes that if retribution is never justified, the absence of free will is not a problem.

What about legal punishment, not as retribution but in pursuit of deterrence and prevention? According to Scanlon, this introduces a different kind of responsibility. So far I have been talking about what Scanlon calls ‘moral reaction responsibility’, by which he means ‘the conditions required to make forms of blame and other reactive attitudes appropriate’. But the conditions that make punishment appropriate fall under what he calls ‘substantive responsibility’, by which he means the conditions required for something a person does (or fails to do) to change his obligations to others and their obligations to him. Examples include making a promise, signing a contract, or undertaking a risk, as well as committing a crime.

The conditions of responsibility in such cases are again not simple, according to Scanlon, but depend on the purposes of the governing practice or institution and the way it allocates benefits and burdens. Having a choice is important in most cases. Legal punishment imposes fines and imprisonment to deter people from harming others. But ‘insofar as a policy of inflicting hard treatment on wrongdoers is justified this must be not only because it is an effective way to protect others from being wronged but also because this policy inflicts this treatment only on individuals who have had a fair opportunity to avoid being subject to it.’ In most cases responsibility does not require actually making a choice. Having the possibility of choice is often enough, even if one fails to take advantage of it through carelessness or neglect.

Let me move now to Scanlon’s contractualist theory. We can see how it works in evaluating some principles proposed by Peter Singer in his paper ‘Famine, Affluence and Morality’. One principle Singer offers is this: ‘If it is in our power to prevent something very bad from happening, without thereby sacrificing anything morally significant, we ought, morally, to do it.’ Singer gives the example of seeing a child who has fallen into a pond and is about to drown. Obviously it would be wrong not to rescue the child, even though it means wading into the pond and getting your clothes muddy. But Singer goes on to say that this same principle also requires us to donate money to famine relief to save the lives of people halfway around the world, if we can easily spare it. The application of Scanlon’s theory to this case is straightforward: starving people in Bengal, whose lives could be saved for modest sums, could reasonably reject a principle that permits affluent Westerners to contribute nothing at all to save them – the gap between the costs to the former and to the latter is too great. So it is wrong for the affluent not to contribute to causes that help the destitute survive.

But Singer doesn’t stop there. He contends that this principle is too weak and that a much more demanding variant is necessary: ‘If it is in our power to prevent something bad from happening, without sacrificing anything of comparable [my italics] moral importance, we ought, morally, to do it.’ He believes this requires us to give until we are almost as badly off as the starving Bengalis. Scanlon’s theory would not support this stronger principle: it would not be reasonable – taking into account the reasons on both sides – for the Bengalis to reject a principle that allowed the affluent to stop giving before they impoverished themselves. Even if there is not a sharp line between what is reasonable and what is not, this seems a clear example of the distinction, and provides a credible explanation of what is wrong with Singer’s ultra-strong principle.

Scanlon makes the same point in response to Derek Parfit’s charge that his own theory is too demanding. Suppose that if you donated one of your vital organs to someone who needs an organ transplant, it would extend his life by many years and shorten your life by a few years. Parfit claimed that Scanlon is committed to the Greater Burden Claim – ‘it would be unreasonable … to reject a principle because it imposed a burden on you when every alternative principle would impose much greater burdens on others’ – and that this implies that you are morally required to donate your organ. But Scanlon replies that if we think of this as a general principle, requiring us ‘to decide what to do in every instance by comparing the benefits and burdens to everyone affected by our choice’, it

could reasonably be rejected because complying with it in general would have a paralysing effect on our lives … We would not be able to give any preference to individuals towards whom we have special relations and would not be able to plan effectively to carry out any project, since other more pressing uses would be very likely to arise.

Scanlon’s contractualism is a reductive theory in the sense that it explains moral rightness and wrongness in terms of more basic normative ideas that are not moral: the idea of what individuals have reasons to want and the idea of what it is reasonable for them to demand or reject, taking into account the reasons other individuals have. According to Scanlon, morality does not emerge when we take an impartial, impersonal perspective on humanity; it emerges from the one-to-one relation each of us has with each other individual. We can justify our conduct to others only if, while being primarily occupied with living our own lives, we nevertheless follow rules that take their interests and reasons into account, and don’t give unreasonably disproportionate weight to our own. That is different from giving their reasons and ours the same weight. What is unreasonable is a matter of judgment: this is not a mechanical decision procedure, but an account of what we need to think about in thinking about moral questions.

Scanlon​ believes that his reductive method is superior to the more common method of relying on our intuition about what is right and wrong in particular cases to determine the structure and boundaries of rights, obligations and duties (what he calls the question of content). Even if the results are similar, the test of what it is reasonable to reject helps explain those boundaries, often in subtle ways. It also explains why we have reason to care about the results (he calls this the question of acceptance): namely, because we want to be able to justify our conduct to others, appealing to their reasons as well as ours.

There is a particularly interesting discussion of the question of acceptance in the essay called ‘Ideas of the Good in Moral and Political Philosophy’. Scanlon shows that there is a tension between the values that guide individuals in leading their own lives and the more abstract and general values that figure in accounts of morality and justice. One example comes from Marx’s discussion of religious toleration in ‘On the Jewish Question’, where he argues that political toleration of multiple religions is consistent with commitment to the truth of only one of them because toleration is held from the point of view of ‘a citizen’ and commitment from that of ‘a man’. A committed believer is not motivated by the general value of ‘religion’, but a citizen must be. Although one can occupy both standpoints, they are in tension, since religious toleration puts limits on what can be done in the service of one’s own religious convictions. Something similar can be said about the tension between one’s commitment to one’s friends and family and the abstract general good of everyone’s being able to fulfil such commitments – which can set limits to what one may do for one’s own. These tensions pose the question of acceptance: whether abstract general values have the authority to place moral or political limits on the claims of corresponding personal values.

Scanlon’s work is situated in highly contested philosophical territory. In my view, his interpersonal foundation for morality is more plausible than the impersonal foundation of impartial benevolence that is its main contemporary rival. He offers a persuasive explanation of what underlies many of the intuitions about individual rights, prohibitions and obligations that are often cited in opposition to utilitarianism.

Scanlon’s position on the subject of freedom and responsibility is sophisticated and coherent; it finds that nothing about the natural order of causation is incompatible with the ways we legitimately hold people responsible. I remain sceptical, partly because I continue to feel that the wicked should suffer, and that this makes sense only if they have some kind of free will, even though I don’t have an account of what that is. Perhaps that retributive desire can be discredited as a retaliatory instinct instilled in us by evolution as a useful deterrent against aggression; but I do not find that such an explanation enables me to distance myself from the feeling and regard it as an evaluative illusion. I also can’t rid myself of the idea that free will is part of my conception of myself as a rational agent and thinker, and is presupposed in my attitudes towards others. Without it, my view of them is hollowed out, even if I continue to interact with them as if they were more than parts of the natural causal order. Scanlon is certainly right that it is because humans can respond to reasons that we hold them, but not tigers, responsible. We differ, I suspect, in our views of what it is to respond to reasons.