When people are asked what is fair, whether in the allocation of rewards and other benefits, or in the allocation of blame and punishment, they often make reference to desert [see Miller (Reference Miller1999, ch. 4) for an overview]. Equally, as many experiments have shown, when they are asked to distribute sums of money to people who have been asked to complete various tasks, they have a strong tendency to give more to people who have performed better – unless they are the better performers themselves, in which case a ‘politeness ritual’ may inhibit them from taking a larger slice of the cake (Mikula, Reference Mikula and Mikula1980). A just world is one in which everyone gets what they deserve – belief in which can also sometimes distort people's judgments by leading them to attribute (without sufficient evidence) better performances to those who have been given higher rewards, so that fairness is seen to be preserved (see Lerner, Reference Lerner1980).

With few exceptions, Joel Feinberg noticeable among them, this enthusiasm for desert has not been shared by philosophers, even those like John Rawls who think that philosophical reflection should at least begin with our intuitive beliefs about what is fair. Why don't philosophers share the public's taste for desert? There are several reasons.

• • Most philosophers who write about justice are egalitarians: they begin with a presumption in favor of equality. From this starting point, desert looks like a tool for justifying inequality. In particular, philosophers are sensitive to the way in which the grossly inflated salaries of CEOs and other top officials are often justified by claiming that they have made an outstanding contribution to the performance of their organization – a desert-based claim. Getting rid of desert would rule out any such justification.

• • Philosophers see desert as entangled with the problem of determinism and free will: If we examine the performances that form the basis of desert – whether doing something valuable in the cases of ‘positive’ desert, or doing something harmful in the case of ‘negative’ desert – it may appear doubtful whether these performances can really be attributed to the performer in the deep way that desert judgments seem to assume. People certainly act, for good or ill. But what causes them to act in the way that they do, and what gave them the powers and abilities that enable them to perform? Desert seems to involve falsely crediting or blaming people for actions whose ultimate causes are outside of their control.

• • At first glance, giving people what they deserve might seem simple: Someone acts, and then there is a response to that act – a reward or a punishment – that matches what the person has done. But matters become much more complicated when we consider the social background against which such judgments have to be made. For other people are already being rewarded and punished in many different ways, and fairness seems to require consistency across the entire group. For example, what counts as a fair wage for me can't be decided without looking at how countless others have performed and been rewarded, which seems to make the concept unmanageable, particularly in the case of ‘positive’ desert. One way of putting this point is to say that social justice must be ‘holistic’ – it must apply to the overall way resources are distributed in a society – whereas desert appears to be ‘individualistic’, in the sense that we claim to be able to know just by looking at how someone has behaved what treatment they are owed (see Scheffler, Reference Scheffler and Scheffler2001). So desert may have a role in micro-contexts, say when we are passing judgment on the behavior of a friend or a colleague – or indeed Sir John Stuffgut – but it has no place in debates about law and social policy.

These are serious charges that deserve a longer response than is possible here – though I will touch briefly on a couple of relevant points later. Instead, I want to explore why these philosophic doubts are unlikely to discredit desert in the eyes of the general public. What is the source of their attachment to it? Adam Oliver suggests that responding to desert is a form of reciprocity, and we have good reason to think that reciprocity is hard-wired into us as social beings: without it, it is difficult to see how sustained social co-operation is possible. Let's grant that premise. Does this give us a way of explaining the widespread (and apparently ineradicable) taste for desert?

Although requiting desert, whether it takes the form of rewarding or punishing, can often look like reciprocating someone else's behavior, I suggest that at best there is an overlap, rather than an identity, between these two practices. Think first about forms of reciprocity that don't seem to involve desert, such as exchanging favors, or giving gifts. Suppose the two of us agree to swap baby-sitting duties to give each other a night off. When it's my turn for a night out, can I really say, on this ground alone, that this is something I deserve (as opposed just to being something that I'm owed as a result of the swap deal)? Or when you send me a birthday card because I always send you one, is any kind of desert being recognized (surely not the labor of going out to buy a postage stamp)? Then think about cases when we should give people what they deserve even though their activities have had no direct effect on us. When the Booker Prize committee gives the award to the author it judges to have written the best book on the shortlist, it isn't reciprocating anything that the author has done for the committee – indeed if it were to be found to be reciprocating some favor that had been dispensed, it should be immediately be discharged for corruption. Much rewarding and punishing of desert has this third-personal character: it isn't the victim of crime who imposes the sentence on the criminal, but a detached magistrate. To assimilate this to reciprocity, you would need to think of the judge as somehow representing the victim and acting on his behalf, but this is not her role – nor, in most cases, should she treat the criminal in exactly the way that the criminal has treated his victim.

So reciprocity can only supply a partial explanation of our eagerness to respond to desert. Moreover, who are we, in this context? Although some kinds of desert, or merit, may have been recognized in human societies throughout history, I believe that our understanding of desert has been shaped by our more specific experience as members of modern, post-Reformation (and especially, perhaps, Protestant) societies. In these societies people are held individually responsible for what they have done, and judged on that basis, and in passing judgment we may unconsciously be doing what Divine justice is supposed to do, namely assess a person's moral character by looking at both their deeds and the intentions behind them. This element of moral evaluation seems to be yet another reason why philosophers such as Rawls want to dismiss desert (Rawls usually thinks of desert as ‘moral desert’). They think that, at least in public, people should be shielded from this kind of character evaluation. On this view, our social institutions should not be constructed with the aim of passing judgment (in the case of the law, this means favoring deterrence over retribution as its main purpose).

If this conjecture about the origins of our contemporary idea of desert is sound, it may explain something to which Oliver directs our attention, using the case of Stuffgut's soup calamity, namely that in thinking about desert we are often pushed back and forth between focusing on what actually happened and what the person intended or didn't intend to happen. This gives rise to various conundrums in the law, such as explaining why failed attempts (such as attempts to kill) are punished less harshly than successful ones: why is a would-be killer whose bullet misses his target less blameworthy than one whose bullet strikes home? There are similar questions to be asked about positive desert. Imagine someone's cat stranded up a tree and two people attempting to rescue it: both try equally hard, but one is agile and the other clumsy, so the agile one succeeds. Should the grateful cat-owner reward them equally for their efforts, or thank the clumsy rescuer politely while handing a bottle of decent wine to the person who has actually retrieved the cat? People's intuitions about this case appear to differ sharply. My late colleague Jerry Cohen was adamant that both rescuers must be thanked equally; I demurred (see Miller, Reference Miller2014 for our exchange of views on this question).

If a person's deserts were entirely a matter of what they had intended to do, not what they actually succeeded in doing, desert would indeed have little relevance to the design of social institutions, not least because intentions are often unobservable, and at best can be inferred indirectly. When institutions do aim to track desert, it is usually desert in the form of what someone has actually achieved, rather than merely what they were trying to achieve. Occasionally it may be possible to do justice to both aspects, as a school does when it establishes prizes both for those who have performed best in a class, and for those who have worked hardest to improve. But this is an exceptional case.

We still need to ask, however, whether desert should have a role in shaping social policy, particularly those areas of policy that are conventionally described as forming part of the welfare state. We might think, following Walzer (Reference Walzer1983), that different principles of justice should apply in different social spheres, and in this sphere, the relevant principles are equality and need. Here it is of some interest to observe that, however strong the normative arguments are for strict separation between the spheres, the public as a whole appears reluctant to accept that desert should be entirely dismissed even where we would normally think that some other principle of distribution is most relevant. In one experiment, subjects were asked to think of themselves as administrators in charge of a warehouse having to allocate different resources (money, prescription medicine, and food) between pairs of individuals on the basis of information about their respective needs and work performance; another variable was whether or not the warehouse was responding to an emergency caused by flooding (Scott & Bornstein, Reference Scott and Bornstein2009). Both variables – type of resource and emergency/non-emergency – had an effect on the preferred allocation, with desert being given more weight in the case of money and need in the case of medicine, as one might expect. But even in the scenario we would expect to be most favorable to distribution according to need – allocating medical supplies in an emergency – it was striking that a sizeable minority of respondents (14.8%) still chose to allocate according to desert, and a larger minority (29.7%) chose a mixed strategy that gave weight to both desert and need; in the non-flood scenario, the proportion allocating on the basis of need alone fell to 34.1%. (To clarify, distribution solely on the basis of desert meant giving the more productive worker twice the quantity of medicine that he needed, and the less productive worker half the quantity.) People appear to feel that unequal desert has to be recognized somehow, even though the only available currency of reward (in these scenarios) was medicine, which a Walzerian would say belonged to a sphere whose governing principle should be need and only need.

We may find the suggestion that people's deserts should have any bearing on the distribution of medical aid somewhat repugnant. There is, however, a different way in which desert might play a role even in a context in which the main principle of distribution should be need. We might require that all recipients should cross some threshold of desert to qualify, but thereafter their relative needs are all that should count in determining how much of the good they should get. Here it is interesting to consider Rawls's response to the suggestion that his (desert-free) principles of justice would imply providing everyone with an unconditional basic income at the highest sustainable level – since this is what it would take to maximize the economic position of the least-advantaged group in society. Rawls recoiled from this, saying that ‘those who live on welfare and surf all day off Malibu’ must instead learn to support themselves (Rawls, Reference Rawls2001, p. 179). In other words, everyone is subject to a prior expectation that they should contribute to the scheme of social co-operation (or at least attempt to do so by making themselves available for work) before they can be included within the ambit of social justice. Although Rawls might prefer not to describe it in this way, this is a threshold version of desert. To be sufficiently deserving, you have only to make some recognized social contribution (or offer to do so); beyond that point, degrees of desert can be ignored. This aligns with Oliver's suggestion that for people in receipt of welfare support, there should be an expectation, if not a requirement, that they undertake some hours of public service in return.

What then, finally, about those philosophical objections to desert? In a short space, a couple of points are worth making. The first is that while desert can justify rewarding and punishing people differently, it doesn't, by itself, tell us what scale we should use – how large the rewards should be, or how severe the punishments. So it's perfectly possible to argue that people should be paid according to the economic contribution they make, while thinking that the salaries of top executives today are absurdly high and out of kilter with what the rest of us are paid. The second (in response to worries about determinism and free will) is that desert should be based on what people intentionally do, and in most case, there is no reason to dig deeper and ask how their performances were possible at all. Usain Bolt deserved a gold medal for running 100 m in 9.69 s. If we say that, we assume that it was indeed Usain who drove his legs forward, intending to win. But we don't need to ask how much of his exceptional performance was down to good genes, and how much down to hard work on the training track. We aren't in the business of dispensing Divine justice.