Desert, fairness, and resentment

Abstract

Responsibility, blameworthiness in particular, has been characterized in a number of ways in a literature in which participants appear to be talking about the same thing much of the time. More specifically, blameworthiness has been characterized in terms of what sorts of responses are fair, appropriate, and deserved in a basic way, where the responses in question range over blame, sanctions, alterations to interpersonal relationships, and the reactive attitudes, such as resentment and indignation. In this paper, I explore the relationships between three particular theses: (i) the claim that one is blameworthy to the extent that it is fair to impose sanctions, (ii) the claim that one is blameworthy to the extent that one deserves sanctions, and (iii) the claim that one is blameworthy to the extent that it is appropriate to respond with reactive attitudes. Appealing to the way in which luck in the outcome of an action can justifiably affect the degree of sanctions received, I argue that (i) is false and that fairness and desert come apart. I then argue that the relationship between the reactive attitudes and sanction is not as straightforward as has sometimes been assumed, but that (ii) and (iii) might both be true and closely linked. I conclude by exploring various claims about desert, including ones that link it to the intrinsic goodness of receiving what is deserved and to the permissibility or rightness of inflicting suffering.

Keywords:

• desert
• responsibility
• accountability
• fairness
• reactive attitudes
• moral luck

Acknowledgements

Many thanks to Derk Pereboom, Sam Rickless, Maureen Sie, and an anonymous reviewer for most helpful comments on earlier drafts, and to David Brink for thought-provoking conversations about these issues. Although they might not all accept the point, they deserve my sincere gratitude.

Notes

At least this is a very natural way of reading Strawson. See, for example, Watson (1987/2004).

See, for example, Fischer and Ravizza (1998), who endorse the biconditional version of Strawson's claim, but also offer a substantive account of the conditions of responsible agency that makes no reference to the responses of others.

Strawson's own framework can be seen to provide a way for allowing the attitudes to be a guide to what features of agents are part of the necessary conditions for blameworthiness. (See Brink and Nelkin, forthcoming.)

It is important to note that I do not thereby assume all of the further theses that Watson takes to be true about this latter notion of responsibility. As will become clear in the text, I take it to be an open question whether each of these is true. My reason for beginning here is to start with a notion of responsibility that clearly goes beyond attributability as Watson understands it, and one that I believe is central to the debate about responsibility.

See Wallace for the thesis that blameworthiness should be understood in terms of the fairness of holding responsible, where holding responsible is understood in terms of “a liability to the responses of blame and moral sanction” (1994, 82).

My argument here draws on that of Otsuka (2009), although he would not accept all of the premises.

Explicitly focusing on punishment by the state, Lewis (1989) considers a kind of overt penal lottery in which all attempted crimes are treated similarly in the sense that the likelihood of more severe punishment depends on the risk of harm incurred. He then imagines that instead of drawing straws, we let the attempt/success ratio be the proxy for the degree of risk of attempts and punish successes more severely than attempts. He presents an argument that there is no injustice in such a system, and claims that he cannot find good enough reason to reject it. At the same time, he suggests skepticism about its soundness.

See Nelkin (2008) for a survey of these and other sorts of explanations for our initial reactions.

Otsuka also rejects a claim similar to the Fairness of Sanctions Thesis, but not because he would accept (4). In fact, he goes on to argue that two people's degree of blameworthiness can vary even when the only difference between them is precisely the kind that appears in the drunk driver cases, a point to which I return later in the text.

It is possible to find a parallel argument in Scanlon (2008), but instead of allowing luck to affect the degree to which it is appropriate to hold the reactive attitudes, luck affects the degree to which it is appropriate to modify one's relationships. There he writes:Briefly put, my proposal is this: to claim that a person is blameworthy for an action is to claim that the action shows something about the agent's attitudes towards others that impairs the relations that others can have with him or her. To blame a person is to judge him or her to be blameworthy and to take your relationship with him or her to be modified in a way that this judgment of impaired relations holds to be appropriate …. (128–9) Thus, rather than understand blame in terms of the reactive attitudes, he understands it in terms of a judgment about how your relationship has been impaired and, it seems, the very modification in your relationship that such a judgment justifies. Later, he adds:As I interpret it, blame is not a mere evaluation but a revised understanding of our relations with a person, given what he or she has done. Blame is therefore a function not only of the gravity of a person's faults but also of their significance for the agent's relations with the person who is doing the blaming. The outcome of D's action may be due in part to bad luck, but it is also due to a fault on D's part. It therefore increases the significance of that fault for those who have been affected by it. (150)

Thus, it seems that luck can affect your degree of blameworthiness, insofar as luck can affect what sort of relationship modification is justified. Now Scanlon offers independent reason to reject a thesis like the Desert of Sanctions Thesis under discussion here (and one I will discuss in Section 5), but, interestingly, this argument that appeals to luck offers another ground.

I think that a similar move might be made to preserve the consistency of the immunity of blameworthiness to luck and Scanlon's account of blameworthiness in terms of relationship impairment. Perhaps we ought to modify our relationships to the exact same degree in cases of success and failure to harm, even if it is understandable that we differentiate. It is plausible, for example, that we ought to modify our relationships in exactly the same way for both swordsmen. Thus, considerations of luck do not give us reason to reject the biconditional linking blameworthiness to relationship impairment. However, on independent grounds, I am inclined to reject it. See Nelkin (2011).

Nor is it connected to a notion of accountability. See also Pereboom (2001).

For example, see Wallace (1994, 56). He does, however take there to be a strong connection between the reactive attitudes and sanction. See Note 5.

Fischer and Tognazzini (2011) recognize a significant distinction between the mere having or targeting others with reactive attitudes and outwardly expressing them. In their subtle paper, they explicitly take questions about the justification of having the attitudes to be distinct from questions about expressing them, and they take these questions to be two different moments or stages on the way to “accountability of the strongest sort” (382). Further, they write, “expressing resentment may raise questions of fairness that are not raised by merely targeting an agent with resentment” (393), which seems to suggest that expressing, but not merely feeling, resentment may be a sanction or harm. Notably, though, they do not commit on this question. And further, they explicitly choose not to define “sanction”, taking it to be a “slippery term best explicated by examples, some interpersonal (such as demands for compensation and rebukes) and some institutional (such as fines or jail sentences)”. With this understanding of sanction, they take it that questions of the justification of sanctions is yet a further stage after the questions of justification of outward expression of the attitudes, a claim which would suggest that at least on their understanding of sanction, the reactive attitudes are not sanctions. But even this conclusion would be too hasty, because the further question directly after the questions concerning the justification of outward expression of the attitudes is expressed as follows: “Is it justified, in the circumstances, to impose some sort of sanction (beyond mere outward expression of the reactive attitudes) on [an agent] on the basis of her [performing an action]?” (394). The parenthetical phrase suggests that the expression of the attitudes may be sanctions of a kind after all. Putting all of this together, I conclude that while their remarks are suggestive of the view described in the text, they can also be understood in a way consistent with a lack of commitment to the claim that outward expression is necessarily a sanction in the sense we have been concerned with here.

McKenna (2012) offers a “conversational” view of our blaming practices, in which reacting with the reactive attitudes is a stage in a kind of moral conversation. He argues that there is at least some reason to accept a version of his view that incorporates a desert thesis. In explaining how the view could work with the idea of desert, he appeals to the fact that the expression of the reactive attitudes in response to wrongdoing that are part of a kind of moral conversation is typically harmful and to the fact that this particular kind of harm can be intrinsically good, and thus, deserved. McKenna's view is complex and subtle, and a full exploration will have to await a future occasion. At this point, one might ask whether the fact that the expression of the attitudes is typically harmful is by itself enough to give a unified explanation of how the Reactive Attitudes Thesis and the Desert of Sanctions Thesis are related. If the relationship between the reactive attitudes and sanctions is that the (expression of) the attitudes are typically, but not essentially, a harming response, then it would seem that if we take the appropriateness in the Reactive Attitudes Thesis to be understood as desert, then a revised version of the thesis may seem called for. For example: A person is typically blameworthy to the extent that it is appropriate to adopt (or express) the reactive attitudes toward her. Now one way of avoiding this sort of weakening of the thesis, but still highlighting the fact that the expression of the attitudes typically – if not always – constitutes a harm, is to return to an idea of Feinberg's that I rejected earlier. The idea is that what is deserved should itself be understood only in terms of what is typically, or “generally”, disfavored. In other words, on Feinberg's view, “X deserves response R” can be true even if X welcomes R, as long as R is generally disfavored. This move seems to me to give up the core idea that what distinguishes desert from at least some other notions of fittingness is that what is deserved in any particular case is good or bad for the person. To illustrate: suppose that someone does something seriously wrong. He then receives a response that is neither unwelcome in his eyes nor bad for him in any other way, but it is a response that is “generally” bad for its targets. It seems odd to say that he got what he deserved. Perhaps the response is fitting in an important way, but it is not what he deserved in the distinctive sense at issue here.

For a similar view, see Watson (1996/2004, 278–9) and Wallace (1994, 56, 72).

See Hieronymi (2004) and Graham (2012), for example, for the defense of the view that questions of unfairness do not arise for the reactive attitudes because holding these blaming attitudes is, in some primary sense, a matter of judging accurately.

Pereboom is willing, however, to settle for the weaker thesis that it is close to psychologically impossible for one who expresses her indignation not to believe that the target is deserving of blame.

This might seem no different from (1), but I believe that it is. To see the difference, consider another example: You are walking home and see that you can continue on your normal route or take a shortcut through a neighboring farm. The fact that your neighbors are out of town does not give you a reason to walk onto their property. However, in other circumstances, this fact can contribute to your having reason: suppose now that as you are walking you notice that the door is open and you hear a child's cry. The fact that your neighbors are out of town might now contribute to your being the only one in a position to hear the child and to attempt a rescue. In a parallel way, one might think that someone's being deserving of harm does not give you reason to inflict it, but that it can in some circumstances give you some such reason. For example, you are in a position in which you have no choice but to inflict harm and you can harm someone deserving of it or someone else. This may give you a reason to harm the person who is deserving rather than others.

See Kant (1797/1996), for example.

I am inclined to take the former approach, but have not argued for that here.

See, for example, Hieronymi (2004) and Graham (2012).