2026/07/01 by Victor Tadros
Environmental Science · Arts and Humanities · Social Sciences · #Climate Change and Geoengineering #Philosophy, Ethics, and Existentialism #Political Philosophy and Ethics
paper · doi:10.1215/00318108-12455462
Generally, harm is bad. Is it always bad? Some retributivists—axiological retributivists—think not. They think that harm to people who are responsible for wrongdoing can be good, for they deserve it. In his excellent, intricate, and humane book, Michael McKenna defends a wimpy and qualified axiological retributivism. His retributivism is wimpy in that wrongdoers, on his view, only deserve proportional harm that arises from blame—the harms of having a hampered ability to engage in some spectrum of social intercourse, of a lack of freedom to live a personal life as she wishes, and the unsettling of her emotional stability (61). These harms are not (just) instrumentally good for McKenna—they are good noninstrumentally as part of an organic unity that is intrinsically good. The organic unity consists in the harm occurring in the course of the wrongdoer responding appropriately to their wrongdoing.McKenna’s retributivism is qualified in that it is good for responsible wrongdoers to be harmed in these ways only as a result of being blamed, where blame is part of a conversation that responds appropriately to the wrongdoing of the responsible person. Wrongdoers deserve appropriate blame, and appropriate blame aims at harming responsible wrongdoers in these ways. And his retributivism is wimpy in another way—McKenna admits uncertainty about whether the conditions for desert can be satisfied in particular cases, or in general (chapter 10). And he has concerns about the role of conversational responsibility and desert in socially unjust circumstances (chapter 9).Nevertheless, McKenna defends a retributivist theory of punishment as well as blame along similar lines, though he is uncertain whether states can punish for retributivist reasons (see 90; I was left a bit unclear about the difference between state actors and nonstate actors that might make the latter but not the former permitted to punish on retributive grounds, and McKenna’s view is the converse of the standard, though I think also mistaken, view that states have the unique right to punish). Punishment extends beyond blame in that it involves intentional harm, which blame need not, and it involves imposing sanctions. These sanctions aim at ensuring that wrongdoers recognize that they have acted wrongly and fostering their understanding that they deserve the consequences imposed on them by the sanction. The sanction provides an external source of the kind of harm that it would be appropriate for the wrongdoer to experience if she felt guilty. If she then feels guilty, she will come to see these harms as deserved.Blame, and the harm that results from it, for McKenna, can be fitting. They are appropriate responses to responsible wrongdoing. And, for him, this is the best way of interpreting the Strawsonian project of understanding responsibility for wrongdoing as a feature of interpersonal communicative and emotional practices (chap. 6). He considers the possibility that the hostile emotions such as indignation and resentment involved in blame are by their nature destructive—they can be destructive, he admits, but suitably tempered and qualified, they need not be (chap. 8).This is a terrific book. McKenna’s conversational theory of responsibility is highly attractive. His discussion of desert is nicely integrated into that view. His arguments are careful, often deeply thought out, and humane. And he shows admirable modesty and doubt about the role of desert, even when he is attracted to it. He shows not only great philosophical skill but also a real connection with the emotional lives of wrongdoers and victims of wrongdoing. He is attentive to the significance of historical injustice and the potential harm that endorsing retributivism in the real world can cause. And the wimpiness and qualifications to his retributivism make it in some respects more appealing and less barbaric than other forms of retributivism.In the remainder of this review, I put pressure on McKenna’s retributivism. One weakness of McKenna’s book is that he does not fully consider the merits of other proposals about the moral implications of responsible wrongdoing that are, like retributivism, essentially backward looking but that do not rely on the idea that harm suffered by wrongdoers is noninstrumentally good. These alternatives strike me as better than the retributivist view.Consider one of the harms that McKenna claims wrongdoers deserve—lack of freedom to live one’s personal life as one wishes. First note that such freedom is not an unqualified good on any sensible view. I might wish to live my life as a serial killer. My freedom to do so isn’t good—not for me or anyone else. Why not? Because where I am required not to v, I am already morally unfree to v. If I am morally unfree to v, there is often no value in my being practically free to v. Practical freedom to do something that I am already constrained by duty not to do lacks value. I wish to lie but don’t have the opportunity, so I lack the freedom to do so. No loss!Now, uncontroversially, wrongdoers acquire duties to respond to their wrongdoing. Their wrongdoing results in their being morally less free. The fact that they are made practically less free when they have acted wrongly is then not bad for them. We need not appeal to the noninstrumental goodness of their being unfree to explain this. So we have a straightforward nonretributivist explanation of the intuitive idea that wrongdoers are not free to live their lives as though they had not acted wrongly.Now compare the idea that it is good that wrongdoers feel the pain involved in guilt with a close alternative. It is good for wrongdoers to understand that they have acted wrongly. Understanding this is painful. But the pain is just an inevitable intrinsically bad side effect, in those who are psychologically typical, of the good thing—understanding. And although pain is intrinsically bad, it is instrumentally good because it helps the person to comply with their duties in the future. As understanding is very important, and pain is instrumentally good, it is typically overall better that the person feels guilty as a result of their wrongdoing, even though the pain they suffer is intrinsically bad.McKenna is aware of this alternative but doesn’t do enough to defend the retributivist view against it. How might we adjudicate between McKenna’s retributivism and this alternative? Intuition is a bit unhelpful. When we react negatively to the fact that a responsible wrongdoer is not suffering pain, are we reacting to the fact that it is bad that they don’t understand, or to the fact that is bad that they are not suffering? And how successful are we at screening out the instrumental value of the pain of guilt to focus on its intrinsic qualities? Not very.How do we make progress? McKenna does this by comparing grief. He claims that it is good when a person feels the pain that is involved in grief from losing a loved one. I didn’t have this reaction—I think that the suffering involved in grief is just bad. It is a bad consequence of something good—understanding loss. When one understands the importance of the loved one, and the role that they played in one’s life, one understands the loss at their passing. And this understanding, which is good, inevitably results in the person feeling pain. But I find it quite implausible that the suffering involved is itself good, even as part of an organic unity. Were it so, we would react by valuing the suffering, but we don’t.Here is another way of making progress. We can imagine psychologically unusual people who can fully understand that they are responsible for wrongdoing but who do not suffer the pain of guilt. And we could also imagine that these unusual people suffer no loss in motivation from their lack of pain—understanding does the trick without any additional psychological support. I don’t find it so hard to imagine people like this, but when I do, I don’t feel the loss of anything good. McKenna, of course, responds otherwise.Is there more to say? One thing in favor of the instrumental view is that it is just intuitive in itself that pain is always intrinsically bad. Of course, the idea of organic unities, familiar since G. E. Moore, shows that retributivism is coherent. But it does no more than that. And there is something metaphysically extravagant about organic unities. Where we find alternative explanations for our intuitions, these should be preferred. Finally, we might explore how understanding that one is responsible for wrongdoing leads to the feeling of pain. I cannot outline how best to do this here, but I have done so in other work (see Tadros 2011: 44–51). My suggestion is that once we recognize the source of the pain, we also find it harder to vindicate the view that pain is noninstrumentally good. McKenna could have done more to explore this line of argument to make progress beyond the bare intuitive response to the suffering of wrongdoers.Finally, let us focus on the relationship between McKenna’s retributivism and his conversational theory of responsibility. McKenna thinks that harms to wrongdoers are good only if they result from the kinds of blaming conversations that follow from wrongdoing. This struck me as in one way appealing, in that desert does not float free from human interaction, but in another way less plausible than unqualified retributivist alternatives.Retributivists often focus on our intuitions about the suffering of serious wrongdoers. But many people have these intuitions (or at least initial reactions) in cases where conversation does not, and cannot, occur. Consider: Hitler’s Island: Hitler did not die at the end of World War II but escaped to a remote island where his diaries have been found by an explorer. On reading the last entries, the explorer realizes that he showed no remorse about his conduct after the war. He is yet to find out about the weather on the island. If it was rainy, Hitler was miserable; if it was sunny, he was happy. As the explorer knows that no one else has found out what happened to Hitler, and that he was alone for the rest of his life, what happened to him had no effect on anyone else.I think it better that Hitler’s island was sunny. But many retributivists find this view absurd. McKenna’s view has a similar implication to my anti-retributivist view. However, most of us recognize at least the initial reaction that it is better that Hitler’s island is rainy. To that extent, both anti-retributivism and McKenna’s wimpy retributivism are revisionist. But if we endorse revisionism in this case, why not go all the way? McKenna implies revisionism in this case but not across the board, without fully explaining why we should be more confident in our retributivist inclinations in conversational cases than in this case.Furthermore, the conversational view has surprising implications. Consider: Hitler’s Island II: You are an explorer who arrives on Hitler’s island and you discover that Hitler has fallen into a deep pit. He doesn’t yet know that you are there. As a result of being in the pit, he is suffering. He might come to recognize that this suffering is good. You can engage Hitler in conversation or leave him alone. If you leave him alone, he is more likely to recognize that he acted wrongly and consequently believe that his suffering is good, than if you engage him in conversation.McKenna’s view seems to provide you reasons to engage Hitler in conversation, for only by doing that can you make his suffering good. If he suffers without the conversation, that suffering is just bad. This strikes me as an unappealing implication for retributivists. At least in this respect either full-blown anti-retributivism or unqualified retributivism fare better.There is, of course, much more to say, and I haven’t done justice to the many excellent arguments and ideas in McKenna’s excellent book. While I am yet to be convinced of McKenna’s retributivism, his is perhaps the most humane and thoughtful version on offer.