I agree with John on a non-unitary account of blame and of the treatment of offenders. In the latter case, I argue that the self-defense right and consequentialist deterrence each have a role (and I set out my case for this in Ch. 4 of Wrongdoing and the Moral Emotions). On the naturalness of attitudes that feature desert, Thomas is right that this varies across cultures and across individuals. But even if they are natural for many of us, it may still be that we ought not to justify action and policy on its basis. At the same time, due to its naturalness, if people are victims of injustice, it often isn’t right to criticize people for having these attitudes, or to criticize them for being motivated to act on their basis if so acting has another justification that’s in the clear (I also argue for this view in the new book). Neal, on the backfire effect, it’s real, but we don’t have a clear and comprehensive picture of its causes. Anger sometimes causes it, and sometimes doesn’t, and it can be caused by non-angry interventions. In a study I mentioned, Fieke Harink and Gerben A. Van Kleef (“Be Hard on the Interests and Soft on the Values: Conflict Issue Moderates the Effects of Anger in Negotiations,” British Journal of Social Psychology 51 (2012): 741–52) contend that while at least in the short-term anger is effective in conflicts about interests, it’s not effective in conflicts about values, and this is because in conflicts closely tied to a person’s values, norms, and identity, expressions of anger occasion the backfire effect, that is, an outcome in which value beliefs and actions targeted for revision are strengthened and not weakened by the interaction. (I discuss this material in “Undivided Forward–Looking Moral Responsibility,” The Monist, 2021.) But non-angry interventions can also cause the backfire effect. In "Going Public in an Era of Social Media: Tweets, Corrections, and Public Opinion," Dino Christenson, Sarah Kreps, and Doug Kriner saw a backfire effect resulting from non-angry attempts to modify political beliefs. The point I want to stress is that because interventions of various sorts can result in a backfire effect, sometimes very pronounced, it’s important to keep this in mind when one intervenes, so as to try to avoid subverting the forward-looking aims of the practice of holding morally responsible.

Justin: Thanks for another great post! On the defense of optimistic free will skepticism, I endorse virtually everything Thomas says in reply, and I won’t add anything to what he says. On settling, I agree with John that settling doesn’t require indeterminist free will. One might think of settling by an agent as a kind of difference-making, and there are good compatibilist accounts of this notion. In Carolina Sartorio’s view (“Making a Difference in a Deterministic World,” Philosophical Review 122 (2013): 189-214), moral responsibility requires difference-making in the sense that the agency-involving actual sequence leading to the action makes an agent responsible for the action only if the absence of that actual sequence would not have made the agent responsible for the action. In a recent paper (The Phenomenology of Agency and Deterministic Agent Causation, on my website) I tweak this account to yield a compatibilist criterion for settling. In that paper I also argue that settling requires agent-causation, but that the deterministic accounts of agent causation that Ned Markosian and Dana Nelkin advocate will do. Here is the criterion: (S-AC) An agent settles whether an action occurs only if she agent-causes it, where the absence of her agent-causing the action would not have caused that action. Put in terms of David Lewis’s semantics for counterfactuals, the idea is that an agent settles whether an action occurs only if she agent-causes it and in the closest or most similar possible worlds in which she does not agent-cause the action the absence of the agent-causing would not have caused that action.

Hi Justin: Thanks for the great post and also for the kind words about my new book! You correctly point out that suggestions are less forceful than demands. However, the view I propose allows for more than mere suggestion. In my exchange with Dana Nelkin in Science, Religion, and Culture, which was on Flickers last month, I develop this point with respect to the specific case of ought judgments that apply in friendship. More generally, I want to argue that the demands of moral obligation have a near functional equivalent in (i) ought claims of axiological recommendation; (ii) the appropriateness of protest in situations in which there is a threat that such an ought claim will not be honored; (iii) in some cases, appropriateness of the kinds of forward-looking invasive measures such as incapacitation and rehabilitation that the skeptical view allows, and of forewarning of such measures. You suggest that oughts of specific agent demand are more motivational than oughts of axiological recommendation. But maybe both kinds of ought generate motivation apart from protest and sanctions only when the agent is committed to act accordingly. On Craig’s point about possibility, in the book I argue that there is a ‘can’ requirement on oughts of axiological recommendation. I think that for such claim to be appropriately directed toward an agent, it must at least be epistemically open that the agent will in fact come to act accordingly.

Peter: It’s commonly thought that supervenience accounts of physicalism are too weak; a seminal article on this issue is Jessica Wilson’s “Supervenience-Based Formulations of Physicalism,” Nous 2005, which has occasioned an intricate discussion. There are plenty of options for strengthening the account, and mine is one attempt. One worry for supervenience-based accounts is that the mere metaphysical necessitation of mental by physical properties doesn’t capture the fact that the mental properties are made up of the physical ones. I take this for free in my account, by making the made-up-of relation basic. This is prima facie unsatisfying, and so I would be happy to consider suggestions on which the made-up-of relation isn’t basic. Condition (c) serves to rule out identity, and this is optional on my view. But these controversies are beside the point when it comes to the objection to your account that I sketched. Looking back at the previous thread, this objection is related to questions regarding the reductionist option asked by the philosopher who goes by the pseudonym “Angra Mainyu” (we can be almost certain that this is a pseudonym, for the reason that Angra Mainyu itself would almost certainly opt for elimination over reduction). Let me address your question regarding identity, the paradigmatic reductionist relation. Causal exclusionary reasoning, as I understand it, and as I think Kim understands it, essentially targets proposals according to which some effect E has more than one cause – it will threaten all but one of the causes. So if on some proposal, a dualist or a nonreductivist one, M and P are distinct causes of E, the threat posed by exclusionary reasoning will be neutralized by any response on which the number of causes is reduced to just one. There are two ways to achieve this: a first is by eliminating all but one of the causes, and the second is by identifying the causes. (That Kim takes the identification option seriously is reflected in the Causal Inheritance Principle that he endorses in his discussion of the exclusion argument: The Causal Inheritance Principle: If mental property M is realized in a system at t in virtue of physical realization base P, the causal powers of this instance of M are identical with the causal powers of P. ) The argument that I sketch in my earlier post aims to show that given that identification neutralizes the exclusion threat, material constitution will do so as well. Your discussion with “Angra Mainyu” indicates that you prefer the eliminativist response where exclusionary reasoning is convincing. But exclusionary reasoning all by itself could not show that; this would require an additional argument. One such additional argument Kim considers for elimination is that there might well be no satisfactory mapping of our mental concepts onto the underlying microphysical reality. These claims are relevant to the conclusion of your argument, viz. (iii) there is no causal work left for MEs as such to do. I think (iiia) “There is no causal work for MEs, conceived as nonidentical to microphysical events, left to do,” would be an appropriate way to construe this conclusion. “The following interpretation, (iiib) “There is no causal work for MEs, conceived either as identical to or nonidentical to microphysical events, left to do” can’t be secured by exclusionary reasoning alone, and thus would require an additional argument.

Peter: Let’s grant that you’re right about an exclusion argument stated in terms of causal closure – indeterministic mental causation escapes it. I’m wondering whether the core idea in Kim’s exclusion argument would nevertheless apply to indeterministic actualia. The core idea, as I’m seeing it, is that on the nonreductivist’s view all mental causation would involve redundant, two assassins-style, overdetermination, which is too wild to be credible. Just as the shots of the two assassins do the same causal work, so the mental cause and the microphysical cause would in each instance of mental causation do the same causal work, and this just can’t be. ‘Causal work’ is a notion you use in your statement of the exclusion argument in the previous post. But note that it’s neutral between deterministic causation and indeterministic causation. Imagine a scenario in which each of ten indeterministic causes (C1 – C10) renders some effect (E) 50 likely to occur. Suppose that in the actual situation each of C1 –C10 occurs at the same time, and E then occurs. Intuitively, this could count as a case of redundancy of causal work. The exclusionary concern would then be: on the indeterministic-actualia version of the nonreductivist’s position, all mental causation would involve the sort of redundancy of causal work that we see in this scenario, and this is too wild to be credible.

Peter: As I see it, the core concern that Kim’s EA raises for the nonreductive physicalist is that on that picture, if there were mental causation, then a mental cause (M) and its microphysical realization (P) would redundantly overdetermine the effect (E), in the way that the shots of the two assassins redundantly overdetermine the death of the victim. The compatibilist about mental and microphysical causation (e.g., Stephen Yablo, Lynne Baker, Karen Bennett, Hilary Kornblith, myself) begins by pointing out that the relation between M and P is metaphysically much tighter than the relation between the assassins’ shots: by contrast with the shots, mental entities are realized by or materially constituted by microphysical entities. I prefer material constitution (I set this out in my 2011 mind book). Here is one account of this relation. Suppose x and y are concrete physical entities. The made up of relation is asymmetric, irreflexive, and directed so that the less fundamental is made up of the more fundamental, while its core is primitive. Entities x and y are materially coincident just in case they, at some level, are made out of the same parts. And suppose ‘D’ designates the y-favorable circumstances—the relational context required for something to constitute y (this idea comes from Lynne Baker's account). Then: (C) x materially constitutes y at t if and only if (a) y is made up of and materially coincident with x at t; (b) necessarily, if x exists and is in D at t, then y exists at t and is made up of and materially coincident with x at t; and (c) possibly, y exists at t and it is not the case that y is made up of and materially coincident with x in D at t. But the specific details don’t matter all that much for the answer to Kim: If, as on Kim’s view, identity and not just (current) material constitution were required to preclude the sort of causal competition between M and P that generates redundant overdetermination, there would be a feature required for avoiding redundant overdetermination that identity has and current material constitution does not. The candidate relevant features that identity has and current material constitution lacks are constitution at all other times and at all other possible worlds. But then it would have to be the actual or possible absence of material constitution at some past time or at some future time, or the possible current absence of material constitution that generates the sort of causal competition issuing in redundant overdetermination. It’s difficult to see how this could be so – how such temporally or modally extrinsic properties could make a difference for this sort of causal competition.

Hi Mark: On your response to Tom and Peter, under 2ii, at the the end of the first paragraph, you make a crucial move: "To this I might respond: “I don’t know; I just did. I had to choose right then, I couldn’t wait, and I was torn, so I just sort of randomly picked.”" I think what you say here fits the phenomenology of certain torn decisions. And when I interpret the "I" in this quotation as referring to an agent-substance, and the picking as agent causation, I can make sense of what you say. But you're an event-causal libertarian. So how do we interpret this quotation, in particular the "I just sort of randomly picked" part, in event-causal terms?

Dana: Thanks for the objection, which I think is hard to answer. One possible route is for me to avoid saying that respecting and exercise of rights – for example the rights to liberty and self-defense-- count as good consequences, and to say only that liberty and self-defense have a particularly significant weight in the calculation. You’re right -- this would be revisionary, but it doesn’t seem to run into the sort of problem you’re posing. Mark: Thanks for the challenge. It’s important to avoid the difference between the first two cases that you develop, and I think it can be done. Due to the direct intervention of the neuroscientists, the egoistic quality of Plum 1’s practical reasoning process is increased relative to what it would have been without the intervention -- not a lot -- but enough so that he’s now causally determined to decide to kill. Due to some environmental stimulus, say hearing some bad news, the egoistic quality of Plum 2’s reasoning process is increased to the same degree relative to what it would have been without hearing the news, enough so that he’s now causally determined to decide to kill. And in neither case does it seem to Plum that he’s acting out of character.

John and Sofia: Thanks very much for this question. The argument against the basic desert sense or aspect of blameworthiness I develop in the book is that for an agent to basically deserve a harmful response she must have a kind of free will or control in action that we don’t have. But I agree with Sofia that there are other reasons to be skeptical of a sense of blameworthiness that involves basic desert. One concern is that for a number of contending general normative ethical theories the notion of desert seems to have the role of an awkward supplement. For instance, despite Kant’s invocation of desert in justifying criminal punishment, that appeal appears unrelated to any formulation of the Categorical Imperative, which he held to be the supreme and comprehensive moral principle. Another issue is that negative basic desert seems to involve the idea of harm as an intrinsic good, which is at least prima facie dubious. On the related question of the defensibility of the retributive theory of punishment, in the book (and also in Living without Free Will) I argue that the concerns about control in action yield a challenge to a theory that’s in jeopardy for other reasons, including those just mentioned. A further skeptical worry is that the intuitions that drive the retributivist theory are at root fueled by vengeful sentiments, and that therefore retribution has no more plausibility than vengeance as a morally sound policy for treating criminals. Acting on vengeful sentiments is, in general, immoral for the following sort of reason. Although acting on such sentiments can bring about pleasure or satisfaction, no more of a moral case can be made for of acting on them than can be made for acting on sadistic desires. Acting on sadistic desires can bring about pleasure, but in both cases acting on the impulse in question aims at the harm of the one to whom the action is directed, and in neither case does acting on the impulse essentially aim at any good other than the pleasure of its satisfaction. But then, according to the objection, because retributivist intuitions have their source in vengeful sentiments, acting for the sake of retribution is also morally wrong. One might respond by arguing that retributivist intuitions are independent of vengeful sentiments, or that vengeful sentiments are legitimate when they are in turn grounded in a plausible sense of justice. But I think that the skeptical worry is strong enough to show that the retributive theory does not meet a plausible epistemic standard for justification of harm, as Ben Vilhauer has also contended. But these alternative considerations are all controversial, and so the argument from the absence of control in action has an important role to play, and to my mind is interesting in its own right.

That’s an interesting new study, Jonathan. In my experience, philosophers have a fairly strong tendency to reject the claim that intentional manipulation is a relevant difference maker, so we don’t have much of a history of philosophical explanations of that proposal. Eddy and Oisin’s paper will break new ground. Here’s an explanation that takes a cue from their interventionist proposal, and also fits with Alan’s luck suggestion. A significant part of our practice of holding morally responsible aims at moral formation. On Manuel Vargas’s and Daniel Dennett’s view, even the ultimate point of attributions of moral desert is a forward-looking goal of this sort (here desert is non-basic). With this in mind, as a subject in Jonathan’s survey I’d attribute less responsibility in cases of intentional manipulation than in the contrast cases. My sense would be that intentional manipulation poses the greater threat to the forward-looking efficacy of blaming, and hence to justified blaming, and on the Vargas/Dennett position, to justified attributions of deserved blame. But this explanation does not affect the core incompatibilist claim – the one I intend my manipulation argument to support -- that justified attributions of basic desert are incompatible with causal determination.

Eddy, I’m looking forward to seeing the final version of the paper by you and Oisin. I agree that it might be hard for subjects, generally speaking, not to think of robots as acting intentionally. But in response to concerns of that type, Gunnar Björnsson constructed a scenario modeled on the Plum examples in which a non-intentional cause—a bacterial infection—slowly makes the agent more egoistic without bypassing or undermining his agential capacities. His thought was that if subjects were prompted to see the agent’s behavior as dependent on this non-intentional cause, this would undermine attributions of responsibility to roughly the same extent as the introduction of an intentional manipulator. This was indeed the case. In a study involving 416 subjects, the infection scenario undermined attributions of free will and moral responsibility to the same degree as intentional manipulation.

As a proponent of a manipulation argument, I would say that it’s a presumption of the argument that its target audience would initially tend to think that agents who are causally determined to act badly in a natural way – without intentional control, deserve to be blamed, while people who behave similarly but are intentionally causally determined do not deserve to be blamed. So the experimental result fits nicely with this presumption. Most people enter into the free will debate with the assumption that ordinarily agents deserve to be blamed when they knowingly do wrong. For the natural compatibilist, the prospect that whenever we act, we are causally determined by factors beyond our control wouldn’t change this assumption. Incompatibilists believe that this reaction fails adequately to face up to the implications of causal determination. The way manipulation arguments aim to address this concern is by first devising an intentional deterministic manipulation case with the hope that it will be more successful at eliciting a non-responsibility intuition it its target audience than ordinary causal determination does. Crucially, the next stage involves having the compatibilist come to believe, upon rational reflection, that such non-responsibility is preserved even when the intentional manipulation is subtracted, on the ground there is no responsibility-relevant difference between the case that features intentional manipulation and one that doesn’t. Josh says: (1) The intentions of the manipulator have no bearing on the degree to which the agent is morally responsible. Now suppose we learn a surprising empirical fact: (2) Given the way human cognition works, people’s intuitions about whether the agent is morally responsible are influenced by their perceptions of the manipulator’s intentions. (C) Together, these two claims would give us reason to think that our intuitions about manipulation cases were systematically mistaken. This would then to some degree undermine a crucial premise in the manipulation argument for incompatibilism. The proponent of the manipulation argument agrees with (and presupposes) (1) and (2), but is wary of (C) -- some reason, maybe, but not a lot. The question here is what the result of rational reflection will be when the target of the manipulation argument rationally reflects on the fact that she accepts non-responsibility in the intentional determination case but denies it in the ordinary deterministic case, despite no difference in the satisfaction of the standard compatibilist conditions on moral responsibility. There are three salient possibilities. The first two involve accepting the no-responsibility-relevant difference claim: (a) either deny deserved blame in both cases (e.g., Patrick and me) or (b) accept deserved blame in both cases (e.g., Michael McK. & Carolina S). I think that accepting deserved blame in the intentional determination case is more of a stretch than denying it in the ordinary deterministic case; Michael and Carolina disagree. It sounds like Josh and Matt accept a third option, that intentional determination really makes a difference to deserved blame. Suppose you know that someone’s bad action satisfies all of the standard compatibilist conditions on deserved blame, but you don’t yet know whether it was intentionally causally determined by an intelligent agent or causally determined by a robot without intentions – and let’s say the robot is not in turn controlled by an intelligent agent. Is it reasonable, upon reflection, to believe that whether he deserves blame depends on which way it turns out? My sense is that it isn’t.

Kristin: I agree that if the basic argument were sound, it might well trump any explanatory considerations that arise from a manipulation argument. But if the basic argument isn’t sound, then a manipulation argument can do explanatory, diagnostic work -- maybe not some austere formulation of a manipulation argument all by itself, but rather conceived as an argument to the best explanation. If we conceive of the manipulation argument as an argument to the best explanation, we shouldn't expect that the sort of formulation of the zygote version you present (for example) will reveal everything that's relevant to the best explanation. In particular, we shouldn't expect that formulation to show that the basic argument is unsound. Best explanation arguments need to rule out the relevant competing explanations, and this task may be wide-ranging. Kip: Here’s the kind of consideration that Randy Clarke raises in his 2005 response to the basic argument, which convinces me. Suppose the agent-causal libertarian conception of agency is coherent, and we’re created as agent-causal libertarian beings, preprogrammed with a set of strong self-interested motivations and set of roughly equally strong altruistic motivations, and no other kinds of motivations (e.g., motivations to do evil for evil’s sake). Basic argument-style reasoning might well show that we’re not responsible for the fact that our actions are all either self-interested or altruistic. But I can’t see that it shows that when it’s up to someone to make either a self-interested choice or an altruistic choice, and she makes the altruistic one, she’s can’t be morally responsible for making the altruistic one rather than the self-interested one. Randy’s response is in “On an Argument for the Impossibility of Moral Responsibility,” Midwest Studies in Philosophy 29 (2005), 13-24.

Hi Kristin: Here’s how I see it. What rules out moral responsibility (in the basic desert sense) generally is that agents don’t actually have responsibility-relevant control. There is more than one way for this control might to be precluded. Determination by factors beyond the agent’s control is one way, indeterminism in event-causal contexts is another. When causal determination rules out moral responsibility, it does so because the action is deterministically produced by factors beyond the agent’s control, and when indeterminism precludes it, it does so because it does not allow for the agent to settle whether the decision will occur. The disappearing agent argument is my favorite way of showing the latter, the manipulation argument the former. One might suspect that the impossibility of self-creation explains the intuitions in the manipulation argument. But for the kinds of reasons Randy invokes, which I mention in the previous thread, I don’t think the basic argument works. So in the deterministic case, I’m left with the intuition that agents are not responsible because they’re causally determined by factors beyond their control. Crucially, in my view it’s the failure of the basic argument that helps confirm this as the best explanation of the intuitions in the manipulation argument. But this does not indicate that the manipulation argument is the façade of another argument (i.e., the façade of an objection to the basic argument).

Hi Kristin, In answer to the questions at the end of your reply to me: On your definition of incompatibilism, “no one (subject to the laws of nature) has free will because deterministic laws obtain,” could someone be a self-creator in Galen’s sense if all of her states and actions are causally determined in this way? So first, thinking about Joe Campbell’s arguments, we might ask whether an agent whose every state is causally determined by a previous state could be a self-creator if that agent has no beginning in time. Maybe – imagine that this agent’s history features appropriate self-creating states at certain intervals, going back in time to infinity. But – inspired by Carolina’s reply to Joe -- given causal determination, an agent whose existence has a beginning in time prior which there was a past can’t be a self-creator, and it seems to me that such an agent can’t be a self-creator because all of her states are causally determined by her remote past. So here the basic argument would seem to yield a case for incompatibilism, given your definition (and Galen's suppositions about the connection between free will and self-creation). There will also be certain kinds of indeterministic histories that rule out self-creation, but this does not preclude deterministic histories, given a remote past, also explaining the impossibility of self-creation.

Hi Kristin: I agree with you and John that the Zygote argument, if successful, does not all by itself establish incompatibilism rather than just incompossibilism. At this point we have to consider what might best explain the intuition that Ernie is not morally responsible, and/or the sense that there are no relevant differences between Ernie and the ordinary human being. There are a number of candidates on the table. Setting aside the compossibilist ones, e.g., that Ernie is not responsible only because he is manipulated by another agent, I agree with what you’re suggesting, that Galen’s explanation is one option; i.e., that causal determination (of an agent with a remote past) rules out moral responsibility only because it precludes self-creation, while the option I favor is another, i.e., that causal determination by factors beyond one’s control, or as Carolina puts it, by factors beyond one’s causal reach, per se precludes moral responsibility because it rules out the sort of control that (basic desert) moral responsibility requires. So at this point the incompatibilist needs to invoke an objection to the basic argument, and here my intuitions are similar to Randy’s. Suppose we’re created as agent-causal libertarian beings, with a set of strong self-interested motivations and set of equally strong altruistic motivations. It seems to me that so described we might well be morally responsible for choices between self-interested and a moral options, despite our not being self-creators in Galen’s sense – such choices could still be up to us. But dialectically, the upshot is, as you suggest, that more work needs to be done to establish the incompatibilist diagnosis of the Zygote argument as opposed to non-incompatibilist competitors.

There might be a sense of ‘guilt’ given which I’m OK with a feeling of guilt being fitting. Suppose that feeling guilt does involve of representation of oneself as blameworthy, as you suggest, but one sense of ‘blameworthy’ is 'being an appropriate target of blame in a largely forward-looking and non-desert-involving sense.' Then I'd agree that a feeling of guilt can be fitting. But I definitely accept the fittingness of a pained or unpleasant response upon having done wrong.

Hi Randy, I think I can agree that in addition to wrongdoer’s recognition that he has done wrong, the unpleasant feeling that you describe makes for a morally fitting addition, without accepting that this gives us a good reason to believe that the wrongdoer deserves to have the unpleasant feeling. In her 1990 book, Bruce Waller proposes that such pain can be explained by the recognition that one did not live up to one’s own moral standards, and is fitting for that reason. Hilary Bok, in her 1998 book, develops a similar story. This explanation doesn’t seem to involve desert, at least not obviously so. By analogy, one might feel bad that on some occasion one didn’t meet one’s standard for chess or piano playing when one completely understands that one’s substandard performance is due to factors beyond one’s control, and one fully believes as a result that no pain is deserved. Alternatively, Ben Vilhauer (2004) develops an account of this feeling that grounds it in sympathy with those one has wronged, on which the feeling is fitting because the sympathy is morally appropriate. This account also need not invoke desert. But intuitively, what Bruce, Hilary, and Ben propose can’t amount to the complete story, for as my grad student Austin Duggan pointed out, we believe that a pained response to one’s wrongdoing is fitting for those who aren’t predisposed to feel bad upon the violation of their moral standards, and who lack sympathy with those who have been wronged. But I’m thinking that this belief can be accounted for by the largely forward-looking goals of moral reform and reconciliation. For those who lack the predisposition to feel pained in this way, feeling pained upon having done wrong would nevertheless be valuable because it is apt to result in moral change and potentially in reconciliation with those who have been wronged, and is fitting for this reason.

Although I think that vengeful and venomous responses to manipulated Ernie, and any that involve the supposition of basic desert, are ruled out, plenty of what’s involved in our practice of holding morally responsible is unaffected. I argue that several forward-looking aspects of the practice, those that aim at protection, reconciliation, and moral formation, are unaffected. And Randy and Dana have pointed out to me that in the case of blame those aims have backward-looking aspects as well – it’s important that the agent in fact acted badly, and in the case of reconciliation addressing the nature of the immoral action is crucial. Moreover, these forward-looking goals may often demand a tough-minded response. And since these aspects of the practice are compatible with causal determination, I’m committed to the claim that they are appropriate for manipulated Ernie, and for Plum in each of the manipulation cases. So for these largely forward-looking aspects, the hard-line response is the right one.

Subrena Smith (Cornell) hired by the University of New Hampshire. AOS: Philosophy of Science, Philosophy of Biology, Epistemology. Raoul Saucedo (Cornell) hired by the University of Colorado, Boulder. AOS: Metaphysics, Philosophy of Logic. Previously Assistant Professor at Yale University. Scott O’Connor (Cornell University) hired by New Jersey City University. AOS: Ancient Philosophy, Metaphysics.

Hi Hanna -- I’m struck by your ‘stark choice’ question: “You can either hold on to the idea that the emotions, thoughts, and actions constituting affective blame are deserved by those who do wrong and indulge in them, or you can give them up and reduce the amount of wrongdoing and harm in the world.” I definitely sympathize with the underlying idea, partly because I want to reject desert attributions. So do you think that your stark choice applies to non-clinical cases? If you do, suppose that Eddy is right, that in non-clinical cases desert attributions are nevertheless true. Even then, would it turn out on your view that such desert attributions, even if true, should have no role in how we in non-clinical cases react to bad behavior, since our aim should be to reduce the amount of wrongdoing and harm in the world?

Why do we think that the retributive emotional reactions are virtuous? Suppose a family member is intentionally harmed by some hostile person. In a Jeffrie Murphy-style thought experiment, we compare the retributive reaction to the emotionless one, and we clearly prefer the retributive reaction. But we do so because the retributive reaction indicates love and loyalty by contrast with the indifference that the emotionless reaction suggests. So supposing these are the alternatives, we strongly prefer that the person in question have the retributive reaction, and this is because it’s an indicator of virtues such as love and loyalty. But the specific content of the retributive reaction still faces the challenge that Tamler specifies in his post – it does not aim at well-being, and in this respect it does not seem to differ from the content of a sadistic desire. There’s a nice solution: consider it a good thing (at least prima facie) when relevantly situated people have the retributive reactions, but when it comes to justification of the way the target of the reaction is to be treated, set the content of the reaction aside.

On Tamler’s conversion, I can definitely feel the force of his reasons. But I can live with the following view. While it’s mistaken to believe that the person who deliberately harms a family member basically deserves to be harmed, we can also affirm that for natural and perhaps historical reasons we’re retribution machines. And although we have an impressive capacity to modify this aspect of our character, it is limited. So while in the kind of harm case at issue the basic desert belief is mistaken, it’s inevitable, and in certain circumstances we may not be able to keep ourselves from acting on it. We then can accept ourselves for the kinds of beings we are and for what we inevitably believe and do as a result. Here I’d resist classifying having the basic desert belief as ‘wrong,’ and given certain kinds of circumstances, I’d be tempted to say the same about the resulting reprisal. But I don’t think that this story makes it legitimate for the state to invoke basic desert in its justification of the treatment of the kinds of violent criminals at issue.