Blog Coordinator

« 'Freedom and Resentment' Turns 50 | Main | APA Commentary »



Feed You can follow this conversation by subscribing to the comment feed for this post.


I think it would be helpful to distinguish two kinds of "desert." First, there is a consequentialist version of desert whereby an able-minded agent deserves to be reprimanded or punished for her bad behavior (however mild or heinous) so long as these responses produce some positive benefits for the victim, the offender, or society more generally (or all three!). Let's call this instrumental desert. Second, there is a version of desert that one finds in retributivists like Michael Moore and Kant whereby it's intrinsically valuable to proportionately reprimand or punish an able-minded agent for her bad behavior (however mild or heinous). Let's call this intrinsic desert.

I am not a skeptic about instrumental desert--although I might prefer a word other than "desert"--while I am a skeptic about intrinsic desert. Moreover, while I don't think instrumental desert requires something like libertarian-style free will, I do think that intrinsic desert requires a kind of ultimacy that is lacking in compatibilist accounts. It's for this reason that I don't think compatibilist free will is compatible with retributivism--properly conceived, of course!--even though it is compatible with consequentialist accounts of punishment and responsibility. But we've had this conversation before, no? :)

I'm not sure I understand what instrumental desert is. Consider one reason to be skeptical about it. On the notion of 'intrinsic desert', we can at least separate the following two questions:

1. Does S deserve punishment?
2. Should we punish S?

These seem to me conceptually distinct questions, even for a staunch retributivist. (They may get the same answer necessarily, if we assume that one should always give others what they deserve. But this would be a further substantive commitment.)

But for 'instrumental desert', the two questions are the *same* question. If S deserves punishment just in case it is best to punish them, then, on a consequential theory, we can't separate the two questions. This makes me think that 'instrumental desert' doesn't really identify anything unique at all. What's left to say about 'desert' on the instrumental reading that doesn't apply to 'should'?


Like I said before, I am not entirely comfortable calling instrumental desert "desert." I guess I was merely trying to get at two different senses in which someone could be said to be an apt target of reprimand or punishment. I only drew the distinction between instrumental and intrinsic desert to remind Eddy that skeptics about free will and intrinsic desert can provide a consequentialist account of why he was an apt target of honking--indeed, I would have honked at him myself given the circumstances!--while remaining skeptical about desert in the deeper sense normally associated with retributivism. I could just as easily have said: Eddy didn't deserve to be the target of someone's honking, but the honking was nevertheless justified on other grounds. Moreover, these "other grounds" depend on several of the cognitive capacities Eddy highlights. Eddy's post made it seem like skeptics about desert must somehow think the honking was unwarranted or unjustified--which is simply not true. The issue's not whether the person should have honked. The issue is what precisely justifies the honking.

Eddy, I do address these "trivial" cases in my book. I pick 2 (a). We have all the compatibilist powers you want, but we can't bring it about that there are desert-entailing differences between them through their exercise. There is no quick story about why this is the case: I will give you different arguments depending on what you think the sufficient conditions for responsibility are. But they will all turn on luck. Either present luck undermined your control such that there is no desert-entailing difference between you and another similarly situated agent who performed a different act, or constitutive luck features in your history to bring it about that a historical condition is not satisfied. I have different arguments for those who reject the historical condition.


Seems to me like Lucas is right, that you did deserve it, though I'd be inclined to interpret his statement as meaning something like this:

"You just did some blatantly rude that you knew would probably lead to other drivers being mad at you, and you've got no good excuse for doing it. If he were to come over to our car and yell at you, all you could say for yourself is 'Yeah, I know. I'm sorry.'"

Does that seem like what Lucas had in mind? If so, I'm not sure it says anything about how you ought to be treated or even how you can appropriately be treated. It seems more like saying, "You've got nothing to say in your defense here, dad."

Do you mean low-level desert, or desert of low-level things (e.g., mild rebuke)? The post seems to concern the latter, in which case we needn't talk of degrees of desert.

Neal, it does seem to say one thing about how he should be treated: "it was appropriate to honk at you, Dad." Right?

Neil, but doesn't it seem implausible to interpret "Eddy deserves that honk" as "it would be appropriate to honk at Eddy unless there is no difference between Eddy and a similarly situated person that doesn't ultimately boil down to luck"?

Tamler, if we understand desert as Neal does, then I'm fine with it. If honking is a signal (notice you're inconveniencing me), then of course it is appropriate only wrt someone situated like Eddy and not someone who is not inconveniencing someone. As usual, I understand desert as entailing something tha would not be justified on consequentialist grounds alone.

Hey Tamler,

I'm not even sure I'd go that far. It seems to me that all Lucas is saying here is a vicarious "Touché."

The situation here seems to me similar to that of a well-informed chronic smoker who ends up getting lung cancer. The smoker could sensibly say, "Well, I guess I deserve this," and simply mean that he has nothing to say in his defense, but that doesn't have to mean anything stronger about whether it really is appropriate for him to have cancer or whether he really deserves to die in any deeper sense.

In Eddy's case, "You deserved that" seems akin to "You asked for that" or "This is your own fault". And that seems right, even if the world is shot through with luck.

Neal, but I think Eddy's point is that there is that desert in a deeper metaphysical sense than "I asked for that" or "I had that coming" exists only in the mind of the metaphysician (to paraphrase Strawson)and is not part of the ordinary way we understand the concept.

About the smoker, doesn't "I deserve this" suggest that it's more appropriate that it happened to me than to a non-smoker?

Neil, what consequentialist justification? The point is that the honking is appropriate even if it doesn't hurry him up now or in the future.

Tamler, a justification can be consistent with consequentialism without being a justification on consequentialist grounds. If "Eddy deserved that" is equivalent ( here) to "Eddy brought it upon himself" ( as I take Neal to suggest), then it is appropriate.

As for consistency with what ordinary people say or believe, I'm a honey badger.


Can you say a little more about what constitutes a legitimate "defense" on your picture? One thought is that without standards of appropriateness, anything (and equally well, nothing) counts as a defense.

A separate thought is that considerations of luck might counts as defenses. Eddy's 'defense' might be that others don't get honked at who, but for luck, didn't hold traffic up; or, alternatively, that he only held traffic up due to luck.

Tamler: I'm not exactly sure what you are saying in your first paragraph. As for the smoker case, I think you're right that it does suggest some deeper sense of "justice" in the world or something, though I also think that need not be what the smoker means.

Matt: I was just thinking of all the ordinary explanations that Eddy might give in other similar situations: it was an emergency, he didn't realize he was sticking out into the next lane, etc. Lucas seems to be saying that none of those things apply in this case, since Eddy knew exactly what he was doing and did it anyway. (The sort of luck that worries philosophers never figures in those sorts of ordinary defenses.)

Does this mean that Lucas must say that the honking was appropriate or warranted? Maybe, maybe not -- it depends on what Lucas's general views about honking as a social practice are. But I suppose he could say this: if the honking is *inappropriate* in this case, it's not because of anything about Eddy.

Neal, imagine that I said: "Well sure, I guess there is a sense of love that can be described in terms of the feelings and relationships people have for each other. But I'm talking about REAL love, which involves the intertwining of two non-natural souls. It's this deep kind of love that I'm skeptical about."

A natural response would be:

"Well yes, you're right to be skeptical about that, but why do you think that real love involves the intertwining of two souls?"

Similarly, when you say that you have no problem with desert in the sense of 'I had this coming' and 'it's my own fault,' but that you're skeptical about the deeper sense of desert (or think it's incompatible with determinism)," a natural response might be: 'why do you think desert involves anything deeper than 'I had this coming?.'

I hope that makes the paragraph a little clearer but maybe it just muddies the water even more. I should probably just let Eddy stick up for himself since he's the anti-skeptic anyhow.


Some (like Thomas) think that the deeper incompatibilist sense of desert is necessary to justify non-consequentialist punishment. But as John has said here a couple of times, compatibilists think that it's an *intrinsically* good thing that offenders suffer, no good consequences need ensue.

To say "I had it coming" might simply be to say that I've evoked an expected response on the part of those I've offended, a response that likely serves to shape my future behavior (which is why we evolved to respond that way). It isn't necessarily to say that I should get honked at whether or not it serves to shape my behavior.

I don't get why compatibilists think that suffering of offenders is intrinsically good. They might very much want to see offenders suffer, but that makes suffering good in their eyes, not intrinsically.

It feels like you posted a desert drive by, Eddy. I think Lucas would agree you deserve some cyber-honking! I know Sam would agree with me!

Yeah yeah, I abandoned my baby (the post, not my real one). I could make excuses, but I deserve to be reprimanded (one should not post unless one knows one will have time to respond to comments). And I mean "deserve" here in just the same sense of the word as I would mean it if I said (about a hypothetical case!), "I deserve to have Lucas and Sam removed from me and to go to jail (or worse) since I abandoned my baby girl [knowing it was wrong and having the opportunity not to, etc.]." The difference would be in the degrees of blame and punishment I deserve for the two acts. (I'm not sure if that answers your question, Randy.)

Alas, assuming the previous paragraph does not satisfactorily answer the rest of y'all's interesting questions and comments, I'll have to put it off for now. Or I can just let my post's adoptive parent, Tamler, take care of it.

In any case, my main goal was to get some action going here at Flickers, which needs it. I clearly deserve credit for doing that!


I think it's important to note that a compatibilist can deny that "the deeper incompatibilist sense of desert is necessary to justify non-consequentialist punishment" without resort to the claim that "it's an *intrinsically* good thing that offenders suffer". Indeed, we can deny that an incompatibilist sense of desert *is* a deeper sense of desert.

Two things are worth thinking about here. First, are there two different senses of desert going round? One, compatibilist; the other, incompatibilist? I'd be surprised if this were the case, but I'd certainly like to hear more about these senses if it is.

Second, one's theory of responsibility doesn't seem to determine one's theory of punishment. I find nothing at all contradictory about being a compatibilist (or libertarian!) about responsibility and adopting a consequentialist theory of punishment. Or, say, a quarantine model of punishment. I can consistently hold that the blameworthy deserve blame without holding that anyone deserves to suffer.

That's pretty weak sauce, Herr Nahmias. No wonder people honk in your direction. It nevertheless makes some sense to credit you for getting the discussion started--e.g., you weren't compelled to write the post, you did it with good intent, and applauding your efforts will hopefully make it more likely you and others will contribute in the future--so at least there's that :)

That being said, earlier I tried to give a very brief account of a way to justify the honking that didn't make any appeal to “deep” or intrinsic desert. Instead, I mentioned the kinds of instrumental reasons that one could appeal to as being sufficient to justify reprimand or punishment in the absence of desert altogether—e.g., you were reasons responsive, you had the capacity to understood the traffic rules both formal and informal, you didn’t have an excuse or justification, honking at you enabled the irritated driver to vent and publicly express both his and other drivers’ annoyance at your driving, the deterrent effect it would potentially have not only on you, but also on others who may be tempted to drive selfishly in the future, etc.—just as I can appeal to these same factors to justify reprimanding very young children and pets. What additional justificatory role then does desert play on your view?

On the traditional retributivist view, desert was something that justified the reprimand and punishment regardless of whether the negative reinforcement would yield any “happy surpluses” in terms of positive consequences. Yet on what I take to be your instrumentalist view of retributivism, it seems like you just want to say that if and when reprimanding or punishing someone is justified on instrumental grounds, they deserve it. This is what Matt was quite correctly getting at earlier with his remarks about my admittedly wrong-headed attempt to use the term “instrumental desert.” It makes it impossible to say something that I think we should say in both low level and high level cases—namely, we can be justified in reprimanding and punishing a person even if they don’t deserve it so long as certain instrumental conditions are satisfied.

But even if you don’t agree that we should say this, I think it’s clear that it ought to at least be something that is conceptually possible to say. Yet, for those of you who are trying to motivate what you take to be retributivist desert-based theories of responsibility and punishment with compatibilism and “instrumental desert,” it becomes impossible to distinguish between someone deserving reprimand and punishment and it making sense to reprimand or punish them (here again, a restatement of Matt’s earlier point). In short, you need “P deserves to be punished” to be distinct from “it makes sense to punish P”, yet if you appeal to instrumental desert rather than intrinsic desert, you can’t draw the distinction (by definition).

Another problem with the way you go about trying to ground “desert” in the honking case is that it looks to me like a slight of hand since you make it sound like the compatibilist is delivering the goods when it comes to desert and the skeptic has to concede that the honking wasn’t justified. After all, on your view, if the honking was justified, you deserved it. And since the skeptic denies you deserved it, the skeptic must on pain of irrationality conclude the honking wasn’t justified. But this simply isn’t the case. I reject the desert of the pure retributivist but I can still avail myself of all of the justificatory tools in the consequentialist tool kit.

If my last comment didn't draw you in, Eddy, then perhaps this longer and even more tedious one will! For the rest of you, if you're already sick of the line I usually push on this front concerning retributivism and the intrinsic value of suffering, please don't read on! You've been warned!

Contrary to what you seem to assume in the post, I don’t think you and the traditional retributivist mean the same thing by desert. The traditional retributivist thinks that desert is something that grounds the reprimand or punishment not something that flows from the fact that someone has been justifiably reprimanded or punished. On this more traditional view, “happy surpluses” and instrumental goals play no justificatory role when it comes to grounding punishment. Desert alone is both necessary and sufficient (think Kant's island). Yet your account can’t possibly capture this intuition since your account defines desert partly (if not entirely) in terms of the punishment’s being justified on other instrumentalist grounds. But that means that an important strand of retributivism isn’t adequately captured by your instrumentalist version of compatibilist desert. I also don’t think your instrumentalist view settles with folk morality.

Since I know how you love data on folk intuitions, consider, for instance, the following explicit statement of retributivism:

People who commit crimes deserve to be punished even if punishing them won't produce any positive benefits to either the offender or society—e.g., rehabilitation, deterring other would-be offenders, etc.

In a project I am working on with Shaun Nichols, Deanna Kaplan, and Saeideh Heshmati, we asked people to state their agreement with this statement on a scale from 1 (strongly agree) to 6 (strongly disagree). The average response was 2.74. More importantly, participants’ agreement with this statement positively correlated both with retributive judgments in response to an economic game theory task and with their scores on the agency and responsibility subscale of the scale that you and I are working on with Chandra and others. In short, there is an important strand of non-instrumentalist desert in folk morality that your view doesn’t capture.

Now, perhaps you don’t want your view to capture folk retributivism. Perhaps you have an error theory at hand to explain away the view. Perhaps you’d prefer to just go revisionist along with Manuel. But then you should be more explicit that that’s what you’re doing rather than making it seem like skepticism and libertarianism are radical departures from common sense while compatibilism captures it (at least in the present case). Either way, it is clear, at least to me, that your view—along with the views of nearly all other compatibilists who don’t follow Manuel down the road to revisionism and who also don’t follow Moore down the road to pure retributivism—doesn’t capture a prominent feature of punishment and desert that has not only been important to traditional retributivists but which is also an important part of folk morality.

As a skeptic, obviously I am not in the business of trying to capture folk morality. Instead, I am trying to subdue it by highlighting both its empirical and conceptual confusions as well as highlighting what I take to be the morally suboptimal policies that flow from it. But for someone who wants to and claims to be rescuing folk morality from its libertarian and skeptical assailants, I think you have a lot more work to do. If I have shown anything in this comment thread, I hope it’s that I can coherently claim that the driver behind you was justified in honking even if you didn’t deeply deserve it—i.e., even if honking at you wasn’t intrinsically valuable. It would be question begging to insist in response that if the honking was justified on the grounds I mentioned, then you deserved it. What I want to know is whether you think you would have deserved it even if positively no good would have come from it (either for you or anyone else). If so, then you are pure retributivist after all, contrary to what you have stated in the past. If not, you were apparently turning left down the road to revisionism that day—which didn’t settle well with the traditional retributivists stuck in their cars behind you. They were happy to self-righteously honk come what may since you deeply deserved to be annoyed in turn (just not merely in that wishy washy compatibilist way!).

Now if that doesn’t pull you into the discussion, Eddy, I throw up my hands. But I won’t defer to Tamler. That guy flip flops more than Van Inwagen’s counterparts who suddenly find themselves in deterministic worlds!

First of all, I like to think of it 'open to changing my mind.' Second of all, my overarching position (and the topic of my new book) metaskepticism allows for flipflopping on first order questions since there is no fact of the matter about the conditions for desert.


OK, here's what I think I believe:
1. You ask the folk: "People who commit crimes deserve to be punished even if punishing them won't produce any positive benefits to either the offender or society—e.g., rehabilitation, deterring other would-be offenders, etc." and most agree.
I think I would agree with that statement too. I don't know if that commits me to saying that suffering is "intrinsically good" for anyone, but if so, I think I'm willing to accept that (if that makes me a bastard, I guess I'll have to accept that too!). I do not think the folk are committing themselves to agent causation, souls, or impossible powers of self-creation by agreeing to that statement. (We need to figure out how to test that.)

2. The point of the post was to suggest that it seems even stranger to say (or suggest the folk think) that such libertarian powers are required to make this sentence accurate:
"People who [compatibilist freely] do douchey things like hold up traffic (or who do impressive things like win the Australian Open in almost 6 hours of tennis) deserve to be reprimanded (or deserve the title) even if such reprimand (or reward) won't produce any positive benefits to either the offender or society—e.g., better behavior from them and others, etc." And if this sentence is accurate, then I want to hear someone explain *why* the type of desert mentioned in it is different in kind than the type mentioned in the punishment sentence (and presumably why they require different types of free will). Notice that it will be extremely hard to tease apart the "instrumental" features of such reprimand and reward from the "pure" desert features, maybe so hard that any such attempt is a mistake and/or can't be properly understood by people (and that will hold for punishment for more serious cases at least as much). But that's an argument for another day.

3. Remember that I am a revisionist to the extent that I think people deserve a lot less than we tend to think, because we possess (and have the opportunity to exercise) the compatibilist capacities required for desert less than we tend to think. And I may even be revisionist about what the folk take to be the requisite capacities and opportunities--they may think it takes less than I do. I don't think anyone deserves to die for their crimes, and I don't think Djokovic deserves however many millions he won today (nor do the 0.1% deserve the ridiculous share of wealth they have).

4. It's an empirical question, but I suspect that if our goal is to get people to be more compassionate (and more realistic about what people deserve in terms of wealth and power), then we are more likely to succeed by advocating this sort of revisionism than advocating skepticism about free will and desert. But that's a bigger post to deliver (and risk abandoning), so I'll leave it at that.

"People who commit crimes deserve to be punished even if punishing them won't produce any positive benefits to either the offender or society—e.g., rehabilitation, deterring other would-be offenders, etc."

I think some people may reject the premise of the question outright, but being forced to answer, turn back to the idea that everyone must be punished since that matches their ontological position to begin with. I am thinking of someone who may maintain that given the nature of human beings and what is necessary for a functioning society, it will ALWAYS be socially beneficial to punish a criminal; there will always be positive benefits, at least for society, in punishing perpetrators of crimes. Finding the hypothetical to be impossible, they answer the question affirmative not from retributivist grounds, but actually from (confused) consequential grounds.


First, that you agree with the explicit retributive statement doesn’t mean you’re a bastard, it just means you’re a retributivist in the traditional sense. It also means that you don’t find pure retributivism as puzzling as you have suggested you found it in the past. Usually, when I follow Moore in claiming that to be a retributivist is to believe in the intrinsic value of giving offender’s their due, you have sworn off this kind of deep desert and denied retributivists are committed to it. You have opted instead to define retributivism and desert in what I take to be less robust (and less puzzling) forward-looking terms. Your former strategy isn’t surprising for a compatibilist since you don’t need an especially robust notion of free will to ground that type of retribution and desert (although I would prefer to call it something other that retribution and desert, as you know). Indeed, I don’t think you need free will at all. But that’s a story for another day.

The important point is that to the extent that you believe that it is intrinsically valuable to make offender’s suffer, you really need them to “own” their actions in a deep enough way—i.e., if you believe in deep desert, you are going to need an account of “deep” free will to go along with it (or so it seems to me).

Second, as for the strangeness of things one might say, I think it is strange to say the following: It is intrinsically valuable for a norm violator to suffer even if his action was the only possible outcome open to him given the physical state of the universe and the physical laws at the moment he decided to violate the norm. I don’t find it strange to think this same person ought nevertheless be reprimanded or punished on instrumental grounds. Indeed, I think compatibilists often misleadingly try to increase the plausibility of their view by conflating these two faces of responsibility. Here is a good example from Dennett (which I discuss at some length in my paper on disillusionism in the volume on Libet’s legacy):

"Is she [i.e., the free will skeptic] going to jettison our system of law and punishment? Is she going to abandon the social leverage by which we encourage people to take responsibility for their actions? Is she prepared to dismiss the distinction between honesty and cheating as just another myth fostered by the traditional concept of free will?" (2008: 255)

There is a sense in which I think your post is guilty of the same slight of hand (intentional or not). In short, you and other compatibilists try to highlight the implausibility of skepticism by making it markedly less plausible than it is (or at least than it can be). The skeptic about free will and deep desert need not be a skeptic about reprimand, punishment, or even the reactive attitudes more generally.

If it could be shown that punishment is more useful than other options for certain norm violators when it comes to deterring and preventing crime, then I think it could be justified even if determinism is true and even if all we have are the cognitive capacities that compatibilists use to ground free will. What I don’t think is justified in a deterministic world—even one filled with agents who have compatibilist powers—is punishing people solely because they deserve it since I don’t think in a deterministic universe the agency of norm violators runs deep enough to make their suffering intrinsically valuable. For someone’s suffering to be intrinsically valuable, I would want it to be the case not only that they could have done otherwise in the unconditional sense, but I also want them to have a kind of control over their beliefs, desires, and actions that isn’t possible for the kinds of creatures we happen to be—creatures with bounded rationality and bounded volition who are largely the product of epigenetic forces we do not control. But fully fleshing that out would, as you point out, require a lengthy post of its own.

The comments to this entry are closed.

Books about Agency

3QD Prize 2014: Marcus Arvan