Blog Coordinator

« Dennett Reviews Waller | Main | Parfit on Desert, Punishment and Fairness »



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

Hi John (and all),

This is a somewhat off-topic, but here's one quick point. John, you regularly rely on the thought that libertarianism fares no better in certain important ways than compatibilism, on grounds that the libertarian will accept (or have to accept) that it a (fully?) morally responsible agent has been ""shaped" so that it was 99% or say 85% probable that he would behave badly". And you've invoked this idea in other ways too -- e.g. that it is odd to suppose that responsibility should hang on the "thread" of the difference between 100% and 99% probabilities. But two things. Can you give an instance of a (prominent) libertarian saying such a thing? That is, can you provide an example where a libertarian has said, "Well, just so long as the probabilities that we'll behave in the various ways are just 99% probabilities, everything is fine by me, and that's all I want, and all I require." I can't think of any libertarians who straightforwardly say something like this -- but maybe there are some? Or is this just meant to be a straightforward and obvious implication of the prominent libertarian theories of responsibility that are out there? But is it indeed such a obvious implication? I don't see that.

More importantly, though, it has always struck me that this is something the libertarian would be extremely well-advised not to say. First, if the libertarian even allows that there are "objective probabilities" (given by...?) associated with free decisions, why shouldn't she say that these probabilities can't be 99% or some such? Imagine God setting up a scenario so that it is somehow 99% likely (given...? the laws of physics?) that you'll do some bad thing, and then you do it -- and then God blames you. Seems a bit odd -- God didn't really give you a fair chance at avoiding the blame, if there was a 99% chance that you'd do the bad thing. Why wouldn't that be the thing the libertarian should say? (I guess she could also say that, in such a scenario, your responsibility would at least have to be seriously mitigated, but really, I don't know how to make sense of this idea here -- if God sets it up that it's 99% likely, you're only responsibility to "degree 1", but if God sets it up so that it is just 50% likely, you are responsible to "degree 50", or something like this? I don't know if this model really makes sense.)

I think that the libertarian should probably reject (or at least not quickly accept) this whole framework of numerical "objective probabilities" in the first place, for what would these probabilities be based on? Where would they come from? From the indeterministic laws governing micro-particles? But why should the libertarian think that free decisions can be "realized" in such a way? Further, what is the objective probability, determined by the laws of physics, that an agent will make a free agent-causal decision to do one thing rather than another, on agent-causal libertarianism? The question makes questionable sense, for presumably the agent-causal theorist is (or anyway could be) thinking of free decisions as essentially not the sorts of things that are governed by the laws of *physics* at all.

All in all, I worry that you are here attributing to the libertarian a view which questionably makes sense, she would do well to reject, and which no prominent libertarians have accepted! :)

I'm definitely not saying that the libertarian has no problems here, though. Obviously, they'll want *some* sense in which we can talk about how likely it is that someone will freely do this or freely do that, or perhaps just do this or do that, and I don't know how that's going to go! So you could just reply, fairly, by saying, "Well, if the libertarian isn't going to say that, then she has the burden of giving a theory regarding probabilities concerning free decisions". And that's a fair point...

Professor Waller,

Allow me to return the compliment. I really enjoy working with Consider Ethics in my ethics courses. I especially like the FW chapter, where each counterexample segues to an upgraded philosophy. It allows students to see how philosophers work.

Robert Allen


Hmmm. You write, "One, the arch of the scientific study of human behavior has always bent towards moral responsibility impossiblism. There has not been one advancement in this field (or any related field, physics, biology, evolution...) that has made moral responsibility make *more* sense." And also: "On a side note, I’m curious why you would think it possible for science to be *more* supportive of compatibilistic moral responsibility than it already is." I don't quite see how these two go together. Maybe you are saying that science is already NOT AT ALL supportive of compatibilism, and that it is just INCONCEIVABLE that science could support compatibilism. Is that your view? That's pretty strong, and implausible.

And I didn't get the point about "dishonesty". Were you saying that I am being dishonest? Perhaps not--I just didn't get it. Thanks in advance for clarifying.

You seem quite certain of yourself--quite certain that the advance of science is leading away from the notion that we can be morally responsible. I don't see this at all! We'd need to do a lot of looking into the science, and the interpretation of it, before we could get anywhere near the pronouncement you make! I think that a bit of intellectual modesty might serve one better here. I think that the scientific picture, as it relates to moral responsibility, is highly complex and unclear, with elements going in different directions. I doubt if it will be intellectually fruitful to be as dogmatic as you appear to be. (But perhaps I misinterpret you?)


Thanks for your (characteristically) sharp and thoughtful--and challenging--comments.

You say your comments are somewhat "off-topic", but they are not--they are perfectly to the point and fair. You also say that you are making a "quick point"--but it is hardly quick, and you raise a number of difficult and complicated issues, some of which I probably [!!--objective prob. of 99%] can't address here.

I can't come up with a quote by a libertarian of the sort you want. Of course, one woulnd't necessarily expect such a quotation. I guess I was thinking more in terms of implications or possible implications of libertarianism, rather than something a proponent would highlight. I suppose I was thinking that the libertarian finds causal determination problematic, but does not have any problem with causal indeterminism. But you are quite right that no libertarian worth her salt will say that indeterminism is sufficient for control and moral responsibility--something more is required.

But then it gets really messy. Isn't just really arbitrary to say that the probabilities "can't be 99% or some such"? I mean, that DOES seem ad hoc. And what does "some such" mean here? I think that many libertarians are indeed willing to say that we can "rise above" difficult formative circumstances, and that even if it is very difficult to do the right thing, as long as we have a chance--as long as we are not determined--we have the power that grounds moral responsibility. But I can't pin that to someone (right now, without some more work).

Also, I don't see how event-causal libertarians, such as Van Inwagen, Kane, Ekstrom, et. al., can easily given up the notion that there are objective probabilities associated with our behavior. I'm honestly not sure how they could do that, and it would seem to be a big price to pay.

Perhaps the rejection of such probabilities fits better with agent-causal theory--I'm not sure. It just seems to be a potentially obscure way to go, in any case. So in the end I DO want to say exactly what you suggest (graciously) is a "fair point": "Well, if the libertarian isn't going to say that [i.e., isn't going to associate objective probabilities with our behavior], then she has the burden of giving a theory regarding probabilities concerning free decisions".

There are other issues you raise, but perhaps I'll stop here. Well--I lied: I'll again thank you for pushing me in legitimate ways, and for suggesting some possible moves by libertarians--moves that would need to be addressed in a more thorough treatment of these issues. As you know, I think that control and responsibility are fully compatible with indeterminism; so unlike Saul and Bruce, I think there could indeed be libertarian freedom. I am no moral responsibility skeptic--indeed, and since I do not think it should hang on a thread, I feel it is important that we could indeed have the requisite control in an indeterministic world (of a certain sort).


Exactly what are you referring to whey you claim that the "arch" [sic] of science has been toward moral responsibility impossibilism? I could imagine various possibilities, but I'd like to know what you are talking about.

I think science itself doesn't straightforwardly imply anything about moral responsibility. Of course, we need to interpret the results in light of philosophical theories of responsibility, and see how the science fits.

Funny: a leading philosopher of free will, Robert Kane, has spent a good bit of his career arguing that his views about free will--which are definitely not skeptical--fit nicely with neuroscience. Tim O'Connor makes similar claims about his (somewhat different, but still libertarian) views. Of course, perhaps these philosophers are just wrong--and wrong in glaring and obvious ways. But I doubt it.

So: what science shows that moral responsibility is impossible?

What has always bothered me about libertarianism is that it is most often cast against a background of ceteris paribus assumption of the possession of a universal norm of human reason that somehow gounds control (why else have insanity defenses?). But what constitutes human reason in this role has several dimensions of what constitutes normativity, including the cultural variability of what supposedly constitutes reason itself and then varying norms within given cultures of degrees of reasonableness from what is judged to be instances of competent reason to assessments of incompetence.

In other words, libertarians seem to think that an assumed incompatibilism of freedom and determinism can underwrite all subtleties of the relation of reason and responsibility. I can't see that at all, and would say that such presumed incompatibilism entails absolutely nothing about the relation of reason and responsibility. This is where (semi-)compatibilism comes in as recognizing something other than a certain metaphysical relationship as grounding claims of responsibility.

Bruce writes:

"And if we keep looking, then we find that Jack was shaped as a cognitive miser, and that Jack is currently in a state of ego depletion. And when we find such causes -- and it seems to me very valuable to find them -- then it casts doubt on the fairness of holding them morally responsible..."

John agrees that there's a complete causal story behind bad behavior that we could perhaps discover were we to keep looking, but suggests we shouldn't. The refusal to keep looking singles out the agent as the primary causal player in the story, thus keeps reactive attitudes maximally engaged, driving the perceived fairness of holding someone MR, that is, of non-consequentialist punishment.

The language of MR aside, what matters in the debate between Bruce and John (and between compatibilists and MR skeptics overall) is how wrong-doers should be treated. And, given human psychology, that treatment can depend on the perceived salience of causes outside the agent, a matter of how persistently we ask "why?" The criminal justice system specializes in cutting short the investigation of causes. This is perhaps a practical necessity but it also conveniently caters to the psychology of desert.


You write, "John agrees that there's a complete causal story behind bad behavior that we could perhaps discover were we to keep looking, but suggests we shouldn't." You then discuss a "refusal to keep looking..."

Well, I did write that we need to figure out--that, indeed, a kind of wisdom and maturity requires that we figure out--when to push the "why" question, and when to stop. Certainly this is not controversial, is it?

But also I have emphasized--or I thought I was emphasizing--that compatibilists are just as interested as anyone in seeking to identify, figure out, interpret, analyzie, and so forth the causal history leading to behavior. Indeed, I pointed out that compatibilists are in a good position to do this, at least in contrast to some (such as certain libertarians, in particular, agent-causal libertarians). Certainly compatibilists are at least arguably in a good position to do this, in contrast to non-causalists.

Although I am not a Strawsonian in certain respects, I do think that some of you moral responsibility skeptics at least sometimes are wiling to put aside the tremendous importance, not just of blame and punishment, but of interpersonal relationships in our lives--love, friendship, and so forth. You focus on what you take to be "deep injustices" and "conveniently catering to the psychology of desert". Well, as we know from Tamler, sometimes even this orientation can well change when you come to really love someone--such as a child. You can indeed come to feel that a brutal torturer and rapist of your child might well--err, dare I say it?--deserve to be punished. But also a sole focus on blame/punishment can obscure the crucial role that moral responsibility plays (through at least some of the reactive attitudes and other related mechanisms) in stitching us together as human beings--as friends, lovers, colleagues, and even "co-deliberators" (in Scanlon's--another notorious an no doubt morally dubious compatiblist's) term.

Now of course I'm not saying that our actual penal system is humane or justified as it is--it is certainly cruel and stupid and problematic in many ways. But that's not the issue. And I would further claim that we compatibilists (well, remember, I'm a Supercompatibilistic Semicompatibilist) are in as good a position as anyone to look into problems and abuses and to insist on changes. Bruce, Saul, you, and I--and I am pretty sure Patrick--would all unite in this project.

Saul says we should "assume that the moral order could not be much improved (i.e. Neil's #1)." I think we can gain insight by comparing our attitudes to the remaining disasters of such a justice system, to the disasters faced by a good doctor.

A good doctor loses patients sometimes. Occasionally, the deaths result from the doctor's actions, where a different treatment would have succeeded. Even an excellent doctor can't stay entirely up to the minute on all the research - she would spend all her time reading, and no time operating (and she'd still fail to keep up). Even an excellent doctor can't personally double-check every judgment call by her radiographers, anesthesiologists, etc. Even an excellent doctor can't afford to run every conceivable diagnostic on every case. And so on.

After the patient dies, the causes will be examined. The doctor will certainly recognize that a tragedy has occurred, and she will feel - in some sense - responsible. Hopefully not at the level that leads to burnout, but she won't just nonchalantly say, "oh well, stuff happens." On the other hand, she won't conclude that she acted unjustly, either. Not even a little bit.

Compatibilists can reasonably take a similar attitude to the failures of a well designed justice system, I say.

Also, I note that it takes a certain sort of metaethics to even begin to make sense of the claim that a social system can be constructed in the best and fairest way possible, yet still be unjust. As someone with sympathies for the kinds of constructivism advocated by thinkers like Scanlon, Velleman, and Habermas, I doubt the existence of the very ground on which Saul wants to stand that claim of injustice.

I'll jump in with some specific replies, and a bit later post again with some general points, just to keep this post under manageable proportions. A good discussion, everyone.

Neil – I am glad we agree about injustice. On victimization, I guess it depends on how one uses the term, so our disagreement here probably doesn't matter much.

Mark – one can distinguish The Trap as an argument/story from my own views, i.e. one could use the argument and not be a compatibility-dualist. In any case, the point is to try to show, particularly to people who see compatibilism as SUFFICIENT, that even under ideal compatibilist conditions matters will be morally dubious. We can argue whether in my waiter case giving him less of a tip is unjust at all, or unjust but doesn't matter, morally, or matters a bit but can be set aside for other considerations without real concern. But when we look at a case of people ending up spending 20 years in prison, their lives practically ruined, then the concern is greater. So, your (3) is important for this story, simply because it makes it effective.

Ben – hell has been historically important in western thought, but I don't think we should attach it to any position on the FW debate, for all of them may (and morally should) reject the permissibility of the idea of eternity in hell. I was merely using the term to describe conditions in actual prisons. Compatibilists like John, let alone a compatibility-dualist (the true semi-compatibilism!) like myself, will be happy with talk about much more pedestrian forms of desert.

As to the reactive attitudes – I don't think that a plausible place for making the cut would be between our reactive attitudes and desert, or between the reactive attitudes and practices such as punishment (if this was what you were proposing). IF you grant resentment, moral indignation, and blame, then spilling over into practice – e.g. punishment, just not up to the level of hell – would be fairly obvious. This does not mean of course that it would be automatic, i.e. that you could punish anyone you resent, further ideas would be needed, but a view which acknowledged that one could perfectly well deserve blame, but never punishment, does not seem compelling.

Paul – excellent points. First, on the doctor analogy. Yes, in some sense or up to some level that is what we should feel. That is an advantage of the compatibility-dualist view, that it can acknowledge a great deal of validity in compatibilism – and even grant that it ought to dominate our practice – without then having to say that this is the end of the matter. If we establish a compatibilistically-just Community of Responsibility, then this is not a trivial moral achievement. But the question still remains whether that sums up all that we would then want to say about the free will problem. Perhaps it would be helpful to see it in the following way. There are two levels of innocence here: on one, if one committed an offence and no compatibilist excuses pertain, then one is not innocent. The ultimate or HD perspective takes the SAME event and says that there is a further problem, for on another level, on a broader view, whether one did what one did compatibilistically-freely is not all that matters. On that broader view, for the familiar incompatibilist considerations, such as that ultimately the criminal could not avoid doing what he did, that he was merely carrying our what his bad luck scripted him to do, and the like – the criminal himself is seen as a victim or the circumstances that molded him. Ought we then to dismiss the first, compatibilist, sense of innocence? No, it is a condition for respect for persons and a civilized society. Compatibilist agency and moral responsibility based on it matter, and ought to be respected. But we also should not avoid acknowledging that the very distinctions we ought to make, on the compatibilist level, are also groundless on a broader view, and result in victimization of the innocent, in the second broad sense. If you cannot ultimately control your motivation set, and as a result of those motivations you just happen to have, you end up doing things for which you are then made to suffer tremendously, then to say that all that matters is the first issue, whether you did the deed (and were not compatibilistically incompetent), is, well, not to say all that there is there to be said.

On the meta-ethics. Well, that's a big topic. I think that it matters a lot what sphere we are talking about. As I argued myself ("Control, Desert, and the Difference Between Distributive and Retributive Justice", Philosophical Studies 131 (2006): 511-524), there are good reasons not to take control-based desert (or its absence) as seriously in distributive justice, as we should in retributive justice. But the idea that there are no valid pre-institutional desert intuitions AT ALL strikes me as very extreme and unconvincing. If we can know anything, morally, it should be obvious that punishing someone who is completely innocent, in every sense, IS unjust. So punishing contemporary Jews for what some Jews did or did not do some 2000 years ago to Jesus is unjust because no one can control events that happened before he was born, and so cannot deserve to be blamed or punished for them. The immorality, to me, does not depend on any construction, it is there. It seems to me analytic that the innocent-in-every-way simply cannot DESERVE to be punished. Tamler's societies that do otherwise and call it justice are morally mistaken (and barbaric). Perhaps there are in certain extreme circumstances other, e.g. consequentialist, reasons, which override desert, and hence we ought to punish (or "punish") the innocent, but doing so would still be unjust. However, once we acknowledge that this makes sense meta-ethically, then the door is opened for taking the broader, ultimate sense of injustice also seriously.

I have two questions.

1. You say "The right kinds of compatibilists ask us to look with just the right level of magnification at just the right range of data: not too much, and not too little. Asking for too much is metaphysical megalomania, and for too little leads to admittedly a shallow and unsatisfactory view." The latter part I can understand but why is too much megalomania? Isn't it better to know more. Why is more knowledge above some point negative?

When discussing Zed you say that is it important to know that there are ".. no coercion, force, manipulation, psychological impairments or distortions to the ordinary human mechanisms or practical reasoning, and so forth." But why isn't it important to know everything about his heritage and environment? If he was sexually abused when young, if he had disturbances in his personality that are not classified as psychological impairments but still caused him to behave as he did etc. I don't see that you answer this question in the discussion with Bruce.

2. A related question: In the post Still A Compatibility-Monist, After All These Years? you said: "My point: I lack control over various factors which are parts of the story of who I am and how I act, and which are such that, if they didn't obtain, I would be very different, or wouldn't exist at all, or wouldn't act as I do. But that seems obviously compatible with my playing the cards that are in fact dealt me in a way that is enough to make me morally responsible." I have read your article The cards that are dealt you. It seems to me that you divide the past events in two parts, those you couldn't influence, how the card are dealt, and those you could influence, how the cards are played. But if the world is deterministic, when you were born you have no influence on who you are and at no time afterwards there is a point when you have a real choice. If there are one hundred deterministic worlds that were identical when Zed was born, they all will be identical when Zed commits his crime. If you or I were one of these Zeds we would also become criminals.

I agree with you that libertarians doesn't fare better but the problem is how does compatibilism fare. The skeptical position implies other problems but I think dualism is part of the solution.


Thanks a lot for these questions; I'm assuming they are addressed to me.

Of course, we just CAN'T know everything about the causal histories that lead to our behavior--there would just be too much information. Much of it is clearly irrelevant. But, yes, it would be important to know whether someone has "disturbances in his personality" that lead to his behavior. And it would be important to know similar features--and in some cases it would be helpful to know the history of these "disturbances". As you perhaps know, I have an explicitly historical account of moral responsibility. And I take it that one attractive feature of my approach is that it points us to which facts about the agent and his/her history we need to know about, and which we don't. So: does the "disturbance" lead to a lack of the required reasons-responsivenss? If so, then there is no moral responsibility; if not, there is moral responsibility. Just saying that someone has a "disturbance"--well, that is not enough. We need to know more, and I have at least offered, in admittedly an incomplete way, the fundamental elements of an approach that could guide us to what we need to focus on (and what we can ignore).

I agree with you that various positions, including my own favored positions--Supercompatbilistic Semicompatibilism--have significant problems. I certainly don't think--and I don't believe that its proponents would contend--that (say) libertarianism is without significant challenges. And certainly responsibility-skepticism has its drawbacks too. For me it is a matter of a holistic cost-beneft analysis, and I choose one way, and others will do the cost-benefit analysis differently and choose differently.

I welcome your openness. It fits nicely with the intellectual humility and modesty of Saul Smilansky and Bruce Waller, who typically seek to highlight the complexity of the issues and the fact that all the major approaches to these issues are latching onto some important truths (or aspects of the truth).


I take back my point on meta-ethics. There's no reason in principle, I suppose, why we couldn't rationally justify a *multi-layered* set of norms. Punishing recidivist burglars with 20 years in jail could, conceivably, be just on one layer (and perhaps overall) yet unjust on another. So it's not about meta-ethics. I just disagree on the object level.

Just as I don't think the excellent doctor acts unjustly when she kills a patient with a treatment that works best overall for those kinds of cases, I also don't think the legislator acts unjustly when he condemns recidivist burglars to jail. Not even a little bit, not at the "ultimate" level nor any other. Similarly for the judge, or the police. None of which is to say that the legislator should declare victory when the burglar goes to jail.

"(H)ell has been historically important in western thought, but I don't think we should attach it to any position on the FW debate, for all of them may (and morally should) reject the permissibility of the idea of eternity in hell."


For someone unwilling to get into theology because it's "too deep," you seem quite sure of yourself here. But, look, it’s really only a question of heuristics. You are using “pedestrian forms of desert” to make your point against Compatibilism. I was only trying to be helpful, should they prove insufficient.

Plenty of philosophers who deny the existence of Heaven and Hell have nevertheless used such scenarios to explain what they mean by LFW: here’s the sort of freedom one would need to possess before God could assign him his place in Eternity based on his deeds/misdeeds. Compatibilistic notions would be grossly insufficient, they maintain, in that context. LFW, in your estimation, is still not enough; but the question is, what would be NECESSARY before ANY punitive measures would even make sense.

A Hard Determinist might deem “pedestrian” punishment unjustified sans LFW. Or he might avail himself of Heaven and Hell scenarios to further dramatize his point, even if on humanitarian grounds he objects to the notion of damnation. Better still, he could avoid altogether talk of punishment and ask instead, what sort of freedom would be necessary to make sense of rewards, mundane or eternal?

Gunnar – I agree with most of what you say and am heartened by your view of the possible merits of a dualism on the compatibility question.

John – I can understand why you oppose a monistic hard determinism, which doesn't recognize at all any of the truths that you find with compatibilism. But why is the alternative you opt for a monistic compatibilism? If as you say there are strong arguments on both sides, then why dismiss one set altogether, rater than try to incorporate them? You have dominant compatibilist intuitions, and that is fine, moreover no one is insisting on a 50-50 division (whatever that would mean exactly, in practice). But why is the alternative a 100% dismissal of any of the importance of the ultimate point of view or of the insights of HD, rather than a partial acknowledgment of the weight of the opposite side, in a way that lets it into your own? In other words, if there is as you say a lot to be said on both sides, why go, methodologically, for a WINNER TAKES ALL position? Do you accept in some way the Assumption of Monism (the idea that we must be completely compatibilists or HDs, if there is no LFW)?

Paul – it might be that defending your position will depend on going meta-ethical. Let me try and explain why I disagree with what you say. First, I think that the doctor analogy may mislead us. As I understand your story, the doctor does his best for all his patients, and there is no one he gives up on for the sake of others (or himself). In The Trap matters are darker – we as a society know that Zed will fall to his doom, but we keep on playing our part of what for him is a trap, for our own reasons. We are not quite the equivalents of the universally-benevolent doctor who is doing his best for everyone. There are various ways in which we can form social relations, after all, even various acceptable ways, and those where Zed ends up an imprisoned criminal are not necessary, they are the results of our choices, based upon our interests. We keep making these choices, even after we know that Zed's interests are in no way served by them. The Trap is, in part, intended to capture that aspect, how we flourish through practices that we choose and that harm people, who cannot avoid their location in a social nexus which leads to their doom. The idea is that we then recognize that, although in real life we cannot see individual trajectories with such predictability, this is what we are in fact doing, even if with less awareness. From the ultimate perspective, we see that Zed ends up playing a certain role because of the way he has been biologically and then socially constructed; this role then leads him to his doom – a doom where our role is crucial not only in the final stages, but all along.

But more importantly, even the idea that we are justified in setting up the social situation in just the way we did, does not eliminate the injustice. I have, after all, built that into The Trap, in order to give compatibilists everything they want. Even if there is a sense in which we have done no wrong, this does not eliminate the fact that those ending up so badly off do not, ultimately, deserve to be so badly off (or indeed any worse off than another). When utilitarians support scapegoating an innocent person for the greater good, then even if you accept that, all considered, this is the right thing to do, then you would still consider it unjust, because the victim does not deserve to be punished. When I tell the story how even the compatibilistically-guilty are, on another, broader level, innocent, and indeed victims of the forces that have molded them, without their control, then you can no more dismiss THAT injustice because it is justified, then you could the injustice in the utilitarian's actions, on the assumptions that they are justified.

Robert – I don't think that I understand. I can see why your heaven and hell analogy can help the incompatibilist case, and did not object to it, except at the extreme, where I think it is open to the compatibilist to say that no one deserves eternal suffering. I do not believe that there is actually a hell where people are punished in the afterlife by a benevolent God, but if there is, I can see why that might be good. I am, after all, in part a compatibilist, and a moral order where those who escape getting what they deserve in this life can still do so in the next world can, in principle, be better, particularly if it would make this world better (that might well require more knowledge here of what goes on there than we currently have).


Nice questions. Well, I think there are important insights in incompatibilism. One such insight is crystallized by the Consequence Argument. As you know, I do explicitly incorporate this into my overall approach, in the sense that I am only a Semicompatibilist, not a full compatibilist! That is, I think that causal determinism might well rule out freedom to do otherwise; I just don't think that such freedom is required for moral responsibility. So in this clear sense I, unlike many other compatibilists--indeed, all classical compatibilists--take seriously and incorporate into my theory insights of another major alternative view.

Further, in my 1982 Journal of Philosophy article, "Responsibility and Control", I identified and took very seriously the view that has come to be known as "Source Incompatibilism". That is, I argued in that early paper--well before the view was presented and defended by others--that even if the Frankfurt cases show that we don't need alternative possibilities for moral responsibility, it does not follow that causal determinism is compatible with moral responsibility. Causal determination, as I wrote, "flows through the actual sequence", and it is at least plausible that this would threaten moral responsibility.

Over the years I have sought to defend the notion that, even so, causal determinism need not rule out moral responsibility, and I have presented and discussed major threats to this view--defenses of PAP, defenses of Source Incompatibilism, and defenses of Galen Strawson's "Basic Argument".

No doubt I have not entirely captured the worries of critics of my views, and no doubt I haven't assuaged these worries, when I have succeeded in crystallizing them!

I think I have taken seriously at least some worries by moral responsibility skeptics. And I certainly agree that there are real worries about our moral responsibility, even in spite of everything I and other non-skeptics have managed to argue. No question: I have no knockdown refutation of moral responsibility skepticism, and if I were not an insomniac for other reasons, this might be a reason that I would lose some sleep. Just for example, it is challenging to reply to Derk Pereboom's Four-Stage Argument. (And that's just for starters...)

But of course I distinguish between taking an insight on board and taking an opposing insight seriously. I have taken some insights of the incompatibilists on board, but perhaps I have only taken seriously (and even then, perhaps not seriously enough!) the insights of moral responsibility skeptics. (By the way, Bruce's very interesting-looking book is on my reading list, toward the top, and I look forward to leaving from it.) But so far I have not felt the need to incorporate the insights into my view; as you very graciously put it in a previous discussion, I prefer a tight (although not tight-assed, simple, and elegant view, insofar as that is possible.

And, although I like to think of myself as at least open-minded and willing to think about and seek both to understand and to address challenges to my view, I'm tougher than a two-dollar steak. I'm not going to complicate or clutter up a beautiful theory, just to please anyone else.

Sorry about the obvious typo in my previous comment: "(By the way, Bruce's very interesting-looking book is on my reading list, toward the top, and I look forward to leaving from it.)" "Leaving" should have been "learning".

Really, really, that was not a Freudian slip...

John - thanks, I agree that your view is a compromise with traditional compatibilism insofar as your reject the claims for a compatibilist "ability to do otherwise" and in other ways. And of course it has been developed out of a lifetime of careful and rigorous thought about the problem. It would probably be too disconcerting for everyone if you changed your mind now anyway, and certainly there would be an aesthetic loss. I guess we each should aim to crystallize the view he or she finds most plausible. A central aim of mine, no doubt because of my own psychological makeup, is to open up possibilities. I even think that compatibilists shouldn't give up a compatibilist interpretation of "ability to do otherwise", across the board; in certain cases that interpretation is plausible. At any rate, I am glad that you see no inherent objection to a complex, dualistic approach, i.e. you do not hold the Assumption of Monism.

Just as I try to nudge incompatibilists into lowering the bar for some of the good things (freedom, moral responsibility, desert, etc), so that they recognize and acknowledge them to some extent even in a deterministic world (or a world without LFW), so I try an nudge compatibilists not to settle for a low bar, and then, once they manage to jump over it, to feel that there is no more to be said. So I agree with the compatibilist that there is an important sense of moral responsibility, and we should by and large establish a Community of Responsibility based upon compatibilist distinctions, and yet see a scenario such as The Trap as showing the limitations of compatibilism, in that a person like Zed (an ideal free compatibilist agent, functioning in ideal compatibilist settings) is still trapped into a life where he is doomed to end up very badly, just because of forces outside of his control, which have molded him in a certain way. Sometimes this embeddedness does not matter (such as my waiter case), and we can be compatibilists without much worry, but, for example, when the price is so high, then we should be concerned that the price of setting up a Community of Responsibility is so harsh for some people, in a way that they cannot be said to fully deserve. As in my reply to Paul above, even if, overall, we hold compatibilist justice to be the best we can do, this does not mean that it captures everything that needs to be said about justice. Injustice can be morally unavoidable but still unjust. A robust, transcendent libertarian free will seems impossible, but we can see why it is an attractive idea, for then the world is not, for people like Zed, a trap.


You haven't identified a disanalogy with my doctor scenario. In the real world, we don't know that Zed will burglarize, any more than the doctor knows that the test battery which performs well overall will fail for (and thus lead to treatments which kill) Mr. Smith.

But you construct a scenario where we know much much more. You say "We, as a society, let us further assume, probably have good reasons not to change our procedures in a major way. Our practices work pretty well, STATISTICALLY." It is a struggle to think how we could know that Zed will burglarize, and that there are ways to prevent it, yet not have compelling reason to change our rules. Perhaps any rule system which saved Zed, without condemning others to a life of crime, would simply be too complex for people to grasp? Leading to other miscarriages of justice, when judges got confused? Or, too expensive to allow us to fund other necessities?

If we can find such a reason, it is extremely plausible that similar reasons would apply to the rules in medicine, such that we *could* change the rules to save Mr. Smith's life, but this would have unacceptable costs. After all, medical procedures are complex already, and there is a limit to the amount of time and work we can demand of doctors and patients. Perhaps there is no set of rules that saves Mr. Smith's life without costing as many or more lives overall. This is at least as plausible as your hypothesis about the rules for punishment working well statistically. But then, the analogy holds.

Paul – thanks for continuing the debate. I was worrying that the medical analogy may mislead us because I was taking it to paint an overly-benevolent picture, whereby the doctor is concerned for all his patients, as the justice system (which e.g. is happy to use the punishment of some to deter others) is not. But I agree that we can make a limited analogy in, say, the use of resources: both systems will spend only so much, even though we know (either statistically, or in my story individually) that some will get a bad deal. However, no one would say, in normal cases, that the patients who end up badly deserved it (as the compatibilists want to say about typical criminals being punished), and that their lot is just. The major question is whether compatibilism is sufficient, i.e. whether there is a sense in which (given that there is no robust libertarian free will) the punished criminal is a victim, is being punished unfairly or unjustly. Your point, I take it, is that we are doing our best and hence there is no problem. But the same sort of reply is available to the utilitarian or interest-based-contractualist, yet we routinely think that the results of utility-optimal or contractual results can be unfair. So, as far as I can see, the medical analogy does not help you there at all. Unless you assume that compatibilism is sufficient (but assuming this would be begging the question), you then need to acknowledge that there is room for the ultimate level or HD concern over what the compatibilist would consider as completely kosher.

My deep psychoanalysis of your post reveals that it was indeed a Freudian slip, but not quite of the sort you might imagine; what you deeply want is “to be leaving from it with my immortal soul intact”; which is a very reasonable concern. After all, you have been carefully and seriously reading moral responsibility skeptics, and with a clear and open mind; but those are skeptics who at least have the basic minimum decency to express some degree of regret about the demise of moral responsibility. In contrast, I want to strangle moral responsibility in its bed, drive a stake in its heart, bury it at a crossroads, and dance on its grave; taking such a view seriously, particularly given your honest willingness to genuinely consider opposing views, might well result in the loss of your immortal soul. And as the director of the Immortality Project, that obviously places you at special risk. Nonetheless, you are clearly obligated – and indeed fated – to go forward with this profoundly risky venture. For you now have the opportunity to perform an experiment of the greatest philosophical significance: Take a complete inventory of Dr. Fischer at this moment, call it pre-AMR; then read Against Moral Responsibility, and take another complete inventory of Dr. Fischer, at post-AMR; whatever difference there is between the two Fischer-states will be the immortal soul that you lost in the reading! Though this may seem a spiritually risky experiment, it is clearly your fate: After all, who is the other great thinker who risked his immortal soul in search of knowledge? Dr. Faustus. Your first name and that of Dr. Faustus are the same (Dr.)! Not only that, but your family names both start with F. And both names contain exactly 7 letters! And there are precisely 7 – not 6, not 8 – deadly sins. So obviously you have no regulative control about whether you will perform this experiment; however, you do have guidance control, so I am confident you will pursue it with your usual charm, vigor, and courage.
Saul, my apologies for wandering off into such nonsense. This is what happens when you spend a week driving, eating too much pumpkin pie, and drinking altogether excessive quantities of Dogfish Pale Ale. I like your new problem, and if I manage to collect my wits all in one place, I hope to make a minimally intelligent comment on it.


Ha! Very good. You are so nice it is hard to disagree with you! And that Dogfish Pale Ale sounds great, although I admit to just having had a pint of great German beer.

btw, are you sure it wasn't "Dogfisch" Pale Ale?


First, I'm glad the cease-fire is holding. With the new developments in Egypt, life in the whole Middle East must be somewhat unnerving. I hope you and yours are well and continue to be.

Second, reading recent comments about the Trap made me think of a very basic question: aren't traps for real creatures? But why in the face of skepticism about absolute MR shouldn't such a skeptic think that the Trap is one similar that I might set for Santa Claus? I bait it with milk and cookies as I'm supposed to by the stories and set it up by the fireplace (a humane trap that ensnares, not kills--I reject capital punishment for Santa). Why should I expect to catch anything except I have posited the existence of Santa? But the Trap you describe seems parallel to me--Zed will only be Trapped if one posits the existence of something like ultimate responsibility. A Santa skeptic sees no real point in the Santa trap. Why would an ultimate MR skeptic be any different with respect to yours?

Bruce - thanks, and I am also curious what you will make of my argument in a previous post, "Is it about time?", that hard determinists cannot really deny their freedom and moral responsibility at the time of choosing and acting. The free will part isn't a problem for your unique view, but responsibility is. But I don't want this to lead you to drink…

Alan - thanks for your concern. We had some bad missile days in Haifa in 2006, but the more recent ones did not reach us here in the north of Israel. In any case I try to take such things philosophically. At least this is so insofar as it applies to conventional weapons - WMD is cheating, and the prospect of Iranian nuclear weapons is very disturbing.

Ultimate injustice isn't Santa Claus, rather denying it is the Flat Earth Society... The Trap tries to capture something, and if one doesn't see it, then I want to hear more. There is after all nothing here which isn't familiar from commonsense concerns, or indeed even compatibilist ones - i.e. control, desert, justice. We simply go further in the same direction.

Hi Saul,
Your questions raise such fascinating issues that I want to take a month to think about them before responding. John manages to write these brilliant arguments with remarkable swiftness -- would that I had such powers. Or perhaps John's German beer does not have the same effects as Dogfish Pale Ale; or more likely, John has now found the middle way of golden mean virtue, and has learned to drink exactly the right amount of German beer to perfectly stimulate the philosophical fibers; sadly, we extreme moral responsibility skeptics tend to be extreme about everything. I had really intended to comment on your time case, but before I could do so, I read the post by Eddy Nahmias, and did not think I could really add anything to his excellent insights and arguments. So in response to your question about my views, I agree with Eddy, and he outlines the argument better than I could. But to add just a couple of points. It does seem to me that I can believe in determinism, believe in my own compatibilist freedom, and believe that I do not have moral responsibility for my deliberative decisions even as I am deliberating (but maybe that takes practice -- or perhaps it's practice in self-deception -- having been a moral responsibility skeptic for many years). I don't feel compelled to make a specific choice, but I do firmly believe there are causal factors (many or perhaps most of which I am not consciously aware of) that influence my choice (the choice is still mine). There is no doubt a difference between being in the middle of the deliberation experience and reviewing the experience at a later date; but I think there is more than just a passage of time involved. For example, it is likely that the later reflective examination of my deliberative process is being done at (what Kahneman calls) a slower system two process of more careful deliberation, while the initial decision may have been swifter; and even if I have not been consciously reflecting on my earlier decision during the interim, my deliberative processes may have continued nonconsciously (just as I continue nonconsciously thinking about a philosophical discussion, and an hour later the perfect counterexample "pops" into my conscious thought). Finally, I thought Robert Allen's response to this question was particularly interesting: that if I could be wrong about the nature of my own freedom of choice, that would open the door to extreme skepticism. But my impression is that there are many instances when we are clearly wrong about the nature of our choices (Daniel Wegner's research provides many examples). To take one that I find fascinating: think of the "facilitators" who aid autistic children in "expressing themselves" by means of typing elaborate messages on a sort of "plachette"; these people sincerely believe that they are only steadying the child's hand, and that the entire message is being composed by the child; but research shows again and again that the facilitators are in fact composing the messages themselves. That seems to me a case in which people are quite mistaken about their own free choices, but it does not invite general skepticism about science. Anyway, thanks for these fascinating and thought-provoking cases, and to John and Robert and Eddy and all the others who have provided such insightful posts. Was recently reading Tamler Sommers' superb RELATIVE JUSTICE, and he acknowledges how much the Flickers discussions helped his work (as well as the special help provided by John, incidentally); and I certainly agree, it is a wonderful resource.


Thanks for your reply. I agree that the burglar's going to jail is a moral problem. Similarly, Mr. Smith's death at the hands of the medical system is a moral problem. But we have to distinguish between a result being morally troubling, versus its being an injustice done to that person. Mr. Smith's death is not an injustice done to him, despite the fact that he was causally *doomed* to die in this way. (Determinism, I take it, is at least as plausible for medical conditions as it is in the case of crimes.) Even when the hospital administrator *knew* that the system would kill Smith, and chose not to change the (tied for optimal) system - which is the parallel to your case - that still isn't an injustice done to Smith.

But if knowable causal doom imposed on someone by a system we could change doesn't demonstrate injustice, then it doesn't demonstrate that a compatibilist legal system constitutes an injustice, either. That takes the sting out of your argument, as far as I can see. It doesn't tend to establish compatibilism as correct, of course. It just leaves the score at zero-zero.

The comments to this entry are closed.

Books about Agency

3QD Prize 2014: Marcus Arvan