Blog Coordinator

« Compatibilism, the Individuation of Responsibility, and The Trap | Main | CFP: Society for the Theory of Ethics and Politics »



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


Ýou say that "commonsense fairness" - fairness as it's ordinarily conceived - requires desert, which in turn requires some sort of non-consequentialist basis. Most compatitiblists think there is such a basis (I've never understood what that is, since compatibilist criteria for responsible agency, e.g., reasons responsiveness, are all forward looking), but some like Dennett in his recent review of Waller suggest that there's a defensible notion of purely consequentialist fair punishment, very much as Parfit would have it: fair warning. The threat of punishment works to sustain norms (a consequentialist rationale), and it's necessary to punish since a warning is only credible when followed through on. All responsible agents are equally subject to this policy, so in principle it's evenly applied, that is, fairly applied. I suspect most folks would call this fair punishment, but if it isn't, what should we call it?

Dennett also thinks that purely consequentialist punishment constitutes "just deserts" but that seems to me a step too far, departing from the commonsense notion of desert. Pace Dennett, we can be fairly punished without desert, but promoting the hard determinist view that we don't deserve it highlights what seems to me the unfairness of non-consequentialist punishment.

Tom. the analogy I use with regard to a system that attaches penalties to actions is the construction industry. Construction is a dangerous occupation: hundreds of people are seriously injured on building sites every year, and not a few are killed. Construction is far from optimally regulated (workers are under all sorts of pressures to take risks) but even if it were, the industry would still be dangerous. The trade offs we have to make ensure that we can't have the good of construction coupled with a genuinely risk-free environment; not without devoting an incredible amount of resources to construction. So, in the ideal case, we have a system justified by its benefits that we know will produce casualties. But saying that the system is justified is not saying that for any worker who is seriously injured, that worker *deserves* the injury. Of course there are disanologies; but I can't see how it is disanologous with regard to the question at issue (which is just to say that I agree with you and Saul that the system doesn't generate desert).

Saul, apart from the fact that I'm a compatibilist - that is to say, my free will scepticism is independent of the question of determinism - I agreed with everything you said against Parfit. That surprised me: I expected you to say that the notion of desert falls away for the sceptic. But you don't: you say that desert constitutes a baseline for the sceptic, from which deviations are not justified. Isn't that inconsistent with your Law & Philosophy paper; don't you suggest in that paper that we can't have the notion of desert if we are sceptical?

Tom - thanks, this is interesting and there is a lot to respond to here. First, we need to distinguish between Parfit's view and what each of us thinks about fairness and punishment. Parfit is clearly, I would say, the greatest living ethicist, and arguably one of the greatest moral philosophers ever. And he has been deeply concerned over the free will problem for decades (just anecdotally, he was one of the examiners of my Oxford D.Phil. back in 1990). So what he thinks matters, and if, as I believe I show, his view is inconsistent, then this matters. I don't think that it is plausible to attribute to Parfit Dennett-like, merely consequentialist views about fairness.

But our main question goes beyond Parfit, and is whether hard determinist punishment can be fair, i.e. whether hard determinists can fairly punish. That depends upon one's idea of fairness, and if one has very weak requirements from fairness, then almost anything can be fair. You seem to think that as long as there is a consequentialist justification for some arrangement, and perhaps also as long as no one is discriminated against for, say, racial reasons, then matters are fair. But that is, surely, a very watered down sense of fairness, as it hardly allows any criticism of consequentialism in terms of fairness. Now, a utilitarian-like consequentialism has many advantages, but sensitivity to fair distribution, for example, or to equality, or to the idea of using some for the sake of others, are not among them. This is all too familiar to repeat; think, e.g., about Rawls' argument in A Theory of Justice that utilitarianism does not take seriously the distinctness of persons. So if hard determinism is fair merely in the sense that every person has to take his or her chances in the overall consequentialist calculus, then that seems to be very little fairness indeed.

Moreover, such a minimal sense of fairness cannot make sense of concern over the free will problem, except to dismiss it as irrelevant. As I have argued in my "funishment" piece which we discussed before on Flickers, I believe that a morally robust hard determinism cannot justify punishment, but only funishment, i.e. incarceration of wrongdoers who are then fully compensated for the deprivation of their liberties and other forms of harm. That turns out to be a practical reductio, but that's a different matter. Hard determinists should be morally concerned whenever people are punished, because (if or insofar as we follow HD) there is no free-will justification for treating people harshly and/or worse than others; whatever one does one is, in the deep sense (the HD claims), morally innocent, and this matters. Becoming unconcerned about FAIRNESS because such punishment can be justified by its overall good consequences is, I feel, very unconvincing.

Incidentally, if I put on my compatibilist hat for a moment, I don't see why you view reasons-responsiveness (or most other compatibilist criteria grounding moral responsibility and desert) as forward looking. The person is culpable, or blameworthy, or deserving of certain reactive attitudes or of punishment, because he WAS reasons-responsive (etc). As you know, I believe that there is a further broader sense of desert, and this is not captured by the compatibilist senses, so that, for instance in my Trap example, the victim deserves punishment according to compatibilism, but it is still in one important sense unfair and unjust TO HIM that he is being punished. But setting aside my compatibility-dualist criticism of compatibilism, I don't see why compatibilism is hard to understand (accepting is another matter). You merely put the bar for moral responsibility very high, beyond the reach of compatiblist freedom, and that's too high even by my HD-sympathetic standards, but I think that you do understand the compatiblist sense. At any rate, I don't find a view which puts the bar for moral responsibility and desert so high, and for fairness and justice so low, morally compelling.

Neil - I am glad that you agree with me. I would have liked to find some way to save Parfit, without interpreting him in a way that makes him merely opt for a non-free will-related notion of fairness, but I don't see that he has any way out. If you are a hard determinist (or incompatibilist lack-of LFWilist), then morally everyone is innocent, whatever he or she has done. If you then take the free will problem seriously, then this innocence matters. But then punishment is unfair.

That the idea of fairness is not limited to the question whether there is consequentialist and/or contractual justification can be seen by a simple thought experiment: ask whether we can criticize social arrangements which are justified in these ways, as unfair. And of course we can. People may even sign actual contracts, which serve their interests, and still we might speak about the contracts as not being fair; say, as being exploitative.

As far as I can see, this paper on Parfit and my earlier one on "funishment" you responded to are of a piece. For hard determinists there can be no free will-based divergence from the moral baseline, and we must take this seriously. AS A RESULT, for HDs no one can be fairly punished, only funished. Funishment is the only moral way in which wrongdoers may be incarcerated, because no one deserves to be worse off than another (for HD's). Hence if incarcerated away from society, criminals need to be fully compensated, by the fun part in funishment.


Thanks, some further thoughts in partial reply:

Hard determinism points up the cosmic unfairness of some people ending up worse off than others – no one, at birth, *deserves* to end up destitute, in prison, insane, victimized, etc. We all want to mitigate this unfairness, hence the importance of progressive social policy to prevent radical inequality.

But compatibilists think that yes, people *do* deserve to end up in prison if they have the requisite capacities for responsible agency, e.g., being sane, reasons-responsive, acting voluntarily, etc. This sense of desert has, compatibilists claim, *nothing to do* with the consequentialist, forward looking considerations that obviously engage these capacities, e.g., being warnable via the threat of punishment. I just want to know what this further thing is that grounds compatibilist desert, that justifies non-consequentialist punishment. We can’t just say that it’s morally important or emotionally significant to us that we be deserving in this way, since that begs the question of whether we actually are.

I don’t agree that the fairness of equal treatment is shallow, since it respects all persons as equally deserving of having their basic rights protected (e.g., not being punished when innocent), where this sense of deservingness derives from our (equal) status as creatures with needs and desires seeking fulfillment. Nothing is more basic. Seems to me that consequentialists can and should take the protection of equal rights (and if we’re liberal, equal opportunities) as a fundamental desideratum in their calculus, a value that functions to limit maximization of utility via victimization (e.g., punishing the innocent). This puts an important sort of fairness within consequentialism, a fairness that has nothing to do, it seems to me, with compatibilist desert, whatever that is.


There you go again (as Ronald Reagan famously said in a debate--was it against Jimmy Carter?) Well, I mean, in the past you have also asked the question what justifies the "desert" in a compatibilist theory that does indeed presuppose deeert. And it is certainly a legitimate question--I was just kind of joking with the reference to the Great Communicator.

Ok, and now tell me what's wrong with this (people on this blog are not shy in this specific respect): An agent deserves to be punished when he does something wrong--violates the law--iff he exhibits guidance control when so acting (assuming that the epistemic criterion for responsibility is met). That is, you ask for "something additional", besides the standard compatibilist ingredients--but why should there be something additional. For me, the freedom-relevant component of moral responsibility just is guidance contro; it is (partly, given the epistemic criterion) in virtue of exhibiting such control that one deserves to be treated in certain ways. Control helps to ground the desert-ascriptions. If you ask me to provide further ingredients, I reply: you're asking for too much. If you ask me to prove or at least render more plausible what I say, I don't know if I can do it, except to say that the account helps to systematize my considered intuitions and bring them into a wide reflective equilibrium. I don't know what else I can say, or what else would need to be said.

Of course, compatibilism is also consistent with non-desert based approaches--in terms of "fittingness", or other notions.

Does this help, or merely reinscribe the problem (or something maybe in-between)?

'I don’t agree that the fairness of equal treatment is shallow ....' Tom Clark

Such a standard, Tom, would not only be shallow, but noxious. Those advocating it would have completely failed to come to grips with the implications of HD. No matter what steps have been taken to promote fairness, they are going to prove NECESSARILY inadequate in the case of criminals, if HD is true: they could not have failed to violate the law even with those measures in place. It was simply impossible for them to have avoided their transgressions. The stern warnings they received not only proved ineffective, there wasn’t even the remote possibility of them having the desired effect.

Echoing Saul, how could you in good conscience punish someone believing full well that there was absolutely NOTHING that he could have done to avoid his transgression? If he began to show remorse for his misdeed would you not instead be obligated to interject: “The bite of conscience, like a dog biting a stone, is a stupidity”? The knowledge of punishment was ineffective in his case- nay; it could not have been anything but otiose. Who knows, it may have in some perverse way even contributed to his waywardness! Is not the only appropriate response ‘There but for the grace of God go I’?

Nietzsche was at least honest enough to draw these conclusions, going one step further and quoting with approval Dostoevsky’s observation that many of the criminals he saw in Siberia ”were carved out of just about the best, hardest, and most valuable wood that grows anywhere on Russian soil."


Are you aware that in a causal universe the words “control” and “guidance” can only be appropriately used in the metaphorical sense? If so, do you admit that your defense of desert is hinged nearly exclusively on debunked psychological buzzwords that no longer refer to anything in particular. Can you really be so frustrated when moral responsibility impossibilists ask for “something additional”? That something more needs to be added seems excruciatingly obvious.

Hi Saul,

Ok, here's one attempt to make Parfit's views consistent. It's undeniable that there are plenty of cases where unequal distribution of punishment can be justified independent of desert. For example:

1. My friend and I agree to flip a coin. If it lands on heads, he punches me in the face. Tails, I punch him in the face.

Now of course, that has nothing to do with knowing a rule and breaking a rule so how about this?

2. My wife and I decide we need to cut down our swearing in the house because our daughter is starting to think she lives in a David Mamet play. So we have a swear jar and make a deal: Whoever swears the most during the month has to do the dishes for the following month. Going into the last day, I'm winning handily. But then through no fault of my own, the Patriots play the Giants in the Super Bowl. As you may or may not know, the Giants are the most loathsome team in the history of the NFL and also by far the luckiest. They recover every fumble, get every call, and their worst receivers will catch terrible passes on their helmets. There is no way not to avoid a horrific filthy unending stream of profanity during one of these superbowls. And so I do (again through no fault of my own, I constitutionally can't avoid it). And I lose the swear contest by one measly swear.

Isn't it fair that I should do the dishes for a month even if I couldn't avoid losing the bet? I think so. And why is it fair? Because I knew the ground rules going in.

Why can't this example be generalized to criminal justice?

(I concede that this type of response works best when the penalties are mild, worst when the penalties are severe.)

'Isn't it fair that I should do the dishes for a month even if I couldn't avoid losing the bet? I think so. And why is it fair? Because I knew the ground rules going in.

Why can't this example be generalized to criminal justice?' TS

No it's not fair, Tamler; and no the example should not be generalized to cover the penal code. And for the very fact you mention- you were unable to avoid losing. Your knowledge of the ground rules was necessarily otiose; it did nothing nor could it have done anything- given your "constitutional" inability- to prevent your tirade. It only gave you the illusion of being able to win; it was merely part of what Saul has called your "trap." The so-called fair warning given to all lawbreakers was similarly bound to go unheeded. So, at the end of the day, lacking humane measures to render would be criminals compliant, we must make do with incarceration sans punishment.

The Giants under Tom Coughlin have been one of the best football teams in recent memory, led by a superb O-line, bruising backs, and a dependable, if not spectacular QB. The Patriots rely way too much on passing for my taste.


You say that "control helps to ground the desert-ascriptions" but you can't say why, which is the obvious next question. It's easy to see why having control grounds the judgment that it's *consequentially appropriate* for compatibilist-responsible agents to be subject to punishment: they have the capacity to anticipate sanctions and thus adjust their behavior accordingly. So it's fair to punish competent agents as a (currently) necessary evil in maintaining moral norms, a forward-looking rationale that Dennett and Parfit endorse.

But of course this is not desert, which is what you're defending. That you can't offer (and perhaps don't see the need to offer) a reason why having control justifies being non-consequentially deserving of punishment - being subject to suffering without it needing to produce any benefit - is the scandal of compatibilism.

In the absence of a reason, one is tempted to conclude that it's the emotional satisfactions of seeing the suffering of offenders that's really driving the compatibilist commitment to desert. Tamler has alluded to retributive emotions as an influence in his conversion to meta-skepticism, and in a post here at Flickers (url below) you seem to endorse the taking of such satisfactions:

"By the way, I just saw the American remake of The Girl with the Dragon Tattoo... I really enjoy the moment when the girl's abuser gets what he deserves. I would even say: it is intrinsically good that he suffers in a way commensurate to the way he made the girl suffer. I think it is hard to argue for this kind of claim, but I do think retributivism taps into something important and deep about human beings.”

Important and deep perhaps, but in my view prima facie morally problematic unless you can supply an argument.

Mr. Nozick (or Mr. Sommers),

I am not sure I like the comparison, as Allen insinuates and as you do. The results about what happens to you was some kind of game you played or agreed to play to do some behavioral work as regards swearing in your house. It seems to me it was both lighthearted and maybe somewhat effective; though obviously totally ineffective. In other words, the punishment that is usually meted to real "criminals" changes their lives in significant ways, damaging even the output they get out of that life. Doing the dishes causes you to miss the first quarter of the football game or the first inning of the baseball game, something you are probably fairly ambivalent about, especially when weighed against the fact that it is a necessary domestic duty that has to get done by someone.

On the last post, John Fischer's example provides, perhaps, better analogy. A husband decides to take strong measures to prevent late night drunk driving, he agrees to his wife's demands of radical personal intervention, e.g., alcohol-induced-ignition or therapy that encourages strong measures to prevent weak willed bad decisions. The husband agrees to make these drastic changes in personal behaviors even though they mean great inconveniences to one's self (and to make this contingent, we could say after the next time he comes home late after drinking yet again, which assuredly ensues). He does this because he understands that the two of them are partners working towards a better life and his changes will help his self and his family. But nowhere would a criminal agree or would it be personally useful to agree to, for example, a 20 year prison sentence (there may be some wiggle room here, maybe, only if some rigid psychological unmovable holds).

As a society, I say following some others, we may demand isolation or a limited state of social interaction because we demand safety for all members, but we should make that known that such is not about punishment but only about overall safety, and where we can find some rehabilitative method to be satisfactory we will take it. I feel like that may back into the idea that only supposedly did the poorest members of a society agree to and agreed to whatever "social contract" they supposedly agreed to, when of course they don't and did not. In other words, there is not fairness nor desert nor contractual agreement in most criminal punishment, today and likely in the near future, and we should stop pretending that there is.

Thanks to everyone for the latest series of comments, I think that we are giving each other a hard time here in productive ways, and making progress.

Tom - Three points. First, I think I can nudge you here in a way that you might be willing to be nudged. I think that you are primarily concerned with the morally best policy, and so are attracted by consequentialism, but you also have some strong "justice" intuitions. In my tirades about ultimate injustice (the injustice which often follows when we implement compatibilist, or contractual, or utilitarian considerations and e.g. punish people, while from the ultimate or HD perspective they are trapped and ultimately do not deserve such treatment or such end results) I am interested primarily in being philosophically clear. I also have some thoughts on policy, but they are not my focus here. I think that once we are clear about this distinction, then you OUGHT to grant "my" ultimate injustice. Let's say you can defend utilitarianism (or some version mixed with equality, or the like - more about this in a minute), even then you still should - it seems to me - grant that since no one deserves to be worse off than another, and THIS MATTERS, any such punishment, even if justified by it's consequences or the like is, on one important level, unjust. I hope you don't find this presumptuous, but it seems to me that distinguishing between the morality of policy and the morality of evaluation, and then agreeing with me on the latter, would not harm what you care about, but on the contrary make it clearer.

Secondly, I don't understand how you can say that utilitarian-like consequentialists can use equal rights as "a value that functions to limit maximization of utility via victimization (e.g., punishing the innocent)". Since you give no inherent weight to compatibilist control and responsibility, I don't see how you have any philosophical resources to limit maximization in this way. In fact you have no reason to care about the "punishing of the innocent" – for you, unless I am missing something, everyone is equally innocent, whatever they have done. That's a direct implication of hard determinism. Then it just depends how the utilitarian chips fall; if the best policy involves punishing the compatibilistically-innocent, your should be for it, if not, then not; you don't (have the philosophical resources to) care about "the innocent" as such. The same applies, as far as I can see, to Parfit, so you are in an unhappy moral place, but at least in good company.

Thirdly, although I cannot enter into the issue of policy in detail here, this last point shows why compatibilist desert might be thought to be important. For compatibilists (or compatibility dualists like myself), innocence DOES have inherent moral value. Perhaps John's saying that reasons-responsiveness is enough might not be compelling, but compatibilists have further resources that can help to make sense of why we need to take compatibilist distinctions seriously - for example, a P.F. Strawson Humean reactive attitudes view, or my attempt to ground compatibilism in respect for persons (which involves respecting their agency-based backward looking desert). Here my point is not that you must accept such compatibilist thoughts, but that you should see and acknowledge why they – unlike your version of consequentialism – would be better IN PRACTICE in safeguarding certain moral reactions and practices.

So, in sum, I think that you have no real need to resist my push towards the recognition of ultimate injustice – and indeed should acknowledge and embrace it, on the level of moral evaluation. On the level of moral policy, you should, I think, recognize that your philosophical commitments (HD and utilitarianism) turn out to have much more radical implications than you think. The innocent (in conventional terms) have a better chance under a compatibilist regime than under a consequentialist one. And if you – again like Parfit – do think that innocence matters in practice, then at the very least your consequentialism needs to be indirect: in other words, you may well want, for your consequentialist reasons, that most people continue to believe in free will, moral responsibility, and desert!

Tamler – as you note, the sense of fairness does seem to track the severity of the price. I don't think that it is accidental that we had this result with the dualism, which might suggest that we are on some level thinking here of compatibilism (e.g. in the choice to join the agreement), and hence the agreement seems fair to us, unless the penalties increase, in which case the ultimate or HD injustice becomes more salient. If that's plausible then it just shows the power of a compatibility-dualist position, and not the fairness of bad deals even under HD.

But in fact I think that I can take an easier line of argument. Even if I grant that in some sense your agreement is fair, under some senses of fairness, my position would not be weakened. For, as far as I can see, Parfit (and you as his attorney here) need to defend the stronger claim, that there is NO significant sense in which when a person ends up in prison – that is (importantly) unfair. And that seems like a very high mountain to climb. The Trap helps to show why. If, as it seems to me, even compatibilist control and responsibility does not suffice to overcome the injustice in a lengthy prison sentence, then such severe punishment when coupled with the absence of such goodies, under HD, surely cannot avoid being unfair and unjust in important senses.

'I don't understand how you can say that utilitarian-like consequentialists can use equal rights as "a value that functions to limit maximization of utility via victimization (e.g., punishing the innocent)". Since you give no inherent weight to compatibilist control and responsibility, I don't see how you have any philosophical resources to limit maximization in this way.'

The utilitarian straw person who carries out show trials to discourage crime can only be successful when there is secrecy. If enough individuals are aware that they are at risk of such treatment, they will opt for a different system. I would imagine hard determinist ethicists generally support democracy and transparency. And hard determinists can utilize negative reinforcement and punishment (in the behaviourist use of the term) for beneficial behaviour change without desertification.

Hi Saul,

I don't think compatibility dualism is the answer here because in the example, I lacked freedom of any kind including compatibilist freedom. The punishment is independent of desert regardless of how one interprets the conditions for it. (Also, Lyndon, this is why I don't think John's example would work since it seems like the husband was MR for behavior). What seems to be doing the work for making the punishment fair is the 'fair warning'--or possibly that we both agreed to the deal and to abide by its rules.

It's true that the punishment is ridiculously light in this case, but one can imagine similar sorts of arrangements with much harsher punishments that still wouldn't ring unfair alarm bells. (Maybe 'unwise alarm bells' or 'that was a dumb deal to make' alarm bells, but not 'unfair' alarm bells.)

Brent: I can't speak for John, but I, for one, was not aware that "in a causal universe [what's a causal universe, btw?] the words “control” and “guidance” can only be appropriately used in the metaphorical sense?"

Suppose everything is deterministically caused by preceding events and states of affairs. That sounds like a causal universe. Now suppose that I go to fly a remote control plane with some friends. using the controller, I steer the plane left and right and do loops; it's great fun. Do I exercise real (i.e., non-metaphorical) control over the plane? You betcha. Do I guide the plane's movements? Yup; I do that, too. All in a causally deterministic universe.

Hi Tamler,

I reject moral responsibility across the board, so do not really see the difference in the contractual sense in the two cases. Given that he was caught drinking by his wife or just coming home late, also against marital rules, the concept of whether he is “morally responsible” in some way to society (or God or Family or just flat our morally responsible) seems beside the point, the punishment that he is accepting follows the agreement he made with his wife.

One of the things that is driving my intuitions here is an ambiguity about “punishment.” In these domestic examples, the individual can clearly see the benefit to their self, their family, or their community; and they have agreed that certain behavioral changes are necessary for a better life and society. That sense of communal agreement seems lacking in most criminal punishment cases. Perhaps the odd case is where individuals acknowledge and agree to drug rehabilitation. When punishment can be clearly seen by the individual as improvement in bettering my self, then of course they (may) agree and find fairness in the behavioral control that is given. Contrarily, when punishment is sanctioned by the state for deterrence of others, or because the law says so, or simply because this individual deserves it, then this type of possible contractual agreement becomes obscured, the punishment is no longer tied to a necessary and personal behavioral problem, but is instead tied to a more abstract notion of social stability or social control or societal justice. To turn around and ground the latter punishments in contractual agreement by that individual would be a stretch, though I am warily thinking about public hangings and deaths of olden times where the individual confesses their failure to uphold their covenants to god and state and prince.

Justin, I actually don’t think you can say you have "control over the plane." That supposedly obvious example is indeed using control in the metaphorical sense. Sure, it might make sense practically to say that you are controlling the plane. Ultimately though, a person who is completely controlled himself (by the laws of the universe and previous circumstances) cannot be said to have control over anything else. That would be akin to saying the ocean "controlled" the coast line (lame example, I know). It might be the closest cause that we can understand, but surely you can see that oceans (and people) are controlled by larger more complex groups of causes

Brent: You say "a person who is completely controlled himself (by the laws of the universe and previous circumstances) cannot be said to have control over anything else." What's the argument for that claim?

Suppose we one day discover that determinism is true. Few people, not even recalcitrant incompatibilists like myself, would start thinking that I don't literally control the plane's movement. What we would (and should) say, rather, is that I don't have ultimate, buck-stopping control over the plane's movements. But control of this latter sort isn't the only kind of real (non-metaphorical) control.

The debate between compatibilists like John and hard-determinists (and other incompatibilists)is not over whether determinism precludes any and all forms of control. The dispute, rather, is whether determinism would preclude the sort of control needed to ground moral responsibility (of a desert-entailing sort).


Fair enough, so then the real question is whether we can make sense of a social contract justification for criminal punishment. Certainly, that's been tried--the benefit/burden approach of Herbert Morris, for example, or an original position style approach defended by Igor Primoratz among others. What's interesting is that these justifications are usually part of retributive theories involving autonomy and desert whereas in this case, it would be employed to justify a theory of punishment that REJECTS desert. If successful, though, it seems like Parfit's views would be consistent.

Of course, one might object that you have to free to make the agreement or contract in the first place in order for the punishment to be fair. But that's where a hypothetical form agreement (like what might happen in the original position)could help--possibly.


That's just it! You wouldn't be "literally", partially or in any other significant way "in control" of anything in a deterministic universe. The word simply cannot be used. The concept of control in a deterministic universe is like a square triangle. And the point I was trying to make in regards to John's post, is that it can't be so surprising that incompatibalists want "something more" than what should be in their minds an incoherent concept. I actually think our difference is merely your willingness to use “control “more leniently than I think the usual and most common definition of the word allows.

David - the problem with utilitarianism isn't limited to Sheriffs and lynching mobs, as I showed in one of my first papers ("Utilitarianism and the punishing of the innocent: the general problem" Analysis 50 (1990): 256-261. Here's an offprint:

There is no reason to believe that the optimal utilitarian (or similarly consequentialist) level will be one where respect for innocence is at the height we intuitively feel it should be - unless we are utilitarians. Your confidence seems to me too great - contemporary democracies are, after all, perfectly capable of approving of unjust social arrangements, e.g. the way the US justice system deals with minorities. The problem in the effectiveness of blame and punishment that is not thought to be deserved but is merely seen as forward looking manipulation is also well known, Bernard Williams pointed out once that this is only more likely to create resentment.

Tamler - I understood that your scenario wasn't dualistic, but merely suggested that we might sometimes feel that arrangements were fair because we were thinking about, e.g., the choice to enter a contract in compatibilist terms. This connects to what Robert and others have said.

But my central reply is that I don't need to deny that there is SOME fairness possible even in HD contexts - rather, it is Parfit who needs to deny that ANY (serious) unfairness follows from punishing the undeserved, and that seems impossible. If a giant captures you and I and is about to eat us, but grants us one last wish between us, then yes, maybe flipping a coin as to who will get this last wish is a fair procedure, but still, being eaten at no fault of our own is unfair.

Hi Brent: I'll just make one last reply. Your claim that we can't literally control anything in a deterministic universe lacks motivation, and on the face of it just seems (to me at least) wrong. The sense of the word control would remain the same regardless of whether determinism is true. But our dispute isn't merely semantic. I think we really can control things in a deterministic universe (even if this control isn't of the sort needed to ground free action and moral responsibility). The toy plane example seems to be straightforward illustration of this fact. I don't see why that should be controversial. As an incompatibilist, I needn't say that ALL forms of agential control are absent in a deterministic universe. All I need to say is that determinism precludes the sort of control (e.g., regulative control) required for free action and moral responsibility.

Saul, you write:

"rather, it is Parfit who needs to deny that ANY (serious) unfairness follows from punishing the undeserved, and that seems impossible. If a giant captures you and I and is about to eat us, but grants us one last wish between us, then yes, maybe flipping a coin as to who will get this last wish is a fair procedure, but still, being eaten at no fault of our own is unfair."

But now it seems like you're not being fair to Parfit--or at least, you're asking too much of any system of punishment. After all, it's unfair in some ultimate sense if someone in the criminal justice system (through no fault of their own) dies of cancer at a young age, but that's not a defect of the system itself. For a theory of criminal punishment to be fair, it just has to be the case that the state is treating people in a fair manner. It doesn't have to be the case that the universe is treating people in a fair manner. So then if we can develop a social contract defense of a desert-independent approach to punishment (no small task), it would seem like punishment could be fair as long we adhere to the terms of the contract. Yes, some people would end up on the losing end--but that's the case in any contract in which there's risk to one or more parties. (Think of a game of poker.)

Tamler – thanks for giving me trouble over this and pursuing our discussion. I don't think that it is accidental that your example is of a game like poker, and I think that its weight depends on its being a low stakes game. If this is true it already rules out the contractual approach as a model for the criminal justice system, unless one believes in libertarian or compatibilist free will, for here the game is for life. Again, my position is the moderate one: I do not need to deny that there are any senses of fairness, in which agreement as such makes things fair. It is Parfit who needs to deny that harsh punishment of the undeserving is unfair. (Both Parfit and myself are speaking about social punishment, not about getting cancer.) The radical nature of Parfit's view can be seeing by noticing that he seems to be saying that IT IS NOT UNFAIR IN ANY WAY TO PUNISH PEOPLE FOR COMMITTING CRIMES WHICH IT WAS IMPOSSIBLE FOR THEM TO AVOID COMMITTING. Being a hard determinist, Parfit of course cannot fall back on semi-compatibilism or some other such compatibilist view, which affirms desert even under the impossibility for doing otherwise. No, he will acknowledge that the criminal couldn't have done otherwise, was not morally responsible for whatever he did, is not blameworthy for doing so, and does not deserve the punishment, but somehow there is still NO significant sense in which it is unfair that he is punished! I simply can make no sense of this.

The comments to this entry are closed.

Books about Agency

3QD Prize 2014: Marcus Arvan