Punishing with a Compatibilist Heart

Inspired by the need to start thinking and learning more about punishment for a conference organized by Gregg Caruso, by Gregg’s interesting paper posted earlier this month, and by this post by Jerry Coyne on the Norwegian prison system, I thought I’d offer a few musings here to try to get some discussion going during the last week of our free month at Flickers. 

I want to suggest that compatibilist approaches to free will are best situated to deal with concerns about criminal punishment.  Here are some reasons why, which I’ll just present as a list of numbered claims, barely argued for, so people can refer to them in comments (and demand more argument!):

1) This claim is true: Retributive punishment is not justified if we lack ‘desert-justifying’ free will (the sort we lack according to Pereboom, Caruso, and others whom I’ll label ‘FW skeptics’).  But that claim does not entail this one:  If we have such free will, retributive punishment is justified.  Hence, compatibilists can argue that we have free will but that retributive punishment (or more precisely, some forms of it) is unjustified. 

FW skeptics do not typically assert that compatibilists (or libertarians) must support retributivism, but they often suggest that FW skepticism gains a significant advantage because, unlike pro-FW views, it rules out retributivism, or they suggest that one of the main reasons people defend free will is to preserve retributivism.  But both of these moves are problematic.

2) Of course, there is a view of free will tied to a history of dualist, libertarian, religious thought that was likely developed in part to serve as a foundation for retribution of the most severe sort—hell for eternity, or nasty purgatory for a long while.  This is the ‘heaven-and-hell-justifying’ free will that Galen Strawson targets as impossible and that may have been invented to protect God from the problem of evil and the problem of damning finite creatures to infinite punishment.  And no doubt a good many people, especially religious people in American and elsewhere, have this view of free will and retribution (though it’s worth noting that some of these religious people are also the most forgiving, as witnessed by the remarkable reactions of many of the family members of victims of Dylann Roof’s racist murders).

Here, compatibilists can (and should) take sides with the FW skeptics against both the existence of this absurd type of free will and any types of retribution that depend on it.  To the extent that one thinks that, by definition, free will means impossible self-creation or dualist interventionism, perhaps because one thinks that’s the view of most people, then compatibilists should brandish the skeptic banner about ‘free will’. But as I often argue, this is not the only (or primary) understanding of free will, so telling people that we lack free will risks throwing too much baby out with the bathwater.

3) I’m not sure what retributive punishment must involve.  It is often associated with causing physical or psychological suffering to criminals (or executing them).  But can’t we argue that these forms of retribution, or at least the most severe of them, are unjustified (for various ethical reasons not tied to questions about free will), while also arguing that criminals can (if they meet requisite compatibilist conditions and depending on the severity of their crime) deserve:

-to be blamed for their crime.

-to be forced to perform restorative measures, ideally including expressions of guilt and repayment to victims and/or society in general.

-to be quarantined, where this includes not just the forward-looking goal of incapacitation but also an aspect of shunning (for some appropriate time) because the criminal violated society’s (‘contractual’) rules and hence does not deserve (or have the right) to participate in that society (for some time).

-to be the subject of our moral anger (and other relevant reactive attitudes), perhaps including expressions of such anger (etc.) by victims or their loved ones.

-to be forced to accept our attempts to rehabilitate them (with ethical constraints on what the system can do to them without their consent, especially in the age of neuro-interventions).

Of course, many of these features overlap with (and may be difficult to distinguish from) the forward-looking aims of the FW skeptics.  But I think it is significantly different when we understand them in terms of the criminals’ deserving each of these features of punishment, and it has some potential benefits, such as:

3) We can lessen the risk of demeaning or devaluing the rights and feelings of victims.  As John Fischer, others, and I have expressed here before, there are significant problems with telling the victims (and victims’ loved ones) who have suffered brutal rapes or killings, such as the recent ones by Roof, that the conscious, planned, controlled actions of the perpetrator were, in some sense, no different from a shark attack or a tornado.  Or telling them that the perpetrator, though he deeply wanted to carry out the crime, thought about it carefully, and rejoiced in its ‘success’, is not really morally responsible, did not have a real choice (see Coyne) or genuine free will, and hence does not really deserve to be blamed or punished. 

Yes, the qualifiers are supposed to carry with them the full weight of the FW skeptics’ careful, valid arguments (unsound though I think they are), along with their distinctions about what kinds of free will and desert they are targeting.  But I think victims (and most others) would prefer the compatibilists’ similarly careful argumentation and distinctions that conclude that the perpetrator is not an uncaused cause or self-created agent, yet he did make choices, he has some measure of free will, and he does deserve our indignation, and the punishments outlined above, even though, for various ethical and practical purposes, he does not deserve to be executed or tortured or driven insane with solitary confinement—indeed, as Norway shows, he may be rehabilitated with more humane punishment, and then more likely to be able to restore his wrongs and rejoin society. 

4) It may be appropriate and beautiful when the victims or their loved ones forgive the criminals who harmed them, but it seems that they must arrive at such forgiveness rather than have others (philosophers, scientists, society) tell them that there is nothing to forgive, not really, since the criminals did not act with a sort of free will that makes them deserving of blame, and hence of forgiveness.  If everyone is already forgiven, then no one is really forgiven.

5) We can avoid the risk of dehumanizing criminals. Gregg’s public-health model (building on Derk’s) addresses these concerns with great care.  Yet my own intuitions (and the way some scientific skeptics talk) suggest that thinking of criminals as lacking free will leads to thinking of them as broken machines needing to be fixed in whatever ways we can.  The ‘objective stance’ may be appropriate for the insane but not for all criminals.  In any case, I’ve never understood why we should predict that a punishment system that drops the idea of desert and retribution will end up more humane than one that maintains those concepts—the former is consistent with the death penalty (as a cheaper incapacitation method for those who cannot be rehabilitated) or at least life-long sentences, with severe forced methods of rehabilitation and restorative payment, etc.  If one rejects these inhumane punishments for ethical reasons, one can use the same reasons to reject them on a compatibilist foundation.

6) If we don’t toss it out, then we can use the concept of free will (and related concepts regarding choice, self-control, or moral reasoning that people might have more difficulty retaining if they accept FW skepticism) in order to make more fine-grained distinctions regarding moral and legal responsibility.  We can recognize that free will comes in degrees, to the extent that individuals possess the requisite capacities for rational self-control and to the extent they have had reasonable opportunities to exercise them.  Since many criminals likely have diminished capacities or opportunities, we can, were we to develop a more humane criminal justice system, more carefully map criminals’ degrees of responsibility and appropriate punishments.  (I take it as a bonus of this view that it allows us to recognize that many white-collar criminals come out as more free and responsible than many of the people we lock up for longer.)

We might also use the concept of free will to better ‘build better beings’ (a la Vargas).

 7) Perhaps the difference between much of what I’m advocating here and what the FW skeptics advocate is razor thin (while the difference between what we jointly advocate and the most retributive bloated free willers is a wide gulf).  But even so, I don’t see the advantages of advocating skepticism over compatibilism.  I don’t think the Norwegians (or other systems with less retributive penal systems) are actually FW skeptics (are they?), though they likely believe in more limited free will than most people in the US, especially more conservative and more religious citizens.

In the end, perhaps deciding which is the correct view of free will and moral responsibility depends on which view works best to create the societal outcomes both (many) FW skeptics and (many) compatibilists seek?  We both want to move away from the worst retributive features of the American system and towards the more effective systems of places like Norway, while not going too far towards a system that dehumanizes victims or criminals (nor suggesting fatalism or lack of capacities for rational self-control we do have).  Perhaps empirical work can illuminate which view is more likely to get us from where we are to a better system.  Or perhaps both approaches can get us there in different ways, much as conservatives and liberals offer some different (and some similar) reasons and methods for reforming the disastrous US penal system.

Comments

[Neil]

Here's one difference between your proposals and those of at least on FW skeptic (I don't presume to speak for others) It is an empirical question whether any sort of harsh treatment is justified in any particular case. Everyone should think that the US prison system is completely dysfunctional; that's not something that should be at issue in the debate at all. But that leaves room for lots of disagreement. I think it is likely that there will actually be cases in which there is no justification for any sort of harsh treatment. In fact, I suspect that the proper response in some cases will be, in effect, to reward the criminal: to offer her help and opportunities. That will depend, in part, on further empirical issues: can the broader community accept that the person is rewarded for wrongdoing (obviously, their response matters, for the stability of the system among other reasons). But if the empirical issues go my way, then I suspect we will see a difference.

[V. Alan White]

"In the end, perhaps deciding which is the correct view of free will and moral responsibility depends on which view *works best to create the societal outcomes both (many) FW skeptics and (many) compatibilists seek*?" (My emphasis.)

Exactly Eddy; I agree. Which is why I think Neil's remarks in general are actually not so far from what you say here. You add in fact: "Perhaps empirical work can illuminate which view is more likely to get us from where we are to a better system." I should think Neil would agree with that too.

Now turning to Neil: then your critique is that empirical studies might show that particular instances, independently of a general system in which they are embedded, might require no punishment qua harsh consequences in order to obtain the best outcomes. You do allow that such outcomes must be within a larger social context accepting them as such (or redefining punishment as non-harsh corrective rewards, or neutral ones, like the so-called oil in the gummed works or the corrective pill). I think the current collective social conscience (not just the US) could not accept such drastic value changes to think of punishment as what would now be viewed as non-punitive, though maybe that could be manipulated over time incrementally with non-harsh corrective measures for lesser crimes instituted first, and then for more serious offenses later (is this Vargas building better beings?). However, such a program would require a value-laden justification for its implementation in the first place. Whether one is a skeptic or compatibilist, that larger issue must be addressed above all. That is one reason I am a pragmatist--we must settle why we are or should be engaging in any social practice in the first place, and that requires a frank discussion about what are workable values.

The practical question of values wags the tail of FW and MR, and not the other way around.

[Eddy Nahmias]

Hi Neil and Alan, thanks for getting the conversation started. I hope others join in.

Neil, I'm not sure we're disagreeing, though I couldn't tell if you were leaving the door open for harsh treatment in some cases (e.g., Dylann Roof). I certainly think my view leaves the door open for the positive treatment you suggest, especially for criminals who (a) had diminished capacities or opportunities to make better choices when they committed their crime and (b) who could be given training, education, etc. that would improve their capacities and opportunities to make better choices. I suspect many people we currently lock up for drug crimes and petty theft (driven by poverty) fit these criteria. I think recent "Smart on Crime" initiatives are moving in this direction, but I also think society will accept them more readily if we say (as I think we are justified in saying) that these criminals *deserve* to be incarcerated and forced (=given the opportunity!) to learn how to make better choices (and then to be monitored when freed).

This response seems to line up with Alan's comment. However, I want to emphasize that, despite my closing comments in the post, I don't think that the correct answers to debates about free will, desert, responsibility, or retribution (and their compatibility with determinism, etc., their justifications, and the meaning of the concepts) depends solely (or even mainly) on the practical effects of various answers. I'm a wide reflective equilibrium guy, who thinks those effects are one consideration, among many, in the process of WRE.

[Tom Clark]

Thanks Eddie, lots to chew on here.

Very glad to hear your recommendation that compatibilists join FW skeptics in taking up the banner against naturalistically impossible varieties of agency that help justify bad, excessive retribution. I think making a big deal about the fact we couldn't have done otherwise in actual situations is a good way to undercut vengeful attitudes that drive capital punishment, solitary confinement, rape in prison, etc. Bringing up determinism, explicitly, widens the consideration of causal factors outside the person, thus mitigating reactive attitudes. But of course compatibilists are not known for taking up this banner. Hopefully they will do more.

Re the milder retributivism you want to hold onto: the role of compatibilist desert is, I take it, to justify inflicting punishment on offenders even if it isn't necessary for achieving any consequentialist good. Offenders, you say, categorically *deserve* blame, shunning, quarantine, being targets of moral anger, etc. even if none of these are necessary for reform, deterrence, public safety, etc. I'm wondering what it is about the capacities that usually figure in ascriptions of compatibilist free will (e.g., reasons-responsiveness, rational self-control) that make it the case that offenders deserve punishment (not treatment). This difference between compatibilism and the denial of desert is not razor thin.

Lastly, I'm curious about what you see as some of the ethical reasons to reject excessive retribution. You say they "are not tied to questions about free will."

Thanks!

[Neil]

Eddy, the cases you mention are the easy ones: diminished capacities especially. I want to suggest that rewarding crime might be the right response in cases of agents with no impairment and regardless of severity of crime. It's a partially empirical question: does any degree of harshness accomplish some good? You mention Dylann Roof: I am committed to the view that in principle his might be a case in which reward is appropriate (I suspect it isn't, because of the role so responding would play in perpetuating inegalitarian institutions, but it's not completely obvious that that's the case).

[Eddy Nahmias]

Hi Tom, I'm glad you are joining in! You ask good (tough) questions. I'll start with the last one. I take it that there are good reasons (ethical and practical) not to torture criminals, cut off their hands, castrate them, execute them, or use neuro-interventions to alter their brain, and that these reasons would often, perhaps always, trump the forward-looking reasons for these 'procedures' EVEN IF it turned out they were very effective ones to, respectively, get crucial information, deter stealing, prevent rape, incapacitate a psychopathic killer, or reform them. Similarly, EVEN IF some kinds of harmful treatments were deserved by a criminal, there are good (ethical and practical) reasons not to use them. For instance, as we know all too well from the US system, EVEN IF we thought some harsh punishments were deserved, they often conflict with the goals of rehabilitation and lowering recidivism. While I would not want to commit myself to the view that anyone (even Dylann Roof) deserves execution, EVEN IF I thought some did, I would still advocate banning capital punishment--banning it is overdetermined by various ethical and practical reasons. I think that's true for most of the harsh treatments that some people defend under the banner of retributivism.

I didn't say offenders *categorically* deserve blame, etc., since they might not if they lack sufficient free will. But one point I was trying to make is that the supposedly backwards-looking features of retribution and desert are not so easily separated from the supposedly forwards-looking features of punishment. Criminals can *deserve* to be quarantined and to carry out rehabilitative tasks, because they freely chose to commit crimes that they knew would (if caught) result in such punishments and because that is what is required for them to be a part of our society. If it turns out (empirically) that some of these incapacitation and rehabilitation measures would involve harsh treatment, the criminals may deserve that (contra the FW skeptic who might face the paradox of funishment).

The cases Neil raises are tricky, if I'm understanding him correctly, because it might turn out that the best ways of rehabilitating some criminals is with pleasant procedures (that look like funishment). I have to think more about what I would say to the prospect that the best way to reform Roof might be with treatment he would enjoy, but the goal of general deterrence seems to conflict with this possibility. Here again, the idea of deserving punishment seems tied to the forward-looking aims, in that we set up rules whose violation rational agents know will be punished such that if they break those rules when they had the opportunity not to, then that's why they deserve the punishment.

And I'm a compatibilist, so I don't think determinism entails that criminals (with the requisite capacities and opportunities) "couldn't have done otherwise in actual situations."

[Bruce Waller]

Great post, Eddy; I was starting to feel Flickers deprived, and your post was precisely the fix I needed. You raise a great question, and bring up very important points. I agree with Neil that “the US prison system is completely dysfunctional,” and you obviously concur (as you said in your response To Katrina Sifferd’s excellent final post on precisely that problem). I had breakfast yesterday with one of our local prosecutors, who was an undergraduate philosophy major and does a superb job of adjunct teaching some applied ethics courses for us; her job involves prosecuting some people who do very bad things, and thus she is involved in sending people to prison – and she was almost in despair about the cruel and counterproductive conditions in our prisons. As both you and Neil have noted, this is probably something almost all of us agree on; and we also agree that Norway’s prison system is much much better. But a disputed question may be WHY there is such a striking difference between the systems. You focus the question perfectly: “I’ve never understood why we should predict that a punishment system that drops the idea of desert and retribution will end up more humane that one that maintains those concepts”.

You state that you “don’t think the Norwegians are actually FW skeptics,” and you are correct, they are not – at least according to sociologists/criminologists who have studied the culture. However, they are much closer to FW skepticism (I think it’s actually MR skepticism, but that’s another issue) than are cultures like the U.S. and Great Britain. As Michael Cavadino and James Dignan state: “It is also possible that in a social democratic culture people are not so ruthlessly held responsible for the offenses they have committed, which are less likely to be attributed to the free will of the individual offender. Without necessarily going so far as to say that ‘society is to blame’ for all crime, there could nevertheless be a greater willingness to assume a degree of collective responsibility for the fact that an offence has been committed.” In contrast, the U.S. emphasis on free will and individual responsibility contributes to a harsher attitude toward those who commit criminal acts: “There is much about the USA in the last three decades which may have made it fertile soil for law and order ideology. To begin with, the ideology’s conjunction of beliefs in free market capitalism and in the free will of the individual expresses a very American rugged individualism, which may help to explain the particularly strong burgeoning of punitive politics and its popular appeal in the USA. The neo-liberalism of American society is predicated upon individual responsibility in a way highly consonant with the law and order attitude.”

Of course not every firm believer in individual moral responsibility is a “law and order” advocate who wants to hold offenders “ruthlessly responsible”; obviously you are not, nor is John Martin Fischer, nor (on the libertarian side) is Bob Kane, and we could list many more. There is a possible world in which strong believers in individual moral responsibility fiercely reject the law and order ideology (for a variety of reasons, including inefficiency). But the point is that holding people morally responsible does CONTRIBUTE to a stronger belief that offenders justly deserve harsh punishment. We could discuss the various philosophical reasons why MR belief contributes to such attitudes (I think the reasons are legion, and Tom Clark suggests one of the most important, namely that belief in moral responsibility tends to block deeper inquiry into causes of behavior, and that deeper inquiry mitigates the retributive desires – or so psychologists tell us); but the relevant point here is that there is empirical grounding (in the work of researchers such as Cavadino and Dignan and many other criminologists) for the claim that belief in individual moral responsibility contributes to support for harsher punitive conditions. While we can’t do a laboratory test on this hypothesis, observation of changes in Great Britain – since it adopted (following Thatcher) more of the individual moral responsibility perspective and has become more committed to it over recent years. We find harsher punishments for juveniles (with John Major advocating that in dealing with juvenile crime we should “condemn more and understand less”); higher imprisonment rates in harsher conditions; a rejection of what was once proudly regarded as the “Golden thread of British Justice”: the right to silence – now refusal to answer questions can be taken as evidence of guilt; and recently, a push to deprive prisoners of the right to vote (a move that has provoked the wrath of the European Union).

One last thing (I promise): certainly there is a difference between deaths caused by a tornado and deaths caused by Dylann Roof. We should tell those who lost loved ones in this racist killing that what was done was a horrific purposeful act; and it is important to acknowledge that Dylann Roof is a very very bad person; and it is essential that we take very seriously the terrible wrong that was done, and make genuine efforts to prevent such wrongs in the future. The execution of Dylann does not begin to examine the real causes of this tragedy; to the contrary, if Dylan is executed (or imprisoned for life) there is a tendency to suppose that the problem is solved, and that we need not look deeper at the causes. The people who lost loved ones in that tragedy obviously realized that the basic source of the problem is not one deeply disturbed kid; rather, the real problem is a racism that has been exploited for political and selfish gain for many decades (Paul Krugman has a great essay on that point in the Monday Times). They realize – profoundly realize, through long sad experience – that (as one example) one way that union organization (that would benefit both black and white working people) was defeated in the South (I observed it directly in the defeat of pulpwood haulers attempting to unionize in Louisiana, and textile workers struggling to unionize in North Carolina) was by fanning racist fears: whites will be forced to work alongside blacks, and blacks would be treated as equals, and – God forbid – you whites might wind up taking orders from a black foreman. It is precisely that deeper understanding of larger social causes that the people of the church brought to this tragedy, and made them more forgiving of the individual; and the focus on individual moral responsibility draws attention away from those crucial factors.

I recently read a wonderful essay -- “Scientific Challenges to Free Will” – in which you conclude by noting how important free will is to our psychological well-being, and why we should cherish and strengthen the resources of free will (a point which I keep trying to convince Gregg Caruso of, to no avail – I hope he reads that paper); but I cannot understand why someone who gives such a brilliant account of free will wants to hang onto an atavistic notion of moral responsibility – a notion that does not lead you to law and order harshness, but which does seem to have that effect on the folk: commitment to humane treatment of the poor, better treatment of criminals, and stronger support for the strengthening of free will resources seems to vary inversely with commitment to moral responsibility.

[Neil]

Right, Eddy - it's an empirical issue. Actualy I wasn't thinkIng only of rehabilitation. I would bet that the majority of murderers dont need rehabilitation: they are unlikely to reoffend. And I doubt that deterrence is particularly important wrt murder (it seems far more important wrt property crimes). I doubt it is fear of the consequences that holds most of us back from killing one another, and it may be possible to distinguish cases, such that those who are deterrable know that they will be punished if they offend, whereas those who are not deterrable will not.

[Paul Torek]

Great post Eddy! Predictably (and freely :) ) I agree with just about everything you say. But I do take exception to "The ‘objective stance’ may be appropriate for the insane but not for all criminals." No, the I-Thou stance is appropriate for the insane, if you hope for them to recover, which you should:
http://www.ncbi.nlm.nih.gov/pubmed/3591992
http://www.tandfonline.com/doi/abs/10.1080/17522439.2011.595819
https://www.ted.com/speakers/eleanor_longden

Full disclosure: I'm married to a psychologist.

[Tom Clark]

Thanks Eddie. Although you don't actually enumerate the "good" and "various" ethical reasons not to exact harsh punishment, whether for consequentialist or retributive reasons, I'd suggest they are broadly speaking individual level concerns about minimizing suffering and maximizing autonomy. These goals are always being balanced against the risk of future offenses (consequentialist considerations) and the desire for payback (retributivist considerations).

In response to my query about why compatibilist capacities justify desert, you say that it’s very difficult to separate retributive (desert-based) considerations from consequentialist considerations: "the idea of deserving punishment seems tied to the forward-looking aims." Just so, which reflects the fact that reactive attitudes imputing desert are naturally keyed to just those capacities, like being sane and reasons-responsive, that make offenders more likely to be reformable and deterrable by the prospect of punishment. In which case, my (non-rhetorical) question is, why hold onto to the idea of desert as it’s usually defined: the moral necessity to inflict suffering independent of consequences?

I guess I'm confused about the compatibilist notion of could have done otherwise. On the assumption of determinism, things would have had to be different in some respect for the agent to have acted otherwise, such that the capacity to do otherwise is conditional. So, barring indeterminism, the agent couldn't have done otherwise in an actual situation as it played out. If we admit some sort of indeterminism into the picture, the agent could have done otherwise, but not in a way that makes her more responsible than under determinism (imo). So this sort of freedom (from determinism), should it exist, doesn't help to buttress ascriptions of desert as far as I can see.

[Tamler Sommers]

Hey Eddy, very nice post. The debate seems to gravitated towards the familiar issue of whether harsh treatment or punishment can be deserved--but I think maybe the biggest advantage of the compatibilist approach is how well it complements restorative approaches. And restorative justice involves more than just restitution and community service. There's face to face meetings between victims, wrongdoers, and the community affected by the crime. Victims can ask the wrongdoers questions about why they committed the offense. The wrongdoer can apologize. They can work together (with mediation and oversight) to try to make things right as much as possible with restitution and maybe some prison time or community service thrown in. When it works, both sides gain respect for the humanity of the other. Obviously this will be less effective and satisfying when it comes to horrific actions like Dylann Roof's. But fortunately those aren't the most common kinds of offenses.

Now maybe a free will skeptic can accommodate this kind of approach too, but it's a much less natural fit. If we don't think that anyone deserves blame, then it's hard to see why the offender is obligated to enter this process and make things right to the victim. The quarantine model or some sort of social insurance policy (we're insured for floods and fires, we can also be insured for being victims) makes more sense for the skeptic. Or do my former partners in non-blameworthy crime disagree?

[Eddy Nahmias]

I’ll try to keep my head above water responding to comments, but forgive me if I don’t respond to everything (and it will be worse on Fri and Sat as I will have limited internet access, though I hope that others will join in and y’all respond to each other!).

Bruce (thanks for your kind words), you suggest the Norwegians are “much closer to FW [or MR] skepticism” than we are. Well, that’s one way to put it, but given that skeptics say we *entirely* lack the sort of FW required to genuinely deserve *any* MR/desert (or just the latter in your case), I am not sure what it means to say a person or culture can be *close(r)* to it—you either accept it or not. That’s one reason I favor a compatibilist view that explicitly focuses on degrees of freedom (and desert). On such a view, it makes sense to say that some people and cultures, more than others, are open to or already accept the facts regarding: humans limited capacities for free will, some individuals’ more significant limitations of these capacities and opportunities to develop them and exercise them, and the causal explanations (societal, genetic, psychological, etc.) for these facts, and the way all of this informs us about the degrees to which we should hold wrongdoers responsible, blame them, and punish them.

And yes, more Americans (and Brits, etc.) need to see the light, but I’m not sure the blinding light of skepticism is as effective as the dimmer switch of cagey compatibilism.

Finally, I prefer my theories of free will and moral responsibility to go hand in hand, so I just can’t get on board the ‘semi’-train, whether it’s semi-compatibilism about MR but not FW (to do otherwise) or about FW but not MR. But maybe I’ll see the light someday—as you can tell, I’m holding onto some forms of retributivism with slippery fingers (see below), but I’m not letting go of desert: while no one may deserve to die or deserve extreme suffering for their crimes, some people definitely deserve to be blamed, to be forced to justify their choices, to make restitution for them, to choose to improve their decision-making, and to be shunned (quarantined from general society) until they do so.

Neil, I’m sure lots of killers are unlikely to kill again and they clearly were not deterred by punishment, but surely most killers made their choices to kill while knowing and accepting the risk of punishment and lots of others choose not kill in part because of the risk of punishment (how many is indeed hard to know, but we do know that general crime rates increase significantly when the risk of getting caught goes down—e.g., because of lack of police presence or enforcement).

Paul, yes, I too quickly followed Strawson in ‘objectifying’ the insane, and obviously ‘insanity’ is used to describe a wide range of cases. Most people we label as insane are not such that “you cannot reason with him. You can at most pretend to quarrel, or to reason, with him” (Strawson). I suspect the participant stance (I-Thou) will be more effective for even the most difficult cases.

Tom, you say “desert [is] usually defined: the moral necessity to inflict suffering independent of consequences.” That’s the definition I’m rejecting—it’s not the way it’s usually defined nor should it be. Perhaps ‘retribution’ is usually defined that way, though I’m not sure how most people think about it. If we say that retributive punishment involves ‘giving people what they deserve’, then that’s open-ended to include all the aspects of punishment I’m saying many criminals deserve that are not aimed at making them suffer, and it also allows that any suffering criminals deserve is *only* appropriate in the service of these other aspects.

For instance, as Tamler says in selling restorative justice, we can say that criminals deserve to be required to try to restore the damage they have done in the ways Tamler describes, and some of these ways will likely include some suffering. If they refuse to try (and have the requisite capacities to make that choice freely when given the opportunity), then they deserve to be shunned, along with the suffering that causes.

Can someone explain what mistake I’m making if I claim that causing pain and suffering need not be the direct goal of *retributive* punishment?

[Tom Clark]

Eddy, thanks for your replies. You say that, on the usual definition of desert (giving people what they deserve), suffering might be an unavoidable side effect of deserved *punishment* that's inflicted on consequentialist grounds, (e.g., victim and community restoration), but suffering is not deserved in itself. If we polled the populace on this, do you think the majority would agree that no one deserves to suffer for their offenses? Isn't the point of deserved punishment, on the common understanding of desert, the retributive goal of inflicting suffering, deprivation, disadvantage, etc., independent of any consequentialist good?

So yes, I think you're making a mistake to claim that causing pain and suffering need not be the direct goal of retributive punishment (giving people what they deserve), at least as it's ordinarily understood. But maybe I'm wrong about how most folks conceive of desert and retribution. You, at any rate, are very close to eliminating any non-consequentialist element of desert, which is great, since on that view suffering becomes only an unfortunate side-effect of deserved punishment, not a morally acceptable end in itself. But I don't think most folks would go along with this. Not sure about philosophers.

[Eddy Nahmias]

Hi Tom, I have been thinking about how we might test people's intuitions about these issues and maybe I'll figure out something I can run by people on the blog. There's been a lot of work showing that people (at least in America) have retributivist motivations in punishment, but I don't think any of it shows that people think retribution must have, as its goal, suffering. And there's been very little on how people understand the concept of 'desert' or what they mean by deserving blame or punishment. My initial intuition is that a lot will depend on the nature and severity of the crime. Particularly awful crimes, like Roof's, will lead people to want to see him suffer or die--I admit I feel that way (I feel that way about some financial and political criminals too). However, I also think most people will *not* think some (less severe) criminals need to suffer, even though they do deserve to be shunned while they are rehabilitated and deterred, etc., even if they will suffer some as a result. And I think information about the limited opportunities (or capacities) of these criminals, or their past suffering, will influence people's judgments. If you or others have ideas for how to test people's intuitions about these issues, let's do it!

[Tamler Sommers]

Tom, I don't think the restorative approach is "inflicted on consequentialist grounds." There may be some consequentialist benefits, but the goal is not to increase net utility. The goal is justice.

[Thomas Nadelhoffer]

Nice post, Eddy, although you’ll be unsurprised to find that I disagree with lots of what you say. So, here are some responses, requests for clarification, etc. (in no particular order):

First, not all FW skeptics are opposed to punishment. I, for one, am not opposed to punishment. I am not even opposed to hard treatment so long as it maximizes the preventive and rehabilitative aims that I think should be the sole concern of the criminal law. So, it’s not punishment per se we reject, but more specifically retributive punishment. While you do make the latter and not the former claim, it’s not always clear that you keep the distinction in place. Because the most prominent free will skeptics have been anti-punishment, there is temptation to think that being a skeptic goes hand in hand with being an eliminativist about punishment. But that is not (and need not) be so. An eliminativist about retributive punishment need not be an eliminativist about punishment more generally (as you seem to assume in places). I could say more about this if you’re interested…

Second, I was surprised you didn’t discuss the recent paper by Clark et al. entitled, “Free to Punish: A Motivated Account of Free Will Belief.” They provide findings from several studies which suggest that our beliefs and attitudes about punishment drive our free will beliefs, not the other way around (as has traditionally been assumed). At a minimum, this motivated cognition approach merits consideration in the present context. In short, it seems that the punitive/retributive tail is wagging the agential dog here (contrary to what you seem to suggest in places in this post).

Third, according to the way most people talk about retributivism in the punishment literature, there are a few core components to retributivism: (a) moral desert is either necessary or sufficient for punishment (depending on the brand of retributivism under discussion), and (b) the suffering of the deserving is intrinsically (and not merely instrumentally) valuable. It is because of (a) and (b) that free will skeptics eschew retributivism as a workable theory of punishment. But here again, just because a skeptic rejects retributivism, she need not reject punishment more generally. As you note in #2, a skeptic can bring on board nearly all of the items you list as virtues of the compatibilist approach where punishment is concerned.

Fourth, you discuss rational self-control in #6. I have a question. What does your revisionist conception of free will add above and beyond this notion of self-control. It starts looking redundant. Given that the skeptic can allow that we are sometimes rational agents who have control over our actions (for instance, there is an difference between someone who does something voluntarily and someone who does something involuntarily), what mileage is gained by insisting that the agents additionally have free will* (as you would define it). Control comes in degrees as it is, as does rationality. Why can’t these do all the work we need for a non-retributive justification for punishment? It seems to me that a skeptic already has the tools to “build better beings” without the need to retain and revise the notion of free will.

Fifth, you say: “But even so, I don’t see the advantages of advocating skepticism over compatibilism. I don’t think the Norwegians (or other systems with less retributive penal systems) are actually FW skeptics (are they?), though they likely believe in more limited free will than most people in the US, especially more conservative and more religious citizens.” As you know, this is an empirical question. We even developed a scale that could help should shed some light on the answer! See Bruce’s comments for some places to start looking for those answers.

Finally, I want to point out that Michael Corrado has been doing some really interesting and important work on this front which should be getting more attention by philosophers than it does. You can check out his chapter in my volume on punishment. He’s got a newer piece as well. He tries to show that he can eschew the concept of free will while avoiding some of the horrible consequences that are often said to follow.

Tamler,
You suggest that a strong suit of the compatibilist approach is that it can best accommodate a restorative approach to criminal conduct. I would like to hear more about this. As you suggest, the skeptic can seemingly bring all of this on board in the name of the instrumental value of this approach when it comes to addressing wrong doing. While the skeptic surely could focus on quarantine, I don’t see why this need be the case. What did you have in mind such that skepticism and restorative justice aren’t a good fit?

[Gregg Caruso]

Eddy, great post! And thanks for referencing my public health-quarantine model--much appreciated! I'm hoping my model gets some traction in the literature since, as you note, it tries to address many of your concerns (and the concerns of others). I'm heading to Greece tomorrow (I know, great timing!) and I just wanted to say that I agree with much of what has already been said by Tom, Neil, and Bruce. I also want to concur with Thomas that people should check out Michael Corrado's work. He participated in the first "Justice without Retribution" conference and will also be participating in the second and his work is worthy of attention.

With regard to your first point, I use to understand the claim but I'm no sure anymore. You write: "This claim is true: Retributive punishment is not justified if we lack ‘desert-justifying’ free will (the sort we lack according to Pereboom, Caruso, and others whom I’ll label ‘FW skeptics’). But that claim does not entail this one: If we have such free will, retributive punishment is justified. Hence, compatibilists can argue that we have free will but that retributive punishment (or more precisely, some forms of it) is unjustified." While I acknowledge that compatibilists can oppose excessive retributive punishment and be for reform of the criminal justice system, it *does* seem to me that they are still committed to some form of retributivism if one considers blame, shame, and the reactive attitudes of resentment and indignation forms of punishment (which I think they are). To the degree that these attitudes, judgments, and treatments are justified on backwards-looking, non-consequentialist grounds, aren't they retributivist? The compatibilist is essentially saying that the agent justly deserves these attitudes, judgments, and treatments as payback for their immoral actions. Aren't compatibilists committed to at least this minimal form of retributivism? If not, I don't see how they differ from FW skeptics.

I also concur with the point that it is largely an empirical question whether free will skeptics are correct about the benefits of relinquishing the notion of basic desert moral responsibility and the suit of retributivist attitudes, judgments, and treatments that comes along with it. That is why I have recently written about some of the empirical work that is already being done and have argued for optimism. (http://philpapers.org/archive/CARFWS-2.pdf)

[Eddy nahmias]

Thanks for your helpful comments Thomas and Gregg. I'm in mountainschool so will respond Monday. Maybe others can jump in.
Meanwhile, it's wonderful to see America get something right: good work (most of) SCOTUS. Let the wedding bells ring.

[Paul Torek]

Thomas,

Although a common definition of retributivism may include "(b) the suffering of the deserving is intrinsically (and not merely instrumentally) valuable," there is at least one obvious alternative. Rather, a retributivist may hold (b2) the administration of the punishment, by appropriately situated members of society, is obligatory, or fitting. On a suitably deontological or virtue-ethical view or one that includes these elements, the intrinsic value of resulting pains may insufficient to decide the case, and may even be negative. I don't see why the retributivist has to be a follower of G.E. Moore, always identifying the right action with the "best" outcome.

Moreover, something like (b2) is compatible with holding that the details of the crime:punishment correspondences are to be settled by social consensus (resentment and indignation), democratic decision (jail time), or both, consistent with some broad moral constraints such as relative proportionality. Punishments will presumably typically be burdensome or inconvenient, for the sake of general deterrence, but if an atypical criminal enjoys the regimented life in prison, that need not be a problem, or somehow "intrinsically disvaluable".

[Mike Corrado]

Hi, Eddy. Intriguing post. I'm new here, and I don't know the views of everyone involved, so I hope you'll forgive those of my comments that reflect that fact. I think I am sympathetic with much of what you say, but there are some things I'm not clear on.
For one thing, although (as you say) one who accepts free will is not bound to accept that retribution is justified, neither is she bound to reject it. So one question is why the compatibilist will want to reject it. Is it just that as a matter of fact no one deserves punishment, though under the right conditions they might? Or is there some deeper reason? (Sorry if you have already answered that.)
The second question I have is this: you list a number of things that ARE justified, even if retributive punishment is not: blame, moral indignation and anger; quarantine and coercive rehabilitation; demands for restitution. It seems to me that these are very different things, and perhaps the rejection of retributive punishment doesn't entail using them all indifferently on a given offender. (Not that you said it did.) Restitution is in the realm of tort recovery. (Even though it is perhaps a useful tool in connection with crime, I don't think it requires any particular attitude toward retribution.) Quarantine and coercive treatment and rehabilitation are appropriate for one class of offender but not another. You distinguish the sane from the insane; it's only the latter, you say, for which the "objective stance" is appropriate. I like that, and I agree wholeheartedly with it. I would add, though, that the objective stance and the methods that go with it are hardly more humane than the harsh treatment associated with (humane) punishment. (I'm pretty sure I am in disagreement here with Gregg and Derk.) The more humane world is a world in which the offender is not subjected to whatever approach will work, even if it involves rewarding him (an intriguing suggestion from Neil which may play a role before we resort to harsh treatment!), but rather, if he is capable of processing it, to a measured harsh response. I would argue that this cannot be punishment, since punishment is by its very nature retributive, but I don't think I'll get many people to agree with me on that. Instead it must be non-retributive harsh treatment. I think that is something you would countenance as well. But part of the justification for this claim is the very fact that it avoids subjecting the competent offender to the preventive methods available to the state. To that extent it does not rest on any empirical claims; at least if it does I don't see what they could be. Of course it has to be humane, and in part the fact that there must be some justification for that, even if we can't settle on what it is, is clear from the very consensus we see in the remarks here.
On top of that, though, even if consequential benefit isn't to be a test of the amount of harsh treatment given, the great harm done by excessive punishment should be enough to rule it out.

[Thomas Nadelhoffer]

Mike, It's great "seeing" you here in the comment threads. Thanks for accepting my invitation to be the latest contributor to the blog. As I mentioned in my previous comment, your work ought to be of interest (on many fronts) to the readers of this blog. So, I hope you don't mind if I plug your research once again--much of which can be found here:

http://papers.ssrn.com/sol3/cf_dev/AbsByAuth.cfm?per_id=215778

Hopefully, we'll have the chance to interact here at Flickers moving forward. We could always use more skeptics around these parts (or so it seems to me)!

[Bruce Waller]

Eddy, your excellent post has not only generated a great discussion, but -- with the encouragement of Thomas -- has prompted a contribution from Michael Corrado, who is a most welcome addition to the discussion (as Thomas notes, it is particularly nice to have more skeptics present -- but I would be delighted to have Michael's comments even if he were a true believer). While almost everyone here will be familiar with Michael's superb paper in Thomas N's great anthology onThe Future of Punishment, and many of his other papers, if you are like me -- and sometimes so overwhelmed by the volume of interesting work in philosophy journals that you miss some of the really good work in the law journals -- I would particularly recommend Michael's "Responsibility and Control," published ten years ago in the Hofstra Law Review. No one cuts the key issues at the joints with the precision of Michael Corrado.

[Tamler Sommers]

Thomas,

Two points in response to your question. First, there are plenty of cases where I'd favor the restorative approach even when there aren't instrumental benefits (or when the instrumental benefits of some other approach might outweigh them). The consequentialist, by definition, can't say that, right?

Second, there's something a little odd an obligation to atone and make amends for an action that you're not morally responsible for. What would ground that obligation on the consequentialist view? Relatedly, it's important for the offender to feel this obligation--otherwise, part of the point to restoration is lost. Again, I'm sure there are ways for the skeptic to formulate this in their own ethical terms, but I imagine it'll be more convoluted, a pitiful intellectual trinket, overintellectualizing the facts, and so forth. I could be wrong though, which is why I'm hoping to get this account from a skeptic.

[Thomas Nadelhoffer]

Tamler,

I took restorative approaches to criminal wrong doing to be approaches that focus on rehabilitating offenders via reconciliation with the victims and the community more broadly. As a consequentialist who believes that rehabilitation and prevention ought to be the chief justifying aims of the criminal justice system, a model that focuses on rehabilitation via reconciliation is easy enough to motivate and defend for me qua consequentialist.

You suggest there are plenty of cases where you'd prefer rehabilitation via reconciliation even when there are no instrumental benefits--but this can't be right so long as one assumes that reconciliation is an instrumental benefit (which is what I will assume for the purposes of this discussion).

So, that leaves the other possibility you mention--namely, that you would prefer rehabilitation via reconciliation even if this instrumental consideration were outweighed by other instrumental considerations. Perhaps you could provide a case where this holds. I am conveniently drawing imaginative blanks!

As to your second point, it will all depend on what you mean by "atone" and "make amends." If *real* atonement requires a recognition that one is deeply morally blameworthy for an action that one could have avoided, rather than merely causally responsible for voluntarily and regrettably causing harm, then obviously the skeptic can't accommodate atonement.

But why can't atonement simply involve realizing and admitting that one caused harm, made a mistake, etc. coupled with a commitment to making it right (that is, with making amends)? The skeptic need not deny that people can appreciate the error of their ways, feel bad for that they did (as necessary as it may have been at some cosmic level or as luck-bound as the action may have been), wish they had behaved differently than they did, regret that their intentional behavior caused harm to others, etc.

Even the lowly free will skeptic--as convoluted as she may be--need not deny the difference between mice and men (or men and asteroids). Why think this difference isn't sufficient for some rudimentary sense of atonement and making amends?

[Thomas Nadelhoffer]

Tamler,

I have been thinking a bit more about this today and I wanted to say more here. Indeed, if there wasn't a new Featured Author coming up in a few days, I might dedicate a separate post to the issue. For now, I will just sketch out some ideas that were motivated by what you said earlier. I apologize in advance for such a lengthy (i.e., long winded) comment. If it helps, this is more for me than anyone else (as I try to get my mind around some of the key issues at play when it comes to compatibilism vs. skepticism with respect to regret, atonement, making amends, lamentation, etc.). That said, here it goes…

For starters, there are two distinct issues at hand: First, there is the issue of whether it is a strong suit of compatbilism that it seems a natural fit when it comes to restorative approaches to wrong doing. The second issue is whether skepticism has a strike against it because it is not a natural fit with restorative approaches. I want to say something about each of these issues.

Presumably, you think compatibilism is well-suited to accommodate restorative approaches because it makes room for some kind of regret that is necessary for atonement and making amends—a kind of regret that is also purportedly in conflict with skepticism. On this view, regret only makes sense when the agent could have done otherwise. When I feel regret, apologize, and ask for forgiveness, I am taking ownership for what I did and acknowledging that I should have done something else instead. I take it this last part is one of the key issues here. Because the compatibilist can appeal to the conditional ability to do otherwise and the skeptic cannot, the compatibilist is purportedly at an advantage when it comes to capturing the attitudes and behaviors I mentioned above.

If it only makes sense to regret things I have done when I could have done something else, then without at least the conditional ability to do otherwise in her metaphysical tool kit, the skeptic cannot properly be said to accommodate the notion of regret (assuming, of course, the skeptic is a determinist--which need not be the case, but I will set that aside for present purposes). And insofar as restorative accounts give regret pride of place for the purposes of atonement and making amends, this is a knock against skepticism.

But I have some reservations. Let’s assume that an agent can have (a) the unconditional ability to do otherwise, (b) the conditional ability to do otherwise, or (c) neither the unconditional ability to do otherwise nor the conditional ability to do otherwise.

Now let’s imagine that an agent P performed some harmful action x.

If P regrets having done x and P takes herself to have the unconditional ability to do otherwise, then she regrets having done x rather than y and she also believes that she could have done x rather than y even if everything leading up to her decision and action had been the same. In this case, the regret is both deep and personal. That is, the agent regrets the fact that she x-ed and she was the ultimate cause of x. Under these circumstances, when P regrets having x-ed and apologizes for having x-ed, she is apologizing for having done something she could have avoided in the deepest sense—something she could have avoided even if everything leading up to her action had been kept constant. Let’s call this libertarian regret.

If P regrets having done x and P takes herself merely to have the conditional ability to do otherwise, on the other hand, the target of her regret seems impersonal rather than personal. Here’s why: P can’t regret not having done something else in the actual circumstances in which she x-ed. After all, under those circumstances, x-ing was the only thing she could have done. So, what P regrets isn’t so much what she actually did, but rather she regrets that things hadn’t been different such that she could have avoided what she did and done something else instead. If we assume that what P did was the inevitable result of the past and the physical state of the universe tracing all the way back to the Big Bang, then her regret starts to look impersonal rather than personal—that is, rather than regretting having done x rather than y (which, under the circumstances, was something ultimately beyond her control), she regrets that the past and the physical state of the universe were such that she x-ed rather than y-ed. In short, her action isn’t something she can personally regret since her action was in the cards given the past and the physical state of the universe. Rather, she can only regret that things hadn’t been different than they were such that she had done y rather than x. It’s in this sense that P’s regret is impersonal. The target of her regret isn’t her actions but rather the conditions that led her to perform those actions (rather than some alternative action she could have performed if only something had been different). Let’s call this compatibilist regret.

There are two things I want to suggest at this point (and then I will stop for now, I promise): First, it seems to me that the libertarian is better situated than the compatibilist to accommodate restorative approaches (as I take you to conceive these approaches) since these approaches hinge on personal regret, atonement, etc. Second, the kind of impersonal compatibilist regret I sketched above is a kind of regret that the skeptic can bring on board. After all, there is nothing that prevents a skeptic from regretting that the universe hadn’t been different such that her actions had been different. What the skeptic may not be able to do is accommodate personal regret. But I am not sure the compatibilist can capture this kind of regret either. Rather, both the compatibilist and skeptic can merely lament the fact that the universe had not been different. After all, both compatibilists and skeptics alike (of a deterministic bent, at least) think that at the moment of action, there was only one possible outcome given the past and the physical state of the universe.

OK, I am not sure how inchoate that was (which probably means it was pretty damn unclear). If it made no sense, I apologize (in the impersonal sense, of course). I am sure, however, that I have already gone on too long. But since I wanted to sketch this out sooner rather than later, I figured now was as good a time as any.