Free will and prescriptive claims

In this post I would like to call attention to a phenomenon that interests me: the lack of prescriptive claims in free willist philosophy.

To begin, consider Shaun Nichols’ division of the free will problem into three different projects: a descriptive project, a substantive project, and a prescriptive project.  The descriptive project involves determining “the character of folk intuitions surrounding agency and responsibility”; the substantive project just asks whether these intuitions are correct; and the prescriptive project asks how we should revise our moral and responsibility practices in light of these other conclusions.  I think Nichols’ distinction is very insightful and I’m embarrassed to say that I have not kept these three projects distinct in my own thinking/writing on the free will problem.

As I’ve noted before, some philosophers have addressed this last intersection between ethics and the free will problem: Michael Slote, Richard Double, Derk Pereboom, and Saul Smilansky (if you know of any others, I would be interested to know who they are).  Furthermore, all of these philosophers approach this prescriptive project from a more skeptical perspective.  One conclusion about this follows: compatibilists and libertarians have not much addressed the ethical consequences of their views.  This, at least, is my own impression of free willist views: they argue that freedom exists but rarely address what would happen if didn’t exist.

This suggests a fascinating possibility: orthodox views do not address the ethical question because it is irrelevant.  Compatibilists and libertarians would act the same towards wrongdoers whether they had free will or not.  The other possibility is that compatibilists and libertarians would act differently towards wrongdoers if they did not have free will.  But, in answer to this question of how they would act differently, I (perhaps through no fault but my own) only hear silence.  So I would put the question to free willists: what does free will secure for you, in the context of our moral responsibility practices, which non-realism about free will cannot, other than the ability to know/believe/say “free will exists”?

In answering this question, I think it’s important to consider two alternatives.  The first is retribution.  On this view, compatibilism or libertarianism secures retributivism and allows the compatibilist or libertarian to punish just because another committed wrongdoing, without anything more.  Pereboom discusses this possibility at the end of Living Without Free Will and notes some problems with it.

Yet, to my knowledge, no prominent and contemporary free willist has expressed sympathy with retributivism or considered free willism as securing retributivism for hir.  But then the question remains: what, if anything, does compatibilism or libertarianism secure for people, in the context of our moral responsibility practices?

This leads to the second possibility: the prospect of what Daniel Dennett has called a “medicalized society.”  For example, Bertrand Russell wrote:

“No man treats a motorcar as foolishly as he treats another human being. When the car will not go, he does not attribute its annoying behaviour to sin; he does not say, ‘You are a wicked motorcar, and I shall not give you any more petrol until you go.’ He attempts to find out what is wrong and to set it right. An analogous way of treating human beings is, however, considered to be contrary to the truths of our holy religion.”

Similarly, Victor Hugo wrote:

"We shall look on crime as a disease, and its physicians shall displace the judges, its hospitals displace the Galleys. Liberty and health shall be alike. We shall pour balm and oil where we formerly applied iron and fire; evil will be treated in charity, instead of in anger. This change will be simple and sublime."

And, as a final example, Robert Wright reports the following quotation from Darwin’s notebooks:

"One must view a wicked man, like a sickly one—We cannot help loathing a diseased offensive object, so we view wickedness.—it would however be more proper to pity than to hate & be disgusted."

All of these intellectuals think it would better to cure than to punish.  They seem to agree that this would be better even if we can cure/treat/rehabilitate all wrongdoing.  To be more precise, we can consider what I have called a Lower Pain Cost Therapy (LPCT) with the following three criteria:

1.    the LPCT has a lower pain cost than traditional punishment
2.    the LPCT achieves the same deterrent effect as traditional punishment
3.    the LPCT’s alters the person’s personal identity no more than traditional punishment does

Alternatively, one might consider LPCT*s, which only include the first two elements, because we may be willing to accommodate some disrespect for persons to prevent other negative consequences (future wrongdoing, suffering through traditional punishment).  Again, all of these intellectuals seem to think it would be more preferable to use LPCTs (or LPCT*s) than punishment.  The question is: do compatibilists and libertarians agree?  As I noted in the “Wanted: Philosopher to serve as Expert Witness in Court” thread, I think there are three possibilities:

(i)    we regard non-excused wrongdoers as free and morally responsible and therefore refuse to "cure" them and persist in punishing, etc.
(ii)    we treat or cure wrongdoers while also regarding them as free and morally responsible or
(iii)    if there is an intolerable inconsistency in ii, we treat or cure wrongdoers and no longer regard them as free and morally responsible.

And so I put the question to compatibilists and libertarians: which is it (i, ii, or iii)?  I would also put the question to mixed views, such as Vargas’ revisionist project and Smilansky’s illusionist view.  Smilansky is one of the few (to my knowledge) to have addressed this prescriptive project and, although he is ultimately a non-realist about free will, he rejects the utilitarian response.  Even Pereboom notes problems with consequentialism and considers Kant’s Categorical Imperative to be compatible with free will denial.  With respect to almost every view in the free will debate, it is not clear to me how that view would answer this question—but isn’t it the most important one we can ask?  For example, if the truth of free will’s existence had no effect on our moral responsibility practices, then doesn’t the question of whether free will exists become hopelessly trivial?

Comments

[Paul Torek]

It's (ii), but it's more than (ii). By morally responsible wrongdoing, perpetrators open themselves up to retaliation that would be morally prohibited against an innocent aggressor.

There has been a shipwreck, and you are struggling to board a lifeboat. Around you are many drowning people, some of whom are literally scared out of their wits, or so blinded with seawater that they cannot see what they are doing. One such victim is climbing on you, pushing you under; call this person an "innocent aggressor". But there are also some wicked people who are deliberately pushing swimmers under so that they can have the lifeboat's provisions to themselves.

You're entitled to defend yourself against the innocent aggressor. But you're not entitled to "make an example" out of him in order to scare off the others. By contrast, you are entitled to "make an example" out of one of the wicked, in order to scare off the others, provided that it reasonably appears to be both necessary and effective.

That's the most glaring difference which responsibility makes; there are others.

[Jonathan Farrell]

Kip, your questions in the first section of your post strike me as odd, but perhaps I've totally missed the point. So I'll try to respond to what I think are the questions you asked, and hopefully this will allow you to point out which end of the stick I should be holding.

First point: I'm surprised that you're surprised that free willists haven't said much about what would be the case if there were no free will. I can see why someone might assume what they take to be false in some situations. For example, you might do this in order to argue for your position (via reductio-type arguments), or to otherwise indicate the strength of your point of view. Or you might do this so as to see what would be lost were you to be incorrect (I think Kane does a bit of this in 'The Significance of Free Will'), so improving your understanding of whatever it is that’s at issue. There may be other reasons why we’d do this, but I’m not sure what reasons you’re suggesting we might do this for. (Sorry, I've tried and tried, but I can’t make that last sentence any less ugly!) Do free will deniers say much about what would be the case if we _did_ have free will? I don't know the answer to this question, but it wouldn't strike me as strange if the answer was 'no.'

Second point: one common characterisation of free will is that - whatever else we can say about it - it's something that is necessary for moral responsibility. Presumably this comes from thinking that 'ought implies can' or something similar. If you think this, then if we don't have free will then it looks as if there is no ethical question that can be (sensibly) asked. (A stronger, more general, version of this point would say something like: without free will then we cannot (in any robust and interesting sense) _do_ anything. So if we lack free will, asking ‘how should we act?’ is not an interesting question.)

[Jonathan Farrell]

Sorry for posting twice in a row, but this one concerns those bits of Kip’s post where I think I _do_ know what’s going on…

First, a general remark. I'm one of those people who think that it's helpful to keep the free will and moral responsibility issues apart (although of course there are many connections between them). To me, Nichol's substantive project seems to sit in the MR box, not the FW box. So it doesn't seem unreasonable to me for someone to have a position with regards to free will issues, but not with regard to the questions concerning punishment you ask. (As a comparison, I don't think there's a problem with holding a particular meta-ethical view (say, being a moral realist) and yet not having strong views about which normative ethical theory you’d endorse.)

Turning now (at last!) to the main part of your post, it’s not obvious to me that the truth of either compatibilism or libertarianism is enough to secure retributivism. But I’m not sure that you were claiming this, so I won’t say anything about this. More importantly, it doesn’t seem to be as if there are only two options available to free willists: punish or cure.

Punishment would seem to involve treating the agent as responsible for their actions but it doesn’t seem to (or at least might not) involve the notion of improving the agent. ‘Cure,’ on the other hand seems to involve improvement, but not taking the agent to be responsible for what they’ve done. Isn’t there room for (at least) a third option, something like reform, which is concerned with both improvement and responsibility?

The differences might be as follows. Reform is like punishment in that we see the agent as responsible for their actions but, unlike punishment, we attempt to improve them rather than just giving them a ‘negative’ reason (i.e. they want to avoid the experience of being punished) to not misbehave again. Reform is like curing in that it involves improving the agent, but does so in a way that involves the agent themselves (you might cure someone by injecting them with a drug which removes or reduces their antisocial/immoral urges, bypassing their agency. Reforming would be more like education – introducing new beliefs and desires in ‘non-deviant’ ways.)

[Kip Werking]

Jonathan, here are some responses:

1. My comments about free willists (as opposed to agency theorists in general) avoiding prescriptive claims is based upon my observation about Double, Slote, Pereboom, and Smilansky. The comments are only supported to this extent, which is not very great, and perhaps I spoke too strongly. The final question of my post ("With respect to almost every view in the free will debate, it is not clear to me how that view would answer this question—but isn’t it the most important one we can ask?") is put to all positions in the free will debate. That question remains interesting, at least to me, even if non-realist and hard determinists avoid it.

2. "Do free will deniers say much about what would be the case if we _did_ have free will?" Actually, not so much (as I hinted above). Some say strange things like "if people had free will, then their actions would be perfectly inexplicable" (Clarence Darrow, among others, said this). Others say that free will involves some godlike power, like being "causa sui", although it is not entirely clear what that means. Most importantly, however, even these are not prescriptive claims. One notable exception would Galen Strawson's work. Strawson not only characterizes free will as being "causa sui" (plus believing that one is causa sui) but also associates it with a robust moral responsibility. This robust kind of moral responsibility would justify eternal punishment in hell.

3. I tend to agree with you that free will and moral responsibility should be kept separate. Free will, as I understand, just involves that amount of control that people rightly or wrongly suppose themselves and others to have, and without which they could not be (entirely) morally responsible for their actions. The two concepts are linked in this way, but the folk could just be wrong, and so the concepts are ultimately independent (or so it seems to me).

4. Your comments about reform are interesting. But I think they are not emphasizing an element that I think is crucial to this issue: pain. You characterize reform as involving something like punishment. But if we consider painless reform, then it is not at all clear to me that it involves something like punishment. I think it is just like medicine. The central issue, I take it, is pain or suffering, and the question is how much should wrongdoers feel? We can conceive (at least in very futuristic scenarios) how technologies would allow us to regulate behavior just as well as we do today, but without any of the pain.

5. The comments above raise an interesting point: the shallowness of compatibilism. One significant barrier to compatibilism is that, as best as I can tell, it leaves no room for sympathy or pity or empathy with fully-free-in-the-compatibilist-sense-wrongdoers. After all, what ground for sympathy would there be: the agent had all of the freedom and responsibility s/he could ever want or ask for? S/he could have done otherwise in every sense that the agent could reasonably want or ask for? At least to my jaundiced eye, this leaves no room for sympathy for the devil, for the Hitlers and Osama bin Ladens of the world. It results in a sort of hardening of the heart, and turning a blind eye to the factors-beyond-their-control that led innocent children like Hitler and Osama to become the monsters they became. I think this would be to our moral detriment.

One can get a feel for the shallowness of compatibilism when one considers the design or Mele's Zygote argument (or the Boys from Brazil argument). Smilansky has also written a great paper on the subject:

"Compatibilism: The Argument From Shallowness", PHILOSOPHICAL STUDIES vol.115 (2003), pp.257-282.

Available:

http://philo.haifa.ac.il/faculty_pages/smilansky/COMPATIBILISM-%20THE%20ARGUMENT%20FROM%20SHALLOWNESS.pdf

[Jonathan Farrell]

Kip, sorry for the delay in replying.

I think I misunderstood what you meant by 'free willist'. From your last post, it seems that you're using it to mean (something like) 'those who discuss free will'; I'd taken it to mean 'those who think we do/might have free will' (as opposed to 'those who deny that we have free will'). So I thought you were saying 'some people (mostly free will deniers) interested in free will do deal with the prescriptive project, but others (free will asserters, mainly) do not. Why is this?' But you weren't. Oh well!

Regarding your fourth point, and the connection between punishment and pain. I take it you are saying that, to count as punishment (as something justified only on retributive grounds, perhaps?), pain must be suffered, right? If that's the case, then I agree that there could be painless reform.

However, reform and punishment still have something in common: they both take the agent as having been responsible for their actions. And (except in the cases of capital punishment and other kinds of punishment that alter what an agent is capable of) the attempt to alter the miscreants future behaviour does not bypass their agency/autonomy. Curing someone, on the other hand, does not seem to involve dealing with them as agents at all. So although neither curing nor reforming might cause suffering to the wrongdoer, I think there is still a clear difference between the two types.

I haven't read Smilansky's paper, but I don't think I agree that compatibilists cannot be sympathetic. Why might we think this? (That someone had 'all the free will they could have asked for' doesn't seem to help to me. Why think that having the greatest amount of free will possible means no one can (rightly) have sympathy for you?)

[Clifford Sosis]

Hello,

I tend to be a reflective agnostic about the compatibilist/incompatibilist debate but here are a few ideas I've been kickin' around which seem to make compatibilists mad. Now this might be because I'm right, or, it is more likely the case it is because I'm dead wrong. In any event I'd appreciate constructive criticism.
It seems to me, that Pereboom has gone over all of this ground. Pereboom suggests that we should not call anyone “moral responsible” because our moral responsibility practices are fundamentally mistaken. He believes they are mistaken ethically and metaphysically.
Pereboom attempts to show how our metaphysical criteria for determining the appropriate target of our retributive moral responsibility practices cannot be met, and so rejects “moral responsibility.”
He also attempts to show that our retributive moral responsibility practices, as they stand, are ethically reprehensible.
This, Pereboom contends, is another good reason to dispense of, “moral responsibility” as we know it, even if we do have free will. One reason why Pereboom might be reluctant to say, “There are morally responsible agents” is because he acknowledges that when we talk about morally responsible agents, we tend to talk about the legitimate target of our moral responsibility practices, as we know them.
The design of our current moral responsibility practices reflects a commitment to some sort of retributive justice, and punishment. I will not bother to describe the structure of our moral responsibility practices in detail, but I will say that if I were to do so, it would be obvious that moral responsibility, retributive justice, and punishment are wedded to one another in practice.
In my opinion, this might explain why hard incompatibilists do not want to say that we are morally responsible agents.When we talk about morally responsible agents, we are implicitly suggesting that said agents ought to be subject to a retributive system of justice and punishment.
Ordinary common-sense intuitions would seem to suggest that if we are not administering punishment in accordance with the dictates of retributivist justice then we cannot be said to be holding people morally responsible, though we may be holding them morally or rationally accountable, proper.
Compatibilists however, have grown conspicuously fond of “moral responsibility.” For one reason or another, they want to continue to say that, “We are morally responsible.” This may pose a problem for compatibilism: our moral responsibility practices are retributive.
So, if we are talking about morally responsible agents we are not just talking about who we have called morally responsible and why as Fischer, Ravizza and Wallace are; we are also talking about the potential targets of retributive justice. This means that if the compatibilist is not talking about the potential targets of retributive justice, the compatibilist is not talking about morally responsible agents, proper.
In other words, the compatibilist is not talking about moral responsibility practices as they stand. It seems to me that compatibilists are forced to defend the foundations of retributive justice because they want to defend moral responsibility as we know it, lest they abuse the notion of “moral responsibility” and depart from the intuitions that they intended to preserve in the first place.
If they do not engage in this dubious ethical project, they are talking about what would be considered something else by the moral community, such as morally or rationally accountable agents. It should be clear that if compatibilists continue to identify individuals as morally responsible, they need to justify retributivism.
Here is the first premise of an argument that I believe will show that this is the case:

I. If (X is morally responsible) then (X is potentially an appropriate target of retributive justice.)

In general, this seems true. As I have pointed out, the truth of this premise is confirmed by the moral responsibility practices of our community. Is the compatibilist going to be able to produce the relevant empirical data that could disconfirm the truth of my first premise?
A compatibilist could show that moral responsibility, and retributivism are divorced, by showing that there are many formally executed moral responsibility practices that do not treat agents as if they are the appropriate targets of retributive justice in some sense.
It may appear that my first premise is an empirical claim that depends on pending data. If so, whether or not my arguments are sound remains to be seen. However, I suspect the compatibilist or libertarian is not going to be able to show that, in general, there are formally executed moral responsibility practices that do not treat agents as if they are the appropriate potential targets of retributive justice, because the retributive picture of justice corresponds to moral responsibility practices as they stand.
However, I also believe there are good reasons to suspect that moral responsibility and retributive justice are analytically involved. I believe that the extent of that involvement is deep. For instance, an anti-retributive compatibilist who holds a utilitarian or rule utilitarian position regarding justice and punishment cannot be said to be holding agents morally responsible: for utilitarians and rule utilitarians, punishment, in general, is justified only insofar as it maximizes utility. From the utilitarian perspective punishment is justified if it prevents the criminal from repeating his crime or deters crime by discouraging would-be offenders; both of these contribute to social utility. In this type of picture moral responsibility, morally responsible agents and our moral responsibility practices as we know them, do not exist. The administration of punishment is a matter of utility, not moral responsibility.

Consider this: Black rapes and murders an orphan for fun. Several members of community X find out and incarcerate the man. The members of community X realize that he is a rational person, in the instrumental sense. He has repugnant desires. Now, suppose the members of community X have the technological resources to literally, that is, biologically, reform the man, and they do. The members of community X administer a painless procedure in which they 'erase' his repugnant desires, and make him a respectable member of the community.

Two questions:
1. In the story, do the members of community X consider Black a morally responsible agent?
2. In the story, is Black being held morally responsible for his actions, by the members of community X?

I suspect that most people would say that the members of community X do not consider Black a morally responsible agent. I also suspect that most people would say that Black isn't being held morally responsible for his actions. Then again, my hypothesis might be wrong.
A compatibilist that defends retributivism, can argue that free will, moral responsibility and determinism are compatible, because they have shown that it is possible for agents to be morally responsible and that we ought to hold agents morally responsible.
For retributivism, punishment is nothing if not a matter of moral responsibility: criminals deserve punishment equal to the harm done or more. In this picture, morally responsible agents would be subject to punishment. Perhaps an illustration will make my point clear:
Suppose that we figure out, somehow, that retributivism is dead wrong, so we decide to shut down our prison system. Now, instead of punishing criminals, we send them to a criminal ‘hospital’ in which they take drugs that create the illusion that they are still members of society.
Now, rapists, shoplifters, pedophiles, bank robbers and murderers are under the impression that they have been released from prison, and are doing those deeds that we find admonishable, though they harm no one. We continue to give people who are members of our society the impression that they will be punished, rather than quarantined in a hallucination, if they do something wrong.
It seems to me, and most people, that these criminals are not being held morally responsible, because they are not being punished for the bad things that they have done. The fact that most people believe that punishment is a necessary condition for moral responsibility, should explain why most people don’t believe that the criminals in the criminal ‘hospital’ are being held morally responsible.
Retributivism and moral responsibility as we know it and as it stands are wed: if you are not punishing people, you are not holding them morally responsible.
So, if we are not talking about retributive justice and punishment, we are no longer talking about moral responsibility as it understood, or, as it stands.
As far as I can tell, morally responsible agents cannot exist outside of the context of retributivism. I believe that this should reveal why retributivism, and moral responsibility are conceptually hitched. Now, if a compatibilist or libertarian grants me the truth of my first premise, on either empirical or conceptual grounds, and argues that:

II. ~ (X is potentially an appropriate target of retributive justice.),

because they find retributive justice unsound, for some reason, then it would follow that:

III. ~(X is morally responsible).

Free will and moral responsibility are incompatible with determinism, if we deny that retributivism is true, because, in doing so, we are denying that people can be held morally responsible. It turns out, as far as I can tell, that compatibilists have to do more work than they once thought.
If compatibilists or libertarians are going to capture our ordinary intuitions about free will and moral responsibility, which is what they set out to do in the first place, then they are going to have to do more than just provide us with criteria for the attribution of moral responsibility.
Compatibilists argue that a morally responsible agent is responsible for her morally significant conduct. Hence, they say, the agent is, under appropriate circumstances, an apt target of moral praise or blame as well as reward or punishment. My arguments suggests that if compatibilists wants to preserve a notion of moral responsibility as it stands, or as we understand it, they need to defend retributivism somehow or they could give it up and stop using the phrase, “moral responsibility.”
If you do not punish or reward agents, you are not holding agents morally responsible. To say that a person is “morally responsible” but not the target of our morally responsibility practices, as they stand, or as we understand them, is just misleading, but, is there any room for a compatibilist, who believes in free will, but denies that agents are morally responsible? I think so.
Hard compatibilism, as I like to call it, is the thesis that free will is compatible with determinism, even though moral responsibility might not be.
I understand that most people want to say that people have free will so that our moral responsibility practices remain justifiable.
That is why I consider this version of compatibilism hard: it is going to be hard for most people to swallow. However, I find this position attractive, and believe that it represents the most plausible form of compatibilism.
Hard compatibilists can suggest that even though we could be said to be the author of many actions, we are not morally responsible for those actions. This position is light at the end of the tunnel for a reasonable compatibilist who would like to reject retributivism, but continue to say, “I am a free agent” even though determinism may be true.
However, it cannot be true that free will and moral responsibility are consistent with determinism if moral responsibilty, as we understand it, or, as it stands is false, then the rest of the conjunct which it is a part of, ‘free will and moral responsibility’is false. If this conjunct is false, then it cannot be the member of a consistent set. A set is consistent if and only if every member of the set can be true. Free will and moral responsibility, can not, by definition, be consistent with determinism, if we reject retributivism.

[Mark Smeltzer]

Clifford,

You said that,

For one reason or another, they want to continue to say that, “We are morally responsible.” This may pose a problem for compatibilism: our moral responsibility practices are retributive.

This may be true for some compatibilists, but how do you see this criticism applying to compatibilists who believe that when we're talking about MR, we are talking about that which would validate our reactive attitudes?

It may be the case that if our reactive attitudes are valid, it might also validate retributive punishment, but I hardly think that connection is necessary.

Fischer's account, for instance, is silent on the question of how we ought to deal with blameworthy or praiseworthy agents.

[Kip Werking]

I want to thank everyone for the great comments on this thread. Here are some short (and not so short) responses:

Paul:

You suggest that moral responsibility justifies general deterrence. I'm intrigued by your reference to "other" differences that responsibility makes (what are these?).

I think a consequentialist can press you by conjuring up scenarios in which we would want to make examples of even innocent people (e.g. to prevent human extiction, or some other great evil). But then the general deterrence distinction does not seem to be so clear cut; it seems to be one of number and not kind. One would want to ask: if we are unwilling to make examples of innocent people in *ordinary* circumstances but are willing to do so in *extraordinary* circumstances, how will you (or anyone) draw that line?

My guess is that there is no precise way to draw that line because there is just not a sufficient consensus about intuitions on these subjects, and indeed the notion of "free will" has its own ambiguities, that we can wrestle with, without needing to explore the ambiguities in moral responsibility and metaethics.

Jonathan,

Thanks for your fascinating comments; it's great to find someone who shares my interest in these more futuristic scenarios. Let me address one *very* important points that you make:

DISTINCTION: Curing and reform might both be painless, but only curing would bypass the agent's interface, while reform (like punishment) would not

Pereboom makes this same point in LWFW. And earlier I mentioned the possibility of LPCT*s to accomodate these concerns.

The first question I want to raise is: how possible is painless but non-bypassing reform (PNBR)? I can imagine how such PNBR might work in some circumstance (Alcoholics Anonymous, for example). But it also seems possible that some wrongdoers are effectively immune to talk therapy or *any* kind of PNBR. Consider the possibility of a perfectly rational sociopath who, because of some freak accident or mutation, loves torturing innocent people. Such a sociopath might satisfy the most stringest compatibilist requirements for free agency but also be immune to PNBR. Suppose, however, that both curing (which bypasses) and punishment (which does not bypass, but causes suffering) are available. Would you, or free willists in general, feel more committed to punishment?

In other words, one fundamental question (and perhaps I am just ignorant of a larger literature on this subject) is:

How should we weight the cost (if any) of bypassing with the cost (if any) of the wrongdoer's suffering-in-punishment?

[A related question: would it soothe any of your concerns about byspassing if the wrongdoer agrees to the curing or moral pills, like a sex offender agreeing to castration?]

For my part, I regard the bypassing-cost as almost zero and the suffering-cost as great indeed. But I realize that most others seem to have very different intuitions on the subject (and much depends upon how what exactly "bypassing" entails; one can imagine *conservative* bypassing which alters the wrongdoer as little as possible).

These comments concerned the futuristic possibility of curing or moral pills. There is another possibility: perfect prevention or designing lives. In this scenario, a supremely intelligent being (perhaps enhanced humans in an advanced civilization) designs a pre-person/zygote *such that* the person never does any wrongdoing throughout their entire life. For example, the person might be strongly reasons-responsive (to use Fischer's terminology). In this case, there would never be any need to reform or punishment. But it seems terribly strange, at least to me, to say that such people are still free and morally responsible. What seems to have happened is this: curing, or something like it, was done *before* the agent's character came into existence. In other words, both reform and punishment, at least in some very futuristic scenarios, might become obsolete. Is there no tension here between that possibility and free willist views?

Finally, you ask:

"Why think that having the greatest amount of free will possible means no one can (rightly) have sympathy for you?)"

But surely there is an obvious and intuitive sense in which, to the extent a person lacks control over what happens to them, the more sympathy we feel for the tragedies that befall them. If a fully autonomous person, in the compatibilist sense of your choice, freely murders X, we tend to feel sorry for X, but *not* for the murderer. If the murderer were to say "feel sorry for me", we would want to say "Why should we? You had your chance! We told you to stop! This was entirely preventable, you had perfect control over the situation." If something like this is right, and our sympathy diminishes with increased control, then all "the free will they could have asked for" would entail a lack of sympathy. What I have written is imprecise, but I do think there is a kernal of truth here. Smilanksy elaborates on these points in his wonderful article.

Clifford:

Although I sympathize with the inclination to tie compatibilism to retributivism, and without such a tie I suspect that there may be no significant difference between the way compatibilists and non-realists treat wrongdoers (and so the question of "free will" becomes somewhat trivial), I tend to agree with Mark that there is no necessary tie. I find Honderich's comments helpful, when he notes that punishment involves something consequentialist and something retributivist. Neither of these aspects seems absolutely essential to the concept of punishment, and again I think there are sufficient ambiguities in the notions of responsibility, punishment, and metaethics that those studying free will might benefit from putting these questions to the side, and focusing just on (i) how much control people actually have and (ii) how much control people *think* they have.

Mark:

It also seems to me that Fischer's account is silent about prescriptive claims, and I had theories like his in mind when I wrote this thread.

On a related note, you again raise a point, which has also appeared in Tamler's Schadenfreude thread, that I would like to address: defining moral responsibility as what justifies reactive attributes (as opposed to practices like praising and blaming). Remember that when you defined mr as what justifies mr practices, I raised problems involving consequentialism and punishing innocent bystanders for the greater good, etc.

What I did not say in our last dialogue (but I will mention now, because it has come up again and is an interesting question) is that: if, in the alternative, you define mr in terms of justifying reactive attitudes, just how much sense does it make to speak of reactive attitudes being justified? Some points:

1. There are already robust arguments for moral non-realism (or anti-realism; it's not clear to me what difference, if any, there is between these terms) about even species-typical taboos such as incest or torturing infants. Smart (if not necessarily right) people like Josh Greene and J. L. Mackie have argued that even these actions are not wrong in the strong sense that people mean when they say they are morally wrong.

Why do I mention this? If it is already so hard to justify species-typical taboos on actions like incest and torture, how much harder must it be to justify *attitudes*? What does it *even mean* to say "it is wrong, or improper, or inappropriate, for you feel angry at an innocent bystander"? It seems that you want to say that it is wrong in the strong sense that Mackie and Greene say cannot even apply to incest and torture, but you would only seem to be justified in saying something like "I, and my culture, disapprove of your feeling angry at the innocent bystander."

2. But there is a greater problem. Even that last sentence above sounds horribly awkward. Why? Because moral responsibility is usually tied to notions of control, but the extent of control we have our attitudes is not very great. Sure, I can imagine how we might have *some* control over our attitudes---a man might say "I get angry too often and for the wrong reasons, so I am going to go to anger management therapy and curb this bad habit"---but this extent does not go very far. Attitudes have a much greater sense of just-happening-to-us than actions do, and moral responsibility seems more properly to belong (if it belongs to anything) to actions and not attitudes. We rarely, if ever, hold a man responsible for feeling anger. Anger largely just happens to be people. But we do often hold him responsible, at least in a compatibilist sense, for whether he *acts* on that anger. This seems to be the more appropriate domain for moral responsibility.

But, once we return to actions, we return to the consequentialist examples involving harm to innocent bystanders for the greater good. These raise the interesting prospect of people who inwardly *feel* very differently than they *act*.

Perhaps, as I've suggested above, it would be better (at least for now) for those studying free will to just focus upon how much control people have and how much control people *think* they have. Nevertheless, I wanted to start this thread and explore those intersections, to see just how murky and confusing this territory is.

[Paul Torek]

Kip,

The "other" differences, reactive attitudes and respect for agency, have already been discussed, and I don't feel that I have much to add there. As Mark points out, many reactive attitudes seem to make a presupposition of responsibility. I do have one thing to add there. Ronald de Souza's book, The Rationality of Emotion, is a good place to start if asking just how much sense it makes to speak of reactive attitudes being justified.

On bypassing a person's agency, I like what Jonathan has to say. But speaking from my own experience, I have to say that painless reform is pretty much purely theoretical. After all, the first step in reform is to admit that you're flawed. And unless you're amazingly enlightened, that's gonna hurt.

Even if the difference between acceptable responses to innocent aggressors vs true perpetrators is/were a matter of degree, it still seems important. After all, at night there are still some photons entering your eyes, but the difference between night and day is like the difference between night and day. I realize this quip is getting a bit hackneyed, but it's y'all's fault for giving me too many opportunities.

[David Hodgson]

One very important reason for maintaining notions of free will, responsibility and retribution is that these notions underpin deeply-held principles of justice and human rights, which are regarded as essential pre-requisites for civilised societies.
These principles give great weight to the autonomy of people and require respect for that autonomy. According to them, a citizen is generally entitled to freedom from interference from the coercive processes of the State unless he or she voluntarily breaches a fair rule of law, publicly promulgated by the State. That is, a citizen should have a choice as to whether to be liable to coercion or not.
This is not to say that criminals should be punished purely as retribution, because that is what they deserve: on the contrary, what I am saying is entirely consistent with the view that the criminal law must be justified largely by its utility, because of the need to protect the majority of citizens from dangerous and anti-social activities of others. But, given that society needs such a system, the question is how should this system identify those persons to whom the coercion is to be applied, and how should it determine what coercion is to be applied to these people.
The solution which has been adopted by civilised societies with respect for justice and human rights is generally along the following lines, which I call the human rights qualification to the power of the State to coerce its citizens: as a general principle, the system should allow coercion to be applied only to people who have been proved, by due process of law, to have voluntarily acted in breach of a public law; and then, no more coercion is to be applied than the person deserves by reason of the gravity of the offence and the degree of the person's responsibility for it.
It is of course argued that there are good utilitarian or consequentialist reasons for this human rights qualification, which do not require acknowledgement of the validity of any idea of criminal responsibility which legitimises punishment. For example, there is the consequentialist argument that there should be parsimony in threatening and applying coercion; and since it is generally only voluntary actions which are susceptible to deterrence by threat of punishment, it makes sense to threaten and to apply coercion only in respect of voluntary actions. But even this argument presupposes a fundamental distinction between actions which are voluntary and actions which are not which is difficult to sustain if one rejects free will and responsibility.
And in any event, the argument does not do away with the need to be able to rely on appeals to justice and human rights, which attribute great significance to the distinction between acts which are voluntary and those which are not - particularly because consequentialist arguments are irredeemably indecisive. Those governments which do not recognise the human rights qualification to the application of the coercive processes of the State generally assert that the good of society overrides other considerations; so that, for example, the detention without trial of political opponents is justified. The strengthening consensus of reasonable opinion throughout the world is that that kind of approach can be justified only in cases of real emergency, and then only as a temporary measure pending resolution of the emergency and as a part of the process of establishing or returning to a normality in which the human rights qualification does apply. However, if one is limited to arguments about consequences, it is impossible to make out a case which would convince anyone who wished not to be convinced: it is impossible to prove what all the consequences of the alternatives would be, let alone to prove which of the totalities of consequences would be ‘better’.
Imagine trying to prove to a previous Soviet government that it would have had better consequences not to confine dissidents in mental asylums.
But acceptance of the independent force of the considerations of fairness which justify the human rights qualification means that there needs to be positive justification for overriding it - in terms of real emergency, clear and present danger, etc. If justice and human rights are given independent weight, then a heavy onus can be placed on governments seeking to deny human rights to their citizens.
So I contend that these ideas of human rights depend crucially upon discriminating between persons whose conduct makes it fair and permissible that their freedom be curtailed, and persons whose conduct does not make it fair and permissible that their freedom be curtailed. And in order that conduct make curtailment of freedom fair and permissible, it must involve voluntary action in breach of a public law, for which the person is responsible.
If the notions of free will and responsibility are discredited, human rights are prejudiced: there will appear to be no rational basis for saying that it is fair, and thus permissible, to curtail the freedom of a person (who has had the bad luck to be caused by genes and environment to become a person) who acts in breach of the law, yet unfair and thus impermissible to curtail the freedom of a person (who, without breaching any law, has had the bad luck to be) regarded by the government as a danger to society.

[Jonathan Farrell]

Kip, I should first say that I have no well-thought out position when it comes to issues of punishment/reform/cure so these are all ‘first thoughts’ – I for one certainly am ignorant of the literature on this topic. (And Paul, I agree that painless reform is unlikely in this context. Perhaps the difference between punishment and reform here is that the pain involved in reform is incidental to its purpose, whereas that is not the case with punishment.) That said, the issue of someone for whom reform could not work is an interesting one. One the one hand, we should strive to treat people as autonomous agents (to the degree that they actually are) and on the other, there may well be problems with justifying retributive punishment. So whichever way we go, we may cause harm (although one that will presumably be outweighed by the benefit of improving the wrongdoer's character or attitudes). Perhaps it might come down to considering individual cases where we compare the harms involved in the two possible approaches (i.e. unjustified infliction of pain and bypassing of agency).

I have very different intuitions to you when it comes to the cost of bypassing agency – I think these costs are likely to be significant. Free will is often taken as being important for moral responsibility, and it seems to me that it is this ‘angle’ on free will that gets most attention in the literature. However FW also seems to be crucial for autonomy. If we don’t have free will, then it’s hard to see in what sense the life we live is our own – it seems that we are more like spectators than participants. So by bypassing someone’s agency, we condemn them to observe their life being lived, rather than allowing them to live it. And, I would claim, considerable harm is done by doing this. (So I strongly disagree with your claim in your response to Clifford that, if our only options are non-realism and compatibilism, the answer to the ‘free will question’ is trivial.)

So it seems to me that whether a wrongdoer assented to being cured would be important. I would argue that if you seek to kick your nicotine habit by getting hypnotised then this is like a cure, and yet you remain responsible for giving up the cigarettes. Drunk-driving might be a similar kind of case too (again, you presumbly freely chose to abdicate (or reduce in some way) your agency).

Regarding perfect prevention, I think that what Watson says (if I remember correctly) is right: compatibilists should be hard compatibilists. As you note, however, it might seem strange to say that such agents would have free will. Perhaps the compatibilist could say that this is because our notion of free will is still ‘infected’ with ideas which involve us being causa sui, and that, once we have ‘cleaned up’ our notion of free will, perfect prevention cases will not seem so strange. I’m not sure how satisfactory this sort of response might be though. And/or the compatibilist might note that it doesn’t make sense to talk of curing (i.e. bypassing the agency of) an agent than doesn’t yet exist. This would probably only deal with one way of stating this issue, however; I think the question of what we should say about perfect prevention would still remain.

On your last point, I agree that there is a connection between feeling sympathy for someone and the amount of control they have over their situation. I disagree with you, though, when you suggest that a compatibilist could claim that a free agent has _perfect_ control over their situation. There will always be things that are beyond the agent’s control that are relevant to their situation. Here’s an analogy (although I’m not sure how good it is): someone might be the best (fastest) sprinter that a human being can possibly be, but this doesn’t mean that there are not speeds at which they cannot run. So there will always be ‘room’ for us to have sympathy with a free agent, even one who is as free as they can be.

[Kip Werking]

Jonathan,

Thanks for your generous reply. You raise enough issues (and many that fascinate me) to write a whole book. This comment will have to suffice.

I, too, am largely ignorant of any literature discussing the (ethical) differences between punishment, reform, and "curing", etc. Indeed, I don't think there is much of a literature on the subject. I know that certain thinkers, like Karl Menninger and Sheldon Glueck, discussed similar ideas before the "retributivist revival."

You wrote:

"I have very different intuitions to you when it comes to the cost of bypassing agency – I think these costs are likely to be significant. Free will is often taken as being important for moral responsibility, and it seems to me that it is this ‘angle’ on free will that gets most attention in the literature. However FW also seems to be crucial for autonomy. If we don’t have free will, then it’s hard to see in what sense the life we live is our own – it seems that we are more like spectators than participants. So by bypassing someone’s agency, we condemn them to observe their life being lived, rather than allowing them to live it. And, I would claim, considerable harm is done by doing this. (So I strongly disagree with your claim in your response to Clifford that, if our only options are non-realism and compatibilism, the answer to the ‘free will question’ is trivial.)"

This is very interesting for multiple reasons.

1. You express concern about bypassing and autonomy. But you are (I presume) perfectly comfortable with the fact that we have prisons for wrongdoers. I am not suggesting, of course, that we should use bypassing with those who do no wrong. Rather, like prison, it is a tool of last resort. So, once we are in this context of wrongdoers, and not just people in general, it does not seem (to me) that autonomy poses any special problem for the therapy/curing I discuss. Both prisons and therapy/curing violate the wrongdoer's autonomy.

Prison and curing might violate autonomy in different ways, however. You can put someone in prison, but they can fume and boil and nurture their love of torturing innocent people. Such a person's physical autonomy has been violated, but their psychological autonomy is less compromised (although it is still compromised to some degree). In contrast, if we could "cure" someone, they go walking free as a public citizen, safe to the community, and in this way they would have physical autonomy but their psychological autonomy might be more compromised.

2. The other (perhaps more important) thing I would say about that paragraph is that it is very curious to talk about giving or taking any one's free will. Free will, as I understand (and I think most people understand it, at least to some degree) involves a certain immunity to accidents of birth. Understood this way, free will is also something we don't have. It is quite distinct from freedom of action or movement, which can be given or taken away.

To help you understand how this concept of free will is relevant to your paragraph, consider this example. Suppose that X is, like myself and yourself, born with certain largely species-typical pro-attitudes or values, like preferences for family members over strangers, and preferences for eating food over starving to death. Now suppose that X is subject to a radical curing-like intervention which, instead of making X safe for society (and more similar to the rest of us), reverses these species typical pro-attitudes or values, making him prefer strangers to family members and prefer starving to death over eating food (and less similar to the rest of us).

It is a strange example, to be sure. But you may see my point by considering this question: what autonomy has this curing-like intervention taken away from X? Sure, he did not freely choose to prefer strangers over family members, and he did not freely choose to prefer starving to death over eating food. But he *also* never chose to prefer family members of strangers, or eating food to starving to death. He was *thrown into this world* with those original preferences, as part of his genetic makeup, and if he is now a slave to the new preferences, he would also have been just as much a slave to the old preferences. In at least one important sense, he seems just as autonomous, with respect to these preferences, as he has always been: not very.

I would also say that we do not need to get tangled in thorny issues involving moral realism to make my point. In referring to X, in this context, I can limit myself just to talking about how free or autonomous he is. I understand that some would like to say, for example, that just because we are all thrown into the world with those original preferences, and they seem to that extent arbitrary, that does not prevent them from being truly, objectively, absolutely moral or right. For now, we can separate the threat that accidents of birth pose to morality from the threat that they pose to freedom or autonomy (although I think they are intimately linked, and I suspect that one who had difficulty grasping the one would have difficulty grasping the other). Regardless of how truly, objectively, absolutely right the preferences and values we are born with are, I hope you can still see how X was already, in one important sense, not free with respect to his original preferences.

In other words, I think the truth of Mele's no-difference principle (discussed in his recent papers posted on the Garden) can apply with equal force to certain curing-like interventions.

If you can understand this threat, then you might see how strange it is, to my ear, when you say that bypassing would undermine a wrongdoer's autonomy. As the X example should show, such wrongdoer's are already significantly un-free or non-autonomous. I certainly don't see any reason for necessarily respecting the wrongdoer's autonomy to commit (uncontroversial) wrongs.

Moreover, I tend to think that autonomy, if it is valuable, is valuable because it is conducive to other things, such as pleasure or happiness, and not in and of itself.

You also raise the issue of revisionism. This is entirely different issue, and worthy of a book on its own. Manuel Vargas has does a great job of highlighting this possibility and adapting it to compatibilism. I think Watson (whom you mentioned), in his critique of Van Inwagen's Essay on Free Will, also expresses revisionist tendencies. Without changing the subject entirely, I will pose this initial problem for would-be revisionists:

In discussing free will and moral responsibility (fw and mr), Vargas gives the example of whales. We thought whales were fishes, but they are mammals. But that doesn't mean whales don't exist. They do.

In contrast, fw and mr seem, to me, more analogous to unicorns. Consider philosophers, and others, who have precious beliefs in unicorns. They look out into the real world and, much to their chagrin, they only see horses. No unicorns. But in a strange, and perhaps misguided, attempt to preserve the existence of “unicorns”, they try to convince people that unicorns don't need to have horns. Then they can point to horses and say "unicorns exist." But unicorns do have to have horns. And they do not exist.

In other words, the danger is that revionists defend fw and mr in name only.

I've talked with Manuel a little about this. I wanted to know: if fw and mr could be like whales, or like unicorns, how we decide which one? He suggested that we need to see if anything in the world serves the role that we thought fw and mr serve. I'm not sure if this is right, but whether it is or not, I think we need to collect much more data on what *precisely* people think fw and mr are and what role they are supposed to play in our lives. For example, if fw was supposed to immunize us from accidents of birth, and/or render disturbing situations (like someone else designing our entire life) impossible, then revisionism may not work. But maybe it will. I certainly don't think there is much of a consensus, with any precision, about what mr and fw are.

Finally, you raise the tantalizing example of the sprinter. You note that there are faster speeds than even the fastest runner could run. And, by analogy, one might have much greater control over one's life than even the most compatibilist-free agent has. And so one might have greater sympathy and pity for the tragedies that befall compatibilist-free agents to the extent that they do not have this greater amount of control.

But do compatibilists ever talk like this? When Watson says that agents with totally-designed-lives should be held mr and regarded as free, he does *not* also say "but notice how much control this agent doesn't have over its life, and we should accordingly feel that much greater sympathy over the tragedies that befall it, even those that it superficially seems to bring upon itself." (He comes close to saying something like this in Responsibility and the Limits of Evil, but note that the murderer discussed there is quite different, psychologically, from ordinary people and ordinary murderers). Similarly, when Fischer insists that Plum could still be morally responsible (but not blameworthy!) for murdering White, he does not also say "but also realize just how much is obviously beyond Plum's control, as seen by the manipulators that are playing him like a puppet, and we should therefore feel so much greater sympathy and pity--and not just anger--for Plum's horrible actions." I've *never* heard a compatibilist talk like that. (One might detect a hint of this sentiment in the fact that Fischer doesn’t regard Plum as blameworthy, but this extent must be very small, because Fischer still considers him morally responsible.) In general, I consider this to be a moral blind-spot in compatibilism. Because it lowers the bar for fw so low, it becomes morally blinded to tragedies that befall people because they do not meet the higher bar for a stronger kind of fw. It is, to use Smilanky’s word, “shallow.”

[Kip Werking]

Here are two short additions to my original post (not necessarily directed at any one of the commentators):

1. In the original post, I inquired about any link between retributivism and free willism. With respect to that, I would note this interesting study I've found:

o PSYCHOLOGICAL REPORTS 93 (3): 1013-1021 Part 2, DEC 2003
o Abstract: Determinists were compared with weak, moderate, and strong libertarians with respect to philosophy of punishment. Data provided support for the contention that determinists are less punitive than libertarians

2. At multiple points in this thread, I've suggested that free will is exclusively about control. That's not quite right. I also suspect free will is essentially about creativity or unpredictability (if not absolute unpredictability, then freedom from absolute predictability). Indeed, I think it in part because free will has *both* of these aspects (control and creativity), and they do violence to each other, that the existence of free will is questionable. Double says something similar in his critique of Kane's libertarian view.

[Jonathan Farrell]

Kip, apologies for the delay in replying. Here’s some remarks on your last-but-one post.

Re your (1). I don’t think that prison violates the kind of autonomy that I mean to be talking about here. I don’t think that locking someone in a room affects their free will (unless they go mad with loneliness or something), although clearly it limits their freedom in some wider sense. So I don’t think that imprisonment is a kind of curing as we’ve been talking about it.

Regarding (2), this is a very interesting topic (in fact I think this is the area of the free will debate that I’m most interested in.) In the example you give, we have an agent, X, who has ‘species-typical pro-attitudes.’ Then radical curing occurs and we have an agent, X*, who has very different pro-attitudes. If X is free – and we’re stipulating that he is – then X* is also free. I agree. I also agree that neither X nor X* were free with regard to (i.e. did not & could not choose) their ‘initial conditions.’

However, it is not clear to me that X and X* are the same agent. There’s (at least) two ways this might be so. Perhaps the changes that occur to X are so radical that X* and X are not even the same person. If this is the case, then I think it’s fair to say that they are not the same agent. But perhaps the changes are not as extreme as this. Now we might want to say that X and X* are the same person, but they are not the same agent. (In fact, I think we can go further, and say that X and X* are not the same agent with regard to actions that depend on the attitudes that have been altered, although they may be with regard to other actions.)

This is why I think that curing can bypass an agent’s autonomy. X did not gain these new attitudes in the right kind of way (however that is to be unpacked) and the resulting agent, X*, is a different agent to X.

On another note, unlike you, I think that autonomy/free will may be valuable for its own sake. I’m sure you can be happy/satisfied without free will, but I’m not sure that you can the kind of ‘good life’ that people like Aristotle talk about without it. This goes back to the spectator/participant distinction I mentioned in an earlier post: I think most people would rather be participants in a not-too-bad life rather than mere spectators on a perfect one.

On unicorns: here are a couple of ways to take your unicorn example. Firstly, before we look out into the world, we know there are horses (and we call them horses). When we find a lack of horned-horses, we have two options: decide that there aren’t any horses, only unicorns, but unicorns are not quite what we thought they were. Or we deny that there are unicorns (and affirm that there are horses). I think it’s pretty clear that the second option is best.

Secondly, we could take it that your story goes this way: before we look at the world, we’ve no knowledge of horses (and, presumably, no term ‘horse’). Now, when find a lack of the horned creatures, I think there’s not much between the two options: affirm the existence of horses and deny that of unicorns, or vice versa. (I suppose it might depend on how important to the notion of unicorns it is that they have horns.) Compare unicorns with salamanders. ‘The mythical salamander resembles the real salamander somewhat in appearance, but makes its home in fires, the hotter the better…. These myths originate in Europe from the fire salamander, Salamandra salamandra, which hibernates in and under rotting logs. When logs were brought indoors and put on the fire, the animals mysteriously appeared from the flames.’ (http://en.wikipedia.org/wiki/Salamander) So there are salamanders, they’re just not quite how we thought they were.

I think the FW situation is more like the second way of reading your story and I think that we should say that there aren’t any horned equine beasties out there, but that a lot of what we were looking for (hooves, manes, tails, whinnying, etc.) is out there, and since there’s no other obvious name for it (we don’t have the term ‘horse,’ remember), calling it a unicorn is the sensible thing to do (just like the salamander).

The sprinter: you say that you’ve never heard a compatibilist note that even the freest of agents does not have total control over their actions and so can be the appropriate object of sympathy. You may well be right that they don’t say this (I certainly don’t have a counterexample to wave about). But I’m tempted to think that this is because this is just an obvious fact. Even Sartre, who makes rather radical claims about what freedom entails, acknowledges that we are limited by our facticity – our initial conditions.