What is the Free Will Problem About? Or The Four Questions

What is the free will problem about? I have found it helpful to think about it as a combination of four very different questions:

1. Is there libertarian free will (LFW)? (Here would go as sub-questions the issue of determinism, the question whether libertarian free will is coherent, and so on.)

2. If LFW does not exist, do we still have moral responsibility and the related things (e.g. desert)? This is, of course, the familiar compatibility question, best asked as the question whether moral responsibility is compatible with the absence of LFW (rather than with determinism).

3. If we do not have moral responsibility (etc.) because of the absence of LFW, or it is at least seriously harmed by the absence of LFW, is this a good thing?

4. What should we do about the conclusions to questions 1-3?

The order seems intuitive, in that pessimism about the first question leads to the second, and so on, while optimism about an earlier question means that we wouldn't be too worried about the question that comes later.

The first two questions are familiar, but I think that we need to think about questions 3 and 4 as well, much more than we commonly do, and in any case be clear which question is being addressed in a given work. Derk Pereboom, for instance (in his Living Without Free Will, CUP 2001) says "no" to the first two questions, but a qualified "yes" to the third, and hence when he comes to the fourth question he is not too worried. In my Free Will and Illusion (OUP, 2000), I say "no" to the first (existence of LFW) question. Then on the second (compatibility) question I give a complex yes-and-no answer, attempting to combine the insights of compatibilism and hard determinism, a position which I call "Fundamental Dualism". Since I think that the absence of LFW and the (partial) harm this does to moral responsibility are negative and highly significant, I answer "no" to the third question as well. Which then leads me to think about question number 4, and my "Illusionism" on free will aims to deal with that question. I would not be able to make sense to myself of much of my disagreement with Pereboom (or, differently, with P.F. Strawson's classic position in "Freedom and Resentment"), if I continued to think only in terms of the first two questions.

Is this helpful? Are these the right questions?

Saul Smilansky

Comments

[Robert Allen]

Prof. Smilansky,

Framing the debate in this way makes it appear as if LFW is the only type of FW; that FW stands or falls with LFW. But any Compatibilist defense of MR worthy of the name, in my opinion, would show that FW is also compatible with CD. Moreover, in refuting Libertarianism, it would show that it’s Indeterminism that’s incompatible with FW (and MR). (Cf. my “Free Will and Indeterminism: Robert Kane’s Libertarianism,” forthcoming in The Journal of Philosophical Research and available @ my website, http://home.twmi.rr.com/robertallen/papers.html.) I have never been able to separate the notions of FW and MR, as some Compatibilists apparently do: if an agent is MR, then (in some sense) she committed her action of her own FW and vice versa.

p.s. FW&I is not only one of my favorite FW books, but one of my favorite books period.

[Mark Smeltzer]

Dr. Smilansky asked,

"Is this helpful? Are these the right questions?"
In the context of the current debates, this is a perfectly natural set of questions, and the ordering principle fits, but I believe that ultimately these are the wrong questions, or at the least represent the wrong place to start.

I believe the question that bears primacy in this domain of inquiry is something to the effect of: what do freedom and moral responsibility look like? If this is correct, then it is problematic to begin one's inquiry with questions that relate to the topics of LWF or the compatibility question.

The reason I believe this starting point to be so crucial is that I do not think we have come to consensus about what freedom and responsibility look like. Thus, it will prove impossible to ever come to consensus on a single theory of freedom and responsibility. Moreover, if one does not have a clearly defined picture of freedom and responsibility, one's views on these will be shaped by the theory one identifies with. In other words, I believe it is erroneous to say things like, "Because I am a Libertarian I see freedom this way..." We should instead be saying things like, "I am a Libertarian because I see freedom this way..."

Ultimately this comes down to the problem of the criterion. I believe approaches such as the one outlined by Dr. Smilansky represent epistemological methodism -- something I strongly oppose. I believe, and would hasten to argue, that we must approach these topics as experimental particularists; lest we commit ourselves to skepticism from the start.

[Derk Pereboom]

I think that Saul’s progression of questions tracks the concerns of many people who think about the issue of free will and moral responsibility. Many of us initially assume that we have libertarian free will – and the claim that many people do so finds support in the Nichols/Knobe paper on this website. We then encounter the threat that general causal nature of the world precludes responsibility-relevant control. It’s then natural to resist this threat, and either retain belief in libertarian free will, or deny that the general causal nature of the world precludes this sort of control. Some of us conclude that the threat wins out. That covers questions 1 and 2. Questions 3 and 4 are primarily a concern for this last group. But there are certainly reasonable participants in the debate who think that it’s so unlikely that we lack responsibility-relevant control that questions 3 and 4 are of merely academic interest to them.

My guess is that for some participants in the debate Saul’s progression of questions is less natural. They don’t begin with the assumption that we have libertarian free will, but rather with the assumption that we have responsibility-relevant control. To them, it is somewhat puzzling or surprising that others would think that there is a serious threat to responsibility-relevant control from the general causal nature of the world. For them, the main tasks are to provide a theory that brings to the fore the rules of our practice of holding people morally responsible, and to explain to the others that they are mistaken in taking seriously the claim that the general causal nature of the world poses a threat to this practice. There are undoubtedly further mind-sets, but I think that these two are well-represented, and that different progressions of questions would be natural for them.

One worry suggested by Mark’s comment is that we are still trying to sort out the concepts of free will and moral responsibility. In my view, we have considerable clarity as to which notion of responsibility is at issue. Responsibility in the sense at issue is the sort of control over our decisions and actions that would make it the case that we deserve blame for the immoral ones, and praise for those that are morally exemplary. Then the question is: what sort of freedom is required for responsibility characterized in this way? The concept of this sort of freedom, on which all of the participants in the debate can agree, is in a way functional or relational: it’s whatever sort of freedom is required for moral responsibility, defined as it just was. Now of course the various participants in the debate disagree about the characteristics of agency that will fit this definition. But this isn’t a conceptual debate – we’re not trying to clarify some concept of freedom. So the concept of moral responsibility is clear, and the relational concept of freedom is also clear. At the same time, there are many different conceptions of freedom that are proposed as the requisite characteristics of agency – freedom as the availability of alternative possibilities for action, freedom as reasons-responsiveness, freedom as voluntariness, etc.. But there is also no conceptual murkiness here, since concepts of these various sorts of freedom have been made clear by their advocates. Sometimes people need to be clear as to which concept of freedom is being invoked in some discussion, but that’s not a lack of clarity about any concept.

It is true that sometimes concepts of responsibility distinct from the one just defined are raised in the debate – for example, Hilary Bok’s notion of the legitimacy of calling an agent to account, and Gary Watson’s aretaic notion of responsibility. But this does not reflect conceptual unclarity either. These are concepts of responsibility clearly distinct from the one at issue in the debate, and from each other. And they are not the concept that is first and foremost at issue in the debate, since incompatibilists generally would agree that responsibility in these other senses is not threatened by the general causal nature of the world.

[Tom Clark]

I think Saul’s questions nicely capture the basic metaphysics and moral significance of the free will debate, and I’m glad to see 3 and 4, getting at the practical ramifications. Reading about the various manipulation scenarios in the “Mele on Pereboom” thread, I was thinking: Ok, so what’s the upshot of any of the positions staked out? So Saul’s questions remind us why all this is important. My take on them is as follows, for what it’s worth.

Question 1 asks about the metaphysics of human agency: are human beings in any sense causal exceptions to the deterministic unfolding of things that science describes at the macro level? To say there is no libertarian free will (LFW, what I’ve called type 1 freedom way back when) is to say that we’re not, and I agree with Saul that indeed we’re not.

Assuming a no answer to Q.1, question 2 asks what happens to moral responsibility if we aren’t causal exceptions to nature. Of course, if MR is conceptually tied to LFW, then we don’t have what we might call LMR if we say no to question 1. I take it that it’s this desert-implying sense of MR that Derk says above is “responsibility in the sense at issue.” But if our conception of MR is, for instance, that of holding rational, reasons-responsive, voluntarily acting agents (that therefore have type 2 freedom) responsible in order to shape behavior, then such MR survives in the absence of LFW and desert. Call this consequentialist MR, or CMR. Praise and blame are justified instrumentally, not from desert.

Question 3 asks whether or not it’s a good thing if we don’t have the sort of MR tied to LFW, but instead perhaps a weakened or alternative conception, such as CMR. Here my take diverges sharply from Saul’s, since MR hasn’t been “harmed,” just naturalized, and that’s a good thing since now we can live in the light of the empirical truth about ourselves. We don’t need supernatural MR (LMR), or belief in it, to ground ascriptions of human agency, dignity, value, meaning, etc. If LFW and LMR are logically and scientifically untenable (as Saul agrees), we shouldn’t continue to use such conceptions as benchmarks for what we think must be the case to be morally responsible, dignified agents. Giving them up can lead, as I, Tamler, Pereboom, Richard Double, Kip Werking, Galen Strawson and others surmise, to a more effective, more humane set of responsibility practices (I’ve called such folks type 2 progressives). We shouldn’t take our current practices as canonical, since aspects of them may presume the faulty metaphysics considered in question 1. And keeping the illusions of LFW and LMR in place only obscures the causal reality of how people become good or bad.

Question 4 asks about the upshot of our responses to questions 1-3. What should we do, for instance, if it becomes clear LMR doesn’t exist? My suggestion is to educate ourselves out of the LMR myth, and move toward the science-based, humane CMR alternative, as for instance in criminal justice and social policy. Cases in point: opposing the death penalty, see “Crime and causality: do killers deserve to die?” at http://www.secularhumanism.org/library/fi/clark_25_2.htm , and on criminal justice more generally, see Joshua Greene and Jonathan Cohen’s “For the law, neuroscience changes nothing and everything,” at http://www.csbmb.princeton.edu/~jdgreene/Greene-WebPage_files/GreeneCohenPhilTrans-04.pdf , which recommends “a shift away from punishment aimed at retribution in favor of a more progressive, consequentialist approach to the criminal law.”

[Kip Werking]

Saul Smilanksy:

I wanted to read some of your work (which others have recommended to me) before commenting upon this thread. I had hesitated because I was familiar with your view about “illusionism” and my immediate reaction was one of distaste. So I am happy to report that your Egalitarianism, Free Will, and Ultimate Injustice greatly impressed me. You write very well and, in short, eloquent prose, capture some of the truths about free skepticism that myself and others struggle to express. For example: “On the other hand, given that the sort of transcendence that libertarian free will is supposed to give us cannot exist, everything that people do, including their compatibilistically-free actions, is ultimately an unfolding of the given.”

From this point I must respond to your remarks about the four questions. My first reaction to the third question is to recoil from it. This is not to say that questions of value are not important or do not belong within ethics. Rather, I agree with Derk Pereboom when he criticizes Dennett, in his introduction to LWFW, for only examining “the varieties of free will worth having.” It seems to me that the focus of Dennett upon those varieties, and your focus upon the third question, while they might not commit the fallacy of adverse consequences (it’s bad, therefore it’s not true), they greatly increase the temptation to do so. For whatever reason, people seem to have tremendous difficulty divorcing the question of free will’s value from the question of its existence. Perhaps this is because, and you show, lacking free will causes an unavoidable and tremendous amount of injustice in the world. Whatever the reason is, I am not so sure that people ask these questions in the order that you pose them. All too often, in my experience, the answers to questions one and two turn upon the answer to question three. Of course, that would be a mistake.

This leads me to some remarks about your “illusionism,” with which I only have a passing familiarity. According to illusionism, we don’t have free will, but we should act like we do. Is this correct? Tom Clark relates that Steven Pinker, Marvin Minsky and Matt Ridley have also suggested that free will is a necessary fiction. That position strikes me as opposed to the spirit of philosophy. It would take us backwards, through the Enlightenment, and into the Middle Ages, when men believed in the liberum arbitrium voluntatis.

You speak of the dangers of abandoning our belief in free will. What of the dangers of continuing to believe in it? What of the victims of the Fundamental Attribution Error? What of those whose executions only serves the purpose, not of deterrence, but of quenching our blood thirst?

For a more careful critique of Smilansky’s illusionism, see Tom Clark’s Is Free Will A Necessary Fiction:
http://www.naturalism.org/fiction.htm

[Neil]

For many people (including me), something like Saul's questions are the ones that grip us and make free will important. But they are not the only ones that matter. Some people don't think that free will is essentially about moral responsibility, but instead about origination or ultimatacy. Saul's questions are good ones, but they are not the essential questions for everyone. There is no single set of questions, or decision tree, that captures all the central concerns captured under the heading "free will".

[Paul Torek]

If the answers to questions one and two turn upon the answer to question three, would that be a mistake? Not necessarily. If the answer to question three is "it doesn't matter one whit," then we have made a mistake in answering question 2, by overlooking a sense of responsibility which is indeed important. Dennett is right to insist on examining varieties of free will worth wanting. (Maybe we should add that if there are any varieties worth fearing, we should examine those too.) Otherwise free will becomes an "academic" question in the pejorative sense.

[Derk Pereboom]

I want to highlight and endorse an idea that Kip and Paul discuss in their posts. Daniel Dennett and Susan Wolf have cast the debate in terms of the sort of freedom or responsibility we would want, or the sort worth having. It’s then sometimes assumed that if what’s incompatible with determinism is only a sort of freedom or responsibility we wouldn’t want, then compatibilism is true with respect to any notion of freedom or responsibility that’s really significant. But as Paul says: “Maybe we should add that if there are any varieties worth fearing, we should examine those too.” The traditional problem of free will, moral responsibility, and determinism is whether determinism threatens the sort of control over our decisions and actions that makes it the case that we fundamentally deserve blame for the immoral ones, and praise for those that are morally exemplary. It’s a pervasive feature of human psychology that we assume that people deserve blame and praise in this sense. As far as I can tell, almost everyone is at virtually all times angry with some person or other because of what he or she has done, and this anger presupposes that this person fundamentally deserve blame. A large proportion of humanity assumes that policies for punishing criminals can legitimately be based on their fundamentally deserving blame. For this reason, regardless of whether this sort of responsibility is something we would want, whether it is compatible with determinism, or with the general causal nature of the world, is highly significant.

This gives rise to a worry about using the term ‘compatibilism’ in the way that Tom proposes. Let me first say that I don’t think that there is any important substantive difference between his position and mine. The difference is just terminological. I call my self a (hard) incompatibilist and he calls himself a compatibilist. Perhaps only substantive views, not labels, are worth discussing vigorously. But if says he advocates compatibilism, it will typically be assumed that he thinks that the sort of responsibility that has traditionally been at issue – which has, as I’ve just argued, a very prominent place in human psychology and policy – is compatible with the general causal nature of the world. This would be misleading. At the same time, there may be good reasons on the side of calling this position ‘compatibilist,’ (and, to my mind, revising the traditional terminology,) but there is also this cost.

[Tom Clark]

Kip: Turns out that Pinker is no longer a free will illusionist – he’s come out four square against the myth of the ghost in the machine in The Blank Slate, see esp. Ch 10, “The Fear of Determinism.” I have to update all my Pinker references to reflect this.

Derk: Thanks for pointing out my non-standard self-ascription of “compatibilist” – I certainly don’t want to be misleading. In The Problem of the Soul, Flanagan calls his position “neo-compatibilism,” saying: “I don’t intend by calling my view neo-compatibilism to suggest that my view is really new. It’s mainly a way of marking and acknowledging the fact that my view, just like the views that inspire me – those of Harry Frankfurt, Gary Watson, John Martin Fischer, Susan Wolf, and Daniel Dennett to name a few – does not in fact agree that the traditional [Cartesian] concept of free will is compatible with determinism. These fellow philosophers think, truly, that some suitably naturalized conception of human agency preserves some, but not all, of what is worth preserving in the traditional concept of free will, and that this concept is compatible with a causally well-behaved universe” (fn p. 127).

I’m taking the same line regarding MR: the compatibilist’s idea of MR simply can’t be what the libertarian is defending, or so it seems to me. It must be consequentialist, not desert-based, broadly speaking. So perhaps I’m best described as a neo-compatibilist.

I’ve been trying, with mixed success, to get those who call themselves compatibilists to put their cards on the table regarding actual responsibility practices and their rationale, e.g., retribution vs. consequentialism in criminal justice. This would illuminate exactly what sort of MR they have in mind and how the characteristics of moral agency they say exist motivate the responsibility practices they endorse. This would speak to your point that “perhaps only substantive views, not labels, are worth discussing vigorously.” But it’s also good to have it pointed out that compatibilism is generally taken to mean that “the sort of responsibility that has traditionally been at issue,” the sort that endorses full-fledged desert that underwrites retribution, is compatible with determinism/no LFW. I don’t see how such compatibility is possible.

[Robert Allen]

"But it’s also good to have it pointed out that compatibilism is generally taken to mean that 'the sort of responsibility that has traditionally been at issue', the sort that endorses full-fledged desert that underwrites retribution, is compatible with determinism/no LFW. I don’t see how such compatibility is possible."

Tom,

I'll tell you how it is possible- NOBODY, not even God, is controlling me. Yes, my actions are causally determined. And, yes, most of their causes, the ones that occurred before I developed the abilities that make up my will, are events over which I never had any control, but they do not CONTROL me. Only a person could do that. Thus, my will is free, making me responsible for what I do. Augustine made this point a long time ago and Keith Yandell nicely established a similar claim at the recent Wheaton conference. For the life of me, I can't understand why it does not resonate with some FW theorists.

[Tom Clark]

Robert,
Yes, I agree that normally no other agent is controlling you, and that therefore you usually have control over your actions and can legitimately be held responsible as a way of shaping your future behavior. The question remains, what kind of responsibility practices are justified by this sort of responsibility? Are you liable for retributive punishment, the kind of punishment that its defenders, such as Michael Moore in his book Placing Blame, claim needs no consequentialist rationale? If so, why? It’s only by answering this question that we can get clear about the actual content of the sort of moral responsibility you endorse.

[Robert Allen]

Tom,

'It' in my first sentence refers back to 'such compatibility' in the quote. So, yes, I believe in retribution up to and including damnation.

Honderich, Strawson the younger, and Pereboom assert that retribution require something that is impossible: that nothing but those things that a person himself has done are causes (deterministic or probabilistic) of his character. For my part, I cannot see why the existence of a causal connection between my vices and some event in the distant past would absolve me of blame for them unless it were an attempt by some person to control me.

[Christopher Knorr]

I think that the biggest problem is the lack of asking the proper questions.Questions or Problems are the catalyst for understanding. Some Q&P maybe percieved in the third-person as being universal in nature; But I postulate that Q&P are unique to an agent. The Following are a series of questions I think should be asked about "Free Will" in order to acheive the proper level of clarity and control over the issue to discuss and move the issue forward.

1.Why do I need to know what "Free Will" is ?
My simple answer is for control of physical and mental process' and survival.

2.What is "Free Will"? (( Free = Choice) ( Will = ?))
(I think we should'nt get the concept of free and freedom mixed up because I interprete them as having important differences.I Define Freedom as being choice and understanding which can lead to progress, but thats a whole different arguement.)
After Question (2) The Concept of 'Will' seemed more than a little important. So I asked another series of questions.

Is Free Will what you 'want' to choose?
(want is a fully intellectualized concept)

Is Free will what you 'can' choose?
('can' is a concept that made me inquire how potential actions must be based upon the agents collective knowledge.) How an agent knows and what an agent knows can be better understood through a truthful and functioning theory of the mind. How an agent acts is judged in a historical context; And currently social judgements do not coherently account for the factors an agent take into account when physical executing mental process. This problem is directly related to the lack of understanding of how an agents mind can work.

Is Free Will what you 'feel' like choosing?
('Feel' is a Non-Intellectual concept)

Is Free Will What you 'have' to choose?
(Having to choose something may sound like its not a free will question, but I could best describe this as social perceptions as interpreted from a first-person perspective. These can be controlling ideas that if an agent does not understand they can control an agents physical an mental processes. The Degree of control is relative to the agent but can be largely based on the magnitude to which a agent does not understand and the agents interpreted relavence of the controlling idea to the agents 'life'. From the FPP I think this would be perceived as stress. An example of a controlling idea would be the concept of Time.

3.How can I apply a working knowledge of Free Will to my life, and should other people apply it to there lives?
I think we should Answer 1 & 2 first.

I see most philosophical arguements as being stuck in one of the questions I just asked, but not seing the other questions clearly or even seeing the other questions at all. And I see a truthful concept of free will as being a unifying answer to questions 1-3.

[Tom Clark]

Robert,
My question is, what is it about being a no LFW moral agent that justifies punishment that need have no forward-looking, behavior-guiding purpose? We agree that moral agents are just those rational agents operating without any other agent being in control of their behavior. But what characteristic of such an agent justifies punishment, up to and including damnation, if such punishment serves no purpose other than inflicting suffering upon the agent? That’s what retribution is: the infliction of suffering simply because someone *deserves* to suffer, not for any future benefit. Why should an agent, who is not self-originated and who could not have done otherwise as the situation arose, suffer if that serves no end other than the infliction of suffering?

[Robert Allen]

Tom,

If they are moral agents who have made others suffer, e.g., capitalists, who deliberately steal their poor workers' time, why shouldn't they suffer in turn? Wouldn't it be an injustice if someone in a position to make them suffer refused to do so? No one else made a capitalist become one; he/she steals of his/her own free will (in my sense, which makes sense, unlike LFW). I gotta tell ya, I look forward to the reading of the story out of St. Luke's Gospel about Lazuras and the rich man. Life would seem truly absurd without the thought that, as Dylan put it, "if God is looking down from heaven (evildoers) will someday get what they deserve." Putting 'deserve' in quotes suggests that you do not believe that the agents in ? really deserve to suffer. But you have agreed that they are moral agents, given their ability to overcome their influences. What else do we need to justify retribution? LFW? But it is false. If my compatibilist definition is true and the true definition entails justification for retribution, then any evildoer to which it applies deserves to be punished. Or do you think that no definition of FW entails that evildoers deserve punishment?

BTW, if Dosteovsky is correct that punishment is redemptive, then it necessarily has a "forward-looking, behavior-guiding purpose."

[Tamler Sommers]

Tom, one thing that seems clear from your dispute with Robert--you should not be calling yourself a compatibilist. Robert: now THAT'S a compatibilist.

Getting back to the four questions. If the answer to (1) is no, then hard incompatibilists would answer (2) by saying that the "related thing (e.g. desert)" is the ONLY type of moral responsibility we don't have. Granted, it's an important kind of MR, but disagreements about desert are really all that seperates the MR skeptic and the compatibilist. If we answer (2) that way, then I think that the answer to (3) will vary from person to person. Some people can't bear the thought that their loved ones aren't deserving of praise, or that evil people are not deserving of blame. Others (like me) aren't bothered by thought at all. It seems to be largely a matter of temperament.

Finally, (4) seems to me to break down into two separate questions:

A) what should WE, as a society, do about the fact that we are not morally responsible in the robust sense (i.e. we are not deserving of blame or praise for our actions)?

And

(B) what should I, as an individual, do about this, even if society stays pretty much the way it is right now?

(A) gets most of the attention in the free will debate, but it seems that (B) is just as important, maybe more so. It’s certainly a more pressing question. It's not like there's a no-free-will-no-moral-responsibility party just waiting to make their move in the midterm elections. It'll be a long long time before American society embraces hard incompatibilist principles, if it ever does. But what can change significantly, right away, are the attitudes, behavior, and beliefs of the individual who comes to believe that there is no such thing as robust moral responsibility. And perhaps, the more we examine the implications for the individual, the more acceptable the position will come to appear for society at large.

[Robert Allen]

Tamler,

You made my day.

[Saul Smilansky]

Two quick clarifications on the way I see questions 1-4, and then I'll jump in with some comments. First, it is not only "legitimate" but a good idea to go sometimes "in reverse", say, ask whether compatibilist notions of moral responsibility are sufficient, and if one concludes that they are not, go on to consider whether we might have LFW. Second, there is every reason to think about these issues also outside of morality and justice, namely concerning self-respect or indeed the meaning of life. Still, the issue of (1) what sort of freedom we have (or can have), on the ontological level, (2) what this implies about notions such as responsibility, desert, blame and praise, self-respect etc., (3) whether we are happy or otherwise about these implications, and (4) what, if anything, we should do about it, will reappear. What this means is that (A) the free will problem is much more complex than was traditionally thought, and that (B) broadly pragmatic concerns play an important role in it.

I agree with Tamler that we need to ask question 4 on a number of levels (although I do think that compatibilist notions of desert (gasp!) - which are not utilitarian-like - are viable, and indeed crucial, although they are also shallow). The complexity again seems to me to be very large here. I am suspicious whether someone can really internalize the absence of LFW, take a hard determinist interpretation of this, and yet stay relaxed about what this means about his or her own achievements, appreciation of loved ones, and the like. I consider this in the section on the possibility for "Unillusioned Moral Individuals"(UMIs) and in other parts of my book. But even if someone can, it is not at all clear that he or she should "spread the word" to others, who might well become very reasonably depressed by all this. Even if real internalization is possible, fundamental competing values (peace of mind and self-respect versus integrity, knowledge and self-honesty) would clash here. And then there is the further question whether a social order where people have "determinism-based" excuses, and know that they have (and will have) them, can generate the sort of moral seriousness and moral behavior we better get. And so on. Of course Illusionism, broadly, the idea that the continued belief in the fantasy of LFW is basically positive, has reactionary potential, goes against the philosophical ethos, and is just demeaning and depressing. Yet this might well be our best alternative, rather than a Brave New World outside of the paradigm of concern for control, responsibility, and desert.

[Mark Smeltzer]

Dr. Pereboom,

You said earlier that, "One worry suggested by Mark’s comment is that we are still trying to sort out the concepts of free will and moral responsibility. In my view, we have considerable clarity as to which notion of responsibility is at issue." However, the idea I was trying to motivate in my first post was not that the conceptual roles filled by said terms are ambiguous -- I'm willing to grant that we know what these things ought to get us in practical terms. Instead, I was trying to motivate the idea that we have no consensus regarding the way these practical concepts ought to be fulfilled; there are many candidates, and many approaches.

If we were all being good experimental particularists we would have before us an assortment of paradigm cases that we could point to and agree, to some majority, that they represent, in concept, cases where agents exercise free will, and/or are morally responsible for their actions. However, what seems to be the case is that skeptics attack the starting points offered by experimental particularists from the epistemological methodist’s perspective – which, ultimately, is question begging. Then, from these cases we could attempt to determine what the necessary and sufficient requirements of FW and MR are. It is in this later phase of the discussion that distinctions between Compatibilists and Libertarians could arise, but no sooner. Moreover, on this line the skeptics really haven’t entered the discussion -- in this story the skeptics have chosen to abstain from the beginning by sticking to their arbitrary epistemological criteria.

My contention is that everyone doing FW/MR research ought to approach these topics as epistemological experimental particularists. I believe that FW and MR skepticism are default positions for those who (for whatever reason, or perhaps no reason at all) do otherwise (a group of which I would argue that yourself and Dr. Smilansky are members).

Dr. Smilansky,

I believe my comments above to be pertinent to this discussion because it seems to me that you are attempting to put forward a method for doing FW/MR philosophy. However, even if your project is not so grand my comments suffice as a criticism to your method, as it has been described thus far.

[Tom Clark]

Tamler,

I’ll take my argument with Robert off list, and will eschew the compatibilist label for myself. You say the disagreement between MR skeptics and compatibilists is about desert. I’m wondering if the following is a fair description: Both sides see LFW as hopeless, but hard-hearted compatibilists (as you once called them) want to retain the robust MR retributive mandate ordinarily bequeathed by LFW, and they attempt to do it by concentrating attention on the proximate control capacities of the agent, some of which end up recursively shaping character, a sort of proximate self-causation. Soft-hearted RMR skeptics such as ourselves point to the lack of ultimate control and lack of ultimate self-origination, and ask why anyone should be punished for being fully caused to be bad (recursively or not), if it serves no consequential purpose. For at least some compatibilists, it’s obvious that proximate control is sufficient for full-blown desert that might justify even death for wrong-doers, and for RMR skeptics it’s equally obvious that it isn’t. Does this capture the disagreement, or is there something I’ve left out?

[Tamler Sommers]

Tom,

I think it nicely captures the disagreement. Maybe some compatibilists think otherwise, I don't know. As a sociological observation, however, I don't know of any compatibilists who support the death penalty (i.e. who believe that certain criminals deserve to die). This may be because, like Eddy Nahmias, they see robust responsibility, or desert, as coming in degrees. And the degree to which an agent may be robustly responsible, according to the compatibilist, almost never reaches the point where he or she could deserve to die. (Although Robert Allen seems to think that an agent could deserve to suffer eternal damnation.)

[Eddy Nahmias]

As Tamler points out, I do believe that moral responsibility (including desert) come in degrees, primarily because the relevant cognitive capacities tied to free will come in degrees. And I tend to think the degree to which most people possess and have the opportunity to exercise those capacities--and hence the degree to which they are morally responsible--tends to be lower than we ordinarily presume (e.g., in criminal law cases). This has nothing to do with the fact that we are ultimately "part of the causal stream" (whether deterministic or indeterministic). Rather, it has everything to do with the fact that we are often subject to causal influences that limit our capacity to exercise the self-knowledge and self-control required for autonomous and responsible agency.

For a nice discussion of one example of such influence, see Philip Zimbardo's discussion at www.edge.org, where he describes his famous Stanford Prison experiment, relates it to the Abu Ghraib case, and explains how situational factors often "overwhelm" our dispositional traits (or, I would add, our capacity to control our actions in light of principles we would or have reflectively endorsed). Given Zimbardo's description, the prison guards should certainly be held less morally responsible and punished less severely than they are, but do they *deserve* *no* blame or punishment? Why think that? (rhetorical question only--we all know why some people think that!)

[Eddy Nahmias]

Please forgive my poor subject-verb agreement in the previous post. My former self (English high school teacher) is very upset with my current self.

Also, just a reminder (as program chair) that the deadline for submissions to this year's Society for Philosophy and Psychology (SPP) is February 1. For info go to:
http://www.hfac.uh.edu/cogsci/spp/wwwanlmt.htm.

[Paul Torek]

Tom, and maybe others too: I'm having a problem with the attempt to define free will (or even one aspect or dimension of it) as justifying retribution. My problem is that in the strictest sense of "retribution", I don't think anything could justify it. Regardless of whether we have LFW or not, it just isn't the case that the world is fundamentally better for the suffering of evildoers. (And in this moral viewpoint, I hardly think I'm alone.) Am I, then, a "free will skeptic"? I think that's an unfair characterization.

I also want to explore the idea of "punishment that need have no forward-looking, behavior-guiding purpose". Suppose I made a commitment to others, in hopes of motivating certain behavior on their part. Suppose it turns out, through my blameless ignorance, that I don't get the behavior I wanted. Still, I feel that it is my obligation to carry through on my commitment. Am I, in carrying it through, acting on a "forward-looking, behavior-guiding purpose"? Just taking the words at face value, I'd have to say no.

You can probably guess how this relates to desert ...

[Robert Allen]

"And the degree to which an agent may be robustly responsible, according to the compatibilist, almost never reaches the point where he or she could deserve to die. (Although Robert Allen seems to think that an agent could deserve to suffer eternal damnation.)"

For the record, I am adamantly opposed to the death penalty as a form of retribution. But how can you treat (say) the war criminals that make up the Despicable administration as morally responsible for the horrors that have taken place in Iraq over the last 2 years, as some Compatibilists do, and not think that they deserve to suffer in some way? Must a compatibilist maintain that it was morally inappropriate to punish those convicted at the Nuremberg trials, that they should only have been rehabilitated, if possible, or incarcerated, but comfortably, if not? Any definition of MR worthy of the name, it seems to this Compatibilist, should vindicate the common sense judgment that war criminals in being punished 'get what they deserve'. In fact, I would go a step further and say that it is psychologically impossible to rehabilitate someone without punishing them, as this is the only way to get them to realize what is essential here, viz., that they are blameworthy. Anything less conveys the wrong message. It's the same way with raising children.

Eternal damnation is a different matter, as it involves rejecting our loving Father's offer of mercy. (It's the Prodigal Son deciding to starve instead of returning home; it's St. Peter drowning instead of clinging to the Lord for all he's worth.) Now I cannot fathom how someone could do such a thing, which is why I subscribe to the Really Good News like Keith DeRose, but, if they did, what can you say?

[Mark Smeltzer]

Tamler,

I am a Compatibilist and I believe that sufficiently evil agents deserve the death penalty. However, I make a big deal out of the epistemic gap between our limited judging faculties and an agent's true character. I have argued that we are warranted in exacting punishments retributively if and only if those we have sufficient evidence that the agent does in fact deserve those punishments. Consequently, I would be very hesistant about dealing out death and judgement because the epistemic requirements are so high.

[Robert Allen]

"However, I make a big deal out of the epistemic gap between our limited judging faculties and an agent's true character."

Mark,

As James Baldwin put it, "you can see a man's fall, but not his wrestling."

[Tom Clark]

Eddy,

You say the prison guards *deserve* at least some blame and punishment for the failure to exercise self-control, which means they should be punished whether or not any future benefit accrues from the punishment. I can understand various functional justifications for sanctions, e.g., as an inducement for the guards to behave better next time, or to deter other guards from similar behavior, or to perhaps rehabilitate them, sequester them if they’re dangerous, or to provide restitution to their victims. But why punishment for its own sake?

If punishment is justified by agents having the “capacity to exercise …self-knowledge and self-control,” this immediately suggests (to me at any rate) consequentialism, since it’s the anticipation of punishment that prompts the exercise of such capacities to prevent future failures of self-control. But of course retribution and desert are non-consequentialist, so this can’t be your rationale.

Why do agents *deserve* punishment by virtue of having a capacity for self-control, irrespective of any consequences such punishment might have in shaping behavior, or other consequential benefits? What’s the link between blame and having this sort of control, if it isn’t the way the prospect of blame can modify the behavior of agents that have it?

In short, I don’t see that you’ve shown what the non-consequentialist rationale is for punishment on your account of moral agency, which is what’s necessary for compatibilist desert and retribution.

[Mark Smeltzer]

Tom,

I agree that the questioned posed to Eddy ("Why do agents *deserve* punishment by virtue of having a capacity for self-control, irrespective of any consequences such punishment might have in shaping behavior, or other consequential benefits?") is a very challenging one. In fact, my take on the issue is that ultimately desert and control are not conceptually linked in the way that responsibility and control are. Praise and blame are attitudes we hold regarding other agents in virtue of the things they do. These attitudes are warranted, on the compatibilist view, if and only if the agent was responsible for those actions – notice this is an epistemic criterion!

However, conceptually speaking, one cannot implicitly infer this principle an agent actually deserves punishment because of one’s negative attitudes toward that agent are epistemically warranted. (This is structurally analogous to arguing that one cannot assert that one knows proposition P simply because one’s belief in P is warranted; P must also be true.) Thus, I believe desert is deeper than responsibility. I believe that desert (and consequently the retributive punishment) is ontological in nature and derives its status according to the degree an agent-as-it-presently-is differs from the agent-as-it-ought-to-be (enter moral realism). Responsibility is merely an epistemic conduit that connects our moral judgments, attitudes, and retributive practices to that ontological reality – responsibility is the mechanism that grounds the possibility for knowledge of an agent’s character. However, due both to the epistemic gap between our perception of an agent's character and the agent's character as it is, and to factors in the physical world which can impede responsibility (manipulation, etc.); our judgments are often going to be incorrect. Consequently, extreme reactive attitudes and retributive practices ought to be handled with extreme care.

So, in the end, I agree with you that the historic notion which directly connects responsibility to retribution is flawed. I believe it really boils down to whether one can successfully maintain a rigid direct realism regarding agent-character-stuff. If we were constantly able to be directly presented with knowledge of an agent's character as it is, that would be sufficient to bind the two concepts. However, I don't see how that position could be supported even if one maintained a rigid direct realism regarding physical phenomena.

[Chris Franklin]

Mark,

You wrote, "These attitudes are warranted, on the compatibilist view, if and only if the agent was responsible for those actions – notice this is an epistemic criterion!" I agree with the first part of this statement. But I am not sure why I should take the force of "warranted" to be epistemological. It seems more appropriate to say one is correct, right, or justified morally in applying certain attittudes to an agent iff the agent was responsible for certain actions - where the attitudes arise on the basis of those "certain actions."

You also wrote, "Responsibility is merely an epistemic conduit that connects our moral judgments, attitudes, and retributive practices to that ontological reality – responsibility is the mechanism that grounds the possibility for knowledge of an agent’s character." Could cash this out for me a little more? How could responsibility *ground* the possibility for knowledge of an agent's character? This claim strikes me as odd. Surely the question, "is this agent responsible?" is different from "what is the character of this agent?" So in what way, if any, does an answer to the former ground the possibility of an answer to the latter?

Finally, I am doubtful that there is such a great gap between our perception of agent's character and the agent's character as it is. Surley one must practice care and discernment when making claims about an agent's character, but that is no reason to think there is some gap. As an extremely intelligent man once said, "By their fruit you will know them."

[Mark Smeltzer]

Chris,

Good comment. I suppose there's two ways to understand my view, and I hadn't intended that. There's the way that you mentioned which casts my view of responsibility and the validity of our attitudes towards others as essentially pessimistic; meaning that we're going to be wrong a lot of the time. That is not my view.

Rather, I am trying to stake out, fundamentally, what I believe to be a valid way of addressing the discord between the concepts of self-control and desert. I do place heavy epistemic requirements upon our right, as it were, to hold extreme reactive attitudes towards others and to enforce extreme levels of retributive punishment. However, that does not entail that I have to be pessimistic about our ability to naturally do so.

In fact, even though I am not a direct realist regarding knowledge of the characters of others (or at least see no reason to be), I do believe we have indirect access to a wealth of knowledge about the characters of others by reviewing what we know about the patterns people engage in, and the conclusions we can inductively justify based on those patterns. (If this sounds a lot like a type of virtue epistemology, that's because it is.)

The practical ramification is simply to stress the importance of maintaining a high burden of proof within the domain of law, and perhaps to rethink some of the habits we have developed for how we ought to feel about and respond to other people's actions on a personal level. With regard to this second aspect, it would simply mean stressing the role of reason and developing the fortitude to restrain oneself from holding certain attitudes about others when the requisite epistemic requirements have not been appropriately satisfied. So really, the practical ramification in both domains is the same: promoting the ideal of erring on the side of caution when judging others.

[Mark Smeltzer]

Chris,

Maybe I should have packaged these separate responses together, but since you raised two distinct issues in your post, I wanted to devote attention to them separately. Here I am going to address the questions you asked about how responsibility and knowledge character connect.

First off, the questions you raised were good starting points. I believe out ability to answer the question, "why kind of character does this person have?" is what ground our right to hold extreme reactive attitudes and engage in extreme retributive practices. If we cannot successfully answer that question on a routine basis with regard to ordinary people, then we really don't have a right to praise, blame, or punish anyone.

So, I believe that question has logical primacy over the question of individual responsibility. My claim is that we can only answer the first question by forming moral judgments that are based upon actions which function as expressions of an agent's character, and in order for actions to function in this way, they must be actions for which the agent was responsible -- enter the topics of reasons-responsiveness, self-control, self-knowledge, identification with one's desires, etc.

The epistemic connection between responsibility and desert is that for a particular action X, if agent A cannot reasonably be held responsible for X, then we are not warranted in believing that A deserves reactive attitude R or retributive punishment P with respect to X. The ontological connection between responsibility and desert is that for a particular action X, agent A deserves reactive attitude R or retributive punishment P if and only if A actually was responsible for X.

While the epistemic connection between desert and character is also linked to responsibility, I do not believe the same is true for the ontological connection: agent A deserves that reactive attitude R or retributive punishment P at time T which corresponds to the quality of A's character at T. Thus, responsibility is the mechanism through which how we come to know an agent's character.

[Mark Smeltzer]

One point of clarification: the point I am trying to motivate is that there is a distinction between desert according to an agent's ontological moral properties and desert according to what we can know about an agent's ontological moral properties because of the knowledge grounding function of responsibility. The benefit of this distinction is that it allows us to clearly connect (and separate) the concepts of responsibility and desert in an intuitive and straight forward manner.

I would be interested to hear what other gardeners think about this distinction and its validity.

[Saul Smilansky]

Mark,

I am not sure that my comments will reflect sufficient understanding of your views, so correct me if I am mistaken. But, if I understand correctly, then I disagree with much of what you say (BTW, I don't think that there is a big methodological divide here, most of us presumably work by some sort of "reflective equilibrium" that combines examining particular cases and theory construction, in interaction).

I think that the moral paradigm in the free will problem connects control-moral responsibility-blameworthiness-desert-(e.g.)punishment. Because we believe (if/when we do) that people have contol over their actions, they are morally responsible for them, hence when they act in wrong ways, they can be blameworthy for so acting, and thus deserve e.g. punishment. Incompatibilists are more demanding and require LFW to fill "control" at the beginning, while compatibilists are less demanding and think compatibilist versions of control suffice.

If this is accepted, then desert is more a description of what can emerge when there is control and moral responsibility than an independent factor (which is why, if we assume the absence of LFW, even though I see the force of hard determinism, I still see some possibility of talking about a shallower, compatibilist notion of desert - namely, one that emerges because people can still be more or less responsible even if they lack LFW).

Character is even less of a player here, IMHO. We blame, praise, punish an so on for choices and actions. If a wrong action is out of character or in character would usually be minor (it is not even clear which is worse). And since character is not (for me) where most of the action is, the epistemic doubts also do not focus there, and perhaps become less pressing.

So, in my view "all" we need to debate (assuming again the absence of LFW) is the moral (and personal) force of compatibilist control and moral responsibility, namely whether it is fair to blame and punish people who although they did what they wanted, ultimately did not have control over the sources of their motivation. Here opinions of course vary (I myself think that we ought to combine the insights of both compatibilism and hard determinism). And, finally, whatever we end up with, we then need to ask (particularly if we are not too happy about the best compatibilists can offer) what we should do with our results.

[Mark Smeltzer]

Here is a perfect (yet horrific) real life case where it seems that it would not be hard to make a case that extreme negative reactive attitudes are warranted.

http://www.local6.com/news/4161751/detail.html

[Mark Smeltzer]

Dr. Smilansky,

There's really only one thing I want to respond to right now, and it is this quote, "BTW, I don't think that there is a big methodological divide here, most of us presumably work by some sort of 'reflective equilibrium' that combines examining particular cases and theory construction, in interaction."

Unless one begins with particular cases that one believes represent the ideas one is studying, then one is logically confined to being a skeptic regarding those things. I'm sure that Descartes thought his method was valid, and frankly I can't find a substantial distinction between his epistemological methodology and your own. If one only accepts clear and distinct cases as valid starting points for creating a realist conception of free will, responsibility, desert, etc., then is all but inevitable that one will end up a skeptic.

Free will skeptics such as yourself and Dr. Pereboom who seem to accept a realist conception of moral values have set the epistemic bar too high, and this is the direct result the methodological paradigm you subscribe too.

[Ignacio Prado]

I have responded to this question, here: http://thewebofbelief.blogspot.com/2005/03/what-free-will-problem-is-about.html