Smilansky, Van Inwagen, and You

Once again, the Indomitable GFP Reading Group of Destiny returns with another action-packed episode, with enough chills, thrills, and spills to keep you bathed in the glow of your monitor just a bit longer. In this episode, the Hero of Haifa - a mild mannered professor usually known as Saul Smilansky-  swings into action to test the the Sultan of South Bend, his excellency Peter van Inwagen. The target? The Sultan's forthcoming edict: "How to Think About the Problem of Free Will."

Will the edict stand? Will the Hero prevail? Who will the people favor? Dear reader, only YOU can decide.

Saul's comments begin below . . . .

(Thanks to Saul and Peter)
---------------
Peter van Inwagen’s paper “How to Think about the Problem of Free Will” is written with his usual admirable clarity, force, and scope. Anyone working or thinking seriously about the free will problem, whatever his or her position, will benefit from reading this challenging paper. It can help us to open up a meta-free will debate, about what it is valuable to discuss at all, and how. And the stakes are high: Peter calls for no less than a radical change in the whole direction of the contemporary free will debate. The fact that this proposed transformation is traditionalist, wishing to take us back to the golden past, should not hide from us the truly revolutionary nature of this paper. If Peter is right, many (and perhaps most) of us have been wasting our time, while if he were to be widely followed and is mistaken about the debate, as I believe that he is, this would set back the philosophical investigation of the free will issue enormously.

Peter does a number of things in his paper, and I will not be able to address them all. After briefly presenting some of his main contentions, I shall make a preliminary point, and then sketch a very different construal of what the free will problem is about.

Van Inwagen’s points (that I will discuss)

Broadly, Peter sees the debate in the following way (and I urge you to read his paper and not settle for my brief outline). There are (a) seemingly unanswerable arguments for the incompatibility of free will with determinism, and (b) seemingly unanswerable arguments for the incompatibility of free will with indeterminism. If all those arguments are indeed unanswerable, there is no free will. But (c) there are also seemingly unanswerable arguments for the dependency of moral responsibility on free will. It is, however, evident that moral responsibility does exist. But all this cannot be true together: it cannot be the case that there is moral responsibility, which is dependent upon free will, while there is no free will. So, at least one of the claims made above must be false. Given that moral responsibility is evident, either (a), (b), or (c) must be denied. The free will problem is about deciding which one.  (pp.1-2)

Free will, according to PVI, is the “power or ability to do otherwise than one in fact does”. And “compatibilists and incompatibilists mean the same thing by ‘able’”. Therefore, “’free will’, ‘incomatibilist free will’, ‘compatibilist free will’ and ‘libertarian free will’ are four names for one and the same thing”. (p.8)

Much of what goes on in the free will debate is, therefore, a mere pernicious confusion, “led astray by bad terminology and confused ideas” (p.14). For example, “It’s simply not true that there are two distinct things, libertarian free will and compatibilist free will, and that libertarians want the one and don’t regard the other as worth having” (p.12). It is therefore not true that libertarians want something else than compatibilists or soft-determinists. “We present-day incompatibilists see the free will that compatibilists believe in as the genuine article; their only mistake, in our view, is to suppose that it, the genuine article, is compatible with determinism” (p.12). In sum, “There is only one variety of free will worth wanting, because there is only one variety of free will: the ability to do otherwise” (p.13). The way to be a compatibilist is to follow David Lewis and discuss the "Consequence Argument" (p.19).

A preliminary point

How should we understand the issue of free will? It seems to me that we should try to formulate the issue so that what are widely accepted as major contributions in the contemporary debate are not excluded outright. Peter’s suggestion clearly doesn’t meet this guidance. Perhaps the two figures that have been most influential in the contemporary debate are Harry Frankfurt and P.F. Strawson. A major insight of Frankfurt was that it seems that people can have free will and be morally responsible, even in situations where they did not have the power or ability to do anything except the very thing that they did do. That is the intuitive force of “Frankfurt-type” examples, namely, that “ability to do otherwise” is NOT required for free will and moral responsibility. This position is of course (and reasonably) contentious. But it is excluded by definition in Peter’s very formulation of free will (see the previous two paragraphs). P.F. Strawson’s contribution to the free will debate was made in terms of the reactive attitudes and of the justificatory force that they have. It seems that for Strawson's "scepticism-proof" neo-Humeanism there is even less room under the constraining guidelines for the debate that we are being offered. It might be countered that it IS after all possible that even such central positions will be completely mistaken, and perhaps Frankfurt, Strawson and the many philosophers who they have influenced are talking about the free will equivalent of the “ether”, and not about reality. But a very strong case has to be made in defence of such a claim, and I don’t see it. In sum, it is not clear why we should construe the issues in such very narrow and exclusionary ways. On the other hand, we immediately see how much that is interesting and fruitful would be lost to the debate, if we adopted Peter’s Draconian constraints. When I shortly set out in outline my own alternative understanding, we will (I think) see further reasons for not accepting these constraints.

A very different way to see the free will problem (and hence what the debate should be like)

The quickest way in which I can sketch an alternative to Peter’s way of seeing things, and show the attractions of this alternative, is to take up the challenge at its critical points. I will sketch the way I see compatibilism, and then show that the notion of “compatibilist free will” is relatively clear, that it is clearly something very different from "libertarian free will", and that what separates the different positions on the free will debate follows from these differences.

COMPATIBILISM is the view that even if everything is determined people can have the sort of control over their actions that suffices for moral responsibility (and concomitant notions).

The “suffices” here is ambiguous, hence it is useful to distinguish between

MODERATE COMPATIBILISTS, who believe that compatibilist control or free will suffices for SOME measure of moral responsibility (but perhaps not for all significant forms or levels of moral responsibility), and EXTREME COMPATIBILISTS, who think that compatibilist control or free will suffices for ALL the worthwhile forms or levels of moral responsibility. Such a moderate compatibilist might hence also be at the same time (gasp!) partly a hard determinist, and think that there are also important forms or levels of moral responsibility (and desert, and moral worth, etc.) that are impossible under determinism, and hence that determinism does matter greatly. I am in fact such a moderate compatibilist and a partial (or moderate) hard determinist, which adds up into making me a “dualist” on the compatibility question. I think that there are at least two and very distinct varieties of free will in some sense worth wanting, the compatibilist and the libertarian. In my book I sought to combine the partial but valid insights both of compatibilism and of the type of hard determinism which morns the absence of libertarian free will. 

Clearly all this is grossly muddled, and disallowed, according to van Inwagen’s way of understanding the problem. But while I may well be mistaken, I don’t see that my way of speaking lies outside of the free will debate, nor that is it trivially mistaken. And (if I may say so), it also seems fruitful for those who have other views on free will to think about such an alternative position. If this is so, Peter’s advice about how to understand the issues should not be followed.

COMPATIBILIST FREE WILL is the capacity and ability people have (even under determinism)   to recognize options, to understand and evaluate reasons, to reflect critically upon their desires, to form their choices in accordance with their reasons and desires, and to act effectively upon their choices.
[or something like that.]

In another way, compatibilist free will is the sort of control that compatibilists investigate. But why do compatibilists think that we can have moral responsibility even under determinism? Clearly, because they are less demanding: they settle for such lowly forms of local control that exist even under determinism. They think compatibilist free will suffices.

Arguably, most people most of the time have compatibilist free will. When people perceive new options, reason better, reflect upon their desires, enhance their self-control, increase their power to pursue their choices, and so on, they (I claim) are increasing their free will (and capacity for behaving responsibly). People who do not have such pedestrian, compatibilist, free will – because they are under the power of overwhelming compulsions and fears, say – ARE MUCH LESS FREE. The kleptomaniac or alcoholic, just as the person who is afraid to leave her house, or must constantly wash her hands, or has hallucinations she cannot distinguish from reality, are not free in the sense that most people are. And – this is very important – these people would typically wish to be released from the grasp of the forces harming their free will.

But notice that the difference between the free and unfree here has NOTHING to do with determinism. The difference between such examples of lack of freedom and normal, more or less free human beings is not that the one group is determined and the other not (if we assume here that determinism rules all human actions). What make the crucial difference are the type of deterministic processes that are going on, and their impact upon compatibilist control, rather than the mere fact of determinism. I think that if Peter himself were to become convinced of determinism he would, on reflection, still see that some unfortunate people have much less free will (reflective control over their actions) than others, on account of such compulsions (and the other factors discussed by compatibilists). He might admit that they have good reason to wish to become more free, in compatibilist terms. And he might even think that some (those who have compatibilist free will) might take part in a Community of Responsibility while others (whose level of compatibilist control is very low) cannot. That being as it may, such considerations show that compatibilist free will is both a clear and an important notion, and (more controversially) that most people most of the time have quite a large measure of it.

What, then, is in contention? Some determinists (the extreme hard determinists) and libertarians put the bar of freedom and moral responsibility much higher: such pedestrian forms of control as we have considered do not suffice for genuine moral responsibility, they claim. They are more demanding, THEY WANT MORE. There is some truth here, and compatibilist free will indeed does not give us everything, morally and personally, that we thought we could have, and that it matters that we don't have (recall that I am also a partial hard determinist). Yet, I also think that such extreme incompatibilists are too extreme when they completely deny the importance of commonplace compatibilist free will. Recall the kleptomaniac and alcoholic. But that is a big issue that I cannot argue here further. What I believe that everyone should recognize, pace van Inwagen, is that those who seek (or mourn the absence of) something beyond compatibilist free will, are seeking (or morning the absence of) something ELSE. This, properly called libertarian free will, is some form of transcendent ability to control one’s actions, which goes beyond anything that is possible within a deterministic world. It certainly goes beyond the models of free will offered by compatibilists. I am not the right person to try to make sense of this ambitious notion, but some such thing is what libertarians try to describe, and in doing so they go beyond the commonplace compatibilist forms of control. This other thing is also what libertarians WANT, in the precise sense that they think that moral responsibility requires it, and hence they hope that it exists – again, distinctly apart from the pedestrian compatibilist forms of control. In order to explain this very different thing libertarians utilize indeterminism (which is of no use whatsoever to compatibilists explicating compatibilist free will), or postulate a new form of (“agent”) causation. 

There is hence, as far as I can see, no problem in making sense of the notion of “compatibilist free will”, and in seeing that it is a very different and much more conventional thing than “libertarian free will”. Similarly, it makes perfectly good sense (and even happens to be true) to say that (extreme) compatibilists do not see why anyone would want libertarian free will (they are happy with mere reasons-responsiveness etc.), while libertarians do want it – otherwise they would be content to lie alongside the lion of determinism, as the compatibilists are. 

To go back to the beginning of Peter’s paper, what, then, is the free will problem about? As I have written in the Garden before, I have found it helpful to think about it as a combination of four very different BIG 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 at all 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 we can have moral responsibility without LFW (rather than with complete determinism).
3. If we do not have moral responsibility (etc.) because of the absence of LFW, or if MR is at least seriously harmed by the absence of LFW, is this a good or a bad thing?
4. What can and should we do about the conclusions to questions 1-3?

In our daily lives as free will investigators, many specific questions will concern us: How can compatibilist free will be enhanced (irrespective of LFW)? How important are things like rationality, or identification, or second-order reflection, or tracking values, or self-expression, within compatibilist free will? If there is no LFW (so that all we can have are pedestrian compatibilist forms of control), how deep can compatibilist notions of free will, desert, justice, self-respect or moral worth be? How much do we need to revise our view of free will and moral responsibility, if there is no LFW? If there is no LFW, is there a difference between the viability of the notions of free will, ability to do otherwise, moral responsibility, blameworthiness, praiseworthiness, self-respect and desert (i.e., can we have some but not others)? To what extent ought our decisions about the degree of compatibilist free will and moral responsibility to be context-dependent? Or affected by contractual or utilitarian considerations? How can we combine moderate compatibilism and moderate hard determinism? What sort of free will do people believe in, and how much can such beliefs (and related reactions and practices) change? What would be the effects of the decline of belief in libertarian free will, and should we encourage it? These are just a few quick examples of the many fruitful questions that we have been discovering relatively recently. It makes perfectly good sense to speak in such mostly van Inwagen-forbidden ways, and I think that doing so is the key to making progress on the free will problem.

Comments

[Neal]

Thanks, Saul and Peter, for your contributions. Here are some thoughts of my own:

I feel like there is no real disagreement here. Van Inwagen outlines one really interesting puzzle, but it is not the only puzzle that we are interested in.

We are interested in a number of important questions, all of which can be stated without the term 'free will'. Van Inwagen understands the phrase 'Neal has free will' roughly as follows -- 'Sometimes, with respect to some of his actions, Neal is both able to perform the action and also able to refrain from performing the action.' Call this power that I supposedly have 'dual-ability'. Now, here are the questions that van Inwagen is interested in:

1) Is determinism compatible with dual-ability?
2) Is indeterminism compatible with dual-ability?
3) Does moral responsibility require dual-ability?

These are all, I think, great questions and they deserve to be studied. But they aren't the only questions that deserve to be studied. Here are some others:

4) Is dual-ability sufficient for moral responsibility?
5) Is determinism compatible with moral responsibility?
6) Is indeterminism compatible with moral responsibility?
7) Is reasons-responsiveness sufficient for moral responsibility?
8) Does moral responsibility require agent-causation?

And there are many more interesting questions besides.

I don't think that van Inwagen denies that there are interesting questions other than (1), (2), and (3), however. I see his project in this paper as an attempt to do some housekeeping and make sure that our terminology isn't getting way out of control. His suggestion is that we use the term 'free will' to mean dual-ability, and that we use the terms 'compatibilism' and 'incompatibilism' to refer to a 'yes' answer to question (1) and a 'no' answer to question (1), respectively. I for one think this is a pretty good idea -- or, at least, I applaud the housekeeping effort. (After all, isn't it slightly annoying that these days one can't use a term like 'compatibilism' without specifying *which* propositions one thinks are compatible? The term has become ambiguous.)

Of course, the terminology has gotten so out of control that I'm not optimistic about the project of tidying things up. But it's a noble effort. And I don't think anything of great consequence hangs on which way we do the tidying -- whether we use the term 'the free will problem' to denote the problem van Inwagen outlines or something else -- just so long as we try to do it.

P.S. I'm also in favor of eliminating terms like 'compatibilist free will', even if defined as Saul does in this post. Here's one reason -- according to Saul's definitions here, someone can both believe that compatibilism is false and that we all nevertheless enjoy compatibilist free will. (!)

[Alan]

I concur that van Inwagen's argument is important--but I concur with Smilansky as well that its importance lies in its challenge, and a challenge to be met with vigorous resistance. For if we surrender to van Inwagen's eloquent and sheparding argument, the free will problem will be surrendered to the ground of the incompatibilist.

Traditionally, the free will problem, as reflected in the dilemma of determinism, is a function of two logically distinct but relatable arguments or issues, both compliant with the law of excluded middle by definition.

One is a metaphysical one about the way minds actually work, the debate about determinism versus indeterminism of human nature. (Never mind that the debate often discourses, perhaps irrelevantly, about absolute determinism ala the consequence argument--although the relevant issue is potentially narrower, about human minds and human action--that is, it is about freedom in that sphere alone irrespective of what is arguably outside goings-on. The consequence argument is of great moment if absolute determinism settles the problem--but it means nothing if absolute determinism is false. And quantum physics seems to say that one level of existence does not involve determinism. The alternative is to constrain the consequence argument to human nature and human actions, but it's not clear that it could be effectively reformulated with such a restriction, since that domain of restriction is the original argument's very point.)

The other issue is the conceptual problem of what "freedom" means--whether freedom is consistent or inconsistent with determinism (of the human mind). But such a problem must leave open questions about (for example) whether relevant freedom abilities require alternative opportunities or do not so require them--essentially, whether (something like) Frankfurt's view or (something like) van Inwagen's view is right about what "freedom" means.

van Inwagen's paper effectively attempts to eliminate compatibilists' proposals that "freedom" is meaningful without some sort of real alternatives. But that is, in my view, an attempt to eliminate a legitimate one-half of the free will problem!

[Kip Werking]

It’s fitting that Smilansky is reviewing Van Inwagen’s article here, because he is the one philosopher I know who most uses the term “libertarian free will.” This is a practice that I have criticized, too, but for very different reasons (roughly: Van Inwagen thinks that there is no libertarian free will distinct from plain vanilla free will; I disagree, there are at least two different kinds of free will here. But Smilansky means something like “magic free will”, not libertarian free will, because indeterminism cannot help—but “magic free will” just doesn’t have the same ring to it).

[First, a note about Smilansky’s comments here: I don’t think Frankfurt or Strawson’s contributions are so essential that they cannot be relegated to the dustbin of history. I think Strawson is just wrong and Frankfurt, if he is right about alternative possibilities, does not show that in his famous paper.]

I think Van Inwagen makes some ambitious or controversial claims which receive less skepticism than they deserve, because he is otherwise so careful and precise. Here are some claims I wonder about:

1. That the existence of moral responsibility is obvious in a way that the existence of free will is not.
2. That there is only one sense of free will and therefore only one sense of ability.
3. That both libertarians and compatibilists are both referring to the same thing.
4. That amongst philosophers who specialize in free will, there are more incompatibilists than compatibilists.
5. That there are more incompatibilist specialists than compatibilist specialists because of the Consequence Argument.
6. That most non-realists find the non-existence of free will to be regrettable (Pereboom, in criticizing Dennett’s “varieties of free will worth wanting” move, certainly leaves open the possibility that it is not regrettable. And, even if the impossibilists like G. Strawson and Smilansky find it regrettable, one wonders how much sense it makes to regret not having an impossible power. As Joe Campbell said “if nobody can have it, I don’t want it.”)

Regarding 1, Smilansky presses Van Inwagen on this point in his 1990 Analysis article, available on his website. Note that, in this Reading Group article, Van Inwagen does not say much about moral responsibility, and Smilansky noted, seventeen years ago, that Van Inwagen was not saying much about moral responsibility then either. In this Reading Group article, Van Inwagen writes: “It is, however, indisputably true that people have sometimes done what they ought not to have done.” But how can it be indisputably true when people dispute it every day—and not just anybody, but intelligent people, in prominent philosophy departments, disputing just this indisputable view at length in published books?

I agree with Smilansky wholeheartedly about 3. Van Inwagen offers his own definition for free will: the simultaneous ability to X and not X. But is this right? Robert Kane, for example, defines free will as the ability to be the originator and sustainer of one’s own ends and purposes (a definition I prefer, but perhaps for the wrong reasons).

Given this controversy, how does Van Inwagen defend his definition? He suggests it will not favor compatibilism or incompatibilism—which is not the same thing as giving sufficient support for it as a definition (many wrong definitions might be silent on the question of compatibilism). And he notes that David Lewis used the same definitions in one paper. And I don’t think that would be sufficient either.

I’ve heard Van Inwagen make what I think is a stronger argument: that philosophers used to always use this definition and the new definitions (e.g. free will as the freedom relevant condition for moral responsibility) are just that—new developments. I have two worries here: the first is, supposing, for a moment, that the definition, as philosophers use it, has evolved in recent decades, why should we privilege the definitions from the 70s over the definitions of the 90s? How long has the definition been evolving in this way? Is the definition of the 70s different than the definition from the 50s? And if it is, should we still privilege the definition from the 70s?

My second worry is that: if a word comes to mean something else than it used to mean, how fair or correct is it to cling to the old definition? There is the danger that someone will end up using the word in a (currently) confusing and idiosyncratic way.

This brings me to a related point: even if I grant Van Inwagen’s definition of free will, I think it admits of more senses than he seems to. Clearly, there is the incompatibilist sense in which nobody has the ability to do other than they do. This is the sense in which a piano player, next to a piano, who decides—on a whim—not to play did not have the ability to play. But there is also, obviously, the other sense in which the piano player, just because he is a piano player, had the ability to play, and did not exercise it. And language like “ability to play” is how we distinguish piano players from non-piano players—a distinction which is important even in deterministic worlds.

Indeed, Van Inwagen seems to be looking at the possibility that there are multiple senses of ability, and multiple senses of free will, with a jaundiced eye (perhaps to make the problem more interesting than it really is?).

Van Inwagen points to things like pain, and God, to suggest that compatibilists and incompatibilists are all talking about the same thing. But, it seems to me, free will is relevantly different than all of these things. At least, Van Inwagen has not shown that they are relevantly similar. It is not sufficient—it is far from sufficient—to show that people were mistaken to think that pain or God referred to different things, in order to demonstrate the same about free will.

Some scattered comments:

1. I think it’s fascinating that Van Inwagen wishes compatibilism was true.
2. Van Inwagen says that his argument for the claim “moral responsibility requires the ability to do otherwise” does not require that moral judgments have truth values. This suggestion is intriguing and I would like to see the argument, so formulated.

Finally, I think it’s ironic that Van Inwagen seems to have the best of intentions but recommends courses of action that, I think, will lead philosophers backwards and not forwards. His project is completely opposed to the one I described in, for example, my Pereboom on the Weirdness of Compatibilism thread (I would be interested to know what Van Inwagen thinks of Pereboom’s argument there). Van Inwagen’s intentions are good: he thinks “we have first raised a dust, and then complain we cannot see” and would like to clear our vision. I suspect free will is more like a mirage than a sandstorm: we think we see something clearly, but the more we try to grasp it, the more we realize there’s just nothing there.

[Joe]

Thanks for your commentary on Peter’s paper, Saul!

And thanks to Peter and to the folks at the Garden for organizing this wonderful exchange of ideas!

Let me admit, first, that I am bit of a van Inwagen disciple. His 1983 book sets the standard for philosophical discussion of the topic, as far as I’m concerned. Van Inwagen’s work marks the pinnacle of the Chisholm/Taylor/Lehrer way of looking at free will, which informed me in my formidable years. Since I’m an unabashed compatibilist, I disagree with Alan’s assessment that van Inwagen’s project prejudices incompatibilism.

One might say that it prejudices the traditional, or leeway, view of free will, which stipulates that free will requires alternative possibilities of action. Certainly, on the face of it source theories get little recognition in van Inwagen’s taxonomy.

But we have to be very careful here. It is not as if source concepts -- like ultimacy, or origination, or (to use a very nice phrase of Saul’s) up-to-usness -- are ignored by van Inwagen. Note that the English version of the Consequence Argument uses the phrase ‘up to us’ (1983, 16). It is only later, well into the formal versions of the Consequence Argument, when the traditional analysis plays a role. Even the N-operator is initially translated with the phrase “no one has, or ever had, any choice about whether” (1989, 404), and the term ‘choice’ is neutral in the traditional/source debate. Lastly, van Inwagen closes his discussion of the Consequence Argument responding to the objection that he privileges a particular conception of free will. He merely redefines the N-operator: “p and, in just the sense of having a choice that is relevant in debates about moral responsibility, no one has, or ever had, any choice about whether p” (1983, 104).

Van Inwagen believes that sourcehood and alternative possibilities of action are both necessary for free will. It is a huge mistake to think that he is merely talking about an outdated concept of free will, e.g., the traditional concept. (I don’t think that Saul interprets Peter in this way but I think that readers of Saul’s post might.)

Saul writes: “COMPATIBILIST FREE WILL is the capacity and ability people have (even under determinism) to recognize options, to understand and evaluate reasons, to reflect critically upon their desires, to form their choices in accordance with their reasons and desires, and to act effectively upon their choices.”

Later he writes: “In another way, compatibilist free will is the sort of control that compatibilists investigate. But why do compatibilists think that we can have moral responsibility even under determinism? Clearly, because they are less demanding: they settle for such lowly forms of local control that exist even under determinism. They think compatibilist free will suffices.”

And still late he writes: “In order to explain this very different thing libertarians utilize indeterminism (which is of no use whatsoever to compatibilists explicating compatibilist free will), or postulate a new form of (‘agent’) causation.”

Why do contemporary philosophers always forget to take account of compatibilist theories of agent causation? There is Kant, Davidson, Ned Markosian, Michael Nelson, and my colleague Harry Silverstein, to name just a few.

I am very firm in holding that compatibilists, in general, have not compromised free will in any way. Show me a positive way of thinking of free will and I will show you a compatibilist who tries to account for it. Of course, I mean to dismiss negative, question begging requirements like the falsity of determinism. I’m talking about positive concepts like sourcehood, alternative possibilities of action, agent causation, etc. I’m not ignoring that in each case incompatibilists have argued that the compatibilist notion is faulty. That is something different. (This is precisely why I think that the action in the debate remains with arguments for incompatibilism.)

Let me close with a challenge to Saul, or anyone else, to provide a description of ‘libertarian free will’ that is indeed distinct from what I, or some other, compatibilist might offer. Saul writes:

“This, properly called libertarian free will, is some form of transcendent ability to control one’s actions, which goes beyond anything that is possible within a deterministic world. It certainly goes beyond the models of free will offered by compatibilists.”

By way of a positive description of what we might be missing should determinism prove to be true, this passage doesn’t say a lot. Certainly the passage advocates incompatibilism but I’m not sure what we get by doing so.

[Alan]

Joe--

Thanks for your clear and thoughtful post. Actually, however, I don't think you and I are on opposite sides of the fence at all with respect to either van Inwagen's claims or Smilansky's post.

I completely agree that compatibilists can deliver on just about any promise that a sufficiently psychologically adequate and morally sturdy concept of FW should be equipped to make. And part of such a promise must be about alternatives in *some* sense--in at least the weak sense of logical/epistemic possibilities of thought. But as I read it, van Inwagen's proposal is to dig deeper and require that the promise must be about something stronger--metaphysical or real possibilities that inhere in the real world and make-up of the mind--a "power or ability to do otherwise that one in fact does." As Saul (if I may be so familiar) appears to conclude as well, this seems to mean something that conflicts with any account of such a power or ability that entails determinism (not just one that is compatible with it, for it might be compatible with indeterminism as well). So *strictly* determinist compatibilist accounts of FW would be put out in the conceptual cold.

[Mark]

Alan,

Are you familiar with Vihvelin's position? She's a classical compatibilist who defends the idea that a "power or ability to do otherwise that one in fact does" is necessary for free will and that such a power or ability is compatible with determinism.

There was a post here on the Garden discussing her view a few months back. Here's the link.

[Alan]

Mark--

Thanks and yes, I did follow the discussion then (though admittedly not as closely as I might have). :-)

van Iwagen in his paper treads, as he usually does, very very carefully. (I have a lot to learn from him in that respect.) On p. 8 he very carefully distinguishes property-ascriptions of "on some occasions able to do otherwise" from the power/ability-ascriptions "the power or ability to do otherwise than what one in fact does". The former property is a generic, blanket one that encompasses all compatibilist accounts (whether they entail determinism or not) and incompatibilist ones as well, but it's not clear to me that the second ascription, with its added proviso of "what one in fact does" is one that compatibilist concepts welcome or at least easily accommodate. (I'm not at all sure Vihvelin would welcome it, for instance, though I'm not positive.) For that proviso is backward-looking or at least counterfactual in some strong sense that would seemingly force the compatibilist into challenging determinism head-on. Note that van Inwagen's very considered use of the tenseless word "does" here, instead of "did", gives him room to claim furthermore that it is up to the compatibilist to meet this challenge in any way their ingenuity can--so long as they retain the proviso.

So why did van Inwagen include this proviso in the ability/power version? And is it absolutely essential for the compatbilist to accommodate the proviso in considerations of alternative-ability claims?

[Mark]

Alan,

Since Vihvelin's account is primarily about what "abilities" are, she should have no problem with any of those phrasings. Whether one agrees with her account is another matter, but she intends to capture the full scope of what ability is and its relation to "the power or ability to do otherwise than what one in fact does" as regards free will.

The main difference between her account and van Inwagen's is that he presumes that the sense of ability required for moral responsibility requires indeterminism since he accepts the consequence argument.

[Alan White]

Still, Mark, even if Vihvelin's account is a compatibilist one that meets with van Inwagen's "otherwise than what one in fact does" proviso, there are any number of others that wouldn't. And PVI's point is that at least some of these others can't be legitimate claimants to a free will position. But that to me is to eliminate contending views by definition--and the question of freedom or responsibility can't be just an a priori matter, or, to put it another way, I don't see why one should dismiss outright the possibility that these questions have at least an a posteriori component. And such considerations might well affect our conceptual grasp of what FW means.

Remember his "God" argument in the piece? Well just because a term reliably and univocally refers to one and the same thing does not entail the same understanding of what that thing is in its referential use. The old view that "dinosaur" referred to extinct cold-blooded reptiles is certainly incompatible with the newer understanding that these were actually warm-blooded avian ancestors. Yet the contemporary use of "dinosaur" certainly refers to those same creatures. But the two underlying concepts are very different. I think the FW controversy on the conceptual side is much the same as this--there are rival, inconsistent concepts that refer to the same phenomena of what we call choice and responsibility.

[Eddy Nahmias]

I hope to be able to write a more significant response after I finish grading (long ago I promised a second post in response to this PvI article in addition to the one where I tried to count up how many incompatibilists and compatibilists there are in order to put pressure on PvI's claim that the consequence argument shifted the ratio significantly).

For now, I just wanted to commend Saul on his excellent response to the article. I appreciate his attempts to prevent certain types of combatibilists from being dismissed from the debate by definition (or declaration).

I also wanted to mention that I can't believe no one here has ever used the term "combatibilists" (typo on one of my student's exams!)...

[Joe]

Alan,

Certainly Lewis is able to accommodate van Inwagen's power/ability-ascription phrases, like "the power or ability to do otherwise than what one in fact does." (Note that at the end of his essay, van Inwagen says that he doesn't think that Lewis was "muddled.") Vihvelin can account for them, too. Thus, there are at least some compatibilists (I would include myself, as well) who wouldn't back away from these phrases, or think that the phrases prejudice the debate in favor of the incompatibilist. Perhaps this is a minority view among contemporary compatibilists, though.

But you go on to say, in response to Mark, "even if Vihvelin's account is a compatibilist one that meets with van Inwagen's 'otherwise than what one in fact does' proviso, there are any number of others that wouldn't." But even here we should be careful.

Here are two separate issues. First, whether, according to van Inwagen's taxonomy, certain compatibilist theories are left out in the cold. Second, whether, according to van Inwagen's taxonomy, certain theories of free will are left out in the cold.

I don't think that the first point is correct. Take Fischer's view, for instance. John believes that free will is what van Inwagen thinks that it is AND he believes that it, as so defined, is incompatible with determinism. He also thinks that free will, as so defined, is not necessary for moral responsibility and that there is some other freedom-relevant condition that is necessary for moral responsibility, one that might be satisfied even if determinism is true. So John has no trouble adopting van Inwagen's taxonomy and explaining his own views accordingly. If fact, his project does just that.

This is not to say that there are not views that get left out in the cold. Most contemporary free will denialists, to name one group, seem to be left out in the cold on van Inwagen's taxonomy -- which is weird since his position is closely aligned with this group’s. Personally, I think that it would be better to broaden the concept of 'free will' to include the source view (and perhaps other views if there are any), and since that is the heart of your position we pretty much agree on this.

[Alan]

Joe--

Very well said. I agree that John's view is a model for the complexities of possible FW positions. I agree as well that van Inwagen's thesis is too restrictive--though you've convinced me that it is not as restrictive as I first thought (see my previous remark on my admiration of PVI's exceeding carefulness of expression--I still need work there!).

Thanks for replying.

[Saul Smilansky]

I haven't been gardenning for a while, for various reasons, some good and some bad (none of course related to our Garden); and it's good to be back. Thanks Manuel for the invitation and Peter for the paper (and everyone for the comments so far). I don't want to take up too much space after my long post, and I won't reply to everything related to what I wrote, but just a few thoughts.

One thing that it's good to think about occasionally (and that Peter's paper helps to focus on) is the "state of the debate". This sort of discussion can get pompus, but risking that I (by contrast to Peter, clearly) think that we are doing just fine. One of the great benefits of the Garden is that it gives us a good indication on this. Some fields are plagued by postmodernism or other such beasts, and perhaps in some countries philosophers discuss free will obscurely, or stupidly, or even dishonestly (I don't know). But at least in discussions of free will within English-speaking analytic philosophy, as also illustrated in the Garden, the standards of clarity, rigor and knowledge in discussion are usually good. So (I claim) we should be mainly concerned with inviting experimental thoughts and being open to as many perspectives as possible.

Second, where things are a bit messy, that's just as likely to be an indication that matters are interesting and we might gain some new insights. A good example is the "compatibility of what" issue. I (for what its worth) think that it's much easier to make a case for free will within the limits of determinism (or, better, for compatibilist free will even if we don't have libertarian free will), than for moral responsibility. The reasons are the one's I sketched in my reply to Peter's paper: it's hard to deny that people suffering from e.g. compulsions are obstructed in their free will in the way that most people are not. The jump up to morality IS a jump, and it's less certain that we can make it. So (and this is also my reply to Neal), as the joke goes, "This isn't a reductio of my view, it IS my view". So, we might have to specify whether we are compatibilists about this or that, but that's not really that much of an effort, and it's in fact good to see that there is a significant question here.

I feel a bit weird wearing my compatibilist hat so much as I do in this thread (I am supposed to be one of the negative and pessimistic guys), but as with other positions, I find that it's just very interesting to see how many different ways there are of getting to the same broad stance. It would be useful to map the great variety of types of compatibilist replies on moral responsibility, but clearly they are very different (e.g. Frankfurt does something totally different than P.F. Strawson, as, differently, does Fischer, Mele by himself does a number of different things which all differ from the above, David Lewis did yet an Nth thing, and so on (the more directly normative/moralistic angle I have tried to develop is based on local free will as above, the notion of "respect for persons", and limiting the compatibilist aims to a shallower "moderate compatibilism"). Where will it all end up? I don't know (haven't read the new Quatro-FW book - congratulations you all), but, seriously, we are (I feel) learning all the time, as we do from the different sorts of hard determinists (whether we agree with them or not), and so on. I am hoping that we will be shown possibilies coming from even wilder places (e.g. utilitarianism or contractualism on free will), in the future. All this just seems to me to confirm J.S. Mill's point about the benefits of variety.

Anyway, this sermon is becoming too long and tacky so I'll be quiet now.