The Garden Reading Group - Comment Now Open

I hope you've all had the opportunity to read Mark's paper by now (if not, it is available here) and are bursting with thoughts about it. Now's your opportunity. But wait, there's more....

My commentary on the paper is here.

Mark's response to my comments is here.

Thanks to Mark for correcting some misapprehensions on my part. And of course for agreeing to participate in the group.

Now over to you.

Comments

[Jim White]

Balaguer wants to argue that the question of whether we have free will is independent of the question of what free will is. The former question, though, is uninteresting without some understanding of the latter. We can agree on all the relevant facts about human nature, but still be at a loss as to their significance. The issue between compatibilism and libertarianism seems to concern this issue of significance: what is at stake if the facts are this way or that way; is anything worth caring about lost if the world is this rather than that way? We can agree that Kobe Bryant is a person without an adequate analysis of personhood (Balaguer's analogy). But in disputed cases (fetuses, for example), agreement on all the facts about an organism doesn't settle questions that matter (how much should we care about damage to this thing?) . (Levy's species analogy also makes this point well.) The questions we're left with may not be questions about "the nonsemantic part of the world," but then, by the same token, the question of whether we have free will (or control or authorship) of the relevant sort when human nature is such and such, isn't either a question about the nonsemantic part of the world.

[Mark Balaguer]

Two points:
1. Response to Jim White: There can of course be borderline cases in connection with lots of our predicates--'person', 'chair', 'mountain', etc. (I probably shouldn't have said in my response to Neil that for all the objects I've actually encountered in my life, I'm very reliable at determining whether they're persons). But this doesn't matter in the present case, because I think it can be argued that the relevant sorts of decisions (i.e., appropriately undetermined torn decisions) involve NON-borderline cases of authorship and control by the relevant agent. (There can also be borderline cases of authorship and control--e.g., cases involving partial determination by external (i.e., non-agential) causal influences.)
2. I realized too late (after Neil had posted his comments and my response) that I should have made one more point about the elephant case, to clarify my view there. The point is this: It could of course be that the terminological question about ‘species’ is settled, or partially settled, by metaphysical questions--e.g., by the question ‘Which concept (SPECIES-1 or SPECIES-2) does a better job of carving nature at the joints?’ But this would be a case of metaphysics being relevant to semantics and not vice versa. For we could answer the latter question independently of semantic considerations and then give our answer to conceptual analysts to use in their deliberations.

[Bob Doyle]

Mark says that the what-is-free-will question is metaphysically uninteresting because it is just a semantic question. I find all the talk of elephants and personhood somewhat semantic.

Neil asks "what is appropriate nonrandomness?"

This term comes from footnote 2 in the paper where Mark defines a special form of L-freedom, designed to handle cases of choices where the reasons are equally balanced (the ancient liberum arbitrium indifferentiae).

"We can say that a person is libertarian free, or L-free, iff she
makes at least some decisions that are such that (a) they are both
undetermined (i.e., not causally determined by prior events) and
appropriately nonrandom, and (b) the indeterminacy here is relevant
to the appropriate nonrandomness in the sense that it generates it, or
procures it, or enhances it, or increases it, or some such thing."

Note that involving both indeterminacy ("free") and an adequate determinism ("will") is critically important.

The indeterminacy can only generate options for action. The appropriately nonrandom will determines the choice from those options. The will is not directly caused by the indeterminacy.
See http://www.informationphilosopher.com/freedom/problem/

Mark's L-freedom is discussed in his earlier paper A Coherent, Naturalistic, and Plausible Formulation of Libertarian Free Will (Noûs 38:3 (2004) 379–406), which, as he says, will be chapter 3 of his forthcoming book.

I have discussed Mark's L-freedom concepts on his page at Information Philosopher.
http://www.informationphilosopher.com/solutions/philosophers/balaguer/.

[Joseph Campbell]

I'm still struggling through the paper, reply, and response but Bob's comment reminded me of a question that I wanted to ask Mark wrt L-freedom.

How does this concept solve the problem of luck other than defining it away? Suppose I give this definition: a person is C-free iff all of her decisions are (a) causally determined and (b) the determinacy here generates, or procures, or enhances, or increases, the up-to-usness of the agent's actions. That doesn't solve the problem of free will and determinism, does it?

[Cihan Baran]

It's nice to see you commenting after a long time, Joe.

Mark, thanks for the paper. As much as I can tell, I agree with you - I would go even further and even claim that since what-kinds-of-freedom-do-we-have question is trivial and what-is-free-will question depends on semantics/intuitions, the debate itself is not very interesting.

Nonetheless, I will go ahead and try to play the devil advocate.

It seems to me that there is an important difference between terms like "elephant", "planet" and "free will". Scientific terms such as "planet" and "elephant" are merely stipulated in a way "free will" can't be.

Thus, it seems to me that conceptual analysis on "free will" can be *informative* in a way the meaning of "planet" or "elephant" isn't - where I'd say *informativeness* as opposed to being "metaphysically interesting" is what we really should care about.

[Bob Doyle]

Hi Cihan,

I would like to suggest a very simple conceptual analysis of the term "free will." It might be "informative."

I believe that "free will" has always been a confusion of two separately analyzable terms - "free" and "will" - two separable concepts.

An act of will is by its nature a determination. We might say we are "determined" by our habits, by our beliefs, and by sufficient reasons in our minds. But this in no way implies that we are determined by strict logical and physical laws going back to the beginning of the universe.

Freedom on the other hand does imply the existence of chance. Chance seems to some to imperil moral responsibility. If the world contains real chance, then luck is real. Tough luck.

But if "free" describes not the willed action - which the standard two-horned dilemma argument claims - but the person, then it means only that random thoughts occur to us that can be considered by our determining wills.
http://www.informationphilosopher.com/freedom/standard_argument.html

Thoughts are free - they "come to us."

Actions are willed - they "come from us."

John Locke said as much centuries ago. He liked the idea of Freedom and Liberty but was disturbed by the confusing debates about "free will".

He thought it was inappropriate to describe the Will itself as Free. The Will is a Determination. It is the Man who is Free.

Locke said, in the Essay, Book II, ch 21, On Power.

"I think the question is not proper, whether the will be free, but whether a man be free." "This way of talking, nevertheless, has prevailed, and, as I guess, produced great confusion."

Indeed it still does.

[Neil Levy]

Joe asks how Mark's account deals with the luck objection. I suppose Mark will reply that the luck objection turns on the action being random or uncontrolled, but the L-free act is controlled and nonrandom. In his response to me, he says we can see that torn decisions are nonrandom without saying what authorship and control are. Now I haven't read Mark's full argument for this claim, but insofar as it rests on his analogy with the concept of personhood, I am unconvinced.

Mark points out, reasonably, that having an analysis of personhood wouldn't help him much in recognizing persons in day-to-life. Indeed, he says, things are really the other way round: it is his ability to pick out persons to which he refers when evaluating analyses of personhood. An analysis would only be useful if he encountered a borderline case, and that doesn't happen often (actually, he suggests it never happens, but that seems false: medical ethics and animal ethics is full of cases...).

But the analogy with personhood is in order only if torn decisions are *paradigms* of controlled events. In that case, we wouldn't need an analysis of control to see that they were nonrandom. Insofar as Mark's conception of torn decisions is like Kane's, though, they are not paradigms of controlled events. Instead, I think even Kane would say that they are (nearing) borderline cases. So we do need an analysis after all.

If all that's right, then Joe's objection is right too: the account simply stipulates the luck objection away, and that's question-begging (to say the least).

[Bob Doyle]

Hi Neil,

Mark's "torn decisions" are as you say similar to Kane's very difficult "self-forming actions." C. A. Campbell said they were decisions that require a great deal of effort.

They are "restrictive," a term coined by John Martin Fischer to describe Peter van Inwagen's claim that only a tiny fraction of our decisions and actions could be free actions. For van Inwagen, like Mark, it is those which have closely balanced alternatives. For Kane, it is those rare and difficult decisions that are deeply moral. They are those moments in which are character is formed. Later decisions made consistent with our character and values can then be traced back to these "self-forming actions." This provides us with what Kane calls ultimate responsibility or UR.

The ancients called freedom in such cases where the alternatives are closely balanced liberum arbitrium indifferentiae, the liberty of indifference. To prove that only humans had such a freedom, they denied it to animals in the classic example of Buridan's Ass.

Now the problem with Kane's L-freedom is the introduction of some chance into the decision. Laura Ekstrom does the same. And so does van Inwagen. See my recent post where I discussed this with Galen Strawson.

At this point Kane and his fellow Libertarians have lost any claim to moral reponsibility, and lost it more assuredly than if luck had suggested an alternative possibility to them - which they then determined to be their best alternative, consistent with their character and values, their habits and current desires.

In my model (which makes "free will" compatible with both indeterminism and "adequate" determinism), although luck may enter "freely," if determination by the "will" is "up to us," then we are the originators of our actions, the authors of our lives and clearly in control, don't you think?

Joe will note that this only solves the "problem of luck" in that it explains how the will can suppress the randomness of a lucky thought.

It does not solve the luck problem as regards moral responsibility.

[Mark Balaguer]

Hi everyone. Thanks for all the comments. Let me try to respond to some of them. First, Joe’s question. I’m not defining the problem away, because I’m not defining L-free choices into existence. The definition of L-freedom gives us a statement of what the world would need to be like in order for libertarianism to be true. But after I’ve given this definition, a very real problem remains: are there really any such things as L-free decisions. We can break this down into two problems: (i) are any of our decisions (and, again, I think the most important decisions here are torn decisions) undetermined in the right way?; and (ii) if any of them are undetermined in the right way, are they also appropriately nonrandom (and does the indeterminacy generate the nonrandomness)? I take these both to be VERY difficult problems. Chapter 3 of the book is entirely dedicated to arguing that the answer to question (ii) is “Yes”. (There was an earlier paper, in Nous, that argued this point as well, but the book version is much improved and expanded.) Chapter 4 is dedicated to arguing that question (i) is an open empirical question--i.e., that right now, we have no good reason for endorsing either answer to it.

So I don’t think I’m defining anything away. Part of the argument in chapter 3 is explicitly concerned with the luck objection, but I actually think the whole chapter is relevant to that objection.

Next, a word about Neil’s remark: I don’t think I really need to argue that these decisions involve PARADIGM CASES of control. I just need to argue that they ARE under our control, in the ordinary sense of the term. And that’s what I do. But I think it’s also implicit in my argument that they’re not borderline cases. (Also, I officially take back anything I said that might have implied that there are no borderline cases of persons. I didn’t mean it. I temporarily lost my mind.)

Next, to Cihan: I agree that an analysis of ‘free will’ can be informative. If we’re clear about what sorts of facts settle the what-is-free-will question, then we’ll be clear about the way in which it’s informative. E.g., if we think that an analysis is right iff it captures ordinary-language usage and intentions, then an analysis of free will would, if true, tell us something about folk usage and intentions. But I disagree with your claim that the which-kinds-of-freedom-do-we-have debate is trivial. I think the question of whether we’re L-free is wide open.

Finally, to Bob: On the libertarian view I’ve got in mind--and I think this is true of just about all libertarians--the most important indeterminacy occurs at the moment of choice, i.e., in the act of choosing between the various live options, or the tied-for-best options. There can also be prior-to-choice indeterminacies, e.g., in connection with what considerations occur to us, but these are less important, at least to most libertarians.

[Neil Levy]

Mark, Do you think I should be convinced by what you in the paper and response that torn decisions are controlled? Or do you think that I am within my rights not to be convinced until I read the book?

Here's something you could say (but I take it you don't; at any rate, Kane wouldn't). Suppose you think all (or a very large proportion) of our actions are torn. In that case, you could say that torn decisions = (to a first approximation) paradigms of controlled actions. In the absence of that claim, I will want to know what features of torn decisions differ from our paradigms. And we are off on the luck debate.

[Mark Balaguer]

Hi Neil. Certainly, you’re within your rights to remain unconvinced, because I didn’t even give the argument. All I really argued was that you shouldn’t think, without even seeing the argument, that I would need to give an analysis of authorship and control in order to argue cogently for the claim that the relevant decisions ARE authored and controlled by us. In fact, I’d say that you SHOULDN’T be convinced until you read the argument.

Also, I wouldn’t use the strategy you suggest. I do argue that if our torn decisions are appropriately undetermined, then they involve ordinary, garden-variety authorship and control. But as I said before, it’s a long, complicated argument. I don’t think there’s any shortcut to the conclusion. And I think that in order for the argument to be successful, it needs to provide a response to the luck objection. My argument does contain a response to that objection, but of course, you shouldn’t just take my word for it. You should remain unconvinced until you read the argument. But once you read the argument, you should be convinced--unless, of course, you want to be irrational.

[Eddy Nahmias]

Is it just me or is there a lot of conceptual analysis going on in this thread?

Anyway, I find Mark's paper interesting but unconvincing. The metaphysical/scientific discoveries don't tell us much unless we know what we are looking for, and we won't know what we've discovered unless we have some conceptual clarity.

Here's one way to illustrate this. Suppose scientists tell us they are discovering that free will is an illusion. Suppose what they mean is that work in neuroscience and psychology gives us good reason to believe that nothing in the human decision-making process is unlawlike or uncaused. They are, we can suppose, working with a libertarian (even agent causal) conception of free will and they are taking their research to show that humans do not have such powers.

Now, for the sake of argument, suppose most ordinary people do *not* believe free will (even of the sort required for moral desert) requires such libertarian powers (e.g., that our decision-making is in part unlawlike or indeterministic, etc.). The folk just think you need compatibilist powers, like rational deliberation, planning, and self-control.

When the scientists say free will is an illusion, if the folk believe them, they'll take it to mean that rational deliberation, planning, and self-control are illusory.

That would be a bad outcome. It might make people behave worse (as suggested in recent research) or change how they feel about themselves and others *and for no good reason* (i.e., based on a conceptual confusion). This would occur even though, we can suppose, no one (not the scientists, folk, or philosophers) disagrees about the scientific and metaphysical facts. It would occur because people are understanding the concept of free will in different ways.

This makes me think that the conceptual debates about free will are not irrelevant and that the scientific study of human decision-making cannot proceed successfully without some clarity about the relevant concepts (e.g., neuroscientists and psychologists use concepts like "decision," and "voluntary" in different ways and it leads them to carry out research and interpret results in different ways...)

[Mark Balaguer]

Hi Eddy. Thanks for the comments. I believe (if I didn’t read your remarks too quickly and miss an implication of something you said) that I agree with all of it and that none of it is inconsistent with my view. There are two really central claims here. The first is this:

(1) Scientific study of human decision-making processes can’t proceed effectively without conceptual clarity.

I completely agree with this. Here’s one way to put this point, in the lingo of my paper: You can’t make much progress on the which-kinds-of-freedom-do-humans-have question until you’ve got some clear kinds of freedom on the table. And to get some clear kinds of freedom on the table, it’s going to take some good hard work by conceptual analysts to figure out the best way to articulate the various kinds of freedom. E.g., if conceptual analysts have worked out five really clear kinds of freedom, then those who are trying to answer the which-kinds-of-freedom-do-we-have question can make much better progress. And on the other hand, if, e.g., (a) scientists are working with a kneejerk, unreflective assumption that free will is indeterministic, so that they’ve got some sort of libertarian freedom in mind, and (b) they haven’t thought through exactly how to articulate L-freedom, so that they’ve got all sorts of false assumptions about what L-freedom is, then this is bad.

But notice that this all has to do with the *articulation of concepts*, and again, I think this is metaphysically important. Indeed, one of the central points of my own work is to provide a clear conception of the exact sort of indeterminacy that’s needed for L-freedom--what it needs to consist in, where it needs to be located, etc.--and how this indeterminacy needs to connect up with authorship, control, rationality, etc.

Note, however, that once we’ve given scientists a set of clear concepts of freedom, so that they can try to determine which ones we possess, it will not help them at all to be told which of these concepts of freedom, if any, provide correct definitions of ‘free will’--and this is the point I was arguing in the paper. In other words, there still doesn’t seem to be any need here to figure out the correct answer to the what-is-free-will question.

The second main point Eddy makes is that it would be bad if (a) scientists announced that we don’t have free will, and (b) what they had in mind was X-freedom, and (c) the kind of free will that ordinary folk care about is Y-freedom, and (d) people become (unnecessarily) depressed over the scientific announcement.

I agree that this would be bad. Now, what does this tell us? How ought we to remedy the situation? Well, one might adopt either of the following two views here:

(2) It would be important in the above scenario to educate scientists and ordinary folk about the difference between X-freedom and Y-freedom and how the scientific discovery is about X-freedom and what people care about is Y-freedom. If we do this, folk depression over the scientific discovery will disappear. Thus, there is no need to answer the what-is-free-will question here. All we need is conceptual *clarity*. I.e., you just need to show people that what they care about is Y-freedom. You don’t need to show them that free will *is* Y-freedom.

(3) In order to get rid of the folk depression, we would need to show people that Y-freedom is *real* free will, and so we would need to answer the what-is-free-will question.

Stance (2) is obviously consistent with my view, but I want to point out that stance (3) is also consistent with my view. My claim is that the what-is-free-will question is essentially irrelevant to the metaphysical project of discovering the nature of human decision-making processes. But I haven’t claimed that it’s irrelevant to the pragmatic job of eliminating folk depression, and it may be that it’s not irrelevant to that job.

That said, I should also add that I find stance (3) dubious at best. In particular, I don’t see why we should think that there’s a very tight connection between what we care about, or what’s depressing to us, and the answer to the what-is-free-will question. One of my problems with the what-is-free-will question is that there is no consensus (implicit or explicit) on the sorts of facts that settle this question. There are numerous factors here. E.g., there are the kind(s) of freedom that correspond to ordinary usage and intentions, the kind(s) that are required for moral responsibility, the kind(s) that humans actually have, the kind(s) that we would be depressed if we didn’t have, etc., etc., etc. Why think that what would depress us is more than one factor among many? And, more generally, what principled, non-arbitrary reason could we give for adopting some particular view of what settles the what-is-free-will question? Finally, I think it’s also worth noting here that different people would apparently be depressed by different things. I don’t think Eddy would be depressed to hear that we aren’t L-free. But I would--and this despite the fact that I lean toward the thesis that L-freedom is not required for moral responsibility.

[Akilesh Ayyar]

This is a really interesting paper & discussion.

My objection to Mark's paper is simple: the what-is-free-will question is important to the what-kinds-of-freedoms-do-we-have question because it explains the kind of analysis that we would find satisfying.

Suppose I asked whether someone was feeling anger at time t and you told me how all the molecules in the universe were moving at time t. Suppose I found this a dissatisfying answer.

Then the question is: why is it dissatisfying? Would not the answer to that question necessitate a closer look at what-is-anger? Would the explication of that question then be "merely semantic"? No. It would point to a possible category mistake.

Similarly, if I ask: how is man free and you answer: "he has this freedom and that freedom and the other under x, y, and z frameworks" and I find that answer dissatisfying -- is not the question: why is this dissatisfying?

And that would lead to a what-is-freedom question, and the answer would come back: "free, as in free will."

[Bob Doyle]

Hi Mark,

Libertarians (you, Kane, Ekstrom, van Inwagen) who want chance at the "moment of choice" should not want the choice to be "made by chance."

I suggest all that you need is chance "during the process of choice," i.e., during deliberation in which you consult your creativity (which employs indeterminism) to generate more alternative possibilities, until you come up with one that meets your rational criteria.

I like cogitation, rather than deliberation, and call my model the Cogito, because it emphasizes the shaking around of new possibilities.

Now I do not deny that you can decide to use chance directly in your choice, to flip a mental or physical coin. But then this is a deliberate decision you can feel is "up to you."

When alternatives are truly balanced, your decision is probably made this way.

But if chance entered the decision without your control, if it "just happened" to you, you would not regard this as your considered decision.

On reflection, you would probably says something like - "I was out of my mind" (as you said above to Neil) - and correct the decision that by chance happened to you.

I have been in personal correspondence with Kane, who has given me permission to cite his views. He basically agrees with you as follows:

"If all of our choices are determined at the time of choice that would not be libertarian freedom even if some chance events in the past were responsible for forming some of the determining factors that now determine our choice because however the determining factors were formed in the past, all of our choices would be determined when they are made."

My Cogito model does not only rely on "chance events in the past" long before our deliberations. We can and always do introduce chance during the process of choice as we ponder the possibilities and generate new ones. Randomness in our brains and in our environment contribute to the generation of new possibilities.

Arguably all biological organisms do something like this. As you said in your paper Libertarianism As Scientifically Reputable, "it might very well turn out that parakeets have free will."

Note that Liberty (and thus the proper aim of Libertarians in my view) is to understand how the world can be indeterministic, thus breaking the chain of strict causal logical and physical determinism (accepting PvI's Consequence Argument), while maintaining an "adequate determinism" that prevents our decisions from being completely random and indeterministic (the concern in PvI's Mind Argument).

As Joe Campbell once put it, we must show that indeterminism can help with freedom, but also that it does no harm to our determinations, to our authorship and control.

If you just expand your "moment of choice" to cover the time of deliberation, then it seems to me that it allows "indeterminism that is appropriately non-random" in a way that should satisfy both libertarians and compatibilists today who accept modern physics.

Determinists and libertarians agree that the determinism we have is only "near determinism" (Honderich), "almost causal determinism" (Fischer), "micro-indeterminism" (Kane), "determined for all practical purposes" (van Inwagen).

Kane is optimistic. He says "the Cogito view I agree is something that could be the beginning of forging a compromise. I've always thought that, because it is an important part of the libertarian view, even if not the whole."

Kane developed his own two-stage model, free - then will, in the early 1980's, but was not happy with it.

The Cogito model is not original with me (I have only contributed a solution to the location of quantum uncertainty, the where and when problem).

Two-stage models of free will were independently developed by William James, Henri Poincaré, Arthur Holly Compton, A.O. Gomes, Karl Popper, Henry Margenau, Daniel Dennett, Robert Kane, Alfred Mele, and Martin Heisenberg.

Mark is urging us to develop a variety of "kinds of freedom" to improve our communications with the neuroscience community.

I believe that young philosophers should consider two-stage models as the most probable explanation for our complex concept of free will. Kane thinks a philosophical consensus might find them a satisfactory "compromise."

To me it is more than a compromise, it is a synthesis of indeterminist/incompatibilist/libertarian views with determinist/compatibilist views.

Please see http://www.informationphilosopher.com/freedom/

[Mark Balaguer]

1. Response to Bob: You make lots of points worth thinking about here. Let me just say three things. First, a rather minor point: I don’t consider myself to be a libertarian because I think it’s an open question whether we’re L-free. Second, I agree that if the moment-of-choice undetermined event were external to the agent--if it just “happened to the agent”--then that wouldn’t give us genuine authorship, control, etc., and so it wouldn’t give us genuine L-freedom. In order to have genuine L-freedom, it needs to be the case that the relevant undetermined event (and keep in mind that it needs to be undetermined in a very specific way) IS the decision, i.e., the conscious, intentional, mental event. This idea might seem a bit obscure, but I develop it at length in the book (chapter 3). Third, as for prior-to-choice indeterminacies, I don’t want to deny that these can be important in various ways. But I agree with Kane that, by themselves, they are not enough to give libertarians (or at least libertarians of the sort I’ve got in mind) what they want. On the prior-to-choice-indeterminacy view, the choice itself will be determined by prior-to-choice events; but what libertarians (of the sort I’ve got in mind) want is a certain kind of moment-of-choice indeterministic freedom. This, I think, is almost definitionally required for the truth of this sort of libertarianism.

2. Response to Akilesh: I guess what I want to say is that I wouldn’t find it unsatisfying to be told, for all the different kinds of freedom that we’ve articulated, which ones humans have and which ones they don’t have. This would tell me everything I want to know (relative to the issue of free will) about the nature of human decision-making processes. It wouldn’t tell me EVERYTHING I might want to know--e.g., it wouldn’t tell me which kinds of freedom are required for moral responsibility--but as I argue in the paper, finding out which kinds of freedom are required for moral responsibility wouldn’t tell me anything substantive about the nature of human decision-making processes.

[Akilesh Ayyar]

Mark,

I don't see how it's possible to ever articulate all the freedoms
humans have, because the very choice of the categorization of
freedoms, and the kind of evidence used to prove their existence, will
be satisfying only against a background concept of free will.

For example I could argue that humans are generally "free" to move
their limbs (a physical freedom), to practice religious ceremonies (a
legal freedom), and so on. Evidence to prove these might be given from
viewing people move and from legal texts, respectively.

These are clearly not the relevant kinds of freedoms we've been
talking about, and their enumeration and proof would be unsatisfying.
The kinds of evidence generally accepted in their proof is similarly
irrelevant.

Why? Because of an underlying belief about what freedom-as-free-will
is which governs our notions of the relevant issue. If
L-libertarianism seems an important freedom to this discussion, it's
due to such a belief. Whether experimental psychology evidence is
viewed as bearing on its existence it is also due to the underlying
free will definition.

The disanalogy with the "planet" example is that there, the relevant
physical phenomena and associated evidence are not in controversy.
Here, the very identity of the thing being discussed is in very
serious question.

A better analogy might be to the debate about whether consciousness can be explained in materialist terms. If one said in response: "let's not worry about whether consciousness is material, let's simply list all the kinds of ways in which we're conscious," this wouldn't work. Qualia and the very nature of the phenomenon are in question. People who disbelieve in qualia would frame "the ways in which we are conscious" and the evidence for it differently. In that case as in this, you can't catalog the kinds without first agreeing on what it is that's being catalogued.

[Devon]

Hi all,

I think that Mark point in the paper is rather simple, and obviously correct if you accept his understanding of what's metaphysically interesting. His point can be put like this: (i) Metaphysically interesting questions are about the nonsemantic part of world, and (ii) the What is free will? question is about the semantic part of the world (e.g. the meaning of English words and the intentions of English speakers), so (iii) the What is free will? question isn't metaphysically interesting.

I think that this argument is compelling, and it contributes to our thinking of how the free-will debate should proceed and, more generally, how metaphysical debates should proceed. I think that this is Mark's aim, and one that should be taken seriously to say the least. Philosophers have been worried about what they're doing for some time now.

The main problem I have is this: There is a good explanation for why the compatibility question has dominated the free will debate in recent years. The explanation I have in mind is that philosophers (I think) are primarily interested in the ordinary notion of freedom or free will, and so want to clarify THAT concept in order to see if we the power that that concept describes. And a natural assumption to be made is that the empirical results concerning the nature of human decision-making processes won't tell us whether we have the sort of power described by the ordinary concept of free will, even though such results may (or may not) tell us whether we have the powers described by other concepts of freedom. I'm interested in whether or not I have the freedom described by the ordinary concept. I'm not (at least not right now) interested in whether we possess other sorts of freedom, if there are any.

Now I think that this is consistent with what Mark says in his paper. But I do think there is a tension between the kinds of freedom Mark is ultimately interested in, and the kinds of freedom that philosophers have been tradionally interested in.

[Paul Torek]

I find that the links to Neil's commentary and Mark's response are broken. Anyone else have this problem?

[Paul Torek]

Of course there's a very tight connection between what we care about and the answer to the what-is-free-will question. That is, once we get in the right neighborhood of cares (we care about the Stanley Cup, too - but we all know that's not at issue here). Free will is not a concept invented in the course of dispassionate scientific observations or logico-mathematical calculations. Our concerns shape the usage of "free will" and thereby severely constrain the range of interpretations of our linguistic practices. A definition of "free will" which failed to connect to those concerns would be as hopeless, as a definition of "soul" which was irrelevant to our concerns about the possibility of an afterlife.

Neil (June 10 5:53pm) makes an interesting point about paradigm cases. Looking back on my own life, all the paradigm cases of free action are highly expressive of my character. None were torn decisions. They were often slam-dunk decisions that, once the to-be-chosen alternative was imagined, were crystal clear.

Which leads me to ask Bob: Why "involving both indeterminacy (`free') and an adequate determinism (`will') [are] critically important"? What good does the indeterminacy do?

Mark proposes that an undetermined event "IS the decision, i.e., the conscious, intentional, mental event." But insofar as it's undetermined, it's undetermined by me. The fact that I did X rather than Y in this case says nothing about the kind of person I was, going into the decision. (It might say more about what kind of person I am afterwards.) Such a thing would be metaphysically interesting, to be sure, but it hardly seems a paragon of agency. I don't see how it could be more free than a clear-cut decision.

[Kip]

This is a great paper that lays down, quite eloquently, some important things to remember.

The principle difficulties I have with the paper are:

1. I don't consider the ideas that new. The chief benefit of the article seems to be, not that it makes huge innovations, but that it crystallizes and makes plain some ideas that have been left assumed or poorly articulated. That, in and of itself, is hugely important though.
2. The article goes beyond its original claim (that the compatibilism question reduces to semantics and doesn't involve metaphysics), with which I wholeheartedly agree, and states that this question is irrelevant to the "do we have free will" question. This goes too far. The inference is based on a strained, or too thin, view of the "do we have free will" question. In particular, it's based on only focusing on the metaphysics of that question, and not the semantics.

I would add that the article seems to go too far in saying that the compatibilism question is relevant to that question in only a trivial sense. It's not trivial at all in just this sense: it will probably decide the question. That is, many people already have grave doubts about the existence of L-freedom, and these doubts seem based on reasonable evidence about the scientific/mechanical working of the brain, etc. In that case, the "do we have free will" question *turns* on the compatibilism question, which can hardly mean that the latter is irrelevant, or only trivially relevant, to the former.

Here's an example: suppose that two law firms are engaged in a huge litigation. Both sides have spent 20 million dollars in attorneys fees and deposing expert witnesses, etc. Near the end, it becomes clear that the entire case will turn on whether the contract in dispute uses the word "predominantly" to mean greater than 60% or greater than 70%.

Now the question "is 60% or 70% 'predominantly'" seems to be just semantics, just words, no metaphysics involved whatsoever. One would be tempted to say that the question is trivial. One would be tempted to say that the question is irrelevant to the "is the contract enforceable" question, or only relevant in a trivial way. That would be a mistake. That tiny, little semantic question decides the whole case--$40 million. The free will debate is like that.

One last point: Paul above suggests that our concerns shape the definition of free will. I think Pereboom's remarks in the intro of Living Without Free Will are decisive against this point: we must acknowledge the possibility that (i) we thought that free will means X; (ii) we thought X was valuable; and (iii) it turns out that X isn't valuable. In that case, free will *still* means X; we don't fudge the definition so that our vocabulary sweeps our value error under the carpet. Our language practices have to leave room for speaking about, and acknowledging, the value error. If the definition of "free will" changes to always be something valuable, then we have no easy way to say "we thought free will wasn't valuable, but we were wrong."

Here's an analogy: just as Kant rightly says that existence is not a predicate, so too "valuable" is not a predicate. We consider concepts, and then we ask whether any instances of that concept exist, and we similarly ask if we value such an instance. But valuable isn't built into the definition of the thing (at least not into free will).

[Bob Doyle]

Hi Paul,

Indeterminacy does these critical things vital to the Libertarian argument for "free will."

1) Indeterminacy breaks the causal chain of strict logical and physical determinism back to the origin of the universe.

This involves what Peter van Inwagen and others call the Laws of Nature and the Fixed Past in his Consequence argument.

What quantum indeterminacy does is show us that the Laws of Nature are not deterministic, they are statistical laws. To be sure, for large objects, these laws approach the classical laws (e.g, Newtonian physics) to the limits of our observational/experimental accuracy.

This is what I mean by "adequate" determinism. It is as much determinism as the physical world contains.

Adequate determinism is all that a compatibilist/determinist could ask for to insure that our decisions are "up to us," that we are the authors of our actions, that we are in control.

2) Indeterminacy generates novelty, bringing new information into an otherwise static, information-preserving universe.

You know the Laplacian super-intelligence that can predict the future of the universe from knowledge of the positions and velocities of all the material particles at any instant. (This is the natural physical equivalent of a supernatural deity that has perfect foreknowledge).

In a word, indeterminacy vitiates both the super-natural and the super-intellingent. In theology, it says the future of the universe is unknown (Open Theism).

3) Indeterminacy generates alternative possibilities for human thought and action that a Frankfurt demon can not anticipate.

Genuine alternative possibilities means that the past may be fixed, but the future is truly ambiguous, logically and physically, depending on your decisions.

Logically, Aristotle's Sea Battle showed that there are statements neither true nor false (three-valued logic) until the future resolves the truth or falsity.

Physically, each of us may have a truly original thought and act in a truly original way that changes the world.

So what is the problem with a concept that helps us to understand Mark's what-is-free-will question, indeed understand it metaphysically?

The problem is its dependence on notorious chance, on randomness, on chaos. As William James said of both hard and soft determinists, they have an "antipathy to chance."

For the religious philosopher, real, irreducible, and absolute chance has been regarded as atheistic. It denies intelligent design in creationism, for example.
______________

The above reasons are why Libertarians embrace indeterminacy.

So far, so good.

But they are way too enthusiastic about it. They want it all the time.

Some of them, including Mark, Bob Kane, Laura Ekstrom, and van Inwagen specifically say indeterminacy must be involved directly in our free decisions.

They want chance to be the cause of our actions!

Now this is absurd, but perhaps it is a leftover residual from the wish for a metaphysical freedom, a Kantian noumenal freedom, a God-given freedom denied to animals like Buridan's ass.

Libertarians think only such a freedom can deny the iron grip of determinism at every instant, including what they mistakenly describe as the "moment of choice."

A decision does not occur at an instant of time. It is a process of deliberation. In the early stages of deliberation, indeterminacy generates those possibilities.

Even as late as the "moment of choice" itself, a decision can be postponed and sent back to the possibility generator for "second thoughts."

Note we can therefore involve chance in our deliberations up to some minimal time (perhaps the Libet time?) before our "final" decision, which will be "adequately determined," based on our character and values, our habits and preferences, our current feelings and desires.

In short, our choice will be everything a compatibilist/determinist could want in terms of feeling we are the authors of our decisons and in control of our actions.

And enough chance will be involved in our decisions to provides the causal chain disconnect needed by PvI's Consequence argument.

So what's not to like about the two-stage model of "free" (chance) and "will" (choice) in a temporal sequence?

Compatibilists have pointed out for centuries to the Libertarians that if chance directly causes our actions, that is no help, only harm, to the idea that we are the authors of our lives.

And van Inwagen's Mind argument confirms these fears. If you do not know van Inwagen's labels for the two horns in the dilemma of determinism vs. indeterminism, please see my discussion of the standard argument against free will.
http://www.informationphilosopher.com/freedom/standard_argument.html

The logical conclusion of the standard argument, first articulated as such by Jack Smart in his 1961 Mind article that led van Inwagen to call it the Mind argument, is this.

Determinism and Indeterminism are logical opposites. They are contradictiories. Only one of them can be logically "true."

The simplistic physical application of this correct logical conclusion, is this.

If determinism is false and indeterminism true, all hell will break loose - causality is impossible, logical arguments are not valid, we can not reason, science can establish nothing, in short, the universe falls apart.

If you think I exaggerate the fears about indeterminism, please read the standard argument page on my Information Philosopher website.

To contain any "harm," we need to show Compatibilists that indeterminacy is limited to the critically important role of providing them with novel possibilities for action.

Do you think we can get compatibilists to calm these fears? Will they consider this? Bob Kane thinks there is room for a grand compromise here between contending Libertarian and Compatibilist views.

As a quantum physicist, I can assure you that indeterminism is "true" in the logical sense above. But you can be equally sure that quantum indeterminacy has a minimum and negligible effect on the macroscopic processes that provide us with our causal everyday world.

We are just lucky that we have chance available to us where and when we need it. It plays a critical role not only in "free will" but in all the creation processes in the universe, from microscopic objects like atoms and molecules to macroscopic objects like galaxies, stars, and planets.

Please see my story of creation.
http://www.informationphilosopher.com/introduction/creation/

[Mark Balaguer]

Hi everyone. Just a few responses to some of the points that have been made…

First, to Devon: You say that people care about the ordinary notion of freedom and not OTHER notions of freedom, i.e., the ones I’m talking about. But the ordinary notion (if there is such a beast) IS one of the kinds of freedom I’m talk about. I’m talking about ALL the kinds of freedom that philosophers have articulated. Now, of course, it’s possible that no one has succeeded in articulating the ordinary notion, but that would just mean that we have some more work to do in terms of concept articulation. If we ignore this possibility, and if we assume that we’ve got, say, five kinds of freedom on the table, then one of these kinds of freedom (perhaps more, but let’s ignore that possibility) IS the ordinary notion. So I’m not ignoring any interesting kinds of freedom.

I’m also not suggesting that we shouldn’t try to figure out which kind of freedom
is ordinary freedom. I’m simply saying that if we already know that, say, we have freedom 1, 2, 3, and 4, and that we don’t have freedom 5, then we’re not learning anything genuinely new about ourselves if we learn that freedom 3 is ordinary freedom.

Kip suggests that in learning this, we would be “settling the debate”. There is, of course, a clear sense in which that’s true. But I would say that if we understand “settling the debate” in this way, then settling the debate doesn’t tell us anything genuinely new about humans. We can also settle the debate about how many planets there are in the solar system by coming up with a definition of ‘planet’, but in so doing we wouldn’t be learning anything substantively new about the solar system.

To Paul: I’m not suggesting that there’s NO connection between the which-kinds-of-freedom-do-we-care-about question and the what-is-free-will question. I’m simply resisting the idea that the former is some sort of trump card. The problem is that if you ask ten philosophers what sorts of facts settle conceptual analysis disputes, you get ten different answers. It also seems unlikely to me that our “cares” zero in on a unique kind of freedom. It also seems unlikely to me that our usage and intentions zero in on a unique kind of freedom. All of this leads me to have serious doubts about whether there’s anything even close to a fact of the matter about the answer to the what-is-free-will question. We might be better off asking more local questions like “Which kinds of freedom correspond to ordinary usage and intentions?”; “Which kinds of freedom are required for which kinds of moral responsibility?”; “Which kinds of freedom do people care about?”; “Which kinds ought they to care about?”; etc.

Also, Paul, you say that insofar as a decision is undetermined, it’s undetermined by me. Well, that’s just the old worry about libertarianism that’s been raised in different forms by Hobbes, Hume, Hobart, and contemporary advocates of the luck objection. And as I said earlier, this is just the objection that I dedicate much of my book to responding to. I wish I could give the argument here, but I am reluctant to summarize a 30-page argument in one paragraph; I would just shortchange it.

Finally, to Bob: I certainly don’t want to claim that chance is the cause of our actions. In fact, I’m not exactly sure what that would even mean. Moment-of-choice libertarians can say different things here; one thing they can say is that (at least in ideal cases) the choice is probabilistically caused by the agent’s reasons for action. Now, obviously they’ll have to say more than this; in particular, they’ll have to say why, if the causation is only probabilistic, there isn’t a freedom-damaging kind of luck present. But, again, I think this challenge can be met.

Also, on the libertarian view I’ve got in mind, we don’t have an every-moment sort of L-freedom. We can go about our business in an essentially Humean sort of way for hours on end; but every so often, we’re presented with a choice, and we feel torn, and we pause for at least a moment, and then we make a torn decision. Most of these (but not all) will be pretty unimportant--e.g., should I work out before going to the office?; should I get off the freeway in a traffic jam or take surface streets?; should I order chicken or beef?; should I vacation in Paris or Rome? But SOMETIMES, I think, we make more important torn decisions, decisions that have serious moral implications, or serious implications for our own personal futures.

[Mark Balaguer]

One more point about Kip’s suggestion that we can settle the free will debate by figuring out the correct definition of ‘free will’. I just said that this is true on a certain understanding of “settling the debate”. But what I meant is that it’s true (on that understanding “settling the debate”) RELATIVE to Kip’s assumption that we can already answer the which-kinds-of-freedom-do-we-have question. But I actually reject that assumption. E.g., I think we have no good reasons right now for endorsing or rejecting the thesis that we’re L-free. This, I think, is a substantively open question.

[Tom Clark]

"I think we have no good reasons right now for endorsing or rejecting the thesis that we’re L-free. This, I think, is a substantively open question."

Mark, I'm wondering: should we refrain from responsibility practices and ascriptions of desert premised on L-freedom until such time as we know for sure we're L-free? This raises the question of whether there are any real differences between libertarians and compatibilists in the sorts of practices and ascriptions they think are justified.

[Mark Balaguer]

Hi Tom. I have serious doubts about the thesis that L-freedom is required for moral responsibility. Maybe it's required for some kinds of moral responsibility but not others. I also have doubts about whether there's even a fact of the matter as to whether "real" moral responsibility requires L-freedom. But I guess I lean toward the view that at least some important part of our moral practices (including our responsibility practices) are inherently compatibilistic. If that's right (a pretty big if, I suppose), then the question of whether we’re L-free wouldn't have such great consequences for morality as some people might suppose. But this doesn't make the question of whether we're L-free less interesting, from a purely theoretically POV, and it doesn't make me CARE about it any less. Perhaps others would care about L-freedom only if it were required for moral responsibility, but that's not how I feel.

[Kip]

Mark,

You write: "But I would say that if we understand 'settling the debate' in this way, then settling the debate doesn’t tell us anything genuinely new about humans."

Did any compatibilist (or incompatibilist) ever suggest otherwise? That is, did any such philosopher ever suggest that answering the compatibilism question will "tell us anything genuinely new about humans"?

[Mark Balaguer]

Hi Kip. I think that’s a pretty standard compatibilist view. If you can show that free will is Humean freedom (or whatever), then since we know that we’re Humean free, it follows that we have free will. And that’s an important fact about us.

[Devon Dickinson]

Hi Mark,

I understand that the ordinary notion of freedom is one of many notions you are interested in. I was only saying that the traditional project concerning free will was a project about the ordinary notion. And since you are interested in any of the notions that have proposed--an interest that isn't limited to the ordinary concept--that your project is just different. So, if someone was only interested in the traditional project, then they should worry about what you're saying only to the extent that what you're saying might pertain to the ordinary notion.

I was just hoping to highlight a difference between what you're doing on the problem of free will and the traditional project. Now someone might not find this all that interesting, if she thought, for instance, that the traditional project was misguided. But if someone were interested in that project, then she should keep asking the What is free will? question even if it isn't a metaphysically interesting one.

[Paul Torek]

Mark,

I only want to press the luck objection so far as to deny than a torn decision is ipso facto more free than a clear-cut decision. I don't deny that torn decisions can be free, nor that the agent is responsible. The choice of X over Y doesn't particularly express her status as an agent, but the choice of (X or Y) over Z does. And on the responsibility question, she's not only responsible for (X or Y) but also for X. Bernard Williams was right. There is such a thing as moral luck, and a torn decision could be an example.

We might not disagree, since you express doubt about incompatibilism. A clear-cut decision may be determined by the agent's reasons, but if you don't see this as a problem, then my alarm bells were misguided.

Kip,

Of course it's possible (ex ante) that we thought free will means X and then X turns out not to be valuable. But in that case we did care about X. Similarly someone might grow up thinking that "success" is highly valuable, but on achieving it discover otherwise. That is not the fate of some analyses of free will (such as simple indeterminism). Instead, they offer us something we never wanted in the first place. It is like discovering that we had defined "success" not only as a mere number in a bank account, but a mere number in somebody else's bank account. Such a definition is simply wrong, and our surprising lack of even momentary delight is the dead giveaway that it is wrong.

Bob,

Aha, you're citing the Consequence Argument at me. Color me unimpressed. But here is not the place to pursue that argument. Instead, I'll come to your Information Philosophy blog.

[Devon Dickinson]

The thought that there are many kinds of freedom might seem puzzling to some. It's puzzling to me, especially when there are good reasons for thinking that many of the compatibilist's analyses of freedom are consistent with intuitively UNfree actions.
So even if we act in a way that satisfies some of these notions of freedom, it is puzzling to take these actions as FREE actions.

Instead of talking about the many ways of acting freely, perhaps we should talk about actions that satisfy various proposals of freedom. For example, instead of saying that we have Humean freedom or Franfurtian freedom and so on, we should talk about the actions that people perform in a way that makes explicit that they satisfy the conditions involved in some analysis of freedom or other. So if I act in a way that satisfies Hume's conditions for freedom, then we might say that some of my actions are H-actions rather than saying that I possess H-freedom, since there is good reason to think that H-freedom isn't freedom at all.

I think that this way of talking highlights an important feature of the What is free will? question, even if it is granted that this question isn't interesting in Mark's sense. Suppose it is deteremined that we often perform H-actions or F-actions or any other sort of "free" action out there, it would still be rather informative, and hence, worthwhile to find out, that H-actions or F-actions are FREE actions.

Again, this is consistent with the main points Mark makes in his paper. I just want to motivate why philosophers are preoccupied with answering the compatibility question, since he says in the beginning of his paper that this is unfortunate.

[Kip]

Mark,

I'm confused by your comment (June 12, 2009 at 04:19 PM). Do you endorse what you wrote there, or are you writing as a compatibilist to paraphrase a view you're attacking?

Here is my point: you argue that compatibilism involves only semantics or conceptual analysis, and not metaphysics. You argue that, therefore, the compatibility question is "trivial" and irrelevant to the "do we have free will" question.

But it is not clear to me that compatibilists ever disagreed with you. As an example, Dennett writes in Freedom Evolves that his compatibilist view is only slightly/terminologically different than free will anti-realism.

Now, I agree that, to the casual reader who glances at compatibilist writings, they might get the impression that compatibilism is more important--more metaphysical--than it is. And all parties could have done a better job making clear your point (and that is the great virtue of your paper, I think: making plain that compatibilism is just about semantics and conceptual analysis). But compatibilists have studied the issue in depth, and I would be surprised if, in all of their studies, they never recognized the point that you make.

[Mark Balaguer]

Paul,
I think we agree. I don’t think libertarians need the result that torn decisions are *more* free than clear cut decisions. They only need the result that our torn decisions can be simultaneously undetermined and appropriately nonrandom and that the indeterminacy procures or generates the nonrandomness. But to say that the right kind of indeterminacy would generate appropriate nonrandomness in our torn decisions is not to say that the same thing holds for other kinds of decisions, and it’s not to say that clear cut decisions can’t be determined and appropriately non-random.

Devon,
I agree that some of the things I call “kinds of freedom” might not actually be kinds of freedom, depending on the answer to the what-is-free-will question. That’s why, in certain places in the paper, I put scare quotes around ‘freedom’. E.g., I talk about kinds of “freedom”.

[Kip]

Paul,

I'm confused by your comment, too. You write:

"Kip,

Of course it's possible (ex ante) that we thought free will means X and then X turns out not to be valuable. But in that case we did care about X. Similarly someone might grow up thinking that "success" is highly valuable, but on achieving it discover otherwise."

I am not sure why you are writing "but." Who are you contradicting? My point is that we must acknowledge the possibility that (i) we believe free will means X, (ii) we valued X and (iii) it turns out that X is actually not valuable. If you agree that we need to acknowledge that possibility, then there is no dispute between us.

Note that, if you do agree with me about that, then I don't think you can say that the definition of free will must track our values or concerns (as you originally wrote at June 11, 2009 at 06:37 PM). It need not so track our values and concerns because, as we've agreed above, we might have been wrong to have those values and concerns.

"That is not the fate of some analyses of free will (such as simple indeterminism). Instead, they offer us something we never wanted in the first place. It is like discovering that we had defined "success" not only as a mere number in a bank account, but a mere number in somebody else's bank account. Such a definition is simply wrong, and our surprising lack of even momentary delight is the dead giveaway that it is wrong."

This example makes no sense to me. If we had defined "success" to mean the amount of money in someone else's bank account, then that's what it meant. That definition of success (in that bizarrro world) is very different than what we mean by success (in this world), but that doesn't mean that we should switch/fudge the definition.

Basically, your example says: "even in hypothetical worlds where success is defined differently than it is in our world, the other worlds should revise their definition to track what it means in our world." Why? What am I missing?

[Mark Smeltzer]

Kip,

I think the story goes something like this:

1. If S values P, then P is valuable to S.
2. S values P, and P is therefore valuable to S.
3. S believes P is defined by {C1, C2, .. Cn}, or simply C.
4. If C is not valuable to S, then S ought not to believe that P is defined by C.
5. S does not value C, and C is therefore not valuable to S.
6. Therefore, S ought not to believe that P is defined by C.

[David Chaffee]

Mark,

Almost without exception, the introduction to a book on free will speaks of the problem’s universal appeal to people. I am one such person, well outside the field, who can attest to that; and I am very appreciative of the opportunity to interact with the professionals at this forum. If you are not altogether exhausted from the discussion thus far, I would sure be interested in your thoughts.

I think your argument, as an argument, is sound. As has been noted, the what-kinds-of-freedom-do-we-have question may not be appropriately considered as a reduction or a more fundamental question, but I think those are minor distinctions and your argument doesn’t hinge upon it. But, I did experience a couple of concerns that, if valid, would be more problematic.

First Concern

However valid the argument is semantically, I wonder if there isn’t some sense of reality that escapes it. The compatibility debate, while it does contain a lot of definitional wrangling, also promotes or refutes a variety of metaphysically possible answers to the what-is-the-nature-of-our-will question. I believe there is a feedback mechanism in the debate for compatibilism or incompatibilism, such that our conceptions of the metaphysically possible are refined. So, I am concerned that throwing out the compatibility debate would be throwing the baby out with the bath water. We may not find the water particularly interesting, but if the baby is going to come clean, it seems like we’ve got to run the wretched little guy a bath. As a venue to define and refine the metaphysically possible, I would say the compatibility debate contains, as an essential element, metaphysically relevant discussion. I do agree, of course, that we’re not going to define our way into metaphysical knowledge.

Second Concern

I also wonder at the argument as a device to motivate consideration of libertarianism and empirical exploration. Ideally, the motivation for libertarianism would be a new species of it that is not vulnerable to the old, and in my view valid, criticisms. In the libertarian accounts with which I am familiar, the torn decisions seem analogous to a pencil balanced on its point. And there seem to be four choices. We can knock it over with strict determinism; we can flip a coin; we can say the pencil decided; or we can say one can’t really balance a pencil on its point. Even when we face a torn decision, and are motivated to think either choice is better than doing nothing, we need a mechanism for that choice, undetermined as it may be at the moment of choice. If we can only reduce it to “the pencil decided”, then it seems like an argument for a soul. Deciding to flip a coin seems equally problematic. If I refrain, via a coin flip, from throwing a rock at a neighborhood kid who is retrieving his ball from the flower garden, is my morality to be held in high regard? Just to be a bit of a contrarian, I would argue that conceptual analysis needs to be done to separate the notions of freedom and control (even independently of the free will problem). How ever cleverly these concepts get shaved in a philosophical work, there seems to be a fundamental prohibition at separating the two, as with wave-particle duality. I very much liked your discussion of rationality and the subconscious in your 2004 NOÛS paper and I would make the bonus comment that however torn a decision feels to our awareness, the “vote” of a few billion neurons would seem unlikely to arrive at a tie.

Similarly, the best motivation for empirical exploration is the description of an experiment which would produce evidence on the nature of our wills. Even an experiment that would reduce some of the meanings we attach to concepts used in the debate would be helpful.

These concerns aside, I am interested in libertarianism and in empirically-based advancements. I look forward to reading more from you.

Regards,
Dave

[Mark Balaguer]

Hi Dave,

I’m not entirely certain that I really understand either of your two concerns, but in connection with the first one, I think the response is simply that I do not want to “throw” the compatibilism question out. I just want to point out what it’s relevant to and what it’s not relevant to. And as for the second, if I understand you, I think you’re once again raising a version of the luck objection--how can a decision be MINE, and how can I be responsible for it, if it occurred indeterministically and, hence, presumably randomly? If that is indeed your worry, then again, I don’t have a response to it in the paper under discussion here, but there is one in my forthcoming book (and there’s an early version of it in the Nous paper).

As for the last comment about the probability of billions of neurons producing a tie, the term ‘torn decision’, as I use it, is defined in terms of the phenomenology of our decisions, i.e., in terms of how they feel to us. So if it feels like a tie and if it feels like a decided without resolving the tie in the weight of the reasons--i.e., if it feels like I “just decided”--then it is a torn decision.

[Devon Dickinson]

Mark,

Just to clear things up (I hope) about my previous post, I was only suggesting a way of putting what you're saying in the paper that didn't mention anything about there being one or many different kinds of freedom. Since the do-we-have-free will question is really about the sort of abilities we have (or don't have), we can talk about those abilities without mentioning anything about freedom or freedoms. That's all I was saying, and I wasn't saying or implying that you thought there was more than one kind of freedom.

I don't think this is a merely terminological objection either. Whether there is one or many kinds of freedom is probably a matter of dispute, so to say what you want to say in a way that avoids that dispute would seem important.

[Kip]

Mark S.:

Right. I think we understand the argument exactly, as you explain it.

I just think it's fallacious.

The problem is premise 4:

1. If S values P, then P is valuable to S.
2. S values P, and P is therefore valuable to S.
3. S believes P is defined by {C1, C2, .. Cn}, or simply C.
4. If C is not valuable to S, then S ought not to believe that P is defined by C.
5. S does not value C, and C is therefore not valuable to S.
6. Therefore, S ought not to believe that P is defined by C.

Premise 4 is false.

Consider an analogy:

We believe dust is *dust* (where *dust* is that annoying stuff that collects on furniture). We also believe dust is valuable. One day we discover that dust isn't valuable. It's just worthless crap that collects on furniture, and requires cleaning. Does that mean "dust" isn't *dust*? No. Dust is still dust. It just isn't valuable: we were wrong to think it was valuable, and our language should not conceal our mistake.

[Paul Torek]

Mark S., you hit the nail on the head. To put it another way, when a concept is strongly tied up with certain concerns, we should pay attention to the clues those concerns provide when evaluating proposed definitions. As Mark B. points out with the concept "person", we can freely and quite accurately use a concept without having anything close to a definition in hand. Nor is this the exception - it's the human condition. Our mental faculties evolved to recognize cases, not spout definitions. Conceptual analysis is a tricky business of hypothesis and explanation, and we need all the data we can get. Concerns provide some of that data.

[Paul Torek]

Kip,

I'm not saying that our definition of P has to track our values, in the sense of evolving in lockstep with them as they change. Rather it has to connect with them in the place they originally reside. If, later on, we decide that P (defined as C) wasn't valuable after all, that's OK, the definition may be correct.

[Mark Smeltzer]

Kip,

How about this?

1. S believes P is defined by the set of conditions {C1, C2, .. Cn}, or simply C.
2. If cases that satisfy C are not as valuable to S as cases S would identify with P, then S ought not to believe that P is defined by C.
3. Cases that satisfy C are not as valuable to S as cases S would identify with P
4. Therefore, S ought not to believe that P is defined by C.

[Kip]

Paul,

I don't know what you mean by "it has to connect with them in the place they originally reside." But we seem agree that arguments of the form:

1. that kind of free will isn't valuable or worth wanting
2. therefore "free will" can't be defined that way

Don't work.

Mark S.:

Your new premise 2 seems just as troublesome as the previous premise. I disagree with the general idea behind these arguments, and so I don't think a careful rewording of the argument will satisfy me.

I tried to explain my objection with the "dust" analogy. I would be interested to hear what you and Paul thought of that analogy.

[Mark Smeltzer]

Kip,

First, I thought you were playing on the fact that the original (4) didn't make any specific reference to the fact that we were only meaning to be talking about things that S would find valuable. The refined version addresses that criticism -- whether or not it was the objection you intended to raise.

Outside that interpretation, I don't see any relevance to the premises of either form of the argument and your case about dust. If you mean to provide a case that demonstrates that the argument's logical form is invalid, please provide a definition for each term, and it should be obvious whether the case succeeds in defeating the argument. Otherwise, I think you would need to provide an argument against the central claim.

Second, I don't see how you can "disagree with the general idea". For, the general idea is that if there is a property P that we hold true of Q, and we have a definition for picking out what does and does not qualify as a Q, then property P must* be true of every x that the definition suggests is a Q. (* If we find a case where an x satisfies the definition for Q, but lacks the property P, there are two possibilities: we were wrong that P is true of all Q's or our definition is wrong.)

The property being used in the argument is the relational property that "S values P". The neat thing about this property is that S is the truth maker for this property. As such, we don't need to have a debate about whether the property is true in a given case -- we just ask S. This is significant because we're interested in the question of how S's values function with respect to what S ought to believe.

The core principle at work here is an old one: (EV) If P is false and S learns that P is false, S ought not to believe P. On the face of it, it may seem that people like Smilansky might argue against this principle, but I don't think so. I think Smilansky would add another principle... something to the effect of: (EV) and If P is false and S learns that belief that P is fails entails horrible outcome Q, S ought to keep the falsify of P to herself and continue propagating belief in P. (Or something like that.) In other words, I think even Smilansky could endorse EV and still maintain Illusionism. Which is to say that I think all rationale persons ought to endorse (EV).

[Kip]

Mark:

Let me address your points in turn.

"Outside that interpretation, I don't see any relevance to the premises of either form of the argument and your case about dust."

Your original premise 4 is "4. If C is not valuable to S, then S ought not to believe that P is defined by C." Clearly, this applies to the dust analogy: dust is not valuable, but it doesn't follow that we ought not to believe that dust is *dust*. Your new premise applies to the analogy in a similar way.

"Second, I don't see how you can "disagree with the general idea". For, the general idea is that if there is a property P that we hold true of Q, and we have a definition for picking out what does and does not qualify as a Q, then property P must* be true of every x that the definition suggests is a Q. (* If we find a case where an x satisfies the definition for Q, but lacks the property P, there are two possibilities: we were wrong that P is true of all Q's or our definition is wrong.)"

As your parenthetical caveat notes, your bolded quote is only true if P is actually a property of Q. My point is that we can be wrong about P. We seem to agree about this. Note that this applies to your premises and my dust analogy in a way that generalizes: if we can be wrong about P, and we find out that X doesn't have P, it does not follow that our definition of X is wrong.

"If P is false and S learns that P is false, S ought not to believe P."

Right! Which is why, if we learn that free will isn't valuable, we should say "free will isn't valuable," and not "free will is still valuable, long live free will!"

[Mark Smeltzer]

Kip,

Okay, so I guessed correctly, which is why the revised argument is not vulnerable to the kind of counterexample you suggest. Check again.

Regarding, (P), (P) has been defined such that S is the sole arbiter of the (subjective) value of (P). In this case, if we ask S whether P is valuable to S, and S says "yup", then that line of the debate is closed. As long as S maintains that P is valuable, S gains the ability to reject definitions such as Q that do not capture the value that S sees in P. (Moreover, careful attention should be paid to the fact that the value S appreciates in P is what fuels S's desire to define P in the first place.)

This is, more or less, the way that analytic philosophy -- the attempt to define the necessary and sufficient conditions for a concept -- is done. So, if you think that this process lacks merit, there is a lot more on the line than just this one line of thought and its application to the analysis of free will.

In fact, fringe skeptical positions, like impossibilism, owe their relative success to this process. If we weren't worried about the difficulties involved in precisely defining the things we care about without tarnishing the meaning those things have to us, it is hard to imagine that anyone finding skepticism compelling. Skepticism is born when someone encounters a difficulty to convince himself that something is true, and without the analytic process I think we would be open to convincing ourselves more readily.

[Kip]

Mark,

Your new comment makes me think that you are hinging your position on the following distinction:

"Well, Kip, you are talking about value as if it is something we can be wrong about. Something like objective value. I'm not talking about objective value. I'm talking about subjective value. S can't be wrong about subjective value, because S is the ultimate arbiter of whether something is subjectively valuable (to S)."

Let's call that s-value, just to be clear. I agree with you that the points I make above don't apply to s-value, for exactly the reasons you give.

Of course, I wasn't talking about s-value then. I was talking about a more objective kind of value, which is what I think most people mean when they talk about free will's definition and its value. For example, I think Dennett's view considers "the varieties of free will worth wanting," and not "the varieties of free will that we currently want, even if we're stupid to do so."

In response to those people, like Dennett, I would say (following Pereboom) that we have to acknowledge the possibility that we were wrong to value free will, because maybe it will turn out to be worthless.

[Mark Smeltzer]

Kip,

Premise (1) in the original argument made it clear what kind of value is at bear in the argument (both arguments actually):

1. If S values P, then P is valuable to S.

As regards the soundness of the argument, it doesn't matter a lick if S adopts different stances toward the value of P throughout her life. What matters is whether there are times for S when the premises of the argument are true. In those cases, if the premises are true (viz., premise (1) and (2) of the revised argument), then the conclusion follows.

If at a later time S stops valuing P, S would be (epistemically) free to entertain C. Likewise, if at an even later time S regains her esteem for P, in a way that cases that satisfy C do not measure up to, S would be back in the position where belief that C defines P is epistemically intolerable. And so forth.

The argument being motived here obviously applies only so long as S values P.

Moreover, to the chagrin of those who would like to convince S to adopt definition C for P, I do not think that S is obligated to give a rational about for what S values in P that is missing from C. S may desire to answer that question, but S's brute impression of P constitutes a fortiori support for S's desire to reject C.

As to the possibility you suggest, I don't see why the mere possibility of being wrong should dissuade anyone from valuing free will.

As a way of illustration, consider someone who values the pleasure of bathing in the sun's luxurious rays. Later in life, this person becomes acquainted with the research that warns of cancer risk, etc., and consequently opts to avoid unnecessary exposure to the sun. Even in this situation it would be hard to make the claim that this person no longer values the pleasure of sun bathing. Rather, that value is interposed in this person's mind against competing values. In the end, the value of reducing the risk of cancer may weigh more heavily on this person than value of sun bathing, but this analysis does not mitigate the value that sun bathing still holds for this person.

Surely we can imagine another person who, after learning about the dangers of sun exposure, develops a thorough going hatred for the sun. The example above is meant to show that it is possible to both love/cherish and fear/hate the same thing in a ways that do not seem to contain logical contradiction (though I am not sure that avoiding contradiction outright is strictly necessary here).

In other words, even if free will can be shown to have certain (even critical) difficiencies, I doubt we can ever be "wrong" about valuing it.

[Kip]

Mark,

I agree, and I missed the point about your first premise in your argument. It is important to note, though, that your argument is limited to s-value (subjective valuation), as I explained above. I think we agree about this (right?). In that case, it is importantly different than claims about the "varieties of free will worth wanting," for example.

Regarding your sunshine example: consider the impossibilist kind of free will that G. Strawson, Tamler, Honderich, myself and Pereboom* have asserted best defines free will (or does so better than the rest). It seems silly to value the logically impossible.

Note that we can tell a story about why people would use "free will" in the impossibilist way, we can tell a story about why they find it superficially attractive... it appears to give them ultimate control, it prevents us from saying "my genes made me do, or my childhood environment made me do it," etc. I actually think that story is the most plausible, and explains why people actually do use "free will" that way. But, on inspection, it's still impossible, and therefore silly to value.

*I regard Pereboom's definition of free will in LWFW as only trivially different than the impossibilist kind that G. Strawson discusses.

[Mark Smeltzer]

Kip,

Sure, I understand what the impossibilists are trying to do. They are able to do so because they do not value "free will" as they define it (and who would??) -- otherwise they would be epistemically obligated to reject their own definitions. However, to people who do value free will, these arguments should provide little to no evidence that the impossibilsts have found *acceptable* definitions of free will.

And this epistemic obligation to resist movement from belief to unbelief is owed to, at least in part, the argument presented above (though certainly a positive account for free will is desirable).

To put it in context, because (1) I value free will and (2) I do not value in the least that which the impossibilists have defined (seeing as their definitions point to impossible powers/situations), I am, therefore, epistemically obligated to reject impossiblist definitions of free will (for as long as I continue to value free will).

You're certainly right that this method of argument does not rely on the objective value of "free will", but look at how much can be accomplished regardless. (Although the argument does seem to be playing on a kind of "objective" value with regard to epistemic duty, but if need be I think this duty could be sufficiently grounded subjectively as a quality of rational beings.)