Blog Coordinator

« SpaceTimeMind Episode 7: The Illusion of Free Will | Main | Winding Down »



Feed You can follow this conversation by subscribing to the comment feed for this post.

I agree with everything you say, Kristin. I'll add a few points.

1. In other words, arguments for incompossibilism tend to merge together, once we recognize that agent causal libertarianism (ACL) is incoherent. Here, it is important to stress that ACL is incoherent in the sense that it cannot provide more control or responsibility that compatibilist powers do. Another way to say that is that, even if determinism is false, the laws of logic (which still apply in indeterministic worlds) prevent the kind of self-creation needed for free will.

2. I would stress that free will doesn't require total self-creation. That sounds unreasonably demanding. Fischer calls this Total Control, and he has a point. He says that we don't need control over whether the sun exists in order to have free will.

Constitutive luck skeptics (CLSs) should respond to Fischer by narrowing what is required for self-creation, while still preserving the original skeptical insight. I think you can narrow it quite a bit, without losing the original thrust of the Basic Argument.

To have free will, you don't need to control the sun. And you don't need to create your own limbs and your own body and your own skeleton, out of nothing.

But, psychologically, for any terminal goal that leads you to make a decision, you need to be able to evaluate that goal. But you can only evaluate goals in light of other goals. So, at some point, you were simply given a terminal goal (or goals), without other goals or values with which to evaluate those goals. And *that* is the problem of constitutive luck.

An important corollary of this (narrower) argument about constitutive luck is what is called the Orthogonality Principle. Artificial Intelligence researchers refer to it when discussing the motivations and actions of computer programs. The basic insight, per Hume, is that reason and passion are separate, so that reason can only do what the passions suggest. In other words: humans, as rational creatures, use goals/values as inputs to their rationality, but rationality-without-goals is powerless to evaluate those goals. At some point in history, we are simply given goals/values, and we live our lives according to them, and there's nothing we can do to scrutinize or stop them.

You can read about the Orthogonality Principle here:

What the Orthogonality Principle shows is that "bad" original terminal values are non-recoverable. If someone is born with "be a serial killer" as a basic terminal value, then no power of rationality is going to be able to persuade that person to revise that terminal goal. Yet, that original zygote/fetus was just given that terminal goal - it was powerless to evaluate and/or filter it. That's the essence of the argument.

Hi Kip,

I must say that it sure is great to have you kick off these threads! It’s hard to tell if I’ve actually persuaded you of something, or if I’m just preaching to the choir—but, either way, you won’t hear me complaining!

As for your summary, we're on the same page. I might push a little bit on whether a logical principle prevents us from being free, but that's just a quibble. I think that the terminology we have been using has been playing into the hands of the incompatibilists, and has obscured that there are actually no *arguments* for incompatibilism.

The Orthogonality Principle you mention doesn’t ring any bells, but sounds interesting. I’ll have to take a look at that. Thanks.

I agree that there is room for debate over the exact sense in which one would have to “self-create ex nihilo” in order to be free, so thanks for bringing up that point. Indeed, I think that it is in exploring this question that we get insight on how the problem of constitutive luck might be overcome.

Specifically, I think once we turn our attention to this topic, we begin to see the importance of epistemic conditions of freedom (think: neo-Aristotelian autonomy; I’m also sympathetic to much of what Eddy Nahmias has said on this issue). I’m not a skeptic (quite yet), and I think a person might overcome constitutive luck through the acquisition of knowledge (that and how). The major problem with going this route to solving the problem of constitutive luck is that ideal freedom looks really boring… I’m currently working on that. I do feel certain, like you, that we aren’t going to solve the constitutive-luck problem on the cheap by appeal to indeterministic breaks in otherwise deterministic causal chains.

What about you? Do you think there’s any real hope of solving the problem—or are you pretty much resigned to “constitutive luck skepticism”?

Hey Kristin -

Some points:

1. Yes, you are just preaching to the choir.

2. I like your terminology "incompossibilism" (vs. incompatibilism), but I am not prepared to disagree with incompatibilists. Traditionally, incompatibilism is simply the view that free will and determinism are not compatible. See the Stanford Encyc. Philosophy, which states "[i]ncompatibilists hold that we act freely in this sense only if determinism is false." You're right to point out that, on many occasions, incompatibilists have regarded determinism as a specific threat or best explanation for free will skepticism. But that does not, in my view, justify redefining incompatibilism to mean something narrower (i.e., that determinism is a specific threat that undermines free will in a uniquely meaningful way - or something like that). In other words, I'm a formalist about terminology. I disagree with definition shifts like the one you propose, for the same reason that I disagree with revisionists and compatibilists about using watered-down definitions of free will. Your raise valid, and very fruitful, insights, but I would not coopt old labels to make them.

3. I am absolutely convinced that constitutive luck is a moral problem and that it cannot be meaningfully solved (i.e., argued away). The only "solution" is to adopt a kind of radical consequentialism, like Pereboom proposes at the end of Living Without Free Will.

I do have some skepticism about free will skepticism, but that is because I'm not convinced 100% that free will refers to the-power-to-choose-otherwise-that-overcomes-constitutive-luck. I'm not sure what free will refers to exactly - I suspect that the term is under-defined and that this vagueness explains why philosophers keep arguing for thousands of years.

Personally, I was always inclined to define free will in terms of the constitutive luck problem (well before I knew the term "constitutive luck"), because that was the most interesting problem to me, and free will was the most elegant term to describe its purported solution.

But I don't know how to prove that my definition is right to a compatibilist. If Dennett wants to say that free will means X, and I say that it means Y, it would take a lot of empirical work to prove Dennett wrong. That work either hasn't been done, has produced muddled and inconsistent results so far (see Eddy and Thomas's work on measuring free will beliefs). So that makes me modest as a skeptic.

Nevertheless, I'm convinced that constitutive luck is the most interesting, perhaps the only interesting, problem in the free will debate, and that it's best to frame the debate in those terms. Einstein, Darrow, Spinoza, Darwin, Pereboom, Strawson... I see all of these people as concerned with constitutive luck, whether they realized it or not. It's just a red herring to focus on other threats, like Libet's gaps in consciousness or neurocompatibilist limits on our rationality.


4. I do encourage you to read the article on the Orthogonality Principles. If I'm right, you will hopefully find it as fruitful as I did in understanding what specifically is the threat to free will, and how skeptics can require less than Total Control, as Fischer suggests.

Hi Kristin,

Let me ask a clarifying question. You'd like to, more or less, group these collected arguments together, as what they all do is point in one way or another to "our shockingly difficult problem with constitutive luck". Could you say a bit more about how you take the basic argument to do this?

You might think its that the basic argument shows necessary that to be responsible for action you need to be responsible for the psychological explainers of action. And since there's just one way to be responsible for something, acting so as to bring it about, you need action to be responsible. Rinse and repeat, and you're all too quickly an infant who's not responsible for anything.

But, of course, the basic argument doesn't *show* this at all. What it does is assert three things: one very plausible claim about psychologically explaining action, and two substantive principles. Famously (infamously?) it is precisely those two principles that Strawson gives zero defense for.

So, I wonder whether you think a defender of BA needs to do more, or whether Strawson's version is sufficient?

I agree that there is a basic problem to be explained. If we can be morally responsible for what we do, then we have to explain how this comes about since we don't start our lives as responsible for what we do. This is an interesting problem, and one that crosscuts the traditional lines of the debate. But as it stands, it also isn't a problem about constitutive luck, since the puzzle it proposes doesn't presume any particular obstacle to the solution.

Thanks, Kip.

You say “Traditionally, incompatibilism is simply the view that free will and determinism are not compatible. See the Stanford Encyc. Philosophy, which states ‘[i]ncompatibilists hold that we act freely in this sense only if determinism is false’.”

I’m not sure which SEP article you had in mind here, but a few points:

1. As I've said, the terms 'compatible' and 'incompatible' are ambiguous and are used by different people to mean different things, so I don't know how to understand claims invoking these terms when not defined by the user. I don't find the follow-up definition to be of much help. It clearly implies incompossibilism, it implies that the falsity of determinism would somehow "help", but is logically consistent with impossibilism and the view that the truth of determinism is irrelevant to free will. I find that confusing.

2. If revisionism is a worry, then it seems that your proposed definition also seems revisionary, insofar as it suggests a positive condition under which persons *can* act freely. ‘Incompatibilism’ is a technical term that was, as best as I can tell, introduced by van Inwagen to name the negation of what was often called “the consistency thesis”, the thesis that determinism and the free-will thesis are logically consistent. This consistency thesis later picked up the name “compatibilism” and “consistency theorists” became “compatibilists”. van Inwagen wanted a name for the view that the consistency thesis/compatibilism is false. He chose ‘incompatibilism’ (and seems to have first used it in print around 1972 in a response to essays in which this term did not appear).

van Inwagen’s definition does not logically entail that free will can exist in indeterministic universes, or at all. All skeptics are clearly incompatibilists in his sense of the word. But consider the characterization of incompatibilism you give: “we act freely in this sense only if determinism is false”. Wouldn't it be quite misleading for a constitutive-luck skeptic to say they hold this view? The CL-skeptic doesn’t think that we act freely *even if* determinism is false. Again, I just think it would be nice to clean up our terminology just to enough to avoid this lack of clarity.

4. I know we have similar views about the big picture, but I’m not sure how to read your complaint. Do you think I’m overstating how widely mischaracterized/misunderstood the major “arguments for incompatibilism” really are? If people are already quite good at tracking the difference between the views I've called incompossibilism and incompatibilism, then why doesn’t Pereboom explicitly invoke his defense of possibilism in the best-explanation argument in his manipulation argument? Why didn’t philosophers immediately demand that he do so? Why don’t incompatibilists take pains to run CA on the assumption of determinism *and possibilism*, and give us a best-explanation argument to close the logical gap? I think it's because incompatibilists think they get incompatibilism “for free” from any defense of incompossibilism. They don't. This is very easy and intuitively said with my new terminology. Could say a bit about what how you would prefer to get the same work done?


I wrote several big paragraphs about several different points, one of them slightly critical - and that's the one you've focused on (with four replies!).

Let me clarify.

I'm a formalist about terminology. I don't believe in fudging definitions for practical reasons - it causes a lot of mischief in philosophy.

I do agree that the SEP article definition I cited is problematic. To give you another definition, see Routledge:

(To be fair, I think Galen Strawson wrote that article!)

Consider also the SEP article on compatibilism:

Both of these articles generally define compatibilism as simply the compatibility of free will and determinism. And incompatibilism is the negation of that thesis.

In other words, I agree with you 100%: "van Inwagen’s definition does not logically entail that free will can exist in indeterministic universes, or at all. All skeptics are clearly incompatibilists in his sense of the word." As far as I can tell, the ViW sense provides the consensus or definition of "incompatibilism."

I also agree that, on occasion, some incompatibilists have suggested something stronger than incompatibilism. They have suggested something like: (A) free will is logically possible and/or (B) determinism is a specific threat that uniquely undermines free will, as a best explanation (or something like that). In my view, these deviations are simply not frequent or explicit enough to suggest that the term "incompatibilism" means anything more than the mere inconsistency thesis. Instead, it means that, in addition to being incompatibilists, these thinkers also adopted some additional thesis - call it "specific-threat-incompatibilism," or something like that.

So I agree with you that there is some confusion about what certain incompatibilists believed caused the threat to free will. Where I (tentatively) disagree is that I don't think this confusion justifies giving "incompatibilism" the specific-threat-definition. Instead, I would simply leave incompatibilism as the term denoting the incompatibility of free will and determinism (however that occurs, even if it occurs in a merely technical sense, because self-creation is impossible). If you wanted to refer to another, more specific thesis (i.e., that determinism is specifically or uniquely menacing to free will in a way that indeterminism is not), then I would adopt another label for that term. Call it Kristin's Incompatibilism, or whatever you like.

At some level, we are just arguing semantics, and that might seem trivial. But as federal judges like to say in patent law (I'm an attorney), million dollar litigations can turn upon the definitions of words...

Hi Matt,
1.I’m not sure how to take your first question. I don’t think take it that CA, global manipulation arguments (MA), and the Basic Argument (BA) are about constitutive luck only if the BA is sound. As I tried to say in the post, I think they are *even on their faces* arguments about constitutive luck—once we get passed all that determinism talk.

“Determinism” appears in CA and MA because there is a worry that constitutive luck is inescapable when determinism is true. My claim is that neither CA nor MA give us reason to think that *only determinism* prevents our escape from constitutive luck. If indetermination doesn’t “help”, it seems hard to argue that deterministic laws “hurt” merely for lack of helping--and where's the independent argument for the conclusion that determinsitic laws hurt?

The Basic Argument comes in because it identifies a ground of non-freedom that suffices to underwrite a Rule Beta sort of transfer principle without appealing to any menacing transfer mechanism. That, I take it, is one of the key steps in closing the case for the conclusion that the diachronic evolution of the universe doesn’t matter to free agency—even if CA and MA are sound.

2. Would you mind stating the two “substantive principles” you have in mind?

3. As for the “psychological principle”, I think that Strawson does defend this, insofar as he sometimes (rightly, I think) summarizes BA as the conjunction of CA and some version of the Mind Argument. When he does this, Strawson is often read as saying that deterministic laws “hurt” and indeterministic laws merely “don’t help”, but that’s got things inverted.

CA concludes only that deterministic laws don’t help (it doesn't pinpoint the laws as a positive threat), but all versions of the Mind Argument tell us that indetermination actually hurts. That is, any attempt to add indetermination into the mix in order to solve the problem of constitutive luck would do nothing more than provide a new ground for non-freedom, making it overdetermined a person lacks free will when he acts. Even a causa sui (in Strawson’s sense) would have her free will diminished/undercut by indetermination in key elements of the deliberation process leading up an action if some version of the Mind Argument is sound. (Notice: a manipulation argument couldn't be run on a causa sui--a causa sui would fill the role of the "manipulator" for herself! And, as Joe Campbell's "No Past" objection points out, CA requires some sort of grounding principle--the poison has to be poured into the cup at some point.)

So, in short, I think the Basic Argument does roughly take care of the business it is supposed to, burden-of-proof wise. Neil Levy has continued to flesh out the Basic Argument, arguing that there really isn’t any more modest or “adequate” sourcehood available--any attempt to avoid constitutive luck sinks you with present luck. I consider this argumentation/clarification to fall under the heading the "Basic Argument".

Now, as I’ve said, I personally think the constitutive luck problem can be solved. I’m working on the relationships between all these arguments because I want to have the clearest possible picture of the battles I’ll need to win in order to defend compossibilism. I think that people who haven't taken constitutive luck seriously before (as I didn't until recently) don't realize that *even their standard arguments for incompatibilism* are arguments about constitutive luck. I was sort of surprised when I saw that clearly... or, er... when I *thought* I saw that clearly.

Since you think I'm off track, would you mind saying a bit more about what I've said wrong or the gaps you think I should fill? I'd really appreciate it!

Thanks, Kip!

I think I’m guilty of falling short on blogging etiquette--sorry! I have much appreciated your support and constructive criticism this month. I’m really sorry if I’ve downplayed that too much. I’m from Garrison Keillor's Midwest--we're trained to publicly downplay all claims of success up here!

As for focusing on your quick criticism of my terminology, that’s honestly just because I was so glad you were willing to bring up the point. To lots of people it seems like a totally trivial detail, but I'm glad you don't agree. So far, many have shared your concern that there’s something awry with using ‘incompatibilism’ as I do (which indicates that people care about the term itself more than they might want to admit). I'd like to fix this, if I can.

I like your “specific-threat incompatibilism” suggestion. I’ll have to think about that a bit more. I get worried when it comes to tweaking names of views like Pereboom's "source incompatibilism" (which is incompatibilism in my sense, i.e. the deterministic-threat sort). I'll keep working on it, though. Thanks for the suggestions!

Hi Kristin,

Ah, the dangers of writing so early in the morning! I certainly could've been clearer. Let me see if I can do better:

So Strawson says three things which generate the regress that underlies the BA. This is as I recall them, but as I'm given to misremembering, let me know if I've gotten something wrong.

1) What we do is at least partially caused/explained by how we are psychologically.
2) To be responsible for what we do requires that we are responsible for how we are.
3) To be responsible for how we are requires an action to make ourselves that way.

Since (3) requires action, which 'depends' on psychological states, we'd better be responsible for those states, but there's only one way to do this, via action, which returns us to some previous psychological states.

I take it that (1) isn't issue. I take (2) & (3) to be substantive claims, however -- at least, their negation doesn't seem crazy. Why suppose that the only way to be responsible or something is to act so as to bring it about? Why suppose that responsibility for action requires responsibility of attitude?

It isn't that these are implausible claims, but they don't strike me as obviously true. And, as I recall, Strawson explicitly gives them no defense. So my clarificatory question was whether you thought any defense of either was necessary, or whether you take those principles to be true on their face.

I may be being overly critical with your uses of "identifies" or "shows", (as opposed to "purports to identify/show") -- if so, I apologize.

I think this connects up with your comments on constitutive luck, but I'll leave things here to make sure I'm not already mistaken.

Hi Matt, and thanks for the follow-up. This is helpful.

Yes, I agree that 2 & 3 are substantive claims and that their negation doesn’t seem prima facie crazy.

But I disagree with you about whether Strawson/proponents of BA offer any/adequate defense of them. As I see things, (2) and (3) are usually given a decent negative defense that shows the cost of denying the truth of these claims, even though this defense (e.g. appeals to the Mind Argument) is not included in the standard formal summary of the argument. Even if proponents of BA haven’t done enough to show that denying these premises is totally absurd, I tend to think that proponents of BA have done enough to shift the burden of proof onto those who would deny that these claims are true. Having seen the positive stories offered by libertarians (e.g. Kane’s “self-forming actions”) and compossibilists (e.g. Fisher’s “ownership” conditions), I don’t think those efforts have panned out very well. Those who think otherwise will disagree with my view about where the burden of proof currently lies. Does that seem fair to you?

I feel like a trap is ready to spring… what trouble have I just made for myself?

Hi Kristin: Maybe this is a bit of axe-grinding on my part (probably so!), but I'm not sure we should still consider the Mind Argument a viable argument to appeal to in discussions of free will. See e.g.:

Say we find incorrigible evidence of how human beings work. Would that solve the FW problem? No. It would merely shift all real questions about choice and responsibility to the realm of axiology--what is valued about human action, and what that implies about assigning responsibility based on value-guided judgments about choice. This is why, for example, Manuel Vargas' revisionism is so important. It works from the empirical/value side of the question to elucidate what FW and responsibility must mean rather than working from the metaphysics side to value judgments. I'd argue as well that one reason Frankfurt is so important is that his metaphysical challenge to PAP effectively placed values of action prior to any given reality of how minds work. Fischer refined that insight on the side of responsibility to dismiss the freedom question explicitly. But I favor Vargas here because his work reveals best what I think is the real methodology at work here overall: pragmatism. The line of Frankfurt--Fischer--Vargas is one that I think reveals the ultimate victory of pragmatism, emphasizing the role of values over any particular metaphysics. Such a pragmatic response to FW questions is the only one that can even hope to reply to the exquisite criticisms by Neil Levy on the issue of constitutive luck.

Thanks, Kristin. That helps. And I'm certainly not trying to set up a trap. :)

One way to take the BA is as a challenge. It says, "Show me how you could be responsible for how you are, when you started off not responsible for anything." This, at it's most basic, is the puzzle I do think every non-skeptic has to explain. And it is a puzzle. We might then ask, what are the prospects for answering it?

There are two ways you might go about answering this question. One is to look at extant solutions and evaluate them. It seems you find extant solutions wanting. That's fair. But that just shows us is that we'll need better proposals to find a solution, and maybe it tells us something about how *not* to try to solve the puzzle.

The BA proposes a different answer to the question, which is to say it can't be solved. The solution is impossible. This is a strong claim, so we should consider what reasons we have to believe it. The claim rests on those two substantive principles. But we can't defend those principles by pointing to failed efforts to solve the puzzle. The failed efforts just show us that we don't have a solution yet. And I don't see how showing that indeterminism is irrelevant helps here, for we needn't have thought that the only way to solve the puzzle is indeterministically.

So I'm still missing how (2) & (3) are negatively defended by those appeals. This seems important to me, because if you reject either (2) or (3), the argument doesn’t go through, and so fails to show a solution to the puzzle is impossible. Am I missing something?

(I should add that I worry that my concerns here are distracting from the central thrust of your post. Apologies!)

Marcus & Alan:

Yes, Marcus, you're very right to point out that the Mind Argument has its problems. I think some version of it (or some nearby argument, depending on your argument-individuation principles) is sound. My view is that I'm permitted to hold onto that hope so long as I haven't seen any positive account which comes close to being a free but undetermined action--which I haven't (to *my* satisfaction). I don't think libertarians have yet delivered a positive theory of freedom that provides contrastive explanations for what I do, and (perhaps for lack of imagination?) I don't see how they ever could.

I am something of a control freak when it comes to freedom, and indetermination just looks antithetical to freedom-as-control. My intuitions have led me down the path to thinking that freedom isn't about the excitement of a future "garden of forking paths" with actual-sequence possibilities or any alternatives at all. It is, though, about doing something to overcome what Levy calls one's "endowment", even if one's endowment was a good one and one ultimately settles on a character that is very much like the one given to him by nature. I'm very Aristotelian in this regard.

Ideal freedom, as I see it, would correspond with the loss of all interest in alternatives. This is why determinism has never struck me as a threat. Even if determinism is true and we have what Levy aptly calls an "endowment", we can LEARN. How we learn, what path we take to knowledge, is unique depending upon our endowment. But interesting things happen when we start to learn A LOT about A LOT, and especially interesting things happen as we approach expertise (even in one domain). Expert reasoning and decisions tend to look very similar, despite diverse backgrounds of the individual experts. I think there's a lesson there.

In short, I want a freedom modeled on expert decision-making--e.g. the free choices made by my surgeon during my open-heart surgery or by Kasparov while playing chess. I want *that* sort of freedom--a freedom from ignorance making bad choices look good, a freedom from mistakes. That's why I develop my self-control, why I study, why I practice. I want to guarantee that my mind is prepared to make the *right* choice in whatever situation I find myself. Sadly, my life will be short, I had a slow start, and I will not likely get very far in overcoming my endowment. But my kids may live much longer than I--maybe hundreds of years more. Might that extra time make all the difference? What should I tell them about how to think about their potential for become free agents? I think we often get so caught up in thinking about finding freedom between now and then, that we fail to draw on intuitions about what might happen *after* then and now.

I'm not sure, Alan, if this will turn out to require some revisionism. I tend to think it won't. I *hope* it won't. I agree that empirical studies on the brain aren't going to deliver an answer to the question of whether we have free will. However, I would say that with every bit of new knowledge about this human-body machine, I learn how to avoid its built-in biases and shortcomings. That sort of knowledge strikes me as a growth in freedom-relevant, endowment-escaping power. Let me extend my mind with some computer chips and give me access to some genetic therapy, and maybe I could speed up my journey to freedom even more? (I'm gonna bite my tongue on current X-Phi on free will.)

I do think that we have to get our metaphysics squared away.
That's one of my biggest complaints about the current theories that talk about freedom-relevant alternatives epistemic terms (e.g. Wolf, Nahmias). I think there are real monsters looming down in the metaphysical depths, and they threaten to jump up and devour any hapless theory of freedom/responsibility that doesn't take them seriously. Of course, as a metaphysician, you'd expect me to say that. ;)

That said, I agree that we are, or at least *I* am, often working back from something of value (for me: excellence) to what a satisfactory theory of free will would have to look like (me: control). I think that's a really good thing to keep in mind. Thanks!

Hi Kristin: Thanks for your reply!

You write: I don't think libertarians have yet delivered a positive theory of freedom that provides contrastive explanations for what I do, and (perhaps for lack of imagination?) I don't see how they ever could...I am something of a control freak when it comes to freedom, and indetermination just looks antithetical to freedom-as-control."

My reply: I think your line of thought here--which is also broadly the line of thought pushed by the Mind and Luck Arguments--(1) begs the question against libertarianism, and (2) there are empirical reasons to think it is unsound. Allow me to explain.

Go back to my earlier comment, on electrons and CAUSA SUI. What *ultimately* explains why electrons behave as they do? Gravity, electromagnetism, etc are all *proximate* explanations for why electrons behave as they do. That is, GIVEN the laws of nature (of gravity, electromagnetism, etc.), electrons behave in such-and-such a way. But what explains the FACT that electrons *obey* those laws of nature (this law of gravity, electromagnetism, etc.). The answer, I believe, is simple: the laws of nature are PRIMITIVE. Nothing can, in principle, *explain* them. They are the "ground floor" which all further explanations must take a basic, or fundamental.

But now let's think about this. If laws of nature are primitive--if *nothing* can explain them--then an analogue to the Mind and Luck Arguments would imply that there can't be any laws of nature. Why? Well, because it is "incoherent to suppose that something--electron behavior--can come from nothing." But this is wrong. Electron behavior doesn't come from nothing. It is explained by PRIMITIVE laws of nature.

But now once we admit that certain types of causation are primitive--and I think there is no way around this--it is perfectly possible that our MINDS are complex CAUSA SUI in a higher reference-frame leading in *this* (our reference frame in the physical world) to indeterminacy and quantum collapse. (Again see ).

On this model, although all physical information is "set", our MINDS--in a higher-level reference frame--may well be entirely self-constituted Cartesian "thinking things", possessing a primitive capacity for thought and free will, which then interacts with physical information to cause very subtle neurological changes (changes to dendrite formation and microtubule resonances) leading to libertarian-free behavior.

Now, of course, you might think this model is preposterous. After all, it affirms a kind of Cartesianism about the mind, and Cartesian interactionism with the physical world, that many people think have been effectively disproven. However, as I have argued in "A New Theory of Free Will", it *hasn't*. The model I defend resolves a lot of physical and philosophical problems, and improves upon existing interpretations of quantum mechanics). Furthermore, your claim--and Kip's claim, whenever he gets the chance;)--is that this is not just preposterous but a *logical impossibility*. I think I've just shown that it's not. Although libertarianism may appear to entail a "disappearing agent" and "luck" problem in OUR (physical-world) reference frame, it is perfectly possible for minds to be SELF-FORMED mechanisms of pure conscious thought in a higher-reference frame. I see no logical impossibility here, any more than it is logically impossible for electrons to obey laws of nature "just because" (with no deeper explanation).

@ Matt:

About the "trap", that's just my philosophical playfulness coming out. A facepalm is a small price to pay for getting closer to the truth, so I love (good-spirited) criticism.

You say "One way to take the BA is as a challenge", and I am in TOTAL agreement! That IS how I take it. I think you're right that the argument depends on some potentially false principles--and thank goodness!--but we seem to disagree about burden of proof. I think that proponents of BA, by pointing to arguments that indetermination hurt, to arguments that determination doesn't help, and "by pointing to failed efforts to solve the puzzle" overall, does provide a negative defense of the principles. These things do not constitute a positive argument *for* these principles, of course. So the proponent of BA has no positive defense of (2) and (3) and I have no successful positive theory to show they are false--but it's not a stalemate because those principles sure look true (at least to me). So, if I want to reject the principles, I feel that I have more work to do. Does that seem fair?

I keep putting my skeptic hat on because I want to see BA's challenge as clearly as I can. I don't want to be fighting a straw man; I don't want a technical victory. I want to meet the spirit of the challenge head on. So, I'm pleased that Kip, a full-blooded skeptic, thinks I've done an okay job of making the case for skepticism.

(I had hoped to hear from someone who would argue that CA is a free-standing defense of incompatibilism (as I understand it), since it is so often described as such. Maybe such incompatibilists figure I'm a lost cause? Or maybe they are fine with my characterization of CA? Can't tell.)

In any case, I too hope that the challenge of BA can be solved, that it rests upon false principles, and that constitutive luck "washes out". I was going to make a final post on that, but it's been such a fun month...maybe I ought to wait on my positive project.

Hi Marcus,

I appreciate that you are drawing on a highly-developed theory, and I am sorry for being unfamiliar with the details.

As you summarize your theory here, though, I'm not sure how it supports your charge that Mind/Luck arguments beg the question against the libertarian's proposed theory of free will. I also don't see how empirical evidence supports that indeterministic causation in the reasoning of the causa sui you described (the decision that led the agent to causally infect a distinct, physical reference frame) is free.

I totally accept that "miracles" relative to a closed physical system are metaphysically possible--laws account for what happens in a physical system so long as the system stays closed, but they don't *keep* the system closed. So, if your reference-frames account allows, in effect, for miracles relative to the physical laws of nature, I'm fine with that. But if we are miracle-workers and causal-closure principles are false, then why think that deterministic laws are a threat? Sure, *determinism* must be false for your story to work out, but determinism rules out miracles by definition. If we buy miracles, then deterministic laws can obtain even when determinism is false. Why should miracles only be possible when the physical laws are indeterministic? Why, that is, should one be a libertarian?

Also: What are the laws like in the reference frame in which the decision is made? No laws? Indeterministic? Deterministic? Haven't you just moved the bump in the rug, as they say?

Also, as I said, I don't accept a realist interpretation of the Standard Interpretation of quantum mechanics. I don't think there is a deep "measurement problem". I take Schrodinger's cat as it was intended: as a counterexample to realist interps of SI. It's a quirk of history that anti-realist philosophy of science was dominant when SI was being developed, and funding for anti-realist research programs has since been hard to come by. Maybe it's for the best--much progress has been made after all. But, the point is that I'm not persuaded that our minds are somehow free because they are responsible for quantum collapse: I don't think that quantum collapse actually happens, and I think that for reasons that are entirely unrelated to particular views I have about free will.

There's quite a lot going on here!

But perhaps I could jump in with a remark about Alan's comment. I do not dismiss freedom as relevant to moral responsibility. I distinguish two kinds of freedom (as does Frankfurt), and I (as also does Frankfurt) hold that we need a kind of freedom (in my case, guidance control) for moral responsibility. So I don't eschew metapahysics in favor of values--but only a certain interpretation of what metaphysics requires. I'm guessing that Manuel Vargas would agree, but I'm not sure. In any case, Vargas does develop a version of a reasons-responsivenss account of moral responsibility (sometimes he calls it a "moral-consideration-responsiveness" account"). This is at a certain level very similar to my account of guidance control I think Vargas thinks that this condition must be met in order for an agent to be morally responsible. I don't see how this is "pragmatism" or the priority of values over metaphysics.

In general, I find it kind of unclear what the distinction is supposed to be between "metaphysical" and "normative" approaches to moral responsibility. I try to address this in one of the essays in *Deep Control*.


I'd be curious as to exactly why you do not think my "ownership" condition "hasn't panned out very well". Granted, there are certainly issues with it, but I'm not sure I believe the notices of its demise. In any case, I was wondering on what basis you worry about it.

Kristin: Thanks for your reply!

You write: Also, as I said, I don't accept a realist interpretation of the Standard Interpretation of quantum mechanics. I don't think there is a deep "measurement problem". I take Schrodinger's cat as it was intended: as a counterexample to realist interps of SI."

My answer: If you read my paper, you'll see what's wrong with this. As I explain (and this is not something only I understand) the standard Copenhagen interpretation is incoherent in two different ways, and moreover, asserts a fact that is unmeasurable and unexplainable in principle (i.e. the quantum wave-function). My new model resolves these problems.

You write: "As you summarize your theory here, though, I'm not sure how it supports your charge that Mind/Luck arguments beg the question against the libertarian's proposed theory of free will. I also don't see how empirical evidence supports that indeterministic causation in the reasoning of the causa sui you described (the decision that led the agent to causally infect a distinct, physical reference frame) is free...I totally accept that "miracles" relative to a closed physical system are metaphysically possible--laws account for what happens in a physical system so long as the system stays closed, but they don't *keep* the system closed."

I reply: On my account, our reality is functionally akin to an ordinary online videogame. Let's think now about (1) the phenomena observers *within* an ordinary online videogame would observe, and (2) the complete system both *within* their frame-of-reference and without (i.e. our frame of reference outside of the simulation).

With respect to (1): existing online games present observers *within* the game precisely what *we* see. Online simulations have physics (for how rocks, bullets, gravity, etc.) behave in the simulation. Moreover, there are often "animals" in simulations, and they behave very much like animals in our world. Scientists within such a simulation would thus theorize that their world is governed by a (seemingly) deterministic physics: gravity, mass, etc. There would be only *two* things in their reality that they could not predict with certainty. First, due to the system link at our reference-frame--the internet architecture--there would be inherent noise/indeterminacy from their reference-point (viz. a quantum wave-function). Second, there would be ONE type of creature whose behaviors would not fully conform to that law: namely, the characters whose actions we (the outside users) are controlling. Our joystick movements *outside* of the simulation lead to non-random behavior that cannot be predicted IN PRINCIPLE within the simulation. Thus, if scientists in such a simulation were to investigate "their brains", they would witness small violations of the "ordinary quantum wave-function." This is where libertarian free will (vis-a-vis their reference frame) would enter the world (via our choices outside of the simulation. FINALLY, and most importantly, the way simulations work, even this aspect of their world will appear to them to be deterministic (you can "rewind" a simulation to the beginning, and replay it so that everything replays itself as if everything were completely *inevitable*). In this way, genuine "libertarian" inputs from outside a system can give rise to an *illusion* of deterministic causal closure in a lower reference-frame.

Finally, you ask: what would the laws of nature be like in the higher reference-frame I'm talking about? Well, in the paper, I say there are two possibilities: (1) the higher realm is deterministic itself, or (2) it's a realm of *pure* Cartesian CAUSA SUI thought, where our minds are literally disembodied souls that SELF-CAUSE. Although in the paper I contend we can probably have no evidence either way (in which case I say no one can ever fully verify *or* falsify libertarian free will from our reference-frame...though my model does make three empirical predictions which might give us some clues), I have now developed an (unpublished) argument in which I argue that it must be the latter: Descartes was right; we are disembodied, fundamentally qualitative, CAUSA SUI souls interacting with quantitative physical information (akin to a laser mechanism reading information off an ordinary DVD from the outside).

I'm sure you think this is all insane. But again, the model (as I defend it in "A New Theory of Free Will") resolves and explains a ton of problems--including why there is quantum indeterminacy, wave-function collapse, wave-particle duality, etc. in the first place--that no other theory presently provides a good explanation for.

Finally, it's worth noting that the more *science* proceeds--what with General Relativity and quantum phenomena--the more it looks like we ARE living in a completely preposterous universe. What's one more preposterous thing, if it explains a bunch of stuff we can't explain now?

Hi John,

I walked right into this one, didn't I?

I will start by saying that I think reasons-responsiveness is where it's at and that compossibilists need ownership conditions that supplant BA's "starting-point condition" when it comes to ultimacy conditions. I think your ownership condition serves as an interesting marker along the road to developing freedom/responsibility. (I think of freedom/responsibility as falling on a spectrum.)

I just think that your ownership condition is very weak--people qualify as free/responsible who have made virtually no progress with regard to overcoming their "endowment". A minimal bar for free-ish action is great, but I'm looking for the upper end: Just how much can we get? I see the methodological reasons for setting the bar where you do, but (from a marketing standpoint at least) I think the plausibility of compossibilism depends upon showing that people subject to deterministic laws can satisfy a much more demanding ownership condition than you offer--and I think they can.

In my view, most people who have ever lived have had barely a glimmer of freedom/responsibility in their daily lives. As such, I think your ownership condition maps onto how much freedom we actually have. So, that's good. But it's not entirely satisfying, you know what I mean?

John, I apologize for my too slick and clumsy characterization of your views. My only point was that I can read a genetic line (as it were) through the Frankfurt-Fischer-Vargas literature (and more) that is consistent with an evolution of value/pragmatic concerns that I am increasingly becoming enamored with. I hope at least folks might see why I find a pragmastic approach a fertile one for dealing with the deep concerns about luck that are fast becoming a centerpiece of FW/MR discussion. Again, I did not mean to misrepresent your views (and I think I do really understand them). And apologies extended to Manuel too if I have twisted his views to unrecognizable form. His "Building Better Beings" has had a considerable influence on my thinking however.

Kristin, you have done such a great job not just stimulating discussion but carefully replying to everyone. Terrific stint as guest blogger!


Hmmm. I actually think most people are fully morally responsible for what they do. And I think my account maps nicely onto that. So we agree with the second, but not the first sentence. I guess that's good--but it is a mystery.


No worries. Most people who invoke the distinction between metaphysical and normative approaches are operating at a pretty vague level, in my view. It is interesting to try to specify the distinction a bit more, though.

And I agree: Kristin has done a very nice job, and is to be commended.

But. I just got an email from Al Mele, who informs me that Diana actually set things up 100 billion years ago so that causal determinism is true, all compatibilist-friendly conditions are met, and Kristin blog exactly as she actually has. I told Al, "Well, that's interesting, and thanks for sharing. But it doesn't change my views about Kristin at all--she's done a great job and she deserves credit for it."

And I'm right!! It doesn't matter at all that Diana had the intention she had 100 billion years ago--we could change the story so that it was accidental, or that she meant to set things up different but did it in this way, and so forth. What difference does it make what some agent intended 100 billion years go, or that that agent set things up to get that result?

So, Kristin, rest assured: you are cool.

Kristin: sorry for the errant capitalizations. I have an enthusiasm problem. ;)

Hi Kristin. I've just been catching up on your (very interesting) posts from this month. While I don't come down strongly on either side, it would be sad if you spent a month trying to spark a debate between incompossibilism and incompatibilism and couldn't find anyone to defend the latter, so the following is a limited attempt, in two parts.

Firstly, suppose physicists found, by studying the dynamical laws that ordinary matter obeys, which happened to be deterministic, that the universe must have always been increasing in volume, from an initial point-like state, with the evolution over time being deterministic at all times and in all places. I don't think this possibility has quite been realised (yet), but it's not very far from actuality. In this case, the laws of physics alone would determine the state of the universe at all times. Let's call laws like this 'superdeterministic'. You could draw up a consequence argument to the effect that, since the superdeterministic laws are not up to us, neither are our acts, which are their necessary consequences - this works without the need to invoke contingent facts about the distant past. (I've phrased everything in terms of up-to-us-ness because that's how you stated the consequence argument before.)

I expect you would still say that this is a disguised argument from luck - the luck is in the laws we happen to live under - but I'm genuinely unsure if you'd think it is a successful argument for a weak form of incompatibilism, viz. the position that up-to-us-ness is impossible in the presence of superdeterministic laws *because* of the superdeterminism. I can see two ways you might go: one is to agree that, if we lacked up-to-us-ness under superdeterminism, we'd lack up-to-us-ness because of it; the other is to say that, if we lacked up-to-us-ness, it would be because of the laws, but not because of superdeterminism. I'm not sure which you'd favour, but I find the first more plausible at the moment.

Secondly, though you may well be right that most philosophers' uses of consequence arguments are really just arguments from luck, I think that they can be used otherwise. Here is an example. Suppose I and a friend both have quasi-Frankfurtian theories of up-to-us-ness, in that we think what makes an act (choice, resolution, omission, etc.) up to us is just some potentially-complicated fact about our beliefs and desires, and general state of mind, at the time we do it. In particular, neither of us care about luck at all. We both accept that many of the acts we think are up to us are the results of things we have no influence over. (We also enthusiastically bite all bullets necessary to ward off manipulation arguments. You can have been created by a Greek goddess 5 seconds ago to perform that very act and we'll still say it was up to you.) But my friend and I differ, in that I think that up-to-us-ness is compatible with determinism, whereas my friend, unusually, believes that the state of mind accompanying an act has to include some indeterministic element for it to be up to the agent.

My friend argues as follows: "We both agree that some of the acts that are up to us are the results of things beyond our control. You accept that, if the laws said that whenever Jupiter is in Pisces, that will cause events such that you will eventually have toast for breakfast, and Jupiter was in Pisces, and you ate toast, that could still be up to you. And I accept that, if the laws said that whenever Jupiter is in Pisces, it will either cause events such that you will eventually eat toast, or it will cause events such that you will eventually eat grapefruit, and Jupiter was in Pisces, and I therefore ate grapefruit, that could still be up to me. But, unlike me, you face a problem. It is an exceedingly attractive principle about up-to-us-ness that, if it is up to you that P, and Q is an unavoidable consequence of P, it is up to you that Q. But, if the laws did say that you'd eat toast after Jupiter was in Pisces, and you didn't eat toast, then it would be an unavoidable consequence of your not eating toast that Jupiter was not in Pisces. But, like me, you think it's (at least potentially) up to you that you don't eat toast, if indeed you don't, and therefore you ought to accept that it could be up to you that Jupiter is not in Pisces, which is crazy."

An argument like the above might not be a particularly good one. But it seems to me that it would not be a concealed demonstration that constitutive luck rules out up-to-us-ness, with the references to determinism only helping to make this clearer. Instead, it would be an argument that determinism rules out up-to-us-ness in a way that indeterminism doesn't. Do you disagree?

(It strikes me now that van Inwagen's use of the consequence argument isn't that far from th hypothetical use I just described. At the time, plenty of compatibilists (to my limited knowledge, anyway), thought that up-to-me-ness consists in my knowing what I am doing and being able to do otherwise, which they analysed in a compatibilist fashion. And van Inwagen basically agrees, I think, albeit in different terminology, except that he thinks the ability to do otherwise is incompatible with determinism. Both van Inwagen and his compatibilist opponents therefore hold or held a non-historicist postion on up-to-us-ness, in that how I came to be in a position where I know what I'm doing and could do otherwise is ostensibly irrelevant. I don't know van Inwagen's exact position, but it's at least consistent with his incompatibilism that many of the things that are up-to-us are the results of things beyond our control, as long as they aren't unavoidable consequences of them. So, you could think of my imaginary friend in the above example as van Inwagen talking to one of his contemporaries.)

The comments to this entry are closed.

Books about Agency

3QD Prize 2014: Marcus Arvan