Blog Coordinator

« Some Varia about Free Will | Main | The Featured Author of the Month: Randolph Clarke »

04/25/2014

Comments

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

I don't think any work in cognitive science threatens free will. Let's put aside libertarianism for a moment. The great free will debate depends on the definition of free will. If free will is defined in a way that requires self-creation, ultimacy, and/or being an independent variable not dependent on antecedent variables (as Pereboom recently wrote), then free will doesn't exist. If free will merely involves exercising rational control and/or responses sensitive to moral considerations, then free will (often) does exist.

Which of these two views is correct does dependent on empirical questions. But those questions involve the common usage of "free will" (to determine its definition) and related questions about appropriate language usage.

Of course, if free will were defined in a libertarian way, then discoveries in cognitive/neuroscience could prove that free will exists (or doesn't exist), just by analyzing the brain at the right level of detail. But I don't see how this argument even gets off the ground because of a priori concerns about the coherence/benefits of libertarian free will, like the luck objection. As Levy and others note: libertarianism is just parasitic on compatibilism.

I consider other empirical work, like Libet's and Wegner's, to be giant red herrings, distracting us from the real problem: constitutive luck and whether free will is defined in a way that survives it.

My view is that the empirical finding in the behavior, cognitive, and neuroscience are relevant to some extent. Unfortunately when people discuss the "scientific threats" to free will too much attention is spent on Libet, Haynes, and Wegner and not enough on situationism, automaticity, and the adaptive unconscious.

As I see it, compatibilism faces two potential lines of attach: external and internal. By external threats I mean the traditional philosophical arguments against compatibilism (e.g., the consequence argument, the basic argument, the luck argument, Pereboom's four case argument, etc.). These arguments question whether compatibilism as a whole is capable of provide a positive account of free will and moral responsibility.

The internal threat, however, while not as potentially devastating as the external one (meaning it is probably not capable of refuting compatibilism on its own) is nonetheless worth discussing. The internal threat is essentially the one of shrinking agency. The question presented by this internal threat is whether the freedom-conferring conditions compatibilist accounts put forward (leading candidates being control conditions, reason-responsiveness, hierarchical identification, etc.) are difficult or impossible to satisfy given what we are learning about the pervasiveness of unconscious processes. I’m not prepared to make an argument one or the other here as to how to answer that questions, but I think it is an empirical question what requires a carefully look at the literature in social psychology. Put differently, I believe compatibilists accounts of free will have empirical constraints (and different compatibilist accounts have different empirical constraints depending on the account), hence they should be subject to empirical analysis. It would be at this point that I propose the empirical literature on situationism and automaticity is relevant.

BTW, Neil Levy has a wonderful new book which defends something he calls the consciousness thesis: Consciousness of key features of our actions is a necessary condition of moral responsibility for them. I will let him speak for himself, but I see it as an important argument and will force compatibilists to take more seriously the internal threat. (BTW, I think Neil and I disagree on how big a threat this internal challenge is and how easy compatibilists will be able to satisfy the consciousness condition—I believe it may be a harder condition for compatibilists to satisfy than Neil does. At least that is the impress I have gotten from Neil.)

I want to be clear, I think the internal threat (as I describe it) is an empirical one and is probably not capable of refuting free will. Depending on how big a threat it turns out to be, however, it has the potential to seriously limit even the kind of free will compatibilists defend.

That’s my two cents—for what it is worth.

Kip: Suppose I were to say to you, "I don't think any work in physics threatens absolute space and time. It all depends on how we define space and time." The right answer to this is: "No, it doesn't. Space and time are *things* the nature of which are independent of how we might wish to define them."

The same is true of free will. You can try to define it however you like, but ultimately, it's a metaphysical issue. Either all of our actions are determined by the laws of physics, or they are not. Either neural-firings cause all of our actions, or we have the brute ability to cause our own actions from nothing (viz. we are "uncaused causes"). These are not definitional issues. They are metaphysical ones.

Now, you may think the kind of libertarian free will I describe here is incoherent, but I think that is plainly wrong, and that it is therefore incorrect to construe the free will debate in the way you do (as merely a definitional issue).

Hi Kip,
So if I read you right, cognitive science couldn't threaten free will because free will doesn't exist. But then you say that it is a question of folk usage whether free will means what compatibilists say it does or not. So are you espousing a position on the x-phi of free will? Or would you rather say that the folk are just wrong because they haven't reflected deeply enough on luck problems? Or something else?
As an aside, you also say that if libertarianism were right then cognitive science might have some work to do. Is that because you think we might find genuine indeterminism in the brain?

Hi Gregg,
I like the way you're carving up the space. I'm always struck by how quickly consciousness comes up when discussing these issues. You mention the pervasiveness of unconscious processes. Do you think that the fact that a process is unconscious makes it less apt to subserve the kinds of mental processes you mention (reason-responsiveness, etc.)?

FWIW, I've read Neil's book too, and I think everyone that reads this blog should do the same. Perhaps later on I'll try to get discussion going about it - I'd like to hear your thoughts on Levy on consciousness.

Hi Marcus,

Nothing you say in your comment commits you one way or another, but your comment makes me wonder what you think about the relevance of x-phi to uncovering the meaning of 'free will.' Suppose, for example, that it turns out that very few people think of free will as requiring uncaused causes. Or suppose Eddy is right and very few people think free will requires indeterminism. Would that be informative regarding the metaphysical issue you highlight?

I recently taped an episode of Philosophy TV with Neil Levy discussing his new book. It's all about the relationship between consciousness and FW/MR--something we are both interested in. I will post it here once it becomes available.

One thing I really like about Neil's book is his use of an access-consciousness view. It's kind of a remake of the old-fashioned "topic-neutrality" view of mind, and really works well for setting the mind-body problem aside to make his point. I read his mss and loved it. Neil's book is indeed a must-read for anyone interested in the neuro-ethics of the FW issue.

Hi Josh,

Thanks for your question. I think the question of what we call things is mostly irrelevant. I don't think it matters much what we (or the folk) are willing to call "free will." I think what ultimately matters are the phenomena themselves. Do we have the capacity to do otherwise than we do, or are all of our actions the consequence of physical law? If all of our actions are the consequence of physical law, does it make moral sense to have certain types of reactive attitudes? These are the questions that matter, in my view. In other words, I'm with Dennett (these days apparently!). We should not get hung up on the term "free will" and look past it to the metaphysical and moral questions that really matter.

Hi Josh -

"So if I read you right, cognitive science couldn't threaten free will because free will doesn't exist."

I should be more precise. What I meant was: the big, genuine, interesting dispute about free will - in my opinion - is the dispute between free will skeptics and compatibilists. And that dispute is, essentially, a dispute about how to define free will - although it is rare to hear the dispute framed in those terms (perhaps because it makes the dispute sound trivial?).

Other commentators have reminded me that: yes, if free will has a compatibilist definition, then it's an empirical question whether we satisfy that definition (and to what extent). Same for (event-causal) libertarianism (which is parasitic on compatibilism). In that way, challenges to rationality can become relevant. But these challenges don't interest me much, because I think the free will debate is poorly framed, if it hinges upon subtle deficiencies in human rationality, just exposed in the last few decades.

And you're right, if free will has a "causa sui" type definition, then evidence is irrelevant because free will is logically impossible.

"But then you say that it is a question of folk usage whether free will means what compatibilists say it does or not. So are you espousing a position on the x-phi of free will? Or would you rather say that the folk are just wrong because they haven't reflected deeply enough on luck problems? Or something else?"

As I wrote above, I think the most interesting dispute is essentially one about the definition of free will. I think that evidence becomes important in deciding what the definition is. I'm not sure that evidence about common usage and language practices is the *only* evidence to show what free will means. But it is certainly extremely important. This evidence is about language practices, and not neuroscience or cognitive science.

"As an aside, you also say that if libertarianism were right then cognitive science might have some work to do. Is that because you think we might find genuine indeterminism in the brain?"

Well, I think the chances of that happening are about 0%. But, in principle, libertarianism should be testable by inspecting the brain. If you open up the brain, and you find something like a Geiger counter, with random events that have no deterministic predecessors, then that could (conceivably) support libertarianism. Actually, I think even this is problematic, because any deterministic world looks exactly like an indeterministic world with the same events/content. But to the extent that determinism correlates with regularities and indeterminism correlates with irregularities, then it should be testable. (Still, as I alluded to above, you can have deterministic chaos and indeterministic orderliness, and this point has not received nearly enough attention in the literature.)

An aside: is it cognitive science that would pose threats to free will or just psychology and neuroscience? I ask because many include linguistics, computer science, and even education under 'cognitive science' and it does seem obvious to me that these fields do work that poses serious threats to (any definition of) free will. But perhaps I am missing something here.

Nick: what is it about cognitive science and these other fields that you think *obviously* poses a serious threat to any definition of free will? I'm just curious if you could clarify -- for perhaps then we could have a conversation! (For my part, as I've argued in some of my own work, I don't think physics, as we currently understand it, is at all inconsistent with full libertarian free will -- and so, since linguistics, cognitive science, etc., all piggy-back on physics, I don't see what the problem is supposed to be).

Marcus,

I agree with you that the laws of physics are consistent with free will (i.e., the ability to do otherwise), and I've stated my reasons numerous times here on FoF. Perhaps you could post a comment summarizing why you also believe that's true. (I'm sincerely interested!)

Hi Nick,

I said 'cognitive science' mainly to include psychology with neuroscience and also to avoid worries about determinism - Adina Roskies has convinced me that theoretical physics rather than neuroscience is the place to look for those worries.

Hi James: Thanks for the question. For my full answer, you should see my recent article in The Philosophical Forum, "A New Theory of Free Will" (http://philpapers.org/rec/ARVANT-2 ). In it, I argue from several theses in metaphysics and contemporary physics to a new model of reality that reconciles libertarian free will with quantum mechanics.

Basically, I make good on Heisenberg's suggestion that quantum mechanics necessitates a kind of dualism, and once we appreciate that, we can construct a new model of reality in which free, libertarian free will in a higher reference-frame results in an illusion of causal-closure in this (our) reference frame -- and in a way that solves many deep problems in the philosophy of mind, time, and metaphysics more generally. Because this new model predicts and explains quantum phenomena and many philosophical problems that otherwise lack an adequate explanation, I argue that it is probably true.

Finally, you might want to see my recent Flickers post showing systematically how the model promises to explain quantum uncertainty, superposition, wave-particle duality, wave-function collapse, etc. (see http://philosophycommons.typepad.com/flickers_of_freedom/2014/02/a-new-theory-of-free-will-and-the-peer-to-peer-simulation-hypothesis.html )

Marcus: I meant to include 'not' in front of 'seem'—so I should've typed "...does NOT seem obvious." How foolish of me. My apologies. I hope that clears things up a bit. And thanks for your work on this subject. I follow you on academia.edu, so I was pleased to find your "A New Theory of Free Will" awhile back.

Josh: Thanks for clarifying. Distinguishing the implications of cognitive science from those of physics seems sensible. I think I agree with you (and/or Adina Roskies) on this.

To be clear: although I do study cognitive science and philosophy I consider myself an amateur on this subject, so I did not mean to take issue with anyone's work here. Even if I did take issue with it, my position would be amniotic and undeveloped at present. So my comment was an attempt for clarification. Sorry that my typographical error implied a polemical tone. That was not the intention.

Thanks for hosting the discussion!

Hi Josh -

I thought I replied earlier, but let me repeat:

"Hi Kip,
So if I read you right, cognitive science couldn't threaten free will because free will doesn't exist."

Let me be more precise: I see the real, genuine, interesting debate to be the debate between skeptics and compatibilists. And that debate is essentially a debate about the definition of free will. You won't usually hear the debate characterized that way (perhaps bc it makes the debate sound more trivial). But that's essentially what it is.

And so, in my view, evidence can be relevant to the free will debate, but it's going to be evidence about the definition of free will - evidence about common usage and language practices, etc. If consequentialist-type considerations can justify adopting one definition over another, at the margins, then evidence about those considerations (i.e., good outcomes) can become relevant too - see Shaun Levy's recent article on this topic.

If you adopt a particular position in the free will debate, then evidence may or may not relevant:

1. Skepticism adopted: evidence usually not relevant, because free will is (for all practical purposes) impossible

2. Compatibilism adopted: evidence can be relevant to show that people do not satisfy criteria for rationality. But compatibilists are typically smart enough to lower their rationality requirements enough that people satisfy those requirements. This debate is less relevant to me because, as interesting as studies in akrasia are (and I find them very interesting), I don't find these studies to be particularly threatening to free will, as the debate is traditionally framed. It's more of a red herring from the debate about the definition of free will, which I find more interesting.

3. Libertarianism: In theory, evidence can show that events in the brain do not correlate with identified regularities or antecedent events. If you looked at the brain and found something like a Geiger counter, or quantum randomness, then you might take that as evidence of libertarianism. I think the odds of this happening are about 0%. Also, one point that is really under-appreciated in the literature is that there is no per se link between regularity and determinism, or between randomness and indeterminism. Indeterministic regularity is possible (in theory) and deterministic "randomness" or chaos is also possible. A sequence of events that happens deterministically looks exactly the same as a sequence of events that happens indeterministically - so, in principle, it's impossible to distinguish between the two.

I think the above answers your other questions too.

Hi Nick: Yes, that does clear things up a bit. ;)

Interesting discussion. Can someone explain to me the definition of luck in the simplest possible way that might make sense to a neuroscientist? I've never been able to fully wrap my head around it.

thanks,
Thalia

Kip--

"Indeterministic regularity is possible (in theory) and deterministic "randomness" or chaos is also possible. A sequence of events that happens deterministically looks exactly the same as a sequence of events that happens indeterministically - so, in principle, it's impossible to distinguish between the two."

The latter case may occur only in the case that an indeterministic sequence plays out according to something like a Humean regularity, though it is indeed metaphysically possible that that might happen even if unlikely. But as I understand it, deterministic chaotic randomness is a complexity constraint on epistemic conditions about predictability only, and is independent of Humean or non-Humean assumptions about determinism. So I'd say the claims about indeterministic regularity and deterministic randomness mixes metaphysical apples with epistemic oranges.

Hi Thalia,

There has been a lot of ink spilled about luck as it relates to free will, and I'm hardly an expert. But since I don't know how many experts are following the thread at this point, here is a quote from a Stanford Encyclopedia article on Incompatibilism by Randy Clarke and Justin Capes:

'If a decision is nondeterministically caused, and if there remains until it occurs a chance that the agent will instead (at that moment) make a different decision, then there is a possible world that is exactly the same as the actual world up until the time of the decision, but in which the agent makes the alternative decision then. There is, then, nothing about the agent prior to the decision—indeed, there is nothing about the world prior to that time—that accounts for the difference between her making one decision and her making the other. This difference, then, is just a matter of luck. And if the difference between the agent's making one decision and her instead making another is just a matter of luck, she cannot be responsible for the decision that she makes.'

I quote them since they put it better than I would have. If any luck experts are reading, feel free to chime in. Hope that helps!

Josh: I have to admit that I have never understood the so-called Luck Argument against libertarianism, and I think the passage your quoted shows what is so wrong with it.

The passage reads: "There is, then, nothing about the agent prior to the decision—indeed, there is nothing about the world prior to that time—that accounts for the difference between her making one decision and her making the other. This difference, then, is just a matter of luck."

The inference here ("This difference, then, is just a matter of luck") is *plainly* fallacious. If two possible worlds are exactly similar up to time t, and in one world *I* choose X and in the other I choose ~X, this difference need not be luck. It may be a *brute* exercise of libertarian free will.

Opponents of this answer may then say, "What do you mean by 'brute' libertarian free will?" if not causation from nothing (which, again, is just a matter of luck). But this is fallacious, too. All kinds of things in the world are brute. The laws that govern the path of electrons are brute. The laws of quantum chromodynamics are brute. They are not explained by anything besides the laws themselves. So, claiming libertarian free will is brute is perfectly coherent, and to say that it is brute is to say that we have the brute ability to choose -- without any explanation in terms of luck or anything else -- between multiple options. That is what is wrong with the Luck Argument. It's simply question-begging against the idea that libertarian freedom is brute (which is of course what the libertarian wants to say).

I bet Randy and Justin (at least Justin) would agree that the Luck problem isn't decisive - there they are summarizing for the SEP article.

But there is a cousin of the Luck worry that might arise when one is considering whether cognitive science threatens free will. One way of thinking about luck emphasizes an agent's lack of control (in some respect): luck is control-excluding. Neil puts it in something like that way in his book Hard Luck. Now, arguably cognitive science suggests that agents do not control much of their activity - many actions are prompted by the environment in a way that largely bypasses deliberation, much action control is the provenance of sub-personal mechanisms, and so on. Since cognitive science suggests we do not have anything like the control we believe we have over our own activity (our actions and omissions), cognitive science threatens both libertarian and compatibilist free will.

To be clear part 1: this is not the Luck problem, obviously. But I do think it taps into some of the same intuitions.

To be clear part 2: that is a (breezy) way of formulating the threat. I do not here endorse the threat's plausibility, nor do I really like such breezy usage of the term 'control.' I think the devil is very much in the details here, and put it breezily just to engage a broader range of concerns, and to get the thread back on topic.

Thanks Josh and Marcus. Putting aside brute libertarian free will for now and considering luck as hinging on neither the prior state of the agent nor world, is there an example of how luck could manifest in a physical sense? Are we talking the level of (putative) indeterminacy of electron spins/wind patterns/etc? Is that putative indeterminacy/noise/randomness efficacious on a scale that would be pragmatically relevant? Is there a better concrete example? That might help me understand it.

I second Kip's request for definitional clarity. Do we have a better understanding of the lay definition of free will? Thomas and Eddy, are the results of your study in? I assume you can't post the exact results because of publishing rules but any hints? Very curious.

V. Alan White:

You wrote:

"The latter case may occur only in the case that an indeterministic sequence plays out according to something like a Humean regularity, though it is indeed metaphysically possible that that might happen even if unlikely. But as I understand it, deterministic chaotic randomness is a complexity constraint on epistemic conditions about predictability only, and is independent of Humean or non-Humean assumptions about determinism. So I'd say the claims about indeterministic regularity and deterministic randomness mixes metaphysical apples with epistemic oranges."

For indeterministic regularity: exactly, a Humean regularity. You say this is possible, so I am not sure how I'm mixing apples with oranges.

For deterministic irregularity: you indicate that this situation only relates to our powers of prediction. In other words, in your view (as I take it), there is some underlying regularity, but we're just not smart enough to figure it out. I'm not so sure. Consider the deterministic law that says: "sequence of events A, B, C, D..." occurs, where that sequence of events is simply random and non-repeating. We wouldn't identify any regularity there, not because we're not smart enough to find it, but because there isn't one to find - it's not repeating, and the sequence itself is random.

Maybe this is a minor quibble between us. My overall point is that deterministic worlds and indeterministic worlds can look exactly the same, and so it's impossible to distinguish between them, in principle. In fact, as the years go on, I question more and more whether the distinction between them is even coherent.

The comments to this entry are closed.

Books about Agency


3QD Prize 2014: Marcus Arvan