More Free Willish Goodness from Science
The Libet studies, which have caused such a stir, apparently demonstrated that conscious choice lags behind by neural activity by around one third of a second. One of my favorite responses to these studies - Dennett's - argues that the notion we can perform simultaneity judgments as to neural events and volitions depends upon the idea of the Cartesian Theatre, a place in the brain where everything comes together. But while that response looks plausible for a lag of one third of a second, it doesn't look quite so plausible when the gap is an enormous 7 seconds. That's the claim of a new study just published in Nature Neuroscience.
In the study, subjects engaged in a free choice task, choosing between pressing a button with their left or their right hands. The researchers found that they could predict with 60% accuracy which hand they would choose, a full 7 (and up to 10) seconds before the subject reported that the decision was made, by analysing activity in the PFC. The popular presentations of this study have not failed to draw the conclusion that this study threatens free will.
One comment: I don't see why we shouldn't interpret the PFC activity as representing a disposition and not a choice (though PFC is involved in high level planning).
This shouldn't trouble compatibilists, though, because most compatibilists say that free will has nothing to do with predictability, right? In other words, all of the major defenders of compatibilism will say that even if someone could perfectly predict every future motion of every atom of your body, you can still have free will (as long as those predictable atoms are arranged in just the right way)?
Based on my understanding above (please correct me if necessary; I'm a little rusty), a compatibilist wouldn't regard reports of human predictability as even prima facie evidence against the existence of free will.
So the fact that many people in the media, etc., regard these reports as prima facie threats to the existence of free will suggests to me (and I'm guessing most non-compatibilists) that compatibilists have the wrong conception of free will. They've revised it, and watered it down.
If there's even a little truth to this, then I find this fascinating. To help see why, consider this analogy:
"We live in the ghetto and have no money."
"Well we have this money, so we're not poor, we're actually right."
"Wait, that's not real money, that's monopoly money!"
"No, no, no, this is real money. I know there's a difference between this money and the money you call 'real money', but I don't think we need all of that stuff for it to be real money."
"But most people say you need this stuff to be real money. Most people think monopoly money is not enough."
"Well, I don't care."
Now most compatibilists (and many libertarians) will say that unpredictable free will, a la being-causa-sui, isn't valuable. If that's true, then the analogy above might not apply.
But it's important to remember that people do value novelty and unpredictability and spontaneity---and for good evolutionary reasons. People who are novel, unpredictable, and spontaneous can (if other circumstances are right) avoid more predators and attract more mates. So there may have been strong evolutionary pressures for mammals, including our species, to implicitly or explicitly value and pursue these things.
The above is one reason why I think "free will", to the extent there is a consensus about its definition (and I think the extent is small), necessarily involves unpredictability. And that is one glaring deficiency of compatibilism in my view.
Posted by: Kip Werking | April 14, 2008 at 09:47 AM
Kip,
You said:
Can you point me towards any arguments for this claim? Surely, from the fact that we value novelty it doesn't follow that free will necessarily involves unpredictability.
I guess I just don't see these findings as even prima facie evidence for the claim that we don't have free will. And while I'm not a libertarian, I can't figure out why they would feel threatened by these findings either. Predictability threatens neither the AP condition nor the sourcehood requirement of libertarian theories.
So without an argument conceptually connecting unpredictability to free will, I'm inclined to think that no one should feel particularly troubled by these findings (merely because of worries about prediction).
Posted by: Justin Coates | April 14, 2008 at 11:14 AM
Justin,
You raise a good question, and draw attention to (what I realized is) a weaker point in my argument.
First, let me answer that the widespread tendency of the media, etc., to conclude that reports of human predictability at least prima facie threaten the existence of fw, is at least one data point.
However, I don't have knock-down evidence, and I'm not sure that unpredictability, to any extent, is a knock-dead feature of fw (in the same way that having four corners is a knock down feature of being a square, for example).
That said, the questions you ask sound funny to me because they give the impression (to my ear) that, in order to resolve the definition of free will, we just need to sit back in our arm chairs and toss arguments back and forth. But we've been doing that, explicitly or implicitly, for millennia and we have little to show for it. The extent that we've made progress since the Greeks, or the Middle Ages (and I would characterize this progress as a small shift away from libertarianism and towards compatibilism), it seems to me more due to the development and advancement of science and a less superstitious view of the world and our place within it, than to any philosophical rhetoric (as Manuel Vargas and others have noted, there appears to be a statistical correlation between libertarianism and religiosity).
So, when I hear "what arguments do you have?", I think "who needs arguments? we need *data*!"
I'm actually no expert in determining the definitions of words. I know little philosophy outside of the topics discussed here, and that ignorance includes ignorance of philosophy of language. I'll bet that the people at Webster's Dictionary know more about determining definitions than I do, but that might not be saying much. I know philosophers bristle when I suggest turning towards the dictionary, or common usage, but I think this largely represents a hesitation towards exposing one's view to empirical fire.
Perhaps what we need is a large, systematic study of what "free will" actually means. It has to be a study, or else there's nothing to prevent armchair-philosopher A from saying "free will is X" and armchair-philosopher B from saying "free will is ~X", back and forth, until Kingdom Come, perhaps in very eloquent articles published in the most prestigious philosophy journals.
To my knowledge, such a study has never been performed, and maybe it never will. But I think it will take a definition-determining project, a small philosophical Apollo mission, to make progress on this ancient problem.
Posted by: Kip Werking | April 14, 2008 at 05:03 PM
This sounds strange to me, but maybe it just reveals different methodological convictions. That said, I'm perfectly willing to concede that data about the folk conception of free will can play some part in an argument for favoring one definition of free will over another (however, I must admit, I doubt that such data will be ultimately decisive).
But back to predictability...
Here are two beliefs that I suspect many, perhaps most Americans hold:
(1) We have free will, and
(2) God exists and foreknows all of our actions.
So one reason for thinking that unpredictability isn't essential for free will is because free will--a concept regularly invoked in religious contexts--doesn't seemed to be ruled out by God's foreknowledge (at least before we do philosophy). This is hardly decisive, but it seems at least* as compelling as invoking the media's obsession of writing about how the scientific discovery du jour proves there is no free will.
Again, I worry about your methodology for analyzing the concept "free will," but my main concern is just the claim that unpredictability is necessary for free will. But even assuming your methodology, I suspect that we could find data that suggest free will is not essentially linked to unpredictability. If so, then by your lights, we wouldn't need to worry about scientists' ability to predict our behavior (even if we need to worry about such findings for other reasons).
Posted by: Justin Coates | April 14, 2008 at 06:50 PM
Justin,
You wrote:
"That said, I'm perfectly willing to concede that data about the folk conception of free will can play some part in an argument for favoring one definition of free will over another (however, I must admit, I doubt that such data will be ultimately decisive)."
To speak of a "folk conception" of free will implies that there are two different conceptions. Which implies that philosophers use the term in an idiosyncratic way. Which implies that either (i) there are two different definitions of "free will", just like there are at least two definitions of a "bank" (e.g. think river or think money) or (ii) philosophers are just wrong.
My whole point in asking people what they mean by "free will" is to see the extent to which they agree with philosophers, because the consensus is what defines the term.
About God's foreknowledge and free will, most people say they believe in both, but also consider:
1. most people, if you point the above out to them, immediately see and agree that there is a prima facia tension or inconsistency here. That is why the conflict between God's foreknowledge and free will is a famous paradox which philosophers and theologians have grappled with for millennia. So your analogy would be much more convincing if people failed to see any tension or conflict between these two. But many or most do.
2. the above suggests that people may be believing in both, despite their apparent inconsistency, because they don't reflect on that inconsistency. Most of these people also say that God is all-loving and all-powerful but does not intervene to prevent evils like cancer and earthquakes, etc.
"This is hardly decisive, but it seems at least* as compelling as invoking the media's obsession of writing about how the scientific discovery du jour proves there is no free will."
If your goal is to simply to put free-will-does-not-require-unpredictability on equal footing with free-will-does-require-unpredictability, then I'm happy to tentatively agree with you. As I said, we need more data.
Posted by: Kip Werking | April 14, 2008 at 07:50 PM
I don't want to turn this thread into an argument over which meta-philosophical position we should adopt to solve thorny philosophical problems. I will say that your claim that consensus is what determines definition seems to be problematic. The referents of "water," "square," "validity," "planet," etc. are not determined by popular use. Perhaps "free will" isn't like these concepts, but perhaps it is--as van Inwagen has suggested. "Free will," he claims, is one of the few terms that started as a philosophical term of art that was transmitted into wider usage because of its role in explaining certain Christian doctrines. If this is right, the definition isn't beholden to general use. I'm not in a position to evaluate the historical veracity of this claim, but it seems like a plausible story (for whatever that's worth).
Moreover, even if philosophers are using it "idiosyncratically," I see no problem. Think Carnapian explication here. We start with a vague, inchoate term and precisify it into something useful for theorizing. But I digress...
At the end of the day, I'm really just looking for an argument for your claim since I can't, for the life of me, figure out why prediction per se (especially prediction that's only accurate 60% of the time!) would render our wills unfree. It's possible I'm just being really thick.
Since neither of us is going to run any experiments on the predictability/free will relation, I'll just leave the issue with this comment. It would surely be an interesting result if the folk concept required that an agent's freely willed actions were essentially unpredictable, but it's much more likely that the folk are deeply divided over the nature of "free will," which is just one more reason to think that "free will," our inchoate, unclear explicandum, is in need of explication.
Posted by: Justin Coates | April 14, 2008 at 09:09 PM
It's interesting that no one is drawing the obvious conclusion here which is that this and other experiments of choice prediction - with 60-70% accuracy - are evidence FOR free will! If we're determined, we should be 90+% predictable.
I have always argued, long before these experiments, that while humans are unquestionably free, we are also - speaking v. impressionistically - something like 60-80% consistent in our behaviour (and therefore, I guess, purely in principle, 60-80% predictable).
Of course, any such figures are highly impressionistic, but there is a basic underlying truth - humans HAVE to be highly (as in 60-80%) consistent, otherwise we'd never recognize each other as we do. Our choices and behaviour do manifest fairly consistent personalities.
But then there's that other 30 odd % - those are the times when you don't choose to work, or exercise, or stick to that diet - all those examples of "akrasia". I'd say those are also, in a v. broad way, consistently inconsistent. I can predict with a fair degree of confidence that most humans will freely break all those resolutions a good deal of the time - be, if you like, now "manic" and positive in their activities, now "depressive" and negative.
What I - and all the sciences (and they've tried) - can't predict is the patterns of people's "manic-depressive", positive-negative, active-passive and v. free choices. We know people will "swing both ways", we just don't, and can't, know WHEN - because they're free.
Oh no, this is all evidence FOR free will. (And why is no one talking - or have I missed something yet again - about Baumeister's self-control experiments and work - he was at the recent Free Will conference)? The evidence for free will is going to mount steadily. There's a lot of it around if people want to look.
Posted by: Mike Tintner | April 15, 2008 at 05:58 AM
It's interesting that no one is drawing the obvious conclusion here which is that this and other experiments of choice prediction - with 60-70% accuracy - are evidence FOR free will! If we're determined, we should be 90+% predictable.
I have always argued, long before these experiments, that while humans are unquestionably free, we are also - speaking v. impressionistically - something like 60-80% consistent in our behaviour (and therefore, I guess, purely in principle, 60-80% predictable).
Of course, any such figures are highly impressionistic, but there is a basic underlying truth - humans HAVE to be highly (as in 60-80%) consistent, otherwise we'd never recognize each other as we do. Our choices and behaviour do manifest fairly consistent personalities.
But then there's that other 30 odd % - those are the times when you don't choose to work, or exercise, or stick to that diet - all those examples of "akrasia". I'd say those are also, in a v. broad way, consistently inconsistent. I can predict with a fair degree of confidence that most humans will freely break all those resolutions a good deal of the time - be, if you like, now "manic" and positive in their activities, now "depressive" and negative.
What I - and all the sciences (and they've tried) - can't predict is the patterns of people's "manic-depressive", positive-negative, active-passive and v. free choices. We know people will "swing both ways", we just don't, and can't, know WHEN - because they're free.
Oh no, this is all evidence FOR free will. (And why is no one talking - or have I missed something yet again - about Baumeister's self-control experiments and work - he was at the recent Free Will conference)? The evidence for free will is going to mount steadily. There's a lot of it around if people want to look.
Posted by: Mike Tintner | April 15, 2008 at 06:00 AM
Kip and Justin, maybe somebody will do a big study to try to get a much clearer sense of what people think free will means, its relationship to responsibility, and various threats to it... if that somebody could just get the time and funding to do it!
Despite the possible counterexample of God's foreknowledge, I'd be surprised if most people do not find the idea of *complete* predictability of at least their own actions to threaten their free will. As someone (Skinner?) said, with predictability comes control.
Mike, I agree that there is a lot of good work (including Baumeister's) that helps to *explain* the existence of free will rather than *explaining away* free will. Of course, much depends on how we understand free will. If we take it to mean a part of us, the agent or soul, that is not subject to natural laws (as some of the "explain awayers" do), then there will be no explaining free will by the natural sciences.
But if we take free will to mean our capacities for rational self-control (and such), then we should expect explanations for these capacities from the sciences ... unless of course the sciences show we don't really have those capacities (as other types of explain awayers do).
Posted by: Eddy Nahmias | April 15, 2008 at 06:51 AM
Justin,
I understand if you don't want to turn this thread into a philosophy of language debate. However, your last comment raises so many interesting points, I can't resist. For example, you wrote:
"Moreover, even if philosophers are using it "idiosyncratically," I see no problem. Think Carnapian explication here. We start with a vague, inchoate term and precisify it into something useful for theorizing. But I digress..."
If I understand you correctly, you mean to say "well, if fw is vague, that's no problem, because we can just talk about fw* and fw**, and so on, and that will help us decide whether fw exists." But, as I hope is clear, that wouldn't help. No matter how much work we do with fw* and fw**, that work can't help us decide whether fw=fw* or fw=fw**, for example.
Posted by: Kip Werking | April 15, 2008 at 09:28 AM
Mike, I'm afraid the 60% predictability might just as easily reflect the difficulty of interpreting patterns of neural activation. We know which brain areas are involved in vision, yet we don't achieve more than 80% accuracy in inferring from fMRIs which gestalt of an ambiguous figure a subject is perceiving. Brains are noisy systems. So the 60% might reflect difficulty of interpretation, not agent's freedom.
Posted by: Neil | April 15, 2008 at 05:25 PM
The results seem consistent with a wide variety of theories of free will. Of course with compatibilist views, but also with, e.g., Kane's or O'Connor's indeterminist views. (O'Connor wouldn't allow that prior brain happenings cause a free choice, but he allows that they structure the choice situation, establishing probabilities of the various options.)
One thing that struck me about the article was the use of expressions such as 'the brain decides'. Should we consider whether it's the brain (and not the person) that is responsible for what's done?
On that point, I don't recall anything in the study about what might cause the brain's decision, or even about any factor that had a greater than chance correlation with which decision the brain made.
Posted by: Randy Clarke | April 16, 2008 at 05:32 AM
Just to touch on Kip's point, the issue isn't predictability per se. But predictability is relevant. If you assume (currently counterfactually) that an action were perfectly predictable at 7 seconds-plus before the onset of a conscious intention, then the idea that when we act, we are acting intentionally seems to be weakened. For as far as we'd know, our conscious "intentions" might just be epiphenomena. At the very least, then, a compatabilist would have to show that these intentions represent or reliably reflect the epistemically relevant grounds upon which the subconscious (or preconscious, or subrational, or...) decision was made by the brain/body.
Posted by: Q the Enchanter | April 16, 2008 at 09:31 AM
The frontal cortex, the putative candidate for such a core-brain-region for free will, its an heterogenous tissue without no clear anatomical boundaries between its ventral and dorsal parts, and its activity is related to intention, future goals planning, shift mechanisms in attention, executive control...
Not only we must have clear what we mean by free will(as Nahmias said above)but what pshychological functions underwent when we express actions de novo.
Posted by: Anibal | April 16, 2008 at 09:38 AM