Free Will and the Bounds of the Self

From a certain sort of theoretical perspective, the problem of free will can seem extremely simple and straightforward. Yes, it may be true that your actions are determined by our beliefs, desires and intentions... but then again, your beliefs, desires and intentions are simply you. So the alleged 'problem' here is just that your actions are determined by your own self. Where then is the difficulty?

Yet, strikingly enough, people generally are not persuaded by this simple argument. They often hold on to a sense that determinism would be a real threat to their freedom. It seems to them somehow that if all of their actions were determined by causal interactions among their mental states, they themselves would be unfree, like mere  marionettes controlled by this complex computational process. 

In a new paper, Shaun Nichols and I suggest that it might be possible to make sense of people's intuitions in such cases if we can just get a little bit more clear on the folk conception of the self. In particular, we argue that people do not simply see the self as a collection of mental states (so that anything produced by one's beliefs, desires and intentions is thereby produced by the self). Rather, people can easily adopt a conception according to which the self is something that stands outside all mental states and can freely decide, in light of those states, whether or not to perform an action. On such a conception, it is almost as though the agent's mental states serve as lobbyists or advisors and the self is a kind of executive that can listen to the various opinions and then make a choice.

In the actual paper, Shaun and I provide a series of experimental studies to support this claim about the folk conception of the self, but regardless of whether you end up taking a look at those studies, we would love to hear any thoughts you might have about this basic theoretical approach.

Comments

[Keith Douglas]

In the graduate seminar I took at UBC a few years back on the FW problem, the immediate rejoinder to the "answer" you provide (which I must say is my favourite one in the debate) is merely that it postpones the problem. Namely, where did the desires, beliefs, etc. come from? Factors outside your control. And so the problem recurs. It thus stands somewhat to reason that your research might disclose a tension in what is considered "me" and "not me", etc.

[Steve Maitzen]

Josh,

Thanks for reporting this interesting research. Compatibilists won't be surprised to learn that incompatibilist intuitions may arise from imagining a "featureless self," a notion that invites the problem noted by Keith Douglas in his comment.

If I may, I'd like to offer the x-phi community an experimentally testable conjecture. Compatibilism is true, but incompatibilists are right that our intuitions are incompatibilist to some degree. I conjecture that incompatibilist intuitions stem from confusing determinism with manipulation. Being manipulated relieves you of responsibility (to at least some extent), but being determined, as such, doesn’t. Our intuitions commonly confuse determinism with manipulation because of a deep-seated, primitive tendency to think that nothing happens that isn't caused to happen by some mind. This anthropomorphizing of determinism, however unconsciously, tweaks our intuitions in the direction of incompatibilism.

Some evidence for this conjecture can be found in Derk Pereboom’s "four-case argument" against compatibilism (Four Views of Free Will, pp. 94-97), an argument that I think crucially depends on conflating determinism with (subtly hidden) manipulation.

[Jonathan Weinberg]

Some pretty cool stuff there! Just out of methodological curiosity, what kind of controls for conprehension, unintended fillings-in (those are short vignettes!), etc did y'all use?

[Joshua Knobe]

Jonathan,

We actually didn't include any comprehension questions, mostly just because our scenarios were so incredibly simple and easy to understand. Do you think there is any kind of misunderstanding here that might be at the root of our effects?

(By the way, it would be great if you could put some of *your* amazing new free will results up here at some point.)

Keith and Steve,

You are certainly right that people's incompatibilist intuitions could be arising from these sorts of worries, but the hypothesis of our paper is that one can actually explain them in a simpler way.

The claim is that people think that if your actions are completely controlled by your own beliefs and desires, then these actions are not truly controlled by your self. There is then no need to introduce anything more complex involving manipulation or tracing back causal chains. The intuition just follows immediately from the idea that we can't act freely if our actions are controlled by something external.

[Josh Weisberg]

Very interesting!

A few thoughts after a quick read.

First, in Study 2, I think there's a class of mental states that does not give the result you claim--what we might call "acts of will." John's act of will did not cause his hand to tremble. Now, maybe these aren't mental states in the relevant sense? Or the whole idea of acts of will is incoherent (as Ryle thought)? Anyway, it seems like this is a line the mental state folks could use.

Second, in studies 3 and 4, do you predict that there should be no difference in assignments of blame for both zoomed out cases? That is, if you ask, "Is John to blame for breaking the glass?" should you get the same answer in both the choice and emotional conditions? That goes against my intuitions--I wouldn't think John is to blame for breaking the glass in the emotional condition (n=1 philosopher sitting in an armchair in Houston). But maybe this isn't the prediction? Perhaps by asking about blame, that forces a "zooming in" again? But if so, that means that all cases where we ask about blame might be zoomed in cases, and we lose the shifting of self concepts.

Third, why did you go with a morally abhorrent example (shocking poor lab critters) in the robot vs. human case? People shouldn't shock poor critters, no matter what. Do you predict the same results in a non-moral case?

In any event, I believe that (most?) people clearly do not think they are computers, in particular because computers "only do what they are programmed to do" while we are not like that. But is this because computers don't have "selves," or because their "selves" are programmed? I tend to think folk intuition goes with the latter, though I don't know how to test this!

Anyway, this is a bit scattered--my apologies. Thanks!

[Fritz Warfield]

I'm not sure what "the" problem of free will is so that a dubious literal identification of an agent with (all? some?) of the agent's mental states would be an answer to it. I do hope that we're down to one problem though; that would be progress.

And looking at the early parts of the paper, I'm not sure what the "surge" of work on van Inwagen's beta principle is referring to -- there was a surge of work on beta in the mid-1990s that included a formal proof that van Inwagen's beta principle is invalid. That's not controversial is it?

[jonathan weinberg]

Btw (and not in lieu of responding to your question), this New Yorker cartoon seemed rather relevant to y'all's paper:
http://tinyurl.com/chainedtoself

[Joshua Knobe]

Fritz,

We certainly didn't mean to suggest that there is just one problem of free will, much less that this one problem could be understood in terms of a theory about the nature of the self. Rather, the idea is that although there are many difficult philosophical problems here, some of these problems are of little concern to ordinary folks. The claim then is that we can use this conceptual framework to address the specific sort of worry that is at the root of folk anxiety about free will.

Josh,

That's a nice point about people's ordinary attributions of blame. It certainly seems that attributions of blame aren't just a matter of attributing the outcome to the agent's self. Even if it is clear that the agent's self is the source of the outcome, there are yet other criteria that have to be fulfilled before the agent can be considered responsible for it.

[Genruk]

In my opinion, free will is a romantic notion that will always be overpowered by man's social nature. As much as we like to think of ourselves as self-determinate, we are driven by our social needs to be accepted by those communities that we find significant and appropriate to our own well-being. We develop personas that are best fashioned to allot for the acceptance of those same groups, personas which oft times do not reveal our true selves, which in itself obliviates any pretensions to free will as we are behaving in a fashion that does not represent our veritable inner character in order to belong.
Genruk