31 March 2013

Responses on free will

There are two topics that seem to reliably bring people of strong opinion out of the woodwork every time they're brought up: feminism, and to a much lesser extent free will. I got quite a few comments on my post last week in which I took Jerry Coyne to task for overstating the implications of a recent study on decision-making. Rather than try to respond to each one with a separate comment (since many of them touch on similar issues) I thought I'd just consolidate my responses in a new post.

If there's any confusion about where I stand on free will, reading that post as well as this one – which was heavily influence by this post by Sean Carroll – ought to clear the air. The following quoted sections are some of the more notable excerpts from comments I received:
Yes, the processes going on in the brain that lead to an action (which we could call a "decision") is a real process, but could another "decision" have occurred given the same state? If not, then even though we cannot predict every "decision," we could not have chosen otherwise. This is what I think Coyne means when he says free will is an illusion. Not that the process we call a "decision" does not occur, but that since it could not have occurred otherwise it was not free.
I've never quite understood the notion that we are choosing, yet we could not have chosen differently. It seems clear to me that to say that we could not have chosen differently is exactly the same thing as saying that we could not have chosen at all, the undeniable implication of which is that all human volition is illusory – we are automatons. This renders moral responsibility and, quite frankly, the entirety of human experience as illusory since all consciousness itself is no more than a set of deterministic processes. I simply don't see a way to avoid that conclusion. To add to the weirdness, if human experience is illusory, then the entire concept of something being illusory is illusory. The whole picture isn't just completely useless, it's absurd.

The problem I have with it is that it confuses levels of description. Some things might be viewed as "fundamental", like quantum indeterminacy or the neurological structure of the brain; others might be viewed as "emergent", like rational volition (or baseball, to use Carroll's example). There may be no model of particle physics that gives us free will. But according to model-dependent realism, it's pointless to ask what is "real" – only what is the more useful description.  Where hard determinists and compatibilists seem to part ways is that compatibilists view free will as a valid description of reality – an emergent construct – while hard determinists argue that if it's not fundamental in some way, it's not real at all. But as I explained above, you can't arrive at such a conclusion without throwing the baby out with the bathwater. That's the reason why when Dan Dennett talks about free will, he does so using the framework of evolution rather than, say, quantum mechanics.
The fact that it may provide a useless model of human behavior doesn't make it untrue. I think our ability to choose *IS* more illusory than consciousness, because it's already known that we are conscious of our choices after the choice is done.
The experiments in question (which were the subject of the post) are only able to predict simplistic decisions with accuracy not much better than a coin toss. All this does is show that certain mechanisms of choice may be subconscious – not that there is no real choice at all. That brings me to another issue, though, which I think reveals more of the inevitable confusion hard determinists bring upon themselves. If hard determinism is true, then it does not matter whether the mechanisms of "choice" appear in the conscious or subconscious mind. The conscious mind is just an illusory construct, only giving us an illusion of awareness and volition. On hard determinism, the subconscious and conscious minds are functionally identical, since they can both be described with deterministic processes. These types of experiments cannot even in principle support hard determinism, since hard determinism holds that even mechanisms about which we may be consciously aware are still wholly deterministic.

From my perspective, that's a huge hurdle for hard determinism. It's rooted in counterfactual claims ("I could not have chosen differently") that can never be falsified. It cannot be scientific, nor can it be a valid basis for understanding human behavior. What good is it then, aside from armchair philosophical masturbation?
If it could be demonstrated that, say, a fifth of time when we think we are able to choose freely we are actually not able to choose freely, all without showing anything in particular about the other four fifths of the time when we think we are able to choose freely, we would have good reason to think that we are wrong at least some of the other four fifths of the time about which we don't have as direct evidence.
In every case, the hard determinist will appeal to underlying processes – but where in the neurological complexity of the brain is there "free will"? I can't help but think that this is confusing compatibilism with libertarian free will. Libertarian free will would say that free will, probably due to substance dualism, exists in the brain in some underlying, fundamental way. Compatibilism views free will as an emergent construct of more fundamental processes – something that's absolutely necessary to understand human behavior, and thus as "real" as anything else in the human experience.

Now, a common straw man leveled at compatibilism is the idea that biological or environmental influences don't affect our behavior. X process in the brain or x environmental factor influences our behavior, thus we're not really making "free" choices. This is only a problem for one who subscribes to libertarian free will. To suggest that we make decisions free of influence – right down to the chemistry in our brains – is obviously incoherent. But to say our choices are influenced by factors we cannot control is hardly the same as saying we do not choose at all.

It's not just worth pointing out that these Libet-style experiments are both unimpressive in their accuracy (60%?!) and dramatically simplistic, but that they don't demonstrate that the person was not actually making a choice. At best, it simply demonstrates that subconscious processes occur that direct or influence our conscious choice. If experimenters were able to predict decisions 100% of the time, or if they were able to predict complex human behavior rather than simplistic either-or decisions, I might be more impressed. But remember – since hard determinism holds that consciousness itself is every bit the illusory construct as choice, it wouldn't matter whether we were aware of any decision-making processes in the brain or not. It just drives home the fact that hard determinism is unscientific and devoid of any pragmatic utility.
I don’t think that determinists are reductionists and deny anything that compatibilists believe; I think that they simply don’t deem it free will. I can’t for the life of me think of one tangible, empirical difference, even in principle, between determinism and compatibilism. So the compabilist says we make choices and determinists agree, we make choices like a computer makes choices by following a certain path determined by its code, but it’s all determined by the underlying structure. So the determinist asks, where’s the 'free' part?
I don't think it's that simple; I think the hard determinist is essentially defining the very concept of choice out of existence, and with it the entirety of human experience. I suppose we could call it semantics, but I think it's worthwhile to do as model-dependent realism suggests and call into question how we define what is real in the first place. If we don't "really" make decisions, then there is no "we" at all – we've defined ourselves as no more than a collection of particles in this or that quantum state.

Over on the aforementioned post by Sean Carroll, I found a couple of comments I liked quite a bit:
Every claim that “I did something,” or that “you do something,” presupposes free will lest those claims make no sense. When I say, “I made a sandwich,” I claim that I am the cause of the sandwich and that I alone am the cause of that sandwich. Free will is when the agent is the sole cause of something, in the sense that without that agent’s acting, the something would not have happened. Most human beings are causal agents such that they choose something without there being any prior cause compelling them to do it.
This is as silly as arguing that computers cannot add numbers. At a certain level, they can't. They can only manipulate electrons so that the relationships between the original and final voltages and currents can be interpreted as adding two numbers. There is nothing that "adds" in a computer circuit any more than in a mechanical adding machine.
This goes to show why levels of description and the concept of emergent properties are so important. Consciousness itself is unquestionably a structure that arises from deterministic processes. However, those underlying processes do not act causally in the environment with which I interact. Understanding that phenomenon requires a new level of description – the rational self. The rational self simply cannot be understood without acknowledging that we are casual agents for a future that, at least in the only frame of reference that is meaningful to us, is indeterminate. That makes our choices as real and valid as any other facet of the human experience.

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.