Meat Computers and Metaphysics

Jerry Coyne and Massimo Pigliucci have engaged in intellectual internet fisticuffs on the problem of free will, which is unsurprising given that they seem to disagree on nearly everything. But their particular disagreement is more interesting than "We have free will." "Nuh-uh." "Yeah-huh." This dispute has brought up some interesting questions about the nature of the free will debate, which I'd like to explore a little bit. What kind of question are we asking when we ask if we have free will? Is it a metaphysical problem? Scientific? Conceptual?

In the posts I've produced intermittently over the last few months on the free will topic, I've presented it as a philosophical puzzle that's at least partially conceptual in nature. Certainly our views on free will should be informed by the sciences, but I haven't assumed that there will be some finding of science that will solve (or dissolve) the problem without requiring philosophical analysis.

Jerry Coyne seems to disagree with this approach, taking the question out of the philosophical and placing squarely--and solely--within the purview of science. But first, his definition of free will:

When faced with two or more alternatives, it's your ability to freely and consciously choose one, either on the spot or after some deliberation. A practical test of free will would be this: If you were put in the same position twice — if the tape of your life could be rewound to the exact moment when you made a decision, with every circumstance leading up to that moment the same and all the molecules in the universe aligned in the same way — you could have chosen differently.

This is a fairly straightforward and I think common concept of free will, which takes the ability to do otherwise as the key factor in determining if an agent's actions count as freely done. Coyne recognizes that the issue of free will is not an "arcane academic debate about philosophy" (why does philosophy always get described as "arcane," by the way?) but touches on issues of moral and legal responsibility as well.

Coyne claims we don't have free will; our sense that we are making choices that could have turned out differently than they actually did is an illusion.

You may feel like you've made choices, but in reality your decision to read this piece, and whether to have eggs or pancakes, was determined long before you were aware of it — perhaps even before you woke up today. And your "will" had no part in that decision. So it is with all of our other choices: not one of them results from a free and conscious decision on our part. There is no freedom of choice, no free will.

Coyne bases this conclusion on two lines of evidence.

First, we are biological creatures, made of physical molecules that are governed by the same physical laws that govern the rest of the molecules in the universe, and thus there is not room for some mysterious, nonphysical "will" to direct your actions in a way other than those that the laws of physics ultimately dictate. "Everything that you think, say, or do, must come down to molecules and physics." Coyne argues that science has shown that our personal identities are "constructs of the brain" and thus not independent of the physical causes by which the brain functions.

Second, Coyne notes recent work in neurobiology specifically that appears to show that conscious decisions are not conscious at all:

Our brains are simply meat computers that, like real computers, are programmed by our genes and experiences to convert an array of inputs into a predetermined output. Recent experiments involving brain scans show that when a subject "decides" to push a button on the left or right side of a computer, the choice can be predicted by brain activity at least seven seconds before the subject is consciously aware of having made it.... "Decisions" made like that aren't conscious ones. And if our choices are unconscious, with some determined well before the moment we think we've made them, then we don't have free will in any meaningful sense.

Coyne thinks the upshot of this scientific evidence is clear: "The ineluctable scientific conclusion is that although we feel that we're characters in the play of our lives, rewriting our parts as we go along, in reality we're puppets performing scripted parts written by the laws of physics."

But are these findings as straightforward and unphilosophical as Coyne would have us believe? Massimo Pigliucci thinks not. He first takes issue with Jerry's "tape of life" thought experiment, arguing that such an experiment cannot ever actually be carried out, which means it's an empirically untestable criterion. A result, the thesis is actually a metaphysical rather than scientific thesis:

Science, if nothing else, is about empirically testable hypotheses, to which the above scenario certainly does not belong. Rather, Jerry et al. are making a metaphysical argument, an approach with which I’m fine, to a point, as a philosopher, but that is strange coming from people who clearly despise the very idea of metaphysics and scorn anything that cannot be approached by the empirical methods of science.

He then argues that Coyne's argument based on physics rests on some implicit philosophical assumptions that he has not argued for: that the physical domain is causally closed, that we have a functional concept of "cause", that determinism is true, and that there are no emergent properties. Pigliucci thinks that none of these assumptions is adequately defended, by Coyne or anyone else, which means that Coyne has more work to do if these concepts will be employed to argue against free will:

I just don’t think any of the above issues has been settled, and since it is Jerry who is making an extraordinary claim — that we are profoundly mistaken in our first person experience about free will, consciousness and morality — it seems fair to point out that he lacks the corresponding extraordinary evidence.

Pigliucci also argues against Coyne's use of the evidence of neurobiology. Focusing on the Libet experiments, he points out that need not be taken to have demonstrated that there is no such thing as free will, since

reporting awareness of an urge (in this case, of pushing a button) hardly qualifies as a conscious decision. The latter is the kind of reflective deliberation that Jerry and I engaged in while composing our respective essays, and it is simply not measured by Libet-type experiments.

Pigliucci goes on to contest some other features of Coyne's article, which I won't comment on. Coyne responds to a number of the issue mentioned here, which I won't discuss either, in part because it appears to me that the disputants are talking past each other. Which actually brings me to the point of this post.

It seems to me that a significant source of the difficulty revolves around two things. First, there is no broad agreement about what kind of question "Do we have free will?" is. This first difficulty is caused by the second, namely that there is no broad agreement about what is meant by "free will."

What is at stake in determining what kind of question the free will problem is? How we view the question of whether or not we have free will is going to significantly alter what sorts of answers are going to seem palatable. Coyne views the question as a scientific question, and thus bypasses the persistent conceptual issues from philosophy of science that Pigliucci points to. Pigliucci, on the other hand, sees the question as a metaphysical problem, which can't be resolved fully from within science.

If the question of free will is one that is squarely within the purview of science, then Jerry Coyne is totally within his rights to adopt the methodological assumptions of science and proceed in just the sort of way he proceeds. He offers a thought experiment, marshals the findings of modern neuroscience, all under the reasonable assumption that the causal closure of the physical obtains, and that we have a rough and ready concept of cause that is good enough to frame the discussion about whether so called free actions have determinate causes or not, and whether those causes work in a deterministic rather than stochastic fashion.

But if the question of whether we have free will is a metaphysical rather than scientific question, then Pigliucci is right to argue that Coyne can't just help himself without argument to the methodological assumptions of science.  Instead, there are many more distinctions that might be relevant to whether or not we want to say that an agent is acting freely or could have done otherwise. If the problem of free will is a metaphysical rather than scientific problem, then there are several notions of necessity that come into play before we can pronounce that some event could not have been otherwise.

If we're speaking strictly in physical terms, then only physical necessity is relevant, which is why under Coyne's scientific description of the problem the answer seems clear cut. Given the state of affairs of the universe at the time of a decision, the agent could not choose otherwise than he has done, since under physicalism and the laws of physics, every event is fully determined by some preceding cause. But physical necessity is not the only kind of necessity that needs to be considered if the question isn't a strictly scientific question. There is logical necessity, but no one really thinks that any particular action is logically necessary except under some contrived set of circumstances.

More to the point here is metaphysical necessity. One way to think about metaphysical necessity is to think about whether or not some state of affairs obtains in all possible worlds. Now, one might argue that in this world the state of affairs preceding a particular choice determined it and the agent could not have done otherwise. But, there is a conceivable state of affairs in some other possible world which might have obtained in which the agent took a different action. The question becomes, is this metaphysical contingency sufficient to ground a robust conception of "free will"? I think it would depend on the details. If I were to say that an agent decides to X in world W, but decides to Y in world Z, which is identical to W except for the agent's desire (i.e., in W he wanted to X but in Z he wanted to Y), then we might be willing to accept such a metaphysical indeterminacy as exemplifying free will on the part of the agent. He's acting in line with his volitions or desires, which might just be enough for some to call it free will, or at least, free enough.

These disagreements about what type of question the free will problem is are born out of a lack of clarity about what the notion of "free will" even means. Even when writers extremely careful to define their terms (like Coyne always is), the discussions wind up going past each other because the definition is unsettled. Coyne seeks to define the notion clearly and in terms of a counterfactual that is at least responsive to empirical evidence. This is fine, if indeed the problem of free will is best addressed by subjecting it to scientific methodology. Similarly, those who would be comfortable with the broad outline of a theory of metaphysical freedom sketched in the preceding paragraph would likely be so because their definition of free will is something more like the Lockean version I discussed previously.

But without a shared definition, it's easy to see how these discussion devolve quickly into participants talking past each other.