A handful of musings about free will have been popping up in my blog reader of late. Jerry Coyne has been discussing the issue with Eric MacDonald in a series of posts (further links therein). Russell Blackford writes a long post that he promises isn’t the post he will eventually write, David Eagleman has an article in the Atlantic, and Zach Weiner also chimes in. So we have a biologist studying theology, an ex-Anglican priest turned agnostic, a philosopher and neuroscientist both of whom write science fiction, and a webcartoonist studying physics. That constitutes a reasonable spectrum of opinion. Still, what discussion of reality is complete without a cosmologist chiming in?
In some ways, asking whether free will exists is a lot like asking whether time really exists. In both cases, it’s different from asking “do unicorns exist?” or “does dark matter exist?” In these examples, we are pretty clear on what the concepts are supposed to denote, and what it would mean for them to actually exist; what’s left is a matter of collecting evidence and judging its value. I take it that this is not what we mean when we ask about the existence of free will.
It’s possible to deny the existence of something while using it all the time. Julian Barbour doesn’t believe time is real, but he is perfectly capable of showing up to a meeting on time. Likewise, people who question the existence of free will don’t have any trouble making choices. (John Searle has joked that people who deny free will, when ordering at a restaurant, should say “just bring me whatever the laws of nature have determined I will get.”) Whatever it is we are asking, it’s not simply a matter of evidence.
When people make use of a concept and simultaneously deny its existence, what they typically mean is that the concept in question is nowhere to be found in some “fundamental” description of reality. Julian Barbour thinks that if we just understood the laws of physics better, “time” would disappear from our vocabulary. Likewise, discussions about the existence of free will often center on whether we really need to include such freedom as an irreducible component of reality, without which our understanding would be fundamentally incomplete.
There are people who do believe in free will in this sense; that we need to invoke a notion of free will as an essential ingredient in reality, over and above the conventional laws of nature. These are libertarians, in the metaphysical sense rather than the political-philosophy sense. They may explicitly believe that conscious creatures are governed by a blob of spirit energy that transcends materialist categories, or they can be more vague about how the free will actually manifests itself. But in either event, they believe that our freedom of choice cannot be reduced to our constituent particles evolving according to the laws of physics.
This version of free will, as anyone who reads the blog will recognize, I don’t buy at all. Within the regime of everyday life, the underlying laws of physics are completely understood. There’s a lot we don’t understand about consciousness, but none of the problems we face rise to the level that we should be tempted to distrust our basic understanding of how the atoms and forces inside our brains work. Note that it’s not really a matter of “determinism”; it’s simply a question of whether there are impersonal laws of nature at all. The fact that quantum mechanics introduces a stochastic component into physical predictions doesn’t open the door for true libertarian free will.
But I also don’t think that “playing a necessary role in every effective description of the world” is a very good way of defining “existence” or “reality.” If there is anything that modern physics has taught us, it’s that it’s very often possible to discuss a single situation in two or more completely different (but equivalent) ways. Duality in particle physics is probably the most carefully-defined example, but the same idea holds in more familiar contexts. When we talk about air in a room, we can describe it by listing the properties of each and every molecule, or we speak in coarse-grained terms about things like temperature and pressure. One description is more “fundamental,” in that its regime of validity is wider; but both have a regime of validity, and as long as we are in that regime, the relevant concepts have a perfectly good claim to “existing.” It would be silly to say that temperature isn’t “real,” just because the concept doesn’t appear in some fine-grained vocabulary.
We talk about the world using different levels of description, appropriate to the question of interest. Some levels might be thought of as “fundamental” and others as “emergent,” but they are all there. Does baseball exist? It’s nowhere to be found in the Standard Model of particle physics. But any definition of “exist” that can’t find room for baseball seems overly narrow to me. It’s true that we could take any particular example of a baseball game and choose to describe it by listing the exact quantum state of each elementary particle contained in the players and the bat and ball and the field etc. But why in the world would anyone think that is a good idea? The concept of baseball is emergent rather than fundamental, but it’s no less real for all of that.
Likewise for free will. We can be perfectly orthodox materialists and yet believe in free will, if what we mean by that is that there is a level of description that is useful in certain contexts and that includes “autonomous agents with free will” as crucial ingredients. That’s the “variety of free will worth having,” as Daniel Dennett would put it.
I’m not saying anything original — this is a well-known position, probably the majority view among contemporary philosophers. It’s a school of thought called compatibilism: see Wikipedia, or (better) the Stanford Encyclopedia of Philosophy. Free will as an emergent phenomenon can be perfectly compatible with an underlying materialist view of the world.
Of course, just because it can be compatible with the laws of nature, doesn’t mean that the concept of free will actually is the best way to talk about emergent human behaviors. (Just because I know the rules of chess doesn’t make me a grandmaster.) There are still plenty of interesting questions remaining to be clarified. At the very least, there is some kind of tension between a microscopic view in which we’re just made of particles and a macroscopic one in which we have “choices.” David Albert does a great job of articulating this tension in this short excerpt from a Bloggingheads dialogue we did some time back.
I don’t generally think that the superior wisdom one acquires via training as a physicist grants one the power to see clearly through complicated issues and make philosophical conundrums dissolve away. But this is a case where insights from physics might actually be useful. In particular, what we are faced with is the task of reconciling effective theories at different levels of description that have apparently incompatible features: the impersonal evolution of the microscopic level (whether we go all the way to atoms, or stick with genes and neurons) and the irreducible possibility of “choice” at the macroscopic level.
This kind of tension also appears in physics. Indeed, the arrow of time is a great example. The microscopic laws of physics (as far as we know) are perfectly reversible; evolution forward in time is no different from evolution backward in time. But the macroscopic world is manifestly characterized by irreversibility. That doesn’t mean that the two descriptions are incompatible, just that we have to be careful about how they fit together. In the case of irreversibility, we realize that we need an extra ingredient: the particular configuration of our universe, not just the laws of physics.
In fact, the connection goes beyond a mere analogy. If you look up arguments against compatibilism, you find something called The Consequence Argument. This is based on the “fundamental difference between the past and future” — what we do now affects the future, but it doesn’t affect the past. Earlier times are fixed, while we can still influence later times. The consequence argument points out that deterministic laws imply that the future isn’t really up for grabs; it’s determined by the present state just as surely as the past is. So we don’t really have choices about anything. (For purposes of this discussion we can ignore the question of whether the microscopic laws really are deterministic; all that really matters are that there are laws.)
The problem with this is that it mixes levels of description. If we know the exact quantum state of all of our atoms and forces, in principle Laplace’s Demon can predict our future. But we don’t know that, and we never will, and therefore who cares? What we are trying to do is to construct an effective understanding of human beings, not of electrons and nuclei. Given our lack of complete microscopic information, the question we should be asking is, “does the best theory of human beings include an element of free choice?” The reason why it might is precisely because we have different epistemic access to the past and the future. The low entropy of the past allows for the existence of “records” and “memories,” and consequently forces us to model the past as “settled.” We have no such restriction toward the future, which is why we model the future as something we can influence. From this perspective, free will is no more ruled out by the consequence argument than the Second Law of Thermodynamics is ruled out by microscopic reversibility.
None of this quite settles the question of whether “free will” is actually a crucial ingredient in the best theory of human beings we can imagine developing. I suspect it is, but I’m willing to change my mind as we learn more. The context in which it really matters is when we turn to questions of moral responsibility. Should we hold people who do bad things responsible for their actions — even if our understanding of neuroscience improves to such an extent that we can identify precisely which gene or neuron “made them do it?” (This is the focus of Eagleman’s article.)
This is a resolutely practical question — who gets thrown in jail? Criminal law has the concept of mens rea, guilty mind. We don’t find people guilty of crimes simply because they committed them; they had to be responsible, in the sense that they had the mental capacity to have known better. In other words: we have a model of human beings as rational agents, able to gather and process information, understand consequences, and make decisions. When they make the wrong ones, they deserve to be punished. People who are incapable of this kind of rationality — young children, the mentally ill — are not held responsible in the same way.
Might we someday understand the brain so well, reducing thought to a series of mechanical processes, that this model ceases to be useful? It seems possible, but unlikely. We know that air is made of molecules, but the laws of thermodynamics haven’t lost their usefulness. Thinking of the collections of atoms we call “people” as rational agents capable of making choices seems like a pretty good theory to me, likely to remain useful for a long while to come. At least, that’s what I choose to think.