The biologist Robert Sapolsky has a new book out where he argues that free will doesn’t exist. I think Sapolsky is wrong and his arguments are bad. Now, if you’d asked me 15 years ago, I would have been totally on Sapolsky’s side. When I first entered graduate school, I was what is known as a “hard determinist” (like Sapolsky). That is, I thought that the laws of nature are deterministic (or deterministic enough; quantum mechanics complicates things a bit), and that determinism implies that we have no free will. Therefore, we have no free will. However, over the course of my studies, I came to appreciate the case for compatibilism, the view that free will is compatible with our actions being determined. I’m now a compatibilist myself. This post explains why. These arguments aren’t necessarily the best possible argument for compatibilism (whatever that might mean). But they’re the arguments that convinced me. Perhaps that counts for something.
The first argument that convinced me is an argument that takes our regular way of speaking seriously. We say that people did things “of their own free will” all the time. And we distinguish this from cases where people do things “against their will.” How do we draw this distinction? Roughly, we do it by evaluating whether or not the person in question wanted and intended to do the action in question. If I really want a donut, and so I decide to go down to Dunkin and get a donut, and then I do go get that Dunkin donut, then everyone will agree that I did that of my own free will. In general, if you do something because you intend to do it, and you intend to do it because you want to do it, then when you do it, you do it “of your own free will.” This is just how we talk.
We distinguish actions that we do of our own free will from actions that we do against our will. Actions that we do “against our will” include actions where we’re literally forced from outside: if a strong wind blows my arm forward, then I didn’t raise my arm of my own free will. We will often also count coerced actions as being “against our will.” If someone puts a gun to my head and threatens to shoot me if I don’t get a donut, then I may get a donut, but not of my own free will. I do it because I intend to, but not because I want to.
There’s plenty of room to try to figure out exactly the way in which these two kinds of actions are different. That’s an open and interesting philosophical debate. But there are clearly a class of action where we are happy to say that people act of their own free will. And this isn’t because they are magically violating the laws of nature or anything like that. It’s just because they do it because that’s what they want to do. But, of course, doing something because you want to do it is consistent with the laws of nature determining your actions. Indeed, it’s what makes “doing something because you want to do it” possible at all. The regular, natural causal relations which ensure that your desires and intentions will make a difference to your actions is an important part of the story of how acting from your own free will is possible. (If intending to do X in no way guaranteed that I would actually do X because my limbs flailed about wildly in an irregular and uncaused manner, I could not act from my own free will.) So acting from my own free will is compatible with the existence of deterministic laws of nature.
Of course, a hard determinist will say that that’s not the point. The point is that what we want and intend is determined by past circumstances, in accordance with inviolable laws of nature. So really we couldn’t have done otherwise. And thus we have no free will. I’ll tackle this argument head-on in a second. But before I do, I’ll say that I think that this argument is the one that misses the point. Judgments of whether something is done “of your own free will” are extremely common, and they seem to be intimately tied up with whether the action was done as the result of a desired intention or not. To ask about the causes of that desired intention is, in an important way, to simply change the subject.
This is the point in the dialectic where people will start to talk about two different concepts of free will: the “compatibilist” conception of free will and the “libertarian” conception of free will. The compatibilist conception of free will is what I’ve been talking about for the last four paragraphs. If you do something because you desire and intend to do it, then you do it freely in the compatibilist sense. But there is another sense of the phrase “free will.” And what really matters is whether or not we have free will in this other sense.
What is this other sense, the “libertarian” sense? It has to do with alternate possibilities. Could you have done otherwise? If so, you acted freely. But if not, you did not. This brings me to the second argument I heard in graduate school that had a big impact on my thinking in this area. There are not two senses of the phrase “free will.” There is just one, and that one sense is the compatibilist’s sense. Libertarian free will, the free will of “could have done otherwise,” is just another way of saying that you did it because you wanted and intended to do it. For, after all, if you hadn’t wanted and intended to do it, you wouldn’t have. That’s just what we mean when we say that you did it because you wanted to!
I bet you’re unconvinced by that. Feels like some sort of a trick, right? But wait, what are we even talking about in the first place? Let’s back up.
The concept of “libertarian free will” is (something like) “You acted freely in doing X only if you could have done something other than X.” In philosophical terminology, libertarian free will is a “modal” concept. It’s about what could have happened, or what would have happened (in certain circumstances), or what is possible (or impossible). So how do modal terms work? That’s an absurdly complicated and controversial question in philosophy. But here are two relevant observations.
First, modal terms have variable meanings. The same sentence containing a modal term can mean different things depending on the context in which that sentence is used. To use an example from the philosopher David Lewis: I am a mature human with an intact and well-functioning brain. As such, I have all of the capacities that I need in order to speak any learnable human language, like English, French, Latin, or what have you. I can speak Swedish. But please don’t ask me a question in Swedish; I can’t speak Swedish! So: can I speak Swedish? Or not? In one sense, I can: my biology does not prevent me from speaking Swedish. But in another sense, I can’t: my actual state of knowledge of the Swedish language prevents me from speaking Swedish. Hold onto that thought, I’ll come back to it in a second.
Here’s a second relevant observation about modality: We talk about what could be the case or what would have been the case all the time. “I could have tried harder.” “If I’d knocked the glass off the table, it would have broken.” “If Lee hadn’t assaulted Meade’s lines directly on the third day at Gettysburg, he wouldn’t have sacrificed a critical portion of his army, and he could have won the war.” These are not bizarre and esoteric ways of speaking. If anything, it would be bizarre to not speak this way. BUT (and this is an important point): the laws of nature are deterministic. If we hold fixed the deterministic laws of nature, together with facts about the far past, then nothing could have happened except what actually happened. I couldn’t have tried harder, I couldn’t have knocked the glass off the table, Lee couldn’t but have sacrificed a critical portion of his army in Pickett’s charge, and he couldn’t have won the war. So are all our ways of thinking about counterfactuals fundamentally flawed? Of course not.
To say that something is possible is to say that it’s consistent with certain constraints. Which constraints? That depends on context. So could I have tried harder? Could Lee have won the war? That depends on what what we mean. (Could I have learned Swedish?) Lee’s winning the war is not consistent with the actual course of events. But Lee’s winning the war is consistent with looser constraints. When we evaluate counterfactuals, we usually try to hold the laws of nature fixed, as much as possible. But we can’t hold the laws completely fixed. When we ask a counterfactual question, we’re asking something like “What if the world worked basically the way it does, but instead of X happening, Y happened instead?” Now, if we assume deterministic laws of nature and assume some particular facts about the way the world was at some point long before X, X would have to have happened, so this question doesn’t make sense. But it does make sense! So, Lewis concludes, when we are asking counterfactual questions, what we’re imagining is that some sort of “divergence miracle” has occurred. What if the laws of nature slipped, for just a moment, and something else was allowed to happen? The big bang and the laws of nature together determined that Lee would sacrifice his army at Gettysburg. But what if there was a hiccup in the laws of nature right when Lee was deciding whether to authorize an infantry charge on Sentinel Ridge, and as a result he decided to hold back? What then? “A hiccup in the laws of nature!” you scoff. “Impossible!” Well, holding fixed the laws of nature, that’s impossible. But we’re not holding fixed the laws of nature. That’s the whole point of assessing a counterfactual. “What if the slip did happen, although of course in a sense it couldn’t have?” (Could I have spoken Swedish if greeted by a Swede? Sure! But I’m glad it didn’t happen, because I couldn’t have spoken Swedish!)1
The point is that, if we hold the laws of nature and some initial conditions fixed, I couldn’t have done otherwise. But if we hold the laws of nature and some initial conditions fixed, nothing could have been otherwise. When we ask what could have been, we always hold some facts about the past fixed. But we never hold everything fixed because, if we did, then nothing that didn’t actually happen could have happened. This is just how the counterfactual game is played.
Back now to that donut. I chose to get it. Could I have done otherwise? Holding fixed the laws of nature and all the facts about the past: no. Of course not. But that’s trivial, and that’s not what we care about. We’re playing the counterfactual game, and when we play that game, we don’t hold everything fixed. We allow for small “divergence miracles.” So I ask again: I chose to get the donut. Could I have done otherwise? Well, sure: if I had chosen not to. My choosing not to would have required a divergence miracle. But all counterfactuals require us to imagine a kind of divergence miracle. We can’t make sense of them otherwise.
So when we do something, do we have libertarian free will? Could we have done otherwise? Well, yes. That would require a divergence miracle; that would require the laws to have been temporarily suspended. But that’s just how counterfactuals work. The laws don’t change, but so what? Counterfactuals ask us what would happen if they did. The argument for hard determinism asks us to play the counterfactual game without playing it. Could things have been different if nothing was different? Obviously not. That’s clearly not what we’re asking when we’re asking what could have happened. So what are we asking when we ask if we could have acted differently? Really? Well, we’re asking whether our will was causally efficacious in bringing about our actions, such that, if we had chosen differently, we would have acted differently. That’s the libertarian conception of free will. And it’s also the compatibilist conception of free will.
No matter how you look at it, the compatibilist has an incisive understanding of what free will is. And that’s why I’m a compatibilist.
This example, and the argument that follows from it, is from David Lewis’s classic paper “Are We Free to Break the Laws?” I learned it from my professor Kadri Vihvelin, who expanded Lewis’s argument in that paper into a whole book, which I read in a seminar that I took with her on the subject of free will. It’s a great book and it was a great seminar. Kadri is amazing. She had a huge impact on my way of thinking about free will. Why am I a compatibilist? Because I was an incompatibilist, and Kadri taught me better.
Very interesting post. However, I want to note that we don't have to say that if I hadn't gotten the donut, the laws of nature would have been momentarily suspended in such a way that I would not have gotten the donut. We can instead say that the microscopic state of the world would have been different (all the way back to initial conditions of the universe if the laws are deterministic) in such a way that I wouldn't have gotten the donut. I know that this is somewhat counter-intuitive, but I find it more intuitive than saying that if I hadn't gotten the donut, the physical laws of nature would have been different. On a counterfactual theory of causation, this latter view implies that the actual laws of nature hold because I decided to get a donut, which doesn't sound right at all.
Thank you for writing such a good defense of compatibilism. I agree with nearly everything you wrote, and yet...
I still don’t believe in free will, and I think compatibilism, not determinism, is the trivial perspective.
Yes, we can play with counterfactuals, and learn from them. Counterfactual thinking and free volition (to distinguish from free will) are critical parts of the matrix of determinism.
But whether or not you eat a donut, with or without a gun to your head, is not the important aspect of determinism. Much more important is the question of whether you have a say in going to hell for all eternity or not...
Consider how many people there are in the world who believe that they will be eternally punished or rewarded for the choices they make.
It makes a huge difference to them if they learn that they and the people they care about couldn’t actually do otherwise. Determinism changes the identity of God and challenges his omnipotence and benevolence. It changes people’s calculations and motivations for doing good and bad things.
Whether or not we believe in libertarian free will changes how we feel about how people treat us, and how quick we are to forgive.
In my opinion, the determinism that is conceded in compatibilism is not trivial at all. It is a huge deal that most people haven’t yet taken onboard, and it has enormous practical, social, and psychological consequences.
The moral dimension is the whole ballgame, and the rest is mostly curiosa – a bit of a semantic game of mostly academic interest, that clouds the larger issues by confusing the language.