Why Free Will is Compatible with Determinism

Recently I was having a fun and interesting conversation with someone I’d just met, a clearly very bright person, and during our talk I commented that I’m a compatibilist determinist – someone who believes that free will and determinism are compatible concepts. My clearly very bright new friend dismissed the idea as obviously false. In fact, he trivially dismissed it; his immediate response was that claiming that free will and determinism are compatible didn’t even merit the weight of consideration he’d given the rest of the conversation.

I get that reaction a lot. Narratives about the future are almost always about how it is an unfixed fog of infinite possibility, crystalizing onto history and becoming solid in the flash of the present moment. Almost all time travel stories include the instance or threat of someone going back in the past and changing the future, positing an assumed indeterminist model of reality. In those very few (and usually very good) time travel stories that do take place in a determinist universe, the discovery of the fixed nature of the future is usually treated as a tragic diminishment of possibility, an attack on free will. But outside of academic writing, free will is very rarely defined. My impression is that most people, even irreligious people, have a sort of vague sense of what they mean by “free will” that is inherently numinous, that free will is some ineffable quality by which we are masters of our own fate. When I’ve asked people to try to explicate to me what they mean when they say “free will,” they usually say something about the existence of choice, and free will being the thing that makes choices possible, and that choice is what gives life meaning. The notion that a fixed future annuls choice seems intuitively obvious.

It isn’t. Here I’d like to offer a fairly straightforward argument for why not.

Prologue: The Predictor

Consider the following hypothetical scenario conceived by Ted Chiang.1 Imagine that I gave you a toy, a little box the size of a remote for unlocking a car, called a Predictor. The Predictor has only a button, a small LED light, and some internal circuitry that I tell you has a built-in negative time delay. When you play with the Predictor, you find that the light always illuminates exactly one second before you push the button.

I give you a Predictor to play with, and at first if feels like a game, like the goal is to push the button right after you see the light flash. But if you try to break the rules, you find you can’t. If you try to press the button without having seen the light, no matter how fast you move it always flashes exactly one second before your finger gets there. If you try to wait for the light to flash, intending to not push the button after, then the light never turns on. There’s no way to fool the Predictor.

So, I’ve given you a toy that conclusively demonstrates that your future actions are already determined. Having played with this toy, the question is, do you stop feeding yourself? Or paying your bills? Or caring for your family? You’ve been shown a demonstration that the events of the future are, in principle, precisely determinable from the present. It either is or is not true, already, that you will feed yourself, get dressed, pay your bills, bathe your children, etc. So do you think that you, the person reading this article, would stop doing those things after I gift you a small toy? If you don’t believe that you would stop, then you have some intuition that a fatalist attitude of defeated meaninglessness is not a necessary consequence of determinism. It remains for me to motivate exactly why that might be, and to convince you that your choice to continue doing things like feeding yourself is, in fact, a choice, despite the revealed deterministic nature of the world. I will try to do that now, then return to this hypothetical example.

1. Definitions

Let’s get the terms out of the way.2 As used here, determinism is the notion that at any given moment the physics of the universe admits only a single possible future.

Choice is a little more complicated. When I use the terms “choice” or “decision,” I am referring to categories of actions undertaken by specific agents. For these choices to be free, they must be caused strictly by processes internal to the agent, but that alone is not a sufficient definition. Taking people to be the agents in question, there are plenty of actions that people perform for reasons that are strictly internal, but are not volitional. (Sneezing, for example, or breathing while asleep.) So let us say that an action is a choice or decision if and only if it is caused by the agent’s beliefs and desires.

This definition still admits some quibbling. How are we to consider, for example, addiction? It motivates action based on beliefs and desires, but we feel in some sense that the desires have been warped by an external factor. Similarly with compulsion, where we feel that desire has been warped by a non-volitional internal factor. So let us even more narrowly define a free choice: a choice is free if and only if (1) the agent would have acted differently had it so chosen, (2) the action was voluntary (unaffected by internal restraint), (3) the action was uncoerced (unaffected by external restraint).

It’s worth noting that if we consider things like addiction and compulsion to be, in some fashion, an adulteration of free will, then seeing such things exist in human beings lends credence to the mechanistic description of free will that I am going to develop. But what do I, specifically, mean by “free will?” Let us describe an agent as having free will if and only if the agent possesses the ability to perform deliberative processes that result in choice.

2. A Brief Hearing for Indeterminism

Let’s look quickly at the notion of free will in the absence of determinism. Imagine, instead, that the future is completely independent of the present, that there is no causal relationship between this present moment and the one to come. What, then, does choice consist of? If there is no causal relationship between the present and the future, then, as David Hume noted, any attempt to make an informed choice as to how you should act in pursuit of a goal is futile. In the purely indeterminist case, past experience is not a logical guide for future behavior. So rather than free will being obviously incompatible with determinism, it is in fact pure indeterminism that trivially excludes the existence of free will. Choice can not exist in a universe where the future is random. Choice requires some degree of causal relationship between past, present, and future.

“Some degree,” though, is a quibble phrase. Perhaps, one might argue, the universe is determinist-ish. It’s predictable enough for choice to exist, but not perfectly predictable. Perfect predictability, one might wish to argue, also excludes free will. If one wishes to take this perspective, then one has to answer the question: at what point does the somewhat predictable universe become too predictable for free will to exist? I will now argue that choice can exist even in a purely deterministic universe, sliding that boundary point right off the scale.

3. Deliberative Process as Physical Event

Posit that we exist in a perfectly deterministic universe, and consider the case where you throw a ball at my face. If I am asleep–eyes closed, largely insensitive to my environment–then you are very likely to hit me. If I am awake–eyes open, watching your throw–then you are unlikely to hit me. (What do I mean though, in this deterministic universe, by “likely” and “unlikely?” I’ll address that more later. For now, let’s say that if we repeated many trials with slightly varying contextual conditions, in most of the cases where you throw the ball at my face while I’m asleep you hit me, and in most of the cases where you do so when I’m awake, you miss.) I am, when awake, able to avoid the ball, using faculties unavailable to me when I am asleep. I am able to avoid the ball, but I don’t have to do so. If, say, you’ve thrown a ball at my face because we are playing baseball, I might perceive some advantage in allowing it to hit me. That is to say, I might avoid avoiding the ball. I’m able to do this because our species has evolved sensory apparatus (sight, proprioception) that allows me to know when projectiles are coming at my head and move out of the way. When those apparatus are nonfunctional, such as when I’m asleep, I can’t move out of the way.

Events that I have the capacity to perceive and avoid are what Daniel Dennett likes to call evitable3 (to distinguish from those that are inevitable). Now, you may be objecting, “the universe was posited to be deterministic. Whether or not the ball hits your face was already determined the moment it leaves my hand. That makes it inevitable.” To which I would respond: inevitable to whom? It is clearly not inevitable to me; as the baseball/non-baseball example shows, whether or not the ball hits me is influenced by my beliefs and desires. Perhaps you mean inevitable to the universe. But that is not a meaningful notion, the universe is not a volitional agent. It does not make sense to speak of the universe avoiding, or avoiding avoiding.

The motions I make with my body after you throw the ball at me result from a cascade of electrochemical events in my brain, which correspond to my weighing the desirability of the ball hitting my face, the position of my body, how I would have to move to avoid the ball, etc. This electrochemical cascade, this physical event, is itself a deliberative process, one that results me choosing to dodge or not dodge the ball. All of my initial definitions for free will have now been met: I am an agent possessed of the ability to engage in deliberative processes that result in the perpetration of a choice: decision to dodge or not dodge the ball. That decision is definitionally a choice, in that it is an action caused by my beliefs and desires, where my beliefs are my sensory/conceptual model of the world, and my desires are my internal preferences. It even meets the more restrictive definition of being a free choice. The fact that the dodge or the hit was already extant in the future when the ball left your hand is irrelevant. The free choice was already extant, too.

I know of no logically coherent way to define freedom of choice that is incompatible with choice as an event that can occur within a deterministic universe. As long as choice is a behavior arising from a deliberative process, it is compatible with determinism. Thus free will, as I have defined it, is also compatible with determinism.

4. The Issue of Counterfactuals

One reason that it seems (incorrectly) to many people that determinism is incompatible with choice is that our deliberative processes which result in choice involve the consideration of counterfactuals. We think to ourselves, “What would be true if I did X? What would be true if I did Y? How likely is it that any action I take will result in Z?” This notion of likeliness seems to be challenged by determinism. How can one meaningfully think of things being likely or unlikely if the future is already determined? But in truth there is no contradiction. When we utilize counterfactuals in our deliberative processes, we are conducting mental simulations based on our beliefs and our understanding of past experiences. Our ability to judge how likely we think something is does not depend on what actually later occurs. The mental events of simulation and prediction are just part of the deliberative process that results in choice.

We talk about counterfactuals in a confusing way, though. If I am standing at the free throw line on a basketball court, and I shoot a free throw that bounces off the rim, I might be heard to say, “I could have made that.” What does that statement mean? I’m not actually saying that if everything about the state of myself and the world were somehow exactly the same down to the minutest detail, and the situation were to recur, I would make the shot. Rather, I’m making a claim about counterfactuals. I’m saying that among the family of possible worlds admitting minute variations of the air, moisture on the ball, potential gradients along the ion channels of the cells in my muscles; in many of those possible worlds I make the shot. Here again, the fact that when I took the shot there was only one physically possible future does not invalidate my counterfactual analysis. Just as it is meaningful to say that it is, was, and always will be the case that I missed the shot, it is also meaningful to say that I (counterfactually) could have made it. It is simply semantic ambiguity that makes these two notions seem to be in conflict.

Epilogue: The Predictor, Again

So let’s go back to the case of the Predictor. If I were to gift to you a tiny toy that happened, by implication, to demonstrate that the future already existed, of course you wouldn’t stop feeding yourself, or paying your bills, or acting in the interest of others you care for. You do these things because you believe they matter, and make choices motivated by that belief. The only reason for your choices to change would be if your beliefs fundamentally changed. Maybe you’ve been previously convinced, for no good reason, that determinism would mean that nothing matters. Then playing with the Predictor might be dangerous. You might then, as some people in Ted’s story do, choose to abdicate all personal responsibilities and never do anything again. But that’s not the Predictor’s fault, nor the universe’s; it’s the fault of you having “determinism = meaninglessness” in your head as a disabling, destructive narrative . That would be a tragedy if it were to happen–but why should it? The Predictor is just a small piece of plastic that you can throw in the trash if you want4, and anyway, reading this has taken a long time and you’re hungry. Might as well go eat something.


  1. In his story “What’s Expected of Us,” Nature, 2005 

  2. I’m indebted for portions of this section to Curtis Brown, my symbolic logic professor at Trinity University. 

  3. Dennett discusses evitable and inevitable events at length in his book Freedom Evolves, which is a much more learned and thorough explanation of compatibilist determinism than this article. 

  4. All of the discussion up to this point has been about a hypothetical universe that I simply posited at the start was deterministic. I haven’t made any claims about the actual reality you and I inhabit, nor have I placed a tiny plastic Predictor in your hand. And, of course, I can’t. They don’t really exist. But I think I can give you something that is very close to the same.

    Special relativity has as one of its basic results that simultaneity does not have any meaning across reference frames. The math for this isn’t too complicated, but instead of writing out equations, here’s a two minute video that clearly demonstrates the phenomenon:

    There are many versions of this thought experiment, but the one in the video is the one Einstein proposed. As you can see in the video, the man on the platform sees the bolts of lightning strike both ends of the train simultaneously, and the woman riding in the train sees lighting first strike the front of the train, and then strike the back of the train. And neither person is wrong. In the man’s frame of reference, the strikes were simultaneous. In the woman’s, one came before the other.

    Special relativity has been experimentally confirmed time after time. As theories go, it’s one we are as sure of as we are sure of anything at all. Consequently it is already widely accepted that simultaneity as a concept has no meaning across reference frames. But lets think through the implications: the man on the platform observes an event that, at the time he observes it, is still in the woman on the train’s future. That is to say, there exist an event–the lightning striking the back of the train–that is in the man’s past, and in the woman’s future. Thus it is possible for a physical event that is already in my past to still be in your future. But the past is unchangeable. Anything that is in the past, for anyone, is necessarily a thing that happened in the universe. But if my unchangeable past can be your future, that means that an event in your future is determined. And this reasoning can, in principle, apply to any arbitrary event. Therefore the future already exists, and the events of the future are already determined.

    Conclusion: special relativity implies that the universe in which we live is, in fact, deterministic. 

1 Comment

Add yours →

  1. How are beliefs not part of a self-ascribed internal restraint? I also think that to argue free choice is meaningless. Choices are made irrationally and are necessarily random. That we have the freedom to change our beliefs is where this thing takes root. our thinking is capable of change through intentionally driven desires whether consciously or subconsciously formed. in my view, I could be wrong.

Leave a Reply