Yes, You Might Have Free Will

by Greg Skiano | 14 min read

Physicists occasionally speak as if physics can answer our most fundamental questions, and if not, then we best not ask. Take, for example, a youtube video created by accomplished theoretical physicist Sabine Hossenfelder entitled “You don’t have free will, but don’t worry.” In the video, which has over 800,000 views, Dr. Hossenfelder argues against the reality of free will, and yet, despite the avalanche of comments beneath the video suggesting it is brilliant, amazing, cool, and basically a knockout, I remain underwhelmed. At one point, Hossenfelder quips that humans having free will “makes absolutely no sense if you know anything about physics.” However, I doubt the opposite idea (that humans do not have free will) makes any sense if you know anything about humans.

The argument according to Hossenfelder is a familiar one and goes something like this. Physicists know that reality consists of a system of particles governed by deterministic laws, which means that the state of the system at any given time is determined by the initial state of that system. Since humans are part of the system, their brains, their thoughts, and all their “choices” are the result of interactions between particles, nothing more. We might point to chaos or quantum randomness to make room for free will, but both fail. On the one hand, chaos implies merely that something is unpredictable, not that it is undetermined; on the other hand, randomness means precisely that something is undetermined, whereas free will usually means having the power to determine something about the future. Thus, concludes Hossenfelder, free will does not exist, and furthermore, the idea that we can choose between alternate futures is totally incoherent—or rather, incoherent so long as we do not question the certainty of physicists, which is something a scientific amateur (such as myself) would never deign to do.

Luckily, David Hume already did. He argued that, ultimately, all our empirical knowledge depends on what is sometimes called the uniformity principle, the idea that the laws of nature remain the same everywhere and for all time. The validity of the law of gravity, for example, depends on the fact that tomorrow, and the next day, and every day after that, every particle in the universe will still attract every other particle with a force described by the universal law of gravitation (or whatever the newer, better post-Newtonian laws would say). But to justify our confidence in this uniformity principle we require something besides physics, and the reason we do is obvious: our confidence in physics depends on the uniformity principle being true. For Hume, this circular dependency shows that all empirical knowledge is essentially a matter of custom. If we take this conclusion seriously, then nothing we “know” from physics can definitively rule out the possibility of free will since we have no certainty that the laws of the universe remain the same everywhere in the universe, including the recesses of the mind. However, even if we dismiss this skeptical conclusion as unwarranted, we still have a problem, and a more interesting one. If we want to say that humans have any knowledge about the way the universe really is, even though we cannot justify the uniformity principle deductively or inductively, we must assert that there is something about humans that makes us able to know, a power the only evidence for which is our experience. And if it is our experience that grounds our claim to knowledge—not some physical law—then we already accept we have a power for which we find no justification in physics. Free will, if it exists, would be another such power. So I ask: are the two really so different that we can claim with utter certainty that humans are able to know but not able to choose?

I think Hossenfelder would likely say yes, the two are critically different in the following way. Unlike our ability to know, our apparent ability to choose is inconsistent with the best evidence we have. The idea here is that any empirically unverifiable proposition must at least cohere with our best empirical evidence, otherwise it must be false—a view I basically agree with. However, I do not agree that the proposition “humans have free will” is (inarguably) incoherent with our observations. Obviously, if we claim that reality simply equals that which we can observe, then there can be no inconsistency between our conception of reality and our observations. But—and this is an objection I doubt will please Dr. Hossenfelder—if our conception of reality includes more than we can possibly observe, that conception need not necessarily be inconsistent with our observations either.

Imagine a chess game. Note that the board, the pieces, and the rules of chess were all created by humans. Now imagine that, somehow, we imbued the chess pieces with consciousness of the board, the other pieces, and (amazingly) themselves. They cannot perceive our world or our fingers moving them. They can only perceive the world of the game. Despite that limitation, they have discovered the rules of chess through observation, noting that each kind of piece can only make certain kinds of moves and that the immediate effect of each move is perfectly predictable. They also realize that even though pieces have a wide range of potential moves open to them on each of their turns, the actual moves they make are fairly predictable based on the likely effect they will have on the future of the game. The pieces agree that the rules are real and almost certainly unbreakable, but they cannot seem to agree on what moves the pieces, what causes them to select this move rather than that one. Some of them ignore the problem, assuming that the pieces move themselves of their own accord, content to simply enjoy the game. Others insist on arguing and split into two general groups. The first argues that since the game follows deterministic rules insofar as they can observe, and since the selection of moves is clearly not random, then the selection of moves must also be accounted for by a more complete understanding of the rules. The second objects that this is a leap in reasoning, that even a complete understanding of the “rules” would not give an account of the existence of the game, nor would it explain the power that drives the changes in the game. This power, they suggest, implies an origin outside the game. Within this second group there is further disagreement, some positing that this external power determines all the moves that are made, others positing that the pieces have some role in determining the moves for themselves, albeit in a way that is still dependent on the external power. Crucially, none of these views contradict what the pieces are able to observe, and all of them parallel views that have been and will continue to be held by serious people about our own reality.

If “nature” according to Hosenfelder means the totality of particles in the universe, and there actually are events in nature that are determined by free will as commonly understood, events not determined by previous events and not totally random either, resulting from a power belonging to agents themselves, then it follows that their cause must be something outside, above, or otherwise separate from the laws we observe, a cause that is (strictly speaking) super-natural. Furthermore, if humans are the cause of such events (i.e. they have will), then they must somehow participate in the supernatural. Stated another way, supernatural will could exist without human will, but human will could not exist without supernatural will. If we claim it does, then Hosenfelder is quite correct that we are speaking nonsense.

At least Hossenfelder’s argument against free will is logical, if not compelling. The same cannot be said for her argument about why you shouldn’t worry. To ease our worries, she tries to address two common objections: 1) that people will not behave morally if they realize they have no free will, and 2) that people cannot understand themselves without reference to this “non-existent” free will.

Regarding the first, she tells us that free will is not needed to explain morality. Since we are all “just running software that is trying to optimize our own well-being,” it follows that we will take action against other people who harm us, preventing people from behaving in ways we would describe as immoral. Those who cause harm must be stopped simply because they “embody the problem,” and this is true even if they do not act freely. The assumption here is that so long as our behavior toward others does not become more harmful upon realizing we have no free will, we have lost nothing, an assumption which requires us to dismiss many questions as meaningless—questions about intention, self-denial, blame, praise, judgment and so on—questions which deserve consideration, but which I set aside for now. Instead, I want to focus on the second concern, the concern that we just have a difficult time understanding ourselves without free will.

If we are bothered by the “cognitive dissonance” that comes from believing we are decision makers, even though we know we cannot make decisions, Hossenfelder offers the following alternative:

I suggest you think of your life as a story which has not yet been told. You are equipped with a thinking apparatus that you will use to collect information and act on what you have learned… The result of that thinking is determined, but you still have to do the thinking. That is your task. That is why you are here. I am curious to see what will come out of your thinking, and you should be curious about it too.

I for one am curious about how I could actually take this advice, and not just because taking advice about what I “should” do implies making a choice. If I think of my life as a story unfolding before me, that makes me like someone watching a movie of my own life. But when I watch a movie, attending to the story with curiosity—as Hossenfelder suggests I should—I am not passive. I am active. If I point a video camera at the movie I am watching, its sensors capture the very same light and sound I receive with my senses, but the camera is passive in a way I am not. My ability to engage with the story requires that I direct my attention, which is an act of will.

I suspect Hossenfelder would protest that, though my response to the movie is more complex than the camera’s, it is still deterministic because it is caused by particles that follow deterministic laws. However, my point here is only that her proposal—that I should see myself as a viewer of my own story rather than an actor in it—offers no alternative at all since viewing implies agency the same way doing does. As a result, her proposal only multiplies the problems. Many of the words she uses to offer her alternative become hard, if not impossible, to grasp if we accept her conclusion. Aren’t expressions such as “have to” and “should” pure nonsense? And shouldn’t the everyday understanding of “you” as well as its distinction from “I” be discarded as misleading illusions once we realize that at best we are only patterns of information that emerge from this system of particles and not two beings in the sense we commonly assume? And, finally, doesn’t the word “do” strongly suggest an agent, which is exactly what we will obliterate if we extrapolate her argument?

Perhaps we can change the common understanding of words like should, you, I, and do, carefully redefining them in a way that is compatible with total determinism so that Hossenfelder’s advice makes sense, but then she would be inconsistent in another way. Earlier in the same video, she firmly rejects all compatibilist attempts to redefine “free will,” declaring them nothing more than “verbal acrobatics.” She clearly thinks that when we realize free will cannot mean what we intuitively think it means we should abandon the idea. But if she uses words like should, you, I, and do as they are commonly used, then she is guilty of invoking free will indirectly. Conversely, if she attempts to redefine them in a way compatible with her conclusion, then she is guilty of the same verbal acrobatics she condemns.

Let me be clear. I am not arguing that humans definitely have free will; nor am I arguing that you should ignore scientific evidence simply because it is doubtable. What I am arguing is this: Even if you believe—as both Hosenfelder and I do—that a satisfactory definition of free will is incompatible with both determinism and indeterminism, there are still at least two views that are compatible with what we know about the physical universe. One view is that free will does not exist. Full stop. Another is that free will does exist but depends on supernatural will for its power. There may be others as well—a not unlikely scenario, given philosophers have been thinking about this for a long, long time… philosophers, I might add, far more capable than Hosenfelder and I—but setting that aside, we can still profitably compare these two views.

There are many reasons to prefer Hossenfelder’s view. First, and most obvious, this view appears as if it emerged from the progressive accumulation of knowledge. People in the past, living as they did in ignorance and superstition, had to content themselves with quasi-religious arguments, but not us! No, we have the scientific method and, armed with science, have escaped the confines of all their quaint beliefs. However, one need look no further than Lucretius and the Epicureans—whose discussion of atoms and their “swerve” bears an uncanny resemblance to recent attempts to save free will with quantum mechanics—to see that two-thousand years ago people were debating this issue in terms we find relevant today. Even Hosenfelder says her view is obvious if you know “anything” about physics; and even though people today know more about physics than people two-thousand years ago, those ancient people still knew something about physics. It is not obvious how incremental improvements in physics could fundamentally change this argument. Therefore, it shouldn’t surprise us that ancient people recognized the deep tension between free will and empirical observation.

Second, this view appears scientific and, thus, falsifiable. Unlike dogmatic assertions of “supernatural” powers, this view is held only so long as the evidence remains in its favor, and its superiority flows from its corrigibility. But is it really open to correction? If the hypothesis is that nothing in the universe acts freely, then one single observation of humans freely acting is sufficient to reject it. And don’t we have countless observations of just that? We observe people acting freely all the time, others as well as ourselves, and we can even observe that some actions are clearly more or less free than others. Hossenfelder restricts the definition of “observation” so as to exclude these kinds of observations, but why? We have many methods of observation, and a slavish adherence to one (say the methods of theoretical physics) strikes me as impoverished. No reason for it is given, and Hosenfleder assumes she can explain away our self-perceived volitions as the result of material causes without giving us an account that closes the massive gap between the level of particles and the level of human action. That’s because no complete account is available. Maybe to Hosenfelder this gap is uninteresting, but if she expects us to reject all that is plainly evident about ourselves and our actions, this conspicuous elision is worrisome. It begins to appear as if her view is, in actuality, totally unfalsifiable. We cannot offer her any examples of free actions that she will accept because she is already certain they are illusions. It is not that evidence supports her conclusion, but rather that her conclusion provides a lens through which all evidence is interpreted.

Third, this view stakes out the high ground in the free will debate. It is a viewpoint that can be absolutely tolerant of its opposite. Why get upset about those who want to believe in free will? It makes no difference anyway, and if it makes them feel better, why bother getting worked up? This kind of detached acceptance of opposition is not just placating. It is powerful too. Compared to the flabbergasted and naive proponent of free will, who incredulously asks how anyone could really accept such utter nonsense as determinism, the cool-headed denier appears mature, elevated, sophisticated. And even if the defender of free will does so with poise and precision, the denier still retains an advantage: for her there is nothing truly important at stake, whereas for the defender everything is. If free will is at the very foundation of your view of what it is to be human, losing this debate risks demolishing everything you have built atop that foundation, and having so much at stake is a real disadvantage in rational debate, where caring too much can cripple you.

But if your goal is not victory but truth, perhaps you should risk adopting the tactically weaker position. It is, I think, better to have something worth losing than to have nothing at all, and that is precisely what Dr. Hossenfelder’s view leaves us with. Nothing. Nothing worth caring about, at least. Free will, by itself, can seem abstract and expendable, but the things that depend on it cannot be dismissed so lightly. Once we give up free will, we give up not just responsibility, but forgiveness, gratitude, care, and finally love. And for what? So that we won’t be deceived by superstition? Kierkegaard, in the opening of Works of Love, warns that “one can be deceived in many ways; one can be deceived in believing what is untrue, but on the other hand, one is also deceived in not believing what is true; one can be deceived by appearances, but one can also be deceived by the superficiality of shrewdness, by the flattering conceit which is absolutely certain that it cannot be deceived.” We would do well to heed this warning, lest we confine ourselves to beliefs so narrow, so confining, that all the hard-won knowledge of humanity we receive through philosophy, art, tradition, and even our own experience, evaporates before our eyes.