The philosopher Christian List has written a new book defending the existence of free will. Has he managed to crack the ancient riddle?
The traditional problem of ‘free will’ is how to reconcile our supposed capacity to ‘have chosen otherwise’ in a universe governed by deterministic physical laws. If one event simply follows another, how can we ever make real choices that affect how the future turns out? Free will seems incompatible with ‘physical determinism’.
In his book, List lays out a new conception of both of these concepts, in the hopes of revealing that they are not inconsistent after all. I highly recommend reading it yourself, regardless of your background knowledge. His arguments are clear and concise, with hardly any tedious philoso-jargon. You can also hear him discussing it on an episode of Michael Shermer’s podcast: Science Salon.
I’m going to briefly summarise his key arguments, before explaining what I think he gets right and wrong.
List thinks that for an agent to possess free will, it must have three capacities:
- The capacity to act intentionally
- The capacity to choose between alternative possibilities
- The capacity to control its actions
The book is structured in chapters laying out what each of these capacities are, and defending why they can and do exist in our universe.
Now, no one really disputes whether humans have the first capacity. We all experience acting intentionally, whether we’re grabbing a cup of tea or yelling at another car in traffic. The reason ‘free will’ has been such a contentious topic in philosophy for thousands of years is that we can’t seem to make sense of the second two capacities.
List acknowledges that the traditional problem with capacity (2), being able to choose between alternative possibilities, is that it seems impossible if we live in a physically deterministic universe. There can be no such thing as ‘alternative possibilities’, if the universe simply unfolds like a the ticking of a clock. Given certain initial conditions and universal laws of physics that determine how one state leads to the next, the future appears to be completely fixed.
There is also an issue with capacity (3): whether we can control our own actions. What this claim really means is that we can control our actions with our conscious intentions. In other words, we wouldn’t have free will if our intentions were simply by-products of unconscious processes, and therefore had no real causal efficacy on the world.
But the problem with this is explaining how a non-physical thing (e.g. intentions) can cause a change in the physical world. Opponents of free will tend to claim that only physical events can cause physical changes, and so conscious intentions must be merely ‘epiphenomenal’ — like the smoke coming out of a train’s engine.
Let’s start by thinking about this capacity, because this is where I think List is highly persuasive.
To simplify, List basically points out that our best explanations of human behaviour cannot avoid treating abstract or non-physical entities as real causes. This is because when we think about causation, we are thinking about what events make a difference to how situations unfold.
To take the tea-drinking example again, if we only consider the micro-physical aspects of the situation, we fail to put our finger on what exactly would have led to a different outcome. Sure, you can tinker with the micro-physical state, by relocating a few atoms, but in most cases the tea will still be drunk. In other words, it doesn’t matter what most of the atoms are doing. What matters is the abstract state of ‘wanting to drink some tea’, which there is actually no satisfactory micro-physical description of.
This is because of ‘multiple realizability’ — the fact that macro-states like ‘wanting to drink tea’ are realisable by many, in fact, infinitely many, micro-states.
What this means is that to get a genuinely causal story about why I drank the tea, you’d have to include my consciously felt desire and intention to drink it. It follows that if such intentions play a crucial role in our causal story, then we have to assume they have genuine causal potency.
List argues this point far more nuanced way, and I recommend reading the whole of Chapter 5 in which he does so.
So let’s get back to free will capacity number (2), because is where I think List’s argument fall short. This concerns whether we can really choose between alternatives in a world governed by physical determinism.
List‘s definition of determinism is fairly standard:
Given the complete physical state of the world at any point in time, only one future sequence of events is physically possible. “Physically possible,” in turn, means “compatible with the fundamental physical laws.”
But he spends Chapter 4 arguing that this ‘low-level’ kind of determinism, that applies to atoms and quarks, need not imply ‘high-level’ determinism, in the realm of beliefs, desires, and intentions.
With the aid of the helpful diagram below, List represents different low-level physical histories, each with their own initial conditions, and each progressing in a strictly deterministic manner from time t=1 up to t=6.
He then shows that when you abstract across these low-level states in different histories to form aggregate high-level states, the high-level states can end up progressing in an indeterministic manner. That is, certain histories branch off into multiple possible future states, instead of only following the one fixed path.
The idea is that while the low-level events (in the first diagram) may unfold deterministically, the high-level events (in the second) need not. This leaves room for us to make choices that affect the course of the future at the psychological level, even though determinism reigns at the lowest levels of reality.
But on his blog, Jerry Coyne rightly questions the relevance of this argument to our own predicament, given we presumably live in a single history / universe, with the low-level initial conditions presumably set for good back at the time of the Big Bang. This means that our universe consists of only one of those low-level histories, as well as the corresponding high-level history that is in turn fixed by it.
Contrary to List, it seems that so long as you grant that the psychological level ‘supervenes’ on the low-level reality, determinism at the low level fixes determinism at the high-level. (For non-philosophers, ‘supervenes’ in this context just means that if you change the high-level facts, then the lower level facts must change as well. For example, if I suddenly cease wanting to drink a cup of tea, then something in the physical world must have changed as well, presumably somewhere in my brain).
So Coyne is right that List’s argument doesn’t seem to work. But I think they’re actually both wrong in how they construe ‘physical determinism’ in the first place.
The definition they both agree upon assumes a ‘clockwork universe’, which unfolds fatalistically according to physical laws. But there’s actually no need to cling to this conception.
There are really interesting challenges against this notion of determinism being developed by physicists, perhaps most notably by David Deutsch and Chiara Marletto at Oxford. With their project of Constructor Theory, they are seeking to re-frame fundamental physics away from the ‘initial conditions’ plus ‘laws of motion’ conception. This would do away with the idea of some defined function that takes as input the current ‘entire physical state of the universe’ and mechanistically outputs the next one.
The right way to think about the fundamental laws of physics, on their view, is as constraints on what kinds of physical transformations are possible and impossible. The point is that this is perfectly compatible with evolved organisms within the universe creating explanatory knowledge and using this to transform the world as they desire (i.e. exercising their free will), without being trapped on a single fixed trajectory. Joe Boswell has some great interviews of Deutsch and Marletto on YouTube discussing these ideas.
Advocates of the scientific method like Coyne, List, and Harris are rightly worried about ditching the concept of determinism because they think the alternative is ‘mysticism’, which would allow for all kinds of silly miracles and supernatural beings. But such concerns are not warranted under the ‘constructor theoretic’ conception. According to it, we still live in a universe governed by timeless, fixed laws — it’s just that these laws do not dictate by themselves how exactly the future will unfold.
The physical laws that make it possible for us to be conscious and creative human beings, making real choices about what will happen next, are the very same laws that rule out Jesus spontaneously converting water into wine, or rising from the dead.
Given this alternative way of thinking about fundamental physics, we don’t need to accept the notion that the universe evolves according to some predetermined plan, set from the beginning of time. Our best theories of physics don’t require it, and our best ethical, psychological, and political theories must reject it.