Did Physics Kill Free Will? (Script) (Patreon)

Physicists have a long history of sticking our noses where they don’t belong - and one of our favorite places to step beyond our expertise is the question of consciousness and free will. Sometimes our musings are insightful, sometimes incoherent, and usually at least somewhat naive. Which a fair description of this show, so of course Space Time needs to weigh in physics and  free will..

…

In recent episodes we explored the notion of determinism, in the context of the block universe idea that we get from Einstein’s relativity. In it, the past and future have a sort of eternal, timeless existence from the point of view of some god-like observer outside both space and time. And this naturally brings into focus the question of how physics relates to concept of free will.

But the most evocative expression of perfect determinism is given by Laplace’s Demon. In 1814, Pierre-Simon Laplace postulated that a sufficiently vast intellect that had perfect knowledge of the current state of the universe - like the positions and velocities of all particles - and perfect knowledge of the laws of nature - could calculate perfectly all future states of the universe.

If you that accept the mind is generated by the brain, which is made of matter, and you’ve already accepted that matter follows the cold, inviolable  laws of nature, then your mind, every single thought, choice, decision, is the inevitable result of the long chain of cause and effect that came before it. And what you do next, too, is already determined. Laplace’s demon know whether or not you’ll decide you believe in free will.

There are many ways of looking at and formulating Free Will. We’re not going pretend to give a definitive answer of its reality, but we do want to look how its popularly positioned in physics. The idea that determinism negates free will. Is it really all that simple?

One note before I go further: yes, we will mention the quantum, but importantly, that is not to say that free will or consciousness are necessarily related to it. We’ll talk about that connection or lack thereof another time. But understanding the implications of quantum mechanics is key to unraveling determinism and its connection to predictability.

Quantum mechanics seems to tell us that one of the following is true: the outcomes of quantum-scale events are fundamentally random or the outcomes of quantum events are perfectly determined by quantum laws, and the apparent randomness comes from our limited perspective. Let’s unpack what these possibilities imply for free will by looking at it in terms of information.

One of the most important rules of quantum mechanics is the principle of conservation of quantum information. It just says that quantum information can never be destroyed or created out of nothing. We’ve talked about why this is such a  fundamental law previously.

Let’s paint a cartoon representation of quantum information in a block universe.  Two dimensions of space and one of time. The slices of the block represent the causal ordering of the universe - lower layers cause the layers above them. Each bit of quantum information can be represented as a thread. Threads are evolving quantum states. These may transform and become entangled with other states, but if information is conserved a thread can never vanish, nor start out of nothing. We’ll ignore what happens at the beginning of time today.

Tangled networks of these threads may spend time as different types of matter, and sometimes as brains. Both as the crude matter that forms the brain and the ephemeral configuration of neural connections, synaptic strengths, and electrochemical dynamics from which we think emerges from the experience of our minds.

So information goes in as the stuff of the brain and the experience of our senses - and information comes out, sometimes in the form of the choices we make. The brain is like a machine for processing information into choices. We can parcel that decision machine off from the rest of the universe and we can ask: what has to happen inside this patch for us to reasonably attribute free will to it.

There are different ways we could define free will in this picture. We could require that  information emerging from the brain in the form of a decision is either 1)  an entirely new thread of quantum information, in violation of the conservation principle, or 2) that this information just not be predictable even in principle by any arbitrarily precise far future brain-scanning technology. There are other requirements we could place, like 3) that the future not be predefined and not singular. Or 4) we could require that the choice not depend on any underlying non-free-willed mechanistic process, whether that process is deterministic or random.

I’m going to argue that it’s reasonable to endow the brain with free will with some combination of the first three. If we demand number 4) - that there’s no underlying non-free-willed mechanics to the process of choice - then free will is indeed dead. But I’ll also argue that this is an unreasonable ask.

Let’s start with 1 & 2 - you are free-willed if new or fundamentally unpredictable information can emerge from your brain’s patch of spacetime.

Creating brand new information explicitly violates conservation of quantum information. That said, the most mainstream interpretation of quantum mechanics has the same problem. The Copenhagen interpretation insists that the apparent randomness of quantum events is really random. The quantum information in the state of the wavefunction before collapse is destroyed, but it’s also created. The outcome of quantum interactions are “chosen” in fundamentally unpredictable ways. Information threads both end and begin at every wavefunction collapse.

Random processes driving our choices doesn’t really sound like free will. So is there any possible way to generate new information that isn’t random? That is somehow intentional? Actually not really. Think about a new thread of quantum information starting from nothing - let’s say, emerging from a packet of spacetime where no information enters. By definition, it has no cause. But the notion of “choice” implies a deliberate cause to that new bit of information being intentionally, say, a one or a zero rather than randomly a one or a zero. But that cause is by definition a thread of ingoing information. If the cause arises within our packet - our brain - then either it itself is random or has its own causing-thread whose origin is either in our brain and uncaused - random - or has an origin outside our brain, meaning our brain didn’t “choose”. You could say that the cause arises outside “the universe” in a supernatural sense - but then all we need to do is expand our definition of the universe to all things causally connected and we reach the same conclusion: The only way for new information to come from an closed region is for it to be generated randomly within that region.

On the surface, the idea of randomly generating new streams of quantum information within a brain doesn’t seem to help the cause of free will. I’ll argue shortly that there’s a notion of free will in which it doesn’t matter the origin of the unpredictability of choice. But first let’s look at another source of that unpredictability.

Even in a deterministic universe, Laplace’s demon is a fiction. It’s not just impossible in practice to perfectly measure all the states in a brain, of neurons, their activity, the molecules involved in their signaling, etc - it’s also impossible in principle. By that I mean there’s a fundamental limit to the amount of information that can be collected about a system.  This is a direct consequence of the Heisenberg Uncertainty principle, which tells us that measuring one property perfectly leaves a counterpart property completely undefined.

So there’s no possibility to perfectly know the state of a brain - of the information threads that go in to make it - either in principle or in practice, either because random processes are at work and/or perfect measurement is not possible for any entity that can possibly exist.

But is the impossibility of perfect predictability enough to endow us with free will? What if some future type of brain scan allowed 98% predictability of your choices? This is actually a really interesting area of research in which there have been some badly misrepresented results that we’re going to talk about another time. But the short story is that even for the simplest decisions, the best people have done is to predict just a bit above chance, at 60-70 percent accuracy, and retrospectively. Which means, you could ask if that is even “prediction,” or really just finding a “correlation” between a brain state and an outcome a few seconds later.  And don’t quote the infamous Libet experiment at me - that’s been thoroughly cast in new light. The short answer is that we don’t know how well we’ll ever be able to forward-model a brain’s behavior even with the best future technology.

There is also evidence to indicate that the brain actively utilizes the random fluctuations of neuronal activity to make certain decisions. In other words, a certain type of randomness may be “a feature” of the brain, an item in its mechanistic toolbox. So  even the all knowing Laplace demon may be out of luck - especially if quantum randomness or quantum indeterminacy is magnified to brain-level activity. Again - not saying consciousness is quantum! We’ll talk more about this soon.

An important consideration is also whether or not the future is actually singular and determined, whether or not any agents of choice exist. In the Copenhagen interpretation the future is singular but undetermined - one future will happen, but it’s impossible to predict what it will be. In the Many Worlds interpretation the future is plural and determined - all possible futures exist and develop according to the laws of quantum mechanics. In either case, there’s a place for some fair notion of free will. Either our brains play a part in “choosing” what future happens, or the choices we experience are the process of choosing which Many-Worlds future we travel towards, even if other versions of us choose otherwise. The case where free will really struggles is in the case of the singular, determined future. You can get this sort of singular, hard determinism with things like de broglie-Bohm pilot wave theory, but fortunately for free will, such interpretations don’t look like they’re correct, at least as they currently stand.

Let’s talk about the 4th argument against free will in my earlier list - that a purely mechanistic process can’t lead to any notion of free will. So we’ve established that there are plausible ways in which fundamentally unpredictable information - perhaps in the form of choices - can emerge from a brain. So this nexus of quantum information processing may truly be an unpredictable black-box decision making machine, at least some of the time. Add to that the fact that its emergent phenomenon - conscious free will - can recursively influence the machine itself. The brain can talk itself into new states.

Could that be enough to designate it free-willed?

Some would argue no - if there’s a mechanistic substrate to generating this phenomenon which is founded in either deterministic or random processes then the phenomenon is an illusion. But I think it’s a reductionist fallacy to deny an emergent phenomenon any properties not also possessed by its parts. A red apple is made of atoms, but atoms possess neither redness nor appleness. Yet we don’t say that apples are an illusion. If it’s not even in principle possible to perfectly predict a brain’s choices, and that brain feels like it’s making choices, is it reasonable to deny meaningfulness to the concept of free will because a brain’s atoms don’t have free will? The same goes with your conscious experience. It’s not an illusion - it’s an emergent phenomenon - and it’s the most directly verifiably real thing you’ll ever observe. You don’t choose the mechanical behavior of your brain’s atoms or the electrical potential that triggers each firing neuron. Those things are the substrate of your choices, and your choices are you.

In the end this is all semantic. There are definitions of free will that we definitely don’t have - but both physics and neuroscience tell us that we can probably call ourselves free-willed by a number of reasonable and functional definitions. Ultimately, many agree that the question is badly posed. When you ask: Is free will an illusion, what do you mean by free or will or illusion. Or Is, for that matter. Well, go ahead and enjoy the very convincing illusion that you’re going to go down a recursive rabbithole deciding whether you have the free will to decide if free will exists. I’m now going to exercise my free will to not end this episode with me saying Space Time.