NB. I was sent this book as a review copy.

From Princeton University Press

There is much to like about this book. Kevin’s discussions on evolution and neuroscience are detailed, clear, and insightful. His writing in general flows nicely and much of the book is definitely a useful text. However, for me, this does not include the sections which actually talk about free will.

The problems start early on in the book, on page 21 where he says (bold font mine throughout):

Meaning became the driving force behind the choice of action by the organism. That choice is real: the fundamental indeterminacy in the universe means the future is not written. The low-level forces of physics by themselves do not determine the next state of a complex system. In most instances, even the details of the patterns of neural activity do not actually matter and are filtered out in transmission. What matters is what they mean – how they are interpreted by the criteria established in the physical configuration of the system. Animals were now doing things for reasons.

That causal power does not come for free: it is packed into the organism through evolution, through development, and through learning. It is encoded in the genome by the actions of natural selection….

There is nothing here that violates the laws of physics; it just demands a wider concept of causation over longer timeframes and an understanding that the dynamic organisation of a system, which encodes meaning, can constrain and direct the dynamics of its component parts.

I disagree that this doesn’t violate the laws of physics. The comment here is that there is indeterminacy in the universe and that this is resolved by brains that have meaning. How?!!

Much of the book details how top-down causation from evolution and society are the things which somehow alter the dynamics, somehow in a way which isn’t at the level of the electrochemical processes.

I agree of course that evolution has shaped our brains. However, my current brain, heavily shaped by evolution, exists in its present state. Its history does not matter. It is in a state, in the world, and unless you think that there are forces beyond those that we understand, the current state of it all in phase space (ie. not just the static, but the dynamics) should be all that one needs to talk about. Either the brain does have free will in this moment, or it doesn’t, independent of how it got here.

And through discussions of quantum physics and chaotic dynamics, the equivalent of the God of the Gaps argument seems to appear. There are aspects of the world which are non-deterministic (through quantum physics) or non-predictable (through chaotic dynamics), but how the brain somehow intervenes in these gaps is glossed over and top-down causation is invoked.

He goes on later:

…They process information in the context of their own internal state and recent experience, and they actively make holistic decisions to adapt their internal dynamics and select actions.

This represents a wholly different type of causation from anything seen before in the universe. The behaviour of the organism is not purely driven or determined by the playing out of physical forces acting on it or in it. Clearly, a physical mechanism underpins the behaviour, which explains how the system works. But thinking of what it is doing – and why it is doing it – in terms of the resolution of instantaneous physical forces is simply the wrong framing. The causation is not physical in that sense – it is informational.

We imbue meaning on top of the behaviour of organisms, of course, in a developmental, evolutionary and learning sense, but that doesn’t mean that the causal mechanisms underlying the dynamics are not simply physical ones.

In a similar vein he says:

In a holistic sense, the organism’s neural circuits are not deciding – the organism is deciding. It’s not a machine computing inputs to produce outputs. It’s an integrated self deciding what to do, based on its own reasons. Those reasons are derived from the meaning of all the various kinds of information that the organism has at hand, which is grounded in its past experience and used to imagine possible futures. The process relies on physical mechanisms but it’s not correct to think it can be reduced to those mechanisms.

Again, I want to know how? What is the mechanism that is causing the neurons to do what they do beyond what is caused by the neurons (and other electrochemically active elements) before them in the very complex sequence? Of course these neurons are all set up with synaptic connections, neuromodulators and glial structures which have been altered through past experience. But the brain is now what it is, based on how the past of the organism, and of its lineage has been encoded in those brains. And now, in this moment the brain is as it is. What does it mean for the organism to decide, beyond the neural circuits deciding?

On this possible gap in the causal chain at the microscopic level:

The contemporary philosopher and mathematician George Ellis similarly argues that physical indeterminacy creates causal slack in physical systems, which opens the door for what is known as “top down causation.” Put simply, this is the principle that the way a system behaves depends on the way it is configured, which can constrain the lower-level components and functionally select among patters of those components. []…This highlights another key principle, which also diverges from a reductive, comprehensively bottom-up view: causation is not wholly instantaneous.

This lack of instantaneousness in the causal chain also seems deeply mysterious and unphysical. Indeed from one moment to another some time later there may be a causal chain which links them, but the causal chain must itself be made up of well-defined causal influences that do happen at the microscopic level. Indeed they may even happen in a way which gives a big picture meaning, and there may be coherence between different parts of the causal chain. As I move my arm, all of it moves at the same time, and thus we see a moving arm. However, the movement is happening because at a microsopic level, each part of my arm is being moved in just the right way at just the right speed. Programmes have developed in my body that are now encoded, that allow this to happen. It doesn’t matter when they were encoded, they are there now and the structures in my brain send coherent signals to the muscles in my arm which allow for this movement to take place.

Regarding indeterminism, he further states::

From this point of view, we can see that the apparent unreliability of neural transmission at the level of (at least some) individual neurons is a feature in the system, not a bug. The noisiness of neural components is a crucial factor in enabling an organism to flexibly adapt to its changing environment – both on the fly and over time. Moreover, organisms have developed numerous mechanisms to directly harness the underlying randomness in neural activity. []…These phenomena illustrate the reality of noisy processes in the nervous system and highlight a surprising but very important fact: organisms can sometimes choose to do something random.

Indeed it is well known that neural signalling is an inherently noisy process, and we can simulate populations of neurons making decisions, for instance this paper by Beck et al. It is just not clear how an organism can choose to do something random beyond what the neural populations are doing. Where exactly is the cause? What is the mechanism of the top-down causation? This remains unclear.

This highlights the concept of “free-will of the gaps”:

Martin Heisenberg [] noted in a 2009 essay that physical indeterminacy opens the door for true agency and free will. The constant jitteriness of neural activity means that the whole system is not predetermined to adopt any particular state: there are degrees of freedom in the system that the organism can exploit. The agent itself has both the power and the time to decide. Indeed, we have time to think, and choose, and change our minds, and think again if we need to.

I fear that I am becoming a broken record here: But how? How does the organism exploit these supposed degrees of freedom? Does it somehow get into the about-to-collapse wave function and nudge the collapse one way or another? If so, how? He hints at one point into the finite nature of the universe and how in chaotic systems, because you need an infinite amount of precision to truly determine anything, there is hidden indeterminacy there as well. The question remains in this case. How is the agent able to fiddle the books, down at the level, that the universe itself doesn’t otherwise have access to? If it is through “meaning” and evolutionary history, then I still need to know the mechanism. Calling it simply top-down causation just fills a gap in my understanding with a word that doesn’t itself tell me how. It all feels very much that because of the desire not to see the death of free will, the author must utilise these sleights of hand.

Anyway, I think that I have said enough on these points.

The other thing that I think absolutely needs highlighting is the fact that there are large chunks of the beginning of the book where prediction, imagination and action are discussed, without once mentioning Active Inference or Predictive Coding. He is essentially describing these topics in quite clear terms, but does not mention that there is a well-known framework for these ideas. Kevin either was unaware of the Active Inference framework or opted not to mention it explicitly. This is a confusing omission.

So, as you might be able to tell, I do not agree with the overall message of the book. The concept of free will emerging through evolution presents significant gaps. The illusion of free will is strong, but if we are going to argue that free will itself exists, there must be very clear mechanisms by which we get to fiddle at the microscopic levels of the universe in ways that apparently the universe doesn’t have access to, and that is something for which a great deal of evidence would be needed.