Free Will Is Real (with Kevin Mitchell)

Kevin Mitchell: Thanks very much. Thanks for having me.

Kevin Mitchell: Yeah. It kind of struck me. I mean, it was a true story: I was literally watching him play this video game.

And it struck me that the game kind of encapsulated this concern that people have had–over millennia, really–that there was the contrast between him as the player, where he seemed to be acting really for his interests–he had goals in the game; he was pursuing those goals; he was making decisions given the scenario that the game presented to him, and so on–and then these other people that he would encounter, these non-player characters. Like, the bartender, where you go and you ask them something. And then the bartender–which is just pieces of code, really–responds somehow.

But, it seems like they’re also characters, they’re also entities, these non-player characters.

The concern is that what you–you have this sort of intuition that there’s a difference, that you say, ‘Well, I’m doing things because I want to for my own reasons, and they are just playing out their programming.’

And then the concern is, like, ‘Well, okay, but what if I’m just playing out my programming? What if I’m just acting out what the genetic code says I should do, generally speaking, or what my experiences have wired into my brain at any moment so that I have an illusion of choice, but maybe I’m fooling myself? Maybe I’m a non-player character, just an automaton.’ And that’s the existential freakout moment that you can have when you start thinking about these issues too deeply.

Kevin Mitchell: Okay. Good.

Kevin Mitchell: Right. The strange thing about this is that it seems like we go about making decisions all the time. That seems to be the kind of bedrock of our experience. And yet, there’s these lingering doubts that people have had for thousands of years about how this could possibly be the case. And, if it was trivial to rebut them, then there wouldn’t have been a need to write the book, right?

So they have some purchase. I mean, these are strong intuitions once you start thinking about them.

And the first level is the level we already talked about: the idea that, ‘Yeah, it seems like we’re freely in control of our actions,’ but actually we’re being driven by our habits. We’re being driven by the psychological programming that’s in our minds based on our past experiences. And I think someone like your previous guest, Robert Sapolsky, would argue for that–

Kevin Mitchell: He would say that all of the experiences that you’ve had to date are shaping your mind and your brain in such a way that when you encounter some new scenario, you develop the intent to do something. And you do that. You do what you want to do. But the emergence of the intent is not up to you. You don’t have a hand in that. Your brain is just suggesting it to you.

Kevin Mitchell: Yeah. The argument is that that’s an illusion. Exactly. That you’re not really free when–the intent just pops into your mind, and then you carry it out. Right? That would be the argument.

Now, I’m going to jump ahead and say: I just don’t buy that argument. There’s lots of evidence that the intent doesn’t just pop into your mind. That, in fact, you are actively involved in making the decision about what to do. The intent is the end point of that product–sorry, of that process–that you are actively engaged in. It doesn’t just appear, most of the time. In some cases it’s true, but not–one of the key things here is you shouldn’t extrapolate from what happens in one kind of scenario to every kind of scenario. And I think that’s a mistake that people have made across this literature, generally.

Okay, but we’ve got this sort of psychological intuition that maybe we are really non-player characters. We’re just driven by all of our past experience to not really have a choice in the moment, just at a psychological, cognitive kind of level.

But, you can go below that. Right? Neuroscientists are very much materialists in that they’re interested in how the machine works. Right? And as neuroscientists, we’ve been incredibly successful in getting in there and studying what’s going on, not just in people’s brains, but in animals as they are making decisions. With things like neuroimaging technologies, where we can scan someone’s brain while they’re making decisions, we can kind of see the cogs turning. We can say, ‘Well, look, I see something happening in your brain when you’re making a decision, and it’s in this circuit, or this circuit. If you make this decision, I’m likely to see more activity in your amygdala, or your prefrontal cortex, or wherever the case might be.’

And that, you know, gives the impression that actually your brain is making the decision–not you.

That’s a weird framing. Right? Because it kind of suggests: on the one hand, we have this very materialist, reductive sort of view of the mechanism. And then on the other hand, you’ve got you.

Well, what’s you? Some free-floating ghost in the machine? It’s a strange framing.

And, I think there’s two ways to look at this neuroscientific evidence. One is to say: Clearly, your brain is making the decision, not you. And the other is to say you are making the decision using your brain. Like, what else would you use? Obviously, there’s going to be activity in your brain while you’re making a decision.

But still, there’s other experiments in animals where it’s not just observing what’s happening: it’s driving activity. We can go in and we can activate specific parts of an animal’s brain, and we can make it go to sleep. We can make it move forwards or backwards, or roll around. We can make it hunt. We can make it do reproductive behaviors. All sorts of things can be kind of remote-controlled in mice and rats, and so on.

We can even change the way that they think. We can change their decision-making so that they are more risk averse, or they’re more impulsive, or something like that.

And once you can do that–when you can control the machinery of decision making–it really just focuses the mind on the mechanistic aspect, and it leads to the kind of intuition that the whole organism as an agent is actually not doing anything. It’s a place where stuff is happening, but it’s really being driven around by what’s happening within its parts. And that’s obviously a concern.

Kevin Mitchell: Yeah. Yeah. A lot has been made of these kinds of cases where, especially in people with neurological damage to some part of the brain, they can do things for reasons that they’re not aware of. And then, when you ask them, they often make up a story. It’s called confabulating. They make up a story. They’re trying to interpret their own behavior because they don’t have access consciously to some processes that they normally do, and then they make up a story post hoc. So, this is an example where people have, I think, overextrapolated from that kind of data. They say, ‘Okay, well, here in these patients with neurological conditions, or under conditions where it’s a particular experimental setup that uses some psychological thing, it gives you subliminal processing that you’re not aware of, a sneaky psychologist design.’ You can also get people to confabulate even when they don’t have any neurological damage.

Kevin Mitchell: That’s true[?]–

Kevin Mitchell: That’s the extrapolation–

Kevin Mitchell: Exactly–

Kevin Mitchell: Yeah. So, it’s a very seductive argument, and it carries a lot of force. The idea is we started at the psychological level. We said, ‘You’re psychologically determined to act in certain ways.’ And then we went down to the neuroscience level and said–the implication is, ‘Look, all that psychological talk is kind of epiphenomenal. It’s not really causal. The fact that you believe something or desire something is not actually the cause of your behavior. It’s these neurons firing in the machine.’ So, it reduces behavior to the output of an electrical engineering system, almost: it’s at the neural level.

But, why stop? Why stop at that level? You can keep going down, because you can say, ‘Well, look, neurons are made of molecules and atoms, and they’re going to obey the laws of physics, so everything that happens within you is actually dictated by the laws of physics.’ There, you get to the deepest level of determinism, exactly the concern that you just articulated, which is that if those laws are fully deterministic, then everything that happened from the Big Bang till now was already predetermined.

Kevin Mitchell: It’s cooked in.

So, let me just explain what that means for the laws to be deterministic. What it means is that if you have the state of the universe at a given time point, and it’s fully defined–fully, precisely defined at some time point–then you just apply those numbers to the laws of whether it’s–it could be Newton’s laws, it could be the Schrodinger equation when it comes to quantum fields and things like that, whatever the level at which you’re doing the calculation. And then there’s just an outcome that’s inevitable as you go along. That’s the argument.

And, we interpret that idea that the laws are deterministic, really, from Newton’s laws of classical mechanics because what he showed, say, for the orbits of the planets, is that when you can really, really predict where there are going to be thousands of years in advance, very, very accurately, and there doesn’t seem to be any chancy element there. There’s no bit of randomness there in those equations.

Kevin Mitchell: Yes. So, there’s this fictional being that was proposed by a French mathematician whose name was Laplace, which came to be called Laplace’s Demon. The demon in this case was this omniscient being. And, the interesting thing for this being was not just that they could predict everything that was going to happen, but that actually time would be meaningless for them. All of time would just be laid out as a whole thing all at once because any moment going forwards, or backwards, would entail every other moment in it. Right? So, that’s the argument.

And in fact, if you look at Einstein’s theory of relativity, he has this idea of a block universe where it’s very much the same ideas: that everything is sort of fixed, and there’s no–there’s actually no definition of the present moment within that. Nothing distinguishes the present from the past and the future. Which is interesting given that that’s such an obviously important part of our experience, that these major physical theories don’t know what to say about it. That’s an interesting thing, a wrinkle that gets overlooked, I think.

Kevin Mitchell: Exactly. Exactly.

Kevin Mitchell: Yeah. And so, the implication is, if–first of all, the implication is the future is fixed. There are no possibilities. Possibilities just don’t exist. That’s not a thing that exists in a deterministic universe. And so the idea that we might be exercising choice doesn’t even arise. It’s just not relevant–

Kevin Mitchell: or applicable, yeah, under those circumstances.

And so, the odd thing about this from my point of view, looking at the debate that has raged for millennia, is that it’s always free will versus determinism. And, determinism, many people just take it that physics says that that’s the case: physics is deterministic, that that’s a result of physics. And it’s absolutely not. It’s just: physics doesn’t say at all that the world is like that.

So it’s really weird that the entire debate in free will–when you’re talking about free will, people just assume that it’s free will versus determinism. In fact, you can make a very good argument from physics that there’s inherent indeterminacy in the way that things evolve in physical systems. That’s true at the quantum level, which is, I think, well known. But it’s also true at the classical level, which is less well-known.

And, the implication of that, then, is, Well, now we’ve got a different challenge, if we’re talking about free will. In a deterministic universe, you have to ask where could the freedom come from? There’s no freedom in that universe. Everything’s just inevitable.

But, in an indeterministic universe where many things could happen, where there’s some probabilistic nature to what goes on based purely on the laws of physics, well, then you have to ask for an organism, where does the control come from? Right? How can an organism have some control over what goes on? Because, it feels like either the universe is deterministic–in which case you have no choice–or at the low levels there’s some indeterminacy. But, if the way that the indeterminacy plays out just in terms of physical interactions, if you can’t affect that and it ultimately manifests in your behavior, then you also have no choice. So it feels like we’re on the horns of a dilemma.

Kevin Mitchell: It is a good argument, but it misses one vital point.

And the vital point is this–and this is really actually kind of the core, I guess, if you want to say metaphysical argument in my own approach. The argument is that the fact that the future is open, that when you have some indeterminacy at the low levels–and that comes from not just from quantum stuff; there’s other arguments for why there must be some indefiniteness to the future if the future is really open–if that’s true, then what it means is many things could happen. It’s just not the case that the current state of the universe says only one thing could happen. We start at time t, the current state; then, at time t+1, it’s actually a bit indefinite. Multiple things could happen.

And then the question is, what other kinds of factors could constrain what happens? Right?

And there you get the possibility that actually the way that a system is organized could do some work to constrain the trajectories that are possible–the possibility space that emerges.

And, so this is actually utterly commonplace, even though people seem to think it poses this huge metaphysical problem. Like, the computer I’m using right now is constraining the paths and trajectories of electrons within it by virtue of the hardware being organized a certain way, and by virtue of the software that is running. And it does that without violating the laws of physics. It’s not going down, pushing individual quantum events around. It’s actually not meddling or micromanaging at that level. There’s still all this noisy improbability happening down there–sorry–indeterminacy happening down there.

What it’s doing is constraining the possibility space of how things could go at a macroscopic level. It doesn’t matter–my computer running this program, it’s not affected by individual electrons zigging or zagging. It’s affected by populations of electrons, and robust, emergent statistical behavior in the circuits that are doing things.

So, you can get macroscopic organization that controls things at a macroscopic level without, like I said, violating anything. There’s no new laws of physics. There’s no overriding the core theory of quantum mechanics or anything like that. It’s just that that level, the core theory of quantum mechanics, is just not causally complete because it’s inherently got probabilistic elements. It opens some causal slack for something else to be involved.

Kevin Mitchell: Yeah. No, that’s great. So, I mean, my approach in the book is–yeah, like you say, to take an evolutionary approach, to try and understand not just how humans can control their behavior, but how any organism can be said to be doing anything. I mean, so before life emerged, there weren’t any doings in the universe. There were just happenings. Things were happening.

And then, at some point, these entities emerged that seemed to be able to do things. Right? They’re causal agents unto themselves. And, that’s a really deep question. And, it felt to me like if we didn’t get a handle on it at that level, we’d never understand humans, this most complicated sort of instantiation of these powers as we see them.

So, what I wanted to do was build up some of the concepts gradually and naturalize them so that it didn’t feel mystical if, at the end, I was talking about us doing things for a reason, things having meaning for us, acting on purpose, things having value. We can get a handle on those ideas in very simple organisms, even at the origin of life, and get a sense of what does it mean?

Like, we want to understand what it means to be a human being. At the core of that, the question is, what does it mean to be a living being? What is life really about? And, for me, the core element of life, which is actually, I think, often overlooked in biology, is agency. It’s the fact that living things do things. They act in the world. So, all of this hinges on this idea of macroscopic organization. And, the one extra element that you need to add to this to understand it is the idea of selection.

So, you could have–in a complex system many things could happen. And, when that’s true, then you could end up going one way or the other. Right?

Now the question is, can you arrive at an organization that is better at persisting as a pattern than another organization?

And, just randomly, by chance, somewhere in there, you’ll get some patterns of activity that are sort of self-reinforcing. So, in our world, on our planet, that emerged as sets of chemical reactions that reinforce each other. So, you get these sets of reactions where the product of one reaction is the substrate for the next, and you get these sort of interlocking cycles that all reinforce each other. And, it’s thought that these probably started out as what we call geochemistry. So, it was chemistry in probably, like, deep sea vents that started to produce organic molecules with lots of hydrocarbons, basically, doing these sorts of complex interlocking chemical reactions that ultimately became a pattern that is stabilizing itself as an entire regime.

Whereas, patterns that just happened to arise that were unstable, well, we just don’t see them. It’s a tautological kind of dynamic where things that are good at persisting tend to persist. They are then the substrate for further evolution, and so on. And, eventually this geochemistry became biochemistry, and we had free-living organisms that are–they’re contained within a little membrane, a little bubble, or a cell wall that separates the inside from the outside. And then, they’re also, in a sense, trying to persist. They’re doing work. They’re doing thermodynamic work to keep that pattern, that dynamic pattern, going, right? Which is–again, that’s a completely new thing. There was nothing like that in the world before that.

Kevin Mitchell: [?]We can argue.

Kevin Mitchell: Yes. So, this is the journey that the book tries to trace, is this trajectory from as if agency to real agency.

And, when you were talking about real agency, you mentioned the word consciousness there. And, as the idea that there’s an extra element here. It’s not just that you’re acting for reasons. It’s that you’re aware of what those reasons are. You would be able to articulate them if somebody asked you why you did something. You might be able to interrogate them yourself and ask, ‘Is this a good reason for doing something?’ And so on.

And so, that’s the highest level that we know of, of agency, which has this level of metacognition where you’re aware of your own mental processes. But, the trajectory evolutionary was a long way to get there. And, it starts out with a much more simple kind of agency, which still is purposive. There’s still a purpose to it. They’re still acting for reasons, even if the reasons are kind of baked in by evolution.

So, even, like, a bacterium that has a tendency to move towards a food source, for example, is acting for a reason relative to its purpose, and its purpose is to persist. And, it’s not aware of that. There’s no awareness, right?

Kevin Mitchell: Exactly–

Kevin Mitchell: Exactly. But, the purpose is a very real thing. You can’t understand what’s going on without recourse to saying that the reason this thing is configured in the way that it is, that that’s a good policy to have that favors its persistence.

And so, I talked about things being this sort of dynamic stability–a living thing as just being a set of dynamic processes, a pattern that’s trying to persist through time. The problem is the world doesn’t cooperate. The world is changing. It’s this dynamic, hostile environment usually. And, what that means is that living organisms, it pays for them to be reactive or adaptive to, as conditions change.

Kevin Mitchell: Even better is to be proactive, exactly, to anticipate possible changes in the environment.

And so, what organisms have evolved–and this includes the very simplest organisms like bacteria–is a way of gathering information about what’s out in the world, and a way of linking that to a policy or a decision about what to do about it.

So, our little bacterium has a policy sort of biochemically wired into it that, when it detects food through this receptor that sits on the membrane of the cell, and it gives a little wiggle inside the cell, then the cell changes its direction of motion. The little bacteria, they swim around with kind of, like,an outboard motor.

So, what’s interesting there is that if you look at that system, you might say, ‘Okay, well, clearly the bacterium is literally being pushed around by its parts.’ You can see individual proteins inside it changing their biochemical confirmation or activity, and the eventual result is: Food out here, bacteria moves this way.

And, that’s basically similar to what I was talking about with the neuroscientists, where they say, ‘You only did X because this part of your brain was active, and it connects to your muscles in this way, and this was the action that was instituted as a result of that.’

However, I think what you can do is actually zoom out a little bit and say, ‘Well, wait a minute. Actually, if I give my bacterium a food source like that, it doesn’t always move towards the food. It depends what else is around. It’s actually integrating lots of things.’ I mean, when we do an experiment that’s the only thing we have in the experiment is there’s some food over here, and there’s a bacterium, then we get a pretty reliable result, even though it’s executed in a probabilistic fashion. But, if we put other things around, now that’s much more naturalistic. Now our bacterium is having to act like a wholistic agent. It’s having to say, ‘Wait, okay, there’s a food source over there, but maybe the pH of the solution is too low over there. That’ll kill me.’ Or there’s other kinds of things. Or maybe their metabolic state is actually fine: they’re not low on fuel, so they don’t need to go towards food. Maybe they’re dividing. Whatever it is.

So, when you take a more wholistic view, what you get, rather than this mechanistic, reductive thing being pushed around by its parts, is you get an organism that’s encountering the world in a proactive way. It’s not sitting there waiting for some stimulus. It’s always active, and it’s exploring the world. It accommodates to new information, it integrates a bunch of things, and then it kind of makes an all-things-considered judgment that’s optimal given multiple goals and multiple parameters that it’s trying to satisfy.

And, I mean, there’s a direct analogy to economy here in terms of bounded rationality, and trying to optimize over multiple conflicting goals and timescales, and so on.

Kevin Mitchell: Absolutely. Absolutely.

Kevin Mitchell: Exactly. Yeah.

So, living systems have had to evolve control systems that allow them to adapt to changing conditions, and that basically give them the power to act in the world as a wholistic entity, as an agent. And the reason is that the agent is acting for reasons that favor its persistence as a whole thing. It’s not like the individual proteins within a bacterium are not trying to persist. The whole pattern is trying to persist. Even if it’s only as if, like, once you get strong enough, there’s actually no difference from as if and really is. It’s–until you get to points where there’s now this meta-awareness of goals and reasons, and so on.

So, I think we can start with the simplest life forms. We’ve got purpose; we’ve got meaning, value; we’ve got informational kinds of causation that are at play that transcend just physics. When a receptor binds a food molecule like a sugar on the outside of the cell, there isn’t any transfer of energy or momentum that’s pushing anything around. There’s just an informational change. There’s a confirmation of the protein inside the cell that is a signal.

So, we have a kind of causation that, again, is not at play in any non-living part of the universe. This was something that came along with life.

Kevin Mitchell: Yeah, it does make no sense. And, I think you’re right to say that even the most ardent free-will skeptic, the most avowed free-will skeptic, doesn’t live like that.

And, the implication–it’s not just about whether we make decisions and should be held accountable for them. It’s about what it means to be a human being. Because if you think that actually you have no agency and autonomy at all, you are not a human being. You’re not a person. I mean, the definition of a person is someone who has enough autonomy to be afforded the dignity of being thought of with personhood. I mean, it just doesn’t mean anything to be a self.

Kevin Mitchell: Yeah, absolutely. But, it doesn’t mean anything to be a self if you don’t have some autonomous kind of control.

So, you touched on a really important aspect here, which is the idea that Keith could learn–or could be trained–not to look at the football game during a conversation or a dinner with his wife. Right? So, that is the next sort of stage in the evolutionary process, is: Once we go from unicellular organisms where they have to figure out what’s out in the world and what should I do about it, we get this transition in evolution to multicellular organisms; and then they had to reinvent this equipment for figuring what’s out in the world and what should I do about it, and the ability to actually move around in the world. So, they invented sensory organs–eyes and ears, and whiskers, and antennae, and so on. And they invented muscles and limbs, and things were moving around.

And, they invented a control system to hook those things up to each other, and to adjudicate over complex sets of possibilities. And, that is, of course, the nervous system.

And, the amazing thing with the nervous system is, evolutionarily you can wire in some preconfigured control policies that are just good things to do, like an escape response of a startled rabbit or something like that. It’s not thinking about anything. That’s wired in by evolution. It doesn’t require any learning whatsoever–

Kevin Mitchell: Exactly. Right? And, we have our own reflexes as well that are just good things to do.

But, nervous systems are amazing because they can learn. So, what that means is that it enables individual organisms over their lifetime to acquire new information. And, most of the information that they’re interested in is based on how things have turned out. They’ve been in a certain scenario, they’ve tried some action, and it either turned out well or it didn’t. And, if it turned out well, then that’s a piece of information that they can use–actually, either way, it’s a piece of information that they can use if they’re faced with the same scenario again. Or a similar scenario.

And, this is where we learn not just about the outcomes of actions, but we also learn about the nature of the world, the kinds of entities that are out in the world. We can categorize them as living or non-living, cats or dogs, human beings, males or females, things you can eat, things that might eat you, whatever it is. You develop these categories where, when you learn something, when you learn one thing about a member of a category, it doesn’t just apply to that one thing: it applies to everything else. If you’ve ever been bitten by a dog, you can be shy of all dogs in the future because they’re the type of thing that can bite you.

So, we can build up this complex kind of model of the world that’s populated with all of these kinds of entities, these categorical relations between them, a sort of[?] network of causal relations; and those causal relations particularly apply to what I can do in the world, and what things in the world can do to me. And then, that’s the array of information over which we’re operating when we’re making a decision and we’re trying to say, ‘What’s the best thing for me to do?’ And, you’re drawing on all that past experience.

And, you know, you mentioned habits, which: habits get a bad rap because there’s a sense in which having a habit is just reducing your freedom. And, in one sense, that can be true, but in another sense, it’s really good. Right? So, I’ve got lots of habits that I apply all the time where a habit is basically just like: I know in a certain situation this is a good thing to do. And, I don’t need to think about it actively now. I don’t need to expend cognitive resources and precious time because I’ve thought about this loads of times before. I know it’s a good thing to do. That’s just a super-super-efficient way to manage your control system and make the most out of the learning-from-experience that you’ve had.

And so, when you’re faced with a scenario in which that’s very, very familiar, there may be a habitual response that you engage in. And then, the question is, okay, for free will, what does that mean? How do you interpret that?

Kevin Mitchell: Yeah. I mean, thank you for reading that out because I think that for me was–when I got that concept, I felt like it was going to be useful. Because, that’s basically what learning is, right? We have learned from the past, and it extends into the future.

But, the other thing that this encapsulates is, again, a kind of a deeper metaphysical question of what it means to be a self. And, for in a physics kind of a frame of mind or even a neuroscience frame of mind, the self is just the physical organization of a thing right now at this moment in time. And, to me, that’s just physical stuff. That’s not the self. The self is the continuity of that pattern through time. And, for us to be a self means not just maintaining this pattern of physical processes, it means maintaining our psychological selves, our biographical information, all of the knowledge that we’ve learned, all of the habits that we’ve picked up, plus all of the projects that we’ve committed to that extend into the future. [More to come, 49:32]