I wish you'd given Sapolsky some more pushback, but you were pretty clear about where you're coming from, so it isn't the worst thing in the world. I'd like to question his premise. Why is it only possible for free will to exist through the breakage of fundamental laws of the universe? Why is that treated as a remotely reasonable expectation from his side? Why does something as fundamental, flexible, and pervasive as causation have to be usurped in order for free will to exist? Mitchell may make dubious claims about how fully he understands the implications of quantum mechanics, but Sapolsky expects us to become literal reality warping gods before he'll allow free will to exist.
To try and reset the frame, try looking at free will as a meter, rather than a switch. A 'Free-will-o-meter' actually makes more sense than a strict binary, in my experience. Amoeba have more free will than rocks, but less than humans. Humans have less free will than omnipotent beings of infinite creation, but that doesn't mean we don't have free will, and it certainly doesn't mean we can't gain more of it based on how we interact with the world. Unlike amoeba or rocks, humans are capable of comprehending our evolutionary history, our hormonal and neuronal state, our culture and upbringing, and incorporating those into our decision making. This means we have more control over the decisions we make than amoeba or rocks. We have more free will. But Sapolsky (if I'm reading him correctly from your description) doesn't view this as evidence in favor of free will, he instead views it as further evidence against it. He only sees more causal lines which force humans to make such and such a decision. That's the issue with his entire premise. He sees free will as a binary, and all of the necessary complexity which contributes towards an increase in the free-will-o-meter only serves to highlight how humans aren't at the absolute maximum. More free will leads to less free will, because a greater capacity to understand and shape the world around us necessarily leads to a greater capacity to understand the things we cannot shape. But an inability to shape certain facets of the universe does not invalidate free will. You cannot ignore all the things that we are able to shape which were previously impossible, and fixate on the things which currently seem to be impossible, and say that free will doesn't exist!
Okay, sorry, wow. It's getting late, I could probably go on, but I need to sleep. This was interesting, thank you for posting it. Hope there was something in my rant up there that was worthwhile, and my apologies if there isn't, I love rambling about this stuff.
Thanks for the great comment! Let me try to address your two main points.
-Why is it only possible for free will to exist through the breakage of fundamental laws of the universe? … Why does something as fundamental, flexible, and pervasive as causation have to be usurped in order for free will to exist?
The type of free will that Sapolsky is arguing against (and which is what I think most people mean when they say “free will”) is the type that construes human agents as independent causal forces in the physical universe, interfering somehow in the chain of causation going on around and within them. So, how is this possible without violating the laws of physics? Mitchell makes a valiant attempt to show how it could be possible, but to my mind he comes up short. Somebody’s got to explain it to me in a way that makes sense, otherwise I’m going to have to stick with Sapolsky’s explanation of a world without free will, which strikes me as perfectly sensible.
-To try and reset the frame, try looking at free will as a meter, rather than a switch.
I think this is a reasonable way to conceive free will, if it’s real. However, we’re in a situation where, in terms of a system's behavior, greater free will would be indistinguishable from greater information-processing complexity. Free will and complexity could look exactly the same from the outside. For this reason, I’m not sure the continuum approach helps much in making the case for free will.
Anyway, this is a very tricky topic and there are still lots of scientific unknowns. For now, I’m with Sapolsky, but I’m openminded. Hope you slept well!
Thanks, I got about as good a sleep as you'd expect, for someone who types up lengthy comments on the existence of free will in the dead of night.
There are a bunch of things I could go into, but I'll try to limit myself somewhat. For reference, I'm approaching this as a free will believer, but one who views the arguments from either side to be in deadlock, mostly due to missing/inaccessible information. Your essay seems to be applying a double standard, placing the burden of proof unduly and primarily on the free will side. Sapolsky doesn't understand all the precise mechanics of neuronal firing in the brain, or evolutionary history, or idea formation, expression and utilization. He assumes that all of these processes are deterministic, even though they haven't been fully explained by anyone, much less by him. Mitchell's response is to point to areas of the universe which, best we can tell, aren't deterministic, and say that those are the source of free will. At which point he is (I'd say rightfully) criticized for epistemic overconfidence. But, why isn't Sapolsky being accused of the same overconfidence? He is making an even more widespread claim about things which we really don't understand at the level of detail which would reveal causal flexibility. Yet, when he assumes it is all deterministic, that's fine. He apparently isn't being overconfident, despite doing what Mitchell has done multiplied by ten thousand.
Please let me know if that seems like an unfair characterization. If there's there some key difference between Mitchell and Sapolsky here that makes it reasonable to believe Sapolsky's claims, while still doubting Mitchell, I'd love to hear it.
Okay, that's a core issue with your essay I sort of skipped over last night. As for your response to my comment, first off, thank you! It's a dangerous game responding to rambly comments, but we appreciate the attention more than you'd think. If you poke at people's basic definition of free will, you'll generally find it to be nuanced in really informative ways. For example, a surprisingly large number of people, despite claiming to have free will in their day to day lives, will also claim to have no free will whatsoever, completely subservient to God, or the universe, or somesuch. That's why I find a sliding scale model of free will to be much more useful when discussing this stuff, even if the gut impulse for most folk is to call it a binary. A scale of free will more accurately captures how people independently assess their own free will; sometimes a lot, sometimes literally none. Depends on the circumstances. If you set those circumstances to macro scale, fundamental universal laws level, then you'll always come up with an answer of none, which seems misguided. The frame really matters here, so I try to give pushback to anyone that's trying to bolt it down.
That said, I haven't fleshed things out as much as I'd like yet. It's a good point that the way I've set things up overly favors information processing. That is certainly a necessary part of free will, but I've made it too central, I agree. Most of free will centers around what you do with the info, choices made in the midst of it, not just the ability to generate. Eh, I'll try putting more thought into it sometime, we'll see if anything interesting pops up.
Two points: First: One is not required to believe in causality as Sapolsky so defines it. In his version, a given state must always lead to one and only one next state, and so can be said to cause or determine it. But this might not be the case. Imagine that God duplicated the entire universe down to the last subatomic particle and its position. Completely identical. And then you wait a bit. If Sapolsky-type hard determinism is true, then the two universes will continue to be identical forever more. But if at many points along the way the next step is only one of several possibilities, which we can only know in a probabilistic fashion, then the two universes will begin to diverge. This isn't enough to get you to free will, but if it is possible to influence the possibilities then you can get somewhere close enough.
Second: The evolutionary argument in favour of the existence of Free Will is the whole existence of natural selection. Various modifications make an organism "more fit", i.e. survive and leave more descendants than those that don't have the modification. A great many of these modifications give better perceptions, and better cognitive abilities. Now, if all of this is in the service of making better decisions -- see the tiger! don't get eaten! -- this all makes perfect sense. But only if you actually could make the decision the other way. A world where every decision had to be what it was, is one where there is no need to develop complex organisms at all. Why would we have them? What's the point of a tiger?
Thanks for your comment, you make really good points. As for the first one, I think the fundamental question you address is whether Sapolsky’s deterministic anti-free will argument can be salvaged in the face of a nondeterministic universe (which is possible at least because of the probabilistic nature of quantum mechanics). I would say yes, it is, because even if the universe is fundamentally indeterminate, our behavior is still determined by the probabilistic events at the quantum or any other level. In other words, randomness in physics does not imply free will, and our behavior is still determined by processes outside our control.
Your second point brings up another fundamental question: how can we distinguish free will from complexity in a cognitive/information processing system? For a system that we assume has no free will (e.g. a robot), its ability to interact with the environment in increasingly complex and purposeful ways could look the same from the outside as the “intentional” behaviors of an organism. So how do we know that there’s any difference? If free will is real, then how do we know when to interpret complex behavior as “intentional” instead of just “complex”?
> FYI, for any free will debate aficionados out there, we’re not concerned with the “compatibilist” view here; it’s either libertarian free will or determinism, nothing in between.
I think the "in between" part is where the interesting thinking occurs. The problem I have with the way Sapolsky and others set up the problem is that you can either claim magic or *prior* cause and effects are all that matter. Those are your only two choices in a complex world where we know knowledge is fallible.
Determinists look at the *past* chain of cause and effect that leads up to a decision and claim that that causal chain caused our decision. From one angle, I agree with that, but I also think it erases the individual (however you construe 'the self') from the chain. In a causal chain, there are alternating antecedents and effects. At some point in the causal chain, we become the antecedent. That isn't magic, or an uncaused cause. Also, I think there's another angle to causal chains that's less linear and includes the individual. Mitchell speaks about it in his book. I think he calls it "circular causality?" I don't remember for sure. Anyway, the idea is, instead of just thinking about causality running in a straight line, we can also imagine it as a feedback loop. In a feedback loop, we aren't just the effect, but also the antecedent, not just an entity being pushed around. We can do some pushing ourselves.
Another way to put this: Our decisions are influenced by prior causal chains, but they are also part of the chain, and they help cause what happens next. The results of our decisions happen downstream; they do change our path. No, we aren't fully in control, but our actions have consequences for us and those around us. Agency and free will are a matter of degree.
If science settles that we live in a hard deterministic universe where everything is predetermined, that would be the end of the free will debate, but I don't think either Mitchell or Sapolsky believes that.
I agree that there’s a lot of insight to be found in the compatibilist perspective, but I don’t think it’s what most people intuitively mean by free will. I think Mitchell’s arguing for more of a libertarian perspective (although sometimes I did find it hard to tell).
What I don’t understand about the circular causality you describe is how becoming an antecedent in the causal chain gives us any real causal power. Even if the causal chain is loopy (and this makes sense to me), why can’t your decisions as an antecedent still be fully determined by prior causes just like everything else. Isn’t the loopyness just a layer of complexity, not some kind of interference from outside the causal chain? You say it’s not an uncaused cause, right? Then doesn’t it follow that it’s determined and not free?
Even if we don’t live in an indeterministic universe, someone still has to explain how the indeterminacy can be harnessed by a human agent. This is what Mitchell attempts to do, and as far as I can tell he comes up short.
Happy to be wrong here. If I do have free will I’d really like to be convinced of it.
> I agree that there’s a lot of insight to be found in the compatibilist perspective, but I don’t think it’s what most people intuitively mean by free will.
First, let me say I’m no expert on this topic. Pragmatically, the question of free will doesn’t matter to me. I think it’s near impossible to live our lives as if they’re determined. Sapolsky says as much himself. During a debate with Mitchell, the moderator asked Sapolsky about writing *Determined*, “Given that in the back of your mind, there was, ‘This is not my merit. I'm not really doing this; I'm just a product of low-level things without agency.’ Do you not feel a tension between those things, or do you think that tension can be argued away?” Sapolsky replied, “Oh, I don’t feel that tension at all because 95% of the time I’m a flaming hypocrite and cannot function as if it's really the case.”
But the question of free will is like an itch that keeps recurring in my life. From time to time, I read and think about it. It keeps recurring because to think that I’m just an automaton sucks the meaning out of my life.
You mention that the compatibilist perspective isn’t what people intuitively mean by free will. I agree, but from Sapolsky’s stance, he gives two choices: magic or causation, and I don’t know what to call the in-between, other than compatibilism. I’m with Mitchell when he says in *Free Agents*, “I do not think that either the strictly determinist or the compatibilist position is satisfactory. Both say free will is an illusion to some extent (p. x, preface).”
> I think Mitchell’s arguing for more of a libertarian perspective (although sometimes I did find it hard to tell).
Mitchell is definitely not a libertarian. He has this to say in *Free Agents*, “My aim is to show that, in thinking about these issues, we are not limited either to a simplistic physical determinism, in which all causes are located at the level of atoms or quantum fields, or to some kind of magical dualism, where we have to invoke immaterial forces to rescue our own agency (p. xi, preface).” As I quoted him above, he’s also not a determinist or a compatibilist. I don’t think he’s concerned with labels or doesn’t know how to label himself.
I’ve read around 60-70% of both *Free Agents* and *Determined*. I need to buckle down and finish them and do a lot more thinking about each.
Mitchell’s book argues how simple life forms evolved agency, and later how free will evolved in humans. This statement can sum up Sapolsky’s approach in *Determined*: “Show me a neuron (or brain) whose generation of a behavior is independent of the sum of its biological past, and for the purposes of this book, you’ve demonstrated free will (p. Chapter 1).” At least that’s how I would sum up the two books’ goals.
I need to revisit Mitchell’s book before I can discuss his “circular causality” and his overall argument. This is my best attempt, and I may be off, but I think he rejects the linear billiard ball kind of causality at the level of our brains that determinism is fond of. Our DNA, past experiences, everything that makes up our self does constrain us but doesn’t control us completely, because our brains are much more complex than a billiard ball. We have aspirations, we set goals, and we can modify our behavior. In essence, our brains have some control over shaping our future. I think Mitchell is saying it’s not all about simple causality and constraints, but also about a person’s ability to control their life. “Circular causality” exists over time. The effects of one billiard ball hitting another happen in a fraction of a second. Determinism only looks from the perspective of simple causality and uses it to explain all behavior, but ignores creativity, awareness, and intelligence as a means of control. It’s been months since I read his book. Give me a few weeks to reply about this, or maybe I’ll write a Substack on it.
Personally, I don’t think there will be a consensus on the topic of free will for a very long time. It will take determining whether our universe is predetermined or not, and if it isn’t, mapping the workings of our brain from the quantum level to the biological level, and mapping that to consciousness. Both are tall orders!
I have always believed in free will (it's the catholic in me) and find the concept of not having it a little scary. Although I do think Sapolsky's argument kind of has me convinced. Would be interested to hear what others think.
I haven't read Mitchell but he seems to have very interesting ideas so thank you very much for this post.
The view of humans as systems, agents with their own language is an interesting scientific framework for formalizing free will, although I'm not sure it is sufficient. It is undeniable that we are agents as we are actors in our own world and can influence it with actions, communication, etc. The same way a bacterium is active and react to its medium by producing and exporting proteins. But does the bacterium has free will? Or is it simply responding to external stress?
On the other hand if there is one view I can't agree of with it is Sapolsky's. There is no way that we can reliably predict one's behaviour according to the laws of physics so the argument against free will falls under the category of speculative to me. And uncertainty in quantum physics is a good example of why we may never be able to. But let us suppose that quantum physics is no longer uncertain, and that human beings are determiniscally driven and can now be realistically simulated by a computer. The most advanced artificial intelligence agents we have all have emergent properties, properties that were not planned as part of the code, such has the so-called hallucinations. What forbids our concept of free will, which is tangible to most of us, to be an emergent property of our humanity? I believe that must be the crux of Mitchell's arguments.
I agree that the scientific perspective can't answer questions about free will, but I would also say it can't answer questions about determinism or causation either. If it were really true that nothing could be otherwise, what sense could we make of causality? I haven't read either Sapolsky or Mitchell, but it sounds like they both seem to miss this point? Has either of them solved the problem of induction? What do they take causality to be?
Sapolsky seems not to understand the difference between correlation and causation. For instance:
"What was happening in your brain a split-second to seconds before you make the decision? E.g. activity of your brain cells that indicates an intended action, which can be detected up to 10 seconds prior to you realizing you’ve decided to do it."
But what WAS happening in my brain a split second before making a decision? Is it identical to the decision I was going to make? This decoupling of the brain state to an "I" that later becomes conscious of making a decision (which is really a brain state) is precisely what he should avoid talking about, I would think, since it smacks of a ghost in the machine. Curious if he mentions LIbet's interpretation of vetoing action potentials.
The rest of his list is typical of reducing all possible explanation to mechanistic physical cause. Well, I don't agree with that understanding of causation, so his argument is weak from my point of view. He seems to be assuming determinism to prove free will doesn't exist, which is begging the question. But granting him his premise, if all of the sciences added together give us a completely deterministic causal picture, then we ought to be able to predict people's thoughts and actions based on current laws and facts. At least sometimes. This is what he seems to be claiming we should be able to do, given that he won't allow for any indeterminacy. But maybe I've misconstrued his argument (which wouldn't be surprising given I haven't read him)?
—But even if Mitchell happens to be right about quantum indeterminacy, we’re still left with the problem of how people harness it to make free choices after it supposedly bubbles up to the level of neuronal activity.
Very interesting comment! I've been curious about how a deeper philosophical view of causation would affect this free will debate (I've got Judea Pearl's The Book of Why on my shelf but I've been afraid to open it). I think you're right that Sapolsky doesn't dig deep into the correlation/causation conundrum. But, to be fair, my summary only skims the surface of his argument, so it would be good to engage with his book directly on this. On the topic of predicting thought/behavior he's quite clear in the book: because of chaos/complexity/nonlinearity, it's possible for behavior to be fully determined and at the same time not fully predictable (at least with reasonable amounts of information).
Thanks! Yeah, causality is a huge topic in philosophy, though it's very easy to get hand wave-y about it. On prediction, I'd have to read what he says, of course, but it sounds like he's claiming epistemic uncertainty to explain why future thoughts can't be predicted, yet arguing for a Laplacian demon-style determinism on the same basis (that each discipline taken alone may be gappy, allowing for free will, but all of them together somehow close those gaps). I've heard similar arguments before, but I'm not sure they work.
BTW, I love the book battle concept of your Substack. What a great way to explore these questions!
So... an important insight from Mitchell might be this: "Mitchell’s secret weapon, his answer to Sapolsky’s challenge, is a human self that’s capable of conscious, rational control."
Perhaps I'm going to be giving the game away to Sapolsky by simply putting a gloss on determinism, but bear with me. What if our satisfaction hinges on what we identify with? We want to identify some outside force that flips switches almost "at random" as if capriciousness was some virtue.
But I don't think Sapolsky (and correct me if I'm wrong, I've only read his previous tome) would deny an effective element in conscious thought. Rather he'd deny that thought being uncaused. Ie, there's no 'you' standing above the causal chain.
But let's draw a circle around this opaque sequence. Regardless of whether the brain is *in principle* determined, I'd assert it's practically unknowable. That is, we will probably never be able to perfectly simulate someone, even if we could we couldn't gather the all the inputs and weigh them correctly, etc. Those ten factors Sapolsky identifies, and whatever else.
So there is, in humans and to a lesser extent other animals, and, when we're feeling generous, perhaps artificial minds, this opaque decision making process that takes inputs and outputs actions. We want to look for the operator of that system and call it 'me'. But could we be comfortable drawing a circle around the entire process and calling *that* the self? A 'soul' then, is a pattern of behavior that makes rational decisions based on beliefs about the world, goals, and values, and acts on the decisions. It does not create itself but it does contribute to its ongoing character by various feedback mechanisms. Every soul is unique and an integral and, from an outside perspective, unpredictable element of the affecting of reality.
Sapolsky would say this is determinism, probably. But it seems sufficient to be satisfied with our role.
"Can you take pride in those achievements as evidence of your own intrinsic effort and grit and self-ordained power to make decisions and act upon the world? Or should you look upon your achievements with humble gratitude and admit that it all, ultimately, boiled down to luck?"
This is tangential, but I wonder how modern, or WEIRD to use an acronym you're familiar with, this question is? It seems to me a lot of people would be perfectly happy to be thought great because of their inherent superiority. But we want to be great in spite of our inner stuff, to have some singular core that can claim credit for steering the ship of our lives against the temptations of time and flesh.
Or maybe that's nothing new, my knowledge of Greek or more foreign philosophy is poor.
I wish you'd given Sapolsky some more pushback, but you were pretty clear about where you're coming from, so it isn't the worst thing in the world. I'd like to question his premise. Why is it only possible for free will to exist through the breakage of fundamental laws of the universe? Why is that treated as a remotely reasonable expectation from his side? Why does something as fundamental, flexible, and pervasive as causation have to be usurped in order for free will to exist? Mitchell may make dubious claims about how fully he understands the implications of quantum mechanics, but Sapolsky expects us to become literal reality warping gods before he'll allow free will to exist.
To try and reset the frame, try looking at free will as a meter, rather than a switch. A 'Free-will-o-meter' actually makes more sense than a strict binary, in my experience. Amoeba have more free will than rocks, but less than humans. Humans have less free will than omnipotent beings of infinite creation, but that doesn't mean we don't have free will, and it certainly doesn't mean we can't gain more of it based on how we interact with the world. Unlike amoeba or rocks, humans are capable of comprehending our evolutionary history, our hormonal and neuronal state, our culture and upbringing, and incorporating those into our decision making. This means we have more control over the decisions we make than amoeba or rocks. We have more free will. But Sapolsky (if I'm reading him correctly from your description) doesn't view this as evidence in favor of free will, he instead views it as further evidence against it. He only sees more causal lines which force humans to make such and such a decision. That's the issue with his entire premise. He sees free will as a binary, and all of the necessary complexity which contributes towards an increase in the free-will-o-meter only serves to highlight how humans aren't at the absolute maximum. More free will leads to less free will, because a greater capacity to understand and shape the world around us necessarily leads to a greater capacity to understand the things we cannot shape. But an inability to shape certain facets of the universe does not invalidate free will. You cannot ignore all the things that we are able to shape which were previously impossible, and fixate on the things which currently seem to be impossible, and say that free will doesn't exist!
Okay, sorry, wow. It's getting late, I could probably go on, but I need to sleep. This was interesting, thank you for posting it. Hope there was something in my rant up there that was worthwhile, and my apologies if there isn't, I love rambling about this stuff.
Thanks for the great comment! Let me try to address your two main points.
-Why is it only possible for free will to exist through the breakage of fundamental laws of the universe? … Why does something as fundamental, flexible, and pervasive as causation have to be usurped in order for free will to exist?
The type of free will that Sapolsky is arguing against (and which is what I think most people mean when they say “free will”) is the type that construes human agents as independent causal forces in the physical universe, interfering somehow in the chain of causation going on around and within them. So, how is this possible without violating the laws of physics? Mitchell makes a valiant attempt to show how it could be possible, but to my mind he comes up short. Somebody’s got to explain it to me in a way that makes sense, otherwise I’m going to have to stick with Sapolsky’s explanation of a world without free will, which strikes me as perfectly sensible.
-To try and reset the frame, try looking at free will as a meter, rather than a switch.
I think this is a reasonable way to conceive free will, if it’s real. However, we’re in a situation where, in terms of a system's behavior, greater free will would be indistinguishable from greater information-processing complexity. Free will and complexity could look exactly the same from the outside. For this reason, I’m not sure the continuum approach helps much in making the case for free will.
Anyway, this is a very tricky topic and there are still lots of scientific unknowns. For now, I’m with Sapolsky, but I’m openminded. Hope you slept well!
Thanks, I got about as good a sleep as you'd expect, for someone who types up lengthy comments on the existence of free will in the dead of night.
There are a bunch of things I could go into, but I'll try to limit myself somewhat. For reference, I'm approaching this as a free will believer, but one who views the arguments from either side to be in deadlock, mostly due to missing/inaccessible information. Your essay seems to be applying a double standard, placing the burden of proof unduly and primarily on the free will side. Sapolsky doesn't understand all the precise mechanics of neuronal firing in the brain, or evolutionary history, or idea formation, expression and utilization. He assumes that all of these processes are deterministic, even though they haven't been fully explained by anyone, much less by him. Mitchell's response is to point to areas of the universe which, best we can tell, aren't deterministic, and say that those are the source of free will. At which point he is (I'd say rightfully) criticized for epistemic overconfidence. But, why isn't Sapolsky being accused of the same overconfidence? He is making an even more widespread claim about things which we really don't understand at the level of detail which would reveal causal flexibility. Yet, when he assumes it is all deterministic, that's fine. He apparently isn't being overconfident, despite doing what Mitchell has done multiplied by ten thousand.
Please let me know if that seems like an unfair characterization. If there's there some key difference between Mitchell and Sapolsky here that makes it reasonable to believe Sapolsky's claims, while still doubting Mitchell, I'd love to hear it.
Okay, that's a core issue with your essay I sort of skipped over last night. As for your response to my comment, first off, thank you! It's a dangerous game responding to rambly comments, but we appreciate the attention more than you'd think. If you poke at people's basic definition of free will, you'll generally find it to be nuanced in really informative ways. For example, a surprisingly large number of people, despite claiming to have free will in their day to day lives, will also claim to have no free will whatsoever, completely subservient to God, or the universe, or somesuch. That's why I find a sliding scale model of free will to be much more useful when discussing this stuff, even if the gut impulse for most folk is to call it a binary. A scale of free will more accurately captures how people independently assess their own free will; sometimes a lot, sometimes literally none. Depends on the circumstances. If you set those circumstances to macro scale, fundamental universal laws level, then you'll always come up with an answer of none, which seems misguided. The frame really matters here, so I try to give pushback to anyone that's trying to bolt it down.
That said, I haven't fleshed things out as much as I'd like yet. It's a good point that the way I've set things up overly favors information processing. That is certainly a necessary part of free will, but I've made it too central, I agree. Most of free will centers around what you do with the info, choices made in the midst of it, not just the ability to generate. Eh, I'll try putting more thought into it sometime, we'll see if anything interesting pops up.
I've read Sapolsky. I haven't read Mitchell.
Two points: First: One is not required to believe in causality as Sapolsky so defines it. In his version, a given state must always lead to one and only one next state, and so can be said to cause or determine it. But this might not be the case. Imagine that God duplicated the entire universe down to the last subatomic particle and its position. Completely identical. And then you wait a bit. If Sapolsky-type hard determinism is true, then the two universes will continue to be identical forever more. But if at many points along the way the next step is only one of several possibilities, which we can only know in a probabilistic fashion, then the two universes will begin to diverge. This isn't enough to get you to free will, but if it is possible to influence the possibilities then you can get somewhere close enough.
Second: The evolutionary argument in favour of the existence of Free Will is the whole existence of natural selection. Various modifications make an organism "more fit", i.e. survive and leave more descendants than those that don't have the modification. A great many of these modifications give better perceptions, and better cognitive abilities. Now, if all of this is in the service of making better decisions -- see the tiger! don't get eaten! -- this all makes perfect sense. But only if you actually could make the decision the other way. A world where every decision had to be what it was, is one where there is no need to develop complex organisms at all. Why would we have them? What's the point of a tiger?
Thanks for your comment, you make really good points. As for the first one, I think the fundamental question you address is whether Sapolsky’s deterministic anti-free will argument can be salvaged in the face of a nondeterministic universe (which is possible at least because of the probabilistic nature of quantum mechanics). I would say yes, it is, because even if the universe is fundamentally indeterminate, our behavior is still determined by the probabilistic events at the quantum or any other level. In other words, randomness in physics does not imply free will, and our behavior is still determined by processes outside our control.
Your second point brings up another fundamental question: how can we distinguish free will from complexity in a cognitive/information processing system? For a system that we assume has no free will (e.g. a robot), its ability to interact with the environment in increasingly complex and purposeful ways could look the same from the outside as the “intentional” behaviors of an organism. So how do we know that there’s any difference? If free will is real, then how do we know when to interpret complex behavior as “intentional” instead of just “complex”?
> FYI, for any free will debate aficionados out there, we’re not concerned with the “compatibilist” view here; it’s either libertarian free will or determinism, nothing in between.
I think the "in between" part is where the interesting thinking occurs. The problem I have with the way Sapolsky and others set up the problem is that you can either claim magic or *prior* cause and effects are all that matter. Those are your only two choices in a complex world where we know knowledge is fallible.
Determinists look at the *past* chain of cause and effect that leads up to a decision and claim that that causal chain caused our decision. From one angle, I agree with that, but I also think it erases the individual (however you construe 'the self') from the chain. In a causal chain, there are alternating antecedents and effects. At some point in the causal chain, we become the antecedent. That isn't magic, or an uncaused cause. Also, I think there's another angle to causal chains that's less linear and includes the individual. Mitchell speaks about it in his book. I think he calls it "circular causality?" I don't remember for sure. Anyway, the idea is, instead of just thinking about causality running in a straight line, we can also imagine it as a feedback loop. In a feedback loop, we aren't just the effect, but also the antecedent, not just an entity being pushed around. We can do some pushing ourselves.
Another way to put this: Our decisions are influenced by prior causal chains, but they are also part of the chain, and they help cause what happens next. The results of our decisions happen downstream; they do change our path. No, we aren't fully in control, but our actions have consequences for us and those around us. Agency and free will are a matter of degree.
If science settles that we live in a hard deterministic universe where everything is predetermined, that would be the end of the free will debate, but I don't think either Mitchell or Sapolsky believes that.
Thanks for the great comment!
I agree that there’s a lot of insight to be found in the compatibilist perspective, but I don’t think it’s what most people intuitively mean by free will. I think Mitchell’s arguing for more of a libertarian perspective (although sometimes I did find it hard to tell).
What I don’t understand about the circular causality you describe is how becoming an antecedent in the causal chain gives us any real causal power. Even if the causal chain is loopy (and this makes sense to me), why can’t your decisions as an antecedent still be fully determined by prior causes just like everything else. Isn’t the loopyness just a layer of complexity, not some kind of interference from outside the causal chain? You say it’s not an uncaused cause, right? Then doesn’t it follow that it’s determined and not free?
Even if we don’t live in an indeterministic universe, someone still has to explain how the indeterminacy can be harnessed by a human agent. This is what Mitchell attempts to do, and as far as I can tell he comes up short.
Happy to be wrong here. If I do have free will I’d really like to be convinced of it.
> I agree that there’s a lot of insight to be found in the compatibilist perspective, but I don’t think it’s what most people intuitively mean by free will.
First, let me say I’m no expert on this topic. Pragmatically, the question of free will doesn’t matter to me. I think it’s near impossible to live our lives as if they’re determined. Sapolsky says as much himself. During a debate with Mitchell, the moderator asked Sapolsky about writing *Determined*, “Given that in the back of your mind, there was, ‘This is not my merit. I'm not really doing this; I'm just a product of low-level things without agency.’ Do you not feel a tension between those things, or do you think that tension can be argued away?” Sapolsky replied, “Oh, I don’t feel that tension at all because 95% of the time I’m a flaming hypocrite and cannot function as if it's really the case.”
But the question of free will is like an itch that keeps recurring in my life. From time to time, I read and think about it. It keeps recurring because to think that I’m just an automaton sucks the meaning out of my life.
You mention that the compatibilist perspective isn’t what people intuitively mean by free will. I agree, but from Sapolsky’s stance, he gives two choices: magic or causation, and I don’t know what to call the in-between, other than compatibilism. I’m with Mitchell when he says in *Free Agents*, “I do not think that either the strictly determinist or the compatibilist position is satisfactory. Both say free will is an illusion to some extent (p. x, preface).”
> I think Mitchell’s arguing for more of a libertarian perspective (although sometimes I did find it hard to tell).
Mitchell is definitely not a libertarian. He has this to say in *Free Agents*, “My aim is to show that, in thinking about these issues, we are not limited either to a simplistic physical determinism, in which all causes are located at the level of atoms or quantum fields, or to some kind of magical dualism, where we have to invoke immaterial forces to rescue our own agency (p. xi, preface).” As I quoted him above, he’s also not a determinist or a compatibilist. I don’t think he’s concerned with labels or doesn’t know how to label himself.
I’ve read around 60-70% of both *Free Agents* and *Determined*. I need to buckle down and finish them and do a lot more thinking about each.
Mitchell’s book argues how simple life forms evolved agency, and later how free will evolved in humans. This statement can sum up Sapolsky’s approach in *Determined*: “Show me a neuron (or brain) whose generation of a behavior is independent of the sum of its biological past, and for the purposes of this book, you’ve demonstrated free will (p. Chapter 1).” At least that’s how I would sum up the two books’ goals.
I need to revisit Mitchell’s book before I can discuss his “circular causality” and his overall argument. This is my best attempt, and I may be off, but I think he rejects the linear billiard ball kind of causality at the level of our brains that determinism is fond of. Our DNA, past experiences, everything that makes up our self does constrain us but doesn’t control us completely, because our brains are much more complex than a billiard ball. We have aspirations, we set goals, and we can modify our behavior. In essence, our brains have some control over shaping our future. I think Mitchell is saying it’s not all about simple causality and constraints, but also about a person’s ability to control their life. “Circular causality” exists over time. The effects of one billiard ball hitting another happen in a fraction of a second. Determinism only looks from the perspective of simple causality and uses it to explain all behavior, but ignores creativity, awareness, and intelligence as a means of control. It’s been months since I read his book. Give me a few weeks to reply about this, or maybe I’ll write a Substack on it.
Personally, I don’t think there will be a consensus on the topic of free will for a very long time. It will take determining whether our universe is predetermined or not, and if it isn’t, mapping the workings of our brain from the quantum level to the biological level, and mapping that to consciousness. Both are tall orders!
I have always believed in free will (it's the catholic in me) and find the concept of not having it a little scary. Although I do think Sapolsky's argument kind of has me convinced. Would be interested to hear what others think.
I haven't read Mitchell but he seems to have very interesting ideas so thank you very much for this post.
The view of humans as systems, agents with their own language is an interesting scientific framework for formalizing free will, although I'm not sure it is sufficient. It is undeniable that we are agents as we are actors in our own world and can influence it with actions, communication, etc. The same way a bacterium is active and react to its medium by producing and exporting proteins. But does the bacterium has free will? Or is it simply responding to external stress?
On the other hand if there is one view I can't agree of with it is Sapolsky's. There is no way that we can reliably predict one's behaviour according to the laws of physics so the argument against free will falls under the category of speculative to me. And uncertainty in quantum physics is a good example of why we may never be able to. But let us suppose that quantum physics is no longer uncertain, and that human beings are determiniscally driven and can now be realistically simulated by a computer. The most advanced artificial intelligence agents we have all have emergent properties, properties that were not planned as part of the code, such has the so-called hallucinations. What forbids our concept of free will, which is tangible to most of us, to be an emergent property of our humanity? I believe that must be the crux of Mitchell's arguments.
I agree that the scientific perspective can't answer questions about free will, but I would also say it can't answer questions about determinism or causation either. If it were really true that nothing could be otherwise, what sense could we make of causality? I haven't read either Sapolsky or Mitchell, but it sounds like they both seem to miss this point? Has either of them solved the problem of induction? What do they take causality to be?
Sapolsky seems not to understand the difference between correlation and causation. For instance:
"What was happening in your brain a split-second to seconds before you make the decision? E.g. activity of your brain cells that indicates an intended action, which can be detected up to 10 seconds prior to you realizing you’ve decided to do it."
But what WAS happening in my brain a split second before making a decision? Is it identical to the decision I was going to make? This decoupling of the brain state to an "I" that later becomes conscious of making a decision (which is really a brain state) is precisely what he should avoid talking about, I would think, since it smacks of a ghost in the machine. Curious if he mentions LIbet's interpretation of vetoing action potentials.
The rest of his list is typical of reducing all possible explanation to mechanistic physical cause. Well, I don't agree with that understanding of causation, so his argument is weak from my point of view. He seems to be assuming determinism to prove free will doesn't exist, which is begging the question. But granting him his premise, if all of the sciences added together give us a completely deterministic causal picture, then we ought to be able to predict people's thoughts and actions based on current laws and facts. At least sometimes. This is what he seems to be claiming we should be able to do, given that he won't allow for any indeterminacy. But maybe I've misconstrued his argument (which wouldn't be surprising given I haven't read him)?
—But even if Mitchell happens to be right about quantum indeterminacy, we’re still left with the problem of how people harness it to make free choices after it supposedly bubbles up to the level of neuronal activity.
Yeah, agreed.
Very interesting comment! I've been curious about how a deeper philosophical view of causation would affect this free will debate (I've got Judea Pearl's The Book of Why on my shelf but I've been afraid to open it). I think you're right that Sapolsky doesn't dig deep into the correlation/causation conundrum. But, to be fair, my summary only skims the surface of his argument, so it would be good to engage with his book directly on this. On the topic of predicting thought/behavior he's quite clear in the book: because of chaos/complexity/nonlinearity, it's possible for behavior to be fully determined and at the same time not fully predictable (at least with reasonable amounts of information).
Thanks! Yeah, causality is a huge topic in philosophy, though it's very easy to get hand wave-y about it. On prediction, I'd have to read what he says, of course, but it sounds like he's claiming epistemic uncertainty to explain why future thoughts can't be predicted, yet arguing for a Laplacian demon-style determinism on the same basis (that each discipline taken alone may be gappy, allowing for free will, but all of them together somehow close those gaps). I've heard similar arguments before, but I'm not sure they work.
BTW, I love the book battle concept of your Substack. What a great way to explore these questions!
So... an important insight from Mitchell might be this: "Mitchell’s secret weapon, his answer to Sapolsky’s challenge, is a human self that’s capable of conscious, rational control."
Perhaps I'm going to be giving the game away to Sapolsky by simply putting a gloss on determinism, but bear with me. What if our satisfaction hinges on what we identify with? We want to identify some outside force that flips switches almost "at random" as if capriciousness was some virtue.
But I don't think Sapolsky (and correct me if I'm wrong, I've only read his previous tome) would deny an effective element in conscious thought. Rather he'd deny that thought being uncaused. Ie, there's no 'you' standing above the causal chain.
But let's draw a circle around this opaque sequence. Regardless of whether the brain is *in principle* determined, I'd assert it's practically unknowable. That is, we will probably never be able to perfectly simulate someone, even if we could we couldn't gather the all the inputs and weigh them correctly, etc. Those ten factors Sapolsky identifies, and whatever else.
So there is, in humans and to a lesser extent other animals, and, when we're feeling generous, perhaps artificial minds, this opaque decision making process that takes inputs and outputs actions. We want to look for the operator of that system and call it 'me'. But could we be comfortable drawing a circle around the entire process and calling *that* the self? A 'soul' then, is a pattern of behavior that makes rational decisions based on beliefs about the world, goals, and values, and acts on the decisions. It does not create itself but it does contribute to its ongoing character by various feedback mechanisms. Every soul is unique and an integral and, from an outside perspective, unpredictable element of the affecting of reality.
Sapolsky would say this is determinism, probably. But it seems sufficient to be satisfied with our role.
"Can you take pride in those achievements as evidence of your own intrinsic effort and grit and self-ordained power to make decisions and act upon the world? Or should you look upon your achievements with humble gratitude and admit that it all, ultimately, boiled down to luck?"
This is tangential, but I wonder how modern, or WEIRD to use an acronym you're familiar with, this question is? It seems to me a lot of people would be perfectly happy to be thought great because of their inherent superiority. But we want to be great in spite of our inner stuff, to have some singular core that can claim credit for steering the ship of our lives against the temptations of time and flesh.
Or maybe that's nothing new, my knowledge of Greek or more foreign philosophy is poor.
Yes, I think most of us probably have that impulse, at least unconsciously