“The physics tells us that approximately 900,000 of the duplicates will water Susan and approximately 100,000 of them will not. If we ran the experiment many times, each time creating a million more duplicates and waiting for them to decide, the physics tells us we would get roughly the same frequencies each time. But if what happens is totally up to each duplicate – in the radical incompatibilist sense – then there ought to be no such predictable frequency. The number that do and don’t water the plant should change each time, as the radically free choices of each individual varies.”
The ‘approximately’ and ‘roughly’ give you all the flexibility you need for the numbers to not be exactly the same each time you do the expreiment.
Also, you can’t exactly duplicate the same situation more than once.
Even radical free will allows for predictability and limits to outcomes. If there is no way to tell what each iteration will do then you have room for free will.
If quantum mechanics is correct, you can prove the universe is not deterministic. Here's how. Get a little lump of radium and a Geiger counter. If your radium is just the right size, then the probability that it makes the Geiger counter beep in any one minute period will be 50%. And quantum mechanics says that is completely and truly random. Over a large number of observations, the 50% holds, but in any single case, whether it happens in one minute is strictly undetermined, strictly random. So set up the radium and Geiger counter and decide that you will eat a muffin if the counter beeps within a minute or less, but otherwise you will put the muffin in the fridge. Your choice to eat is delegated to your radium/G counter setup, so you are not making the choice. And yet according to quantum mechanics, whether you eat or not is strictly random. Hence, the universe is not deterministic.
If you carefully inspect your conscious choices in normal life (no radium lumps on hand), you can never catch yourself making a free choice. At least, I can't. If I act according to a desire, I didn't choose that desire, I just find it in me. Even if I created a habit of having a certain desire by hard work, what motivated me to do that hard work? It was another desire that, inspected carefully, I didn't choose (e.g. preferring an orderly life over the disorder of frequently giving in to a short term desire to drink rum).
Another example: if I am persuaded that trying to always act in love is the best way to help steer the universe toward maximal flourishing for sentient beings, what made me like the idea of maximal flourishing? Another desire, and the knowledge from my experience that I don't like to suffer and it appears that other creatures don't like to either. But I didn't choose to have a type of body that can suffer physical pain. I just found myself in one. This can go on endlessly, and I have never been able to see that I have actually make a truly free choice.
If panexperientialism or panpsychism is true, maybe smaller constituents of myself, my organs, cells, neurons, etc. make truly free choices and flush them up into my consciousness in the form of desires. But again, I can't detect that process so I, my conscious mind, does not seem to be able to make truly free choices as far as I can definitively identify.
So can it be that the universe is not determined, yet humans don't have the ability to make truly free choices?
I'm on board with a lot of this, assuming for the sake of argument an incompatibilist position.
But I don't think it makes sense for *The* Laws of Physics to be ceteris paribus laws. The laws of physics as currently understood could be ceteris paribus laws, but, to me, *The* Laws of Physics are the rules by which the world works. Even if there is a patchwork of different effective laws for different domains, to me, *the* laws of physics describe the overall patchwork, including rules to decide which rules govern which domains.
So, if there is strongly emergent libertarian free will, such that the laws of physics as currently understood are just ceteris paribus laws, then I think this means that *The* Laws of Physics will have to describe how free will works, and then *The* Laws of Physics are not ceteris paribus laws.
The problem is that people who want libertarian free will don't usually seem to want it to be describable in a lawlike way. That would take away the magic and reduce it again to the kind of mechanistic account they're trying to avoid. But either there are rules by which it works or there aren't. If there are, then it's ultimately going to be too mechanistic to satisfy libertarians. If there aren't, then we're handwaving and giving up any hope of understanding it. That's unscientific and against the spirit of methodological naturalism.
I also think it's incoherent. I don't think there can be any kind of fundamental process at play which cannot be described precisely in principle, and a precise description is all you need for a mathematical law.
I think that the demand that any kind of fundamental fact about reality is described by a mathematical law is a very, very strong claim and will get into problems not only with free will but also with a lot of conceptual thought (including mathematical thought). There is no mathematical law that bachelors are unmarried, and there is no mathematical law that mechanistically describes the conceptual steps that a usual mathematical proof takes. Or did I misunderstand the demand?
I think you misunderstood the demand, which is not necessarily your fault!
I think that bachelors being unmarried is neither a fundamental fact about reality (your phrase) nor a fundamental process at play (mine). It's not fundamental because we don't need to make changes to the fundamental laws of physics to accommodate bachelors being unmarried. Having higher-level rules that say that bachelors are unmarried doesn't mean that the laws we do have are ceteris paribus laws. There is no conflict between the laws of physics and rules about bachelors being unmarried. So all our observations of unmarried bachelors etc are consistent with laws of physics which are describable mathematically.
Whether the steps taken in a mathematical proof can be reduced to (at least computable) physics is controversial. Roger Penrose would interpret Godel's incompleteness theorems as showing that that they cannot. Most other experts (I believe) disagree with him on this. If Penrose is wrong, then we don't need special physics to account for how mathematicians (or LLMs, now) produce mathematical proofs. There is no fundamental process at play which cannot be given a standard mathematical description.
Processes at higher levels of description (involving concepts, truth, insight and so on) are not fundamental, and need not be describable mathematically, because these concepts are loose, vague and ambiguous, and are convenient ways of generalising and abstracting useful commonallities between distinct physical processes which are each describable as arising out of more fundamental processes which are describable mathematically.
Thanks! I still feel like the claim that any kind of fundamental process needs to be understood in terms of mathematics is a strong limitation on what kinds of truth you are admitting. Nearly all philosophy papers make arguments and claims, but very few philosophy papers give mathematical laws to mechanize their arguments. Would you say that this rules out the arguments in almost all philosophy papers as describing a fundamental process of reality?
It doesn't necessarily need to be understood in terms of mathematics. I'm not saying that art is a waste of time, and we need to concern ourselves exclusively with writing theorems.
My claim is that it is incoherent to posit causal forces at play in the world which have no underlying mechanisms, unless those causal forces are already mechanistic (describable mathematically). But that's not to say that we can't talk about things without understanding the underlying mechanisms. I just don't think it's coherent to deny that such mechanisms exist.
Ah, right, now I get your comment - you were only talking about causation and whether it is mechanistically described or not, and I went all overboard and thought you were talking about all of reality. For causation I understand your point much better.
I still think it is wrong, though, because I think that sentences like "the sight of a loved one made me happy" describe a process of reality that I have (imperfect but good) epistemic access to without needing a mathematical description of it, and if there ever will be a good mathematical description for that, perhaps in terms of a reductive explanation of neuro-physics, that is great but that is not a prerequisite for the sentence to be true, nor do I believe that such an explanation is forthcoming with certainty.
But I agree that non-mechanistic/mathematical causation happening in the case of a free choice is a bullet that you need to bite if you want proper free will. And I also agree that we need more understanding how causation works in a free choice, in order to be less handwavy. But I think such better understanding is possible in philosophical terms without needing recourse to mathematics.
I have no problem with sentences like "the sight of a loved one made me happy" describing a process of reality that you have (imperfect but good) epistemic access to without having a mathematical description of it.
Say sentences like this all day and I'll mostly assume you are saying true things.
But say "Maybe there is no underlying mechanism for how it is that the sight of a loved one makes me happy" and I'll think you are not making sense.
It's bad enough for agency that incompatibilists want to get rid of determinism, but if they get rid of probability as well, we would be completely disabled. If I really, really want to do A and not do B, then hitched to physical determinism, I will reliably do A. If determinism is false, then I can at least say that I will probably do A. But if I can't even say that I will probably do A, what does this mean? That repeating an A-loving and B-hating situation a million times I might do A anything between 0 and a million times?
We do have evidence of top down causation in physics. We have the Cambrian Explosion. We have the Genesis event. We have the point of human consciousness. We have the beginning of the universe.
Philip, I think your critique of Mitchell holds, but you then make the same mistake one level up. Both of you look for freedom in the wrong place.
Mitchell looks for it in the room that quantum indeterminism supposedly opens; you show correctly that this room does not exist, because the objective probability of the Born rule constrains just as much as a deterministic law. You then relocate freedom to a layer of "higher-level causal powers" that complement physics through ceteris paribus laws. But this is simply a displacement of the same move: you are still looking for a place inside or beside physical causation where freedom can be inserted. The search itself is the mistake. Free will is not a kind of causation that competes with physical causation, escapes it, or is added to it from outside.
Your duplicates experiment shows this. You reduce the decision to water Susan to a position of particles over which the Born rule fixes a frequency, and conclude that the recurrence of that frequency rules out freedom. But that configuration has meaning only in the physical regime. Free will is not a position of particles; it is self-modeling - the level at which a system produces a model of itself and acts from it. Asking how many of a million identical duplicates will water the plant simply asks how many identical configurations yield identical outcomes. That stays closed within the physical plane and says nothing about self-modeling, because it has already cancelled that level in the very framing.
Of course free will obeys physics - but only on the physical plane, the way we obey gravity. I cannot defy gravity without expending extraordinary energy; I defy it only through technology that overcomes it on its own terms. This obedience does not touch the level where free will operates, because it does not operate as a physical force competing with gravity, but as self-modeling that occurs through the physical configuration once complexity crosses a threshold - not against it, not beside it.
And free will is not exclusively human. An animal that decides how and when to hunt its prey, and that fights for its young, acts on the basis of a model of itself; it has an "I", feels emotion, sustains social life. What separates us from it is not the presence of free will but its extent, and the extent comes from the capacity to project the imagination - to model possible, counterfactual, not-yet-existing states. That is an increase of complexity, not an added substance.
The only constraint on my will is rational and arises from ethics. Ethics is the natural form of rationality; the two are interdependent. I understand what is good and what is bad because I grasp the consequences of my actions, and this arises from the same capacity that widens freedom: the projection of the imagination. To model an action before I carry it out and to see that it undermines the conditions that keep me or another in existence - this is what makes the ethical limit internal to rationality, not a norm imposed from outside.
So the determinism-indeterminism debate as a source of freedom seems to me misplaced at the root. Both sides look for freedom in the regime where physical causation happens. It is not there, and it never was.
Can someone explain to me how indeterminacy opens the door to agency though? It seems at best it allows for an open, non-determined future, but still says nothing about how those states are achieved and how mental properties affect physical states?
This caught my eye immediately because I wrote on this very topic just some months ago! While I affirm the libertarian position, the goal of the essay was more to consider the potential implications for free will in light of the various interpretations of QM.
1) If quantum physics giving objective probabilities rules out free will it rules out quantum physics being ceteris paribus. If objective probabilities constrain our neurobiologic choses they constrain any higher level phenomenon. I'd make what I think what you are saying is that if probabilification gives enough slack to allow free will then it can only do so by making quantum physics ceteris paribus laws.
Imagine 1000 cases where some human agent has the same psychological state (wants the same things has the same value standards etc.) and is in various quantum states, and is deciding on what to order for lunch. Take it that free will is when what he wants to eat (as a biological agent) determines what he orders. In which case since we've stipulated the 1000 are in the same psychological state they all order the same thing and in so doing probably exemplify a different outcome distribution than their objective quantum probabilities imply. So yeah the macro neuropsychological causes have to be extra causes that redefine the probabilities that can't merely be consistent with them and still do the job (determine what we order for lunch).
Whether this is just a friendly redescription of Mitchel's position or contradicts it depends on rather fine grained analysis of his account that I can't make from your description of it or reading the abstract of his paper.
2) Is somewhat pedantic, but I think worth pointing out. Depending on how one defines incompatibilism, any form of determinism is incompatible with free will, for example what you do being determined by what you want would be antithetical to free will. You can disprove that people have free will simply by observing that they do what they want. This seems silly to me and rather fatal to the idea that incompatibilism is a sensible much less the default way to define the dichotomy between personal responsibility and defining how decisions are made, but here we are. So one needs to stipulate not merely an incompatibilist frame, but which incompatibilist frame one is using.
I was worried I would be accused of constructing a straw man with my second point, but I notice someone who replied to this does indeed insist that doing what we want is proof we don't have free will. Again my point would be that Mitchel's definition of free will is just that our neuropsychological states are causal, so we do what we want and this is somehow independent of our microphysical constitution, it is incompatible with microphysics having causal closure. Tim Miller who I'm quoting is defining free will as being ruled out by causal closure of the neuro-psychological, if what we want is determined then we can't do what we want because doing what we want implies wanting what we want (or rather wanting from some prior place outside of the causal chain). I don't see any evidence that we (English speakers, or humans etc.) actually have a well agreed upon definition of free will, so I feel free to reject both of those different incompatibilist definitions as not fit to my purposes. I mean I think we can try and motivate different definitions of free will/moral responsibiltiy and so on, but we need to do so starting from the understanding that we are all working with slightly and even grossly definitions of each part of the debate.
I wonder if you're underrstimating how weird it is that we can't predict the individual electron's behavior in a single measurement.
Suppose we have a system that works less like "if the electron measures spin up you water Susan, if it measures spin down you don't" and more like "if the electron successively measures ↑↓↑↑↓ you water Susan, if it measures ↓↓↑↑↑ you don't, and if it measures anything else you hesitate until one of these patterns appears".
The statistical analysis wouldn't initially notice anything "wrong" with this electron, as (let's say) both options fall well within standard predictions, being about 50% up and 50% down; it would only become apparent that something peculiar was happening if the analyst noticed that the "watering Susan" pattern was strangely common. This discrepancy would not show up in bulk statistical analysis.
Having not seen this discrepancy in nature, we might posit a new rule that says it isn't allowed either — in effect, that the electron isn't allowed to generate code — but at this point we're another layer of statistical analysis deep. A defender of Mitchell's thesis might force us to dig a few layers deeper still by positing other subtle ways for the electron to make its "will" known across measurements. At what point then are we saying "electrons aren't allowed to buck the trends we've identified because that would buck the trends we've identified"?
What I've just said is almost just a version of "what we call universal natural laws are only universal until (maybe) they aren't", but I do think it's worth noting that a statistical law has an additional layer of conceptual slipperiness that a deterministic law lacks. Afaict we don't run into this problem of needing multiple (infinite?) layers of statistical analysis to forbid a physical entity from being smart/choosy if we've acccepted that it obeys simple deterministic laws.
The lack of a closed-form solution to the three-body problem, by induction, proves that determinism is only a philosophical construct, not unlike 'objective reality'.
The question is does determinism mean 1) for every A that there is some X, Y, and Z are sufficient for the consequent A (A iff X,Y and Z), or 2) that for every A that there is some X, Y and Z are sufficient to predict the consequent A.
Universal prediction (option 2) ) is definitely impossible just from the uncomputablability of certain outcomes (like the halting problem etc.) and so on that the second definition determinism is false, if one constrains it further and demands practically implementable prediction one can invoke the 3 Body Problem etc. to get even more cases of failure of 2) type determinism .
One might stipulate that the only definitive proof of 1) is 2). If 2) is the case 1) is definitely the case. In the absence of 2), technically 1) is unfalsifiable and unverifiable, there will be A's where you can not falsify that X, Y, Z is sufficient for A by showing that X, Y, Z imply (predict) not A and you can not verify that X, Y, Z is sufficient for A that it implies (predicts) A.
However as you say that seems a very thin reed, our ignorance is no evidence against the sufficiency of X, Y, and Z for A. We could have pretty good evidence of the law of causality (for every A there is an X, Y, Z sufficient to bring it about) or at least the law of probable (?) causality (for every A there is an X, Y, Z that gives it a definite probability).
OK: Claude said "it's a proven structural limitation. Poincaré showed in the late 19th century that the three-body problem is non-integrable in general: the series solutions diverge, and the phase space exhibits chaotic structure. Bruns and Poincaré essentially proved there are no new independent algebraic integrals of motion beyond the classical ones (energy, momentum, angular momentum). That's not a failure of technique — it's a theorem about what technique can achieve."
Actually it's more interesting than that. This is the first example I've seen of "strong emergence" - where the emergent system is known to result from the underlying physics, but it's been mathematically proven that it's not just very complicated.
“The physics tells us that approximately 900,000 of the duplicates will water Susan and approximately 100,000 of them will not. If we ran the experiment many times, each time creating a million more duplicates and waiting for them to decide, the physics tells us we would get roughly the same frequencies each time. But if what happens is totally up to each duplicate – in the radical incompatibilist sense – then there ought to be no such predictable frequency. The number that do and don’t water the plant should change each time, as the radically free choices of each individual varies.”
The ‘approximately’ and ‘roughly’ give you all the flexibility you need for the numbers to not be exactly the same each time you do the expreiment.
Also, you can’t exactly duplicate the same situation more than once.
Even radical free will allows for predictability and limits to outcomes. If there is no way to tell what each iteration will do then you have room for free will.
If quantum mechanics is correct, you can prove the universe is not deterministic. Here's how. Get a little lump of radium and a Geiger counter. If your radium is just the right size, then the probability that it makes the Geiger counter beep in any one minute period will be 50%. And quantum mechanics says that is completely and truly random. Over a large number of observations, the 50% holds, but in any single case, whether it happens in one minute is strictly undetermined, strictly random. So set up the radium and Geiger counter and decide that you will eat a muffin if the counter beeps within a minute or less, but otherwise you will put the muffin in the fridge. Your choice to eat is delegated to your radium/G counter setup, so you are not making the choice. And yet according to quantum mechanics, whether you eat or not is strictly random. Hence, the universe is not deterministic.
If you carefully inspect your conscious choices in normal life (no radium lumps on hand), you can never catch yourself making a free choice. At least, I can't. If I act according to a desire, I didn't choose that desire, I just find it in me. Even if I created a habit of having a certain desire by hard work, what motivated me to do that hard work? It was another desire that, inspected carefully, I didn't choose (e.g. preferring an orderly life over the disorder of frequently giving in to a short term desire to drink rum).
Another example: if I am persuaded that trying to always act in love is the best way to help steer the universe toward maximal flourishing for sentient beings, what made me like the idea of maximal flourishing? Another desire, and the knowledge from my experience that I don't like to suffer and it appears that other creatures don't like to either. But I didn't choose to have a type of body that can suffer physical pain. I just found myself in one. This can go on endlessly, and I have never been able to see that I have actually make a truly free choice.
If panexperientialism or panpsychism is true, maybe smaller constituents of myself, my organs, cells, neurons, etc. make truly free choices and flush them up into my consciousness in the form of desires. But again, I can't detect that process so I, my conscious mind, does not seem to be able to make truly free choices as far as I can definitively identify.
So can it be that the universe is not determined, yet humans don't have the ability to make truly free choices?
I'm on board with a lot of this, assuming for the sake of argument an incompatibilist position.
But I don't think it makes sense for *The* Laws of Physics to be ceteris paribus laws. The laws of physics as currently understood could be ceteris paribus laws, but, to me, *The* Laws of Physics are the rules by which the world works. Even if there is a patchwork of different effective laws for different domains, to me, *the* laws of physics describe the overall patchwork, including rules to decide which rules govern which domains.
So, if there is strongly emergent libertarian free will, such that the laws of physics as currently understood are just ceteris paribus laws, then I think this means that *The* Laws of Physics will have to describe how free will works, and then *The* Laws of Physics are not ceteris paribus laws.
The problem is that people who want libertarian free will don't usually seem to want it to be describable in a lawlike way. That would take away the magic and reduce it again to the kind of mechanistic account they're trying to avoid. But either there are rules by which it works or there aren't. If there are, then it's ultimately going to be too mechanistic to satisfy libertarians. If there aren't, then we're handwaving and giving up any hope of understanding it. That's unscientific and against the spirit of methodological naturalism.
I also think it's incoherent. I don't think there can be any kind of fundamental process at play which cannot be described precisely in principle, and a precise description is all you need for a mathematical law.
I think that the demand that any kind of fundamental fact about reality is described by a mathematical law is a very, very strong claim and will get into problems not only with free will but also with a lot of conceptual thought (including mathematical thought). There is no mathematical law that bachelors are unmarried, and there is no mathematical law that mechanistically describes the conceptual steps that a usual mathematical proof takes. Or did I misunderstand the demand?
I think you misunderstood the demand, which is not necessarily your fault!
I think that bachelors being unmarried is neither a fundamental fact about reality (your phrase) nor a fundamental process at play (mine). It's not fundamental because we don't need to make changes to the fundamental laws of physics to accommodate bachelors being unmarried. Having higher-level rules that say that bachelors are unmarried doesn't mean that the laws we do have are ceteris paribus laws. There is no conflict between the laws of physics and rules about bachelors being unmarried. So all our observations of unmarried bachelors etc are consistent with laws of physics which are describable mathematically.
Whether the steps taken in a mathematical proof can be reduced to (at least computable) physics is controversial. Roger Penrose would interpret Godel's incompleteness theorems as showing that that they cannot. Most other experts (I believe) disagree with him on this. If Penrose is wrong, then we don't need special physics to account for how mathematicians (or LLMs, now) produce mathematical proofs. There is no fundamental process at play which cannot be given a standard mathematical description.
Processes at higher levels of description (involving concepts, truth, insight and so on) are not fundamental, and need not be describable mathematically, because these concepts are loose, vague and ambiguous, and are convenient ways of generalising and abstracting useful commonallities between distinct physical processes which are each describable as arising out of more fundamental processes which are describable mathematically.
Thanks! I still feel like the claim that any kind of fundamental process needs to be understood in terms of mathematics is a strong limitation on what kinds of truth you are admitting. Nearly all philosophy papers make arguments and claims, but very few philosophy papers give mathematical laws to mechanize their arguments. Would you say that this rules out the arguments in almost all philosophy papers as describing a fundamental process of reality?
It doesn't necessarily need to be understood in terms of mathematics. I'm not saying that art is a waste of time, and we need to concern ourselves exclusively with writing theorems.
My claim is that it is incoherent to posit causal forces at play in the world which have no underlying mechanisms, unless those causal forces are already mechanistic (describable mathematically). But that's not to say that we can't talk about things without understanding the underlying mechanisms. I just don't think it's coherent to deny that such mechanisms exist.
Ah, right, now I get your comment - you were only talking about causation and whether it is mechanistically described or not, and I went all overboard and thought you were talking about all of reality. For causation I understand your point much better.
I still think it is wrong, though, because I think that sentences like "the sight of a loved one made me happy" describe a process of reality that I have (imperfect but good) epistemic access to without needing a mathematical description of it, and if there ever will be a good mathematical description for that, perhaps in terms of a reductive explanation of neuro-physics, that is great but that is not a prerequisite for the sentence to be true, nor do I believe that such an explanation is forthcoming with certainty.
But I agree that non-mechanistic/mathematical causation happening in the case of a free choice is a bullet that you need to bite if you want proper free will. And I also agree that we need more understanding how causation works in a free choice, in order to be less handwavy. But I think such better understanding is possible in philosophical terms without needing recourse to mathematics.
I have no problem with sentences like "the sight of a loved one made me happy" describing a process of reality that you have (imperfect but good) epistemic access to without having a mathematical description of it.
Say sentences like this all day and I'll mostly assume you are saying true things.
But say "Maybe there is no underlying mechanism for how it is that the sight of a loved one makes me happy" and I'll think you are not making sense.
It's bad enough for agency that incompatibilists want to get rid of determinism, but if they get rid of probability as well, we would be completely disabled. If I really, really want to do A and not do B, then hitched to physical determinism, I will reliably do A. If determinism is false, then I can at least say that I will probably do A. But if I can't even say that I will probably do A, what does this mean? That repeating an A-loving and B-hating situation a million times I might do A anything between 0 and a million times?
We do have evidence of top down causation in physics. We have the Cambrian Explosion. We have the Genesis event. We have the point of human consciousness. We have the beginning of the universe.
You get the idea 💡.
Philip, I think your critique of Mitchell holds, but you then make the same mistake one level up. Both of you look for freedom in the wrong place.
Mitchell looks for it in the room that quantum indeterminism supposedly opens; you show correctly that this room does not exist, because the objective probability of the Born rule constrains just as much as a deterministic law. You then relocate freedom to a layer of "higher-level causal powers" that complement physics through ceteris paribus laws. But this is simply a displacement of the same move: you are still looking for a place inside or beside physical causation where freedom can be inserted. The search itself is the mistake. Free will is not a kind of causation that competes with physical causation, escapes it, or is added to it from outside.
Your duplicates experiment shows this. You reduce the decision to water Susan to a position of particles over which the Born rule fixes a frequency, and conclude that the recurrence of that frequency rules out freedom. But that configuration has meaning only in the physical regime. Free will is not a position of particles; it is self-modeling - the level at which a system produces a model of itself and acts from it. Asking how many of a million identical duplicates will water the plant simply asks how many identical configurations yield identical outcomes. That stays closed within the physical plane and says nothing about self-modeling, because it has already cancelled that level in the very framing.
Of course free will obeys physics - but only on the physical plane, the way we obey gravity. I cannot defy gravity without expending extraordinary energy; I defy it only through technology that overcomes it on its own terms. This obedience does not touch the level where free will operates, because it does not operate as a physical force competing with gravity, but as self-modeling that occurs through the physical configuration once complexity crosses a threshold - not against it, not beside it.
And free will is not exclusively human. An animal that decides how and when to hunt its prey, and that fights for its young, acts on the basis of a model of itself; it has an "I", feels emotion, sustains social life. What separates us from it is not the presence of free will but its extent, and the extent comes from the capacity to project the imagination - to model possible, counterfactual, not-yet-existing states. That is an increase of complexity, not an added substance.
The only constraint on my will is rational and arises from ethics. Ethics is the natural form of rationality; the two are interdependent. I understand what is good and what is bad because I grasp the consequences of my actions, and this arises from the same capacity that widens freedom: the projection of the imagination. To model an action before I carry it out and to see that it undermines the conditions that keep me or another in existence - this is what makes the ethical limit internal to rationality, not a norm imposed from outside.
So the determinism-indeterminism debate as a source of freedom seems to me misplaced at the root. Both sides look for freedom in the regime where physical causation happens. It is not there, and it never was.
Can someone explain to me how indeterminacy opens the door to agency though? It seems at best it allows for an open, non-determined future, but still says nothing about how those states are achieved and how mental properties affect physical states?
This caught my eye immediately because I wrote on this very topic just some months ago! While I affirm the libertarian position, the goal of the essay was more to consider the potential implications for free will in light of the various interpretations of QM.
Two things to note.
1) If quantum physics giving objective probabilities rules out free will it rules out quantum physics being ceteris paribus. If objective probabilities constrain our neurobiologic choses they constrain any higher level phenomenon. I'd make what I think what you are saying is that if probabilification gives enough slack to allow free will then it can only do so by making quantum physics ceteris paribus laws.
Imagine 1000 cases where some human agent has the same psychological state (wants the same things has the same value standards etc.) and is in various quantum states, and is deciding on what to order for lunch. Take it that free will is when what he wants to eat (as a biological agent) determines what he orders. In which case since we've stipulated the 1000 are in the same psychological state they all order the same thing and in so doing probably exemplify a different outcome distribution than their objective quantum probabilities imply. So yeah the macro neuropsychological causes have to be extra causes that redefine the probabilities that can't merely be consistent with them and still do the job (determine what we order for lunch).
Whether this is just a friendly redescription of Mitchel's position or contradicts it depends on rather fine grained analysis of his account that I can't make from your description of it or reading the abstract of his paper.
2) Is somewhat pedantic, but I think worth pointing out. Depending on how one defines incompatibilism, any form of determinism is incompatible with free will, for example what you do being determined by what you want would be antithetical to free will. You can disprove that people have free will simply by observing that they do what they want. This seems silly to me and rather fatal to the idea that incompatibilism is a sensible much less the default way to define the dichotomy between personal responsibility and defining how decisions are made, but here we are. So one needs to stipulate not merely an incompatibilist frame, but which incompatibilist frame one is using.
I was worried I would be accused of constructing a straw man with my second point, but I notice someone who replied to this does indeed insist that doing what we want is proof we don't have free will. Again my point would be that Mitchel's definition of free will is just that our neuropsychological states are causal, so we do what we want and this is somehow independent of our microphysical constitution, it is incompatible with microphysics having causal closure. Tim Miller who I'm quoting is defining free will as being ruled out by causal closure of the neuro-psychological, if what we want is determined then we can't do what we want because doing what we want implies wanting what we want (or rather wanting from some prior place outside of the causal chain). I don't see any evidence that we (English speakers, or humans etc.) actually have a well agreed upon definition of free will, so I feel free to reject both of those different incompatibilist definitions as not fit to my purposes. I mean I think we can try and motivate different definitions of free will/moral responsibiltiy and so on, but we need to do so starting from the understanding that we are all working with slightly and even grossly definitions of each part of the debate.
https://substack.com/@timbmiller/note/c-266046416?utm_source=notes-share-action&r=27ibk6
We don't need to discuss quantum physics because this part is already wrong!:
>> If the laws of physics are deterministic, then free will is ruled out
https://ramblingafter.substack.com/p/to-understand-free-will-flip-a-coin
https://substack.com/@hawkeyespeaks/note/p-199430506?r=2p93cq questions are answered here
Thanks for this!
I wonder if you're underrstimating how weird it is that we can't predict the individual electron's behavior in a single measurement.
Suppose we have a system that works less like "if the electron measures spin up you water Susan, if it measures spin down you don't" and more like "if the electron successively measures ↑↓↑↑↓ you water Susan, if it measures ↓↓↑↑↑ you don't, and if it measures anything else you hesitate until one of these patterns appears".
The statistical analysis wouldn't initially notice anything "wrong" with this electron, as (let's say) both options fall well within standard predictions, being about 50% up and 50% down; it would only become apparent that something peculiar was happening if the analyst noticed that the "watering Susan" pattern was strangely common. This discrepancy would not show up in bulk statistical analysis.
Having not seen this discrepancy in nature, we might posit a new rule that says it isn't allowed either — in effect, that the electron isn't allowed to generate code — but at this point we're another layer of statistical analysis deep. A defender of Mitchell's thesis might force us to dig a few layers deeper still by positing other subtle ways for the electron to make its "will" known across measurements. At what point then are we saying "electrons aren't allowed to buck the trends we've identified because that would buck the trends we've identified"?
What I've just said is almost just a version of "what we call universal natural laws are only universal until (maybe) they aren't", but I do think it's worth noting that a statistical law has an additional layer of conceptual slipperiness that a deterministic law lacks. Afaict we don't run into this problem of needing multiple (infinite?) layers of statistical analysis to forbid a physical entity from being smart/choosy if we've acccepted that it obeys simple deterministic laws.
If Gödel had lived longer, could he have solved the free will problem?
The lack of a closed-form solution to the three-body problem, by induction, proves that determinism is only a philosophical construct, not unlike 'objective reality'.
Just cos we don't know the solution doesn't mean there isn't a fact of the matter. The universe knows the answer!
Actually we know the solution instantaneously, our failure is in the integration and that's just algebraic incompetence.
The question is does determinism mean 1) for every A that there is some X, Y, and Z are sufficient for the consequent A (A iff X,Y and Z), or 2) that for every A that there is some X, Y and Z are sufficient to predict the consequent A.
Universal prediction (option 2) ) is definitely impossible just from the uncomputablability of certain outcomes (like the halting problem etc.) and so on that the second definition determinism is false, if one constrains it further and demands practically implementable prediction one can invoke the 3 Body Problem etc. to get even more cases of failure of 2) type determinism .
One might stipulate that the only definitive proof of 1) is 2). If 2) is the case 1) is definitely the case. In the absence of 2), technically 1) is unfalsifiable and unverifiable, there will be A's where you can not falsify that X, Y, Z is sufficient for A by showing that X, Y, Z imply (predict) not A and you can not verify that X, Y, Z is sufficient for A that it implies (predicts) A.
However as you say that seems a very thin reed, our ignorance is no evidence against the sufficiency of X, Y, and Z for A. We could have pretty good evidence of the law of causality (for every A there is an X, Y, Z sufficient to bring it about) or at least the law of probable (?) causality (for every A there is an X, Y, Z that gives it a definite probability).
Righto - but 'definite probability' is already a construct or approximation since the initial conditions can never be identical over t.
That's a good story!
OK: Claude said "it's a proven structural limitation. Poincaré showed in the late 19th century that the three-body problem is non-integrable in general: the series solutions diverge, and the phase space exhibits chaotic structure. Bruns and Poincaré essentially proved there are no new independent algebraic integrals of motion beyond the classical ones (energy, momentum, angular momentum). That's not a failure of technique — it's a theorem about what technique can achieve."
So now I know. Thanks.
Actually it's more interesting than that. This is the first example I've seen of "strong emergence" - where the emergent system is known to result from the underlying physics, but it's been mathematically proven that it's not just very complicated.
Motivation rules out free will. Prove me wrong: give me one example where you did something out of free will that you weren't motivated to do.
I could be hungry. That is motivation. But I can still choose to eat or not eat. And I can choose what to eat, if I eat.
Complete red herring. But i'm at work, so.