A computer scientist finds the question of free will uninteresting for bad reasons

January 15, 2019 • 10:30 am

UPDATE: Scott Aaronson has emailed me and pointed out that his views on this matter are set out in a clearer and longer way in a publicly available paper he wrote called “The ghost in the quantum Turing machine.”  It’s 85 pages long, I wasn’t aware of its existence, and it is probably above my pay grade, but perhaps not for some readers who are physics-savvy and also willing to read the paper. If you do, weigh in below. Please consider this post a response to Aaronson’s 8.5-minute explication of free will on the “Closer to Truth” interview and not to the totality of his published views.

I’ll add, with Scott’s permission, a clarification that he emailed me along with the link to his paper:

“Briefly, you can make any theory “deterministic” by the addition of hidden variables, which is exactly what de Broglie and Bohm did for quantum mechanics.  In that sense, to ask whether the world is deterministic is not even to ask an interesting question about it; it’s only to ask about a particular description of the world.  But whether you can predict someone’s behavior without destroying them IS a question about the world.  So focusing on that means you’re led to actual meaty empirical questions about the world, rather than endless and boring semantic debates about the exact meanings of terms like “free will” and “compatibilism.”  I’d hope anyone with a scientific mindset would find that to be a feature.”


I may be remiss in not knowing who Scott Aaronson is, as I gather he’s quite well known. He’s the David J. Bruton Centennial Professor of Computer Science at The University of Texas at Austin, and works on quantum computing. He also has a popular website called Shtetl-Optimized (“Shtetl” refers to small Jewish towns in eastern Europe), which deals largely with computation but also personal revelations and amusing tales (see here about his quirks and then here for a story in which airport police arrested him because he absent-mindedly took change from his purchase of a smoothie—on a debit card!—from the tip jar in an airport food court.

In this 8½-minute video, Aaronson is quizzed about free will by Robert Lawrence Kuhn for the “Closer to Truth” public television series. (I’m weighing an invitation to participate in this series.) Do watch the video if you have time.

Aaronson thinks there’s a real and important question in the free-will debates, but argues that that question is not whether physical determinism of our thoughts and actions be true, but whether they are predictable. He thinks that the question of whether there is free will could in principle be solved by answering the following question:

Would one be able to construct a machine that, if it was programmed with the immense knowledge about your brain and your environment that would be required to do such a job, would be able to predict your actions?”

If the answer is “yes,” he says, then the question is solved: we have no free will and are “automatons”.

(By the way, Aaronson doesn’t appear to be a compatibilist, as he says that if such a machine were possible, and all our choices were predictable, we would certainly not have free will, even though some people would say we would still have free will. Those “some people” are surely compatibilists, who find the notion of free will compatible with physical determinism.)

Because, Aaronson says, the answer to the Big Machine question is not known (I would argue that in principle, because behavior and thoughts are the results of physical processes, the answer is “yes”), he finds the question of free will to be moot.

As he says at 4:15,

“My view is that I don’t care about determinism if it can’t be cashed out into actual predictability.”

This seems to me misguided, conflating predictability with the question of determinism. Surely it will be impossible, at least in our time, to gather the requisite information to accurately predict someone’s behavior, for such a computer would have to model not just a person’s brain, but also the entire universe, for the universe impinges on a person’s brain in ways that affect their behavior.

Further, insofar as fundamentally unpredictable quantum events may determine behavior, no machine could ever model those: at best it could give probabilities of different behaviors. But those quantum effects do not violate physical determinism, and cannot give us free will in the sense that most people think of it. (Surveys show that most people think of free will as “contracausal”: the you-could-have-done-otherwise form of free will.) In fact, insofar as Aaronson espouses physical causation of behavior in his answer—he says he sees the brain as a kind of computer running a program—he’s already admitted physical determinism.

Do Aaronson’s lucubrations, then, make the question of free will uninteresting?

Not to me; I find the issue not only philosophically interesting but a question that has practical ramifications, even if we can’t build the Aaronson Machine. If we are truly biological automatons, which I think is true on first principles (viz., we are made of molecules), then that has huge implications for religious thought and dogma, which of course depend on assuming contracausal free will. You are free to choose your saviour, your faith, your actions, and, for gay Catholics, whether to commit homosexual acts. Because you make free choices, making the wrong choice will send you to perdition, and making the right one to God, Yahweh, or Allah.

There are ramifications for the justice system. I firmly believe that if we grasped that nobody, including criminals, has a “choice” in whether or not to do something, like mugging someone, we would structure the justice system differently, concentrating less on retribution and more on keeping baddies out of society, trying to reform them, and using punishment as a deterrent to improve society.

There are ramifications for politics. Once you realize that people’s acts solely reflect the physical consequences of their genetic endowment and environment, you (or at least I!) become more sympathetic to the plight of those who drew a bad hand in the poker game of life. The notion of the “Just World”, in which people get what they deserve, depends on accepting contracausal free will. But that view must be tempered by realizing that neither the successful nor the downtrodden freely chose their paths.

I’ve always said that I don’t really care if you say people have “free will” if you define that term in some compatibilist way like “the inputs to creating a human behavior (output) are complex”.

What I care about is whether determinism be true. And I think it is, though of course I can’t prove it. All I can say is that the laws of physics don’t ever seem to be violated, and, as Sean Carroll emphasizes, the physics of everyday life is completely known.

What we see above are the ruminations of a man whose life is devoted to computing, and his profession shows in the way he thinks about free will, turning it into the question of whether machines could predict behavior.

166 thoughts on “A computer scientist finds the question of free will uninteresting for bad reasons

  1. I use the software comparison to dismiss the “but I make decisions” objection to no-free-will. So, while not making free will questions uninteresting, I think it resolves them pretty rapidly making further discussion pointless.

  2. Jerry, you argue like a Compatibillist. There cannot be ramifications, because you employ counterfactual reason there (where without the determinist insight, reality/society would go one way, and with the insight, it goes another way).

    Those ramifications are only trivially true effects of causes. They simply happen as determined.

    1. Sorry, but saying that doing one thing has different consequences from another is not a violation of hard determinism. You can in fact test these predictions, though of course whether or not you do the test is determined. But different people can do different tests, and if the results comport then the ramifications are real.

  3. I think we know enough now about human behavior to have gone beyond the point of using retribution i the treatment of people who break the law. To me, retribution is revenge and torture.

          1. I want tRump to spend at least 5 years in prison. Am I being aggressive, do you think? Retributive? Or are there ramifications which should make me regret those thoughts and settle for him spending his last decade at Mar-a-Lago with his 4th wife?

          2. Excuse me, that’s blatantly unjust. What has tRump’s 4th wife done to deserve that?


  4. This is really a question about simulating the universe before it happens. That does not preclude determinism. In my view, it’s just not physically possible to construct such a machine. It would have to be the very universe it is trying to predict.

    The Free Will Problem becomes the ‘Big Machine Problem’. If no such machine can exist, given the laws of physics, then determinism, itself, removes the possibility of predicting future actions, or states, as physicists would call them.

    1. This is more or less Laplace’s being with sufficient intellect(demon) problem.

      My thought on this nevermind that such a machine is impossible, it would have to be so large it would start affecting the universe around it.

      The only way to predict with any accuracy is to survive long enough to experience. Of course this is no longer a prediction.

    2. I disagree. It’s only necessary to construct a sufficiently accurate simulation of a person’s brain and get live feeds of all their sensory inputs.

      I think that is perfectly possible in principle, but we are centuries away from being able to do it in such a way that we don’t have to throw away their real brain.

      1. That might be true if there are no decisions made in the brain that hinge on a single quantum state. In practice and in principle, all simulations, in general, are inaccurate. This is a result of every event in the universe being unique and no event can be repeated. Only comparatively trivial systems can be modeled with 100% accuracy. The “halting problem” says, in general, we can’t create an algorithm that can examine another algorithm and tell if it halts. All these things are really one and will frustrate your perfect simulation of the human brain and its predictions.

        1. Human brain cells are vastly larger than the quantum scale. I’d bet my house that there brain doesn’t rely on quantum level effects.

          You also don’t need perfect accuracy, only sufficient accuracy which, I agree, is probably a long way off in the future but not fundamentally unobtainable.

          I don’t see how the halting problem is relevant at all. You don’t have to know if an algorithm halts in order to simulate it. In fact, unless you are claiming that the human brain is something fundamentally different to a computer, we know it can (in principle) be simulated by another Turing complete machine.

          1. “In fact, unless you are claiming that the human brain is something fundamentally different to a computer,”

            Isn’t it rather obvious that that is the case? The computer upon which I type is 100 million times as fast as the human brain but my dog is smarter.

          2. I don’t think that “the speed of the human brain” is a meaningless concept. Brains don’t have a speed. Brains have complexity. Neurons have attributes like speed when talking about the rate of transmitting electrical pulses. But saying your computer is faster than your brain doesn’t make sense to me.

          3. (I should have pointed out that talking about the speed of a computer is also rather loosely tied to reality. We can talk about the speed of various components of a computer, but speed at which a CPU can compare a couple of values has little to do with the speed at which signals are sent between miscellaneous components when the computer is doing “work”.)

          4. I disagree, Paul. I don’t even think the concept makes sense. What could it even mean? Speed to do what? Brains do all kinds of things simultaneously, what could you even be measuring? (I realize now that my first sentence, above, has a superfluous “don’t” that confuses things. Apologies to all for that. I should have simply said I think it is meaningless to to talk about measuring something that we can’t even define.)

          5. I am only pointing out that we can measure whatever we need to measure. Certainly, research is full of measurements of reaction times, response times, etc. which involve the brain. At the neuron level, response times are measured all the time. If we don’t have a measure called “brain speed”, it is only due to our not defining what is to be measured.

            An analogous situation exists with respect to digital computers. There are many different measures, some of which may even not be applicable to a specific computer. There is no one accepted measure of “computer speed” but this is not because it is a meaningless concept or because we can’t measure things. It is only because we can measure so many things and they all give different values. While this is true for brains and computers, it is somewhat trivial, IMHO.

          6. We can, of course, measure many things about brains. And we can measure many things that involve brains. Some of those measures will involve timing. But that doesn’t correspond to a speed measure for brains. I think this is more than just a matter of defining something to measure. We can define it as anything we like but that doesn’t make the definition meaningful. (I define it as the rate of travel of African Swallows flying with coconuts.)

            Computer speed is also meaningless. That’s exactly why we don’t measure it. Hell, even some of the more sensible measures aren’t all that meaningful.

            I don’t think this is a trivial matter when “the obvious” is asserted because (it is claimed) “a computer is 100 times faster than a human brain”.

          7. “A computer is 100 times faster than a human brain”

            Such a statement is not wrong if we measure both on a specific task: adding columns of numbers, for example. Clearly a fair test involves testing both the computer and the brain on the same task. It also goes without saying that both have to be able to perform the task. As with most statements, the meaning depends greatly on the context.

          8. But that isn’t “brain speed”, Paul. It is the speed of addition of a column of numbers. You’re just using the “I define it as” gambit to arbitrarily whip something up.

          9. This is ridiculous. I was just giving an example. Define whatever you want to be “brain speed” so we can measure it. We all agree that “brain speed” is meaningless if we don’t define what exactly we want to measure more precisely. Can we leave it at that?

          10. I’m happy to leave it but I don’t think the matter is ridiculous because of the context in which the issue began.

            “Brain speed” was used to support “the obvious” (not by you). I don’t think “the obvious” is obvious (that a brain is unlike a computer).

          11. I’m not sure what “speed of the brain” means either. However, there is a maximum speed of transition of signals through synapses and along nerve bundles that move signals and data between different modules. FMR is probably capable of computing these values. Current imagine methods can monitor individual neurons, if my memory serves.
            Another way speed might be a useful term is in comparing how quickly subjects can accomplish mental tasks, such as add a sum. Thus, two subjects could be compared.

          12. “Another way speed might be a useful term is in comparing how quickly subjects can accomplish mental tasks, such as add a sum.”

            In fact, I think it is the only way to usefully compare brains to non-brains at the moment. It is also the most meaningful speed measure. In computing, it usually only makes sense to measure the speed of whatever it is you actually care about. In other words, proxies may be easier to measure but they are often misleading.

          13. Oh, and one more thing…
            ” arbitrarily whip something up.”

            Arbitrary whipping is, as far as I know, a misdemeanor offense on this forum. Even accusing someone of arbitrary whipping can be considered a way to practice arbitrary whipping. Just my 2 cents.

          14. No it isn’t obvious to me.

            It’s certainly massively parallel and the connectivity of each neutron is huge in comparison to our normal efforts but it is not (IMO) fundamentally more powerful in an abstract mathematical sense (i.e. it can be simulated by a universal Turing machine).

          15. Each neurone in the brain accepts inputs, and processes them in a fairly simplistic way to produce outputs. At pretty much any higher level, the different parts of the brain can all do processing at the same time.

            I’ve read The Emperor’s New Mind. I think Roger Penrose is wrong. Even if he were right, what’s to stop us from harnessing the same quantum phenomena he thinks are at work in the brain in our computers?

          16. “the different parts of the brain can all do processing at the same time.”

            There is no evidence that they do.

            “I’ve read The Emperor’s New Mind. I think Roger Penrose is wrong.”

            The reference is to his “Shadows of the Mind” (1994). As to your opinion that Penrose is wrong, without a justification, I’m inclined to agree with Penrose.

            “Even if he were right, what’s to stop us from harnessing the same quantum phenomena he thinks are at work in the brain in our computers?”

            Nothing stopping us except we haven’t done it, which means it’s vaporware – science fiction at this point.

          17. “the different parts of the brain can all do processing at the same time.”

            I’m guessing he only means that the brain exhibits massive parallelism which is true.

            “I’ve read The Emperor’s New Mind. I think Roger Penrose is wrong.”

            Most people involved in brain research and AI believe Penrose is wrong. It doesn’t mean he’s wrong, of course, but I think he is. While Penrose has done good work, most think he’s a quack in this area.

          18. There is no evidence that they do.

            Can you hear things and see things at the same time?

            The reference is to his “Shadows of the Mind”

            I know. His arguments in The Emperor’s New Mind were unconvincing. Perhaps you could say how he mends them in “Shadows of the Mind”.

            Nothing stopping us except we haven’t done it, which means it’s vaporware – science fiction at this point.

            Nobody is denying that we do not have the technology right now. And I believe I did say it would be centuries before we do possess the necessary technology.

            What I am denying is the idea that there is something magic about the human brain that means it can’t be simulated in principle by a human made machine.

          19. There is no evidence that they do.

            Can you hear things and see things at the same time?

            The reference is to his “Shadows of the Mind”

            I know. His arguments in The Emperor’s New Mind were unconvincing. Perhaps you could say how he mends them in “Shadows of the Mind”.

            Nothing stopping us except we haven’t done it, which means it’s vaporware – science fiction at this point.

            Nobody is denying that we do not have the technology right now. And I believe I did say it would be centuries before we do possess the necessary technology.

            What I am denying is the idea that there is something magic about the human brain that means it can’t be simulated in principle by a human made machine.

          20. I believe the brain can be simulated on computer but this does not mean simulation to the level that it is identical to some particular human being. This kind of deep simulation would be possible only in principle because it would likely require simulation of every chemical and electrical value in every cell and even in its immediate environment. We have no evidence that the brain works like a digital computer in that it records state using massive amounts of charge/magnetism which is the basis for computer reliability and reproducability. In fact, there is a great deal of evidence that the brain doesn’t work that way. And, as I already pointed out, where analog input is involved, regardless of human brain or computer, there is uncertainty. A digital computer hooked to a photocell that outputs discrete colors (red, blue, green, etc.) is not predictable if the color of the light hitting the photocell is right on the border between those discrete colors. Same with a brain.

            I do believe we will someday build a digital brain and achieve true Artificial General Intelligence but we won’t claim that it simulates a specific human being. That would be a totally different beast. And, to loop back around, that level of simulation is required to predict the actions of a human.

          21. I would agree that the brain doesn’t depend on quantum mechanics is the way that Penrose, et al suggest. My point was that there are decisions the brain makes that are right on the edge between possibilities that can change via just a nudge at the quantum level either way. This is certainly the case for our senses which receive analog data. The brain may or may not be a digital computer but we can be certain that at least part of it is analog and, therefore, subject to uncertainty of this kind.

      2. What is sufficiently accurate? Supercomputers can not accurately describe the spin state of 1000 atoms after a second. Simulated brains sufficiently accurately is going to fail for general cases that require predictability.

  5. Some think if you give a computer enough input/information it can do or predict anything. Others think it unlikely the computer can ever replace or replicate much of the human mind’s abilities.

    1. There’s a big gap between those two things. It is impossible for a computer or a human to predict everything or anything in the sense Aaronson puts it. However, making computers think like humans seems likely though it is not “just around the corner”. The human brain is not magic. Claims that the brain can compute things in ways that computers cannot must be proven. IMHO, Penrose and others are simply wrong.

    2. Imagine a future solar max causes a geomagnetic storm that wipes out a third of GPS and all electrical infrastructure in Siberia. Most children in a village grow up in economic disorder and unrest, cause a Russian revolt and start WWIII. How on earth is a computer going to predict any of that?

      Humans are easy, what they do with external inputs is not easy. Find a computer that will predict the identity of a murderer in 50 years and there’s a bridge from Earth to Pluto worth investing in.

      1. Why Russian? How about a storm in Montana and a redneck revolt?

        In fact, what computer could have predicted that tRump would ever have been elected? Anything social is an incredibly complex system, chaotic in the mathematical sense.


        1. Many people predicted Trump would be elected. And now that he is, it is very easy to predict what his cult-followers will do. They will do whatever he asks, and believe whatever he says. Most of the time, anything social is an incredibly complex system…the Trump effect negates this complex system.

  6. “What I care about is whether determinism be true. And I think it is, though of course I can’t prove it. All I can say is that the laws of physics don’t ever seem to be violated, and, as Sean Carroll emphasizes, the physics of everyday life is completely known.”

    This is also what I care about – if determinism is true in the first place. And as you, Prof. Coyne, correctly say, you can’t prove it. As far as I know, nobody can prove it. I, on the other hand, don’t think that determinism in the sense of radical determinism (which would obviously be the same as predestination,no?)is true. I also can’t prove it, of course. But I do think that if a fully determined world (the block universe) is completely indistinguishable from this experienced world where predictability a n d unpredictability together are present everywhere and chance and randomness are concepts that are so ingrained in every language that it is unthinkable to do without it (how to talk about evolution without the concept of true randomness?)- well then, determinism is just useless as argument against a contra-causal free will. The contra-causal free will is easily rejected by naturalism alone. To replace much of what is called “determined” with “conditioned”, makes it already easier to think about it.

    I do not think that I have contra-causal free will, but I think that in a real sense I am, like any other biological machine that is able to interact with the world, participating in the ongoing development of the history of the universe – which is basically unpredictable, but of course limited by what is physically impossible.

    Is there anything truly unreasonable about my stance? I can’t see it myself, of course.

    1. I agree. We are just playing our roles in the universe. Our sense of free will is just as valid as anything else. If we can’t use determinism to predict what people will decide, or anything else for that matter, then determinism has no impact on our daily lives and we would best move on as we were.

      1. You are giving your own opinion here, with which I obviously don’t agree. And determinism has certainly had an impact on MY life. So blanket statements that “determinism” has no impact on our daily lives are simply untrue.

        We can’t prove that the sun won’t rise tomorrow, either, but I’d bet my savings on it. All the evidence we have supports the view that the laws of physics apply everywhere in the Universe, and certainly on Earth, and that’s good enough for me.

        1. My point was that we live our lives the way we do whether or not determinism is true. Obviously it impacts our lives. In fact, if determinism is true, it determines everything. As Aaronson points out, if knowledge that the universe is deterministic doesn’t allow us to predict anything, then it is of no use to us. This observation is similar to the one leveled at the multiple universe hypothesis. If it can’t be tested, is it really a scientific hypothesis? Perhaps we can test whether the universe is deterministic but we can’t test what it determines.

          1. Seems like you are claiming that having the awareness that determinism is true makes no difference to they way we live our lives. I don’t know how that position is defensible. All you need is one person to come forward and say “it has for me” (as Jerry did) and you’ve lost your case.

          2. At that level, you are right of course but that is a trivial argument. One can be inspired by anything. I could say my belief that a turtle holds up the world has guided my entire life. That can be a true statement regardless of whether a turtle in fact holds up the world.

          3. “Seems like you are claiming that having the awareness that determinism is true makes no difference to they way we live our lives.”

            GBH, I’d say that’s true for me. I can’t say that contemplating determinism has made any difference to my actions. I’d say any tolerance I have towards others comes more from a sense of ‘fairness’ and Kant’s moral imperative than from determinism (and long predated any acquaintance with either of those topics).

            Evidently it has made a difference for Jerry, and presumably many others. But it’s not universal, in fact I doubt that most people have ever directly considered the question.


          4. Ohmigods, my humble apologies, GBJ. I really did not intend to make that typo! Diana’s (or was it Diane’s?) slip from the other day must have got stuck in my brain. I should have caught it before I posted.


          5. I see my typo has attracted more responses than all my bloviating about Free Will or determinism.

            There’s a moral in that somewhere…


    2. “I do not think that I have contra-causal free will”

      Once you realize that contra-causal free will would of necessity be contra-self free will, the value of the idea evaporates.

      A meaningful concept of self requires determinism. No matter that many people think contra-causal free will is desirable; it is an incoherent concept.

  7. Dr. Coyne, you have said repeatedly that determinism leads you to be more empathetic to other people, including criminals, but why should it lead society to be more empathetic instead of less? For example, the thought might instead be: “criminals are malfunctioning meat robots, so let’s just decommission them all.”

    In other words, I’m curious where you get the idea that this would lead to more humane punishments, instead of harsher and more severe punishments on the grounds that human dignity is after all an illusion.

    (No, I don’t personally think we should do this.)

    1. I don’t think that “human dignity” is a real thing and doesn’t affect anyone’s motivations.

      I agree with Jerry that determinism *can* lead to more empathy if one explores the consequences. The only thing people truly own is their subjective experience, and that experience is determined by events outside of their control.

    2. Because robots can be deprogrammed and their behavior changed. Human dignity, whatever that is, has little to do with my views on the justice system, which are largely about how to make society function better without causing unnecessary pain to people.

      At least in my case, my acceptance of determinism, which wasn’t that long ago, did indeed lead me to favor more humane workings of our justice system.

      1. Right, but there are good arguments for those reforms that don’t involve, or are not inspired by, determinism. Compatibilists and incompatibilists should unite for justice reform.

        1. I agree too. So then what is argument for keeping the concept of free will? My will is a result of causality, be it deterministic or indeterministic.

          I look at someone like Trump, and find myself having a dislike of that particular President. I realise that my dislike is primarily a result of my environment past and present. Also trump is a product of his environment past and present. We both share a greater environment, this world and the universe in general.

          We both cannot do otherwise, we cannot envisage other alternatives other than the alternatives we imagine in the moment.

          1. The term “free will” still has its everyday meaning that a decision or action was made free of coercion by another or under the influence of drugs or other incapacity.

          2. I agree we have this everyday meaning of not having a gun to my head.

            But this is not what the discussion is about.

  8. The research Aaronson refers to at the beginning is probably that of Benjamin Libet. While Libet’s work is interesting, Aaronson’s extension of it into a brain prediction machine is just silly. Basically, Libet found that he could detect what his subjects were going to decide fractions of a second before the subjects reported they made a decision. Somehow people thought this was some kind of groundbreaking result. Instead all it shows is that the decision-making process takes time and that the decision is made by processes not accessible by conscious thought shortly before the decision reaches our consciousness. This has no metaphysical significance, IMHO.

    Aaronson’s Prediction Machine really has nothing to do with Libet’s work and can never exist. Aaronson is right, though, that even if the universe is deterministic, it doesn’t make any difference unless we can predict decisions before they are made. Free will is simply the feeling of agency we have when we make decisions. It is free for all practical purposes.

    1. Libet’s experiment actually convincingly shows how people, if asked to do so, make random decisions (of which the outcome is completely irrelevant). They probably just try to surprise themselves and let their unconsciousness decide instead of themselves.
      A decision about which house to by would not work in this way…

      1. “let their unconsciousness decide instead of themselves”

        In what sense is “their unconscious” separate from “themselves”? In no way that I can see.

        1. Exactly. It’s just another level of “themselves” – it’s always themselves. Is there a problem for the discussion?

          1. It seems to me that there is if on asserts that house purchases operate in a way that is different from other “decisions”.

        2. Certainly it is…people’s sense of identity is their conscious selves; something that comes from the unconscious might as well come from another person.

      2. Some believe that ALL decisions are made by unconscious brain processes and the only role for consciousness is as a sort of storyteller. Personally, I find this interesting but the division between conscious and unconscious processes is likely a complex one. Currently we can’t even define “conscious” and “unconscious” well enough to even have the discussion.

        1. There is very good reason to think that unconscious decision making evolved first, and therefore occurs first, and therefore conscious decisions are an after the fact awareness addition that follows the already unconsciously made decision and is embellished with story telling. It seems to me that unless we have empirical evidence to the contrary, the default assumption, given what we know about biology (and physics), should be that consciousness is after the fact awareness embellished with story telling. Furthermore, insofar we currently have empirical evidence about this question my understanding is that it favors this same perspective, so we also now have a non-default basis for this perspective. However, I am skeptical that this is eithor or, maybe our conscious thinking also influences our unconscious thinking to some extent. But any such feedbacks cannot loop continuously, it must fade out or time out.

          1. But story telling is thinking too. How can we know how much of that is done unconsciously? Until we know more, consciousness is simply how it feels to think.

    2. I agree that the Libet result is very limited. The subjects have no rational reason to pick one time or another to press the button. They in fact are compelled to make a random decision, and therefore leave it to twitches in their nervous system, since no thought process will produce a “good time to act” as as a rational response.

  9. I’ve decided in my scholarship that it’s best to avoid placing too much weight on terms like “free will” and “determinism” because thinkers over the centuries, in their efforts to win arguments with anything they can, have re-defined those terms nearly into semantic oblivion.

    My preference is to focus on the question of which factors could possibly affect behavior and which ones could not. The ones that cannot almost certainly include mental states, such as intentions, beliefs, etc.

    We already have a robustly evidence-based physiological explanation of how behavior is generated, and few would deny that it fully describes the causes of behavior in almost every species on earth that has a brain. The only alleged exception is human beings, whose behavior supposedly can also be caused or influenced by subjective mental states. There is however no evidence, except fallacious inferences from certain common correlations, that such “mental causation” ever actually occurs or even possibly could. (See my website for working paper of article on this topic forthcoming this spring, if you’re interested).

    If mental causation cannot occur, arguments for the possibility of free will are, it seems to me, doomed.

    1. Placing weight on determinism is important when you think about generating or determining future behaviors out of people in order to minimize suffering.

      Engineering controls, regulations, laws can all be implemented based on evidence based observations. We can minimize tendencies toward murder, for example, by improving our educational, political, and cultural systems.

      Using determinism for good. In many ways, it puts morality out of business because, in an ideal world, everyone is determined to do good.

  10. I did not watch the video … Aaronson’s argument is strange.

    Take an atom of uranium 238, we can’t predict when it will fission, only its probability. Does this mean it has free will.

    Strange argument.

  11. “What I care about is whether determinism be true. And I think it is, though of course I can’t prove it. All I can say is that the laws of physics don’t ever seem to be violated, and, as Sean Carroll emphasizes, the physics of everyday life is completely known.”

    Of course as you’ve said on many occasions, any *in*determinism that played a role in everyday life wouldn’t give us more control and responsibility than we already have under determinism (proximate control but not ultimate). So universal determinism needn’t be the case in order to access the compassion and understanding and policy implications that as you rightly say follow from debunking contra-causal free will. So I don’t think we need to worry about whether determinism always holds with respect to human behavior.

    1. What I meant was “physical determinism” in the sense of “our behavior obeys the laws of physics”, not that it is always PREDICTABLY determined in advance. I thought you had grasped that point, but I apologize if I was unclear.

      1. No worries. In the philosophical literature causal determinism is widely understood (to quote Dan Dennett) as the thesis that “There is at any instant exactly one physically possible future [and] this set of worlds has just one member, the actual world, the world in which Austin misses [his put].” (p. 75 of Elbow Room)

        I found this quote in Tufts philosopher Erin Kelly’s great new book The Limits of Blame which you’ll be happy to hear is a refutation of retributivism.

      2. “our behavior obeys the laws of physics”

        So are you saying that we can freely choose between various courses of action as long as none of those choices violates the laws of physics, or that the notion of free choice per se violates the laws of physics? The former seems self-evident; the latter not so much. Please don’t interpret this as “snark”: I’m genuinely trying to understand and don’t think I’m alone in this.

        1. The notion that we could have done otherwise: that we can, in exactly the same circumstance, use or will to make a different choice, is what I meant. I don’t know what you mean by “freely choosing between various courses of action.” And yes, the notion that there is some non-material “will” that can, in identical physical circumstances, lead us to make different choices, does violate the laws of physics. This is what Sean Carroll says, in effect, in The Big Picture, though he’s a compatibilist.

          Most people, when asked, think that in identical physical circumstances, we can have a kind of free will that enables us to make alternative choices (i.e., choose both), and that that kind of free will gives us moral responsibility. In a deterministic world, most people think that we would lack moral responsibility.

          I’m doing my best to explain what seems obvious to me: we are material creatures made of atoms; our behaviors and actions stem from the arrangement of those atoms in our brains, and those atoms must obey the laws of physics. Therefore, our behaviors and actions must obey the laws of physics, and are “deterministic” in that sense. We are, in effect, robots made of meat, with a really sophisticated onboard guidance system. I know many people don’t like that notion, but I think that, given the laws of physics, it’s ineluctable.

          1. Thanks for the reply, Jerry, especially because I know you don’t like doing follow-up. I grok that any choice we make represents the best we can do with what we’ve got to work with, which includes our genes, our temperament, our entire history up till now, etc. etc. etc. In that sense, I agree that we couldn’t have chosen other than we did. This is why the phrase, “If I were you. . .” has never made sense to me: if you were me you’d do exactly what I’m doing.

            So besides free will being subject to the objective laws of physics, it’s also limited by what we might call the subjective laws of personal identity and all that entails. That said, it still comes into play. I can’t choose to play the cello like Yo Yo Ma, but I can choose to play it like yours truly—or I can choose not to play it at all. My freedom of choice may be highly conditional, but it’s freer than whatever freedom the atoms that execute those choices have.

            In short, I’m still confused. Perhaps other determinists can elucidate based on the above as a starting point. Thanks.

          2. I think confusion regarding this subject is a perfectly normal and common state of mind. 🙂

            “My freedom of choice may be highly conditional, but it’s freer than whatever freedom the atoms that execute those choices have.”

            Even in a strictly physical sense, you are absolutely correct. In physics any system is said to have some number of degrees of freedom.

            “Degrees of Freedom: each of a number of independently variable factors affecting the range of states in which a system may exist, in particular any of the directions in which independent motion can occur.”

            Generally speaking the higher the number of elements that make up a system the higher the number of degrees of freedom the system has. And the less accurate predictions of future states become. Compared to a complex, organized blob of matter like a human individual atoms have very little freedom.

          3. An aside – I actually think one can be open to all manner of ‘woo’ and it still doesn’t change much about free will. Even if true randomness does exist and it’s possible one ‘could have done otherwise’ in that sense (in that if you replayed a dice roll five times you would get five different numbers, replay a choice five times and you would get five different choices,) it would be due to randomness, not moral responsibility. Even if a metaphysical substance called ‘will’ existed in people or was piped in from another dimension or however you want to frame it, that still has little to do with an individual being’s personal culpability. It adds a lot of weirdness to the picture, but it doesn’t create a single egoic agent that is in some sense removed from all causes and conditions and is constantly creating its own ‘first cause’ in choosing to want to want to want to want to want a particular thing (to an infinite degree).

          4. The key to appreciating Aaronson’s paper, I think, is his take on Newcomb’s Paradox. That paradox has been a litmus test for this entire question for decades. For myself, I had already reached Aaronson’s conclusion about it years ago and thought he does a fine job of explaining his view from a computer scientist’s perspective, one which is heavily influenced by a computer scientist’s understanding of “algorithm” and “computable.” Beyond that, I have a hunch that as time passes Bell’s Theorem will have an increasing relevance to the resolution, if indeed there even is one.

        2. we can freely choose between various courses of action as long as none of those choices violates the laws of physics

          The freedom in choice that think we have is simply the ignorance of the chemistry/physics that goes into making that particular choice.

  12. I’ve always loved this show! Jerry, please take them up on their invitation. What topics would you discuss? ( I’ve gotten the impression that Kuhn does long interviews on several topics and then chops them up for different shows)

    1. I’m worried, from looking at past shows, that it’s too religiously oriented and obsequious towards faith. On the other hand, there have been some good science segments, like the one with Sean Carroll, so I’m leaning towards doing it.

  13. One can expect that predictability would be problematic in the absence of determinism. But consider the reverse of “cashing out” determinism into predictability. If a person’s actions can be reliably predicted determinism must be true (and contracausal “free will” most likely an illusion).

    Anyone with a close friend or family member or a significant other surely can predict (with high reliability) that person’s response to certain situations, remarks, etc. That ability to predict is the basis for “button-pushing”, passive-aggressive behavior, bullying, cajoling, courtship, and so on. Without determinism, prediction would be much less reliable, and most social interaction as we know it wouldn’t exist.

    1. With high reliability, yes, but not absolutely accurate in all cases. People can sometimes surprise you.

      But I agree, 99% (or maybe 90%) predictability is all that is needed for social interaction to work.


  14. It’s ironic that the more free will you think you have, the easier you are manipulated by exploiting your cognitive errors (f.i. advertising).

    Also, the more (non trivial) choices you make or moral rules you follow, the result will be that you have less freedom.

  15. I think predictability is relevant in the sense that if one’s actions could be shown to be as predictable as say next week’s lunar eclipse, that would end (finally) the debate about whether we have free will.

    1. Certainly that kind of free will. But that would still leave everyday free will, the sense of agency we get from making decisions. Knowing that determinism is pulling the strings behind the scenes doesn’t change much. “Could have made the decision otherwise” is a meaningless statement in a universe where nothing is ever repeated.

      1. Paul, the point of the “could have done otherwise” argument is precisely the supposition that “under the exact same circumstances”,” I could have done differently” (“holed the putt”). A free will beyond causes is its point.
        I don’t know if we live in universe where “nothing is ever repeated”. It seems like a lot of stuff —from bad music to the orbit of the planets — happen repeatedly.
        I guess it depends on how fine toothed you want to get. I think we are designed to ignore a lot of minute changes. A lot in my life seems ‘the same’.

        1. Actually Greg, no that’s not the point of “could have done otherwise”. The point is to separate facts that *depend on me* from those that don’t. If a circumstance *depends on me*, then it’s allowed to vary.

          So if my golf putt was too short, the scenario in which I “could have done otherwise” would be one in which *my arms and club were swinging a little faster* just before I hit the ball. That’s a circumstance that’s allowed to vary, because it depends on me.

  16. Scott Aaronsson is an excellent computer scientist and a sharp analyst. Recently he explained the problem with the work that seemed to say that century old classical quantum mechanics was inconsistent. He also had an enjoyable paper on how computer science may show that time travel is impossible along the line of “where are the time travelers”. Essentially the result was, I think, that a computer with access to time travel would make all problems of the same difficulty class. And that cannot become real physics. (This is akin to how time travel would tear down the stability of light cones in our fundamental physics theories. That cannot happen in real physics as we know it.)

    As for the topic, I will grant the social uses but not the means since “free will” seems to be a religious invention adopted by philosophers (or maybe the other way around). We should not encourage people to take revenge, including revenge on criminals, since it may feed criminality at the very least. And we should encourage empathy since our species is (but perhaps mostly because recent decades have seen this strategy successful in economics). With that I can paraphrase Scott:

    My view is that I don’t care about philosophy if it can’t be cashed out into actual predictability.

  17. I agree that Aaronson misses the key point about freewill. Prediction by a machine is a side issue. His head has been in the machine too long.

  18. Contra-causal free will is an imaginary solution to an imaginary problem. The “problem” of determinism (not determinism itself, but the alleged problem it poses for human action) comes from projecting human intuitions about causality onto physics, where they often don’t belong.

    I didn’t read Aaronson’s long paper, but I did read a companion piece on his blog. There, he stipulates a definition of free will that relates to predictability. That is an unfortunate choice of term, in my opinion, to label his stipulated concept. In his defense, physicists sometimes use free will in a closely related way, to talk about the ability of scientists to impose a condition that has no relation to the past variables of the system being studied. E.g., “we will measure polarization along THIS axis! Hah!” If the system in question is inclusive enough, that equates to scientists behaving unpredictably. But then, that too was an unfortunate labeling. Independence would be a better word for that.

  19. Aaronson has a computing background. The thing about a computer program is that the outcome is, in theory, absolutely deterministic. In practice, modern computers and programs are so complex that unpredicted results can easily occur. (Hackers notoriously take advantage of this).

    This is quite aside from internal timing errors or race conditions, or even gamma rays flipping bits.

    Much computing effort is devoted to avoiding this or correcting for it.

    There’s an obvious parallel to Real Life (TM).

    I tend to side with Aaronson. I think determinism is correct, that Free Will is probably an illusion, but for practical purposes we might as well assume it operates. This doesn’t preclude me from having some sympathy – not unlimited – for individuals who mess up. But that’s based more on a feeling of “There, but for the grace of D*g, go I” than any conviction that they were pre-programmed to offend.

    Unlike PCC, I can’t say that determinism has had any influence on my thinking (other than giving me a headache).


    1. This is a good way to put it and the computer program analogy is a good one. Ability to predict seems like a crucial component here. There’s a big difference between a process that is predictable in principle and one that is predictable in practice.

      “… but for practical purposes we might as well assume it operates.”

      I would go further to claim that we can’t do otherwise. As has been pointed out by many, even with determinism and without Free Will, we are still try to argue our points in order to change minds.

    2. “For practical purposes”, that’s a good idea and the practical purpose is an epistemic one. If we want to learn more, we follow and believe in good Reasons. If only more criminals were adequately socialized and mentally healthy enough to think about Reasons for their behavior, maybe they wouldn’t do it. Oops, that sounds like Free Will.
      It sounds like computers who are programmed to play chess and write music (they are given a general sense of it) and in one way that is deterministic but in another they beat grand masters and write original music. They get the general form and then try variations—like evolution does.
      Free will is not an illusion, it is an everyday necessity; the more illusory thing is our intellectual understanding that all things are determined.

  20. Sorry I’m late to this. I follow this blog in a desultory way. I don’t usually pay attention to arguments about free will because as a lifelong determinist (& atheist) I agree with Marx’s comment that to be an atheist is to still buy in to the God argument. I’m generally just not interested.

    I am, however, surprised that Godel doesn’t appear in the discussion. I seem to recall (from 30+ years ago) that there were pretty successful arguments that determinism doesn’t lead to predictability because the system is too complicated…

    However, I’ve always been struck with Robert Paul Wolff’s reading of Kant’s “Groundwork of the Metaphysics of Morals,” which argues that the only way to reconcile Kant’s determinist epistomology & his morality is that the universe had to be created from the beginning to take everyone’s (all & forever) willful decisions into it in creating a determinist system.

      1. Of course the “oracle” will take everything into account. But one can simply program another, smaller machine to turn a light bulb on or off, precisely in the opposite way that the “prediction panel” of the oracle says it will.

        The oracle can take all the variables into account it wants to, but it still cannot change the fact that the smaller machine will do the opposite of whatever it says.

        Notice that this works precisely BECAUSE the Universe we considered is deterministic, and the laws of physics are never broken (this frustration thought experiment works if you generalize it to probabilities to, of course).

    1. Scott Aaronson supposes that the machine puts its predictions in a sealed envelope, and you don’t get to see them until later.

      I think that limits the interest of the prediction. If the prediction could be *shown to the person in question* and still always be correct, *that* might show something about free will. This, not so much.

      1. Right, or if the machine never told anyone the prediction, or if the machine somehow existed outside of our universe (almost like a deity, haha), etc. etc. etc.

        This shows that the machine cannot possibly print the prediction before the action happens – otherwise, one could “intercept” the sealed envelope and give it to the frustrator. So the oracle is either useless, inexistent, imaginary, or abstract.

        And this thought experiment works in any deterministic universe, or in our quantum-mechanical world.

    2. Thanks for the interesting link. This demonstrates something I’ve mentioned in other comments. Events in the universe, including decisions we make, can’t ever be repeated. When considering the question, “Could I have done otherwise?”, the answer under determinism is “no”. However, it means less than most people might assume as a second similar decision will not be identical to the first as conditions will have changed. In the video, the first event at 11:55 affects the second event at 12:00, making this explicit.

  21. I disagree with the idea that a deterministic free will that is sufficiently obscured by unknown variables is more or less not worth considering. The same could be said of many topics (neuroscience, physics, etc.,) and if the sacred cow of free will wasn’t involved, it’s not an argument anyone would think to make. (Essentially saying something like – “When you get right down to it, we’ll never be capable of understanding the brain fully, so why bother knowing anything about it at all?”, or “I don’t know if a meteorite will hit my house tomorrow, so since I can’t account for every possible variable, why plan at all?” for example. Neither of those intuitively make sense to us, at least to my mind.) What we *do know about how the human mind and will works, in terms of causality, is at the root of pretty much any psychology, (obnoxious buzzword trigger warning) ‘mind hack’, or ability to reverse engineer ourselves that our culture has developed. All of those presuppose the idea that ‘will’ is deterministic and that the causality of choices can be studied with at least a reasonable chance of finding predictable patterns.

    Of course it doesn’t have to be an all-or-nothing situation. One can think will is to some degree deterministic and to some degree metaphysical or random. But to the *degree that one sees will as deterministic, then to that same degree the causal chains behind it are worth knowing about if we want to understand how to make better decisions and choices.

  22. As a poor stand-in for WEIT commenter Ben Goren, MIA (should anyone phone 911?)
    there is a reference to the person, Alan Turing,
    who is regarded by many as the father of computer science.
    If you pop his name and “halting problem” into your fave search engine it will point you to his paper in 1936 where he showed that in principle you cannot have a program that accepts two parameters (a program to check and its inputs) and decide if that program will halt or not.
    NB this runs what we now call a Turing Machine. This is a theoreitcal machine and not subject to halting on the heat death of this universe or even running out of memory.

    It would appear that his result means that even in a deterministic universe that not all things can be determined in advance?
    But, throw in some quantum woo, synaptic vesicles etc and BINGO!

  23. Many of the arguments in this thread refer to the “laws of physics”. The unstated assumption is that they are deterministic, and thus the future is predictable. This is not the case.

    “Anyone not shocked by quantum mechanics has not yet understood it.” ~ Niels Bohr

    “Nobody understands quantum mechanics.” ~ Richard Feynman

    “[T]he atoms or elementary particles themselves are not real; they form a world of potentialities or possibilities rather than one of things or facts.” ~ Werner Heisenberg

    “The doctrine that the world is made up of objects whose existence is independent of human consciousness turns out to be in conflict with quantum mechanics and with facts established by experiment.” ~ Bernard d’Espagnat

    “In the beginning there were only probabilities. The universe could only come into existence if someone observed it. It does not matter that the observers turned up several billion years later. The universe exists because we are aware of it.” ~ Martin Rees

    Confused? You’re not the only one. There are 14 common interpretations of the “quantum enigma”.


    All of them deal, in one way or another, with the observer problem. Consensus or conclusions? Nope.

    The following is a video of a Richard Feynman lecture. It is cued to his declaration that quantum mechanics is “mysterious”. There’s about 3 minutes of really interesting stuff here.

    1. I think most of the arguments argue that the laws of physics are simply lawlike, and that the molecules in the brain obey them, whether they be deterministic or probabilistic. In both cases, though, there is no room for contracausal free will.

    2. There may be Laws of Physics but I don’t think we have any idea of what they are. There are “previously accepted Laws of Physics” (Newton’s Laws) some of which have been superseded by “currently accepted Laws of Physics” (General Relativity) that are likely to be superseded by “future accepted Laws of Physics” perhaps infinitum.

      Will we ever have more than “Currently accepted approximations to Real Laws of Physics”? How will we know?

      1. I wouldn’t say we don’t have any idea of what they are. The accepted laws may not be ultimate, but they are a refinement of previously accepted laws.

        1. Yes, the new laws are refinements of the old laws but only in terms of predictions they make. New laws can completely change our view of the universe. The classic example is relativity. It didn’t change predictions much from those made with classical mechanics but it radically changed how we view the universe. I could easily imagine some future physics revelation changing how we view determinism and, perhaps, Free Will of the kind that relates to it.

  24. It’s an absurd premise, being the universe is NOT predictable under any circumstances. Even systems on the scale of a human can be non-deterministic, and are, if you look at them with enough resolution. Quantum mechanics is real and the current science supports that it yields non-deterministic results that can be leveraged into randomness on the macroscopic level.

    In terms of computer science, a feature of digital computing is that your logic levels are robust enough that you can disregard random quantum effects and get repeatable and predictable outcomes. While it is possible that a 1 can turn into a 0 randomly it is so improbable we never expect to see it. In contrast there is no theory supporting that the human brain will not change course due to one photon of energy or one molecule that randomly either reacts or does not react. This does not prove free will, but it excludes the use of determinism for negating it.

    1. As I’ve said repeatedly in my posts on free will (and in this post too, I think), pure quantum indeterminacy, which does lead to a lack of unpredictability, does not support what most people think of as free will. By “determinism” (as I said in the comments), I mean “obeys the laws of physics.” Insofar as the laws of physics apply to the human brain, there is no free will in the contracausal sense.

      Finally, we do not know whether quantum indeterminacy has any effect on human decisions and behavior. In fact I think there’s at least one paper suggesting it does not. You are assuming as the null hypothesis that a photon can change human behavior. Don’t you need a theory to support that? Even so, quantum indeterminancy does not support the existence of contracausal free will.

      1. Sorry, my statement was ambiguous, I was not implying that your article was supporting contracausal free will, just clarifying that my response to it contra Aaronson was not doing that either.

        The question of free will itself it a different one entirely. The Calvinist theologians addressed it exhaustively and came to the conclusion that it has no bearing on criminal justice, because if the criminal does not have free will, neither does his judge or executioner. Calvin of course had a different name for his “Big Machine,” knower of all outcomes in the universe, and came to remarkably similar conclusions on the non-existence of free will. But the Reformed theologians on both sides of the free will debate ultimately agreed that even if the answer to free will is unknowable, there is no benefit to anyone in living as though we don’t have it.

    2. I think I agree with your penultimate sentence but the double negative makes me uncertain. Are you saying that it is possible that the human brain can change course due to one photon of energy, etc? If so, then I agree. Given the nature of stimuli that the brain has to work with, it seems obvious to me. If we present an object to a subject for which the color varies continuously between red and green, there must be a color setting that is right on the edge between triggering a “red” response vs a “green” one.

      Even if the brain is mostly digital, the senses are not. Also, as I pointed out earlier, even if we imagine the brain is digital, it would be subject to the Halting Problem, meaning no other process could predict its behavior in all cases.

      1. Paul, I strongly agree with you. In past arguments about contra-causal free will, this kind of quantum uncertainty didn’t matter, but in Aaronson’s formulation, it does matter, and the answer to his question is unambiguously “no”, there can never (even in principle) be a machine that can predict human behavior over long time scales.

  25. Hi Jerry. I almost recommended “The Ghost in the Quantum Turing Machine” to you a few years back after a blog post on free will. You should read it, there is lots of very thought provoking stuff in there! Also you should read Scott’s blog, you’ll learn a lot 🙂

  26. My compliments for your Oped. It makes a good and interesting read. Yet I would propose we take a different approach to the issue. I think science shows “free will” and “determinism” are not the core issues. I think the core issue is “individual responsibility” and the assumption we function as mature and concious individuals.

    The debate now seems to focus on the wuestion of determinism vs free will. And about what we not know (due to quantum probability). The computer metaphore misses the social aspect of human behaviour.
    ANd although I agree there is a relation between free will and crime. I think we should not ignore that modern day society has become far more demanding than that. And I hope I can show that many of us are not able live up to our own expectations that most of the time we act as mature and responsible adults.

    Biology and social science have shown that human behaviour is for more than 95% unconcious behaviour.
    In addition our behaviour is very much determined by our position and relations to our social group. Some suggest most of our behaviour is directed by “social” factors as status, certainty (control), autonomy, relatedness and fairness (if in doubt google “SCARF” behaviour).
    When we feel we gain on these factors we feel reassured and safe and are capable of taking our time and think things over. If we experience or fear any loss on these factors we tend to flee these issues or attack the messenger who confronts us. Our concious thought follows this instinct (cognitive consonance).
    So we should not compare our behaviour with an individual computer program. A better comparison would be with a computer that is programmed that seeks constant confirmation of his team mates that he is doing nicely.

    Furthermore we know that adults will fail to take responsibilty for about 25-40% of “civic duties” modern day society expects from them. This 25-40% includes people the period we suffer from mental issues or old age. But also lack basic skills and routines, eg discipline, litteracy etc, etc. This is underlined by a recent study in the NLs. It showed that most people entering debt restructuring programs get into trouble because they lack the basic skills to full fill the administrative duties government imposes on them. One of the (suprising) findings was that about 10% of those people had at least a college education and had very responsible and hi paid jobs in which they were considered to be highly efficient, but still they lacked the skills to manage their own ‘private’ duties. But even those “responsible” people rationalized their behaviour in fleeing or attacking their basic administrative duties.

    The issue of free will is therefore is not a question of physics. The discussion showes that issue is still debatable and its consequences are only theoretically relavant and have no practical significance. The real issue is that we should question our (western) assumption that most of us are fully competent adults (individuals) who are free to choose our behaviour. We should recognize most of us have our “dependencies” from time to time. We are only able to function thanks to other people helping and/or reminding us. This “social safety net” of family and friends is even in welfare states the best indicator of social success. Or if we lack it, of our troubles.

    1. Those are interesting thoughts, although I don’t agree with your conclusion that the discussion of free will “has no practical significance”. (I do sometimes say trying to talk about it can, in purely pragmatic terms, hurt more than it helps in that people seem to really misunderstand the idea – but that doesn’t mean it doesn’t *have practical significance, just that you have to be very careful in that people can often understand that significance in ways that are incorrect or even harmful. “Why bother trying, I’m just a robot?”, etc. Erroneous understandings of a concept can be worse than no exposure at all, to my mind.) *Any explanation one gives that invokes predictable cause-and-effect – whether that is social, physical, etc. – *does speak to the importance of determinism in will. Even if our will was 100% determined by social factors, those social factors would still presumably boil down to physics if you could look at every neuron and neurotransmitter involved. Or, on the contrary, if someone was 100% able to resist any temptation to stay on the perfect diet, follow the perfect financial plan, and so on, that would also happen as the result of factors that said individual didn’t directly choose, via a long chain of cause-and-effect.

      I do agree that I think the amount of emphasis we put on encouraging and rewarding the role of individual will (not *free will, but will,) vs. viewing will as a constellation of factors is important. On the one hand, of course some percentage of will has to be considered individual, because if people weren’t having personal subjective experiences it would make no sense to give them prescriptives in the first place. On the other, any time we try to reverse engineer something, we start by understanding how it works and then seeing where in the chain of cause-and-effect change is possible, and in human thinking, I think we find much of that chain of cause-and-effect *outside the individual. (Even for feelings that are purely subjective, any external advice on how to best deal with them – the temptation to eat a chocolate bar while on a diet, for example – will by definition come from external sources, not individual ones.)

      I think some degree of focus on individuals is important in that individual autonomy and human rights should have a certain sanctity in liberal democracies. That said, I think the flip side of the coin is understanding that the creation of each individual depends upon a conglomeration of forces, and creating societies that produce happier individuals means understanding that web of elements. (None of that, again, speaks to ‘free’ will, but it does speak to ‘individual’ will. We can appreciate that there is something unique about individual consciousness even if it is not self-created.)

      1. @Roo
        I do agree on most your saying and most of all your last paragraph. And yes the individual responsibilty remains in center even in a more social context.
        My point was to tip the balance from purely individual and pure physics to a more balanced approach that honours the social context a bit more. I am glad you understand that.
        If I talk about practical significance that is in contrast to what I would call academic significance. (And when I mean academic I mean that litteraly, it is significant for research).
        I disagree with the idea our behaviour can be understood by understanding the physics or chemistry that runs our neurons. These processes are required and facilitate our program, just like silicium and electricity facilitate how program runs. But my educated opinion as a molecular biologist is they do not determine the “machine language” the individual uses to run its processes, assess its surroundings and react on that.
        I think the basic feedback mechanism that facilitate our abilities and learning are already known. Although we do not fully understand the “machine language” or information architecture life uses. Recently for instance it was discovered that our intestinals (gut bacteria) have an important influence at our mental state. And I do hope we are up for some more suprises along the way.

        1. I think the idea of whether or not pure physics maps on to large scale social interactions gets into the concept of emergence. I think this is a possibility, although emergent properties would then, *themselves, be dependent on causes and conditions that could eventually be traced back to physics. So while I think the idea of emergence adds an interesting layer philosophically, it seems to me that the important factor is whether or not one believes a process is causal or not. (Strangely, I think this actually true of both people who do and don’t believe in free will. Even groups who believe in free will believe strongly in causality in pretty much all *other realms surrounding said wills. Religion, for example, is generally not meant to be a pell mell celebration of quantum randomness, quite the opposite, it rests almost entirely *on the idea of cause and effect. You need to do X if you want Y to happen. I think the only philosophies that really eschew the idea of determinism all together are those that involve deep ‘it’s all beyond words’ mysticism. If some other force that ‘makes things happen’ in the universe exists and it is NOT based on either causality or randomness, then I don’t actually think our human minds were made to comprehend such a thing. In every way we are conceivable of understanding, events happens because of other events. To say they happen ‘freely’ is pretty much semantically incoherent, to my mind.)

          1. Yes, I am talking about “emergent intelligence”. I am not necessarily speaking of free will.Although I am not religious I found the ideas Paul Davies put forward on these matters interesting.
            But unlike him I do not believe the laws of physics determine we end up with intelligent life. But can not deny that a univere without life would never be noticed.
            To me it is comparable with evolution. Evolution is not deterministic. It has no goal. I think the way we learn resembles that selection process a lot.
            Still one could argue that evolution in hindsight leads to more complexity. That is correct but does not imply the proces is deterministic, only probabalistic which is (to me) very different. But I agree probabalistic does not imply free as well. In my view the choice is nor deterministic nor free, but the ‘excluded third’ probalistic and emergent.
            (I mean this option was excluded in all theological debate about free will and predestination.).

          2. (Random fun fact after a quick Google search – there are actually two Paul Davies who appear on the first search page talking about free will. I assume you were talking about the professor in California, there is another at William and Mary.)

            I agree that the development of self-consciousness is an undeniable (in that it’s self-evident) phenomenon that really does escape any real scrutiny at this point. (Whether it’s largely probabilistic or not I’m agnostic on – I’m not sure that this would make much difference other than making people more or less consistent / reliable actors. It seems split-second adaptivity would be more the variable that makes our will appear ‘free’. You can add the tiniest sliver of new information [that ‘candy’ bunny is made of wax!] and no how firmly resolute a previous decision was, we can still change course mid-action.)

  27. SA I believe works in computational complexity theory, which *just maybe* can help to tell what sort of resources are needed to do the aforementioned predictions, which might be interesting.

    That said, the continual replacement of epistemology for metaphysics is a habit that many scientists and technologists are prone to – Jerry seems immune here, which is good.

  28. This discussion is like an elaborate joke. Let’s say someone can predict exactly what one person can do. Life might get hard for that one person, but what chance is there that someone will have the computer firepower to model everyone, or even just a significant minority. And if they do, who is going to quit making decisions, or whatever decisions will be renamed.

    I think one of the most amazing things about humans is the ability to fake anything. If that’s true, determinism has a high hurdle.

    In short, proving deterministic behavior won’t change anything.

  29. He is asking the wrong question. The real question regarding free will/ determinism is if you had a machine that could send signals backward in time, would you know the future? Or would you only know possible outcomes and not a definite future? If there are multiple futures, then the universe is not deterministic. However, it still could be random and not allow free will. If given knowledge of possible futures by cross-temporal communication, you could influence the trajectory of history, then you have free will.

  30. It seems obvious that there is no such thing as free will. Every effect has a cause. All causes are themselves effects of prior causes.

    That is unless one’s consciousness spans some dimension where cause and effect aren’t necessarily linked. It seems like a waste of time to contemplate that though.

  31. I have posted this comment on previous occasions when this topic came up: how can you know anything without free will? Apparently Lequier put it (in the 19th century) as “I seek a first truth, therefore I am free. Freedom is the first truth I sought, since the search for knowledge implies freedom, the positive condition of the search.” That sounds a little abstruse, but animals and humans always try to maximise their freedom to vary their actions, given their opportunities. One cannot know things reliably unless one can freely (and ideally randomly) test them – see R.A. Fisher’s justification of randomization that it gives you knowledge even if everything else in the experiment has been rigged to deceive you.

  32. Another thought – sometimes predictability is a *good* thing. When I meet a friend for lunch, I want to know reasonably certainly that she will not show up 2 hours late, blow my head off with a shotgun, etc.

    1. Yes. I want to be predictable to my friends, and unpredictable to my enemies. Not that predictability is the same as control, or unfree will. I’d rather not be predictable to enemies, but that’s better than being under their control.

  33. I finally read the Aaronson paper Jerry cited in the Update to this post. It was well worth it. In spite of Aaronson’s over-emphasis on prediction, what he got right was far more important than where he went astray. The extremely right part comes in section 2.9, where he rejects step (ii) of a well-known incompatibilist argument, and insists on the distinction between microphysical facts and macro-facts.

    Since Aaronson is an expert on entropy, it’s not surprising that he noticed that irreversibility is associated with entropy, and thus with macroscopic facts. But it’s still a smart move.

    1. I’m also starting in on Aaronson’s paper. I am gratified to see that he will distinguish between a brain-like simulation that can pass the Turing Test and a simulation of a particular brain, a point I made here as well. Judging from the TOC, it should be an interesting read.

  34. I wish that one day someone who makes the claim “You have no free will” would provide a suitable definition of the term “free will” and then provide a reason why we dnn’t have it.

    Just saying ‘because we can’t escape the laws of physics’ is clearly not enough. How would ‘escaping the laws of physics’ (whatever that means) give us free will (whatever that means).

    And what does it have to do with social issues? Taking the retributive aspect out of justice is a good idea in any case. Tying it to some obscure metaphysical point only confuses the issue.

    1. You don’t seem to have noticed that I provided my definition of free will IN THE POST. I’m sorry to say that you haven’t even read enough on this site to credit me for what I have said (like describing it connection to social issues), much less thought clearly enough of the issue to articulate your own questions. I wish that some day people who criticize my posts, as you do here, would read the damn things before they shot their bazoo off.

  35. “Would one be able to construct a machine that, if it was programmed with the immense knowledge about your brain and your environment that would be required to do such a job, would be able to predict your actions?”

    – Machine, when and where exactly am I going to die?

    The only way for that machine to do such a job is to kill you instantly. Otherwise, all you have to do is not to be in a certain place.

Leave a Comment

Your email address will not be published. Required fields are marked *