Stephen Wolfram (and I) on free will

July 7, 2012 • 3:00 am

The European has an interview with Stephen Wolfram, (Many of you will know Wolfram as the designer of the famous computer program Mathematica.) The interview is mostly about progress in  computer technology and its social benefits, but there’s one question about free will.

The European: This equation of nature and technology raises a few very hard questions about the definition of life, about free will, and about intelligence. What are the mathematician’s answers to those questions?
Wolfram: Let’s talk about free will first. The problem with human free will is that deterministic stuff is going on underneath, like the chemical processes in our brain, but that we don’t seem to act in a deterministic way. People used to think that deterministic processes must result in deterministic behavior, and that belief has underpinned much of the debate about free will. It’s the reason why the science fiction robots of the 1950s often speak very logically and behave very stupidly. The main scientific discovery is that it must not be like that. We can have simple deterministic underpinnings that result in very complex and seemingly random behavior.

Computational irreducibility is a key feature of life: We cannot grasp life through a formula, but must really simulate and observe it to see what happens. That’s how we as humans end up freeing ourselves from the deterministic rules. I tend to think that the concept of computational irreducibility is probably the answer to the philosophical debates of the past two thousand years about the relationship between free will and determinism. Philosophy is always at a certain distance to human behavior, so a lot of questions really get answered by science.

This answer seems to me misguided, because deterministic underpinning that can result in “unpredictable” behavior doesn’t bear on the classical issue of free will. First of all, “deterministic underpinnings” mean that behavior (including “choice”) is still determined, even if “complex and seemingly random.”  After all, chaos theory, which is purely deterministic, yields systems whose outcomes are unpredictable if you don’t know the exact starting point. But chaos in our brain doesn’t give us free will.

As for this sentence,

I tend to think that the concept of computational irreducibility is probably the answer to the philosophical debates of the past two thousand years about the relationship between free will and determinism.

I simply don’t understand it.  It seems to mean that if we can’t reduce behavior to predictive equations, we don’t have free will.  But that doesn’t bear on the question of whether all our behavior is determined purely by our genes and our environments, and that we have no real choices.

Now my definition of “real choice”, i.e., free will, has always been this:  if we could rerun the tape of life to the moment when we made a decision, with every particle in the universe in the same place at that moment, and yet we could have made a choice different from the one we did, then that is free will.

But, as some readers have pointed out, if quantum indeterminacy obtains, then perhaps rerunning the tape could yield a different decision, and that contradicts my assertion that I don’t accept quantum effects, which are random, as part of free will.

I then modified my definition to include “if you rerun the tape of life and could have consciously made a different decision at that moment, then you have free will.” But that doesn’t work either, for quantum effects might manifest themselves by making your decision seem conscious—especially if, as some experiments show, a conscious decision is simply how our brain perceives a decision that was really made unconsciously.

So for me (and I realize this isn’t true for others), whether we have free will or not boils down to whether quantum indeterminacy holds at the level of the brain and human behavior. Some people say yes; others say no. This is a matter for empirical study. But whether or not those effects exist won’t affect the conclusions of philosophical compatibilists: folks like Dan Dennett who claim that pure determinism is compatible with free will. To compatibilists, free will is completely independent of the issue of physical “causation.”

101 thoughts on “Stephen Wolfram (and I) on free will

  1. When talking about free will with people, I find that a useful gauge for determining how they perceive it (as your last few paragraphs show, defining it is a difficult thing to do) is to ask whether or not they think a dog has free will. If they don’t, but think that we do, then the discussion moves to asking them what they think we have that makes us different. If they think dogs have free will as well as us then I think it’s really just down to semantics (as few people tend to think there are any canine ghosts in their furry machines!)

  2. “It seems to mean that if we can’t reduce behavior to predictive equations, we don’t have free will.”

    The other way around. It means if we can’t get at the underlying model, for all practical reasons we have free will (as in, we’re responsible for our actions) whether we actually do or not.

    A stance I agree with. We don’t have freedom from determinism, and if “replayed” we would make identical decisions (barring a few quantum errors), but in practice that makes zero difference.

    We’re still responsible for our actions and when we act outside of moral and legal bounds we should be subject to corrective action. Punishment and vengeance are still barbary.

    In short, that we’re deterministic as noumenon is immaterial in practice.


    1. That makes no sense. If everything is determined, how can we be responsible for our actions. The point is that we need to assume free will exists (even though it doesn’t) so that we can hold people responsible. Otherwise the law of the jungle will apply.

      1. There’s an important difference between holding people responsible, and enacting accountability. It does not follow that no free will means the law of the jungle – in fact, a widespread acknowledgement of no free will could (given the correct, non-anarchic interpretations) leads to a more compassionate judicial system. We would realise that criminals are not evil scum, but malfunctioning machines. Voters and politicians would still recognise the necessity of dealing with people who don’t play by the rules, but our attitude would be to either fix or isolate the broken machines. There is no law of the jungle here.

  3. Jerry, now, should we say determined volition instead of free will, then might we say that we make choices limited by our determinants that don’t coerce us but neither do we have that random free will.My determinants made me seed therapy and thereby, new determinants arose to make me better off. As we live both nature and nurture change, so that we might become quite different
    I cannot see my next thoughts coming in detail.Yes,that unconscious acts before I’m aware.
    So,how does all that interplay so that maybe I could or not could have chose differently? Can it be that we do have choices along the way that determine that finally we really don’t have a choice as when criminals start being criminals? Some people say that at the last minute that they changed their minds about something. How did their determinants change for that to happen?
    The main point touches moral responsibility.
    So, I lean towards compatibilism but not the one stating we have free will. Is that more than terminology?
    WEIT, thanks. And I wonder have I posted so ornately that it’s hard to fathom me?
    We have intent, but no divine intent appears, so how can God have free will to cause stuff? The Coyne-Mayr-Lamberth argument stems from Mayr’s use of the term teleonomy and Coyne’s magnificent refutation of teleology in that should be famous essay and Lamberth’s reading of what Paul B.Weisz maintains in “The Science of Biology.” Backwards causation, putting the future before the past, putting the effect before the cause,thereby negating time so that scientists never could do experiments as all would be pre-planned for the same outcomes as Weisz would probably say would negate science itself. Thales of Miletus and Strato of Lampsacus are ever right to note for no one to use teleology regarding science unlike Aristotle whilst otherwise a leader of naturalism, here set science backwards with teleology.
    Jerry, thanks for setting the record straight on free will and teleology!
    http://naturalistgriggsy.wordpress invites WEIT and others to post there,perhaps correcting my mistakes, if any.
    I’m recommending WEIT on my blogs.

  4. Predictability proves absence of free will. Wolfram seems to make the elementary logical mistake to posit that therefore unpredictabilty proves free will.

    P -> not(F) does not imply not(P) -> F.

    1. Does predictability disprove free will? I’ve seen this assertion used to answer the question of why God didn’t provide us with more conclusive proof of Christianity. If it didn’t take any faith to believe in Christianity, then everyone would be a Christian, so people wouldn’t have free will. But it seems to me that simply because everyone makes the same choice, that doesn’t mean they don’t have free will. If you give one man the choice between being with a hot woman or being mauled to death by a tiger, and you give another man the choice between two doors, one of which leads to a hot woman and one of which leads to a tiger, and don’t tell him which is which, which one of them has a more predictable choice? Which one of them is more clearly exercising “free will”?

      In fact, free will requires some level of determinism, because if we have no way of predicting the consequences of our action, then there is no meaningful sense in which we are making a “choice”. “Choose between rolling random die one or random die two” is not a choice.

      1. Some actions are clearly more predictable than others. If I could predict all your actions with perfect accuracy, would you still maintain that you have free will?

        1. So, to be clear, what you meant to say was “Universal, perfect predictability disproves free will”? Also, I don’t understand why you said “still” maintain that I have free will, since I don’t recall saying that I do. But if you could predict everything I’m going toe do, then you would know better than I do whether I will claim that I have free will, wouldn’t you?

  5. I just love the paradoxical nature of free will, and in particular, trying to prove that we don’t actually have free will. Are those who argue that we don’t have free will exercising free will by making the argument, or do they have no choice?

    And then, if they have no choice then what are the implications about the past and the future? Although a complex and incalculable interaction, does it imply that the whole history of humanity – including wars and the like – is already determined?

    Since we are all subjects of the experiment (of whether we have free will) then we cannot observe ourselves objectively and so I guess the question remains and always will remain unanswerable?

    1. I don’t think *proving* the absence of free-will is really claimed. It’s more that, given our understanding of determinism+undeterminism there is nothing left that explains exactly what free-will could be, in the traditional sense. It’s more a case of a challenge to those that assume free-will to explain its mechanism.

      Responsibility (and blame, culpability, etc.) for an action is an association, a relation, we make between an entity that, as a whole, is the most localised cause of the action; the collective focul point of the all the deterministic (and non-deterministic) causes that resulted in the entity performing the action.

      If a rock rolls down a hill and lands on my foot, then the rock ‘did it’, but because I don’t construct a strong relation between the rock’s motion (and the cuases of its motion) an the pain it caused I don’t construct a lasting attribution of blame – though momentarily I may well curse the rock (and not, incidentally, myself, for cuasing myself to be in the rock’s path).

      Similarly, we blame criminals for their acts, because they are the most localised focused causes of the acts they perform. Even if we can identify serious social precursors: abused as children, abandoned by the state and society as a whole, etc. The difficulty of constructing personalised rules of attribution of blame, by taking into account all precursors to a persons acts, means we use broad rules.

      A bit like speed limits, which we set in bands: 30, 40, 55, etc. Clearly not every road requires every car, of whatever kind, and every driver, of whatever skill, and every stretch of such a road, or whatever condition, to stick to these limits in order to achieve a particular level of safety.

      So with morality. We have broad rules within which we collectively agree to, even if we vary individually in our ascent to those rules. And when someone breaks those rules (or in non-free-will language, when an entity is caused to breach those rules, from whatever collective and historical causes) then we identify that individual as th cause of the offending action, and only then start to consider mitigating excuses.

      Responsibility, culpability, encarceration for the protection of others, are all reasonable under a non-free-will understanding of human behaviour. What is harder to justify, under non-free-will, is retribution – other than as an understanding of the psychology of the victim.

      As to the past and future: The past is what is is, but what we can discover of it is another matter, which has its own elements of epistemological indeterminism due to our lack of access to complete information. The future is similarly epistemolgically indeterminate, whether the underlying reality is determinate or indeterminate – the consequence being that, even if we feel intuitively that our futures are determined precisely, or are randomly variable, we can’t get at them. So our view of our futures, as being freely willed or not, is indistinguishable. So, I agree with part of your last point: “we cannot observe ourselves objectively”. I’m not sure it will always remain unanswerable, though there seems to be a limit to the extent to which any system can have complete knowledge of itself.

      Clearly, if this view is right, then I too have no realy ‘choice’ in coming to it. In this matter my views are a product of what I’ve learned, who I’ve interacted with.

      1. So are you saying that within the realm of our knowledge of the properties of the physical universe we cannot find anything that allows that free will to occur, and therefore if someone is to claim that we have free will they need to demonstrate the mechanism? If they cannot, then we have to conclude that free will is a delusion?

      2. I must get the hang of this, my response below was only to the first paragraph, which was displayed in the email…
        I try read your discussion through the lens of you having no choice in writing it, and it seems to me (a delusion) that it becomes meaningless. There is no opinion, because you have no choice in what you write. Your agreement on the question of objectivity of our observation is not a considered opinion because you have no option but to write it – as you say in your last paragraph. And therefore I suggest there is really also no ‘you’ – the ‘you’ that is being deluded cannot be deluded as it has no choice in accepting or otherwise what is being experienced as real or no. (I don’t think I’ve put the last sentence very well – I must spend some time later trying to clarify…)
        Hence the wonderful paradox.

  6. I think Wolfram’s answer isn’t complete, but it helps us understand the actual question we should be asking: why do we not behave like Wolfram’s science fiction robots: speaking logically and behaving stupidly (and, per Dennett, “sphexishly”)?

    The philosophical question that you ask is, as I’ve noted before, not scientific. There’s no way at all we can use evidence or observation to determine that we live in a world in which we could or could not “rewind the tape” and make a different decision; we cannot ever actually rewind the tape and observe whether the same or different decisions obtained.

    Randomness doesn’t help, you cannot ever prove randomness; indeed, randomness isn’t even definable. Furthermore, randomness does not address the question in a philosophically satisfying way, especially regarding moral accountability: we are no more philosophically “accountable” if our decisions are the outcome of truly random processes than we are if they the outcome of rigidly deterministic processes.

    Wolfram asks the scientific question: what is it that we actually observe about human behavior that needs to be explained? What needs to be explained is not the presence or absence of “free will” (since we can’t observe its presence or absence), and the “free will” or its lack cannot be an explanation without something observable that needs to be explained.

    Philosophy is, I think, much like theology: divorced from any method that differentiates between true and false. Logical validity (apparently difficult, from the appalling frequency of professional philosophers who make elementary mistakes of logic) is not enough, and the kind of “intuition” philosophers often appeal to to justify their premises is at best a backhanded and awkward attempt at experimental linguistics and psychology and at worse no better than “divine revelation.”

    The best you can say, I think, is that the question that you personally pose is, fundamentally, not interesting. It directly references nothing that we can observe, and there is nothing that we can observe that would decide it. At best, the definition can serve as a “springboard” to a discussion in the humanities about what individuals think about personality, morality, accountability, and the role of judgment, praise, blame, reward, and punishment in society and culture. But as a matter of truth, it’s a non-starter.

  7. This kind of thing makes my brain hurt. Now, I’m pretty sure people who believe in libertarian free will are whacked. The real contest seems to be between the compatibilists and the determinists. When I read through the debates they seem to be using the same words in different ways. My perception is that they don’t really see things so differently, they just define the terms differently.

    1. Pretty much. From what I’ve seen, compatibilists consider unpredictability as sufficient for free will, while determinists don’t. That’s why the discussion never seems to go anywhere: we’re using different definitions, not debating evidence, because we already agree about the evidence.

      1. Yes! Now,how about discussing my points about changed nurture and nature and determined volition. How does that pre-choice subconscious decide for us anyway? I hold for no free will but a volition that can choose somehow.Anyway,fatalism rules itself out.
        Enough with the semantics!
        Tom, thanks again for!
        WEIT,thanks for telling us about determinism.

    2. No, The contest is between noncompatibilists and compatibilists. A compatibilist, by definition, believes and free will and determinism are compatible. You’re right about the important part, though. It is about the definition of free will. The biggest sticking point has been the loaded adjective “free”, so I’ll drop it.

      I’m pretty much convinced by the compatibilist view, but that’s because I think that labelling the emergent quality of will as an “illusion” a mistake of extreme reductionism. A chemist,when discussing of intermolecular forces, doesn’t think the concept of “dispersion forces” is not useful because dispersion forces arise from more fundamental concepts of electrostatic forces acting in quantum mechanical systems. I think that the leap Jerry makes from determinism to “no free will” just throws out a particular emergent property of the complex living systems that are “us”. I doubt that he would dismiss all of evolutionary biology as an “illusion” just because biology is ultimately determined by chemistry, which is a consequence of physics.

      1. I’m a couple days late to the party, but I wanted to highlight this point:

        I doubt that he would dismiss all of evolutionary biology as an “illusion” just because biology is ultimately determined by chemistry, which is a consequence of physics.

        This is exactly the right way to look at this. Humans make “real choices” in exactly the same way a few genetic mutations confer “real evolutionary advantages.” I think if you deny one you have to deny the other as well.

        In some sense, the compatibilist version of “could have done otherwise” is just “would have done otherwise if _____” for the relevant counterfactuals, and the compatibilist project is to show that, rather than being trivial, it is really all we could hope to say, and is in fact the basis for saying anything scientific about the world in the first place. The point is to find out what the relevant counterfactuals are, and how fine-grained they have to be to make a difference.

        It is not nonsense to say that a Mini Cooper, say, has more “degrees of freedom” than a Model T, and that this real freedom is due to differences in design. We can compare humans, dogs, and slugs in the same way without needing to deny determinism.

        This is an epistemic point, but I don’t think it’s merely an epistemic point — something like the Laplace Demon really has no need at all to distinguish between horses and trilobites, because the patterns of regularity with which it is concerned are far more fine-grained. It has no need of the “natural selection” hypothesis because patterns of regularity like “species of life forms” are not in the class of epistemic problems it needs to solve or explain. It could dismiss the rationale for calling the collection of these lumps of matter over here “the set of all African elephants” as a “mere illusion” because it has direct access to what is going on from a physical standpoint, and each of these “elephants” is “in fact” extremely different from each other physically. To my mind, calling the classification scheme “elephant” an “illusion” is very much the same as calling “choice” an illusion.

        So yes, it is “just” a semantic disagreement between compatibilists and incompatibilists, but it hinges on rather staggering different senses of words like “real,” “I,” “choice,” “property,” “organization,” “illusion,” and the like.

  8. “To compatibilists, free will is completely independent of the issue of physical ‘causation.'”

    Well, compatibilists hold that you’re morally responsible and deserving of punishment if *you* caused your behavior, not someone or something else, so too much quantum indeterminacy or other noise in the mix would render you not responsible. Compatibilists wanting to defend retributive punishment (such as Dennett) like to single out the agent as the most salient cause of behavior, backgrounding her causal history and context, which of course incites the attribution of blame,

    Incompatibilists think that to be deserving of punishment one would have to be a libertarian causa sui, in some sense a self-caused first cause, hence the attraction (for libertarians) of claiming a causal disconnect from prior circumstances via quantum indeterminacy. But of course as is often pointed out, an indeterministic source of behavior can’t be attributed to the agent’s character or motives, so can’t be the causa sui basis for moral responsibility. And since consciousness and the neural processes associated with it don’t transcend cause and effect, it too fails as a causa sui.

    Wolfram’s “computational irreducibility” is yet another (failed) attempt to establish some sort of naturalistic causa sui, such that the person alone is singled out as causally responsible and thus deserves blame in the way our traditional notion of moral responsibility requires.

    1. If we don’t have free will, then those who engage in retribution can’t help it, so it makes no sense to criticize them. Furthermore, if all our choices are determined by prior circumstances, then the existence or absence of people engaging in retribution will be one of those prior circumstances, so engaging in retribution will cause people to act better.

      1. Of course it makes sense. Our own deterministic brain processes cause us to criticise them for holding this false view. And the practical result is that another falsity is corrected. As was determined!

        1. First you claim that it makes sense to criticize them, then you present an argument for why it makes sense that people do so. “Criticizing makes sense” and “It makes sense that people criticize” are different claims.

  9. ‘So for me (and I realize this isn’t true for others), whether we have free will or not boils down to whether quantum indeterminacy holds at the level of the brain and human behavior.’

    I don’t see what difference quantum indeterminacy makes, except that it possibly adds an element of randomness. It doesn’t give us control over our decision making. It’s like throwing dice to help you choose an ice cream flavour.

      1. How could that possibly work. His idea is that quantum indeterminacy throws up random items to choose from (he likens that to randon mutation in evolution) and that we choose from this list (he likens that to natural selection in evolution). But on what basis do we make that choice. There must either be determinists for our so called “choice”, or we “choose” randomly. We are back where we started!

        The above actually demonstrates the illogicality of free will.

        1. Billy Joe, Are you concluding that within the constraints of a universe that strictly follow the laws of physics, including quantum level phenomena there is no possibility of free will? If so, since we all, for all practical purposes (even this discussion) behave as if it exists then does that not lead to the inevitiable conclusion that there is more to us than simply deterministic objects which follow the law of physics?
          As Sherlock Holmes says “How often have I said to you that when you have eliminated the impossible, whatever remains, however improbable, must be the truth?”

          1. That bit of Sherlock has always annoyed me. After you eliminate the impossible what you are left with is a set of possibilities, some more and some less probable.

            It is one of those statements that sounds like a profound insight but actually leaves you no wiser after pondering it than before.

  10. It sounds to me like you and Wolfram agree on the big point. He’s just saying that even though everything is deterministic, it doesn’t mean we can predict exactly what people are going to do at any given instant.

    I think he sees that as one of the big stumbling blocks to people accepting that free will doesn’t exist—if everything is deterministic, why can’t we predict the future? Why can’t you tell me what I’m going to do next, or why I did X when I should have done Y? He’s saying the answer is because it’s still a very complex system that defies simple understanding.

    1. PS. The reason I say he would agree is that I know of some of his other writing, and basically he sees the whole world as being composed of finite deterministic automata. That view is not compatible with any notion of free will except that it doesn’t exist.

    2. We can’t predict the weather or the exact number, size, and travel path is a container of carbonated water, either, but nobody is arguing that heat waves or cans of soda have free will.

      Heck, just point out to people how often their computer doesn’t do what they expected it to do.

  11. If we accept that quantum indeterminacy affects human decision making, and therefore we have free will, don’t we also have to accept that quantum indeterminacy affects the decisions of essentially every organism with any sort of decision-making capability? So, not only do humans have free will, but so do chimps, cats, bees, and tapeworms. If using quantum indeterminacy to make a decision is our only criteria, does the contraption used to kill Schrodinger’s cat have free will? Is free will still a meaningful concept if it applies so broadly?

  12. (and me)

    I can deal with the simple grammatical matter. But, like mikespeir, my brain hurts when free will is on the table.

    1. Surprisingly, Steven Pinker says it’s perfectly grammatical to use “I” as part of a compound object the way Jerry did, as in, “Win with Al Gore and I!” He has an interesting discussion of why this is so in his classic book The Language Instinct.

      1. I quite enjoyed The Language Instinct when I read it a number of years back. I don’t remember his defense of the incorrect 😉 use of “I” in this way, though. Was it along the lines of “so many people say this that we shouldn’t consider it wrong”?

        1. I’m noticing more and more in discourse, wherever, that “people that” is replacing “people who.”

          Is this due to a predictable-enough tendency to generally conform with others in our more-or-less daily and immediate presence, or is there possibly a growing (sinister?) tendency to purposefully objectify, and devalue, human beings with the word “that”? I.e., a human being is a “who,” and a “human resource” is a “that.”

          If yuh see what Ah mean.;)

        2. On p 391 of The Language Instinct, Pinker notes, “People tend to misremember the advice” not to say “me and Jennifer are going to the mall,” and so “they unthinkingly overapply it — a process linguists call ‘hypercorrection,’ resulting in ‘mistakes’ like ‘give Al Gore and I a chance.’…But if the person on the street is so good at avoiding….’Me is going,’ and ‘give I a break,’ and if even Ivy League professors…can’t seem to avoid ‘give Al and I a chance,’ might it not be the [language] mavens that misunderstand English grammar, not the speakers? The mavens’ case about case rests on one assumption: if an entire conjunction phrase has a grammatical feature like subject case, every word inside that phrase has to have that grammatical feature too. But this is just false.” Then he goes on to explain why.

          1. Bummer. I thought you were going to explain why and not make me go find my copy of The Language Instinct. OK… I’ll go find it.

            1. I wouldn’t have been able to summarize it–it’s somewhat complex–so I would have had to do a huge block quote, on a side topic to this thread, which seems impolite. Hopefully [another usage Pinker champions], you’ll find yourself, as I did recently, delightedly pulled into re-reading the book!

          2. ” “People tend to misremember the advice” not to say “me and Jennifer are going to the mall,””
            It is somewhat ironic that while discussing hypercorrections, you engage in one yourself, following the misguided prohibition against “split infinitives” and therefore avoiding the correct form of “advice to not say” rather than “advice not to say”.

            “and if even Ivy League professors…can’t seem to avoid ‘give Al and I a chance,’ might it not be the [language] mavens that misunderstand English grammar, not the speakers?”
            The obvious conclusion is that language processing is a specific brain module that operates primarily on inborn propensities and inductive processing, and is largely independent of other general intelligence and higher education, Ivy League or not. This module is capable of distinguishing between objects and subjects, learning rules regarding those categories, and applying those rules at the speed required for human speech, even if the person hasn’t been explicitly exposed to the concepts of “object” and “subject”, nor is consciously aware of the rules. Since the conscious brain is not involved in language production, the conscious knowledge that “me” is an objective pronoun and that what follows “give” is the object of that verb does not affect whether “me” is used in the phrase that follows “give”. The language module, which normally applies grammatical rules, can be easily confused, resulting in usages that are inconsistent with a rational analysis of the structure. A common piece of advice is to remove the rest of the compound phrase, such as changing “Give Al and I a chance” to “Give I a chance”. This tactic involves using conscious processes to transform the sentence into one that is simpler and therefore easier for the language module to operate on; the result of the language module’s processing is then consciously applied back to the original sentence. This is takes much to much time to be applied during an actual conversation, which is of course why the faster, but less rational, language module is used in the first place.

    2. While I would prefer the objective case, given that it’s a sentence fragment to begin with, and the clause is operating as neither a subject nor a object (and, indeed, there is no verb for which the clause can be a subject or object), I don’t see how it’s valid to definitely state that one particular case is correct.

    3. I agree.

      “Stephen Wolfram (and I) on free will” is grammatically incorrect. If there was no Stephen Wolfram, the heading would be “Me on free will” not “I on free will”.
      (Steven Pinker notwithstanding)

  13. Now my definition of “real choice”, i.e., free will, has always been this:  if we could rerun the tape of life to the moment when we made a decision, with every particle in the universe in the same place at that moment, and yet we could have made a choice different from the one we did, then that is free will.

    This, of course, would mean that there’s some sort of “you” outside of your physical self that’s making decisions. And I think we can all agree here that cognition is something that goes on within the brain.

    I still maintain the concept is incoherent. “Free” of what? Determinism? Determinism simply means that there are rules that are followed. If you’re “free” from determinism, then there’s nothing left that’s in anyway meaningful or distinguishable; you’re just flailing about. Free from (quantum?) randomness? So, even though stuff is happening at random all about you, you’re still following your game plan.

    The classical Christian viewpoint is that there’s some sort of supernatural spirit pulling the strings of your brain, and that it’s this otherworldly parasite that’s got the freedom. But is it flailing about at random, or is it following some sort of guidelines? Christians tend to get upset when you ask those kinds of questions…and, of course, if you really want to see some cranial assplosions, ask if they’ve got free willies in Heaven….



    1. “Free” means free from unwanted influence by other minds. Evolutionists of all people should not forget that we are conscious animals living in a social environment. That means we evolved to deal with other intention-holders. Free Will is a concept that is important only in a battlefield of conscious intentional agents. That’s why it is important to us. Free will is not about being free from causation. Free will is about being free from coercion by others. I’ve always been pointing this out that this is the meaning that we use in everyday usage – “Did you sign the contract of your own free will”, etc.

      1. Someone who falls into a coma due to natural causes clearly does not have free will, while a child subject to parents’ rules does, so your criteria for “free will” don’t make sense.

          1. So, does “free will” apply to a PERSON, a SITUATION, or a CHOICE? If a child wants both chocolate ice cream, and strawberry, but is forced to choose just one, does the child have free will?

      2. “Free will is about being free from coercion by others.”

        What others do and say are inputs into your brain and are therefore part of the deterministic cause and effect relationships in your brain that determines your brain’s output.

        So, no, fail.

        1. Get out of the reductionist level of thinking. Otherwise your everyday usage of language will be incoherent. Freedom, intention, choice, responsibility – should have no meaning to you if you are consistent with your philosophy of lowest reduction.

    2. “I still maintain the concept is incoherent.”

      Incoherent is a better description. I have generally said illogical, but I will use incoherent from now on.

      But here’s the point. Consider the two words separately. Free? Free from determinism? Therefore random? Random would certainly be free. But Will? Where is the Will in random? And if not random, then the Will must be based on something. And therefore the Will is not Free.

      It’s incoherent as you say.

      1. And yet you don’t act like free will, choice, responsibility are incoherent. In fact, I content without even knowing you personally, that you act in your everyday life as if you know perfectly what those words mean.

        Saying something while acting otherwise – there is a word for that.

  14. The way I read the troublesome sentence is that computational irreducibility explains the relationship, that we have something that is determined at the bottom, but still looks enough like free will at our level for philosophers to have been debating it.

    That said, while Wolfram has an interestingly skew approach to mathematics and makes some spiffy software, I don’t consider him all that great at philosophy of science.

  15. We do make choices, in a sense. (See Dan Dennett on the evolution of “evitability”). Our choices are determined by our brain state at the moment of choice, including the emotions we experience that motivate those choices. And that brain state is in turn the product of the interaction of our gene-environment wiring-up and all our previous experiences from conception through the instant before the choice. As Sam Harris has pointed out, we don’t even actually experience ourselves as having free will–we only think we experience our choices this way. But when we observe our experience closely, we find that our choices simply arise in our minds, or manifest themselves in our behavior, and we don’t have access to what happened in our brains to produce that choice. We explain to ourselves why we made the choice, but our explanation is only an inference, at best–a mere post-hoc rationalization at worst. Causality is difficult to show when there are so many variables, so many of which are hidden from our awareness.

    1. Yes, great comment.

      It seems that even though it’s determined by the result of all of the brain states preceding it, the same instantaneous brain state is both unique and can never be repeated.

      If for no other reason, as Peter Ozzie pointed out below

      “Given that our wetware is even more complex and based on electro-chemical neurones etc
      it is improbable that the same computation could be achieved reliably.”

      So it seems that we don’t really have “contr-causal” free will but we are all on our own unique trajectory though time, our own trips so to speak.

  16. I’m sorry, but the entire conversation concerning “free will” takes place in the world of mind, i.e., the world of “free will.” There is no scientific theorizing without mind. There is no truth without mind. Free will is thus fundamental to any sort of conversation. The supposed conflict between mind and atoms is a misformulation. To understand this, look to, e.g., Wittgenstein, not Einstein. (Although I’d guess Einstein would probably agree with Wittgenstein on this point.)

  17. Dear Prof Coyne
    First, I appreciate reading your thoughts all the way down here in Oz.

    You say:

    . . . if we could rerun the tape of life
    to the moment when we made a decision,
    with every particle in the universe in the same place at that moment,
    and yet we could have made a choice different from the one we did,
    then that is free will.

    But that is clearly impossible to achieve.
    That is, getting everything back to that state.
    I assume this is some Gedanken experiment you are postulating?

    As someone who knows diddly squat about most things I hesitate to tell you how to suck anything.

    But, from my background in computing–even simple digital switches are
    anything but boolean switches. Real devices exhibit metastability, ie not just 0-1.

    At a lower level, a zener diode exhibits non-deterministic behaviour as it is taken from
    one state to another, depending on temperature etc.
    See “A zener diode random number generator circuit” at

    As you climb up the digital levels to microcode
    then that too is non-deterministic.
    Then actual computers with myriad of simultaneous events
    are never deterministic.
    Eg interrupts, network traffic, user inputs etc occurring seemingly randomly.
    Any real-time system, for eg, is hard to test because of this lack of repeatability.

    Even simple memory devices will
    fail with alpha particles randomly emitted by their very fabric altering their state.
    See “Alpha-particle-induced soft errors in dynamic memories”,
    IEEE Transactions on Electron Devices,
    V: 26, #1, Jan 1979.

    Given that our wetware is even more complex and based on electro-chemical neurones etc
    it is improbable that the same computation could be achieved reliably.
    And would our decision circuits have access to the same information
    on the second attempt?

    Me thinks you have too much faith to be a physicist!

  18. You talk about quantum indeterminacy,but what about the Everett interpretation of QM? The one that says that we live in a multiverse of parallel worlds. So to take your example, rewinding the tape to the beginning of time and starting again would give indeed you a different result in a different branch of the wavefunction than the one we live in. Thoughts?

    1. That is, as all philosophy or it wouldn’t be philosophy, not a problem for physics. You can’t do that experiment.

      And you can’t even get close in principle if you try a less ambitious rerun. See how deterministic chaos means no finite resources suffice to reposition as you would need infinite precision in parameters.

      I would posit that a gedanken experiment needs to be testable in principle, that is how Einstein used them with effect. Say, rejecting non-relativity by positing an observer at the wavefront of a light signal. You can’t do that with a massive observer, which was the telling point. It was his loose gedanken experiments (“no dice playing”) that got him into trouble.

      If anything, the above problem shows that the “free will problem” has not even a physical template. In other words, “free will” Isn’t Even Wrong.

      We can have a “will” description if we will, without any problem of specific physics. (So called “compatibilism”.)

  19. i dunno.. call it whatever you want. but the only “free will” i would be interested in having, is one where my choices are made after I consciously choose to do them… and so far, the science seems to point in the other direction, that my choices are done long before i become aware of them, which if true, makes any “free will” we have rather moot imo.

  20. It all boils down to “Are we the dreamer or the dream?” and there is no math, science, philosophy or religion that can determine that conclusively. So any and all of us can decide which answer is correct and proceed with our happy little lives. That is Free Will.

      1. That makes eminent sense. Philosophy as an area admit anything for everything, so explains nothing on no thing.

        The difference is that postmodernism is explicitly designed to be a branch achieving “anything goes” directly and explicitly.

  21. My reading is that Wolfram agrees we don’t have free will but the problem of working out how we actually come to following a course of action is impossible. So the only way of finding out what course we will end up following is to “run the program” and see. SO he would agree with Jerry that re-running the tape would produce the same answer but that doesn’t mean, even in principle, that we can predict ahead of time what course of action will be followed.

    What this means from an ethical point of view? dunno.

  22. Apparently most determinists nowadays rely on neuroscience – but are unaware (ignorant?) of neuroscientific ignorance, so they vastly overgeneralize from current neuroscience. They think that all we think and do is predetermined by brain states. Well, maybe, but what sort of evidence might clearly support that? Here’s one:

    Show neuroscientists brainscans of the most sophisticated types currently available of an individual’s brain – without their knowing who the person is or in what circumstances the scans were taken. If they can then describe EXACTLY what that individual was thinking and perceiving (seeing, hearing, feeling, pondering etc.)when the scans were taken – that would be pretty convincing evidence of determinism. But current neuroscience is about 1000 miles away from being able to do anything close to that. Until that IS done, I’m skeptical that it could be done. I guess I’m a radical empiricist – I’m skeptical that something can be done until it IS done. For more on ignorance, Google for: Ignorance Firestein.

  23. I’ve always seen Wolfram as a hair’s breadth away from being a Kurzweil. He may be rolling in dough but in general I just can’t take him seriously.

  24. I still think that if the universe were replayed from the Big Bang up to the instant you posted this, you would title it “Stephen Wolfram (and me) on free will.”


  25. Love these debates about free will or is it “free willie” (willie or won’t ‘ee?). Here’s something to insert into the discussion. Does thought influence action? Or, is action all just a part of deterministic processes? Recently saw this news item on TV: Researchers have developed a robotic arm they claim can be operated by thoughts alone. They fitted this arm to a woman amputee who is paralyzed from the neck down. They say she is able to control the arm (make it do what she wants) by thinking alone. Unfortunately, no explanation was given as to how her thoughts are transmitted to the arm. A helmet that picks up electrical signals from the brain? I don’t know. Can it be said she “wills” the arm to do her bidding? Is this free will? Anyway, the company that invented the arm “gave” it to her (for free) so they could continue their research.

    Any thoughts on this regarding the free will debate?

  26. I seems to me that free will is a social construct. In the courts for example, judges and juries try to determine degrees of “responsibility” of the accused. One of the most known is, during a crime like murder, did the accused understand the difference between right and wrong?

    To me, that means–did the accused learn the social rules we live by well enough to understand was is acceptable and unacceptable in our society. Learning that is due to training in conditional behavior, or as Skinner said, operatant conditioning.

    It has nothing to do with free will, but with the measure to which the accused integrated, or not, this training into his social behavior. If he is thought to have understood social restraints, he is considered “responsible”. Some people seem to have mental disfunctions, for whatever reasons, which prevent them from absorbing and maintaining acceptable social behavior.

  27. As for this sentence,

    I tend to think that the concept of computational irreducibility is probably the answer to the philosophical debates of the past two thousand years about the relationship between free will and determinism.

    I simply don’t understand it.

    Wolfram thinks that computational irreducibility is probably the answer to everything. Yes, most know him as the creator of Mathematica, but but recently he’s been peddling his idea that cellular automata are the key to everything, as in his book A New Kind of Science. I think he is (rightly) pretty much of an outsider here.

  28. Wolfram is simply describing an effective “compatibilist” view, which goes as much into the biological details that one can do at the moment. I.e. (IIRC) neuroscience shows that rats simulate what they do up to and from the position they are in as they navigate a maze. Something similar to what we seem to do when we dream, likely in order to fixate memories.

    In that sense folk psychology models of “free will” correspond to the biological function of how to see ourselves with our distributed brain functions making choices.

    It is interesting to see Wolfram making a cogent description of anything, since he is very much like the generic woo futurist of the Kurtzweil form. Specifically for Wolfram he is crackpotting on trying a cellular automata base for physics, which are known to be inconsistent with relativity and so can never work.

    Shows that even a broken watch can be right once in a while.

    1. An explanation is that like Kurzweil, Wolfram labors under the problem of being mostly something other than what he fancies. He is mainly a computer and math specialist.

  29. I just think the term “free will” should be abandoned. No matter where you stand on the issue, the term has just become laden with so much unsightly baggage. You have to wade through a mountain of crap just defining what you mean by it before you can begin a discussion, and often there are still misunderstandings. The question really just ought to be, “Are human behaviours deterministic?” Then, with either a yes or no answer, we move onto the question of how to deal with moral issues. What’s in the issue in deciding whether any of this is deserving of the label “free will”? We gain nothing from it.

    For example, I think Dan Dennett’s compatibilist approach to “free will” is really just an explanation of how we can maintain moral accountability in a determinsitic world. He appropriates the term “free will” for this, but it’s unnecessary and potentially misleading. I often wonder if this is because of his concern for truly dangerous ideas – he may secretly believe that a widespread belief in no free will could be damaging, and so is using the term in this manner. Just get rid of it, I say.

  30. I’ve been giving the question of free will a lot of thought and the more I think about it the more I realize that no one really knows what free will is.

    Like everyone else here I started from determinism vs non-determinism and I could see where someone could have free will in a deterministic universe – or not have fre will in a non-deterministic universe. For example – a leaf floating down a turbulent stream in a non-deterministic universe has no free will.

    Eventually I concluded that determinism doesn’t matter in the free will question. Or self awareness is sort of a self modifying software program that runs on a meat computer known as the brain. Our software (which I refer to as the soul – but not an immortal soul – unless Ray Kurzweil saves us – is a program that creates an identity we call the self. I think therefore I am.

    So – my intuition is that self awareness and free will are 2 aspects of the same question and that both represent an evolved form of software that is different than any software humans have developed so far. And therefore humanity might not yet know enough about reality to even understand what free will is yet.

Leave a Reply