Some Thoughts on Free Will

Thoughts on Free Will

I love watching Christopher Hitchens. Although we don’t agree on everything, he’s probably my favorite debater. He has the important skills: passion, eloquence, command of facts. And, he favors the Oxford debate style, where the goal is to be right, rather than to build consensus.

But I always cringe when he gets asked about free will. Hitchens is, I think, at core an anti-totalitarian of the Orwell cast. So, despite being an avowed Atheist, he bristles at the idea of agreeing that there is no free will, even though it is the consequence of believing in a deterministic materialistic reality. He inevitably gives the same answer – “I have no choice.” It’s clever, but it’s clearly a dodge.

It got me thinking on the question, and I think I’ve reached some interesting conclusions. This essay will probably be a bit involved, so please get a cup of tea or coffee before you jump down the rabbit hole with me.

Got it?


As I see it, there are both necessary and sufficient conditions for free will. These can be made more specific, but the general idea is this:

The necessary condition for free will is that the universe (or multiverse) be non-deterministic.

The sufficient condition for free will is that human beings have some sort of access to the necessary condition.

That leaves 5 possibilities:

1) Universe is deterministic, so we have no free will. (Determinism)

2) Universe is deterministic, but we still have free will. (I would call this a religious argument)

3) Universe is non-deterministic, and we have free will. (Hooray!)

4) Universe is non-deterministic, but we have no free will. (That is, for example, humans are predictable but electrons aren’t).

5) There’s some other possibility I’m not considering.

Let’s look at the necessary condition first. For a long time, I was troubled by this thought experiment: Let us suppose that the universe is deterministic, meaning that each action is simply the result of the prior moment. Now, let us suppose that in this universe, you took a subsection (say a 20x20x20 meter cube) and built a computer that knew all the parameters of everything inside. Let’s also suppose there’s a person inside. Now, suppose the computer predicts to the human, that today for lunch he’ll have a ham sandwich. What’s to stop the human at that point from defying the prediction?

This confused me because it seems like you arrive at a contradiction by assuming a deterministic reality. But, I thought about it more, and I eventually arrived at this conclusion: The thought experiment is faulty, because the monitoring computer has to be considered part of the system. That is, in order to know it’s predicting correctly, the computer must know the state of all of its own parameters in addition to the parameters in the cube. This seems to me to be impossible, because the computer would have to perfectly simulate a system more complex than itself! This leads to even deeper problems, since (if true) it’d mean that even if reality is deterministic it could never be determined, even by a being who exists outside of reality.

After that, I started to wonder if philosophy wasn’t the right way to approach this problem, at least not yet. After all, free will is essentially an empirical question, at least as posed above. Does the universe meet the necessary condition, and do humans meet the sufficient condition?

For the necessary condition,  it’s useful to look to physics. Although determinism makes intuitive sense (since it seems impossible to have effects without causes and vice versa), there are strange things in physics that make you wonder if reality is actually non-deterministic. So, it seems to me that we don’t  have an answer for the necessary condition, as it is an empirical query without an answer.

But let’s not stop there. This opens up another possibility. What if there is no answer to the necessary condition? That is, suppose we discover that for good scientific reasons it is never possible to determine whether reality is deterministic or not. Does that mean reality may as well be non-deterministic? After all, if we can never know the answer, pragmatically speaking any answer is appropriate.

I think that, if the necessary condition is impossible to get it, it may be worthwhile to consider the sufficient condition. For, if the sufficient condition cannot be met, the necessary condition is moot.

I think it’s conceivable that you could have a non-deterministic world that has deterministic stuff in it. For example, it’s possible to imagine a complex molecule whose components have some quantum wobble to them, but which in aggregate interact at a purely chemical level.

If that’s the case, we have to ask ourselves whether there’s an evolutionary advantage to free will that’d cause access to it to be a fitness advantage. An simple example of animal access to free will would be a machine in the brain that performs the double slit experiment every time your brain needs a random number for something. Okay, bit of a lame example, but I think it’s a decent proof of concept. You can imagine how a creature that is capable of truly random action might strategically outdo an opponent who behaves in very recognizable patterns. Anyone who’s ever played a single player video game long enough to exploit the AI knows this to be true.

I don’t know if we possess abilities like this or not, but it’s interesting to contemplate. What’s even weirder is that you can imagine a species that lacks free will, but which is at some point able to install a machine into itself that gives it access to free will! I’ll have to write a sci fi story about that one day.

So, where does this leave us? Right back at the beginning, I’m afraid. As far as I’m aware, we can’t make a final judgment on either the necessary or sufficient conditions. Perhaps then Hitchens’ irony is entirely appropriate. We have a question we can’t answer. A question we may never be able to answer. So, we might as well have a laugh at it.



This entry was posted in Philosophy. Bookmark the permalink.

39 Responses to Some Thoughts on Free Will

  1. Zach says:

    About the infinitely complex computer system:

    There’s always the question of the man hearing the prediction and trying to prove it wrong, but what if he was the only person who could NOT see the prediction? What if everyone outside of the 20×20 system could see the prediction, but the man inside must remain isolated. Or what if it only reports to one person or another computer which is isolated in practical terms from the human in the system? Then if the computer could still make accurate predictions, wouldn’t the universe be deterministic?

    At very least that could prove it would be semi-deterministic, leaving only the variable of a human to whom the report was delivered.

    • Evan says:

      (first off, sorry for spamming this thread, I’m just feeling procrastinatey)

      3 points.

      1) if it’s theoretically possible to perfectly predict the human’s behavior from physical laws, then that’s straight up, full on determinism, nothing “semi” about it.

      2) If this toy universe is deterministic, and outsiders predict what the human will do, and then they poke their noses into the system to whisper in the human’s ear “hey, I dare you to not eat the ham sandwich”, or whatever variant (“I dare you to not touch your boner” doesn’t count because noone can reasonably argue that not touching the boner is physically possible), that act affects the system. You can’t communicate that information to the human without changing the system, fundamentally. And once you’re changed the system, you can run your simulation again and find out that, lo and behold, the human doesn’t eat the ham sandwich. This doesn’t violate determinism or predictability or anything. It’s simple exactly what would happen in a completely deterministic universe with a god-computer.

      3) In case you haven’t come across it yet, you might enjoy reading about quantum interrogation. It’s a technique that can allow you to (in some circumstances) receive information about a system without observing it (the wavefunction doesn’t collapse). It’s a little tangential to free will, but if you want to wiggle out of the argument of “how would an outside observer get information about the system?”, then that’s one way you could theoretically do it.

  2. Dave says:

    I guess I was wrong in thinking that QM already pretty well shows that the universe is non-deterministic?

    • Evan says:

      No, you’re right. There is no scientific debate regarding this at this point, the universe is non-deterministic at a fundamental level, although you can still make perfectly legitimate probabilistic predictions.

      • Matthew says:

        Actually I had a quantum professor in college who would say that given the experiments of quantum mechanics you have to give up one of three ideals, that the universe is non-local, non-deterministic, or non-temporal, and you would still be able to explain all phenomena. It just happens that for humans the easiest one to grasp a universe without is the non-deterministic bit, but if you wanted to dearly hold onto that in a quantum system you could, but at the expense of one of the other two.

        • Evan says:

          Hah, I haven’t heard much by way of “non-temporal.” Is that the idea that we can only perceive of one instant of time (at a time), and therefore this may be the only instant of time that exists? But yeah, you’re right, there’s always good old non-locality. But nobody really wants to go there. It’s just an ugly place.

          More to the point, regarding this discussion of free will, you can play a pea-and-cup game with non-locality or non-determinism all you want, but that’s effectively a cosmetic difference. The experiments don’t care how you phrase it. I stand by the feeling of my original statement. There may be one or two people out there seriously working on non-local theories, but they would be in the extreme minority. Which isn’t to discount their work, I think it’s a terribly cool topic, but there is a general consensus of interpretation at this point and it’s solidly in the non-deterministic camp.

      • matthiasr says:

        Actually, there is. Put a bunch of physicists in a room with a few bottles of beer and they will inevitably start debating about the interpretation of quantum mechanics.

        The most popular interpretation, the one taught in most schools, is the – indeed non-deterministic – Kopenhagen interpretation that postulates the non-deterministic “waveform collapse” upon “measurement”. IMHO it is a little dodgy about what distinguishes a measurement from any other ordinary interaction, considering that all the instrument and the experimenter all consist of particles obeying the same quantum mechanics.

        The second most popular interpretation is Everett’s Many Worlds theory, which basically says that there is no such distinction, that the wave-like rippling of cause & effect continues, encompasses our brain and it just looks like some particular outcome happened to the observer in that branch of the universe – but actually every outcome happened to some part of the observer.

        Then there are theories that say that the collapse is actualy triggered by outside influences (e.g. Penrose: by the self-gravity of an object). This may or may not be deterministic.

        There are a ton of interpretation, a pretty long list is in Wikipedia.

        Nonetheless, in order to actually predict anything on the microscopic level where the universe might be predictable someone (Laplace’s Daemon) would have to know the full state of all particles in all of the visible universe. And if it can actually act on it’s knowledge, it is part of the universe and must consider it’s own state too.

        • Evan says:

          I’m as big a fan of physicists drinking as anyone.

          Yes, you’re right, the Copenhagen Interpretation is famously untested (possibly untestable). But I feel that, especially in this discussion of free will, that bringing up quantum mechanics in general is a big red herring. Unless you’re prepared to argue that QM is “where free will is found”, then it’s just a way of muddying up the issue of determinism without substantively addressing the paradoxes of determinism.

          Also, and again, you won’t find a bigger fan of drunk physicists than myself, I feel that until you actually have studied QM to the point where you truly understand the “debate” surrounding the Copenhagen Interpretation, have the canonical experiments engrained in the core of your being, have bled onto the pages of Bell’s Theorem and can recite the narrative of thought experiments as fluently as a childhood fairy tale, that until you actually know what the hell you’re talking about, I believe that pointing out that there are other interpretations is downright misleading. I feel it’s misleading (despite being literally more accurate) because it suggests that the other interpretations might be more reasonable, and the fact is that they aren’t. The Copenhagen Interpretation is absolutely the time-tested Occam’s razor of QM interpretations.

          It may well be fundamentally untestable, and thus the “debate” surrounding it would be entirely inconsequential (hence the quotes) but, for my money, the bottom line is that any other interpretation must be at least as unswallowable, and generally they are much worse. As you say, the “many worlds” interpretation is perhaps second to bat. It’s an interpretation that literally says the entire universe multiplies itself a hundred million times every time you open your eyes. Now maybe I’m just old fashioned with a fond heart towards energy conservation, but good god if that’s the second most popular interpretation, I’d hate to go down the list (in truth the many-worlds interpretation is just easier to state than non-locality, and surely that’s as much to owe its popularity, but still).

  3. Oliver says:

    Thanks for the post, it’s interesting to read a science guy being open about this topic. I still remember the link you posted last month to a blogger asking if the soul had a physical particle, and proudly concluding there was no free well – appalling.

    I’ve studied philosophy for a while, and I think I could add a bit more info about your 2rd hypothesis, the “religious” one :
    2) Universe is deterministic, but we still have free will. (I would call this a religious argument)

    This is an extremely irritating argument : deciding that the basis of free will resides OUTSIDE of physics definitely makes the question unanswerable, it’s a matter of believing or not. It can NOT be proven wrong, nor can it be proven right either.

    But it’s not really specifically religious. Look at Kant’s moral agent theory.
    In short : if we are deterministic, without free will, our civilization and rules are moot, and everything making the whole system work is hollow.
    Take justice : why punish a guy stealing bread if he’s got no free will.
    Take love : that girl doesn’t love me, it’s the atoms, stupid.

    All the things on which our lives are based are non-material, they’re beliefs and faiths, all of them relying on the faith there’s free will at some point.

    Kant’s conclusion : we can’t prove it because it’s not material, outside of physics, but we’ve got two choices, either believe in something « a priori » (call it the soul, free agent, who cares, it allows free will anyway) – or jump into the nearest river with a heavy stone attached to our neck.

    As I said, this is utterly irritating.
    I didn’t think the scientific side of the question could be equally irritating – thanks once again for the post !

    • Another Zach says:

      But don’t you think its possible that a girl doesn’t love you because your pheromones and her receptors aren’t in sync? Its definitely part of what makes love happen. Maybe it isn’t the be-all end all, but maybe it is. Maybe by matching signatures of atom frequencies more accurately we can turn marriage into a scientific field, thus significantly reducing divorce.

      Not the most romantic idea, I know, but something worth examining.

  4. bnetter12 says:

    Whether or not the universe is deterministic becomes even more confusing when including the quantum foam, heisenberg uncertainty and observation. We can in effect create effects without cause.

  5. Nicolás says:

    Nice post, Zach.

    I’ve been thinking on this problem for a while, and arrived at some interesting (and, for some, frightening) ideas.

    Whether or not reality is deterministic, it’s risky to claim freedom:

    1) if it’s deterministic, and our decisions are predictable, then we all agree that we wouldn’t have free will.

    2) on the other hand, if reality isn’t deterministic, what do we have? random decisions? what kind of freedom is that if even myself cannot predict what I want!

    If our decisions are truly randomised (because I cannot conceive another way of avoiding predictaility) we are slaves to some mysterious multiversal dice which decides if our choices roll twenties or plain ones. That’s for sure not free will.

    What’s more, our actions become statistics!!! What a way to undignify humanity!

    The only possible solution to this problem, is to make a concession to some unknown state of things (souls?) by which we can be simultaneously dignified by controling our own lives AND having our honour intact by not letting anyone else predict it with scientific certainty. The main defenders of such contradictory things already gave this idea a name: religion.

    I said it would be uncomfortable. I, as an atheist, find it hard to make this concession to the not-yet-discovered, but it’s the only way to allow freedom to exist (whatever freedom may mean).

  6. Nemo says:

    I think we live in a deterministic Newtonian world built on a possibly non-deterministic quantum foundation, but it’s the first part that matters. Mostly.

    Anyway, yes, physics can do more to answer this than pure philosophy. Great post.

  7. Jos says:

    I don’t think that free will would entail being unpredictable, nor that the absence of unpredictability would mean there is no free will.

    The brain researchers say that the proof of the absence of free will can be found it the registration of certain brain patterns before the cognitive processes set in. So the subject of the test thinks he made a choice at a certain moment, but the brain has already decided before that moment in time.

    In essence, you’d need a good definition of free will first. I personally think that free will is the ability to not follow primary instincts and the ability to modify behavior and thinking without a direct evolutionary drive. As such, it might be very closely tied to the discussion about religion.

    Anyway, In short, I think free will exists, because one can “will” behavior to change, over time. Predictability cannot proof the absence of free will alone.

  8. FrangenS says:

    I don’t think time travel (predictions!) should be mixed in a debate about free will. (Outsiders may well do that, but for insiders I’d refer to the computer problem you mentioned.) This would leave me with an open mind for the possibility of free will AND a deterministic universe. You see, free will is not something that strikes from the sky and makes you behave irrationally. The deterministic universe would just have to know what you WANT to do, thus having you want it to turn out like it secretly hoped. “Problem solved!”

  9. Evan says:

    When you talk about free will and physics, you need to be careful how you phrase things: in particular, determinism. In the face of quantum mechanics, we know that it is no longer possible to predict the future except in probabilities. That simply must be accepted. It’s true that often enough the probabilities have massive statistical biases (which is why you CAN predict a ball whereas you can’t predict the electrons that make it up), but ultimately it’s all subject to these probabilities in a profound and meaningful way. So what I might call “literal determinism” is out of consideration.

    So now some people will try and use that foothold of quantum mechanical uncertainty to say that “Aha! That’s where free will must reside.” Unfortunately, that turns out to be absurd. We can’t predict (beyond very real probabilities) a low-level physical interaction like electrons, and we also haven’t been able to piece together a high-level system like consciousness, and yes ultimately the consciousness must rely on some 10^25 low level interactions at any given moment, but trying to connect the two is profoundly misguided. To wit, until a human can predict a double slit experiment by “thinking” about it or “willing” it to be some way, that argument carries no weight.

    The main thrust of my view on free will then, is that it’s an idea developed by humans to describe a clear, intuitive facet of a high-level system of consciousness, like qualia. It’s fundamentally a part of the human condition, by semantical design. When I choose from a menu, it’s true that at some level it may be possible to piece together enough of my mental state and preferences to come up with a good probabilistic prediction of what I’ll decide to order before I figure it out, but ultimately I choose to order whatever I feel like. It’s my choice, and nobody else’s. You might say “but isn’t it rather physics’s choice?” Well, fine, but my consciousness is a manifestation of a high-level system of physics. At the end of the day, I’ve made my own decisions, from my own myriad of motivations, and although you may be able to reasonably predict my actions, they are still fundamentally and meaningfully my own.

    What I would ask YOU is “what is a sufficient condition for an act of free will?” Must we operate outside the bounds of physics? Indeed, if our actions had so much as a probabilistic bias, that would immediately allow some predictability. But then, even if a being behaves purely randomly, we know a thing or two about randomness, so we could make mathematical predictions about that too. So I ask again, “what is a sufficient condition for an act of free will?” Really, I’m just asking you what do you really mean when you say “free will”?

  10. I really don’t see how a non-deterministic universe would do to solve the problem of free will. There are interesting things one can say about free will, e.g. discussing to what extent we are controlled by biology or the subconscious, but this general grand philosophical question is in my opinion nonsense. If an individual decides to act in a certain way there are three options:
    a) The decision was the result of a deterministic process, given the subjects past experiences and, for lack of a better term, personality.
    b) It was a random decision, which does not seem to be compatible with our intuition of free will either.
    c) A mixture of the previous two.

    There doesn’t seem to be any other option. None of them seem to satisfy our intuition about what free will is. The only logical conclusion then is that our intuition is misleading.

    When we think about something we did and we think about free will, our thoughts are probably expressed in some way similar to ‘I could have made a different decision’. But that is simply not true. What is true, and what I think is what our intuition is actually trying to tell us, is that we can learn from the experience and become a different person. A person that would have acted differently in that situation.

    On the same line, I think we can define free will as the capacity to consciously learn from past experiences and react accordingly. This definition would be in general compatible with how free will is used in our legal systems (not with the part that seeks to punish, not reeducate, but that part is in my view obviously immoral in any case). It also leaves the other problems with free will that I named at the beginning as interesting questions, based on some definition of consciousness, which we are going to need anyway if we are talking about the essence of being human.

  11. Jhadur says:

    I’d suggest that you’re asking the wrong question – not “is free will, defined this way, possible/real?”, but “how would free will need to be defined in order to be possible/real?”. Hume has some interesting things to say in this direction (falling broadly into your category 2), but tend to be ignored by science types*, who often see that kind pf redefining terms as cheating.
    (Btw, thst story, inevitably, has already been done by Greg Egan :op)

    * – for reference, my background is in maths.

  12. Deno says:

    Zach, glad you’re posting with more frequency here!
    I think that the question of whether we have free will is a moot point. The universe is probability-driven on the microscopic scale, and deterministic on larger scales. Free will arises for the same reason weather does: Many complex factors are non-linearly related.
    As an astrophysicist, I sometimes make derisive comments about sociology/psychology, simply because these fields are based entirely on a statistical footing, and quite frankly, Freud, Jung and other founding members were as far from deep statistical analysis as you can get (even today, I think psych/soc researchers run the risk of small numbers phenomena). However, just as we can make broad-brush arguments about long range climate forecasts, we can also make serial killer profiles (albeit the data set on serial killers is orders of magnitude smaller than what climate scientists work with).
    Bottom line: Free will, as usually defined, is half-true as our brains work by pushing ions around. Determinism, when dealing with large agglomerations of people or particles, can give an accurate (if imprecise) model of what goes on in people’s brains.

    • Evan says:

      As a cosmologist myself, I wonder, if you were trying to study the human mind, how would you go about it any differently than psychologists do (as a scientific community, not one or two bad ones, dare I go through the rich anthology of astrophysicists in search of a couple with a poor sense of statistics)?

      More to the point of free will, I think the brunt of the issue isn’t with realistic predictability, but rather with theoretical predictability. Unless you want to argue that free will is found in quantum uncertainty (it surely isn’t), you’re faced with a situation in which, for all meaningful purpose in this discussion, your actions in life are theoretically entirely predictable.

      I love quantum mechanics dearly, but for the sake of this discussion I wish it wasn’t around because it so heavily clouds our ability to think coherently about free will. So what I’d like to argue is that if we can agree that quantum mechanical uncertainty is not “where free will comes from”, then it must come from other laws of physics which do not have a principle of uncertainty. Predicting actions from the information of all the particles in a brain is prohibitively complicated, obviously, but nonetheless consciousness (and thus free will) must be completely describable by laws which are theoretically predictable with complete precision.

      So that’s what’s at stake. If (QM aside) you’re entire life could have been precisely predicted, every decision you ever made, from information about the state of the universe a hundred million years ago, then a rock falling to the ground is, in some sense, equally responsible for its actions as a person committing murder.

      The line of reasoning then goes on to try and paradoxically conclude that we’re all basically fleshy robots of sorts, and it doesn’t make much sense to punish a robot if its committed a murder, and also all passion in life is somehow fake (being robots).

      So that’s the challenge that the question of free will asks, it’s not a question of “are our minds realistically predictable?”, rather it’s “are we soulless, fleshy robots who are not responsible for any of our actions?” So it’s a question that requires a rather different approach.

  13. theorrhea says:

    I think the best way to approach this problem is to focus on the WILL aspect of the equation – the question of the universe’s determinacy only becomes relevant once we start bringing up metaphysical characterizations of a prosaic psychological subject.

    Our motivation for these metaphysical characterizations, as far as I can tell, always comes down to something missing in our experience – John Searle, who is libertarian on free will, calls it ‘the gap’. Given a decision, we have reasons for making one choice or the other, but we don’t experience the reasons as causally sufficient – instead there’s a gap between the reasons and the decision, that we explain as free will.
    In most arguments for free will I’ve read, this gap tends to be brought up almost universally in the context of decision-making, which is already a little weird. Does free will only ‘turn on’ when we have a conscious decision to make? None of our other conscious experiences seem to give us evidence for free will – which to me makes it likely that something else is going on here.

    Instead of characterizing the gap as metaphysical, I think it should be talked about in epistemic terms – the same faculty that makes the decision is the faculty we gather important information with, and self-reflection falls under that umbrella. Experience of the gap is just that – a gap in our knowledge.
    People reliably make an objection along the lines of ‘If my decision is caused by preceding events, then it’s not my decision.’ Again, this is just a consequence of the very specific function of decision-making: it would be silly to say “if my hunger is caused by preceding events, then it’s not my hunger.”

  14. Tab Atkins says:

    Reality being non-deterministic does *not* satisfy the folk definition of “free will”. If a decision is made solely based on a coin flip, people won’t say it made of someone’s own free will. (Substitute an actual non-deterministic process for the coin flip, of course.)

    The definition of “free will” is much closer to “goal-directed, but not entirely predictable, behavior”. This can exist in deterministic and non-deterministic universes, because it’s a side-effect of our own predictive capabilities and lack of knowledge. As your predictive capacity increases, less things seem to have free will.

    For example, to primitive people the weather seems to have a will (explaining at least part of the whole weather-god thing everyone had going on). Our knowledge and predictive capacity is much greater now, though, so we no longer consider the weather to have a will.

  15. A deterministic universe will be non-deterministic within the effective light cone and scale detail that are important for your determination of what “free will” is.

    Assuming that:

    1) You could have your 20x20x20m cube;
    2) and that it would simulate an isolated (for all intents and purposes, a limited light cone) system;
    3) and that you could create an instant snapshot of said reality at a level deemed enough for a deterministic outcome within a timeframe tight enough to not affect said determinism;
    4) and that you had enough computing power outside that system that could simulate said snapshot at a rate faster than “real” time without affecting determinism;
    5) and that you could receive results from that simulation without affecting it.

    You could then put yourself inside that cube with all the books, food, air, etc. needed to sustain this learning marathon of yours, snapshot it, leave, run the data set faster than real time, and see these blog posts showing up from your simulated self, and decide if the results were good enough for you to launch yourself (or not) into this endeavor.

    a) Your outcome would depend not only on an isolated 20x20x20m cube, the rest of the planet could change the outcome. Solution: simulate the whole planet on a bigger computer outside earth.
    b) At a meta level, it’s still deterministic, because you can’t escape your later decision. See the “what’s your definition of free will”.

    Basically, “free will” is an indeterminate concept, depending on the light cone and the level of detail (if determinism depends on the whole universe and at the most basic level, you cannot have enough computing power to escape determinism, *within the system*). Ipse Deus non potest computari.

    • Evan says:

      the snapshot would invariably include that you were going to leave the cube at some point to check the computer’s findings. Basically, either you never leave the cube, in which case the computer will have the outcome of your experiment before you complete it, but you won’t have access to that information. OR you do leave the cube, in which case the computer KNEW you were going to leave the cube at that point.

      And if the computer knows that you leave the cube at some point, but it doesn’t know anything about the world outside the cube, then it wouldn’t be able to predict what would happen if you went back to the cube to continue studying. In order to make that prediction the computer would need to be larger than itself and so you have the same problem as before.

      What you could do theoretically, is program into the computer a model of someone very much like you except that this clone of sorts will actually contentedly live out its life studying science in the cube, then you could get information from that. You could even open up the cube door and let him out if you wanted to, it’s just that HE couldn’t let himself out.

  16. Faré says:

    In my essay fmftcl I explain why physical determinism is wholly irrelevant to the problem of free will.

  17. Stephen Ware says:

    I think that “free will,” as the term is often used, does not even make sense. Those who believe in free will would probably not be comfortable equating it with random will.

    Randomness is no more “free” than determinism… indeed, it is less so because at least with determinism a decision can be said to come from the agent itself.

    Before we can ask whether or not humans have free will, I think we need a definition for it that is something other than random will. And I don’t know what that definition would look like… hence, I don’t think “free will” is even a meaningful thing to discuss.

    I find the idea of deterministic decision making much more comforting than decision making that is influenced by randomness.

  18. Rylan says:

    The best way I can think to explain my idea of free will is that we look at it the wrong way. We say “The universe is deterministic, so we end up at the same point no matter what,” or, “we have free will, so we have no determined end to our lives.” Both of these are wrong.

    The concept of free will and determinism stem from humanity’s lack of and want of foresight. We don’t know with certainty what will happen, and we find that frustrating. We deal with that with the idea of fate. We like to think that everyone is fated to something (or like to not believe it, depending on the person.) However, stating that there is or isn’t free will absolutely cannot be true, entirely because of our lack of foresight.

    As an individual, when you commit to an action, you lead yourself down an irreversible path. Once I eat a sandwich, I cannot un-eat it. This has some level of effect on the rest of the actions in my life, to an undefined degree (maybe I get food poisoning, maybe it’s so good it inspires me to be a chef, maybe nothing.) Since we do not know the outcome of the action until after it has occurred (hindsight is 20/20,) we cannot say that there *is* a predetermined outcome. That is not to say that there isn’t a predetermined outcome, only that we, as humans, cannot name the outcome accurately before the event. What this leads us to is this: We have free will, in that we make choices that can lead us to an infinite number of possibilities, none of which we can predict beforehand. Fate still exists though, because actions collapse us into a determined and irreversible reality. I believe Zach called this the “Religious Argument.”

    • rash says:

      the point is not whether you can yourself predict it, it’s whether it is *theoretically* predictable.

      going with the 20x20x20 box thingy, if the computer manages to always perfectly predict what the person inside will do (and this would be possible in a deterministic universe once the computer has data on everything in the box and all the laws of physics perfectly, provided nothing from outside the box interacts with what is inside the box and vice versa) then you have no free will.

      the way i see it, we probably don’t have free will, but if we do, i don’t see where it can come from except QM (and i don’t know the mechanism for it if it could come from QM)

      whether it is practical to be able to predict something doesn’t stop it being predictable.

      obviously just being able to act randomly wouldn’t be considered free will, so i’m not sure how a non-deterministic universe could lead to free will exactly, but as far as i can see, a completely deterministic (and so predictable) being CAN’T have free will.

      • dlorde says:

        Determinism does not necessarily entail predictability in chaotic systems, and there is now good evidence that there are chaotic systems at work in the brain; simple feedback circuits can easily produce chaotic outputs. So if we accept the suggestion that free will is a function of unpredictability, we could have free will in a wholly deterministic universe.

        But are we saying that free will is simply an expression of unpredictability? I think not – expressions of free will may be unpredictable, but it seems to me that, more often than not, they are predictable. Given some knowledge of an individual and their past behaviour, we can generally predict their ‘free’ choices – to jump off a cliff or not to jump, to pick chocolate or vanilla ice-cream, etc.

        We need to consider what people mean by free will. As best I can make out, ‘free’ suggests not constrained or coerced (as in ‘of your own free will’, rather than ‘against your will’), and ‘will’ suggests ‘what you want or wish to do’. So to exercise free will would be take the action you wish to take without coercion or constraint. I believe this is the ‘libertarian’ interpretation of free will.

        What we want or wish to do can be described in terms of our internal physical and mental state, our programmed drives, our learned preferences and goals, etc. These influences are the internalized result of multiple interactions of external input with previous internal state (and ongoing internal processing, involving feed-forward and feedback). So ‘will’, in this respect, is a somewhat fuzzy concept, depending on the degree to which we accept certain internalized concepts (e.g. morals, legalities, social niceties, etc.) as valid constituents of it.

        Freedom from constraint or coercion seems to be to do with judgement, or feeling. It is about how we judge our current options, or the consequences of our current options, with respect to our ‘will’. Precisely how we judge our freedom depends on how we distinguish between self (internal) and other (external). The demarcation isn’t clear because there are conflicting ‘loyalties’ between, for example, self (individual/personal) and other (e.g. partner, family, group, society).

        The boundary between freedom and constraint doesn’t necessarily correspond to the boundary between internal and the external – for example, the desire to take an action may conflict with an internalized moral imperative (‘I want to do X, but I know or believe that X is wrong’).

        In the appreciation of both ‘free’ and ‘will’ there is also the implicit requirement of sufficient awareness of our desires and of the implications or consequences of our actions, so that we know the potential constraints upon them. For example, if we know an action will be judged illegal or immoral in a particular context, that knowledge may act as a constraint on our freedom to act in accordance with our predilections.

        However, If we refrain from an action we desire to take because we know it is ‘wrong’, are we exercising our free will to make a moral choice, or suppressing it in favour of an external imperative? – or does it just depend on how you view it?

        I suggest that free will is little more than the feeling that you can take some action without constraint or coercion, but that it doesn’t bear close examination (because of the confusion between internal and external motivation, etc) – it’s an superficially convenient but actually meaningless abstraction.

  19. Harvey says:

    I find that even if you didn’t bother reading zach’s posts, he’s a phenomenal curator of thought-provoking wikipedia articles.

  20. Michael says:

    When did the definition of free will become “magical causality defying powers”? When you define being “free” as being unpredictable and uncaused, all you’ve got a meaningless definition of freedom.

    Free will is being able to do what you want without any intervention. Whether or not it can be predicted by a third party is absolutely irrelevant. Meanwhile, the fact that it is caused is a good thing, because it means your decisions are bason on your beliefs and desires, rather than being random (the alternative to causality). If I eat ham sandwiches because they’re my favourite food, I don’t have an existential crisis when my friend says “based on that fact, I predict you’re going to have a ham sandwiche for lunch today”. I say, “that’s fine, my action is easily predictable because I have a good reason for it.”

    Furthermore, consider the alternative. “What’s even weirder is that you can imagine a species that lacks free will, but which is at some point able to install a machine into itself that gives it access to free will! I’ll have to write a sci fi story about that one day.” This shows to me that you don’t properly understand how free will works. So, what? They would go from eating ham sandwiches because they’re delicious, to eating BLTs because BLTs don’t reflect their preferences in any way, and therefore can’t be predicted. Is that free will?

    I suspect this whole paranoia surrounding predictability is a psychological thing. We’ve learned/evolved to be incredibly averse to being manipulated, to the extent that if we’re told we’re being manipulated, we’ll go out of our way to falsify whatever we were supposed to do, even if going along with the manipulation would in fact be in our best interests. And, being predicted and being manipulated have enough overlap that a lot of negative reactions to prediction got to come along for the ride.

  21. Pingback: Free Will Is as Real as Baseball | Cosmic Variance | Discover Magazine

  22. Pingback: Free Will Is as Real as Baseball | Cosmic Variance – Will

  23. M. says:

    First, an addition to your analysis of the “computer that predicts future actions.” Besides the reason you stated, there is another issue: if the computer adds the prediction into the system, that changes the system, and changes the outcome. In other words, I would have had a ham sandwich for lunch; but the computer’s prediction of that event, once added to my deterministically contrarian mind, changes the pattern, making it deterministically certain that I will not eat a sandwich.

    Second, yes, we do have a randomness generator in our brains. I can’t find the reference now, but there was a paper a few years ago that found essentially that, in the parietal lobe IIRC.

    Third, something under 5) in your option selection: a nondeterministic input into a deterministic universe. This is an unlikely, but not impossible option. Essentially, we have a very hard time explaining consciousness; one option, again possible but unlikely, is that it is a product of a completely separate process within the universe, one that could be indeterministic in principle. If so, this could, conceivably, spill into decision-making processes. But it is a tall order, which becomes that much taller the more we learn about the universe.

  24. Pingback: Free Will Is as Real as Baseball | Cosmic Variance | Discovery

  25. Pingback: Rum and Reason » Free Will Is as Real as Baseball | Cosmic Variance

  26. Pingback: Evolutionary hysterically hysterics (rather the mechanisms of natural selection) « El Guapo Blog

  27. Trey says:

    When embroiled in an unanswerable philosophical question, go deeper. Cause and effect is the root of these ruminations. Being that it causes (!) so much confusion, maybe the answer lies in a deeper understanding of cause and effect itself. Lots of past masters of Buddhism say there is no such thing. Which supports neither of the embattled models.

  28. Nick Beat says:

    I’m just gonna spew some loosely relevant Existntialism here:

    I’m 100% certain that the future does not exist, nor does the past. All we have are memories and predictions. We exert free will in our every action, though our choices are influenced by the predictions of others. Someone may say, “Don’t jump.” But you do it anyway cuz YOUR predictions tell you that you can make it across. If we choose not to, it is because we CHOSE to listen to the universe.

    It is our ability to make these predictions that make us a moral race, for if we can’t postulate the consequences of our actions then we cannot be held accountable for them. We are not alone of course; Dogs know if they are gonna get smacked for eating a shoe, Cats learn to fear the spray..

    Certainly a psychopath will have trouble making good decisions because their grasp of the present is faulty. A sociopath may be incapable of predicting the fate of others, only themselves. Or perhaps they can but they just don’t care.

Leave a Reply

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

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>