Free will and deterministic predictions
There are many who believe ardently that free will is just an illusion, if not for any convincing positive proof, but merely for lack of convincing proof that it is real. So, in an effort to understand the problem better, I have constructed a thought experiment that might help the case of free will just a little bit.
Imagine, hypothetically, that the world did indeed operate fully deterministically, meaning that, in theory, should all the laws of physics be absolutely and fully understood, it should be possible to plot out the exact history of the universe from the Big Bang (assuming that was the actual beginning) to the present. This model would seem to admit of only one possible universe, and we would be able to see retrospectively that all our seemingly free choices were in fact not free, and that the illusion of freedom only came from some outside factor that was beyond our control or even knowledge, which itself was rooted in some other factor. So far so good. It would seem hard to deny, given that we were witness to what this model showed us, and that it did indeed correlate with the real history of the universe as far as we know it, that the universe must indeed operate deterministically. (This of course is false, because we began the thought experiment by imagining a deterministic universe, but bear with me.)
Now the model, apart from only being able to show us the past with absolute accuracy, of necessity must be able to show us the whole future as well. Here lies the rub, because we will be able to see all our future actions mapped out before us, all the consequences of those actions, and so forth. When we continued our daily life, either the illusion of free will would have to be completely removed, or we will indeed be truly free, and be able to prove those deterministic predictions completely wrong. How the illusion of free will can ever be taken away is unimaginable, and the idea that simply seeing the future is enough to instantaneously change the way our minds work is preposterous. And of course, if it is possible to prove deterministic predictions wrong, then clearly those predictions can’t be deterministic in the first place, since they should be logically impossible to disprove.
There is one powerful objection that I know of that might be raised against the above reductio. That is, it is logically possible to disprove a deterministic prediction without free will entering the picture—for instance, via the use of a binary inversion machine that, upon seeing that its predicted outcome is “1″, would simply invert that, and output “0″. This, I think, can be bypassed quite easily, because there is no reason that any prediction whatever must be made of something that only operates on the input of predictions. This can be complicated with the addition of an intermediary who tells the inversion machine the prediction. In this case, though, the machine poses no problem, because the prediction need only speak of what the intermediary tells the machine (I’m assuming here that the intermediary has no free will).
However, this does not take away from the power of the objection itself. After all, it is not true a priori that a deterministic prediction can only be proven wrong with free will. I would argue, though, that the defender of determinism (or at least fatalism) has the burden of proof here. As I have shown, the problem of the simple machine that inverts its predicted outcome can easily be overcome (in fact it isn’t really a problem). If one wants to disprove free will, then one has to take the first step of devising a way to disprove a deterministic prediction in a way that does not invoke free will. Until anyone does this, I will have to tentatively conclude that I do indeed have free will.
on 
on 
13 comments
This is a fun one. Quantum and chaotic limitations on predictions may give the anti-free will person lots of leeway here, but I won’t focus on that as it is murky and unclear.
When we continued our daily life, either the illusion of free will would have to be completely removed, or we will indeed be truly free, and be able to prove those deterministic predictions completely wrong.
This is a false dichotomy. I would still probably have the illusion of free will, just as knowing a visual illusion is an illusion doesn’t make it go away.
I could still have the illusion, think I’m free, but pace the predictions of the model perfectly. The prediction machine would predict our reactions to seeing it (it has to include itself in its predictions, of course). It would predict we would fight its predictions. It would be spooky, but we would be there doing everything it predicted, thinking we were doing it freely. It would predict our desire to outsmart it, and we’d have the spooky experience of “freely” trying to outsmart its predictions, but no matter what we did, it would be right.
My point is that you can’t just assume that we would be able to act differently than its predictions, as that begs the question. That is, the bald claim that we could show it wrong begs the question. I don’t share that intuition, in the logically possible universe you have described.
But if a system is properly deterministic, it should be linear: it should be possible from any point in time to be able to tell what happened at any time before, and what will happen at any point in the future, in such a way that it just couldn’t be otherwise. If it’s linear, then the prediction can only speak of one possible event at any one time. So no matter how much foresight it has about my reaction to the prediction, once it has made it, it can’t take it back. I would then be made aware of the prediction, and am free to prove it wrong.
I think the problem with your response is that it assumed a sort of “live” prediction machine which is always making new predictions, which is not what I envisioned. Imagine instead that the machine does its calculating in a single instant, including all the past and all the future. In that case, I don’t see how it could possibly factor in all the reactions to its predictions: it seems that it would spend an infinite amount of time calculating the infinite amount of possible human reactions, and so we can never in principle know whether human action is governed deterministically.
I also don’t think it’s much of a leap to assume that we would be able to act differently than its predictions. After all, we can contradict predictions in daily life with ease. Why should the prediction machine make us suddenly unable to do that?
I was assuming it spit out predictions far into the future, that they are fixed. It wouldn’t make a difference either way in a deterministic universe.
The machine, when turned on with all the right initial conditions set in its paramaters at time t1, will include your psychology, all of your neurons, in its predictive apparatus. So of course it will predict all of your reactions to learning about its outputs. It will correctly predict every episode in which you look at its predictions, and your reactions, your attempts to work around it, for the next fifty years. If the initial conditions are set up correctly, then why should it ever be wrong?
Of course, if you were to loosen the assumptions, and assume for the reductio that the machine took into account quantum mechanics (so it was a statistical prediction machine), then it’s not clear your argument would work either, because in the very near future the predictions would be of very different worldlines. I guess you could say we could try to violate those predictions, but things get more murky.
I think your idea is fun, and does push on intuitions in an interesting way, but it’s not clear it can do more than act as a touchstone for our preexisting biases. I would predict we wouldn’t be able to get out of its predictive prison. You don’t. How are we to mediate this disagreement without further argument?
One thing I should clarify: I tend to be a compatibalist. Free will is a psychological category, is a psychological reality. E.g., animals can “choose” from among different alternatives offered, and work toward them. I don’t see the biology being any different in kind from neuropsychological explanations of other psychological phenomena. I only disagree with the metaphysical claim that free will implies that we act according to more than is proscribed by the laws of physics, basic biology, etc..
Hence, I don’t like when these debates are posed as between ‘free will’ and ‘determinism.’ We already have good reasons to beleive determinism is false. The contrast should be between those that believe we are bound to the nomic regularities in physics, and those that believe we can violate the laws of physics in certain circumstances.
Also, ‘linear’ is not quite the word you want. There are plenty of nonlinear deterministic systems (e.g., neurons). ‘Deterministic’ captures what you want just fine, I think.
You’re right. “Linear” isn’t the word, but I can’t for the life of me remember what the right word is. And I completely agree that my idea doesn’t really disprove determinism, but I feel it could be the starting point for a more developed argument against it.
Part of the problem with using the word “deterministic” is that it is of course possible that determinism and free will are true at the same time. Like you, I also don’t like the idea of a free will (or anything else) that completely contravenes the laws of nature. It is perfectly reasonable, though, that there can be a law of nature that allows two or more possibilities of choice, neither of which can be predicted in principle (my very basic knowledge of quantum and chaos theories tells me this). But what I want to argue against (call it determinism, fatalism, or what you will) is the idea that free will is an illusion. Something, I think, must account for the pervasiveness of the illusion (if that’s what it is) of free will. I think it’s fair to discount any notion of a Cartesian deceiving demon.
By the by, I think that one possible solution to the dilemma might come from evolutionary psychology, specifically in its potential to explain why we feel we are making free choices. If this can be shown to be real, then the anti-free will position will be further corroded.
Ramana Maharshi pointed out that the only choice you really have is whether to identify with the body or not.
And that the moment you identify with the body you become helplessly bound to all of its yes/no/maybe dichotomies and all of the universal patterns of relationship within which the body is entangled, including the karmicly determined/inherited patterns of the particular body-mind-psyche that you have chosen to identify with.
Which of course is going to disintegrate and die, and you know very well that this is so.
The world of klik-klak and its relentless pattern patterning. Klik-klak being completely indifferent to the well-being and survival of any and every biological form that arises within the pattern.
That sounds interesting but very confusing. I looked up Maharshi on Wikipedia, but still don’t quite understand…
This is interesting, but I’m not sure the thought experiment is actually possible. In creating a model of the future based on the laws of physics, etc, we will have to take into account the impact that the predictions will have on people’s behavior (as has been noted). So imagine a simple case – a stubborn person who is determined not to do anything that it is predicted he will do. The predicting machine would simply be *unable* to make accurate predictions to this stubborn person (at least in many cases) – but the machine would know that it couldn’t, based on its deterministic calculations of how the person would react to any predictions he was given. So, it’s not clear how any of this undermines determinism…
Yes, I see your point. But what do you mean by “the machine would know that it couldn’t” make an accurate prediction? Remember, it’s just a machine that’s been programmed to map out the universe from beginning to end. If it gets to the point that the machine simply can’t work out what someone will do, then this seems to me to show that our actions can’t be overdetermined. If they were, it would have to make the correct prediction, would it not?
Consider a situation in which there are only two options for the stubborn fellow, X and Y. The machine can calculate that if it predicts X, the person will do Y, and if it predicts Y, the person will do X. There is no failure on the machine’s part to calculate what the person will do, nor is there a failure of determinism – the inability to make an accurate prediction to this person is simply a result of the fact that the person’s knowledge of the prediction will be a factor in determining his behavior.
I agree completely, and I mentioned in an earlier comment that I think free will and determinism are compatible. I am less concerned with disproving determinism than I am with proving that free will exists. In the example you gave, the person may or may not be free. But suppose there’s a situation in which the machine predicts the following: “if I predict X, then the person will do A or B, and if I predict Y, then the person will do C or D.” There is nothing un-deterministic about that (as far as I can see), and it seems likely that if there is no way to distinguish in the prediction between two equally likely options, there is room for freedom of will somewhere. Still not proven a priori of course.
Now I am a bit confused. If it’s impossible for the machine – which by hypothesis knows every detail about the laws of physics and all the physical details of the universe – to calculate whether A or B will follow from X, then that by definition seems to be an indeterministic universe.
But back to your original thought experiment, the point I am trying to make is that there won’t be any deterministic predictions being proven wrong by anyone’s freedom. Rather, no accurate prediction can be made – which on the face of it has nothing to do with freedom (unless you have a particular way you want to spell out what freedom means), but with the simple fact that the prediction itself is a causal factor that determines the person’s behavior.
Forget the example I gave. It is a bit confusing. (My instinct is that it’s possible in principle for that example to be deterministic, but only if true randomness is possible, and I don’t know if it is…)
The fact that no accurate prediction can be made is pretty much the crux of it, and you’re correct in saying that my thought experiment doesn’t succeed in showing that free will exists. (In my defence, I wrote in my post that I had to “tentatively conclude” that I have free will. You’ve made that conclusion slightly more tentative.) However, all the myriad possibilities of predictions, predictions of predictions, reactions to predictions, etc., make the waters a lot murkier, and I think they’ll have to be waded through a lot more thoroughly before we can come to a conclusion either way.
I’ve always maintained a “soft deterministic” stance when thinking about free will. Basically, I hold that the person you are has been determined by your biological makeup and everything that has happened to you. So you’re actually making a choice “freely,” it’s just that the person you are, making the choice, has been largely determined already.
People looking at this from a revamped version of religious theories of free will (if it has to happen, or if God knows it will happen, is it free?) are going to confuse the forest for the trees here.
Leave a Comment