Saturday, January 15, 2005

Free will and determinism: a physicist's perspective

This is an amusing topic I've discussed with many people over the years. My opinion: if our current understanding of physical laws is correct, humans have only the illusion of free will.

What do I mean by free will? Well, that is hard to define. But the absence of free will or deterministic behavior is not hard to define. In classical physics the future evolution of a system is completely determined by the state of the system (including time derivatives) at any instant. For example, suppose that at some instant t I know all of the positions and momenta of every molecule or atom in the universe. Then, in a classical universe, I could in principle predict with certainty its future evolution. Another universe prepared in the identical state would evolve identically. So, in a classical universe everything that has happened since the big bang could (again, in principle) have been predicted at that moment. Such a universe is deterministic, and no subcomponent of that universe (i.e., a dog, a human, a robot, or a martian) can have free will.

Note that chaos does not help. Suppose the classical evolution is nonlinear, and exhibits chaos. Then initial states which are very similar may, after only a short time, have evolved into radically different configurations. This makes the job of predicting the future behavior of some initial data very hard, but nevertheless possible in principle.

Does quantum mechanics help? Well, a quantum universe is not deterministic. (In the many-worlds description, the evolution of the wavefunction describing all possible universes is deterministic, but only a single branch of the wavefunction is perceived by any macroscopic observer, and the path followed by a particular observer at each branching is not predictable.) However, it seems that the functioning of our brains is almost entirely classical. The number of atoms involved in the firing of a synapse, or other chemical reactions in brain functioning, is quite large, and there is little quantum coherence. As far as we know, our brains are readily simulated by purely classical processes.

Still, the inputs received by our brains may be non-deterministic in a quantum world. For example, the clicks of a geiger counter (which detects the photons from individual radioactive decays) are intrinsically random according to quantum theory. But, it is hard to see how introducing occasional random inputs amounts to granting free will to an otherwise deterministic machine. A classical robot which is fed occasionally random data does not seem to me to have free will. If prepared in an identical state and fed the same inputs, it would behave the same way every time.

If the previous discussion is correct, all we really have left is the illusion of free will. It seems like I chose what to have for lunch today. The conscious part of my brain didn't know what the answer was until going through what seemed to be a decision process. But, given that there are lots of lower-level brain processes that my consciousness has only a vague awareness of, it seems quite plausible to me that the "decision" was made in a deterministic way without my knowledge.

20 comments:

Russ Abbott said...

Choosing what to have for lunch is one thing. Deciding what to write in a blog piece seems like another. How does one talk about the latter in terms of the illusion of free will?

I suppose a distinction between free will and consciousness helps. One might have illusary free will without giving up consciousness, in which case consciousness is simply a presentation to oneself of what one has already decided. (I think that's Daniel Dennet's position, but I didn't read his book on the subject.)

But if that's the case, what's the evolutionary purpose of consciousness? Why are we constructed to create this illusion for ourselves?

Steve Hsu said...
This comment has been removed by a blog administrator.
Steve Hsu said...

Well, if a deterministic robot could decide at some point to write a blog, I suppose I could too... As you point out, I am finessing the issue of consciousness here. I am assuming that it can arise within classical physics from a sufficiently complex information processing device ;-)

I did read Dennett's book many years ago, and I believe I agreed with him at the time.

I think consciousness is the evolutionary by-product of requiring more and more complex decision-making ability from organisms. There is utility in having a creature that is aware that it is hungry, injured, scared, etc - and as a byproduct, bored, fascinated, sad?

Anonymous said...

Russ Abbott

"I suppose a distinction between free will and consciousness helps. One might have illusary free will without giving up consciousness, in which case consciousness is simply a presentation to oneself of what one has already decided."

Nicely phrased. My father would have agreed with Steve, but he would have added strongly that the illusion is quite enough to hope we might behave morally. Darwin's finch has had a hatch late in summer and now it is fall and the nestlings are not fully fledged. The weather is turning, finches are flying south. The mother bird sits on the nest but is pulled south, pulled south. At last the mother leaves the nest and the nestlings and begins to fly south. Yet, yet, there is the memory for a time of the nestlings. In the memory there is waking conscience. The mother finch might even return.

Anne

Anonymous said...

Wonderful post as usual. We must have a post of the arts now and then :)

Anne

Anonymous said...

Note: William James, Does Consciousness Exist? and Will to Believe. Pleasing reads.

Anne

Steve Hsu said...

Anne,

The Web is really great - I was able to find the James essay here:
http://psychclassics.yorku.ca/James/consciousness.htm

The problem with discussing free will is that hard-headed physicists will ask "How do you define free will?", which is why I approach from the opposite direction of determinism. Consciousness is similar to free will in that we all know what it means intuitively, but have trouble defining it in any rigorous way.

I'm surprised when talking to philosophers that they don't seem to be interested in the strength of the classical physics point of view and what it says about determinism. My feeling is that people are happy to accept that "quantum mechanics gets us out of that problem" without examining whether it does in more detail.

Anonymous said...

William James, Does Consciousness Exist?

'Thoughts' and 'things' are names for two sorts of object, which common sense will always find contrasted and will always practically oppose to each other. Philosophy, reflecting on the contrast, has varied in the past in her explanations of it, and may be expected to vary in the future. At first, 'spirit and matter,' 'soul and body,' stood for a pair of equipollent substances quite on a par in weight and interest. But one day Kant undermined the soul and brought in the transcendental ego, and ever since then the bipolar relation has been very much off its balance. The transcendental ego seems nowadays in rationalist quarters to stand for everything, in empiricist quarters for almost nothing....

I believe that 'consciousness,' when once it has evaporated to this estate of pure diaphaneity, is on the point of disappearing altogether. It is the name of a nonentity, and has no right to a place among first principles. Those who still cling to it are clinging to a mere echo, the faint rumor left behind by the disappearing 'soul' upon the air of philosophy. During the past year, I have read a number of articles whose authors seemed just on the point of abandoning the notion of consciousness, and substituting for it that of an absolute experience not due to two factors. But they were not quite radical enough, not quite daring enough in their negations. For twenty years past I have mistrusted 'consciousness' as an entity; for seven or eight years past I have suggested its non-existence to my students, and tried to give them its pragmatic equivalent in realities of experience. It seems to me that the hour is ripe for it to be openly and universally discarded.

Anne

Anonymous said...

It may be that consciousness doesn't exist as a thing in itself but that it is a necessary bi-product of the structure of our brains. The fact that it seems important to us (indeed it seems to be the purpose of us) doesn't make it so.

Anonymous said...

"It may be that consciousness doesn't exist as a thing in itself but that it is a necessary bi-product of the structure of our brains. The fact that it seems important to us (indeed it seems to be the purpose of us) doesn't make it so."

Please develop this interesting comment. Bertrand Russell argued along this line to a degree, if I understand the comment properly. I take the comment as arguing that a sense of mind body dualism is part of what it is to be human and introspective, even though the sense is deceptive for we are all body. "Feelings" seems as though not connected with body, but this is not so as William James pointed out initially. We cry and feel sad, we smile and feel happy. Feelings are physically generated. Interesting.

Anne

Anonymous said...

Very fascinating.

The only "weak link" in the argument IMO is the following.

"However, it seems that the functioning of our brains is almost entirely classical. The number of atoms involved in the firing of a synapse, or other chemical reactions in brain functioning, is quite large, and there is little quantum coherence. "

Even if there are no macroscopic quantum effects (as in superconductivity or Bose-Einstein condensation)
what about the firing of the synapse itself? In what way is synaptic process "classical"? In what way is it different from radioactive decay, for instance?

"As far as we know, our brains are readily simulated by purely classical processes."

I guess I do not know enough about brain science. If true, hmm... very thought provoking, to say the least!

MFA

Steve Hsu said...

MFA,

Your point is well taken - when I said simulated by classical processes I meant classical statistical mechanics, which nevertheless isn't completely determininstic, as there may be a random element. (I say may be because if the number of molecules involved is large enough the behavior is pretty predictable.) So, perhaps the most honest model is deterministic machine with small amount of stochastic noise.

But it seems to me that small bits of internal randomness in the brain do not give rise to free will any more than small bits of external randomness (i.e. inputs from a Geiger counter). Thanks again for your comment!

Steve Hsu said...

Max,

It seems to me there are lots of moments when decisions just happen. I honestly can't trace back specific reasons for many small decisions during the day, like whether to click a link or stand up and stretch at a particular moment.

The reasoned decisions like cold vs delicious food seem to have good evolutionary purposes of optimizing several factors which impact gene survival.

This is where issues like "can consciousness arise inside a deterministic machine?" are relevant. If it can, I imagine there might be higher level processes (consciousness) which are unaware of lower level processes which impact what appears to be decision making.

Carson C. Chow said...

The neuroscientist's perspective would agree that there is no free will. However, arbitrarily small perturbations probably do alter decisions. Thus, unless we can monitor every single input, including thermodynamic fluctuations in the movement of ions and neurotransmitters, I don't think we can predict too far into the future what a person will do exactly. We can probably predict what a person will do on average though.

Anonymous said...

Actually, the notion of free will is highly problematic on purely philosophical grounds. The problem is very simple and is clearly stated in many texts, but in my experience very few people, even among the highly educated, are aware of the problem. Here goes:

Every action we perform must ultimately be traceable to an external cause (genetics, environmental influence, divine intervention, whatever) or else there is something in human behavior that occurs _for no cause at all_. In the former case we have determinism (which everyone sees rules out free will). But in the latter case we have that something in human behavior is random: we do things for no reason at all. That's indeterminism, and quantum mechanics could conceivably introduce it into the human brain, but it's not free will either, at lost not in the way that any ordinary person thinks of free will.

Steven Weinberg says (and I agree) that Charles Sanders Pierce said all that can be said about the problem of free will when he claimed that all that matters is that we have _the subjective experience of choice_.

Steve Hsu said...

Wait, did I miss something? If you are arguing on purely philosophical grounds (not within the context of physics) you are not restricted to either an external cause for each of our actions or simple randomness. Why not a "soul" existing in some higher realm which is not governed by physical laws? This is the "traditional" (albeit intrinsically unscientific) response to the free will question ;-)

Anonymous said...

"Steven Weinberg says (and I agree) that Charles Sanders Pierce said all that can be said about the problem of free will when he claimed that all that matters is that we have _the subjective experience of choice_."

Nicely argued.

Anne

Anonymous said...

Well, I'd say that the traditional conception of free will is something more than "unscientific." It requires us to suspend logic beyond a certain point. Because if you claim you have a soul, the same questions that applied to the brain will logically apply to it too. Do all the decisions of the soul have an ultimate external cause (for instance, God's will) or do they not?

The Calvinists, for instance, traditionally don't believe in free will, because they hold that God predetermined every human action. I guess most other religious people think that at some level the soul somehow "moves itself." But logically that would imply that the soul chooses _without a cause_ (i.e., randomly). Which, religiously, is not very satisfying.

But I guess if you're religious you can always at some point adopt a mystical rather than logical outlook.

P.S.: That should have been Charles Sanders Peirce (not "Pierce") in my earlier post.

Anonymous said...

Charles Peirce was ultimately more insecure than William James and longed for encompassing or absolute answers to philosophical or psychological questions where James was content with conditional observations and answers. Remember "the Fox and the Hedgehog?" Well, Peirce knew he should be a fox and know many things but oh so wanted to be a hedgehog and know one big thing. For James the subjective experience of choice was enough, Peirce wished wished wished there were more.

Anne

is72 said...

you fail to consider the compatibilist view, on which even a strictly deterministic universe may be endowed with free creatures just in case their desires and their actions line up properly. but even if incompatibilism is true, and free will does not exist if determinism is true, your post fails to show that free will is an illusion.

you say: "A classical robot which is fed occasionally random data does not seem to me to have free will. If prepared in an identical state and fed the same inputs, it would behave the same way every time."

so? why does that preclude free will?

libertarian free will requires that the future not be determined by the past and the laws of nature - and that the agent is responsible for the performance of an action. the sentence of yours that i've referenced does nothing to undermine this view.

actually, it speaks to another, also important, question: whether this type of freedom is more valuable than compatibilist freedom or than none at all. i believe this freedom is more valuable. but this was not the question your statement meant to answer, and so i won't expand upon this, different, question. i only mean to expose what is wrong in the comment you did make.

Blog Archive

Labels