Is free will really "free"? Your actions can be predicted from your brainscan | Hacker News Reader

Is free will really "free"? Your actions can be predicted from your brainscan

opinionator.blogs.nytimes.com

10 pointsJun828 comments

These neuroscientists are very confused if they think they can say anything useful about free will, or evil, or the like. Unless they are going to provide empirical evidence against determinism itself, there's nothing those brain scans are going to tell them about these issue.

Will Wilkinson argues the point here:

http://bigthink.com/ideas/40538

I must confess, I don't understand your argument, nor the one at the linked blog post. Why do they need to show evidence against determinism?

There is a growing body of evidence that choices are "made" by the brain long before we are consciously aware of them. The most parsimonious explanation for these data is that the brain comes to a decision through a mechanistic process of integrating evidence and then the conscious self is informed of that decision. It's at this point that we feel that we have made a choice.

One nice prediction of this theory is that brain disorders that disrupt the communication of the decision making circuitry and the parts of the brain that are normally informed about it should produce the impression that ones choices are being externally controlled. And indeed there are a wide variety of disorders with these features (schizophrenia, alien hand syndrome, etc).

Some of the pieces are missing--what is this pathway and which brain structures are involved in the perception of agency--but I think it makes a lot more sense than believing in non-causal free will.

EDIT: Another compelling piece of evidence comes from brain stimulation. When you electrically stimulate the surface of the brain (say, as part of a surgical procedure to localize a tumor) you will find areas that evoke movements of the arm, eyes, legs, etc. The experiment is: you do this without telling the patient exactly when it will happen and then afterwards ask them, "why did you move your arm?" Many people will confabulate and say, "I felt like it" or "you asked me to". This isn't always the case: sometimes people will report that their arm felt like it moved "on it's own", but I think it is very suggestive.

I am the thing making decisions. To the extent that the thing making decisions didn't exist exactly when and where either I or the neuroscientists thought it was, that's merely one in a very long list of things that we were wrong about, and that our definitions were wrong. Declaring that this disproves "free will" is really overdoing it; it's akin to declaring "gravity" disproved after astronomers established that gravity works in space as well as on Earth. Yes, sure, "gravity pre-better-astronomy" was disproved, but that did not cause things to start suddenly floating off the Earth. First you draw a dichotomy where there shouldn't be one, then you demonstrate the dichotomy doesn't exist, then you declare that the whole thing is all wrong; it's just "A -> B, not A, therefore C" when you get down to it.

I am already well aware that the-thing-making-decisions is more than just the conscious mind itself, but while we have the words "subconscious" and "conscious" in reality there's no clean line between the two.

  > I am the thing making decisions.

Well if that's your premise, then I think we're already done.

If someone makes a decision, but doesn't tell you about it, is it your decision? What if they make a decision and then inform you about it later, is it yours now? Finally, what if they make a decision, tell you about it, and then you actually enact it? Can we even call this yours?

>Well if that's your premise, then I think we're already done.

Why call one part of my brain me, but not the other part of my brain me? I took jerf's point to be that the part doing the awareness/recording can be different from the part deciding, but that doesn't mean the part deciding isn't me.

I don't personally think there is a single, unitary self-- a man in the head so to speak. It may be a convenient useful fiction but it probably misleads our intuition on these matters.

You're injecting the term "someone" as if it has some sort of meaning. What is this "someone" who isn't me? Define it, precisely. How do I determine whether some neuron in my brain is either mine, or this "someone"'s?

I'd contend this "someone" is an artifact of your thought processes, not a real thing that you can actually define, and furthermore when neuroscientists disprove the existence of this "someone", it is their own artifact that they are disproving, not anything that corresponded to reality in the first place. This is an interesting result, just as in the gravity example I gave earlier, but it certainly seems to be very common for scientists to get a little too expansive about the artifact they have disproved.

The point is that if physical law is deterministic, then none of these experiments are going to tell you anything about free will that you didn't already know. Yes, they tell you that people are confused when they say "I decided to do X" when, in fact, the normal decision making part of the brain wasn't involved with doing X because X was actually artificially stimulated by the experimenter's electric probe. But if you believe in determinism already, what does this tell you about the possibility of free will? Given determinism, we already know that all actions and choices can be traced back to previous causal events, so we already know people are confused when they think that their choices are somehow not caused by previous events but come from their soul or whatever.

So none of the empirical facts are under dispute. The question then is a purely philosophical one: is there a sensible notion of free will given determinism? Should (normative claim!) we treat people like they have free will, i.e. punish/reward them for their actions?

> "The point is that if physical law is deterministic..."

Under current QM, it doesn't appear that things like alpha decay happen predictably or at one time vs another for any reason. And that isn't simply a problem of not knowing the state of the system; hidden variable theories have not met with empirical success.

So, imagine we could have two universes, both with our laws of physics, and identical up to time=0. Current QM allows that at t=1 in one universe radioactive decay occurs which then produces chain of events A, while in the other universe at t=1 there is no decay and thus chain of events B. Why then call these universes and their physical laws deterministic? They seem rather not.

I'm a physics grad student doing work in quantum decoherence, so I'm aware of all this. When discussing free will, "determinism" really is shorthand for the idea that the cause of all events can be traced back to previous events in known physics, including fundamental probabilistic events. The key claim is that there is no "soul" (or any other causal source which cannot in-principle be explained by known physics) producing choices.

(Yes, you bundle up all the probabilistic quantum events which occur in your brain and identify that as "free will", but I don't think most people would find that compelling. Not the least of which because then it seems like a plutonium atom has just as much right to the term "free will" as you do.)

Well, I'm glad then that we get to skip the bit were I have to spell out the facets of the argument I made. (Over on reddit bringing this up usually means me having to gently prod the other person to read up on hidden variable theories so they can tell me if that is what they meant by phrases like "something determines it we just can't see what".)

I'm actually a compatibilist about "free will", so I've no need for a ghost in the machine or to be flipping quantum coins in my head. The point of my comment was simply to point out that "our universe is deterministic", full stop, isn't a warranted premise by current evidence. If you only meant "adequately deterministic" by "deterministic", fine and dandy. However, I don't know anything that actually stops the two causal chains A and B in my thought experiment from being macroscopically very different. So skipping the "adequately", I feel, gives those not familiar with the details a misleading impression that we can unproblematically think of ourselves as just in the old clockwork universe. (Thus the kinds of phrasing I see a lot on reddit from the loosely-informed on such matters.)

(Edits: silly grammar and spelling stuff.)

OK, I support being precise with our terminology. What I was trying to emphasize is that, with regards to discussion of "free will", the important thing about "adequate determinism" isn't predictability. It's that we can trace back all causes to known (possibly intrinsically probabilistic) physical "stuff", with no room for a soul.

Something I've been wondering, and I'm way out of my depth here - would determinism be something that could eventually find an explanation in information theory?

My layman's argument against determinism was something like: OK, even if my biological self is making my decisions for me, how does it get the inputs that drive those decisions? Imagine a simple exercise study, conducted without face-to-face contact. Half of the participants are told to exercise, and half are told not to. At the end, the half that is told to exercise will have exercised more often than the half that did not. Therefore, if the world is deterministic, what's "sending" that determinism from the study to the participant? My conclusion was: I dunno, information is flowing?

Though I guess that's where quantum mechanics comes in, and I guess there's probably also an information-theoretic aspect to that.

Not sure I know what you're asking.

If we ignore quantum mechanics, then all of our empirical evidence strongly suggests that the universe is strictly deterministic. (Although this conflicts with our intuition about our own consciousness.) What that means is that if we were to have the position and velocity of every particle at a given moment in time we could (in principle) calculate the position and velocity for any future moment. The important thing is that the known and understood physical "stuff" is just particles, and the only thing that can or need be known about them is their position and velocity. There is no thing-which-is-NOT-made-of-particles which we might call a soul or whatever.

In particular, you could in-principle predict what experimental subject Bob would chose based on the configuration of the particles in the universe before he was born.

So, is your question "How is that pre-Bob configuration of particles causing Bob's choice?" If so, then you just look to the fundamental laws of classical physics: particles obey certain rules which govern how they move around. Given a certain pre-Bob configuration, we would compute that the particles would all follow certain trajectories which eventually would lead to a configuration where many particles were assembled into Bob himself, and these particles moved in such a way (according to known physics) that Bob (who is composed of the particles) made a certain choice.

Or is your question about quantum mechanics?

I guess my question is about "how does that propagate?" I have no physics background, so maybe the answer is obvious, but I have trouble intuiting how, in my thought experiment, the experimenter and the subject have any shared state.

Oh, ok. That's just a manner of understanding how humans work (when viewed as a machine). The experimenter has some idea in his head which is encoded in the connections between neurons (although no one knows how...which is why it's hard to be totally confident that people are deterministic). This idea is translated into vocal cord vibrations and lip movements, which vibrate air in a certain way to product sound. This sound wave travels to the subject's ears. It is translated into electrical signals traveling on neurons by the ear, leading into the subject's brain. This changes the state of the subject's brain (the connections between the neurons) in such a way that a day later, while going through his normal routine, the subject's brain directs him to exercise.

D'oh. Of course. The state is shared along exactly the same media as the perceived, "intentional" communication. Thanks.

It should be pointed out that Bell's inequalities only rule out the existence of classical hidden variables being able to explain quantum effects, which if one understands the history behind the question makes sense. Technically it doesn't (and can't) rule out hidden variables that are "quantum" in nature (if you'll permit me the shorthand), but from our perspective those are the very definition of a needlessly multiplied entity in the Occam's Razor sense. Nevertheless, that simply means they aren't useful to us, not that they don't exist. It is certainly possible to imagine that we exist in a simulation in which the simulation is "quantum", but whenever a random number is needed, a psuedorandom and deterministic source is used instead. We'd never be able to tell from the inside, unless there was something very, very broken about the psuedorandom number generator, which given that we are talking about something with the capacity of simulating our entire universe seems unlikely (unless intentional).

Basically, our universe could be deterministic from the outside while being nondeterministic on the inside, in the sense that those of us on the inside will never be able to collect enough information to predict our own universe's evolution from the inside. (In other news, I'm in the camp that considers "randomness" and "lack of knowledge" to be two phrasings of the same thing, so it's sensible to talk about things being deterministic(/"not random") from one viewpoint but nondeterministic from another.)

This is just FYI since it's relevant, not disagreement.

> It should be pointed out that Bell's inequalities only rule out the existence of classical hidden variables being able to explain quantum effects

No, Bell's inequality rules out all local hidden variables, whether those variables are generated deterministically or not. Bell's inequality does not say anything about the existence of intrinsic randomness in the universe; it says something about locality.

(Of course, the fact that quantum mechanics is intrinsically random as far as we can tell made it conceptually very difficult to formulate the correct inequality, although the math was simple. That's why the original EPR paper couldn't quite nail down what was so weird about QM, even though they were on the right track. It was an astounding feat by John Bell.)

  > So none of the empirical facts are under dispute. The  
  > question then is a purely philosophical one: is there a
  > sensible notion of free will given determinism? Should 
  > (normative claim!) we treat people like they have free
  > will, i.e. punish/reward them for their actions?

Thanks for your reply jessriedel. In my opinion the two outcomes that make sense here are to declare that free will makes no sense to apply to any entities at all, or to apply it to literally all entities in the universe. I'm not sure which is more palatable.

Regarding the normative / moral / legal issues, I don't think rejecting non-causal free will necessarily has any immediate implications. Patricia Churchland's writings on this topic very persuasive to me.

>Patricia Churchland's writings on this topic very persuasive to me.

Any particularly accessible works of hers you would recommend?

Here is a short article of hers that was published in New Scientist: http://www.philosophicalturn.net/intro/Freedom/Churchland_Fr...

Here is a lecture of hers on the topic:

http://www.youtube.com/watch?v=opM7E6ty28A

Thanks.

I don't see many arguments in Wilkinson's article. All I see are assertions, spiced with a unsubtle comment that neuroscientists are fabricating their results.

Neuroscience can shed light on the philosophical study of free will via empirical investigations, specifically the study of human brains. After all, if a philosopher wants to describe a property of reality, it must submit to empirical results. Philosophers can abstract away from reality if they wish, but they should come clean and rename themselves logicians.

None of the empirical facts relevant to the question of free will are disputed. See my response to joeyo.

I'm just going to put this here http://wiki.lesswrong.com/wiki/Free_will

EDIT: Ok, I'll just say that it's more useful when facing a difficult question like this that isn't even guaranteed to have a correct answer to first ask what causes the question. That is asking, "Why do we think we have free will", before going back to the question of whether or not it exists.

"We must believe in free will -- we have no choice."

Issac Bashevis Singer

"[A recent article in Nature] show that, prior to the moment of conscious choice, there are correlated brain events that allow scientists to predict, with 60 to 80 percent probability, what the choice will be."

It's amazing how correct the Architect was and at the same time how limited: "But we already know what you're going to do, don't we? Already I can see the chain reaction, the chemical precursors that signal the onset of emotion, designed specifically to overwhelm logic, and reason." Evidently, at the neurological level, there is no "logic" or "reason" as he uses those terms.

Reply on news.ycombinator.com