The Strong Free Will Theorem [pdf] | Hacker News Reader

The Strong Free Will Theorem [pdf]

ams.org

23 pointsTriinT17 comments

For those encountering this paper for the first time I thought it would could helpful to point out some stuff:

"The Strong Free Will Theorem" is a successor paper to an earlier paper called "The Free Will Theorem" (easily found via a search engine or Wikipedia). Both papers are fairly accessible to the typical hacker. Both are somewhat profound. It might be easier to start with the earlier paper.

Traditionally, people debate "interpretations" of QM especially and relativity to an extent. E.g., the "many world" interpretation of QM. These papers take a novel and interesting approach to analyzing such questions. They are quite refreshing, in terms of avoiding needless metaphysical claptrap.

Finally, I don't recall if the papers themselves mention this but at least in speaking appearances Conway has said that he doesn't mean for people to read too much (or too little) into the word "Will", here. Fundamentally, the theorems are about how the history of the universe up to some point (taking into account a relativistic view of history) does or does not relate to the future action of the universe. They formulate things in particularly poignant anthropomorphic terms but the result is more general. If you would rather not locate "will" in these results, Conway is happy to concede that the theorems could as well be dubbed the "free whimsy" theorems... implying meaningless rather than meaningful randomness in certain human choices.

What is the difference between meaningful and meaningless randomness?

One reflects our ignorance of the underlying mechanisms, and could be predicted with more information and/or a better model, one is intrinsically stochastic and theoretically impossible to predict.

You mean it's "meaningful" if it wasn't actually random in the first place.

ssl/ssh vs. /dev/random :)

Or your favourite application for the job.

What if there's nothing deterministic in this universe? The notion of determinism deduced from God, Nature, or whatever is the key behind the concept of free will. Otherwise, how can we identify free will, which is essentially a fancy term for "non-determinism"? The concept of free will cannot stand by itself without the notion of determinism. In other words, if everything is truly random, then there's no room for both determinism and free will. However, I don't think the structured randomness described by quantum mechanics counts for the randomness that truly goes beyond both determinism and free will. It may be incomprehensible for any of internal observers of this universe.

Lack of determinism is not the same as free will. Being constrained to the past to include the future (i.e. some probability distribution over possible choice states) is still a mechanism embedding stochastic processes, and can be modelled algorithmically.

To really get to free will we also need teleological causation:

danielnolanphil.googlepages.com/TeleologyHawthorneNolan.pdf

Also, here are empirical results from my research showing we are capable of teleological causation and thus have free will:

http://www.box.net/shared/u13u3agxqg

To quote Douglas Hofstadter: There is no free will.

Since he didn't write it by free will, I wonder what it's worth.

"When a male dog gets a whiff of a female dog in heat, it has certain extremely intense desires, which it will try to extremely hard to satisfy. We see the intensity only too clearly, and when the desire is thwarted (for instance, by fence or a leash), it pains us to to identify with that poor animal, trapped by its innate drivers, pushed by an abstract force that it doesn't int the least understand. This poignant sight clearly exemplifies will, but is it free will?" - Douglas Hofstadter (I Am A Strange Loop)

Very interesting.

If I'm reading this correctly, he says that the idea that human will is deterministic is obsolete. The idea came because it was thought the universe is deterministic, but modern quantum mechanics shows the universe is not deterministic, and neither is human will.

I'm not sure how you got that. In both the original theorem and the new "strong" theorem, the theory is a conditional statement (if-then), and that humans have free will is part of the condition, not part of the conclusion.

In describing the original theorem they write: "It asserts, roughly, that if indeed we humans have free will, then elementary particles already have their own small share of this valuable commodity."

And in describing the new theorem, they write: "The axioms SPIN, TWIN and MIN imply that..." Where the axiom MIN is: "Assume that the experiments performed by A and B are space-like separated. Then experimenter B can freely choose..." That "then" clause occurs inside the axiom, so is part of what is assumed. The axioms are assumed - in particular, the free will of experimenter B is assumed - and a conclusion is drawn.

"It asserts, roughly, that if indeed we humans have free will, then elementary particles already have their own small share of this valuable commodity."

He goes on to show the reverse of this. This line is a teaser, not a conclusion.

It doesn't really make sense than human free will should affect elementary particles, but the reverse does make (some) sense.

The experimental setup of having a human choose something is to correlate the two, not to mean that a human choosing the setup give the particles free will. (Unless I totally misunderstood this article, which is certainly possible.)

There's a difference between humans not being able to guess the future beyond a probability measure, and the future not being determined from the "beginning" (or some time far enough in the past that humans don't really have a say in what will happen).

In other words, it's quite possible that humans can't guess the future beyond a probability measure and at the same time have the universe (including their own actions, of course) being completely determined.

I think you've essentially reduced free will to the Halting Problem.

Hume pretty much refuted this point of view 300 years ago. Lack of determinism does not create free will; if your actions were totally random and uncaused, then they could not be said to be free. They'd just be the equivalent of unconscious muscle spasms.

His belief (Compatibilism) was that free will was not only compatible with determinism, but required it; your actions can be said to be free if they proceed from your personal thoughts and characteristics. Example: if pg had a choice between using Java and Lisp, we could quite confidently predict that he would use Lisp, even if we ran the experiment infinite times. It would still be his free choice, however.

It's a quibble over the definition of 'free', but then most philosophical arguments are over definitions of words.

Reply on news.ycombinator.com