Live data from Hacker News

Free will is back, and maybe we can measure it

aeon.co

61–63 of 63 posts

Re: Free will is back, and maybe we can measure it

#61

Earlier quoted context omitted.

> The fact is whether or not we have free will is largely irrelevant. I think perhaps more proper would be to say, whether there is an underlying determinism or not is irrelevant (and that's what Dennett is saying). Free will (as in freedom) is relevant. And about that question, it's quite easy to: - Build a deterministic system on top of non-deterministic one. Just take a lot of indeterminism and average it. For exa…

> For example, we do that in operating systems, where operating system itself is deterministic but each application lives in non-deterministic environment, because it doesn't know what other applications or processes are running inside the system. This may be a bit nit-picky, but if i understand correctly, the fact that other agents in the system influence the one you understand in ways you don't is not the same as n…

> Arguably, if you knew the exact system date and all other software installed, you could predict exactly how your program would behave.

First of all, there is a difference between what you can predict about the system and what you want the system to be. Since even on nondeterministic system you can build a deterministic one, it doesn't matter what you can or cannot predict about the underlying system. If you want a deterministic system, you can simply ignore the unpredictable parts by averaging over them.

Although, if you know the other software, you could predict the system. But normal application, unless it has root privileges or there are privilege escalation loopholes, cannot inspect the whole system and cannot see all the processes that go on. Therefore, deterministic modelling of the system from a perspective of one application is impossible.

But there is actually a simpler example. Just take an application using the pseudorandom number generator. You can't know if there are other processes using the generator too, so once you use pseudorandom generator, your application becomes non-deterministic. If on the other hand the application wants to be deterministic, all it has to do is not use system-wide generator of randomness, regardless of whether the generator is actually deterministic or not.

The point is, for free will to exist, it doesn't really matter if the universe is deterministic. Just as it doesn't matter in D&D game if the result is determined by dice rolls or a pseudo-random number generator with a known starting seed. People cannot calculate the latter in their heads, so the difference won't affect their reasoning (these things are conditionally dependent on the observed reality); that's all that matters.

Re: Free will is back, and maybe we can measure it

#62
post #39

Earlier quoted context omitted.

Exactly. I have never understood why people use determinism to excuse bad behaviour, but fail to see the resulting punishment is just as determined.

They don't fail to see it, they just object to calling it "fair".

If determinism is true then nothing (or everything) is “fair”.

Re: Free will is back, and maybe we can measure it

#63
post #8
post #7

Earlier quoted context omitted.

ISTM that the hundreds of years of debate comes from a unfortunate misunderstanding provoked by many centuries of religious thought: that to be free we need some kind of "soul" apart from physical reality. (The article mentions that) But I don't fully understand why so many smart people, including non-believers, can't see the obvious: that determinism has nothing to do with free will. Free will is just a mechanism of…

that determinism has nothing to do with free will The theory goes that, on an atomic level, interactions between particles is a deterministic process - it's wholly and completely predictable because the outcome is determined by a knowable set of rules. Given the fact that our brains are just[1] a huge set of particles and chemical processes it should then follow that a perfect computer model of someone's brain would…

>>that determinism has nothing to do with free will

>The theory goes that, on an atomic level, interactions between particles is a deterministic process - it's wholly and completely predictable because the outcome is determined by a knowable set of rules. Given the fact that our brains are just[1] a huge set of particles and chemical processes it should then follow that a perfect computer model of someone's brain would make the exact same decisions as that person.

yes. Absolutely true. In theory. The coin head/tail outcome can be modeled and predicted even more easily than brain. Thus the coin is an example of determinism. In theory. On practice though it is the thing that comes among firsts to mind when we talk about in-determinism.

Giving the amount of computations necessary to perfectly model even a coin, less a brain, and the minimum energy cost of a unit of computation, it may be so that such perfect computer is just not possible inside our Universe (it may happen that Universe is the smallest computer that can run such simulation and that exactly that it is doing right now :)

Post reply on HN