Wednesday, July 18, 2012

Free Will vs. God of the Gaps

My initial experiment was watching defences of determinism submitted to New Scientist's letters section, which I noticed had to keep retreating.

My ultimate conclusion is that in practical terms, these supposed opposites are indistinguishable. The question has been pushed out of physics entirely. Things like the Newcomb's Box situation, which depend on one or the other, are not only prohibitively expensive, but actually forbidden. In turn it means that things like responsibility and punishment should function indistinguishably under the two possibilities; which I have previously worked out to be true. More generally, you can believe you're deterministic or not at your pleasure.

However, it is the determinist who must alter their beliefs about the world, because the libertarian was correct about the universe being unpredictable. Even though the libertarian case is pending a good definition, the determinist-physicalist position has been pushed off the table entirely.
"Although compatibilism, the view that determinism and free will are not logically incompatible, is the most popular position on free will amongst professional philosophers"
Epistemology is not democracy. There's no reason for libertarians to look for a way to make their views compatible with something that has been proven wrong, especially when the views, properly understood, have no known actionable conclusions that their beliefs could misguide them into.

At the bottom, I investigate the epistemic implications.




At first, supposedly if we knew everything about neurons, human behaviour was supposed to be precisely predictable. This turned out not to be sufficient, so it became cells, which became particles. Particles has since fallen, (via) which incidentally has implications for many-worlds interpretations.

Indeed there appear to be multiple ways to reach this conclusion.
"The point is: predicting the system's behavior requires us to compute the future state more quickly than nature does. For that, we need to build a computer that outpaces nature. If we would attempt building a computer capable of doing so, we would soon discover that this computer collapses and forms a black hole long before it reaches the required size."
Let me put it a third way. Let's say to predict the outcome of your brain requires ten information. However, crossing the boundary of your skull is only six information. To find out the other four would mean investigating the circumstances of your brain's creation, which means gathering, what, like 100? 1000? (No, way more than that.) By which time I need an entirely new ten information, because your brain has moved on. But it is even worse than this, because;
"Key factor in the above argument is that the system needs to be 'sufficiently complex'. In practice this means that he system needs to have a strong tendency of amplifying tiny causes into large effects. A human brain - with eyes and ears connected to it - certainly has that tendency."
I need not just the ten information about your neurons, but all the information in your environment that those neurons may be fed with. To do that, I need the information about the environment's environment, so I can predict the environment, and to do that...well, you see where the black hole comes from.

That said, there must be some shortcuts, because some predictions can be made. The fact that you aren't daily surprised by your confederates disproves this;
"The thing is: given the fundamental laws of physics, for a system of sufficient complexity you can not predict its future. For that to happen you need a shortcut to describe the deterministic evolution of our universe. However, such a shortcut does not exist."
You don't need the universe as a whole because effects diminish with distance. Just as 99% of an electron's probability can be found within a few angstroms, 99% of the contributions to human action can be found on Earth or its immediate surroundings.




There's no physical difference between an unpredictable system and an indeterminate system. Simply put, the prediction computers are part of physics, made of physics; if physics wants to know what it'll do, it's out of luck too. For me to be sure of what you'll do I have to wait for you to do it, and for physics to know what physics will do, it has to go and do it. As an example, imagine two electrons, one of which is unpredictable, the other, random. From any point except the electron in question, can you think of an experiment to tell the difference?

(Hmm, I think I just figured out how to precisely define free will.)

Even if physics is ultimately deterministic in some sense, it is not possible to notice. That said, there may still be a philosophical line of inquiry.

Though these lines of evidence seem independent, since they all point to the exact same conclusion, they must be related. They are likely different facets of some underlying truth.




Epistemics

Given that we now know the two are indistinguishable, was there an easier way we could have found that out?

Causality.
For free will to exist, it must be causally linked to physics. Let's use the example of going for a walk to buy milk, or not, as the case may be. From the perspective of physics, which is implementing the transaction, is there any way to tell the difference between a chosen cause of a milk run and an unchosen cause? There isn't; there can't be. The thing, whatever it is, must plug the same way into the same socket.

(Just a tad shorter than doing it the hard way.)

That the two were indistinguishable was a conclusion available to Socrates and Aristotle. Charitably, I'm going to say they were busy and that's why they didn't get around to figuring it out. Though modern understandings of physics are certainly a big help; that just means Hume could have figured it out and didn't, and he wasn't as busy as Aristotle. Spent too much time arguing with people and not with reality.

This is also corroboration. The upward and downward conclusions are matching.

4 comments:

Anonymous said...

Between relativity (bounded light cones) and QT (stochastic particles), the case for indeterminism seems rock-solid for any other observer in our universe.

If, on the other hand, we imagine that our universe is a very robust simulation, does it not stand that our actions could be entirely predictable – if not determined outright – without us knowing? I think the problem then becomes purely epistemological: how would we ever know?

Alrenous said...

The epistemic impinges on the ontologic.

If we can't ever know, that means there's no experiment to tell the difference. Which is equivalent to saying the possibilities are indistinguishable, which means identical. Again, we take supposed opposites, and find they work out to mean the same thing.

Anonymous said...

Indeed. "Free Will" now simply means that *you* don't know what you'll be doing moment-to-moment, regardless of functionalism.

Anonymous said...

Related:

http://www.smbc-comics.com/index.php?db=comics&id=2672#comic