Wednesday, January 21, 2015

More on Smolin's Free Will

I've been musing more on the interface of choice and predictability. Experiment has shown that our choices are much more determined and limited than we tend to feel, and that many are made unconsciously even when we think they are conscious. Experiment has also shown that nearly all (perhaps all) motivation is emotional, and that reason follows after. Experiment has also shown the physical universe to be very nearly deterministic. Even seemingly random quantum events (which the Free Will Theorem suggests may be arbitrarily chosen rather than random) don't change deterministic outcomes on almost anything (and maybe nothing) by the time you get to the level of just a few atoms interacting. The realms in which quantum events play a significant, unpredictable role are tiny even compared to our brain cells. So it seems within grasp to understand living things as deterministic systems. Very complex, yes, but still predictable (at least in aggregate, if not individually). Lee Smolin has suggested that the choice being made by matter is to copy other, similar matter, thus resulting in very nearly deterministic laws. The only time truly new and unique choices will be made is when a system doesn't have a sufficiently similar system to copy, and in our vast universe such systems are vanishingly rare, since so many things can be explained through reintegrated, reductionist models. The only ones that may be amenable to experiment are new, quantum entangled systems that we may be able to artificially create with big physics in the future (I don't really understand what this means, either, but it does suggest that nothing we experience day to day falls in this category of the untested).

I can't tell what all this means. Can a system become un-predetermined simply by being complex enough? Hydrogen atoms copy hydrogen atoms, water molecules other water molecules, proteins other proteins because the atoms making the proteins are copying similar atoms, and that continues to larger structures giving us the laws of nature that we observe. This predictability isn't because Free Will is absent, but because everything is exercising its Free Will to be a copy cat. If matter hadn't long ago decided to follow this course and stick with it there would be no coherence to any structures or laws. So in some sense matter doesn't have another viable choice, but we know of people who choose dissolution and destruction all the time, so other stuff existing that never chose to be a copycat is plausible to me. But does the emergence of new interactions and relationships among the parts result in limited amounts of unpredictability? Even though the only truly unique quantum events may be in rare, not yet attempted quantum entangled states, and seeming randomness has no effect even at the size of mechanisms sending messages in our brains, is there a level of complexity at which we begin to face real, unpredictable choices?

Remember, I'm not arguing that choices only happen at certain unique points in time and space, the way we are typically inclined to think of and talk about Free Will. Choices are going on constantly and everywhere, but most of them are simply the choice to relate and continue to exist as coherent structures. Is there a complexity point at which choices qualitatively shift and turn into what we more typically think of as Free Will, conscious choice, or agency?

I don't think this is a stupid question. Even as I attribute choice to everything that exists, I don't imagine that Pluto has much agency in the typical sense. Biology and computer science are continually probing the boundaries between consciousness, conscience, choice, instinct, programming, and other traits we readily use to differentiate between human, animal, and robot, and trends seem to point toward "human" characteristics as emerging from simpler characteristics as complexity increases. Intelligence and self-awareness are neither a perfect continuum from "pure instinct" to rationality, nor are they an inexplicable step. Qualities may be absent until several simpler qualities come together in a relationship, so there is a true condition of presence versus absence, but we can understand many of the relationships and use them to explain the origins of the new qualities.

If there is such a level of complexity, is it meaningful in our daily lives, or is it a level so much greater than human complexity that we are, for all intents and purposes, living determined lives? I reject a purely deterministic universe quite confidently, but still wonder if we manage to escape determinism. Is the community of Gods sufficiently complex that they escape total determinism, but we are not? Are the choices of Gods only not predetermined when they make choices that involve new and unique quantum states? Is our future only not predetermined when we interact with Gods based on rare non-predetermined quantum events? I don't think so. I suspect if we could quantify it we would find that the "choice threshold"--the level at which new, unique, only partially predetermined events happen at semi-frequent intervals--happens at approximately human complexity. We should be right at the low end of how frequently we have to deal with such choices. That's my guess. Wish I knew how to model it.

Tangential Thoughts

If these assumptions are correct, we should expect artificial intelligence to become "human" for all intents and purposes when it attains the right kind of complexity.

Smolin's model seems to require something like:
1. Matter searches the entirety of existence space and behaves as other sufficiently similar matter has in sufficiently similar situations.
2. Matter interacts with nearest neighbor matter by a different set of rules.
3. This different set of rules should arise because of shutting down connections between certain bits of matter until connections are limited to three dimensions.
4. Time is real (although it appears to be malleable or warped through interactions in space dictated by nearest neighbor relationships).
Does someone want to try to build a new "Game of Life" and try to make "laws" emerge out of some form of these "meta-laws"?

No comments:

Post a Comment