From: Eric Hawthorne <[EMAIL PROTECTED]>
Subject: Re: Why no white talking rabbits?
Date: Fri, 09 Jan 2004 10:36:41 -0800

Hal Finney wrote:

What about a universe whose space-time was subject to all the same
physical laws as ours in all regions - except in the vicinity of rabbits?
And in those other regions some other laws applied which allow rabbits
to behave magically?

While this may be possible, we seem to have found so far that the universe admits of many
simple regularities in its complex systems and its fundamental laws. Therefore many of the
essential properties (future-form-and-behaviour-determining properties) of these complex
systems admit of accurate description by SIMPLE, SMALL theories that describe these
simple regularities in the complex systems.

But that's an empirical observation about our universe, it doesn't tell us anything about *why* this should be true if you take seriously the "everything that can exist, does exist" theory that this list is meant to discuss. For example, if you consider the set of all possible Turing machine programs, then for any given complexity, there are an infinite number of programs that are more complex than that but only a finite number less complex. So it seems like you need to assign progressively less measure to the more complex programs in order to get a high likelihood of living in a universe defined by a simple program (assuming you believe in 'universes' at all, which advocates of TOEs that deal with first-person probabilities might not). One solution might be that more complex programs tend to run simpler ones inside them somehow, increasing their measure (like a detailed physical simulation which contains, among other things, a simulated computer running a simpler program), but then you have to address the problem in that Chalmers paper I posted about how to identify instantiations of a given program in a way that doesn't imply that every program instantiates every other possible program.

Also, the problem with taking the "white rabbit" example too literally is that programs that create orderly phenomena like talking white rabbits would almost certainly be very rare unless you had a measure that was specifically picked to make them likely--this is why I prefer examples where the laws of physics break down in a region in a more random way, like getting a completely wrong pattern of photons hitting the screen in the double-slit experiment. Among the set of all possible distributions of photons you could get in this experiment, the number of possible "wrong" ones should vastly outnumber the number of "right" ones that quantum mechanics assigns a high probability to, so why do we never see such violations? This is another form of the "white rabbit problem", but without the misleading orderliness of examples like an actual talking white rabbit, a man walking on water, etc.


Worried about inbox overload? Get MSN Extra Storage now!

Reply via email to