From: Darin Sunley (rsunley@escape.ca)
Date: Mon Mar 06 2000 - 22:23:39 MST
David Fendrich <f96dafe@dd.chalmers.se> wrote:
[snip]
>But what if I make every chip record its own state and after
>a while tell all my neurochips to play what they have
>recorded.
[snip]
>Since each chip has a record of its own states, we can
>actually cut the wires connecting them and still get
>the neurochips to play its recordings like nothing
>has happened, the consciousness would still be there.
[snip
>In fact we could put them all in pile
>and the two individuals would still be there, feeling.
>But every linear combination of the two sets would also
>be an individual, since it's not possible to say which
>neurochip belongs to what.
I don't think this thought experiment is very damaging to the matieralist
theories of consciousness. Here's why.
Let's apply this to a computer:
Design a computer with smart memory and smart registers. Like your
neurochips, they record their own history. After multiple runs of this
computer, we now have multiple histories.
As you said, playing back these histories results in something that looks an
awful lot like a computation.
Not only that, we can do, as you suggest, and throw the histories in a pile,
and we can't tell which particular run any particular history of an
individual memory address came from.
Here's the problem.
While it IS true that we can assemble any given linear combination of
histories into a record, the vast majority of the records will bear no
resemblence whatsoever to a computation. They will just be a picture of a
random electronic spasm. We don't even have to solve the philosophic problem
of "what is the difference between a computation and a record of a
computation", because this randomly fluctuating collection of bits is
clearly neither. 99+% of the time, it's current state will have no
relationship whatsoever to it's previous state. Anybody even vaguely
familiar with the assembler language of this computer will be able to tell
instantly that this is random noise because (for example) the program
counter is jumping around wildly without any kind of jump instruction moving
it.
Richard Dawkins said "Howerver many ways there are to be alive, there are
infinitely many more ways of being dead." He was referring to the fact that
the state space of biological organisms is infinitesimally tiny compared to
the state space of matter. There are many fewer living organisms with n
atoms then there are dead configurations of matter with n atoms.
The same is true of computation, and, assuming the matieralist viewpoint,
consciousness. The state space of brain states with coherent memories and
perceptions is WAY smaller then the state space of POSSIBLE brain states.
Your current mental state, in general, has a causal relationship to your
prior mental states (assuming of course you believe in the passage of time,
that your senses at least SOMEWHAT adaquetely reflect an objective outside
world, that the amazing internal consistency of your memories isn't a
coincidence, etc, etc).
The thought experiment does not imply that the matieralist model is flawed,
per se, but does empahsize that the matieralist model depends on time and
causality.
My current brain state seems to remember both causality and time functioning
in a reasonably traditional manner, so it holds the belief that Moravec's
uploading should still work. On the other hand, Descarte's Evil Genius could
just be messing with me :)
Darin Sunley
rsunley@escape.ca
This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 15:27:13 MST