X-Message-Number: 15006
Date: Thu, 23 Nov 2000 20:58:18 -0700
From: Mike Perry <>
Subject: Brains, Turing Machines, Survival.

Thomas Donaldson asks:

>So come on, guys. I have asked a very simple question: has anyone got
>a proof that human brains can be imitated (even if only abstractly)
>by a Turing machine.

What, Thomas, do you mean by "even if only abstractly"? I'm not going to
claim that I can "prove" that something like a classical computer can
imitate a human brain in realtime. (That's a problem for the future!) If you
don't insist on that but allow longer times, it seems reasonable that a
"Turing machine" (do you mean the mathematical abstraction or some kind of
real device?) can do the trick, by modeling isomorphically the state changes
at the quantum level. Or to put it another way, if quantum mechanics is
correct, and again you aren't concerned with realtime performance (though as
far as I can tell you are, but then why say "even if only abstractly"?), you
in effect have your proof. 

Some clarification for my posting yesterday: I was considering a scenario
where there were N copies of me made, and I said things like "all but one
would be destroyed." What I meant was that, after the N copies were made,
there would be a total of N+1 of us, and all but one of the N+1 copies would
be destroyed, i.e. the original was at risk too, on an equal footing with
the others.

On another subject: I do differ from some others in not thinking of a copy
of me as a different person, until some actual difference emerges. So if I
am frozen, and a copy is somehow made and animated, "I," meaning the copy,
would *not* particularly want the original reanimated, any more than right
now I would particularly want a copy of me made and animated. (No, it
doesn't bother me in the slightest that I could be frozen but the original
never reanimated because of a decision made by a copy.)

Next, Pat Clancy, #14993, responding to Ettinger, says:

>Calculating successive states is not the same as implementing quantum 
>reality - so your "[calculate?]" in my statement I would not agree with.
Sure a 
>computer can calculate quantum states - that's the kind of thing computers 
>were invented to do.

But calculating the states does capture the important part of what is going
on. If you did this for a person, and supplied your emulation with
appropriate inputs from the outside, you could talk to it, etc. As for
collapse of the wave function, it is interpretation-dependent; in
many-worlds it doesn't happen. What does happen, though, would be no small
challenge for a conventional Turing machine with a 1-D tape. The number of
separate lines of computation would grow exponentially with the time step,
thus be completely impractical, though possible in principle.

On universal Turing machines and computers: the universal Turing machine is
indeed a special case since not every Tm is universal. It happens that most
real computers are computation-universal, however; it's relatively easy to
make them that way and there's no percentage in going to the trouble of
handicapping them, even when they are dedicated to specialized tasks that
would not require universality. This, I was told, is true even in such cases
as a pocket calculator.

Mike Perry

Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=15006