Re: A nerw idea to play with

From: Gilles HENRI <Gilles.Henri.domain.name.hidden>
Date: Tue, 14 Sep 1999 16:03:17 +0200

>
>>I know that you present comp as a working hypothesis- I just insist that
>>comp is indeed the very strong hypothesis that there is a computable TOE.
>
>This is a misleading way to tell it. Remember that there are no computable
>theory of the 'computerland'. Machine's fate, viewed from the 1-person but
>also from the 3-person point of view, is 'the' non computable.

I'm not sure this is relevant here. I agree that it is impossible for any
machine to predict its own fate, and in particular it is obviously
impossible for each brain to know its own state and predict its own future.
But it does not prevent it to know about general physical laws, and about
external conditions of external systems. The problem is that, in our
present knowledge of physics, even this is impossible, even for a simple
single electron! You cannot program a Turing Machine to predict correctly
the result of experiments for which you have no right theory. Schroedinger
equation, or even Wheeler-DeWitt equation, is not enough. So it is
impossible to write a program to simulate the evolution of a brain doing
an experiment about testing quantum gravity, wheras it is possible - in
principle- to perform *really* such an experiment. As long as you do not
possess a TOE, you will always find (by definition) such an example.

>
>>Recalling that the quantum state of a single particle is not observable ( a
>>fact that seems to have been neglected by some computer scientists...) ,
>
>Remember that the 1-person point of view (of a machine) is also
>not 'observable'.

I agree, but the weirdness of QM lies in the fact that the 3-person
description of the outer world is also non directly observable, which is
not in my knowledge a prediction of computer theory?

>
>>...unlike for example the magnetic moment of a computer memory cell, I
>>maintain that there is an objective distinction between digital systems and
>>analogical ones, and that the brain is not obviously digital :
>
>Sure. That is an important point. The brain is not obviously digital.
>One of the reason I am insisting on the 'crazy' consequence of comp
>(like the disparition of substance) is to help people to realize that,
>indeed, comp is a highly non trivial hypothesis. It implies a total
>reversal of the common 'naturalistic' and materialistic view of the
>world.

ok

>
>
>>the clearest
>>difference between an analogical system and its digitized computation is
>>that the relevant information (quantum state) is fundamentally hidden in
>>the first case, but is known in the second one. This is as fundamental
>>physically as Church thesis for computability!
>
>Be careful, I suspect you suppose some interpretation of QM. Remember that
>comp entails that any machines looking at itself below its level of
>substitution
> will be force to infer the existence of unobservable
>'quantum' information. (I call it here 'the comp-quantum-principle'
>for reference below).

again the problem exist for any machine looking at anything (not only
itself) at the quantum level, whichever the interpretation of QM you adopt.

>> In fact all proposed
>>"simulated brains", Chinese room, Mauldlin experiment and so on imply the
>>use of classical systems whose state IS measurable - i.e. are digital
>>following the definition I propose.
>
>The PE-omega (UD argument) does NOT depend on it. The thought experiment
>are more easy with it, but clearly, even if my 'brain-state' is an
>unobservable and non measurable state, the UD will duplicate it for ever,
>and that is the point I need for my proof.

My point is that if any physical Turing machine is unable to replicate
"really" a brain (in the sense that it will never "think" like it), the
PE-omega argument is itself doubtful!

>In fact QM is 'very' computable, even linear ...

I agree, but it is must be still considered as an incomplete theory, such
as Newton's gravitation. So the point is that computing Schroedinger
equation for a brain does not obviously insure that the computation is
actually thinking. Only the use of a computable TOE would escape this issue.


>To use the phenomenological quantum weirdness as a argument against
>comp will not work with me because I take the QUANTUM as the most
>convincing (a posteriori, I confess) confirmation of the DIGITAL ultimate
>nature of reality.
>Of course a confirmation is not a proof, and computerland would
>be a sad country if all machines are betting comp :-)
>
>Bruno

Sure! :-)
I still have the impression that in the absence of a TOE, you are not
allowed to identify any (practical or "theoretical") computation with any
physical phenomenon, including the thought...and again even if it were
true, the problem of the emergence of consciousness would not be easier.

Cheers

Gilles
Received on Tue Sep 14 1999 - 07:05:57 PDT