Re: Some comments

From: Gilles HENRI <Gilles.Henri.domain.name.hidden>
Date: Tue, 11 May 1999 10:58:46 +0200

>Even if we stipulate that a computer cannot simulate itself, and cannot
>simulate the whole universe which contains it, it would not follow that
>it could not simulate me.
>
>I am not in interaction with the entire universe. For one thing, due to
>speed of light limitations, nearby stars may have exploded, and I am not
>aware of them.
>
>More prosaically, I don't know what is going on on the other side of the
>Earth, or even outside the walls of this room.
>
>If I spent some time reading a book in a small, enclosed room, I would
>still be conscious. Wouldn't it be possible to design a computer which
>could perfectly simulate me during that time interval?
>
>Hal

OK, I agree that you can imagine well defined conditions with very few
interactions where it should be possible to emulate your brain with a
machine. The dreaming state belongs also to this category. It remains that
this computer could not replace you in ANY circumstance (which for me is
mandatory to think of a real "duplication") The fact that you must appeal
to very specific circumstances where the influence of environment is
minimized (dream, closed room) is for me an evidence that the duplication
is NOT perfect! If you think just a little of what made you what you are
now (including the possibility of reading books and give a meaning to the
words you are reading), you will admit that the interaction with an open,
not predictible environment (including other people) has been absolutely
crucial. You may think possible to implement this memory in a machine at
some time t=0, but the machine will not be able to continue its evolution
like you in a proper way after a "reasonable" delay.

Usually when people (and SF authors) speak of thinking machines, they mean
machines capable of interacting with such an open environment "like" a
human being ("like" is not very precise of course, but like Turing we can
admit we are able to agree on what is a "human-like" behaviour). What I
would stress is that this requirement is INCOMPATIBLE with the possibility
of duplication, because this capacity includes a proper self-representation
in space-time which is by definition unique.

Gilles
Received on Tue May 11 1999 - 02:21:42 PDT