Re: Turing vs math

From: Marchal <marchal.domain.name.hidden>
Date: Wed Oct 27 10:01:17 1999

Here is a piece of a dialog between Gilles Henri and
Juergen Schmidhuber I want to comment.

>>Gilles:
>>
>>>But the assumption
>>>that the reality itself is a computation is indeed a very strong,
>>>restrictive and unneccessary one.
>>
>>But it is compatible with the data! The restriction makes the explanation
>>of the universe less complex. Why add more than necessary?
>>
>>>Occam's razor deals with the world of
>>>approximate theories, not with the physical world itself.
>>
>>Why assume the physical world is non-computable, in absence of any evidence?
>
>sorry I don't agree. Again there is no up to now formal exact description
>of the world. The hypothesis of computability is a very strong one, since
>it strongly restricts the number of possible worlds. Even if current
>approximate theories are computable, it is not necessary (or even useful)
>to assume that the reality is at a finite level...


I agree with Gilles. Computationalism is a VERY STRONG hypothesis.
Sufficently strong to give highly non trivial constraints for the
mind/body problem. It needs something akin to an act of faith as we can
understand with the simple duplication thought 'experience'.
It is incompatible with materialism and empiricism, and most of
the so-called naturalism.
It implies a complete reversal of the current paradigm following which
matter is more primitive than "personal" histories. That is locally true
(even with comp), but false ultimately.
Is that not a simple consequence of PE-omega ?

Now, unlike Gilles, I don't postulate the existence of a "physical
universe" and I prefer to work with comp which seems to me
much less arbitrary, and much more conceptually simple. But that is
a question of taste.

Bruno
Received on Wed Oct 27 1999 - 10:01:17 PDT