Re: Information content of the brain

From: Marchal <marchal.domain.name.hidden>
Date: Wed Mar 24 09:27:28 1999

GH
>If you mean that a string has a unique "interpreter" encoded in itself, I
>like to know where it is and how to find it!


BM
Suppose the string is a recursive encoding of UD*, the complete
*splashing* of a fixed Universal Turing Machine M.
If you are patient enough you will find in UD* all possible
"interpreters", which BTW are just other universal Turing Machine.

(If you are using a mechanical device to filtrate the "interpreter" you
will miss an infinity of interpreters. This is a consequence of RICE
theorem in Th-Comp-Sc, but this doesn't change anything at this step).

All executions are encoded in UD*, in particular some interpreter will
interpret collections of SRCM (Self Referentially correct machines)
sharing some rather long and deep computationnal history.

It is not important that YOU are able to recognize these discourse. From
an ontological point of view it is enough that the relatively encoded
machines
*recognize* by themselves their computationnaly generated most probable
consistent neighborood (*worlds*). That is why such reasonning is immune
to Rice-like theorems).

(Note the similarity with Everett MWI, where the criteria of correctness
are given by the discourse of the machine-observer (the interpreter
here)).

A priori for any computationnal state there will be to much
computationnal continuations. We must explain the absence of the "White
Rabbit" ! ... or the *presence* of physical laws.

Some people in the list seems to agree that the *presence* of physical
laws in "our history" can be explained by the role of a little program.
I agree, but I argue elsewhere it is not enough.

You ask interesting questions, Gilles. But these are difficult question.
With comp we have no other choice than to confront them.

Remember also the *goal*. The goal here (it seems to me) is not to find a
"new physics" giving better prediction and unifying the older physics.

The goal of an "everything theory" is to explain why we expect physical
laws at all.

(to my knowledge Wheeler is the first and may be the only (?) physicist
to aknowledge clearly that no "new physics" can do that).

... and evidence appears that an explanation of THAT will rely on a
subtil mixture of number theory and its intensional variants: information
theories, computation theories, computability theory, proof and
provability logics, communication theories, *artificial intelligence*,
philosophy of mind, etc.

To sum up, when you say :

>To summarize I have the feeling that if you want a string to describe
>something, you have to put restrictions on the way you choose the
>representation ;

I'm afraid I do think quite the contrary. The problem is that we must
*integrate* on all possible (relatively consistent) representations in
there.

Bruno
Received on Wed Mar 24 1999 - 09:27:28 PST