Re: on simply being an SAS (and UDA)

From: Marchal <marchal.domain.name.hidden>
Date: Tue Jan 18 02:50:07 2000

Hi Russell,

I will answer your questions when I will have more time. Also
I'm not sure I understand all of them.
Here I present just some easy comments and questions.

>See my heirarchy above. COMP \equiv NEURO. One's expectations are
>constrained by the laws of the QM multiverse. Within that, naturally
>there is 1-indeterminism. Not all brain states are reachable by
>computational continuations from a given state.

Remember that I show comp => physics is derivable from machine's
psychology (= computer science including 1-person machine's discourse).
So I cannot rely on QM. QM shoulb be derivable from comp, or
either comp or QM is false.

Note: I eliminate the NEURO hypothesis to show that my result
does not depend on the heavy argumentation (in the philosophy of mind
community) between so-called externalist (mind supervenes on brain +
environment) and internalist (mind supervenes on "biological" brain
only). I just don't care.

>I was hoping for a transparent explanation of why you *believe*
>COMP<=>Schmidhuber.

This really astonishes me. I make a big effort, though!
To be sure I don't understand the difference you see between
Schmidhuber's assumption and mine. I am not talking of the
*consequences* we derived from comp.
I'm not postulating more than Schmidhuber, the computationalist
evrything is UD*, = the great programmer's work.
With this I show that machines will be confront with continua.
Schmidhuber miss that point because he confuses 1 and 3 point
of view. That's all.
Schmidhuber'plenitude => comp, because if comp is false, then
with or without NEURO, I cannot exist in that plenitude and then
it is not a plenitude. Is that not obvious?

Also, if by Tegmark you mean only the axiomatisable formal theory
then, from the *ontological* perspective it is easy to show
that Tegmark = me = Schmidhuber.
Some of the (seldom) remark by Tegmark in the list gives me
the feeling that he is quite open to that equivalence.

About "digital device". You seem to think that there exists
digital device which are not Turing emulable, and you gives
me as an exemple some radioactive (quantum) device.

If you where true Church Thesis would be violated.
Schroedinger equation and Quantum Discrete devices are
Turing emulable. The randomnes is seen by the observer when his
coupling with the 'superposed object" is emulated by the UTM.
This is quite in the spirit of Everett and any no-collapse
interpretation of QM. If you believe that a radioactive device
is not Turing-emulable, then you believe in some form of
non-computational collapse, and you are again slipping toward
Penrosian form of non-comp.

More importantly, please tell me exactly the difference you see
between "my comp assumption" and Schmidhuber's one.
Is there is a difference I agree that different names should be
used, but introducing differences where there are none, is still
more confusing.

Bruno
Received on Tue Jan 18 2000 - 02:50:07 PST