On 07 Oct 2012, at 16:45, Stathis Papaioannou wrote:

On Mon, Oct 8, 2012 at 1:23 AM, Bruno Marchal <marc...@ulb.ac.be> wrote:

One theory is that existence of platonic entities such as numbers is
not ontologically distinct from actual existence. In that case, all
possible universes necessarily exist, and the one that has the laws of
physics allowing observers is the one the observers observe.

That is "Tegmark error". It cannot work. First it is obvious that numbers have a distinct existence than, say, this table or that chair, and secondly, once you accept comp, whatever meaning you give to the existence of numbers
as long as you agree that 2+2=4 is independent of you, the global
indeterminacy on arithmetic, or on the UD, has to be taken into account, and physics has to be explained in term of *all* computation. That is what
Tegmark and Schmidhuber have missed, and which I have explained when
entering on this mailing list.

Even in the case one (little program), like DeWitt-Wheeler equation for
example,  would be correct, so that indeed there would be only one
computation allowing consciousness, such a fact has to be justified in term of the measure taken on *all* computation. I thought you did grasp this
sometime ago. Step 8 is not really needed here.

Computation necessarily exists, computation is enough to generate

OK. At least comp can explain 99,9 % of consciousness, and can explain entirely why a tiny part of consciousness cannot be explained, but that is a detail here.

and physics,

But how? The whole problem is there.

Tegmark would answer here:

<<by assuming that the physical reality is a computation or a mathematical structure among others, and if it sustain consciousness then it will work.>>

 But this does NOT work. (besides making physics trivial).

With comp, the physical reality is not a computation, nor a priori a purely mathematical structure a priori, physics has to be the invariant pattern in the first person indeterminacy on the computational continuations, and in their resulting statistical interferences.

therefore no need for a separate physical

It is not so much that we don't need it, it is that we cannot use it, even if true. And that we have to explain what it looks true, from a sum on all computation.

Can you explain the subtlety I've missed?

I think you are currently forgetting the first person indeterminacy. You seem to forget the fact that if you drop a pen on the ground, it will not fall on the ground because you are in a computation which support you and a falling pen, but because the pen fall in the majority of the computations going through you actual state.

This is what makes physics reducible to machine psychology, and physics is not a structure among others: it is a sum on all (comp) structures. That is what AUDA translate in math. Now, as it might be FALSE that the pen drop in the majority of continuation, or even that such a majority does not make sense, the whole thing is testable. Comp might really be refuted here (well comp + definition of knowledge based on Theaetetus + Gödel-Löb self- reference).

Deutsch criticized Schmidhuber as being trivial: explaining too much, and I gree with Deustch. But comp, when we take the first person indeterminacy into account makes this explanation not trivial at all, as we have to derive physics by extracting a measure on all computation. This is far from done. AUDA does show, though, that the measure one exists, and already obeys to quantum logic.

Comp, in that way, makes the "everything" idea into an operative theory capable of precise predictions. The physical reality is not a mathematical structure among possible other: it is the unique (quasi) mathematical structure which makes our subjective experience persistent.

Normally this should be clear at the step seven (as some Occam razor, stronger than usual but not that much, can be used to avoid step 8).




You received this message because you are subscribed to the Google Groups 
"Everything List" group.
To post to this group, send email to everything-list@googlegroups.com.
To unsubscribe from this group, send email to 
For more options, visit this group at 

Reply via email to