AUH/MWI/UDA (Was: Everything is Just a Memory)

From: Alastair Malcolm <amalcolm.domain.name.hidden>
Date: Tue, 1 Feb 2000 20:28:58 -0000

----- Original Message -----
From: Marchal <marchal.domain.name.hidden>
> >I (and others) have always tried to keep AUH's, and MWI's of QM
(including
> >Everett), as two distinct concepts. For my own favourite AUH (which is
quite
> >close to Tegmark's hypothesis), there are three possibilities, where in
each
> >case one or more MWI's are encompassed within the AUH, provided they form
> >consistent theories in their own right.
> >(a) If all MWI's are inconsistent, they are not represented in the AUH -
> >this would appear unlikely.
> >(b) The simplest consistent MWI (cet. par.) does not provide the highest
> >measure for SAS's of our kind - again this would appear unlikely, because
we
> >would count ourselves unlucky if the theory that gives the highest
measure
> >for our kind (and therefore the theory almost certainly describing our
> >universe) happens to look indistinguishable from QM-based many worlds
> >theories.
> >(c) The simplest MWI *does* give the highest measure for our type of
SAS -
> >this seems the most likely scenario out of (a), (b) and (c), but is not
> >*necessarily* the case.
>
> I agree that a priori AUH is more general than MWI (Everett's one to
> "fix" the things). Do you agree that if AUH > MWI then QM should be
> empirically false,
> unless you can explain from your AUH why empirically we "see" only MWI.

I'll rephrase that question if I may (I'm sure my explanatory powers are
independent of the veracity of QM!):
Do I agree that if AUH > MWI, then either (I) QM is false, or (II) MWI is
derivable from AUH as providing the highest SAS-measure?

Let's assume for simplicity that Everett's is the only consistent MWI, and
consider only SAS-containing worlds. Then I agree to the extent that either
MWI dominates the AUH (case (II) and my (c)) (in which case MWI should be
derivable (in principle) from AUH as the most prolific SAS-creator per
information unit), or MWI does not dominate the AUH, in which case QM is
much less likely to apply to our universe (unless there is some sufficiently
simple single-universe QM theory) - close to cases (I) and (b).

> IF AUH = MWI, the white rabbit problem is "solved" by the decoherence
> phenomenon, OK?. And with pure comp, if AUH = MWI, the white rabbit
> problem is "reduced" to the
> derivation of MWI (QM) from psychology/computer-science/number-theory.

I agree that if AUH = MWI (which I don't accept), or MWI dominates AUH, then
the essential WR problem is solvable. But all the above is predicated on the
contention that simple universes are more numerous than complex ones.
Do you accept the arguments given for this contention (given in my web site,
Russell's Ockham paper, and earlier posts of ours)?

>
> Of course here we are vague. With comp we are warned: the day we give
> a name to our favorite plenitude: BAM! Wellcome the many-plenitudes
> theory.

Then we have to go back to underlying principles to find the overarching
plenitude.

> My problem with your (and others) type of everything approach is that you
> seem to take the "worlds" for an intuitively clear notion. But even with
> Everett MWI, I am not sure that this is clear enough.

I mainly find intuitive definitions sufficient for worlds, but agree that
there are occasions when one must be more careful/explicit.

> In a nutshell the UDA shows that
> how I cope with the WR problem = how is the appearance of physics
> generated.
> In my thesis (and in all my papers) there is two parts:
> 1) a proof that comp implies that physics is ontologically and
> epistemologicaly
> reducible to the psychology of machines (which is itself a branch of
> computer science). That is the Universal Dovetailer Argument (UDA).
> 2) an attempt toward an explicit derivation of the logic of the physical
> propositions based on a formalisation of the UDA in pure arithmetic.
> (This is more technical: modal, quantum logic, provability logics...)
> Strangely enough I got rather easily a sort of quantum logic.
> Now my problem is to derive a probability calculus from that logic. This
> is
> difficult: I'm still lacking a semantic for my "quantum logic".
> Conceptually the problem is to explain the quantum probability rule: to
> explain why we must add amplitudes and not probabilities in the case
> we ignore in which computational history we belong ...
>
> For the point 1), look at the UDA post I send today.

One thing I don't understand about this (its nice to have it all together,
by the way) is in point 14: how is comp refuted for a physics deviating from
traditional empirical physics (would comp be refuted if we found ourselves
in a classical-physics universe; or an ordinary (QM) universe + a WR event?)

Alastair
Received on Tue Feb 01 2000 - 12:38:55 PST