I honestly do not see where Jim's comments add anything cogent to the
discussion that was not covered in the Wolfram's article that I referenced
previously. :_( But I do appreciate that you brought it to my attention.
Please forward this post to Jim.
What I am trying to figure out is if it is possible to salvage some
aspects of the computation idea, for example most of Bruno Marchal's work on
1st and 3rd person aspects, and dovetail them into a " our experiential
world is a simulation" model. I think that this is possible but it requires
that the computations that are both "ongoing" (not one that is timelessly
preexisting like a "Platonia"), "updatable" and implemented in the quantum
mechanical realm itself. Of course this requires that we grant ontological
"reality" to wavefunctions and their attendant mathematical objects, such as
Hilbert spaces. ;-)
My idea is to identify the unitary evolution of the wavefunction itself
as the computation that is generating the simulation of the world. But
instead of trying to have a single computation simulating a single classical
world we have a potential infinite number of QM systems (Hitoshi Kitada's
Local systems www.kitada.com ) each generating a repertoire of simulations
of classical systems. BTW, in Kitada's theory the observers themselves are
taken to be the QM systems and their observations are classical, an
inversion of the treatment of observers and observables by the Copenhagen
The simulated classical systems are taken to be the possible worlds of
experience, similar to Barbour's "time capsules" but without any kind of
prespecified arrangement or "best matching" in an a priori sense. Think of
these simulated classical systems as finite "patches" of space-time with
particle trajectories, fields, etc. embedded in them and have some duration
or "thickness in time" associated; Qcomps have been shown (by D. Deutsch!!)
to be able to simulate not just static "portraits" of classical systems but
can simulate the dynamics and kinematics, movies instead of snapshots if you
The "coordinating" of the simulated worlds is another issue that I am
also working on using a generalization of the notion of periodic gossiping
over graphs. ;-)
----- Original Message -----
From: "CMR" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Monday, January 26, 2004 6:05 PM
Subject: Is the universe compressable?
> > The problem is that there is a large class of physical systems that
> > are
> > not "computable" by TMs, i.e., they are "intractable". Did you read the
> > Wolfram quote that I included in one of my posts? Please read the entire
> > article found here:
> > Another way of thinking of this is to concider the Laplacean notion
> > where given the specification of the "initial conditions" and/or "final
> > conditions" of the universe that all of the kinematics and dynamics of
> > universe would be laid out. The modern incarnation of this is the
> > so-called 4D cube model of the universe.
> > Again, these ideas only work for those who
> > are willing to completely ignore the facts of computational complexity
> > the Heisenberg Uncertainty principle.
> Am I correct that you're essentially saying that our universe is
> algorithmically incompressible? If so I would agree and, interestingly, so
> does my friend Jim in a parallel thread I sparked from this very thread on
> the infophysics list a week or so back; thought I'd post it because he
> represents the "hard" info physical view on this subject much better than
> From: "Jim Whitescarver"
> Subject: Re: [InfoPhysics] Fw: Is the universe computable
> In so far as the universe is logical it can be modeled as a logical
> information system. The information nature of the quantum makes such a
> model convenient. It seems surprising how closely nature obeys logic
> granting validity to science.
> If we suppose that it is indeed logical and has no other constraints
> outside that logic, we then find it is an incompressible computation, that
> cannot be represented with fewer states. The universe is computably as it
> is a computer, but only a computer larger than the universe itself could
> model it. In this sense, the universe is not technically computable in
> practical terms.
> Intractability, however, is not exclusive of there existing good
> solutions. Unknowability is inherent in complex systems and we can
> capitalize on the the uniformity of the unknowable in the world of the
> Consider a pure entropy source, e.g. a stationary uncharged black hole.
> It effective eats all the information that falls in irretrievably
> randomizing it into the distant future. It is not that systems falling in
> stop behaving determistically, it is that we no longer care what their
> state is effectively randomized and outside our window of observation.
> Nothing in our world covaries with what happens inside the black hole but
> we know that there would be correlations due to the determinism that
> exists independently on the inside and the outside.
> I am not saying we can compute all of this. What happens at any point is
> the result of the entire universe acting at that point at this instant.
> Clearly this is not knowable. Causes are clearly not locally
> But we can represent the black hole as a single integer, its mass in Plank
> action equivalents. From this all it's relevant properties to our
> perspective are known in spite of however complex it is internally.
> All participants, modeled as information systems, are entropy sources like
> black holes, but we get samplings of their internal state suggesting a
> finite state nature and deterministic behavior. The distinction is
> whether we can determine what that deterministic systems is or not. We
> cannot without communicating with all the participants and that is not
> always possible.
> But given a set of perspectives, there is no limit to how closely we can
> model them. Where no model works randomness may be substituted and often
> we will get good, if not perfect, results.
> Even legacy quantum mechanics, misguidedly based on randomness, yields
> deterministic results for quantum interactions shown accurate to many
> dozens of decimal places. This suggests that simple deterministic models
> will most likely be found.
> <-- insert gratuitous quotation that implies my profundity here -->