X-Message-Number: 11587
Date: Wed, 21 Apr 1999 01:55:36 -0700
From: Mike Perry <>
Subject: Parallel Worlds, etc.

Thomas Donaldson, #11577, writes:
...
>While I agree with the arguments about "computable functions" it's not
>clear at all that they work the same in the case of possible worlds.

Basically, the point of view of quantum mechanics is that all processing is
computational or its equivalent. There is, for example, the universal
quantum simulator which is a refinement of Deutsch's universal quantum
computer. It simulates (again, really emulates) any other bounded
(localized) quantum process under suitable conditions.

Bob Ettinger, #11579, writes
...
>Mike Perry says:
>
...
> 
> >Starting on p. 41 [Deutsch's book], you can read about an interesting >
>variation of the famous
> >two-slit experiment, that is, a four-slit experiment, in which half of the
> >interference bands that you get with two slits are virtually cancelled out.
> >The explanation of this is, again, that "ghost" photons from parallel
> >universes are nudging "our" photons so they don't strike where they
> >otherwise would. So how is this "no explanation for interference"? (The
> >"ghosts" by the way, are just as "real" as the "real" particles, just parts
> >of other "real" universes than our own.)
> 
>Again, "nudging"is not an explanation. One would need details of the 
>mechanism of interaction. Deutsch's book provides none, as far as I can see.
>

Certainly more details could be supplied but are omitted from this
non-technical book. The basic way photons "nudge" one another is by linear
superposition of the complex-valued probability amplitudes. It isn't easy,
beyond a certain point, to say what this means physically, but there it is
and the consequences have been verified experimentally. It also ties in well
with the many-worlds theory, i.e. the interference effects that are seen can
be viewed as a limited sort of interaction of particles across boundaries of
parallel universes.

I thank Bob for the references to quantum mechanics in #11582, none of which
I have read yet, but intend to take a look at.

Bob also says in this posting,

>Now a reminder about the Turing "universal" computer, which enters our 
>discussions from time to time in several ways. It isn't universal. On a 
>practical level, as Deutsch notes, all classical computers (all current 
>computers, as well as the abstract Turing computer) are far too slow to 
>address effectively many real problems, including those of quantum mechanics 
>beyond the very simplest systems.

The universal Turing machine or computer does appear to fail the test of
efficiency, compared to a quantum computer (which, however, also has a
"universal" version), but is otherwise as capable, which still has
philosophical significance. 

> The Turing computer also fails on a matter 
>of principle, which never seems to be acknowledged--viz., it cannot produce 
>e.g. the value of a vector, let alone a field of tensors, at a moment in 
>time. 

It can't produce "real" effects, that is, but only a process that is in some
sense isomorphic. As far as I can see, though, that would still allow it to
simulate a person, at least in principle, because, once again, such effects
as "real" parallel processes could be simulated sequentially, and the
simulated beings would have no way of knowing the difference, therefore it
effectively wouldn't exist--for them.

>Next a repeat to Mike Perry about the alleged need to invoke many-worlds to 
>explain quantum computation. I suggested that quantum computation "merely" 
>involves harnessing physical effects and then interpreting observations as 
>the results of computations--in other words, it is just another kind of 
>analog computation. Mike responded--if I understood him correctly--to the 
>effect that he knows of no other interpretation that explains in detail the 
>success of Shor's algorithm and similar tasks.

I would say that contrived explanations are possible, but I know of no other
straightforward explanation besides literally having a lot of processors
running in parallel. Otherwise, how do you get all that computation done?
How do you explore all those pathways, to find the factorization you are
looking for? If this is possible in a single-world domain, then it seems
that the factorization problem must be less difficult than it seems to
be--but so far that hasn't been demonstrated.

> Now I repeat that the 
>interpretation doesn't matter; if quantum effects exist, and can be exploited 
>or interpreted as calculations, it is totally irrelevant whether "really" 
>there are parallel processes going on in other universes, or whether "pilot 
>waves" of single particles are producing self-interference effects, or 
>whether hidden variables in the classical sense are at work, or any of the 
>other still-debated possibilities.

It's quite possible that more than one mechanism could yield the same
result, so in this sense the mechanism "doesn't matter." But for other, in
this case philosophical reasons, it does matter. I think that a successful
implementation of Shor's algorithm or some similar process surpassing the
efficiency of a classical computer would constitute a powerful argument for
many-worlds. Again, the grounds are simply that there is no reasonable,
alternate explanation for what is going on, though this does not rule out
other explanations entirely.

> If we need an example, you can do definite 
>integrals by measuring the charge on a capacitor and relating it to the 
>current and time--and this could have been done before anyone knew what an 
>electron was. The success of this analog computation proves nothing 
>whatsoever about the basic nature of electricity. 
>
Okay, in theory someone could build a device that happens to produce a
factorization of numbers without knowing anything about quantum mechanics or
even wondering how or why it works. But when explanations are sought, some
clearly are better than others.

On a final note, I've been trying to use informative headers in my CryoNet
messages, which I assume are not of interest to everybody. But last night I
goofed and the generic header referring to the previous day's CryoNet
messages is all that came up. That's because I inadvertently typed the real
header, "Broderick, String Theory, Deutsch" in the cc field rather than the
subject field--sorry about this & will try to do better in the future.

Best forever,

Mike Perry

Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=11587