X-Message-Number: 8061
From: 
Date: Sun, 13 Apr 1997 14:44:55 -0400 (EDT)
Subject: simulation/emulation

As I understand it, the Information Paradigm people say something like this:

"We agree that a simulation of water is not wet, a simulation of a flame is
not hot, and so on. But if a person is simulated in a corresponding manner,
then a simulated person will perceive the simulated water as wet and the
simulated flame as hot; and if simulated water is poured (by simulation) on a
simulated flame, the simulated flame will go out (by simulation); and so on.
Simulated life can be as full and real as our own." 

I'll indicate why I think this is probably wrong, after some preliminaries
concerning "simulation" vs. "emulation."

Not all info people use these words the same way, which impedes discussion;
but in general, as I understand it, an "emulation" is a physical system unto
itself with an appropriate isomorphism; e.g. an electric circuit with
resistance, inductance, and capacitance can be used as an analog of other
systems or of mathematical transformations or operations. (Accumulation of
charge can represent a definite integral; a current, dq/dt, can represent a
derivative; etc.) 

A "simulation," on the other hand, can be a mere description or predictive
train of statements, essentially a data store and program, as in a Turing
Tape. Or--if I had the time and resources--I could simply WRITE DOWN in
longhand, on a zillion reams of paper, a complete description/prediction of
you and your actions over a certain period of time in a certain environment,
all of your quantum states and transitions.

Anyone who maintains that the last mentioned type of simulation would BE
you--well, he has more courage than sense. After all, if mere information is
all you deem important, why bother even to write it down? The requisite
information ALREADY EXISTS in the universe; you and all your parts have
physical connections to the outside and to the past; you have left indelible
foot prints on the sands of  time; therefore in principle someone with enough
technology and resources COULD describe/predict you, at any time in the past,
present, or future. (Cf. Tipler, Perry, etc.) Not everyone agrees that
information is conserved as a strict physical law; but if it is (and I think
it is), then no infofreak need worry, since he is immortal already. John
Brown's body may be moldering in the grave, but his info goes marching
on--and that's all that matters to the info people, right?   

Now let's talk about emulation:

1. You cannot FULLY emulate anything unless/until you know EVERYTHING there
is to know about it (or at least everything you need to know). For example,
we cannot fully emulate quantum phenomena because we do not fully understand
them; likewise gravitation; likewise the "passage" of time. Since there are
wide gaps in our knowledge of ourselves and this universe, how could we
possibly do any broad emulations in a virtual universe? The virtual universe
would necessarily reflect (at least) all of the deficiencies in our knowledge
of the original.

2. Even if you knew everything, it seems impossible that you could get a
genuine ONE-TO-ONE isomorphism between the original and the emulation. By
one-to-one I mean that for EVERY feature in the original there is exactly one
feature in the emulation, and nothing extra; and vice-versa. 

For example, a Turing tape cannot be an emulation for many reasons, including
the fact that on the tape setup only one thing can happen at a time, while in
the world many things happen simultaneously. 

Now ask yourself, what kind of physical system could emulate a carbon atom?
Remember, we want ALL the features of a carbon atom, and nothing extra; and
we want the same economy of space and time as in the original. (That last
requirement might, at first thought, be modified by allowing scale-up; but
you can't scale up quantum effects, so forget it.) It seems to me "virtually"
certain it can't be done. There is no substitute for carbon...and very
possibly, if you want a self circuit,  there is no substitute for meat.

Robert Ettinger

Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=8061