X-Message-Number: 4979
From: 
Date: Fri, 13 Oct 1995 14:23:37 -0400
Subject: limits of inference

Although I can't do justice to it in a short piece, I do want to add a bit to
Joe Strout's reminder on homogenates (#4970). 

The point, in extreme brevity, is that in-principle reversibility is probably
MUCH closer to perfection than to zero, even in the worst cases--even if you
believe that quantum uncertainties or/and chaotic extremes represent the last
word. The reason is that quantum theory and chaos theory (or classical
sensitivity to initial and boundary conditions) are highly misleading.

They are misleading because they usually focus on over-simplified text-book
cases, these  being the easiest to understand and the only ones that are
mathematically  tractable. That is, they deal with simple, isolated systems;
or else with systems (such as weather patterns) that are somewhat tractable
with present capabilities. But in fact there is no such thing as an isolated
system, as far as we know--at a minimum, according to present understanding,
there are gravitational interactions between ALL particles, bodies, and
systems. Further, every particle, body, and system has a HISTORY of prior
interactions (and also, if you believe in a block universe, or even just in
Gell-Mann's totalitarian principle, a future history). 

This means that--for example--if we want to infer past or future positions of
an asteroid, we are not (in principle) limited to use of the laws of motion
and gravitation as applied to the accessible parts of the solar system,
tracing trajectories. Each asteroid--and each part of each
asteroid--interacted in many ways with other bodies and systems, establishing
a history which is NOT compatible with most of the solutions of the n-body
problem. 

The implications are profound. By  using ALL of the theoretically available
information, we can in principle discard all the garbage and  find a unique
solution. In practice it is a different story--especially in these primitive
times--but in a reasonable future we could surely discard MOST of the
"solutions" related to the over-simplified trajectory-tracing procedure.
Similar reasoning applied to biology probably means we will be able to
reconstruct a person to an acceptable degree of fidelity, even under
circumstances now regarded as extremely unfavorable. (See also Ralph Merkle's
pieces on cryptography to get some feeling for what can be done with
"homogenates.") 

It shouldn't need to be said, but probably does need to be said, that this is
not a "Rosy Scenario" offering. Whatever may be possible in principle, no one
knows what will be attainable in practice, and we must always do our utmost
(after setting appropriate priorities) to minimize risk and maximize our
chances. Research emphatically yes--but also realism about the probabilities
in an open-ended future. 

Robert Ettinger
Cryonics Institute
Immortalist Society


Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=4979