X-Message-Number: 9809
From: 
Date: Fri, 29 May 1998 15:51:24 EDT
Subject: Merkle & Donaldson

Thomas Donaldson seems to have lost a bit of focus, in saying that Ralph
Merkle's post #9762 (and previous related work) is "irrelevant" to the
question of information theoretic loss of memories and other vital structures
in the brain.

Dr. Donaldson's point was that, now matter how detailed a nanotech mapping may
be--even if it can locate every atom--that will not tell us what to look for.
To know whether memories have been preserved or destroyed, we need to know the
structure of memory, including higher levels of organization.   He says, " any
argument which only talks very generally about how we can locate every atom
and molecule simply fails to tell us just what those locations may say."

But Dr. Merkle's argument did NOT consist only in showing the great diagnostic
power of future nanotech. Much of his discussion related to evidence of
limitation of damage. 

As one very simple example, he notes that fractures in frozen tissue are
usually relatively clean, so the jig-saw approach is feasible, suggesting
little loss of information. As another example, he considers turbulent flow in
partly frozen tissue, which would degrade information, and shows that very
little turbulence is likely. This is RELEVANT, although of course not
conclusive.

In my opinion, Ralph's writings have not only been highly relevant and useful,
but also conservative in the sense of understating reasons for optimism. For
example, as far as I recall, he has given no actual criteria for information
theoretic death; he has implicitly conceded the point of the pessimists that,
when disorder reaches some unspecified discouraging level, we should assume
the needed information is irretrievable.

As I have noted several times, we have other means at our disposal besides a
simple trajectory-tracing approach, or a simple jig-saw puzzle approach.
Cross-referencing internal information, and exploiting external
information--while requiring enormous data bases and computing power--would
nevertheless constitute an extremely powerful inference engine, to which
hardly anyone has paid any attention. (If Lee Corbin is looking, greetings.)

Still another, slightly different, approach I have suggested--again demanding
enormous computing power--is to work backwards in a sense: Start with a large
variety of normal structures, then simulate a huge variety of degradatory
processes, find the final configurations, and compare them with the case at
hand. This would greatly narrow the field of possibilities.

But regardless of the merit of my suggestions, Merkle's work deals explicity
(although not exclusively) with questions of likelihood of serious damage of
various kinds--kinds of damage applicable to structures in general, hence
including those of memory. Yes, Thomas, understanding memory is highly
desirable, but studying various kinds of potential damage--and the likely
limitations of damage--is relevant in any case.

P.S. Dr. Donaldson does a valuable service in his frequent reminders that we
cannot presume to place arbitrary limits on the powers of future technology.

Robert Ettinger
Cryonics Institute
Immortalist Society
http://www.cryonics.org

Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=9809