X-Message-Number: 9571
Date: Thu, 30 Apr 1998 14:09:40 -0400 (EDT)
From: "Kevin Q. Brown" <>
Subject: Threshold of Death

In msg #9566 Paul Wakfer replies to Bob Ettinger's msg #9558:

>>3. Almost any kind of biostasis--cryogenic freezing, non-cryogenic freezing,
>>vitrification, freeze-drying, chemical fixation, whatever--will preserve more
>>structure and information than would be preserved in the grave, hence offer a
>>better chance of rescue.
>
> This is where we start to differ. IMO, this is not true as stated.
> There is a "threshold" of damage obove which recovery of sufficient
> mental attributes to constitute "identity" is impossible by *any*
> means. At present we have no good idea where that threshold lies.
>   . . .

Paul,
Information theory gives a way to view how such a threshold can exist.
A simple model is a Binary Symmetric Channel (BSC), which correctly
passes a bit with probability "p".  The quantity representing how
well bits "survive" the process of passing through this channel is
known as the channel capacity.  The channel capacity of a BSC is:

    C(p) = 1 - H(p) = 1 + p log p + (1-p) log (1-p)

where "log" denotes the base 2 logarithm.  Note that a channel
that makes a mistake every time is just as good as one that is
correct every time because in both cases the information is preserved.
What is striking about this formula, though, is how rapidly the
capacity deteriorates with the probability "p":

    C(1.0) = C(0.0) = 1.0
    C(0.9) = 0.531
    C(0.8) = 0.278
    C(0.7) = 0.1187
    C(0.6) = 0.029
    C(0.5) = 0.0

For example, if one considers "death" to occur between a capacity
of 0.25 and 0.5, then the "threshold of death" is a narrow
probability range of about 0.8 to 0.9.

How does this relate to humans?  The "signal" for humans is the
information/personality/etc. in our brains.  The "channel" is a
process that inflicts hours of ischemic injury during the process
of dying, followed by perfusion with cryoprotectants and slow
freezing that causes enormous mechanical and chemical damage,
maintenance in that state for many years, and eventual "best-guess"
reconstruction by nanobots (or whatever technology has to offer).
How good is the "signal" coming out of this process and how could
the formula above apply to us?

Here is one approach.  Suppose that your nanobot-reconstructed
memories are correct with probability "p" and are lost, garbled,
or totally fictional with probability "1-p".  Did "you" survive?
Did that roll in the hay with Mary Lou really happen or was it
just a wish or dream?  Here is what a future revivee might say
about his situation:

  "At the best of times, I'm amused by the funny memories of
  another age, but often I wonder why the old fart inflicted them
  on me.  Most of them are BS invented by nanobots trying to
  reconstruct his personality; I never can trust what I remember
  from back then.  Worse, the baggage, the preverbal knee-jerk
  reactions and unresolved neuroses that I inherited, and which
  seem to have been preserved better than anything else, usually
  just get in my way and must be laboriously UNlearned to get
  ahead in today's world.  Sometimes I'm tempted to just flush
  all that crap, but I still get a kick out of telling people
  I'm a reincarnated programmer from PQC (Pre Quantum Computers)
  and amazing them with my vivid, although largely fictional,
  recollections of how crude life was then."

This is _not_ a successful scenario.  Let's make sure that we do better.

    Kevin Q. Brown
    

Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=9571