X-Message-Number: 11797
From: Thomas Donaldson <>
Subject: To Mike Perry, with a word for Kennita Watson at the end
Date: Sat, 22 May 1999 23:05:25 +1000 (EST)

Hi Mike!

Yes, I'd be interested in reading your book. I'll even reimburse you for
the cost of sending it to me, and if you wish, send it back.

I will also note one major change in what you seem to be saying. A robot
living in the world is not the same as a subroutine or other program
structure in a computer --- even if that program structure in the computer
seems to act as if it is a human being in a virtual world. The first does
not perform any purely symbolic actions: it acts in the world, and thus
is not just a symbolic structure (the same as a book). The second is a 
symbolic structure alone, and acts like a human being only in the sense
that a REAL human being, reading or seeing its activities in its virtual
world, would agree that they are like those of a human being. But it has
no more real existence than any other symbolic entity --- like Donald
Duck. (A computer virus is a REAL life form because it acts on the
computers in reality. If it only made virtual changes, it would only be
a virtual computer virus, and again no more real than Donald Duck).

There is a subsidiary question which Bob Ettinger has brought up several
times. It deserves attention, but at present I would not claim it is
decisive in any way. The neural nets of which our brain is composed do
not work like those in current computing: they grow new connections.

Whether this becomes a critical issue would need a good deal of thought;
I will say, though, that it just might mean that such neural nets, and
the brains made up of them, are not finite state machines in the normal
sense of a finite state machine ie. a machine with a FIXED and FINITE
set of possible states. Yes, at any given time, they would have a finite
set of possible states; but since they can grow new connections, the
finite sets will change over time in response to new learning. That is,
their set of states is FINITE, but it can change considerably over time.
And the source of those changes does not come from within the neural 
net but from extra influences in the world. (This means that those sets
of states cannot be predicted merely by examining the existing set of
states).

Naturally this means that they cannot be imitated by any finite state
machine in the normal sense, and so may not be Turing machines. But then
I would hope that our notion of computing in broad enough that it doesn't
simply evaporate if we discover non-Turing machines.

Finally, a word to Kennita Watson:
If you think about it a bit, when we discuss issues such as whether or not
we could be read off in a computer, we are discussing not just our
(possible) long term futures, but even more than that just what cryonic
suspension must preserve and what it need not preserve (ie. our
consciousness and how it may work). Some people actually like discussing
such issues, others do not; but they DO relate to cryonics. Speaking for
myself alone, I don't feel like I am beating my head against anything.
Perhaps I need not say this, but there is no reason that you or anyone
else who wants to discuss more concrete issues need bother to read our
discussions. And because I have met you, I'll add that such an attitude
in no way suggests that you lack intelligence or other such features.
It's just a personal choice.

			Best and long long life to all,

				Thomas Donaldson

Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=11797