Consciousness of Turing machines

From: Russell Standish <>
Date: Wed, 2 Feb 2000 11:45:28 +1100 (EST)

I've been a little quiet over the last couple of weeks - blame it on
the dreaded lurgi.

I am posting a followup to the conversation I was having with Bruno
about COMP. The first thing is that I agree with his conclusions, that
the ontology between psychology and physics is reversed. (At least in
the sense that Bruno is using this phrase!) This is the conclusion of
my Occam paper. Obviously, Church's thesis is assumed implicitly, and
the strong form of arithmetical realism taken, Schmidhuber
style. However, where I differ is Bruno's assumption 1), that is of
survival under replacement by a Turing emulation. Since this
assumption is independent of the other two (I had been calling it
COMP), it should have a different name, SURV, say.

Instead, I make two additional assumptions: 1) time, and 2)
projection. I try to normalise the comparison by making a stronger
assumption, ie that I am actually a Turing emulation already - clearly
satisfying SURV. Clearly, time will be implied by this, but I'm not
sure what the status of projection is.

Now consider a slight variation of Bruno's copying machine, one which
has two buttons marked "W" and "M". Pressing the relevant button sends
me to Washington or Moscow. A Turing emulation walks into the machine,
closes the door. Before that emulation can press the button, the
outcome is decided. Knowing the precise state of the Turing emulation
prior to entering the enclosed chamber which admits no firther
stimulation from the outside world, one can predict with certainty
which button will be pressed.

Now a conscious entity equipped with a real random number generator -
eg amplifying thermal noise across a synapse by means of a chaotic
network - enters the chamber. This time, we're allowed to read all the
potentials present in the brain, but we cannot predict the result of
the random number generation. We can only assign a probability
distribution between the "W" and the "M" button being predicted.

So, from the "God's eye" view, looking at this particular quantum
history, in one case, there is only one continuation, in the other
there is two. So one is not the same as the other. However, look at it
from the first person point of view, ie the being that is entering the
teleporter. Since one can efficiently construct a pseudo random number
generator that will fool any particular observer (as has been pointed
out already in this list), the Turing emulation may well believe that
it has a free choice in deciding which button to press. It has the
"illusion" of free will only. The question is whether we actually have
real free will, and whether this really matters!

Certainly in the everyday environment, we are swamped by randomly
generated information. It is a moot point whether this dominates our
internal sources of randomness. If it does, then a Turing emulation
would have as much free choice as do we.

So clearly, if I actually made use of this randomness, I would be
losing something by being turned into a Turing emulation. On the other
hand, it may not really matter. As a consequence, I would have no
faith to enter one of Bruno's machines, knowing I could be potentially
losing a source of my creativeness.


Dr. Russell Standish Director
High Performance Computing Support Unit,
University of NSW Phone 9385 6967
Sydney 2052 Fax 9385 6965
Room 2075, Red Centre
Received on Tue Feb 01 2000 - 16:44:01 PST

This archive was generated by hypermail 2.3.0 : Fri Feb 16 2018 - 13:20:06 PST