From: O'Regan, Emlyn (Emlyn.ORegan@actew.com.au)
Date: Wed Jun 23 1999 - 00:29:26 MDT
Harvey wrote:
> The replay is not conscious. To get the replay to work, the Turing Test
> administrator has to do the exact same test. If the question is delayed a
> second, the "brain" will answer a question that was not asked. If the
> speaker to the "brain" burned out, it will answer without hearing the
> question. If the questions are not identical, the wrong answers will be
> given. The really-conscious brain would correctly react to these
> situations. The fake-conscious brain will fail to react.
[etc]
Here I think that there was a more subtle point to the examples of patterns
as pretend causal mechanisms. I concede your points about a recording above,
of course, it's like asserting that someone's voicemail message could be
mistaken for the person themselves.
A consciousness+intelligence exists in an environment. The environment
includes the consciousness+intelligence (yes I think that you can remember
the experience of the colour red).
In the examples, the environment is as controlled as the brain. We think of
the brain as a set of initial states (axioms) plus a set of rules of
derivation - a formal system. The entire environment is part of the state of
the system (ie: the whole universe) in that everything effects everything
else (no matter how minutely), and the rules of the system are based on the
environment.
Then the theorems of the system are all reachable states of the environment
(states of the whole universe including the internal state of the brain).
The existence of causality here is that there are rules of derivation.
Rather than starting with all theorems as axioms and no rules of derivation
(pattern based), we have a small set of states and methods of navigating to
all other states.
The idea of replacing causality with pattern comes from the observation
that, even though the rules of derivation look like causality, they can be
precalculated to determine what they would produce. Then we can have an
identical system (from inside) which looks exactly as causal as the system
with rules of derivation. Except that now all the calculations made from the
rules are taken as axiomatic, and there is no derivation. A pattern based
system.
One objection raised was that the rule based system was necessary to create
the pattern based system. But you can work the other way round. Take this
new pattern based system, and you can create a new causal system. Start with
the initial patterns as axioms. Theorems have a sequence. Now we can create
rules of derivation X->Y for each adjacent (in time) theorem pair X and Y.
Then remove all the theorems. All you have left is the initial axioms and a
LOT of rules of derivation. A causal system. But it looks more like a
pattern based (lookup table) system.
There really isn't a good difference between a system of derivation rules
and a lookup table. For someone sufficiently familiar with the derivation
system, it IS a lookup table, because they can "see" (in their mind's eye)
all the theorems by contemplating the system. For example, chess looks a lot
like a formal system of few axioms and rules of derivations to a beginner,
and a lot more like a lookup table to a chess master. The ability to see
past the axioms and derivations to the theorems, with calculation, would
seem to constitute what it is to understand something.
What I am claiming is that a set of theorems derived from a formal system is
an isomorphism of a formal system. A pattern based system is an isomorphism
of a causal system.
But if someone says "Non-deterministic", I might vanish in a puff of smoke.
Emlyn
This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 15:04:16 MST