Re: Emulation vs. Simulation

From: Robert J. Bradbury (bradbury@aeiveos.com)
Date: Tue Mar 27 2001 - 16:37:00 MST


On Tue, 27 Mar 2001, Mikael Johansson wrote:

> Of course a human brain without neural activity is not concious at that
> moment -- then again, that human brain won't be using these 'lookup tables',
> since a human brain without neural activity hardly is having any activity at
> all (ex definitio).

We need a working term for a 'potential' consciousness (this is kind of
the logical complement of a zombie). I.e. an informatic state, that if
'runs' on a computer that is designed to execute that "bit collection"
that when executed results is a "consciousness". This is perhaps similar
to the 'potential' human that exists in a developing embryo or fetus.
(It is not yet self aware, but has the potential to become so given
sufficient resources that are not normally under its control.)

> But I still don't buy -- myself -- the argument that the consciousness or
> not of an entity relies entirely on their internal representation; that is,
> whether they react according to a lookup table, or according to some
> mystical other.

I think 'consciousness' is a combination of things, not the least of which
include:
  (a) The ability to "grab onto" the behavioral (director) bit-stream;
  (b) The ability to create a sim into which the bit-stream is redirected;
  (c) The ability to view yourself as the 'actor' in the sim;
  (d) The ability to run the sim forward into future time with reasonable
      accuracy;
  (e) The ability to evaluate the results of the sim;
  (f) The ability to choose to execute or not execute the sim in 'reality'.

> This mini-sim, how is it built? What does it do?

You have to read William Calvin on this. He explains it far better
than I can. It essentially involves the development of an internal
recreation of what needs to be done to manifest something in reality.
(Imagining throwing a ball, "dreaming" about what you want to do, etc.)

> What tells you that it cannot be equivalent to some sort of 'lookup table';
> to some sort of function mapping the input (that is output from the first
> function) to some sort of output (the desirability rating)?

Once you have run the mini-sim, I'm sure the 'evaluation' of the
results probably does involve some kind of lookup in your experience
table to grade the result. Imagine 'proposing' marriage to a woman
(or for the woman, imagine being 'proposed to'). How do you decide
the 'merit' of the quality of the proposal or the answer? It isn't
like most of us have dozens of experiences under these conditions.

On the other hand we do have dozens of experiences crossing the
street and evaluate our risk of being hit by a car with an estimate
of the distance, vehicle velocity, our rate of crossing the street,
etc. The lookup table approach works fine for many behaviors.

> It still seems to be two functions; that are composed in order to form
> 'consciousness' according to you -- so what says that this composite
> function isn't conscious in itself?

I would agree that 'consciousness' is a combination of the lookup-tables
and the evaluation of the results of the tables. The process of
developing 'consciousness', is perhaps the primary thing that babies
learn. They must accumulate a lot of information of what works and what
doesn't and develop the ability to use it. However the information
"itself" does not provide consciousness. Minimal consciousness probably
requires an (a), (b) and (e) above; then (f) develops; and finally
one gains sufficient experience for (d) to work reliably and if
one is moderately enlightened (c) begins to play a role.

Good points though, I wouldn't have broken it down like I
do above, if you hadn't asked the questions.

Robert



This archive was generated by hypermail 2b30 : Mon May 28 2001 - 09:59:43 MDT