Re: Emulation vs. Simulation

From: Robert J. Bradbury (
Date: Wed Mar 21 2001 - 02:11:09 MST

On Tue, 20 Mar 2001, Lee Corbin wrote:

> I think that none of the people I've seen who have
> written at length on this topic will disagree with:
> 1. A "zombie", defined by Dennett and others as an
> entity that can behave the same way that a
> feeling and conscious creature does, but which
> is not feeling or conscious, is impossible.

Well since I'm unsure whether I'm dealing with the "real" Lee
(in which case I would say, "no you silly dimwit, thats simply
wrong") or the sim/zombie Lee (for whom I have much more sympathy),
to whom I would respond, "No kind sir, I would suggest you
think about this a bit further")... [The lack of cycles
available to the sim'ed Lee could explain this error.]

I haven't read Dennett, so all I have to go on is my
impressions about what you are claming -- which as
far as I can assess is that "unconscious zombies are
impossible." Phoooey!

Lets assume for the moment that "conscious" humans are 50%
"genetic" and 50% "environment" (i.e. their behaviors are sourced
from those ultimate foundations). For the genetic behaviors, once
I know concretely the genetic makeup of the individual (and
potentially vis brain wiring that may put a spin on that)
it should be feasible to predict with very high accuracy
what will happen if that individual is exposed to a microbe,
a crowd of raging thugs on Mardi Gras, gets a needle stuck
in their finger or the response to any other stimuli that
generates an "automatic" response. For the "environmentally"
determined behaviors you simply build up a database of situations
and responses for individuals with similar genetic makeups
and/or life-histories.

Humans are *remarkably* consistent for the most part in how
they do things. Why? Because *not* being consistent is risky.
You learn what works and you do it over and over and over again.
When your database doesn't match the situation you pick a response
that is "reasonably" close and substitute it. An overwhelming
majority of people are going to be clueless. Why? Because
sometimes people do stupid things for which they themselves
have no explanation. In case you *HAVEN'T* noticed most
of those people out there that you assume are "conscious"
are executing "unconscious" patterns. When you do something that
an "unconscious" behavior pattern determines is the successful
strategy in a particular situation and then subsequently ask
the "conscuous" individual, "Why did you do that?", the answer
is either "I don't know" or something the made up on the fly
that happens to make sense.

You want to consider that the interesting aspects of recent
"Turing Test" contests are not that the machines are more
often being voted on as humans but that the humans are
more often being voted as machines!!!

It is pure fantasy to believe that zombies could not be
successfully emulated using a combination of the knowledge of
the underlying biology of humans and a very large statistical
database of how humans act in various circumstances.
"Consciousness" is not a primary determining factor in the behavior
of the large majority of people on the planet.

A very clever panel of philosophers & psychologists might be
able to pierce the curtain, but the average person on the street
doesn't have a chance.

> 2. the term "simulated", when applied to an entity,
> is ambiguous; better is "emulated", to refer to
> entities whose mental processing is on the order
> of a human being's or greater, and "portrayed",
> to refer to a very shallow implementation which
> is not much more than the image of a real entity.
> Thus, for example, please be kind to emulated
> beings, but it doesn't matter how you treat
> portrayals.

You don't have to be kind to "emulated" entities if you
can determine for certain that they have more cycles at
their disposal than you. If they have been programmed
as moral beings, who will never do you any harm -- off
them and take the cycles. That is what natural selection
is all about. If they too are pawns in the
game try to work out a cooperative strategy. If they
have significantly less cycles than you, they are irrelevant
so simply ignore them. [I think Game Theory dictates
the behavior patterns here, but someone more competent
than I in the complexities of that should comment.]

If its a "portrayal" (as I outline above) that has
significantly more cycles than you, then before you
disregard it entirely you need to determine whether it
views having access to any of your resources to its advantage.
If so, then you need to be very very careful around it.

> History is so complicated, I maintain, that only emulations can
> prove anything.

This is something we should be able to statistically answer
in the not too distant future. The question becomes to
what degree is the course of history determined by the
large scale statistical behavior of the masses and to
what degree is it determined by small twists of the
paths based on critical individuals. If the behavior
of the masses is consistent, as I believe it is, then its
the small spins that specific individuals put on it that
are most significant. If that is true then portrayals may
substitute for emulations for the masses.


This archive was generated by hypermail 2b30 : Mon May 28 2001 - 09:59:42 MDT