Newsgroups: comp.robotics
Path: brunix!cat.cis.Brown.EDU!agate!howland.reston.ans.net!pipex!sunic!trane.uninett.no!eunet.no!nuug!EU.net!uunet!hobbes!earth.armory.com!rstevew
From: rstevew@armory.com (Richard Steven Walz)
Subject: Re: 100 Billion Nuerons
Organization: The Armory
Date: Sun, 17 Jul 1994 17:22:19 GMT
Message-ID: <Ct3HLF.83M@armory.com>
References: <Csxrrs.74n@armory.com> <R1wgPc5w165w@sfrsa.com>
Sender: news@armory.com (Usenet News)
Nntp-Posting-Host: deeptht.armory.com
Lines: 85

In article <R1wgPc5w165w@sfrsa.com>, Ed Severinghaus <eds@sfrsa.com> wrote:
> 
>uh, sorry folks, it seems in my haste I hit the wrong key and copied all 
>of Steves post back onto the net <chagrin evident>.
>
>What I meant ot do was simply to ask Steve (and Jeff, please no flames 
>about philosophy) what he means by awareness.  Do you propose an 
>operational definition of it?  Sorry if I missed it, but your bost was 
>pretty long :)
> 
>I want an editor!!!!
>Ed
-------------------------------
I have read a bunch of philosophy of AI arguments for human-like awareness,
but I'm not so interested in that. However, I Do think that awareness can be
defined as any input output system which obviously includes in its implied
rationale a repeatable and diligent consideration for itself in its milieu.
Now I think that many small life-forms have that in a remarkable amount.
There are those which hunt and consume, but it is hard to say they realize
their own existence. The trick is to begin with an attempt at a cognitive
program in a limited milieu. If you can "talk" to it with simple data
entry, and if it can discuss, no matter how primitively, it's own existence
as a preferred "third person", then even if its awareness is minimal in
scope, it is still to be regarded as aware! What else could one call that
facility? Does that need to be a fully complex understanding? I don't think
so. It can be as simple as understanding and resisting it coming to an end.
If it strives to survive in its milieu against parameters which could cause
it to end, then if it also can convey output to another about that process
recognizing that it can end and wishes not to end, then we have a type of
awareness. I think that a good logician may be able to write these simpler
than I and recognize as aware, by my definition, even better than I do! 
It needn't even conceive of another, if it can be shown with ouputting
breakpoints in the program which inform as to its cognitive processes. In
othrwords, at some point we should be able to see from within the entity
that it does make mental models and conceive a reluctance to be shut off or
"defeated" in this self-same inner dialog. I feel like I could almost write
such a thing, but I may not get enough chance to get anywhere with it. I
simply wonder whether awareness as a small scale phenomenon has been
discussed in AI circles, as then all we would have to do is add ons to
increase its scope up to conversational level. We may not have human-like
awarenesses for a very long time, but how about a bird like awareness??
Many parrots have recently been shown to have self-awareness and an ability
to synthesize hypotheses and use verbal constructs. It turns out that they
do not do this simply out of imitation of sounds, as we had thought!!!!
This is really what got me thinking about it. With a brain the size of an
acorn or less, this animal had learned to talk and to mean things!!! I saw
it myself on a special on animal intelligence. The parrots actually
impressed me more than the dolphins did, although the great apes are the
most compelling surely. But still, that tiny brain understanding all sorts
of things when asked, and in English! And responding in English!!! And
being able to understand past and future and interact! This was
unbelievable! When something THAT small can do it, I begin to believe that
it is NOT the feat of some giant brained animal, but simply a structural
difference in a rather simpler brained animal, just a trick, like so many
other things our brain does, but that no other animal had needed as much as
the animals which have it! It seems then a matter of specific architecture
rather than of massive processing. The parrot may know only 200 words, or
so, but it can relate and talk to us in a way which, if we had a robot do
such a thing it would be a marvel so surpassingly great, that it might
overwhelm even the moon missions in their scope for the meaning of our
existence here! And the parrot could bring things and count, and it could
speak! It would be suddenly like having a dog you could talk to!!! Weird. 
This is what I'm getting at. The trick we are looking for in massiveness
may be nothing more than an architectural or algorithmic trick!!! Our much
vaunted "AWARENESS" no more than a simple trick, as simple as others we
already are bored with, but in this case simply have not found, partly
because of our silly pride at imagining it to be hard!! Now you and I know
that the hardest thing the parrot probably did was to convey thought I/O
in English!!! This leaves HOW tiny a remainder of a brain, which also by
the way knows how to FLY!!! How small a piece is the part in which
awareness is centered or could do that job without the others if they were
disabled??? It gets scarier the more you think about it!!! With all our
memory and sophisticated logic, we might be only 5% more aware than
anything's awareness that has had one thus far! We just have a bunch of
mass storage and visual-manual processing going on which might awa well
even be able to do it itself without being aware!!!! COULD there be a being
which was not aware, but which was a tool maker and builder of cities and
starships as a reflex, a learned trick, and not really have the same kind
of awareness we do! A being which did not fear because it did not know it
existed?? Is our awareness just a "trick" we needed for our evolutionary
niche? Do all humans even think now??? Are THEY really aware. Might we even
enhance "awareness" well beyond that of humans? Is it the "trick" we are
looking for so hard? I don't pretend to know.
-Steve Walz   rstevew@armory.com

