Newsgroups: comp.ai.philosophy
Path: utzoo!utgpu!watserv1!maytag!watdragon!violet!cpshelley
From: cpshelley@violet.uwaterloo.ca (cameron shelley)
Subject: Re: Testing for machine consciousness
Message-ID: <1990Oct31.023922.13795@watdragon.waterloo.edu>
Sender: daemon@watdragon.waterloo.edu (Owner of Many System Processes)
Organization: University of Waterloo
References: <3499@media-lab.MEDIA.MIT.EDU> <1990Oct4.154655.23004@canon.co.uk> <oliphant.4676@telepro.UUCP> <1990Oct30.091654.25318@canon.co.uk>
Date: Wed, 31 Oct 90 02:39:22 GMT
Lines: 69



  I'd like to inject a few comments regarding testing for machine
consciousness.

  Firstly, why do we accept the belief that other humans are conscious?
(I use the word "belief" advisedly, since I think that knowledge of
another's subjectivity is problematic.)  I would argue that we use a
genetic analogy: I am human (which is now a genetic term), and I am
conscious; therefore since this other individual is human, he or she
is also conscious.  In other words, we believe ourselves to be conscious,
and we believe that the genetic connection between ourselves and other
humans is 'close' enough to preserve that property.  (I should also
point out that "human" wasn't always this way - animistic religions
ascribed "human"-type consciousness to animals such as wolves and bears
and other "totems".  But I digress. :)

  So what stands in the way of our belief in machine consciousness?  If
the above is true, it predicts that we will have problems because our
connection with the computer/program pairs we create is not genetic (or
of the same order of strength as genetic).  Our connection with any
machine/program pair is that we have brought it into existence to fufil
a specific purpose - normally one we have formally defined, or are
capable of formally defining.  This implies that the only way we will
accept the belief that any pair is conscious, is if we  can formally
describe consciousness and verify that the program involved matches
the definition.  This also helps to explain why people are more
reluctant to accept a connectionist approach, since the connection
between us and the eventual behaviour of the pair is even more distant
than under traditional circumstances.

  If you do not believe that we can define consciousness, then according
to this line of reasoning you must either give up on ever accepting a
belief in machine consciousness, or come up with a new criterion.  The
obvious way of getting around the "genetic" analogy is to try and
generalize the notion of consciousness to avoid anthropomorphisizing
terms (which seems difficult considering the lack of other examples
of 'conscious' to go on).  Even if we manage to avoid loaded terminology,
I wonder if we can avoid loading the concepts we imagine when we use
it?  

  Anyway, many definitions of conscious are being offered in another
thread and I have nothing new to offer at the moment.  But how about
the "new criterion", a new analogy?  (Bear in mind that I think 
analogy is the right idea here, since I am addressing belief, and not
proof - the arguement for which is in the first paragraph above.)
Analogies will not give definitive statements and can be misguided,
but it ultimately seems to be what we are using anyway.  The analogy
should not rely on some biological factor unless you really wish to 
rule out machine or alien consciousness by fiat.  Since, personally,
I do not subscribe to any form of dualism, I would regard it as 
possible to require that the analogy make reference to the conscious
Whatever's structure as well as its behaviour.  Well, this is as far
as I've gotten it, so all I can do now is open the floor!  Any
suggestions out there?

  Btw, my last post "Public Apology" was an ironic attempt at showing
that a machine consciousness right now would have great trouble
percieving us as we really are given its sensory environment.   I was
being ironic about other things, but it doesn't matter.  The point
is: I *don't* have the code!  I have yet to build a sentient posting
reader-responder, so I cannot honour any requests for one!  Sorry! :>


--
      Cameron Shelley        | "Fidelity, n.  A virtue peculiar to those 
cpshelley@violet.waterloo.edu|  who are about to be betrayed."
    Davis Centre Rm 2136     |  
 Phone (519) 885-1211 x3390  |				Ambrose Bierce
