Newsgroups: comp.robotics
Path: brunix!sgiblab!sgigate.sgi.com!olivea!charnel!yeshua.marcam.com!news.kei.com!uhog.mit.edu!nntp.club.cc.cmu.edu!honeydew.srv.cs.cmu.edu!IUS5.IUS.CS.CMU.EDU!jkr
From: jkr+@CS.CMU.EDU (Julio Ken Rosenblatt)
Subject: Re: Self Awareness Experiment
Message-ID: <Cn8svB.Bq8.3@cs.cmu.edu>
Sender: news@cs.cmu.edu (Usenet News System)
Nntp-Posting-Host: ius5.ius.cs.cmu.edu
Organization: Robotics Institute, Carnegie Mellon University
References:  <25MAR94.03102200@nauvax.ucc.nau.edu>
Date: Fri, 25 Mar 1994 22:52:22 GMT
Lines: 71

In article <25MAR94.03102200@nauvax.ucc.nau.edu>, boone@nauvax.ucc.nau.edu writes:
 > 
 > A biologist friend of mine believes chimpanzee self awareness can
 > be proven by the red ribbon experiment. Take an assortment of
 > primates and on by one stick a red ribbon on their heads without
 > them noticing it. Then let them view themselves in a mirror. Most
 > will show no extra interest in their reflection.  A chimpanzee on
 > the other hand will reach up and pluck the foreign object from
 > his or her head using the mirror as a guide.
 > 
 > My friend contends this proves the chimp is "self aware" and
 > there for conscious. I argue that red ribbon experiment only
 > demonstrates visual-spacial skills and a machine could be built
 > which would pass the ribbon test. Such a machine would be far
 > from conscious.
 > 
 > It would have to
 >      1. poses vision
 >      2. be able to differentiate reflected and unreflected views
 >      3. identify its own image
 >      4. identify changes in its own image
 > 
 > Or have I missed something? Would this be so terribly difficult
 > or dose the red ribbon experiment really demonstrate self
 > awareness?
 > 
 > ---------------------------------------------------------------
 > Dan Boone                           Northern Arizona University
 > BOONE@NAUVAX.UCC.NAU.EDU            Flagstaff AZ


I'm not one to try to define when a machine is "conscious" or
"intelligent", but there I do see one fundamental difference between
the chimpanzees and your robot.

I would suppose that chimps do not have any knowledge of what a mirror
is or even a reflection for that matter, so they must be doing some
sort of causal reasoning involving their own actions and what they see
in the mirror. Furthermore, they must have a model of themselves in
order to notice that there is a foreign object on their head. What
does it mean for an object to be "foreign" unless there is an
understanding of what is an intrinsic part of the "self"?

The robot, however, would simply be programmed to achieve this one
specific task of removing a foreign object. One could counter by
saying that the chimp is likewise programmed for that task. The
difference is that if a new identity test were developed based on
something other than visual differencing, then the robot would surely
have to be reprogrammed to pass that test as well, indicating that it
does not truly have a sense of self.

Of course, one could keep adding new capabilities to the robot, such
as the type of self-checking and correcting behavior mentioned with
the anti-viral program, and eventually one might call the robot "self
aware". The problem is that when one decides to make that statement is
highly arbitrary since the concept is not well-defined. So I'll just
go back to saying it's meaning less to try to decide if a robot is
"conscious", "intelligent", "self aware", etc.

Enough blathering for now.

	Julio

__________________________________________________________________________
Julio Kennedy Rosenblatt			Carnegie Mellon University
jkr+@cmu.edu					5000 Forbes Ave.
Phone:	(412) 268-6880				Smith Hall, Robotics
Fax:	(412) 621-1970				Pittsburgh, PA 15213-3891

  There is only one difference between a madman and me. I am not mad.
                                                           - Salvador Dali
