X-Message-Number: 4320
From: 
Date: Sun, 30 Apr 1995 11:49:46 -0400
Subject: re #4318

This is way past the point of diminishing returns, but just a couple of quick
corrections to John Clark's latest (#4318):

John keeps repeating--which I have repeatedly denied--that I claim a computer
"can never be intelligent." Computers ARE intelligent in some respects and in
some degree; they will almost certainly eventually surpass human intelligence
in most ways and perhaps all ways; and they MAY  even develop feeling and
consciousness. I merely say that, until we know the nature of feeling and
consciousness in mammals, we cannot ASSUME that inorganic
computers--intelligent or not--can ever be conscious.

As for the alleged "tautology" between BEING intelligent and ACTING
intelligent, John himself admits that the Turing Test is not infallible. But
if there were a genuine tautology, the TT WOULD be infallible. Fallibility of
the TT (as well as John's reported experiences with people) PROVES that
"intelligence"--or degree of intelligence--is not necessarily easy to judge.
 (John has also contradicted himself by saying, on occasion, that he knows
intelligence when he sees it, even if he can't define it; and on the other
hand admitting that he has sometimes been mistaken.)

Furthermore--this is tiresome, but what ought to be obvious apparently isn't
always--it is commonplace for someone to think one thing and do another, or
to adopt the worst of several contemplated courses of action. All of us, one
time or another, have thought of the right response but uttered the wrong
one, thought of the right decision but acted out the wrong one. What does
this do to John's simplistic "tautology" between "being intelligent" and
"acting intelligent"?

John keeps repeating, and I keep denying, that I have said or implied that
one should ignore behavior in studying consciousness. I'm far from perfect,
but I could never say anything THAT  stupid, and never have. To understand
feeling and consciousness we need to study BOTH externals and internals. John
seems to think the internals have so little likelihood of being relevant or
useful that we should ignore them, and this I find VERY hard to
understand--except possibly as an extreme expression of abhorrence of meat
chauvinism.

How do you test a theory of consciousness? There are many possibilities. One
of the most obvious is to look for correlations between a human subject's
reported feelings in an experimental situation and his concurrent physiology
or brain states as determined by EEG-type or NMR-type (or something better)
scans; then you use one to predict the other and see how well it works. You
might even be able to create an inorganic analog and test that. 

In the end, I guess we just psychoanalyze each other. John thinks I am a
closet dualist, which of course I deny. I suspect that people who share John
's attitudes are victims of PC--terrified of meat chauvinism or of any threat
to the information paradigm.

But if we are on Cryonet, we are probably more alike than different, and we
share many goals.

R.E.


Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=4320