From newshub.ccs.yorku.ca!torn!utcsri!rutgers!uwm.edu!wupost!usc!elroy.jpl.nasa.gov!swrinde!mips!cs.uoregon.edu!nntp.uoregon.edu!news.u.washington.edu!milton.u.washington.edu!sanelson Tue Jul 28 09:41:28 EDT 1992
Article 6458 of comp.ai.philosophy:
Path: newshub.ccs.yorku.ca!torn!utcsri!rutgers!uwm.edu!wupost!usc!elroy.jpl.nasa.gov!swrinde!mips!cs.uoregon.edu!nntp.uoregon.edu!news.u.washington.edu!milton.u.washington.edu!sanelson
>From: sanelson@milton.u.washington.edu (S. A. Nelson)
Newsgroups: comp.ai.philosophy
Subject: Re: Defining intelligence
Message-ID: <1992Jul15.233344.6478@u.washington.edu>
Date: 15 Jul 92 23:33:44 GMT
References: <1992Jul8.092458.3088@otago.ac.nz> <1992Jul15.013626.24984@dcs.qmw.ac.uk> <BILL.92Jul14224037@ca3.nsma.arizona.edu>
Sender: sanelson@u.washington.edu
Organization: University of Washington, Seattle
Lines: 42

>   I'd say that the ability to solve problems is a RESULT of intelligence,
>   and certainly that there's more to intelligence than logic. If we look
>   at human production in the fields of literature, art, music, (you get
>   the picture) I think we'd all agree that they require intelligence, but
>   they're not exactly what we'd call problem-solving.
>
>The goal, in, say, writing a novel, is to produce *something of
>value*.  Since the space of novel-length word sequences is enormous,
>and the subset of high value is relatively small, producing a good
>novel is a search problem of formidable proportions.  Finding a
>solution demonstrates a great deal of intelligence.

	I would have to disagree that the goal here is to produce something
of value. I wrote a novel because I had an idea that I wanted to put on paper
and show to others later, and the idea was a fictional one. "Value," whatever
that is, had nothing to do with it. I, for one, am not ready to say that so-and-
so's real goal is to create something of value and they just *think* it was to
make money or whatever. We have no concept of "goal" that is more reliable or
enlightening than the common sense one. There's this section of Wittgenstein's
"Philisophical Investigations" (I forget the number) in which he points out
that a major problem in thinking about the mind is the tendency to ascribe an
inner process to an outward behavior and then say "Later we'll investigate the
details of this process. Eventually it must become clear." The fact is, we have
no idea what goes on when one writes a novel, and no way (that I can see) of
finding out. Calling novel-writing a "search problem" adds nothing to our
understanding of the behavior and will probably prove misleading later on.

>I don't think it's useful to identify intelligence with everything
>that's happening in the brain.  Most people are comfortable speaking
>about a dog or a bear as having behaved intelligently in certain
>circumstances.

	I tend to think that they are comfortable with that because the word
intelligence is so indefinable. If you ask them if the dog reasons, they are
likely to say no. Then if you ask them if something that doesn't reason is
intelligent, I'm betting on another no. Furthermore, I've noticed that people
think animals are intelligent when they act like people, and this includes many
behaviors that are not problem-solving. I'm with the Turing-Test crowd in that
I think the closest thing we have to defining "X is intelligent" is "X behaves
like me."
					-Sunny Nelson
					sanelson@u.washington.edu


