Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/5/84; site gloria.UUCP
Path: utzoo!linus!philabs!cmcl2!seismo!rochester!rocksanne!sunybcs!gloria!colonel
From: colonel@gloria.UUCP (Col. G. L. Sicherman)
Newsgroups: net.ai
Subject: Re: aesthetics and A.I.
Message-ID: <737@gloria.UUCP>
Date: Tue, 14-May-85 11:59:21 EDT
Article-I.D.: gloria.737
Posted: Tue May 14 11:59:21 1985
Date-Received: Thu, 16-May-85 21:03:47 EDT
References: <2836@sdcc3.UUCP>
Distribution: net
Organization: The Jack of Clubs Precision Instruments Co.
Lines: 42

>     This business of sonnets and concertos can certainly be overlooked.
> Just because I cannot write a concerto, and I probably could not even if I
> wanted to, does not mean that I am not intellegent or not human.

We'll grant that you're human and computers are not!  There are many
kinds of intelligence: evidently your musical intelligence is not high.
But when we ask whether somebody or something "is intelligent," what do
we mean?  (I ignore the colloquial use of the word: "most of my students
this year are not intelligent.")

>     However, I think the issue raises some valid questions that cannot be
> dismissed so easily.  Just how central to intelligence is this matter of
> artistic ability?  Even though I do not consider myself overly artistic, I can
> appreciate the "beauty" of a particularly elegant or insightful mathimatical
> proof, or of a well-writen computer program.  This capacity for appreciation
> also seems to be related to my ability to perform in these domains (I am
> considered excellent at both).  Is it possible that to make computer
> programs that posses true expertise in a given field, we will have to
> address these seemingly "artistic" problems?

I'm glad you put quote marks around "beauty." Just what is "beauty," anyway?

Take the computer-discovered proof of the isosceles triangle
proposition: showing that the triangle is congruent to itself.
To a classical geometer, the proof is ugly; to a high-school student it
may be inscrutable; but a logician may find it elegant and beautiful.

And the Four-Color Theorem!  There's expertise without beauty for you!

>     Perhaps so.  Lakoff and Johnson have shown that an ability to understand
> metaphor, something usually considered a peripheral, aesthetic ability, may be
> central to the matter of understanding natural language.  Perhaps we do not
> fully grasp the depth of connection between what we normally consider
> aesthetic issues and matters of real intrest to AI.

I do not regard metaphor as aesthetic.  The listener is, unavoidably, a
metaphor for the speaker.

> 	Stuart Ferguson   --	ucbvax!ferguson@ucbcory.Berkeley.ARPA
-- 
Col. G. L. Sicherman
...{rocksvax|decvax}!sunybcs!colonel