Newsgroups: rec.arts.books,comp.ai,sci.cognitive,sci.psychology.theory
Path: cantaloupe.srv.cs.cmu.edu!rochester!udel!gatech!news.mathworks.com!tank.news.pipex.net!pipex!uknet!newsfeed.ed.ac.uk!festival!castle.ed.ac.uk!cam
From: cam@castle.ed.ac.uk (Chris Malcolm)
Subject: Re: AI needs Lit to make Agents Intelligent
References: <41qn5n$jdf@Mars.mcs.com> <41t9md$k5r@Venus.mcs.com>
Message-ID: <DE8J32.EwB@festival.ed.ac.uk>
Sender: news@festival.ed.ac.uk (remote news read deamon)
Organization: University of Edinburgh
Date: Fri, 1 Sep 1995 16:28:13 GMT
Lines: 44
Xref: glinda.oz.cs.cmu.edu comp.ai:33042 sci.cognitive:9345 sci.psychology.theory:497

In article <41t9md$k5r@Venus.mcs.com> jorn@MCS.COM (Jorn Barger) writes:

>Among the goals of AI, one extreme possibility is to build a program
>that can read and comprehend the classics....

I can imagine the researcher filling out the grant application form.
She comes to the section entitled "Economic relevance of proposed
research".

"Let's see now. How about `Falling standards of higher education
[Throstlethwhyte 1989] have resulted in an ever diminishing proportion
of the human population able to read and comprehend the classics
[Haughlmsleigh 1991], therefore, er, ah, this program will, um,
......'"

>Probably there are some AI people who believe that this goal can be
>achieved by a program that starts from some simple axioms, and 
>uses *logic* to build the complex understandings necessary for the
>classics...

Fascinating! I thought this idea was confined to non-AI people! Where
might one find such AI people?

>But AI also supplies another paradigm, used by Lenat in his Cyc
>project, that assumes that if you want your program to understand
>Plato, you may have to explicitly spell out for the computer what
>it is that Plato's saying...

If you substitute "student" for "program" and "computer" this paradigm
was in use in higher education around the world long before the
computer was even thought of.....

>I think it's important for the lit side to be aware of this
>encroachment from AI, and that the AI side acknowledge that an agent
>isn't intelligent if it draws a complete blank on literature.

It seems a bit unfair, when setting a criterion of intelligence for AI
to aim for, to set one which excludes so much of the human race, not
to mention all animals, including dogs, dolphins, and pongids.
-- 
Chris Malcolm    cam@aifh.ed.ac.uk         +44 (0)131 650 3085
Department of Artificial Intelligence,    Edinburgh University
5 Forrest Hill, Edinburgh, EH1 2QL, UK                DoD #205
"The mind reigns, but does not govern" -- Paul Valery
