From newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.ecf!utgpu!jupiter!morgan.ucs.mun.ca!nstn.ns.ca!aunro!ukma!asuvax!cs.utexas.edu!convex!cash Tue Nov 26 12:32:22 EST 1991
Article 1584 of comp.ai.philosophy:
Xref: newshub.ccs.yorku.ca rec.arts.books:10637 sci.philosophy.tech:1115 comp.ai.philosophy:1584
Path: newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.ecf!utgpu!jupiter!morgan.ucs.mun.ca!nstn.ns.ca!aunro!ukma!asuvax!cs.utexas.edu!convex!cash
>From: cash@convex.com (Peter Cash)
Newsgroups: rec.arts.books,sci.philosophy.tech,comp.ai.philosophy
Subject: Re: The Philosophical Foibles of John McCarthy
Message-ID: <1991Nov25.194234.10663@convex.com>
Date: 25 Nov 91 19:42:34 GMT
References: <1991Nov25.164015.13499@leland.Stanford.EDU>
Organization: The Instrumentality
Lines: 36
Nntp-Posting-Host: muse.convex.com
Poster: usenet@convex.com (news access account)

In article <1991Nov25.164015.13499@leland.Stanford.EDU> francis@hanauma.stanford.edu (Francis Muir) writes:
...
>But back to the point. What makes AI and Turbulence so interesting
>for me, and, apparently, so dangerous to some others, is their shared
>sense of misdirection. It is not the solutions that are troublesome but 
>the feeling that the problems are improperly posed.

That may be--I've certainly never gotten entirely clear about what "AI" is
supposed to be. If the endeavor of AI is directed to building something
that functions in some ways like a human being, then the problem is pretty
clear--and even uncontroversial. For example, there are impressive "expert
systems" in existence now (e.g., for medical diagnosis), and there is the
reasonable expectation that such systems will be even more impressive in
the future.

I suppose that such an "expert system" or a chess-playing program could be
regarded as being a sort of "artificial intelligence". If one understands
this in a metaphorical way, then it is not terribly problematic. Such
software could be thought of as a metaphorical doctor or chess player, much
as an early steam shovel could be thought of as a metaphor for John Henry,
the "steel-drivin' man".

The project of AI becomes much more muddled when it is seen as the attempt
to duplicate human beings in more general--and therefore vaguer--ways. If
one asks, "Can a machine think?", then the question has become
controversial (and philosophical). But why is it controversial? I think the
question is controversial because it is--as Francis Muir notes--"improperly
posed". The assumption that underlies this question is that the cognate
question--"Can humans think?" is clearly understood and uncontroversial.
But that is a mistake.

--
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
             |      Die Welt ist alles, was Zerfall ist.     |
Peter Cash   |       (apologies to Ludwig Wittgenstein)      |cash@convex.com
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~


