Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!ames!elroy!orion.cf.uci.edu!uci-ics!venera.isi.edu!smoliar
From: smoliar@vaxa.isi.edu (Stephen Smoliar)
Newsgroups: comp.ai
Subject: Re: Question on Chinese Room Argument
Summary: scepticism meets naivite
Message-ID: <7785@venera.isi.edu>
Date: 15 Mar 89 21:40:45 GMT
Sender: news@venera.isi.edu
Reply-To: smoliar@vaxa.isi.edu.UUCP (Stephen Smoliar)
Organization: USC-Information Sciences Institute
Lines: 64

In article <2573@crete.cs.glasgow.ac.uk> gilbert@cs.glasgow.ac.uk (Gilbert
Cockton) writes:
>  I'm all for scepticism about meanings carried in public
>language, but this scepticism should apply equally to the products of
>AI, but it's all hope, "don't smother the budding flower" and "astronomy
>took centuries" there.  All I can say from your account of scholarship
>and intuition is that it smacks of hypocrisy: the products of AI come
>in for far less searching scepticism than the everyday intuitions in
>our language (and as we don't speak the same dialects of English, I
>don't know if our common sense understandings are shared).
>
Gilbert, it sounds to me like your impressions of AI are based solely
upon claims of various hucksters who try to sell their products in its
name.  I really wonder how many of us who have participated in this recent
argument over Searle regard any of these products as any sort of a benchmark
of our progress in the study of mind.  You don't, that is clear;  but while
I spend the better part of my working day worrying about designing models of
memory and reasoning, I do not feel I am "betraying my profession" by rejecting
those products as strongly as you do.  I get the impression that you have
chosen to ignore the whole issue of "idols of the market," rather that
recognize my words as a "limp restatement."

Here is a little bit of slightly oversimplified history:

	In 1957 Allen Newell published a paper in which he listed
	all sorts of wonderful "intelligent" things computers would
	be able to do in ten years.

	In 1967, Herbert Dreyfus proclaimed a loud "Aha!" since none
	of Newell's predictions had been achieved.  He then wanted
	to throw out the baby with the bathwater, claiming that the
	rest of us should stop wasting our time.

	In 1977, Ed Feigenbaum was telling us about the wonderful
	future which expert systems would bring us.

	In 1987, we started reading papers about what was wrong with
	expert systems.  This time, however, the criticism was coming
	from voices within the AI community.  Some of us, for example,
	were beginning to grasp what Marvin Minsky was talking about
	in questioning our very assumptions about how to model memory
	or how to process language.

Humberto Maturana describes come provocative uses of language as "triggering
perturbations."  What he means is that we should not take these phrases at face
value.  Instead, we should treat them as incentives for our own thought.  Where
you see hypocrisy, some of us see triggering perturbations.  Stop wasting bytes
preaching against the huckters.  They don't waste their time reading this
stuff anyway.

>If you know what you're on about, if you know how you judge progress
>in AI, if you know what marks out a good AI research proposal from a
>poor one, then you should share it with us.  You are doing AI, so tell
>us how you do it, what it tells us, and why we should believe you.
>
As my former composition teacher used to say, "While I may be incapable
of laying an egg, I know when one is fresh."  Like Dreyfus, you would throw
the baby of our curiosity out with the dirty bathwater of admittedly worthless
commercial AI products.  Remember the challenge in THE LAST UNICORN:

	Your true task has just begun, and you may not know in your
	life if you have succeeded in it, but only if you fail.

Walk for a while in our shoes before you consign our scholarship to the flames.