Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!cornell!uw-beaver!ssc-vax!bcsaic!ray
From: ray@bcsaic.UUCP (Ray Allis)
Newsgroups: comp.ai
Subject: Re: Question on Chinese Room Argument
Keywords: Chinese Room understanding
Message-ID: <10288@bcsaic.UUCP>
Date: 21 Feb 89 22:56:14 GMT
Organization: Boeing Computer Services ATC, Seattle
Lines: 40

>From: sher@sunybcs.uucp (David Sher)
>Subject: Re: Question on Chinese Room Argument
>
>Now that we've posted megawords on "understanding" and whether a machine
>can or can not posses it, can I ask: what is the advantage of a 
>machine with "understanding"?  Assume that HAL doesn't understand 
>anything.  He merely manipulates symbols so that he creates an illusion
>of understanding in his correspondents.  In what way does that inhibit
>HAL as a useful tool.  What could an "understanding" machine that a 
>merely intelligent (the symbol manipulator that merely gets the right answer)
>machine could not?  Unless someone can show me an advantage to it I'm not
>going to waste much time designing it into my programs.
>
>-David Sher
>ARPA: sher@cs.buffalo.edu	BITNET: sher@sunybcs
>UUCP: {rutgers,ames,boulder,decvax}!sunybcs!sher
>

Of course if the symbol manipulator "gets the right answer", the answer to
your question is "There IS no difference!"  I am one of those who doubt,
however, that it is possible for either a person or a machine to "manipulate[s]
symbols so that he creates an illusion of understanding in his correspondents".
I don't think the Chinese Room could fool a perceptive human for very long.

"Understanding" language is (at base) the evocation of experience in the
receiving organism.  Translation between languages requires
language1-to-experience followed by experience-to-language2.  You can't go
directly from English symbols to Chinese symbols.  Computers can't translate
languages because they can't experience.  (Yet.)

Searle's Chinese Room is doing transliteration, and as pointed out by an
earlier poster, rec.humor.funny just had several pages of hilarious examples of
the results of that, e.g. a sign in a furrier's shop, "Here ladies can have
coats made from their own skins".  It might not always be funny.
Understanding is more than language translation.  Suppose I instruct a computer
to "Eliminate crime in Detroit".  It returns the next day with "Done!  And it
only took one 20 megaton device, Boss."

*Lack* of understanding is THE major flaw in 30 years of "Physical Symbol System
Hypothesis" AI.