From newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!rpi!think.com!spool.mu.edu!uunet!psinntp!scylla!daryl Tue Jan 21 09:26:56 EST 1992
Article 2863 of comp.ai.philosophy:
Path: newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!rpi!think.com!spool.mu.edu!uunet!psinntp!scylla!daryl
>From: daryl@oracorp.com
Newsgroups: comp.ai.philosophy
Subject: Re: Table-lookup Chinese speaker
Message-ID: <1992Jan18.135808.5622@oracorp.com>
Date: 18 Jan 92 13:58:08 GMT
Article-I.D.: oracorp.1992Jan18.135808.5622
Organization: ORA Corporation
Lines: 32

Frank Boyle writes:

>>Don't be so sure of that.  Consider, where P is a person and C the computer:
 
>>P: Hello.
 
>>C: Ah, so we meet again, Professor Chung!
 
>>P: I'm not Professor Chung!
 
>>C: My mistake.  I'm sorry, Professor Han!
 
>>P: But I'm not Han either!  Are you blind?
 
>>C: I didn't think so, but I'm certainly having some trouble.

>Actually, on a slightly different thread, given my discussion with Daryl
>McCullough about his digestion analogy and the requirements one
>stipulates for testing (and since he brought up the database of 
>possible conversations), the above does not seem like a "sensible"
>(Daryl's characterization) conversation with a computer given the 
>input/output limitations. Through a teletype, the computer (or a human 
>at the other end) could not tell who was saying hello. So, a more likely
>sensible conversation that would be stored would be: 
>"Hello." "Hello, could you please identify yourself?"
>Sensibility also means conforming with the communication channel.

Thank you. That is exactly what I meant.

Daryl McCullough
ORA Corp.
Ithaca, NY


