From newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.ecf!utgpu!csd.unb.ca!morgan.ucs.mun.ca!nstn.ns.ca!aunro!ukma!wupost!uunet!mcsun!uknet!edcastle!aiai!jeff Thu Jan 16 17:19:43 EST 1992
Article 2646 of comp.ai.philosophy:
Path: newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.ecf!utgpu!csd.unb.ca!morgan.ucs.mun.ca!nstn.ns.ca!aunro!ukma!wupost!uunet!mcsun!uknet!edcastle!aiai!jeff
>From: jeff@aiai.ed.ac.uk (Jeff Dalton)
Newsgroups: comp.ai.philosophy
Subject: Re: The Robot Reply (was Re: Searle, again)
Message-ID: <5953@skye.ed.ac.uk>
Date: 10 Jan 92 20:23:11 GMT
References: <5825@skye.ed.ac.uk> <309@tdatirv.UUCP> <5908@skye.ed.ac.uk> <1992Jan08.231644.39424@spss.com>
Reply-To: jeff@aiai.UUCP (Jeff Dalton)
Organization: AIAI, University of Edinburgh, Scotland
Lines: 36

In article <1992Jan08.231644.39424@spss.com> markrose@spss.com (Mark Rosenfelder) writes:
>In article <5908@skye.ed.ac.uk> jeff@aiai.UUCP (Jeff Dalton) writes:
>>There is no equivalent supposition that humans have no understanding
>>without sensors.  
>
>Wanna bet?  You are claiming that a human could develop understanding
>without ever having had sensory input? 

No at all.  Go back and reread my message.  Note also that in your
sentence above you said almost the same thing as the part of my
message you quoted -- except you've added the word "develop".

>>Of course, sensors help in learning.  But if a
>>person was in a Turing Test, the person can ignore everything except
>>what's coming in on the teletype and still understand what's being
>>said.  
>
>Yes, by virtue of the incredible mass of information about the world
>acquired by living in it.

Just so.  Have I said otherwise?  

>>Of course, a computer might learn, with the aid of sensors, and
>>then be put in the TT.  But the same kind of information could
>>have been in the program from the start.  
>
>In which case, there are grounds to argue that the program does have
>a real semantics, since symbols can be referred to that mass of information
>about the world.

Searle argments are about programs in general.  So, if correct, (as I
keep saying in this thread), he's right about programs that have all
that mass of information.  Nothing in his argument depends on the
program not incorporating (to the extent that programs can) all kinds
of things about the world or on any other details about what's in
the program.


