From newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!rpi!uwm.edu!linac!att!rutgers!rochester!cantaloupe.srv.cs.cmu.edu!crabapple.srv.cs.cmu.edu!andrew.cmu.edu!fb0m+ Sun Dec  1 13:06:41 EST 1991
Article 1757 of comp.ai.philosophy:
Path: newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!rpi!uwm.edu!linac!att!rutgers!rochester!cantaloupe.srv.cs.cmu.edu!crabapple.srv.cs.cmu.edu!andrew.cmu.edu!fb0m+
>From: fb0m+@andrew.cmu.edu (Franklin Boyle)
Newsgroups: comp.ai.philosophy
Subject: Re: A Behaviorist Approach to AI Philosophy
Message-ID: <AdBfkmC00WBME1JqUw@andrew.cmu.edu>
Date: 29 Nov 91 22:16:18 GMT
Organization: Cntr for Design of Educational Computing, Carnegie Mellon, Pittsburgh, PA
Lines: 58

Brian Yamauchi writes:

> True, but as many, many people have pointed out -- Searle isn't right.
> I still have yet to see a convincing rebuttal to the "Systems Reply".

Well, let me try the "Nonsense Room", which I briefly outlined at
the end of a previous posting.

Suppose we have what might be called the "Nonsense Room", which is similar
to the Chinese Room except that the symbols fed into the room and those
which form the rules in the rule book for processing the input (that is, 
the part of the rule book which is not part of the instructions on how to 
use it) were intended by the creators of that book to refer to 
nothing. That is, they were never intended to have any meaning whatsoever!  
Such a rule book would be similar to the rule book in the Chinese Room 
to the extent that the person using it to process the input would go through 
the same sort of rule following behavior that the person in the Chinese Room
does -- matching to one or more rules initially which may then produce a
chain of matchings until an output form (the answer to a question, for
example) is reached.  And the Nonsense Rule Book can be made as complex
as necessary, that is, as complex in terms of the number of different inputs
it can process as well as the average amount of processing (rule following)
per input as the rule book in the Chinese Room. Now those outside the room 
feeding it input have a similar rule book in order to check that the right 
answers are forthcoming (that is, to check that the person in the room is 
doing the rule following correctly) since there is no meaning associated 
with the symbols being output.

So, the only difference between these two situations is that the Chinese
symbols have referents for someone who "knows" Chinese.  But these
referents are unknown to the person in the Chinese room *and* to the room + 
person + book, etc.  All of these entities "see" only symbols whose forms are 
arbitrary with respect to the forms of their referents.  *By definition*, 
the symbols in the Nonsense room have no referents, so obviously the room, 
person, etc. cannot know the referents of them.  In both cases there is no 
"understanding" by anyone or thing.  If you claim there is for the Chinese
Room, then it has to be coming from the Chinese characters themselves or the
particular set of rules.  But in what way (besides their literal forms) 
are these different than the nonsense characters and rules?  If you think
there is some difference other than literal form, then where is it?

I wouldn't think understanding can come just from the forms of
the characters or the rules, because the Chinese characters could have
been encoded by those inputing the questions (which, of course, would change 
the rules in the rule book -- the latter could, of course, also be changed 
independently of a change in the forms of the characters) so that the input
forms and rule forms would be completely different. And the person, as well 
as the room, doesn't have the encoding!

Unless you can come up with a fundamental difference between the Chinese Room
and the Nonsense Room, I would say that there is no *understanding* of
Chinese (of the kind we normally attribute to someone who speaks Chinese
and of which Searle was referring to) by the person, room, book, or anything
else that isn't beyond the walls of the room.

Frank

 


