From newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!rpi!usc!sdd.hp.com!spool.mu.edu!uunet!mcsun!uknet!edcastle!fofp Thu Jan 16 17:21:40 EST 1992
Article 2711 of comp.ai.philosophy:
Path: newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!rpi!usc!sdd.hp.com!spool.mu.edu!uunet!mcsun!uknet!edcastle!fofp
>From: fofp@castle.ed.ac.uk (M Holmes)
Newsgroups: comp.ai.philosophy
Subject: The English Room
Message-ID: <16523@castle.ed.ac.uk>
Date: 14 Jan 92 15:05:07 GMT
Article-I.D.: castle.16523
Distribution: comp
Organization: Edinburgh University
Lines: 42

This has been bothering me for a while.

The setup of the Chinese Room is that an English person with no
"understanding" of Chinese is given a set of rules for manipulating
chinese symbols. These are so complete that a chinese person conversing
with the Room will conclude that they are talking to another chinese
person (or technically won't be able to differentiate between a chinese
person and a machine doing the conversing).

Now using the same premises that Searle uses in this thought experiment
suppose that the chinese interlocutor has a subtle grasp of the problem
he faces in the Turing test and also knows the Chinese Room argument.
He arrives with a book of rules, in Chinese, which allow the Chinese
Room to manipulate English symbols in such a way that an English
interlocutor could converse with the "English Room" and would be unable
to tell if they were conversing with a machine or an English person,
i.e. the English Room would pass the Turing test in exactly the same way
as the original Chinese Room.

I'm curious as to whether under Searle's argument, the English Room
would be said to "understand" English. 

I'd expect the answer to be that the English Room did not understand
English since this would be consistent with the Chinese Room not
understanding Chinese. However, we know that the person inside the
English/Chinese Room does understand English so this would then be
eliminated as an important criterion for whether the English Room understands
English. This is a problem because the Chinese Room argument appears to
rely on the fact that it can be said that the person in the Chinese Room
does *not* understand Chinese (and therefore by implication neither does
the Chinese Room).

If Searle were to argue that the English Room *did* understand English
that would clearly be inconsistent with the argument that the Chinese
Room did not understand Chinese. After all, in both situations the
person inside the room is simply manipulating Chinese symbols and it is
an axiom that he does not "understand" these.

I'd be interested in any thoughts on this. I'd be *really* interested in
Searle's thoughts on this.

Mike Holmes


