Re: Searle's Chinese Room Argument

From: Fox, SJ (
Date: Wed Feb 19 1997 - 11:36:29 GMT

The Chinese room argument is a very good example of a computer program,
and shows how we as humans do not treat language as a kind of
"stimulus-response" system.

The man in the room (who can be thought of as the computer) has his set
of rules (which can be thought of as a computer program), and he
receives his messages, in Chinese, which he responds to by consulting
his rules.

This is all well and good, provided that the person giving him the
messages does so in the same way as the person who wrote the rules for
the man; what I mean is that if the messages are not given in perfect
grammar, for example, the response from the man after consulting the
rules would either be nonsensical or non-existent. This is where the
analogy with the computer becomes most apparent: most people who use
computers will realise that they often blurt out nonsense, or nothing
at all, in response to an imperfect input. This is what would happen to
the man in the Chinese room faced with a grammatically imperfect
sentence - he'd either reply wrongly, or concede that his rules did not
cover a sentence of ungrammatical structure.

"We don't need no education" (Roger Waters, 1977). This line from the
Pink Floyd song goes some way to prove the above - the sentence is
technically ungrammatical. We all know what the implied meaning is, but
it is not said in the correct way, which would be a double negative,
and give the opposite meaning to that which was intended.

This is an interesting point, and one which confuses me; we recognise
that the grammar is wrong for one meaning, but we also recognise that
the grammar would be workable (if a little disjointed) for
another...imagine an "English Room Argument": a Chinese man in an
English room with English rules etc.. Given the above song lyric, his
rules would tell him that the sentence was double negative, and should
be replied to as such. But, as the use of double negative is accepted
in certain circumstances, how could any rules distinguish the response
when there could be two possible responses?

Please clarify... Simon Cox.

This archive was generated by hypermail 2b30 : Tue Feb 13 2001 - 16:23:50 GMT