From newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!rpi!usc!cs.utexas.edu!uunet!psinntp!scylla!daryl Tue Jan 21 09:27:28 EST 1992
Article 2922 of comp.ai.philosophy:
Newsgroups: comp.ai.philosophy
Path: newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!rpi!usc!cs.utexas.edu!uunet!psinntp!scylla!daryl
>From: daryl@oracorp.com
Subject: Re: Intelligence Testing
Message-ID: <1992Jan20.191915.14755@oracorp.com>
Organization: ORA Corporation
Date: Mon, 20 Jan 1992 19:19:15 GMT

Jeff Dalton writes: (in response to Barbara H. Webb)

>> The reason for using the test now is because it is an acceptable
>> hypothesis that "anything that could converse like a human must
>> have understanding". Not a proven hypothesis. An acceptable one.

> But why?  Why should we use the Turing Test?  There aren't any robots
> running around that can pass it, so that we could use it to decide
> whether or not they had intentionality (or understanding or consciousness
> or whatever).  It can't be used to refute someone who questions
> whether TT behavior always involves understanding, because then it
> would be begging the question.  For the same reason, it can't be used
> to show that, in the Chinese Room, "the system understands".

One reason for using the Turing Test is that it provides an objective
measure for computer competence. It gives AI researchers a goal to
shoot for in developing machines that have intelligence comparable to
humans. You and Searle (as well as many AI researchers) may believe
that passing the Turing Test is not sufficient for intelligence, but I
think that everyone agrees that if we *cannot* build a machine that
can pass the Turing Test, then we can't build a machine with
human-like intelligence.

A second reason for using the Turing Test is that, for the practical
purposes that one might want Artificial Intelligence, it is good
enough to have a machine that can pass the Turing Test. If the desire
is to have machines that can receive instructions in English and carry
them out (asking clarifying questions if necessary), then developing a
machine that can pass the Turing Test (behave *as if* it understands
English) is most of the problem. The distinction between "real
understanding" and "simulated understanding" doesn't make a difference
for most applications.

A third reason for using the Turing Test is the enormous reflective
power of human language. A corresponding Turing Test for dogs, which a
computer would pass if it barked at the appropriate times, would not
convince anyone that someone had somehow captured the essence of a
dog's mind. However, in the case of human beings, we know that we can
talk about our inner mental states. We can talk about being hungry, or
our plans for the future, or our preferences in art and music. Of
course, there are some things that we don't know words for, for
example, it is very difficult to put into words the difference between
the smell of a rose and the smell of lilacs, even though the
difference is clear to us. But these differences do make a difference
in future conversations, since we would say things like "Susan's
perfume that night smelled more like lilacs than roses" that make
correlations between undescribable things.

By asking a being (machine, human, or Martian) to talk about his
feelings, his plans for the future, his preferences in art, etc. we
are probing very deep connections between the being's past experiences
and his thoughts and feelings about them. Because of the reflective
nature of human language, the Turing Test is *not* a superficial,
surface-level test of understanding, but is as deep as can be conveyed
by language. Of course it is only a hypothesis that human language can
tell us enough about our inner thoughts to give evidence of
understanding, but I think it is a plausible hypothesis.

The sufficiency of the Turing Test is not an argument against Searle,
it is an assumption that may be wrong. The argument against Searle is
that he has not shown that the Turing Test is not sufficient, and there
are good reasons to believe it is sufficient.

Daryl McCullough
ORA Corp.
Ithaca, NY




