[lit-ideas] Re: The Answering Machine

  • From: John McCreery <john.mccreery@xxxxxxxxx>
  • To: lit-ideas@xxxxxxxxxxxxx
  • Date: Thu, 12 Jan 2012 09:01:20 +0900

On Thu, Jan 12, 2012 at 12:22 AM, John Wager <jwager@xxxxxxxxxx> wrote:

> I don't see that the "Turing test" is really a test of whether machines
> can think; the person who is communicating to me BY MEANS OF that machine
> is still a person, not a machine.
No. We need to distinguish between a Turing machine and a Turing test.

A Turing machine is simply the most elementary form of a computer; like
other computers, it can generate an infinite number of different results in
response to input. It isn't restricted to a small finite set of answers
(usually with one member) the way an answering machine is.

A Turing test is based on the idea, you've got this right, that if you talk
with the machine can can't tell the difference between talking with a
machine and talking with a person, you call the machine intelligent.

The mistake is in thinking that by talking to the computer you are
interacting in the same way that you are with an answering machine. The
answering machine is only storage, a place where you leave a message,
intended for the person who reads it. Anything you hear is prerecorded and
will not be modified by your interaction with the machine.

In the case of the Turing test, you aren't just leaving a message; you are
interacting with the machine. There is no one else receiving your message.
Other scientists may read the record of the interaction to see how closely
it mimics interaction between humans. But this is only listening in. You
aren't talking to them. They aren't receiving and responding to your
message. The machine is.

John McCreery
The Word Works, Ltd., Yokohama, JAPAN
Tel. +81-45-314-9324

Other related posts: