[lit-ideas] Re: The Answering Machine

  • From: John McCreery <john.mccreery@xxxxxxxxx>
  • To: lit-ideas@xxxxxxxxxxxxx
  • Date: Thu, 12 Jan 2012 11:31:15 +0900

Further to Robert's comment: The Turing Machine as Turing conceived it was
a primitive computer whose machine language, zeros and ones, were printed
on a single strip of paper tape that moved back and forth under the head
that read and wrote them. This is the thought experiment that, except for
pedagogical experiments has never been realized in hardware. That said, one
of the odd things that happen when you study computer science is learning
that machines can be software as well as hardware, i.e., the logic of the
machine can be implemented either in lines of code or in physical circuits.
This is, for example, why I can run Windows XP on my Macintosh using
Parallels, a program that translates the Windows XP code into code that Mac
OSX understands, with the Mac OSX code translated ultimately into the ones
and zeros, the machine code specific to the Intel hardware. The the
processing unit, the "Intel inside" chip that processes the ones and zeros
is, in effect, an enormous cluster of tiny (now approaching nanoscale)
Turing machines that individually do nothing but read and write ones and
zeros taken from and replaced from memory, which functions as the
equivalent of the Turing thought experiment's paper tape. In this sense,
all modern digital computers are Turing Machines.

Early experiments in computer-human interaction with the computer
simulating a human being included, for example, Eliza, a computerized
non-directive therapist. When you turned on Eliza, Eliza would ask, via
words that appeared on the screen, "What's  bothering you today?" If the
human responded, "I feel very tired all of the time," Eliza would reply,
"You feel very tired...." Basically, all the computer was doing was picking
words and phrases out of the input and feeding them back to the human in
plausible linguistic frames, the sort of thing one expects to hear from a
therapist. The interesting thing was that, while Eliza's capabilities were
very primitive, indeed, many people were observed using Eliza as if Eliza
were a real therapist. The absence of intentions and the other sorts of
things that philosophers worry about didn't seem to bother them at all. In
fact, in this case, Eliza was arguably the ideal Rogerian therapist, one
who puts aside her own feelings and intentions to simply listen intently to
what the other is saying.



On Thu, Jan 12, 2012 at 9:32 AM, Robert Paul <rpaul@xxxxxxxx> wrote:

> Let me add to John McCreery's clarification, that the Turing Machine was a
> thought experiment, not an actual computer or other device although some
> people have tried to build them for (I guess) pedagogical purposes: <
> http://www.loebner.net/Prizef/TuringArticle.html>
> In Turing's paper, 'Turing Machines and Intelligence,' MIND, 1950, Turing
> does describe the kinds of machines (including digital computers) that
> might have some chance of fooling the interrogator into thinking she was
> engaged in a conversation with a human subject; and human subjects will, of
> course, be interrogated along with 'machines.'
> Robert Paul,
> caught in an almost infinite loop...

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

Other related posts: