From newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!rpi!think.com!mips!mips!munnari.oz.au!uunet!psinntp!scylla!daryl Wed Feb 26 12:53:54 EST 1992
Article 3952 of comp.ai.philosophy:
Newsgroups: comp.ai.philosophy
Path: newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!rpi!think.com!mips!mips!munnari.oz.au!uunet!psinntp!scylla!daryl
>From: daryl@oracorp.com
Subject: That damn humongous table again
Message-ID: <1992Feb24.064728.23742@oracorp.com>
Organization: ORA Corporation
Date: Mon, 24 Feb 1992 06:47:28 GMT

Mark Rosenfelder writes:

> When you interact with the lookup-program-and-table ensemble, you are
> in effect having a predetermined dialogue with the human(s) who
> created the table. (I think Toby Kelsey made this point a few days
> ago.)

> So, the table lookup program is neither a failure of the Turing test
> (there _is_ an intelligence there), nor an example of non-human
> intelligence. (Sure, the interaction with it is occuring _via_ a
> machine, but so is the interaction with the real human in a Turing
> test.)

Mark, there was no requirement that the table's outputs be the same as
any particular human would make, only that they pass the Turing Test.
I suppose in the case where the responses are the same as those of a
particular human--call him Fred--then conversing with the table could
be considered the same as conversing with Fred. However, there are
obviously many more possible tables that pass the Turing Test than
there are human beings, so therefore there exist (in the mathematical
sense) tables that (a) pass the Turing Test, and (b) produce responses
that don't correspond to any human being who has ever lived. How do
you view those cases?

Daryl McCullough
ORA Corp.
Ithaca, NY


