Newsgroups: comp.ai.philosophy
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!howland.reston.ans.net!swiss.ans.net!solaris.cc.vt.edu!insosf1.infonet.net!internet.spss.com!markrose
From: markrose@spss.com (Mark Rosenfelder)
Subject: Re: Bag the Turing test (was: Penrose and
Message-ID: <D1HLn2.8py@spss.com>
Sender: news@spss.com
Organization: SPSS Inc
References: <322280663wnr@luptonpj.demon.co.uk> <3cl4dq$dr8@cantaloupe.srv.cs.cmu.edu> <429205102wnr@luptonpj.demon.co.uk> <1994Dec23.023515.8055@news.media.mit.edu>
Date: Tue, 27 Dec 1994 20:45:01 GMT
Lines: 24

In article <1994Dec23.023515.8055@news.media.mit.edu>,
Marvin Minsky <minsky@media.mit.edu> wrote:
>Help.  Would someone please explain to me in simple terms what is
>supposed to be the point of the HLT?  I presume that it's a table that
>duplicates the behavior of a certain person P for up to a certain N bits of
>input, presumably by storing 2^N entries.  After that, presumably it
>produces outputs that have "forgotten" what happened earlier--and no
>longer acts much like P or anyone else--that is, if N is less than a
>person's memory capacity.  So it will "pass the P-person test" for
>that long.  
>
>Ok.  You can fool all of the people for some of the time, and you can
>fool some of the people for all of the time, but you can't fool all
>possible memory-sized people all of the time.
>
>Now, for the life of me, I can't figure out where there is an interesting
>philosophical problem in this.  

I presume it's a _reductio_ aimed at the idea that intelligence can be
operationally defined as passing the Turing Test... if anyone actually
believes that... and doesn't simply respond by accepting the HLT as intelligent.

Beyond that, it's just a mental construct that's fun to talk about despite
(or because of) its impossibility, like the Library of Babel or the Ringworld.
