From newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.ecf!utgpu!cs.utexas.edu!sun-barr!ames!elroy.jpl.nasa.gov!usc!zaphod.mps.ohio-state.edu!sol.ctr.columbia.edu!emory!gwinnett!depsych!rc Tue Jan 21 09:26:29 EST 1992
Article 2811 of comp.ai.philosophy:
Path: newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.ecf!utgpu!cs.utexas.edu!sun-barr!ames!elroy.jpl.nasa.gov!usc!zaphod.mps.ohio-state.edu!sol.ctr.columbia.edu!emory!gwinnett!depsych!rc
>From: rc@depsych.Gwinnett.COM (Richard Carlson)
Newsgroups: comp.ai.philosophy
Subject: Re: Table-lookup Chinese speaker
Message-ID: <qFVNeB1w164w@depsych.Gwinnett.COM>
Date: 16 Jan 92 21:12:37 GMT
References: <1992Jan14.172756.46284@spss.com>
Lines: 55

markrose@spss.com (Mark Rosenfelder) writes:

> In article <1992Jan14.045132.19208@oracorp.com> daryl@oracorp.com writes:
> >The fact that there are a huge number of conversations consistent with
> >what has been said so far is irrelevant: the computer only needs to
> >select a response that occurs in *some* conversation that agrees with
> >the current conversation so far. If more than one response is possible,
> >the computer just picks one. Conversations don't have to have a unique
> >outcome.
> 
> Oh, come now.  Such a system would almost immediately fail the Turing test.
> Picture it in operation.  Your first utterance is "Hello."  The computer
> races through its database looking for conversations that begin with this
> sentence.  There are rather a lot of them; it picks one at random:
> "Ah, so we meet again, Professor Chung!!"  You are taken aback, not least
> because your name is Daryl.
> 
> Now, the conversation "Hello."  "Ah, so we meet again, Professor Chung!"
> (or rather its translation into Chinese) is indeed a possible conversation,
> so it lives in the database.  But it's not appropriate.  The table lookup
> algorithm is incapable of any sense of context.  There are simply too many
> possible conversations at every point; all its replies will be nonsense.
> 
> And no, don't talk to me about assigning probabilities to the conversations,
> either.  The next conversation you have is almost guaranteed to be unique,
> and thus can't be distinguished by its probability from any other conversatio
> 
> This algorithm is hardly worth discussing, except that I already see you 
> using it in another reply to argue that a machine that can converse in
> Chinese is theoretically possible.

Wouldn't something like a table lookup procedure work for a dumb
person trying to simulate high intelligence?  In other words if
he, Bart II for example, has _some_ semantics and _some_
understanding_, he can eliminate the obviously inappropriate
conversational "scripts," as well as jump from one script to
another if one doesn't seem to be working.

This sounds trivial until you look at the reverse, namely that you
seem to need _some_ semantics to use the table lookup, so it is
not a valid way for a machine to cheat, as you say.

Finally, however, don't even intelligent people "cheat" in this
way, not necessarily to appear more intelligent or more learned or
more cool (or whatever), but just because it is the path of least
resistance, consumes the least energy, is "efficient" in some
cost-benefit, survival sense?  Hence table lookup becomes
interesting again and the kinds of considerations you advance,
such as about database size, do become "worth discussing."

--
Richard Carlson        |    rc@depsych.gwinnett.COM
Midtown Medical Center |    {rutgers,ogicse,gatech}!emory!gwinnett!depsych!rc
Atlanta, Georgia       |
(404) 881-6877         |


