Newsgroups: comp.ai.philosophy,comp.ai,alt.consciousness
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!howland.reston.ans.net!gatech!psuvax1!news.ecn.bgu.edu!siemens!princeton!flagstaff.princeton.edu!schechtr
From: schechtr@flagstaff.princeton.edu (Joshua B. Schechter)
Subject: Re: Thought Question
Message-ID: <1995Jan29.034937.24962@Princeton.EDU>
Originator: news@hedgehog.Princeton.EDU
Sender: news@Princeton.EDU (USENET News System)
Nntp-Posting-Host: flagstaff.princeton.edu
Organization: Princeton University
References: <3fmba3$nmf@vixen.cso.uiuc.edu> <1995Jan19.214105.19878@Princeton.EDU> <3g5mqm$l8n@tuegate.tue.nl>
Date: Sun, 29 Jan 1995 03:49:37 GMT
Lines: 49
Xref: glinda.oz.cs.cmu.edu comp.ai.philosophy:25050 comp.ai:26871

In article <3g5mqm$l8n@tuegate.tue.nl> Fred Schenk <f.a.schenk@stud.tue.nl> writes:
>schechtr@flagstaff.princeton.edu (Joshua B. Schechter) wrote:
>
>>Turing test roughly defines thinking as conversing in a manner indistinguishable
>>from a human being.  Thinking is what humans do.
 
>> I always interpreted the Turing test to be a test which gives evidence
>> for thinking, and neither proves or defines what thinking is. In other
>> words, if some machine passes the Turing test, it gives us a good
>> indication that it may be carrying out processes which we define to be
>> "thinking." This evidence is about as strong as the evidence we get
>> when interacting with someone else (minus reasoning by analogy, which
>> is fairly weak anyway.) The strength of the Turing test, IMHO, is that
>> it DOES NOT define what thinking is.
>
>Hello Josh,
>In his reply on the Turing test (the Chinese room) Searle tried to show that
>the Turing test roughly defines thinking as conversing in a manner 
>indistinguishable from a human being. From your comment I make that you don't
>agree with Searle. Why not?

Hello Fred,

To clarify my point on the Turing test (and I must admit, I haven't
read his original paper in quite some time,) I disagree with Searle,
but perhaps only on a technicality. The Turing test only can test to
see if something converses "in a manner indistinguishable from a human
being," no more. However, I don't view the test as showing that this
something conclusively is "thinking" or "intelligent." I would define
intelligence (if I could come up with a coherent definition) to be
other than the ability to engage in a human-like conversatioon. In
many ways, it seems that it should be defined more broadly (and with
more qualifications,) including the abilities of analysis, synthesis,
etc. And, this definition seems tro too narrowly restrict intelligence
to what is familiar, forgetting about the possibility of completely
alien modes of intelligence. What this test does provide is a
reasonable criteria for believing that there is intelligence somewhere
in the system. After all, this seems to be all the evidence we humans
use in judging each other intelligent.

I would say that the Turing test is just a tool. If a computer could
manage to pass it sucessfully, it would give us evidence that it
perhaps was "intelligent" (however that word should be defined.)
Taking it as the definition of intelligence seems to me to be too
narrow, and something almost nobody would agree upon.

Anyway, I hope this helped clarify what I meant.

I welcome any comments/criticism/questions, etc.
