Re: Turing Test Meaningless?

[rubywand@aol.com (RUBYWAND)]

In article <4lgnj3$n9@nntp.seflin.lib.fl.us>,
z007400b@bcfreenet.seflin.lib.fl.us (Ralph Silverman) writes:

>
>: >This (the Turing Test) is meaningless, in my opinion.  I think we need
a
>new 
>: >test.  What do you think??
>
>: I think the Turing Test is an interesting idea, and great fun at
parties,
>but
>: you are right, it has fundamental flaws. Still, When Turing thought of
it,
>it
>: was a stroke of genius. It has caused much speculation and drawn
attention
>to
>: the field of AI.
>

     It may be that Turing was having a little fun with 'serious' students
of intelligence. After all, the test merely requires that a machine frame
seemingly responsive replies. The trick is that the machine's
"intelligence" is, in effect, created by those receiving the replies. Even
worse, a dull dumbo is more likely to spot the computer than someone who
is very bright and imaginative. Evidently, Turing had a sense of humor.


>: A flaw in the TT is that it only deals with human intelligence. A
martian
>would
>: have a hard time convincing anybody that it is really human.

     This asserts that something called "human intelligence" exists as
distinct from "martian intelligence". What about "intelligence
intelligence"?

     If you sat the martian down in front of a TV set for a month, he/she
might lose a few IQ points; but, the martian would probably do fine as a
Turing Test "human" substitute. 


>
>: The nature of intelligence makes it very hard to determine whether
>something is
>: intelligent or not.

     On the other hand, one might observe: "The nature of intelligence
makes it very easy to determine whether something is intelligent or not."


>
>: "If the human mind was so simple that we could understand it, we would
>: be too simple to understand it." - Pat Bahn
>

     True enough. Indeed, if we were smart enough to understand the human
mind, it would be too simple to permit understanding.


				Morten.
>: --
>: Morten Lauritsen       | Student at the Department of Computer Science
and
>: Spobjergvej 38 vaer 10 | Programmer at the Department of Genetics and
>Ecology
>: 8220 Brabrand          | at the University of Aarhus
>: Denmark, Europe        | Email: moffe@daimi.aau.dk Web:
>www.daimi.aau.dk/~moffe
>
>--
>***********begin r.s. response************
>
>"...the nature of intelligence makes it very hard to
>	determine whether something is intelligent or not..."
>
>	what if it can beat the pants off you at chess?
>		
>***********end r.s. response**************
>

     Easy. A computer that beats the pants off you at chess is not
intelligent, it's just lucky!
 
  



Rubywand