From newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!rpi!think.com!samsung!crackers!m2c!wpi.WPI.EDU!johndunk Mon Jan  6 10:30:19 EST 1992
Article 2470 of comp.ai.philosophy:
Path: newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!rpi!think.com!samsung!crackers!m2c!wpi.WPI.EDU!johndunk
>From: johndunk@wpi.WPI.EDU (John Dunkelberg)
Newsgroups: comp.ai.philosophy
Subject: Re: Intelligence testing
Message-ID: <1992Jan2.024621.22198@wpi.WPI.EDU>
Date: 2 Jan 92 02:46:21 GMT
References: <1992Jan1.115429.2331@arizona.edu>
Distribution: world,local
Organization: Worcester Polytechnic Institute
Lines: 43

Let's see if this survives posting unlike my previous post..

In article <1992Jan1.115429.2331@arizona.edu> bill@NSMA.AriZonA.EdU (Bill Skaggs) writes:
>
>The Turing test has often been criticized as too weak, but in
>my view it is actually much too stringent to be a good test
>for machine intelligence.  Suppose, instead of applying it
>to a computer, we apply it to an alien creature from the planet
>Zeta Galactase -- we call the creature intelligent if and only
>if it can imitate a human being on a teletype.  Obviously this
>is human chauvinism of the rawest kind.  If it is unfair to
>apply such a test to an alien creature, how can it be fair to
>apply it to a computer?
>

	The Turing test only applies to machine-intelligences created
by humans to replicate the human mind.  While this is the intention
some some AI enthusiast, it is not the intention of all.

	The question of intelligence tests versus knowledge tests is a
big one, as one can see over the perpetual war over standardized testing.
The common human "IQ" test is a quotient between one's chronological
age and mental/knowledge age (as applied by the biased testing group).
Thus, for instance, as an avid reader as achild I had a tremendous "IQ",
since it was compared to the average american child.  My "IQ" now would
be considerably lower... have i become less intelligent?

	As an additional point, one of my favorite side-effects of the
Turing test is that to pass the test, we must teach the intelligence to lie.
A fundamental part of intelligence?

	Though without a basis, I think we have *no* way to measure the
intelligence of an alien mind.  Any proposals for such a test?  ET searchers
are hoping that the prime numbers and mathematics are shared by all, and
thus might consititute the beginning of such a test...

John

-- 
-----------------------------
|  John S. Dunkelberg Jr.   |
| (Dimensionally Different) |
|  johndunk@wpi.wpi.edu     |


