Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Computer Chess Programs & Intelligence

Author: Enrique Irazoqui

Date: 09:24:40 03/15/01

Go up one level in this thread


On March 14, 2001 at 16:05:46, Bruce Moreland wrote:

>On March 14, 2001 at 12:30:27, Enrique Irazoqui wrote:
>
>>To beging with, they don't play. They make exclusively mechanical moves
>>following a pretedermined and invariable set of instructions. An illustration:
>>
>>[D]3k4/1r2p3/r2pPp2/b1pP1Pp1/1pP3Pp/pP2K2P/P7/8 b
>>
>>Give this position to programs time and again and until the end of times they
>>will evaluate is as a crushing win for black. That's not intelligence any more
>>than the talk of a parrot.
>
>I don't think that pefection is mandatory.  The programs handle general cases
>very well.  They are optimized to handle practical cases, and you of all people
>know how well they handle general cases.
>
>If I find a case you can't figure out in reasonable time, does that mean that
>you are a parrot, too?
>
>They aren't exactly like humans but that shouldn't matter as far as the use of
>the term "intelligence" goes.
>
>bruce

Dear Dr. Frankenstein, :)

British Encyclopedia: "Intelligence, mental quality that consists of the
abilities to learn from experience, adapt to new situations, understand and
handle abstract concepts, and use knowledge to manipulate one's environment."

"Although definitions of intelligence vary, theorists agree that it is a
capacity or potentiality rather than a fully developed attainment."

Like this one, all definitions I found of intelligence are more descriptive or
intuitive than theoretical, but all of them include the ability to learn as a
required condition of intelligence. If you decompose semantically the word
intelligence in units of significance, "learning", "understanding", "relating",
"projecting", will appear in the scheme as necessary conditions and they are all
totally or partially alien to chess programs.

Chess programs do not learn, can't apply knowledge they have to a new context,
and much less to a new field. Therefore there is no intelligence in them.

Because we know that to humans chess requires some degree of intelligence, we
may assume mechanically that a program able to play chess shows intelligence,
which is an illusion since all it does is to behave, predetermined, "as if",
without reaching the "capacity or potentiality rather than a fully developed
attainment" mentioned above.

Incidentally, it is not that I am afraid of machines becoming intelligent. On
the contrary, this "as if" is what I always found fascinating in computer chess.

Enrique



This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.