Re: Kosko Profile in IEEE Spectrum

Dave Jones (djones@news.tuwien.ac.at)
Tue, 20 Feb 1996 23:08:27 +0100


dleckhar@eos.ncsu.edu (Douglas Leo Eckhart) wrote:
>Fred A Watkins (fwatkins@hyperlogic.com) wrote:
>
>: Hello all,
>
>: The latest issue (Feb 96) of IEEE Spectrum contains a very nice
>: profile of Bart Kosko, one of the brightest lights in fuzzy theory. If
>: you get a chance read the article. It's well worth it.
>
>Anyone read "Fuzzy Thinking" by Kosko?

Yes.

> Is it worth reading?
>

Could be. Could be a tenth of that.

If you enjoy an egotistical polemic against the entire tradition of
western philosophy and logical inquiry, it may be just your cup of
gunpowder green. I've got to admit I liked parts of it and found them
thought-provoking. Other parts were, well, just provoking.

The lead blurb on the jacket back is by Marvin Minsky. Make of that
what you will.

Turning to a random page I find, "Modern science and math stand on
exactly this rare case of all answers black or white -- Aristotle
boxed into corners."

Flipping again I find, "Kant might ask: Are fuzzy numbers possible?
The old answer was no."

How old is old?

David Hume talked of degrees of belief and degrees of similarity over
three hundred years ago. I re-invented what I now know as the GRNN after
reading an essay by Hume. I have since discovered that the mathematics of
"fuzzy" was investigated very extensively under the name of "kernel systems"
beginning at least as early as the 1950's, and is still being actively
researched today. The literature is very deep compared to the literature of
"fuzzy" by that name.

Just for the fun of it, I'm going to open the book to one more random spot.

Okay, this one is not so bad. A chapter called "Life and Death". I guess it
is hard to be very pretentious when speaking of mortality. There's even a
reference to Richard Feynman, my hero, who did in fact die.

I don't want to stop on that note. Let's try just one more.

"Heisenberg showed that even in physics the truth of a statement is a
matter of degree." Hum. How true is that I wonder? It's certainly not what
I thought he showed! Maybe it is only 1/3 true that he showed that. Could
Bart Kosko possibly be two thirds wrong? Perhaps it is only 1/3 true
that Heisenberg showed it is 3/4 true that the truth of a statement is
a matter of degree. Could be. Could be a tenth of that.

Half sincerely yours,
Dave