Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site ucbvax.UUCP
Path: utzoo!linus!decvax!tektronix!ucbcad!ucbvax!dambrosi
From: dambrosi@ucbvax.UUCP
Newsgroups: net.ai
Subject: Re: Zadeh's paradox
Message-ID: <195@ucbvax.UUCP>
Date: Wed, 19-Oct-83 18:28:57 EDT
Article-I.D.: ucbvax.195
Posted: Wed Oct 19 18:28:57 1983
Date-Received: Sun, 23-Oct-83 05:17:01 EDT
References: <12579@sri-arpa.UUCP>
Organization: U. C. Berkeley Computer Science
Lines: 12

I agree with Ken Laws interpretation of the paradox.
His resolution, however, is somewhat problematical, in that it assumes
a finite subcategorization of "rarity" can handle the paradox.
I think part of the point of the paradox is that there does not
exist any finite subcategorization which will handle all possible
paradoxes of the form proposed. This suggests, then, that then 
statement "rare things are expensive" is inherently "fuzzy" and
that a statement of this fuzziness should be stored with the base statement
itself. Then, when a paradox is encountered, the "fuzziness" estimates
can be used as a guide to which statements to explore in attempting to
construct a non-paradoxical interpretation.