Date: Sun, 19 Jan 1997 12:10:45 -0700 From: lgorbet@unm.edu Subject: Re: Logic, etc Sender: owner-linganth@cc.rochester.edu To: linganth@cc.rochester.edu MIME-version: 1.0 Precedence: bulk This whole discussion of "logic" in natural language is pretty amusing to me in its almost apologetic tone. As though the "non-Classical" logic suggested by natural language form were a survivable defect, a legacy of our mere humanity. Or, at best, irrelevant to real reasoning.

To be just a little crude, I'd say the proof of a logic is in the "eating". A logic is a means for reaching valid conclusions.

The fact is that the best (i.e. fastest, most effective) *machines* for reasoning don't use "classical" logics. In two senses. First, they are highly complex and irregular --- i.e., they look at a big picture and send the logic down various often special-case paths; moreover they often begin with a *guess* and see how it pans out. Second, many make use of fuzzy logics (witness the circuitry in current camcorders etc.) to guide "decisions". They do this not to be humane but to work better.

The fact is that "classical" logic is not just something we don't have built in to natural language (as a whole). It would be a hindrance to effective reasoning if it were there.

Larry Gorbet lgorbet@unm.edu
Anthropology & Linguistics Depts. (505) 883-7378 University of New Mexico
Albuquerque, NM, U.S.A. Go Back to Aaron's Remarks