Message-id: <9208250012.AA10188@SMI.Stanford.EDU>
Date: Mon, 24 Aug 92 20:07:57 EDT
From: sowa@watson.ibm.com
Subject: Ontology

Please don't take offense at my suggestion that you should avoid some
words that you like to use.  My criticism was intended in the same
spirit as the kind of criticism that I get from reviewers, editors, and
helpful readers about things that I write.  My general attitude about
such criticism can be summarized in two sentences:

 1. Whenever reviewers say that any passage in my writing is unclear,
    confused, misleading, or ambiguous, they are almost always right.

 2. But whenever they make a suggestion for fixing the passage,
    they are almost always wrong.

I said that the word "ontology" was being used in a way that can cause
problems.  It is a rare word that is being used in a sense that is not
part of its usual meaning among the small minority who actually use it.

I think that Bob MacGregor's point that "ontology" is a word we may
have to live with is probably correct.  But one solution is to avoid
using the term except as a vague description of the whole enterprise.
Although the purpose of the project is to work on ontology, in the
actual output of the project, define certain collections as "terms",
"definitions", "axioms", "vocabularies", "theories", etc.  Then there
would never actually be any specific product called an "ontology" --
that would just be the overall name of the whole project.

There are lots of precedents for that.  Even though the activity we
are engaged in may be called "artificial intelligence", the things that
people actually produce usually have a much narrower, more precise
definition.  And speaking of confusing terms that we have to live with,
"artificial intelligence" is probably one of the worst.