Re: CG: Re: Top level ontology

JFDelannoy (
Tue, 9 Dec 1997 10:58:52 -0400

> Date: Mon, 8 Dec 1997 22:44:21 -0500
> From: (John F. Sowa)
> To: jmc@cs.Stanford.EDU
> Subject: Re: CG: Re: Top level ontology
> Cc:,,,
>, onto-std@KSL.Stanford.EDU,

> John McCarthy writes
> >1. Maybe polysemy can be divided into accidental polysemy and
> >essential polysemy. The distinction between a river bank and a
> >financial institution is accidental. Different languages won't have
> >the same accidental polysemies. However, it seems to me that either
> >concept may give rise to polysemies in a systematic way. For example,
> >someone may refer to banking a boat meaning to approach the bank of
> >the river and to bank money meaning putting it in the bank. AI will
> >involve itself in an endless chase if it proposes to tie down all
> >these polysemies in advance. I suppose they can be avoided for
> >systems that don't have to tolerate human ad hoc invention of
> >polysemies.
> The "bank" vs. "bank" ambiguity is caused by a historical accident,
> but the kind of polysemy that arises in science is of the same nature
> as the polysemy that arises in kitchen utensils or any other topic.
> For example, just consider the word "file" in operating systems.
> In Unix, the records are separated by line feeds, but in MS-DOS,
> they are separated by a carriage-return + line-feed combination.
> In IBM mainframes, there are as many different kinds of files as
> there are "access methods" such as BSAM, QSAM, ISAM, VSAM, VTAM,
> or my favorite when you get tired of all the rest, BDAM.
> You can take almost any term in computer science and find similar
> multiplicities of meanings that can only be distinguished by context.
> The origin of the ambiguities is not a property of language, but of
> the fact that there are more possible variations of entities in the
> world than there are distinct words (or predicates) in our vocabulary.
> So inevitably, the same words are going to be used to cover different
> things and acquire different meanings.
> Another example: consider the word "automobile", which was synonymous
> with "horseless carriage" when it was first introduced. But today's
> entities called by that name are very different from the originals.
> In French, the word "voiture", which used to mean a horse-drawn
> carriage, now means automobile. They didn't pass through a stage
> when they felt a need for a separate word for the version "sans cheval."

Presumably, it was originally called "voiture automobile". (to check).

Voitures (and, respectively, IBM files) are consistent as to their main
functionality, if not their aspect. Don't forget this FUNCTIONAL or
instrumental aspect (cf Winograd and Flores, who wax philosophical about

John McCarthy's suggestion is in keeping with lexicographic practice of
distinguishing variants at two levels: HOMONYMS vs SENSES of one word.
I think this is necessary to maintain for knowledge representation.
bank1: side of a river
bank2: institution for financial operations; its building; the whole
occupation; etc.

The chase for polysemy/homonymy is endless in a philosophical sense, maybe
(computer-equipped Achilles chasing the language productivity tortoise)
but there is considerable previous work in the dictionaries.

> John Sowa

JF Delannoy

(I wish I had my Quirk grammar with me)

JF Delannoy
RES International, Ottawa