Re: clarifying clarifying ontologies"Jorn Barger" <email@example.com>
Subject: Re: clarifying clarifying ontologies
To: firstname.lastname@example.org (Fritz Lehmann)
Date: Tue, 8 Aug 1995 17:37:35 -0500 (CDT)
From: "Jorn Barger" <email@example.com>
In-reply-to: <9508082057.AA10449@rodin.wustl.edu> from "Fritz Lehmann" at Aug 8, 95 03:57:43 pm
X-Mailer: ELM [version 2.4 PL24]
> This example, along with Sowa's recent upper ontology, Parker-Rhodes'
> base-domains, specialized KL-ONE "BOXES", and the "conceptual scales"
> of Ganter & Wille, all illustrate the important and generally unduly-
> neglected _product_structure_ of a taxonomy. Much of the large-scale
> structure of conceptual hierarchies can be analyzed into a lattice or
> other poset which is a direct-product (or my fret-product where there
> are symmetries) of independent factor-hierarchies. [...]
(Is this my cue? ;^/
I've been arguing on comp.ai and elsewhere that taking the product of
a hierarchy with itself (which I think is *more* general than taking
the product of separate hierarchies) requires a 'fractal thicket' in
which each node contains a (partial) image of the whole hierarchy...
For example, if your hierarchy includes:
then you want your product to include (person tool) as well as
(female thing). To handle all these combinations a fractal thicket
is most efficient. (You don't instantiate meaningless combinations,
but you do allow products of three or more elements as well.)
Semantically, these products can be used to index all *relationships*
between the given elements...