Re: The semantics of new-ediFritz Lehmann <firstname.lastname@example.org>
Date: Fri, 23 Dec 1994 01:15:52 -0800
From: Fritz Lehmann <email@example.com>
Subject: Re: The semantics of new-edi
References: <firstname.lastname@example.org> <D122yL.JCL@cix.compulink.co.uk>
Organization: University of California, Irvine
In article <D122yL.JCL@cix.compulink.co.uk>, on usenet bit.listserv.edi-l,
Chris Johnson <email@example.com> wrote:
>I am curious as to why you choose to refer to 'ontology-based EDI'. I
>seem to remember from student days that ontology was the study of systems
>of being, and epistemology the study of systems of knowledge.
>Superficially I would have expected 'epistemology' to be more appropriate
>in this context. The choice of word has interesting implications and I
>think an explanation of the reason for its usage could be informative.
The new use of "ontologies" in AI is based on the original
meaning in philosophy. It means a system of concepts and
relations which correspond to useful ones in the real world, and in
particular includes "base-level" subjects such as space,
time, shape, change, causality, etc. More specialized
concepts have formal definitions in terms of the more primitive
base-level entities. Thus the computer "knows" that a
CONTAINER is a PHYSICAL-OBJECT which fully or partially
ENCLOSES another object or SUBSTANCE for the PURPOSE
of TRANSPORT, etc. An ontology then is represented
by a large formal representation (in Conceptual Graphs,
KIF, HOL, SNePS, CYC-EL, logic, etc.) of a generally useful
set of real-world concepts, defined using primitives, with
extensive constraints and type-restrictions on the content.
This supplies the answer to "What things, concepts and
relations are there?" which is ontology, rather than "How
can we know this?" which would be epistemology.
Yours truly, Fritz Lehmann
GRANDAI Software, 4282 Sandburg Way, Irvine, CA 92715, U.S.A.
Tel:(714)-733-0566 Fax:(714)-733-0506 firstname.lastname@example.org