RE: SUO: RE: Comment #12 - 'Definition of Ontology'
Chris,
. Thanks for your reply and explanation. I can see that just
"language" doesn't adequately specify
our task.
. I have just returned to the print-outs of my look at KIF when this
project was first raised with me some 3 months ago. With the benefit of the
subsequent dialogue, KIF strikes me as fitting in with what I was referring
to yesterday as a "simplified language", etc. I didn't take in all its
definitions in the 5 minutes I just devoted to it, and will return to them
to take these on board (For what it's worth, my own background is as an
engineer, including having written knowledge based decision making software
in ForTran).
. Bearing in mind your position at A&M, do you have, or have access
to, any definitions for "ontology" that cover what we are aiming at? My
Pocket OED is totally inadequate , just talking about the field of
metaphysics involved.
. I would see it as desirable that a brief definition, or at least a
description with a relatively comprehensive scope, of "ontology" be included
in the early portions of the project documentation. I believe this would
increase the acceptance and usage of the resulting product.
. My thoughts at this stage are that a general purpose ontology is:
* a vocabulary of words, each with a unique meaning, and
comprehensively covering the known range of concepts, objects, processes,
and their attributes; in
* a precisely defined language in which the grammar allows for only a
single interpretation of any sentence structure.
. From this I suggest a definition of an SUO as perhaps:
* a vocabulary of words, each with a unique meaning, and
comprehensively covering the known range of "classes or types" of concepts,
objects, processes, and their attributes; in
* a precisely defined language in which the grammar allows for only a
single interpretation of any sentence structure.
. BTW, where is A&M? (I have never been to Texas).
Cheers Graham Horn
Australian Institute of Health and Welfare
================================================
Phone: 02.6244.1094
Fax: 02.6244.1199
Email: Graham.Horn@aihw.gov.au <mailto:graham.horn@aihw.gov.au>
-----Original Message-----
From: Chris Menzel [mailto:cmenzel@philebus.tamu.edu]
Sent: Wednesday, September 13, 2000 11:18 PM
To: Horn, Graham
Cc: Standard-Upper-Ontology list; Christensen, Joe
Subject: Re: SUO: RE: Comment #12 - 'Definition of Ontology'
> .. As a newcomer to the ontological field, I agree the original
definition is incomplete. However, Chris's suggestion still doesn't explain
enough for me to get a clear understanding.
I would think not! It wasn't meant to be an explanation, or a suggestion of
any particular wording for the Scope and Purpose, but only an observation of
an inadequacy in the original statement. The original S&P stated that an
ontology is "a set of terms and formal definitions". This itself is vague,
but it suggests that an ontology consists in a set of primitive, undefined
terms (which, of course, there must be in any basic ontology) and a set of
terms that are formally defined in terms of the primitives. This is
woefully inadequate, on several counts. Most seriously, so understood,
ontologies are meaningless, devoid of content. Note that on the standard
notion of definition, definitions are "conservative", in that (roughly) they
do not enable you to prove anything that you couldn't prove before you had
the definitions. Hence, a mere list of primitives plus definitions gives
you no more than just the list of primitives, and a list of primitives
*alone* is meaningless. What's missing, of course, are the *axioms* (or
principles, or whatever you want to call them) that capture the fundamental
properties of, and logical relations among, the chosen primitives.
Here's a simple example. Suppose I give you the primitives "Number", "0",
and "s" (aka "successor"). Now I'll give you some definitions:
1 =df s0
2 =df s1
3 =df s2
...
10 =df s9
There's my ontology. Some primitives and some formal definitions. Nice and
clear, and completely useless until I add some axioms that characterize
their intended meanings, e.g.,
0 is a Number.
No two distinct Numbers have the same successor.
0 is no Number's successor.
And so on.
> I would still like to simplify the language we are using, so we can more
simply and clearly articulate the project.
>
> . I suggest such a language needs to include:
> * a simplified vocabulary - in terms of both lexicon and semantics,
basically with clear and precise definitions and without metaphors or other
devices that add to ambiguity; and
> * a logical grammar or syntax.
That ontologies would be expressed in a logical language (and perhaps
informally in less rigorous languages as well) I believe has been a
presupposition all along. There seemed to be some fairly broad agreement to
use the simplified verions of KIF that can be found by following the link on
the SUO homepage.
Regards,
Chris Menzel
--
Christopher Menzel # web: philebus.tamu.edu/~cmenzel
Philosophy, Texas A&M University # net: chris.menzel@tamu.edu
College Station, TX 77843-4237 # vox: (979) 845-8764