John, (01)
I will first confess that I don't understand much of what you say below.
But I would try to point out the difference between "exceptions" and
"conflicting microtheories". (02)
Mathematically, an "exception" to a universal statement is a
counterexample -- a demonstration that the universal statement is false.
For ontologies that means: if you are aware of exceptions, your ontology
doesn't contain the universal statement, it contains a guarded one. The
guard has the form of an implication: For all x, if not
exceptional-situation (x), statement (x). My favorite example from the
business rules domain is the sign that reads: "No admittance except for
authorized persons." (This carefully leaves open the question of how
membership in the class "authorized person" is determined.) (03)
Microtheories, on the other hand, represent different perspectives on
reality, as held by different stakeholders with different concerns.
They model the world as they understand it, or as it is useful to the
purpose of their model. The understanding of others, or the purposes of
other models may lead to conflicting ontologies. The term "microtheory"
arises from communities of practice in which there is agreement on the
broad "universals", but disagreement on details that are relevant to
different sub-communities. As a young colleague of mine, Antoine
Lonjon, observed a few years ago:
"A fact is an actuality interpreted by a model."
Insert "mental" before "model", and you come to the fundamental
difficulty of knowledge engineering for broad usage. (04)
I don't know that I am disagreeing with John's position. I just want to
be careful to distinguish the ideas. (05)
-Ed (06)
'Speak when you're spoken to!' The Queen sharply interrupted her.
'But if everybody obeyed that rule,' said Alice, who was always ready
for a little argument, '...nobody would ever say anything.'
C.L. Dodgson (Lewis Carroll), "Through the Looking Glass" (07)
John Bottoms wrote:
> "FORM AND FUNCTION"
>
> "Form and Function" are the top level elements (Universals) for fashion
> and industrial designers. They spend endless hours in discourse on this
> topic. Yet it is solely the first level of their ontology. Thet have not
> progressed much beyond that point since the Bauhaus movement in 1919 and
> the indications are that this will continue.
>
> I had thought that the way to start with ontology design would be to fix
> the universals, but even that is not possible as Lakoff ("Fire and
> Dangerous...") points out. If we must allow for alternate universals
> then it appears that we are resigned to use one of the proposed
> incomplete lattices, or to a set of universals that allow exceptions. An
> alternate approach would be to provide associative links from a set of
> references for universals that could be selected based on a context.
>
> Exceptions (or other predicates) in universals may not be such a bad
> thing. Intelligence has always used exceptions to provide coverage for
> widely varying, real world situations, and exceptions are an under
> represented area of study and representation. I know the logicians
> wouldn't like that approach, but I believe it is possible using FOL.
>
> -John Bottoms
> FirstStar Systems
> Concord, MA USA
>
> _________________________________________________________________
> Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
> Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
> Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
> Shared Files: http://ontolog.cim3.net/file/
> Community Wiki: http://ontolog.cim3.net/wiki/
> To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
> To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx
>
> (08)
--
Edward J. Barkmeyer Email: edbark@xxxxxxxx
National Institute of Standards & Technology
Manufacturing Systems Integration Division
100 Bureau Drive, Stop 8263 Tel: +1 301-975-3528
Gaithersburg, MD 20899-8263 FAX: +1 301-975-4694 (09)
"The opinions expressed above do not reflect consensus of NIST,
and have not been reviewed by any Government authority." (010)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx (011)
|