[Top] [All Lists]

Re: [ontac-forum] What should be in an upper-level ontology

To: mitioke@xxxxxxxxxxxx, ONTAC-WG General Discussion <ontac-forum@xxxxxxxxxxxxxx>
From: "John F. Sowa" <sowa@xxxxxxxxxxx>
Date: Mon, 22 May 2006 18:42:22 -0400
Message-id: <44723E4E.1040006@xxxxxxxxxxx>
Ken and Matthew,    (01)

This discussion seems to be converging.    (02)

KE> Just the knowledge of the upper level made things in the
 > lower and middle layers fit -- that, in my mind, may not have
 > fit before; I learned.  I did not alter my way of thinking
 > in that I adapted to new facts.    (03)

That is an important point:  It's necessary to have guidance
on how to organize the categories of an ontology and how to
associate axioms with those categories.  But that kind of
guidance could be obtained from a textbook, a set of design
tools, or a collection of examples.    (04)

I think of the collection of theories as a library or
registry of formal specifications that people have found
useful for various projects.  The registry would include
a list of contributors and users who helped develop those
theories and the kinds of applications for which they were
used.  All of that would be very useful to anyone who needs
some guidelines on how to proceed.    (05)

MW> Well this at least I can broadly agree with.
 > The main thing I would note is that you are starting with an
 > assumption that an Upper Ontology is/should be. You are assuming
 > it should be that which should common regardless of any ontological
 > commitments one might choose to make. This of course fits with
 > your ideas for a "Lattice of Theories" (which again I support).    (06)

At this stage, broad agreement is sufficient for making progress.    (07)

MW> Let us take person (or more generally organism). You need to be
 > careful not to say it is a subclass of physical object or process,
 > both, or even that there might be two organisms (organism as physical
 > object, and organism as process). Equally, we can all agree (I hope)
 > that there are people (whatever they are). I don't know how you can
 > leave yourself that uncommitted (can you explain how to do that?)    (08)

I agree that any useful ontology for business applications should be
able to talk about people as well as plants, animals, and other living
organisms.  But it is possible to have a biological taxonomy whose
top node is directly linked to Entity at the very top.  Any microtheory
that treats living organisms as physical objects could add an additional
axiom that says LivingOrganism < PhysicalObject.  And a microtheory for
a Whiteheadian ontology could add the axiom PhysicalObject < Process.
But those assumptions would not be in the general-purpose hierarchy.    (09)

Many different taxonomies could be introduced as needed by linking
them directly to Entity at the top.  I am assuming that no axioms
are asserted for the category Entity other than the obvious:    (010)

    If x exists, then Entity(x).    (011)

I wouldn't even assume the converse, because there might be things
in an ontology such as airplanes that are still on the drawing board.
It would be safe to say that they would be entities, if they existed,
but there could be a slot for them even before they exist.    (012)

MW> I think there is another interpretation of Upper Ontology
 > which says that it is enough to specify the key ontological
 > commitments for an ontology. So precisely whether it is 3D,
 > 3+1D or 4D. Whether/when the vase and clay are the same object,
 > etc.
 > I guess in your terms these would be top level micro-theories,
 > and if we can agree on a Lattice-of-Theories approach, then this
 > would fit well. I would still call these Upper Ontologies however
 > (as opposed to THE Common Upper Ontology).    (013)

Yes, I would be quite happy with that.  That means we could have
one giant hierarchy that was very underspecified.  It would be
used primarily as an index for the common set of labels for the
relations that are used in all the theories of the lattice.    (014)

Since it would be so highly underspecified, it would be on the
borderline between an ontology and a taxonomy.  In fact, it would
even be less detailed than many taxonomies, since it wouldn't make
a commitment to calling a living organism a process or a physical
object.  Any microtheory that required further commitments could
make such assertions as needed.    (015)

MW> I think that it would be very interesting to pursue what could
 > be a Common Upper Ontology. I agree with you that it would be
 > surprising how much had to be left out, even at the subtype/supertype
 > level.    (016)

I agree.  And note, by the way, that there is no limit on the size
of any theory in the lattice.  Anyone who needed a large, detailed
hierarchy could take the general underspecified hierarchy and add
many more subtype/supertype links and axioms.  There could be
several such theories, which could become de facto standards for
various communities of users.    (017)

But anyone who needed a different set of assumptions could start
with the underspecified hierarchy and specialize it in a different
way for a different community.    (018)

John    (019)

Message Archives: http://colab.cim3.net/forum/ontac-forum/
To Post: mailto:ontac-forum@xxxxxxxxxxxxxx
Shared Files: http://colab.cim3.net/file/work/SICoP/ontac/
Community Wiki: 
http://colab.cim3.net/cgi-bin/wiki.pl?SICoP/OntologyTaxonomyCoordinatingWG    (020)
<Prev in Thread] Current Thread [Next in Thread>