[Top] [All Lists]

Re: [ontolog-forum] What is the role of an upper level ontology?

To: ontolog-forum@xxxxxxxxxxxxxxxx
From: John F Sowa <sowa@xxxxxxxxxxx>
Date: Fri, 17 May 2013 00:14:01 -0400
Message-id: <5195AE89.1090805@xxxxxxxxxxx>
Gary and Tara,    (01)

Some comments:    (02)

> In Ontoclean Guarino & Welty makes the point that there are what
> we might  consider upper or foundational semantic theories that
> are tools that can be leveraged.    (03)

That is a good summary of the arguments for an upper ontology.
Unfortunately, there are few, if any clearly defined, widely accepted
semantic theories that (a) all scientists and philosophers agree on,
(b) are consistent with the way terms are used in natural languages,
and (c) are universally accepted as suitable foundations for computer
science and applications of computer systems.    (04)

> Examples are Semantic Theories of:
>   parts & wholes,
>   essence & identity,
>   composition and constituency.    (05)

I agree that many people who propose an upper ontology like to include
such things.  Unfortunately, those are among the most complex issues
that have been debated in philosophy for millennia -- with no consensus.    (06)

For parts & wholes, the following is a classic:    (07)

    Simons, Peter (1987) _Parts: A Study in Ontology_,
    Oxford: Clarendon Press.    (08)

But look at the following list of 16 more recent books on mereology:    (09)

    http://philpapers.org/rec/BURHOM-2    (010)

For briefer analyses, see the following by Ingvar Johansson:    (011)

    Natural science and mereology    (012)

    Mereology and natural science    (013)

    Continua in biological systems    (014)

For essence, the situation is even more hopeless.  Plato and Aristotle
couldn't agree on how to define essence, how to recognize it, or how
to reason about it.  Today, all the debates between P & A are just as
hot as ever.  There's much more detail, much more debate, and even
less agreement.  There are also skeptics from Sextus Empiricus to
Quine who debunk the very idea.    (015)

Identity is another swamp.  The = sign in logic and mathematics looks
very clear and simple, but outside of mathematics identity is *never*
fundamental.  You can observe similarity, but identity is *always*
a context-dependent inference for a particular purpose.    (016)

In logic, it looks so simple to say, For all x, x=x.    (017)

But Whitehead, asked "Is the child in the cradle identical to the adult
30 years later?"  For some legal purposes, yes.  For most practical
purposes, no.  And how can you recognize identity?  By a continuous
trajectory in space-time?  Perhaps in theory, but certainly not
in practice.    (018)

And that is just for the human beings.  When you get to trees, worms,
rocks, and electrons, all bets are off.  Some philosophers deny that
the notion of identity can be applied to artifacts -- you can remove
and replace parts indefinitely.  But now, we're getting closer to
doing that for humans.    (019)

As for composition and constituency, all the problems of parts,
essences, and identity are multiplied by the huge numbers of ways
of constructing, growing, and processing anything.    (020)

Fundamental principle:  All the options are important and can be useful
for various applications.  But they belong in an open-ended family of
*microtheories*.  None of them belong in an upper level ontology.    (021)

> it is always possible to simply add a condition to the quantification
> in order to restrict it to a certain subset of the domain    (022)

I agree.  The point I was trying to illustrate was Johansson's remark
that the two categories of particulars and universals correspond to
the distinction in predicate calculus between instances and relations.    (023)

But you can, if you wish, partition the relations in many disjoint
classes to represent as many categories as you need in any ontology.    (024)

> I would say that CL has an implicit five-category ontology. There are
> the three categories of entity, function and relation.    (025)

Actually, that's only one category.  Every entity in CL has three
personalities:  it can be used as an individual, as a relation,
and as a function.  Many applications don't use all those options,
but it always permissible to use the same entity in all three ways.    (026)

> The functions and relations can also be partitioned into discourse and
> non-discourse.    (027)

That's not really a difference in ontology, but a difference in the
way the entity can be used.  In a segregated dialect, some names are
restricted in the way they can be used.  But the entities that those
names refer to have all three personalities -- individual, relation,
and function.  An unsegregated dialect can refer to the same entity
and use it in ways that are not permissible in an segregated dialect.    (028)

John    (029)

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    (030)

<Prev in Thread] Current Thread [Next in Thread>