W3C home > Mailing lists > Public > www-rdf-interest@w3.org > April 2001

Re: Can Resource be the top of our ontology ?

From: Seth Russell <seth@robustai.net>
Date: Thu, 12 Apr 2001 09:50:30 -0700
Message-ID: <018b01c0c370$b0c492c0$b17ba8c0@c1457248a.sttls1.wa.home.com>
To: "Aaron Swartz" <aswartz@swartzfam.com>, "RDF Interest" <www-rdf-interest@w3.org>
From: "Aaron Swartz" <aswartz@swartzfam.com>

> Seth Russell <seth@robustai.net> wrote:
>
> > But I would like to observe that Resources as defined above do not
function
> > nicely as the only valid top of our ontology;  whereas Thing(s) do. Here
are
>
> We don't need anything at the top of our ontology. As TimBL said: you
can't
> have something that's the root of all knowledge. Thing may be at the top
of
> your ontology, but Thwablog is at the top of mine. ;-)

Wooh!  It's a giant (erroneous) leap between talking of "top of an
ontology", and talking of "the root of all knowledge".   The former is a
formal (read mechanical) specification, and the latter is a natural language
term with all kinds of political associations.  The only consequences to
choosing a top to any particular ontology is that in that ontology would
have a single class that doesn't have a subClass arc to anywhere else.  That
is all, nothing else. Also, please note, that the composition of any
ontology *always* entails some subjective choices which can be changed over
time ... so there is nothing to prevent someone from coming along and
asserting in their local view of such an ontology something like {Thing
subClass SpiritObject} ... one merely needs a way to draw such a
distinction.  My point is that when you define a class is such a way that it
is already distinguished from a more general class, then that does not
qualify as a node which cannot have a subClass arc on it.

> > 1) The definition itself implies that there are things which can have no
> > identity by saying: "A resource can be anything that has identity".  So
what
> > happens when we must talk of things with no identity?  Are these things
to
> > have no ontological status?  Can I not describe a dust mite that was
present
> > in the room in which I was born; or would I have to name the bugger
first?
>
> Well, unless you can identify something, I don't think you'll be able to
> talk about it on the Semantic Web. Unless I'm missing it's hard to talk
> about something that's not identified...
>
> "Yeah, it's purple."
>     "What is?"
> "It!"

They would be anonymous.

      [;  hasColor Purple]

Seth
Received on Thursday, 12 April 2001 12:54:52 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:51:48 GMT