W3C home > Mailing lists > Public > www-rdf-interest@w3.org > July 2003

RE: Do resources have representations?

From: <Patrick.Stickler@nokia.com>
Date: Mon, 28 Jul 2003 18:54:21 +0300
Message-ID: <A03E60B17132A84F9B4BB5EEDE57957B5FBC0D@trebe006.europe.nokia.com>
To: <b.fallenstein@gmx.de>
Cc: <jon@spin.ie>, <www-rdf-interest@w3.org>



> -----Original Message-----
> From: ext Benja Fallenstein [mailto:b.fallenstein@gmx.de]
> Sent: 28 July, 2003 18:25
> To: Stickler Patrick (NMP/Tampere)
> Cc: jon@spin.ie; www-rdf-interest@w3.org
> Subject: Re: Do resources have representations?
> 
> 
> Patrick.Stickler@nokia.com wrote:
> > I've found it useful to think of URIs as being bound to
> > particular intensions of resources.
> > 
> > E.g. several URIs may all denote the same thing, i.e.
> > 
> >    http://example.com/MorningStar 
> >       owl:sameIndividualAs
> >          http://example.com/EveningStar ;
> >       owl:sameIndividualAs
> >          http://example.com/Venus .
> 
> ...but do not *connote* the same thing, would be my term for 
> it. E.g., 
> the representations received when we enter each of the above in a 
> browser may be different.
> 
> > but each URI, and the statements made about the resource
> > using that URI, are specific to a particular intension of 
> > the thing denoted.
> 
> Each URI is specific to a particular intension: ok.
> Each statement using the URI: Hm.
>
> OWL specifies that if A owl:sameIndividualAs B, then all 
> statements you 
> can make about A entail a similar statement about B. E.g., if I say
> 
>      <http://example.com/EveningStar>  rdf:type  foo:BeautifulThing
> 
> then I'm implying that
> 
>      <http://example.com/Venus>   rdf:type   foo:BeautifulThing
> 

Not that *you* are implying that. But that it is nevertheless
true. If the thing denoted by http://example.com/EveningStar
is in fact beautiful and is the same thing denoted by
http://example.com/Venus, then that too is also beautiful.
How could it not be. It is the same thing. 

I agree that this is pretty tricky, and not, I think, fully
captured by RDF or OWL machinery as presently defined.

There is a certain lack of resolution with regards to URI
denotation. To what degree, e.g. to Dr. Jeckell and Mr. Hyde
both denote the same individual. One may in fact be beautiful
yet the other not? How does one use RDF or OWL to capture
intension vs. denotation vs. partial commonality? Tricky.

The names (URIs) we use certainly do carry along with them the 
intensions of the things denoted which go beyond their denotation.
And owl:sameIndividualAs has a pretty corse, absolute/complete
view of equivalence of denotation that would probably not be
appropriate for equating Dr. Jeckell and Mr. Hyde.

This indicates that even though URIs which are in fact
related by an owl:sameIndividualAs assertion are not necessarily
arbitrarily interchangable.

E.g., given 

   :Superman owl:sameIndividualAs :ClarkKent .
   :Superman :can :fly .

it is fair to infer that

   :ClarkKent :can :fly .

and in fact, he can, though he doesn't usually do so
when acting in the persona of Clark Kent.

However, once you start talking about belief, perspective,
scope of knowledge, intension, then you can run into
challenges. E.g. while the above is quite OK, you
wouldn't want your system inferring that, if

   :Superman owl:sameIndividualAs :ClarkKent .
   :Superman :can :fly .
   :LoisLane :believes :x .
   :x a rdf:Statement;
      rdf:subject   :Superman ;
      rdf:predicate :can ;
      rdf:object    :fly .

then

   :ClarkKent :can :fly .
   :LoisLane :believes :y .
   :y a rdf:Statement;
      rdf:subject   :ClarkKent ;
      rdf:predicate :can ;
      rdf:object    :fly .

If the URIs were freely interchangable based on equivalent
denotation, we'd muck up all kinds of stuff relating to
intension, knowledge, perspective, etc.

I think this remains a problem for RDF and OWL and certainly
an area needing further research. I'm sure those having more
experience in logic and intensional semantics could make this
alot clearer (and even correct ;-)

Its stuff like this that motivated the under (or un)defined
semantics for RDF reification in the latest specs. Lots of
rat holes there, it seems...

Patrick

--
Patrick Stickler
Nokia, Finland
patrick.stickler@nokia.com
 
Received on Monday, 28 July 2003 11:54:24 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:52:00 GMT