W3C home > Mailing lists > Public > www-webont-wg@w3.org > June 2003

Re: draft response to #owlref-rdfcore-owl-class-denotation

From: pat hayes <phayes@ihmc.us>
Date: Fri, 13 Jun 2003 18:42:00 -0500
Message-Id: <p0521061abb100c1cb133@[10.0.100.24]>
To: "Jonathan Borden" <jonathan@openhealth.org>
Cc: www-webont-wg@w3.org

Jonathan: sorry, I missed the point when responding to your message 
at the end. This is what I should have said::

>
>Understood, but what about 'translating' between OWL-DL and OWL-Full -- same
>problem??? e.g.
>
>_:x rdf:type _:x .
>_:x rdf:type owl:Class .
>
>logically true in OWL-Full, but contradiction in OWL-DL?

Unsayable in OWL-DL: syntactically illegal.  OWL-DL's attitude 
towards all the RDFS assertions which would violate its semantic 
assumptions is like a three-monkeys parable: it just refuses to allow 
itself to even speak of such things.  But in RDFS we have no such 
restrictions, and if we assume that owl:Class refers in RDFS to the 
same kind of class that it refers to in the OWL model theory (and 
about which we are, in RDFS, allowed to say more than OWL-DL syntax 
allows us to say) then yes, a contradiction; for those classes 
cannot, by DL semantic fiat, contain themselves.  So apparently the 
OWL-Full identification of owl:Class with rdfs:Class changes the 
meaning of owl:Class; this OWL URI means different things in OWL-DL 
and OWL-Full. As long as you stay inside the OWL-DL syntax, you can't 
tell the difference; but you can tell the difference in OWL-Full 
itself.  I would prefer to allow OWL-Full to use rdfs:Class distinct 
from owl:Class; this would not disturb the OWL-DL meaning of the OWL 
terms while allowing them to be used in RDFS more generally. With 
this understanding, the above would not be true even in RDFS, since 
there would be no reason to assume that the RDFS properties of 
rdfs:Class also applied to owl:Class; in fact, one could prove that 
they do not.

>(I am assuming that
>since owl:Class and rdfs:Class have the same extensions in OWL-Full and
>RDFS,

Well, they don't necessarily  have the same extensions in RDFS. As 
far as RDFS is concerned, owl:Class is just another RDFS class name. 
If we were to accept the OWL-DL theorems about owl:Class (and 
owl:Thing, etc) as true in RDFS, then the distinction between 
rdfs:Class and owl:Class would be vividly clear, as noted. The fact 
that it has been rendered invisible in OWL-Full means that the 
meaning must have changed somewhere along the line.

Pat


-- 
---------------------------------------------------------------------
IHMC	(850)434 8903 or (650)494 3973   home
40 South Alcaniz St.	(850)202 4416   office
Pensacola			(850)202 4440   fax
FL 32501			(850)291 0667    cell
phayes@ihmc.us       http://www.ihmc.us/users/phayes
Received on Friday, 13 June 2003 19:42:02 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:58:01 GMT