- From: Jeremy Carroll <jjc@hplb.hpl.hp.com>
- Date: Fri, 25 Oct 2002 14:13:01 +0200
- To: "Brian McBride" <bwm@hplb.hpl.hp.com>, "Jos De_Roo" <jos.deroo.jd@belgium.agfa.com>
- Cc: "Jeremy Carroll" <jjc@hplb.hpl.hp.com>, "pat hayes" <phayes@ai.uwf.edu>, <w3c-rdfcore-wg@w3.org>, <w3c-rdfcore-wg-request@w3.org>
> >1:[[ > > > >eg:prop rdfs:range eg:A . > >eg:A rdfs:subClassOf eg:B . > > > >entails > > > >eg:prop rdfs:range eg:B . > >]] > Brian: > Wierd! I think I had better explain it in English, by means of the example Jos gave of rdfs:Resource. The informal meaning of eg:prop rdfs:range eg:A . is that "any object of an eg:prop triple is of type eg:A". We happen to know that "any object of an eg:prop triple is of type rdfs:Resource". (ignoring Literals for now) that is we know a priori that eg:prop rdfs:range rdfs:Resource . ----- To get to the actual entailment, we are given: > >eg:prop rdfs:range eg:A . > >eg:A rdfs:subClassOf eg:B . i.e. "any object of an eg:prop triple is of type eg:A" "any object of type eg:A is of tpe eg:B" therefore "any object of an eg:prop triple is of type eg:B" that is eg:prop rdfs:range eg:B . --- or: eg:bananaTrifleContains rdfs:range eg:banana . eg:banana rdfs:subClassOf eg:fruit . entails eg:bananaTrifleContains rdfs:range eg:fruit . the conclusion is always true if the premises are true, unless the conclusion is seen as having intensional force. ==== This is an example of a wider issue that has come up in webont, which is: "Are properties of properties understood extensionally or intensionally?" Decoding that, in this example, is being a range of a property something which a class has because certain class membership relationships hold (extensional), or is something which a class has because somebody said so (intensional). If you take the extensional view, then you should support the "Jeremy entailment", if you take the intensional view then you should oppose it. As I said before the name reflects authorship not belief. This is a genuine question, and both points of view are legitimate. The relationship with WebOnt is that the majority on webont would like to treat all properties of properties in the same way (either all extensionally or all intensionally). Which is the preferred way is yet to be decided. If RDF Core either don't care or are split without strong opinions, then a way of deciding would be to follow the webont pref. If RDF Core care, then that is a significant input into the WebOnt decision on the other properties of properties. (eg. owl:TransitiveProperty). Jeremy
Received on Friday, 25 October 2002 08:13:12 UTC