RE: Reflexive properties in OWL

Bikash Gyawali wrote:

>I see your point, Marco. So, there's no way to model properties that are
>both locally reflexive and transitive in OWL? 

Just to clarify: it won't be a syntactically valid OWL 2 /DL/ ontology if
you combine these features (I wasn't aware that you want to do this when
writing my last answer). 

But it is perfectly possible to write this in terms of the OWL 2 Structural
Specification [1], which is more general than OWL 2 DL (see Sec. 3 of the
spec for the restrictions defining the OWL 2 DL syntactic fragment). Also,
the OWL 2 Direct Semantics [2], which is the semantics of OWL 2 DL but is
actually specified for the whole of the OWL 2 Structural Specification, will
give you a well-defined meaning for such a non-DL'ish combination of
features. Further, a reasoner will even be OWL 2 DL(!) conformant when
giving you the results as defined by the OWL 2 Direct Semantics to the
/whole/ of the OWL 2 Structural Specification, since this will automatically
meet the conformance criterion for "OWL 2 DL entailment checkers" given in
Sec. 2.2.1 of [3]: an OWL DL reasoner just must not give wrong results on
OWL DL ontologies. 

Apart from this, OWL 2 /Full/ [4] allows this as well, and also gives a
precise (though slightly different) meaning to such an ontology.

So saying that it is "not possible" or "not allowed" in "OWL" is not quite
valid. The problem is rather that all OWL 2 DL reasoners that I know will
either deny to reason on such a non-OWL 2 DL ontology, or will ignore
certain parts of it while doing reasoning. Of course, this is a real
/practical/ problem, which makes my discussion above pretty moot... :-(

Michael

[1] <http://www.w3.org/TR/2009/REC-owl2-syntax-20091027/>
[2] <http://www.w3.org/TR/2009/REC-owl2-direct-semantics-20091027/>
[3] <http://www.w3.org/TR/2009/REC-owl2-conformance-20091027/>
[4] <http://www.w3.org/TR/2009/REC-owl2-rdf-based-semantics-20091027/>

In real life, there are
>properties like that, for eg: the hasReferenceTo property or 'Knows'
>property.
>
>Bikash Gyawali
>
>-----Original Message-----
>From: Marco Colombetti
>Sent: Monday, December 27, 2010 7:21 PM
>To: Bikash Gyawali ; Michael Schneider
>Cc: Public Owl Mailing List
>Subject: Re: Reflexive properties in OWL
>
>Bikash, if I understood well another message of yours, you
>want a property to be both transitive and locally
>reflexive. This is not possible, because transitive
>properties are composite, and you cannot apply a
>self-restriction to a composite property.
>
>Marco
>
>On Mon, 27 Dec 2010 18:26:05 +0100
>  "Bikash Gyawali" <bikashg@live.com> wrote:
>> Hi Michael,
>>
>> So good to get an exact answer for what I looked for. I didn't know
>that
>> "new" feature before.
>>
>> Thanks.
>>
>> -----Original Message----- From: Michael Schneider
>> Sent: Monday, December 27, 2010 6:12 PM
>> To: Bikash Gyawali
>> Cc: Public Owl Mailing List
>> Subject: RE: Reflexive properties in OWL
>>
>> Hi Bikash,
>>
>> to define a property :p to be "locally-reflexive" w.r.t. class :C,
>make :C
>> a
>> subclass of the self-restriction on :p (new feature of OWL 2):
>>
>>    :C rdfs:subClassOf [
>>        rdf:type owl:Restriction ;
>>        owl:onProperty :p ;
>>        owl:hasSelf "true"^^xsd:boolean ] .
>>
>> This can be read as follows: For every instance x of class :C the
>> reflexive
>> property assertion x :p x holds.
>>
>> If you want to additionally ensure that the only reflexive triples for
>:p
>> are those with values from :C, then replace "rdfs:subClassOf" by
>> "owl:equivalentClass" in the above axiom.
>>
>> Michael
>>
>>From: public-owl-dev-request@w3.org [mailto:public-owl-dev-
>request@w3.org]
>> On Behalf Of Bikash Gyawali
>> Sent: Monday, December 27, 2010 1:52 PM
>> To: Public Owl Mailing List
>> Subject: Fw: Reflexive properties in OWL
>>
>> Hello All,
>>
>> The specification of domain and range for any property doesn't act as
>> constraint.
>> It just acts as an axiom in OWL.
>> So, I am just creating properties (without specifying their domain and
>> roles). Later,I create restrictions on the property belonging to that
>> respective class.
>>
>> However, I see a problem in reflexive properties.
>> The reflexive property is automatically being assigned to all classes
>> in the ontology. I don't want that to happen. When I create a
>reflexive
>> property,
>> I intend to act it so only for my particular class of choice.
>>
>> With other types of properties, this is not a problem.
>> The reflexive property is "smarter" because it always knows its domain
>> and role beforehand. So, its getting attached to all classes in the
>> ontology.
>>
>>
>> How can I restrict the reflexive property to a particular class?

--
Dipl.-Inform. Michael Schneider
Research Scientist, Information Process Engineering (IPE)
Tel  : +49-721-9654-726
Fax  : +49-721-9654-727
Email: michael.schneider@fzi.de
WWW  : http://www.fzi.de/michael.schneider
=======================================================================
FZI Forschungszentrum Informatik an der Universität Karlsruhe
Haid-und-Neu-Str. 10-14, D-76131 Karlsruhe
Tel.: +49-721-9654-0, Fax: +49-721-9654-959
Stiftung des bürgerlichen Rechts, Az 14-0563.1, RP Karlsruhe
Vorstand: Prof. Dr.-Ing. Rüdiger Dillmann, Dipl. Wi.-Ing. Michael Flor,
Prof. Dr. Dr. h.c. Wolffried Stucky, Prof. Dr. Rudi Studer
Vorsitzender des Kuratoriums: Ministerialdirigent Günther Leßnerkraus
=======================================================================

Received on Tuesday, 28 December 2010 14:34:56 UTC