Re: Possible issue in the OWL 2 RL specification

Hi

Did you mean to write "but not ex:aaa ex:p1 ex:bbb", an A-Box inference?
Because "ex:p2 owl:inverseOf ex:p1" suggests you expected a T-Box
inference, but that's not what the rule entails.

Otherwise, if you do not indeed see the triple "ex:aaa ex:p1 ex:bbb", it
sounds more like an implementation problem, and you should get in touch
with the GraphDB support team.

Best,
Rashif

On Wed, May 20, 2026 at 3:31 AM Gilles Falquet <Gilles.Falquet@unige.ch>
wrote:

> Hello,
>
> We are sending this message to the RDF-DEV community group, as the
> public-owl-comments@w3.org address has been deactivated.
>
> Apparently the OWL 2 RL profile, as defined in
> https://www.w3.org/TR/owl2-profiles/ , does not consider that
> owl:inserseOf is symmetric. Applying the RL rules to
>
> ex:p1 a owl:objectProperty .
> ex:p2 a owl:objectProperty .
> ex:p1 owl:inverseOf ex:p2 .
> ex:aaa ex:p1 ex:bbb .
>
> yields
>
> ex:bbb ex:p2 ex:aaa .
>
> but not
>
> ex:p2 owl:inverseOf ex:p1 . (*)
>
> Is this intentional or is there a rule missing :
>
> T(?p, owl:inserseOf, ?q) => T(?q, owl:inserseOf, ?p) ?
>
> Best regards
> Hélène de Ribaupierre and Gilles Falquet
>
> (*) We first observed this behaviour when using GraphDB with a OWL 2 RL
> repository.
>