Given an ontology O, and let A, B two classes and object properties P and Q such that:
- P domain A
- P range B
- A ⊑ =1 P.⊤
- Q ⊑ P
- A ⊑ =1 Q.⊤
Thus, we to prove P ≡ Q, we only need show that P ⊑ Q, because we already have the other direction, Q ⊑ P, in (4). To conclude that P ≡ Q are equivalent; let
- (x,y) ∈ P
then x ∈ A and y ∈ B and we have from (5) that Q relates each individual of A to exactly one individual of B; then there must exist y′ ∈ B such that (x,y′) ∈ Q; and by (4), we can infer that
- (x,y′) ∈ P
Then, from (3), (6), and (7), we can infer that y = y′. Thus, if (x,y) ∈ P, then (x,y) ∈ Q, which means that:
- P ⊑ Q
Then, by (4) and (8):
- P ≡ Q
Questions:
- Is this conclusion true?
- Reasoners (e.g., Pellet, via the Protégé plugin) are not inferring P ≡ Q, however each time I assert P(a,b), the reasoner infers Q(a,b), and vice versa!