The translation of the OWL ReflexiveObjectProperty(R) axiom to the FOL ![X0]:'R'(X0,X0) axiom contradicts the following object property and background axioms
% background axiom: object domain and data domain are disjoint
![X]:('thing'(X)=>~'literal'(X)))
% background axiom: there are literals
?[X]:'literal'(X)
% Declaration(ObjectProperty(<http://example.com/R>))
![X,Y]:('R'(X,Y)=>'thing'(X))
Consider this simple ontology in Manchester Syntax:
Prefix: : <http://example.com/>
Ontology:
ObjectProperty: R
Characteristics: Reflexive
The translated FOL theory is inconsistent. EProver outputs
This is indeed a problem in the specification:
For reflexive object properties, it states that $\forall x: R(x, x)$.
Instead, it should say: $\forall x: (owl:Thing(x) \rightarrow R(x,x))$.
The translation of the OWL
ReflexiveObjectProperty(R)
axiom to the FOL![X0]:'R'(X0,X0)
axiom contradicts the following object property and background axiomsConsider this simple ontology in Manchester Syntax:
The translated FOL theory is inconsistent. EProver outputs