Coreference via substitution rules by Bernard Vatant develops two interesting notions:
- Using substitution to test interchange of references
- Using operational rules rather than declarative assertions
See his post for the full details.
He uses context to define when one reference could be substituted for another.
Also observes that any mapping, such as owl:sameAs can be abused.
As with many things, semantic integration may not be as much a technical issue but a human one. Semantic integration tools aren’t going to lead to semantic integration unless we use them with semantic integration as a goal.