Have, mostly by accident, been having an interesting discussion on the Jena-dev mailing list on the understanding of ‘implication’.
Given A -> (B -> C), and understanding that -> represent defeasible implication, I read this as “A suggests that B implies C”. Dave’s point, correctly I think, is that some of the time, you need either A to _definitely_ say B -> C, or for A to suggest that B is _definitely_ C (although I think this is harder, as then I would try and put this in the ontology). Looks like I (ideally) need two forms of inference (maybe version 2.0….)
It important because of course it shows that I can’t depend on everyone else having the same prejudices as me….