I think Emily is correct in her assessment. I think unless someone comes up with a really awesome parser that understands english in all its context sensitive glory, this issue will just be a typical consequence of using a natural language parser. At that moment though, I don't see anyone coming up...