Is it rash to assume that when a practised writer says a thing, he is more likely to mean what he says than what his commentators think he means?