SCIENCE AND SEMANTICS: THE CASE OF VAGUENESS AND SUPERVALUATION*