On the Nature of a Physical Theory