On the Nature of Induction