An Aristotelian work often goes on continuously at first, and then becomes disappointing by suddenly introducing discussions which break the connexion or are even inconsistent with the beginning; as in the Posterior Analytics, which, after developing a theory of demonstration from necessary principles, suddenly makes the admission, which is also the main theory of science in the Metaphysics, that demonstration is about either the necessary or the contingent, from principles either necessary or contingent, only not accidental.