And the same in hypotheses and hypothetical arguments; for it is necessary sometimes to demand the granting of some hypothesis as a kind of passage to the argument which follows. Must we then allow every hypothesis that is proposed, or not allow every one? And if not every one, which should we allow? And if a man has allowed an hypothesis, must he in every case abide by allowing it? or must he sometimes withdraw from it, but admit the consequences and not admit contradictions? Yes; but suppose that a man says, "If you admit the hypothesis of a possibility, I will draw you to an impossibility."