AbstractThe paper questions the common assumption that rational individuals believe all propositions which they know to be logical consequences of their other beliefs: although we must acknowledge the truth of a proposition which is a deductive consequence of our beliefs, we may not genuinelybelieveit. This conclusion is defended by arguing that some familiar counterexamples to the claim that knowledge is justified true belief fail because they involve propositions which are not really believed. Beliefs guide conduct or issue in assertion by answering questions which arise in the course of deliberation and conversation, but the troublesome cases present propositions which do not present the agent's answer to any question. The paper concludes by sketching the conditions under which the deductive consequences of our beliefs can be believed.1