[Rumsfield's] reply included a complex formulation that would become inextricably associated with him: 'There are known knowns. These are things we know that we know. There are known unknowns. That is...