In this paper I discuss some of the implications of viewing questions rigorously in the light of Austin's conditions for performatives. I argue that this cannot be done without overt reference to the speaker's knowledge state. First I briefly review how questions as interrogatives have been dealt with by linguists since Bolinger 1957. Next I comment on Austin's theory of performatives in relation to questions and briefly compare his main points to those of Searle. I then discuss a number of sentences and contexts in support of my analysis of questions as epistemic requests. In a futher section I comment on the disagreement between Belnap and Åquist as it relates to this paper. The final part reviews the main points.