Does a healthy respect for the deliverances of experience require us to give up any claim to non-trivial a priori knowledge? One way it might not would be if the very admission of something as an episode of experience required the use of substantive a priori knowledge — if there were certain a priori standards that a representation had to meet in order to count as an experience, rather than as a memory or daydream. What is surprising is that we can find elements of this essentially Kantian line about experience even in the work of empiricists such as Locke and Bas van Fraassen.