Näive Realists have recently proposed that illusions occur in circumstances that are ‘non-paradigmatic’ or with which we are insufficiently familiar. While this proposal may work for many of the illusions philosophers normally discuss, I argue in this paper that there are other illusions that do not fit this pattern. In particular, the vertical-horizontal illusion (VHI) occurs in circumstances that are both familiar and paradigmatic, while disappearing (or becoming attenuated) in more unusual circumstances.

Develops an empirical argument against naive realism-disjunctivism: if naive realists accept "internal dependence", then they cannot explain the evolution of perceptual success. Also presents a puzzle about our knowledge of universals.