Lydia Maniatis (2009) adroitly argues our theories of a perspective illusion to do with angles are rigged in favour of our conclusions. Nice point. She develops useful alternative accounts of our illusion. However, her rigging argument involves the Bishop Berkeley induction problem. In response, we point out 2D features can specify 3D shapes. Counter to her induction problem, we argue specification requires a 3D environment of a particular, restricted kind, resulting in meaningful 2D features, as in the landmark Pizlo (2008) treatise.