Gregory's well-known account of the visual illusions depends on the observer's misconstruing certain lines as cues for depth. Proving the existence of these 'inappropriate' depth cues has been difficult. It is shown here that Gregory's major example of a 'primary' depth cue --converging lines --can be tested independently of the illusions by using a triangle configuration but that the outcome does not support the theory of primary constancy scaling.
Read full abstract