Abstract
The occlusion illusion refers to Kanizsa's (1979) finding that the visible portion of a partly occluded object (e.g., a half circle whose other half is hidden behind a rectangle) is perceived to be significantly larger than the same region when it is not occluded (e.g., exactly the same half circle alone). The partial modal completion hypothesis implies that the occluded region appears larger because the visual system “fills in” a thin strip along the occluded border. The apparent distance hypothesis implies that the occluded region appears larger because it is the same retinal size as the unoccluded region, but is perceived to be farther away due to the occluding surface, thus leading to the perception of greater size by Emmert's Law, as in apparent distance theories of the moon illusion. We measured the magnitude of occlusion illusion psychophysically in several experiments to investigate these and other hypotheses about its cause. One experiment shows that the magnitude of the illusion varies with the strength of the evidence for occlusion. For example, the effect is larger when a rectangle partly occludes a circle than when it partly occludes a square, presumably because the half circle is perceived as more incomplete than a half-square (i.e., a rectangle). Other experiments support modal completion explanations over apparent distance explanations. A critical test between these hypotheses is whether the occlusion illusion produces its effect via a simple change in overall size of the occluded region (as predicted by apparent distance) or via a change in shape due to the occluded edge's perceptual extension behind the occluder (as predicted by modal completion). Observers judged the occluded region to be more similar to an extended-edge version of the actual stimulus than to an enlarged version, both of which were unoccluded and psychophysically matched to the occluded region.