Abstract
Introduction: A recent study (Rauschenberger, Peterson, Mosca, & Bruno, 2004 - Psychological Science) argued for an “ambiguity” account of amodal completion. This account holds that once an incomplete 2-D figure has been amodally completed, either the complete or the mosaic representation is accessible. The representation accessed may depend on external factors like set or context. In a visual search experiment, they used two types of target - a complete disk next to a complete square (“separate”) and a notched disk abutting a complete square (“adjacent”). Each non-target was a notched disk and complete square pair, where the notch was aligned with, but a short distance from, the corner of the square. A subject was instructed (“set”) to respond whether a complete disk appeared. Two display durations were used: 100 and 250 ms. They found search to be comparably efficient between the separate and adjacent target conditions at 100 ms, but not at 250 ms, where search in the adjacent condition was inefficient. They concluded that “set” influenced perception toward the complete representation for adjacent targets at 100 ms, but that “context” of the non-distractors influenced perception toward the mosaic representation at 250 ms. However, due to the alignment of the notch with the corner of a square for a non-target, a partial illusory square boundary may appear to occlude the notched disk. Thus, notched disks in the non-targets may (weakly) appear complete at 250 ms, rendering them similar to an adjacent target. This may explain inefficient search for adjacent targets at 250 ms.
Method: We replicated their experiment, but gave explicit boundaries to all disks and squares, in order to inhibit formation of illusory square boundaries for non-targets.
Results: Search was comparably efficient between separate and adjacent targets at both 100 and 250 ms.
Conclusion: The “ambiguity” account of amodal completion and the new effect of “context” are not supported by these results.
We are grateful to Robert Rauschenberger for use of his experimental software.