Abstract
Subjects viewed displays containing many line segments that were identical in length and orientation except for one line, and their task was to indicate where they saw a difference by placing a mark on the display. Ss' localization responses tended to center on the one disparate line when the lines were scattered randomly across the display. When the lines were spaced so as to group into pairs by proximity, however, their responses tended toward the geometric center of the whole pair rather than to the midpoint of the one oddly oriented line (the singleton). This result is in keeping with the Theory of Basic Gestalts, which holds that singleton popout occurs not because the target is unique but because it breaks a symmetry or other emergent feature in the display, in this case parallelism. The result complements a configural superiority effect (CSE) found with oriented line segments, wherein the parallelism existing between two lines can be detected faster and more accurately than can the orientation of either line alone. Combining that CSE with the present result, we suggest that parallelism is more likely than orientation to be a basic feature in vision, and that orientation singletons pop out not because they are unique but because they break that emergent feature of parallelism. Pop out does occur in all these displays, but the thing that pops out in displays where items do not group into pairs is the literal singleton of one differently oriented line, whereas what pops out in displays that group into pairs is the whole pair that is non-parallel, in contrast to the other pairs, all of which are parallel.
Meeting abstract presented at VSS 2014