Abstract
Convexity and symmetry are well-established cues to figural status. These properties were originally thought to operate locally; for example, that the convex side of any curved border was likely to be perceived as figure. Peterson and Salvagio (2008) found that convex regions were increasingly perceived as figure as the number of alternating convex and concave regions increased. These "context effects" revealed display-wide influences on figure assignment, and required homogeneously colored concave regions. Goldreich & Peterson (2012) proposed that multiple same-color concave regions could be unified into a background, thereby allowing the convex regions to be perceived as figures in front of the background. The original context effects were obtained with displays having rectangular frame-like borders. There was no reason to think frame type mattered until Mojica and Peterson (2013) tested for symmetry context effects. Originally, they used displays with articulated borders to preserve the symmetry/asymmetry of the outermost regions. Symmetry context effects were not found, but later were found when frame-like borders were used. Mojica and Peterson reasoned that articulated outer borders were perceived as intrinsic to the display and consequently prevented the interpolation of concave regions into a background, whereas rectangular borders were perceived as extrinsic to the scene and allowed amodal continuation of the background. To test the generality of frame effects, we used 2- and 6-region convexity displays with articulated borders. Consistent with the view that articulated borders prevent the emergence of context effects, convex regions were no more likely to be seen as figure in 6-region (62%) than in 2-region displays (64%), p > .80, and were significantly less likely to be seen as figure in 6-region displays with articulated borders than rectangular borders (77%), p<.03. These results show that figure-ground cues to do not operate locally.
Meeting abstract presented at VSS 2014