Abstract
Convexity is a well established figural cue for figure-ground organization. However, it has an inherent implicit confound with closure based on the observation that if the shared-contour is extended on the convex side it will form a closed region. Thus, the relative influence of convexity and implied-closure is unknown. Here we describe our attempt to minimize this confound and empirically investigate whether convexity or implied-closure has a greater impact on determining figure-ground organization. The critical segment, “wings”, of the contour is the part located near the top and bottom of the shared-contour that slightly curves around to intersect the perpendicular-borders of the bipartite image. We created conflict “incongruent” stimuli by flipping the wings about the vertical axis to imply closure on the side opposite to convexity. Control “congruent” conditions were created by moving the position of the flipped-wings away from the perpendicular-edge towards the center of the contour. This resulted in net convexity matched to the incongruent-condition while congruency between convexity and implied-closure was not affected. Stimuli consisted of 128 bipartite black-white images with: 2 shapes of shared-contour (circle/triangle) of 2 sizes each (small/large); 2 wing sizes depending on the base shared-contour size (range 7%–20% of entire length); position of wings along the shared-contour (1 incongruent “UpDown” and 3 congruent “Up”, “Center”, “Down”). They were black/white and left/right counterbalanced. The observers made a figure-ground judgment after viewing the stimulus for 1 second and responded with a mouse click on left/right response buttons. Data from 40 participants show that the percentage figural response for the side with both convexity and implied-closure is 70%. However, in case of conflict, figural response fell to 50% for convexity-only-side vs. implied-closure-only-side. We conclude that the convexity reported as the figural cue is effective only when present with implied-closure and not otherwise.