Figure – Ground Effects on Shape Memory for Objects versus Holes

The circumstances under which the shapes of figure-versus-ground regions are perceived and remembered were investigated in three experiments that replicate, extend, and clarify Rubin's [1921 Visuell wahrgenommene Figuren (Copenhagen: Gyldendals)] classic study on this topic. In experiment 1, observers reported which of two regions they perceived as figure within ambiguous, bipartite, 2-D displays. In a later shape-recognition test, the shapes of regions previously seen as figures were remembered well, but the shapes of regions previously seen as grounds were remembered no better than novel distractor regions. In experiment 2 we examined the same question about memory for the shape of figure-versus-ground regions in nested displays in which the central region could be perceived either as a closer figure surrounded by a farther ground (ie as a solid object) or as a farther ground surrounded by a closer figure (ie as an empty hole). Unlike experiment 1, the shapes of regions initially perceived as grounds (holes) were remembered as well as those of regions initially perceived as figures (solid objects), and much better than those of novel distractor regions. In experiment 3 we further demonstrated that this outcome did not depend on the figure – ground instructions employed in experiment 2, because the same result was obtained with unambiguous 3-D cardboard displays of objects versus holes with no figure – ground instructions at all. The present findings support an account of hole perception in which the shape of an intrinsic hole is encoded as a shaped, immaterial (or virtual) surface where the absence of matter is coded by a functional ‘missing’ symbol (analogous to a minus sign in mathematics) to represent its non-material status.