On what is bound in figures and grounds

All else being equal, regions providing a quick, good match to object memories are likely to be seen as shaped figures rather than as shapeless grounds. Good matches to object memories occur only if the parts are properly bound spatially. If object memories are accessed in the course of perceptual organization even for regions ultimately seen as shapeless grounds, then binding of parts cannot require attention to figures. Novel silhouette primes were shown immediately before real or novel line drawings. The silhouette contours of half of the primes preceding real line drawings sketched a known shape on the outside, which nevertheless appeared to be a shapeless ground. At short prime-line drawing SOAs, observers took longer to categorize real line drawings following these experimental primes than control primes, suggesting that parts are properly bound spatially for regions perceived to be shapeless grounds, and further that object memories matching grounds are inhibited.