Abstract
Four visual search experiments tested whether there is amodal completion in early vision. In our experiments, the search rates slowed down both for displays where amodal completion would and, critically, for displays where amodal completion would not interfere with the distinction between target and distractor. Any effect on search rates previously attributed to early amodal completion (Rensink and Enns, 1998) might therefore be better described by a more parsimonious account of early vision. In this account, early vision is not capable of assigning relative depth to objects that are seemingly occupying the same position in space (as indicated by T-junctions), and focal attention is needed to determine what is in front and what is behind. It might be that once attention is more focused, amodal completion gets the opportunity to exert its influence and reduce the distinction between target and distractor, slowing search down further. Note however that under this account it is the presence of T-junctions that is responsible for the initial need for more focused attention, and that the influence of amodal completion is contingent on this.