Abstract
When an object extends across two different illuminations, observers can make two separate lightness matches for the two patches of the object that lay in the two illumination levels and these patch matches are well predicted by Anchoring theory. However, observers can also match the lightness of the whole object and these object matches cannot be predicted by anchoring theory in its present form (Gilchrist et. al., 1999). Previously we have reported that the object match coincides with the match given for the patch in the prevailing illumination, or the illumination with the largest area. But this prediction fails in certain conditions. To find a more effective definition of prevailing illumination, we conducted five experiments. A display consisting of five adjacent rectangles, spanning from black to white was mounted either in midair or on a wall and either a spotlight or a shadow was projected across the five rectangles. Size and position of the special field of illumination was also varied. Two modifications of the definition of prevailing illumination can account for all the lightness matches obtained in all of our conditions. First, the level of prevailing illumination (functionally) is influenced by the highest illumination and second, and it is influenced not only by the largest area in the global framework, but also by the largest area within the test display itself. It is interesting to note that the same two factors, highest and largest, parallel the two factors found to anchor lightness. That is appropriate because the question of distinguishing prevailing illumination from special illumination is just the question of anchoring within the illumination domain.