Can face recognition really be dissociated from object recognition?

We argue that the current literature on prosopagnosia fails to demonstrate unequivocal evidence for a disproportionate impairment for faces as compare d to nonface objects. Two prosopagnosic subjects were tested for the discri mination of objects from several categories (face as well as nonface) at di fferent levels of categorization (basic, subordinate, and exemplar levels). Several dependent measures were obtained including accuracy, signal detect ion measures, and response times. The results from Experiments 1 to 4 demon strate that, in simultaneous-matching tasks, response times may reveal impa irments with nonface objects in subjects whose error rates only indicate a face deficit. The results from Experiments 5 and 6 show that, given limited stimulus presentation times for face and nonface objects, the same subject s ma demonstrate a deficit for both stimulus categories in sensitivity. In Experiments 7, 8 and 9, a match-to-sample task that places greater demands on memory led to comparable recognition sensitivity with both face and nonf ace objects. Regardless of object category, the prosopagnosic subjects were more affected by manipulations of the level of categorization than normal controls. This result raises questions regarding neuropsychological evidenc e for the modularity of face recognition, as well as its theoretical and me thodological foundations.