Using Tulving's (1985) remember/know procedure, the present research investigated the experiential concomitants of person recognition. Noting basic differences in the manner in which the mind processes expectancy-related material, it was anticipated that facial typicality would be a critical determinant of people's recollective experiences (i.e., remembering vs knowing). In particular, it was expected that whereas remember responses would be more prevalent for distinctive than typical faces, know responses would reflect the opposite pattern. The results of two experiments provided general support for these predictions. In addition, the recollective advantage for distinctive faces was traced to the availability of attentional resources at encoding. These results are considered in the context of contemporary issues in person recognition and social cognition.