Subjects aged 54-84 performed 5 separate tasks involving various aspects of face processing: structural decisions (1), familiarity decisions (2), semantic decisions (3), first-name decisions (4), and name retrieval (5). For the categorization tasks (1-4), the mean reaction times for the older subjects (over 65) were plotted against the corresponding means for the younger subjects (under 65). This produced a linear function (slope greater than 1, intercept less than 0), providing only partial support for a simple, multiplicative model of cognitive slowing with age. Reaction time distributions were also plotted for each of the 5 tasks (older vs. younger subjects). The resulting functions were almost perfectly linear, with the exception of name retrieval, which was exponential with respect to age. This was attributed to the increased probability of a tip-of-the-tongue state with age caused by insufficient activation at the level of the name information (the final state of face identification).