Ninety-eight healthy participants were examined with a new test of Famous Name Comprehension which, in the framework of a serial model of person processing, sequentially assessed Name Recognition (i.e., the ability to classify items as familiar or unfamiliar) and Person Identification (i.e., the ability to provide biographical knowledge of recognised items). Names were presented in a written format. A perfectly equivalent face version of the test allowed a comparison of familiarity and identification of people from name and from face input. Furthermore, the effect of the "age" of the items, i.e. the time elapsed from the presumed first exposure to the stimulus to the time of testing, was also investigated. Normative data are provided. Education was the only significant variable for recognition, while education, age and gender turned out to be significant for identification. Recognition was significantly better with name than with face input, while on identification names and faces did not differ significantly. "Oldest" items were both recognised and identified significantly worse than recent ones. The results of face-name comparison are interpreted in terms of the different opportunities to be exposed to names and faces, the relevance of visuoperceptual attributes linked to faces and the evidence of shared knowledge from different inputs. The relative advantage of recent celebrities supports the semantic characterisation of knowledge of famous people.