Identity priming in the recognition of familiar faces.

Morton (1969) proposed that word recognition is mediated by logogens--threshold devices whose thresholds would be lowered after 'firing' and only slowly return to near their original levels. Extension and revision of the original logogen model has occurred in response to results obtained in experiments which examine the effect of previous exposure to an item on later recognition thresholds. On the basis of results obtained in such identity priming studies, Warren & Morton (1982) argued that picture recognition might be mediated by pictogens--threshold devices which respond when any picture of an object is seen. Warren & Morton found that recognition of a pictured object was facilitated by earlier exposure to the same or a different picture of the object, but was not facilitated by earlier exposure to the object's name. Recently, several authors have suggested that face recognition might be mediated by units somewhat analogous to logogens or pictogens. In order to explore this analogy, Warren & Morton's procedure of identity priming in picture recognition was adopted to examine identity priming in the recognition of famous faces. When the recognition test involved naming a celebrity's face, identity priming was obtained from prior exposure to the celebrity's name, a different picture or the same picture as that used in the test. If the procedure was modified so that subsequent recognition did not require retrieval of the name, a pattern consistent with the word and picture recognition literature was found. Prior exposure to either a different or the same picture produced a priming effect but prior exposure to the celebrity's name did not. These results are generally consistent with the hypothesis of face recognition units. Differences between face recognition units and pictogens are also discussed.