Categorization Method Affects the Typicality Effect: ERP Evidence from a Category-Inference Task

The typicality effect during categorization describes a phenomenon whereby typical items are more easily judged as members of a category than atypical items. Prior studies of the typicality effect have often used an inclusion task, which asks participants to assess whether an item belongs to a category. However, the correct exclusion of non-members is also an important component of effective categorization, which has yet to be directly investigated. Thus, the present study investigated how categorization method (inclusion vs. exclusion) modulates the typicality effect via behavioral and electrophysiological measures. Thirty-two participants (16 in the inclusion and 16 in the exclusion group) were shown six consecutive words that all shared one feature. Then, a seventh word was presented. The inclusion group judged whether the seventh word also possessed the feature, whereas the exclusion group judged whether the word did not possess the feature. The seventh word could be typical, atypical, or a nonmember of the category. Behavioral and event-related potential (ERP) data were collected. Behavioral results showed that the two groups did not differ in accuracy. However, typical items elicited shorter response times than atypical items, and this effect was more pronounced in the inclusion than the exclusion group. With regard to ERPs, interactions between item type and group were shown for the P2, N2, and N400 periods. Within the inclusion group, a typicality effect (indicated by a main effect of item type) was present in the P2 and N400 time windows, while the exclusion group elicited a typicality effect only in the N2 time window. These results provide electrophysiological evidence that an inclusion judgment task is more sensitive to category typicality than is an exclusion task.