Familiarity and Lexicality as Determinants of the Generation Effect

Two experiments addressed the question of whether representation in the mental lexicon is a sufficient condition for obtaining the generation effect. Contrary to an earlier report (Nairne, Pusen, & Widner, 1985), our experiments showed that low-frequency words did produce significant retention advantages when generated, but only when items were rated as highly familiar to the subjects; when low-frequency words were recognized as words but were rated low in familiarity, no generation effect emerged. Overall, our results support the position that lexical representation is neither a necessary nor a sufficient condition to produce the generation advantage.