PRIMING EFFECTS IN SEMANTIC CLASSIFICATION TASKS

Six experiments were conducted on priming in semantic classification tasks that allow free play between perceptual and semantic processes. Priming was greatest when words were repeated on the same semantic task at study and test but was absent when repeated words were classified on different semantic tasks (size and man-made; Experiment 1). Thus, merely repeating perceptual information is not sufficient to produce priming. Priming was obtained, however, when items on the same semantic task were repeated in different formats (words and pictures; Experiment 2). Consistent with stage models of single-word reading, priming was obtained when a semantic classification task was followed by a word form task (i.e., lexical classification or naming) but not when it was preceded by the word form task (Experiments 3 and 4). Priming was also found across lexical tasks that both involve the word form (Experiment 5) and across classification tasks that refer to the same semantic domain (overall size and relative dimensions; Experiment 6). Results suggest that priming is determined by the overlap in the component processes of the study and test tasks.