Spreading activation within semantic categories: Comments on Rosch's "Cognitive representation of semantic categories."

Abstract

In the letter-matching task of Posner andMitchell (1967), a subject looks at a pairof letters and decides whether they are thesame or different. The typical finding isthat subjects take longer to respond toname-identical pairs such as Aa than tophysically identical pairs such as AA. In avariation of this procedure, Beller (1971)presented his subjects one of the pairs ofletters to be matched in advance and foundthat this advance information (or prime)shortened the reaction time for both physicaland name matches. This result supportsother data suggesting that when a subjectis provided in advance with part of the in-formation needed to make a timed re-sponse, his reaction time upon reception ofthe remainder will be reduced (cf. Cohen,1969; LaBerge, Van Gelder, & Yellott,1970).Rosch's (1975) procedure is a variationof Beller's in which subjects matched pairsof words or pictures instead of letters. InExperiments 2, 3, 5, 6, 7, and 8, Rosch'ssubjects were shown two category members(e.g., orange, pear) and their task was topress a same key if both members of the pairbelonged to the same natural category and,if not, to press a different key. In somecases, the pair to be judged was preceded oraccompanied by a priming stimulus. Thispriming stimulus was a category name, suchas fruit, which informed the subject that atleast one of the two stimuli he would seewould be a fruit. In Experiments 4 and 5,physical identity instructions were used: Thesubject pressed a same key if the stimuli