TESTING A PREDICATIONAL MODEL OF COGNITION - CUEING PREDICATE MEANINGS IN SENTENCES AND WORD TRIPLETS

Predication is a logical process in which meaning is extended from a b roader context to a narrower, targeted referent independently of synta x and the passage of time. Since predicates lend meaning to their targ ets in cognitive processing, it follows that when unrecalled sentences are cued with their predicate-word meanings there should be greater r etrieval than when these unrecalled sentences are cued with subject wo rds. Three experiments (combined N = 164) tested this hypothesis and f ound data in its support (p - levels ranging from . 05 to . 001). A fo urth experiment (N = 48) removed syntax from consideration by employin g triplets in which one word out of three sharing a common topic was t he broadest in meaning, and hence was the expected predicate for cuein g triplets when they were not initially recalled. As predicted, it was found that when the expected predicate of unrecalled triplets was use d as a cue there were twice as many retrievals occurring as when the l ess broadly meaningful words were used as cues (p < .001). The finding s are discussed in terms of logical learning theory's claim that ongoi ng cognition involves the continual ''taking of a position'' within a sea of opposite possibilities.