Weak contextual constraints in text and word priming

Abstract

The effects of text priming and word priming have often been confounded in the literature. Four lexical decision experiments are reported here that clarify this distinction. In the first two, priming effects are observed for words related to a prior text even though they immediately follow an unrelated word. Two further experiments dismiss the possibility that this effect is attributable to sustained priming, stemming from associatively related words occurring earlier in the text. Overall these findings run counter to predictions from Kintsch's (1988, Psychological Review, 95, 163–182) construction-integration model and favor the lexical distance model Sharkey, 1989a, Sharkey, 1990a. However, an absence of associative priming for prime target pairs embedded in text rather than scrambled sentences (Experiments 3 and 4) is a difficult finding to align with any current model of word recognition.