Sentence Context Effects on Lexical Access

Abstract

One of the most robust findings in the word recognition literature is that responses to words are faster when a word is preceded by a congruent context than when it is preceded by a neutral or incongruent context. For example, a word such as "treasure" would be recognized more quickly in the sentence "The pirate found the treasure" than in the sentence 'The person liked the treasure", or worse yet, ':The house was destroyed by the treasure" (e.g., Balota, Pollatsek, & Rayner, 1985; Ehrlich & Rayner, 1981; Fischler & Bloom, 1979, 1985; Foss, 1982; Schuberth, Spoehr, & Lane, 1981; Simpson, Peterson, Casteel, & Burgess, 1989; Stanovich & West, 1979). No one disputes the fact that context plays a role in the processing of individual words in a sentence. The controversy arises when we begin to ask more specific questions about which contextual factors are influencing the construction of which representations. There are two questions at issue here: What is the locus of the context effect?, and What is the processing mechanism that accounts for it? There is widespread agreement that the structure of language can be described in terms of a number of subsystems such as morphology, syntax, and semantics, and furthermore, that these systems are fairly independent in that the rules and representations used to describe each of these systems are different. However, there are still widely divergent views about the extent to which this structural independence translates into processing independence, with modular theories of language processing at one end of the continuum and interactive theories at the other.