Effects of word- and sentence-level contexts upon word recognition.

In the present experiments, we investigated the facilitatory effects of auditorily presented prime words in neutral sentence contexts upon visually presented target words. It is shown that when the target task is a lexical decision, facilitation is obtained when the relationship between the prime and the target is one of synonymy (and low association strength). When instead the target is an antonym (again of low association strength), there is no priming effect; lexical decision is facilitated only when the prime word is presented in isolation. In further experiments, it is shown that primes in sentence contexts can produce facilitation of antonyms if they are strongly associated, or in the absence of association if the target must be named. The results are explained in terms of an integration process, which checks for the coherence between an upcoming word and the highest available level of representation of the context, but which affects only responses in the lexical decision task.