Ness, T. & Meltzer-Asscher, A.
Tel Aviv University
Studies have shown that during the processing of a sentence, listeners/readers make predictions regarding upcoming words (van Petten & Luka, 2012). The current study sought to examine the mechanisms that come into action when such predictions are not met. We hypothesized that if the most predictable lexical item is pre-activated (and perhaps integrated into the sentence), inhibition of this lexical item would be needed in order to integrate an unexpected word that appeared instead. In contrast, such inhibition would not be needed if an anomalous word appeared, since this word cannot be properly integrated into the sentence's representation.
To test this hypothesis, 42 participants read sentences (84 sets) presented word-by-word and made an auditory lexical decision regarding the predicted word or a matched unrelated word, in one of three positions: after the verb (were the prediction is made), after a congruent-unexpected word, or after an anomalous word. Results support the aforementioned pattern of inhibition: a priming effect measured after the verb was reduced to a significantly greater extent when a congruent-unexpected word appeared, than when an anomalous word did. These results indicate prediction at the verb, followed by lexical inhibition only when a congruent-unexpected, and not an anomalous, word appears.