Integrative Processes in Utterance Resolution

Abstract This study investigates how listeners integrate spoken utterances into a discourse representation. We covary discourse focus (whether the discourse structure foregrounds a particular protagonist), pragmatic inference, and type of anaphoric expression, using a cross-modal naming task where subjects hear a short discourse followed by a sentence fragment. Immediately at the offset of the fragment, a visual probe word is presented. This probe is either an appropriate or inappropriate continuation of the fragment, depending on how the sentence fragment has been interpreted in its discourse context. Naming latencies to the visual probe show that all three factors (pragmatic inference, discourse focus, and type of anaphor) can effectively link utterances to discourses, and that the language comprehension system is highly flexible in its use of different kinds of processing information. We discuss the implications of these findings for Fodor′s modularity hypothesis.