Speakers often imply more meaning than is stated directly in a sentence. What are the exact cues that listeners employ to construct implied meanings? One useful cue is sentence prosody, which includes the pitch pattern of the sentence, known as the intonational contour. This dissertation investigates how adult listeners of spoken American English interpret intonational contours that imply contradiction. An example of this is "The cookie jar WAS full?", where a strong, contrastive accent on 'WAS' is followed by a rise at the end of the clause. To most listeners, this tune suggests that the cookie jar is no longer full (as verified by preliminary research for this project). However, the cognitive processes of interpreting such cues and meanings are underexplored.
Three sets of experiments will exploit distinct methodologies -- Sentence Continuation, End-of-sentence Picture Naming, and Visual Search with Eye-tracking -- to investigate a set of carefully matched intonational tunes (e.g., neutral, emphatic, and contrastive contours) allowing detailed comparisons of the relationship between specific parts of the tunes and resulting meanings. The Sentence Continuation studies will establish the extent to which each type of contour induces contradictory interpretations, as well as the exact type of contrast that is conveyed. The Picture Naming studies will explore the time course of constructing mental representations for both the literal meaning (the full cookie jar) and an implied contrast (an empty cookie jar). The Visual Search studies will track listeners' eye-movements to allow a detailed analysis of the time point at which the pragmatic contrasts emerge from each component of the intonational tunes. This project will help explain the role of prosody in higher-level meaning processing. The results should aid speech engineers working on the improvement of automatic speech generation and recognition and second language learners of English, among others.
