Memory and context effects on representations of variation in phrasal intonation

Northwestern University, Evanston IL

Investigators

Abstract

This award is funded in whole or in part under the American Rescue Plan Act of 2021 (Public Law 117-2). Meaning is communicated in spoken language through words and their arrangement in sentences, and also through the way the words and sentences are said — their intonation. In English, intonation conveys discourse meaning, marking the distinction between a question and a statement, expressing the speaker's uncertainty or surprise, or highlighting informative words. These meanings are conveyed through complex patterns in pitch and other acoustic features, across the words in the sentence. Understanding how an intonation system works requires a detailed description of the intonation patterns that speakers produce, and an account of the elements that are perceptually salient: Is the pitch pattern of every word in a sentence potentially meaningful, and if not, which intonational features and which regions of the sentence are more (or less) critical for conveying discourse meaning? A further question is whether listeners pay attention to subtle variation in intonational features as marking corresponding distinctions in broad categories like 'question' or 'information focus', or alternatively, whether they attend to coarse-grained distinctions in intonation as cues to the broad meaning categories, while perceiving fine-grained variation as cues to speaker identity or psychological state. Understanding the mapping between intonation and meaning is important for theories of speech communication, cognition, the brain, and language learning. It has applications in clinical domains related to speech, hearing, and the brain, and for enhanced computer speech technologies that can process expressive spoken language. The goal of this project is to identify the elements of phrasal intonation that are perceived in heard speech and encoded in long-term memory representations. These representations are tapped through tasks where listeners reproduce the intonation of a heard utterance. Experiments will test 24 distinct intonational "tunes" proposed for English in the Autosegmental-Metrical theory of English intonation. In eight experiments, the participant's task of reproducing intonation is manipulated in three ways: using model sentences with more complex intonational patterns and variable lexical and syntactic content; introducing a time delay or an intervening speech task between hearing the model sentence and reproducing the intonation pattern; with information about the immediately preceding discourse and social context that support the interpretation of intonational meaning. Analyses of dynamic acoustic measures will test the similarity of reproduced intonation patterns to corresponding stimuli, revealing the phonetic granularity of intonational representations along dimensions related to pitch, loudness, tempo, and voice quality, and whether individuals vary in the intonational features they reproduce. The findings will resolve long-standing questions about intonational representation, informing linguistic theories of intonation and its meaning, and advancing intonation research with a novel experimental paradigm that can be used in studies of other languages, with no requirements on written language or the explicit definition of intonational meaning. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.