The shape of a spoken phrase

Abstract

When someone speaks, you don't just hear a string of words. Usually, you also hear how the words are grouped, and how the word groups are related; you hear where the speaker wants to focus our attention; and you hear a lot about what kind of person is talking, what the situation is like, how the speech fits into the situation, and how the speaker feels. You may hear hesitation or fluency, haste, or relaxation, a happy grin or a mouthful of chewing gum. Nothing will be said here about grins or gum, and relatively little about social setting, discourse structures, or the projection of emotions in speech. The topic is the shape of a spoken phrase, that is, the ways in which its fluent performance reflects the structure of the information it presents. Specifically, some of the ways in which the pronunciation of words in various languages reflects their grouping and their relative prominence will be examined. However, it is important to keep in mind that only a few apsects of potentially meaning‐bearing variation in linguistic performance will be treated, and that most real speech will also show effects of the many phenomena that are being ignored. The study of phrasal shape in speech sheds light on the investigation of linguistic information structure, is of practical importance in the design of algorithms for speech synthesis, and is relevant to the study of musical performance as a source of comparison and contrast.