A new mathematical model for intonation: Historical motivation and application

Abstract

Early attempts to characterize English intonation were theoretical descriptions in terms of tones, tunes, and segmental pitch levels. Recent acoustic descriptions of intonation have involved the perceptual determination of F0 contours in terms of global declination lines or as breath groups with terminal and nonterminal slopes. Lieberman et al. (1985) showed that F0 contours could be described objectively by all‐points linear regression lines, but this did not yield a compact characterization of the contours. Hirst (1983) developed a model based on nonlinear (parabolic) function, but it is not unified in that it resets parameters for syllables independently. Attempts to compare intonation contours in different languages and dialects have been based on the existing models. In the following abstract a mathematical model capable of compactly quantifying intonation contours using four parameters is proposed [Drumright and Rekart (1987)]. This paper demonstrates how the model can be applied to analyze foreign language, non‐native, and dialectal differences in the intonation of statements and questions in spontaneous and read speech.