Knowledge about intonation in ordinary speech is increasing rapidly. In part, this is due to the practical motivation of producing synthetic voices that do not sound like robots. But the main reason is theoretical: researchers have begun to study pitch TARGETS (peaks and valleys) rather than pitch MOVEMENTS (rises and falls). This project builds on a previous ESRC project, on Modern Greek, which discovered that pitch targets align precisely with specific points in a word, and that pitch movements are squeezed or stretched depending on phonetic context. Previously most researchers assumed that pitch movements have constant slope or duration. This finding appears to apply to English and Dutch (the two languages for which most is known about intonation), and provides a focus for research. The project should yield clear answers to some long-standing questions, such as how the pitch falls during the last accented syllable of a statement. Detailed answers to such questions are crucial for authentic-sounding synthetic intonation, as well as for theoretical research on the abstract structures underlying speech. Most of the project will consist of production experiments (acoustic measurements of carefully selected spoken utterances), followed up by perceptual experiments (where speakers judge utterances with experimentally modified intonation)

Start date
01 February 1998
End date
29 June 2001
Grant holder
Professor D Ladd
Professor Ineke Mennen
Grant amount
Grant reference
Grant type