Response to Fred Cummins: Looking for Rhythm in Speech

Abstract

This commentary briefly reviews three aspects of rhythm in speech. The first concerns the issues of what to measure and how measurements should relate to rhythm's communicative functions. The second relates to how tonal and durational features of speech contribute to the percept of rhythm, noting evidence that indicates such features can be tightly language-specific. The third aspect addressed is how bodily gestures integrate with and enhance the communicative functions of speech rhythm.Submitted 2012 January 6; accepted 2012 July 13. KEYWORDS: rhythm, prominence, grouping, prosody, turn-taking IN his contribution, "Looking for Rhythm in Speech," Fred Cummins begins his introduction with the observation that rhythm is notoriously difficult to define and means very different things to different people. This is perhaps one of the main reasons why rhythm in speech has proved so difficult to study and why it remains such a fascinating and challenging topic. It is also why we must restrict ourselves to covering only a few sub-topics when dealing with the concept of rhythm in speech. Cummins reviews some of the history of empirical research which has frustratingly and unsuccessfully sought to find isochrony in speech. He then moves on to argue for a more promising and fruitful, holistic approach which embraces rhythm as engaging the entire body during speech. This view of rhythm encompasses bodily gestures as well as gestures of the speech articulators and integrates rhythm into a whole-body activity. I am very sympathetic to this approach and would like here to comment on three related topics. The first concerns what to measure and why, and how these measurements should relate to the communicative functions of rhythm. The second is how tonal and durational aspects of speech contribute to the percept of rhythm. The third is how bodily gestures integrate with and enhance speech rhythm in its communicative function.Fred Cummins describes what he fittingly terms "the great isochrony safari" as the hunt for measuring and finding regular temporal intervals in speech related to syllables or stresses. Measurements of acoustic regularity are meant to capture the regularity of articulatory dynamics, but as Cummins points out, the motivation for these measurements stems from intuition or at best a perceived regularity, the structure of which seems to differ across languages. This perception of regularity has proved difficult if not impossible to capture and quantify by means of purely temporal measurements leading to a widespread conclusion that rhythm in speech is a perceptual phenomenon rather than an acoustic one (e.g. Kohler, 2009;Niebuhr & Wolf, 2011). This somewhat paradoxical definition of rhythm as based in perception is not very satisfactory alone. I feel that a more useful and constructive definition should take into account the communicative functions of rhythm. This approach concentrates more on the reasons why we have the percept of rhythmical groups in speech and how this facilitates speech processing...