Exploring the speech rhythm continuum

Ann Marie Olivio

doi:10.20396/joss.v1i2.15022

Abstract

In current speech rhythm research, the traditional search for isochrony and speech rhythm classes has been replaced with a focus on uncovering the acoustic correlates of rhythm in the speech signal (Ramus, Nespor, and Mehler 1999, Grabe and Low 2002, Cummins 2002). In this paper, I present findings from a study in which I describe speech rhythm in a language whose rhythm had not been previously studied—Ashanti Twi (Niger-Congo, Kwa). Additionally, I test the validity of claims made about the utility of various rhythm metrics. Two native speakers of Ashanti Twi participated in the study. Each speaker was recorded while reading a translation of “The North Wind and the Sun”. Vocalic and consonantal intervals were measured in Praat using auditory and visual cues. Various rhythm metrics (interval measurements and Pairwise Variability Indices) were then computed and compared to results from prior studies. Results show that Ashanti Twi is rhythmically more similar to languages that have been traditionally described as syllable-timed, such as French and Spanish. However, it does not fall clearly into the traditional stressor syllable-timed categories, supporting the claim that speech rhythm should be studied as a continuum rather than a categorical distinction.

Highlights

Current trends in speech rhythm research have focused on uncovering the acoustic correlates of rhythm in the speech signal
This categorical distinction was initially based on researchers‘ own perceptions of stress and beats as well as a timing assumption that was based on chest-pulses—the amount and intensity of air flowing from the lungs during speech: ―Speech rhythm is essentially a muscular rhythm, and the muscles concerned are the breathing muscles‖ (Abercrombie 1967:96)
The interval measurements ΔV, ΔC, %V, VarcoV, VarcoC, and speech rate collected from the Praat script are presented in Table 1 below along with the Pairwise Variability Index (PVI) scores calculated by the Excel worksheet

Summary

Introduction

Current trends in speech rhythm research have focused on uncovering the acoustic correlates of rhythm in the speech signal. This is a departure from the traditional search for isochrony characteristic of earlier work that was concerned with categorizing languages into either syllable- or stress-timed rhythmic patterns. Pike 1945, Abercrombie 1967) This categorical distinction was initially based on researchers‘ own perceptions of stress and beats as well as a timing assumption that was based on chest-pulses—the amount and intensity of air flowing from the lungs during speech: ―Speech rhythm is essentially a muscular rhythm, and the muscles concerned are the breathing muscles‖ (Abercrombie 1967:96).

Objectives

Results

Conclusion