Vowel duration before voiced and voiceless stops in the Indian English of Assamese speakers

Abstract

Speakers of many English varieties produce longer vowels before voiced obstruent codas than before voiceless (e.g., House and Fairbanks, 1953; Chen, 1970; Umeda, 1975; Van Santen, 1992), though the extent can vary (Jacewicz et al., 2007; Tanner et al., 2019; Deterding, 2005). While phonetic motivations have been proposed to explain this pattern (Ohala, 1983; Kluender et al., 1988), the effect in some varieties of English is observed to be larger than phonetically motivated (Ohala, 1983; Keating, 1985); furthermore, some languages have less of a voicing effect (e.g., Korean, Russian, and French in Chen, 1970) and others may have none at all (e.g., Polish and Czech in Keating, 1985). This study investigates the phenomenon in the second largest variety of English by population, Indian English. Twenty Indian English speakers, all with L1 Assamese, were recorded reading a word list including 40 monosyllables that place vowels before voiced and voiceless versions of the same stop. We measure vowel duration using a Praat script to determine whether and to what extent this variety shares in the lengthening pattern, and in addition to providing the first measures of the phenomenon in Indian English, our results will be compared with other varieties of English and other languages.