Canadian raising and the perception of consonant voicing

Abstract

Studies have shown that monophthongal VC sequences with higher F1 and shorter durations result in more [-voice] consonant responses than those with lower F1 and longer durations [Moreton, E. 2004 J. Phon. 32(1), 1, Nearey, T. M. 1997, J. Acoust. Soc. Amer., 101(6), 3241]. However, certain diphthongs in varieties of English with Canadian Raising show the opposite pattern, namely that before C[+voice], /aɪ/ diphthongs will have higher F1 in the nucleus than before C[-voice]. This study will investigate the interaction of interpretations of vowel tokens as monophthongs or diphthongs on the perception of the voicing of the following consonant. Measurements will be made from the speech of 10-20 speakers of Canadian English from the Edmonton area to be used in the resynthesis of hVd and hVt tokens. This will involve construction of a multi-dimensional continuum varying F1 values of the vowel nucleus as well as vowel duration and trajectory from nucleus to offglide. The resulting continuum will span both the four phonetic vowel categories [^, a, ^ɪ, aɪ] (where [^ɪ] and [aɪ] are viewed as allophones of a single phoneme /aɪ ), and also the final consonants /t/ and /d/.