Speech Science Forum 14th January - Dr. James Brand

Title: Identifying systematic co-variation of vocalic variables across speakers of New Zealand English

Abstract:

The study of sound change tends to concentrate on trajectories of particular variables in isolation, but it has proven challenging to move beyond individual or small groups of variables, towards a better theoretical and empirical understanding of sound systems. We introduce here, a large scale analysis of how full sound systems co-vary in different ways, demonstrating how constellations of vocalic variables operate together across multiple speakers. By statistically modelling F1 and F2 values for 10 monophthongs of New Zealand English with GAMMs, we were able to use the by-speaker intercepts as an estimate of how advanced each speaker was in our corpus, irrespective of known predictors of sound change (i.e. year of birth, gender, speech rate). By using Principal Component Analysis on the intercepts, we were able to investigate the underlying structural co-variation that exists across the vocalic variables. Our results demonstrate the inter-relatedness of vowel productions and vowel changes.  Within a large subset of vowels, we see 'leaders' and 'laggers'.  However there are also vowels which appear to be linked together through structural relationships, and some sets of vowels which appear to carry opposing social meaning, such that if you are innovative in one, you tend to be conservative in the other. Our results offer a means to overcome long-standing methodological challenges in the study of phonetic co-variation, offering novel insights into the structure of sound systems across large groups of speakers.