Studies of ethnic orientation (EO), wherein scales are used to measure the degree to which people position themselves with respect to their ethnic heritage, have proved beneficial in explaining interactions between dimensions... [ view full abstract ]
Studies of ethnic orientation (EO), wherein scales are used to measure the degree to which people position themselves with respect to their ethnic heritage, have proved beneficial in explaining interactions between dimensions of ethnic identity and sociolinguistic variables (e.g., Bharadwaj, 2013; Hoffman & Walker, 2010). We contribute to this growing body of research by investigating the ways the subscales of an EO scale covary with voice onset time (VOT) in word initial stops in Australian English spoken by people with Lebanese heritage. In Australia, people with Lebanese heritage comprise the 9th largest non-Anglo-Celtic ethnic group (ABS, 2016).
VOT varies as a function of linguistic and extra-linguistic factors and its patterning is language- and variety-specific (e.g., Abramson & Whalen, 2017). It is attractive as a sociophonetic variable because it varies as a function of numerous social factors, including speakers’ ethnic identities and heritage language backgrounds (Nagy & Kochetov, 2013).
Analysing data collected in Melbourne from thirty Australians with Lebanese heritage, our aims are to consider:
- the phonetic factors that condition VOT variability;
- how dimensions of EO (namely, identification, social networks, culture, language proficiency and domains, commitment, and exploration (Anonymous, 2013)) covary with VOT variability; and,
- interactions between these linguistic and social factors.
In this quantitative sociophonetic study, we use algorithms trained on a subset of these data in the corpus phonetics tool, autoVOT (Keshet, Sonderegger, & Knowles, 2014). We measured positive VOT for 8,190 tokens of word initial stops /b, d, ɡ, p, t, k/ in stressed syllable-initial position. Results are modelled using linear mixed effects regression. VOT patterns by place of articulation, following vowel height, and EO. There is a mean positive VOT of 77ms for VOICELESS and 20ms for VOICED stops. Grand mean EO is 3.1 (out of 5) and EO subscale means range from 2.5 to 4.4. In sum, ethnic orientation scores are good predictors of VOT in these speakers, and the interactive effects of the subscale scores on VOT highlight the dynamic interplay between the social dimensions of ethnic identity and the ways in which these dimensions are indexed in fine phonetic detail.
