Computational Sociophonetics and the Linguistics of Diaspora: Some Methodological Implications

This presentation is about a study of language in the Jamaican community in Toronto. As a sociolinguistic look at a diasporic group, it exemplifies some of the theoretical stakes in the sociolinguistics of globalization. In essence, a study of language variation in diaspora highlights modern urban diversity, and thus forces us to abandon the notions of relative ethnic and linguistic homogeneity, which twentieth-century sociolinguistics is founded on. 

Professor Hinrichs will demonstrate the methodological implications for a computational analysis of sociophonetic variation among members of the community. To study a primary corpus of interviews, he tapped into two baseline corpora of (i) non-Jamaican Toronto English and (ii) mesolectal Jamaican Creole. Formant estimates were spectrographically obtained for more than 50,000 vowels using both semi- and fully automated procedures. Most critically, the suite of methods is centered on semi-automated transcription, forced time-alignment of transcripts and recordings, automated formant estimation, and statistics using current data-scientific methods. 

Using recent methodological advances in computational sociolinguistics, we can overcome some of the obstacles that have so far hindered a more fully empirical sociolinguistics of diaspora.