Liminality as a lens on social meaning: A cross-variable analysis of gender in New Zealand English

The variationist sociolinguistic enterprise has been successful in developing models of structural (i.e., language-internal) drivers of variation and change, but one of the barriers preventing the development of a parallel model accounting for social drivers is the difficulty in operationalising salience. Using a corpus of sociolinguistic interviews collected in Auckland, New Zealand between 2013 and 2015, this project examines the relationship between language variation and gendered identity, and proposes an analytical approach of liminality – of examining the linguistic practices of people who have crossed a culturally-reified category boundary – as a possible solution to the problem of identifying socially salient variables.  The participants in this study are straight, queer, and transsexual native speakers of New Zealand English (NZE), representing two age groups that straddle the period of social change in the 1980s that saw a destabilisation of traditional gender roles in New Zealand. Variation in three linguistic systems is examined: adjectival modifiers (intensifiers and moderators), sibilants (/s/, /z/ and /ʃ/), and the vowels of NZE. The project uses established variationist and sociophonetic methodologies, as well as introducing a new metric for making comparisons across the vowel space as a whole. The findings show that speakers are able to encode their gendered identity across multiple variables, and that subtle linguistic signals can be used to affiliate with (or distance from) particular groups. The two most ‘distant’ groups are older straight men and younger [straight, queer, and transsexual] women, which this thesis argues is a reflection of the general societal criticism directed at young women by hegemonic masculinities. There is also evidence that the mainstreaming of queer identities in the wake of the 1986 decriminalisation of homosexuality has decreased the linguistic distance between straight and queer urban New Zealanders, demonstrating the relative rapidity with which social changes can be incorporated into the linguistic system.  This study highlights the utility of studying variation across linguistic systems rather than in variables in isolation, and proposes a typology of variation based on the boundedness and dimensionality of the linguistic systems under investigation. The framework of liminality is found to be productive with respect to the role of gender in language variation, although future research will have to test whether this framework can be generalised to other dimensions of social identity that are perceived as categorical (e.g., ethnicity and heritage culture in immigrant communities).