https://orcid.org/0000-0003-3642-5812
![Sample our Language & Literature journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5941/TOC-Subject-Sample-2021-Language-Literature.jpg"})
ABSTRACT
Increased global migration to international urban centres has motivated a growing interest in ethnolects and the role migrant communities play in language variation and change. Here, we consider ethnolectal variation in real and apparent time, by examining the realization of word-final (er) (e.g. teacher, remember) in Australian English. We capitalize on sociolinguistic interview data collected by Barbara Horvath in Sydney in the 1970s as a benchmark against which to compare newly collected recordings with Sydneysiders in the 2010s. Approximately 15,000 tokens of word-final (er) were extracted from the speech of nearly 200 people, including Anglo-Australians, and second-generation migrants of Italian, Greek and Chinese background. Acoustic analyses of vowel duration and position in the vowel space reveal incremental lengthening with concomitant lowering and backing over time for (er), though only in prosodically final position. This change was led by Greek and Italian teenagers in the 1970s, then taken up by working class women, and today, has been adopted across the community. Tracking this change in real and apparent time provides evidence that ethnolectal features may be adopted by the wider community, with ethnic minorities playing a leading role in language change.
Acknowledgements
We gratefully acknowledge support from the ARC Centre of Excellence for the Dynamics of Language and the Sydney Speaks team. We thank Benjamin Purser for his work in establishing the class clusters, and Benjamin Purser, Ksenia Gnevsheva and two anonymous reviewers for valuable feedback on the paper.
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes on Contributors
James Grama is a Research Fellow and Lecturer in the Sociolinguistics Lab in the Department of Anglophone Studies at the University of Duisburg-Essen. From 2017-2020, he was a Postdoctoral Fellow at the ARC Centre of Excellence for the Dynamics of Language at the Australian National University, where he worked on the Sydney Speaks project. He holds a Ph.D. in linguistics from the University of Hawaiʻi at Mānoa. His research focuses on sociophonetic questions of language variation and change, with an emphasis on vowel behaviour. He has worked largely on varieties of English (in e.g. Australia, Hawaiʻi, and California), English-based creoles (e.g. Hawaiʻi Creole, Bislama), and minority languages (e.g. Matukar Panau).
Catherine E. Travis is Professor of Modern European Languages in the School of Literature, Languages and Linguistics at the Australian National University, and a Chief Investigator in the ARC Centre of Excellence for the Dynamics of Language. Her research addresses questions related to linguistic and social factors impacting on language variation and change, in particular in socially diverse communities. She leads the Sydney Speaks project on which this paper is based, and has also worked extensively on the study of variation in Spanish and in bilingual speech.
Simon Gonzalez is a Research Officer at the ARC Centre of Excellence for the Dynamics of Language (CoEDL) at the Australian National University (ANU), where he works on the Ku Waru Child Language Socialization Study. From 2017-2020, he was a Postdoctoral Fellow on the Sydney Speaks project and the Transcription Acceleration Project at CoEDL at the ANU. His research focuses on acoustic phonetics, empowered by computational technologies. He develops computational tools (scripts and online apps) for more efficient and practical analysis of phonetic and phonological phenomena. He also makes use of programming languages to facilitate the forced alignment of speech data, in both majority and minority languages.
Notes
* Seven of the ten languages shown here were among the eight most widely spoken in both years; of those that were not, German and Polish were in the top eight in 1991 only, and Hindi was in the top eight in 2016 only.
1 The ethnic makeup of second-generation migrants cannot be readily ascertained from census data. The Australian census asks about ethnicity for Aboriginal and Torres Strait Islander people, but does not ask specifically about ethnicity more broadly. It also does not ask about parents’ language, nor (from 2001 onwards) parents’ country of birth. While it does ask about ancestry, the responses to this question are difficult to interpret, and the ABS recommends considering these results alongside language, religion and country-of-birth data (Australian Bureau of Statistics, Citation2016a).
2 Six of the 43 Anglo participants in the 2010s corpus included here report having one grandparent who was born outside Australia in another English-speaking country; one participant reports having an Aboriginal-Australian grandparent. We do not have information about the grandparents for the 1970s participants.
3 Of the 14 Young Adult Italians, nine are second-generation and five are third-generation migrants.
4 In brackets following each example are the corpus name, the speaker social code (community, age group, sex) and speaker number, and time stamps of the beginning and end of the excerpt. All names given are pseudonyms.
5 Systematic checking was deemed unnecessary due to the high quality of boundary placement by LaBB-CAT (cf. Gonzalez et al., Citation2020), as confirmed by spot checking of the (er) alignment.
6 Approximately 20% of the (er) tokens in the prosodic transcription were not matched with the corresponding LaBB-CAT tier (due to differences across transcription versions in LaBB-CAT and ELAN, for example when the timestamps were not identical), and were excluded from analysis.
7 Speech rate was calculated based on pause-to-pause intervals in LaBB-CAT rather than IUs because of the convention in the prosodic transcription method of assigning all of a sound file (including silences) to an IU, meaning that pauses are included in IUs, rendering them less reliable as a measure of speech rate.
8 The general inhibiting effect of following segment is evident in IU-medial position, where the marginal lengthening observed is primarily due to the small number of IU-medial pre-pausal tokens (n = 86), which go from a mean of 101 ms for 1970s Adults to 176 ms for 2010s Young Adults. Interestingly, IU-medial pre-vocalic tokens (e.g. remember it, whatever else) lengthen similarly to pre-consonantal tokens (pre-vocalic means: 51 ms 1970s Adults; 61 ms 2010s Young Adults; pre-consonantal means: 51 ms 1970s Adults; 57 ms 2010s Young Adults). This appears to be due to hiatus resolution, which, based on auditory observations on a subset of the data, includes preservation of /ɹ/, but also glottalization, and glide insertion.
9 Appeal intonation accounts for 3% of the IU-final (er) tokens produced by 1970s Adults, 10% for 1970s Teens, 5% for 2010s Adults and 6% for 2010s Young Adults.