News across time: a sociophonetic study of Australian broadcast speech

Broadcast speech can be a mirror for linguistic change occurring in the wider community. The 'British sound' willingly adopted by Australian radio in the early 1920s was testament to a nation still under the influence of 'Empire'. By measuring longitudinal vowel change in the speech of 12 Australian newsreaders in real time, this sociophonetic study traces the development of a national identity that was moving away from its colonial heritage. Auditory impressions suggested that for some speakers, certain vowels considered to be 'markers' of Australian English had shifted from an accent approximating the British broadcast standard of Received Pronunciation (RP), to the majority accent of General Australian. Acoustic analysis was used to quantify these impressions, by comparing archival news material from the 1950s and 1980s with new recordings made in 2005. Transcribed excerpts of the speakers' own archival news bulletins were re-recorded at interview in 2005, providing two almost identical sets of data for analysis. Three groups of four speakers matched for age and gender were selected to form the corpus. In addition to the first aim of confirming diachronic vowel shift, changes relating to gender and the effects of ageing were identified. Analyses conducted include the first and second formant frequencies (F1/F2), the fundamental frequency (FO), duration, Euclidean distance and variability. Statistical tools employed to confirm and clarify the results were Welch two sample t-tests, Levene's test for equality of variance and linear mixed-effects models. Acoustic measurements confirmed auditory assessments of pronunciation changes. In particular, comparison of the two male speaker groups' archival vowel data demonstrated a change in the broadcast standard accent from 'RP-like' in the 1950s, to General in the 1980s. However, same-speaker comparisons at group level showed that the Females provided the most robust evidence of both accent broadening (decreased F2 values) and certain phenomena associated with ageing (i.e. lower F1 and FO values). The middle-aged Male group's results showed more ageing effects than the Elderly males, suggesting that younger speakers might be used as a reliable alternative to old or very old participants in gerontological language research. Results of the Male and Female group comparisons implied there were close links between identity, prestige and the professional construction of gender. The findings have added to the existing body of research on vowel change and age effects in Australian English. Most importantly, news speech is confirmed as reflecting social change occurring in the community, with the 'trend' study (comparing the 1950s and 1980s data) enabling the emergence of a more 'authentically Australian' identity to be acoustically quantified. Although results demonstrated the speakers' ability to modify their speech past young adulthood, no group displayed the vowel quality of the incoming younger generation identified in previous research on Australian English by Cox (1996).