Regionality in white South African English : an acoustic dialectometric investigation

Abstract

There is general consensus in the literature that White South African English (SAE hereafter) historically had a number of regional dialects; but the regional component to this variety was levelled in the mid-20th century. As a result, most speakers of SAE born after roughly 1930 no longer index regional affiliations through their speech habits. However, the topic of regionality has been gaining in consideration over the past decade or so. Literature on SAE is surveyed to frame the development of the variety from a historical perspective. Moreover, key points in its history are correlated to Schneider’s (2007) Dynamic Model by way of accounting for the development of SAE. In his model, Schneider outlines five phases; a given variety of English may proceed through each until it achieves the fifth – and final – phase. A re-proliferation of regional variation is the defining feature of the final phase. Accordingly, the hypothesis this dissertation seeks to test may be articulated in two components: regionality is increasing in SAE – which evidences the entry of the variety into phase 5 of the Dynamic Model. To test this hypothesis, standard Labovian sociolinguistic interviews were conducted with participants from three Anglophone urban centres in South Africa: Cape Town, Durban, and Johannesburg. The interviews were digitally recorded and subsequently analysed with the FAVE program suite (Rosenfelder et al., 2011). Results obtained from this analysis were subjected to traditional vowel-by-vowel analysis to determine whether individual vowels vary from one region to the next in terms of quality, duration, or differential behaviour with regard to style shifting. Additionally, data have been analysed with the Gabmap toolkit for dialectometry (Nerbonne et al., 2011). This latter analysis is novel in two respects: dialectometry has neither been applied to SAE directly nor has it been incorporated into a study which proceeds from a Schneiderian perspective. The inclusion of dialectometry is motivated by its explanatory power for drawing divisions between varieties that differ significantly from each other. Following data analysis, a brief profile of variation in the vowels of SAE is adduced to demonstrate the entry of the variety into the final phase of the Dynamic Model. That is, results as obtain in both vowel-by-vowel and dialectometric analysis support the hypothesis that speech habits pattern in a way that correlates to region in the SAE speech community. Moreover, the vowel-by-vowel analysis supports earlier accounts of nascent regionality in the variety as well as certain diachronic changes in the indexical value of variants (as proposed by Mesthrie et al., 2015; Wileman, 2011). Finally, it emerges that the so-called Standard Model of the Formation of SAE is supported by present data. Results obtained by Wileman (2011) are replicated for Cape Town and Durban: Durban favours centralised variants for the KIT vowel and monophthongal articulations for the PRICE vowel. Results obtained by Chevalier (2015) for the front vowel pull chain, which involves the TRAP, DRESS, and KIT vowels, are also replicated – with the addition of possible involvement from the STRUT vowel. Novel findings include a preference in Johannesburg for ‘hyper-diphthongs’, i.e., glide-strengthening, a trend for males to lead most of the innovations which promote regionalisation, and a change in the former indexical value of the GOAT and MOUTH vowels – particularly in Durban.