Abstract
Vowel spaces of the varieties of Bruinafrikaans*
The speech of “coloured” speakers of Afrikaans, an important variety of Greater Afrikaans, offers a particularly rich field for answering the question of whether we can speak of Bruinafrikaans as a homogeneous entity, or whether, alternatively, there could be talk of two or more subvarieties of it. Wissing (2022 and 2023) endorses the use of the designation Bruinafrikaans based on limited investigations. Given that these investigations were limited in scope, a broader and deeper look at this matter is justified and necessary. This is the main objective of the research that this article reports on, which involved comparing the vowel spaces of ten areas where Bruinafrikaans is concentrated.
The general approach regarding the choice of surveyed speech areas and method of investigation is the same as that followed in Wissing (2022 and 2023), as is the method according to which the present investigation was carried out.
We analysed and interpreted speech recordings of the reading of a standardised word list by 346 speakers from ten recording areas in Southern Africa to achieve the stated objective. Of these areas, four are in the Western Cape, two in the Northern Cape, a few in the Eastern Cape, one each in the Free State and North West, as well as one in Namibia. More specifically, the areas in which the recordings were made were the following: Western Cape (Cape Flats: Delft, Blikkiesdorp, Bonteheuwel, Elsiesrivier, Langa, Mitchells Plain, Ravensmead and Steenberg); the Boland (Robertson); the southern Cape (De Rust, George, Dysselsdorp, Oudtshoorn, Ladismith and Mossel Bay), and the district of Matzikama on the West Coast. Eastern Cape’s participants come from Cradock, Dispatch, Graaff-Reinet, Humansdorp, Janseville, Joubertina, Middelburg, Port Elizabeth (Gqeberha) and Uitenhage. In the Northern Cape participants were recruited in Kakamas and its close vicinity, as well as the district of Springbok. Participants from North West were inhabitants of Promosa (Potchefstroom). Residents of Bergmanshoogte, Philippolis in the Free State were also involved as readers. Finally, in Namibia’s case, participants were from Gobabis and surrounding areas.
The Afrikaans vowels involved in the present study are /i/, /y/, /e/, /ø/, /ɛ/, /ɑ/, /a/, /ə/, /œ/, /ɔ/, /o/ and /u/ (in kies, nuus, bees, reus, ses, bas, baas, kis, bus, bos, boot and koek respectively; the three basic diphthongs, /əi/, /œy/ and /œu/ are those in byl, uit, and oud respectively. An acoustic description of these segments was elicited to be used in the present study.
The method of data processing and analysis that we followed in this study is the same as that of similar previous studies. Here it is sketched in broad strokes (see Wissing 2022 for details). In the present article the emphasis is not on the basic acoustic information regarding the individual vowel segments, nor on how the vocal spaces are constructed, but on the extent to which these spaces can be compared with each other – such a comparison is not possible with the statistical methods that have not been used so far. For that purpose, a set of new statistical methods has been employed here, namely those of network analysis and cluster techniques, which enable one to compare complete vocal inventories with one another. As pointed out above, such a comparison is essential for achieving the objective concerning the issue of homogeneity.
Distance measurements between two vowel spaces were done using the method of Pastore and Calcagni (2019). First, overlap dissimilarity of the vowel frequency formants F1 and F2 of all vowels was calculated separately. Gliding vowels and diphthongs are characterised by sliding formant tracks that do not run horizontally as monophthongs do, making the comparison of vowel spaces considerably more difficult. In this study, we therefore treated the non-monophthongs as monophthongs, in the sense that we used the average hertz values of their F1 and F2 values, as in the case of the monophthongs. For each pair of recording areas, a row of 15 overlap dissimilarities was found for F1 and likewise for F2 (only vowels that have data points for all locations in the dataset were taken into account). Since an overlap dissimilarity represents a standardised distance measurement, we used an average of all 30 scores in the calculation of a summed statistical index D of the distance between two vowel spaces.
Vowel spaces of these recording areas were determined, after which pairwise relationships between all vowel spaces of all ten recording areas were calculated and visualised using a neural network. This was followed by cluster analyses. Constructed dendrograms show that two subgroups can be distinguished, namely the southern Cape, Delft, Robertson, Matzikama, and Eastern Cape on the one hand, and Bergmanshoogte, Springbok, Promosa and Kakamas on the other; Gobabis forms its own, separate area, but can also be classified in the latter group. This is largely consistent with a geographical division into south and north. The geographical distribution can be explained to a large extent historically, considering the migration of Afrikaans speakers from Cape Town and surroundings northward and also eastward. A similar comparison of the vowel spaces of both sexes (female and male) and both age groups (young and old) show only small to medium differences, which means that Bruinafrikaans and its variants are indeed to a great extent a single, and also stable entity.
* The literal translation of Bruinafrikaans is Brown Afrikaans, but we prefer to use the Afrikaans name in English as well, also in preference to the other possible translation, Coloured Afrikaans. (Please note, incidentally, that while the term coloured may have negative connotations in other parts of the world, it is commonly used in South Africa to refer to this specific ethnic group without anything negative being implied by it.)
Keywords: acoustic characteristics; age; Bruinafrikaans; designation; ethnicity; gender; geographical location; language typology; sociophonetics; spectral properties; vowel; vowel frequencies; vowel space; vowel system