Introduction

This article offers a brief but general overview of the evolution of South African English (SAfE) as well as its current characteristics, both from a descriptive point of view as well as from the point of view of what might be referred to as the “social life” of this dialect, i.e., the linguistic system’s diachronic and synchronic relationships with social factors and forces. In line with the volume of which this chapter forms a part, emphasis will fall on the pronunciation features of SAfE. Details will be provided not only for the standard variety (General White SAfE) but also the various sociolects (e.g., Broad SAfE), ethnolects (e.g., South African Indian English) as well as L2 varieties (especially Black South African English, the numerically strongest and best researched of the nonancestral SAfE dialects).

In what follows, the social history of SAfE is first sketched, detailing its emergence via a complex nineteenth century koineization process and then focusing on subsequent developments. The process of the transmission of English to nonancestral communities also receives attention. The next section then provides an overview of the various varieties’ pronunciation features while the chapter ends with a section overviewing current developments in the field and a conclusion.

The historical sociolinguistics of South African English

The history of English in South Africa begins with the first British occupation of the Cape in 1795 (Giliomee and Mbenga 2007: 85). On the standard account it is not, however, until the arrival of the 1820 settlers in the Eastern Cape (see Figure 16.1) that a new dialect of English is born.

This episode in the colonial history of South Africa constituted what Trudgill (2004: 26) refers to as a “tabula rasa” context, i.e., “those in which there is no prior-existing population speaking the language in question, either in the location or nearby”. In other words, koineization took place among the various English dialects that served as inputs, the output of which was a new variety of English, which has been referred to as Cape English (CpE) in, for example, Bekker (2012). The standard picture, for example, in Lanham and Macdonald (1979) or Lass (1995), is that the 1820 Settlers were mainly of lower class origin and predominantly from the south-eastern part of England (including London). The (over)simplistic picture is, therefore, of a CpE reflecting many of the trends of early nineteenth century Cockney (and similar in many linguistic respects therefore to Australian English). However, according to contemporary historians such as Welsh (1998: 127) and Gilliomee and Mbengwa (2007: 85–86), the eventual group of about 4000 settlers, who were selected from among approximately 80 000 applicants, included a higher proportion of middle class, educated settlers, many of whom had some means upon their arrival in the Cape, and who did not intend to become farmers or laborers. The received view among linguists of the predominantly lower class origin of the settlers is thus challenged by historians, and an updated view may help to explain why SAfE, unlike the other Southern Hemisphere varieties, does not display some typical Cockney features, e.g., the use of -in for -ing for the present participle (talkin’ for talking). Another complication lies in the fact that the settlement area had already been populated to a degree by Cape-Dutch/Afrikaans speakers. There was much intensive contact (e.g., intermarriage) between the English and Afrikaans groups, even if political relations were often strained (Branford 1996: 38–39) and there is some debate in the literature as to whether SAfE (and thus by implication CpE) was influenced by Afrikaans on more than just a superficial level (i.e., on a structural as opposed to purely lexical or lexicogrammatical level), with Lanham and Macdonald (1979), Jeffery and van Rooy (2004), and Wasserman (2014), for example, supporting the notion, while Lass and Wright (1986) and Mesthrie (2002a) argue against it.

The second phase in the formation of SAfE was a period of settlement during the 1840s to 1850s and focused on Natal (now KwaZulu-Natal – see Figure 16.1). Here the standard picture is that the relevant settlers were of a middle to upper class origin, that there was virtually no Afrikaans influence on the koineization process and that there was a distinctly North of England bias, although this bias was no doubt tempered, although not completely, by the use of Standard English (and thus an early form of Received Pronunciation) by many of these middle class to upper class individuals. The output of the koineization process can usefully be termed Natal English (NE) and for many commentators the formation of SAfE ends here. This standard model of the formation of SAfE is, for example, made explicit in Schneider (2007: 176) who explains, with respect to the Eastern Cape and Natal periods, that “in both cases a recognizable founder effect is worth noticing: despite their relatively small numbers … these two groups laid the foundations for the main accents of present-day SAfE”. Bekker (2012), however, argues that an important third phase took place during the birth and development of Johannesburg, which was itself based on the discovery of gold on the Witwatersrand. A discussion of the technical details is not appropriate here, but in essence the argument is that Johannesburg constituted yet another tabula rasa context (Trudgill 2004: 26) and that a third koineization process took place, inputs into which included CpE, NE, and a whole gamut of other English accents (British as well as colonial, e.g., Australian and American) as well as L2 varieties such as the English spoken by L1 Afrikaans and L1 Yiddish speakers. With respect to the later, the immigrants to early Johannesburg included a sizeable number of mainly Eastern European Jews, particularly from Lithuania and Latvia (Kaplan and Robertson 1991). This group became the first nonancestral population to become fully incorporated into the White SAfE speech community.

As argued by Bekker (2012), the output of this last and third koineization process was a sociolectal continuum that many refer to as “South African English”, i.e., that variety still spoken primarily (although certainly not exclusively) by white L1 speakers of English in South Africa and henceforth referred to as White SAfE (WSAfE). This sociolectal continuum is traditionally broken up into three units, referred to by Lass (1995: 93) as “the great trichotomy” (a feature shared with other Southern Hemisphere Englishes):

1. A standard with an external British reference: in terms of pronunciation this is near-RP in Wells’ (1982: 297–301) sense and often approximates an older form of RP. This variety is hardly used among young speakers any longer (Lass 2002: 110). This is referred to in the literature as either Conservative or Cultivated SAfE (henceforth CWSAfE).
2. A more local standard has progressively become the most widely spoken sociolect of WSAfE; in terms of accent, lexicogrammar, and lexis, this standard is distinctive in relation to other varieties of English. It is either referred to as Respectable or General SAfE (henceforth GWSAfE). According to some commentators, such as Lanham and Macdonald (1979), GWSAfE is, very roughly speaking, NE absorbed into the Johannesburg mixing process and reanalyzed as a sociolect. In Lanham and Macdonald’s (1979) time at least both CWSAfE and GWSAfE were associated with “rejection of South Africanism in favour of links with the wider Anglo-Saxon world, a low level of patriotism, and hostility towards Afrikaners” (Jeffery 1982: 254). We suspect, however, that in the intervening 30 or so years, and in the case of GWSAfE, these associations have largely dissipated, partly as a result of the spread of GWSAfE at the expense of the other sociolects and partly because of the ideological effects of the political change to a fully democratic society in 1994. Still, while Coetzee-van Rooy and van Rooy (2005) find that black participants in their attitude study revealed a slight preference for the most educated (but still distinctively black) accents, the GWSAfE speaker was regarded very highly too, and certainly more highly than the Broad SAfE speaker (see below).
3. A variety alternatively known as Extreme or Broad SAfE (henceforth BWSAfE): the indexicality of this variety is more than just working class, an observation that, we suspect, remains as valid today as it was in Lanham and Macdonald’s (1979) time. As explained by Jeffery (1982: 253–255), BWSAfE is associated with attributes such as being “tough, manly, sport-mad, sociable, patriotic and other things beside …. Ext SAE is loaded with political-ideological meaning as well as social: the South African tradition is to be not only tough etc. but also conservative, right-wing, authoritarian, unsympathetic to African aspirations …. Ext SAE speech reliably predicts such views … which are a significant part of the stereotype of the ‘typical local man’. And indeed you do not have to be LC [Lower Class] to conform to the stereotype”. It should also be noted that “the more extreme the variety is, the harder it becomes to distinguish it from second-language Afrikaans English” (Lass 2004: 373). For Lanham and Macdonald (1979) and other commentators, the idea is, very roughly again, that CpE was absorbed into the Johannesburg mix and reanalyzed as BWSAfE.

During the twentieth century this sociolectal continuum has dispersed geographically, largely doing away with the original regional lects (CpE and NE) and creating a typical Southern Hemisphere level of regional homogeneity. Generally, GWSAfE has spread at the expense of both BWSAfE and, in particular, CWSAfE.

While WSAfE was undergoing its formative process, English also spread to other communities in the country, giving rise to nonancestral varieties that are widely encountered in contemporary South Africa. These include South African Indian English (ISAfE), Cape Flats English/Colored English (C[f]E) and Black South African English (BSAfE). Of these varieties, ISAfE has become the native language of the vast majority of its speakers, while a substantial minority of the Colored community has also adopted English as its home language – 21% according to the official 2011 census (Statistics South Africa 2012).

English is the home language of 86% of South Africans of Indian ancestry according to the 2011 census (Statistics South Africa 2012). Indentured laborers were recruited from India to work on the sugar plantations of Natal in the second half of the nineteenth century, while a number of free Indians, mainly traders, also emigrated to South Africa during this period. A total of about 150 000 Indians moved to Natal between 1860 and 1911, and about half of them stayed in South Africa upon the expiry of the indentured contracts (Mesthrie 1995). These immigrants spoke a variety of Indian languages, both Dravidian and Indo-European, some features of which have determined the linguistic nature of ISAfE. English was introduced very gradually into the linguistic repertoire of these immigrants and their descendants, with limited education until the 1950s, alongside some informal contact beyond the classroom (Mesthrie 1992). After the introduction of general schooling, however, language shift was very quick: Mesthrie (1992: 31) notes that older siblings brought English home from the school playgrounds, enabling younger siblings to enter school with a fair command of English. In the period from the early 1950s to the 1970s, English became the first language of virtually the entire school-going population in the Indian community (Mesthrie 2002b: 340), as first documented by Bughwan (1970), who found that 90% of her 547 respondents claimed English as their strongest language.

English came to share a place with Afrikaans in the linguistic repertoire of the Colored people of the Western Cape, where “Colored” refers mainly to descendants of slaves (who were emancipated by the British in the first half of the nineteenth century), as well as children of inter-racial marriages and some descendants of the Khoi who lived in the area prior to the arrival of Europeans. During the course of the nineteenth century, widespread Afrikaans/English bilingualism developed in this community, although a distinctive variety of Afrikaans remained the dominant language for most. In the latter part of the twentieth century, and even more so in the early years of the twenty-first century, however, English has become the dominant language of individuals entering the middle classes (Malan 1996; McCormick 2002). In practice, while differences can be observed in the English pronunciation of English and Afrikaans native speakers respectively, there is a shared core of pronunciation features, some of which can be related to the Cape Vernacular Afrikaans dialect spoken in the same community (Finn 2004).

BSAfE is the most widely used form of English in contemporary South Africa. The roots of the variety can be traced back to nineteenth century mission education (Beck 1997; Hodgson 1997; Shepherd 1941). Mission education provided excellent opportunities to acquire native-like competence by the end of the nineteenth century and continuing into the first half of the twentieth century (see, for example, De Klerk 1999), and was responsible for almost all education among black South Africans until the 1950s (Elphick 1997: 1). However, Hirson (1981: 220) notes that by the beginning of the twentieth century, the total enrolment of Africans in mission schools was still very small. It gradually grew to about 45% by the middle of the twentieth century (Booyse 2011b: 245), but the majority of children did not proceed beyond the second school year, while poor resources and overcrowding were the order of the day (Booyse 2011a: 202–205).

At this point in history, however, the situation changed dramatically: the new government, the National Party, implemented the Bantu Education Act. The government took control of all African schools (Hartshorn 1992; Hirson 1981; Booyse 2011b). This had two effects: “Under the new regime more children were accepted into schools, but the education was even inferior to that provided by the independent schools” (Hirson 1981: 227). This created a situation in (racially segregated) black schools where English was taught by Bantu-language speakers who themselves had limited training and command of English, and resulted in low levels of achievement, reinforcing features of home-language transfer (Lanham 1966). Just as the situation started to stabilize by the 1970s (as evidenced by the improved performance in school examinations), the final phase of political protest against the apartheid government was kicked off by the protest action in Soweto in June 1976. From this point onwards, education in townships reserved by law for black South Africans never quite returned to stability until the political transition in the early 1990s (Booyse 2011b: 257–262).

The cumulative effect of the twentieth century educational and political history of South Africa on the development of BSAfE is that a small elite close to the variety spoken by native speakers was removed from society, and a much more numerous group of relatively poorly educated speakers, with limited contacts beyond their own communities, developed in the second half of the century. However, English remained an important asset to the black community and it continued to be used in a range of functions (De Klerk 1999). Renewed claims about ownership of English started to emerge in the wake of the 1976 protest action in Soweto. This was articulated forcefully by public figures such as Mphahlele (1985) and Ndebele (1987) at addresses to the English Academy of South Africa. Since the political transformation of 1994, English has only increased in importance in the black community, while access to the language has also increased. Hence, in the present generation, significant changes are likely to occur.

The pronunciation features of South African English

White SAfE pronunciation has a number of distinctive characteristics. The area that has attracted the most attention is its vowels. A few consonantal properties have been identified, but nothing unique has so far been recorded in the literature as far as its suprasegmental features are concerned.

The following vowel features have been identified in WSAfE:

• WSAfE displays what has been commonly (and egregiously) referred to as the KIN-PIN Split by Wells (1982: 612–613). As shown in Bekker (2009), this is not a phonemic split at all but rather the entrenchment of allophonic variation in the KIT vowel. Basically in certain restricted contexts (e.g., after /h/) KIT is pronounced [ɪ], before tautosyllabic /l/ it is [ɤ], while in all other contexts it is [ə].
• Unlike Australian English and New Zealand English, WSAfE does not have a diphthongized FLEECE vowel (i.e., [əɪ] or thereabouts); even in BWSAfE it is a categorically monophthongal [iː] – a possible influence from Afrikaans.
• WSAfE does not participate as fully in the Diphthong-Shift and MOUTH-PRICE Crossover as do the other two Southern Hemisphere varieties (Wells 1982); i.e., at least in GWSAfE, MOUTH often has a similar starting point to PRICE (i.e., [ɐʊ] and [ɐɪ] respectively), FACE has a narrow diphthong (i.e., [eɪ]), while GOAT in GWSAfE is often fronted as opposed to lowered (i.e., [øʉ]). There is also much evidence of monophthongization in GWSAfE: FACE, as mentioned, is often narrow, GOAT is often subject to glide-weakening and PRICE is in fact considerably fronted and monophthongized in certain prestige varieties within GWSAfE (i.e., [pra:s] for price). This, however, only underlines the notion that a PRICE-MOUTH Crossover is not a particularly prominent feature of SAfE. It is only in the broader idiolects that one finds a relatively fronted MOUTH onset (i.e., [æʊ]), backed PRICE onset (i.e., [ɒɪ] or monophthongal [ɒ:]), and lowered onsets for FACE and GOAT (i.e., [ɐɪ] and [ɐʊ] respectively).
• WSAfE is often recognizable in terms of its substantially backed BATH vowel, which in the broader lects also shows lip-rounding (i.e., [ɑː] or [ɒː]); SAfE differs from Australian English and New Zealand English in this respect, both of which have a fronted BATH vowel, i.e., [aː]. Bekker (2012) makes a direct link between this feature of WSAfE and the importance of Johannesburg in the formation of SAfE.

Few unique consonant features have been identified, but the following is known:

• WSAfE displays allophonic variation between a clear and dark /l/, but there is no evidence of /l/-vocalization in the coda position (i.e., [jeɫ] not Cockney-like [jeʊ] for yell), and also Yod-Assimilation (e.g., [ʧʉːn] not RP-like [tjuːn] for tune). According to Bowerman (2004: 935), aspiration is not consistently present in voiceless plosives in syllable onsets.
• Broad WSAfE often displays features that can be linked to early Afrikaans influence (via CpE), e.g., obstruent (tapped) /r/ (e.g., [ɾeːliː] for really), semi-rhoticity, and epenthetic schwa (e.g., [fələm] for film). The L2 English variety spoken by Afrikaans speakers (i.e., Afrikaans English) also shows evidence of syllable-final devoicing ([dɒk] for dog), although some of the contrast is retained by lengthening the previous vowel (van Rooy and Wissing 1996).

South African Indian English displays a variety of dialect-specific phonetic features, many of which are traceable to the original Indian substrate languages. However, Mesthrie (2004) simultaneously observes that many of the phonetic variants are similar to older (Cultivated) WSAfE values, which may suggest something about the early- to mid-twentieth century when much of the input to ISAfE was transferred from white native speakers, with subsequent isolation due to apartheid legislation.

• Indian SAfE shares the allophonic variation associated with the KIN-PIN “Split” in WSAfE, but in general shows less evidence of glide-loss. Characteristic vowel features include an unrounded RP-like NURSE vowel (i.e., [ɜː], different in this respect to WSAfE, which has [øː]), a GOOSE vowel that tends to be more back than in WSAfE, and a short diphthong in GOAT (in the region of [oʊ] rather than [ɐʊ] or [øʉ], as found in BWSAfE and GWSAfE respectively) (Mesthrie 2004: 956–959). The backer values for GOOSE have been retained by younger speakers even after the advent of a more integrated society (Mesthrie 2010).
• Consonantal features include occasional retroflexion of /t, d, n/, the realization of /f, v/ as [ʊ̥, ʊ] and /θ,ð/ as [t̪,d̪] (i.e., [d̪] for then), and unaspirated voiceless plosives in some environments (Mesthrie 2004: 959–962).

Cape Flats English is likewise characterized by certain conservative values, but shares the KIN-PIN “Split” with WSAfE and ISAfE (Finn 2004):

• Raised vowels, front and back, are characteristic of CfE, becoming more extensive as one moves further away from the prestige variety along the dialect continuum. Woods (1987) observes this for front KIT, DRESS, and TRAP as well as back LOT and THOUGHT. By contrast, he observes that STRUT is lowered to [a]. These features are also characteristic of all dialects of Afrikaans in the Western Cape. Wood (1987) also notes that unstressed vowels are not consistently reduced, but are often realized as peripheral. Finn (2004) points to the prevalence of “Canadian Raising” of PRICE and MOUTH with non-low onsets (i.e., [əɪ] and [əʊ]) in pre-fortis environments.
• Consonant features include an antedental /f/ (lower lip advanced beyond the top teeth), final-nasal elision ([pl] for plan), and /h/ as voiced, i.e., [ɦ], the influence being conceivably both of a historical nature (in terms of language contact) and synchronic (in terms of L1 interference in the case of Cape Vernacular Afrikaans speakers (Finn 2004)).

Black South African English has been studied more extensively than any other nonancestral variety of SAfE. The general picture is one of a number of distinctive vowel and suprasegmental features, attributable to transfer from the native languages. Differences in consonants are fewer, and mainly due to phonotactic and syllabification differences. However, in recent years, there are clear indications of a gradual change in the pronunciation of some BSAfE speakers, in particular the so-called acrolectal group, which regularly interacts with native speakers and other acrolectal speakers of nonancestral varieties. Nevertheless, there is as yet no overriding evidence of large-scale homogenization of upper class Black speakers and native (White, Colored, or post-acrolectal Indian) speakers.

• The picture that emerges from older descriptions of BSAfE vowel contrasts (e.g., Hundleby 1964; Lanham 1966; Adendorff and Savini-Beck 1993) is that the contrast between tense and lax (alternatively long and short) vowels is neutralized, and central vowels tend to be replaced by their closest front vowel alternative. Typical consequences of such mergers include the homophony of pairs like sit and seat (no length contrast) or bird and bed (central vowel replaced by front vowel). This is attributed to the constraint imposed by the Southern Bantu languages, which have five- or seven-vowel systems, but no phonemic length or tense/lax contrast. Van Rooy and van Huyssteen (2000), analyzing a small number of speakers acoustically, still largely confirm this picture, at least as far as monophthongs are concerned.
• Due to the relative absence of central vowels, there is no allophonic WSAfE-like KIN-PIN ‘Split’, although in the results of van Rooy and van Huyssteen (2000) mid-front [ɛ]-realizations for the vowels from the PIN-set occur more frequently than for the KIN-set (even if [i]-realizations remain most frequent for both sets). This finding can be taken to suggest emerging awareness of the allophonic variation in the speech of BSAfE, without yet translating into a consistent articulatory replication.
• There is less agreement as to the realization of diphthongs in the older literature. While some sources simply claim the absence of diphthongs, Hundleby (1964) and Lanham (1966) in particular reported the breaking of diphthongs in bisyllabic sequences through glide insertion, resulting in PRICE being realized as [ajɪ] or MOUTH as [awu]. Van Rooy and van Huyssteen (2000) use acoustic data to show that in CHOICE the diphthongal realization is general, and some gliding is found in MOUTH, but that with the other traditional English diphthongs there is insufficient evidence for anything other than monopthongal realizations. However, drawing on a larger dataset of similar speakers, van Rooy (2004) adds PRICE and GOAT as potential diphthongs in BSAfE.

Research since the early to mid-2000s points to a new group of BSAfE speakers that exists alongside the speakers of older forms of BSAfE. Following a widespread practice in research on New Englishes, these two varieties are termed acrolectal and mesolectal BSAfE respectively. Such research began to observe changes in the speech of black South Africans about a decade after the political transformation of the early 1990s. The relevant observations about new phonetic realizations of vowels are the following:

• Starting with van Rooy (2004), researchers have observed the presence of lax vowels in the phonetic output of acrolect speakers, although at the time van Rooy studied these (drawing on data from 2000 to 2003), the lax vowels were observed for both traditionally tense and lax monophthongs, i.e., FLEECE and KIT both showed the realizations [i] and [ɪ]. Da Silva (2007) also observes the emergence of central vowels in the speech of a subset of BSAfE speakers, especially for the KIT and NURSE vowels. Most recently, Mesthrie (2010), studying speakers at the very top of the socioeconomic spectrum who have been to fully integrated multiracial schools, reports that Black speakers in his sample approximate the WSAfE speakers’ fronted GOOSE vowel to a stronger degree than Colored or Indian speakers of the same socioeconomic status. Like van Rooy (2004), however, Da Silva (2007) still observes a large group of speakers who produce vowels such as [ɛ] for NURSE or [a] for STRUT, thus not showing evidence of a tense/lax contrast.
• Diphthongs seem to have undergone a more extensive change in the speech of BSAfE speakers. Apart from the CHOICE diphthong already reported by van Rooy and van Huyssteen (2000), Da Silva (2007) reports that almost all of her BSAfE speakers share diphthongal realizations of MOUTH and PRICE, while many also have diphthongal realizations in GOAT and FACE.

Like the other varieties of SAfE, there are fewer unique consonantal features in BSAfE. The following features have however been reported in the literature:

• Consonant cluster simplification is attested in complex codas, especially when the syllable has a final plosive following another obstruent. If the following syllable starts with the same or a similar obstruent, the deletion of a coda obstruent is almost categorical. Resyllabification of coda plosives occurs in just less than 50% of the cases where the following syllable has no consonantal onset, but considerably less frequently in acrolect speakers. To a lesser extent, and only in the speech of mesolect speakers, the sonorant /r/ in onset clusters is deleted occasionally (van Rooy 2007).
• Final devoicing, without compensatory lengthening of the preceding vowel, is quite widespread (van Rooy and Wissing 1996, 2001).
• Like most other varieties of SAfE, BSAfE is nonrhotic. It also shows extensive aspiration of voiceless plosives, including in syllable codas (van Rooy 2000).

Suprasegmental features have also received some attention and are often the main target of prescriptivist comment on BSAfE. However, available data pertains only to mesolect speakers:

• BSAfE displays syllable-timed rhythm rather than stress-timed rhythm (Coetzee and Wissing 2007). In consequence, vowel reduction is not particularly prominent (van Rooy 2004). There is some debate as to what the typical realization is of vowels that are otherwise unstressed in WSAfE. Van Rooy and van Huyssteen (2000) find that in many cases the pronunciation is the mid-front vowel [ɛ/e], while low vowels such as [a/ɑ] are also common in final syllables, especially when the final syllable is open. Mesthrie (2005) argues for a more complex system of realizations.
• Stress patterns are different in mesolectal speakers. While there is a strong tendency to stress the penult, as noted by Lanham (1984), van Rooy (2002) finds that a super-heavy final syllable (or even a heavy final syllable) attracts stress to the final syllable, e.g., “realize” with final stress.

Recent developments and research into SAfE

There is growing evidence to suggest that SAfE might be undergoing a process of nascent regionalization, i.e., that speakers in the different English-speaking urban centers of South Africa (Cape Town, Port Elizabeth, Kimberley, Durban, and Johannesburg) are developing their own manner of speaker and indexing regional provenance. This appears to be true both of WSAfE (Bekker 2007; Bekker and Eley 2007; O’Grady and Bekker 2011) and other varieties such as ISAfE and C(f)E (Mesthrie 2010).

Of perhaps greater interest, however, are the linguistic reflexes of the growing racial integration that has taken place since the advent of full democracy in 1994 in South Africa. What integration exists has been mainly the result of a burgeoning black middle class, so it is particularly at this level of the social class continuum that new developments in SAfE have been noted. Van Rooy (2004, 2007) already identifies the presence of new variants in the speech of acrolect speakers, many of which are closer to WSAfE than the variants attested to in older/mesolectal BSAfE. Da Silva (2007), following Horvath (1985), uses a Principal Components Analysis to analyze the accents of students at the University of the Witwatersrand in Johannesburg and provides evidence for various changes within the English used by black individuals. More recently, Hartmann and Zerbian (2009) have shown that while middle class (particularly female) black South Africans often approximate GWSAfE, they are also, it would appear, creating new means for indexing ethnic identity; in this particular case Hartmann and Zerbian (2009) found evidence for neo-rhoticity (GWSAfE being a nonrhotic variety) in the speech of many such subjects. Research currently underway is investigating whether or not young white female South Africans are attempting to emulate their black peers in this regard. Mesthrie (2010) has broadened the investigation to include all ethnic groups (white, black, colored, and Indian) and concludes, in his study of GOOSE-Fronting among young middle class South Africans and with a few “ifs and buts”, that “middle-class, L1 English-speaking South African students of all backgrounds are fronting the GOOSE vowel”; this is a sign of the possible development of a new, deracialized, middle class variety of SAfE.

At the same time, however, there are a number of similar features across the nonancestral varieties, ISAfE, C(f)E, and BSAfE, that may, with mutual reinforcement, remain resistant to convergence with the GWSAfE variety. The presence of a syllable-timed rhythm is reported for both BSAfE and ISAfE. Stress shifts to the right edge of the word are reported for BSAfE (van Rooy 2002), ISAfE (Mesthrie 2004a) and C(f)E (Finn 2004), with relevant examples being realíze, intoxicáted, and participáte. While these authors use different terms, the actual examples they provide show how similar the process is across all three varieties.

Conclusion

South African English was transmitted to South Africa early in the nineteenth century by settlers from predominantly the south-east of England. While koineization was frequently interrupted by new waves of settlement, a stable form of SAfE must have been in place by the first half of the twentieth century. English spread very gradually and slowly to other communities, with elite bilingualism being a very noticeable part of early spread. However, after the introduction of general education by the apartheid government in the early 1950s, two major types of changes took place: the spread of English to other communities was accelerated considerably, even to the point of becoming the home language for the vast majority of the Indian community, but due to the isolation apartheid enforced between communities, distinct ethnolects developed. Only since the early 1990s have the boundaries that kept groups apart been removed, although the majority of especially the black community still lives in segregated areas, with limited contact with other speech communities. For those individuals of the South African community who are in the middle class or otherwise have access to more integrated educational facilities and an integrated workplace, there are early signs that within the first two decades of an open society, some degree of convergence between the various accents can be detected.

REFERENCES

1. Adendorff, R. and Savini-Beck, M. 1993. The teaching of English vowels and consonants in the new South Africa. Journal of Language Teaching 27: 232–248.
2. Beck, R.B. 1997. Monarchs and missionaries among the Tswana and Sotho. In: Christianity in South Africa: A Political, Social and Cultural History, R. Elphick and R. Davenport (eds.), 107–120, Cape Town: David Phillip and Oxford: James Currey.
3. Bekker, I. 2007. Fronted /s/ in General White South African English. Language Matters 38(1): 46–74.
4. Bekker, I. 2009. The vowels of South African English. PhD dissertation, North-West University. Available from: http://dspace.nwu.ac.za/handle/10394/2003.
5. Bekker, I. 2012. South African English as a late 19-century extraterritorial variety. English World-Wide 33(2): 127–146.
6. Bekker, I. and Eley, G. 2007. An acoustic analysis of White South African English (WSAfE) monophthongs. Southern African Linguistics and Applied Language Studies 25(1): 107–114.
7. Booyse, J.J. 2011a. The provision of education during the first half of the 20th century. In: A History of Schooling in South Africa: Method and Context, J.J. Booyse, C.S. le Roux, J. Seroto, and C.C. Wolhuter (eds.), 169–214, Pretoria: Van Schaik.
8. Booyse, J.J. 2011b. Education provision during the period of National Party rule. In: A History of Schooling in South Africa: Method and Context, J.J. Booyse, C.S. le Roux, J. Seroto, and C.C. Wolhuter (eds.), 215–268, Pretoria: Van Schaik.
9. Bowerman, S. 2004. White South African English: phonology. In: A Handbook of Varieties of English: vol. 1: Phonology, E.W. Schneider, K. Burridge, B. Kortmann, R. Mesthrie, and C. Upton (eds.), 931–942, Berlin and New York: Mouton de Gruyter.
10. Branford, W. 1994. English in South Africa. In: The Cambridge History of the English Language, vol. V: English in Britain and Overseas: Origins and Development, R. Burchfield (ed.), 430–496, Cambridge: Cambridge University Press.
11. Branford, W. 1996. English in South African society: a preliminary overview. In: Focus on South Africa, V. de Klerk (ed.), 35–51, Amsterdam: Benjamins.
12. Bughwan, D. 1970. An investigation into the use of English by Indians in South Africa with special reference to Natal. PhD dissertation, University of South Africa.
13. Coetzee, A.W. and Wissing, D. 2007. Global and local durational properties in three varieties of South African English. Linguistic Review 24: 263–289.
14. Coetzee-van Rooy, S. and van Rooy, B. 2005. Labels, comprehensibility and status in South African English. World Englishes 24(1): 1–19.
15. Conradie, C.J. 1986. Taalgeskiedenis [Language History], Pretoria: Academica.
16. Da Silva, A.B. 2007. South African English: a sociolinguistic investigation of an emerging variety. PhD dissertation, University of the Witwatersrand. Available at: http://wiredspace.wits.ac.za/handle/10539/4955.
17. De Klerk, V. 1999. Black South African English: where to from here? World Englishes 18: 311–324.
18. Elphick, R. 1997. Introduction: Christianity in South African history. In: Christianity in South Africa: A Political, Social and Cultural History, R. Elphick and R. Davenport (eds.), 1–15, Cape Town: David Phillip and Oxford: James Currey.
19. Finn, P. 2004. Cape Flats English: phonology. In: A Handbook of Varieties of English: vol. 1: Phonology, E.W. Schneider, K. Burridge, B. Kortmann, R. Mesthrie, and C. Upton (eds.), 964–984, Berlin and New York, Mouton de Gruyter.
20. Giliomee, H. and Mbenga, B. 2007. New History of South Africa, Cape Town: Tafelberg.
21. Hartmann, D. and Zerbian, S. 2009. Rhoticity in Black South African English. Southern African Linguistics and Applied Language Studies 27(2): 135–148.
22. Hartshorne, K. 1992. Crises and Challenge: Black Education 1910–1990, Cape Town: Oxford University Press.
23. Hirson, B. 1981. Language in control and resistance in South Africa. African Affairs 80: 219–237.
24. Hodgson, J. 1997. A battle for sacred power: Christian beginnings among the Xhosa. In: Christianity in South Africa: A Political, Social and Cultural History, R. Elphick and R. Davenport (eds.), 68–88, Cape Town: David Phillip and Oxford: James Currey.
25. Horvath, BM. 1985. Variation in Australian English: The Sociolects of Sydney, Cambridge: Cambridge University Press.
26. Hundleby, C.E. 1964. Xhosa-English pronunciation in the South-East Cape. PhD dissertation, Rhodes University.
27. Jeffery, C. 1982. Review of The Standard in South African English and its Social History (Varieties of English Around the World: General Series Volume 1), by L.W. Lanham and C.A. Macdonald. Heidelberg: Julius Groot Verlag, 1979. Folia Linguistica Historica 3(2): 251–263.
28. Jeffery, C. and van Rooy, B. 2004. Emphasiser now in colloquial South African English. World Englishes 23(2): 269–280.
29. Kaplan, M. and Robertson, M. 1991. Introduction. In: Founders and Followers: Johannesburg Jewry, 1887–1915, M. Kaplan and M. Robertson (eds.), 10–16. Cape Town: Vlaeberg.
30. Lanham, L.W. 1966. Teaching English in Bantu Primary Schools. Final Report on Research in Johannesburg Schools. Publication No. 4 of the English Academy of Southern Africa, Johannesburg: English Academy of Southern Africa.
31. Lanham, L.W. 1984. Stress and intonation and the intelligibility of South African Black English. African Studies, 43(2): 217–230.
32. Lanham, L.W. and Macdonald, C. (1979) The Standard in South African English and its Social History, Heidelberg: Julius Groos Verlag.
33. Lass, R. 1995. South African English. In: Language and Social History: Studies in South African Sociolinguistics, Rajend Mesthrie (ed.), 89–106, Cape Town: David Philip.
34. Lass, R. 2002. South African English. In Language in South Africa, R. Mesthrie (ed.), 104–126, Cambridge: Cambridge University Press.
35. Lass, R. 2004. South African English. In: Legacies of Colonial English: Studies in Transported Dialect, R. Hickey (ed.), 363–386, Cambridge: Cambridge University Press.
36. Lass, R. and Wright, S. 1986. Endogeny vs. contact: “Afrikaans influence” on South African English. English World-Wide 7(2): 201–223.
37. Malan, K. 1996. Cape Flats English. In: Focus on South Africa, V. de Klerk (ed.), 125–148, Amsterdam: Benjamins.
38. McCormick, K. 2002. Language in Cape Town’s District Six, Oxford: Oxford University Press.
39. Mesthrie, R. 1992. English in Language Shift: The History, Structure and Sociolinguistics of South African Indian English, Cambridge: Cambridge University Press.
40. Mesthrie, R. 1995. Language change, survival, decline: Indian languages in South Africa. In: Language and Social History: Studies in South African Sociolinguistics, R. Mesthrie (ed.), 116–128, Cape Town: David Philip.
41. Mesthrie, R. 2002a. Endogeny versus contact revisited: aspectual busy in South African English. Language Sciences 24: 345–358.
42. Mesthrie, R. 2002b. From second language to first language: Indian South African English. In: Language in South Africa, R. Mesthrie (ed.), 339–355, Cambridge: Cambridge University Press.
43. Mesthrie, R. 2004. Indian South African English: Phonology. In: A Handbook of Varieties of English: vol. 1: Phonology, E.W. Schneider, K. Burridge, B. Kortmann, R. Mesthrie, and C. Upton (eds.), 953–963, Berlin and New York: Mouton de Gruyter.
44. Mesthrie, R. 2005. Putting back the horse before the cart: The “spelling form” fallacy in Second Language Acquisition studies, with special reference to the treatment of unstressed vowels in Black South African English. English World-Wide 26(2): 127–151.
45. Mesthrie, R. 2010. Socio-phonetics and social change: deracialisation of the GOOSE vowel in South African English. Journal of Sociolinguistics 14(1): 3–33.
46. Mphahelele, E. 1985. Prometheus in chains: the fate of English in South Africa. English Academy Review 2: 91–104.
47. Ndebele, N.S. 1987. The English language and social change in South Africa. English Academy Review 4: 1–16.
48. O’Grady, C. and Bekker, I. 2011. Dentalisation as a regional indicator in General South African English: an acoustic analysis of /z/, /d/ and /t/. Southern African Linguistics and Applied Language Studies 28(1): 77–88.
49. Roberge, P.T. 1994. On detecting a prior linguistic continuum in Cape Dutch. In: Nuwe Perspektiewe op die Geskiedenis van Afrikaans [New Perspectives on the History of Afrikaans], G. Olivier and A. Coetzee, 153–165, Johannesburg: Southern Books.
50. Schneider, E. 2007. Postcolonial Englishes: Varieties Around the World, Cambridge: Cambridge University Press.
51. Statistics South Africa. 2012. Census 2011: Census in Brief, Pretoria: Statistics South Africa, Report 03-01-41.
52. Trudgill, P. 2004. New-Dialect Formation: The Inevitability of Colonial Englishes, Edinburgh: Edinburgh University Press.
53. van Rooy, B. 2000. The consonants of Black South African English: current knowledge and future prospects. South African Journal of Linguistics Supplement 38: 15–33.
54. van Rooy, B. 2002. Stress placement in Tswana-English: the makings of a coherent system. World Englishes 21(1): 145–160.
55. van Rooy, B. 2004. Black South African English: phonology. In: A Handbook of Varieties of English: vol. 1: Phonology, E.W. Schneider, K. Burridge, B. Kortmann, R. Mesthrie, and C. Upton (eds.), 943–952. Berlin and New York, Mouton de Gruyter.
56. van Rooy, B. 2007. Consonant clusters and resyllabification in Black South African English. Language Matters 38(1): 26–45.
57. van Rooy, B. and van Huyssteen, G.B. 2000. The vowels of Black South African English: current knowledge and future prospects. South African Journal of Linguistics Supplement 38: 35–54.
58. van Rooy, B. and Wissing, D. 1996. Degrees of neutralization during syllable-final devoicing: evidence from second language phonetics. South African Journal of Linguistics Supplement 33: 77–98.
59. van Rooy, B. and Wissing, D. 2001. Distinctive [voice] implies regressive voicing assimilation. In: Studies on Distinctive Feature Theory, T.A. Hall (ed.), 295–334, Berlin: Mouton.
60. Wasserman, R. 2014. Modality on Trek: modality in South African English. PhD dissertation, North-West University. Available at: http://dspace.nwu.ac.za/handle/10394/13042.
61. Wells, J. 1982. Accents of English, Cambridge: Cambridge University Press.
62. Welsh, F. 1998. A History of South Africa, London: HarperCollins.
63. Wood, T.M. 1987. Perceptions of, and attitudes towards, varieties of English in the Cape Peninsula, with particular reference to the “Coloured Community”. MA dissertation, Rhodes University.