Language shift or maintenance? Factors determining the use of ...

Text-only Preview

Stellenbosch Papers in Linguistics, Vol. 38, 2008, 49-72

Language shift or maintenance? Factors determining the use of
Afrikaans among some township youth in South Africa

Charlyn Dyers
University of the Western Cape, Private Bag X17, Bellville 7535, South Africa
[email protected]


The aims of this paper are as follows:

• to investigate how high school learners in a township school in South Africa report on
their use of, and attitudes toward, their first language, Afrikaans; and
• to use the findings to show how, given their particular situation, the language
continues to be strongly maintained in this community, and remains an important
index of both their group and individual identity.

Much has already been written about the dominant role of English in South Africa today (cf.,
e.g., Ridge 2000, 2004; Webb 2002; Alexander 2004), and the perception that there is a large-
scale shift towards the use of English at the expense of other mother-tongues and home
languages (cf. De Klerk 2000; Kamwangamalu 2003). This large-scale shift, however, may
either not be taking place or be taking place at a much slower rate in many communities of
practice in South Africa. It is the contention of this paper that this shift is far more marked in
middle-class and upwardly mobile black and mixed-race families in South Africa, and that
various factors work against this shift in working class and rural communities. Some of these

50 Charlyn Dyers
factors include (i) the environment in which these people live, (ii) a desperately burdened and
poorly-functioning state education system in poorer rural and urban areas (Soudien 2007;
Taylor 2007), (iii) historically low levels of literacy, and (iv) relatively low mobility for the
poor and working class in post-democratic South Africa.

At the same time, one also needs to take account of the factors that support the maintenance
of home languages and mother tongues in South Africa. Central to these maintenance factors
must surely be the role of language in defining people's ethno-cultural identities, in an era
where belief in a common South African identity is still a vague dream, and ethnic/tribal
identities have re-emerged as a central anchoring point for many groups like the Afrikaners,
Xhosas and Zulus. The focus of the present work is on a particular working-class township
where 75% of the population are members of the most diverse ethnic group in South Africa,
the mixed-race "Coloureds".1 A number of sociolinguistic studies have been carried out in this
township, namely Wesbank, which forms part of Greater Cape Town in South Africa. These
studies have looked at the language attitudes of high school learners (Dyers 2004), the
peripheral normativity in the literacy skills of such learners (Blommaert, Muyllaert,
Huysmans and Dyers 2005), and the degree of multilingualism among some of these learners
(Dyers 2008). The interest in the township stems from its position as one of the first racially
integrated housing developments in post-democratic South Africa.

This paper continues the longitudinal exploration reflected by the above studies of how
"Coloured" South African high school children in Wesbank negotiate their individual and
collective identities through language. The longitudinal study aims to find out how high
school children from diverse backgrounds and marginalised by poverty, location and race
reflect a sense of individual, collective and national identity through language.


There can be little argument against Appel and Muysken's (1990:23) contention that "the
identity imposed by one's group membership is a crucial factor for language choice". The
population of South Africa is made up of many diverse groups speaking different languages,
and the diversity of identities was engineered to be regarded as even more separate and
distinct both by 300 years of colonialism and by the apartheid regime of more than 40 years

Language shift or maintenance? Factors determining the use of Afrikaans 51
(1948-1994), whereby there was separate development for all races under white rule. As a
language, Afrikaans has suffered historically because of its association with the 40-odd years
of apartheid. It was the language of the rulers of this period, the Afrikaners, and became
known during the years of struggle against this system as "the language of the oppressor".

At the same time, however, Afrikaans is the mother tongue of the majority of the group which
was classified as "Cape Coloured" during the apartheid era in South Africa. The majority of
this group reside in the Western Cape Province of South Africa, where 55% of the population
speak Afrikaans as mother tongue, followed by 23% whose mother tongue is Xhosa and 19%
whose mother tongue is English (South African Population Census 2001). The group also
displays varying degrees of bilingualism in Afrikaans and English (from "mainly English" in
some suburbs to "mainly Afrikaans" in others), as well as multilingualism in Afrikaans,
English and Xhosa, depending on their levels of literacy, education and location. Other
languages that play smaller roles within this group are Arabic (particularly in the Muslim
community) and German (which was a popular school subject in the previous dispensation).
However, despite the close identification of many members of this community with
Afrikaans, there are also signs of a shift towards English, particularly among middle-class
"Coloured" families (Malan 1996; Anthonissen and George 2003; Warner 2008). The children
in these families are educated in English, and despite strong exposure to Afrikaans in the
family and community environment, tend to use only English in conversation with others.

It is indisputable that the "Coloured" population of South Africa participated in the
development of Afrikaans out of the Dutch spoken by the first European colonists, who
arrived in South Africa in 1652. Many of the "Cape Coloureds" are descendants of South
Africa's earliest inhabitants, the San and the Khoe, as well as unions of members of these
tribes with the European settlers and people from Asia and other parts of Africa who were
brought to South Africa as political prisoners and slaves by the Dutch, French and English
colonists. Cape Dutch (later Afrikaans) developed out of the need for these extremely diverse
linguistic and ethnic groups to communicate (Malan 1996: 127).

Language plays an important role in defining who we are, and makes us instantly recognisable
to other members of our particular speech community. As Joseph (2006: 39) puts it, "we read
the identity of people with whom we come into contact based on very subtle features of

52 Charlyn Dyers
behaviour, among which those of language are particularly central". Despite the shift to
English in the middle class families reported above, most "Coloured" people in the Western
Cape continue to speak Afrikaans as their mother tongue, and the language remains a key
component of their ethnolinguistic identity (Titus 2008). The sociopolitical history of South
Africa has led to the "Coloureds" developing as a group with a particular identity which sets
them apart from white Afrikaners who share their language (Stone 1995: 277-281). Unlike the
latter group, they have never displayed the same "emotional investment in keeping the
language pure" (McCormick 1989: 206). But the majority identify closely with the vernacular
variety of Afrikaans which they use every day, regarded by McCormick (2000), Malan (1996)
and others as a mixed code, which incorporates many English loanwords. During a focus
group interview carried out in the third year of the longitudinal study of Wesbank teenagers,
one girl had the following to say about the variety of Afrikaans spoken in the township:

Os praatie yntlik regte Afrikaans nie. Os praat Kaapse Afrikaans, Engels en Afrikaans
(We don't actually speak proper Afrikaans. We speak Cape Afrikaans, a mixture of
English and Afrikaans.)

For this informant, her variety of Afrikaans was not proper, standard Afrikaans, but an
informal mixture of Afrikaans and English. According to Hendricks (personal
communication, 2006), it would be more accurate to use the term "Kaaps" (reportedly first
used by the writer Peter Snyders) for this variety instead of the more commonly used "Cape
Flats Afrikaans", as there are variants within this variety. These variants will display either
more or less code-mixing with English as one travels from the centre of Cape Town to its
suburbs, peripheral townships and surrounding rural areas. The two extracts from the focus
group interviews in year three of the study in (2) and (3) are typical examples of the
spontaneous, unmarked conversational code-mixing prevalent in Wesbank. (Code switches
are marked by the use of italic script for English and regular script for Afrikaans.)

(2) S1:
Well… met my ma is ek meer relaxing, like ek en my ma praat, ek sal my ma alles
(Well…I'm more relaxed with my mother. When my mother and I chat, I will tell her

Language shift or maintenance? Factors determining the use of Afrikaans 53
S2: Ja… maar my ouma wat nou oorlede is... Ek en sy was eintlik close aanmekaar.
(Yes… but my grandmother who died recently… She and I were actually close to each

The linguistic convergence of English and Afrikaans in Cape Town and other parts of South
Africa is visible in the local varieties of both languages (McCormick 1995: 203). In addition,
location, education and role models all have a role to play in the varieties of vernacular
Afrikaans spoken by "Coloured" people all around South Africa, and there are a number of
studies of these different varieties (e.g. Combrinck 1978; Webb 1989; Roberge 1995;
Hendricks 1996).

Further evidence of the close identification with Kaaps is provided by the strong attachment
of many "Coloureds" to the works of poet and playwright, Adam Small, who writes in
vernacular "Cape Flats" Afrikaans, the immense popularity of the music of rap and hip-hop
artists who use it (e.g. Brasse vannie Kaap, Prophets of Da City), and the success of theatrical
productions using the vernacular, like Joe Barber (Petersen, Isaacs and Reisenhöfer 1999) and
Suip! (Petersen and Reisenhöfer 1999). As Stone (1995: 280) puts it, "the dialect is beloved
by its speakers as the sacramental marker of communal membership and a vehicle of intimacy
and love". While its speakers acknowledge its low status in relation to standard Afrikaans,2 it
is fair to say that it enjoys a certain status as well as strong vitality in the poor, working-class
townships of the Cape Flats. This is a large area on the periphery of Cape Town, which
includes Wesbank Township, to which the "Coloured" population of Cape Town was forcibly
moved at the height of the apartheid regime.

Wesbank Township, which has only been occupied since 1999, is by all accounts a peripheral
township marked by poverty, unemployment and crime (Dyers 2004). The current
demographics of the area (Havenga 2006) show that, of a total population of approximately
29 000 people, approximately 73% are mixed-race, mainly Afrikaans-speaking "Coloured"
people and 25% are Xhosa people, many of whom have recently migrated to the Western
Cape Province of South Africa from the Eastern Cape Province, where this ethnic group
constitutes the majority population. A further 2% are White, Asian or foreigners from other
parts of Africa like Somalia, Nigeria and The Congo. The township consists of small housing
units, a high school, three primary schools and a supermarket. There are no public amenities

54 Charlyn Dyers
such as a community centre, parks or sports fields and those who work mainly travel to the
more prosperous suburbs of Cape Town by minibus taxi, which is the most common and
cheapest form of public transport.

The question therefore arises: How does this new space organise people and their patterns of
language use? The issue of space and multilingualism has been investigated extensively by
Blommaert, Collins and Slembrouck (2005), whose overarching interest is in how "people are
positioned and [in] the communicative potential they display and have attributed to them in
diverse, scale-sensitive situations and practices” (Blommaert, Muyllaert, Huysmans and
Dyers 2005: 210). Their research has led them to conclude that multilingualism is not what
people have, or do not have, but what their environment enables or disables them to use. As
the study by Blommaert, Muyllaert, Huysmans and Dyers (2005) has shown, such enabling
and disabling environmental forces are also present in Wesbank.

Terms of reference

One of the most common definitions of language shift is that it takes place when the younger
members of a minority speech community no longer speak the language of their parents, but
speak a dominant majority language instead. The language of the parents is therefore not
passed on to the next generation. Conversely, language maintenance occurs when a language
continues to be used across all generations despite the presence of other languages also being
used by a community – the kind of stable diglossia defined by Fishman (1972). According to
Myers-Scotton (2006: 89), two generalisations may be drawn from all studies on language
shift and maintenance, namely –

• There is "always a combination of factors at work" supporting either shift or
maintenance; and
• In a bilingual community, patterns of maintenance and shift can be measured on a
continuum with some individuals using only the first language (L1) at one end of the
continuum and others using only the second language (L2) at the other end. In
intergenerational shift, for example, we might find the older members of a family
using the L1 (but having some competence in the L2), while the children, despite

Language shift or maintenance? Factors determining the use of Afrikaans 55
having almost perfect comprehension of the L1, speak only the L2 (cf. also Warner

In addition, Myers-Scotton (2006: 90) lists the following societal, in-group and individual
factors as being among those factors central to language maintenance:

• demographic factors – large numbers of speakers of the same L1 living together;
• occupational factors – working with fellow speakers of the L1, with restrictive socio-
economic mobility;
• educational factors – e.g. official provision of the L1 as a medium of instruction;
• social networks and group attitudes about the L1 as an ethnic symbol; and
• psychological attachment to the L1 for self-identity.

The above factors are certainly present in the predominantly Afrikaans-speaking "Coloured"
community of practice living in Wesbank. I use the term "community of practice" here to
distinguish it from the wider concept of 'speech community' and to show that differences in
language practices exist in different "Coloured" communities. For Lanza (2007) this concept
captures the reality that even smaller groups can have their own ways of speaking, acting and
believing. Although the "Coloured" inhabitants of Wesbank may have migrated to the city
from various urban and rural communities, most members of this community of practice
experience no difficulty in communicating in Afrikaans with one another. Their children
attend the township schools, where Afrikaans is taught as L1 and is one of the official media
of instruction (the other being English, principally to support large numbers of Xhosa learners
also attending these schools). In most of the churches in the township, Afrikaans also plays a
leading role, although a switch to English may occur when non-Afrikaans speakers are
present in the congregation. Poverty, unemployment and low levels of literacy further restrict
the socio-economic mobility of this working-class community. As a particular space
organising people's linguistic choices, the township therefore appears to play an enabling role
in the maintenance of Afrikaans. Historically, too, Afrikaans has always played a significant
role in the group identity of the "Cape Coloured", constructed as this identity might be for
some. It is perhaps the main marker of a "Cape Coloured" identity, particularly in the absence
of a clear group culture and identity (Dyers 2004: 31), given the huge diversity of origin in
this group. This possibility was borne out by the data on attitudes towards Afrikaans collected

56 Charlyn Dyers
for the present study, which revealed a psychological attachment to the language for self-
identity as well as group identity.

Thim-Mabrey's (2003) distinction (cited in Schmidt 2006:15) between linguistic identity on
the one hand and identity through language on the other is of relevance here. Linguistic
identity means the features of a given language which distinguishes it from other languages,
but also the identity of a person with regard to his/her – or in fact any – language. Identity
through language, on the other hand, describes "the identity of persons insofar as it is
constituted or co-constituted through language and language use".

However, the combination of factors supporting either shift or maintenance may also
influence individual language choices. An individual may be part of a community of practice
where one language is strongly maintained, but may for personal and subjective reasons
reveal different patterns of language use and attitude to those of the dominant group. In other
words, individual identity in terms of language could in some circumstances differ from group
linguistic identity. As Tabouret-Keller (1998) contends (cited in Mills 2005), identity is both a
social construct, characterised by objective features (such as language), and a personal,
subjective construct, characterised by individual mental processes and choices. Individual
choices in terms of language can therefore be far more varied than those of the group to whom
the individual belongs, depending on his/her particular circumstances and environment. An
individual's linguistic choices can –

• reflect almost exactly those of the group to which the individual belongs;
• show degrees of overlap, but with varying degrees of difference, e.g. sounding
more/less accurate than the source group or exhibiting greater use of code-switching
and experimentation with the language; or
• reveal varying degrees of rejection of his/her group's dominant language/s – this could
include increased borrowing of terms from a preferred language, adopting a way of
speech that makes the individual sound different to the rest of the source group, e.g.,
more/less polished, or refusing to speak the language despite having grown up with it,
like the middle-class "Coloured" children in Warner's (2008) study.

Language shift or maintenance? Factors determining the use of Afrikaans 57
In the Wesbank survey, most of the individual linguistic choices of the respondents fitted into
either the first or the second category listed above. There were also individuals who were
multilingual in Afrikaans, English and Xhosa as a result of growing up in multilingual
families which use all three languages (Dyers 2008). However, even in these multilingual
families, Afrikaans played a dominant role in their intimate domains. On the whole it is
probably fair to say that, with the majority of the respondents, there were only finely-nuanced
shades of difference between group and individual identity through Afrikaans.

Among the "Coloured" learners at Wesbank High, there were a few exceptions to the
Afrikaans-dominant group. We identified a small percentage (2%) of "Coloured" learners
who defined themselves as English, rather than Afrikaans-speaking. They were in the school's
English medium classes which were attended by a majority of Xhosa-speaking learners. On
closer inspection, we found that while a small number were recent arrivals from other parts of
the country, like Johannesburg, the rest lived in the middle-class suburb of Highbury, not far
from Wesbank. This tended to confirm our perception that English, as an instrument of socio-
spatial mobility (cf. Blommaert, Muyllaert, Huysmans and Dyers 2005), was largely
associated in the minds of our respondents with the more up-market spaces outside the
confines of Wesbank. The case study of John in section 5 bears out this association.


I began my longitudinal research at Wesbank High School in 2004 with the assistance of
student research assistants, electing to work with high school learners for two reasons, namely –

• as one type of language shift occurs when young members of a community no longer
speak their parents' L1, these learners would be able to provide evidence of such a
trend; and
• high school learners would be better equipped than primary school learners to
articulate viewpoints on often-subjective language issues.

This paper is based on data collected from Afrikaans L1 learners in Grades 8 to 10 in the first
three years of our research (2004-6) at Wesbank High School. The data consist of classroom
responses to a questionnaire on their use of Afrikaans in different domains (see table 1),

58 Charlyn Dyers
written responses on the importance of the language in their lives, focus group and individual
interviews, as well as informal playground observation. The classroom and written responses,
as well as some individual interviews, were collected in the first two years. The respondents
were 70 Grade 8 (first year of high school) and 34 Grade 9 learners. In the third year, the
focus shifted to individual and focus group interviews with 12 selected Grade 10 learners, as
well as informal playground observations and conversations.

English and
Afrikaans only
English, Xhosa and
Domains of use
English only
*with classmates
N = 70 (2004)
** with teachers
N= 37 (2005)
Gr 8
Gr 9
Gr 8
Gr 9
Gr 8
Gr 9
Gr 8
Gr 9
16% 0 0 14% 84% 3% 0
37% 41% 9% 5% 20% 54% 9% 0
School (in class)

70% 0 0 17% 24% 3% 5%
80% 38% 6% 5% 6% 57% 3% 0
68% 0 0 17% 32% 3% 0
62% 0 0 9% 38%
3% 0
Table 1.
Domains of use over two years (2004-2005) for Grade 8 and 9 learners

Findings and analysis

Reported language use in different domains
I begin here with the analysis of language use in different domains, as reported by the Grade 8
and 9 Afrikaans L1 learners. Domains of language use are certain institutional contexts in
which one language or language variety is more likely to be considered appropriate than
another. Fishman (1965, 1968, 1971) did some of the principal research into domains of
language use, and was mainly concerned with the question: who speaks what language to
whom and when? Domains are taken to be constellations of factors such as location, topic and
participants, and include the domains of work, family, school and other educational
institutions, circle of friends and wider communication. As situations change, so do the
choices of language, variety and register. A key influence in both formal and informal
domains is the relationship between the interlocutors (Dyers 2008). Our examination of the