Skip to main content Accessibility help
×
Home

Information:

  • Access
  • Cited by 1

Actions:

      • Send article to Kindle

        To send this article to your Kindle, first ensure no-reply@cambridge.org is added to your Approved Personal Document E-mail List under your Personal Document Settings on the Manage Your Content and Devices page of your Amazon account. Then enter the ‘name’ part of your Kindle email address below. Find out more about sending to your Kindle. Find out more about sending to your Kindle.

        Note you can select to send to either the @free.kindle.com or @kindle.com variations. ‘@free.kindle.com’ emails are free but can only be sent to your device when it is connected to wi-fi. ‘@kindle.com’ emails can be delivered even when you are not connected to wi-fi, but note that service fees apply.

        Find out more about the Kindle Personal Document Service.

        Language and the geopolitics of (dis)location: A study of Zimbabwean Shona and Ndebele speakers in Johannesburg
        Available formats
        ×

        Send article to Dropbox

        To send this article to your Dropbox account, please select one or more formats and confirm that you agree to abide by our usage policies. If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your <service> account. Find out more about sending content to Dropbox.

        Language and the geopolitics of (dis)location: A study of Zimbabwean Shona and Ndebele speakers in Johannesburg
        Available formats
        ×

        Send article to Google Drive

        To send this article to your Google Drive account, please select one or more formats and confirm that you agree to abide by our usage policies. If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your <service> account. Find out more about sending content to Google Drive.

        Language and the geopolitics of (dis)location: A study of Zimbabwean Shona and Ndebele speakers in Johannesburg
        Available formats
        ×
Export citation

Abstract

The Zimbabwean diaspora is a well-documented phenomenon. While much research has been done on Zimbabwean migration to South Africa, the role that language plays in this process has not been well researched. This article draws on South African census data and qualitative fieldwork data to explore the manner in which Zimbabwean migrants use languages to appropriate spaces for themselves in the City of Johannesburg. The census data shows that African migrants tend to concentrate in the Johannesburg CBD, and fieldwork in this area reveals that Zimbabwean migrants are particularly well established in two suburbs—Yeoville and Hillbrow. The article explores migrant language repertoires, which include English, Shona, Ndebele, and a variant of Zulu. While many contributions to the migration literature tend to assume a strong association between language and ethnicity, the article shows how this relationship is mediated by geographic location and social positioning within the city. (Language, migration, Johannesburg, South Africa, Zimbabwe)*

Footnotes

Gugulethu Siziba died on 14 February 2017. A website commemorating his life and work can be found at http://www.gugulethusiziba.org/. The final draft was adapted by the second author.

*

This article draws on Ph.D. research (Siziba 2013) that was funded by a Stellenbosch Graduate School scholarship and a Next Generation Social Sciences in Africa Fellowship, funded by the Social Science Research Council. The South African Census-based research formed part of a project—Language and Urban Social Space (2014–2016)—funded by the National Research Foundation (NRF) of South Africa (Grant 91563). We would like to thank Stephanie Rudwick and an anonymous reviewer for comments made on the initial draft of this article.

INTRODUCTION

Among African immigrants in South Africa, the presence of Zimbabweans is particularly marked. This phenomenon is well documented, but despite a growing literature on ‘the Zimbabwean diaspora’, very few articles have given serious attention to the manner in which Zimbabweans use language to negotiate their outsider status in various receiving states. As the title suggests, we explore the manner in which Zimbabwean Shona- and Ndebele-speaking migrants use language and the manner in which the social positioning of language repertoires affects their subjective sense of location and dislocation in Johannesburg.

Zimbabwe has been described as a ‘nation in motion’ and the first section of the article is therefore dedicated to an exploration of the literature on Zimbabwean emigration and diaspora. Particular attention is paid to the manner in which language has been treated in this literature. We argue that in much of the prevailing literature, Shona and Ndebele have tended to be used as relatively inert ethnonyms. Our effort to deepen the analysis of the relationship between language and migration has both empirical and theoretical objectives. Empirically we explore patterns of ‘otherization’ on both a macro and a micro scale. The macro analysis takes the form of a census-based study of the distribution of ‘other’ language speakers in Johannesburg, focusing particular attention on the official census categories of ‘isiNdebele’ and ‘other’. It is commonly assumed that the ‘other’ category alone measures ‘foreign’ or ‘non-South African’ languages, and we explain why this assumption is misleading. The analysis of census data is followed by a more fine-grained qualitative analysis of the experiences of Zimbabwean migrants in five neighbourhoods.

Building on the preceding discussion of neighbourhood and city-wide language trends, our concluding analysis focuses more explicitly on our theoretical objective: to show how the capacity to use language (or rather specific language repertoires) in the process of in-group construction is mediated by social and geographic position. Drawing on Bourdieu's (1983, 1986, 1991) work—and specifically his (1985) distinction between ‘nominal classes’ and ‘probable classes’—we show how Zimbabwean migrants’ positioning affects their capacity to mobilise language repertoires and thereby to constitute themselves as an effective social group.

A NATION IN MOTION: ZIMBABWE AS A SYNONYM FOR MIGRATION, DIASPORA, AND ‘THE FOREIGN’

Migration is nothing new to Zimbabweans. Indeed, it is part and parcel of Zimbabwe's history, and, as some have noted, by 1980—the year of Zimbabwe's independence—migration had become thoroughly institutionalized (Mlambo & Raftopoulos 2010; Pasura 2011:148). There is a long history of migration, dating back to precolonial times, between the territories that constitute Zimbabwe and South Africa (Sisulu, Moyo, & Tshuma 2007; Mlambo & Raftopoulos 2010). The two countries share very similar colonial histories: both formed part of the British Empire; both emerged as mining-based industrial economies with labour importing regimes; and both were European settler colonies that evolved elaborate forms of racial segregation and discrimination. The more recent migration histories of the two countries have nevertheless diverged sharply. For while South Africa has remained the dominant regional economy and a net-migration importer, Zimbabwe's profile has shifted from an intercalary position—as both a receiver and sender of migrants—to a position marked by significant levels of emigration.

The notion of a ‘Zimbabwean diaspora’ and of Zimbabwe as a state renowned for exporting its people is now well-established in contemporary global migration discourse. Crush, Chikanda, & Mwaswikwa (2012:4), for instance, state that ‘Zimbabwe has become a major global migrant sending country over the past two decades’. Commenting on the unrelenting crises confronting the country, other scholars argue that, in the absence of political change, Zimbabwe‘s ‘biggest export will remain its people’ (Solidarity Peace Trust, in Mlambo & Raftopoulos 2010:6). Crush & Tevera (2010:1) argue that ‘Zimbabwe has now joined the list of “crisis-driven” migrations’, which include crises in Angola, the Democratic Republic of Congo (DRC), Rwanda, Somalia, and Sierra Leone. These observations succinctly capture how, in the current context, narratives about Zimbabwe tend to focus on migration and population displacement. And these narratives conjure an image of Zimbabwe as a nation in motion.

The precise number of people who have left Zimbabwe is contested (Crush et al. 2012). For reasons of historical and cultural proximity, the literature on Zimbabwean migration has tended to focus on the United Kingdom and South Africa. Zimbabweans are nevertheless globally distributed. Crush and colleagues (2012:4) state that ‘[i]n 2001, 192 of the 222 countries reported in the UN Migration Stock database had at least one Zimbabwean migrant’. Emigration peaked in the years after 2000, and hence references to ‘the Zimbabwean crisis’ commonly refer to this period. Hammar & Raftopoulos (2003) have, however, argued that the Zimbabwean crisis is not a single crisis, but rather multiple crises that date back to a few years after independence.

While acknowledging that Zimbabwe's crisis was ‘rooted in the long-term structural political–economic legacies of colonial rule’ (Mlambo 2010:2), more recent triggers of the crisis have been located in the context of a ‘major threat’ to the political future of the ruling party, ZANU-PF (Mlambo & Raftopoulos 2010:2). The late 1990s and the years after 2000 have been widely described as a context in which the law was suspended and replaced by ZANU-PF fiat. The exodus of Zimbabwean migrants must therefore be set against the backdrop of a general breakdown in law and order, economic meltdown, infrastructural decay, and state-sanctioned violence. During this period the Zimbabwean economy operated according to a ‘Kukiya-kiya logic’ 1 (Jones 2010), in terms of which Zimbabweans resorted to unorthodox—and often unruly—strategies to make a living in an abnormal situation. Tendai Biti—a member of the opposition MDC-T party and the Minister of Finance in the 2009–2013 Unity government—has described the economy during this period as ‘ginyanomics’. Ginya is a Shona slang word meaning ‘force’ and ginyanomics therefore refers to the use of force and coercion to maintain a semblance of a proper functioning economy (for example, by using threats of arrest to force industries to lower their prices).

Given the post-2000 domestic context, marked by escalating socioeconomic decline, Zimbabweans who were able to move out of the country did so. Those with relatively scarce skills were sought after in Anglophone countries such as Britain, Canada, Australia, and South Africa. Movement to the United Kingdom intensified around the late nineties and continued thereafter (Mbiba 2005; Pasura 2010, 2011). Although there are no exact numbers of Zimbabweans in the UK, references to the UK as ‘Harare North’ (Mbiba 2005) speak to the large numbers of Zimbabweans. The literature also suggests that Zimbabwean migrants in the UK are better educated when compared to migration streams to other parts of the globe (Bloch 2008). Social class would therefore seem to have facilitated migration to the UK (Bloch 2008; McGregor 2007). Given proximity, porous borders, and affordability when compared to other popular destinations, the social profile of migration to South Africa has been more complex. Makina (2010) notes that in the decade of 2000–2010 most Zimbabweans migrating to South Africa headed for Johannesburg.

Zimbabwean migrants have a pronounced presence in contemporary South Africa and a lot of research has been done to try and understand Zimbabwean experiences of—and in—South Africa. Crush, Chikanda, & Tawodzera (2012:4) have noted a ‘flurry of research’ stimulated by the mass movement of Zimbabweans into South Africa. Much of this research frames the Zimbabwean migrant experience within the discourse of a generalized xenophobia—seen as the overriding logic informing how foreign migrants, and particularly black Africans, experience South Africa. Zimbabwean migrants are, like their other-African counterparts, reconstituted in South Africa as amakwerekwere—a derogatory label that many South Africans use to refer to African foreigners. The term is putatively onomatopoeic (like babbler in English), signifying that the languages spoken by foreign Africans are indecipherable babbling. However, this appellation can also be seen as a denial of voice to amakwerekwere, or what Morreira (2007:434) terms the ‘displacement of voice’. Zimbabwean migrants appear to be a unique migrant category, as—over and above the long history of migration that they share with other African migrants—they also share cultural and historic ties with South Africa (Polzer 2008; Muzondidya 2010; Worby 2010).

Language stands out as one of the most salient cultural artefacts that Zimbabwean migrants use to navigate and negotiate the politics of identity confronting their status as amakwerekwere in South Africa. Language is salient for a number of reasons. First, it has been identified by researchers as one of the markers used both by the general South African population and state agents (notably the police and officials from the Department of Home Affairs) to identity foreigners (Morris 1998; Nyamnjoh 2006). The boundary-marking capacity of language was graphically displayed during the May 2008 ‘xenophobic attacks’, when (mostly Zulu) shibboleths were used to identify amakwerekwere (Hassim, Kupe, & Worby 2008). Second, although most studies of Zimbabwean migrants in South Africa do not take language as the central object of analysis, their arguments often centre on language as a primary vehicle or resource—which Zimbabwean migrants use to negotiate contexts of inclusion and exclusion. While migration studies have contributed to our understanding of the lived experiences of Zimbabwean migrants in South Africa, many have been premised on simplistic assumptions about language.

The literature on migration in southern Africa tends to employ relatively underdeveloped notions of ‘language’. Typically, generic language categories are treated as ‘groups’ and issues associated with sociolinguistic group formation are elided by means of implicit assumptions about the correspondence between ‘language’ and ‘ethnic identity’. This is a problem with deep roots in the institutionalization of linguistics as an autonomous discipline, where—as Beck (2016) has noted—the term ‘language’ tends to mask a

reified language-ethnicity-territory nexus [which] enables consequential ascriptions and valuations of difference highly functional until today, not only in the sense of Africa as Europe's ‘other’, excluded from the project of modernity, but also in Africa with regard to the production of knowledge, identity politics, the formation of elites, and the allocation of resources in colonial and post-colonial administration for nation building, citizenship, political participation and education.

With respect to Zimbabweans in South Africa, this orientation is evident in the tendency to assume a sharp ethnic division between Shona speakers and Ndebele speakers, and correspondingly distinct strategies associated with language deployment. Zimbabwean Ndebele forms part of the Nguni language continuum, the largest category of related languages in South Africa (discussed below). It is therefore often argued that Ndebele-speaking migrants tend to assimilate into local Nguni-speaking communities (and notably Zulu-speaking communities), while Shona-speaking migrants are presented as unable to assimilate, except among the Venda-speaking communities of the northern Limpopo province where they are said to have a linguistic advantage over their Ndebele-speaking counterparts (Muzondidya 2010; Ndlovu 2010; Sibanda 2010; Worby 2010). Shona-speaking migrants are therefore presented as the archetypal amakwerekwere, existing on the margins of South African society.

What these studies ignore is that people are not equally endowed with any given ‘language’ (Bourdieu 1991; Blommeart 2005). Generic language categories tend to mask the fact that most South Africans and migrants from the region have complex language ‘repertoires’. Furthermore, studies that fail to address the issue of repertoire variation typically treat ‘geographic space’ as a relatively simple and inert backdrop, against which migrants deploy ‘languages’ as more or less universal currencies. The notion of ‘repertoires’ breaks from this abstract conceptualization of language as a uniform resource, presumed equally available to all users and unaffected by domains of use. It ‘encompasses both the “means of speech” available to speakers, and the “speech economy” these speakers participate in’ (Johnstone & Marcellino 2010:4). Blommaert & Backus (2011:2) underscore the utility of the concept in analyzing empirical linguistic exchanges noting that it ‘belongs to the core vocabulary of sociolinguistics’. According to Hymes (1996:33) ‘a repertoire comprises a set of ways of speaking. Ways of speaking in turn comprise speech styles, on the one hand, and contexts of discourse, on the other, together with relations of appropriateness obtaining between styles and contexts’.

The concept of ‘repertoire’ chimes well with Bourdieu's concept of ‘habitus’, whose analytical utility depends on a relational way of thinking that locates cultural agency within a specific ‘field’ of practice (Bourdieu & Wacquant 1992). Focusing specifically on language, Bourdieu (1991:17) notes that ‘[t]he fact that different groups and classes have different accents, intonations and ways of speaking is a manifestation, at the level of language, of the socially structured character of the habitus’. Bourdieu (1991:67) goes on to assert that ‘[u]tterances receive their value (and their sense) only in relation to a market, characterized by a particular law of price formation’. At issue therefore is the extent to which geographic ‘places’ (at different scales) can be reconceptualised as social spaces (‘fields’ in Bourdieu's terminology) with varying degrees of ‘agency’ (explored through notions of ‘habitus’), that is, that have an constraining effect on the manner in which migrants ‘choose’ to deploy their language repertoires. ‘Field’ is the key spatial metaphor in Bourdieu's work and our analysis proceeds with a discussion of the City of Johannesburg as a complex stratified space in which Zimbabwean migrants’ status as ‘the Other’ varies according to their position within the city-wide field and their capacity to access and redefine specific places.

In terms of this approach, we need to examine the bundles of linguistic resources that Zimbabwean migrants have at their disposal, and the manner in which these are mobilized within different contexts and within the two ostensibly discrete ‘ethnolinguistic groups’. Furthermore we have to ask ourselves: what is the functional utility of these repertoires in different interactional spaces and domains? The language repertoires available to different migrants and their functionality in different interactional contexts mean that they ‘are restricted as to what they can do with and in language’ (Blommaert 2005:13). Notably, individuals’ language competencies do not have to be equal, that is, proficiency may vary across these resources. To this end, a number of scholars speak of ‘truncated multilingualism’ to capture the functional utility of individuals’ language repertoires that range from full competence to mere abstract recognition (Pillai 2002; Blommaert, Collins, & Slembrouck 2005).

Given that Zimbabwean languages are not officially recognized in South Africa, census statistics can tell us little about these languages. However, as discussed below, the South African census language category ‘isiNdebele’ is ambiguous, as it includes both South African and Zimbabwean communities. Census data can therefore give us a sense of the location and distribution of Zimbabwean immigrants in Johannesburg. The functional utility of Zimbabwean repertoires is explored in the final section, which is based on participant observation and interviews conducted in selected suburbs of Johannesburg.

THE GEOPOLITICS OF ‘OTHERIZATION’ IN JOHANNESBURG

This article is primarily concerned with the sociolinguistic experiences and urban survival strategies of Zimbabwean migrants in Johannesburg. It is, however, necessary to situate these experiences within the wider context of post-1994 sociolinguistic ‘otherization’. A fundamental feature of the post-1994 political dispensation—as codified in the 1996 Constitution—was the official recognition of eleven languages. The official languages are Afrikaans, English, isiNdebele, isiXhosa, isiZulu, Sepedi (also referred to as Sesotho sa Leboa), Sesotho, Setswana, siSwati, Tshivenda, and Xitsonga (SA Constitution, 1996). As official categories, language and race were closely intertwined in the discriminatory logic of the colonial and apartheid periods. Christopher (2004:146) notes that the first postapartheid South African census in 1996 was also the first to formulate ‘a uniform language question for the entire population regardless of race’.

The decision to extend official status to nine indigenous African languages was therefore a core component of the wider process of extending citizenship to formerly excluded South Africans. But this formal or de jure extension of citizenship within the new republic should be distinguished from the de facto capacity to enjoy the benefits of citizenship. Citizenship is not a state-wide given. In the twenty years following the inauguration of democracy, it has become increasingly apparent that even the most formal aspects of citizenship (so called ‘negative’ freedoms) are contingent upon official status and geosocial positioning. ‘Otherization’ is the process of legitimating this contingency, or validating categories of exclusion. In this section we provide a brief overview of sociolinguistic otherization, based on social positioning within the City of Johannesburg.

The City of Johannesburg is a relatively new entity. It is one of eight metropolitan municipalities—or ‘metros’—established in terms of the Municipal Structures Act of 1998. 2 Table 1 draws on 2011 census data to provide a brief profile of the eight metros. These metros are sociolinguistically complex focal points within the post-1996 domain of local governance. The table provides a basic sense of this complexity, by listing the top two to four language categories measured by the census.

Table 1. Metropolitan municipalities in South Africa.

Situated below the level of the national state and the provinces, the metros nevertheless constitute important political entities, which—by virtue of their demographic complexity and fiscal capacity—enjoy a significant level of socioeconomic autonomy. They also constitute dynamic but highly stratified new spaces. Over the last twenty years, high levels of international migration and regional in-migration have complicated, but not erased, the architecture of inequality that dates back to the colonial and apartheid periods. Competition for material and symbolic resources is fierce, and an indication of sociolinguistic contestation is evident in the nomenclature presented in Table 1 above. The older territorial associations with English (Cape Town, Port Elizabeth, East London, Durban) and Afrikaans (Johannesburg, Pretoria, Bloemfontein) names remain, but overlaying these are new administrative footprints with new labels. Four of these new metro names reflect the dominant African language spoken in the metros: Tshwane (Pretoria), Ekurhuleni (the eastern region the Witwatersrand), Mangaung (Bloemfontein), and eThekwini (Durban).

The City of Johannesburg is the most populous of the eight ‘Category A’ or metropolitan municipalities in South Africa. The current metropolitan boundaries and administrative capacities are relatively recent (Municipal Structures Act, 1998). The eastern region of the old Witwatersrand is now the separate metropolitan municipality of Ekurhuleni. Together with the City of Tshwane (Pretoria), these metros constitute the urban heartland of Gauteng—South Africa's most industrialized province and the economic hub of southern Africa. This region has a long history of in-migration from surrounding regions and neighbouring states (notably Mozambique and Zimbabwe). Since 1994 it has also increasingly drawn migrants from the central and western regions of the continent. The three Gauteng metros are therefore ethnolinguistically more diverse than the other five metros, and a crude sense of this can be gleaned from the ‘major language’ column in Table 1 above.

Census data on language is crude in that it provides an overall sense of the official language categories that people recognise and claim as their ‘household language’. But to the extent that the census questions on language privilege the home domain, they overstate linguistic homogeneity and reinforce common-sense notions of relatively discrete ‘home languages’ or ‘mother tongues’. Unlike the 2001 census, the 2011 census asked for two languages, and these provide a better approximation of sociolinguistic complexity at different levels. The question on language reads as follows: ‘Which two languages does (name) speak most often in this household?’

While nationally 52% of respondents claim to speak a second language, in Gauteng and Johannesburg this figure increases to 69%. Moreover, many South African residents have at least a rudimentary command of three or more official languages, and there are two main reasons for this. First, within the group of nine indigenous African languages that received national recognition in 1994, there are two large dialect continua or groupings of mutually intelligible languages. Thus, isiZulu, isiXhosa, isiNdebele, and siSwati constitute the Nguni languages, while Sesotho, Setswana, and Sepedi are referred to as the Sotho languages. Second, in the metros we see a significant level of ‘language contact’, or language acquisition, that typically takes place as a result of interaction outside of the home domain.

The official language categories therefore mask as much as they reveal about sociolinguistic complexity in South Africa, and this is particularly true of Johannesburg. One important aspect of the problem is that five of the official African languages—or rather the names used to designate them—are also recognised by neighbouring states. These include Setswana in Botswana, Sesotho in Lesotho, siSwati in Swaziland, and isiNdebele and Tshivenda in Zimbabwe. IsiNdebele is particularly relevant to this article.

‘IsiNdebele’ is ambiguous to the extent that it is used to refer to two relatively distinct language communities: one in Zimbabwe and one in South Africa. In Zimbabwe ‘isiNdebele’ speakers constitute between sixteen and twenty percent of the population, 3 and are concentrated in the southwestern part of the country that covers the Midlands and Matabelelend provinces. In South Africa the 2011 census puts the number of isiNdebele speakers at approximately 1.1 million, or about 2.1% of the population. The largest concentrations of isiNdebele speakers are located in the provinces of Gauteng and Mpumalanga. While Shona speakers in South Africa are submerged under the residual language category ‘other’, the distribution of Zimbabwean isiNdebele speakers can to some extent be tracked by means of the census ‘isiNdebele’ attribute and the citizenship variable. Table 2 therefore presents census data for the four official language categories that serve to contextualize the subsequent discussion of Zimbabwean migrants in Johannesburg. In this table, ‘yes’ indicates South African citizenship.

Table 2. Language and citizenship (2011 Census, Statistics South Africa).

Table 2 shows that among respondents who claimed a nonofficial language (‘Other’) as their first ‘household language’ in 2011, about 70% did not have citizenship. There is little variation in this figure when one shifts from national level to provincial or city level. But the equivalent figures for isiNdebele are interesting: at national level noncitizens make up 10%, but this figure increases to 24% in Gauteng and 55% in the Johannesburg Metro. This contrasts with the more or less constant figures (10% or less) for isiZulu and English—the two largest language groups in Johannesburg.

An analysis of the language data collected during the 1996, 2001, and 2011 censuses reveals that ‘isiNdebele’ speakers have grown more than any of the other official language categories. This growth is in fact too rapid to be explained in terms of internal population growth. It is therefore clear that the census figures conflate two relatively distinct language populations: South African and Zimbabwean isiNdebele speakers. In-migration therefore explains both the national increase in isiNdebele speakers—between 1996 and 2011—and the particular concentration of noncitizen isiNdebele speakers in Johannesburg. The 2011 census data also reveal that Johannesburg isiNdebele speakers who answered ‘no’ to the citizenship question tend to be concentrated in a number of areas: Hillbrow, Berea, Diepsloot, Ivory Park, Yeoville, Cosmo City, and the Central Johannesburg subplace. 4 As the following table, maps, and fieldwork account suggest, these are clearly Zimbabwean isiNdebele speakers, living in suburbs that have become focal points for African migrants.

Table 3 presents all of the subplaces with noncitizen populations of 5,000 or more. These areas are then mapped twice: Figure 1 shows subplaces in the City of Johannesburg as a whole, while Figure 2 focuses on inner-city Johannesburg.

Figure 1. Distribution of noncitizens in the City of Johannesburg.

Figure 2. Distribution of noncitizens in the City of Johannesburg CBD.

Table 3. Johannesburg subplaces with large concentrations of noncitizens.

Census data—when disaggregated at both city and subplace level, and combined with field interview data—provide a useful means of exploring the relationship between social space and language repertoire. Subplaces are particularly useful, because they are the smallest named territories in South African census geography. People attach contrasting significances to these names, so, while the names and territorial boundaries of these places persist over time, their ethnic associations do not. Thus, given widespread assumptions about the language-ethnicity-territory nexus (Beck 2016), an analysis of census data at the level of the neighbourhood or ‘subplace’ can form the basis of a more nuanced depiction of the relationship between languages and territories.

The following discussion of Zimbabwean (dis)location in Johannesburg focuses specifically on two constellations of subplaces. The first is the ‘old CBD’, where the focus falls specifically on the suburbs of Hillbrow, Yeoville, and Newtown. The second is the Diepsloot-Fourways nexus on the northern fringe of the city. The inner city suburbs and Diepsloot subplaces are relatively poor regions of the city, and the term ‘old CBD’ reflects the intercalary status of the Johannesburg CBD. This historically white core is still a hub of economic activity, but ‘white flight’, and the out-migration of affluent residents and commercial activity more generally, have coincided with the emergence of a new financial hub, Sandton, to the north (Selzer & Heller 2010). A study of migrant language repertoires is therefore simultaneously a study of movement across shifting racial and socioeconomic geographies.

ZIMBABWEAN (DIS)LOCATION: SHONA AND NDEBELE SPEAKERS IN JOHANNESBURG

In this section we explore the situated language repertoires of Zimbabweans living in Johannesburg. The focus falls on two regions in the city: one in central Johannesburg and one straddling wealthy and marginalized areas in the north. In the Johannesburg CBD we look at the suburbs of Yeoville, Hillbrow, and Newtown. In northern Johannesburg we explore Zimbabwean transit between the impoverished Diepsloot region and the wealthy Fourways suburb. The discussion draws on sociological fieldwork conducted by the first author in 2012 (Siziba 2013), which included participant observation and interviews with seventy-four Shona- and Ndebele-speaking migrants. As we show below, the repertoires of these migrants include Shona, Ndebele, English, and a local convergent ‘sociolect’ that the respondents refer to as isiZulu Semaflethini ‘Zulu of the flats’. In order to facilitate the analysis of the relationship between social spaces and repertoires, the suburbs are discussed in descending order of subjective ‘situatedness’—or from ‘location’ to ‘dislocation’.

The situatedness of Zimbabwean migrants in Johannesburg can best be captured by a ‘low hovering over data in order to give expression to the communality of voices across participants’ (Anderson 2007:1). Research participants across the five neighbourhoods consistently evoke the notion of ‘home’, that is, Zimbabwe, in articulating a broad set of experiences associated with ‘location’ and ‘dislocation’ in Johannesburg. Yeoville and Hillbrow in particular emerge as spaces of ‘location’, in which the structuring of power and the general outlook on the neighbourhoods permit a reproduction of ‘home’ away from home. Compared to Yeoville and Hillbrow, Newtown is constructed and presented as somewhat elite and exclusionary, while life in Diepsloot is viewed as the embodiment of dislocation from ‘home’. The nature and value of migrant repertoires vary across these spaces, and this variation reflects diverse meanings and attachments to place.

Yeoville: A cosmopolitan home away from home

“Hapana kana problem tinotaura Shona. Hapana anoti what what” ‘There is no problem. We speak Shona all the time here and there is no-one who can object’. This carefree attitude expressed by Moses, 5 a vegetable vendor at the Yeoville market, articulates a sentiment shared by many—if not most—Zimbabweans in the suburb. In Yeoville, Moses’ ‘we’ signifies Shona speakers in the first instance, but the sentiment also resonates with other African migrants. Yeoville, Berea, and Hillbrow have a dense concentration of foreign African immigrants, with Zimbabweans and Nigerians particularly prevalent. Pholile confidently asserts that much of the inner city area that includes Yeoville, Berea, and Hillbrow constitutes ‘Zimbabwean colonies’ and ‘playgrounds’. This label that is evoked by Zimbabwean migrants conveys the particular concentration of Zimbabwean Shona- and Ndebele-speaking migrants in this region of the CBD, and this concentration is particularly evident in Yeoville. But the label also belies the ethnic complexity and cosmopolitan character of these suburbs. The market in which Moses conducts his business reflects this, and it is worth focusing on the human presence that animates it. The market is a large structure that has a number of vending stalls or bays where vendors are able to display their wares. There are further subdivisions, with vegetable sellers like Moses located at the front part of the market. The outer part—occupied by the vegetable sellers—encases an inner part of the market. In these sections can be found people selling different wares: there are Congolese women selling lotions and other assortments of cosmetics; there are people selling West African cuisine; there are cobblers and tailors, some of whom are ‘Ndebele’ speaking; and there are also Shona-speaking women selling Zimbabwe's staple of thick porridge, meat, and chomolia (isitshwala in Ndebele, sadza in Shona—a type of vegetable that is very common). My ‘talk’ with Moses constitutes one of many conversations in different languages. Our conversation takes place in the market, in Moses’ stall. In between our conversation, Moses speaks to another vendor in a stall that is some distance from his. Their ‘talk’ occurs by way of intermittent shouting in Shona. This does not seem to bother anyone in the vicinity. Around us the market is buzzing with people speaking different languages. People walk around and inquire about the prices of the different products on sale.

In this diverse environment Zimbabweans are adept at negotiating difference and similarity. A key component of this process is the use of ‘code-switching’ and ‘code-mixing’, or what one respondent described as “making do”. Code-switching is frequently used to accommodate different interlocutors (e.g. a switch to English), but it frequently also marks an intra-Zimbabwean status differential, to the extent that Ndebele speakers tend to shift to Shona when addressing Shona speakers. There is therefore, notwithstanding sociolinguistic complexity, an unmistakeable ‘reproduction of Zimbabwe’ in Yeoville. Vuliwe believes that foreigners are in the majority here and Zimbabweans are the most dominant. She states that: “There are too many of us [Zimbabweans] here. It's almost as if it's us only [who are here]”. In addition to the Zimbabwean languages, the Zimbabwean presence is manifested in the remnants of posters advertising Zimbabwean political parties, elections, and meetings on walls in the area.

Moses communicates in Shona, English, and ‘Ndebele’ depending on his interlocutor. Our dialogue began in Shona, but with time I learned more about his repertoire. While clearly most comfortable in Shona, he nevertheless claimed to speak good Zulu—which he claimed was similar to Ndebele. In reality his Ndebele was quite rudimentary and heavily Shona-accented. Despite this, Moses’ bundle of linguistic resources—his repertoire—suffices in Yeoville, where the linguistic market is characterized by multiple languages and where no overarching South African language is hegemonic. On a different occasion I watch an elderly woman, on a pavement opposite Shoprite, competing with other vendors for business. She attempts to catch the attention of a young woman walking by and calls out, “Hallo… hallo makoti” ‘Hallo Hallo bride’. Later when I speak to Mai Chisi and ask her how she communicates with people in Yeoville she states, “Tinongo janukajanuka” ‘we just make do’, which is a reference to her being able to master just a few bits and pieces of ‘Zulu’ and ‘English’. Mai Chisi states that she is proficient in Shona, but does not have a good command of English, Zulu or any other South African languages.

Yeoville typifies an image of Johannesburg as a place of ‘colliding worlds’, ‘divisions’, and ‘fragments’ (Beall, Crankshaw, & Parnell 2002; Bremner 2004; Murray 2011). This is typified by various migrant communities (both from within and without South Africa), diverse nationalities, languages, occupations, relationships, as well as forms of sociality. Sikhumbuzo, an Ndebele-speaking migrant, offers a very useful description of Yeoville, which resonates with other migrant narratives. He asserts “I can describe Yeoville as a cosmopolitan place because there are people from all over the world. I can't just say all over Africa because we have Pakistanis and Jews doing business here”. In this cosmopolitan environment there is a discernible Zimbabwean character, and this manifests most obviously in the use of English, Ndebele, and Shona as local lingua francas. A somewhat different situation is evident in the nearby suburb of Hillbrow.

Hillbrow: Bulawayo-2 and intra-Zimbabwean tensions

Hillbrow shares Yeoville's congestion, with streets full of people selling different wares. Like Yeoville, Hillbrow has various symbolic markers of Zimbabweaness, which again suggest the prominence of a Zimbabwean community. These markers include the vegetables (chomolia), amacimbi ‘mopane worms’, umfusha ‘dried vegetables’ and other foodstuffs that respondents associate with food from ‘home’. It is relatively easy to access Zimbabwean cuisine in the cafes in Hillbrow. In many ways Hillbrow mirrors the situation in Yeoville. There is a dense concentration of foreign African migrants, whom respondents claim outnumber ‘locals’ (referring to South Africans). There is also a thriving presence of informal trade visible in different spaces, including shop pavements. Along Pretorius and Claim Street, Zimbabwean women engage in hairstyle enterprises (street salons). Nigerian men approach and flag you down as you walk by: “Hey, my broda wh'at you want? I can help you now?” The heavy Nigerian-accented English marks them, as does their attire. These are some of the visible marks of foreignness, knitted together to produce a complex and multilayered tapestry. Respondents assert that Nigerians and Zimbabweans are the most dominant groups in Hillbrow, although there is mention of Cameroonians, Malawians, and Congolese, among other nationalities. Focusing specifically on the Zimbabwean presence, Pholani asserts “In fact, I can say Yeoville all the way to Berea and Hillbrow, it's a Zimbabwean colony. It's our playground”. Zephaniah echoes a common refrain, stating that “[In Hillbrow] There are a lot of foreigners and a few locals. There are mostly Zimbabweans and Nigerians here. But I think Zimbabweans are the largest group. There are a lot of Ndebeles and Kalangas”. 6 Another resident, Samukile, states “Hillbrow is Bulawayo-2. 7 It is full of people from home. Shonas are there but they are not that many here. A lot of them are in Yeoville and Rossetenville. I'm not really sure why they don't like Hillbrow… maybe they are scared of Hillbrow”.

Phathisa asserts that Hillbrow is “Bulawayo's townships of Njube, Nkulumane, Mpopoma, etc.”. Violence is seen as one aspect that informs the ethnic texturing of the Zimbabwean presence in Hillbrow (Siziba 2013) . The respondents’ narratives suggest that the fear of violence and mugging informs the ethnic composition of the Zimbabwean population in Hillbrow. Prince evokes the notion of Gukurahundi 8 to describe ethnic divisions in Hillbrow, a suburb he views as the equivalent of Bulawayo's townships. He asserts: “Ah in Hillbrow, it's Gukurahundi. You know what happened during Gukurahundi, right?” Prince's Gukurahundi metaphor can be read in different ways. On the one hand, it simply refers to the reversal of the exclusion Ndebele people encounter in Zimbabwe. On the other hand, it may refer to ethnic tensions between Ndebele- and Shona-speaking people. In this way, Zimbabwe's ethnic cleavages and divisions are reproduced in Johannesburg. These are difficult to authenticate but they speak to the existence of ethnic tensions: of imaginaries and fantasies that are extensions of prejudices and narratives from ‘home’. These tensions are mediated by language in complex ways, as discussed below.

Newtown: Pronounced cosmopolitanism and an exclusionary outlook

It is worth discussing Zimbabwean presence in Newtown by way of comparison with the two ‘Zimbabwean colonies’. According to the 2011 census the populations of Hillbrow and Yeoville were 74,131 and 18,884 respectively (Census 2011). Population densities were 69,000 and 19,000 people per square kilometre respectively. At this time, Newtown's population was estimated to be 2,505, while the population density was placed at 2,900 people per square kilometre (Census 2011). Both Hillbrow and Yeoville consist predominantly of flats or apartments. Hillbrow, however, has more high-rise apartments when compared to Yeoville, and this accounts for the very high population density. There is a general consensus among respondents in Newtown, both resident and nonresident, that the area is not as crowded or as noisy as Hillbrow and Yeoville. The residential spaces in Newtown are anchored in a cultural precinct, which is widely known for its heritage, cultural, and art facilities. These include places such as ‘Museum Africa, the Market Theatre, Kippies jazz and several performing arts and music organizations’ (Delmont 2004:34). Vuliwe of Yeoville classifies Newtown as an area that is quite different from Yeoville. He claims that Newtown typifies esisaladini, which means an ‘abode for those who eat salads’, and commonly refers to places associated with affluence. Vuliwe asserts that such areas are characterized by individuality and a lack of the communality which typifies Hillbrow and Yeoville.

Newtown is also characterized by a cosmopolitan, multiethnic and multinational population. NaNcenga, a female Ndebele-speaking migrant, describes the population in her complex as follows: “There are mostly black African people who speak different languages such as Zulu, Xhosa, as well as Ndebele and Shona.” In addition to these people, she notes that there are also many Nigerians who run small businesses. Nkululeko, by contrast, notes that “there are Sotho, Zulu, Venda and Nigerian guys, as well as Shonas and Ndebeles. It's a mixture of different African people”. Nkululeko states that Newtown has very few Zimbabwean people compared to Hillbrow and Yeoville. He attributes this to the fact that access to and use of space are structured differently. In Newtown there are fewer residential spaces, but these are also expensive compared to Hillbrow and Yeoville. In Hillbrow and Yeoville there is a pervasive innovative use of space. Flats are partitioned in order to accommodate larger numbers of people. This practice allows people to share space as well as rentals. In Newtown, by contrast, space is more tightly regulated by means of formal leases. While Newtown also forms part of the inner-city area, it is structured in a more or less similar fashion to closed communities elsewhere in the city. Furthermore, rentals in Newtown are higher. At the time of the research, respondents paid between R3,500 and R4,000 per month (or roughly $412–470 at the time of the fieldwork in 2012) depending on the size of the unit, which was on the lower end of the spectrum. The initial rental instalments including deposit ranged between R14,500 and R15,000. Flats are situated securely behind walls or security fences and access is controlled, that is, the research had to produce an identity document and be signed in and out. In Hillbrow and Yeoville, such mechanisms of regulating entry are generally absent.

Respondents imagine and construct Newtown as inhabited by social categories that are, to a great extent, different from those in the ‘colonies’. Mbadza utilizes classifications and categories based on ‘economic class’. He asserts “the fact that what qualifies one to be here is one's ability to service rentals regardless of one's identity produces a certain type of community. Although people will be people there is at least a certain level of predictability with regards to people's behaviour”. Nancenga in a way echoes Mbadza's perspective, but also provides another dimension of community. She asserts that the forms of relationships, social ties, and sociality are loose and not centred on a single or coherent form of identification. She opines that people in Newtown—including South Africans—are all sojourners. She states “even the South Africans here do not have much in common. They come from different places and are in Johannesburg pursuing different things. As such I think it's difficult for them to even conceive of themselves as one common group”.

Johannesburg is home to a significant number of ‘internal’ and ‘international’ migrants (see Landau 2006; Landau & Freemantle 2010), who follow different social and economic trajectories. Newtown is presented as somewhat exclusive, when compared to Hillbrow and Yeoville. While Hillbrow and Yeoville are convenient points of entry into Johannesburg for newcomers, Newtown tends to be exclusive. High rentals are the most obvious feature of this exclusion, but patterns of language use also set this suburb apart from other regions of the inner city—as the discussion below demonstrates.

Fourways-Diepsloot: Navigating extreme dislocation

The Fourways-Diepsloot nexus proved to be a laboratory for studying the relationship between the Zimbabwean presence in Johannesburg and the shifting connection between space and language. While Fourways was initially selected to access affluent Zimbabweans, these proved to be elusive and fieldwork in Fourways eventually focused on migrants from Diepsloot engaged in informal activities in Fourways.

Respondents in Diepsloot assert that Diepsloot typifies ekasi ‘township’ and is thus different from both the ‘colonies’ of Yeoville and Hillbrow, and Newtown. First, migrants reveal that Diepsloot does not have the same densities of foreigners as the other three neighbourhoods. This impacts on the migrants’ ability to mobilize and secure accommodation-based networks and the enclave-like lifestyle of the ‘colonies’. Instead, in Diepsloot foreigners have to negotiate with ‘local’ landlords to secure accommodation. Respondents do not experience the same degree of freedom to speak their languages in the open, when compared to the residents of Yeoville and Hillbrow. The notion of ‘playground’ or ‘colony’ does not apply and is not evoked in Diepsloot. For the majority of respondents the density of ‘locals’ and the fact that this is a volatile, under-policed area, which is prone to violence and vigilante justice, makes them feel insecure. Consequently, language and the negotiation of identity is a major priority.

Nomalanga describes Diepsloot as characterized by a strong South African identity—a kasi ‘township’ identity—which is evident in the behaviour, manner of dress, and language usage. She states “most black South Africans in Diepsloot hate speaking English”. Ndebele and Shona-speaking respondents therefore negotiate identity in similar ways. Some hide and pass for Zulu where possible: navigating the public spaces in the area and using taxis entails negotiating contact and situations by means of various forms of avoidance and the maintenance of distance. Others converge or attempt to speak local languages in order to placate locals and make relations amicable.

Moving from Diepsloot to Fourways, however, exposes migrants to a different field of practice, which depreciates the value of Sepedi and isiZulu and frees Zimbabwean migrants from the strictures of bodily performances built around working-class township South Africanness. Given the limitations of space, we briefly draw on the experiences of female respondents working in a car wash in Fourways. Ntombi captures the car wash women's experiences in Fourways, where most of their clients speak English and are generally described as ‘whites’. She states that “English is the most important language here at work because we use English for everything. All our business is conducted in English, for example, asking a white man for business. You cannot speak Zulu [to the white man] but you have to use English. Even when you cannot speak another language you can speak English”. The car wash has male managers who are both Zimbabwean and South African. It is strategically located at the Fourways Mall, which according to Nomhle means that “most of our clients speak English and we communicate with them in English”.

Township, suburb, and inner-city orders of indexicality

Zimbabweans in Johannesburg constitute a complex ‘community’, with diverse aspects of division such as age, ethnicity, gender, educational qualifications, occupations, religion, sexual orientation, and political beliefs. Here we have primarily focused on differences associated with location in Johannesburg, and we end with a city-wide discussion of the relationship between location and language repertoire. The suburbs discussed above suggest a more general patterning with respect to the location and dislocation of Zimbabwean migrants in Johannesburg. The inner-city ‘colonies’ of Yeoville and Hillbrow constitute the epicentre of subjective ‘location’ for Zimbabwean migrants. These suburbs manifest specifically Zimbabwean ‘orders of indexicality’ (Blommeart 2005), or places dominated by Zimbabwean markers of belonging and authenticating ‘enough-ness’.

A number of issues can help us to think through the nature of the Zimbabwean communities in Yeoville and Hillbrow. While quite a few professionals live in these suburbs—notably teachers, business people, researchers, and university students—the majority of residents are semiskilled and have informal types of employment. The density and diversity of foreign African migrants in this area structures the linguistic market in a particular way: allowing the widespread use of migrant repertoires and sanctioning English as the bridging language. Migrant repertoires in Yeoville (and to a lesser extent in Hillbrow) are unmarked, because no significant black South African language predominates in these areas.

Respondents in Yeoville and Hillbrow also note the use of a local sociolect, which they call isiZulu semaflethini ‘Zulu of the Flats’ and which is marked as different from the Zulu spoken by South African residents in Diepsloot, Soweto, and other ‘townships’. The notion of ‘Zulu of the flats’ suggests the extent to which this variant is associated with a foreign community in a specific area. IsiZulu semaflethini is described by respondents as a variety of Zulu that is different from that of locals with regards to intonation, rhythm, vocabulary range, and other stylistic features. This variety tends to distinguish foreigners from the township locals—and particularly home language Zulu speakers—whose language is termed isiZulu sekasi or ‘Zulu of the township’.

While isiZulu semaflethini facilitates communication, it also complicates ethnic divisions within the Zimbabwean communities. This is particularly evident in Hillbrow, which is seen as an Ndebele enclave with Shona people restricted to certain spaces. In Yeoville isiZulu semaflethini also tends to mark ethnic lines. Vuliwe, for instance, reveals that she stays with people who speak her language and she does not speak Shona because “my teeth will rot”. When migrants describe isiZulu semaflethini, they often use it interchangeably with isiNdebele, but are also quick to point out it is neither Ndebele nor isiZulu: “it is Zulu of the flats”. While ethnic subjectivities asserting ‘us’ and ‘them’ exist, these often blur as Ndebele- and Shona-speaking migrants are forced to live together. Sikhumbuzo, for instance, says that he does not mind staying with any Zimbabwean but hates what he perceives to be ‘the Shona’ mindset of domination. He states “now they are forced and struggle to learn Zulu, yet in Zimbabwe, you will not hear them speaking Ndebele”. By contrast, many Shona-speaking migrants reveal a desire to learn other languages, a situation some note is difficult in Zimbabwe, given the dominance of Shona and a Shona-centred habitus. In general, the colonies allow migrants to reproduce the comforts of home, manifested in Zimbabwean social networks, living arrangements, food, and other aspects of culture.

Newtown reveals continuities and overlaps with Hillbrow and Yeoville, notably in the role that English plays as neighbourhood lingua franca. Madumbe describes Newtown as a space that displays continuity with neighbouring Braamfontein and other places in the CBD. She refers specifically to shopping malls in the city, which are characterized by an ‘English normativity’, because of the diversity and relative status of the people who frequent these spaces. As in the colonies, here we also see much code-switching and code-mixing. The city centre is the midpoint in the gradation of English use in Johannesburg: white English-speaking cultural dominance in the northern suburbs (such as Fourways) contrasts with both the cosmopolitan inner-city suburbs and the ‘locations’ or ‘townships’ on the city's eastern fringe. The Zimbabweans interviewed in the north worked in Fourways but could not afford to live there. They depict Fourways as a white space where English predominates.

At night their return to Diepsloot corresponds with a shift across a significant indexical boundary. Here the use of English is marked in a linguistic market dominated by black South African languages—Sepedi and to a lesser extent isiZulu—and Zimbabwean migrants describe Diepsloot as radically different from the inner-city suburbs. Unlike Yeoville—where difference is accommodated, and diversity is negotiated through the use of English—in Diepsloot respondents assert that people “do not like to be made to speak English”. Instead there is a heavy investment in local languages, and a concomitant disdain of ‘other’ languages. Sepedi (or Northern Sotho) is the most widely spoken local language, but isiZulu also serves as a lingua franca. Takunda's situation, feeling compelled to speak in the ‘local’ languages, captures the general position of Zimbabwean migrants in Diepsloot: “When I speak to locals in the local languages it's something which makes me feel good. It's like if you cannot speak a local language it's very difficult to stay with these people. As such you have to sacrifice to speak these languages so that you are well placed in society. It helps when you can speak a local language and the first thing is safety”.

In township spaces Zimbabwean repertoires are devalued and migrants are consequently compelled to learn local languages or maintain a low profile. The presence or absence of Ndebele in the repertoires of Zimbabwean migrants tends to be associated with contrasting strategies for passing as unmarked residents in townships or ‘local’ spaces. On the one hand, Zimbabweans with a command of isiZulu semaflethini (predominantly Ndebele-speaking migrants) tend to adopt a strategy of ‘cross-identification’ (Siziba 2015), which is to say they appropriate ‘Zuluness’ when among South African interlocutors who are unable to distinguish the ‘Ndebele’ and ‘Zulu’ varieties. On the other hand, many Shona-speaking migrants rely on bodily comportment, limited interaction, and silence—withholding their marked linguistic resources—as a means of blending unobtrusively into public spaces (Siziba 2016).

CONCLUSION

In this article we have explored how language influences patterns of migrant location in Johannesburg and how social spaces structure language use. In the first part we noted how African migrants tend to cluster in the CBD. English is the main lingua franca in this diverse inner-city region, and this is particularly evident when African migrants are compelled to mobilise as a group. The patterning of languages in Johannesburg is extremely complex and the language repertoires of Zimbabwean migrants in particular reflect new processes of social transformation. Thus, while English is the main lingua franca in the CBD and other linguistically diverse areas, it is by no means a uniform lingua franca throughout Johannesburg. Among Zimbabwean migrants, English stands in a ‘triglossic’ relationship with Zulu and the Zimbabwean languages. English is most noticeably valued in the northern suburbs (such as Fourways)—which are associated with whiteness and affluence—and in the cosmopolitan inner city. As one moves towards the margins of the city, such as Diepsloot, the currency of English dissipates and other languages in the repertoire define sociality in these spaces.

The value of Shona and Ndebele is also mediated by the location of migrants within the wider Johannesburg area. We noted how many contributors to the literature on Zimbabwean migration tend to treat Shona and Ndebele as relatively inert ethnonyms. This usage corresponds with what Bourdieu (1985:725) calls ‘classes on paper’, that is, it tends to draw on widespread assumptions about the intrinsically ethnic nature of named languages, rather than on actual empirical trends. As theoretical entities, ‘languages’ begin as classes on paper, but their potential to serve as a means of mobilisation is an empirical issue, which Bourdieu articulates in terms of the potential to function as a ‘probable class’. He argues that statistical analysis is ‘the only means of manifesting the structure of the social space’ in which probable classes emerge. In this article we have therefore combined census statistics with qualitative data, in an attempt to explain how specific places mediate the use of Zimbabwean language repertoires in ethnic mobilisation.

Movement across significant geographical boundaries—from the inner-city to a township, or from a township to one of the predominantly white northern suburbs—corresponds with shifting orders of indexicality and these in turn structure the deployment of cultural repertoires. The ‘colonies’ of Yeoville and Hillbrow constitute spaces in which Zimbabweans feel most at home, but this sense of ‘location’ varies across the two suburbs. In Yeoville most obviously the relative comfort of location is associated with the freedom to speak both Shona and Ndebele, whereas Hillbrow is presented as a more Ndebele-orientated space. Diepsloot and other township areas tend to be spaces of relative ‘dislocation’, where the use of English and the Zimbabwean languages marks migrants as outsiders. In these situations, migrants with Ndebele in their repertoires use isiZulu semaflethini to communicate with locals. Many Shona speakers, however, do not have this capacity, and for them restricted use of English, silence, and avoidance strategies tend to define engagement with locals in public spaces.

The situation of Zimbabwean migrants shows that there is a feedback loop between local place and the wider field of Johannesburg. In the City of Johannesburg Zimbabwean migrants tend to concentrate in the CBD, an intercalary space between historically black ‘townships’ and white ‘suburbs’. In this hierarchical and unequal urban setting, language repertoires are resources for both locating and translocating. But the resource-value of Zimbabwean languages varies considerably as the migrants strategically negotiate ‘face-to-face’ encounters in different spaces. Resource-value is subject to the logic of location and dislocation, and Zimbabwean repertoires consequently function as both ‘capital’ (Bourdieu 1986, 1991) and ‘stigma’ (Goffman 1963), depending on their situation. The capacity of Zimbabweans to define themselves as ‘a group’ therefore also varies: migrant identity shifts from active national and ethnolinguistic assertion in the inner-city ‘colonies’ to careful dissimulation in township spaces. Key factors in this variation are individual language repertoires and the ability to use and/or adapt these repertoires across an extremely complex and highly stratified urban environment.

NOTES

1 Kukiya-kiya denotes a process of repeatedly unlocking something and is a term used to describe the period of unparalleled economic crisis in Zimbabwe, which occurred between 2000 and 2008.

2 Local government handbook (2014), available at http://www.municipalities.co.za/.

3 According to the 1982 census, isiNdebele speakers constituted 16.5% of the total Zimbabwean population. A more recent source puts that figure at 20% (Ndlovu-Gatsheni 2009).

4 Subplaces are the smallest named units of South African census geography. As such, the terms subplace, suburb, and neighbourhood are treated as equivalent.

5 All of the names used in this section of the article are pseudonyms.

6 Kalangas are an ethnolinguistic group from the Matabeleland areas of Zimbabwe. In most cases, Kalangas are able to speak Ndebele, having gone through the same process of socialization in Matabeleland. However, in recent times, Kalanga speakers have begun mobilizing for their language to be included in the school curriculum.

7 Bulawayo is Zimbabwe's second largest city. It is situated in the predominantly Ndebele-speaking region of Matabeleland.

8 Gukurahundi is a Shona term that translates as ‘the early rain that washes away the chaff before the spring rain’. This is the name Mugabe designated for a North Korean trained section of the army—the Fifth Brigade—that was only answerable to him. This Gukurahundi section was deployed in Matabeleland massacres (which have also been called to an ‘attempted genocide’) that occurred after independence in 1980. An estimated 20,000 Ndebele-speaking civilians were brutally massacred in the provinces of Matabelelend and Midlands (The Guardian, 19 May 2015; https://www.theguardian.com/world/2015/may/19/mugabe-zimbabwe-gukurahundi-massacre-matabeleland). In Zimbabwe the term Gukurahundi is often used to refer to the period of this attempted genocide. Readers are referred to Gugulethu Siziba's poem ‘Ode to the nameless’, which can be found on the commemorative website http://www.gugulethusiziba.org/.

REFERENCES

Anderson, Rosemarie (2007). Thematic content analysis (TCA): Descriptive presentation of qualitative data. Online: http://rosemarieanderson.com/wp-content/uploads/2014/08/ThematicContentAnalysis.pdf.
Beall, Jo; Crankshaw, Owen; & Parnell, Susan (2002). Uniting a divided city: Governance and social exclusion in Johannesburg. London: Earthscan Publications.
Beck, Rose Marie (2016). Africa, language research. In Callan, Hilary (ed.), International encyclopaedia of anthropology. London: Wiley Blackwell.
Bloch, Alice (2008). Gaps in protection: Undocumented migrants in South Africa. Migration Studies Working Paper Series 38 . Johannesburg: Wits University Press. Online: https://pdfs.semanticscholar.org/f52c/b824a3c7d4342ce57836035e12167c2131c8.pdf.
Blommaert, Jan (2005). Discourse: A critical introduction. Cambridge: Cambridge University Press.
Blommaert, Jan, & Backus, Ad (2011). Superdiverse repertoires and the individual. Tilburg Papers in Culture Studies 24 . Utrecht: Tilburg University.
Blommaert, Jan; Collins, James; & Slembrouck, Stef (2005). Spaces of multilingualism. Language & Communication 25(3):197216.
Bourdieu, Pierre (1983). Erving Goffman: Discoverer of the infinitely small. Theory, Culture & Society 2(1):112–13.
Bourdieu, Pierre (1985). The social space and the genesis of groups. Theory and Society 14:723–44.
Bourdieu, Pierre (1986). The forms of capital. In Richardson, John (ed.), Handbook of theory of research for the sociology of education, 241–58. New York: Greenwood Press.
Bourdieu, Pierre (1991). Language and symbolic power. Cambridge: Polity Press.
Bourdieu, Pierre, & Wacquant, Loïc (1992). An invitation to reflexive sociology. Chicago: The University of Chicago Press.
Bremner, Lindsay (2004). Johannesburg: One city, colliding worlds. Johannesburg: STE Publishers.
Christopher, Anthony (2004). Linguistic segregation in urban South Africa, 1996. Geoforum 35:145–56.
Crush, Jonathan; Chikanda, Abel; & Maswikwa, Belinda (2012). Heading north: The Zimbabwean diaspora in Canada. Migration Policy Series 62 , Southern African Migration Programme, Cape Town.
Crush, Jonathan, & Tevera, Daniel (2010). Exiting Zimbabwe. In Crush, Jonathan & Tevera, Daniel (eds.), Zimbabwe's exodus: Crisis, migration, survival, 149. Ottawa: South African Migration Programme.
Delmont, Elizabeth (2004). Re-envisioning greater Johannesburg: South African heritage in the first decade of democracy. African Arts 37(4):3035.
Goffman, Erving (1963). Stigma: Notes on the management of spoiled identity. Englewood Cliffs, NJ: Prentice-Hall.
Hammar, Amanda, & Raftopoulos, Brian (2003). Zimbabwe's unfinished business: Rethinking land, state and nation. In Hammar, Amanda, Raftopoulos, Brian, & Jensen, Stig (eds.), Zimbabwe's unfinished business: Rethinking land, state and nation in the context of crisis, 147. Harare: Weaver Press.
Hassim, Shireen; Kupe, Tawana; & Worby, Eric (eds.) (2008). Go home or die here: Violence, xenophobia and the reinvention of difference in South Africa. Johannesburg: Wits University Press.
Hymes, Dell (1996). Ethnography, linguistics, narrative inequality: Toward an understanding of voice. London: Taylor & Francis.
Johnstone, Barbara, & Marcellino, William (2010). Dell Hymes and the ethnography of communication. In Wodak, Ruth & Kerswill, Paul (eds.), The SAGE handbook of sociolinguistics, 5766. New York: SAGE.
Jones, Jeremy (2010). ‘Nothing is straight in Zimbabwe’: The rise of the Kukiya-Kiya economy 2000–2008. Journal of Southern African Studies 36(2):285–99.
Landau, Loren (2006). Transplants and transients: Idioms of belonging and dislocation in inner city Johannesburg. African Studies Review 49(2):125–45.
Landau, Loren, & Freemantle, Iriann (2010). Tactical cosmopolitanism and idioms of belonging: Insertion and self-exclusion in Johannesburg. Journal of Ethnic and Migration Studies 36(3):375–90.
Makina, Daniel (2010). Historical perspective on Zimbabwe's economic performance: A tale of five lost decades. Journal of Developing Societies 26(1):99123.
Mbiba, Beacon (2005). Zimbabwe's global citizens in ‘Harare North’: Some preliminary observations. In Palmberg, Mai & Primorac, Ranka (eds.), Skinning the skunk: Facing Zimbabwean futures, 2638. Uppsala: Nordiska Afrikainstitutet.
McGregor, Joann (2007). ‘Joining the BBC (British Bottom Cleaners)’: Zimbabwean migrants and the UK care industry. Journal of Ethnic and Migration Studies 33(5):801–24.
Mlambo, Alois (2010). A history of Zimbabwean migration to 1990. In Crush, Jonathan & Tevera, Daniel (eds.), Zimbabwe's exodus: Crisis, migration, survival, 5276. Ottawa: South African Migration Programme.
Mlambo, Alois, & Raftopoulos, Brian (2010). The regional dimensions of Zimbabwe's multi-layered crisis: An analysis. Paper presented at the conference on Election Processes, Liberation Movements and Democratic Change in Africa, CMI and IESE, Maputo.
Morreira, Shannon (2007). Seeking solidarity: Zimbabwean undocumented migrants in Cape Town, 2007. Journal of Southern African Studies 36(2):433–48.
Morris, Alan (1998). ‘Our fellow Africans make our lives hell’: The lives of Congolese and Nigerians living in Johannesburg. Ethnic and Racial Studies 21(6):1116–36.
Municipal Structures Act (1998). Local Government: Municipal Structures Act (no. 117 of 1998). Government Gazette 402: 19614. Cape Town, South Africa. https://www.gov.za/documents/local-government-municipal-structures-act.
Murray, Martin (2011). City of extremes: The spatial politics of Johannesburg. Cape Town: Cornell University Press.
Muzondidya, James (2010). Makwerekwere: Migration, citizenship and identity among Zimbabweans in South Africa. In McGregor, JoAnn & Primorac, Ranka (eds.), Zimbabwe's new diaspora: Displacement and the cultural politics of survival, 3757. New York: Berghahn Books.
Ndlovu, Thabisani (2010). Where is my home? Rethinking person, family, ethnicity and home under increased transnational migration by Zimbabweans. African Identities 8(2):117–30.
Ndlovu-Gatsheni, Sabelo (2009). Do ‘Zimbabweans’ exist? Trajectories of nationalism, national identity formation and crisis in a postcolonial state. New York: Peter Lang.
Nyamnjoh, Francis (2006). Insiders and outsiders: Citizenship and xenophobia in contemporary Southern Africa. London: CODESRIA Books/Zed Books.
Pasura, Dominic (2010). Competing meanings of the diaspora: The case of Zimbabweans in Britain. Journal of Ethnic and Migration Studies 36(9):1445–61.
Pasura, Dominic (2011). A fractured diaspora: The case of Zimbabweans in Britain. International Migration 50(1):143–61.
Pillai, Nadaraja (2002). Teaching communicative competence in diglossic Indian languages. Language in India 2(2). Online: http://www.languageinindia.com/april2002/nadarajapillaidiglossia.html.
Polzer, Tara (2008). South African government and civil society responses to Zimbabwean migration. SAMP Policy Brief No. 22, South African Migration Project, Cape Town.
SA Constitution (1996). The Constitution of the Republic of South Africa. Cape Town: Government of South Africa. http://www.justice.gov.za/legislation/constitution/SAConstitution-web-eng.pdf.
Selzer, Amy, & Heller, Patrick (2010). The spatial dynamics of middle class formation in postapartheid South Africa: Enclavization and fragmentation in Johannesburg. Political Power and Social Theory 21:171208.
Sibanda, Octavia (2010). Social ties and the dynamics of integration in the City of Johannesburg among Zimbabwean migrants. Journal of Sociology and Social anthropology 1(1–2):4757.
Sisulu, Elinor; Moyo, Bhekinkosi; & Tshuma, Nkosinathi (2007). The Zimbabwean community in South Africa. In Buhlungu, Sakhela, Daniel, John, & Southall, Roger (eds.), State of the nation: South Africa, 552–73. Pretoria: HSRC Press.
Siziba, Gugulethu (2013). Language and the politics of identity in South Africa: The case of Zimbabwean (Shona and Ndebele speaking) migrants in Johannesburg. Stellenbosch: Stellenbosch University dissertation.
Siziba, Gugulethu (2015). ‘Cross-identification’: Identity games and the performance of South Africanness by Ndebele-speaking migrants in Johannesburg. African Identities 13(4):262–78.
Siziba, Gugulethu (2016). The body as a site for (un)making the ‘Other’: Shona speaking migrants’ negotiation of identity politics in Johannesburg. Journal of Immigrant & Refugee Studies 14(2):121–40.
The 2011 Census, produced by Statistics South Africa, is available online at http://www.statssa.gov.za/.
Worby, Eric (2010). Address unknown: The temporality of displacement and the ethics of disconnection among Zimbabwean migrants in Johannesburg. Journal of Southern African Studies 36(2):417–31.