I-Dollar EYI One!1

1. IsiXhosa: ‘A dollar for all’.

–Ethnolinguistic Fractionalisation, Communication Networks and Economic Participation–Lessons from Cape Town, South Africa

Abstract

The relationship between ethnolinguistic fractionalisation and development has long been of interest to economists and linguists. While econometric analyses have shown relatively stable interactions between high levels of fractionalisation and low indices of development, the mechanisms underlying this relationship are still unclear. This paper explores the importance of fragmented versus unified communication networks for socio-economic development, using data from Cape Town, South Africa. Like other cities in low- and middle-income countries, Cape Town shows growing linguistic diversity due to high levels of rural–urban migration. Two aspects of the city's economy will be discussed on the basis of specialised survey data and anthropological fieldwork: (a) the labour market, and (b) informal entrepreneurial activities. The analysis shows the importance of language as an explanatory variable in the study of economic life, and allows us to advance our understanding of human and social capital formation in ethnolinguistically fragmented societies. In the conclusion the authors discuss the policy implications of the study.

Acknowledgements

The work reported in this paper draws on a larger research programme on language, migration and development which has received support from the Institute for the Study of Global Movements, Monash University, Australia (2003–2005), the South Africa Netherlands Research Program for Alternatives in Development (SANPAD; 2006–2007), and the National Research Foundation South Africa (2006–2008). An earlier version of the paper was presented at the DPRU/TIPS conference ‘Accelerated and Shared Growth in South Africa: Determinants, Constraints and Opportunities’ (Johannesburg, October 2006), and the authors would like to thank Haroon Bhorat, Brett Inder, Rajend Mesthrie, Volker Schoer, and two anonymous referees for critical and helpful comments. The usual disclaimers apply.

Notes

1. IsiXhosa: ‘A dollar for all’.

2. The ELF is calculated as follows (where p is the proportion of speakers of a given language and/or ethnic group relative to the entire population of the country or region in question): ELF = 1−(p₁ ² + p₂ ² + p₃ ² + p₄ ²…).

3. Under apartheid South Africa's population was classified into four groups: Asian, Black, Coloured (people of mixed African, Asian and European descent), and White. These racial classifications still linger on in post-apartheid South Africa and are elicited regularly on forms, in surveys and census data to trace the advances which have been made since 1994 towards political, social and economic equity. The racial terms are usually capitalised to denote the historically constructed status of these terms, and their continuing (for example, Black Economic Empowerment, Employment Equity legislation) legal implications.

4. Haldenwang's (2001) (medium) projections assume continuing rural–urban migration as well as HIV/Aids-related population decline. Speaker numbers for English and Afrikaans are likely to be relatively stable, although language shift towards English will continue in Afrikaans communities (Deumert, 2005). Speaker numbers for ‘other languages’ will increase due to growing intra-African migration trends: between 1996 and 2001 the number of African migrants in South Africa grew by 200 per cent (Statistics South Africa, Census, 1996 and 2001). Yet, the overall percentage of the population of international migrants (and their languages) will only increase marginally as this growth proceeds from a very low base: an increase from 0.56 per cent to about 1.6 per cent would be projected for 2006 if one were to assume a continuation of the 200 per cent growth rate for the period 2001 to 2006. However, these calculations do not take account of the fact that the category ‘other’ also includes European migrant languages (for example German, Dutch, Portuguese) with stable or declining speaker numbers as well as Indian languages (Tamil, Hindi, Gujarati). Unfortunately, Statistics South Africa does not provide individual statistics for the many languages coded as ‘other’.

5. At a national level the South African constitution recognises 11 languages as official: Afrikaans, English, isiNdebele, isiXhosa, isiZulu, Sesotho, Sesotho Sa Leboa, Setswana, SiSwati, Tshivenda, and Xitsonga.

6. In South Africa, the term ‘township’ traditionally refers to a low-quality and poorly serviced residential area for non-White residents. Under the Group Areas Act (1950), urban settlement mirrored racial classifications with each population group being given designated areas. The Act was repealed in 1991.

7. The survey used the ‘official’ or ‘strict’ definition of unemployment, that is, individuals who are not employed, want to work and have taken active steps to look for work in the four weeks prior to measurement. The survey classified all individuals who were engaged in some kind of work at the time of the interview as ‘employed’. This included employment in the informal economy as well as various forms of casual, part-time employment.

8. Imizamo Yethu has received considerable public and academic attention since the late 1990s. Its set-up (wealthy suburb in walking distance to long-standing, non-transitional squatter area) remains singular in the South African context. In 2003, the Development Action Group carried out a survey in the area (over 3800 household surveys were completed) which estimated the total population at 7874 residents, and provided an unemployment estimate similar to the Monash survey (28% unemployment for Imizamo Yethu). In the 2001 census, the official unemployment rate is considerably higher for Imizamo Yethu (46%), yet still lower than the rate in Guguletu (51%). It must also be noted that the census data for Guguletu comprises an area much larger than the highly impoverished NY1 settlement which was the focus of the Monash survey. The census suburb of Guguletu includes lower-middle class neighbourhoods with considerably higher employment rates than NY1.

9. The survey distinguished three categories of employment: ‘regular wage employment’ (full-time work which provided respondents with a reliable monthly income, although not necessarily with social benefits), ‘casual employment’ (part-time work which usually, although not necessarily, provided respondents with sufficient income to cover their basic daily needs), and ‘self-employment’ (in the informal economy, see Section 4).

10. For international readers Rand values cited in this paper have been converted into US $ for reference purposes. The currency conversion is based on exchange rates in March 2007 and should only be taken as an approximation.

11. The survey was conducted in two areas: Mitchell's Plain, an English/Afrikaans-speaking neighbourhood with mostly Coloured residents, and Khayelitsha, a historically Black township (similar to Guguletu) with isiXhosa as the majority language. The survey is available on-line from the Data First Research Unit at the University of Cape Town.

12. In this context, literacy in especially English plays an important role: the newspaper market in Cape Town is firmly in the hand of the English press (Cape Times, Cape Argus), and although an isiXhosa community newspaper exists, it does not carry job advertisements.

13. We distinguished five age groups in the analysis: 0–19 years, 20–30 years, 31–40 years, 41–50 years, older than 50 years.

14. In the survey there was considerable over-reporting of English proficiency, while self-assessments of Afrikaans proficiency were more reliable (see Deumert et al., 2005 for a detailed discussion and examples, see also Dustmann & van Soest, 2001, Schmied 1991: 28–29). Use of English with friends was therefore used as a proxy for English proficiency (that is, assuming that people with high proficiency in English are likely report some use of English in peer-group communication, typically in combination with the home language, that is, isiXhosa).

15. In the South African context, Cornwell (2006), using labour force surveys (1996–1998), also showed that black South Africans who speak English at home are more likely to be employed and to earn a higher income than black South Africans who use an African language at home.

16. South African term for ‘barbeque’.

17. Vetkoek: a type of deep-fried bread similar to doughnuts; Koeksuster: a syrup-coated doughnut.

18. In addition, English is becoming more important within the informal township economy as a provider of employment: the growing number of small businesses started by international migrants generally operate through the medium of English (on international migration to South Africa and small business development, see Maharaj, 2001). This is an area where more research is urgently needed.

19. The link between bridging social capital and development was already noted by John Stuart Mills (1848) in his Principles of Political Economy (p. 581): ‘It is hardly possible to overrate the value … of placing human beings in contact with persons dissimilar to themselves, and with modes of thought and action unlike those with which they are familiar … Such communication has always been, and is peculiarly in the present age, one of the primary sources of progress.’

20. In the absence of community multilingualism, multilingual individuals (language ‘brokers’) can establish themselves as linkages between different language-based networks. The literature on language brokers has so far focused on children of international migrants who often serve as interpreters for their non-fluent parents and extended family members (Morales & Hansen, 2005). More research is needed regarding non-child language brokering.