South Africa


Afrikaans and the university language debate (Part 2): The ‘costs’ of institutional bilingualism

Stellenbosch University. (Photo:

Within the domain of higher education, the decline of Afrikaans as a medium of instruction comes on the back of a more fundamental shift: The gradual erasure of the longstanding white institutional division between Afrikaans universities and what used to be called English liberal universities.

Part 2 in a two-part series: Part 1 can be read here.

On the eve of South Africa’s democratic dispensation, a language deal was reached between the incumbent National Party and the aspirant African National Congress. The nine African languages that had previously received subnational recognition within the “grand apartheid” Bantustan system were granted official status, alongside Afrikaans and English. The compromise was articulated in Section 3(2) of the interim Constitution (1993), which came to be known as the “non-diminution clause”:

Rights relating to language and the status of languages existing at the commencement of this Constitution shall not be diminished, and provision shall be made by an Act of Parliament for rights relating to language and the status of languages existing only at regional level, to be extended nationally in accordance with the principles set out in subsection (9).”

Neither this clause nor the Section 3(9) commitment to the “promotion of the equal use” of languages made it into the new (1996) Constitution, for good reason. Presented as the core principle underpinning a progressive commitment to multilingualism, the non-diminution clause was in fact the constitutional equivalent of a perpetual motion machine. The 1993 compromise was unworkable, because the concept of “languages” that underpinned the commitment to “equal use” was fundamentally flawed. Ideas about “language equality” have, however, been central to the post-1994 debate on language in higher education.

The five recent Constitutional Court pronouncements on language policies in the university context, discussed in my previous article, constitute an important new attempt to grapple with entropy in the system. While I am in broad agreement with the “lead” Constitutional Court arguments, ie the majority judgment (University of the Free State) and the consensus judgment (Stellenbosch University), I am also sensitive to the complex new sociolinguistic class dynamics that they manifest.

These two judgments suggest a rather clear general orientation: an expectation that expanded English-medium instruction will provide greater access to a very unequal university system. And from the specific arguments relating to the political and economic costs of the different bilingual forms of instruction at each university emerges a more general statement on the political economy of language in our higher education system. The economics of language is not very developed in South Africa, in part because economists and linguists tend to think of themselves as “hard” scientists with very clearly demarcated and mutually exclusive domains of expertise.

In this article I therefore continue to develop the theme of “costs” — broadly defined to include both material and cultural costs — using three important conceptual distinctions to frame my argument.

Thinking institutionally about languages

To begin with, we need to make a distinction between languages as individual and institutional phenomena. There is no easy way to make this distinction, but it is a useful way of dealing with the common tendency to use a language label (eg English, Afrikaans, isiXhosa) to refer to a clearly defined group of people who share a “mother tongue” (eg English speakers, Afrikaners, amaXhosa). The term “mother tongue” tends to convey a sense of “languages” as discrete cultural containers. When used in this way, “language” papers over other dimensions of cultural identity (as I show below, with respect to race).

To think about “languages” as institutional phenomena is to recognise them as cultural and political projects. This recognition breaks with the organic analogy: the comparison with living things that underpins both popular ideas about language and an established scholarly tradition that treats linguistic diversity as a form of biodiversity.

In the South African context this notion underpins the idea that our official languages ought to develop to equality — rather like pot plants in an inert and greenhouse-like national space. This is problematic, first because “mother tongue” has a historical association with apartheid education and the Bantustan system. The conceptual problems run deeper than this, however, to missionary-mediated ideas about language that originated in 19th-century European philology and nationalism. Most South Africans use more than one language and this use is mediated by political, economic and cultural institutions.

Law is the most obvious source of institutionalised languages. During our democratic transition, the 1993 and 1996 constitutions re-established common-sense language categories that had already enjoyed 50 years of legal recognition.

Another historically profound institutional source is the national census. One reason why race is still a “proxy for disadvantage” in South Africa is that our census categories for “race” and “language” are joined at the hip. “Race” is far older — dating back to the first census conducted by the Cape Colony in 1865. English and Dutch were first enumerated in 1918, ironically, the year in which our three oldest universities (Unisa, Stellenbosch University and the University of Cape Town) were formally established. But our current official African languages were first included in the census of 1946. It is significant that the 1996 census was the first to present independent questions on race and language to the whole South African population. In other words, all previous censuses presented four different questions on language to the four “population groups” — the census term for race. While this is an improvement of sorts, these census categories still perpetuate a widespread sense that both “races” and “languages” are natural things.

The way in which our census statistics are used also tends to reinforce the idea that our official languages (read as “mother tongues”) correspond with 11 distinct “ethnic groups”. The most recent (2011) census asked “which two languages does (name) speak most often in this household?” The following table shows the counts for both census categories.

The table is ordered in terms of first language statistics, as is the common practice. Typically, only the first response is counted, but if we count both responses we get a better sense of (1) the fact that most South Africans speak more than one language, and (2) the socioeconomic status that English has as a second-language lingua franca. The latter is no doubt weakly indicated as the question specifies “household” use and thus excludes the key domains of work and education. While the census is a powerful source of commonsense “languages”, it nevertheless tends to reinforce a de-institutionalised sense of languages as aggregates of individual or household spoken competencies.

In their response to the recent Constitutional Court judgment on the Stellenbosch University language policy, Gelyke Kanse is reported to have argued that the ruling “will affect all mother tongues”. The argument for mother-tongue instruction in higher education is particularly problematic, as it conflates the individual and institutional “uses” of languages at complex public institutions. At universities, this ignores the changing nature of academic work as a student moves from school to undergraduate studies (the domain that has tended to be the focus of attention in the Afrikaans taaldebat), and from undergraduate to postgraduate study and research.

Moreover, this argument trades in a political market for popular ethno-nationalist stereotypes, as evident in the following extract from a letter that I came across in a local newspaper:

For what reasons, other than ulterior motives, would I as an Afrikaans speaker enrol at a predominantly Tswana, Zulu [or] English-speaking university…?” Like the hermetically sealed languages with which they are associated, these “universities” do not exist.

Dual and parallel medium instruction

In order to fully reckon with the institutional “costs” of language in the university system, we need to understand the recent history and political economy of university-level institutional bilingualism. The second key distinction is therefore between parallel and dual-medium instruction. At historically Afrikaans universities the former has involved the duplication of a lecture and/or tutorials, while the latter involves the use of both English and Afrikaans in one class. These two teaching strategies come with very distinctive costs.

As Stellenbosch University argued and the Constitutional Court has accepted, parallel medium instruction is more expensive — in terms of money and human resources. But why then, in the disputed 2014 language policy that Gelyke Kanse sought to reinstate, had the university suddenly opted for parallel-medium instruction? Because the cultural cost of the university’s preferred dual-medium model (involving the so-called “T-option”) had become politically untenable. By “cultural cost” I mean the fact that dual-medium instruction assumes, as a condition of undergraduate enrolment, a high level of English-Afrikaans bilingualism. This tended to exclude students who did not have matric-certified competencies in English and Afrikaans, most of whom were black.

The wider politics of institutional bilingualism can be traced to the conflicts over strategy that emerged at Afrikaans universities in the post-1994 period. The newly elected ANC government favoured a shift to parallel-medium instruction at Afrikaans universities and by the end of the decade, three universities had followed this route: Rand Afrikaans University (now the University of Johannesburg); the University of Pretoria; and the University of the Free State.

In the years after the 2002 passing of the national Language Policy for Higher Education, parallel-medium instruction was “ideologically unacceptable” at Stellenbosch University, because it was seen as unsustainable. While the demography of the Western Cape was used — rather expediently by white academics — to justify this position, many Afrikaans language activists were also concerned that a shift to parallel-medium instruction would prove too expensive and thus presage the decline of Afrikaans as an academic language. While the recent and long-term historical evidence seems to supports this concern, I would argue that this is one aspect of a more complex transition.

Within the post-1910 context of powerful English colonial institutions, the Afrikaner nationalist struggle tended to prioritise parallel-medium over dual-medium instruction — at schools. At universities however, scale and complexity made the situation somewhat different. After 1951 only two universities sustained institutional bilingualism for a prolonged period of time: Unisa’s reorientation to distance education involved parallel-medium instruction, while the University of Port Elizabeth (UPE, established in 1964) opted for dual-medium instruction.

UPE provides an interesting case study of the economic trade-off between apartheid language planning and racial segregation. The discipline-based model of dual-medium instruction that prevailed between 1964 and 1994 was initially not the preferred choice of local white Afrikaans-speakers. The National Party government’s post-1959 extension of apartheid to higher education rendered the parallel-medium option too costly. The social and political costs of a racist system were, of course, discounted — until after 1994. Since 1994, however, both the social and economic costs of various models of English-Afrikaans institutional bilingualism have become increasingly apparent.

Translation and simultaneous interpreting

Third, my evaluation of the institutional costs of tertiary-level language use is premised on a distinction between simultaneous interpreting and translation. This in turn rests on a lower level distinction between language as a spoken medium and “writing” in the broad sense of the term, ie including a wide range of analogue and (increasingly) digital language artefacts.

At Stellenbosch University the ill-fated 2014 policy sought to make English and Afrikaans “equal” mediums of undergraduate instruction, and the principal mechanism for this was a language plan specifying either parallel-medium instruction or the simultaneous interpretation of lectures for most undergraduate modules.

The latter was a third model of institutional bilingualism developed at the Potchefstroom campus of North-West University, in an effort to retain Afrikaans as a medium of instruction and open the campus to students from the Mafikeng and Vaal Triangle campuses. Theuns Eloff, the first Rector of NWU has reacted to the consensus judgment and defended simultaneous interpreting by arguing that this is “international practice at the UN and the EU”. Fair enough, but is it international university practice?

The main advantage of simultaneous interpreting is that it is relatively cost-effective when compared with parallel-medium instruction. First, interpreting relocates the embodied cost of bilingualism from academic staff to a professional corps of interpreters. Second, interpreting creates a new verbal rendition of a lecture, and as such it also a cheaper — but qualitatively inferior — alternative to the translation of the academic material on which the lecture is based. Furthermore, translation is not “simultaneous”, which suggests the importance of time as an additional cost factor.

The recent Constitutional Court case provides an example of this: the case was postponed for a year to allow all documentation to be translated into English.

Unlike simultaneous interpreting, translation has been pivotal to the development of trans-language equivalence in the numerous domains of science. Ironically, while translation has been the great capacitator of transnational science, the economics of translation presents one of the major barriers to “language equality”. As the historian Michael Gordin has shown, translation was central to the post-World War II rise of English as the dominant language of science. The United States established new economies of scale with respect to translation (downgrading the social status of human translators in the process) and used these to systematically translate Soviet scientific journals.

There is an important historical correlation here, which speaks to the role of English in our emerging national class structure and the global stratification of higher education. The 20th-century development of Afrikaans as a university-level medium of instruction coincided with the emergence of English as the dominant global medium of scientific publication. South African universities were part of this latter process and for this reason, institutional bilingualism was both a means of transitioning to Afrikaans (most of our universities began as Anglophone “colleges”) and a process of benchmarking within a wider Anglophone environment.

Even at the height of apartheid, English universities consolidated the status of educational hotspots (in Cape Town, Johannesburg, Durban, Pietermaritzburg and Grahamstown) and white English speakers were culturally dominant within what was effectively a racialised parallel-medium higher education system. The post-1994 rise and decline of bilingual English-Afrikaans language planning models at historically Afrikaans universities must be situated within this wider historical context, which I have explored in a recent academic article.

The rise and decline of post-1994 institutional bilingualism

After 1994, university-based language planning — as both academic discourse and institutional practice — emerged by way of a fragile alliance between Afrikaans and African language interest groups. Both reacted to the tacit status given to English in the ANC-led nation-building project, but the unity of this alliance has been undermined by ambivalent government support, contextual differences between historically black and white institutions, and the growing national and international status of English. There has, however, also been an intrinsic problem: the manner in which language planners have tended to conceptualise the role of language in science.

Afrikaner nationalism was not as successful at establishing Afrikaans as “a language of science” as is generally supposed. What was effectively institutionalised was not the “equality” of English and Afrikaans, but rather relatively equal access to an elite and bilingual national educational system for white English and Afrikaans speakers.

Consider these two statements, reflecting widely held sentiments:

English is the global language of science” and “Afrikaans is a standardised scientific language.” The italicised articles highlight a propensity to see discrete languages as conduits of science or knowledge more generally. The mistake is to assume that named languages are (equal) channels for scientific communication. What we call “languages” are rather the emergent products of communication. And at university level they are expensive products, ie more or less institutionalised in the particular use-contexts of the various domains of science and technology.

The different models of institutional bilingualism that emerged at historically Afrikaans universities therefore reflect different institutional positions within the changing post-1994 university context. An institutional analysis runs against the grain of individualist explanations of the decline of university-level Afrikaans.

This mode of thinking produces clear culprits: FW de Klerk, Roelf Meyer, Kader Asmal, Blade Nzimande and — of course — “the Rectors.” In a recent article, Johann Rossouw argues that “Afrikaners often blame the ANC government, but this Anglicisation was mainly driven by Afrikaans rectors.” There is some merit to this argument, if “Rectors” is used metonymically, to refer to universities and the local contexts in which they are embedded.

The 2002 national Language Policy for Higher Education effectively dis-established Afrikaans universities when it expressed concern “that some individuals have equated institutional responsibility for promoting Afrikaans as an academic medium to the establishment of ‘Afrikaans’ universities. The notion of Afrikaans universities runs counter to the end goal of a transformed higher education system…”. This was underpinned by the position adopted by the Rectors of the Historically Afrikaans Universities, ie “that the sustained development of Afrikaans should not be the responsibility of only some of the universities.”

With hindsight it can be argued that — in their failure to defend the institutional status of Afrikaans — the Rectors exercised poor judgment. I nevertheless think that this ignores the systemic nature of the problem that Afrikaans universities faced: how to sustain Afrikaans in the face of the structural transformation of a racially fragmented post-apartheid system. Frederik van Zyl Slabbert had a sense of this, when in March 1999 he addressed a meeting of the Afrikaans Consultation Forum in Stellenbosch:

I am somewhat confused as to whether we are speaking about the survival of universities or the survival of Afrikaans. I think issues that concern the survival of a particular university differ from those that have to do with the survival of Afrikaans as a language, and in this respect I would say that a university is one of the possible structures that can facilitate the survival of the language. Second, I agree with Neville [Alexander] that there are too many universities. One should remember that many of these universities were created in terms of the so-called homeland need, the consequence of which is that these universities would empty once the society was opened and people were free to choose which university they wished to attend.”

This comment was made before the tumultuous institutional mergers of 2002-2005. But it tacitly reflects the fact that the various Afrikaans interest groups were not simply striving to “save Afrikaans”, in an abstract sense. At Afrikaans universities “language” tended to mask contrasting local-level interests and class positioning, which were channelled into various forms of institutional bilingualism.

The ‘costs’ of racialised institutional bilingualism

It should now be clear why I have shifted the emphasis from Afrikaans to institutional bilingualism. Within the domain of higher education, the decline of Afrikaans as a medium of instruction comes on the back of a more fundamental shift: the gradual erasure of the longstanding white institutional division between Afrikaans universities and what used to be called English liberal universities.

This is occurring within a wider context in which the division between historically white and black universities not only remains, but also overlaps a division between postgraduate-orientated and research-intensive institutions, and predominantly teaching institutions. Within this context, it strikes me that English plays both an integrating and a stratifying role. Given this paradoxical status, the potential advantages and opportunity costs of using other languages need to be considered carefully and context-specifically.

The Constitutional Court judgments on university language policy — along with the Court’s 2017 decision to make English the language of record — bolster a long-running national trend. English is increasingly the language of “record” — more generally — and our dominant educational medium. This is not a uniformly “good” or “bad” trend, but it is a complex fact. English has a colonial legacy and is inextricably entwined with what Nick Srnicek calls our current phase of “platform capitalism”.

But it is also the language that our academics, researchers and journalists invariably have to read to interrogate the evidence for claims about “the Fourth Industrial Revolution” or “Our Burning Planet”. We need to think critically about English, but underpinning this is the need to think critically about “languages”. DM

Dr Lloyd Hill is a lecturer in the Department of Sociology and Social Anthropology at Stellenbosch University.


Please peer review 3 community comments before your comment can be posted