Fifty years after the Soweto uprising, are we still failing learners because of language?

On a bitterly cold morning on 16 June 1976, schoolchildren in Soweto took to the streets. Not simply against apartheid, but against an education system that imposed Afrikaans as the language of instruction, erasing the linguistic heritage of millions of South African children, in the service of white minority rule.

The placards they carried said it all: “Away with Afrikaans”. “We are being certified but not educated”. And my favourite, “If we must do Afrikaans, Vorster must do Zulu”.

The schoolchildren of 1976 understood something we are still grappling with today: language and learning are inseparable. Fifty years later, we are still failing too many children through the medium of language, not through prejudiced legislation, but through weak policy implementation and a belief that English should be introduced as early as possible to children who speak it as a second language.

Is this what the class of 1976 stood and died for?

The cliff at Grade 4

South African policy rightly promotes mother-tongue learning in the Foundation Phase. In practice, many learners experience a dramatic shift at Grade 4 when English becomes the primary language of learning and teaching. For millions of children, this is not a language switch. It is an academic cliff that is a barrier to learning and achieving success.

Abruptly, learners are expected to grasp increasingly abstract concepts in mathematics through a language they are still acquiring. An English home-language learner can focus entirely on understanding the concept. A second-language learner must simultaneously decode the language and the concept. That double burden is not a measure of ability. It is a measure of a system that could have done better.

The consequences are not abstract. They show up in our mathematics results. They show up in the rising number of learners redirected to mathematics literacy between grades 10 and 12. They show up in the 400-odd schools that fail to offer pure mathematics at all. A learner who never had the chance to understand mathematics in a language they think in will not suddenly grasp it when the medium changes. The damage is done long before Grade 10.

What the evidence tells us

Research from Stellenbosch University’s Research on Socio-Economic Policy unit has found that learners who build strong literacy foundations in their home language are better positioned to acquire literacy later in a second language. The reason is straightforward: children do not think in English. They think in understanding.

A child who grasps a mathematical concept in Sepedi, isiXhosa or Sesotho does not stop understanding it when they encounter it in English. What changes is the vocabulary. The understanding is already there, or it isn’t. Language does not create mathematical comprehension. It either enables it or obstructs it.

This is not a new insight. Many countries, including Finland, Japan, China and Singapore, have long taught mathematics in their home language with excellent results. Like South Africa, Singapore has a multitude of languages and cultures (including Mandarin, Tamil and Malay), yet consistently ranks first on international tests like the Programme for International Student Assessment. This has largely been attributed to its foundational curriculum, which ensures learners are taught in their mother tongue. In South Africa, mathematics was, and continues to be, taught in Afrikaans for generations. There is no principled reason why the same commitment cannot be extended to all of the country’s official languages.

In a recent JumpStart classroom, an educator assistant explaining division to Grade 3 learners used the Sepedi word ripagana, meaning “to tear apart”. The concept immediately came alive. Learners could picture it, grasp it and apply it. Division is not an English concept. It is a cognitive one. Once a child understands it, language becomes a bridge. Without that understanding, language is a wall.

Good policy, with encouraging early success

In non-English-speaking schools, many teachers switch between home language and English earlier than policy requires, with the best of intentions, attempting to protect learners from the shock that comes in the Intermediate Phase. Parents often demand it too, and understandably so. English remains the language of universities, boardrooms and economic opportunity. Wanting more of it for your child feels like common sense.

But common sense and evidence are not the same thing.

The Department of Basic Education’s Mother-Tongue-Based Bilingual Education (MTbBE) programme points in the right direction. It began national implementation in 2025 with Grade 4 learners in 12,000 schools. Its logic is sound: children need a language in which they can build deep conceptual understanding, and they need English because it remains central to higher education and opportunity. These goals are not in competition. A strong mother-tongue foundation does not close the door to English. It opens it.

The test is whether learners understand more, participate more and move into the intermediate phase with stronger conceptual foundations than they had before. And in this regard, the MTbBE programme has delivered encouraging early results.

Pilot projects suggest that when learners are taught subjects like mathematics and science in a language they understand best, they grasp concepts more easily and perform better academically. This is important because we have already lost far too many capable learners to mathematics literacy or to abandoning mathematics altogether. Not because they lacked the ability to succeed, but because there were too many factors against them: a weak mathematics foundation in early grades and the added burden of switching to English as the language of instruction.

So, the challenge is not whether the policy works in principle. The challenge is whether we are willing to implement it properly and at scale with the supporting materials, teacher capacity and integration into the education system itself.

The real question

For years, the instinct has been to introduce English earlier in the hope of easing the intermediate transition. The evidence suggests the answer lies elsewhere: stronger conceptual foundations in a language children already understand, with English built alongside, not instead of that foundation.

Technology presents a further dimension. Digital tools have significant potential to support early mathematics learning, but that potential remains unrealised for learners in rural and under-resourced communities where technology is inaccessible. Without deliberate investment in both language and digital access, we risk confining another generation to consumers of the digital economy rather than producers of it.

The schoolchildren of 1976 knew that language could be used as a weapon against them or as a tool to help them thrive. Fifty years on, we should be honest about which of those remains for too many South African children.

We have good policy. We have growing evidence. We have classrooms where, with the right support, the right language and a torn piece of paper, a child’s understanding of mathematics develops appropriately.

The question is no longer about what we know will work. The question is whether we are willing to do this for all our learners everywhere, because we can. DM