Mother-tongue education is failing

The numbers tell a politically uncomfortable truth. A survey of all Grade 4 children in Makhanda’s no-fee public schools this year found that 51% of isiXhosa learners taught in English can read for meaning. 

However, for those taught in isiXhosa — despite it being their mother tongue and despite decades of policy emphasising the value of home language instruction — only 25% reached the same benchmark.

“Children who speak African languages at home and go straight to English schools do better,” says Kelly Long of Makhanda-based NGO, GADRA Education. These are children from the same neighbourhoods, the same socioeconomic backgrounds, often living on the same streets. The only difference is the language of instruction.

The recently released Funda Uphumelele National Survey confirms this pattern nationally – only 29% of isiXhosa learners read for meaning by Grade 4, compared to significantly higher rates in English and Afrikaans.

The data doesn’t prove mother-tongue education is wrong. International evidence supporting the grounding of early learning in a child’s home language is robust. Instead, it shows a gap between policy ideals and classroom realities.

Long acknowledges the global research as irrefutable but says we are setting children up to fail by insisting on mother-tongue instruction without giving teachers the specific tools to teach it. 

“We’ve got to stop being scared of saying the controversial thing,” Long says. “We’ve tried it for 30 years. We haven’t made it work.”

Language differences

The failure begins with a misunderstanding of the languages. For years, the South African education system has largely “copy-pasted” English reading methodologies into African language classrooms.

African languages are agglutinative — complex words are built by stacking prefixes, infixes and suffixes onto roots, allowing a word to represent an entire sentence. Rhodes University linguist Paige Cox says the complexity has implications for teaching reading, yet phonics instruction often ignores them.

The result is a phenomenon researchers call “barking at text”, where learners can sound out words perfectly but don’t comprehend what they’re reading. “You would have children reading fluently, and then you ask them a straightforward literal comprehension question and they can’t answer,” Cox says.

Her research revealed something unexpected: unlike in English, where similarly spelt words help learners read faster through pattern recognition, no such effect exists in isiXhosa reading. Many learners aren’t reading whole words but are stuck decoding syllable by syllable, never building the automaticity needed for comprehension.

Decolonising teacher training

The director of reading at the Department of Basic Education, Dr Nompumelelo Nyathi-Mohohlwane, points out that many universities train Foundation Phase teachers in English and then expect them to teach literacy in African languages “when they were not trained in isiXhosa literacy methodologies”.

The National Framework for Teaching Reading in African Languages confirms “the influence of reading methodologies used in English is so strong that it overrides the development of reading methodologies and pedagogies that are appropriate for African languages”.

African language children taught in English from the Foundation Phase have a paradoxical advantage: their schools benefit from decades of refined, global reading research and a wealth of reading materials. English-medium schools are more likely to have book corners, smaller class sizes, and libraries. And English text is everywhere — on screens, signs and in shops. By contrast, isiXhosa-medium schools often have a workbook and perhaps a couple of graded readers.

Additionally, for millions of children, their meagre “mother tongue” reading materials often don’t correspond with their home language. For Nyathi-Mohohlwane, this is because we have “poor translations, or purist, outdated forms of isiXhosa, or literal translations that have no meaning”.

Most tellingly, South Africa’s standardised official African languages mask dozens of dialects.

Systems, not languages, are the problem

Nyathi-Mohohlwane and Cox reject the idea that African languages are “harder” to read. “The key is that these characteristics are language-specific — and so teaching reading needs to be language-specific. We have not yet cracked this,” Cox says.

Nyathi-Mohohlwane points to our legislation: “Everything about teaching time, resources, assessments — but very little about language. African languages have never been centred in the system.”

Ways forward

“We need to separate the political argument from the practical one and ask: What actually works for our children?” Nyathi-Mohohlwane says. “Everybody talks about language. Nobody talks about the system that stands behind that language — the training, the materials, the assessments, the support.”

The government has allocated a mere R57 million for the national rollout of Mother-Tongue-Based Bilingual Education. But Nyathi-Mohohlwane says South Africa has finally begun treating early-grade African-language books as a priority. For the first time, the DBE has issued a formal national call for Grade 1–3 African-language reading materials and will catalogue approved books.

Yet materials alone are not enough. The foundational gap begins even earlier, in Grade R, where curriculum support is less visible and interventions are rare. 

Makhanda teacher Bridget Swift is addressing this through an isiXhosa adaptation of “Letterland”, a letter-sound recognition programme she developed initially for rural KwaZulu-Natal classrooms 20 years ago. Funded by the Solon Foundation, her “Mountain of Sounds” features characters like grannies and aunties rather than parents, reflecting the realities of migrant labour in rural communities.

The case for coaching

Since 2015, the Department of Basic Education’s Early Grade Reading Studies have rigorously tested methods to help children learn to read. Sending teachers to centralised training workshops had almost zero impact. What worked was on-site coaching — when teachers were provided with high-quality materials and visited by expert coaches, learners made gains, surpassing about  40% of a year’s worth of learning compared to control schools.

The NGO Funda Wande has replicated this success in the Eastern Cape through termly training and classroom visits. “The coaches observe lessons, give feedback, and run afternoon training sessions,” explains coach Nomalungelo Zono. The DBE’s recent FUNS data validated this: interventions with embedded coaching consistently outperformed training-only models.

However, coaching is expensive — estimates suggest R730 to R2 500 per learner per year. And the “human factor” is fragile. In Makhanda, when Funda Wande’s external support ended, test scores began to decline.

Quality digital materials

GADRA Education’s QondaRead programme offers a different model. Rather than relying on human experts, QondaRead “bakes” pedagogy into multimedia lessons projected onto the classroom wall. The meticulously sequenced phonics lessons are delivered alongside culturally resonant graded readers — 24 titles per grade, drawing from children’s everyday lives.

The 2025 QondaRead Report reveals an impact equivalent to a full year’s learning gains — at approximately R172 per learner per year, a fraction of the coaching cost.

However, QondaRead is available only in English. Developing an African language version would require dedicated research and development, with linguists, teachers and designers collaborating to ensure the embedded pedagogy is appropriate — not simply “translating” English approaches that could compound past errors.

Building capacity

In Makhanda, much of this research and development work is underway. Rhodes University’s two-year Advanced Certificate in Foundation Phase Literacy, developed in partnership with Funda Wande, has created phonics scope and sequences tailored to the structure of each language.

Staff member Tabisa Booi emphasises linguistic flexibility: “Allow the languages to be as fluid as they are.” Course designer Sarah Murray says the goal is to generate a cohort of African-language literacy practitioners who can implement programmes and develop materials. Unlike pre-service university programmes that front-load theory, the Advanced Certificate targets experienced teachers and subject advisors who can implement learning immediately.

The uncomfortable truth

The research is clear. Mother-tongue education is effective but only when the system supports it. For decades, we’ve asked teachers to deliver African-language literacy without the training, materials or pedagogical frameworks the task requires.

That may finally be changing. A national materials catalogue, language-specific teacher training and emerging models that embed expert pedagogy into classroom resources suggest the gap between policy and practice can be closed. However, it will require a political will that matches the scale of the crisis.

Until then, African language children taught in English will continue to outperform those taught in their mother tongue — not because the policy is flawed but because we have never truly attempted to make it effective.

This feature was made possible by the Henry Nxumalo Foundation which funded the Between the Lines series

.