How a country chooses the language for its education system is not an easy process. The decision is usually influenced by multiple factors: colonial history, origins of immigrants, legal recognition of minority languages, cultural diversity, political interests – to mention but a few. In some cases, instruction is provided in more than one language; in others the medium of instruction may vary between primary and secondary education.

Underneath this tangled and evolving web of policies and priorities, however, lies an undeniable truth: teaching and assessing children in a language they understand will result in better learning. This is something that has been recognised now for decades. It is written into the 1989 ILO Convention and Convention on the Rights of the Child and the 2001 Universal Declaration on Cultural Diversity. Our new paper out today, ‘If you don’t understand, how can you learn?’ confirms this basic principle, and yet reports that, despite the overwhelming evidence supporting this claim, 40% are still not able to access education in a language they understand.  It is clear that the complex nature of factors affecting language-education policy still take precedent over the accumulation of evidence.

Countries with colonial histories often find that shifting to bilingual education is complicated, as can be seen in many Latin American contexts that continue to use Portuguese, or Spanish, or in many Francophone African countries, where French remains the predominant language of instruction. Our World Inequality Database on Education (WIDE) shows that this trend seriously hampers students’ chances of learning. In Côte d’Ivoire, for example, 55% of grade 5 students who speak the test language at home learned the basics in reading in 2008, compared with only 25% of those who speak another language.

Nor is this confined to sub-Saharan Africa. In Iran, around 80% of grade 4 students who spoke a language other than Farsi at home reached the basics in reading, compared with over 95% of Farsi speakers.

Similarly, in Honduras, in 2011, 94% of students who spoke the language of instruction at home learned the basics in reading in primary school compared to 62% of those who did not.

Part of the repeated emphasis over teaching children in a language they understand is because linguistic barriers only serve to exacerbate the divides caused by other disadvantages such as poverty or gender or location. Students from poor households who speak a minority language at home are among the lowest performers. In Turkey in 2012, around 50% of poor non-Turkish speakers among 15 year olds achieved minimum benchmarks in reading, against the national average of 80%. As such, a mono-lingual education system can unwittingly promote educational disadvantages and economic inequalities from one generation to the next.

The decision over language policy in schools is highly contested since the choice of which language to use for instruction can divide just as it can unite; it forms a group’s identity, and as such can be the glue to bring people together, or the barrier that divides them.

In multi-ethnic countries, in particular, the imposition of a single dominant language as the language of instruction in schools, while sometimes a choice of necessity, has been a frequent source of grievance linked to wider issues of social and cultural inequality.

Our paper lays out several of these examples. Disputes about using Kurdish in schools have been an integral part of the conflict in eastern Turkey. In Nepal, the imposition of Nepali as the language of instruction fed into the broader set of grievances among non-Nepali speaking castes and ethnic minorities that drove the civil war. Guatemala’s imposition of Spanish on schools was seen by indigenous people as part of a broader pattern of social discrimination. In Pakistan, the continued use of Urdu as the language of instruction in government schools, even though it is spoken at home by less than 8% of the population, has also contributed to political tensions

The GMR 2005 argued that there can be no discussions of quality in education without consideration of the language of instruction. With a renewed focus on quality in the post-2015 education agenda, our new paper helps policy makers find a way through the issue, and lays out some key recommendations to ensure that children are taught in a language they understand.

1. At least six years of mother tongue instruction is needed, if the gains from teaching in mother tongue in the early years are to be sustained.
2. Education policies should recognize the importance of mother tongue learning. A review of 40 countries’ education plans finds that only less than half of them recognize the importance of teaching children in their home language, particularly in early grades.
3. Teachers need to be trained to teach in two languages and to understand the needs of second-language learners. Teachers are rarely prepared for the reality of bilingual classrooms. In Senegal, only 8%, and in Mali, only 2% of trained teachers expressed confidence about teaching in local languages. The paper suggests hiring teachers from minority language communities as one policy solution to the problem.
4. Teachers need inclusive teaching materials and appropriate assessment strategies to help them identify weak learners and provide them with targeted support.
5. Provide culturally appropriate school-readiness programmes: Locally recruited bilingual teaching assistants can support ethnic minority children from isolated communities as they make the transition into primary school.
6. Second-chance accelerated learning programmes in local languages can help the disadvantaged to catch up.