The most widely read GEM Report publication is our paper last year on language policy in education. Why? Because there are about 6,500 languages spoken in the world today, and, as we showed in that study, a staggering 40% of the global population are learning in a language they don't understand.
It doesn't need explaining perhaps, that being taught by a teacher in a language you don't speak at home will negatively impact your school performance and test results, but it continues to be a hotly contested topic around the world - as recent headlines from Argentina, India and Uganda illustrate.
Politics and ideology are two reasons the issue is contested, but cost is another. Multilingual teaching, the training needed to support it, and the revision of existing learning materials result in a hefty bill. This is why we should loudly celebrate when countries do make the leap to multilingual schools: The Ministry of Education in Malaysia made such an announcement after the GEM Report's policy paper was released last year. Referencing our recommendation for children to receive at least six years of education in a language they understand, they announced they would be making 300 schools bilingual.
The language of instruction can exacerbate inequalities in learning achievement
New data from our World Inequality Database on Education (WIDE database) show the powerful influence that language of instruction has on learning.
A 2013 assessment among primary school children in Latin America showed, for example, that in Colombia 71% passed the second level of a reading assessment when they spoke the same language at home, but only 11% passed when they did not. In Nicaragua 48% passed when they spoke the same language at home, but only 7% passed when they did not.
And it's not just reading, at the primary level learning gaps also exist in science and mathematics. In Guatemala for example, 38% of children passed the second level of a mathematics assessment among those who speak the language of instruction at home compared to just 8% of those who do not. In Peru, 64% of children passed the second level of a mathematics assessment, compared to 23% who did not.
Data from a 2012 UNICEF study also show that our recommendation of 6 years of mother tongue instruction presents challenges in many settings. Where many languages exist, as in Vietnam, it can be complicated finding a teacher who can meet all students' needs. One survey showed that Vietnamese was the first language of 70% of teachers but it was the strongest language for only 1 in 5 students.
It should come as no surprise that learning gaps also exist in wealthier countries, and are likely to grow as more families seek to improve their life chances and those of their children by moving across borders. In the 2012 PISA survey run by the OECD, nearly 15% of 15-year-old students did not speak the language of instruction at home, but among first generation immigrant students, this rose to 63%. Even as many as 38% of second-generation immigrant students spoke a different language at home too. In Finland, for example, 94% of primary school children who speak the language of instruction at home passed the second level a science assessment that just 58% passed among those who did not. Indeed, in many of these contexts students from immigrant families are among the most disadvantaged.
Overall there is a growing body of evidence that supports our recommendation: at least six years of mother tongue instruction is needed in order to reduce learning gaps for minority language speakers.
Getting to grips with mother-tongue instruction through data
The important role of language for achieving our global goal for equitable, inclusive and quality education is recognized in target 4.5, the target that looks at equity in education. One of the indicators set up to measure this target is the percentage of students in primary and secondary education, whose first or home language is used as language of instruction.
A good starting point to measure this indicator is to map language policies in official policy documents. A recent study in sub-Saharan Africa showed that at the time of countries' independence, 43% of sub-Saharan African countries used local languages in primary education, compared with 80% now.
However, it's not enough to look at policies, we must also monitor whether these policies are implemented, as Mali demonstrates. In Mali, in 2002, the government introduced a multilingual curriculum, introducing 11 national languages in addition to French. Yet, a decade after this reform, school-level implementation problems were considerable. In 2010, in the Mopti region, for instance, only 1% of schools were providing bilingual instruction by a trained teacher, in the appropriate language throughout primary school.
In many other multilingual countries few, if any, reading books are available in the languages children speak. The content of education, as the 2016 GEM Report has reiterated, must urgently be included in monitoring efforts if we are to reach our global targets in education and realize progress in other spheres of development.
Mother-tongue instruction doesn't just benefit learning but also protects knowledge
We recently celebrated International Mother Language Day, with the theme of Sustainable Futures through Multilingual Education, in alignment with our 2016 Report Education for people and planet: Creating sustainable futures for all. Our Report showed the importance of learning sustainable practices from indigenous communities in schools, which otherwise are in danger of being lost. Teaching in local languages is an effective way to impart traditional knowledge.
This is a significant concern. Research has documented how formal schooling has resulted in a significant loss of knowledge about nature, culture and values among indigenous children. Examples from Australia, Canada and the United States show an significant loss of indigenous knowledge from the beginning of the 20th century, when indigenous children were sent to residential schools or put up for forced adoption.
Yet by respecting local cultures and plural knowledge systems, and providing instruction in local languages, relevant knowledge can be shared among generations and communities both within and outside of school, and promote conservation of indigenous knowledge. For example, in Botswana, the Bokamoso preschool programme provides teacher trainees with a system of nature based educational tools incorporating the traditional knowledge of the San, a major indigenous group in the region.
Governments must do more to promote mother tongue learning
The continuing neglect of mother tongue instruction in linguistically diverse countries partly accounts for large disparities in education outcomes. While tracking language of instruction is fraught with technical and possibly political challenges, it is a key issue that governments must tackle head-on to protect indigenous knowledge and ensure no one is left behind.