In January 1970, the University of Ife in Nigeria started an experiment in primary education. Most schools that teach in a home language do it for a year or two, as a bridge into English. This one built no bridge. It taught in Yoruba for all six years of primary school, across every subject, and it taught English as a subject of its own, with teachers trained for that purpose. Textbooks were written in Yoruba. Researchers followed the children through 1978.
I start there because the experiment holds every question this season will carry. When a school chooses a language of instruction, what goes into the choice? The language children already understand, the language teachers are prepared to teach, or the language the examinations require? And who decided which of those counts most?
Fifty-five years later, the question is still open in Nigeria. In 2022 the federal government made home-language instruction national policy for the early grades. Then in November 2025 the next government cancelled it and restored English from pre-primary through university. Nigeria’s education minister pointed to poor results on the WAEC, NECO, and JAMB examinations. The Nigerian Academy of Education said the evidence showed the opposite.
The decision under the decision
A language policy is not a statement about words. It is a statement about whose knowledge the school expects a child to start from.
UNESCO estimates that about forty percent of the world’s people are not taught in a language they understand well. The number is a measure of how many countries have faced the question. They have answered it in different ways. Ethiopia adopted a policy in 1994 that let regions choose the language of their primary schools, so where Amharic was not the home language, a region could use another. Tanzania chose Kiswahili for public primary schools, a language widely shared though not every child’s first. Kenya’s policy supports home-language teaching in pre-primary and the early primary grades, and the practice has been harder than the policy.
On the Kenyan coast, researchers studying reading found that many teachers did not speak their students’ home languages. Lessons were taught in Swahili and English, though the policy called for something else. The language on the page of the policy and the language in the room were not the same language.
What does that tell us? Choosing a language is one part of the decision. A government also has to ask whether teachers speak it, whether books exist in it, and how teachers are assigned to schools. The research supports starting with the language a child already understands: the reading skills built there can carry into a second language. A policy still needs teachers, books, and time. Without them, what happens in the classroom may look nothing like what the policy intended.
Think about your own start for a moment. When you learned to read, was it in the language you spoke at home, or were you learning a new language and a new skill at once? If the two matched, you may never have had reason to notice the decision. Where they did not, you may remember carrying both.
Whose knowledge gets written into the lesson
Language is one question. A second sits beside it, and it is easy to run the two together: whose history and experience appear in what students learn.
In 2003, Brazil passed a law requiring African and Afro-Brazilian history and culture to be taught in elementary and secondary schools, public and private. The law changed the formal requirement for whose history belongs in the curriculum. Passing a law and carrying it out are different stages of the work.
The Moroccan case shows the gap between them. Morocco began teaching Tamazight, the language of the Amazigh people, in public primary schools in 2003, and its 2011 constitution recognized Tamazight as an official language. By 2025 the government reported that the language had reached about forty percent of primary schools, with a goal of all schools by 2030. Staffing is where the effort strains. Reports described roughly a thousand teaching positions filled since 2003, against an estimated need of seventeen to twenty thousand.
No child learns from a recognition. Children learn from teachers, and that means recruiting them, preparing them, and keeping them.
Spain developed its approach region by region. Catalonia’s 1983 language law made Catalan the main language of instruction, and later court decisions required at least a quarter of instruction in Spanish. Since 2004, welcome classrooms have helped students new to the region learn Catalan before they join regular classes. The Basque Country established three school-language models in 1982, and about eighty-three percent of children entering school now enroll in the Basque-language model.
In Aotearoa New Zealand, the push came from families. By 1975 only about five percent of Māori schoolchildren spoke te reo Māori. Families opened the first kōhanga reo, a Māori-language preschool, in 1982, and Māori-language primary schools followed in 1985.
These cases are not interchangeable. Some began in national legislation, some in regional authority, and some in families organizing a school of their own. The show has used a pair of terms for the difference: the grand narrative is the one a state writes about what children should know, and the mini narrative is the one a community keeps about what it already knows. Aotearoa New Zealand is the case where the mini narrative built the school. What the cases share is a lesson about timing. A law takes a year. Teachers take a generation.
What the decision asks of a child
Using one language for every child does not give every child the same experience.
Take two children in the same row of a second-grade room. The first speaks the school’s language at home. She can spend her attention on the lesson. The second is still learning that language, so she has two jobs: understand how the lesson is being said, and learn what it says. Nothing is wrong with the second child. The work she does is simply work the first child never has to do.
I have called this extra work the misrecognition tax, the cost students pay when a school does not readily recognize the knowledge they bring. Language is one place the tax is collected. A student can understand an idea and still struggle to explain it in the language the school requires, and that struggle can show up in a teacher’s judgment or a test score as a lack of ability. The home language is a fund of knowledge, in the sense the research gave that phrase. A school either draws on it or leaves it at the door.
The United States has been through this, and the record is not distant. In 1940, George Sánchez published Forgotten People, a study of Taos County, New Mexico. Among the problems he documented was the use of English-language intelligence tests with children who spoke Spanish at home. Schools read the scores as a measure of the child’s mind. They did not ask whether the child understood the language of the test.
Decades later, families in San Francisco challenged a related problem. About eighteen hundred Chinese-speaking students were receiving instruction in English with no added support. The case reached the Supreme Court as Lau v. Nichols in 1974, and the Court ruled unanimously for the students under the Civil Rights Act. The ruling did not require one particular language of instruction. It held that providing the same materials and the same teachers was not enough when students could not understand the language being used.
Equal materials are not equal access. They are the same lesson offered to children who are not standing in the same place.
This connects to the curriculum debates from Season 2. Language access decides how a student takes part in a lesson. Curriculum decides whose knowledge the lesson contains. They are separate decisions, and both shape who can take part in school at all.
Back to Nigeria
With that in view, return to the reversal.
The minister’s concern deserves a fair hearing. WAEC, NECO, and JAMB are the gates students must pass to move on, and poor results are a real cost to real children. A government that sees those results has reason to act.
The difficulty is the explanation. Those examinations are given in English, so the results raise more than one question. What did students learn? How well were they prepared to show it in English? And had the schools received what the language policy required of them? The Nigerian Academy of Education objected on those grounds. Participants at a national language-in-education summit called for the policy’s restoration, and named a short implementation period and a shortage of teacher preparation, textbooks, and other materials. The dispute was never only about whether home-language instruction is a good idea. It was about whether the conditions existed to judge the idea fairly.
The Philippines followed a similar course. Home-language instruction in the early grades became policy in 2009 and law in 2013. A law passed in October 2024 ended the universal requirement and made its use optional under specified conditions. The cases in Kenya, Nigeria, and the Philippines point to the same thing from different sides: what a school can offer depends on whether it has teachers who speak the language and the resources to support them.
No single reversal proves a policy failed, and no single success proves it will travel. But across these cases it is hard to separate a language policy from the system around it. When instruction changes and teacher preparation, materials, and examinations do not, the results have to be read with that gap in view.
Would it take root?
Each episode this season ends on one question: would it take root? It comes from Michael Sadler, who argued in 1900 that a country cannot lift a practice out of another country’s schools and expect it to work the same way without understanding what holds it up.
Three considerations come out of this episode. Children’s existing language knowledge belongs in the plan for how they learn to read and how they learn additional languages. Language policy depends on sustained investment in teachers and materials, and on enough time to see how implementation is going. And examinations need to be read alongside instruction, or a student’s performance in the examination language gets mistaken for a full measure of what the student has learned.
How a country meets them depends on where it stands. Ethiopia relies on regional decisions. Tanzania relies on a widely shared language. Catalonia works through its own institutions and laws. We can learn from each without assuming another country can copy the result by copying the policy.
If you teach, one place to begin is small. Find out which languages your students speak at home, and which of those languages are spoken by the adults in the building. That will not answer every question here. It will show where the school’s language resources meet its students and where they do not.
Language is one thing a teacher and a child can share. It is not the only one, and next week the question is what else counts when a teacher and a child are matched.
