Bilingual Education: Learning in Two Languages

Bilingual education means using two languages to teach school subjects — mathematics in Spanish and history in English, biology in French and literature in the home language. Language is not merely one more subject on the timetable; it is the medium through which content is learned. That single design decision separates bilingual education from ordinary foreign-language teaching, and it conceals a surprisingly political question: is the learner’s first language a temporary crutch to be discarded as quickly as possible, or a resource to be developed alongside the new one? Wallace Lambert named the two answers subtractive and additive bilingualism, and nearly every controversy in the field comes down to that distinction.

This entry is about the school models and what half a century of research says about them. For bilingualism as a property of an individual mind — its varieties and cognitive effects — see the entry on bilingualism. And despite the name, the bilingual method is something else again: a technique for using the mother tongue inside an ordinary language lesson, not a way of running a school.

The models: from transitional to two-way

Colin Baker’s standard typology sorts programs into weak and strong forms — weak forms use two languages but aim at monolingualism in the majority language; strong forms aim at lasting bilingualism and biliteracy. The main types:

Transitional bilingual education — the home language is used for the first year or three, purely as a bridge, then dropped as instruction shifts entirely to the majority language. The most common model for immigrant children in the United States, and a weak form: its intended endpoint is subtractive.
Maintenance (developmental) programs — minority-language children continue to study partly in their home language throughout schooling, so that literacy grows in both. A strong form, aimed at additive bilingualism.
Dual-language (two-way immersion) programs — classes deliberately mix majority- and minority-language children, and instruction is split roughly half-and-half between the two languages, so each group acquires the other’s language while its own is developed. The flagship American strong form since the 1990s.
Immersion — children of the majority language are schooled largely or entirely in a second language, as English-speaking Canadians are schooled in French. Because the home language is dominant outside school and never at risk, immersion is additive by design.
CLIL (content and language integrated learning) — the European branch: a few subjects taught through a foreign language inside an otherwise ordinary school. It has its own entry on content and language integrated learning.

CLIL deserves the separate word because its logic differs from classic bilingual education. The term was coined in 1994 (David Marsh and colleagues) as a deliberately broad umbrella, and the European Commission’s Eurydice survey of 2006 found some form of it in nearly every European country. The teaching language is usually a foreign language — overwhelmingly English — rather than a language of the local community, and it typically covers a slice of the curriculum rather than half of it: immersion’s machinery, scaled down and pointed at a lingua franca.

The Canadian experiment: immersion and its lessons

Modern immersion has a precise birthplace: St. Lambert, a Montreal suburb, in 1965. A group of English-speaking parents, convinced that conventional French lessons would leave their children functionally monolingual in an increasingly French-speaking Quebec, persuaded the school board to open an experimental kindergarten taught entirely in French — and had the foresight to invite the psychologists Wallace Lambert and G. Richard Tucker to study it. Their book, Bilingual Education of Children: The St. Lambert Experiment (1972), reported results that reassured everyone: the children’s French was far beyond anything ordinary teaching produced, their English was intact, and their achievement in school subjects matched their monolingually-schooled peers. Immersion spread across Canada, where it now enrolls hundreds of thousands of students, and became the most intensively studied program in the history of language teaching.

The fine print emerged after years, and it reshaped language pedagogy well beyond Canada. Immersion students’ receptive skills — listening and reading — approached native levels. Their production did not: they spoke fluent, confident French with persistent, fossilized errors in grammatical gender, verb morphology and idiom that further years of input did not repair. Studies by Birgit Harley, Merrill Swain and colleagues in the 1980s documented the gap, and Swain drew the theoretical conclusion in her comprehensible output hypothesis (1985): understanding messages does not force a learner to process grammar precisely — only being pushed to produce language does.

Immersion thereby became the decisive natural experiment for the era’s biggest theoretical claim — that comprehensible input alone suffices for acquisition. Thousands of hours of meaning-focused input produced excellent comprehension and permanently approximate grammar. The remedy that emerged — brief, systematic attention to form inside content teaching, elaborated for immersion in Roy Lyster’s “counterbalanced approach” — is the same focus on form that the immersion evidence pushed into communicative language teaching at large.

What the long-term research shows

For minority-language children the key evidence is longitudinal, because program effects reverse over time. The largest studies are Wayne Thomas and Virginia Collier’s analyses of hundreds of thousands of student records across U.S. school districts (1997; 2002). Their finding: English-only instruction and short transitional programs — the options that look most direct — closed at most about half of the achievement gap between English learners and native speakers by the end of schooling. The only programs whose graduates reached full parity with native speakers, and held it into secondary school, were the enrichment forms: one-way and two-way dual-language education. The strongest single predictor of eventual achievement in the second language was the amount of formal schooling children received in their first language.

That counterintuitive result has a mechanism, articulated by Jim Cummins as the interdependence hypothesis (1979): beneath surface differences, academic and literacy skills form a common underlying proficiency that transfers across languages. A child who has learned to read, reason and handle abstract vocabulary in Turkish does not relearn those skills in German — she relabels them. Schooling that develops the first language is therefore not time stolen from the second; it is investment in the shared foundation both languages stand on.

Cummins added a second distinction that explains why weak programs look better than they are: conversational fluency (BICS) arrives in a year or two, while academic language proficiency (CALP) — the language of textbooks, abstraction and argument — takes five to seven. A child exited from support after two years sounds fluent, and then sinks slowly against peers whose academic language keeps compounding. Early-exit programs harvest the easy gain and miss the one that matters.

Translanguaging and the contemporary turn

Classic bilingual programs guarded a strict separation of languages — one subject, one teacher, or one day per language — on the theory that mixing breeds confusion. Cummins mocked the result as the “two solitudes”: two languages living in one school, forbidden to speak to each other. The counter-movement began, fittingly, in Welsh schools: Cen Williams coined trawsieithu (1994) for a deliberate technique in which pupils receive input in one language and produce output in the other — reading in English, discussing and writing in Welsh — forcing deep processing of both.

Ofelia García’s Bilingual Education in the 21st Century (2009) generalized the idea into translanguaging as a theory: a bilingual does not house two sealed monolingual systems but one integrated repertoire, and pedagogy should draw on all of it — let the new language lean on the strong one, compare them explicitly, plan when each does what. The pendulum has not swung without resistance; researchers of minority-language schooling point out that a small language may still need protected space in which the majority language cannot crowd it out. But the field’s center has clearly moved: a learner’s other language is a resource to be orchestrated, not a contaminant to be excluded.

What this means for language learning

Three lessons travel from the school models to any adult learning a language. First, content works: learning something real through the language supplies the thousands of hours of meaningful input and genuine communicative need that no course of grammar exercises can match — the strongest practical form of what communicative teaching and input theory recommend. Second, input alone is not enough: immersion’s fossilization result shows that comprehension-first learning leaves production permanently approximate unless forms are noticed, practiced and corrected. Third, your first language is an asset — the interdependence evidence says the skills and knowledge behind it transfer, so using it deliberately (as translanguaging does) beats pretending it isn’t there. The practical synthesis is immersion plus system: surround yourself with content in the target language, and pair it with systematic production practice — active recall and spaced repetition of full sentences, which supplies exactly the pushed-output-and-form component that pure immersion lacks. What lasting bilingualism does for the mind is the subject of the companion entry on bilingualism.

Frequently asked questions

What is the difference between bilingual education and CLIL?

Scope and target language. Bilingual education is the umbrella: schooling that delivers subject content in two languages, often including a community or minority language, sometimes for half the curriculum or more. CLIL is its European specialization — usually a few subjects taught through a foreign language (most often English) inside an otherwise ordinary school. Every CLIL program is a form of bilingual education; the reverse is not true.

Does learning school subjects in a second language harm subject knowledge or the first language?

The evidence says no — with a condition. Canadian immersion students matched monolingually-schooled peers in subject achievement and English; Thomas and Collier found dual-language students ultimately outperforming comparable peers schooled in one language. The condition is that the first language is supported rather than abandoned: subtractive programs that drop the home language early forfeit the transfer effects that make the whole arrangement profitable.

Is immersion enough to reach native-like mastery?

For comprehension, very nearly; for production, no. The consistent immersion finding is near-native listening and reading alongside fluent but persistently non-native speaking and writing — errors that years of additional input do not fix. Native-like production requires what input alone does not force: being pushed to produce precisely, attention to form, and feedback. That is why modern immersion and CLIL build focus on form into content teaching rather than trusting exposure.

Sources

Wallace E. Lambert, G. Richard Tucker, Bilingual Education of Children: The St. Lambert Experiment, Newbury House, 1972.
Jim Cummins, “Linguistic Interdependence and the Educational Development of Bilingual Children”, Review of Educational Research 49(2), 1979.
Merrill Swain, “Communicative Competence: Some Roles of Comprehensible Input and Comprehensible Output in Its Development”, in S. Gass, C. Madden (eds.), Input in Second Language Acquisition, Newbury House, 1985.
Wayne P. Thomas, Virginia P. Collier, A National Study of School Effectiveness for Language Minority Students’ Long-Term Academic Achievement, CREDE, 2002.
Eurydice, Content and Language Integrated Learning (CLIL) at School in Europe, European Commission, 2006.
Roy Lyster, Learning and Teaching Languages Through Content: A Counterbalanced Approach, John Benjamins, 2007.
Ofelia García, Bilingual Education in the 21st Century: A Global Perspective, Wiley-Blackwell, 2009.
Colin Baker, Wayne E. Wright, Foundations of Bilingual Education and Bilingualism, 7th ed., Multilingual Matters, 2021.