Skip to content

Preserve Heritage Language While Acquiring an Additional Language

Principle · All principles
Evidence · 32 claims (27 for, 5 mixed) · 23 studies (5 review, 5 theoretical, 4 quant-synthesis, 4 associational, 3 qualitative, 2 causal), q1–q3 · 3 of 23 report an effect size · 30 claims rest on one study

Conditional relationship

This page owns the decision about what happens to a learner's heritage (home) language while they learn the language of school or work. For a learner who arrives speaking a language other than the school or majority language, and who is acquiring that additional language, continuing to use and develop the heritage language (through instruction or support in it, literacy in it, and its continued use at home) is expected not to slow acquisition of the additional language, and, where the heritage language is developed to an academic level (reading, writing, school vocabulary), to give the learner something to transfer: concepts, literacy skills and word knowledge already held need not be learned again from nothing. The contrary policy, replacing the heritage language with the school language as quickly as possible, is expected to cost the heritage language without a matching gain in the school language. Outcomes are the learner's school-language achievement (general-achievement, usually standardised reading or literacy tests in pre-k-12), heritage-language proficiency, and the accuracy with which a bilingual learner's needs are identified (assessment-accuracy).

What tests this, and how far. No claim in the wiki tests heritage-language maintenance against its absence for comparable individual learners. The "no cost" half rests on associational findings reported second-hand in reviews: in a Manitoba study the share of instruction in French was unrelated to English achievement and strongly related to French achievement, and Spanish use at home raised Spanish skills with no English cost (Promoting minority children's L1 raises L1 achievement at no cost to L2 achievement [+M]); young Spanish–English bilinguals who knew more Spanish at school entry were more likely to have higher English skills at the end of the year (Spanish use at home does not limit English learning [+M]). The "transfer" half rests on Cummins's interdependence account and the studies it cites (L1 and L2 CALP are interdependent [+M]; older immigrant students with established L1 literacy developed L2 academic proficiency faster [+M]). The one randomised test in the wiki is programme-level: in the strongest study the What Works Clearinghouse reviewed, students assigned to dual language programmes in Portland scored higher on grade 5 literacy (effect size 0.14) (Dual language programs improved English literacy achievement [+M]), but not the English learners in that sample, whose literacy, science and mathematics differences were not significant (Literacy benefits of dual language programs did not extend to the English learner subsample [~M]). A non-significant difference is consistent with "no cost" but does not establish it. Several of these claims trace to the same source: the Finnish-migrant findings (Skutnabb-Kangas and Toukomaa, 1976) stand behind three of them (mother-tongue preservation, semilingualism and academic interdependence) through Cummins's reviews, so they are not independent confirmations.

The relationship is conditional on how far the heritage language is developed (the transfer argument needs academic, not only conversational, proficiency, and limited proficiency in both languages is reported as harmful: semilingualism [~M]), on whether the school language also gets strong, continuing support (minimal school-language literacy followed by an unsupported transfer is argued to be inappropriate: early-exit transitional programmes [+W]), on whether the planned allocation between languages actually happens (in one dual language programme English was given priority in practice despite guidelines: English priority over Spanish [~M]), and on the learner's age and literacy on arrival.

How the converted siblings sit beside this: Support Early Language Development owns responsive adult–child talk for building a first language in the early years; its routines apply in whichever language the family uses, and this page decides which languages they run in. Cultural and Life Experiences Connections owns using the learner's own experience as material for tasks; the heritage language is one of the things that page draws on. Accessible Vocabulary and Syntax owns keeping the school language of materials within reach while the learner acquires it.

Default design, while the relationship is untested

The earlier page's guidance, kept as a concrete default a teacher, programme lead or specialist can act on and revise. Each step says what its basis is; the minutes and frequencies are untested proposals.

  1. Find out the languages and how they are used, before planning. In the first week, ask the family (in their language, through an interpreter or bilingual staff where needed) which languages the learner hears and uses, with whom, and for what (talk, reading, religion, school work), and whether the learner reads or writes the heritage language. Record it; it sets every later step. Untested proposal; from the earlier page's requirement of "meaningful, ongoing opportunities to actually use the heritage language".
  2. Assess in both languages before deciding on support, placement or referral. Give a content assessment (mathematics, a story retell) in the heritage language as well as the school language, and do not read conversational fluency in the school language as academic proficiency. Basis: Spanish–English bilinguals performed better on kindergarten mathematics assessments when tested in Spanish [+M]; assuming conversational fluency indicates proficiency has led to misdiagnosis and premature exit from support [+W] (an author's argument, no data).
  3. Tell families to keep using the heritage language at home; never advise switching to the school language. Explain that home use of the heritage language has not been found to cost school-language learning, and that a shared language keeps parents able to help with school work. Basis: Spanish use at home does not limit English learning [+M]; Promoting minority children's L1 at no cost to L2 [+M] (home English use deteriorated Spanish with no English gain); the parent-involvement subclaim of the heritage-language hub [+M] (a practitioner account).
  4. Choose the strongest programme model the setting can staff. Where enough learners share a language and bilingual teachers exist, a dual language programme (both languages used for content, enrolment balanced between speakers of each) is the default. Basis: Dual language programs improved English literacy achievement [+M], qualified by the null for English learners [~W] and by how few studies meet standards (only two of 122 [~M]). Otherwise, school-language instruction with planned home-language support (steps 5–7). Untested as a comparison of the two.
  5. Give the heritage language a planned, daily role in school-language lessons. A working default: let learners plan, discuss or note-take in the heritage language for 5 to 10 minutes before each school-language writing or discussion task; provide bilingual texts or a bilingual glossary for each new topic; pair learners who share a language for the first talk about a hard text. Untested proposal for the doses; the strategies are Primary Language Support with Bilingual Texts and Bilingual Books and Cognate Attention.
  6. Develop heritage-language literacy, not only talk. Where the heritage language has a written form the learner can be taught, plan two or three sessions a week of 20 to 30 minutes of reading and writing in it (a heritage-language class, a community school, a bilingual aide, family reading with books sent home). Basis for literacy over talk: L1 and L2 CALP are interdependent [+M]; L1 academic proficiency speeds L2 academic proficiency [+M]. The doses are untested proposals.
  7. Teach the connections between the languages explicitly. In each unit, point out cognates, compare how the two languages express one idea (word order, gender, tense), and ask learners to say the same thing two ways. Basis: Fluent bilingualism is linked to metalinguistic awareness of how the two languages relate [+M] (case studies and a synthesis); the vocabulary-savings subclaim of the heritage-language hub [+M]. Untested as an instructional routine.
  8. Deliver specialist language intervention in both languages. For a learner with a language impairment that shows in both languages, have the speech-language specialist use the home language alongside the school language, with a bilingual assistant or the family where the specialist does not speak it. Basis: the Kohnert subclaim of the heritage-language hub [+M] (bilingual preschoolers, clinical study, no effect size).
  9. Do not withdraw school-language support at the moment of transition. When a learner moves from a bilingual or support programme into mainstream classes, keep school-language academic support (and heritage-language literacy) running through the next year at least. Basis: early-exit transitional programmes that transfer students with no support are inappropriate [+W] (an author's argument); teaching in a language students understand benefits language and broader learning [+M] (a World Bank narrative review). The one-year figure is an untested proposal.
  10. Check the allocation you planned is the one delivered. Once a term, observe or log the minutes actually spent in each language and which language carries the hard content; a programme can drift towards the school language without anyone deciding it. Basis: English priority over Spanish in practice [~M] (one qualitative observation).

Progression across years (untested arc). Early years (pre-kindergarten to about age 7): oral heritage-language development and early literacy in it where possible, school-language talk through play and stories, assessment in both. Middle primary: literacy in both languages, with weekly explicit cross-language comparison (step 7) and content previewed in the heritage language. Upper primary and secondary: content learned mainly in the school language, the heritage language kept as a resource for planning and reading and, where offered, as a subject taken for credit. A learner moves on from language support on academic school-language measures (reading and writing grade-level content), never on conversational fluency alone, and keeps a lighter form of support for a year after (step 9). For a newcomer aged about 10 or older who already reads in the heritage language, start content reading in it from the first week and expect faster progress than with a younger beginner (L1 academic proficiency speeds L2 academic proficiency [+M]).

When a condition is not met: if no adult in the setting speaks the learner's language, use bilingual texts, translation tools for glossaries, peers who share the language and the family as the heritage-language partners, and keep steps 2, 3 and 9, which need no bilingual staff; if the heritage language has no written form or no materials, develop it orally (retelling, discussion, family stories) and make the cross-language comparisons orally.

Fitting the design to a situation

The three facts that most change the decision:

  • How far the learner's heritage language is developed, and in which skills. A newcomer who reads and writes the heritage language at grade level can transfer that literacy; a child who understands but no longer speaks it, or who has had interrupted schooling, needs it developed before it can carry much. If unstated, ask: does the learner speak, read and write the heritage language, at what level, and when did schooling in it stop?
  • What the setting can staff and allocate. A dual language programme needs bilingual teachers and enough learners sharing one language; a class with one teacher and ten home languages can still assess in both, advise families, use bilingual texts and keep support after transition. If unstated, ask: who in the setting (teachers, aides, peers, families) speaks the learner's language, and how much time could go to it each week?
  • What the decision is for. Placement, referral for a suspected disorder, exit from support and day-to-day teaching each need a different move, and the evidence differs for each. If unstated, ask: is this about how to teach, how to assess, whom to refer, or when to end support?
If the brief says… Then change… Basis
Young child in pre-kindergarten or early care, family speaks another language at home Develop the heritage language orally through stories and play in it; tell the family to keep using it; assess early mathematics and language in both languages before any judgement Spanish use at home does not limit English learning [+M]; tested in Spanish, bilinguals scored higher in kindergarten mathematics [+M] (Spanish–English children; carried to other languages untested)
adolescent newcomer who already reads and writes the heritage language From the first week, give content reading and note-taking in the heritage language alongside the school language; expect faster academic progress than a younger beginner; keep heritage-language literacy going older immigrant students with established L1 literacy developed L2 academic proficiency faster [+M] (ages 10–12, reported second-hand)
Learner with interrupted schooling, low confidence or low literacy in both languages Start with oral work in the heritage language and the school language's most needed vocabulary; build literacy in the language the learner is closest to reading; do not withhold the heritage language to "save time" semilingualism [~M] (limited proficiency in both is reported as harmful; which language to build first is untested)
adult learners in work settings (newcomers learning the workplace language) Let them plan and take notes in the heritage language; build a bilingual glossary of the job's terms in the categories the workplace uses; treat their heritage language as a resource, not something to stop using at work untested; every claim here studied children or adolescents in school
A dual language programme is available or proposed Default to it where staffing allows; measure English learners' outcomes separately; log actual minutes per language each term Dual language programs improved English literacy [+M]; no significant literacy gain for the English learner subsample [~M]; allocation drifted to English in practice [~M]
classroom with one teacher who does not speak the learners' languages, or many languages in one room Pair learners who share a language for first-talk on hard tasks; supply bilingual texts or glossaries per topic; recruit family or community members to read in the heritage language; keep assessment in both languages through an interpreter for decisions untested proposal; the earlier page's requirement of "adult and peer models fluent in the heritage language"
Suspected language disorder or learning disability Assess in both languages before referral; separate attrition (loss of a language no longer used) from a disorder present in both; if intervention is needed, deliver it in both languages Kohnert subclaim of the heritage-language hub [+M] (bilingual preschoolers); conversational fluency misread as proficiency led to misdiagnosis [+W]
Deciding when to exit a learner from bilingual or language support Decide on academic reading and writing in the school language, not conversational fluency; keep lighter support for a year after exit premature exit into unsupported mainstream classes [+W]; early-exit without support is inappropriate [+W] (both an author's argument); the year is untested
Family wants English only, or the heritage language is treated as low-status Explain the no-cost findings and the value for parent help with school work; respect the family's decision and record it; avoid any school message that the heritage language is inferior L1 at no cost to L2 [+M]; treating the mother tongue as inferior impedes movement between home and school language [+W] (an argument, no data)
Instruction is remote, mixed (some at home, some in the room), or the family has only a phone Send heritage-language stories or bilingual audio by phone for family reading; let remote learners submit spoken heritage-language planning as voice notes; keep the same assessment-in-both-languages rule untested proposal
One session only (a workshop or single lesson) One move: let learners plan in the heritage language for 5 minutes, then do the task in the school language, with a bilingual glossary of the session's key terms; claim nothing about language maintenance untested proposal
Scarce facilitator time (volunteers, a busy specialist, a tutor checking in briefly) Keep the moves that need little time: families advised to keep the heritage language, one bilingual book bag sent home a week, assessment in both languages only for placement and referral decisions untested proposal
A fixed short time budget per lesson (e.g. 30 minutes) 5 minutes heritage-language planning in pairs, 20 minutes task in the school language, 5 minutes comparing one expression across both languages; minutes add to 30 untested proposal
The brief names its own routine (its own home-language survey, language allocation, rubric or entry assessment) Keep it; change only what the evidence speaks to: add a heritage-language version of the content check, add one cross-language comparison prompt, and log the minutes actually used in each language untested; the logging rests on allocation drift [~M]
The brief gives its own categories or criteria (a workplace's job terms, a programme's proficiency bands) Build the bilingual glossary and the assessment in those categories, not new ones untested proposal
The learner's language background, immigration or refugee status is sensitive or private The learner and family decide who learns which languages they speak; do not mark learners publicly; ask in private and keep records to those who need them untested proposal
High stakes: a placement or accountability test in the school language only Keep heritage-language instruction; add explicit practice in the test's language and format in the weeks before; read a low score as a school-language measure, not a measure of what the learner knows untested; the reading of the score rests on tested in Spanish, bilinguals scored higher [+M]
A product or performance for a real audience (a family evening, a community publication) Let learners produce it bilingually or in the audience's language, kept whole; use the heritage-language version for family audiences untested proposal; see Authentic Audiences and Purposes

Nine of the eighteen rows cite a claim; of those, the adolescent, early-years and assessment rows carry findings from Spanish–English or Finnish–Swedish learners to other language pairs, and every row that rests on an author's argument (exit, family attitudes, transition) says so. The adult, one-teacher classroom, remote, single-session, scarce-time, time-budget, categories, privacy and audience rows are untested proposals.

Observation, state and explanation

What can be observed is the learner's performance on a stated task in a stated language: what they say, read or write in each language, to whom, with what help, and how it compares with the same task in the other language. "Is losing the heritage language", "has academic proficiency" and "has a language disorder" are inferred states. Performance in the school language alone shows what the learner can do in that language, not what they know.

Response Plausible explanations Observation that could change the interpretation
Fluent in school-language conversation, weak on grade-level reading and writing Conversational proficiency ahead of academic proficiency; gaps in content knowledge; a reading difficulty Give the same content task in the heritage language and a matched grade-level school-language task; compare.
Says little or nothing in class An early stage of acquiring the school language; anxiety; the task is not understood; a language disorder Observe the learner talking with a fluent heritage-language partner; ask the family how the learner talks at home.
Low scores in both languages Little schooling or input in either; a disorder present in both; tests in a variety or dialect the learner does not use Check schooling history and the language variety; teach a short new item and see how quickly it is learned (a dynamic check).
Answers parents in the school language only, understands the heritage language Attrition through disuse; a status preference; peers' language Ask the learner what they use with whom and how they feel about it; test heritage-language production separately from comprehension.
Strong heritage-language writing, weak school-language writing on the same topic Transfer has not happened yet; school-language vocabulary missing; different genre expectations Ask the learner to plan in the heritage language and draft in the school language with a glossary; compare with an unplanned draft.

These rows are proposals, untested as diagnostics. Each probe is itself language use and changes what it measures, so record each one and keep explanations open until an observation separates them.

Evidence and qualifications

No claim in the wiki compares learners who kept and developed a heritage language with comparable learners who did not, on later school-language outcomes. The claims below test neighbouring parts of the relationship.

Learners, languages, countries and decades differ across these claims (Finnish children in 1970s Sweden, Franco-Manitobans, Spanish–English kindergarteners in the United States, a Portland lottery), so do not rank them by effect label. None measures heritage-language proficiency and school-language achievement in the same learners over time with a comparison group.

The evidence by question

Further evidence, not yet read against this model

Claims this page cited before, or found while rewriting it, that bear on it less directly.

Objective and learner-valued goal

The designer's objective is usually the school language: grade-level reading, writing and subject achievement in it, by a stated year. Elicit separately what the learner and the family want from the heritage language: to talk with grandparents, to keep a religious or community practice, to belong, to have a qualification in it, or, for some adolescents, to be seen as a speaker of the school language among peers and to leave the heritage language behind. Agreement and divergence are both observations. A learner who wants to fit in may refuse heritage-language work in front of classmates; a family that values the heritage language may withdraw from a school that seems to discourage it.

For example, a secondary school's objective for a 14-year-old newcomer may be passing an English-medium science course within two years, while the learner wants to keep reading and writing Arabic well enough to finish schooling abroad later and the parents want the family's language kept. Science notes and glossaries in Arabic and English serve both; record the learner's heritage-language reading as well as the English science result, so that progress in English is not read as the whole of what the learner came for.

What would revise this model?

The expectation should weaken if studies following comparable learners over years, with heritage-language use or instruction recorded and the same total instructional time, found that more heritage-language use or instruction predicted lower school-language achievement; or if the English-learner null in the Portland lottery study were replicated as a loss rather than no difference. If independent replications of the Finnish-migrant findings did not show the association, the "transfer" half rests on Cummins's account alone and should be stated as a hypothesis. If dual language programmes help native speakers of the school language but not heritage-language learners, programme choice should be justified on maintenance and identity grounds, not on school-language gains. Do not protect the model by calling every null result "too little heritage-language development" after the fact; record how much heritage-language use and instruction there was before the outcomes are known.

Heritage-language proficiency, school-language conversational fluency, school-language academic achievement, subject learning, accurate identification of disorders and the learner's own goals are separate outcomes. The present evidence sets no dose of heritage-language instruction, no age at which to add the school language and no exit criterion.

Claims

Examples

Key Sources

  • Hansen, L., Umeda, Y., & McKinney, M. (2002). Ragged interlanguage morphology and orders of acquisition issues. Journal of Multilingual and Multicultural Development, 23(1-2), 6-20.
  • Jimenez, R. T., Garcia, G. E., & Pearson, P. D. (1995). Three children, two languages, and strategic reading: Case studies in bilingual/monolingual reading. American Educational Research Journal, 32(1), 67-97. doi:10.2307/1163214
  • Francis, N. (2006). Bilingual competence and bilingual proficiency in children. In J. Kroll & A. De Groot (Eds.), Handbook of bilingualism (pp. 137-163). Oxford University Press.
  • Kohnert, K., Yim, D., Nett, K., Kan, P. F., & Duran, L. (2005). Intervention with linguistically diverse preschool children: A focus on developing home language(s). Language, Speech, and Hearing Services in Schools, 36, 251-263. doi:10.1044/0161-1461(2005/025)
  • Meyers-Sutton, C. (2005). Multilingualism. In J. L. Berko Gleason (Ed.), The development of language (pp. 407-444). Allyn & Bacon.
  • Arduini-Van Hoose, N. (2020). Language development. In Educational psychology. Retrieved from https://edpsych.pressbooks.sunycreate.cloud. CC BY-NC-SA 4.0.
  • Denise Santos & Gláucia Silva. (2015). Exploring Portuguese Heritage and Non-Heritage Learners' Perceptions of and Performance in Listening. The Canadian Journal of Applied Linguistics 18(1), 63-86. https://journals.sfu.ca/cjal/index.php/cjal
Page metadata
Field Value
type principle
id heritage-language-maintenance
aliases assess-hll-nhll-listening-differences-first
title Preserve Heritage Language While Acquiring an Additional Language
description For a learner acquiring a school or majority language, continuing to use and develop the heritage language is expected not to slow that acquisition and, at an academic level, to transfer to it; the support is associational and second-hand, and the one randomised programme-level test in the wiki found no significant difference for English learners.
status review
canonical True
generated by claude/unspecified on 2026-10-15
trust unverified — no one has confirmed this page's content
sources 1 — denise-santos-gláucia-silva-2015

Frontmatter is the page's machine-readable record — see the Schema & Guide.