How Content-Language Integration Works | Why English Learners Need Language Practice Inside the Subject

A student can understand a Science idea and still be unable to explain it in the language the classroom demands.

They may know what changed in an experiment, but not yet know how to say because, therefore, increased, remained constant, evidence suggests, or in contrast. They may recognise the answer in a diagram but struggle to write the sentence. They may follow a Mathematics procedure while missing the language that distinguishes more than from times as many. In History, they may know the event but not yet command the language of cause, significance, reliability or comparison.

This is why separating “English time” from “subject time” can create a false choice for multilingual learners. Academic language is not a detachable course that can be mastered first and then carried untouched into Science, Mathematics, Humanities and the Arts. Much of it is learned while doing those subjects.

The What Works Clearinghouse gives strong evidence to integrating oral and written English language instruction into content-area teaching for English learners, and 2026 IES resources have returned to that recommendation with implementation tools for classroom teachers and leaders.

The key question is not “Should every Science teacher become an English teacher?” It is more precise: which language features are necessary for students to participate in the actual intellectual work of the subject, and how can those features be taught without diluting the content?

The 50-second answer

Content-language integration means teaching academic English inside subject learning rather than treating language development as a separate preliminary stage. The teacher identifies the language students need to access and express the disciplinary idea, then builds structured opportunities to hear, say, read and write that language while learning the content.

The mechanism is use with meaning. A word or sentence structure becomes more learnable when it is repeatedly connected to a real conceptual job: comparing two quantities, explaining a causal mechanism, evaluating evidence, describing a trend, justifying a method or qualifying a claim.

The What Works Clearinghouse rates the recommendation to integrate oral and written English language instruction into content-area teaching as supported by strong evidence. This does not mean adding generic grammar worksheets to every subject. It means making the language of the discipline visible, modelling it, rehearsing it, and giving students repeated chances to use it in increasingly independent ways.

The destination is not simplified content. It is fuller access to the same content through stronger language.

1. Language is part of the task, not packaging around the task

Academic content is expressed through language. Even when the central idea is mathematical or scientific, students must interpret instructions, relationships, conditions, explanations and evidence. The language demand can therefore become a hidden prerequisite.

A learner understands that heating increases particle motion but cannot construct a sentence linking temperature, kinetic energy and collision frequency. The Science idea is partly present; the language for expressing the mechanism is not yet secure.

Teachers should not assume every wrong answer is a language problem. Content misconceptions and language gaps need different repairs.

2. Disciplinary language is more than vocabulary

Students need words, but they also need sentence structures, connectors, reference patterns and ways of organising explanation. Academic English includes how subjects make relationships visible.

In Science, because links mechanism; therefore marks consequence; whereas contrasts conditions. In History, although can hold concession and evaluation in one sentence.

Teaching isolated word lists without the relationships those words perform produces recognition without usable language.

3. Oral language is a rehearsal space for written language

Speaking allows learners to test ideas before they also manage spelling, punctuation and extended written structure. Structured oral practice can make later writing more independent.

Before writing a causal paragraph, students rehearse: ‘As temperature increases, particle motion increases, so collisions occur more frequently.’ Partners then vary the sentence with a new example.

Unstructured talk may let the most fluent speaker dominate. Oral practice needs roles, prompts or turn-taking when individual production matters.

4. Writing reveals whether language and content are integrated

A student may nod through an explanation yet struggle to produce the relationship independently. Short writing exposes what language is actually available.

After a graph lesson, ask for one sentence describing the trend and one sentence explaining a possible cause. The difference between description and explanation becomes visible.

Writing fluency can mask content weakness or vice versa. Teachers need to read for both dimensions.

5. A focused vocabulary set is better than a glossary dump

The WWC separately recommends intensive teaching of a focused set of academic vocabulary. Integration works best when teachers identify a small number of words that unlock the upcoming content rather than pre-teaching every unfamiliar term.

In a Geography lesson, density, distribution, concentrated, and sparse may be more important than twenty peripheral words in the textbook.

Too narrow a focus can also limit access. Students still need support with incidental vocabulary when it blocks comprehension.

6. Words should be taught across several encounters

One definition is rarely enough for academic vocabulary. Students need examples, non-examples, morphology, oral use, reading use and written use across several days.

Teach justify by showing what counts as a justification, comparing it with description, using it in teacher questions, and asking students to justify a fresh answer.

Repetition without variation can create memorised phrases. Change the context while preserving the word’s conceptual job.

7. Morphology can make technical language more transparent

Prefixes, roots and suffixes help learners connect unfamiliar words to known meanings. Disciplinary vocabulary often forms families that reveal relationships.

Connect predict, prediction, predictable, and unpredictable so students see how the base meaning survives grammatical change.

Morphology does not replace direct meaning. Some word histories are irregular and some roots mislead if taught mechanically.

8. Sentence frames can reduce linguistic load

A well-designed frame gives students grammatical architecture while leaving the important thinking unfinished. It can help learners express relationships they understand but cannot yet formulate fluently.

‘The evidence supports ___ because ___’ preserves the need to choose evidence and explain the connection.

A frame becomes harmful when it supplies the reasoning or remains permanent. Scaffolds should fade and vary.

9. Sentence stems should match disciplinary reasoning

Generic stems such as ‘I think…’ are often too weak for subject learning. The stem should encode the type of relationship students need to practise.

Mathematics: ‘This method is valid when ___ because ___.’ Science: ‘When ___ increases, ___ changes because ___.’ History: ‘Although Source A suggests ___, Source B indicates ___.’

Do not turn complex reasoning into rigid formulae. Stems are entry points, not final style.

10. Teacher modelling should expose language choices

When experts speak and write, many language decisions are automatic. Think-aloud modelling can make those choices visible.

The teacher says, ‘I am using whereas because I need to contrast two conditions in the same sentence. I am avoiding but because the relationship is more formal here.’

Do not overload the model with grammar commentary. Name only the features that support the learning goal.

11. Reading can become a language model

Content texts show how disciplinary ideas are packaged. Teachers can use short excerpts to notice vocabulary, sentence structure and information organisation before students produce similar language.

Students underline causal connectors in a Science paragraph, then rewrite one explanation using a different connector without changing the mechanism.

Texts should first be understood for meaning. Language analysis detached from comprehension can become empty feature spotting.

12. Listening matters too

Students encounter academic language through teacher explanation, peer discussion, video and oral instructions. Listening support can include previewing key terms, pausing at dense relationships and asking learners to reconstruct what they heard.

After a short explanation, students state the main causal link to a partner before the teacher continues.

Constant interruption can fragment meaning. Listening tasks need enough continuity for the idea to form.

13. Content teachers need language objectives that are small and useful

A lesson can hold a content goal and a language goal without becoming two lessons. The language objective should be the minimum linguistic capability needed to perform the content task well.

Content: compare two energy transfers. Language: use comparative structures accurately enough to explain one similarity and one difference.

Too many language objectives can dilute subject teaching. Choose the feature with the highest leverage.

14. The language objective should not replace the content objective

A multilingual learner deserves access to serious subject knowledge, not a simplified curriculum dominated by language exercises. Integration means language supports the content rather than displacing it.

Students analyse the same ecosystem relationship as peers, with a diagram, key vocabulary and oral rehearsal before writing.

Some learners may need temporary adjustments in text complexity or response length, but the conceptual destination should remain visible.

15. Background knowledge and language interact

Sometimes a learner appears to have a language problem because the content is unfamiliar. Vocabulary becomes easier when it attaches to a knowledge structure.

Teach evaporation inside an understood water-cycle model rather than as an isolated dictionary entry.

Do not assume rich language alone can compensate for missing subject knowledge. Both need development.

16. Home language knowledge can be an asset

Multilingual learners may possess the concept in another language before they possess the English label. Allowing students to connect languages can reveal knowledge and accelerate mapping.

A learner explains a scientific process first in the home language to a peer or through bilingual notes, then works toward the English terminology required in class.

Home-language support depends on context and available resources. Teachers should avoid forcing public use if students are uncomfortable.

17. Translanguaging can support conceptual continuity

Learners naturally use their full linguistic repertoire to think, plan and make meaning. Strategic translanguaging can keep the concept active while English expression develops.

Students annotate a diagram with mixed-language notes during planning, then produce the final explanation in the target English form.

The final language expectation still matters when the educational goal includes English proficiency. Translanguaging is a bridge, not a reason to abandon the target language.

18. Discussion should require academic work

Simply increasing talk time does not guarantee language growth. Students need a reason to use the target relationship and vocabulary.

Pairs compare two explanations and must agree which is more precise, using at least one target term and one causal connector.

Over-policing language during discussion can suppress ideas. Feedback should prioritise the language feature linked to the lesson.

19. Corrective feedback should preserve the idea

When a student expresses a strong concept with weak language, the teacher can recast or prompt without treating the whole response as wrong.

Student: ‘More heat, particles faster, more hit.’ Teacher: ‘Yes—the mechanism is there. Say it with as… increases… so…: As temperature increases…’

Correction should not erase student voice or turn every contribution into a grammar lesson.

20. Pronouns and reference chains deserve explicit attention

Academic texts frequently use pronouns and noun phrases to carry ideas across sentences. English learners may lose the referent even when individual words are known.

Track what this process, these results, or it refers to in a dense Science paragraph.

Reference instruction benefits all readers. Do not assume difficulties are unique to multilingual learners.

21. Nominalisation can hide actions

Academic writing often turns verbs into nouns: evaporate becomes evaporation, react becomes reaction, decide becomes decision. This compresses ideas but can make text dense.

Unpack ‘the expansion of the gas’ into ‘the gas expands’, then rebuild the nominalised form and discuss why a textbook might use it.

Students do not need linguistic terminology for every feature. The teaching should make meaning clearer.

22. Command words are language and assessment content

Words such as compare, infer, justify, evaluate, and describe encode what kind of answer is expected. A learner can know the topic and still lose marks by misreading the command.

Show two responses to the same content: one describes, one evaluates. Ask students to identify the difference in the thinking job.

Command-word drills should not become formulaic. The subject-specific meaning of the command matters.

23. Mathematics has language traps

Mathematical language contains ordinary words with specialised meanings and relational phrases that strongly affect the problem.

‘Difference’ can mean subtraction; ‘of’ can signal multiplication; ‘at least’ creates an inequality; ‘per’ encodes a rate relationship.

Language teaching should remain connected to mathematical structure. Memorising keyword tricks can produce wrong operations in unfamiliar problems.

24. Science requires mechanism language

Scientific explanations often need causal chains, conditions, variables and careful distinctions between observation and inference. The language makes those relations explicit.

Students move from ‘the water disappeared’ to ‘the water evaporated because energy transferred to the liquid increased particle motion, allowing particles to escape the surface.’

Longer sentences are not automatically better. Precision and causal structure matter more than impressive vocabulary.

25. History and Humanities require stance

Students must distinguish evidence from interpretation and express degrees of certainty, comparison and significance. Language choices such as suggests, indicates, may, largely, and however carry epistemic weight.

Compare ‘The source proves…’ with ‘The source suggests…, although its purpose limits…’ and discuss why the second claim is more defensible.

Excessive hedging can become vague. Students need to calibrate certainty to evidence.

26. English lessons also contain content-language integration

Even in English, students learn disciplinary language for analysing texts, constructing arguments and discussing grammar. Multilingual learners benefit when terms are attached to actual reading and writing moves.

Teach connotation by comparing word choices in a passage and asking students to use the term while explaining the effect.

Metalanguage should serve interpretation and craft, not become an end in itself.

27. Structured writing opportunities should be frequent

The WWC recommends regular structured opportunities for written language development, though the evidence rating differs across recommendations. Frequent short writing lets students practise academic forms without every attempt becoming a major assignment.

After a lesson, write two sentences: one describing the evidence and one explaining what it shows.

Quantity alone does not improve writing. Students need models, feedback and chances to revise.

28. Small-group intervention may be needed

Whole-class integration will not meet every learner’s needs. Some students require additional small-group instruction in literacy and English language development.

A learner who cannot decode the content text reliably receives targeted reading support while classroom teachers continue language-integrated subject instruction.

Additional intervention should connect to classroom learning rather than operate as an unrelated parallel curriculum.

29. Language supports should fade with proficiency

Students need opportunities to become less dependent on frames, word banks and teacher recasts. Fading reveals whether the academic language is becoming independently available.

A paragraph frame becomes a list of connectors, then disappears in a fresh task while the same reasoning demand remains.

Fading too quickly can turn support into a test of language rather than content. Use current evidence.

30. Assessment should distinguish language from concept when possible

A weak English response can hide strong reasoning; fluent English can hide weak content. Teachers should inspect both.

During a Science discussion, allow oral explanation with a diagram to diagnose the concept, then separately teach the written form needed for the final assessment.

Some assessments intentionally combine language and content. In those cases, both are part of the performance and should be taught accordingly.

31. Integrated language teaching benefits many students

Academic language is not only difficult for formally identified English learners. Native speakers may also be unfamiliar with disciplinary vocabulary, syntax and command words.

A whole-class lesson on the difference between describe and explain improves clarity for everyone while multilingual learners receive additional rehearsal.

Universal support should not erase targeted needs. Some learners require more intensive language development.

32. Teacher planning should identify language bottlenecks

Before a lesson, teachers can ask which words, sentence relationships or text structures could prevent students from accessing the concept. This makes language support proactive rather than remedial.

A History teacher identifies cause, consequence, immediate, long-term, and the contrast structure needed for the final paragraph before choosing scaffolds.

Do not annotate every sentence. Focus on bottlenecks that genuinely constrain the intellectual task.

33. School-wide language consistency can reduce friction

Students benefit when departments use stable language for recurring reasoning moves: claim, evidence, explanation; compare, justify, evaluate; variable, trend, limitation. Shared terms reduce relearning of labels.

Science and Humanities teams agree on a common meaning of evidence while preserving subject-specific examples.

Consistency should not flatten disciplinary differences. The same word can carry specialised meanings in different subjects.

34. The destination is disciplinary participation

The point of content-language integration is not perfect grammar before participation. Students learn academic language partly by using it to do real subject work.

A learner joins a Mathematics discussion with a sentence frame, explains a method, hears a peer’s alternative and revises the language in the next attempt.

Accuracy matters and should grow. Participation and correction can coexist rather than being sequenced as ‘language first, content later.’

35. Language routines should recur across lessons

Academic language becomes usable through repeated production, not one successful sentence. A stable routine can make practice cheap: listen to a model, rehearse orally, write briefly, compare, revise. The content changes while the language move remains familiar.

For several Science lessons, students use a cause–mechanism–effect routine with different phenomena. The structure stays stable, but the vocabulary and reasoning change.

Routine should not become formula. Once students can express the relationship flexibly, vary the structure and remove the scaffold.

36. Contrastive examples clarify language choices

Students often need to see two nearly acceptable sentences to understand why one is more precise. Contrast directs attention to the linguistic feature that changes meaning.

Compare ‘The population increased because there was more food’ with ‘The population increased as food availability rose, reducing competition for limited resources.’ Ask what explanatory work the second version adds.

More complex language is not automatically superior. The stronger sentence should be better because it expresses the relationship more accurately.

37. Teacher questions can build language without announcing a language lesson

Questioning can prompt students toward more precise forms during normal content teaching. The teacher can ask for comparison, cause, qualification or evidence rather than merely correctness.

Instead of ‘Is that right?’, ask ‘What evidence lets you say that?’ or ‘Can you state the contrast using whereas?’

Prompts should not interrupt every answer. Choose moments where the language form directly supports the disciplinary thinking.

38. Peer interaction should distribute language production

Pair and group work can increase opportunities to speak, but only if participation is structured. Otherwise one fluent student may produce the academic language while others listen.

Each partner must give one explanation, then paraphrase the other person’s reasoning before the pair writes a joint sentence.

Rigid turn-taking can feel artificial if overused. The structure is useful while equitable production habits are being built.

39. Recasts can preserve flow

A recast reformulates a learner’s response more accurately without stopping the conceptual conversation. It can model a stronger form while affirming the meaning that was already correct.

Student: ‘The metal got bigger because hot.’ Teacher: ‘Yes—when the metal was heated, it expanded because the particles vibrated more and occupied slightly more space.’

Recasts are subtle and may go unnoticed. Important language targets sometimes need explicit follow-up rather than indirect correction alone.

40. Visuals can support language by stabilising reference

Diagrams, timelines, tables and labelled models give students something visible to refer to while constructing sentences. This reduces memory load and can make pronoun and relational language easier to control.

Students describe a graph while pointing to the relevant interval, then write the same relationship without the visual support.

Visuals should not replace language production. The learner eventually needs to express the relation independently.

41. Content-language integration should include assessment preparation

High-stakes tasks often require students to demonstrate content through particular written forms. Teaching those forms is part of fair preparation, not coaching around the curriculum.

Before an examination requiring evaluation, students practise how claims, evidence, counterevidence and judgement are organised in the subject.

Preparation becomes weak when students memorise generic phrases without understanding how evidence changes the judgement.

42. Newcomer learners may need more explicit modelling

Students with very limited English proficiency can participate in rigorous content when language support is carefully sequenced. More modelling, gesture, visuals, bilingual resources and short oral repetitions may be necessary.

A newcomer learner points to the correct stage in a diagram, repeats the key relationship with a frame, then gradually expands the sentence over several lessons.

Support intensity should respond to proficiency and knowledge. One scaffold does not fit every multilingual learner.

43. Advanced multilingual learners still need academic language growth

Conversational fluency can hide persistent difficulty with dense academic syntax, disciplinary vocabulary and extended writing. Teachers may stop supporting learners too early because social English sounds fluent.

A student speaks confidently in discussion but struggles to qualify claims in formal History writing. Instruction shifts from basic access toward precision and register.

Do not treat advanced support as remediation. Academic language continues developing for all learners.

44. A strong school asks where language failure is predictable

Departments can map recurring language bottlenecks in their curriculum. Which command words are misread? Which sentence relationships collapse? Which technical terms create false familiarity? This turns language support into curriculum design.

A Science team identifies that students repeatedly confuse observation and explanation language, so the distinction is reinforced across units rather than repaired only before exams.

Maps should guide teaching, not produce an ever-growing checklist. Prioritise the language that unlocks important reasoning.

45. Language growth should be monitored through increasingly independent use

Progress is visible when students can use the target words and structures with less prompting, across different content and with greater precision. A copied sentence is not the endpoint.

Track a learner from sentence frame, to keyword prompt, to independent explanation across three new examples.

Do not require perfect grammar before recognising growth. The relevant question is whether language is becoming a more reliable tool for disciplinary thinking.

46. Families should not be asked to replace specialist support

Home discussion can enrich language and knowledge, but responsibility for evidence-based language instruction belongs with schools. Families vary in time, confidence and English proficiency.

Provide simple prompts or bilingual resources that help families discuss concepts without requiring them to teach grammar lessons.

Equity suffers when school success depends on parents reproducing professional instruction at home.

47. The strongest outcome is fuller participation

Language support has succeeded when learners can enter the same intellectual conversation with increasing independence: understand the text, ask a question, explain a mechanism, challenge evidence, justify a method and write for the subject.

A student who once contributed only isolated words now explains a comparison, responds to a peer and revises a paragraph using the same concept.

Participation is not merely speaking more. It is gaining access to the knowledge-making practices of the discipline.

Practical route for teachers

Before each lesson, write the content goal and then one language question: What must students be able to understand or say in English to show this content well? Select only the highest-leverage words, connectors or sentence relationships.

Model the language inside the content. Give students short oral rehearsal before extended writing. Use reading excerpts as language models. Check whether students can use the form in a fresh example, then fade support. When a response is weak, diagnose whether the bottleneck is concept, language, or both before deciding what to reteach.

Practical route for students

When subject learning feels difficult, ask whether the problem is the idea or the language used to express it. Collect useful sentence patterns, not only vocabulary words. Practise saying the relationship before writing it.

Build a small bank of disciplinary moves: The evidence suggests… because…; In contrast…; This method is valid when…; As X increases, Y… Then vary the content so the language becomes a tool rather than a memorised answer.

Practical route for parents and families

Families can support academic language by asking children to explain what they learned in complete relationships rather than recite definitions. Useful prompts include: What changed and why? How are these two things different? What evidence supports that? What would make the answer stronger?

Home languages are an asset. A child can discuss the concept in the language they know best and then work toward the English vocabulary required at school. The goal is connection, not replacement.

Common failure modes

Teaching long vocabulary lists without repeated use in real subject tasks.
Adding generic grammar worksheets that are disconnected from the content.
Lowering conceptual demand because English proficiency is still developing.
Using sentence frames that supply the reasoning instead of supporting expression.
Allowing the most fluent students to do all the oral language work.
Correcting every grammar error during discussion and suppressing participation.
Assuming every wrong answer from an English learner is a language problem.
Ignoring home-language knowledge that could reveal strong conceptual understanding.
Keeping scaffolds permanently after the language has become secure.
Planning content objectives without noticing the language bottlenecks hidden inside them.

Frequently asked questions

Should every subject teacher teach English?

Every content teacher teaches through language, so they should make the language needed for their subject’s intellectual work visible. That does not make every teacher a specialist language teacher or replace dedicated English-language support.

Is vocabulary instruction enough?

No. Vocabulary is important, but students also need sentence structures, connectors, text organisation, oral rehearsal and written practice that express disciplinary relationships.

Do sentence frames lower expectations?

They can lower or raise access depending on design. A good frame supplies linguistic structure while leaving the important reasoning to the learner, and it is faded as independence grows.

Should students be allowed to use their home language?

Strategic use can support conceptual continuity and planning. Final English-language expectations still matter when English proficiency is part of the learning goal.

Does this approach help students who are not English learners?

Yes. Academic language is unfamiliar to many students. Whole-class language support can improve access while targeted learners receive additional practice or intervention.

The final idea

Academic language is not a decorative layer placed on top of knowledge. In school, language is one of the main systems through which knowledge becomes visible.

A learner may understand more than they can currently say. The educational response should not be to lower the idea until it fits the language. It should be to strengthen the language until the learner can participate more fully in the idea.

That requires precision. Teach the words that matter. Model the sentence relationships that carry disciplinary reasoning. Let students rehearse orally. Give them structured writing. Correct without silencing. Fade support when the language becomes theirs.

The governing principle is: teach the language of the subject while students are doing the subject.