How Instruction Language Shapes What Students Learn

A student can carry a full conversation in math class, hand up, correcting herself, asking sharp questions. In the next lesson, taught in a language she is still learning, the same student may go quiet for the whole period, though she is no less able than she was an hour before.

Anyone who has used a second language has probably felt a version of this: a conversation where the thought was ready but the words weren’t. In a classroom, that gap does something bigger than cause discomfort. It can decide who gets to show what they know.

A Classroom Goes Quiet

The drop can happen within a single class period.

a group of men sitting at desks in a classroom

Photo by Museums of History New South Wales on Unsplash

In that kind of lesson, hands that were up every few minutes can go down and stay down, and answers can shrink to one word, then to a nod.

What happens next matters more than the silence itself. A teacher who reads the quiet as a knowledge gap may respond by simplifying the content: easier problems, fewer steps, lower expectations. A teacher who reads it as a language of instruction issue, meaning the language the lesson is actually delivered in, can respond by supporting the words instead.

One response lowers the ceiling on what a child is asked to think about. The other keeps the thinking intact and repairs the channel it travels through.

Why Language and Content Get Confused

Subject knowledge and the words used to carry it are tangled together.

Young student focused on studying with textbooks, calculator, and notes on a desk.

Photo by https://kaboompics.com/ on Pexels

A student working out why ice floats has to hold the idea and the vocabulary at once: density, mass, volume. If the vocabulary is still new, the effort spent decoding words comes out of the effort left over for reasoning.

Assessment tightens the knot. Most tests ask students to show scientific understanding through language, so a shaky sentence can read as a shaky concept. Two separate skills get scored as one:

Content mastery: does the student understand the idea?
Language fluency: can the student express it in the test language?
What the grade usually records: a single number that blends both

What the Evidence Actually Shows

Research on this question is not unanimous, which is what makes it useful.

Person using calculator at desk with computer charts

Photo by Jakub Żerdzicki on Unsplash

A 2023 systematic review of studies in low- and middle-income countries found that teaching in students’ mother tongue improved letter knowledge, sentence reading, reading comprehension, and writing in that first language, while its meta-analysis found no effect on word reading1. The same review reported gains in literacy in the national language too, along with oral language, reading comprehension, and writing in a language learned later.

A 2025 review of 34 studies on multilingual teaching methods in primary schools found that the most prominent cognitive benefit was access: students could engage with the curriculum content more easily2. On the human side, the same review found the most frequently reported outcome was self-efficacy, a student’s belief that they can do the work, with engagement and identity affirmation also common2.

Achievement data points the same direction in some systems. A 2024 study using international test results from Emirati primary students found that children taught math and science in English had lower odds of reaching higher achievement scores than peers taught in classical Arabic3.

For a general reader, this means the language on the whiteboard is a variable in the score, sitting alongside effort and teaching quality.

The Case Against a Single Rule

The evidence doesn’t hand anyone a universal policy.

Students in a lecture hall working together, embodying diverse educational learning.

Photo by Yan Krukau on Pexels

A 2024 study of Spanish university business students taking an accounting course found the group taught in English performed better than peers taught in Spanish and concluded that, in that setting, English did not compromise learning of the academic content4. A small 2025 study of 34 primary pupils in Austria found children in a partial-immersion multilingual program had significantly better metalinguistic awareness, meaning the ability to think about how language itself works, along with stronger English5.

So the honest reading is about conditions, not slogans. Qualitative studies in the 2023 review pointed to high-quality teaching and learning materials in the students’ own language as a requirement for mother-tongue programs to work1.

Cost matters too: the 2023 review’s positive results came mainly from high-quality, more costly programs, and its authors describe the evidence base as small1.

Smaller Than an Overhaul

The 2025 review notes that, in most cases, bringing students’ languages into teaching does not require a major overhaul or a change in the language of instruction, and can build on existing resources and curricula2. The approaches it describes include textbooks with bilingual glossaries and lessons that compare vocabulary and cognates across languages2.

Across the studies here, the language a lesson was taught in was linked to what students learned and how much they took part. The direction of that link depended on the setting: English instruction went with lower odds of high scores for Emirati primary pupils and with better results for Spanish university students.