Skip to content
Try Edumo!

Resources > AI Guide > Chapter 3

Chapter 3: Using the Learner's Native Language (L1) with AI

Language teaching went through a long period where the learner's native language was treated almost like an enemy. The idea was: maximum exposure to the target language, minimum use of L1. Some teachers were trained to never use the learner's L1 in class at all. Things have changed.

The shift: L1 as a tool to utilize

The research has moved on. Modern approaches, often grouped under the term translanguaging, recognize that a learner's native language isn't a separate box in their brain that competes with the new language. It's part of a connected linguistic system. Using L1 strategically can actually accelerate learning, not slow it down.

The key word is strategically. Nobody is suggesting you conduct entire lessons in the learner's L1. The point is that there are specific moments where L1 support makes learning faster, less frustrating, and more effective:

  • Explaining complex grammar. Some concepts are genuinely hard to understand in a language you're still learning. Also, few learners are interested in learning the vocabulary surrounding grammar in the target language, except maybe linguists and teachers :).

  • Vocabulary acquisition. Research consistently shows that L1 translation helps with initial word learning. Cognates (words that are similar across languages) are a shortcut; false friends are a trap that's best addressed explicitly.

  • Reducing anxiety. Especially at lower levels, knowing that L1 is available as a safety net helps learners take risks in the target language.

Contrastive analysis: predicting errors before they happen

One of the most valuable applications of L1-aware teaching is something linguists call contrastive analysis: comparing the structures of two languages to predict where learners will struggle. The logic is simple: where L1 and L2 work similarly, learning is easier. Where they differ, errors are likely. Research suggests that teachers can predict 60-70% of learner errors this way.

A Spanish speaker learning English might drop subject pronouns or place adjectives after the noun. A Japanese speaker might struggle with articles entirely, since Japanese doesn't have them. Each pair of L1 and target language creates its own pattern of predictable errors. It extends to pronunciation and listening too. Speakers of different languages may struggle to hear and produce different sounds, e.g. Japanese speakers with L and R ("light" vs "right") or Spanish speakers with short and long vowels ("ship" vs "sheep").

If you share your learner's L1, you probably know many of these patterns already. AI can help you be more systematic about it, generating a comprehensive overview instead of relying on what you notice lesson by lesson. If you don't share the L1, or if you have students with many different native languages, AI makes this kind of analysis possible. You don't need to be a linguist to use this. You just need to ask:

Example prompt:

I'm teaching English to a B1-level student whose native language is Turkish. Next week we're covering the present perfect tense.

What aspects of present perfect will be particularly difficult for Turkish speakers, and why? Briefly explain how Turkish handles similar concepts, and suggest 3 specific errors I should watch for.

You can use the result to inform your own explanations and exercises, or take it further and ask the AI to generate those directly. For example, exercises that specifically target the errors it identified.

Cognates and false friends

Cognates are words that look or sound similar across languages and mean the same thing. They're a learner's best friend. English "telephone" and Spanish "teléfono." English "music" and German "Musik." These similarities give learners a head start.

False friends are the evil twin. Words that look similar but mean something different. English "sensible" and French "sensible" (which means "sensitive"). English "gift" and German "Gift" (which means "poison"). These are responsible for some of the most memorable (and most embarrassing) learner errors.

AI can generate comprehensive cognate and false friend lists for any language pair, complete with example sentences showing the correct usage in each language. If you know the language pair well, you can probably list the obvious ones yourself, but AI will surface ones you haven't thought of. If you don't know the L1, AI can list them and help your learners.

Example prompt:

Create a list of 10 false friends between English and Italian for a B1-level learner.

For each one, provide:
1. The English word and its meaning
2. The Italian word that looks similar and what it actually means
3. An example sentence in English using the word correctly
4. A brief note on how to avoid the confusion

The honest caveats

If you share or are fluent in your learner's L1, you can verify and refine what the AI produces, much like you would any other generated material. The picture changes when you don't speak the L1. You can't check whether an explanation is accurate, and the AI might make subtle errors: a wrong translation, an oversimplified comparison, a cultural nuance that's off.

A few things that help when you can't verify the output yourself:

  • Be transparent with learners. Say something like: "I generated this explanation in your language to help you understand the concept. If something doesn't sound right, let me know and we'll figure it out together." This honesty actually strengthens the teacher-student relationship rather than undermining it.

  • Let learners be the quality check. Asking a learner "does this explanation make sense in your language?" is itself a useful exercise. It requires them to engage with the grammar concept and think about it critically. If you have many learners with different L1s, this collaborative approach may also provide learnings from one L1 useful to another. 

  • Know that AI quality varies by language. AI is generally strongest in major world languages (English, Spanish, French, German, Chinese, Japanese) and less reliable in languages with less training data. If you're generating L1 content in Tagalog or Georgian, be extra cautious and rely more on the learner's feedback.

  • Use L1 as a bridge, not a destination. The goal of L1 support is to help the learner understand something faster, so they can then practice it in the target language. If L1 explanations start replacing target language practice rather than supporting it, you've gone too far.

Because L1 use touches nearly every teaching skill, you'll find "L1 Support" sections in most of the chapters in Part 2 with specific prompts for each skill area.