Native Language Detector — Guess the L1 Behind the English
Every language leaves fingerprints on the English its speakers write. Paste a paragraph and this playful tool guesses the likely native-language family behind it — from articles, prepositions, and sentence shape.
How the guess works
When you learn English as a second language, the structure of your first language leaks through. Speakers of article-free languages (Russian, Mandarin, Korean, Japanese) tend to drop «a», «an», and «the»; Arabic speakers often over-use «the»; Romance-language speakers (Spanish, French, Italian) commonly pluralise uncountable nouns («informations», «advices») and write longer comma-joined sentences. This tool scans your text for a handful of those signature patterns and guesses the most likely language family. It is a heuristic party trick, not linguistics — a short or careful sample will often fool it.
Why it is just for fun
Real native-language identification is a hard research problem that needs far more text and far better models than a browser widget. This tool deliberately keeps it light: it reports a guess, shows you which patterns it spotted, and reminds you not to take it seriously. If your English is clean enough that it cannot decide, treat that as the compliment it is. The point is to make the invisible fingerprints of translation visible — which is the first step to smoothing them out.
From fingerprint to fluent
Whatever your first language, the patterns this tool detects are exactly the ones that make English read as non-native — and they are learnable, one at a time. Articles, prepositions, countable nouns, and sentence length each have clear rules. Diglot is built for precisely this: it understands the transfer patterns of your specific first language and helps you write English that no longer gives itself away.
Take it further with Diglot
This free tool runs in your browser. When you want to act on what it shows — fix the grammar, tighten the phrasing, or make your English read like a native wrote it — that is what Diglot is built for. The free tier is meaningful for daily writing, no card required.
Frequently asked questions
- Is the native-language guess accurate?
- No — it is a playful heuristic, not a real assessment. It looks for a few signature patterns (article use, pluralised uncountables, sentence length) and guesses a language family. A short, careful, or heavily edited sample will easily mislead it, and it can only guess broad families, not specific languages.
- What patterns does it look at?
- Mainly article usage (dropped or over-used «the»), pluralised uncountable nouns like «informations» or «advices», preposition choices, and average sentence length. These are common first-language transfer signals, but none is conclusive on its own.
- Why does it sometimes refuse to guess?
- If your English does not show strong transfer patterns — clean article use, natural sentence length, no telltale phrases — the tool has nothing to grab onto and will say so. That usually means your writing already reads close to native.
- Is my text stored or uploaded?
- No. The detection runs entirely in your browser. Nothing is uploaded or saved, so you can paste any text you like.
