Glossary
Non-native speech
Non-native speech is speech in a language that is not the speaker's first language, often called L2 speech or foreign-accented speech. Recognizers can transcribe read non-native English well, but they handle hesitations, repetitions and revisions differently, and results depend on the system.
Updated October 17, 20263 cited facts, each traced to its source
Also called L2 speech, foreign-accented speech.
Non-native speech is speech in a language that is not the speaker's first language. Researchers often call it L2 speech, for second language, and the everyday term is foreign-accented speech. The label describes the speaker's history with the language, not their skill. A fluent speaker who learned English at school still produces non-native speech when they talk in English, and the speech can carry a foreign accent and different patterns of pauses and word choice.
Non-native speech includes more than an accent
Non-native speech differs from native speech in several ways at once. Speakers may pause more, repeat words, revise a sentence in midstream or choose a word from their first language. These features are called disfluencies when they are not part of the message. A recognizer must decide whether to keep them, drop them or correct them, and that decision affects the text you get back.
Recognizers do well on read sentences, and results vary by system
A 2025 study used the L2-ARCTIC corpus of non-native English. On read speech, Whisper reached a mean match error rate of 0.054, and AssemblyAI reached 0.056, which the author describes as approaching human-level accuracy 1. Match error rate is not the same measure as word error rate, so compare it only with other match error rates.
Read sentences are a controlled test, and they leave out most of the hesitation found in real speech. In the same study, the systems varied significantly in how they handled filler words, repetitions and revisions 2. A tool that performs well on one test can behave differently on your spontaneous speech.
Pronunciation modelling was an early fix
Recognition of non-native speech was studied long before today's models. A 2007 study of foreign-accented English found that the best combination of pronunciation modelling and acoustic adaptation cut word error rate by 46 to 71 percent relative 3. A relative reduction is measured against the old error rate, so it does not mean the error rate fell by that many percentage points. The systems in that study are much older than current recognizers.
A concrete example: a second-language sentence with a repeat
Suppose a speaker who learned English as an adult dictates "I want, I want to send the file before noon", with a short pause and a repeated phrase. One tool may keep the repeat, another may drop it, and a third may rewrite the sentence as "I want to send the file before noon". Each choice is defensible, but the text differs. Check the output against what you meant, and fix the repeat by hand if it matters.
Why it matters for dictation and typing
If you dictate in a second language, expect more edits than you would make in your first language, especially for filler words and false starts. Set the locale to the language you are speaking, dictate short passages, and read the text back before you send it. Some learners also use dictation as practice, since a misrecognized word can show where a pronunciation differs from what the recognizer expects. Our guide on speech recognition and accents covers the fixes, and the guide on dictation for language learning covers the practice method.
Systems also treat filler words differently, as the 2025 study above found, so the same speech can come out with or without them.
Related terms
- Accent: the pronunciation pattern that often marks non-native speech.
- Dialect: a regional variety with its own vocabulary and grammar, which is a separate question from second-language use.
- Locale: the language and region setting that tells the recognizer what to expect.
- Automatic speech recognition (ASR): the system that turns the speech into text.
Guides and questions that cover it
- Guide: speech recognition and accents: accuracy and fixes.
- Guide: dictation for language learning: how and why it works.
- Question: how can non-native speakers get accurate English voice typing?
- Question: how do you practice pronunciation with voice typing?
- Overview: dictation languages, accents and voice translation.
Evidence
Every number on this page, with where it comes from. Open a fact to see the exact words in its source.
- 1.
- 2.
- 3.
Related terms
- Dictation Languages, Accents and Voice TranslationThe topic this term belongs to
- AccentAn accent is the way a speaker pronounces a language, shaped by where they grew up, their first language or their community. Accents change pronunciation, not vocabulary, and speech recognition error rates can differ sharply from one accent to another.
- Character error rate (CER)Character error rate (CER) is the number of character insertions, substitutions and deletions needed to turn a recognizer's output into the correct text, divided by the number of characters in the correct text. It is the usual measure where words are not separated by spaces.
- Code-switchingCode-switching is the habit of moving between two or more languages within a conversation or a single sentence. Speech recognizers built around one language often misread the switched words, so mixed speech is harder to dictate than speech in one language.
- DiacriticsDiacritics are marks added to a letter, such as the acute accent in é, the tilde in ñ or the vowel marks in Arabic, that change how a letter sounds or which word it spells. Speech tools and everyday writers often leave them out, so a dictated word can come out spelled differently from what you meant.
- DialectA dialect is a regional or social variety of a language, with its own vocabulary, grammar and pronunciation. Speech recognition is usually less accurate on dialects than on the standard form, because labeled training data for dialects is often scarcer.
- Input method editor (IME)An input method editor (IME) is software that supports text input for characters that cannot be typed easily on a standard keyboard, such as many East Asian characters. Speech input can also be an input method, and Android allows only one input method to be enabled at a time.
- LocaleA locale is a language combined with a region, and sometimes a script, such as en-GB for English as used in the United Kingdom. Dictation and speech settings use locales because the same language can differ in spelling, punctuation and speech patterns from one region to another.
- Low-resource languageA low-resource language is one with little digital speech and text data for training speech recognition and other language tools. There is no fixed cutoff, and a language can be well resourced for one task and poorly resourced for another.
- Modern Standard Arabic (MSA)Modern Standard Arabic (MSA), also called Fusha, is the standardized form of Arabic used in writing and formal speech, such as news broadcasts and official documents. Speakers use regional dialects in everyday conversation, and speech tools recognize formal Arabic more accurately than dialect speech.
- PinyinPinyin, also called Hanyu Pinyin, is the system that spells Mandarin Chinese syllables with the Latin alphabet and marks the tones with diacritics. People type Chinese by spelling the sound in Pinyin and choosing the character they want, and speech tools must choose among characters that share one Pinyin spelling.
- Right-to-left (RTL) scriptA right-to-left (RTL) script is a writing system, such as Arabic, Hebrew, Persian or Urdu, whose letters run from right to left. Text that mixes these letters with digits or Latin words is bidirectional, so software needs a specific algorithm to order and display it.
- Speech translationSpeech translation converts speech in one language into the same meaning in another language, as text or as synthesized speech. Most systems first recognize the speech and then translate the text, so a recognition error carries into the translation.