Many learners aim to "lose their accent". It is an understandable goal, but research over the last thirty years suggests it is the wrong target for most people, and pursuing it can mean spending months on features that barely affect whether anyone understands you.
This article summarises what applied linguistics research has found about accent and intelligibility, and turns it into practical priorities. Sources are listed at the end for anyone who wants to read the studies themselves.
Three different things
A key study by Murray Munro and Tracey Derwing (1995) separated three things that are often treated as one:
| Term | Question it answers | How it is measured |
|---|---|---|
| Intelligibility | How much of what you said did the listener actually understand? | Listeners write down what they heard; the transcription is scored |
| Comprehensibility | How much effort did it take to understand you? | Listeners rate difficulty on a scale |
| Accentedness | How different does your speech sound from the listener's own variety? | Listeners rate strength of accent on a scale |
Their finding, confirmed in many later studies, was that these are related but partly independent. Some speakers were rated as having a strong accent yet were perfectly intelligible, with every word correctly transcribed. A strong accent did not predict poor intelligibility with any reliability. What mattered for understanding was a narrower set of features.
The nativeness principle and the intelligibility principle
John Levis (2005) described two competing goals in pronunciation teaching. The nativeness principle says learners should aim to sound like native speakers. The intelligibility principle says learners should aim to be easily understood, and that an accent is acceptable as long as it does not interfere.
Research on age and second-language speech helps explain why the second goal is more realistic. Very few people who start learning a language as adults become indistinguishable from native speakers, even after decades. But intelligibility and comprehensibility can improve substantially with focused practice, at any age. Derwing, Munro and Wiebe (1998) found that adult learners who received instruction focused on rhythm, stress and intonation improved in comprehensibility, while a group trained only on individual sounds improved on reading tasks but not in spontaneous speech.
A review by Thomson and Derwing (2015) of 75 pronunciation-instruction studies found that most reported improvement, but that many focused on accent features rather than on what affects intelligibility, and relatively few measured change in spontaneous speech.
Functional load: not all sounds are equal
Some sound contrasts distinguish thousands of common words; others distinguish almost none. This is called functional load. The contrast between /p/ and /b/ separates many everyday pairs (pie, buy; cap, cab; pack, back). The contrast between /θ/ and /ð/ separates almost nothing: thigh and thy, and a handful of others.
Work by Catford (1987) and Brown (1988) ranked English contrasts by functional load. High-load contrasts include:
- /p/ and /b/, /t/ and /d/, /k/ and /ɡ/
- /l/ and /r/, /l/ and /n/
- /f/ and /h/, /s/ and /ʃ/
- /e/ and /æ/, /iː/ and /ɪ/, /æ/ and /ʌ/
Low-load contrasts include /θ/ and /ð/, /ʊ/ and /uː/, and /θ/ and /f/. Later research by Munro and Derwing (2006) found that errors in high functional load contrasts affected comprehensibility more than errors in low-load ones. This is one reason the TH sounds, despite their fame, are rarely the best place to start.
English as a lingua franca
Most English conversations in the world today take place between non-native speakers. Jennifer Jenkins (2000) studied what caused misunderstandings in these conversations and proposed a Lingua Franca Core: the features that needed to be right for non-native speakers to understand each other. Her list included:
- most consonant sounds, though substitutes for TH were generally not a problem
- aspiration of /p t k/ at the start of stressed syllables
- not deleting consonants at the start of words (street must keep its /str/)
- the difference between long and short vowels, including shortening before voiceless consonants
- placing the main stress of a phrase on the right word
Features she found less important in this context included exact vowel quality, weak forms, the dark L, and stress-timed rhythm. Her conclusions are debated, especially for learners who mainly speak with native speakers, but they have been influential.
What this means for your practice
Pulling the research together gives a practical order of priorities for most learners:
- Word stress and sentence stress. Misplaced stress is one of the strongest predictors of comprehension difficulty. Start with word stress and sentence stress.
- High functional load contrasts that your first language merges, such as /iː/ and /ɪ/ or /l/ and /r/. Check the minimal pairs guides for the ones that apply to you.
- Consonants at the beginning and end of words, including clusters and grammatical endings like -ed and -s.
- Aspiration and vowel length as cues to voicing.
- Everything else, such as TH, dark L, weak forms and a specific accent model, if you want to sound more native-like after the first four are solid.
Some honest caveats
Intelligibility depends on the listener as well as the speaker. Listeners who are familiar with a particular accent understand it more easily, and listeners' attitudes can affect how much effort they are willing to make. Being intelligible does not protect anyone from accent discrimination, which is real and well documented.
And there is nothing wrong with wanting to sound native-like. Actors, some teachers, and many learners simply enjoy it. The point of the research is not that accent work is pointless, but that it is a different goal, and that intelligibility work should come first.
Frequently asked questions
Does a strong accent make you hard to understand?
Not necessarily. Studies by Munro and Derwing found that some speakers with strong accents were fully intelligible. Accent strength and intelligibility are related, but partly independent.
What pronunciation features matter most for being understood?
Research points to word and sentence stress, sound contrasts that distinguish many common words, and consonants at the beginning and end of words. Features like TH and exact vowel quality usually matter less.
Can adults lose their accent?
Very few adult learners become indistinguishable from native speakers, but most can become significantly easier to understand with focused practice. Intelligibility is a more achievable and more useful goal for most people.
Sources and further reading
- Munro, M. J. and Derwing, T. M. (1995). Foreign accent, comprehensibility, and intelligibility in the speech of second language learners. Language Learning, 45(1), 73–97.
- Derwing, T. M., Munro, M. J. and Wiebe, G. (1998). Evidence in favor of a broad framework for pronunciation instruction. Language Learning, 48(3), 393–410.
- Jenkins, J. (2000). The Phonology of English as an International Language. Oxford University Press.
- Levis, J. M. (2005). Changing contexts and shifting paradigms in pronunciation teaching. TESOL Quarterly, 39(3), 369–377.
- Munro, M. J. and Derwing, T. M. (2006). The functional load principle in ESL pronunciation instruction: An exploratory study. System, 34(4), 520–531.
- Thomson, R. I. and Derwing, T. M. (2015). The effectiveness of L2 pronunciation instruction: A narrative review. Applied Linguistics, 36(3), 326–344.
- Brown, A. (1988). Functional load and the teaching of pronunciation. TESOL Quarterly, 22(4), 593–606.
- Catford, J. C. (1987). Phonetics and the teaching of pronunciation: A systemic description of English phonology. In J. Morley (ed.), Current Perspectives on Pronunciation. TESOL.
Disclosure: this page links to Preply, a tutoring marketplace. If you book a lesson through that link we may earn a commission, at no extra cost to you. It does not change what is recommended here. See our disclaimer and privacy policy.
Measure your intelligibility
Read a passage and see how much a speech recogniser understood, word by word.
Take the voice test