🔊 Minimal Pairs
Train your ear to distinguish between similar sounds. Click a word, listen to its pronunciation, and practice the difference.
Complement with Other Techniques
Minimal pairs training builds excellent listening discrimination. Combine it with tongue twisters and shadowing for comprehensive pronunciation mastery. See the complete training framework.
What a minimal pair is, and why it works
A minimal pair is two words differing in exactly one sound and meaning different things — ship and sheep, thin and tin, right and light. Because only one sound changes, every other variable is removed, which makes the pair a clean test of whether you can hear and produce that specific contrast rather than guessing from context.
Test perception before you practise production
If two English sounds belong to a single category in your first language, the difference is inaudible to you in your own speech as well as in other people's. Production practice under those conditions rehearses the error.
A two-minute diagnostic. Have someone read twenty words from a pair set in random order while you write down what you hear. Below 70% correct, you have a perception problem and listening drills come first. Between 70% and 90%, train both. Above 90%, the remaining work is motor and you should be recording yourself.
Core pair sets by contrast
| Contrast | Pairs | Typically hard for |
|---|---|---|
/iː/ vs /ɪ/ | sheep/ship, beat/bit, leave/live, feel/fill | Spanish, Italian, Japanese, Greek |
/æ/ vs /e/ | bad/bed, man/men, sat/set, had/head | Spanish, Portuguese, Japanese |
/ʌ/ vs /ɑː/ | cup/cop, luck/lock, nut/not | Most five-vowel languages |
/uː/ vs /ʊ/ | fool/full, pool/pull, Luke/look | Spanish, Japanese, Korean |
/r/ vs /l/ | right/light, rock/lock, pray/play | Japanese, Korean |
/θ/ vs /s/ vs /t/ | thin/sin/tin, thought/sought/taught | French, German, Russian, Portuguese |
/v/ vs /w/ | vine/wine, vest/west, veil/whale | Hindi, German, Mandarin |
/v/ vs /b/ | very/berry, vote/boat, van/ban | Spanish, Japanese, Korean |
/ʃ/ vs /tʃ/ | ship/chip, share/chair, wash/watch | French, Spanish, Portuguese |
Use several voices, not one
Training on a single speaker teaches you to identify that person's vowels rather than the English category, and the improvement does not transfer to new speakers. Using varied voices — different genders, ages, speaking rates and regional accents — produces slower initial gains but far better transfer and much better retention. This is the central finding of high variability phonetic training research, and it is the easiest thing to get wrong.
Move to sentences early
Word-level accuracy predicts conversational accuracy poorly, because the contrast has to survive while your attention is on grammar and meaning. Use frames where the wrong sound changes the sentence:
- I need a sheet of paper against I need a ship of paper.
- He's living in London against He's leaving in London.
- Can you fill the glass? against Can you feel the glass?