A minimal pair is two words that differ in exactly one sound and have different meanings. Ship and sheep. Bat and vat. Think and sink.
Minimal pairs are how phoneticians prove that two sounds are distinct phonemes in a language. They're also how learners discover that what sounds like "the same sound" to them is actually two different sounds that natives hear instantly.
Here are the 30 English minimal pairs that trip up learners most often, organized by the phoneme contrast they expose.
Why minimal pairs work for training
Most pronunciation practice gives you a target word and asks you to imitate it. That trains production — your mouth learning to do the thing.
Minimal pairs train perception — your ear learning to hear the difference. If you can't hear the difference between ship and sheep, your production will never be reliable. You'll guess.
The drill is simple: listen to both words said by a native speaker, identify which is which, repeat until your accuracy is above 90%. Then switch to production.
/ɪ/ vs /iː/ — short and long i
The single most common contrast English learners struggle with. /ɪ/ is short, lax, and slightly lower; /iː/ is long, tense, and higher.
- ship /ʃɪp/ vs sheep /ʃiːp/
- bit /bɪt/ vs beat /biːt/
- fill /fɪl/ vs feel /fiːl/
- list /lɪst/ vs least /liːst/
- rich /rɪtʃ/ vs reach /riːtʃ/
L1s that don't distinguish: Spanish, Italian, Greek, Arabic. If you say "feesh" for fish or "sheep" when you mean ship, this is the pair to drill.
/æ/ vs /e/ — open a and e
/æ/ has a wider mouth opening and the tongue further forward.
- bad /bæd/ vs bed /bed/
- man /mæn/ vs men /men/
- land /lænd/ vs lend /lend/
- had /hæd/ vs head /hed/
- pan /pæn/ vs pen /pen/
L1s that merge these: German, Russian, Turkish, Japanese. The trick is forcing /æ/ open — drop your jaw further than feels comfortable.
/ʌ/ vs /ɒ/ — central and back
/ʌ/ is the u in cut; /ɒ/ (British) or /ɑː/ (American) is the o in cot.
- cut /kʌt/ vs cot /kɒt/
- bus /bʌs/ vs boss /bɒs/
- luck /lʌk/ vs lock /lɒk/
- fun /fʌn/ vs fawn /fɔːn/
- sung /sʌŋ/ vs song /sɒŋ/
L1s that struggle: most non-Germanic. /ʌ/ doesn't exist in most languages — it's central, mid, unrounded. Closest to a quick "uh" but precise. It's also easy to confuse with its unstressed cousin — see schwa /ə/ vs /ʌ/.
/θ/ vs /s/ — the th problem
/θ/ has the tongue tip between the teeth. /s/ has the tongue tip near the alveolar ridge.
- think /θɪŋk/ vs sink /sɪŋk/
- thin /θɪn/ vs sin /sɪn/
- thumb /θʌm/ vs sum /sʌm/
- thigh /θaɪ/ vs sigh /saɪ/
- bath /bɑːθ/ vs bass /bæs/
L1s that substitute /s/ for /θ/: French, German, Russian, Japanese, Chinese. Look in a mirror. If you can't see your tongue between your teeth, you're saying /s/.
/v/ vs /w/ — the lip contrast
/v/ involves the lower lip touching the upper teeth. /w/ involves only lip rounding, no teeth contact.
- vest /vest/ vs west /west/
- vine /vaɪn/ vs wine /waɪn/
- very /ˈveri/ vs wary /ˈweri/
- vow /vaʊ/ vs wow /waʊ/
- vet /vet/ vs wet /wet/
L1s that merge these: German, Polish, Russian, Indian English, some Slavic languages. The fix is mechanical — lip to teeth or no lip to teeth. Practice with your finger on your lower lip.
/l/ vs /r/ — the East Asian challenge
In English, /l/ has the tongue tip touching the alveolar ridge with airflow around the sides. /r/ has the tongue tip raised but not touching.
- light /laɪt/ vs right /raɪt/
- long /lɒŋ/ vs wrong /rɒŋ/
- load /loʊd/ vs road /roʊd/
- lock /lɒk/ vs rock /rɒk/
- play /pleɪ/ vs pray /preɪ/
L1s that merge these: Japanese, Korean, some Chinese dialects. The merged sound in these L1s is /ɾ/ — neither /l/ nor /r/.
How to drill minimal pairs efficiently
A 10-minute daily routine:
- Pick one contrast (e.g., /ɪ/ vs /iː/). Don't bounce between contrasts.
- Listen to native audio of both words in a pair. Native voice, not synthetic if possible.
- Identify which is which when played in random order. Target 90% accuracy before moving on.
- Produce each word, recording yourself.
- Compare your recording to native audio. Pay attention to duration AND vowel quality, not just one.
- Repeat with the next pair in the same contrast group.
After one week of one contrast per day, switch contrasts. After 30 weeks (roughly two contrasts per week), you've covered the major English perception challenges.
IPAtics' Speech Analyzer generates Minimal Pair Anki cards automatically — both IPA forms on the front, both meanings on the back. Drilling becomes spaced repetition instead of active practice time. The compounding is significant after a month.
What minimal pairs don't fix
Two things:
1. Connected speech. Knowing /ɪ/ vs /iː/ in isolation doesn't help if you can't produce them at conversational speed. After perception is solid, drill them inside sentences — the shadowing technique is the most efficient way to train this.
2. Stress. Minimal pairs only address segmental contrasts. Word-stress and sentence-stress problems are separate. See the IELTS pronunciation guide for stress drilling.
The short version
If your English pronunciation feels stuck and you don't know why, run an audit on these 30 pairs. The ones you can't identify by ear are your weakness. Drill them with native audio until your accuracy is above 90%. Then production follows.
Download IPAtics free for macOS and Windows and the Minimal Pair card type makes this automatic. Or try the web version for one-off IPA lookups.
Related reading: IELTS pronunciation: the IPA sounds examiners listen for · How to read IPA — beginner's guide · How I built a pronunciation Anki deck in 10 minutes