Learn one sound at a time — and repeat it until it sticks

The biggest mistake in learning pronunciation is trying to fix everything at once. The effective method is the opposite: pick one sound, master it until it's automatic, and only then move on. Here's why it works — and how to do it well.

Pronunciation is a motor habit, not knowledge

Knowing that /θ/ is said with the tongue between the teeth is one thing. Doing it by reflex, in the middle of a fast sentence — that's something else entirely. Pronunciation isn't knowledge you "understand". It's a muscle habit, like riding a bike or a tennis stroke.

And a motor habit is built in one way: by repeating the same movement until you stop thinking about it. Not through variety. Not by "sampling" ten sounds in one session.

Why "a little of everything" doesn't work

When one session has you practise /θ/, then /æ/, then intonation, and finally two diphthongs — your brain doesn't have time to set any of them. Each sound gets a few repetitions and vanishes. It's an illusion of progress: you feel you "covered a lot", but nothing became a habit.

On top of that there's interference — freshly practised sounds blend into each other if you learn them at the same time. Spacing them out gives the brain a chance to set each one separately, without disruption.

What an effective session looks like

A good session is short, narrow and repeatable. It looks roughly like this:

• Pick one sound. Ideally the one that gives you the most trouble — for many learners that's /θ/, /æ/, /ɜː/ or the /ɪ/–/iː/ pair.
• Understand the mechanics. Where the tongue goes, what the lips do, whether there's voice. One minute of theory is enough — the rest is practice.
• Repeat in isolation. First the sound alone, then in simple words, finally in minimal pairs (think–fink, ship–sheep).
• Record and compare. Set your recording next to a model. The ear catches a difference your mouth can't feel.
• Come back tomorrow. The same sound, a short session. Repetition over time beats the length of a single workout.

How long it really takes

The most common question is: when will there be a result. The honest answer depends on what we count as a result. Producing the sound correctly in an exercise usually comes within one or two sessions — you just have to understand the mechanics and repeat it a few times. That's the easy part.

Harder is automation — the moment the sound comes out on its own, without thinking, in a random word at a normal speaking pace. That's a matter of days and weeks of short, regular repetition — and it's exactly the stage most people skip, treating a sound as "done" after the first successful attempt.

So it's worth separating those two thresholds in your mind. The first only means you know how to do it. Only the second means you no longer have to remember it.

The exception: contrast pairs

There's one situation where you practise two sounds at once — and it's deliberate. That's minimal pairs, where the contrast is precisely the point. Practising /ɪ/ vs /iː/ (ship–sheep) or /θ/ vs /f/ (think–fink), you learn to hear and produce the difference between them.

This isn't "two sounds at once" in the sense of splitting your attention — it's a single learning unit, where two sounds define each other by contrast. Contrast sharpens both hearing and pronunciation faster than practising each one separately.

When to move to the next sound

A simple test: when a sound comes out without conscious thought, in a random word you didn't practise before — it's ready. Then you move on. If you have to stop and "set" your tongue, stay with it for a few more sessions.

This method seems slow, because you're doing "so little". In reality it's the fastest — because you don't drill anything blind and you don't have to unlearn mistakes later. A drilled-in mistake costs twice as much work later as learning it right the first time.

Three mistakes that break this method

The principle itself is simple, but it's easy to break without noticing. The most common slips look like this:

• Practising without a model to listen to. Repeating a sound from memory reinforces what you think you hear — not what's in the original. The model has to be on hand at every session.
• Stopping at single words. A sound mastered in isolation can vanish in a sentence, because neighbouring sounds change the tongue position. Without moving on to phrases, the habit stays unfinished.
• Switching sound after the first success. One successful repetition isn't a habit yet. If you move on right away, a week later you're back to square one.

There's one common denominator: all three come from rushing. The one-sound method works precisely because it deliberately gives up that rush.

FAQ