Chinese tones: why learners plateau, and how to get moving again

Here is a pattern we see constantly. A learner drills the four tones for months. In isolation they are perfect. Then they say a five-word sentence to a native speaker and get a blank look.

This is not a discipline problem. It is a curriculum problem. Tones in isolation and tones in connected speech are close to two different skills, and most material only teaches the first.

What a tone actually is

Mandarin has four tones plus a neutral tone. The usual description:

The critical point, and the one most often skipped: a tone is relative, not absolute. A first tone is not a specific musical pitch. It is high relative to the rest of your voice. This is why people with deep voices and people with high voices both sound correct, and why trying to match a recording's exact pitch is a dead end.

Plateau cause 1: the third tone is a lie

Textbooks teach the third tone as a dip-and-rise, falling low then coming back up. In real speech that full shape appears mainly when the syllable is alone or at the end of a phrase.

Everywhere else it is a low tone that stays down and never rises. Learners who produce the full dip-and-rise on every third tone sound slow and strange, and they lose track of the sentence because the shape takes so long to say.

Try it: say 很好 (hěn hǎo). If you produce two full dips, it takes about twice as long as a native speaker needs. Now say the first syllable as simply low, and the second with the rise. That is much closer.

Plateau cause 2: tones change next to each other

Chinese has tone sandhi: rules where a tone shifts because of its neighbour. Three of them matter enormously and are often taught as footnotes.

Third tone plus third tone

When two third tones meet, the first becomes a second tone. 你好 is written nǐ hǎo but said ní hǎo. This is not optional or regional; it is how the language works. Learners who say two third tones sound noticeably foreign.

不 (bù)

Normally fourth tone. Before another fourth tone it becomes second: 不是 is bú shì, not bù shì.

一 (yī)

On its own, first tone. Before a fourth tone it becomes second: 一个 is yí gè. Before first, second or third tones it becomes fourth: 一天 is yì tiān.

None of these are exotic. They appear in the first hundred words you learn. If nobody has told you about them, that alone can explain years of being asked to repeat yourself.

Plateau cause 3: practising syllables instead of pairs

Drilling single syllables trains a skill you never use. Words are mostly two syllables, and what your listener hears is the contour across both.

There are twenty combinations of the four tones in a two-syllable word. Practising those twenty pairs is far more useful than practising four tones. A few will feel impossible; those are the ones worth your time.

Find the two or three tone pairs you personally cannot produce, and drill only those. Most learners have a specific weak pair, often second-plus-second or third-plus-fourth, and fixing it changes their intelligibility more than another month of general practice.

Plateau cause 4: you cannot hear it yet

Production problems are frequently perception problems wearing a disguise. If you cannot reliably hear the difference between second and third tone in fast speech, you cannot correct yourself, and no amount of speaking practice will fix that.

Test yourself honestly: have someone say random syllables and write down the tone. If you score well below perfect, your bottleneck is your ear, not your mouth. Minimal-pair listening drills will move you faster than speaking drills.

What actually helps

• Learn the tone with the word, always. Never store a word in memory without its tone. A word learned toneless has to be relearned later, which costs more than learning it right.
• Record yourself and compare. Not to a native speaker's pitch, but to their shape and rhythm. Most learners are shocked by the gap the first time.
• Shadow whole sentences. Play a sentence, speak along with it, and match the melody rather than individual syllables.
• Exaggerate early. Overshooting the tones while learning is a much smaller problem than flattening them. You can dial it back later.
• Slow the audio, not your speech. Slowing playback keeps the tone contours intact. Speaking slowly yourself distorts them, because you insert full third tones everywhere.

The honest timeline

Tones take longer than anyone tells you, and they improve in steps rather than smoothly. Learners typically go from unintelligible to mostly-intelligible fairly quickly, then sit at "understandable but obviously foreign" for a long time.

That long middle is normal and is not a sign you are doing it wrong. It ends when tones stop being something you apply to words and become part of how the word is stored, which happens through volume of listening more than through drills.

Every word in KoiSpeak carries its pinyin with tone marks and audio, because we think the tone should never be optional at the point where you learn the word. Getting it right the first time is far cheaper than fixing it later.