Why are Chinese tones so hard?

Tones carry as much meaning as vowels, and they rarely sound like the chart. We counted how often.

By Mandarin The WildUpdated 13 October 20265 min read

We make Mandarin The Wild.

Tones carry as much as vowels

Mandarin has four tones and a light neutral tone. To an English speaker, pitch is for feeling: a question, surprise, doubt. In Mandarin it is part of the word. 买 mǎi is "buy" and 卖 mài is "sell".

我想买一本书。

mǎi, third tone: “I want to buy a book.”

买mǎi

Verb

to buy; to purchase

Also written 買

他在网上卖衣服。

mài, fourth tone: “He sells clothes online.”

卖mài

Verb

to sell

Also written 賣

A study that measured how much each sound contrast does in Mandarin found that getting a syllable's tone right is at least as important as getting its vowel right. Surendran and Levow (2004), The functional load of tone in Mandarin is as high as that of vowels, Speech Prosody 2004.

We counted the same thing in words. Of the most common words in Chinese web text, many have a common twin written the same in pinyin without tone marks. FineWeb-2, the web text we counted (its Mandarin part, cmn_Hani).

The most common words	Share with a common twin told apart only by tone
100	74%
500	46%
1,000	40%

买 is the 268th most common word and 卖 the 938th. 时间 shíjiān "time" is 111th, and 事件 shìjiàn "incident" is 675th.

The chart is not what you hear

A tone chart shows each tone said alone. In speech, a tone bends toward the tones around it. A study of two-syllable sequences found that a tone is pulled by the one before it and pushed by the one after it, so the same tone has different shapes in different places. Xu (1997), Contextual tonal variations in Mandarin, Journal of Phonetics 25.

Some tones change outright. A third tone before another third tone is said as a second tone, so 你好 is said "ní hǎo". 一 and 不 change with the tone that follows them. And many syllables lose their tone and are said light and short, like the 了 in 好了.

你好！我叫小明。

nǐ hǎo, said ní hǎo: “Hello! I’m Xiao Ming.”

你ní好hǎo

Adjective

hello; hi

We counted how often this happens inside words alone, before counting the changes between words such as 我很好.

On a page of Chinese web text	Share of the syllables
Neutral tone	7.0%
A third tone said as a second, inside a word	1.4%
一 and 不 changing tone, inside a word	1.1%
Not said with the tone the dictionary gives	9.5%

So about 1 syllable in 10 is said differently from its dictionary tone even before the changes between words. Our count reads each word with its dictionary's first reading, so the neutral tone, which speakers also use in words like 朋友 péngyou, is if anything undercounted.

Second and third tones look alike

The second tone rises. The third tone, said alone, dips and then rises. Both start lowish and end higher, and they are often called the most confusable pair. Listeners seem to tell them apart by when the pitch turns upward: early for a third tone, late for a second. Shen and Lin (1991), A perceptual study of Mandarin tones 2 and 3, Language and Speech 34(2).

The third tone rarely sounds like the chart, either. Before any other tone, it stays low and doesn't rise, and before another third tone, it becomes a second.

Which tone pairs you'll hear most

Half the words on a Chinese page have two syllables, so most listening is tone pairs, not single tones. Here are the five most common pairs and the rarest pair of full tones.

Tones in a two-syllable word	Share of two-syllable words	Example
4 + 4	13.8%	现在 xiànzài
1 + 4	10.3%	工作 gōngzuò
2 + 4	8.1%	学校 xuéxiào
3 + 4	7.4%	比较 bǐjiào
4 + 2	6.6%	问题 wèntí
3 + 1	3.2%	老师 lǎoshī

The fourth tone is the most common by far: 35% of syllables, next to about 20% each for the first, second and third.

Training works, and it lasts

In a well-known study, eight American learners of Mandarin practiced telling the four tones apart in real words spoken by several native speakers: eight sessions over two weeks. Their accuracy rose by 21% on average, the gain carried over to new words and new voices, and it had largely held six months later. Wang, Spence, Jongman and Sereno (1999), Training American listeners to perceive Mandarin tones, Journal of the Acoustical Society of America 106.

A follow-up asked native speakers to judge those learners' own tones. After the listening practice alone, their spoken tones were 18% easier to identify. Wang, Jongman and Sereno (2003), Acoustic and perceptual evaluation of Mandarin tone productions before and after perceptual training, Journal of the Acoustical Society of America 113. Hearing them better helped them say them better.

What to do with this

Learn words with their tones, from the start. 40% of common words have a tone-only twin, so a word learned without its tone is half learned.
Practice pairs, not single tones. Most words are two syllables, and tones change next to each other. The pinyin trainer drills tone pairs with real recordings, and the tone test shows the tones you hear best and worst.
Hear many voices. The training studies used several speakers on purpose. Listen to two voices for every word: every syllable in the pinyin chart has a male and a female voice.
Read and listen at the same time. The Library has graded stories from HSK 1 with recordings, and pointing at any word, on any page, shows its tones in that sentence. See the extension.

How we counted

We counted the Mandarin part of FineWeb-2, a large collection of web pages, cutting it into words with the engine behind our word cards. FineWeb-2, the web text we counted (its Mandarin part, cmn_Hani). Each word is read with the first reading in our dictionary (得 as the particle de, its usual reading), which covers 93% of the Chinese words counted; most of the rest are names. A "twin" is another word among the 5,000 most common with the same pinyin but different tones. Changes between words, such as 我很好, depend on the sentence and aren't counted here.

Questions

Why are Chinese tones hard for English speakers?

In English, pitch shows feeling. In Mandarin it is part of the word: 买 mǎi is "buy" and 卖 mài is "sell". Tones also change next to each other, so speech rarely sounds like the tone chart.

Which Chinese tones are the most confused?

The second and third tones. Both start lowish and end higher; listeners seem to tell them apart by when the pitch turns upward.

Can adults learn Chinese tones?

Yes. In one study, American learners improved their tone listening by 21% in two weeks of practice, and the gain lasted six months. Their spoken tones improved too.

Sources

Checked on 13 October 2026. Prices and features change; tell us if one has.