The Six Tones in Cantonese

In Brief

Cantonese uses six contrastive tones: the pitch pattern of a syllable helps determine the word itself. Jyutping writes the tone after each syllable as a number from 1 to 6, as in si1 詩 “poem” and si6 事 “matter.” A different number can mean a completely different word, even when the consonants and vowels stay the same.

Overview

A tone is a pitch pattern attached to a spoken syllable. In English, pitch often expresses attitude or marks a question, but it normally does not turn one dictionary word into another. Cantonese works differently: tone is part of a word’s pronunciation. Learning si without its tone is therefore like learning only part of the word.

This is an A1 foundation. You do not need a musical voice or perfect pitch; you need to hear and reproduce relative differences within your own comfortable speaking range. Cantonese speakers do not all use identical absolute pitches. What matters is whether a syllable is high, middle, or low compared with nearby syllables, and whether its pitch stays level, rises, or falls.

The six-tone analysis used here is the one represented by Jyutping, a romanization system for Cantonese. A Jyutping spelling separates the initial consonant, the vowel and final sound, and the tone number: soeng2, for example, ends in tone 2. Tone numbers are pronunciation information, not part of Chinese writing; ordinary Cantonese text is written without them.

How It Works

The six-tone system

Linguists often describe pitch with a five-point scale: 5 is the top of a speaker’s normal range and 1 is the bottom. These values are a map of relative pitch, not musical notes that everyone must match exactly. The labels below give a practical starting point.

Jyutping tone	Common pitch description	Approximate contour	Basic movement	Example
1	high level	55	stays high	詩 si1 “poem”
2	high rising	25	rises from mid-low to high	史 si2 “history”
3	mid level	33	stays in the middle	試 si3 “try”
4	low falling	21	begins low and falls	時 si4 “time”
5	low rising	23	rises within the lower range	市 si5 “market”
6	low level	22	stays low	事 si6 “matter; affair”

The six si words make a useful tone set because the consonant and vowel are held constant. Say them slowly in order: si1, si2, si3, si4, si5, si6. The spelling si remains the same, but the pitch pattern and meaning change.

Three level tones and three moving tones

A useful first division is between level and moving tones:

• Tones 1, 3, and 6 are level. Keep the pitch fairly steady at a high, middle, or low height.
• Tones 2 and 5 rise. Tone 2 rises farther and ends high; tone 5 remains in the lower-to-middle part of the range.
• Tone 4 is low falling. Start low and let the pitch fall slightly lower. It can sound less dramatically falling in connected speech, but it remains distinct from low level tone 6.

English speakers often make every stressed word rise or fall according to sentence melody. In Cantonese, the lexical tone—the pitch pattern belonging to the word—must remain recognizable even while the whole sentence has its own intonation. A question does not require changing every final tone into an English-style rise.

Reading Jyutping numbers

The number follows every romanized syllable:

• 香港 hoeng1 gong2 — “Hong Kong”: tone 1, then tone 2
• 廣東話 gwong2 dung1 waa2 — “Cantonese”: tones 2, 1, 2
• 唔該 m4 goi1 — “please; thank you” in service situations: tones 4, 1

Never read the number aloud as a separate number. goi1 means that goi carries tone 1; it is not pronounced as “goi one.” In study materials, some syllables may be separated by spaces and others grouped into words. Either way, each tone number belongs only to the syllable immediately before it.

Tone and stress are not the same

Stress means giving a syllable extra prominence, often through greater loudness or length. Tone means using a particular pitch pattern. Cantonese can emphasize a word, but emphasis does not remove its lexical tone. Speaking louder will not repair a wrong tone, and reducing an unstressed syllable as heavily as in conversational English can make the Cantonese vowel or tone unclear.

For a beginner, the safest habit is to learn each new item as one package: character, Jyutping syllable, tone number, and meaning. Learn 時 as si4, not merely as si plus a tone to be added later.

Examples in Context

The bold item is the main tone focus, but every Jyutping syllable has its own number. Read the Cantonese sentence naturally after first isolating the highlighted word.

Cantonese	Jyutping	English	Note
呢首詩好靚。	ni1 sau2 si1 hou2 leng3.	This poem is beautiful.	詩 uses high level tone 1.
我鍾意讀歷史。	ngo5 zung1 ji3 duk6 lik6 si2.	I like reading history.	史 rises in tone 2.
我想試吓。	ngo5 soeng2 si3 haa5.	I want to try it.	試 stays mid-level in tone 3.
我冇時間。	ngo5 mou5 si4 gaan3.	I don’t have time.	時 begins low and falls in tone 4.
街市喺前面。	gaai1 si5 hai2 cin4 min6.	The market is ahead.	市 has low rising tone 5.
呢件事好重要。	ni1 gin6 si6 hou2 zung6 jiu3.	This matter is very important.	事 stays low in tone 6.
佢今日返工。	keoi5 gam1 jat6 faan1 gung1.	They’re going to work today.	Notice the contrast between tones 5, 1, and 6.
你叫咩名？	nei5 giu3 me1 meng2?	What’s your name?	The question keeps the lexical tones 5–3–1–2.
我哋聽日飲茶。	ngo5 dei6 ting1 jat6 jam2 caa4.	We’ll have dim sum tomorrow.	A natural sequence containing all six tone numbers.
唔該，俾杯水我。	m4 goi1, bei2 bui1 seoi2 ngo5.	Please give me a glass of water.	Practise the low tone 4 before high tone 1.
佢講廣東話。	keoi5 gong2 gwong2 dung1 waa2.	They speak Cantonese.	Several rising tone-2 syllables occur here.
香港好熱。	hoeng1 gong2 hou2 jit6.	Hong Kong is very hot.	The sequence moves high, rising, rising, then low.

Do not expect the romanization to show English-style word stress. In gwong2 dung1 waa2, all three written tone numbers matter. Start slowly, then repeat the whole phrase without inserting pauses between its syllables.

Common Mistakes

Treating tone numbers as optional annotations

• Wrong: Learning 詩, 史, 試, 時, 市, and 事 simply as si.
• Right: Learn si1, si2, si3, si4, si5, si6 as six distinct pronunciations.
• Why: The segmental sounds—the consonant and vowel—are identical here. Tone carries the contrast in both pronunciation and meaning.

Confusing tone 2 with tone 5

• Wrong: Pronouncing 史 si2 with only a small low rise, like si5.
• Right: Let tone 2 rise clearly toward the top of your range; keep tone 5’s rise lower.
• Why: Both tones rise, so “rising” alone is not enough. Their starting area and especially their ending height differ.

Turning tone 4 into tone 6

• Wrong: Holding 時 si4 completely level at a low pitch.
• Right: Begin low and allow a fall; keep 事 si6 low and comparatively level.
• Why: These tones share a low register, but tone 4 moves downward while tone 6 is level in the basic citation form (the careful form of a word said by itself).

Using loudness instead of pitch

• Wrong: Saying 詩 si1 louder while keeping it at the same low pitch as 事 si6.
• Right: Place si1 high in your range and si6 low, without forcing either one to be loud.
• Why: Volume and pitch are different features. Cantonese tone contrasts depend primarily on pitch height and movement.

Adding English question intonation to the last word

• Wrong: Changing the final lexical tone drastically upward whenever a sentence is a question.
• Right: Keep the final word’s tone recognizable, as in 名 meng2 in 你叫咩名？
• Why: Cantonese questions use their own words and particles as well as intonation. Sentence melody may affect the overall delivery, but it does not license replacing the word’s tone.

Usage Notes

Tone descriptions are idealized targets. In fluent speech, neighboring tones, speed, emphasis, and sentence intonation influence the exact pitch you hear. A tone may occupy a smaller range than it does in a careful recording. Judge it within the phrase rather than expecting every token to match a tone-chart recording exactly.

Pitch range also varies by speaker. “High” does not mean one fixed frequency, and a lower voice does not have to imitate a higher voice. Establish three comfortable reference points—high, middle, and low—within your own range. Relative spacing is more useful than trying to copy another person’s absolute pitch.

Jyutping’s six numbers are especially convenient because the same numbers are used consistently in open syllables and in syllables ending in -p, -t, or -k. Other books or dictionaries may use tone marks or a different numbering convention. Do not assume that “tone 2” in every romanization system refers to the same Cantonese contour; check which system the source uses.

Beyond the Basics / Advanced Use

You do not need to master the details in this section at A1, but they explain forms you will meet later.

Checked syllables and the traditional nine-tone count

You may hear that Cantonese has nine tones rather than six. This is not necessarily a contradiction. Traditional descriptions count three additional checked tones on syllables ending in the unreleased stop consonants -p, -t, or -k. In Jyutping, their pitch categories are written with the existing numbers 1, 3, and 6, so the practical Jyutping system still uses six tone numbers.

For example, 一 jat1 “one,” 八 baat3 “eight,” and 六 luk6 “six” end abruptly in -t or -k. Their short, closed shape comes from the final consonant; their Jyutping tone numbers still identify the relevant high, mid, or low pitch category. Beginners should pronounce both parts: the final consonant and the tone.

Changed tones

Some words take a different tone in particular compounds, colloquial forms, or expressive patterns. This is often called tone change: a word’s expected dictionary tone changes in a recognized context. It is not permission to vary tones freely, and the patterns cannot all be predicted from spelling alone. Learn common changed forms as vocabulary first; study the broader rules only after the basic six-tone system is stable.

Tone in connected speech

Fluent Cantonese does not consist of six isolated pitch shapes pasted together. Pitch transitions are continuous: the end of one syllable leads into the start of the next. A sequence such as m4 goi1 moves from the bottom of the range to a high level tone, while hoeng1 gong2 moves from high level into a rising tone. Clear rhythm and transitions eventually matter as much as producing perfect isolated contours.

Practice Tips

1. Build contrast sets. Record si1–si6 in random order, then compare your recording with a reliable native-speaker model. Focus on one contrast at a time: 1 versus 3, 3 versus 6, 2 versus 5, and 4 versus 6.
2. Mark tones from the first review. Write Jyutping numbers on every new flashcard and say the whole syllable aloud. Cover the number later and retrieve it from memory; do not postpone tone learning until after vocabulary.
3. Move from words to short phrases. Practise a word alone, then in a two- or three-syllable phrase, and finally in a sentence. Shadowing—listening and immediately repeating—helps you copy both lexical tones and natural sentence rhythm without turning the exercise into singing.

Related Concepts

• Foundation: Jyutping Romanization — learn how initials, finals, and tone numbers represent Cantonese pronunciation.
• Later study: Tone Change Rules (Changed Tones) — explore contexts in which a word appears with a changed tone.
• Advanced reference: Dialectal Phonology and Sound Changes — examine regional pronunciation and broader sound variation.