Key takeaways
- Cantonese and Mandarin are not mutually intelligible when spoken, so study material written for Mandarin will not teach you Cantonese.
- Jyutping numbers six tones from 1 to 6, and the number is part of the word rather than an accent you can add later.
- Traditional characters are standard in Hong Kong and Macau, while Guangdong uses simplified characters — pick the set that matches where you are headed.
- Written Cantonese and Standard Written Chinese are two different reading registers, and beginners meet both within weeks.
- The US Foreign Service Institute places Cantonese in its hardest category at 88 weeks of full-time instruction, which is a useful scale check even though it describes intensive diplomatic training rather than app study.
Two things put people off Cantonese before they have written a single character. The first is being told it is "a dialect", which makes learning it sound like a detour off the main road. The second is the tone system, usually introduced in a way that suggests you need perfect pitch to order lunch.
Both deserve a straight answer, because the decision you are actually making — Cantonese or something else, now or never — is worth making on accurate information.
"Dialect" is the wrong word for what you are choosing
Cantonese and Mandarin sit inside the same language family and share an enormous amount of written vocabulary. What they do not share is mutual intelligibility in speech. A monolingual Mandarin speaker dropped into a Cantonese conversation does not follow it, and the reverse is equally true.
That has one concrete consequence for you: material written for Mandarin will not teach you Cantonese. Pinyin does not encode Cantonese sounds. Mandarin's four-tone system is not Cantonese's six. Mandarin grammar words — the ones you use in every sentence for "is", "not", "he" and possession — are frequently different words in spoken Cantonese, not just different pronunciations of the same word.
This is the practical reason a Cantonese course has to be built as a Cantonese course rather than adapted from Mandarin, and it is the design decision behind Learn Cantonese for Beginners: the sound system, the grammar sequence and the character set are all built around Cantonese rather than borrowed and relabelled.
Where Cantonese sits on the difficulty scale
The most defensible public benchmark for language difficulty is the US Foreign Service Institute's training categories, which group languages by how long its students typically need to reach Professional Working Proficiency — a score of Speaking-3/Reading-3 on the Interagency Language Roundtable scale. FSI places Chinese – Cantonese in Category IV, its hardest group, at 88 weeks of full-time instruction (2,200 class hours), alongside Arabic, Mandarin, Japanese and Korean.
Read that carefully before it discourages you. It describes intensive, full-time, small-group classroom training for diplomats — 2,200 hours across 88 weeks is about 25 class hours a week, before homework. It is not a statement about how long a phone app takes, and it is not a target for someone who wants to greet their partner's family properly. What it is good for is calibrating expectations: Cantonese is at the demanding end, so a course that promises speed is promising something the evidence does not support.
The tone system, described honestly
Here is the part that gets over-dramatised.
Cantonese uses pitch to distinguish words. The same string of sounds, said with a different pitch pattern, is a different word — not a differently-emphasised version of the same one. Say the syllable si six ways and you get six unrelated meanings. Tone is not decoration you can add once your vocabulary is bigger; it is part of the word, in the same way that the vowel is.
Modern descriptions of Hong Kong Cantonese count six contrasting tones, and Jyutping numbers them 1 to 6:
| Jyutping number | Usual description |
|---|---|
| 1 | high level |
| 2 | rising |
| 3 | mid level |
| 4 | low falling |
| 5 | low rising |
| 6 | low level |
Descriptions differ slightly between references — some transcribe tone 1 as high falling in certain speakers, and the exact pitch heights vary by region and individual — so treat the table as a map rather than a measurement.
You will also see Cantonese described as having nine tones. That figure comes from an older tradition that counts syllables ending in -p, -t or -k as three extra "entering" tones. Those syllables are cut short by the final consonant, but their pitch levels are the same ones already used by tones 1, 3 and 6, which is why Jyutping writes them with those numbers instead of adding three more. Six and nine are two ways of counting the same system, not a disagreement about the facts.
What this means for how you practise
The useful takeaway is not "tones are hard". It is that tone has to be learned with the word, from the first exposure, because relearning a word's tone later is much harder than learning it right the first time.
That is why a sensible Cantonese course front-loads pronunciation instead of treating it as polish. Learn Cantonese for Beginners works through initials, finals and syllables and then tone contrasts before conversation starts, and where your device provides a Cantonese voice the app uses it and falls back clearly when it does not. Hearing a contrast is the part you cannot get from a page.
Two honest caveats. Recognising a tone contrast in a quiet room is easier than producing one mid-sentence, and both are easier than either in a noisy restaurant. And no app can hear you the way a fluent speaker can — this article and the course are a foundation, not a substitute for a teacher or a conversation partner.
Jyutping, and why you should pick one system and stay there
Jyutping is a romanisation scheme published by the Linguistic Society of Hong Kong. Its defining feature is that tone is a digit written after the syllable: nei5, hou2, m4. That makes it typeable on any keyboard and unambiguous to read aloud.
The main alternative you will meet is Yale, an older Cantonese system that marks tone with diacritics and spelling changes instead. Yale is not wrong; a lot of good older material uses it. The problem is mixing them. Half-learning two romanisations produces a learner who is confident about neither and who quietly mispronounces words that the two systems spell differently.
Pick one and commit. Learn Cantonese for Beginners uses Jyutping throughout, which is also the system you are most likely to meet in current Hong Kong teaching material and input methods.
Three ways Cantonese gets written down
This is the part that surprises most beginners, and understanding it early saves a lot of confusion.
Traditional characters are the standard script in Hong Kong and Macau. Guangdong province in mainland China, where Cantonese is also widely spoken, uses simplified characters. The Unicode Standard encodes both, which is why your phone renders either without complaint, but a learner still has to choose which set to read fluently. Learn Cantonese for Beginners teaches traditional characters, which matches Hong Kong and Macau usage.
Beyond the character set, there are two written registers to be aware of:
- Standard Written Chinese is the formal written language used in newspapers, government documents, textbooks and business writing in Hong Kong. Its grammar and vocabulary are closer to Mandarin, and a Cantonese speaker reads it aloud using Cantonese pronunciations.
- Written Cantonese is spoken Cantonese committed to the page. You see it in messaging, social media, advertising, comics and dialogue in fiction. It uses characters that do not appear in Standard Written Chinese at all — 係 (hai6, "to be"), 唔 (m4, the common negator), 佢 (keoi5, "he/she"), 嘅 (ge3, the possessive and attributive particle).
A beginner meets both within weeks, usually without being told they are different things. Knowing that the formal notice on the wall and the message on your phone are written in two different registers is genuinely clarifying, and it is one of the things the course flags rather than glossing over.
A small number of well-established phrases
Rather than a long list, here are three that repay learning precisely, with the character, the Jyutping and the register they belong to.
你好 (nei5 hou2) — "hello". Universally understood, and the standard opener in teaching material. Worth knowing that in casual Hong Kong speech people often reach for a time-of-day greeting such as 早晨 (zou2 san4), "good morning", or simply address the person directly, so nei5 hou2 can sound slightly more formal in the street than it does in a textbook.
唔該 (m4 goi1) and 多謝 (do1 ze6) are both translated "thank you" in English, and they are not interchangeable. M4 goi1 thanks someone for a service or an action — the waiter who refilled your tea, the person who held the door. It also does duty as "excuse me" and "please". Do1 ze6 thanks someone for a gift or a compliment, something given rather than done. English collapses this distinction; Cantonese does not, and getting it wrong is one of the most audible beginner errors there is.
Numbers behave the way you would hope. 一 (jat1), 二 (ji6), 三 (saam1) — one, two, three — and they combine into larger numbers on a regular pattern, which makes prices, dates and phone numbers unusually good early practice.
The grammar that actually catches English speakers
Cantonese grammar is often sold as "easy" because verbs do not conjugate for tense or person. That is true and it is only half the story. The work moves elsewhere.
Classifiers sit between a number and a noun, and which classifier you use depends on the noun. English has a trace of this — two slices of bread, three head of cattle — but Cantonese requires one every time you count. The general-purpose classifier 個 (go3) covers a lot, and getting a specific one wrong is understandable rather than incomprehensible, so this is a place where beginners should proceed and correct later.
Aspect markers replace tense. Instead of changing the verb, Cantonese attaches markers that say whether an action is completed, ongoing, experienced before, or habitual. English speakers reliably try to map these onto past and present, and reliably get tangled.
Sentence-final particles are the feature with no English equivalent at all. Cantonese carries a large inventory of them, and they carry meaning that English handles with intonation, extra clauses or facial expression — softening a request, marking surprise, signalling that something is obvious, turning a statement into a question. Beginners tend to omit them entirely, which is grammatical but sounds oddly flat.
These are exactly the areas Learn Cantonese for Beginners covers explicitly — word order, pronouns, classifiers, possession, negation, aspect markers, questions, comparisons, adjective placement and sentence-final particles — along with a section on the mistakes English speakers reliably make, which is more useful than it sounds.
What the first month realistically looks like
Week one is sounds, not sentences. Initials, finals, syllables, then tone contrasts. It feels like a delay and it is not. Every word you learn afterwards is stored with whatever pronunciation you had when you learned it.
Week two you start recognising characters inside words. Not memorising a list of 200 in isolation — meeting high-frequency traditional characters inside phrases and sentences you are already saying, which is how the course introduces them.
Week three the tones start colliding. You knew a word yesterday and today it will not come out right. This is normal and it is the point at which most people quit. The structural answer is review that pulls from your own history rather than a fresh list: Daily Review, Mistake Rescue and the missed-questions list exist precisely for the words that will not stick. That approach has a name and a research literature behind it — spaced repetition for language learning explains why reviewing something at the point you are about to forget it beats reviewing it while you still know it.
Week four you build sentences instead of reciting them. Sentence Builder, Dictation Mode and Listening Blitz all attack the gap between recognising a phrase on a page and producing one out loud.
The twelve units of the course run from greetings and introductions out through family, food, shopping, numbers, money, dates and time, transport, work, school, health, emergencies and social situations. Working through them gives you a beginner foundation. It does not give you fluency, and the app does not claim it will.
Who this language is actually for
Cantonese rewards a specific kind of learner. If you are heading to Hong Kong, Macau or Guangdong, if you are reconnecting with a family language, or if you want access to a century of Cantonese film and music on its own terms, the effort has an obvious payoff. If you want the largest possible number of conversation partners for the least effort, other languages answer that better — Malay sits two full FSI categories lower, for instance, and starts in the Latin alphabet.
There is no wrong answer there, only a mismatch between what a language costs and what you want from it. If you are still weighing options, the other beginner guides in our language learning apps collection cover scripts with very different demands: the Thai script asks you to learn tone through spelling rules, Cyrillic is the shortest hard part of a hard language, and Khmer has a demanding script but no tones at all.
If Cantonese is the one you want, start with the sounds, keep Jyutping and traditional characters together from the first lesson, and judge your progress in months rather than days.
Good to know
Frequently asked questions
Is Cantonese a dialect of Mandarin?
How many tones does Cantonese have?
What is Jyutping and why not Yale or Pinyin?
Do I need to read characters to speak Cantonese?
Should I learn traditional or simplified characters for Cantonese?
Is Learn Cantonese for Beginners free?
Sources
- Jyutping FAQ (opens in a new tab)The Linguistic Society of Hong Kong — accessed
- Foreign Language Training (opens in a new tab)U.S. Department of State, Foreign Service Institute — accessed
- Unicode Character Code Charts (opens in a new tab)The Unicode Consortium — accessed