How different are Cantonese, Shanghainese, and Hokkien from Mandarin? A clear guide to Chinese dialects and why most learners should start with Mandarin.
One of the most common points of confusion for new learners of Chinese is the relationship between "Chinese" as a single entity and the bewildering variety of spoken forms across China and the Chinese diaspora. Cantonese speakers in Hong Kong cannot understand Mandarin speakers from Beijing without shared study. Shanghainese speakers and Fujianese speakers are mutually unintelligible. Yet all of these forms are often called "Chinese dialects" — a term that radically understates their differences.
Understanding the landscape of Chinese languages and dialects has practical implications for learners: it affects which variety you choose to study, how you interact with Chinese-speaking communities, and how you interpret the enormous diversity of Chinese media and communication you will encounter as your Chinese improves.
The linguist Max Weinreich famously observed that a language is a dialect with an army and a navy — meaning that the distinction between a language and a dialect is political rather than purely linguistic. Nowhere is this more clearly illustrated than in the Chinese-speaking world.
By the criteria linguists typically use to distinguish languages from dialects — mutual intelligibility (can speakers understand each other without study?) and structural distinctiveness (how different are the grammatical systems?) — many of what are called Chinese dialects would be classified as separate languages. Mandarin and Cantonese are about as mutually intelligible as Spanish and French — which is to say, not at all without learning. A native Cantonese speaker who has never studied Mandarin cannot understand spoken Mandarin, and vice versa.
Despite this linguistic reality, the official Chinese government position, and the common usage both in China and internationally, is to call all of these forms dialects (方言, fāngyán, literally "regional speech") of a single Chinese language. The political motivation for this position is the association of language unity with national unity — a single Chinese language supports a narrative of a unified Chinese people and state in a way that acknowledging ten or more separate languages would not.
For learners, the practical position is to understand both realities: these are linguistically distinct enough to be considered separate languages by many standards, but the shared written tradition and political framing means they are consistently treated as dialects in most contexts. What matters for your study is less the categorisation than the specific features of the variety you are learning.
Linguists generally classify Chinese languages into seven main groups, each containing multiple varieties. Understanding the major groups provides context for the diversity you will encounter.
Mandarin (官话, Guānhuà): The largest group by speakers, encompassing the varieties spoken across northern and southwestern China. Standard Mandarin (普通话, Pǔtōnghuà) is the official standard language of the People's Republic of China, and Guoyu (國語) is the standard in Taiwan. Mandarin is what most learners mean when they say "I'm learning Chinese." Native speakers: approximately 920 million.
Wu (吴语, Wúyǔ): Spoken in the Yangtze River Delta region — Shanghai, Zhejiang province, and southern Jiangsu. Shanghainese is the most well-known Wu variety. Wu has approximately 80 million speakers. Shanghainese has more tones than Mandarin (the exact number depends on analysis) and vocabulary and grammar that diverge substantially from Standard Mandarin.
Cantonese (粤语, Yuèyǔ): Spoken in Guangdong province, Guangxi, Hong Kong, and Macau, and is the dominant language of overseas Chinese communities in many parts of the world, particularly North America, Australia, and Southeast Asia. Cantonese has six or more contrastive tones (compared to Mandarin's four), has preserved Middle Chinese sounds that Mandarin has lost, and has a rich written tradition including a large body of colloquial written Cantonese. Native speakers: approximately 80 million, with very significant diaspora presence.
Min (闽语, Mǐnyǔ): A diverse group spoken in Fujian province and by the Fujianese diaspora across Southeast Asia and beyond. Hokkien (spoken in southern Fujian, Taiwan, and large parts of Southeast Asia) and Teochew are the most widely known Min varieties. Min languages have preserved archaic features of Chinese that other groups have lost, making them linguistically distinctive. Taiwan Mandarin is heavily influenced by Hokkien. Native speakers: approximately 70 million.
Hakka (客家话, Kèjiāhuà): Spoken by the Hakka people, historically a migrant group from central China who settled across southern China and Southeast Asia. Hakka has approximately 45 million speakers and is notable for being relatively well-preserved across geographically dispersed communities due to the Hakka people's distinctive migration history.
Gan (赣语, Gànyǔ) and Xiang (湘语, Xiāngyǔ): Spoken in Jiangxi and Hunan provinces respectively. Less well-known internationally but each has tens of millions of speakers. Mao Zedong's native variety was a form of Xiang.
Even within Mandarin — the group that most learners are studying — there is significant regional variation. Standard Mandarin (普通话) is based on the Beijing dialect, but spoken Mandarin in Sichuan, Yunnan, northeastern China (东北话, Dōngběihuà), and even Taiwan differ from the Beijing standard in accent, vocabulary, and to some extent grammar.
Taiwan Mandarin (國語) has distinct vocabulary influenced by Japanese loanwords, Hokkien substrate, and the historical separation of Taiwan from the mainland since 1949. Many everyday words are different: the word for motorbike, the word for taxi, and dozens of food words differ between mainland and Taiwan usage. Pronunciation also differs — Taiwan Mandarin does not use the Beijing retroflexion (the "r" sound added to many words in Beijing speech, called erhua).
Northeastern Mandarin (东北话), spoken across Liaoning, Jilin, and Heilongjiang, is close to Standard Mandarin phonologically but has distinctive vocabulary, intonation, and colloquial expressions that appear frequently in northeastern Chinese comedy and entertainment. The variety is highly intelligible to Standard Mandarin speakers.
Southwestern Mandarin, spoken in Sichuan, Chongqing, Guizhou, and Yunnan, has a distinctive tonal system and vocabulary influenced by historical contact with non-Mandarin populations. Sichuan Mandarin in particular sounds strikingly different from Beijing Mandarin to learners, even though both are within the Mandarin group.
The factor that unifies all Chinese language varieties linguistically and culturally is the written language. Written Mandarin (文字, wénzì) uses the same characters across all Chinese-speaking communities, regardless of how those characters are pronounced locally. A Cantonese speaker, a Shanghainese speaker, and a Mandarin speaker all read the same written text, even though they pronounce each character very differently.
This creates a distinctive cultural situation: written Chinese functions as a kind of super-language that enables cross-dialect communication through reading, while spoken communication requires either study of a shared variety (Standard Mandarin) or physical proximity within the same dialect region.
The practical implication for learners: the character knowledge you build while learning Standard Mandarin is broadly applicable. Characters you learn while studying Mandarin are the same characters used in Cantonese writing, in traditional script in Taiwan and Hong Kong, and in Sino-Japanese vocabulary. The investment in Chinese character knowledge has cross-variety applicability that spoken knowledge does not.
The one complication is the Simplified/Traditional split. Mainland China uses Simplified characters (simplified forms introduced through a government reform process in the 1950s and 1960s). Taiwan, Hong Kong, Macau, and most diaspora communities outside Southeast Asia use Traditional characters. Many characters are the same in both systems; some differ substantially. Learning to read both is possible and common among serious students of Chinese.
The question of whether to study Cantonese as well as or instead of Mandarin comes up regularly among learners, particularly those in communities where Cantonese is the dominant spoken Chinese variety (parts of North America, Australia, and Southeast Asia).
The case for Cantonese: if your primary motivation is community connection with a Hong Kong or Cantonese diaspora community, Cantonese is more appropriate. If you love Hong Kong cinema, Cantopop, or Cantonese culture specifically, Cantonese provides more direct cultural access. Cantonese-English code-switching is common in many Western Chinese communities and understanding it requires Cantonese, not Mandarin.
The case for Mandarin first: Mandarin is the lingua franca across all Chinese communities in China and the world. It is the official language of China, Taiwan (as Guoyu), and Singapore. Its learner resources vastly outnumber Cantonese resources. Learners of Mandarin can communicate with speakers from every Chinese language background who have also learned Mandarin (which in mainland China is essentially the entire educated population).
For most learners outside communities where Cantonese is the dominant spoken Chinese variety, Mandarin is the practical first choice. Character knowledge from Mandarin study transfers to Cantonese reading (with the Simplified/Traditional adjustment). Learning Cantonese as a second Chinese variety after Mandarin is very feasible — the character foundation is in place, and the grammar diverges in manageable ways.
For learners of Standard Mandarin, the main practical implication of China's linguistic diversity is exposure management. As your Mandarin improves, you will encounter regional accents, dialect influence, and code-switching in Chinese media, conversation, and real life that departs from the standard you have been learning. This is normal and does not mean your standard is wrong — it means the standard is a learned norm that coexists with enormous spoken diversity.
The best preparation for this encounter is broad listening exposure from early in your study. Graded audio stories that use clear Standard Mandarin are the foundation; as your level increases, exposure to news, drama, film, podcasts, and natural conversation will introduce the variation you need to become a genuinely versatile listener.
The stories at Dumplingo — like the Huangshan Adventure series from HSK 1 through the advanced Bondi Beach story at HSK 6 — use clear Standard Mandarin, making them ideal for building the phonological and grammatical foundation that all subsequent listening development builds on. The Grammar section covers the standard grammatical structures that remain consistent across regional varieties of Mandarin.
Standard Mandarin opens the door to over a billion speakers. Start building your foundation with graded stories on Dumplingo. Explore the story library →