Cantonese and Mandarin are two different languages

« previous post | next post »

From Mengnan Zhang:

I found this very interesting image on Facebook. The three columns stand for how to write various terms in Cantonese, their pronunciation, and the meaning of the words listed. As a native speaker of Mandarin, I have no idea what these words are talking about even after reading the meaning of each. Linked back to what our professor had talked about in class, Cantonese is a language, which both script and speech have no correspondence with Mandarin at all.

It's interesting that the so-called dúyīn / duk6jam1 讀音 ("pronunciation") column doesn't really help much with the pronunciation if you aren't already familiar with the pronunciation of the items in the first column.  Many of the pronunciation keys in the second column repeat characters in the first column or replace them with equally or more opaque characters.  The only pronunciation keys that are truly helpful are the eight syllables written in roman letters.

Also, the first character of the second item is very strange.  I've never seen it before, and it doesn't even look like a possible Chinese character to me.  However, using the four-corner code (37772), I did manage to track it down.  It turns out that it has the hypothetical pronunciation kǎn in Mandarin, but is actually a special Cantonese character pronounced gam2 ham6 kam2 and meaning:

(ham6) all, whole, everyone

(kam2) [1] (police) raid on illegal dens
[2] a cover, lid; to cover
[3] to slap; hit with the palms

Sources here and here.

The first character of the sixth item is 丼, the Japanese character for donburi, which will be familiar to Japanese foodies.  It seems like they may have borrowed it from Japanese for the sound of the first syllable of the Cantonese expression, though it is also said to be an old character for jǐng 井 ("well").  Phonologically and semantically speaking, however, the Japanese usage has nothing to do with jǐng 井 ("well").

The first two rows of the following table are Sino-Japanese readings and are related to the Sinitic meaning of "well".  The third row are purely Japanese style readings and convey the following two meanings:

  1. a simple large bowl for serving food
  2. a one-bowl meal served in a donburi bowl, consisting of a large portion of rice, covered in a meat or fish

To disambiguate between 1. and 2., speakers may refer more specifically to donburibachi 丼鉢, where hachi 鉢 means "bowl", and donburi mono, where mono もの / 物 ("thing") refers to the food.


So far, I have only covered two of the unusual characters used to write the Cantonese terms in the first column.  I haven't begun to touch upon the Cantonese pronunciations of the individual items, nor have I delved into their meanings, which, though given in column three in Mandarin, are not entirely transparent in all cases.


  1. KAMiKZ said,

    September 25, 2015 @ 4:59 pm

    The first character of the second item is quite usual in subtitles in HK films. To me, a Cantonese from Guangzhou, it pretty much means "all; entire". Or even more precisely, as with the emotion normally exhibited when you use the phrases that use this word, "leave none alone/untouched"

    The swear word that begins with said character plus the word "family" plus the word "shovel" is the most heard swear word in my experience. It reads phonetically "Hum Ga Tsa'an". I guess it means to murder the whole family of the offender. The shovel probably means to bury the body. Not a very nice phrase I guess, now that I come to think it.

  2. Norman said,

    September 25, 2015 @ 5:35 pm

    Thanks! Definitely brought back a piece of my childhood saying these out loud. The first one made me LOL. My grandma and mom used that all the time!

  3. Steven Marzuola said,

    September 25, 2015 @ 5:42 pm

    I don't know Chinese, but as a Spanish-English translator and member of several organizations, several people have explained to me — for years! — that there is one written Chinese language (well, Traditional and Simplified). Also, that every Chinese speaker, regardless of which language they speak, writes in that language. This is the first time I've read otherwise. Help!

  4. Victor Mair said,

    September 25, 2015 @ 5:59 pm

    @Steven Marzuola

    You need to spend more time in the Language Log archives.

  5. Victor Mair said,

    September 25, 2015 @ 6:00 pm

    From Don Snow:

    Most of these are unfamiliar to me, but that may say more about the limits on my Cantonese than anything else. Like you, I don't find the second column very helpful – no easier than those in the first.

    The second item is familiar to me, and I've seen that first character used in it before – though there are quite a few different ways to write hambanglang.

    I assume the sixth item is "dump garbage" with 丼 used phonetically for dum = dump, borrowing via Japanese as you noted.

  6. Nathan Hopson said,

    September 25, 2015 @ 8:25 pm

    丼 in Japanese probably comes from 井. The meaning of 丼 is a deep bowl, which some believe to be derived from "well." One authoritative dictionary (日本国語大辞典) gives the "don" sound (どん) of donburi (丼) as related to the sound of something being tossed into a well, from which the character was chosen.

  7. K. Chang said,

    September 25, 2015 @ 9:18 pm

    Oh boy. I can't write any of those alternate characters. Don't ask me to pronounce them either. I can speak them, but a few of those are foreign even to me.

  8. JS said,

    September 25, 2015 @ 9:28 pm

    丼 is given the secondary fanqie 都感 (= tomʔ) in rhyme books like Ji yun 集韻; this word is said to be onomatopoeic: 'the sound of something being thrown in a well' 投物井中聲. No idea how this is related to the identical tale in Japanese…

  9. APOLLO WU said,

    September 26, 2015 @ 12:27 am

    Being a native Cantonese speaker born in Hong Kong, I want to point out that although I have no problem speaking those Cantonese terms in the List and understand them perfectly, I have difficulty with the strange characters used. These characters are usually not used in Hong Kong Chinese newspapers, which only print Chinese text. Chinese text is supposed to represent two languages, namely Cantonese and Putonghua. In reality, Cantonese does not have a well-developed written script. An alphabetized Cantonese certainly could not be readable by Putonghua speakers. However, Chinese script which supports Putonghua can be readily understood by both the Cantonese and Putonghua speakers. However, Cantonese speakers who do not speak Putonghua can readily read all printed Chinese in Cantonese speech, and in this way, they feel not so linguistically handicapped. This rather strange bilingual situation is highlighted by the unique goal of educating Hong Kong people in Two Scripts and Three Languages 两文三语。

  10. Chas Belov said,

    September 26, 2015 @ 2:42 am

    What is the relationship between the first word in the second row and the character meaning "all" and apparently also pronounced "hum" in the curse "岑家拎" (source being the LMF hip-hop song as imported from the song album database that iTunes uses). Being user-contributed, this might not be the right character.

  11. Bloix said,

    September 26, 2015 @ 8:10 am

    OT: can someone tell me if the name of the new panda cub, Bei Bei, really means "precious treasure"? If it does, how does that work? Does Bei mean treasure and the duplication is emphasis? Or what?

  12. JQ said,

    September 26, 2015 @ 8:20 am

    The 9th on the list is a transliteration of the English word "fluke"

  13. Victor Mair said,

    September 26, 2015 @ 8:34 am


    I will make a quick post about that within an hour or so.


    Thank you very much for the identification. I will spell it out for everyone within an hour or two.

  14. phspaelti said,

    September 26, 2015 @ 10:51 am

    @Hopson: One authoritative dictionary (日本国語大辞典) gives the "don" sound (どん) of donburi (丼) as related to the sound of something being tossed into a well

    I'm not sure I believe that kind of explanation. The one notable thing about the word 'donburi' is that is has the exact pattern of '-ri' Onamtopoeia ('shonbori', 'unzari', 'pittari', etc.). If that is correct, the 'dobu' root would be suitable for something that is big and hollow.

  15. Victor Mair said,

    September 26, 2015 @ 10:59 am

    The Hong Kong government has a policy of “biliterate trilingualism”:

    (Cant.) loeng5 man4 saam1 jyu5 / (Mand.) liǎng wén sān yǔ 兩文三語 ("two scripts and three spoken languages")

    The two scripts are the Roman alphabet and the three spoken languages are Cantonese, English, and Modern Standard Mandarin (MSM; Putonghua).

  16. Victor Mair said,

    September 26, 2015 @ 12:39 pm

    To clarify the relationship between Cantonese and Mandarin (both in their spoken and written forms), here are some relevant earlier Language Log posts:

    "Is Cantonese a language, or a personification of the devil?" (2/8/14)

    "Cantonese as Mother Tongue, with a note on Norwegian Bokmål" (12/22/13)

    "Spoken Hong Kong Cantonese and written Cantonese" (8/29/13)

    "English is a Dialect of Germanic; or, The Traitors to Our Common Heritage" (9/4/13)

    "Token Cantonese" (5/16/15)

    "A quick exit for Cantonese" (7/22/15)

    "Cantonese novels" (8/20/13)

    "Eighty-one Cantonese proverbs in one picture" (2/27/14)

    "Hong Kong Multilingualism and Polyscriptalism" (7/26/10)

    "Thick toast: another new Cantonese pun" (12/11/14)

    "Signifying the Local" (9/28/13)

    "Mutual Intelligibility of Sinitic Languages" (3/6/09), from which I take the following notes:


    See Donald B. Snow's remarkable Cantonese as Written Language: The Growth of a Written Chinese Vernacular (Hong Kong: Hong Kong University Press, 2003). Appendix 1 gives 14 Cantonese texts, each of which Snow carefully analyzes for the degree to which it adheres to the norms of spoken Cantonese rather than of written Modern Standard Mandarin (MSM). The 14 texts, which cover a wide range of genres, date from around the 17th century to the contemporary period. It is striking that the percentages of overtly marked Cantonese (and Snow is referring here not just to special Cantonese characters) in these 14 texts range from only 3% to 36%: 3, 4, 6, 7, 10, 11, 12, 20, 23, 23, 23, 28, 32, 36, for an average of 17%.

    See also the important works of Robert S. Bauer on Cantonese phonology and orthography and the essential works of Stephen Matthews and Virginia Yip on Cantonese grammar.


    There are many other posts on Cantonese and its relationship to Mandarin and other Sinitic topolects and languages.

    By way of summary:

    1. As is true of the other Sinitic topolects, Cantonese is seldom written down.

    2. When full-blown Cantonese is written in Chinese characters, it is necessary to use a thousand or more special graphs that are distinctive to written Cantonese and not known to speakers of Mandarin or other Sinitic topolects and languages. Furthermore, it is very common in full-blown written Cantonese to use characters from the standard set for written Mandarin in ways that are unique to Cantonese and whose Cantonese meanings are not comprehended by Mandarin speakers and readers, e.g., máidān 埋单 ("pay the bill") and the item that I will explicate in my next comment.

    3. It is possible to write Cantonese fully in Romanization, and that is how Cantonese language is normally taught.

    4. If you want to be literate in "Chinese", you must learn Mandarin vocabulary, grammar, and syntax. That is called Zhōngwén 中文 ("Chinese"), but you may pronounce the characters in Cantonese. Mandarin speakers will not be able to understand you, even though you are reading Zhōngwén 中文 ("Chinese") aloud.

    5. Mandarin speakers do not understand spoken Cantonese.

    6. People who are literate only in Zhōngwén 中文 ("Chinese") cannot comprehend full-blown written Cantonese.

    Mutatis mutandis for the other Sinitic languages and topolects.

  17. Victor Mair said,

    September 26, 2015 @ 1:52 pm

    As promised in an earlier comment addressed to JQ, here is a transcription and translation of all three columns for item #9:

    Column 1 — Cantonese term

    (Cant.) fu4luk6 / (Mand.) fúlù 符籙 (lit., "talisman" — this meaning is unrelated to the Cantonese term, which is a borrowing from English "fluke")

    Column 2 — pronunciation as transcribed in Chinese characters

    (Cant.) fu2,6luk1 / (Mand.) fùlù 父碌 (lit., "father + commonplace; busy; mediocre") — the meanings of both characters are unrelated to the Cantonese term

    Column 3 — meaning in "Chinese"
    (Cant.) kaau3 hou2wan6/ (Mand.) kào hǎo yùn 靠好運 rely on good luck

  18. Victor Mair said,

    September 26, 2015 @ 1:54 pm

    From Abraham Chan:

    The 14 items listed are typically pronounced as follows (in Jyutping, except for the first two items, which are normally pronounced with "double initials"):

    1. klik1 klaak1
    2. ham6 blaang6
    3. le2 gam3 he3
    4. deoi2 keoi5
    5. seon2 pun2
    6. dam2 laap6 saap3
    7. long1 lai2
    8. ce1 daai6 paau3
    9. fu4 luk1
    10. man1 naan3
    11. zit1 am3 cong1
    12. coeng3 cin2
    13. diu4 diu2 fing6
    14. soe4 waat6 tai1

    "冚" is a common Cantonese character used for either ham6 ("all") or kam2 ("cover").

    The use of the character "丼" as dam2 ("throw away") may have its root in the 《集韻》「都感切,投物井中聲。」The following Baidu webpage offers a more comprehensive overview of the character:

  19. Victor Mair said,

    September 26, 2015 @ 1:56 pm

    In Abraham Chan's comment, it is extremely interesting to note that the first two items illustrate initial consonant clusters in Cantonese.

  20. Victor Mair said,

    September 26, 2015 @ 8:18 pm

    Speaking of trilingualism in Hong Kong, a growing number of Hongkongers have become trilingual, with more than 90 per cent of residents under 30 now speaking English and Putonghua, as well as their mother tongue, Cantonese.

  21. Jongseong Park said,

    September 26, 2015 @ 10:36 pm

    Seeing the character 丼 (which I was not familiar with) reminds me of a joke told by one of my teachers in elementary school in Korea. He wrote the character 井 (meaning "well") and added a "dot" stroke in the middle, so that the result was something that looked like 丼. "What character is this?" he asked. "It is 퐁당 퐁 pongdang pong," he answered his own question. 퐁당 pongdang is an onomatopoeic word for a small splash—the implication is that someone dropped a stone into the well and now there is a splash. Korean names for Chinese characters are composed of the meaning part and the reading part, so here the meaning is 퐁당 pongdang "small splash" and the reading of the character is 퐁 pong.

    So I find it amusing that 丼 is an actual character that is in use.

  22. Lai Ka Yau said,

    September 27, 2015 @ 8:12 pm

    To be honest, I think this is likely the latest offering from the 本字 gang. The left column, as Apollo said, is quite opaque even to native speakers, with a few exceptions, 冚 being one. They're probably dug up by the 'original characters' people.

    Searching 虢礫緙嘞 on Google confirms this. I don't know how they came up with the odd orthography, but many of them seem to be rebus writing. 虢礫緙嘞, for example, clearly isn't the sum of its constituent characters, as none of them contribute semantic content to the word. In fact, considering that the four vowels of this word are high front-high front-low back-low back, the word seems to be part of the larger crosslinguistic phenomenon described in Pinker (1994), which also explains vowel placement in tit-for-tat, criss-cross, King Kong, etc. This makes it even more unlikely for the constituent characters to have been semantically meaningful morphemes.

  23. Sara Foster said,

    September 28, 2015 @ 12:49 am

    What a huge jump to conclusions! So you found a table on Facebook, which expressly displays the *differences* between Cantonese and Mandarin, and then conclude that hence the both are totally unrelated? I could easily build such a table for Spanish/Portuguese, American/British English etc.; or even Portuguese vs. Brazilian Portuguese. Would you, when finding that "pineapple" is "ananás" in Portuguese Portuguese, but "abacaxi" in Brazilian Portuguese, conclude that one has "no correspondence" at all to the other?

    yes, Cantonese and Mandarin are not super-close, but still largely similar.

  24. Lai Ka Yau said,

    September 28, 2015 @ 9:02 am

    @Sara Foster: I don't think Zhang jumped to the conclusion based on the table alone… Zhang only 'linked back to what [his/her] professor had talked about in class'. Phonologically, Cantonese and Mandarin are by no means similar. Lexically, they have drifted apart significantly as well. Syntactically, they are more similar, but the similarities have been exaggerated because of Y R Chao's influential pre-Chomskyan work on the varieties of Chinese. Although extremely important in linguistics, Chao had a Mandarin bias when he analysed the various dialects because of his assumption that the varieties of Chinese have a 'universal grammar' that differ insignificantly. In reality, syntactic differences between the topolects are comparable to those of the Romance languages – see Prof. Stephen Matthews' 'Y. R. Chao and Universal Chinese Grammar', for example.

  25. Norman said,

    October 1, 2015 @ 12:20 am

    Just remembered another consonant cluster that my parents used to describe a runny nose: fi-li feh-let which often sounded like fli-flet to my ears. Am I remembering this right?

  26. Victor Mair said,

    October 2, 2015 @ 7:51 am

    @Sara Foster

    "…Cantonese and Mandarin are not super-close, but still largely similar."

    Please provide some evidence for both parts of your assertion.

    A tremendous amount of evidence for the difference between Cantonese and Mandarin has been adduced in this post and the comments thereto, esp. this one:

    How distant is "not super-close"? How similar is "largely similar"?

  27. Eidolon said,

    October 6, 2015 @ 1:00 pm

    I think @Lai Ka Yau @Sara Foster that it is necessary to remember that "Cantonese" and "Mandarin" are not static, monolithic blocs, but are diverse groupings in and of themselves. There is also a need to distinguish between what "Cantonese purists" might consider Cantonese to be, and the actual everyday spoken language of Guangdong, Hong Kong, etc., which today are full of Mandarin, English, and other languages' lexical terms, and are also continuously interacting with them phonologically and syntactically. Language as a medium of communication is by its very nature constantly changing and shifting to meet new needs. As much as it'd simplify the issue otherwise, linguistics is fundamentally a historical exercise, as opposed to the study of a constant object/principle. The "Cantonese" that was understood by Y R Chao fifty years ago is not the "Cantonese" of today and won't be the "Cantonese" of tomorrow.

RSS feed for comments on this post