Written Cantonese

A good source for well documented written Cantonese words can be found in the scripts for Cantonese opera. Readings in Cantonese colloquial: being selections from books in the Cantonese vernacular with free and literal translations of the Chinese character and romanized spelling (1894) by James Dyer Ball has a bibliography of printed works available in Cantonese characters in the last decade of the nineteenth century. A few libraries have collections of so-called "wooden fish books" written in Cantonese characters. Facsimiles and plot precis of a few of these have been published in Wolfram Eberhard's Cantonese Ballads. See also Cantonese love-songs, translated with introduction and notes by Cecil Clementi (1904) or a newer translation of these by Peter T. Morris in Cantonese love songs : an English translation of Jiu Ji-yung's Cantonese songs of the early 19th century (1992). Cantonese character versions of the Bible, Pilgrims Progress, and Peep of Day, as well as simple catechisms, were published by mission presses. The special Cantonese characters used in all of these were not standardized and show wide variation.

• 
  A Hong Kong billboard in Written Cantonese with a mixture of English words in the typical code switch style of Hong Kong speech.
• 
  A Hong Kong political advertisement in Written Cantonese
• 
  Political banner in Written Cantonese

Written Cantonese contains many characters not used in standard written Chinese in order to transcribe words not present in the standard lexicon, and for some words from Old Chinese when their original forms have been forgotten. Despite attempts by the government of Hong Kong in the 1990s to standardize this character set, culminating in the release of the Hong Kong Supplementary Character Set (HKSCS) for use in electronic communication, there is still significant disagreement about which characters are correct in written Cantonese, as many of the Cantonese words existed as descendants of Old Chinese words, but are being replaced by some new invented Cantonese words due to the Hong Kong Government's lack of knowledge about some of the Cantonese words.

General estimates of vocabulary differences between Cantonese and Mandarin range from 30 to 50 percent. Donald B. Snow, the author of Cantonese as Written Language: The Growth of a Written Chinese Vernacular, wrote that "It is difficult to quantify precisely how different" the two vocabularies are.[3] Snow wrote that the different vocabulary systems are the main difference between written Mandarin and written Cantonese.[3] Ouyang Shan made a corpus-based estimate concluding that one third of the lexical items used in regular Cantonese speech do not exist in Mandarin, but that between the formal registers the differences were smaller. He analyzed a radio news broadcast and concluded that of its lexical items, 10.6% were distinctly Cantonese.[3] Here are examples of differing lexical items in a sentence:

In the above table the two Chinese sentences are grammatically identical, using an A-not-A question to ask "Is it theirs?" (referring to some prior mentioned thing). But the characters are all different, though they correspond 1:1.

There are certain words that share a common root with standard written Chinese words. However, because they have diverged in pronunciation, tone, and/or meaning, they are often written using a different character. One example is the doublet 來 lòih (standard) and 嚟 lèih (Cantonese), meaning "to come." Both share the same meaning and usage, but because the colloquial pronunciation differs from the literary pronunciation, they are represented using two different characters. Some people argue that representing the colloquial pronunciation with a different (and often extremely complex) character is superfluous, and would encourage using the same character for both forms since they are cognates (see Derived characters below).

Some Cantonese words have no equivalents in Mandarin, though equivalents may exist in classical or other varieties of Chinese. Cantonese writers have from time to time reinvented or borrowed a new character if they are not aware of the original one. For example, some suggest that the common word 靚 (leng), meaning pretty in Cantonese but also looking into the mirror in Mandarin, is in fact the character 令.[4]

Today those characters can mainly be found in ancient rime dictionaries such as Guangyun. Some scholars have made some "archaeological" efforts to find out what the "original characters" are. Often, however, these efforts are of little use to the modern Cantonese writer, since the characters so discovered are not available in the standard character sets provided to computer users, and many have fallen out of usage.

In Southeast Asia, Cantonese people may adopt local Malay words into their daily speech, such as using the term 鐳 /lɵy/ rather than saying 錢 /tsʰiːn˨˥/ which would be what the Hong Kong Cantonese would say, meaning money and written 錢.

Cantonese particles may be added to the end of a sentence or suffixed to verbs to indicate aspect. There are many such particles; here are a few.

• 咩 – "mē" is placed at the end of a sentence to indicate disbelief, e.g., 乜你花名叫八兩金咩？ Is your nickname really Raymond Lam?
• 呢 – "nē" is placed at the end of a sentence to indicate a question,[5] e.g., 你叫咩名呢？ What is your name?
• 未 – "meih" is placed at the end of a sentence to ask if an action is done yet, e.g., 你做完未？ Are you done yet?
• 吓 – "háh" is placed after a verb to indicate a little bit, i.e., "eat a little bit"; "há" is used singly, to show uncertainty or unbelief, e.g., 吓？乜係咁㗎? What? Is it really like that?
• 緊 – "gán" is placed after a verb to indicate a progressive action, e.g.,我食緊蘋果。 I'm eating an apple.
• 咗 – "jó" placed after a verb to indicate a completed action, e.g., 我食咗蘋果。 I ate an apple.
• 晒 – "saai" placed after a verb to indicate an action to all of the targets, e.g., 我食晒啲蘋果。 I ate all the apples.
• 埋 – "maàih" is placed after a verb to indicate an expansion of the target of action, or that the action is an addition to the one(s) previously mentioned, e.g., 我食埋啲嘢就去。 I'll go after I finish eating the rest. ("eating the rest" is an expansion of the target of action from the food eaten to the food not yet eaten); 你可以去先，我食埋嘢先去。 You can go first. I'll eat before going. (The action "eating" is an addition to the action "going" which is previously mentioned or mutually known.)
• 哇/嘩 – "wā" 嘩！ Wow!
• 㗎啦 – "ga lā" is used when the context seems to be commonplace, e.g., 個個都係咁㗎啦。 Everyone is like that.
• 啫嘛 – "jē ma" translates as "just", e.g., 我做剩兩頁功課啫嘛。 I just have two pages of homework left to do.

Some Cantonese loanwords are written in existing Chinese characters.

Some characters already exist in standard Chinese, but are simply reborrowed into Cantonese with new meanings. Most of these tend to be archaic or rarely used characters. An example is the character 子, which means "child". The Cantonese word for child is represented by 仔(jai), which has the original meaning of "young animal".

Many characters used in Cantonese writings are formed by putting a mouth radical (口) on the left hand side of another better-known character (e.g. 㗎), usually a standard Chinese character. This indicates that the new character sounds like the standard character, but is only used phonetically in the Cantonese context. (An exception is 咩, which does not sound like 羊 (sheep), but sounds like the sound that sheep make.) The characters which are commonly used in Cantonese writing include:

There is evidence that the mouth radical in such characters can, over time, be replaced by a Signific, which indicates the meaning of the character. The new character is then a semantic compound. For instance, 冧 (lām, "bud"), written with the signific 冖 ("cover"), is instead written in older dictionaries as 啉, with the mouth radical.

Other common characters are unique to Cantonese or are different from their Mandarin usage, including: 乜, 冇, 仔, 佢, 佬, 俾, 靚 etc. The characters which are commonly used in Cantonese writing include:

• 冇 móuh (v. not have). Originally 無. Standard written Mandarin: 沒有
• 係 haih (v. be). Standard written Mandarin: 是
• 佢 kéuih (pron. he/she/it). Originally 渠. Standard written Mandarin: 他, 她, 它, 牠, 祂
• 乜 māt (pron. what) often followed by 嘢 to form 乜嘢. Originally 物也. Standard written Mandarin: 什麼
• 仔 jái (n. son, child, small thing). Originally 子.
• 佬 lóu (n. guy, dude). Originally 獠.[8]
• 畀/俾 béi (v. give). Standard written Mandarin: 給
• 靚 leng (adj. pretty, handsome). Standard written Mandarin: 漂亮
• 晒/曬 saai (adv. completely; v. bask in sun)
• 瞓 fan (v. sleep). Originally 困. Standard written Mandarin: 睏, 睡
• 攞/拎 ló/ling (v. take, get). Standard written Mandarin: 拿
• 脷 leih (n. tongue). Standard written Mandarin: 舌
• 攰 guih (adj. tired). Standard written Mandarin: 累
• 埞 dehng (n. place) often followed by 方 to form 埞方. Standard written Mandarin: 地方

The words represented by these characters are sometimes cognates with pre-existing Chinese words. However, their colloquial Cantonese pronunciations have diverged from formal Cantonese pronunciations. For example, 無 ("without") is normally pronounced mòuh in literature. In spoken Cantonese, 冇 (móuh) has the same usage, meaning, and pronunciation as 無, except for tone. 冇 represents the spoken Cantonese form of the word "without", while 無 represents the word used in Classical Chinese and Mandarin. However, 無 is still used in some instances in spoken Cantonese, such as 無論如何 ("no matter what happens"). Another example is the doublet 來/嚟, which means "come". 來 (lòih) is used in literature; 嚟 (lèih) is the spoken Cantonese form.

Though most Cantonese words can be found in the current encoding system, input workarounds are commonly used by those not familiar with them. Some Cantonese writers use simple romanization (e.g., use D as 啲), symbols (add a Latin letter "o" in front of another Chinese character; e.g., 㗎 is defined in Unicode but will not display if not installed on the device in use, hence the proxy o架 is often used), homophones (e.g., use 果 as 嗰), and Chinese characters which have different meanings in Mandarin (e.g., 乜, 係, 俾; etc.) For example,