Sound rules

« previous post | next post »

Stephen Halsey, who is spending the year in Taiwan doing research, observed an interesting linguistic phenomenon that shows the predominance of sound over symbol, even in the writing of Chinese, where the symbols are complex and semantically "heavy" in comparison to phonetic scripts like the Roman alphabet or bopomofo / zhuyin fuhao (Mandarin phonetic symbols), where the symbols are simple and semantically "light".

My landlady was showing me the Chinese input system in her smartphone–zhuyin fuhao in this case–and complained that it was simply too slow.  Dozens and dozens of characters popped up for each discrete input, and so she often uses characters for their phonetic rather than ideographic meaning.  For instance, she wanted to instruct someone to meet her at Guāngfù nánlù 光复南路 ("Guangfu South Road"), but hunting for the proper nán 南 ("south") took too long, so she instead wrote Guāngfù nánlù 光复男路 (lit., "Guangfu Male Road").  Her e-interlocutor knew exactly what she meant.

I gather something similar happened with the Japanese indigenization of Chinese characters in the Nara and Heian periods; they began to represent phonetic meanings–sounds–rather than abstract ideas or concepts.  A minor example of new technology driving linguistic innovation.

This sort of thing happens all the time in Chinese, and for different reasons, e.g.:

1. when you can't remember the character you really want to write, you use a homophonous character instead

2. when you are in too big of a hurry to find the character you really want to write in a long list of homophones called up by your phonetic (N.B.) inputting device (like Stephan's landlady)

3. for playfulness (especially on the internet and in advertisements)

4. when you want to mimic or transcribe the sounds of another language or topolect

5. when you borrow a morpheme from Cantonese or Taiwanese or other topolect that doesn't exist in Mandarin

The first four categories are obvious, but, for the last category, let's look at the ubiquitous expression mǎidān 买单 (lit., "buy the bill") // máidān 埋单 (lit., "bury the bill"), which actually has nothing to do with "buying" or "burying", but simply means "bring / settle the bill / check" in Cantonese (maai4 daan1).  Daan1 单 is a Sinitic morpheme that does mean "check; bill" (as well as lots of other things), so we don't need to worry about it in the rest of this discussion.  It's the Cantonese morpheme maai4 for which there is no known Chinese character (perhaps it came from a substrate language).  So Cantonese have borrowed the mái 埋 character as a way to approximate the sound of their morpheme maai4, which means:

[v] finish up; close out; settle (an account)
[v] move closer/nearer to; close in/up
[adv] closer together; nearer to
[adv] even; too; also; as well

When the Cantonese term maai4 daan1 ("bring / settle the bill") was borrowed into Mandarin, it came along with the conventional sinographic representation máidān 埋单.  Of course, that left Mandarin speakers in a state of bewilderment, so either they tried to rationalize the expression by explaining that, when they paid the bill, they were in essence "burying" it.  That didn't convince the skeptics, who thought it would make more sense to change the mái 埋 ("bury") to the near-homophone mǎi 买 ("buy").  At least it was a financial transaction!  But it doesn't make sense to "buy" the bill either, does it?  In such cases as maai4 daan1 埋单 ("bring / settle the bill"), we simply need to accept the fact that sound overrides symbol.

The ascendancy of sound over symbol in Chinese is borne out in many other ways which we have examined on Language Log.  Here's just a small sampling:

Naturally, the paramountcy of sound over symbol in writing is not absolute, and sometimes — whether due to orthographic error or for other reason — symbol stands above sound, as in this case:

"'Double Happiness': symbol of Confucianism as a religion" (6/8/15)

Then again, some people wouldn't even countenance 囍 as part of the writing system, but would consider it simply an ornament or decoration.

[h.t. June Teufel Dreyer]


  1. Anschel Schaffer-Cohen said,

    September 8, 2015 @ 7:28 pm

    I would have thought the example of mǎidān 买单 show'd precisely the opposite. That the process of folk etymology in Chinese, in favoring a somewhat meaningful character representation for a loanword (and even changing the pronunciation of the word in the process) demonstrates an affinity for symbol at the *expense* of sound.

    I would say the same about, for example, the "y" that English puts in the word "syllabus" to match it with "syllable", despite the words' being unrelated etymologically.

  2. Michael Watts said,

    September 8, 2015 @ 9:08 pm

    This phenomenon doesn't show any primacy of sound over symbol in Chinese writing; it shows that the people involved are using sound-based input methods. I communicate regularly with someone who habitually uses a stroke-based method, and her typos are different. Unfortunately, the only example I can quickly find is of one time when I had to ask for clarification (she meant 公司 but wrote 公园), so it's not a great example of the communication not stumbling when the wrong symbol is used, but that happens too. People using sound-based input methods (as I do) will substitute characters with the same sound as the target character; people using stroke-based input will substitute characters with similar symbols and unrelated sounds. Understanding won't really suffer if you write 打 when you should write 订, any more than it will suffer if you write 男 when you should write 南.

    I agree that in reality sound is more significant than symbol, but this isn't evidence.

  3. Michael Watts said,

    September 8, 2015 @ 9:16 pm

    There's a chinese movie in which someone who was caught peeking through the wall at a bathing girl is sentenced to stand in the center of town wearing a big sign saying 耍流氓 shuǎ liúmáng (a verb phrase, but essentially equivalent to a sign saying "pervert"). The kid who makes the sign writes 要流氓 yào liúmáng. Although everyone makes fun of him for this, there's no pretense that the mistake caused any difficulty in understanding what was meant.

  4. Guy_H said,

    September 9, 2015 @ 6:48 am

    I can think of a sixth reason (particularly in mainland China) – to evade the censors!

    As for the mai2dan1/mai3dan1 example – agree with the first reply. Surely the change in character is an example of symbol over sound?

  5. Jim Breen said,

    September 9, 2015 @ 7:30 am

    Stephen Halsey wrote: "I gather something similar happened with the Japanese indigenization of Chinese characters in the Nara and Heian periods; they began to represent phonetic meanings–sounds–rather than abstract ideas or concepts. A minor example of new technology driving linguistic innovation."

    As I understood it that "represent phonetic meanings–sounds–rather than abstract ideas or concepts" was a case of applying the hanzi which traditionally had been used in China for phonetic purposes. I may have misremembered it – it's been years since I've read any detailed descriptions, but as I recall it the kanji that were used in phonetic transcriptions systems such as 万葉仮名 (manyōgana) were pretty much in that category.

  6. Victor Mair said,

    September 9, 2015 @ 7:42 am

    People realize that neither mǎidān 买单 (lit., "buy the bill") nor máidān 埋单 (lit., "bury the bill") really make sense, but they stick with them because they approximate the sound of the well-known Cantonese expression for settling / paying the bill. Ditto for Cantonese words like "taxi", "steak", and "bus" — dik1si6 的士, si6dik1 士的, baa1si2 巴士 — that are common in the Sinophone world: the characters are being used purely to represent the sounds of the words borrowed from English. The surface significations of the characters make no sense whatsoever, yet people use these expressions all the time without worrying about the meanings of the characters.

    Unintentionally miswriting the strokes of characters and calling up the wrong ones, thus not making sense in what one is writing, is a very different kettle of fish from doing what Stephen's landlady did: knowingly choose a wrong character with the right sound, confident that her e-interlocutor would understand perfectly well the meaning she wished to convey.

    The sixth reason for substituting a homophonous character mentioned by Guy_H is indeed extremely common on the mainland, as often documented on Language Log.

  7. Victor Mair said,

    September 9, 2015 @ 7:49 am

    From an American professor who is fluent in Chinese:

    When I was taking Chinese 101 ab, a friend from high-school years introduced me to a Chinese acquaintance whose father ran the “in” upscale Chinese restaurant in New York at the time. James [VHM: not his real name] had been subjected to Chinese language school in his childhood, but had little real interest in it. His notes to me more than once had characters that had the same sound as the characters he meant, but not the same meaning. So apparently this is pretty common.

  8. Brett said,

    September 9, 2015 @ 8:45 am

    Fifteen or twenty years ago, I heard a fluent, but not native, English speaker from Europe use "bury" to describe the act of paying off a large restaurant bill.

  9. Victor Mair said,

    September 9, 2015 @ 8:45 am

    From a colleague:

    People just enjoy playing with words. In China now, there’s an additional reason: to evade the censors. When, before the 18th Party Congress but after the Bo Xilai purge and Xi Jinping’s mysterious two-week disappearance, use of the term Shíbā dà 十八大 (= Zhōngguó Gòngchǎndǎng Dìshíbācì Quánguó Dàibiǎo Dàhuì 中国共产党第十八次全国代表大会 ["Eighteenth National Congress of the Communist Party of China"]) was proscribed, there was a lot of social media chatter about the internal politics of Sībādá 斯巴達 ("Sparta"; pure transcription — the characters make no sense whatsoever). Since the filters were set to catch characters rather than sound, they didn’t pick up the reference to the 18th Party Congress. Note, moreover, that the sound correspondence (Shíbā dà || Sībādá) was not even that close, but people still got it instantaneously, plus there was the added humor of doing it with a cute code word like "Sparta".

    Very, very cool use of sound over symbol.

  10. Coby Lubliner said,

    September 9, 2015 @ 8:59 am

    Isn't this the Chinese equivalent of eggcorns?

  11. Michael Watts said,

    September 9, 2015 @ 4:38 pm

    Why do you say the sound correspondence between Shíbādà and Sībādá isn't that close? To many, many Chinese, those differ only in the tones.

  12. Victor Mair said,

    September 9, 2015 @ 11:36 pm

    @Michael Watts

    I don't know which "many, many" you're talking about, or how "many, many" you're talking about, but I would suggest you take a look at Hànyǔ fāngyán zìhuì 汉语方言字汇, 2nd ed. (Beijing: Wénzì gǎigé chūbǎn shè 文字改革出版社, 1989), compiled by a research unit of the Chinese Department of Peking University. The characters are arranged by Middle Sinitic pronunciation, but their modern pronunciations are given in IPA for twenty representative topolects from north to south.

  13. K Chang said,

    September 10, 2015 @ 3:59 am

    @Coby Lubliner — only if the substituted expression "makes sense"… even if it's "Huh?" kind of sense.

    @ProfMair — speaking of si6dik1 士的, while it's used it's not that often. Usually 士的 is used as "steak cubes", i.e. small pieces of steak. For a full steak, most restaurants seem to use 牛扒. if you want more vernacular… 鋸扒 (literally, sawing the steak), as in "我地去鋸扒!" (Let's go have steak!)

  14. Mr Punch said,

    September 10, 2015 @ 8:34 am

    This seems to me (as to others) very much like the process of transliteration and translation of terms from other languages into English – which is somewhat obscured by the often retroactive "correction" of the orthography – complicated of course by a non-phonetic writing system. It's like how "ecrevisse" became "crayfish," "crawfish," and "crawdad."

  15. Jamie said,

    September 10, 2015 @ 4:11 pm

    Didn't John DeFrancis argue that the Chinese writing system was largely phonetic? This effect would seem to be consistent with that.

  16. Brendan said,

    September 10, 2015 @ 5:55 pm

    And the "我地" in K Chang's comment above is another nice example: the word for "we/us," ngo5dei6, is also written as the homophonous 我啲. The second character is being used solely for its sound value in both cases.

  17. Chas Belov said,

    September 12, 2015 @ 4:14 am

    Even in handwriting, when I order brown rice at a Chinese restaurant they write 王反 for 黃飯.

  18. Alexis Van Gestel said,

    September 14, 2015 @ 3:07 am

    Has anyone of you heard about the startup that is offering a solution for Chinese input? It seems to combine PinYin and Chinese characters in order to solve the issue of inaccuracy embedded in the PinYin method but at the same time it keeps the encoding of Chinese characters simple and playful.

    It seems like they are launching their first keyboard app for smartwatches, which is pretty amazing in itself, but they are announcing the same keyboard for smartphones and tablets as well in the near future.

    The company is called iBeezi. Check it out on their website,

RSS feed for comments on this post