Annals of airport Chinglish, part 3

« previous post | next post »

Carley De Rosa spotted this sign in the Kunming airport on her way to Laos. Dumbfounded by the Chinglish, not least because what it called an "elevator" was actually an "escalator", on her way back from Laos she made sure to get a photograph of the sign and send it to me for analysis:

The sign actually says:

shǒutuīchē 手推车 ("cart; trolley [Brit.]; pushcart")

qǐng zǒu duìmiàn diàntī, chéngzuò diàntī, qǐng gùkè lā hǎo fú hǎo, értóng zuò diàntī xū jiāzhǎng péihù, bùdé zài diàntī shàng dǎnào.


"Please walk over to the escalator on the opposite side. Upon getting on the escalator, passengers should hold [their cart] tightly and support [themselves] securely. Children who take the escalator must be accompanied by a parent and are not permitted to roughhouse on the escalator."

The substitution of "beard" for "must" is easy to explain, since the simplified character 须 is used both for its original meaning of "must" (which was what the Chinese sign intended to say) and for the traditional character 鬚 ("beard"), which the translation software mistakenly invoked here.

As for the mixup of "elevator" and "escalator", that is a much longer story and is the result of the fact that diàntī 电梯 (lit., "electric ladder / stairs / steps") can mean either and is habitually used for both. When individuals wish to make a distinction between "elevator" and "escalator", there is great variety in the terms that may be used.

N.B.: The following lists are not exhaustive for either "elevator" or "escalator".


diàntī 电梯 (lit., "electric ladder / stairs / steps") — used as the default term for "elevator" in contrast to some other term for "escalator"

shēngjiàngjī 升降机 ("elevator; dumbwaiter; lift") — usually reserved for construction sites (especially when attached outside a building that is under construction), factories, etc., not for regular passenger elevators


zìdòng fútī 自动扶梯 (lit., "automatic [i.e., 'self-moving'] support-ladder / stairs / steps") — this is the "correct" term for "escalator", but people tend to avoid it in speech because it seems concocted and insufficiently vernacular

fútī 扶梯 ("support-ladder / stairs / steps")

shǒufútī 手扶梯 ("hand-support-ladder / stairs / steps")

diànfútī 电扶梯 ("electric-support-ladder / stairs / steps") — this and the previous term seem to be favored by speakers of Taiwan Mandarin

gǔntī 滚梯 (lit., "rolling ladder / stairs / steps")

diàndòng lóutī 电动楼梯 ("electric-driven stairs")

zìdòng diàntī 自动电梯 ("automatic electric ladder / stairs / steps")

Sample sentences:

Zǒu lóutī shàngqù tài lèile, háishì zuò nà biān de diàntī shàngqù ba. 走楼梯上去太累了,还是坐那边的电梯上去吧。("It's too tiring to walk up the steps, so let's take the escalator over there instead.")

Děng diàntī de rén tài duōle, wǒmen zǒu nà biān de zìdòng fútī ba. 等电梯的人太多了,我们走那边的自动扶梯吧。("There are too many people waiting for the elevator; let's take the escalator over there instead.")

To recapitulate, in Mandarin, most people strongly prefer to use diàntī 电梯 (lit., "electric ladder / stairs / steps") for both "escalator" and "elevator", but will only use some other term for "escalator" when forced to make a distinction between the two modes of going up and down in a building.

There are still additional options in Cantonese. In speech the most common word for 'elevator' is lip1 車+立 which is a borrowing of British English "lift". This special Cantonese character (車+立) is sometimes used in writing and is a Hong Kong creation, but the English word "lift" is also written directly or the romanized Cantonese "lip" may also be used. In Cantonese, there is a bewildering variety of terms for "escalator". The Hong Kong Civil Service Bureau's bilingual website lists seven different lexical items with this meaning. Cantonesesheik lists six lexical items.

Based on preliminary Google searches by Bob Bauer, it appears that the most commonly-used items for "escalator'"in HK are haang4 jan4 din6 tai1 行人電梯 ("pedestrian electric ladder / stairs / steps") and fu4 sau2 din6 tai1 扶手電梯 ("hand-support electric ladder / stairs / steps")

Abrahman Chan explains the historical reasons for the plethora of terms for "escalator" in Cantonese:

The use of the three terms "車+立 (lip1)," "電梯 (din6tai1)," and "升降機 (sing1gong3gei1)" in Hong Kong have a rather complicated history.

Between the elevator and the escalator, the former has a much longer use in Hong Kong. The company Otis claims to have erected the first Hong Kong elevator in 1888, but the first escalator only in 1957.

As you know, an elevator is known as a "lift" in Britain, and in Hong Kong we have coined the character "車+立" to represent the loanword "lip1." But a Chinese term "電梯 (electric ladder)" was also used to represent the elevator. I haven't looked into the primary sources, and have no idea whether "電梯" was a local term or borrowed elsewhere (say, from Shanghai). There is no doubt, however, that both "車+立" and "電梯" date from before the Second World War and both were used to represent an elevator.

The introduction of the escalator brought considerable confusion to the use of the term "電梯," however. For one thing, the staircase is known as "樓梯 (lau4tai1)" in Chinese, and the escalator definitely looks a lot more like a staircase than an elevator does; so for a long time (since the late 1950s) the newer escalators have been competing with the older elevators for use of the term "電梯". It is not unusual for one person to use "電梯" to refer to both devices.

The "升降機" is a latecomer to the game, in direct response to the "電梯" confusion. The "升降機" is a translation of the "elevator," and the escalator is often renamed "扶手電梯 (fu4sau2din6tai1)" to avoid confusing with "電梯," which even today can still mean either an elevator or an escalator.

I personally almost never say "升降機," but generally use "車+立" to refer to an elevator. I usually use "電梯" for escalator, although sometime adding "扶手" to ensure there is no confusion with an elevator.

But the use of "車+立" is of course a no-no to Chinese purists, who insist on using either "電梯" or "升降機" for elevators.

In Japanese there is no confusion between "elevator" and "escalator" because they are directly transliterated from the English words, hence erebētā エレベーター and esukarētā エスカレーター. In written Japanese, one may use the technical term hōkō-ki 昇降機 ("rise-descend-device").

Incidentally, there is a special Japanese graph for "elevator girl" (erebētā gāru エレベーターガール), a type of service person that is still common in Japan, namely, the radical 女 ("woman / female") on the left and 上 ("up") written above 下 ("down") on the right, which is pronounced as erebētā gāru. The short form of erebētā gāru is erega エレガ, and the Sinitic-style term is sōjō or shōjō 箱嬢 ("case / box / chest / trunk girl"). The male counterpart is erebētā bōi エレベーターボーイ ("elevator boy"), but one doesn't often encounter such individuals. The sole task of the "elevator girl" is to welcome customers into the elevator and to push the buttons for the desired floors of the depāto ("department store"). Such spiffily dressed girls are often hired to do this work as an extra job (arubaito アルバイト [< German "Arbeit"]). In other industrialized countries the elevator operators were abolished after push button operation was invented to replace the earlier and more complicated switches. The Japanese are so attached to the elevator girls ("lift operators") that it is not uncommon to see in elevators without such personnel a sign politely requesting customers to press the buttons themselves.

And then there was "Erebētā akushon" エレベーターアクション ("Elevator Action"), a popular arcade game that debuted in 1983.

Earlier posts on airport Chinglish include "Drawing a line in the noodles" and "Potable and Non-Slip."

[Thanks are due to Robert S. Bauer, Hiroko Sherry, Zhou Yunong, Zhou Ying, Nathan Hopson, Cecilia Segawa Seigle, Grace Wu, Melvin Lee, Gianni Wan, Jing Wen, Wicky Tse, Jiajia Wang, Zhao Lu, Miki Morita, Rebecca Fu, Sophie Wei, and Mandy Chan]


  1. Chad Nilep said,

    January 29, 2012 @ 10:00 pm

    Is the special Cantonese character meant to look like a question mark, or is that just my browser telling me I lack the necessary font?

  2. Victor Mair said,

    January 29, 2012 @ 10:07 pm

    @Chad Nilep

    That's a custom character which consists of 車+立 side-by-side in the space of one character. I have just gone in and inserted 車+立 everywhere it occurs in the post.

  3. AntC said,

    January 29, 2012 @ 10:53 pm

    Annals of airport 'Englese': @Victor, it seems lacking in balance to be giving examples only of Chinese poorly translated into English. (Should that be "Chinese translated into poor English"??)

    Your examples are getting so run-of-the-mill, I'm failing to raise even a wry smile these days.

    Are airport signs in English-speaking countries always paragons of translation into Chinese? If so, it wouldn't be so hard to snoop round those airports with a camera, and a quick photo-edit when you get back to China to paste the English under the Chinese.

    Could we discuss examples of howlers going the other way? After all, English has plenty enough opportunities for ambiguities, homonyms and homographs. ("In a series of Language Log posts, Geoff Pullum has called attention to the prevalence of polysemy and ambiguity")

    Or is the management at English-speaking airports so scrupulous they never get caught out by Google translate? That seems unlikely, but if you really can't come across instances, we could have a discussion about why.

    For all the gazillions of dollars it takes to establish and operate an airport (and the aversion to loss of face), you would have thought that a few extra bucks (of one-off cost) for some decent signage and translations would be soo worth it. What is it about Chinese airport management/officialdom/attitudes to language/whatever that your Chinglish examples are so easy to come by? (Or are they? Are in fact the majority of signs perfectly understandable, perfectly grammatical, even hyper-correct?)

    Are safety announcements and signs as garbled as non-critical messages? Is signage in international airports (with larger proportions of English speakers) better than in domestic? Some of those questions could form decent post-grad research topics.


  4. fs said,

    January 29, 2012 @ 11:23 pm

    @ AntC

    An example of Chinese translated into "poor English" would be

    Yun still ain't no gone home yet.

    "poorly translated" would be more:

    small cloud still not return home

    — the poorness of the translation is not recognizing that 'xiao yun' is a person's name, whereas the sentiment is entirely correct in the first one, just expressed in nonstandard English.

    The voice / sign announcements (DWI, I think, memorably had Japanese and Chinese announcements) in US airports have all sounded like they were done by competent native speakers.

    A suspicion (nothing more): signage is delegated to a lower official who squirrels away the translation budget and hires a free automatic translator, and higher officials don't remember to check; in the US, there's less (?) of a culture of pocketing budget, or management remembers to check.

  5. Jim said,

    January 29, 2012 @ 11:35 pm

    I'd love to read more about such "custom" characters like the elevator girl one. It seems like there would be a lot of interesting examples.

  6. Matt said,

    January 29, 2012 @ 11:54 pm

    Incidentally, there is a special Japanese graph for "elevator girl" (erebētā gāru エレベーターガール) … namely, the radical 女 ("woman / female") on the left and 上 ("up") written above 下 ("down") on the right, which is pronounced as erebētā gāru.

    Are you quite sure of this? I've never seen or heard of this character before, and the closest I could find after a quick web search was someone asking for (but failing to receive, natch) confirmation of a rumor they'd heard that this was how you wrote "elevator girl" in Chinese (here).

  7. Victor Mair said,

    January 30, 2012 @ 12:26 am


    Of course, I've seen it many times, but always handwritten. There are plenty of such characters that don't exist in any established fonts for Japanese. We have the same phenomenon in Chinese — there are thousands of characters that cannot be typeset. The one for Cantonese lip1 that I wrote about in the above post gave me LOTS of trouble because I couldn't get it to reproduce in my blog. It kept coming out as a question mark (?). instead, I had to write it this way: 車+立.

    The HANZI / KANJI / HANJA writing system is open-ended, so there are new characters being coined every day. It is impossible for font makers to keep up with the new characters that are being devised left and right.

    As a matter of fact, I first saw the elevator girl character about 40 years ago, but I still have never seen it typeset.

  8. AntC said,

    January 30, 2012 @ 12:40 am

    @fs re your suspicion of the pocketed budget: do no high-power officials (with a smattering of English) come to the grand opening ceremony? Do no pilots (for whom English is the lingua franca of their trade) walk the escalators/elevators? Do the information kiosk staff not get asked repeatedly? Do high-power salarymen with planes to catch not get annoyed by the knots of foreigners milling around at the escalators? Do they notice the barmy language but say nothing (the Emperor's clothes)? Does the pocket-budgetting official never get found out? Or do they have no shame?

    (Or, as I queried before, is the phenomenon relatively rare?)

  9. Observation said,

    January 30, 2012 @ 1:55 am

    '行人電梯' being commonly used for escalators is news to me. I think it's only used for particularly long ones like the mid-levels one, or maybe only outdoors ones, like the one in Ocean Park, and not the ones commonly found in shopping centres.

  10. fs said,

    January 30, 2012 @ 2:05 am

    [with HUGE warning of CITATION NEEDED (beyond my meager anecdata)]


    It's not rare.

    It's just very hard to hold anyone accountable for the mistakes, much less have them corrected. I dunno; I just accept it as the charm of China. It's as if they want me to try to read the original text, and give enough hints through the botched translation. Clearly, people infer the meaning just fine (info kiosks don't get flooded with confusion).

    As I understand, slight underhandedness is totally accepted, even celebrated as a life skill. If people here pull strings, it's pretty standard to embroider and cross-stitch yonder.

  11. DW said,

    January 30, 2012 @ 2:17 am

    The "elevator girl" character" can be seen in a published book dated 1976 at Google Books: Yuen Ren Chao, "Aspects of Chinese Sociolinguistics", p. 187.

  12. Matt said,

    January 30, 2012 @ 2:47 am

    Fair enough, thanks for the reply! And thanks for the reference, DW. I found another on p158 of "A computational theory of writing systems" by Richard William Sproat that summarizes my thoughts: "[characters like this] are puns and are not seriously considered part of the writing system", even if there is no orthography-related reason why they couldn't be taken as seriously as 峠.

    There are plenty of such characters that don't exist in any established fonts for Japanese. … The HANZI / KANJI / HANJA writing system is open-ended, so there are new characters being coined every day. It is impossible for font makers to keep up with the new characters that are being devised left and right.

    I will cheerfully defer to your expertise re hanzi and hanja, but (with deepest respect and apologies in advance for replying like an irritating know-it-all to a very stimulating post and generous response to my comment on said post!)… this is not an accurate description of the kanji situation in modern-day Japan*. The rate of kanji creation, if you exclude jokes and gimmicks (e.g. new characters used in one advertising campaign to make a point and then forgotten), is basically zero. If the graph for "elevator girl" isn't in font sets and dictionaries (and it might be in the big ones! I haven't checked), that isn't because the editors are so swamped with new kanji that they can't keep up — it's because they didn't judge it to be a "serious" character seeing non-trivial use. (Although, sure, in the computer age this is a vicious circle: if it's not in font sets, people can't type it, so they don't use it, so it sees no non-trivial use, so it's not in font sets…)

    * Concession: since the systems are obviously connected in some ways, any new hanzi could be imported as-is and assigned a Japanese reading. We might see a lot more of this (as in days of old!) as China and Korea continue to grow in importance to Japan. But at the moment it's more or less limited to proper nouns.

    This wasn't the case as recently as the mid-20th C, when you can certainly see new kanji being built from existing parts (most notably for use in words for concepts and objects imported from Europe/the US), but this tradition is moribund now.

  13. AntC said,

    January 30, 2012 @ 4:05 am

    @fs It's not rare. … It's as if they want me to try to read the original text, and give enough hints through the botched translation.

    Your very possibly right, I don't think we even have to posit a corruption/conspiracy theory.

    I lived in Hong Kong for a year in 1990 (so Google translate wouldn't have come in to the picture). Even in reasonably up-market hotels, restaurants and shopping malls, the Chinglish was often execrable. And they'd been ruled by the English for nearly a hundred years, goddammit! I must say the 'charm' wore off pretty quick.

    I didn't "infer the meaning just fine". I often didn't infer the meaning at all. I just hoped it didn't matter, kept calm and carried on. (I was a Brit, after all, there's an Empire to uphold!)

    There were planeloads of English-speaking backpackers passing through all the time. Surely an entrepreneur (of which there were a few ;-) could have get them together with some passing school kids and a dictionary and for the price of a few drinks, gotten something that was grammatical English?

  14. huixing said,

    January 30, 2012 @ 4:15 am

    i doubt 箱嬢 read sōjō or shōjō. if it is a modern thing at a time in japan, probably people more likely call hako jo , a box girl that is. and also when just listen this hako jo more easyly understand than sojo. on lookof it, it looks like scintic word but actuualy not, its a coinage hako plus jo or ojo, ojosan for girl.

  15. Dan Hemmens said,

    January 30, 2012 @ 4:29 am

    Surely an entrepreneur (of which there were a few ;-) could have get them together with some passing school kids and a dictionary and for the price of a few drinks, gotten something that was grammatical English?

    I suspect the answer to this is "probably not". More precisely, you'd get a different *sort* of bad translation: one that rendered into perfectly comprehensible English but which completely missed the point of the original sign.

    Doing accurate translations is actually quite hard.

  16. Victor Mair said,

    January 30, 2012 @ 6:26 am

    From Bob Bauer:

    Much enjoyed reading your very interesting blog about 'elevator' and 'escalator'.

    We should note that the Central to Mid-Levels Escalator 中區至半山行人電梯 (a.k.a. 中環至半山自動扶手電梯系統) is not really a prototypical escalator but a "moving walkway", since it doesn't have any moving steps that one usually associates with an escalator.

    I'm still trying to figure out what is the most common word in use in HK that means 'escalator. At any rate, here are some words that I have found (maybe you have listed all of them in your blog):

    電動扶梯 din6 dung6 fu4 tai1,
    電樓梯 din6 lau4 tai1,
    行人電梯 haang4 jan4 din6 tai1,
    自動電梯 zi6 dung6 din6 tai1,
    自動梯 zi6 dung6 tai1
    自動扶手電梯 zi6 dung6 fu4 sau2 din6 tai1

    By the way, have you noticed that the DeFrancis dictionary [VHM: ABC Chinese-English dictionary] distinguishes between 升降機 'freight elevator' and 升降梯 'elevator'?

  17. Victor Mair said,

    January 30, 2012 @ 6:30 am

    Brendan O'Kane tried to post this yesterday, but for some reason it didn't go up:

    I had an elevator/escalator mixup myself recently when reading reports of the nasty little nationalist Sima Nan's accident — getting his head stuck in a 电梯 — at a US airport. My mental image of what must have happened had been contingent upon it being an elevator; upon finding out that it had been an escalator, I was simply left wondering how dumb Sima could possibly be.

  18. Terry Collmann said,

    January 30, 2012 @ 9:51 am

    AntC – – "I lived in Hong Kong for a year in 1990 (so Google translate wouldn't have come in to the picture). Even in reasonably up-market hotels, restaurants and shopping malls, the Chinglish was often execrable.

    Really? I've been in HK three months now, and the amount of Chinglish I have seen has been tiny to almost non-existent.

    Victor – while parts of the Mid-Levels system are travelator-like, parts definitely have moving stairs – see here.

  19. Victor Mair said,

    January 30, 2012 @ 10:02 am


    What with your deference and concession and numerous qualifications, you've described the situation with regard to the ever-burgeoning HANZI / KANJI / HANJA system fairly well. Just a couple of additional observations at this point:

    1. One of the greatest achievements of the First Emperor of the Qin Dynasty was to rein in the mushrooming writing system over 2,200 years ago. There were countless local variants and special forms (can't call them Hanzi then because the Han Dynasty hadn't even happened yet!). Thus began the endless battle between the perpetual proliferation of characters and the efforts of the authorities to limit their number. But it was essentially a fruitless task, since no sooner had the First Emperor brought some order to the system and reduced the rampant number of characters than it started to soar again.

    In 1999, I published an article with the very long title "On 'Transformationists' (bianjia) and 'Jumbled Transformations' (laza bian): Two New Sources for the Study of 'Transformation Texts' (bianwen), with an Appendix on the phonotactics of the Sinographic script and the reconstruction of Old Sinitic." It was published on pp. 3-70 in Alfredo Cadonna, ed., India, Tibet, China: Genesis and Aspects of Traditional Narrative (Firenze: Leo S. Olschki). In the appendix, I included Fig. 1, "The graphs and words of Sinitic correlated with literary versus vernacular forms of writing", which shows an ever steepening curve in the number of graphs in the writing system. I predicted that the number of graphs would become so huge that it would be in danger of collapsing of its own weight. I think that the danger of that now is even greater because Unicode and other character management programs permit the preservation of all recoverable characters that have ever been invented or devised (even if they were only used once in the whole of history). Thus, what I call "junk characters" (cf. junk DNA) are preserved the same as useful, frequently occurring ones, clogging up the system and making it cumbersome and complicated. "Character amnesia" and other related phenomena brought about by electronic information processing and transmitting devices are hastening the demise of the HANZI / KANJI / HANJA writing system as it is presently constituted, thus giving the lie to the old hope that it would be computers that saved the characters.

    2. I remember about 15 or 20 years ago there was a female author in Taiwan who gave herself the name 女+利 (one character composed of these two elements squashed together side-by-side) (I forget her surname). This caused a bit of controversy because conservatives complained that there was no such character. But she published books and articles using that name, so the printers had to deal with it. I searched high and low for that character but couldn't find it anywhere. Still today, I'm not certain how it is pronounced (probably lì) or what it means (perhaps "female benefit / advantage"?). Imagine my astonishment just now when I was searching around in my computer that I was actually able to bring up this character: 娳! Apparently, this author was successful enough that she has forced compositors to include her once sui generis name in their fonts. This is but the story of how a single graph is invented and becomes incorporated in the writing system, but it is a story that has been reprinted countless times throughout history, and is still being repeated today. Do not let anyone tell you that the creation of new characters has come to an end! Far from it; what with the cleverness of advertising and the spontaneity of the internet, the number of graphs that are being devised is probably occurring at a higher rate than ever before.



    Boy dies in escalator accident in Beijing

    A nine-year-old boy died Sunday after getting stuck in an escalator at a shopping mall in Beijing, an official said.

    The accident took place around 12.30 p.m. at the Xinyidai (or "New Generation")…

    I wonder whether the accident happened on an escalator or in an elevator.


    You begin your first comment dismissive of research on Chinglish, but you finish it by asking some reasonable questions. Thanks to the intervention of fs, who brings some much needed sense and perspective to the issue, you start to take the phenomenon of Chinglish seriously in your second comment. Rest assured, Chinglish is very common in China, and it is source of much embarrassment to people who care about China's image in the world. Witness the huge effort to clean up Chinglish signs before the Olympics in 2008. If you want to learn more about Chinglish, you can read the books and blog of Oliver Lutz Radtke, watch the play called "Chinglish", go back through the old posts on Language Log, and so forth (there are plenty of resources if you're truly interested).

  20. Ellen K. said,

    January 30, 2012 @ 10:09 am

    I'm pretty sure that when people speak of translating into "poor English", they don't mean English of poor people, or English of an informal register, or anything of the sort. Any sentence with "ain't" ain't the sort of "poor English" that is meant.

  21. Ellen K. said,

    January 30, 2012 @ 10:20 am

    Oops, forgot to indicate, the above is a reply to Fs's first post.

  22. fs said,

    January 30, 2012 @ 10:28 am

    @Ellen K

    Oh, of course not — I was struggling to come up with malformed English, so fell into simply using nonstandard English associated with socioeconomically-poor people. This was a cop-out, and a better example might have used a few homophonic mixups, or a richer sentence in Chinese from which to translate.

  23. JMU said,

    January 30, 2012 @ 10:29 am

    The character with the elevator-girl gloss is a Japanese joke at least 50 years old, not a genuine character–though the question of whether it is possible to draw a line between "real" characters and phony ones, like the thousands of non-characters purposely invented by the artist Xu Bing, is of more than academic interest. Japanese often make characters and their (mis)reading the butt of little jokes, often during evening drinking. Long ago, before the jaded Bubble decade, in a tiny Tokyo "sunakku" far away, I occasionally won beers by challenging young fellow regulars that to write the traditional form of various simplified kanji.

  24. Alan Chin said,

    January 30, 2012 @ 11:27 am

    You can type the Cantonese character "lip" by inputting Cangjie "JJYT" :

  25. Victor Mair said,

    January 30, 2012 @ 12:26 pm

    JMU adds another reason for the invention of new characters: they might start out as jokes, but then some of them stay around and become part of the writing system. Other causes attested throughout history: taboo forms, unintentional miswriting, drinking games, imperial megalomania (empress Wu Zetian), personal vanity or idiosyncrasy, what I call "rebranding" (e.g., the creation of the character for “tea” 茶 by intentionally distorting the character for "bitter weed" 荼 — this monumental change, which consisted of erasing one small stroke, probably occurred sometime around the second half of the 8th c. AD; see The True History of Tea, Appendix C), religious ritual, secret societies…. The sources of new characters are endless, and jokes are one such source.

    Now, it must be emphasized the the invention and proliferation of new characters in the HANZI / KANJI / HANJA writing system is an entirely different matter from the creation of new words in a language. The former increases the complexity of the script, the latter increases the richness of the vocabulary.

    Naturally, a system of writing that permits the simultaneous existence of tens of thousands of discrete graphs would be both unlearnable and dysfunctional, so obviously the Chinese characters beyond about the 23,000th in terms of frequency are essentially useless, YET they keep popping up in people's names, in literature, and so on. Moreover, as I mentioned before, new ones keep getting created for all sorts of reasons.

  26. Victor Mair said,

    January 30, 2012 @ 1:06 pm

    Several people have pointed out to me in private communications that they believe the special Japanese character for "elevator girl" was formed by analogy with tōge 峠 ("ridge", i.e., "mountain" on the left + "up" above "down" on the right). They accept tōge 峠 as a legitimate Japanese character, but it must be observed that tōge 峠 does not exist in Chinese, so somebody had to invent it at some time in Japan. The same may be said for the "elevator girl" character.

  27. paula franklin said,

    January 30, 2012 @ 2:51 pm

    YOUR "very possibly right," from ANTC 1/30 ??????

  28. arthur waldron said,

    January 30, 2012 @ 3:35 pm

    Note that at Philadelphia 30th St. Station an old sign directs passengers to "moving stairs." I have never seen that anywhere before and wonder whether perhaps the word "escalator" was originally copyright? Of course putting moving stairs into Kanji is going to lead to all sorts of alarming coinages.

  29. Wentao said,

    January 30, 2012 @ 3:51 pm

    When need arises, apart from calling an escalator 扶梯 or 滚梯, I've also seen the usage of 直梯 to refer to an elevator, although it's quite rare. To me, 升降机 seems ubiquitous in Hong Kong, especially MTR.

  30. AntC said,

    January 30, 2012 @ 4:04 pm

    @ Dan Hemmens "Doing accurate translations is actually quite hard."

    Dan, I fully acknowledge that; I wasn't trying to denigrate the noble craft of translation in any way.

    But we're not talking about literature or exquisite poetry or subtle nuance or abstract ideas – or even a contract for widgets. Neither are we talking about divining the meaning from a remote/inaccessible mind.

    These are instructions of a few (up to a dozen) words in English. We're talking about kids not getting killed in escalators/elevators. You can stand on the left or the right; you can take on a cart or you can't. I suggested back-packers because they'll have gone through a lot of airports/hotels/shops; they'll know the range of meanings. I suggested school-kids and a dictionary so that the translation can be interactive; the team can do cartoon drawings to confirm.

    I repeat: it can't be that hard. There must be some cultural effect going on. It's so prevalent (according to posters here), and so long-standing that I have to suspect it's not a lack of caring (in that case why bother offering any translation? — British airports offered only French 30 years ago). At risk of further upsetting @Victor, it seems to me there's some kind of calculated insult going on: Hey guilao! the middle kingdom doesn't give a ratsarse about foreigners: we want your kids to get their fingers torn off in escalators. We want you to eat chicken's feet and sea slug ha! ha!

  31. AntC said,

    January 30, 2012 @ 4:18 pm

    @ Terry Collmann "I've been in HK three months now, and the amount of Chinglish I have seen has been tiny to almost non-existent"

    Thanks Terry, that's good to hear. (I've not been back since 1990.) Perhaps Google translate is improving matters.

    So who's going to sponsor their post-grads to probe the issues based on evidence more scientifically gathered than anecdote/personal experiences?

  32. AntC said,

    January 30, 2012 @ 4:32 pm

    @Victor "You begin your first comment dismissive of research on Chinglish .."

    I apologise if you took it that way; I didn't intend to be dismissive; I was asking whether (in effect) there's an equivalent 'Englese' phenomenon, and suggesting that researchers into Chinglish might be in the best position to know. I still feel I've not got an answer to that.

    Is there a literature on 'Englese'? I don't expect it uses that term, so what is the term? If there's not a literature, why not? Is it because there's no such phenomenon?

  33. AntC said,

    January 30, 2012 @ 4:44 pm

    @Paula Franklin "YOUR "very possibly right," from AntC "

    Thank you Paula for pointing out my solecism. So it doesn't need a botched translation to make poor English. (@Ellen K/fs: Poor in the sense of ungrammatical.)

    Exactly what I said: "English has plenty … homonyms …". Did my slip confuse the meaning for you? Had you been translating me into Chinese, would you have ended up with incomprehensible nonsense?

  34. John said,

    January 30, 2012 @ 4:49 pm

    We should note that the Central to Mid-Levels Escalator is not really a prototypical escalator but a "moving walkway", since it doesn't have any moving steps that one usually associates with an escalator.

    Yes it does. See the wikipedia article for pictures.

    @AntC 4:04 pm
    I don't think "guilao" is the right term there!

  35. John said,

    January 30, 2012 @ 4:53 pm

    Maybe the ubiquity of 升降机[sic] is because the MTR adopted it as the official "formal" translation. In HK, the Cantonese term is often just pronounced in Mandarin instead of "translating" it to a proper mainland (or occasionally, Taiwan) usage. When speaking Mandarin, I often find myself using the "wrong" term because of prolonged exposure to it in HK!

  36. Dan Hemmens said,

    January 30, 2012 @ 5:42 pm

    I repeat: it can't be that hard.

    And I repeat, it really, really can.

    Do you have any idea of the kind of nonsense you'd get if you tried to get two groups of people who don't speak each other's language to try to do bilingual translation by committee? How would your hypothetical Chinese Airport Administrator even explain this idea to your hypothetical native English speakers? If they can speak English well enough to explain this rather oblique and complex plan to people who speak no Chinese, they can do their own damned translations. If you can say "excuse me, would you mind checking that the grammar on this sign reads naturally to a native speaker of English" then you can say "please stand on the left."

    To put it bluntly, there are three options. (a) Producing accurate Chinese-English translations is harder than you, a person who as far as I know speaks no Chinese, and has never worked in translation, assumes it to be (b) Chinese people are stupid and lazy (c) Chinese people deliberately insult the West with hilariously bad translations.

    I *honestly* don't see how you can look at those three options and consider (c) the most probable.

  37. Brendan said,

    January 30, 2012 @ 9:48 pm

    The thing about translation is that people who haven't done it have got very little appreciation of what's involved. Clients (many of them Chinese) have suggested to me that I could speed up a translation project because it wasn't important to do an especially good job — 差不多就行了 — as if it would only take half as much time to produce a half-as-good translation. Other people have a charming faith in the power of bilingual dictionaries to solve any issue — I remember once seeing a Chinese novelist who spoke no English accusing his American translator of incompetence after having gone through the translated novel with a pocket dictionary. And for most people — especially when it comes to signs like this one — translation is simply an afterthought and the English is one step above ornamental.

    Which is to say that there's plenty of good-faith Chinglish out there, stuff that was generated by well-intentioned and under-qualified people working on tight deadlines for miserable pay, no "calculated insult" necessary. This sign is not that, though: this pretty clearly the product of machine translation, because it contains mistakes that humans generally wouldn't make, but that MT programs make all the time. The heading and first bit of text are enough to illustrate this:

    手推车 ("Carts")

    请走对面电梯 ("Please take the escalator opposite")

    The MT program is understanding this all to be a single sentence: it takes 手推车 to be the subject and the rest to be the predicate, which is why it assumes (bizarrely) that 请 should be "pleases" rather than "to request." "The cart pleases walk opposite elevator" is the result of a mindless algorithm doing a word-by-word translation of text that it doesn't understand.

  38. Victor Mair said,

    January 30, 2012 @ 10:29 pm

    For those who are dubious about the bona fides of the elevator girl character, remember that Lewis Carroll playfully made up words like "chortle" ("chuckle" + "snort" [?]), and some of them stayed in the English lexicon. Zhuang Zi, his illustrious literary ancestor, was also fond of making up words, and many of them remained in the Chinese lexicon long after he was gone.

  39. Keith M Ellis said,

    January 30, 2012 @ 10:49 pm

    I appreciate Brendan's attempt to answer AntC's questions (and cast strong doubt on his intuition about this) because I'm sure I'm not alone here in sharing AntC's curiosity about the prevalence and sociology of Chinglish. Some kind of comparative examination of the quality of translation in these environments would be interesting, though of course fraught with all sorts of complicating factors.

  40. Ellen K. said,

    January 30, 2012 @ 11:14 pm

    Seems to me, translation isn't necessarily needed, as far as correct signage.

    It's not a case of having a particular text that needs translated.

    It's a case of, we need the right words to put on a sign.

    Translating does not have to be the starting point.

    Someone doesn't need to know any Chinese to know the correct English for an airport sign. What's needed, as far as coming up with the correct English, is someone who knows what the sign is for, what it's meant to convey, and can write that in English.

    So, those backpackers someone suggested could help, if we are contemplating whether that's a reasonable way to get the right English for the signs, we should realize that translation is not the goal. Translation is merely one possible way to meet the goal, which is (presumably) English language text that will help travelers who read English find their way through the airport.

  41. Matt said,

    January 31, 2012 @ 12:08 am

    Thanks for the even more detailed reply, Victor — I should be annoying more often. It seems that we mostly disagree on where to draw the lines between admittedly fuzzy areas ("past" vs "current" practices, "real characters" vs "jokes", "Japanese" vs "Chinese" or "non-Japanese" usage).

    They accept tōge 峠 as a legitimate Japanese character, but it must be observed that tōge 峠 does not exist in Chinese, so somebody had to invent it at some time in Japan. The same may be said for the "elevator girl" character.

    Sure, I have no orthographic basis for rejecting the elevator girl character — or any other litmus test except vague handwaving in the direction of usage patterns and "I know it when I see it". 峠 may have been meant as a joke originally, too, for all I know. But even if so, it turned out to be useful and so people used it. This doesn't seem to have happened in the case of "elevator girl". People might *mention* it, but they don't *use* it. Similarly, among Carroll's inventions, "chortle" hit the big time, but "uffish" and "tulgey" didn't. Sometimes the magic just isn't there.

    You could certainly construct a respectable definition of "real" that was broad enough to include "uffish", [elevator girl], "ghoti", Prince's use of an eye icon to spell "I", and so on. But I think that, in the interest of de-fantasifying the discourse around Chinese characters, it's better to maintain a distinction between self-conscious japes and, well, real real characters.

  42. AntC said,

    January 31, 2012 @ 2:34 am

    @John "I don't think "guilao" is the right term there!" has:
    guǐlǎo (Chinese: 鬼佬) Borrowed from Cantonese "Gweilo", "ghost" or "ghost guy", a slur for white people

    That's what I meant. Apologies that I didn't put on the diacritics. In HK, I experienced it as a fairly good-natured term, but some of the middle-class/college educated people I worked with were embarrassed that I knew it.

    So @Brendan, I'm not saying that the "under-qualified people working on tight deadlines for miserable pay" are perpetrating the insult, of course far less the MT. (Channeling Bierwisch's "My typewriter has bad intentions.") They clearly don't know enough English to know whether they're being insulting.

    I did read back some of the Chinglish signs to people in the street (not translators, but taxi drivers, stall-holders, hotel receptionists: the signs on the desk in front of them). They knew enough English that it was garbage. They knew it didn't say what the characters said. They smiled ruefully.

    How about the people who type-set the signs? They at least are careful enough not to introduce mis-spellings.

    I repeat: it can't be that hard to at least know that the sign is garbage. Perhaps the authority gradient is so steep people know but say nothing?

    I am ascribing bad intentions (or at least culpable negligence) to the administrators who cut so little in costs (if that's the explanation). Perhaps they're also cutting costs on airline safety?: China Airlines Always Crashes. Two CAAC planes fell over the end of the runway at KaiTak while I lived in HK.

  43. Chris Waugh said,

    January 31, 2012 @ 5:40 am

    @Victor Mair: "I wonder whether the accident happened on an escalator or in an elevator."

    Definitely an escalator, judging by the graphic and the article in 新京报. 'Escalator' as in "staircase-shaped thing on which the steps move diagonally upwards or downwards conveying people to the floor they wish to move to regardless of whether they stand on one step and allow themselves to be moved or treat it as a staircase and use the movement of each step to accelerate themselves in their chosen direction", and certainly not "box that moves people and goods up and down a shaft".

    @AntC: Listen to Brendan. He's right on the money, as always.

    Oh, and Kai Tak closed a long time ago and China's internal aviation has improved drastically since. Time to update your stereotypes.

  44. Victor Mair said,

    January 31, 2012 @ 9:21 am


    Come to think of it, Shakespeare invented a lot of words (over 1,700) that still exist in our language today, many of them quite common ones that we use every day ( Why did he do this? Well, sometimes it must have been simply because he felt the need for a particular word, but other times he undoubtedly did it out of playfulness. You'll probably dismiss what I'm trying to tell you by saying that "Shakespeare hit the big time".

    I guess you and I are coming at language from two entirely different planes: I from the vernacular / colloquial / topolectal / popular / descriptive plane (sorry, it's my background in Chaucer and Dunhuang bianwen ["transformation texts"], and probably also Osnaburg Township, Ohio), you from the standardized / normative / authoritative / elitist / prescriptive plane.

    I like to think hard about why we have a character like the ONE (forgive me for not being able to type it) that stands for the two characters of púsà 菩萨 (Bodhisattva) — consisting merely of the two grass radicals stacked one upon the other. That started out as an ad hoc timesaver, and for centuries must have existed only in handwriting. Even now it is considered substandard, but it was invented by *someone* and has been used by thousands upon thousands of individuals who don't give a fig for how many characters are supposed to exist in a normative / standardized / authoritative / official / prescribed set.

    BTW, púsà 菩萨 actually comes from pútísàduǒ 菩提萨埵, so púsà 菩萨 itself is a kind of miswriting, but you accept it because it "hit the big time". Oh, yeah, I should also mention that people are very confused by the last syllable of pútísàduǒ 菩提萨埵, "miswriting" it as 捶, 锤, 缍, etc. and "mispronouncing" it as chui, zhui, and so forth. For a really fun experience, count the strokes of the phonophore in the last four characters, and try to figure out the radical when it stands alone.

    It might be interesting (but also at times very exasperating for both of us) to have you in a class sometime. I could go on like this for years. Matter of fact, that's what I do all the time on Language Log! It's like having you in a class.

    To conclude this class / comment, Chinese and Japanese (and Koreans and Khitans and Vietnamese) like(d) to create characters and words; English speakers like to invent words too, but they're not into inventing letters of the alphabet.

  45. Nick Lamb said,

    January 31, 2012 @ 10:26 am

    OK, since it seems no-one else asked, I will be the boring technical nerd and ask:

    Has this elevator-girl character been assigned a Unicode code point? If not, is anybody trying to get that to happen?

    The Unicode/ ISO 10646 standardisation process looks kindly upon every sort of invention so long as they're not proprietary (so no Apple Computer logo) and are actually being used, for example U+29CE is a pair of left/right triangles, which no doubt began as a scribble on a blackboard in a maths discussion; U+2623 is the universal Biohazard sign and U+1F021 is the symbol used on a Mahjong title for the nine of circles.

    If the elevator-girl character actually appears as a distinct glyph in written materials, even if only as a sort of recurring joke, it may well be justified in appearing in the next edition of 10646. But it will need a champion, and that champion should probably be a native Japanese.

  46. Boris said,

    January 31, 2012 @ 12:40 pm

    When I was first learning English back in Russia I was told that "Elevator" can also be used to mean "Escalator" (though "Escalator" was taught too, what with it being the same word as in Russian). I never heard this usage in the US and Russian has separate words for them, so I don't know where this came from. Is there some variety of English where this is the case?

  47. michael farris said,

    January 31, 2012 @ 1:55 pm

    "English speakers like to invent words too, but they're not into inventing letters of the alphabet"

    But I was under the impression that most or all of these more recent inventions are combinations of well known character elements in new ways.
    It doesn't seem as much like creating new letters of the alphabet as creative spelling which many English speakers …. luv.

    I understand the technical difference in printing these new creations but the mechanics don't seem that different from noo wayz uv spelling English wurdz.

  48. Victor Mair said,

    January 31, 2012 @ 2:34 pm

    @Michael Farris

    No, it's like creating a new letter. You need a whole new code point in Unicode. That's why Chinese characters take up the vast proportion of the entire Unicode. All the alphabets and syllabaries and mathematical symbols and what not of the whole world occupy a miniscule amount in comparison.

    And here's the elevator-girl character in all its glory, designed in red-and-black by Richard Cook, a Unicode specialist (the glyph was made using software):

  49. michael farris said,

    January 31, 2012 @ 2:54 pm

    My point wasn't the technical difficulties of integrating new characters into computer displays or printing (I realize the headaches).

    What I wanted to way was that the elevator-girl character is made up of three elements that anyone with even the most basic knowledge of characters should be very familiar with (like me). In terms of its internal structure it's a new combination of familiar elements and not an entirely new creation.

    The problem of course is that there's no way (at present) to easily allow for that kind of creation in any medium except handwriting.

    Theoretically I think that characters could have been reformed so that that wouldn't have been a problem and so that characters weren't the unicode space hogs that they are now. But they weren't and most likely won't be.

  50. Dan Hemmens said,

    January 31, 2012 @ 3:04 pm

    It doesn't seem as much like creating new letters of the alphabet as creative spelling which many English speakers …. luv.

    And which many others h8?

  51. Dan Hemmens said,

    January 31, 2012 @ 3:30 pm

    With apologies for double-posting…

    What's needed, as far as coming up with the correct English, is someone who knows what the sign is for, what it's meant to convey, and can write that in English.

    This might sound like the end of a not-very good joke, but this immediately raises the question of how you explain to your English speaker what the sign is supposed to convey.

  52. Victor Mair said,

    January 31, 2012 @ 4:56 pm

    @ michael farris

    I think you're starting to grasp the enormity (in two senses) of the problem.

    I will write a separate post about this, but can't do it today because I'm trying to finish up a paper.

  53. Terry Hunt said,

    January 31, 2012 @ 7:19 pm

    @ Victor Mair

    "Come to think of it, Shakespeare invented a lot of words (over 1,700) that still exist in our language today . . . "

    Going off at a complete tangent (for which apologies), but regarding this fairly well-known observation – how do we know in each case whether Shakespeare himself actually made the word up or whether he was using a word already current (though perhaps of recent coinage) and was merely the first to use it whose usage has survived in print or documentation?

    Given that Shakespeare was primarily writing plays whose words were intended to be spoken, and heard fleetingly, rather than be in print to be pondered over, a high neologism rate would surely have bewildered much of his audience. If however he was in many cases merely an early (and inadvertant) recorder in (subsequent) print of expressions already going the verbal rounds in London or the Court, the groundlings would be impressed by his modernity but not nonplussed.

    I ask from the perspective of an SF and Fantasy fan: writers in these genres are particularly prone to neologisms for obvious reasons, but I would think few even of them rival such a degree of original coinage.

  54. Victor Mair said,

    February 1, 2012 @ 8:34 am

    From Hiroko Sherry:

    There are a lot of games/riddles using Kanji in Japan.Asking how to read a newly created fun Kanji is one of the patterns. The one I like best is replacing three 木s  with three 人s in 森. How do you read this? The answer is ' (三人寄れば、)文殊の知恵’,'(san nin yoreba,) Monju no chie'. This is based on a Japanese saying which means 'if three people(brains) get together(cooperate), you can expect an intellectual power/wisdom of Manjusri Bodhisattva.' (文殊菩薩:もんじゅぼさつ/monju bosatsu)

    The Kanji for 'elevator girl, '女+上+下", is one of those 漢字なぞなぞ(riddles) examples, and unfortunately the future of this Kanji is dismal just like the elvator girls' job itself. Different from 峠、I don't think '女+上+下' will ever make it into a Japanese Kanji dictionary entry especially because it's from English (Romaji). Fortunately or unfortunately, we have Katakana. 'エレベーターガール' will be kept in Katakana, and the best it can do is appear in a Japanese Language dictionary (国語辞典).Once the elevator girls disappear from Japanese department stores, which is very likely, I think, the word will be dropped, too.

  55. Victor Mair said,

    February 1, 2012 @ 8:48 am

    from D. Neil Schmid:

    An interesting series of posts. I'm not entirely sure of your interlocutor's (Matt) point. Chữ Nôm provides an excellent lens to refocus some of these issues. The creation of novel characters in Japanese (and Korean I imagine) is nil because other writing systems in those languages allowed for the creation of new words without blending characters or their parts. It's not because those blended characters wouldn't somehow be 'real' or legitimate. The antithesis is Chữ Nôm which only had characters as a basis and so exploited those in a recombinant fashion. However, once an alternative – the Latin alphabet – came along, the entire system could be dumped and was (with a variety of profound consequences). But people certainly did use these "non-standard" Hanzi and they were certainly very real to them.

  56. January First-of-May said,

    February 6, 2012 @ 11:12 am

    For the question of "Shakespearean words": see no further than Wikipedia's list of "English words first attested in Chaucer" [would've put a link, but have no idea how – first time poster]. There's a huge amount of them – from "army" to "create" to "annoyance" to "galaxy" to "bribe" to "digestion" (and I'm only listing some of the ones that still remain common now); and most of them (including all six of the above) are incredibly unlikely to have actually been invented by Chaucer himself (I doubt he really invented any words, actually; Shakespeare certainly did, but most of those that can realistically be attributed to him specifically are spur-of-the-moment coinages*, and thus unlikely to actually be used in the modern word other than as a specific reference).

    I wonder, indeed, had some dialectal words from Huckleberry Finn entered major vocabulary, would we now credit Mark Twain with those words' invention? Not as unlikely as it sounds – for example, have you noticed all those cries of how the entire Ukrainian language (or at least half of it) is only based on the works of Shevchenko? Well, that's mostly because Ukrainian was still (considered) a dialect by Shevchenko's time, so hardly anyone actually wrote in it (and not in Russian) – and Shevchenko wrote enough to include about half of the actual language.

    On-topic: yeah, the Russian word for "escalator" is "escalator" (well, technically it's эскалатор, but even here I doubt all readers know Cyrillic). I didn't know, however – until this discussion at least – that the English word was also "escalator" (I thought for some reason that there was some other term in English – like, I don't know, "moving stairs" or something).
    Coincidentally, the Russian word for "elevator" is "lift" (лифт), and an "elevator" (элеватор) would be what Wikipedia calls a "bucket elevator".

    *) Not least because this is just about the only plausible way we can know a word's inventor reliably enough: if that word only makes sense in context.

  57. Marian Rosenberg - Haikou #1 Translation Agency said,

    February 8, 2012 @ 12:58 am

    Since no one else has mentioned it, the hospital near where I live has a note on the escalator 电梯 telling old people, minors, pregnant women, and those who may easily get dizzy and fall down that they should go and use the elevator 箱式电梯 at the other side of the lobby.

    This is the only place where I have seen this construction.

    I checked Baidu and got 21,000 hits on 箱式电梯. After seeing someone ask the question is it "箱式电梯" or "厢式电梯", I got 24,000 hits on "厢式电梯".


  58. B.Ma said,

    March 3, 2013 @ 5:31 pm

    I think most people realize that the English on many signs is poor. But nobody really cares enough to change anything. The only time this can be fixed is during the design phase. There are plenty of people who can speak two languages well enough to fix all the signs. If I ever opened a restaurant or something, the English and Chinese would all be in line with regular usage. But that's not likely to happen.

  59. web surfer said,

    March 12, 2013 @ 4:41 am

    Excellent work, though 'roughhouse' is slang and probably alien to Commonwealth nations, at least it's not used in England.

    Ironically, by using a peculiar word you've done what the Chinese often do when they translate into English. You can just use the word 'play', I think.

RSS feed for comments on this post