His Coffeeness

« previous post | next post »

Kendall Willets had long ago noticed that Korean honorifics show up disproportionately in commercial settings, but this article brought up something new.  The -si- 시 infix is only supposed to apply to the verb if the subject has higher status, but in service settings it's expanding to everything, including coffee.

The big LOL sentence for me, was when the Coffee 알바 (short for 아르바이트, “Arbeit(work)” from the Japanese-German arubaito/baito which denotes part-time workers in Korea) in the video says,

그 사물 들에게 우리는 존경의 마음을 억누를 수 없습니다. 커피 나오셨습니다. 커피가 제 시급보다 더 비싸거든요.

roughly translated as:

“We cannot control the boundless respect we have for these things. “Here’s your coffee.” (this is the kind of sentence they are talking about, which to my ears, can only be translated into English as (with a little bit of exaggeration) “His Coffeeness has graced us with his presence.” Then she goes on to say “It’s because (a cup of) coffee is more expensive than my hourly wage.”

[VHM:  sic (punctuation and all); emphasis in the original]

One might prefer "unable to suppress" over "cannot control" in the first sentence but that's nitpicking.  The second sentence reads literally "the coffee has come out" (i.e., "here's your coffee") with an honorific infix added to the verb.  The third sentence is translated correctly.

The video that is linked to in the article has been shared among many linguists and language practitioners to alert them to the inappropriate use of honorifics in Korean. This overuse of honorifics seems to have become a new tradition, especially in customer service areas, and has stirred up quite a controversy in Korea. This satirical video was produced to exemplify how inappropriate it is to use honorifics for objects, because this implies that the speaker is inferior to the objects that are so honored.

The coffee shop worker (alba) says:

커피 나오셨습니다. 이쪽이 라테십니다.

keo-pi na-o-syeot-seum-ni-da. i-jjo-gi ra-te-sim-ni-da. (RR)
k'ŏ-p'i na-o-syŏt-sŭm-ni-da. i-tcho-gi ra-t'e-shim-ni-da. (MR)

Here is (your) coffee. This is (your) latte.

She uses the Korean honorific affix 으시/으셨 (-usi [present, RR], -ushi [present, MR] / usyeot [past, RR], usyŏt [past, MR]) to honor the customer's coffee and latte, which is a major no-no in Korean grammar. The honorific affix "으시" is attached to a verb or adjective stem to express the speaker's respect to the human subject of the sentence.

After the coffee shop worker, the car sales person also uses 으시 to honor the engine and the tires in the video:

엔진은 터보이시고요. 타이어는 광폭이십니다.

en-jin-eun teo-bo-i-si-go-yo. ta-i-eo-neun gwang-po-gi-sim-ni-da. (RR)
en-jin-ŭn t'ŏ-bo-i-shi-go-yo. t'a-i-ŏ-nŭn 'gwang-p'o-gi-shim-ni-da. (MR)

This car has a turbo engine (literally, the engine is a turbo). It also has wide tires (literally, the tires are wide).

This attitude toward honorifics in Korean seems to differ from that in Japanese where it is obligatory to speak of "honorable tea", and so forth.

[Thanks to Haewon Cho, Bob Ramsey, and Bill Hannas]



  1. dainichi said,

    April 10, 2014 @ 10:31 pm

    "This attitude toward honorifics in Korean seems to differ from that in Japanese where it is obligatory to speak of "honorable tea", and so forth."

    Well, we're talking about honorific verbs, right? Even in Japanese, using honorific verbs (or verb forms) for objects is considered non-standard

    * ネクタイが汚れていらっしゃいますよ。
    nekutai-ga yogorete irrasshaimasu yo
    necktie-NOM become-dirty is(hon) modal-particle.
    Your necktie is dirty.

    As for honorific forms of nouns, there's two kinds, the one honoring the noun itself, like お茶, o-cha, honorable tea, and the one honoring the owner (or otherwise affiliate) of the noun, like お車, o-kuruma, car belonging to someone honorable. You can use o-cha for your own tea, but not o-kuruma for your own car.

    But even when honoring the tea itself, it is not standard to use honorable verbs with it, say

    * お茶がいらっしゃいました
    o-cha-ga irasshaimashita
    HON-tea-NOM came(hon)
    The tea is here

  2. Nanani said,

    April 11, 2014 @ 1:05 am

    Adding on to Dainichi's comment, the honorific prefix o- (or go-) in Japanese is never used with loanwords, so you would never use it with coffee, but you would with tea.

    HON koohii (where koohii is a straightforward borrowing of "coffee")
    HON tea (where 茶 (cha) is a native Japanese noun).

    So, using honourifics to serve a coffee would raise additional eyebrows.

  3. valency said,

    April 11, 2014 @ 3:02 am

    It's worth noting that, in Japan at least, "baito keigo" (part-time worker keigo) is considered by some older traditionalists as an utter bastardisation of the fine nuances of "true" sonkeigo language. So who can say what weirdness is possible — even some native speakers raise eyes at some usages.

  4. Jongseong Park said,

    April 11, 2014 @ 3:12 am

    This phenomenon of inappropriately overused honorifics in the service industry has been remarked upon for several years at least, safely over a decade. But it seems to be getting more widespread if the increasing number of articles and news reports devoted to the subject each hangul day in recent years are to be believed (for most Koreans the holiday is seen to be about the Korean language in general, not just the alphabet hangul).

    Allegedly, it started in department stores, and has spread to other stores, cafés, restaurants, service centres, and indeed any and all areas with interactions with customers as you can see in the satirical video.

    Language purists can't be the only the ones peeved about the inappropriate honorific use. The following article in Korean has the results of a couple of surveys on this:

    "문의하신 상품은 품절이십니다" 어색한 사물존칭은 왜 생겼을까

    In summary, 86% of the call centre agents surveyed were aware that 사물존칭 samul jonching "object honorifics" was "bad grammar", and 69% reported using this at least from time to time even though they were aware that it was wrong.

    In another survey, respondents chosen from the general population were asked about these sentences:
    요금은 2000원이세요. 손님에게 어울리는 색은 파란색이세요.
    yogeum-eun icheonwon-iseyo. sonnim-ege eoulli-neun saeg-eun paransaeg-iseyo
    "The fee is 2,000 won. The colour that goes with you (literally: the customer) is blue."
    Here, honorifics are applied to fee and the colour blue respectively in the sentences.
    Of the respondents, 32.4% said that these sentences were felt to be uses of honorifics but somewhat awkward, and 22.5% said they were honorific but very awkward. Only 19.3% said they were felt to be honorific and somewhat natural, and 15.5% said they were honorific and very natural. 10.3% replied that they didn't even see this as honorific at all, meaning that overall, 65.2% of those surveyed had a negative attitude to "object honorifics".

    If most of the users as well as the general public see "object honorifics" as incorrect, then why is it becoming so prevalent in the first place? From anecdotal evidence, people who deal with customers say that they are forced to use these honorifics because there is always a customer who will get upset if they are perceived to use the familiar form. So they are forced to add honorifics even where they are inappropriate.

    According to this explanation, it's a vocal minority that is changing the language in a way contrary to the wishes of the majority.

  5. flow said,

    April 11, 2014 @ 5:36 am

    "You can use o-cha for your own tea, but not o-kuruma for your own car.
    But even when honoring the tea itself, it is not standard to use honorable verbs with it, say * お茶がいらっしゃいました"—"Adding on to Dainichi's comment, the honorific prefix o- (or go-) in Japanese is never used with loanwords"

    acc to Roy Andrew Miller's 'The Japanese Language' (1967, quoting from the 1993 German edition) it's not quite clear whether the o-/go- in o-cha is the 'true' honorific prefix. he distinguishes several variants of the prefix: one where it is used for semantic distinction as in naka 'middle' vs o-naka 'belly', one where it is optionally prefixed as in o-cha, o-biiru (sic), o-kashi (and found more frequently in the speech of females than in the speech of males), and one where it is used to indicate that the noun belongs to or is intended to 'speak to' the listener / reader (cf. the headline 'go-annai' 'information (for You)' that graces informational placards in subway stations).

  6. Jamie said,

    April 11, 2014 @ 5:37 am

    I wonder if the use of honorifics in Japanese (such as お茶) was also considered non-standard, initially.

  7. Victor Mair said,

    April 11, 2014 @ 7:12 am

    From Bob Ramsey:


    None of those romanizations are right. Why? All the syllables are separated by hyphens! MR never sanctioned such a thing. As I wrote earlier, Gari Ledyard bemoaned the fact that some Koreans use hyphens to indicate the syllabic nature of the script, because he said it made the language look primitive to Western eyes, reminiscent of 19th-C transcriptions of Injun talk. And they really do look awkward, don’t you think?


    My reply:


    I quite agree with you, BUT, in romanizations of Korean, after we remove the hyphens, should we:

    1. separate each syllable with a space?

    2. try to join syllables into words the way we do with Hanyu Pinyin in its proper orthography?

    If the latter, I suppose that we may use the hyphens in the romanizations given in the original post as a guide for aggregation. My preference, of course, is for joining the syllables into words, but is that more or less standard practice in the romanization of Korean nowadays?


    Bob's response:


    Join into words, of course! After all, hangul orthography itself also has spaces these days. The spaces are not always used consistently by most Korean writers, but Korean word processing programs specifically have spell-check that corrects those inconsistencies.

  8. Victor Mair said,

    April 11, 2014 @ 8:25 am

    Another interesting comment from Bob Ramsey:


    Koreans are much, much quicker to adapt to such differences than Chinese and Japanese—actually, sometimes too quick to make changes, at least for my tastes (since I’m basically a bit conservative about traditions). But seriously, hangul spacing is not always well thought through. Martin always liked more liberal spacing, and I think he was wise. For example, he always separated nouns from from enclitic particles, whereas Koreans write solid. But anyway, take a quick look at almost any Korean publication, and you’ll see what I mean! Sometimes to save print space, newspapers will squeeze the words together, but just about all books have spaces between the words.

    Oh, and I probably should note that North Koreans use a lot fewer spaces than is customary in South Korea. Anyway, please forget spacing and hyphens just to separate syllables. That usage is really just a lingering relic of Sinographic writing–you know, where the only spaces are those between the characters.

    Ah, another note. The next winter Olympics will be in Korea, as you know, and the town where it will be held is 평창, which I see the Korean government has been romanizing as PyeongChang, with the capital C being used to indicate the syllable break. Unfortunate and awkward, in my opinion.


    My reply:


    The InTerCaPiTaLiZaTion of PyeongChang is really ungainly and unfortunate.


  9. Jongseong Park said,

    April 11, 2014 @ 8:39 am

    Bob Ramsey is quite right. Speaking for Revised Romanization (RR), it does allow hyphens in cases where the syllabic division is ambiguous, e.g. 세운 se-un (c.f. *슨 seun) and 중앙 jung-ang (c.f. *준강 jun-gang).

    It also exceptionally allows hyphens to be used to separate syllables in given names. So 홍길동 is supposed to be written Hong Gildong, but Hong Gil-dong is also allowed.

    Hyphens are also used to separate the words used for administrative divisions, e.g. 제주도 Jeju-do ("Jeju province"), 의정부시 Uijeongbu-si ("city of Uijeongbu").

    The rules are not very thorough, but the principle should be to use spaces between words and to leave syllabic divisions unmarked. A crucial advantage compared to Chinese or Japanese is that Korean orthography already marks most word boundaries with spaces.

    Korean orthography, however, puts no spaces in front of particles, although for transcription it is useful to mark them separately from the words they are attached to. So in my comment above, I used hyphens to separate particles as a compromise solution, e.g. 요금은 yogeum-eun where 은 eun is the particle.

    This does lead to cases like 색은 saeg-eun /sɛ.ɡɯn/ where the particle leads to the voicing of the coda of 색 saek /sɛk/, which might not be the most elegant solution. (I do note that in the post, 엔진은 enjin-eun and 광폭이십니다 gwangpog-isimnida are treated inconsistently regarding the placement of the hyphen before or after the coda.)

    I did hesitate with 어울리는 eoullineun, which I wrote as eoulli-neun above to separate the verb stem and the ending but which I now regret. Upon further reflection, there is no reason to mark the internal divisions of verb forms.

    The other difficulty is that Koreans are often liberal in writing compound words without spaces, and often dependent nouns are also written without spaces despite the recommendations of official orthography (in my experience, most Koreans make these errors where they omit obligatory spaces, even those who rarely commit other spelling errors).

    My personal system would be based on the principle of separating particles (I count all forms of -이다 ida for this purpose) from the words they are joined to with hyphens, and of using space to separate compound nouns and verbs that are often written without spaces in Korean (e.g. 사물존칭 samul jonching, 도와주다 dowa juda).

    By the way, in the version of MR I know, 라테 would be written lat'e because the ㄹ represents an original /l/ in "latte".

  10. Jongseong Park said,

    April 11, 2014 @ 8:44 am

    To point out some transcription errors, it should be: "으시/으셨 (-eusi [present, RR], -ŭshi [present, MR] / eusyeot [past, RR], ŭsyŏt [past, MR])"

  11. Victor Mair said,

    April 11, 2014 @ 9:25 am

    From Bob Ramsey:


    Jongseong Park’s comment is excellent: a good, clear explanation of what is going on in your original posting. As he explains, it’s not the coffee that’s being honored; it’s the person to whom the coffee is being served! That’s a very natural direction for a language to change.


  12. leoboiko said,

    April 11, 2014 @ 9:45 am

    Recently a young Japanese woman told me of a dialectal utterance which amused her. It was something she heard from a Kyōto speaker: Ara, niwatori-san wa aruite-har-imasu "Look, Ms Chicken is walking!"

    -haru ( < naharu < nasaru) is a respect (sonkei) suffix in Western dialects, but they say in modern Kyōto its frequent use lowered the impact, and now it only shows neutral, everyday politeness (teinei). The (Eastern) woman who told me this probably wasn't aware of that, and was impressed with the use of -haru when talking about an animal, and also with the -san. Of course this all feeds into the stereotype of Kyōto people being overly polite.

  13. Neil Dolinger said,

    April 11, 2014 @ 1:07 pm

    I note that in both examples given by Dr. Mair above, the isolated -usi [present, RR] and usyeot [past, RR] undergo changes in the romanization, becoming i-si and o-syeot respectively (with similar transformations occurring in the MR versions).

    Does this reflect (1) phonological change (vowel harmony?) rules that depend on the words preceding -usi / usyeot, and (2) if so, are these rules also embedded into the romanization rules? I ask as a point of comparison with Mandarin, where some phonological harmonization seems to occur, but is not usually reflected in the romanization. One example for this are words like 什么, which is romanized as "shenmo" but pronounced "shemme". Another is the change in words like 老板lǎobǎn comprised of two syllables with third-tone pronunciation in isolation, but pronounced as if the first syllable had a second-tone. The romanization of lǎobǎn does not change to reflect the pronunciation.

  14. Jongseong Park said,

    April 11, 2014 @ 1:53 pm

    @Neil Dolinger:
    Note that the correct romanizations are -으시- -eusi- and -으셨- -eusyeot- in RR, with eu corresponding to MR ŭ. In the examples above, it is not the -으- -eu- part that is changing—in fact, this part is not used at all.

    In fact, it is misleading to give the forms starting with -으- -eu- as the isolated forms. I would say that the primary forms are simply -시- -si- and -셨- -syeot-.

    In 나오셨습니다 naosyeotseumnida, 나오- nao- is the verb stem. In 광폭이십니다 gwangpog-isimnida, -이- is the stem of the "to be" particle (at least, it is traditionally treated as a particle although it conjugates like a verb). So these simply join -시- -si- and -셨- -syeot- to the verb or particle stems.

    Rather, it is better to say that -으시- -eusi- and -으셨- -eusyeot- are the regular variant forms that are used when the stem has a consonant coda other than ㄹ -l (which drops off before -시- -si-). The -으- -eu- /ɯ/ functions as the epenthetic vowel in these cases.

    So, to illustrate,
    1. stem ends in vowel
    하- ha- "do": 하시다 hasida honorific "does", 하셨다 hasyeotda honorific "did"
    2. stem ends in ㄹ -l
    알- al- "know": 아시다 asida honorific "knows", 아셨다 asyeotda honorific "knew"
    3. stem ends in consonant other than ㄹ -l
    받- bat- "receive": 받으시다 badeusida honorific "receives", 받으셨다 badeusyeotda honorific "received"

    Finally, -셨- -syeot- is merely the contracted form of -시었- -sieot-, which is a straightforward combination of -시- si and -었- -eot-, which indicates past tense.

    Coming back to your question on romanization, there are lots of regular phonological changes in Korean of this sort, where you get different forms of endings depending on whether the stem has a coda, or depending on the vowel harmony (though this does not apply in the case discussed here). The great thing about Korean being written in an alphabet is that these alternations are directly reflected in the spelling, just like in English you use different spellings for "a" or "an". So romanization simply follows the Korean spelling and automatically reflects these various context-dependent forms.

  15. KWillets said,

    April 11, 2014 @ 1:57 pm

    @Neil Korean affixes usually undergo a lot of contraction or deletion. The vowel prefix here is deleted if the syllable preceding it ends in a vowel; it's basically a placeholder to prevent difficult consonant combinations.

    na-o-syeot-seum-ni-da = na + o + usi + eot + seum-ni-da
    (out+come+honorific+past+polite formal tense)

    There's no easy way to map these combining rules into a romanization.

  16. Jongseong Park said,

    April 11, 2014 @ 2:17 pm

    Either -으시- -eusi- is the form of -시- -si- with an epenthetic vowel -으- -eu- inserted, or -시- -si- is the form of -으시- -eusi- with the -으- -eu- deleted. Alternations of 으 eu and zero are frequent in Korean, and Korean linguists have argued for both positions—insertion or deletion. Lately, it's been suggested that the alternation is due to phonologically conditioned realizations of an underlying empty nucleus (I admit that this is a bit too technical for me).

  17. Stephan Stiller said,

    April 11, 2014 @ 3:21 pm

    @ Jongseong Park

    [RR] also exceptionally allows hyphens to be used to separate syllables in given names. So 홍길동 is supposed to be written Hong Gildong, but Hong Gil-dong is also allowed.

    I find the look of a lowercase letter immediately after a hyphen in a given name odd.

    Note though that you see this in (many?) Taiwanese names too, eg: Chen Shui-bian, Ma Ying-jeou. I don't know how Taiwanese passports are romanized, but this topic has come up a couple of times for me, and the problem is at least that the civil servants in charge are inconsistent.

    In HK names, the most common thing seems to be to separate the two syllables of the largely disyllabic given names by a space: "CHAN, Yiu Lun". I have seen this on Hong Kong ID cards, but don't know whether this is obligatory (for names first recorded there). I know some HKers who run their given name together (that would be "CHAN, Yiulun"). I've seen spellings of the form "CHAN, Yiu-Lun" and "CHAN, Yiu-lun" as well. Needless to say, I prefer the "Yiulun" and "Yiu-Lun" variants. By the way, HK name romanization is generally inconsistent. I'm wondering whether they simply have a character→romanization table in some database (and I'm definitely wondering whether it's possible to get hold of that table – if it's at least consistent for each character, it should be possible to reconstruct it from names in HK). If it is "identical" to what Wikipedia terms the Hong Kong Government Cantonese Romanisation, it is inconsistent even on the character level (樹 is C syu6 and M shù and is written "shue" and (puzzlingly) "shuo" on the two signs shown there within the same name 榕樹下). Finally, many HKers with "English names" find it weird to be addressed via their Chinese/Cantonese name: for them the adopted (often/?mostly unofficial) English name is their de-facto given name in all contexts except administrative ones.

  18. Cecilia Segawa Seigle said,

    April 11, 2014 @ 4:19 pm

    His “coffeeness” was very amusing and entertaining.
    I do not know Korean and Korean honorifics, so I cannot offer any comment.
    As for the Japanese Keigo, if you begin to discuss it, you will have to write a whole book.
    There are many classification in keigo, such as 尊敬語、謙譲語、鄭重語、丁寧語、美化語。I think many of you can guess what they are; if not, you can look them up in Google.
    Today, many Japanese have difficulty using honorifics correctly because for many people, keigo is just the matter of habit, upbringing, picking up through hearing.
    The case that “dainichi” commented, “nekutai-ga yogorete irrasshaimasu yo” or “o-cha-ga irasshaimashita” are simply the wrong application of sonkei-go.
    On the other hand, “Ieoboiko” pointed out that applying the Kyoto dialect of “-haru ( < naharu < nasaru) is a respect (sonkei) suffix in Western dialects, but they say in modern Kyōto its frequent use lowered the impact, and now it only shows neutral, everyday politeness (teinei).” This is an endemic phenomenon and it does not apply to the standard Japanese language or to the entire Japan.
    “Nanani” said “the honorific prefix o- (or go-) in Japanese is never used with loanwords,” and he is basically correct. But they do say “o-koohii”(almost strictly wome). As “flow” said, “O is optionally prefixed as in o-cha, o-biiru (sic), o-kashi (and found more frequently in the speech of females than in the speech of males”
    Some men say ocha (tea) or okashi but never “o-biiru (sic)” More emphatically male-male will say “Cha o kure!” (Give me tea!) “biiru o kure!”(give me beer!) .
    They will NEVER say “onaka ga itai” (my stomach (or belly) hurts) they will say “hara ga itai” whereas women will not likely say “hara ga itai” or “hara ga hetta” (I’m hungry), unless she is trying to assume a sexless or transgender status.
    Honorific “O” is used abundantly by women, and in most cases they are quite natural. But an overuse can happen from a mistaken notion for gentrification or beautification 美化. I don’t know the newest contemporary trend, but there was a period within the last 50 years or so, when certain women added O- on everything, e.g. every kind of fruit (O—ringo (apple), O-budo (grapes), O-ichigo (strawberries), O-banana, etc), even O-sandwich, which sounded very silly. This was a limited phenomenon, which came from the notion that putting O on everything will make the speaker sound and appear more elegant (this was prevalent, for example, among the mothers of children who were trying to enter prestigious kindergarten or elementary school, to make them sound more elegant and aristocratic). They might have discontinued this habit – after they have gentrified everything to death.
    I don’t think people said “O-kōhī” in the early modern age when Coffee was first introduced to Japan.

    Honorific O for gentrification had a special place during the Edo-period, around the Edo castle – the shogun’s residence and government. Many objects, action, structures that had to do with the shogun, and especially with his wife’s living compound (Ōoku) had the prefix of O or Go or Gyo. Some words required O for the shogun’s wife but not for the shogun.

  19. Neil Dolinger said,

    April 11, 2014 @ 4:49 pm

    @Jongseong Park, @KWillets,
    Thank you both so much for the explanations for the changes in romanization. Although I visit Language Log every day and have been interested in the architecture of languages from the time I entered elementary school, I earn my keep in another field. Its these kinds of discussions that let me totally geek out and pretend that I am a real linguist!

  20. Stephan Stiller said,

    April 11, 2014 @ 5:00 pm

    @ Neil Dolinger
    Actually, Mandarin has very little morphological changes induced by adjacent syllables. Tonal change is one of the few case where such a thing occurs. 什么 works as an example too, but it is romanized as "shénme" and (imo) pronounced "shémma".

  21. Stephan Stiller said,

    April 11, 2014 @ 5:05 pm

    very little -> very few
    few case -> few cases
    [tired, tired ...]

  22. Stephan Stiller said,

    April 11, 2014 @ 5:31 pm

    (Finally I've read through the whole thread up till now.)

    @ Jongseong Park

    The other difficulty is that Koreans are often liberal in writing compound words without spaces, and often dependent nouns are also written without spaces despite the recommendations of official orthography (in my experience, most Koreans make these errors where they omit obligatory spaces, even those who rarely commit other spelling errors).

    I think this reflects something fundamental: People want to spell words as words, without spaces in the middle. I don't know in which cases it increases legibility to orthographically separate compound words into their components, though. And there are differences between orthographies which I won't get into here and now.

  23. Victor Mair said,

    April 11, 2014 @ 5:42 pm

    From Sung Shin Kim:


    I also noticed that odd practice over recent years when watching Korean talk show programs. Your article mentioned that this phenomenon usually happens in commercial settings, and someone also commented underneath that it started in department stores. I think it started more specifically with so called Myŏmgp’um (명품, 名品). Originally, Myŏmgp’um was the term for a finely crafted article or master-piece, but it has come to be used for extremely expensive (Western brand) luxury items. Nowadays big department stores usually have a floor exclusively for Myŏmgp’um. Perhaps it started with Myŏmgp’um not just for the fortune they cost, but also because these were relatively unique, quasi individual objects, which in a way elevate the status of consumer. By now however it has trickled down, using -si-infix to more common products, say coffee, that are marketed as distinct and unique. It's really striking how it breaks the boundary in a sense between people and objects.


  24. Jongseong Park said,

    April 11, 2014 @ 6:38 pm

    @Stephan Siller:
    What I refer to as "compound words" in Korean would often look more like noun phrases in English and would be written as such, e.g. 국제^음성^기호 gukje eumseong giho "International Phonetic Alphabet" or 자기^공명^영상 jagi gongmyeong yeongsang "Magnetic Resonance Imaging". German does write the latter as a compound noun without spaces as Magnetresonanztomographie.

    I used ^ above for optional spaces in Korean, meaning that you are allowed to write both 국제 음성 기호 and 국제음성기호. (I think only the latter style without spaces is used in North Korea, though.)

    People may want to spell words as words, but what constitutes a word can be confusing, especially for a language like Korean. You know those examples in English like "every one" vs "everyone" or "any more" vs "anymore" that confuse people and divide opinion? There are incomparably more examples in Korean.

    Most piece of writing in Korean I see that has not been professionally edited contains spacing errors. For example, in the original blog post, there is an extraneous space in *사물 들에게 samul-deurege "to (the) objects". Here, the plural marker -들- -deul- is a particle, so there should be no space. It should be 사물들에게.

    However, in 풍경, 인물, 사물 들에게 punggyeong, inmul, samul deur-ege "to landscape, people, and object", the plural marker 들 deul functions as a dependent noun, so the spacing is obligatory. I know, this is why Koreans get confused all the time with spaces.

  25. Stephan Stiller said,

    April 11, 2014 @ 7:35 pm

    @ Jongseong Park

    Your example with -들 is interesting because it illustrates the orthographic difficulties of consistently dealing with elements that can attach to words as well as phrases. This is a bit similar to cases in English where you use hyphens throughout or en-dashes.

    About "everyone" etc, of course the degree of lexicalization and compositionality matters; see also my comment to another LL post and also the comment of "JS" which I'm referring to there. Maybe the agglutinativity of Korean matters, though this can all be addressed with sensible conventions.

    About 국제^음성^기호 and similar words: Independent of what they would be in English, the question is what they should be construed as in Korean. 국제 does not exhibit adjectival morphology here (ie: you don't have the -ㄴ or -은 ending characteristic of attributively used adjectives, and an analysis as a compound noun seems to me to be the correct one (the part of speech of 국제 here would be simply "attributive"). (English does have pure attributives, but they are rare.)

    (I would also note – for the other readers, because your usage here is correct, plus it doesn't apply to your two English examples – that something like "swimming pool" is really one linguistic word but two orthographic words in English orthography.)

  26. Victor Mair said,

    April 11, 2014 @ 7:58 pm

    From Toni Tan:

    Your Coffeeness

    From Infected Korean Language, Purity versus Hybridity: From the Sinographic Cosmopolis to Japanese Colonialism to Global English by Koh Jongsok and Ross King


    “… in a language like Korean with complicated and intricate honorifics and speech levels, complete identity between speech and writing is impossible from the very outset. And even if it were not, forcing colloquial speech onto written language would only render Korean literary style all the poorer, despite the best intentions of those advocating such a style.”

  27. Jongseong Park said,

    April 12, 2014 @ 5:39 am

    @Stephan Stiller:
    In the comment thread you linked to, you suggested that in all languages there were grey areas in how words were defined, with no obvious reasons for one language being richer in grey areas than the other.

    You're probably right, but spacing does cause a lot of problems in Korean spelling because it's not a simple matter of putting spaces between orthographic words. Due to its agglutinative nature, Korean uses a lot of particles, which Koreans would find unnatural to treat as independent words and to put spaces before them.

    In English, the vast majority of words that you see separated by spaces in a sentence are headwords that you can look up in a dictionary, or simple inflected forms with -s, -ed, etc. In Korean, you won't find 사물들에게 samul-deurege in a dictionary, but will have to look up 사물 samul, 들 deul, and 에게 ege separately. So while in English, the naïve simplification that spaces are put between words can still take you quite far, the logic of spacing in Korean orthography is not immediately apparent to Korean speakers. The existence of forms like 들 deul that can either attach to words as particles or suffixes or stand separately only adds to the confusion.

    Like many technical terms, 국제^음성^기호 gukje eumseong giho 國際音聲記號 "International Phonetic Alphabet" and 자기^공명^영상 jagi gongmyeong yeongsang 磁氣共鳴映像 "Magnetic Resonance Imaging" are Sino-Korean so they are again different from the native morphology. 국제 gukje "international" functions as an adjective in Sino-Korean compounds and even in mixed compounds like 국제^로터리 gukje roteori "Rotary International".

    If we treat all such cases as orthographic words, then we could end up with long, unwieldy forms like 국제연합교육과학문화기구 gukje yeonhap gyoyuk gwahak munhwa gigu 國際聯合敎育科學文化機構 "United Nations Educational, Scientific, (and) Cultural Organization". Allowing spacing between the constituent words as 국제^연합^교육^과학^문화^기구, one could write it in more digestible chunks as 국제연합 교육과학문화기구, for example. I personally like having the choice, although I concede that it causes confusion and creates the impression that all spaces are optional.

  28. Jongseong Park said,

    April 12, 2014 @ 5:52 am

    The excerpt from Koh & King quoted by Toni Tan in Victor Mair's comment above makes a good point. But I don't see the relevance here, unless this weird honorific style has already made its way into written Korean unbeknownst to me. At present, it remains the exclusively spoken jargon of service industry professionals. And precisely because of the point made by Koh & King, it's not going to move into written Korean anytime soon.

  29. Stephan Stiller said,

    April 12, 2014 @ 9:24 am

    @ Jongseong Park

    (Tautologically speaking, it's always a matter of putting spaces between orthographic words. I think you meant something like "unlike in German, it's not simply a matter of putting spacing between linguistic words", with some simplifications.)

    I agree that there are practical considerations according to which it might be sensible to put spaces in some or more positions within your particular examples. (And one can calibrate the rules to make them optimal. There are reasons why people intuitively do certain things with spaces, hyphens, etc in a language, and I mean "in general", not just in Korean or English.)

    I sense that we agree on a number of things. In your (second-to-last) post, there is one place where I'd like to add something: In Chinese, 國際 (C gwok3zai3, M guójì; "international") would be considered an "attributive". (The ABC dictionary gets its part of speech right. Oxford claims it's an adjective.) Many compounds involving it are simply long compound nouns. In general, Chinese often uses such compound nouns where English would use noun phrases (your earlier IPA and MRI examples are representative). On the sentence level, your compounds can certainly be treated as single, big words. You are correct that some parts "function [semantically] as adjectives", but I'd be hesitant to call 국제 (= 國際) an adjective. I guess you could call it something like "uninflected, attributive-only adjective", but the point is that they're a separate class (and them being Sino-Korean has something to do with it; there are also non-Sino-Korean words like that) and very restricted in where they can appear.

  30. Neil Dolinger said,

    April 12, 2014 @ 11:09 am

    @ Stephan,
    I don't disagree with you re the relative rarity of phonological changes in Mandarin induced by adjacent syllables. My point was really about the effect of those changes on the romanization. Pinyin rules do not seem to allow for the same flexibility as Korean romanization.

    Re "shenmo". I don't doubt that it is often romanized as "shenme". Unfortunately I can't find my DeFrancis textbook from the early 80s, but I could swear he used "shenmo" because that reflects the base pronunciation of the second character.

  31. Jongseong Park said,

    April 12, 2014 @ 2:38 pm

    @Stephan Siller:
    I did mean to say that it wasn't a matter of putting spaces between orthographic words in Korean, because my sense was that it was rather meaningless to talk about orthographic words in Korean. But I see now that it was badly expressed and tautological if taken literally. Anyway, the point which we agree on is that if you talk about orthographic words as any sequence of letters with white space at each end, then in Korean, "orthographic words" often do not correspond to linguistic words.

    국제 gukje 國際 "international" is listed as a noun in Korean dictionaries, glossed as "the quality of relating to several nations" or the like, with the qualification that in this meaning, it is only used before certain nouns. I was not familiar with the term "attributive", so thanks for introducing this to me. I was thinking of the Korean part of speech 관형사 gwanhyeongsa 冠形詞 when I said "adjective", which was incorrect—in fact, one of the English translations of 관형사 gwanhyeongsa seems to be "attributive".

    I'm a bit surprised that 국제 gukje is listed only as a noun and not an attributive in dictionaries, but I guess it has something to do with its rather restricted usage.

  32. Jongseong Park said,

    April 12, 2014 @ 2:59 pm

    Sorry for misspelling Stephan Stiller's name in that last comment.

  33. Michael Watts said,

    April 13, 2014 @ 12:18 am

    @Neil Dolinger:

    > Re "shenmo". I don't doubt that it is often romanized as "shenme". Unfortunately I can't find my DeFrancis textbook from the early 80s, but I could swear he used "shenmo" because that reflects the base pronunciation of the second character.

    To me, the character 么 has a few notable uses: it appears in 什么, 这么, 那么, 怎么, and it's used solo as a substitute for 吗. I've never heard anything to suggest that it should be pronounced mo instead of me, though… it's always me. Those uses are so extremely common that it's hard for me to think some hypothetical other use might reflect the "true pronunciation". Is the "true" pronunciation of 的 de or di?

  34. Michael Watts said,

    April 13, 2014 @ 12:32 am

    It's probably also worth pointing out that all the textbooks I've ever seen (more recent than the 80s, but still) use "shenme" (I agree that it's pronounced "sheme"), that putting "me" into a pinyin input system will bring up 么 as the first suggestion, and that putting "mo" in will get you substantially worse results. In google's pinyin IME for android, "mo" brings up 么 as the 36th suggestion, not even on the first page. In Windows 8's pinyin input, which I'm using to type this, 么 is not even reachable by typing "mo". Typing "shenmo" will get you 什麽 but will never get you a 么, even if you select 什 as a first step.

    The biggest case of textbook disagreement with reality that I've personally encountered is all the textbooks transcribing 那个 as "na ge", when anyone with ears could tell you it's pronounced "ne ge". Applying the same IME-based tests, Windows 8 correctly suggests 那 as #6 when I type in "nege", and google pinyin won't, but does suggest 哪个 (?).

  35. John Chew said,

    April 13, 2014 @ 8:16 am

    I am constantly having to remind my (non-native Japanese speaking) wife to put the "o" in front of words like "koohii", especially when in the presence of older relatives. I tell her it's because traditionally, as a woman she is of lower social status than a cup of coffee, a pot or pan, a toilet, etc. Then she usually hits me.

  36. Stephan Stiller said,

    April 13, 2014 @ 9:02 am

    @ Michael Watts

    I think you mean "nei ge". And "shemme" ("sheme" would be different; shé immediately elicits 蛇 (snake) and 舌 (tongue)).

    But in general it is correct that there are systematic cases of "textbook disagreement with reality", as you put it. I think that that's because the educators (incl textbook writers) don't generate their own pinyin but just cling to dictionaries and previous works. Because they themselves often don't speak standard Mandarin.

  37. Michael Watts said,

    April 13, 2014 @ 1:27 pm

    I do not mean "nei ge", I mean "ne ge". The phrase is evidently common enough that 那 takes on the vowel of the 个 that follows it. If it's relevant to you, nearly all my exposure to chinese speakers is in Shanghai. Similarly, I don't think the m sound in 什么 is geminated? But I have only weak feelings about that one.

  38. Victor Mair said,

    April 14, 2014 @ 6:00 am

    There was a stage fairly early on in my Chinese learning career where I often encountered 那个 transcribed as nèige instead of as nàgè (Google Translate allowed me to call up 那个 by typing in "neige", but then yielded nàgè for the romanization!). Mutatis mutandis for 哪个 and 这个. (At the time when I experienced this, the romanizations would have been in Wade-Giles, not yet Hanyu Pinyin.)

    If we're talking about discrepancies between canonical (or standard) transcription and actual pronunciation, one of the students from China in my Literary Sinitic class pronounced 已 as yǐye — two syllables!! — with the second syllable being pronounced in a high, neutral tone. She said it very clearly. I was stunned, and asked her to repeat it several times to make sure I was hearing it right, and each time she pronounced exactly the same way. Moreover, no matter how hard I tried, I couldn't get her to tell the difference between yǐ and yǐye. What is even more astounding is that all the other Mainland students in the class:

    a. pronounced 已 as yǐ

    b. said that the first girl was "right / correct" to pronounce 已 as yǐye even though they pronounced it as yǐ

    When coaxed and coached, I finally got the other Mainland students to tell the difference between yǐ and yǐye, but it was impossible for the first girl to:

    a. hear the difference between yǐ and yǐye

    b. pronounce 已 as yǐ instead of as 已 yǐye

    For the record, I should note that the first girl was born in the south but spent most of her time growing up and maturing in Beijing, one of the other girls was actually born and grew up in Beijing but went to college in Wuhan, and the others were all from somewhere around the Yangtze or places to the north of it.

    I could tell you similar stories about zhèr 这儿, nàr 那儿, and nǎr 哪儿 as zhèhér (two syllables) or zhèhé'er (three syllables) — some people even write it as 这合儿! — etc. for the other two.

    There are also many other striking discrepancies between the standard transcriptions of terms and what people actually say, although it is amazing how strongly standardized transcription impacts speech for the bulk of Chinese students who grow up learning how to read, write, and speak Mandarin through Hanyu Pinyin.

  39. KWillets said,

    April 14, 2014 @ 1:55 pm

    I learned nèige back in the 80's, but I didn't keep up my Chinese enough to notice it changed. Mais où sont les nèige's d'antan?

    Also, claims that 으시 affix use for inanimate subjects hasn't made it into the written language don't seem correct, as I've found it on some shopping websites.

  40. Michael Watts said,

    April 15, 2014 @ 12:23 am

    Some further thoughts:

    Interestingly, exactly one of the dictionaries I use includes an entry for 那 "ne", glossing it as "a variant pronunciation for 那 'na' in colloquial speech". (It's the Pleco self-branded dictionary; ABC and CC-CEDICT disappointingly have no corresponding entry.) But as far as I'm aware (and I could be wrong about this), what's going on isn't that people freely vary between 'na' and 'ne' in informal settings — I would say that 那 'na' is obligatorily pronounced 'ne' when followed by 个.

    It reminds me quite a bit of when I observed a Chinese high school student's English-Chinese dictionary's entry for some word ending in -ate, which I will here call "legitimate". The entry noted that the last syllable might be pronounced with the FACE vowel or with a reduced vowel, and also noted that the word might be an adjective or a verb. What seemed like crucial information to me, but which was totally omitted by the dictionary, was the fact that those aren't free variations — the adjective is necessarily pronounced with the reduced vowel, and the verb is necessarily not pronounced that way. You'd never catch an English-English dictionary confusing those… having checked just now, Merriam-Webster and the cambridge dictionary use separate entries for the different words (as they should!).

  41. Michael Watts said,

    April 15, 2014 @ 12:33 am

    @Victor Mair, since I believe he's influential with the ABC dictionary

    This might be a crazy dream, but I dearly wish dictionary entries would indicate which prepositions a verb can be used with. For example, the ABC entry for 保护 glosses it as "protect; safeguard". I can intuit pretty easily that the transitive object is the protectee, but I'd really like to know how to mark the threat (the object of 'from' in "I will protect you from [whatever]"). I don't see this as being a huge stretch from the existing practice of indicating which measure words a noun can take.

  42. Victor Mair said,

    April 15, 2014 @ 5:20 am

    From San Duanmu:

    Interesting observation. If [nei] is the abbreviation of [na yi] 'that
    one', then it is reasonable for the change. Similarly, [zhei] is the
    abbreviation for [zhe yi]. Thus, one can say [na ge] (omission of
    [yi]), [na yi ge], [nei ge] (merger of [na] and [yi]), but not [nei yi

  43. Jongseong Park said,

    April 15, 2014 @ 10:28 am

    Also, claims that 으시 affix use for inanimate subjects hasn't made it into the written language don't seem correct, as I've found it on some shopping websites.

    I would love to see examples as this is something I've never noticed and I have trouble imagining such usage. People don't really write the way they speak. Granted, I don't look at Korean shopping sites that often, so I went into one at random and clicked at random items (a "health ring", brand name bags), but didn't find any inappropriate honorifics in this tiny sample—honorifics were correctly used for the customers and not for the items themselves.

    By the way, according to this article in Korean, the first complaint about the inappropriate use of honorifics appeared in 2004 in the Voice of Customer feedback for Hyundai department store. It is said that there were quite a few call centre operators who already used "object honorifics" in the late 1990s. But it seems to have become noticeable in the middle of the first decade of the 21st century.

  44. KWillets said,

    April 16, 2014 @ 1:45 am

    @Jeongseong I found that if you search for 가능 하십니다 you get a moderate number of product websites and so on.

    My first thought was to look at some home shopping channel websites (I have some experience with these due to family connections), but I didn't see any major examples. The fancier imported goods appear to have paid for professional translations with more conventional grammar.

  45. Jongseong Park said,

    April 16, 2014 @ 3:36 pm

    I don't see a problem with the first few results I see for 가능하십니다 ganeung-hasimnida "is possible", though:
    대출 가능하십니다 daechul ganeung-hasimnida "loan is possible"
    착용 가능하십니다 chagyong ganeung-hasimnida "wearing is possible"
    입점 가능하십니다 ipjeom ganeung-hasimnida "entering a shop is possible"

    These all illustrate cases where honorifics are technically applied to inanimate objects (loan, wearing, entering a shop), but semantically they refer to actions by the person being elevated. Saying that something that you can do is possible is just another way of saying that you can do something, so people feel that honorifics are not out of place when "you" would ordinarily get the honorific.

    There is no problem with applying honorifics to parts, qualities, actions, or even objects intimately related to the person being elevated, as in 주소가 어떻게 되세요? juso-ga eotteoke doeseyo "How [what] is your address?" where "address" gets the honorific. The same applies for name, age, etc.

    The reason most Korean speakers find "object honorifics" of the kind we have been discussing wrong is that the objects in question, like coffee, are not felt to be intimately connected to the person being elevated, unlike address, name, or age.

RSS feed for comments on this post

Leave a Comment