Kimchee is Korean

« previous post | next post »

Not Chinese.  Do you understand?

This has long been a cabbage of contention, but make no mistake about it:  fermented kimchee / kimchi  (gimchi 김치 (IPA [kim.tɕʰi]) (lit., "soaked [in their own juices of fermentation] vegetables") is not the same thing as pickled paocai / pao tsai 泡菜 (lit., "soaked [in brine] vegetables").

Kimchee and paocai are made differently, have different ingredients and spices, and taste different.  To call "kimchee" "paocai" would be like calling "wine" (pútáojiǔ 葡萄酒) "beer" (píjiǔ 啤酒).

Linguistically, kimchee has its own pedigree, of which I will here give an extended account.

Borrowed from Korean 김치 (gimchi), ultimately composed within Korea of Chinese-derived morphemes (chén, submerged, soaked) and (cài, vegetable), i.e. "fermented vegetable". Doublet of kimuchi.


Nativisation of the Sino-Korean term 침채 (沈菜, chimchae, “soaked vegetables”). First attested in the Hunmong jahoe (訓蒙字會 / 훈몽자회), 1527, as Middle Korean 딤ᄎᆡ (Yale: timchoy).

In most dialects, Middle Korean 딤ᄎᆡ (Yale: timchoy) was regularly palatalized to 짐츼 (jimchui), whence modern southern dialectal 짐치 (jimchi). However, southern Korean dialects also palatalized /ki/ to /t͡ɕi/, which the Seoul prestige dialect did not. Seoul speakers hence hypercorrected 짐츼 (jimchui) to 김츼 (gimchui) to avoid sounding "southern", whence modern Standard 김치 (gimchi).

This word displaced the older Middle Korean term 디히〮 (Yale: tìhí) (whence (ji), now only used in compound words and the southwestern Jeolla dialect), probably because regular sound change rendered it monosyllabic and caused it to have a large number of homophones.



The term ji (), which has its origins in archaic Korean dihi (디히), has been used to refer to kimchi since ancient times.[5] The sound change can be roughly described as:

    • dihi (디히) > di () > ji ()

The Middle Korean form dihi is found in several books from Joseon (1392–1897). In Modern Korean, the word remains as the suffix -ji in the standard language (as in jjanji, seokbak-ji), and as the suffix -ji as well as the noun ji in Gyeongsang and Jeolla dialects. The unpalatalized form di is preserved in P'yŏngan dialect.


Kimchi (김치) is the accepted word in both North and South Korean standard languages. Earlier forms of the word include timchɑi (팀ᄎᆡ), a Middle Korean transcription of the Sino-Korean word (literally "submerged vegetable"). Timchɑi appears in Sohak Eonhae, the 16th century Korean rendition of the Chinese book, Xiaoxue. Sound changes from Middle Korean to Modern Korean regarding the word can be described as:

    • timchɑi (팀ᄎᆡ; 沈菜) > dimchɑi (딤ᄎᆡ) > jimchɑi (짐ᄎᆡ) > jimchui (짐츼) > gimchi (김치)

The aspirated first consonant of timchae became unaspirated in dimchɑe, then underwent palatalization in jimchɑe. The word then became jimchui with the loss of the vowel ɑ () in Korean language, then Kimchi, with the depalatalized word-initial consonant. In Modern Korean, the hanja characters 沈菜 are pronounced chimchae (침채), and are not used to refer to kimchi, or anything else. The word Kimchi is not considered as a Sino-Korean word.  Older forms of the word are retained in many regional dialects: jimchae (Jeolla, Hamgyŏng dialects), jimchi (Chungcheong, Gangwon, Gyeonggi, Gyeongsang, Hamgyŏng, Jeolla dialects), and dimchi (P'yŏngan dialect).

The English word "kimchi" perhaps originated from kimch'i, the McCune–Reischauer transcription of the Korean word Kimchi (김치).


But, wait!  This is not all uncontested.  First of  all, we've previously covered much of this ground in earlier Language Log posts.  I especially recommend Bob Ramsey's guest post, "Who owns kimchi?" (5/5/21) and the valuable comments thereto, many by the author himself.  See also "Kimchee " (1/2/14), written seven years earlier, with 69 informative comments appended thereto, of which i here reproduce several of the more significant ones.

Bob Ramsey:

You ask in this most interesting link if your etymology of the word kimchi is correct: "it would seem that the word kimchee is derived from the pre-modern term chimchae 沉菜 (lit., 'soaked vegetables'), so there is a ready-made, etymologically exact Sinographic written form available for use." And the answer is yes, that's right—with one little caveat: As far as I know, that Sinitic form was never used in China or anywhere else outside Korea. (Maybe you know better, and if so, please let me know!) And so, if I'm right (which I'm pretty sure I am), it means that Koreans simply made up a Chinese-style word, and that was probably because Chinese-character words were seen even then as more elegant and desirable than native vocabulary. (That's certainly the sociolinguistic import of Sinitic vocabulary today in both Korea and Japan, as you and I both well know. It's kind of like English and Americans using French words for food terms, isn't it?)

And one more caution. That Wikipedia article you link to says that: "The term ji was used until the pre-modern terms chimchae (hanja: 沉菜, lit. soaked vegetables), dimchae, and timchae were adopted in the period of the Three Kingdoms of Korea." That statement (which is taken from a pretty unreliable Korean source) is unprofessional and problematic on so many levels. First of all, it's sort of on the right track in identifying ji as the earlier Korean word for pickles, but really not. What should have been said is that the earliest attested native word for pickles (first seen in texts from the 16th century) is tihi, which is where ji, the form used today, comes from (through regular phonological changes). And then: "the pre-modern terms chimchae (hanja: 沉菜, lit. soaked vegetables), dimchae, and timchae were adopted in the period of the Three Kingdoms of Korea." Huh? "terms"? These three variant readings were used in the Three Kingdoms period? That totally doesn't make any sense. And then there's the idea that the fake Sinitic name was used in the Three Kingdoms period. That's a highly suspect assertion. I haven't seen any reference to 沉菜 in materials referring to foodstuffs in the Three Kingdoms period. There are almost no Korean writings of any kind preserved from the Three Kingdoms period itself, which ended around 668. The oldest Korean history that has been preserved was compiled well after that, in 1145, and though that text might have some reference to Korean pickles, I haven't yet seen it. It may be there, but I'd like to know where. We need the citation. To tell you the truth, I kind of suspect that like so many other "facts" about ancient Korea–Korean nationalists are fond of ascribing virtually everything about native Korean culture to the Three Kingdoms period (or earlier—something Dangun came up with maybe?)–this "fact" also represents imaginary history. The source Wikipedia cites is just too sketchy to take seriously.

Jongseong Park, quoted by JS from "Dynamic stew" (10/24/13):

Kuiwon: Do you know of any other non-Sino-Korean Korean words (i.e., "pure" Korean) that originate from older pronunciations of Chinese characters? I know "Kimchi" is from 沈菜.

There are plenty of such cases, though "kimchi" is a somewhat murky example. It is first attested as 딤ᄎᆡ timchoy and then a bit later as 팀ᄎᆡ thimchoy in 16th century texts (using the Yale romanization here since it is Middle Korean, with o representing the archaic vowel ㆍ). The theory is that these represented the then-current Sino-Korean pronunciation of 沈菜, which must be a local coinage since it was not used in this sense outside of Korea.

The trouble is, 沈 is known to have been read as 팀 thim, but the reading 딤 tim is not attested for this character.

In any case, it appears that both 딤ᄎᆡ timchoy and 팀ᄎᆡ thimchoy were in use in Middle Korean. Then, identified with the Sino-Korean word 沈菜, 팀ᄎᆡ thimchoy followed the regular development into Modern Korean 침채 chimchay [ʨʰimʨʰɛ], which is apparently an obscure term used in ancestor worship rituals (so obscure that it doesn't appear in most dictionaries). But the form 딤ᄎᆡ timchoy became 짐ᄎᆡ cimchoy, then 짐츼 cimchuy, then Modern 김치 kimchi.

How to explain this? If the form 딤ᄎᆡ timchoy ever corresponded to the Sino-Korean word 沈菜 (it has been suggested that 딤 tim was an older reading of 沈, otherwise unattested), it may have been decoupled from the regular development of Sino-Korean pronunciation because it was no longer felt to be Sino-Korean, and was free to follow its own phonetic development. An alternative view is that 딤ᄎᆡ timchoy was the original Korean word, and the form 팀ᄎᆡ thimchoy came about because it was reinterpreted as the Sino-Korean 沈菜. What seems clear is that the original term for 김치 kimchi was associated with the Sino-Korean 沈菜 from early on, but it doesn't seem conclusive to me that 沈菜 is indeed the origin of the word.

Bob Ramsey:

Great posting just now from JS. Jongseong Park is absolutely justified in being suspicious of the suggestion that 沈菜 is indeed the origin of the word. In a way, such an idea turns things upside down, because kimchi is not from Chinese. In my posting (actually posted by Victor), I meant only that the word was probably created, in Korea, with the intention that it should like a Chinese word. As Jongseong Park points out, the character 沈 never had, as far as we know, the unaspirated reading 딤. (Complicating the matter slightly, though, is that in Sejong's day prescriptive readings like 띰, 땀, 담, etc. were made up for it.) Still, even though 딤 is undoubtedly a ghost reading, the point remains that the word for kimchi was made up with the intention of giving it a Chinese-style cachet.

There are many other excellent comments to the "Kimchee" thread, including one by julie lee that hints at what I think may be the most significant contribution of this current post, which is the distinction between fermented versus pickled alluded to in the second paragraph above.

Meanwhile, here's the Korean newspaper article that led to this post:

"Seoul pushes xinqi as Chinese translation of kimchi", Korea JoongAng Daily (9.27/23)

Xinqi will be pushed as the preferred Chinese translation for kimchi at restaurants in Seoul, instead of pao cai.
The Seoul Metropolitan Government said Wednesday it will inspect stores and restaurants in major tourist spots such as Myeong-dong in Jung District, central Seoul, to see if they have used the correct official Chinese term for kimchi.

The city government had registered the Chinese translation of kimchi as xinqi in its official foreign language dictionary in September 2021, which was followed by an announcement by the Ministry of Culture, Sports and Tourism.

The Culture Ministry in July 2021 amended its official guidelines for the foreign language translation and notation of the official terms, changing the Chinese term for kimchi to xinqi.
The term, with characters meaning spicy and unique, was chosen through an expert review partially due to its similar pronunciation to kimchi.
The decision was made as kimchi became the subject of a heated cultural feud between Korea and China in 2020 when the nationalistic Chinese tabloid Global Times published an article claiming that Chinese pao cai, pickled vegetables from Sichuan Province, had become the international standard for the kimchi industry.
The article sparked an uproar in Korean society, with some referring to the article as a Chinese attempt to steal Korean culture….

I love kimchee (spicy hot and squishy, plus garlicky) and I love paocai (sour and scrunchy, plus fragrant), but they sure aren't the same thing.  I think they deserve their own names.  Hats off to English for having separate, distinctive names for them!  We don't just call them generic "pickled vegetable" and "fermented vegetable".


Selected readings

[Thanks to Mark Swofford]


  1. Alec Story said,

    September 28, 2023 @ 11:53 am

    If it's interesting, there are some lacto-pickled vegetables in Qimin Yaoshu, the 544 CE farm manual that's our earliest major record of agricultural food production.

    Here's one recipe with my translation below, although I have not put specific effort into cross-referencing this with the various analyses of this work. This might help spot early words for this kind of food. Depending on your browser, be sure to look for the characters that don't render in most modern fonts.

    (full Chinese text at

    When you take vegetables, take the good ones, and bind them with kangaroo grass [Themeda gigantea] or sweet-flag [Acorus calamus, possibly bulrushes Typha orientalis]
    Make salt water, making it very salty, and wash the vegetables in the salt water, and then put it into a weng [an urn].
    If you wash with fresh water, the pickles will rot.
    As for the washing of the vegetables in salt water, when it clarifies, pour it into the weng, just to cover the vegetables, do not re-mix them.
    While the pickles are still green, use water to wash off some of the salt, and boil it to eat, and eating them with fresh vegetables is not peculiar.
    For the field mustard and the Shu mustard, ladle them out after three days.
    Make porridge-clear [the clear, starchy liquid that accumulates on top of porridge] out of proso millet powder. Pound wheat/barley corns to dust, and sieve it through a thick tabby-woven silk cloth.
    Spread the vegetables as a group, and dust them with the flour, and then add them to the cooked porridge-clear.
    Layer them repeatedly like this, and fill the whole weng.
    其布菜法:To spread out the vegetables:
    Every layer must alternate stems and leaves.
    Still add the old salt water to the weng.
    When the pickles are yellow, the flavor will be good.
    To make mild pickles, use Proso millet porridge-clear, and wheat/barley flour, and the flavor will also be excellent.

  2. ohwilleke said,

    September 28, 2023 @ 1:04 pm

    Fun fact: The spices now used in kimchee are a New World domesticate that only arrived in Korea in the 1500s.

    As pertinent to the linguistic question, just as "Kimchee and paocai are made differently, have different ingredients and spices, and taste different." Pre-1500s references, if any, to Kimchee, would be referring to a kind of food only remotely similar to the post-1500s Kimchee that continues to be the Korean national dish.

    Another fun fact: Mostly due to kimchee consumption, Koreans consumer far more garlic per capita than any other country in the world by almost an order of magnitude.

  3. sehhe said,

    September 28, 2023 @ 11:37 pm

    >Seoul pushes xinqi as Chinese translation of kimchi
    辛奇 is indeed pronounced xīnqí in Mandarin, and it does sound somewhat similar to Korean 김치 (gimchi), but its readings in other languages do not resemble gimchi at all. Compare Korean 신기 (singi), Cantonese sānkèih, or Hokkien sinkî. I wonder why they did not use the alleged etymon, 沉菜 (K chimchae, M chéncài, C chàhmchoi, H tîm-chhài).

    I can see that the "Mandarin-centrism" in matters concerning the 'Chinese' has been growing stronger and stronger in South Korea, more than in the North or even the Yeonbyeon region in PRC, home to many ethnic Koreans. For example, 'Xi Jinping' is usually called 시진핑 (Si Jin-ping) in South Korea, while in the North his name is read in Sino-Korean, as 습근평 (Seup Geun-pyeong). Or for 'Beijing', the traditional Sino-Korean reading 북경 (Buk-gyeong) is being phased out by the Mandarin transliteration 베이징 (Be-i-jing) in the South, but not in the North. Now it seems that even Korean cultural terms cannot escape this Mandarin-centric approach.

  4. David Morris said,

    September 29, 2023 @ 2:57 am

    Where does the 'ee' spelling come from? Yale, MR and RR all use 'i' for hangeul 'ㅣ'. I'm trying to think of another similar word. I have pondered the alternative between 'i' and 'ee' and 'u' and 'oo' in transliterations of various languages with no firm conclusions.

  5. Scott P. said,

    September 29, 2023 @ 7:37 am


    I expect that kimchee/kimchi as a term in English dates to the Korean War, and in that era 'ee' was used to spell that phoneme. as in Syngman Rhee.

  6. Benjamin Ernest Orsatti said,

    September 29, 2023 @ 8:26 am

    An open letter to the Korean people:

    Many of us are thinking it, but I'm going to put it in writing: Your romanization system makes me sad. Please stop putting "e"s where they don't belong. In English, we pronounce "e"s unless they're at the end of a word (even a schwa occasions a syllabic pause). When I see "Seoul", I hear "Say-oh-ul". Fercryinoutloud, the system's very name even makes Greeks and Hawaiians scratch their heads with its employment of 3-4 consecutive vowels.

    Not to worry, though; there is a solution. Yes, I get that you don't want your capital city spelled "Soul", but what about those apostrophes you're not using? If you don't like diacritics, let ô–> o' and ʉ–> u' (or "w" in diphthong). And, while you're at it, go ahead and throw them in after your aspirated consonants too! That way, when you want to say, "국어의 로마자 표기법", you don't have to say, "Gugeoui Romaja Pyogibeop", you can say instead, "Kugo'wi Romaja P'yo'gibo'p."

    Please understand the importance of this grave matter. You can't let this thing get away from you too much — look at what happened to French!


    A Guy Who Took Taekwondo and One Year of College Korean

  7. David Marjanović said,

    September 30, 2023 @ 1:18 pm

    "Kugo'wi Romaja P'yo'gibo'p."

    I'm reminded of the Kazakh apostrophe catastrophe (mercifully not implemented so far) and… if the English-speaking world is your concern, well, the custom there is to ignore all apostrophes (and diacritics) as mere ornaments.

    Maybe a German perspective helps: interpreting eo & eu as "the opposite" of oe & ue (ö & ü) makes some sense. They're back (or central) unrounded vowels instead of front rounded ones.

  8. Aotearoa said,

    October 1, 2023 @ 5:56 am

    @Benjamin Ernest Orsatti.

    Multiple vowels are not a problem in Te Reo Maori. Try “eaoia”. Every letter is pronounced. We don’t have apostrophes as our sister language Hawaiian does.

  9. Benjamin Ernest Orsatti said,

    October 2, 2023 @ 11:53 am


    Well, that certainly beats the Greek invocation of Dionysus (εὐοῖ), which, before it was pronounced, "eˈvi", had been pronounced "eu̯.ôi̯" in the 5th century B.C. But the point is, as you say, "[e]very letter is pronounced" (Irish, are you listening?). I get that spoken language "drifts" from the written variety over time, but if you're creating a writing system ex nihilo, why hamstring yourself by _starting off_ with silent letters?

    — B.

  10. Jongseong Park said,

    October 3, 2023 @ 2:00 am

    I would personally rather use apostrophes to mark boundaries that are ambiguous potential digraphs, e.g. in 가을 ga'eul vs 개울 gae'ul. Officially, hyphens can optionally be used in this situation, e.g. 가을 ga-eul vs 개울 gae-ul (both being gaeul by default), but I would reserve hyphens for actual morpheme boundaries, especially to separate particles. Here is how I write 국어의 로마자 표기법: Gugeo-ui Romaja Pyogibeop

    Here, the particle 의 ui is separated by a hyphen. Meanwhile, the Library of Congress version of McCune–Reischauer just treats particles as separate words:
    Kugŏ ŭi Romacha P’yogipŏp

    If we write 의 ui /ɰi/ [ɰi~i, e (as a particle)] as wi, then we will no longer be able to distinguish it from 위 wi /wi/ [y~ɥi]. That just to avoid consecutive vowel letters?

    The suggestion to use o' and u' instead of eo/ŏ /ʌ/ and eu/ŭ /ɯ/ reminds me of the use of in the Uzbek Latin alphabet to represent /o/. Perhaps o' and u' could be seen to evoke Vietnamese-style ơ and ư, but I don't think using the apostrophe as a modifier letter is advisable in general.

    The biggest drawback to the use of eo and eu is that speakers of English or Romance languages might be tempted to use the soft g for 거 geo or 그 geu. But the use of digraphs is really widespread in Latin-based orthographies and romanizations and it is hard to think of a reason to single out eo and eu as being egregiously offensive. As David Marjanović says, we can just think of these as representing unrounded versions of o and u respectively.

    I am curious about the choices in Kugo'wi Romaja P'yo'gibo'p. Initial ㄱ is written as k to reflect its allophonic voicelessness, which means that a separate aspiration marker is needed for ㅍ p'. But why write the initial consonants of 자 and 법 as j and b as if they were voiced, when they are pronounced as in the tense series as if written ㅉ jj and ㅃ pp and therefore voiceless? Are we ignoring the tensification because it is not indicated in the orthography? I think the official romanization got it right by mapping all initial lenis stops and affricates to voiced letters and ignoring tensification, making it easier to romanize based on the orthography.

  11. Benjamin Ernest Orsatti said,

    October 3, 2023 @ 8:28 am

    [In which a native speaker responds to the musings of one who took one year of Korean in college 23 years ago]

    Jongseong Park said,

    I am curious about the choices in Kugo'wi Romaja P'yo'gibo'p. Initial ㄱ is written as k to reflect its allophonic voicelessness, which means that a separate aspiration marker is needed for ㅍ p'.

    — So far, so good, no?

    But why write the initial consonants of 자 and 법 as j and b as if they were voiced, when they are pronounced as in the tense series as if written ㅉ jj and ㅃ pp and therefore voiceless?

    — Are they? To an American ear, medial "자" _sounds like_ "ja"; it's difficult to "hear" "ㅉ" or "ㅃ" because American English doesn't make that phonemic distinction. And, if romanization were intended to be used by _native_ speakers of Korean, I'd say, "sure, write it how the native speaker would write it", but it's not. Romanization is a plea for non-native speakers to approximate the sounds of the target language as closely as possible. So, if you represented medial 자 by "jja", what you'd get from an American English speaker would simply be a _lengthening_, not tensing, of the consonant.

    Are we ignoring the tensification because it is not indicated in the orthography?

    — Sort of. But mostly because we don't "hear" it, see above.

    I think the official romanization got it right by mapping all initial lenis stops and affricates to voiced letters and ignoring tensification, making it easier to romanize based on the orthography.

    — But initial consonants _aren't_ voiced; it's only the tensed initial consonants that we "hear" as voiced.

    — 벤자민/베냐민/뻔자민

  12. Jongseong Park said,

    October 3, 2023 @ 10:29 am

    In my haste, I made a mistake. 국어의 로마자 표기법 should actually be Kugŏ ŭi Romaja P’yogipŏp in McCune–Reischauer. It should be romaja, not romacha. I made that mistake because in my idiolect, I tensify the 자 in 로마자 so that I pronounce it as if were 로마짜, but this is not standard.

    한자 meaning Chinese characters is written hancha in McCune–Reischauer because in this case the tensification is standard and it is pronounced as if it were 한짜. Meanwhile, 한자 as in the Korean rendering of German Hansa is pronounced as spelled, with a voiced medial ㅈ, so it is written hanja in McCune–Reischauer.

    And 표기법 also has tensification so that it is pronounced as if written 표기뻡.

    The point is that medial ㅈ and ㅂ in 한자 and 표기법 are not voiced. They sound not too much unlike ch and p in the English words inches and upper respectively. The tensification leads to these being pronounced as voiceless, not voiced as one might naively assume from the spelling, which is a phonemic distinction that English speakers consider important. English speakers definitely hear it. That is why McCune–Reischauer goes through the trouble of representing it, rendering 한자 "Chinese characters" as hancha and 한자 "Hansa" as hanja respectively.

    I should have specified that I was talking about word-medial but syllable-initial consonants (because syllable coda consonants are neutralized). Word-initial lenis stops and affricates are allophonically devoiced, of course.

    By the way, I followed South Korea's official romanization in writing ㅉ as jj and ㅃ as pp. This romanization tends to be geared towards English speakers, but it is not entirely based on the sound values used by them either. American English speakers are far from the only ones that would need to use romanized Korean.
    What do you think about Hanyu Pinyin, I wonder?

  13. Jongseong Park said,

    October 3, 2023 @ 10:31 am

    Wait, you hear tense initial consonants as voiced? They are definitely voiceless, and are simply unaspirated, like Spanish, Italian, French, or Russian voiceless stops. Do you hear those as voiced as well?

  14. Jonathan Smith said,

    October 3, 2023 @ 12:49 pm

    re Benjamin Ernest Orsatti;'s remarks, Romanization of Korean etc. is not in particular a tool to allow a non-native speaker to read off text such that it sounds (to the non-native speaker) like Korean sounds (to the non-native speaker). We really need systems that represent actual Korean etc., whether tending more to transliteration or more to phonemicity or compomising in some way(s), and for non-native learners to make use of the facts expressed via these representations (among other facts) to learn better.

    While knowing little to nothing about Korean or Romanization thereof, I suspect from a distance that the oft-debated details don't matter much and the real problem is just lack of an agreed-upon convention. In parallel fashion, arguments about Tai-lo/Peh-oe-ji for Taiwanese are IMO a complete waste of time — or worse, as one result of the impasse is that everyone who wants to approach Taiwanese texts has to learn both (+ more). By contrast, the exactly parallel arguments about the relative merits of Pinyin/Yale etc. are now fortunately (in a limited sense) entirely academic…

  15. M said,

    October 24, 2023 @ 5:41 am

    Back on topic, I'm personally fond of 辛菜 as a compromise. It acknowledges the root of the word as being derived from vegetables, it's descriptive("spicy vegetable" in both Chinese and Japanese), and it even sort of sounds like kimchee, depending on the Chinese subgroup/"dialect". If marketing it to Japanese users as well, all you need are some ruby/furigana(キム•チotherwise, it would be something like "kara-sai"), and it's easy to identify on a shelf, even without a picture of the contents. From the merchants' pov, it saves cost, by allowing them to use a single label design for all 3 markets(and a sticker in hastily Google translated English/French/etc. for international markets) When trying to get others to change their own language, it helps to meet them in the middle/compromise.

RSS feed for comments on this post