« previous post | next post »

Michael Watts just wrote this comment on another post, and I thought it was interesting enough to deserve a post of its own:

I've been wondering about a claim that appears on wiktionary. The entry for the Japanese word "midori", spelled 緑 or in older form 綠, states that the word is from Old Japanese, originally referred to buds and shoots, and experienced semantic shift into its modern meaning of the color green.

What bothers me is that the character 綠 is already defined in the shuowen jiezi, which is significantly older than Old Japanese, as referring to a color and not to a plant. So for the Japanese word to be spelled 綠, it seems to me that it must already have lacked reference to plants by the time it was being written down at all.

So… how do we know that it originally referred to buds and shoots? What kind of evidence might we have for that? If it's true, why wasn't the word spelled 芽?

If you want to see what the beautiful color midori looks like, go to the Wikipedia article on "Midori" liqueur.  Incidentally, you will find there that the liqueur is flavored with Yubari (a type of cantaloupe) and muskmelon.

See here for disambiguation of the many uses of "midori".


Selected readings

"Ask Language Log: Are East Asian first names gendered?" (1/14/18) — in the comments

"Grue and bleen: the blue-green distinction and its implications" (10/4/19)



  1. anon said,

    October 27, 2023 @ 10:52 am

    The Daijisen and Kōjien dictionaries also have this etymology. Kōjien connects it to "midumidusi" ("mizumizushii" in Modern Japanese, meaning "young, fresh, lively"). Since 緑 is used to represent "midori" already in the Man'yōshū, if there was a semantic shift it occurred before the introduction of writing to Japan.

  2. Jonathan Smith said,

    October 27, 2023 @ 1:29 pm

    One kind of contemporary evidence re: etymology is polysemy / "range of usage" — so Jdic has items such as midori no kurokami みどりの黒髪 'glossy black hair ([of] young woman)' and midorigo みどり児 'infant' that suggest freshness/newness as maybe a more fundamental meaning, FWIW much more similar to modern and historical Ch. qing1 青 than to lv4 綠.

    As usual in Japanese midori can be written in several ways it seems… including 綠 翠 碧 etc.

    Incidentally lv4 綠 in early Chinese texts is often a kind of silk, maybe or maybe not of some particular color… thus the Shuowen definition ('silk of qing-huang color' 帛青黃色也), qing1 here being the older color term. lv4 'the color green' and midori 'the color green' thus both look like traceable secondary developments in Ch. and Jp. respectively

  3. J.W. Brewer said,

    October 27, 2023 @ 1:40 pm

    The English adjective "green" has extended metaphorical non-color senses that can mean "new" in either a "fresh and vigorous" sense or in a "inexperienced and naive" sense. But you wouldn't think those were much evidence that that's a "more fundamental" meaning than the color. Of course the English color word supposedly descends from PIE *gʰreh₁, which is glossed as "to grow," maybe especially in the contexts of plants. Which gets you back to the "buds and shoots" and their characteristic color.

  4. CuConnacht said,

    October 27, 2023 @ 2:18 pm

    The idea that even a simpleton could think that the moon was made of green cheese puzzled me in my childhood, since it obviously was not green at all. It turns out that "green cheese" = fresh cheese, curds newly placed in a hoop for pressing and aging, to which the full moon does indeed bear a resemblance.

  5. Lucas Christopoulos said,

    October 27, 2023 @ 7:05 pm

    as with the Chinese, here the Japanese often confuse blue with green, though for "green Tea" they use Ryoku-cha (緑茶) and for blue sometimes aoi (青) as in あおりんご 青林檎.

  6. Jonathan Smith said,

    October 27, 2023 @ 7:41 pm

    True @J.W. Brewer but IMO 'glossy (of hair)' in particular seems unlikely to be such an extension. But of course actual etymology would be better…

    @Lucas Chirstopoulos aoi and qing1 must be part of earlier more "basic" color schemes as opposed to involving confusion per se.
    (Cognate of the latter, Tw. tshenn1, survives as a color term… where hard to say if the same item as tshenn1 'raw')
    Also surely not 'blue apple'?

  7. Chris Button said,

    October 27, 2023 @ 10:06 pm

    @Lucas Christopoulos

    Not confusion but a remnant of past usage. There may still be “ao ringo” (blue apples) today, but wasn’t there also a “midori no sora” (green sky) in the past too?

  8. Lucas Christopoulos said,

    October 28, 2023 @ 6:20 am

    @Chris and Jonathan

    And not to forget the Apple-pen…

  9. Chris Button said,

    October 28, 2023 @ 7:21 am

    how do we know that it originally referred to buds and shoots?

    The expected verdant origin of 綠 seems to come from its ancient association with 菉 "kind of carpet grass" (as noted by others, the 糹 in 綠 references dying as in 紫 "purple", etc.). It's a shame that it's not clear what the oracle-bone form for the phonetic 录 in 綠 represents, although it does look like something is being squeezed out. There's a somewhat decent semantic overlap with Indo-European *dhal- "sprout, bloom" from whence Armenian dalar "green" (cf. 綠), Albanian dal "leave, exit" (cf. 逯 "wander off"), Irish duille "wealth, glory" (cf. 祿 "emolument, prosperity"), and perhaps also Welsh dail "leaves, shoots" (cf. 剝 "peel off")

  10. Lucas Christopoulos said,

    October 28, 2023 @ 8:05 am

    @ Brewer
    PIE *gʰreh sounds a bit as Old Chinese *[ɡ]ˤəʔ, *ɡɯːʔ (亥)

  11. Jaka said,

    October 28, 2023 @ 11:43 am

    @Jonathan Smith
    I think you're mistaken '(glossy) black hair' which is 黒髪, for みどり which translated to 'young women' in this case. So the sense freshness/newness is still valid.

  12. Jonathan Smith said,

    October 28, 2023 @ 2:36 pm

    @Jaka if that's the case then yeah, my point dies…
    awaiting info, if there is any, on etymology proper of midori…
    re: lv4 'green', whatever plant lv4 菉 was could be relevant, but the fact remains that lv4 綠 first names a kind of silk… perhaps metonymically due to color of course but hard to say

  13. Jonathan Smith said,

    October 28, 2023 @ 2:37 pm

    lu4 菉*

  14. Chris Button said,

    October 28, 2023 @ 2:52 pm

    I should have noted that 菉 and 綠 are not homophonous anymore. Qiu Xigui has a brief discussion of the fanqie and the relationship in vol. 4 of his collected works.

  15. Chris Button said,

    October 28, 2023 @ 8:47 pm

    The lü reading still remains weird though.

  16. Chris Button said,

    October 28, 2023 @ 8:58 pm

    @ Jonathan Smith

    whatever plant lv4 菉 was could be relevant, but the fact remains that lv4 綠 first names a kind of silk

    The Shijing seems to prove this incorrect on two counts.

    Firstly 綠 seems to mean "green" without necessarily modifying silk, and secondly 綠 is also used as the name of a plant to be gathered. If you need an easy reference, take a look at the glosses in Schuessler's Dictionary of Early Zhou Chinese.

    It's probably worth also noting the much later (albeit close in time to the Shuowen) Shiming "sound" gloss of 綠 as 瀏. The sound of course takes primacy over any semantics here, but it notably further adds the comment 瀏然綠色.

    By the way, there is an isolated oracle-bone form that has been glossed as 綠 in terms of structural similarity (for another easy reference, take a look at Xu Zhongshu's Jiaguwen Zidian). But it's impossible to know what it meant, and I highly doubt it had anything to do with "green" back then.

  17. Jonathan Smith said,

    October 29, 2023 @ 10:38 am

    re: lv4 綠 as simply 'green' in early texts, I'm skeptical… reference seems again and again to be to silk / jacket / cords / coffin linings… including Shijing. It is hard to pin down when this is a color per se.

    Which is the point… our cultural (?) expectation for "colors per se", that is, for discrete color-wheel categories named with etymologically untethered terms, isn't a good guide to what is going on in these contexts.

    (To some extent "color per se" and "incidental color" is a marked difference in modern Chinese languages… Mand. hong(se) 'red', huang(se) 'yellow', lan(se) 'blue', etc. belong to the former at this stage, whereas words for purple, brown, orange, etc., pattern differently.)

  18. Chris Button said,

    October 29, 2023 @ 11:41 am

    @ Jonathan Smith

    From the Shijing:

    綠竹 "green bamboo"

    綠衣黃裳 "green upper garment, yellow lower garment" (here 綠 "green is being used just like 黃 "yellow")

    朱英綠縢 "vermilion plumes, green bands" (here 綠 "green is being used just like 朱 "vermilion")

    采綠 "gather 'lu' grass"

  19. Jonathan Smith said,

    October 29, 2023 @ 2:50 pm

    Of course we can project 2000+ years backwards from modern understandings and be right, but why not be careful? We could read, e.g., early Shijing commentaries and see whether people at that time-depth-ish thought lv4 zhu2 綠竹 was 'green bamboo', etc.

    To make my point just above more simply, if something like 'green silk' works for early usage, it is tough to insist that in particular cases it's simply 'green'. Could be but…

    (the grass lu4 could of course be etymologically relevant, who knows, but in either case being used to write it is totally unsurprising. This comes up in the commentaries noted above. We can find 'color green' written a bunch of different ways even as of right here right now… Chinese character ≠ etymon.)

  20. Jonathan Smith said,

    October 29, 2023 @ 2:51 pm

    *…in either case 綠 being used…

  21. Chris Button said,

    October 29, 2023 @ 4:27 pm

    @ Jonathan Smith

    In the words of the Shijing, you seem to be 采-ing at 綠 (at least its dried variety) right now.

  22. KC said,

    October 30, 2023 @ 2:09 pm

    I was recently at a Japanese ryokan named "Kanamidori" 金みどり

    For my Chinese reading parents, I wrote them that I was staying at the 金綠旅館, even though all of their literature writes "midori" in hiragana みどり, and for all that I know, it could be 金碧, 金翠, or really, none of the above and enigmatically un-kanji-able.

  23. Eiríkr Útlendi said,

    November 1, 2023 @ 7:18 pm

    Interesting thread. I can't speak to the Chinese side of things, but I have spent (too much?) time digging around in Japanese etymologies. I just posted over at Wiktionary about this, for anyone interested: https://en.wiktionary.org/wiki/Wiktionary:Etymology_scriptorium/2023/October#%22original_meaning%22_of_%E7%B7%91_%5Bmidori%5D_in_Japanese

  24. Victor Mair said,

    November 1, 2023 @ 8:50 pm

    Despite all the erudition displayed in the above comments, Eiríkr Útlendi's learned comment is worth a look, a good look.

  25. Chris Button said,

    November 2, 2023 @ 8:57 am

    @ Eiríkr Útlendi

    On Japanese matters, I like the 取る part of 見取る and the idea of a “midoru/mitoru” origin, but I’m not so convinced by the 見 component.

    On Sino-Tibeto-Burman matters, Starostin’s suggestion of an association of 綠 with Burmese rwak “leaf” works well phonologically (and the form with rw- does seem to be in the Lokahteikpan inscriptions) albeit not so well semantically, but it seems to go back to an earlier shift of p(h)- to w-, presumably under the influence of the r-, from pʰak “leaf” (Matisoff talks about this in a few places).

  26. Michael Watts said,

    November 3, 2023 @ 9:30 am

    I must admit to bafflement on how any "silk" semantic connection wound up relating to senses of "leaf" and "green".

    It may be worth noting that the modern Chinese words for red 红 and purple 紫 also sport silk radicals. Chris Button says above that the semantic connection between silk and colors is that silk gets dyed, which does sound plausible.

    Blue (again, in modern Chinese) is 蓝, which has a grass radical!

  27. Jonathan Smith said,

    November 3, 2023 @ 5:12 pm

    "bafflement on how any 'silk' semantic connection wound up relating to senses of 'leaf' and 'green'."

    For again, (Mand.) lv4 didn't/doesn't mean 'silk'. Rather, it is used in early texts to name a KIND OF silk — yes, obviously, one which was dyed some kind of green color, precisely as the SW says (thus the development towards a color name proper.) One again obvious possibility is 'kind of grass, call it LV4 grass' > 'silk LV4 grass in color' >> 'green'. Again IMO, the early texts support 'silk/fabric LV4 grass or sth. in color', not 'green' per se. Re: this point, one must go off and actually read texts, not dictionaries or the like, and develop one's own view.

    See, e.g., Shuowen "綟, 帛戾艸染色" where it is explicitly said that LI4 (or whatever) is silk, dyed with LI4 grass (thus yielding LI4-grass-color-ish-silk). Incidentally, this item is perhaps related to lv4 'green'. At any rate, same process of semantic development.

    This is in general how color terms tend to emerge. I described the Mandarin situation above, in which there is a category of color words proper and an open-ish category of (differently-behaving) words which are used in combination with se4 'color' in a semi-ad-hoc manner — thus ka1fei1se4 and many many others. The Taiwanese case is interestingly different in that (some?) "purists" say that the color words proper are not to be used with sik 'color' at all; rather, the X-sik formula is *reserved* for the ad hoc "coffee-color"-type color names. Whatever the language, the ad hoc items move in some cases towards the "color proper" category, rinse, repeat.

    Yeah in general the "silk radical" tends to suggest (color of) dyed textile. Many other examples. Famously zi1 緇. And given relationship to fabric/dying technology, color names so derived will be relatively late arising. As with lv4. The connection to this "radical" is of course just a tendency though; application of characters to words is capricious like that.

  28. Jonathan Smith said,

    November 3, 2023 @ 5:19 pm

    FWIW, the pre-Qin-ish color of grass/sky/etc. seems often to be cang1 蒼. Certainly not lv4 綠.

  29. Eiríkr Útlendi said,

    November 3, 2023 @ 7:57 pm

    Interesting additional comments. Thank you all for laying out a likely semantic development for the color words spelled with the 糸 radical, that makes sense and seems reasonable enough.

    Re: modern Chinese "blue" spelled as 蓝 / 藍, I note that this means "indigo" in Japanese, and checking a ZH-EN dictionary, it looks like that sense might be applicable to the Chinese as well? If so, the "plant" radical makes sense — indigo the color comes from the use of indigo the plant in dyeing.

    Broadly, and in my ignorance as not-a-scholar-of-Chinese, it seems like we seem to have a group of "secondary" color words with the 糸 radical, likely derived semantically from ancient dyeing practices, and another group with the 艸 radical, likely derived semantically from the colors of the plants themselves (or parts or products thereof).

    … But this diverges from the main topic of the thread, the Japanese word midori and its spelling in Chinese characters. :)

    @ Chris Button, you mentioned, "…but I’m not so convinced by the 見 component." Could you expand on that? I'd be happy to hear any alternative hypotheses for the derivation of Old Japanese mi₁do₂ri. Other phonetic matches for OJP ⟨mi₁⟩:

    * 三 ("three")
    * 水 ("water")
    * 海 ("sea, ocean") — although there is speculation that this ⟨mi₁⟩ might be from the ancient word for "water"
    * 御 (honorific prefix for nouns), derived in turn from 霊・神 ("god, deity, spirit")

    There might be more, that's all I can find at the moment.

    FWIW, the NKD entry available here via Kotobank for the Japanese term midori in reference to both "green" and "new shoot, bud" lists sense [1]-④ as 「みとり(見取)③」. Looking at that entry, sense ③ is about performing a kabuki or jōruri play not all the way through, but rather only doing a selection of notable acts or scenes — aligning with the "looking-taking → choosing from a selection of things" sense of the 見取り spelling. However, this sense isn't attested until the early 1800s, and is thus probably not all that relevant to the Old Japanese terms.

  30. Chris Button said,

    November 3, 2023 @ 9:10 pm

    The word represented by 緇 seems to be related to 彩 "colored" and likely also 色 "hue". As a side note, its graphic form in the oracle bones is interesting. The 甾 "earthen jar" phonetic (which now graphically appears to have the same phonetic as 災) is actually the result of a merger with the form for 西 "west". There's a good discussion in Takashima's Bingbian.

    The word represented by 蒼 is interesting. A possible relationship with 青 is challenging because of its palatal feature that does not appear in 蒼. And the lack of a palatal form in Old Burmese ʰraŋ (which is often cited as evidence for a possible link) is actually the result of the simplification of ʰrj- (a cluster that does not occur in the inscriptions) to ʰr- prior to Old Burmese. So we still need that palatal component.

    Regarding "midori", the idea that "mi" could be 見 just seemed semantically too far removed. I sort of liked the idea of "toru" since 取る as 採る chimes well with the 采 in 采綠 above, but that probably has nothing to do with anything. It just struck me as a nice connection perhaps worth exploring.

  31. Chris Button said,

    November 3, 2023 @ 9:39 pm

    *As a side note, the graphic form of its phonetic in the oracle bones is interesting …

    (To be clear, I didn't mean to imply that the word represented by 緇 was in the oracle bone inscriptions. Rather I was tangentially noting how the oracle bone form for 西 merged with 甾)

  32. Chris Button said,

    November 4, 2023 @ 10:39 pm

    It occurs to me that 生 and 青 might not be related. After all 青 also has 井 (via graphic confusion with 丹 and merger with 月) as phonetic.

    So Old Chinese 生 sraˑɲ from earlier *srjaˑŋ correlates with Old Burmese ရှင် ʰraŋ from earlier *srjaŋ.

    Meanwhile Old Chinese 青 *cʰaɲˑ correlates with Old Chinese 蒼 *cʰaŋˑ albeit with -ɲ ~ -ŋ alternation.

    The -ɲ is the reflex of a palatalized -ŋ, but a difference between non-palatalizing c- and palatalizing cj- is a tall order. And its voiced counterpart ɟ- surfaces as ʝ-, but ʝj- is an even taller order. Cases like this (and others like 戔 with its variant non-palatal reading 殘) suggest to me that the palatal onsets c- and ʝ- inconsistently palatalized the rhyme. In this regard, it's worth noting Nishi Yoshio's (1974 "Birumago no -ac ni tsuite") separation of Old Burmese c- into ts- and tsj- based on whether the coda palatalized or not.

  33. Chris Button said,

    November 5, 2023 @ 7:19 am

    And in that regard, perhaps the difference between 青 and 蒼 is rather one of 青 *tsʰaɲˑ and 蒼 *tsʰaŋˑ both from earlier tsjʰaŋˑ with tsj- simply merging with ts- in some cases before the spread of the palatalization.

  34. Peter said,

    November 7, 2023 @ 12:03 am

    According to the Jidaibetsu Kokugo Daijiten volume on the language of the eighth-ninth centuries (時代別国語大辞典 上代編)the definition is given as being the name of a colour. It gives examples from the Manyoshu amongst others and notes that this word is said to not be the name of a colour originally but referred to young branches or shoots. It further notes that such usage is to be found in the Japanese-Portuguese Dictionary from 1603-4 as well as in some present day dialects.

  35. Jonathan Smith said,

    November 7, 2023 @ 2:10 pm

    @Peter indeed Vocabvlario da lingoa de Iapam via Internet Archive gives the primary definition as "raminhos novos, ramos das arvores" (new sprigs, tree branches [?])… so this applies metaphorically to infant/young woman… should have looked here first ofc.

RSS feed for comments on this post