List of humdrum words that look cool in kanji (by: Sei Shônagon)

For this section of The Pillow Book, it would be interesting to know precisely which characters Shônagon was thinking about, so we should be wary of how exactly our chosen edition arrived at its kanji. Unfortunately I know next to nothing about the different manuscript lineages, but since the 11th-century original was kana literature (wabun), I suppose that’s a bit of a moot point anyway—we can’t have primary evidence of her intentions. Therefore, for this post, I’ll just use the University of Virginia’s E-text version, which is based on the 1929 Yûhôdô 有朋堂 edition, itself based on the Nôinbon manuscripts:


  • 覆盆子。
  • 鴨頭草。
  • みづぶき。
  • 胡桃。
  • 文章博士。
  • 皇后宮の權大夫。
  • 楊梅。
  • いたどりはまして虎の杖と書きたるとか。杖なくともありぬべき顏つきを。

Here’s William Morris’ translation, courtesy Dorothy Disse:

Words That Look Commonplace but That Become Impressive When Written in Chinese Characters:

  • Strawberries.
  • A dew-plant.
  • A prickly water-lily.
  • A walnut.
  • A Doctor of Literature.
  • A Provisional Senior Steward in the Office of the Emperor’s Household.
  • Red myrtle.
  • Knotweed is a particularly striking example, since it is written with the characters for “tiger’s stick.” [i.e. 虎杖] From the look on a tiger’s face one would imagine that he could do without a stick.

And here’s Meredith McKinney’s newer translation:

Things that look ordinary but become extraordinary when written

  • Strawberries.
  • The dew plant.
  • The prickly water lily.
  • Spiders.
  • Chestnuts.
  • Doctors of literature.
  • Acting Master of the Empress Dowager’s Palace.
  • The arbutus tree.

People write the name ‘knotweed’ with characters meaning ‘tiger’s staff’. A tiger doesn’t look as though it would need a staff!

Let’s assume, as a working hypothesis, that the characters from the Yûhôdô edition above are plausible interpretations of the ones Shônagon meant.

  • 覆盆子: This actually is the Chinese word for “raspberry” (modern Mandarin fùpénzǐ). In Japan the three characters were borrowed together as a way to write the word “strawberry”. If read as Sino-Japanese (SJ), it would result in something like *fukubonji; and, if interpreted semantically, the characters would suggest cover-tray-[suffix]; neither option has anything to do with the Japanese word ichigo (I don’t know if the word was etymologically transparent to Shônagon, but it’s thought to derive from ichibigo, of unclear meaning but definitly unrelated to covers or trays). Giving the courtly taste for complex writing and Shônagon’s notorious affinity with Chinese characters, I’m guessing that it must be precisely this mismatch that drew her attention.
  • 鴨頭草: This can be read as tsukikusa (moon-herb?), an alternative
    e name for tsuyukusa (literally “dew-herb”), Commelina communis. The characters mean “duck-head herb”, which again has nothing to do with the Japanese morphemes, much less the sound (Mandarin: yātócǎo, SJ: ôtôsô). Apparently “duck-head” was a poetic expression for a shade of green (though, judging from Wikipedia, the current Chinese name is “duckfoot herb” 鴨跖草).
  • みづぶき: Online dictionaries agree with Morris that midzubuki “water-butterbur” is an older name for what’s now called onibasu “devil lotus”; i.e. the prickly water lily, Euryale ferox (“ the stems, flowers, and leaves which float on the surface are covered in sharp prickles”; therefore “prickly”, “gorgon”, oni, ferox). Unfortunately this text doesn’t suggest us which Chinese characters were used, and why Shônagon would be interested in them.
  • 胡桃: C. hútáo “barbarian-peach” (SJ would be *kotô or *utô), with the writing borrowed for Japanese kurumi. “Barbarian” here refers to Central Asians. I don’t know where kurumi came from (obvious guess: kuri “nut, chestnut”, mi “fruit, seed, nut”?), but it’s surely unrelated to both barbarians and peaches. The characters are used in both modern Chinese and Japanese to denote walnuts (Juglans spp.), not chestnuts (Fagaceæ spp., J. kuri), so I don’t know why McKinney chose the latter (there’s a footnote but I can’t read it in the Amazon preview).
  • Bunshô Hakase and Kôgôgû-no-Gondayû: Perhaps these ritsuryô formal positions were not prestige enough to Shônagon’s tastes, and yet they sounded impressive when written with a string of kanji?
  • 楊梅: Myrica rubra. Not a true myrtle (Myrtaceae) but a “bayberry”. McKinney has instead the arbutus (“madrones” or “strawberry tree”, Ericaceæ). C. yángméi “willow-plum” (SJ yôbai, attested), while the native J. is yamamomo “mountain-peach”.
  • By now the reader must have guessed how it goes, right? Fallopia japonica is J. itadori, possibly coming from “pain-stopper” due to its medicinal properties. The Chinese name, from which the Japanese kanji ortography, is 虎杖 hǔzhàng (SJ *kojô), “tiger cane”, which again is utterly incongruent with Japanese sounds and meanings.

In conclusion, Shônagon probably found obscure borrowings to be very wokashi.

17 thoughts on “List of humdrum words that look cool in kanji (by: Sei Shônagon)

  1. Re Bunshô Hakase and Kôgôgû-no-Gondayû: Maybe the issue here is that in the actual (spoken) words, the sounds (notably those corresponding to the “impressive” kanji like 博士 and 太夫) have evolved to the point where they no longer have an obvious direct relationship to the then-current “on-yomi” of the kanji.

    So when you just say /dayū/ it’s totally ことなき, just an unremarkable set of sounds, but when you see it written down and you realize it was originally 太夫 /daifu/ “big guy”, you’re like “whoa” (“…kashi”).

    The difference between Morris and McKinney’s translation of the title is interesting. I wonder if SS considered 仮名 a separate (non-overlapping) category from 文字.

  2. Re morphemically opaque words—makes sense, yes. The overall theme then fits the title—“wow, this word is written like this? Cool!”

    What I’ve read so far suggests that kana was even more “the same” category as mana/moji than today, if anything (judging from Seeley, LaMarre, the Utsubo Monogatari, certain passages in the Genji); as far as I know, at that time, it was more of a calligraphic/stylistic distinction than a linguistic one. I’d guess “things” (mono) here imply speech: boring things(-and-names) that are written in an impressive way, as McKinney’s heading has it.

    And thanks for the contribution, guys, enriching as always. (whoa-kashi groan)

  3. Actually, this post is quite relevant to my rant at the previous post. It seems that Sei Shōnagon is making a very clear distinction between ‘words’ and the ‘characters’ that are used to write them.

    This is the opposite of the approach I was complaining of, that is, characters came first, language is just a matter of learning how to read them.

    The opposite tack is taken at Jack Halpern’s page on what he calls ‘Japanese homophones‘. For Halpern, 上る, 登る, and 昇る are all ‘homophones’ in the same way that 橋, 端, and 箸 are. I would suggest that his approach is quite confused. 橋, 端, and 箸 are different words that are (quite logically) written differently. They are real homophones. 上る, 登る, and 昇る, on the other hand, are ‘false homophones’ — i.e., the same word that happens to be written differently. In this case I would submit that Japanese tries to distinguish among different senses of the same word by using different Chinese characters, which are available to them by virtue of the fact that they have borrowed Chinese characters to write their language. It’s not a matter of ‘different words that happen to be pronounced the same’, which is what Halpern’s use of ‘homophone’ suggests.

    The failure to realise that speech should be treated as prior to writing (or at least realise that writing and speech should not be jumbled together) is one of the problems that keep intruding when naive native speakers (including so-called ‘experts’) talk about language in Chinese-character using languages.

  4. The relevance was intended :) And of course I entirely agree with you, and language teachers talking about how “Japanese has many different miru” etc have long been a pet peeve.

    A similar peeve are people who try to explain the cultural nuances of a word by analyzing the elements of the Chinese characters used to write it. I had a lit teacher say that scorn 侮蔑 in a story represents a cold emotion because, unlike dread 恐怖, it has no heart 心 in it…

    So yeah, I very much agree with the linguists that writing represents language, language is essentially/primarily spoken, and that there’s a ideographic myth that needs debunking. It’s just that I think their rhetoric gets too strong sometimes, and that there are some important caveats.

  5. 橋, 端, and 箸 are different words that are (quite logically) written differently. They are real homophones.

    This is totally a side issue, but are they even really homophones? This is the notorious “same kana, different pitch accent” trio that all learners encounter, after all.

  6. It’s customary in linguistics to ignore Japanese tone whenever doing so is convenient to the discussion at hand.

    No seriously, the pitch accent (isn’t it rather a lexical tone? could anyone convince me that the last H in a HHHL is actually stressed?) is kind of accessory IMHO. After all, it varies a lot with dialects, and Eastern/Edo wasn’t the standard until recently, and even the current standard is somewhat artificial/normative—adults have to study the NHK guide for jobs, etc. What was the tone like in Old Japanese is anybody’s guess (I think the Ruiju myôgisho has Middle J tones, I should acquire a copy…). I think, as far as etymology is concerned, we can say they’re “close enough”.

  7. Answering my own question:

    Although it has been claimed that “pitch accent” is not a coherently defined term,[1] it is commonly understood to refer to a language that uses phonemic tone, but where only one or two syllables in a word can be phonemically marked for tone, and many words are not marked for tone at all. In such languages, the syllable with phonemic tone typically is acoustically prominent, in a similar fashion to the dynamic stress of languages such as English or Spanish.

    The reference has more info:

    This issue has arisen in a number of cases, the most prominent of which concerns Tokyo Japanese. As seen in (4), Tokyo Japanese has been analysed both accentually and tonally (Haraguchi 1977, McCawley 1978, Poser 1984, Pierrehumbert & Beckman 1988, etc.)

    Apparently I’m not the only one bothered by the notion of “pitch accent”; but from a glance at the paper, there seems to be legitimate reasons not to describe SJ simply as “tonal”.

  8. Whatever you call it, Matt is right, 橋, 端, and 箸 aren’t even true homophones, except in certain dialects that have lost pitch (e.g. Ibaragi). Even if it varies between dialects, the distinction is still there. (The same goes for Mandarin Chinese, where tones differ greatly between dialects but is still distinctive phonemically.)

    Getting the pitch wrong will pretty much invalidate your careful efforts to acquire a Japanese dialect accent. Kansai-ben isn’t just a matter of adopting weird sentence pitch, changing a few endings, and using a few non-Kanto words. If you haven’t got the pitch right, you’re basically speaking fake Kansai-ben.

  9. By the way, Bathrobe, I didn’t mean to pick on you — I was just wondering out loud, basically about whether Boiko’s Custom as cited above is actually a Thing in the literature or just something we all do

    I’m not convinced by the “it varies by dialect” argument either because, as Bathrobe says, although that is true, the variation is systematic. Re normativity, there are two issues here. One is that accent-dropping from nouns is a standard “Nihongo no midare” gripe; in the speech of many young people, you hear nouns that were once HLL becoming LHH (no drop), and so on. I don’t know how systematic and complete this change is, but at least in the Kanto region we seem to be in the middle of a Great Accent Shift of some sort, and that would obviously make the standards of the previous generation appear more normative by contrast.

    Second issue: actual changes aside, there are no doubt a lot of “edge case” words that people don’t use often and have to study the NHK dictionary to get right, but I bet you could say the same of English. Surely many people who want to work at the BBC go through something similar in an attempt to normalize their pronunciation of rarely-used vocabulary. (Not to mention the regional accents thing, although I understand that the BBC is more forgiving about that these days.)

  10. Is that so? I thought the tone was a lot less important than in languages like Mandarin. I mean, if you speak Mandarin with the wrong tones, it’ll be downright unintelligible (right?), whereas Standard Tokyo Japanese with the wrong tones is just funny gaijinspeak. The point about accent varying with dialect is that, even though speakers of different areas have totally different tone contours, they’re still mutually intelligible (are they?), so the tone can’t be too necessary (in an information-theory sense).

    But what I was trying to get at is that I don’t know if the words weren’t originally homophones. I wonder if there’s any work on the origin and evolution of Japanese tones? It might be difficult, since the writing system doesn’t bother with them…

  11. From Wikipedia:

    日本語のアクセントの歴史については、京都のアクセントの記録が平安時代から残っており、今の京阪式アクセントになるまでにどのような変化をしてきたのかが明らかになっている。… 金田一春彦は、東京式アクセントは京阪式アクセントが変化して生まれたと推定し、これがほぼ定説に近いものとなった。

  12. As for Chinese tones, different varieties of Mandarin are often mutually intelligible despite the tones being weirdly different. For one thing, the tones are different in systematic ways, so you can set up ‘equivalences’. Secondly, Chinese can often be understood even when the tones are wrong, as long as the contour of the word is identifiable. Of course, mutual intelligibility is often quite low among extreme examples of Mandarin, but that’s not necessarily the fault of tones.

  13. The point about accent varying with dialect is that, even though speakers of different areas have totally different tone contours, they’re still mutually intelligible (are they?), so the tone can’t be too necessary (in an information-theory sense).

    It’s true that in SJ you can almost always tell what someone means from context, so in an information-theory sense tones aren’t vital. BUT… you would get the same result by speaking English in an accent where all the vowels are replaced with [æ], or voiced and non-voiced obstruents are swapped. And we don’t say that vowels or voicing are unnecessary or ignorable in English; they’re all just examples of multiple redundancy.

    I bet that if Japanese had evolved a “natural” way to show pitch accent, along the lines of dakuten, calling 橋 and 箸 homophones would seem as odd as treating “bit” and “bat” that way. (It would also have greatly affected the course of Japanese literary history, splitting West and East like that…)

  14. Martin does mark the accent in his grammar, and I’d certainly like if more people did that (especially in dictionaries and such). A dakuten-like mark for the accented syllable could be a solution for kana/kanji writing… Well diacritics in kanji sound unusual, but at least it would be more elegant than the NHK発音アクセント辞典 notation.

  15. Kenkyusha has a pocket Japanese-English dictionary that marks accent. I was quite annoyed that our 1st-year Japanese teacher (many many years ago) assigned the Sanseido dictionary, which doesn’t. If pitch accent is taught from the start it’s not so hard to pick up bad habits. Coming back to words you’ve learnt incorrectly is always difficult.

Leave a Reply

Your email address will not be published. Required fields are marked *