In the course of the discussion thread at this No-sword post, Leonardo Boiko decided to create a List of resources on kanji/hánzí character origins that is useful even for those of us who do not read Japanese or Chinese. As he puts it, “Here’s a few sources I use when investigating characters. They all have etymological theories, and can be opinionated—do apply scholarly skepticism.” Among the “Offline” links, this book looks worth investigating, and the mention of Morohashi took me back thirty-five years to when I constantly heard his name on the lips of the wonderful scholar Susan Cherniack (and my goodness, it seems he lived to be 99!). In Leonardo’s words, “Morohashi is God and the Daikanwa [his dictionary] is the Bible.” But for that, you have to know Japanese.


  1. Bathrobe says:

    I think that should be hànzì.

  2. Correct of course, and fixed now. I try to keep all pīnyīn marks but not knowing Chinese can get you.

  3. By the way, the reason I don’t mind using the word “etymology” for Chinese character analyses is that the word “etymology” is etymologically incorrect anyway (a cool factoid in itself)—as I’m sure you guys know, it would be “the study of the true sense of words” and not the study of their origin and history. The kind of investigation scholars make about the construction of hànzì reminds me strongly of the work of etymologists, complete with the arguments, folk theories, makeshift anthropology &c., so I feel comfortable in extending the term.

  4. Yes, I agree.

  5. the word “etymology” is etymologically incorrect anyway (a cool factoid in itself)—as I’m sure you guys know, it would be “the study of the true sense of words” and not the study of their origin and history.
    Suppose you hold the view that the true sense of words is to be found in their origin and history – as was widely believed up until the 18th century. Then “etymology” is exactly the right description of that activity which studies those origins and histories.
    Nowadays there are few people who believe that words have “true senses”. They believe instead that words have “etymologies”. But the various kinds of thing “etymology” have been taken to mean, over its history, are just part of the etymology of “etymology”.
    Nevertheless, the word “etymology” itself involves no claims about the etymology of “etymology”, so I don’t see that it makes sense to say that ” ‘etymology’ is etymologically incorrect”.

  6. In the second paragraph of my last comment, instead of “But the various kinds …”, please read “The various kinds …”.

  7. I mean the current sense of “etymology” would be “etymologically incorrect” according to the previous definition of “etymology”, which is the only context in which an expression like “the meaning is etymologically correct” makes sense. In other words, the current word “etymology” is in itself a refutation of the old concept of “etymology”, and I find that amusing.

  8. With Chinese you have two parallel histories, the history of the spoken word and the history of the written form. The relationship between the two histories is complex, erratic, and hard to figure out. Effectively the “earliest written form” of a word is just the most durable of the erratic choices ancient scribes made about how to represent a given word. These choices were not entirely haphazard, but they were remarkably more haphazard than we would wish.
    Before the standardization of script ca. 200 BC the same word could have half a dozen script forms, which might be entirely unrelated. Even after the standardization there were often variations.
    There also were historical changes and dialect differences in pronunciation, but these were also hard to untangle since the script is only weakly phonetic.
    Morohashi (Zhongwen Da Cidian in Taiwan) is mostly a historical dictionary of Chinese, with little attention to phonetic history and only some attention to graphic history. I wouldn’t call it etymological.
    Over the years I’ve owned at least ten etymological dictionaries of Chinese (from 3 or 4 periods) keyed to pronunciation, and a couple more keyed to graphic forms, and there were several more which I might have bought if I’d wanted to. I never had any way of comparing or judging the differing reconstructions of pronunciation, especially since no two dictionary used the same notation.
    For a long time I mostly used Karlgren’s obsolete and flawed Grammata Serica Recensa, just because everyone was familiar with it and I couldn’t judge between the others. Recently Axel Schuessler produced “Minimal Old Chinese and Later Han Chinese”, which is simply an updating, correction, and reorganization of Karlgren. (Schuessler has also done original work, especially on the earliest period.)
    As far as I can tell, there are a lot of massive etymological questions still on the table, many of which may never be answered. It’s

  9. For those who read Chinese: there is a very intereting discussion in 陈泽平 (Chén Zépíng)’s /A Study of Foochow Dialect/ (fúzhōu fāngyán yánjiū), about the characters 冰 and 凝. To recapitulate: there are two words in (Common) Chinese, one somewhat like [piŋ] (bīng in Mandarin), the other [ŋiŋ] (níng), that means respectively “ice” and “to freeze”. In Modern orthography, piŋ is 冰; ŋiŋ is 凝. In Shuōwén however, the first word is written 仌, the second 冰. For a long time during this transition, the graph 冰 can stand both for piŋ the noun and ŋiŋ the verb. But where does the modern graph 凝 come from? Turns out that it has an attested variant reading, transcribed simplistically ŋik, which has given the modern Foochow word for “to freeze” ŋiʔ.

  10. 凝 is a very common character in Japanese, with the meaning “solidify” or “coagulate.” Freezing, naturally, is the solidification of water, but where is the mystery? Isn’t 凝 just as common in China?

  11. That the graph 凝, originally for a synonym ŋik, is taken to write the word ŋiŋ — because its original graph 冰 is now occupied by the word piŋ, originally written 仌.

  12. I bought a pirated copy of Morohashi in Taiwan in the ’80s for about a hundred dollars and used it throughout grad school. It served me well, but it was superseded by a twelve-volume dictionary from the mainland, the Hanyu dacidian, in the early ’90s. I have both on my shelf and I almost never look at Morohashi anymore.

  13. Ken, I have wanted to buy the Hanyu dacidian for some time and can’t find it anywhere. Amazon lists it but the supposed supplier doesn’t seem to actually have it in stock. Do you have any ideas?

  14. David Marjanović says:

    Turns out that it has an attested variant reading, transcribed simplistically ŋik

    Incidentally, such an alternation between -k and -ŋ is common in reconstructed Old Chinese and Proto-Sino-Tibetan; apparently it marked a morphological distinction that has been lost.

  15. Bathrobe says:

    I have the Hanyu Dacidian, bought from a bookshop in Beijing circa 2001. Is it really far superior to Morohashi? Used for looking up the history of bird names, for instance, the HYDCD is rather frustrating. 鶇 emerges virtually as a character without a history (it gives only the modern sense of ‘thrush’), and I was hoping that Morohashi might have some light to throw on this. Even my cheap Chinese edition of the single-volume Obunsha Kanwa Jiten notes that 鶇鵐 means ‘Wryneck’, which is more than HYDCD gives.

  16. @John Emerson: Thanks. As you can see I’m working with laypeople and students’ resources; when this post got picked up by the Hat I immediately expected knowledgeable opinions to show up. In any case, even with these less scholarly sources, it was already clear that there are no easy answers.

Speak Your Mind