I knew Chinese typewriters were big and complicated, but to see them in action is a real eye-opener. See Victor Mair’s post at the Log and the photos and videos thereto appended; here’s his description of the process:

The main tray — which is like a typesetter’s font of lead type — has about two thousand of the most frequent characters. Two thousand characters are not nearly enough for literary and scholarly purposes, so there are also a number of supplementary trays from which less frequent characters may be retrieved when necessary. What is even more intimidating about a Chinese typewriter is that the characters as seen by the typist are backwards and upside down! Add to this challenging orientation the fact that the pieces of type are tiny and all of a single metallic shade, it becomes a maddening task to find the right character.

I’m glad they exist, and I’m glad I never had to learn to use one.


  1. Prof. Mullaney, who weighs in to the LL post, does not mention that his own blog is called The Chinese Typewriter, documenting among other things his collection of them.
    All I’ve got is a Kanji tablet; but it does still work with a little Emacs code.

  2. Note that the Wired piece referenced by Wikipedia claims that its illustrations are from the November 1947 issue of Popular Mechanics.
    They are from Popular Science. The similar article in Popular Mechanics the following month is actually better at illustrating the Mingkwai’s operation.

  3. Years ago, when publishing a book in Japanese, I encountered a nearly identical machine to the manual model in the video at the LangLog link. The operator was far defter than in the video.
    What really got to me at the time was the sight of elderly Japanese men setting type manually for a newspaper. Racks and racks of tiny stacked bins, each column of bins about five feet high, and the lot stretching 15 or more feet across two walls. Although printing was invented in the East, it never caught on the way it did in the West because of the difficulty of setting type in ideographic writing systems. It wasn’t until offset printing came along that, say, a nicely calligraphed page could be easily printed without resorting to woodblocks or similar. The advent of computers made printing easier yet.
    I recall inquiring about word-order in Japanese dictionaries. I learned that there are two systems: Either by sound, or by the number of strokes in a character and their arrangement in a notional grid. I subsequently read that the sound order loosely follows the sound order of the Western and Semitic alphabets. This is due to Buddhist texts written in Indian alphabets reaching Japan and the adoption there of those alphabets’ sound orders.

  4. The Indian alphabet order is ordered quasi-phonologically, first the stops (nasal included), from the innermost in mouth (k), to the outermost lips (p), then the fricatives and approximants. quite different to the Semitic one. The Devanagari order, for example, goes ka kha ga gha nga, ca …, ṭa …, ta …, p pha ba bha ma, ya ra … The Japanese order is indeed a subset of it: a i u e o, ka …, sa … (ts- in the ancient times, to correspond with c-), ta …, na …, ha … (anciently p-), ma …, ya …, ra ….
    The Semitic/Western order should probably be regarded as purely historical. I, for one, can’t see the logic to begin with ʔ, then b, then g, then d (why not b d g or g d b?), then a fricative h, then a semivowel w, then a voiced fricative z, then a voiceless fricative ḥ.

  5. The Semitic/Western order should probably be regarded as purely historical. I, for one, can’t see the logic
    Has anyone ever claimed it was logical? Of course it’s “purely historical” in the sense that it developed (like most things human) according to historical contingencies rather than logic; the fact remains, however, that it is inextricably embedded in all Semitic/Latin/Greek/Cyrillic writing systems, and we’re all used to it.

  6. The Semitic/Western order should probably be regarded as purely historical.
    Joseph Naveh, in his “Early History of the Alphabet,” writes (p.30) “(T)he earliest evidence for the present-day alphabetic sequence was found at Ugarit, where the thirty cuneiform alphabetic letters were inscribed on clay tablets. The transcription of these tablets is as follows (Fig. 23):”
    ‘a b g h(with inferior caron) d h w z h (with inferior dot) t(with inferior dot) y k s (with a superior caron) l m d(with inferior line) n z(with inferior dot) s ‘ p s(with inferior dot) q r t(with inferior line) g(with superior dot) t ‘i ‘u s(with subscript 2)
    He elaborates no further.

  7. J.W. Brewer says:

    I never knew that about the order of the kana! Happy to be enlightened after all these years.

  8. @hat: I was responding to Paul, trying to say that “Indian and Semitic orders are different”, in the context that “the sound order loosely follows the sound order of the Western and Semitic alphabets” because of Indian influences.

  9. The seven-day week also goes back to Babylon or thereabouts, as I understand.

  10. I once sent a telegram from Nanjing to Jingdezhen in China to tell our friends in J. which train we were arriving on. The poor telegram office clerk had to translate each character into a 4-digit numeric code, although I imagine she knew the codes for the most common telegram words by heart.
    During that same year (87-88) in South China I experimented with various computer input methods then available on Chinese computer systems. It was the era of ‘character at a time’ rather than ‘word at a time’, but I could type pinyin faster than most of my Chinese colleagues could input characters by other methods. Cantonese were terrible with pinyin!

  11. The telegram people know ALL the random four-number codes. Indeed, after the institution of telegram has known its demise, many changed work as computer-based typists, and they use telegram codes to input, while we common people use Pinyin or Wubi.

  12. Nicholas Ostler, on p210 of “Empires of the Word,” notes that “Japan owes the order of symbols in its syllabary, the so-called kana, or go-ju-on, ‘fifty sounds’, to the order of letters in Indian alphabets. The order of Sanskrit letters is conventionally [...].
    “This is not an arbitrary order like our ABCD . . .*”
    Footnote: “* The motivation for this is purely historic. It ultimately goes back to an equally arbitrary ‘aleph beth gimel daleth . . .’ specified by the Phoenicians.”
    The Semites who first toyed with alphabetic writing did so under the influence of Egyptian hieroglyphics. Perhaps the letter-order derives from there.

  13. Japan also has a native ordering, the iroha, which is very poetic and charming for enumerating short things but much less pleasant when you’re trying to find your row in a theater or look something up in an early Meiji index.

  14. Er, short _lists_.

  15. The description of the Chinese typewriter made me think of Chinese and Japanese Braille alphabets. I have heard they exist, but haven’t had the chance of reading them with my fingers. As a blind person, I have always found it fascinating to think of an entire alphabet which is represented through pictures. I studied a bit of Japanese at an International program for blind students in America, but we were urged to write the words in Latin letters, since no one in the vacinity was familiar with the Braille presentation of the language. Now that my curiosity has been reignited, I am going to look into it.

  16. The iroha order is fine for labeling the 48 curves on the mountain road up from Nikko to Lake Chuzenji in Tochigi Prefecture, Japan, although it’s a shame each kana symbol there doesn’t start its own rhyming slogan like the old Burma Shave signs along U.S. highways.

  17. It would be great if the Japanese had just started to write in Devanagari or Siddhaṃ or something with the adoption of Buddhism. The mismatch of Chinese writing with the language is so clunky. Unfortunately Han writing had (and still have) too much prestige.
    Don’t get me wrong, I love them kanji, but one has to admit that writing こころよい as 快い or おとな as 大人 totally ruins the one-character-one-morpheme principle.

Speak Your Mind