RE: Ideograph?!?

From: Kenneth Whistler (
Date: Mon Nov 29 2004 - 19:36:16 CST

  • Next message: Robert Finch: "Keyboard Cursor Keys"

    Allen Haaheim provided some further detailed clarification:

    > Note that Han characters are logographic, not ideographic. That is,
    > they are graphemes that represent words (or at least morphemes),
    > not ideas.

    This correctly states the situation for the normal case for
    Chinese characters used writing the Chinese language in most
    instances. But as is not unusual for real writing systems, the
    situation gets blurred all around the edges.

    For one thing, Chinese has characters which are simply used for
    their sound, as syllabics. In some instances, they are characters
    in dual use, as logographs *or* as syllabics, but in either
    instance they are used to "spell out" foreign words irrespective
    of the morphemic status of the orginal characters -- or the
    morphemes of the foreign word, for that matter.

    And the situation is also not so clear when considered in
    the dynamic context of the historical borrowing of the Chinese
    writing system to write unrelated languages such as Japanese,
    Korean, and Vietnamese. Much of the writing system borrowing
    was *attached* to words -- in other words, the vocabulary itself was
    borrowed in from Chinese, using the Chinese characters to
    write it. But Japanese and other languages faced the problem of how
    to adapt the writing system for preexisting, *native* vocabulary,
    as well as for all the borrowed words from Chinese. And a
    variety of strategies evolved, some of which involved
    abstracting the *meaning* of a Chinese character, and then
    reapplying the character to write an unrelated word in Japanese
    (for example) which had a similar meaning. This semantic-based
    transference of Chinese characters completely ignored
    morphemic status in Chinese, as the whole point was to simply
    find the appropriate character to express the lexical semantics
    of the historically unrelated (but semantically similar)
    word(s) in the borrowing language.

    During such a borrowing transition, you can conceive of
    the process as many Chinese characters temporarily "floating off"
    their morphemic anchors in Chinese, being considered
    purely semantically, and then reattaching to a new set of
    morphemic anchors in Japanese, where they subsequently
    evolve with new lexical histories in another language.

    > But somehow "ideograph" has become the standard term in use outside
    > the field of experts in Chinese linguistics (because of Ezra
    > Pound et al., perhaps?).

    I don't think you have to look to Ezra Pound's poetic
    misrepresentations of the nature of Chinese to find
    reasons here.

    "East Asian ideograph" and "CJK ideograph" caught on as
    acceptable compromise alternatives for "Chinese character"
    or "Japanese character", which were language-specific and
    misleading (in the Japanese case), or for transliterations
    such as kanji or hanzi (also language-specific), or for
    sinogram or sinograph, which were too little known (and
    also too Chinese-biassed for some). "East Asian logograph"
    would have been technically a little more correct, but
    not absolutely right, either. "Ideograph" wasn't used because
    the standardizers were confused about how Chinese and
    Japanese writing systems worked, but simply because it
    was a usable term in the right ballpark, available for
    a specialized technical usage, and less objectionable
    than most of the alternatives.

    As Asmus and Richard implied, "ideograph" should simply
    be treated as polysemous now. It has a narrow technical
    sense applying to the character encoding world, where it
    effectively is equivalent to kanji/hanzi/hanja. And it
    has a separate graphological sense where it refers to
    signs (like symbols marking restroom doors) that represent
    ideas directly without being attached to specific words
    or morphemes of a particular language.


    > I hope this doesn't confuse matters.

    This archive was generated by hypermail 2.1.5 : Mon Nov 29 2004 - 19:37:31 CST