RE: Ideograph?!?

From: Richard Cook (
Date: Tue Nov 30 2004 - 04:36:44 CST

  • Next message: Michael Everson: "Re: Relationship between Unicode and 10646"

    On Mon, 29 Nov 2004, Kenneth Whistler opined contemplatively:

    > 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.

    As usual, Dr. Whistler has hit upon some key and interesting ditinctions.
    The association of a Chinese-derived character with any particular
    morpheme is a slippery slope, as is morphemics or semantics in general, I
    reckon. Much less slippery is phonology, and one is on firmer ground to
    assert that Chinese characters (and characters derived from the Chinese
    tradition, either directly, or indirectly) are regularly associated with
    specific quantifiable pronunciations, in whatever language. And even more
    specificially, they are associated with one or more monosyllabic readings.
    This is why I call the Chinese character an element of a heterographic
    syllabary. In a particular dialect one can identify a set of possible
    syllables, each of which may have one or more writings via a single
    character. In an isographic syllabary, on the other hand, each syllable in
    the canon has but a single graphical rendering. It is the case in Chinese
    that a given morpheme with a certain well-defined pronunciation may in
    fact be representable with more than one character. This is due to
    character variants, or simply orthographic variation. And the graphical-
    semantic- phonological complex shows variations in all three of these
    dimensions, in time and locale ...

    > > 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.

    Pound was famous for being very *bad* at Chinese, but at least he was
    enthusiastic (and only mildly fascist). Prior to Pound some would argue
    that early use of "ideograph" was due to a Jesuit misunderstanding. The
    Jesuits are in fact the ones who also brought us "radical" due to a
    creative analogy with Hebrew "root". It is however less clear to me that
    early use of ideograph is always a mistaken interpretation of Chinese
    writing as "more semantic" and "less phongraphic". It is true that Chinese
    is poorly phonographic. But is it better at ideography? The term ideograph
    needn't be understood as "idea" writing (skipping speech completely),
    which interpretation I think follows one etymology of the "ideo"
    component, associating it with the Greek counterpart of English "idea".
    But the Greek root of "idea" is actually the same as that of English
    "vision", a root related to 'seeing'. Greek "eidon" is also related, and
    means 'image' or 'form' (that which is or has been seen). So, the term
    "ideograph" can be understood in some sense as a pure Greek equivalent to
    Greco-Latin "pictograph", which is to say, a drawing of a real-world form
    which one has witnessed. The components of Chinese characters in the
    Eastern Han tradition are in fact explained in such terms, as representing
    real-world objects. So, "ideograph" in the hyper-literal sense is really
    not so far from the mark, though unfortunately I think many rightly fear
    and/or abhor the idea-writing misconception.

    > "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.
    Yes, one might ask why was the term "logograph" not used instead? I
    believe that again the status of "word" (logo-) is sufficiently insecure
    to have argued against using logograph. Words in Chinese are sometimes (I
    should save often, but I'm unsure how to quantify this) polysyllabic
    rather than monosyllabic. Some like "morphosyllable", though as I said,
    semantics is a slippery slope, and identification of a morpheme, let alone
    association with it is difficult nay impossible. I think
    heterosyllabograph is by far the best term, but thankfully we do not have
    CJK Unified Heterosyllabographs, since, for one thing it takes to long to

    > 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.

    Yes, I think we have witnessed language change here, in development of
    info-tech usage of ideograph, driven by inter-cultural contact, which is
    of course natural in standards work bridging cultural gaps. As one Apple
    ad campaign once said in big letters, "embrace", and in small letters
    "change". Like it or not (and I admit that I did not like it much for a
    long while), this usage is here to stay, and we must all recognize the
    function it served and the circumstances in which it arose, and embrace
    the change. Funny how standards change things ... hopefully for the


    This archive was generated by hypermail 2.1.5 : Tue Nov 30 2004 - 04:38:26 CST