Re: Measuring a writing system "economy"/"accuracy"

From: David Starner (
Date: Wed Jun 29 2005 - 16:56:27 CDT

  • Next message: Jony Rosenne: "RE: Measuring a writing system "economy"/"accuracy""

    On 6/29/05, Eric Muller <> wrote:
    > Yes, the question is "when a writing system is viewed as a mechanism to
    > record sounds, how good a job does it do?", where "good" is to be
    > defined. I chose "economy", because arguably a writing system that has
    > 10 symbols or symbol combinations for the same sound is not "as good" as
    > one that has only 1.

    You write like you assume that every writing system should be some
    form of alphabet. Why should sounds corresponding to writing be a good
    thing? In many ways for English, it wouldn't be good if our writing
    had a one-to-one correspondance between letters and sounds, since then
    I would have to try and understand British and Indian accents in
    writing as well as speech. Is it better for Cherokee to be written
    with 84 letters or 21 letters that need twice as much space? (Or with
    all the accents and stuff that linguists transcribe Cherokee and
    ordinary writers go without?) You're begging the question.

    > IPA, as least when restricted to the set of symbols used for the writing
    > of a given language, is presumably both an economic (there is a single
    > sign for a given sound) and accurate writing system for that language.
    > Hence the idea of measuring by comparing to IPA

    The words economic and accurate are misleading. IPA is not maximally
    economic in the number of symbols it uses--digraphs have fewer
    symbols, and represent the sound as unambigiously. By many other
    standards of economic, like the number of characters it takes to write
    something, the IPA is less than economic.

    The important, IMO, standards of economic relate to how fast can it be
    written, how fast can it be entered into a computer, how fast can
    printed material be written, and how fast can written material be
    read. The IPA is not a winner on several of those, as several symbols
    are similar and hard to write distinctly and clearly distinguish on

    If you judge accurate by the exact representation of sounds by
    letters, IPA will be more accurate, but IPA is not more accurate in
    representing the history of the word, nor is it more accurate in
    representing a hypothetical compromise between competing dialects, nor
    is it more accurate in representing the underlying forms that
    underwent phonetic changes when prefixs and postfixes and what not
    were added.

    I don't see any point of measuring by the IPA; it's fairly obvious
    from the basic knowledge of a script, and I don't think tells us very
    much at all about how good a script is, especially as the mapping from
    glyphs to (completely abstract) characters is as much an important
    part of the script as the mapping from characters to sounds.

    This archive was generated by hypermail 2.1.5 : Thu Jun 30 2005 - 01:50:11 CDT