Re: How is the glottal stop used in some languages?

From: Richard Wordingham (
Date: Thu May 10 2007 - 16:27:00 CDT

  • Next message: Marnen Laibow-Koser: "Re: glottal stop"

    Chris Harvey wrote on Thursday, May 10, 2007 4:08 PM

    > I’m concerned with the addition of characters which are visually
    > identical, and only differ in that one is punctuation and the other is
    > meant to be an orthographical letter. As in the case for U+02BC

    > Ojibwa can be typed on a
    > US-English keyboard as long as the apostrophe is understood to be
    > U+0027 or U+2019 (for those programs using auto-quotes). To introduce
    > U+02BC would be very confusing to Ojibwa speakers; why is ' one thing
    > in English but another in Ojibwa? I have had no success in
    > communicating the practical need for two apostrophes, one for English,
    > one for the Native language with speakers and language educators.

    I must confess I am puzzled as to why the 'punctuation apostrophe', as in
    English "can't", should be U+2019 rather than U+02BC. There must be an
    explanation somewhere. It may be simply that it is too much to expect
    people to make the corrrect distinction between U+2019 and U+20BC in
    English. There are a few examples, such as "must've" and non-standard
    "wa'er", and alien names like "Vl'hurg", where it is clearly letter-like,
    but they are probably not enough.

    > We could go further, Squamish writes its glottal stop with a 7, Tlingit
    > with a period . , Arapaho writes /θ/ with the number 3. These
    > orthographies were developed so that as few exotic characters as
    > possible would be required, and that these languages could be typed on
    > an English keyboard. Should new MODIFIER NUMBER SEVEN, MODIFIER NUMBER
    > THREE characters be introduced?

    In theory, yes. A hypothetical Arapaho *3a3a would be title-cased, by
    default, to "3A3a", and, if I have not misinterpreted rule LB24 in UAX #14
    Line Breaking Properties Unicode 5.0.0, there would be no line break in
    *3a-3a from standard line-breaking unless hyphenation rules cut in. (I may
    have misunderstood them - I'm seeing automatic line-breaking break at the
    ASCII hyphen without trouble, but also getting line breaking at the
    hyphen-minus of '20.0e-3'.). However, the postulated Arapaho hyphen problem
    should go away if you use U+2010 HYPHEN for the hyphen function, instead of

    However, it it were worth the trouble of implementing extra characters, I
    feel it would probably be better to try to move to more conventional letter
    shapes, rather than add characters that will probably cause endless trouble.

    > Perhaps I’m alone in thinking this, but users cannot be expected to
    > differentiate between two visually identical characters, one for one
    > language, one for another.

    That's probably easier than distinguishing two identical characters in the
    same language, and we often do it, albeit unreliably. However, I may be
    underestimating the importance of the unpredictable distinctions made by
    fonts, e.g. '0' v.'O'.


    This archive was generated by hypermail 2.1.5 : Thu May 10 2007 - 16:28:43 CDT