That's good feedback, although it differs from what I have heard from some
other sources. I would appreciate it if people on this list would consult
their Thai sources to get more feedback on this issue.
It is important when asking questions like this to divorce the issue from
the current implementions on computers, and not prejudge the issue by the
form of the question. If you asked a "(wo)man on the street" how many
characters there were in the headline of a newspaper, what would the
response be? From the answer you can tell whether (s)he is counting
syllables or a finer granularity; which might include non-spacing marks on
a base, but not spacing combining marks). This is also tricky because it
depends how the question is translated, of course.
Mark
peter_constable@sil.org wrote:
> Mark Davis wrote:
> >- A combining character sequence is a type of "grapheme" (aka
> "user character"). Besides CCSs, graphemes also include Indic
> syllables, Thai/Lao syllables and Hangul Jamo syllables.
>
> I have no objection with most of your message, but some concern
> about this paragraph. I can't speak with any authority on Indic
> or Hangul, but I'm pretty sure for Thai and Lao that syllables
> would not be considered a grapheme. By structural and
> functional criteria, preceding and following vowel characters
> have as much claim to being graphemes (consider the character o
> ang, which functions either as a consonant or as a following
> vowel, and is certainly a grapheme regardless of the function),
> and I have no reason to doubt that Thais consider these to be
> as much graphemes as consonants. With combining vowels, tones
> and other diacritics, it might be argued on structural grounds
> that these are not graphemes, but the vowels, at least, are
> functionally on a par with baseline characters, and again I
> believe that Thais would consider them to be graphemes. It is
> certainly true that the notion of syllable is significant in
> the writing systems based on each of these scripts, but at
> least for Thai and Lao, I don't think grapheme should be
> equated with syllable.
>
> Peter
This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:53 EDT