Re: THEORY Ideal system of writing

On Monday, August 9, 2004, at 07:23 , John Cowan wrote:

> Ray Brown scripsit:
>
>> B. Obvously, 170 to 200 is too small an inventory for all the morphemes
>> of a language, yet it seems rather high for a syllabary. Y.R. Chao does
>> not elaborate on what each of the 170 to 200 symbols would represent,
>> except the brief reference to monosyllables in (3). Any ideas?
>
> That's because you're thinking of syllabaries for vowel-rich,
> syllable-poor languages like Japanese, Cherokee, and perhaps Old Persian
> cuneiform.
No, not really. True, I am thinking of these, but also of others which
were used for syllables with consonant cluster onsets and at the end of
rhymes, e.g. Mycenaean Greek, which was written in the Linear B syllabary
or Cyrpriot Greek, written the Cypriot syllabary, inter_alia. Old persian
also BTW had consonant combos.

> Among the writing systems encoded as syllabaries in Unicode,
> Ethiopic has 317 syllables (not all used in any one language),
Yep - but, as you say below, it ain't a syllabary. It's an abugida; indeed
it's the one that gave us the noun abugida!

By syllabary, I mean a system of writing where each grapheme denotes a
syllable as, e.g. in the Japanase kanas, linear A (presumably) and Linear
B (certainly), Cherokee etc. All the syllabaries I know of seem to have
less than 100 graphemes.

> Yi has
> 1164 syllables (including tone information -- each tone has up to 345
> syllables), and Canadian Syllabics has 630 syllables (not all used in
> any one language).
But the latter also is not, as you acknowledge, a syllabary.

> Of these, Ethiopic and CS are actually abugidas under the covers, which
> makes them easier to learn.
Quite so, they are certainly easier to learn than syllabaries in the
proper sense, and also have less graphemes. Y.R. Chao was talking about an
ideal system of signs. I would think if we want to write in CV syllables,
an abugida is a more ideal system than a syllabary. But it cuts down the
number of graphemes. Of course, it's not clear whether by symbol, Y.r.
Chao meant simply graphemes or whether he would count the bigraphemic
signs of an abugida as different symbols.

But once we start forming syllables in an abugida, then it would be quite
simple to extend this to represent CVC syllables, and the numbers could
easily go way above the 200 limit, with the need to learn far less symbols
than 200.

> But Yi really does have 819 distinct
> and graphically unrelated syllabograms distributed over three tones;
> the fourth tone is written using an inverted breve over the characters
> used for one of the other tones.
>
I'd forgotten about this one, I must admit. Now we've gone way beyond 200!

> In a toneless language with 12 initial consonants, 5 vowels, and 3
> possible codas (zero, -n, and -ng, as in Mandarin), 180 characters
> would be required for a syllabary.
It would - but I don't think 180 distinct signs would be best. I'd think
something like the Korean Hangul would be better. But maybe this is the
sort of thing Y.R. Chao had in mind.

>  Mandarin itself would need about 400
> characters without tone information, or about 1200 with tone information.
I think the Bopomofo script with different signs for syllable onsets and
rhymes are better for Mandarin
 =========================================================================


On Monday, August 9, 2004, at 09:31 , Dirk Elzinga wrote:

[snip]
> What about a demisyllabic system which has one set of characters for
> syllable onsets, and another for syllable rhymes? You'd need fewer
> characters than for a fully fledged syllabary, but more than for an
> alphabet. Hmong has such a system with 60 onsets and 104 rhymes
> (including tonal information); at 164 symbols that is squarely within
> Y. R. Chao's preferred range.
It is indeed - but I'm not sure he'd have all the tonal stuff in an ideal
system of communication.
 =========================================================================

On Monday, August 9, 2004, at 11:20 , John Cowan wrote:
[snip]
> The bopomofo system of writing Mandarin uses an initial and 1-2
> characters for each final, plus a tone diacritic.
Yep - but even if we count the compound rhyme characters as single symbols,
  we're still way below 200 mark.

Ray
===============================================
http://home.freeuk.com/ray.brown
ray.brown@freeuk.com    (home)
raymond.brown@kingston-college.ac.uk   (work)
===============================================
"A mind which thinks at its own expense will always
interfere with language."         J.G. Hamann, 1760