Re: Onset/Rhyme Scripts

John Vertical, On 08/07/2006 11:40:
> Hello List,
> With the script discussion goïng on currently, I thought I'd raise a
> question: does anyone have / kno' scripts based on two sets of caracters
> - onsets, and rhymes? Those seem to be vanishingly rare compared to
> alphabetic or syllabic systems; in fact, I don't think I've ever come
> across any others than the Chinese zhuyin system. Yet surely such a
> system would, for many languages, be more practical than a plain
> syllabary or abugida; in some cases, perhaps even more practical than a
> plain alphabet?
>
> The real-world rarity is probably due to historical reasons, but I'd
> expect to see at least some around in conscripts...
This principle, which I've called 'hemisyllabic', is used by Livagian native
script, at least for abbreviatory purposes, rather in the way that in English
<x> and <q> sort of serve as abbreviations for <cs> and <cw>.

I once asked the same question you have, about hemisyllabaries in natlang
scripts, & had useful replies from Matt Pearson and John Cowan. I'll quote two
of John Cowan's, one of which quotes Matt's:

*****************
Matthew.Pearson scripsit:
> Most South Asian and Southeast Asian scripts approach this, though they
> don't quite make it all the way. In these scripts, letters represent
> consonants or combinations of consonants, with tautosyllabic vowels and
> diphthongs represented by diacritics above, below, and/or next to the
> consonant letters (Tolkein's tengwar operates on the same principle).
Except for Ethiopic script, one vowel has no representation (the
"inherent vowel").

> However, as far as I know, many (most? all?) of
> the languages which use such scripts do allow coda consonants, as well
> as onset consonant clusters not represented by a single letter.
Buginese doesn't write coda consonants -- you just have to infer them.

However, Bopomofo is nearly hemisyllabic: one letter for each legal
Mandarin initial and one or two for each legal rhyme. (A semivowel in
the onset is a separate character, but a semivowel or nasal in the coda
is not.)

********************
Here's more detail on Bopomofo, or rather the subset of it used to write
standard Mandarin:

Mandarin is normally analyzed as having 21 initial consonants plus zero.
You can reduce this by three if you are willing to play tricks elsewhere
in the system, but both Pinyin and bopomofo recognize 21. There are 35
finals in the system, of which /i/ is in complementary distribution with
zero (both are written -i in Pinyin). Each non-zero final is of the form
(S)V(F), where S is /j/ /w/ /H/, V is /a/ /e/ /i/ /o/ /u/ /y/, and F is
/j/ /w/ /n/ /N/, with complex co-occurence constraints.

Bopomofo has 21 letters for each of the 21 initial consonants. It has 16
letters for each of the possible V(F) parts of the finals; the S part,
if present, is written with the corresponding V letter. A zero initial
or final is not written.
**********************
--And.