Morae (was: Re: Lurkers, poetic forms), and more

Dirk Elzinga wrote:
>It is now generally recognized that syllable weight can be based
>on the length of the vowel as well as on the presence of a coda
>consonant (creating a closed syllable); languages will vary with
>respect to this option>
(snip)
>
This came up years back when I was analyzing Buginese and its relatives.
Background:  Proto Austronesian had 4 vowels, *a,i,u and "e" (= schwa --
yes, I know that's confusing, but it's been that way for years;  a certain
journal was too cheap, apparently, to invest in a type-font with the correct
symbol), and canonic form CV(N)CV(C).  Interestingly, CV(N)Ce# seems to have
been impermissible., Placement of proto-accent unknown.  But in a wide
variety of langs. of Indonesia that retain the *e and the canonic form, the
penult is stressed unless it contains *e, in which case the ultima in
stressed. Ml. BAsah 'wet',  beSAR 'big' etc.
The ancestor of Buginese et al. apparently developed fixed penult stress.
Result: geminate consonants following schwa (or its modern reflex).  So AN
*telu 'three' > Bug. TELlu.  It's entirely automatic following schwa;
unfortunately, there are geminate consonants from other sources, so by some
analyses, geminates are "phonemic".

Further on geminates (I can't find the original posting):  In this group of
langs., all cons. can be geminated except /w/ and /y/ (IPA  j).  When first
exposed to Buginese in a field methods class, we all pressed the informant
for /ww/ and /yy/-- he finally came up with two words which we all heard as:
[sew:a] 'one' and [iy:a] '3d sing. pron.'.  Since the instructor had
sequestered the dictionary, we couldn't check; long after, I found "sewua"
was the correct form for 'one'.  The pronoun is indeed transliterated "iyya"
by Matthes, but historically is prefix i- + i(y)a.  So much for geminate w
and y.
From what I've seen of Arabic in transliteration, /ww/ and /yy/ are possible
there.