Re: Grammatical metathesis

This thread prompted me to re-read one of my sources on Timorese-- Hein
Steinhauer 1996, "Morphemic metathesis in Dawanese (Timor)", in: Papers in
Austronesian Ling. No 3, Pacific Linguistics A-84 (publ by the ANU).

There is not much in the bibliography-- several studies by Indonesians,
which he deprecates, and mentions of met. in languages like Sierra Miwok
("related to the patterning of short and long syllables within the word"),
Rotuman; mentions of other Austronesian langs. Ririo (Solomon Is.), Kwara'ae
(?), Rowa (?); and a ref to Russell Ultan 1978, "A typological view of
metathesis" in: Universals of Human Language, ed. Jos. Greenberg, Stanford
Univ.Press.

Further exs. to show that metathesis is at least partly motivated by
phonological/phrase conditions, showing the preference for ...VCCV...:

mata 'eye'+bilu 'blue' > mat bilu vs.
mata + ?naek 'big' > mat ?naek (1)

noni 'money'+ kase 'foreign' > noin kase vs.
noni + mnatu 'gold' > noni mnatu
===============================
(1) The /?/ on ?naek (and other forms) is odd, and must be a prefix; neither
of my other sources indicate it. It would certainly be easy to overlook, but
it's surprising that Middelkoop, the Dutch missionary/translator in Timor
for many years, didn't spot it. Perhaps it's dialectal.