Re: Conlanging with constraints

A lot of my conlanging (since Kash, begun in 1976) has been influenced by my
work in Austronesian (proto and modern) languages. For ex., the constraints
against w/u and y/i in either order; occasional dissimilation of r--r,
assimilation of l--r~r--l sequences etc.

Kash was developed without a proto-stage, but I have vague ideas, and could
probably produce one. A lot of "exceptions" will have to be explained away
as borrowings between related languages/dialects. The only oddity is that a
single species, several hundred million strong, all speak languages of the
same family. That I think will take some explaining.........

Gwr started out modelled on Chinese/Viet-- isolating, monosyllabic and
tonal. At first I created forms out the thin air, knowing all along that
there had to be rules, but... Once I actually got around to formalizing the
rules, it turned out that maybe 1/3 of the original forms were wrong or
impossible one way or another. Again, I have a single species :: single
language family, but that will be easier to correct (I think....someday...).

Prevli has started off with several prime criteria: 1. lotsa metathesis 2.
resulting allophony 3. some V-harmony ~more allophony 4. VSO 5. compound
subj+obj pronouns with transitive vbs. 6. Realis/irrealis distinction (which
I'm mainly guessing at...) 7. extensive required use of passive voice (like
Indonesian, but more complicated).  I originially played with ergativity,
and lots of noun cases, but decided it was getting too much like Basque. But
it's nowhere near the complexity of Salishan lgs., after all I want to be
able to pronounce things.