Seediq Inflection

I was at the university's linguistic library today and I found a=20
language with quite an interesting phonotactic method of=20
inflectional marking. It uses a manipulation of stem shape for=20
inflectional marking in almost (but not quite) the same way as in=20
Arabic. This language is called Seediq, an Austronesian language=20
spoken by some aborigines of Taiwan. Maybe one of you guys may=20
want to be inspired.

I'll try describing it below from memory:
--------------------------------------------------

Its consonantal inventory is:
p   t   k   q   ?
b   d   g
    c
    s           h
m   n   N
   l r
w       y

Vowels:
i       u
  e   o
    a

The underlying root form is: CVCVC, with stress is regularly on the=20
penultimate (next-to-last) syllable. Both CVCe and CVCo are possible=20
in surface representations, but are interpreted as CVCay and CVCaw=20
respectively.

What's interesting is the restrictions imposed on the surface forms.=20
Vowels in the final unstressed syllables can only be either; <i>, <a>,=20
or <u>. Thus, forms like *CVCeC and *CVCoC are non-existent in the=20
surface forms, but are possible in the underlying forms. Furthermore,=20
word/syllable-final consonants can only be; <t>, <c>, <s>, <n>, <l>,=20
<r>, <w>, <y>, <q>, or <h>. Any, syllables found before the stressed=20
syllable only has [u] as the vowel, but this is generally left=20
unrepresented in the orthography.

In order to keep in line with the restrictions, gradation rules=20
exist for word/syllable-final consonants:

p,k,? -> t
b,d   -> c
g     -> w
m,N   -> n

Similar for the vowel in a word-final syllable:
e     -> a
o     -> u

Here are some examples of inflection (using made-up roots):
     UNDERLYING                SURFACE
     /qitah/                =3D  <qitah>
          + -n- infix       =3D  <qnitah>
          + -an suffix      =3D  <qtahan>

     /tulek/                =3D  <tulat>
          + -n- infix       =3D  <tnulat>
          + -an suffix      =3D  <tlatan>

     /bugaw/                =3D  <bugo>
          + -n- infix       =3D  <bnugo>
          + -an suffix      =3D  <bgawan>

     /paleg/                =3D  <palo>
          + -n- infix       =3D  <pnalo>
          + -an suffix      =3D  <plegan>

---------------------------------------------------
Pretty neat, eh?

This sort of stem manipulation is perhaps doable in Boreanesian.=20
I'm inspired, so its back to the drawing table. ;)

-kristian- 8)