Quapaw wa-a- '1pl patient'.

[R. Rankin]

Rory asks about Quapaw and Kaw.  My Quapaw conjugation data are more easily
accessed than the Kaw right now, but all data are from Dorsey, who failed to
record vowel length.  Here are some paradigms.  There are distinct reflexes of
the instrumental i-.

be thirsty, dry
1sg   aNda'Nbize
2sg   i'dibize
3sg     i'bize
1pl   we'bizawe   (we'- from *wa + i' + a-)

be anxious to do sthg
1sg  aNna'Nhitta
2sg  i'dihitta
3sg    i'hitta
1pl  we'hittawe

swell up
1sg  aNna'Npa
2sg  i'dipa
3sg    i'pa
1pl  we'pawe

And the final paradigm, o'z^eda 'be weary', which clearly shows that there are
two /a/ vowels associated with the 1st pl, one preceding and one following the
instrumental o-.

1sg  o'Nz^eda    (o'N- from *o + a'N)
2sg  odi'z^eda
3sg     oz^e'da
1pl  o'waz^edawe  (o'wa- from *wa + o' + a-)

It is important in the last paradigm to note carefully the *underlying sequence
of morphs.  This can be determined from the accentual pattern of the form.
Recall that MVS *wa+o'  >  accented o'.  And we have accented o'- here, which
shows that the wa- portion of the 1st pl pronominal in fact precedes the o'-.
Then the phonological sequence of o- + a- (where -a- is the second /a/ of wa-a-
'1st pl pat.) entails an epenthetic /w/ between the two vowels.  It is easy to
misanalyze this form as locative o- followed by pronominal wa-(a)-, but this is
wrong, and the accentual pattern proves it.

That's about it for now.  We can see that there is evidence that there are two
/a/'s in the 1st pl. patient in Quapaw, just as in the other languages of the
subgroup.  We can also see that locatives occur inserted between the two /a/'s
of wa-a-.

None of this substitutes for making the necessary measurements to determine just
how vowel length asserts itself in these paradigms and those with 3pl objects.

FWIW

Bob