butterfly

[R. Rankin]

> Reduction of medial lax stops is pretty standard
in Osage, Kaw, and, I
> think, Quapaw, especially in compounds.  I think
the progression is sape >
> sae > sa.  I think this is old, since some of
the earliest known forms
> probably of Osage origin are ethnonyms like
Paniassa, possibly ppadhiN
> (w)asa 'Black (Painted?) Pawnee' = 'Wichita'.

Medial lax stops in Osage seem to do this, at
least with the last generation of speakers, but
I'm not so sure what you're describing here is an
example of that.  I don't find it in Kaw or
Quapaw.  There is a small set of stative stems
that have no "root extension" when used as the
second member of compounds.  These typically
include
ttaN    augmentative (rather than 'big')
z^iN    diminutive  (rather than 'little'),

and colors like
tto     blue
sa      black
s^a    dark
zi       yellow

I think this phenomenon may be morphological
rather than phonological, with the Osage VCV > VV
a separate, and much more recent, occurrence.  In
OS, the two vowels are preserved as distinct from
one another if they are different vowels.

You fnd the truncated form of these verbs in
languages like Tutelo, Ofo, etc. also, suggesting
that these particular items have participated in
the only-part-time relationship with their
respective root extensions pretty much all along.
I tend to look upon the languages that show only
the forms with the RE's as having undergone
analogical leveling.  My view on this sort of
solidified while working through a number of
Carolyn's Osage examples recently.  But it's hard
to be certain . . . .

Bob