Pitch Accent

[Koontz John E]

On Mon, 26 Jan 2004, Koontz John E wrote:
> This is precisely what I noticed in Omaha-Ponca.  Falling pitch on CV## in
> CV' and CVCV' words, e.g., kke' 'turtle' or ttabe' 'ball', and in
> CV'-V-C...  words (typically inflected) across morpheme sequences, e.g.,
> a'-a-gdhiN 'I sat on it' or mu'-a-se 'I cut it off by shooting' (more like
> [mwaase] with falling on aa).  You don't get falls in bisyllables with
> initial accent (CV'CV) in longer words with second syllable accent,
> typically inflected (CV-CV'CV), or in words with initial accent like
> a'gdhiN 'he sat on it' (CV'C...) or mu'se 'he cut it by shooting'.  If we
> write long vowels as two moras, each V, and assume that initial accent
> requires an initial long vowel, then these work out as CV(V)^, CVCV^,
> (C)VV'-V(^)-C..., CVV'CV, CV-CV'CV, (C)VV'-V(^)-C...

I should add that theorizing of this nature is obviously secondary to
listening to vowel length vs. accent, however manifested.

Also, I've noticed that Ken Miner's Winnebago data treat oppositions like
OP aa'agdhiN 'I sat on' vs. aa'gdhiN 'he sat on' (V^CV vs. V'CV, with ^
for falling pitch) in terms of VVCV' vs. VCV' (after the accentual shift)
and also that V1V1 + V2 is always shown as reducing to V1V2 there.

Also, though it's not specifically relevant to your Kaw investigations, it
looks to me as if each of the branches of MVS has to have distinct
adjustments of length and accentual pattern.  I think only Dhegiha treats
some locatives as long (even when not fused with wa) and Dakotan and the
others have distinct ways of treating *CV(V)'Ce or accent in compounds,
etc.  Also, it looks like Dhegiha supports accent on enclitics in some
contexts.