Kaw and Osage 'rain', 'stand' and 'boy'

[Koontz John E]

On Sun, 31 Oct 2004, Rory M Larson wrote:
> CQ:  In Osage, 'to stand' has a long nasal a:N.  'To rain', on the other
> hand, is either ni'z^u or nu'z^u, where the first syllable vowel is not
> long, or at least is not so long as the a:N in 'to stand'. I don't write
> 'rain' with a long vowel.
>
> RLR:  Ditto in Kansa. naaNz^iN' 'stand' (1st syll. long, second
> accented) nuz^u' 'rain' (1st syll short, second accented, /u/ is a
> front-rounded V as in French)
>
> Hmm.  Well, I asked our other Omaha speaker on Friday, and after mulling
> it over for a while, she denied that the first syllable in naNz^iN',
> 'stand', was long.  I wonder how Ponka handles it?

I hope the non-Dhegihanists and non-comparativists on the list will excuse
us as we wrestle with this single form and what must seem like something
we ought to know by now!  Unfortunately, I have to admit that I don't!
It's really important for those of us working with Omaha - and in a larger
sense Dhegiha - to understand how the phonology of forms like this works!

I have the impression that Bob is saying that these forms in Kaw, modulo
vowels, contrast CVVCV' and CVCV', whereas Carolyn is mentioning CVVCV
(CVV'CV?) and CV'CV.  Rory seems to have CVVCV' and CVCV'.  I decided on
Friday that I couldn't rely on my memory for this and had to check the
data - my recordings, Dorsey, etc.  Unfortunately I never got to email
over the weekend and so forgot the whole question!

It strikes me that it would be a bit odd for there to be second syllable
accent if this was the third mora.  Of course, I'm not positive that
"second mora"  accent survives the discovery of vowel length, but I
suspect there is something in it, as it works so well in accounting for
patterns in both Dakotan and Winnebago, albeit in the former case the rule
is somewhat discretized by the absence of vowel length, and in the latter
case it only appears if you reverse a couple of major sound changes, "in
underlying forms."  I also meant to check Miner's Winnebago version this
weekend!

Just out of curiosity, Rory - or anyone, for that matter - what is the
pitch contour of naNaNz^iN in Omaha or anywhere else, not necessarily in
Dhegiha?  I'm wondering, frankly, if the pitch contour isn't

  H H   L
naNaNz^iN

Is there any trace of falling pitch on the final syllable?

  H H   HL
naNaNz^iN

I always thought I tended to hear some bisyllabic H L forms as finally
accented, though I think that they should really be classed as initially
accented.  I'd expect pattern one above with initial accent and pattern
two with final accent.  The fall might be absent, however, in cases where
something additional followed, e.g.,

  H H   H? H?    L
naNaNz^iN=i(N)=the

It might also be difficult to judge length, if the basis for the judgement
was relative length and the form in question had an underlying
plural/proximate =i that was deleted finally with, say, compensatory
lengthening of the final syllable.  This might be controlled for by
looking at a form that moved the plural/proximate slot to a later
position, e.g., naN(aN)z^iN=naN=(i) or didn't require it, e.g.,
naN(aN)z^iN=akHa.  And, of course, in =i=the, the =i should resurface.
Still another way of controlling which is not an option here is that
ablauting final reveals whether the form under consideration has =i by
changing -e to -a.  If you ask for a form like "to walk" or "walking,"
it's not necessarily clear what you'd get in a language that lacks
infinitives (?) and/or may not have a standard citation form.

In fact, it might also be worth looking at your perception of length in a
cross-section of forms:  first, second, third, inclusive and also
imperative and "embedded" under a governing predicate.  The personal forms
won't work with 'rain', of course!  But I think we tend to make the
simplifying assumption that length is constant.  A good deal might
actually depend on factors like inflection or foot structure.