Crow-Hidatsa Willow (Re: etymology of MANDAN)

[Koontz John E]

Since Bob was a bit dubious of the interpretation I offered for Hidatsa
(and Crow) 'willow', let me offer a little support.  Basically, I
interpreted the Hidatsa form as 'yellow bark tree' and the Crow form as
'yellow tree' remodelled as 'yellow water'.  Or, actually, both forms are
uninterpretable wholes that seem to be traceable to these interpetations.
The restructuring in Crow demonstrates that the forms are not
interpretable as they stand today.  I'm deleting all comparisons to
Hidatsa, because there's nothing there that's quite a different matter,
and still opaque to me.

On Thu, 19 Aug 1999, Koontz John E wrote:

Hidatsa > mira  ha'a ci
Crow    > bilii      chi
> The real problem here is not the *wiri 'water' vs. *wira 'wood', though
> that's a serious enough issue.  The real (or bigger) problem is haatsi vs.
> chi. ...
>
> It occurred to me, however, that MVS might elucidate this situation, and I
> think it does.  ...

> Osage for 'yellow willow' is dhuxe-zi < *ruxe-zi(hi), where
> *ruxe is 'willow' and *zi(hi) is 'yellow'.  The (hV) is an extension
> common in MVS color terms, but not universal.
>
> I think that *zi(hi) would appear regularly as *chi in Crow, *tsi in
> Hidatsa.  What we do get is shiile in Crow and tsiri in Hidatsa, with a
> different extension *-re.  The Crow form looks, furthermore, like it's
> from the s^-fricative grade. Since fricative gradation is common in color
> terms in Siouan, this isn't a big problem.  Presumably *chii(le) < *zi(re)
> was replaced by shii(le) < *z^i(re) at some point.  In any event, this
> makes the forms look like
>
Hidatsa > mira  ha'a tsi
Crow    > bilii      chi
        > ???   ???  yellow
>
> which definitely recalls the Osage formulation.  To me it suggests,
> furthermore, that the mira 'wood' is correct, and that the puzzling haa
> element is related to PS *ha 'skin, hide, bark', even though this is not
> attested in Crow or Hidatsa (or Mandan) as far as I know.  Thus:
>
> mira   ha'a  tsi
> bilii        chi
> wood  (bark) yellow
>
> So it's Crow that has modified the vowel of the first part in this
> interpretation.

This etymology is phonologically and morphologically regular except for:

- Lack of CH *-re extension on color terms.
- Crow has the *s (*z) grade of 'yellow' actually expected, instead of
  the *s^ (*z^) grade it normally substitutes as the free term for
  'yellow'.
- I don't think *ha 'bark' is otherwise attested in Crow-Hisatsa or
  Mandan.
- Crow has changed bilaa 'wood' to bilii 'water', presumably by fiddling
  with the phonology of an uninterpretable form rather than by any kind of
  semantic reanalysis, though an association of willows with water might
  be at work (as Randy suggested).

I don't think this is really any more problematic than interpreting Mandan
koxaNte 'corn kernels' as ko(r) 'squash' (attested) + xaNte 'grass' (not
attested, but reconstructable in Miss. Valley), or haNxurar 'bat' as haN
'night' (not attested, but reconstructable in Miss. Valley) + xura(r)
'eagle' (not attested, but reconstructable in Miss. Valley).  Naturally,
such analyses would be stronger if the constituents were all attested in
the languages where the compounds exist, but if the morphosyntax is
correct, the correspondences regular (or we are willing to admit that the
irregularity is secondary), and the sense reasonable, I think the
etymologies can be considered fairly reasonable.

Note that in the case of 'willow' and 'corn kernels' we have also the
suggestive circumstance that the forms in question parallel constructions
found in other Siouan languages, e.g., Osage 'willow + yellow', and the
general pattern in Mississippi Valley of deriving 'corn' terms from
'squash' terms.