Correction on sun/moon 'RE: Colors in Dakota'

[Koontz John E]

On Wed, 26 Mar 2003, Koontz John E wrote:
> ... Though in principle *R and *W appear as n and m only and *r and *w
> appears as n/m or dh/w depending the nasality of the next vowel, there
> are some exceptions, in the sense that some *R + V sequences appear in
> Dhegiha or sometimes just Omaha-Ponca as nVN.  The 'sun/moon' word is
> one of the former.

Oops, the root shape here is not *RV, but *WV, cf. Dakotan wi-, IO bi (I
think?), appearing in OP as mi(N), though I think 'moon' has innovated n
(niaNba for expected miaNba) in OP. It's not clear in OP that the m is a
problem, but the root also appears in Osage as mi(N), and here we would
expect pi, so it seems that Proto-Dhegiha had *wiN (or *WiN?)
irregularly instead of expected *Wi.

Confusingly, R and W represent actual sounds, while V means 'any vowel',
and N indicates nasalization.  Three different expedient uses of
capitalization!

The R and W symbols are particularly awkward expedients for representing
unknown sounds in the r-to-t and w-to-p ranges.  We usually read R and W
as "funny r" and "funny w," with "funny" in the sense 'odd, unusual'.
(This is the Siouanist answer to proto-Algonquian "theta.")  The
correspondences involved are quite regular, in spite of this particular
exception, but seem to sit half way between the sets for *r and *t or *w
and *p in terms of reflexes.  *R and *W were recognized as sets by Dorsey,
but get mixed with *r and *w in Wolff.  I believe Kaufman had sorted them
out again, and recopnstructed *?r and *?w.  I forget how Matthews treated
them.  I think he may have followed Wolff in this, trying to provide a
contextual account for the exceptions.