Cranberries (Re: ho 'circle')

[Koontz John E]

On Wed, 14 Jul 2004, R. Rankin wrote:
> Just to muddy the waters a bit more, bear in mind that the Ocean Spray cranberry
> company has a big, big operation right up there around Wisconsin Rapids near
> Marshfield, WI.  So that particular *ho: is just as tempting. . . .

And the Winnebagos used to work as cranberry harvesters when that was done
by hand.  Still, I expect this more along the lines of a coincidence.

> > I hadn't worried too much about the "extra" -k in Hooc^aNk, because there
> > are some cases of c^aN < *htaN - right? - even if xaN... forms seem to be
> > the rule for 'big' in Winnebago-Chiwere.  But I guess technically, the -k
> > might be otherwise unprecedented in (modern) HochaNk.
>
> It's not so much an "extra K" that bothers me as much as the lack of the
> expected vowel on the end.  All WI *ht become c^ (the only real "t", actually
> [d], in WI is from *R).  The same is true for plain *t in final position.  So
> we're no worse off here than we are with the etymology of baxoje 'Ioway', but
> that's not saying a lot.

The lack of the expected vowel has a phonologically regular solution.  It
appears to me that *a > *e / [velar] ___ # in Winnebago-Chiwere, and all
*e > 0 (that's zero) / [simple C] __ # in Winnebago.  Final e, original or
from *a, is saved in Winnebago only in monosyllables and after clusters.
Chiwere (Ioway-Otoe) is a God-send because it keeps all the final *e
from final *a after velar, showing the vowel shift occurred.

So, if Winnebago has a final -e, then you are dealing with either a (C)Ce
monosyllable or -CCe.  So -ke in ho(o)c^ake has to be from -Cke,
presumably -hke here.  Or, putting it another way -ke in Winnebago
corresponds to Dakotan -kha and Dhegiha -kka and IO -khe.

It's a bit obscure without the Dakotan and Dhegiha and without knowing
about postvelaric final a becoming e, because IO sources tend to conflate
-ke and -khe and Winnebago has k for *hk/__V and *k/_#.  And this bit of
business also complicates understanding a ~ e in final position, because
it makes cases where there is genuine alternation in a vs. e between
branches harder to see.  Plenty of -e in IO and -C# in Winnebago are cases
of alternating a ~ e materializing as e in WC (and being deleted finally
in Winnebago).   But not all.  Some are plain invariant *a.

To apply the rules above in the context, PMV *htaNka 'big' > PWC *thaNke >
Wi c^aNk, but (PMVS *htahka ??? > PWC *thakhe ???) > Wi c^ake ??? (in
ho(o)c^ake 'cranberry').

By contrast a form like *s^uNk(e/a) 'dog' comes out s^uNuNk because either
*s^uNke or *s^uNka would become *s^uNke and then s^uNuNk.  I suspect,
however, based on the patterns in IO and Dhegiha that most of these words
had -e as the more or less invariant final grade in PWC (and PDh), since
there is no *a > *e after prefinal velars in Dhegiha (OP *s^aNge), and
since IO has *e in forms like *suN<ng>e.  Moreover, in non-velaric words
like *siNt- 'tail' and *yaNp- 'beaver' you always have -e (or Wi zero from
*e#), too.