Irregular "to eat" in Dakotan

[Koontz John E]

On Tue, 21 Aug 2001 BARudes at aol.com wrote:

> The root of the Catawba verb to eat is -raN-, which appears in Woccon with
> the regular replacement of /aN/ by /a:/ as -ra:-.  The -te in Eraute is a
> modal suffix.  -raN- itself is a mutating verb.  More often than not, it
> cooccurs with the mutating instrumental ru:# by hand, in such forms as
> du:raNre:  one eats.

This means that *rut(e), maybe *ut(e), is attested only in Siouan, and
might conceivably involve the *ru instrumental, if the r is organic to the
stem.  The Woccon and Catawba forms actually look more like the first part
of *raNpyE, even though the phonotactics of that suggest an analysis of
*raNp-yE.

Although it would be, I think, a unique instance of this pattern of
suppletion, perhaps the Siouan 'eat' stem is just *t(e) (or *tE), and the
third person has the instrumental, while the first and second (and
sometimes the inclusive) lack it.  In this case the occurrence of
instrumental forms instead of non-instrumental forms in the first and
second, etc., persons would be an old alternative, rather than
regularization.  I'm not sure but what Blair was suggesting as much.

JEK