Stem 'to come' (was Re: Archaic A1 p- in Dakotan)

[Robert L. Rankin]

[Rood]:
> I don't understand how the analogy would work.

[Rudes]:
> extent it is relevant, the corresponding stem for 'arrive, come' in
> Catawba is -uu?-, not -huu?- as sometimes listed elsewhere (e.g.,
> Siebert 1945, which contains a number of underanalyzed forms).  The
> conjugation is: c^uu?- 'I come', yuu?- 'you come', huu?- 'he comes'
> (where h- is the 3rd singular marker), etc.

[Rankin]:
> That's an interesting observation in light of the fact that 'come' is a
> verb for which the sound correspondences are irregular.  Several
> languages treat it as {hu:}, but Dakotan (perhaps others) has {?u}.  And
> there are also interesting correspondences between syllable-initial and
> syllable-final glottalization across Siouan.  If the Catawban 3rd person
> can definitively be shown to have been h- and Siebert proved wrong on
> this, we would have a good analogical model to explain the Siouan
> reflexes.

I guess what I meant here was that, at present, we can't account for why
some languages have reflexes of a verb *hu while others, like Dakota, seem
to have reflexes of *?u or just *u.  Or they may have a mix.  It seemed to
me that either the *h- was there in the proto form or it wasn't.  If it
was a 3rd person allomorph in Proto-Siouan-Catawban, as Blair's data
suggest, then it's the 3rd person that is providing the model in the
languages/persons where h- occurs.  In verb paradigms, if analogical
extension occurs, it is usually the 3rd person form that provides the
model adopted in other persons, e.g., English, where 3rd person -s is
generalized for some speakers "I sees, you sees, we sees," etc.

Bob