Information / nouns vs. verbs

[Koontz John E]

On Mon, 13 Jan 2003, Jan Ullrich wrote:
> Oh, and I should add that in the same collection of texts (Deloria's Dakota
> texts) there are 5 occurences of
>
> (huNku / thaNkaku) thi kiN ekta (hi / i / ya)

It looks like Jan confirms the existence of the thi kiN ekta vs. thathipi
alternatives back to the 1800s.  It occurs to me to wonder if some people
liked one and some the other even then, of if there is a stylistic or
semantic distinction.

Unless I've missed something, nobody has confirmed (or explicitly
disconfirmed) ?thathipipi for 'their house(s)'.  I was pretty careful to
refer to pi in pluralizing "or other"  uses, glossing over the issue of
whether the other uses reflected homophony or polysemy, because I knew
someone would bring up the issue in those terms.  I have to admit that my
instinct is that the various uses of pi (or *pi) in Dakotan and Dhegiha
are polysemy, and that, in spite of rather varied uses, "the language"
would tend to resist placing two instances of pi in a row, because it does
when two pluralizing uses are possible.  What exactly would be the
structure of a speaker's grammar or other motivation in this cases I'm not
sure I would care to say, but in historical terms it would be an
inheritence of this reluctance to use two pluralizing pi's in a row.

Apart from Connie's ahaNhepipi example, I've not seen doubled "pi" markers
(in any sense).  I think that's based on haNhe(tu), right?  I couldn't
figure out why the pi was doubled there.  Does anyone have any further
elucidation of that example?