Historical Explanation for *pi as Plural and Proximate and Nominalizer

[Rankin, Robert L]

Most of this makes really good sense to me.  I think one could do a
convincing step-by-step analysis of the semantic shifts.  Although we
only have a pluralizing morpheme with the shape */-api/ in Mississippi
Valley Siouan, there are good cognates in Crow and Hidatsa with the form
/aapaa/ 'with' in Crow and /aapi/ 'with' in Hidatsa (the Crow is
actually /aap + haa/, so the corresponding morph is just /aap-/, and
there is really no problem with the vowel correspondences).  Both have
exactly the semantic associations John predicts for the earlier (and
current?) meaning of the grammaticalized MVS cognates, but the CR and HI
forms are still independent words as far as I know.

The CSD file for this item also mentions the Dakotan particle /ob/
'with' as a possible cognate, in which case Dakotan would have developed
doublets.

John Koontz writes:
> Another way of looking at it is that if you start with proximity or
focus, as the meaning of *pi, you can easily get to plural, while it's
rather harder in the other direction.

> Of course, some of you may wonder what happens to the plural category
in this case, but notice that the *pi "plural" is only reconstructed for
Proto-Mississippi Valley.  Other branches have other plurals.  Also, it
is rather weird to have the plural marker off at the other end of the
verb, working idependently of the pronominals at the front, if it really
has anything closely connected to do with them.  But it is maybe not so
weird to have a focus marker over there - a marker that does something
independent of the pronominals, though involving them, and only later
becomes a plural.