Archaic A1 p- in Dakotan

[Robert L. Rankin]

> Actually, I would like to modestly remind those interested in this issue
> that I wrang two pages out of these forms in:
>
>   Koontz, John E.  1985.  A syncopating conjugation *k-stem in Lakota.
>       IJAL 51.4:  483-4.

Sorry, John, ...just goes to show how my memory is doing these days!

> I'm not sure that the ?*wke 'turtle' form is relevant, but... Assuming
> Mandan pke matches Da kheya and OP kke, it looks like *pku should yield
> ?khu regularly in Da, presumably via *hku, like other hypothetical
> preaspirates that yield aspirates in Da and IO and preaspirates or tense
> stops throughout Dh and voiceless stops in Wi (Ho).

It's more than just hypothetical.  Root mah- 'earth' with the noun
formative -ka (i.e., mah-ka) gives Dakota makha, so hC > Ch was definitely
a Dakotan sound change.  There are a couple of other parallel cases in the
CSD maybe.

'Turtle', like many other animal names, presumably had a wa- or wi- prefix
at one time (I make a case for wi- in a couple of papers).  Like the 1st
person pronominal, it lost its vowel in 'turtle, bison cow, otter', etc.
giving a secondary cluster, again like the 1st person (if we're right
about the underlying 1st person).

Dheghiha has both patterns for secondary pk- clusters.  Usually they go
through the changes:

*w-k > pk > kp > hp > pp (like in vertitive ku we've been discussing).
In Dakotan the last stage is different: hp > ph as we've seen.

'Want' in proto-Dhegiha is ko~ra and it is doubly conjugated. 'I want',
*w-ko~-w-ra, where both w's are 1st person markers, you get 1st sg.
kko~bla in Kansa (with exact analogs in the other Dhegiha lgs.)  So this
verb for some reason apparently didn't undergo the metathesis that the
other verbs in its class did:

*w-k > pk > hk > kk

It remains unclear what the chronology of morphological (as opposed to
phonological) changes was with the noun prefix wi- or wa- and the 1st
person wa-.

Bob