*t-stems in Dakotan

[Koontz John E]

On Wed, 14 Aug 2002, Constantine Xmelnitski wrote:
> Two more words with -kik- (both seemingly obsolete):

> ikikcu "take one's own" < icu

Boas & Deloria (1941:102) mention this one, too.  I omitted it only
because it represents - I think - a different stem initial, *y or maybe
*k.

Background:  It's clear that *y becomes c^h in Dakotan when it's alone in
the syllable initial, but it seems to become just c^ in clusters. Examples
in clusters are (wi)kc^emna 'ten' cf. OP gdheb(dh)aN, apparently from
*kyepraN (or *kyewraN, if one prefers), or c^hetaN 'hawk' cf. OP gdhedaN,
from *kyetaN, or c^haphuNka 'mosquito' cf. Os laphoNke < *kraphoNke, from
*kyaphuNk-.  Sometimes Dakotan loses the initial stop.

The only reconstructed *y-stem is 'think', as in the Dakotan verb epc^a 'I
thought it' (*e-p-yE), only known in the first person in Dakotan.
Omaha-Ponca has all persons, in the first it's ebdhe (< *e-p-yE) + egaN
apparently always requiring the egaN 'like that'.  I interpret this as an
obligatory "sort of":  'I sorta thought ...'.  Since *y and *r merge to
such a great extent, it's hard to identify *y-stems in opposition to
*r-stems.  Dakotan seems to be the only environment in which *Cy is
distinguishable from *Cr in verbs, and it has mainly 'I thought'.  Maybe
i-ki-k-c^u is an additional example.

However, the base form here is apparently ic^u, not ic^hu (fide B&D), and
that tends to suggest that ic^u is from something like *iku, not *iyu.
Unfortunately, I haven't been able to identify a cognate of this stem or
root that would clarify matters.  If the source is a form *iku, then
ikikc^u is unusual in keeping the c^ when separated from the i.  I'd
expect ikikhu < *iki-h-ku < **i-ki-k-ku.

> akicikcita "hunt one's own" < akhita (not sure about aspiration in c's
> in akicikcita).

Buechel (1970:71a), right?  This gives akic^ic^ita 'hunt a thing for
another', and then akic^ikc^ita 'hunt one's own' under that, calling the
latter a possessive of akhita (aspiration clear in main entry a few pages
later).  So I suppose the first form is akic^ic^hita < *a-kiki-khita,
which would be 'to hunt for something for someone'.  I'd expect the
regular possessive of this to be akic^hita, if I'm not confused, and I
often get that way faced with Dakota ki-things.  This form is in the next
column, so I'm probably OK.  Checking a little further, it looks like the
first two forms (akic^ic^hita and akic^ikc^ita) date to Riggs.

Also, I don't know any reason to expect the ki-k- of a pleonastic double
possessive to be separated by the second dative sequence kic^i.

I wonder if the strange second one is being handled as a reduplication of
ac^hita:  a-c^hik-c^hita > a-ki-c^hik-c^hita.  Of course, there isn't any
ac^hita ... so I'm at a loss.  A hapless hapax legomenon.

> Talking about kikta this could be from ki- + kta stem (non-ablauting),
> occuring in a number of other verbs: wakta "expect", akta "respect,
> regard, give heed to", ihakta.

The semantics seem a bit of a stretch.  What's your suggestion for them?

> Here's also from Boas & Deloria's "Dakota Grammar". Page 88
>
> § 101. The use of ki- for back again
>
> "A peculiar use of ki is probably reducible to the dative ki. The
> possessive forms iki'kcu he takes his own, i.e. he takes it back 47.1,
> 48.8; kichu' he gives his own, i.e., he gives it back; ophe'kithuN he
> buys his own, i.e., he buys it back; kicha' he asks for his own, i.e.,
> he asks it back; kikta' to get up from a lying position, i.e., to be
> up again; all imply a return to a former state. The first person has
> the regular possessive form we'. The k does not change to c after e
> and i.

Isn't the root sense of ic^u 'to receive, to accept'?  Of course this is
just a different analysis of 'to take', and the *rusE stem has both senses
in the rest of MVS.  Anyway, I'd argue that 'back' here comes from the
logic of 'to accept one's own', which implies a return.  'Back' is
secondary, not primary, here.

In fact, I think that this paragraph in B&D is simply a speculation that
doesn't pan out.  It also appears to confuse the dative and possessive and
then actually goes on to discuss dative forms with similar readings.

> "A number of other forms which render the idea of return to a previous
> state are expressed by forms corresponding to the first dative ki,
> with first person waki...