*ki and -i(N)##

[Koontz John E]

On Fri, 21 May 2004, R. Rankin wrote:
> I think Blair's comment on Catawban pretty much clinches the argument,
> especially since it shows the same nasal/non-nasal allomorphs as
> Dakotan.

There are at least three morphemes that show final i ~ iN in Dakotan.  It
is intersting to see this alternation so far afield.  Besides =ki (Teton
and/or modern?) and =kiN (Santee and/or old?), I have run into:

PMV =xti(N) 'real, very'

Te ec^he'=xc^i just so'
Sa ec^he'=xc^iN ~ ec^he'=xtiN 'just so'
(In Da more or less fossilized where it occurs, I think.)

OP =xti ~ =xc^i

IO =xj^i

Wi =xj^iN

PMV =s^i(N) adversative

Te/Sa =s^ ADV; =s^=niN NEG
As/St =s^iN NEG

OP =z^i NEG

Wi =z^i 'at least'

> There definitely is a deictic {ki} in Dhegiha -- in fact, in that very
> term, dheGIha -- and it recurs in other deictic clusters.

I've never been quite sure what to make of this stem.  Dorsey gives a set
of cognate or comparable forms along the lines of Dhegiha in his Phonology
of Five Siouan Languages.  Unfortunately I haven't a clue which of 10+
boxes this xerox is in.  A few of the forms are in other sources, quoted
from here, perhaps.

OP dhe'giha 'those here, the people of this place, those on our side (in a
game)'

Ks yega 'right here', yega'ha 'hither; those here, the people of this
land'

Os ???

Qu deka' 'first (people)'

IO Ji'were 'Otoe'.  I seem to recall that this has a variant jegiwere.

Unfortunately these forms aren't regularly cognate and seem to be merely
forms Dorsey had collected for 'the ones here', not necessarily parallel
in morphology.

All the forms clearly begin with the proxmimal demonstrative (OP dhe, Ks
ye, Os dhe, Qu de, IO j^e).  On this basis the Ks form may be parallel
with the remote demonstrative forms gago 'enough', gago'ha 'over that way,
over yonder'.  These seem to involve *ka + *ko, where *ko matches OP gu.

This gu takes a bit of explaining, especially since I don't really
understand it ... OP has as its principle demonstratives dhe 'this', s^e
'that (near you)', ga 'that (yonder, not near you, maybe out of sight)'.
These parallel Teton le, he, ka, except that whereas he is the more common
of the distal demonstratives in Teton, ga may be a bit less marked than
s^e in OP, though both are common enough.  Anyway, paralleling dhe, s^e,
ga are du, s^u, gu.  These occur with motion verbs as prefixes, especially
s^u, with the gloss 'toward you'.  You also get forms like du'=akha 'this
one' or du'diha 'this way, hither', or gu'diha 'that way, yonder' in which
=di is a locative postposition and =ha is perhaps another.  It occurs in
forms with demonstratives and numbers and a few other things prefixed.
These forms have readings like '(to/from) X directions, in X places'.
Maybe it's a multiplicative?

I think the OP 'where' form agudi is essentially a- INTERROGATIVE + gu
YONDER + di LOC, so there's another -gu.  I call this the "Where Away?"
hypothesis.

You also find locatives of the form dhe=dhu, s^e=dhu and ga=dhu.
Sometimes I wonder if the du, etc., forms are old contractions of dhe=dhu,
etc.

However, one thing you don't seem to find is combinations of the first
series of demonstratives with the second.  No *dhedu (or *dedu) or *s^es^u
or *gagu.  However, I think Osage has a locative suffix =ki, and there is
certainly a locative suffix =gi in Winnebago and I think ditto in IO.
So, perhaps yega and gago are reformulated from *yegi and *gagi

In that case, I suspect Ks yegaha and gagoha are from *yegiha and *gagiha.
You also find something similar with doda, dodaha 'this way, etc.' and
goda, godaha 'yonder; that way', which ought to match OP dudi, dudiha and
gudi, gudiha.  In all of these the second vowel has /i/ replaced with /a/
(or /o/ after ga-).

In OP the form dhegiha seems to be something of a relict.  It doesn't
occur in the Dorsey texts and neither does *dhegi, or *gagi or *gagiha or
any -gi locative of any sort, as far as I am aware, except perhaps egiha,
which is variously rendered as 'headlong; into, through, under the
surface'.

The IO forms involve j^e 'this', sometimes with =gi LOC, and then we-re.
IO /we/ seems to parallel OP /he/ 'lying', as in k-he 'the lying', or
ihe=dhe 'lay, place lying' and 'suddenly' auxiliaries like thi=he 'to
approach suddenly' which are often a sequence of a motion verb plus a
positional root.  The function of IO -re here might be adverbial, but, in
any event, the IO forms differ from the Dhegiha ones in adding a
positional fomation of some sort in lieu of the Dhegiha locative suffix
=ha.  In fact, Bob lists Ks forms like yegakhaN 'those here, the people of
this land' alongside yegaha.  Insofar as I can pierce the impenetrable fog
of Kawness here, I take this to be an analog of (unattested) OP
*dhe=gi=thaN or *dhe=gi 'here' + thaN 'the (animate, standing)'.  There's
also ye'gaya 'here, in this place' in which the -ya might be analogous to
the IO -re.

I could be very wrong here, because for a student of OP to make sense of
Kaw is somewhat like using an American guide in the Scottish outback. I
once heard a funny story about some French folks and an American friend
who attempted this.  Before long the French couple were convinced that the
American did not speak the English, and while the American thought that he
did, he was certain that the Scotts did not.

> My preferred reconstruction would be as a broader deictic of some kind
> rather than as a definite article.

For the present, at least, I prefer the -gi in Dhe-gi-ha as a locative (in
Missisippi Valley), over a braide deictic.  However, it is easy to get
from demonstrative to locative and back, and, as Wes Jones has
demonstrated in several contexts, there is considerable evidence that
Siouan does this a lot.  It would be possible for the MV source of Dakotan
ki(N) to be either morphologically independent of the *=ki locative in
Mississippi Valley or the same, and it would certainly be possible for
these one or two forms in MV to derive from a single historical source as
the Tutelo and Catawba deictic elements.  Wes's term was "morphological
constellation."  Anyway, my quibble is no more than a quibble and in no
way a disagreement in principle.  Broadly speaking I agreee with Blair and
Bob.