vowel initial stems are not glottal initial in Lak

[Koontz John E]

On Tue, 6 Apr 1999, David Rood wrote:
> 	Naturally it would behoove me to do some homework before plunging
> into this, since I haven't memorized even the regular sound laws for
> Dakotan or Siouan yet, but I jump in recklessly anyway.  I need to point
> out, in the light of the 'come' discussion with Blair et al., that there
> is a CONTRAST synchronically in Lakhota between vowel-initial and
> glottal-initial stems.  It shows up in the dual forms by contrasting uNkV
> with uNk?V.

Yes, this is true, of course, in Dakotan.  I think Bob Rankin and I are
phrasing things very carefully because there is some disagreement on the
status of the ?-stop stems in Proto-Siouan.

> 	I haven't memorized all the right examples, but 'come' is
> definitely vowel-initial (uNku pi), whereas 'use' and 'be' are both
> glottal initial (uNk?uN pi).  "i" 'to arrive going' is vowel initial, I
> think, and I'm not sure about 'shoot' or 'wear around the shoulders'
> (stems "o" and "iN", respectively).  Buechel gives "uNk?o" for shoot, and
> that sounds right, ...

The V-initial examples in Dakotan that are cited here are h-initial in
Dhegiha and Winnebago-Chiwere, except to the extent that 'come' behaves in
Dakotan and elsewhere as V-initial (?-initial).  The original *h-initial
verbs have reflexes of phV in the first person.  They may be reformulated
as synchronic V-initial in Dakotan where verb initial *h disappears ('say'
is something of an exception).  The Dakota h-stems are usually reinflected
regularly as waV, 'say' being the exception.

The original *?-initial stems are usually nasal and have first persons in
mVN, though *u (?) 'come', which is oral, has bV in older Dakotan
(wa)(hi)bu.  I seem to recall that something reasonable occurs with ?o
'shoot' in Winnebago, but I forget what.

I've suggested that *?-initial stems may actually be *V-initial, because
there's no trace of ? except in the third persons (word initially) and in
the Dakotan and Winnebago inclusives (as I remember it).  The last I knew
Bob really did think that these stems were *?-initial in Proto-Siouan, so
I think that by referring to them as ?-stems or V-stems he's being polite
as much as anything.  In my analysis I pointed out that some Dakotan
?-stems seemed to have unglotalized inclusives, at least as variants.  I
believe the source was Boas & Deloria.  I think that a lot of *?-stem
paradigms are heavily contaminated by *r-stems, e.g., the Dakotan second
persons in n.  In Dhegiha the w-stems have *r-stem second persons
and the first persons tend to have *r-stem alternatives.  The w in the
third persons (or so I argue) is epenthetic, all of these verbs have a
rounded vowel preceding or following the pronominal slot.  The standard
example is 'to interrogate':  imaNghe, i(s^)naNghe, iwaNghe.