Different /e/ phonemes in Siouan?

[Koontz John E]

On Thu, 14 Aug 2003, Rory M Larson wrote:
> Length may be a factor (see below).  I think I've been hearing /E/ in
> various intonational positions.  The "louse" term was originally
> volunteered with an article:
>
>   HE' ama' oNnoN'shkoNshkoN
>   ["head lice"] are making my head itch.
>
> > What other forms might exhibit one or the other of the two e's?
>
> Well, the positionals /tHE/ and /kHE/, as you note below.  I've also
> got one of my techno-terms with a definite /E/ sound on the first
> syllable.  This one does not follow an aspiration.
>
>   nE'xEtti
> ...

Let's see ... The environments in which vowels tend to lower (perceived as
laxness by English speakers) are (thinking of Eskimo and Afro-Asiatic) -
finally and next to consonants at certain articulation points, like next
to uvulars.

So Greenlandic qimmik 'dog' (hope I have this right!) sounds
like qemmik.  And I seem to recall that the standard orthography for
Greenlandic writes e and o for final i and u.  (I hope Willem will correct
this, if I am off track!)

In Arabic and Berber, aiu tend to sound like <ae>iu except next to
pharyngeals where they sound like aeo.  So when I was writing my final
report for field methods (with Tarifit Berber) and I discovered I had
written "Fatima" /fadhma/, not /f<ae>dhma/ I suspected I had misheard
/fa<dh.>ma/ and, sure enough, Arabic had /fat.ima/.  (I hope Bruce will
correct this, though I think he's wandering the Plains of North America at
the moment.)

I think that the "low next to uvular/pharyngeal" phenomenon is repeated in
the Pacific Northwest languages, and it is also essentially comparable to
the basis of "tongue-root retraction vowel harmony" systems as exhibited
in West Africa, Chukchee, Nez Perce, etc.

I believe that long vowels are another environment in which there tends to
be lowering (and/or nasalization).

Anyway, E in nexe (neghe?) might well be explained by x or gh.  In he vs.
hE it might be a matter of length, though the details are not clear to me.

It is true that Winnebago lengthens all monosyllables, and that some of
these (most, actually) revert to short when additional syllables are
added, as in compounds.  I suspect something similar might happen in OP,
and that adding an enclitic like an article might well be a "shortening"
environment.

> Is it possible that the Dakotan form is a compound of *hE + *a, where
> *a is some classifier like the /ama'/ used by our speaker?  If so, the
> /y/ would be epenthetic, but could still cause the preceding /E/ to
> shift to /e/.

That's pretty much what I've been arguing, i.e., that the -a is
essentially a partical forming independent noun stems (an absolute marker,
as they say), comparable to the -a added to s^uNka, and inducing the
epenthetic y after the e of he:  he + a => heya.  I also agree that the
initial a of OP akha and ama might well have the same historical origin as
this absolute marker -a exhibited in Dakotan.

I think the main problem with the absolute marker hypothesis in Dakotan is
that Dakotanists are used to thinking of -a in CVC nouns as epenthetic,
given the sainted memories of Boas and Deloria, who proposed that
analysis, and the "non-epenthetic" cases (-ya) are not very numerous at
present (since 1850 or so), so it's easy to treat them as an unconnected
oddity of four or five stems.  Also, regarding -a as a sort of enclitic
absolute marker requires one to deal with -e' in similar terms, and a lot
of analyses of Dakotan accent would have to be redone somewhat.  The
related inserted -a- in forms like thiyata is also uncommon and easily
regarded as an oddity of the postpositional system which is already an
array of oddities.  In short, you don't get much mileage out of this
absolute marker analysis in Dakotan and you have to rethink some "solved
problems."