More plural.

[Koontz John E]

On Tue, 17 Jun 2003, Rory M Larson wrote:
> Bob wrote:
> > I went and checked the Tutelo form.  There is a 3rd
> > plural suffix /-hele/ that has numerous allomorphs
> > including -hele, -hle, -hela, -hne and others.
>
> Is it in the complementary distribution with an
> *=(a)pi suffix that Dr. Helmbrecht describes for
> Winnebago?

No, but I think the second person plural enclitic is *=pV, where V might
be u (from memory).  The plurals outside MV are rather different, though
they always seem to condition the a-grade in ablaut, and to involve an
enclitic element following the stem.  Sometimes there's no enclitic with
the inclusive.  Mandan and Tutelo have different forms in different
persons.  Biloxi and Ofo both, if I recall, seem to have *ru or *tu.
Things are odder still in Crow and Hidatsa.

I don't think there's any trace of the *k of putative *krE in Winnebago,
but one might argue that -ire is an irregular outgrowth ire, from -ere
from -kErE from -*kre, i.e., with generalized loss of the initial k in
some sort of stem-enclitic juncture problem resolution.  Or one might look
at =ne in IO and conclude that Winnebago =ire was a combination of *=i and
*=re.  This would leave us with *=i to explain separately, if it wasn't
taken to be some sort of reduction of =wi.

Both the IO and Tu forms suggest some sort of nasalization apparently
missing in the other members of this (possible) set.  Tu seems to have l ~
n depending on the nasality of the following vowel.  Or in some cases in
might be the preceding vowel, since here it doesn't seem to depend on the
final vowel!  It is, of course, difficult to tell.

This might be a good point to note that final -i in certain enclitics
seems to vary between being oral and nasal across Dakotan dialects and
MV languages, e.g., xti(N), =s^i(N).

I notice we don't have a model of how enclitic pluralization (or
augmentation: [+someotherguys]) arises in Siouan languages as opposed to
proclitic or prefixed pronominals.  Is this pattern attested elsewhere?
The only likely sort of candidate that comes to mind would be some sort of
positional, and, of course, *=kre does look rather like a positional, and
positionals sometimes condition ablaut.  On the other hand, *=pi and *=tu
don't especially look like positionals.

And why ablaut?  If the a- is thought of as part of the enclitic, why do
all plural enclitics have it?  Has =*api (or *=atu or *=aX) been reduced
to =a in many places (as ..a=i reduces to a in modern OP in many contexts)
and then been supplemented by other things in various languages?  This
would parallel cases of multiple personal prefixes on the same stem, e.g.,
OP a-t-taNbe or wi-b-dhitta (from daNbe and dhitta).

Dhegiha has pluralizing positionals in its progressive (or continuative?)
forms en lieu of the plural enclitics, and some positionals
(interestingly, I think) start with a-, e.g., ama (and singular akha),
which seem potentially to be from simpler *ma and *kha, cf. =ma 'animate
obviative collective' and =khe 'inanimate or animate obviative supine'.
In addition, the inclusive of positionals like thaN 'animate obviative
standing' or dhiN 'animate obviative moving' have a sort of extra or
"locative" a, e.g., aNgathaN, showing another trace of an a-prefix on
positionals.  Whatever this a- is, it might elucidate the a of the plural.
If progressives tended to generalize into aspectually unmarked forms, a
progressive with a positional *=a=pi or *a=tu or *=a=kre might end up an
ablauting plural marker.  Of course, I still don't know why or what the
-a-.  I have at times considered that it might be some sort of
nominalizer.  Progressives might reasonably derive from noun forms,
essentially 'his going'.

Along these same lines, as Rory (and earlier, Boas) have noted, in some
contexts positionals ablaut a preceding stem, e.g., with the future, even
when the positional seems to lack an initial a-, cf. =tta=miNkhe 'I will',
or =tta=(s^)niNkhe 'you will'.  I wonder if those Hidatsa and Crow futures
(which look like just a reduced positional as I recall, with *ktE lost)
ablaut the preceding stem?

Note that Dhegiha positionals following nouns or nominalized (relative)
clauses don't seem to induce ablaut.