Dhegiha Plurals and Proximates

[Rory M Larson]

JEK said:
>> > Anyway, there's no doubt that both the true Osage
>> > plural/proximate and the Omaha plural/proximate are
>> > cognate with the Dakotan plural.  They simply have
>> > different modern patterns of allomorphy, and LaFlesche's
>> > Osage Dictionary, for unknown reasons, but undoubtedly
>> > reflecting the Omaha ethnicity of its compiler, has the
>> > Omaha pattern.  This does not seem to reflect anything
>> > about actual Osage usage at any point.
>>
>> The "name defense" and the "song defense" both rested
>> on the assumption that =i was a recent derivation from
>> =bi.  If these riders go away, then I'm not sure that
>> that assumption itself is necessary either.

> Mississippi Valley Siouan *=pi
>
> Te   OP                   Os                    IO   Wi
>
> =pi  =i ~ =bi ~ =b(=az^i) =pi ~ =p(=a) ~ =p(=e) =wi  =wi
>
> All of these forms condition the a-grade of ablaut.
> All occur with first, second, and third person as well
> as imperative plurals.  It has been mentioned that =pi
> can be considerably reduced in Teton fas speech, and in
> some other Dakotan dialects I believe it is sometimes
> reduced fairly regularly, though I'm not positive I
> recall the details correctly.  In Dhegiha the alternants
> also occur with certain third person singulars, as has
> been seen.  The =i alternant is the less marked form in
> OP.  The =bi occurs in the marked context or contexts
> under discussion.  The loss of the final vowel in
> Dhegiha is essentially due to elision of i before
> another following vowel.  I think all Dhegiha languages
> lose final i before the negative =(a)z^i, in favor of
> the a.  Osage happens to have male and female
> declaratives =a and =e that provide quite frequent
> contexts for loss of i.  In fact, I think =p=e is far
> more common than =pi (because most of the last speakers
> are female).

> Why Omaha-Ponca so regularly lost the b instead of the i,
> and came to retain or restore b in a grammatically
> conditioned context is not clear to me, but loss of the b
> (or *p) is not especially exceptional in Siouan
> developments of =pi.  The w in IO and Wi is the regular
> development of *p in those languages.

> Ironically, most modern Omaha speakers have lost =i itself
> in the environment ...a=i##, which is simply ...a##.
> If there is any change to -a, like lengthening or a if
> there is a voiceless =i in this context, I missed it.

This is what we're finding with our speakers too.  They
don't have much patience with me trying to stick an =i
in after the final vowel.  Needless to say, I've been
very frustrated in my attempts to get them to illuminate
the finer grammatical points of =i and =bi!


> I might add that occasional real Osage forms occur
> in the Dictionary itself, for example in the appended
> text, or in entries like a-xo-ba-bi 'inviolable'.

>> In OP, we are fortunate enough to have a very rich
>> literature recorded from fully fluent speakers in the
>> late 19th century, which provides a wide variety of
>> grammatical usages.  In that language, =i and =bi
>> are radically distinct elements which contrast with
>> each other, while simultaneously signalling several
>> different semantic implications, not just plurality.
>> Given how deeply and subtly embedded these particles
>> are in 19th century OP, can we really be sure that
>> related languages like Osage did not have a comparable
>> contrastive pair?

> I really don't see any evidence to the contrary.
> Osage reaction to the use of OP =i in the LaFlesche
> dictionary is pretty decisive.  "This is not Osage!"
> about sums it up.  I admit I have this information
> at second hand from Carolyn Quintero, and may have
> misunderstood her, and that many other factors,
> including such simple things as writing ptk as bdg
> and using anachronistic and rather florid definitions
> enter in to this, but I think there is no doubt on
> that score . In addition, though I have not gone into
> it, my understanding is that Kaw has forms similar
> to Osage (i.e., mostly =b(=)e, the last speaker being,
> again, female).  Quapaw has =wi or =we, if I recall,
> but my unpacking hasn't progressed far enough to
> provide me with something that I can look that up in!

I certainly wasn't calling Carolyn's or your understanding
of modern Osage into question.  My suggestion was that the
language may have changed in the past century between
LaFlesche's time and our own, with some of the less common
and more subtle grammatical usages simply falling out.
Given that independent =i has pretty well disappeared
from modern Omaha, I can get a reaction almost as
decisively against its being Omaha from our speakers,
but we know it not only existed, but was rather ubiquitous,
in the 19th century.

Also, I don't doubt that *pi forms have shifted to
phonological [i] in some contexts in various MVS
languages.  I'm just not convinced that that is
necessarily the way that OP =i originated.  [i] is
about as short and undistinctive as a word can be,
and any number of words could easily reduce to it,
including some modal particle other than *pi.

What I'm suggesting is that the =i and =bi particles
go back separately at least to proto-Dhegiha, if not
to MVS.  Their exact usage and meaning would be
somewhat different from what we know today, and would
have evolved in parallel or in contrast in the various
daughter languages.

MVS:        *=pi        *=i## (?)

Dhegiha:    *=pi        *=i

In most languages, the *=pi form came to signal
plurality as its most important function, and
spread to the exclusion or near-exclusion of *=i.
In OP, however, =i was retained as a reality
signaller, in contrast to =bi, which signalled
allegedness.  In Osage and other Dhegiha languages,
=i was retained, but the circumstances for its
usage was very rare.

19th century:     OP: =i and =bi
                  Osage: mostly =pi, rare =i

20th century:     Omaha: loose =i and =bi dropped
                  Osage: =i dropped, =pi retained for plural

So Francis LaFlesche, an Omaha, visits the Osage
in the early 20th century.  As a native speaker
of a closely cognate tongue, he approaches Osage
through that language.  He is aware that there
are phonological and semantic differences between
the two languages, and he is careful to record
Osage correctly.  In collecting his verb paradigms,
however, he innocently misrepresents the standard
we-forms.  In Omaha, the pluralization of these is
done in =i, and this is the way he proposes his
verbs to the elderly Osage speakers.  "Can you say
oN[verb]-i?" he inquires in Dhegiha.  In Osage, the
=i form is specialized, rare and archaic, but the
elderly speakers recall their elders using it in
their youth.  To them, it is technically grammatical,
so they accept it, and it goes in the dictionary.
LaFlesche and his informants get habituated to
this routine, and the =i forms pile up with little
criticism.

In the late 20th century, however, when modern
linguists such as Carolyn Quintero go to work with
the Osage, the current generation of elders has
grown up in the 20th century and has never heard
the old =i form used.  They flatly reject this
form as not being Osage.

I think the above hypothesis would exonerate
LaFlesche, explain the =i forms in his dictionary,
reassure us of the basic value of that dictionary,
and explain some possible fossilized =i forms in
Osage, without contradicting the findings of our
modern lingists.

Rory