Osage plural

[Rory M Larson]

(John wrote:)
> Note that *pi remains bi in Omaha in (a) names, (b) in songs
> (at least in the 1890s), and (c) in certain morphological
> contexts, e.g., before =ama the quotative (so-called - it's
> more of a marker of reporting used for things that can't be
> personally vouched for), or in the negative plural =b=az^i.
> Rory Larson has pointed out that is generally associated with
> indirect/unvouched for contexts, even when nothing follows it,
> and that =i alternates with =bi in some morphological contexts
> to mark that distinction.

I'd like to revise my position slightly from what I was
arguing when first grappling with this.  First, the dichotomy
between =i and =bi is quite regular in the Dorsey texts.  If the
verb is followed by =i, the speaker is asserting it on his own
account as the straight goods.  If the verb is followed by =bi,
it means that the speaker is absolving himself of responsibility
for the implication of what he has just said.  Thus, =bi is
regularly used in reporting hearsay, or in describing a former
hypothesis.  In the latter role, it may cover supposition or
expectation ("supposed to").

In third person declarative statements, neither =i nor =bi
normally has anything to do with plurality.  They do indicate
that the concept is complete rather than progressive, and that
it is independent of outside influence.  In commands, and in
statements and exhortations that use the potential particle
/tte/, =i at least signals plurality.  There are a few very
rare, but illuminating cases, however, in which =i is replaced
by =bi in these contexts.  Usually, you command a group of
people in the form: V=i ga!  But if you are conveying someone
else's command, you can cast it in the form: N V=bi ga!, where
N is the name of the party whose command you are conveying!
Similarly, [Concept] tta=i conveys the speaker's assertion that
[Concept] is to take place, while [Concept] tta=bi says that
[Concept] was supposed to take place, but perhaps didn't.
In these cases, I don't know whether =bi would be used in the
singular or not.  I'm also not clear yet on the use of these
particles in declarative you- and we- statements.  Tragically,
both particles seem to be moribund in modern Omaha.

When we first discussed the =i vs. =bi issue, I offered defenses
for three situations in which my distinction of =i, fact, vs.
=bi, hearsay, seemed to run into trouble:

      a) The name of the giant killed by Rabbit:
            Ttaxti-gikhida=bi, glossed as
                  He-For-Whom-They-Shoot-Deer
            Trouble: By this translation, the particle should
                  be =i, not =bi.
            Defense: Names are conservative; this one dates
                  to a time when the =i vs. =bi dichotomy
                  had not yet been made.  At that time,
                  there was only *bi < MVS *pi, and it
                  indicated plurality.  Bound up in a
                  name, the /b/ sound was retained, even
                  as it was lost in comparable derivations
                  from active speech.

      b) The song in "The Lament of the Fawn over its Mother":
            In the narrative portion of this story, =i is
                  used in the dialogue over whether the
                  beings were men or crows; yet in the
                  song portion, =bi appears in the same
                  positions that =i had just above.
            Trouble: If =i and =bi appear in the same place
                  with the same meaning, then maybe they
                  are merely speech variants of the same
                  thing.
            Defense: Songs, like names, are conservative,
                  because they are memorized.  The narrative
                  part, however, is not memorized, and
                  hence reflects standard speech at the
                  time the whole piece was recorded.

      c) Sentences with =i in the narrative:
            Normally, the narrative (non-dialogue) portion of
                  the myths has the =bi form.  Every now and
                  then, however, the narrative shifts to the
                  =i form for a sentence or two.
            Trouble: This is a counter-example to the rule of
                  =bi for reporting hearsay.
            Defense: This shift is stylistic.  It may occur
                  for the same reason that we sometimes
                  shift from past to present tense in telling
                  a story: to make it seem more immediate to
                  the listener.  Also, the =i form is surely
                  the most common one in daily speech, and
                  requires less vocalization; shifting to it
                  may just indicate a moment of laziness.

Of these three "defenses", I think only the third is still necessary.
In the "fawn" story (Dorsey p358), the fawn and its mother argue as
follows:

Fawn: NaNha', dhe'ama ni'ashiNga =i     ha.
      Mother, these   men        (decl) (emph)
      Mother, these are men.

Mother: AN'khazhi, ni'ashiNga =ba'zhi,    kka'gha =i     he.
      Not so,    men        not-(decl), crows   (decl) (emph)
      No, they're not men, they're crows.

The fawn proves correct; the mother is shot and butchered by the
hunters.  The fawn returns and sings the lament:

NaNha', nia'shiNga' =bi   ehe',  kkagha' =bi   eshe'   dhaN'shti.
Mother, men         (hyp) I say, crows   (hyp) you say (past).
Mother, I said they were men, you said they were crows.

In fact, the alternative use of =i and =bi here is exactly as it
should be.  In the first pair, the two are directly asserting their
respective claims as to the identity of the beings, which requires
=i.  But in the song, the fawn is recalling these assertions as
hypotheses that had been stated.  As both the correct and the
incorrect assertion are reviewed as former hypotheses, =bi is the
appropriate particle.  Thus, the "song defense" is not needed here.

The "giant" story revolves around a giant for whom hunters shot
deer, but dared not butcher for themselves, as the giant claimed
all kills for himself.  The story suggests a thread of outrage over
this oppression.  I think the inspiration for this story is almost
certainly some ancient chief who imposed tribute upon the ancestors
of the Omahas, and whose overthrow is here celebrated in a somewhat
garbled way as one might expect after a few hundred years of
retelling.  The giant himself is outraged that Rabbit presumes to
cut up a deer without the giant's permission.  The name
Ttaxti-gikhida=bi is glossed "He-for-whom-they-shoot-deer",
but it might actually be better understood as
"He-for-whom-deer-are-supposed-to-be-shot".  That is, it could
be the expectation that tribute is to be rendered to him, rather
than the bland observation that people shoot deer for him, that
is the true force of this name.  In that case, =bi rather than
=i would be grammatically appropriate.

I've looked through the Omaha names listed in Fletcher and
La Flesche.  There seem to be about two dozen that use =bi and
none that use =i.

      U'nizha=bi        (Meaning uncertain)

      INshta'dha=bi     INshta', eye; dha, cause, bi, he is.
                        Appointed eyes.  Refers to the
                        appointed leader of the chase.
                        This name belonged to one who was
                        hereditary leader of the chase.

      I'nikasha=bi      Refers to tribal pipes-- objects by
                        which the tribe is identified as a
                        people.

      Nia'dishtaga=bi   Ni, water; adi, there; shta, from
                        iNshta, eye; gabdha, to open.  (See
                        Legend of the Sacred Pole, p. 70),
                        where the name appears without
                        elision.

      Te'hutaN=bi       Te, buffalo; hutaNbi, bellowing.
                        (See ritual, p. 298.)

      DhispaN=bi        To feel of.  Refers to corn.
                        (See ritual, p. 266.)

      WanaN'shekhidha=bi      One who is made soldier.

      Gi'dhikaN=bi            He to whom a place is yielded.

      DaN'a=bi          (Meaning uncertain)

      Mi'naNda=bi       The only sun.

      GiaN'ha=bi        Gi, from him; aN'ha, to flee;
                        bi, who is.  One who is fled from.

      EzhnaN'gidha=bi   EzhnaN', only; gidhabi, who is
                        favored-- gi, possessive sign;
                        dha, favored; bi, who is.
                        The favored son(?)

      Wahu'dha=bi       One of whom permission is asked.
                        Appears in treaty of 1815.

      Sigdhe'naNpha=bi  Sigdhe, footprints; naNphabi, to fear.
                        One whose footprints, even, are feared.

      GdhedaN'naNpha=bi Hawk who is feared.

      Udha'ga=bi        Refers to wolf.

      Uma'a=bi          Cut into pieces and spread (scattered?).

      NaN?aN'=bi        NaN?aN, to hear; bi, who is.
                        One who is heard.

      I'iNga=bi         I'iNga, rejected; bi, who is.

      Dha?e'gidha=bi    Dha?e, from dha?edhe, liked or
                        beloved; gi, passive; bi, who is.
                        Refers to a calf that is caressed
                        by its mother.

      I'kuha=bi         I, is; kuhe, fear of the unknown;
                        bi, who is.  One who is feared.

      I'bahaN=bi        I'bahaN, to know; bi, he is.  He
                        is known.  Refers to a chief's
                        son.

A striking feature of most of these names is that the
subject referred to is passivized.  The person being named
is known, feared, beloved, rejected, heard, deferred to,
appointed to office, or has his eyes opened.  Usually the
subject's status depends upon the attitude or behavior of
others toward him.

In contrast, names in which the focus is the subject of an
active or stative verb normally are not followed by either =i
or =bi.  Names like this are numerous, but a good example pair
can be found in the names of the magnificent dogs in the story
"WahaNdhishige and WakaNdagi", p. 109.  In these cases, no
entity class noun is specified for them, but only their action
with respect to an object:

      MaN'ze  =dhaxaN'
      Iron    breaks-by-mouth
      Breaks-iron-with-his-teeth

      IN'?e   =dhashi'zhe
      Stone   shivers-by-mouth
      Shivers-stones-with-his-teeth

In a 19th century Omaha statement, these verbs would have to
be followed by =i or =bi, unless the implication was that
someone else was responsible for their action.  That would
be a reasonable supposition for dogs, but this is regular
for human names as well, and even so, the latter expression
should probably come out as IN'?e=dhashi'zhe=e.

So whereas a statement uses =i and =bi to signal that the
concept is complete rather than progressive, and independent
of outside influence, with =i indicating assertion and =bi
indicating hypothesis, a name generally uses =bi to signal
that the subject is in a passive state relative to the
operation of others.  It is likely that this =bi imparts
a normative claim rather than a factual one.

In the texts, there seems to be another use of =i as well.
It is certainly most commonly used as described above for
statements, but sometimes =i also seems to be used to
indicate passivization, much in the way that =bi is used
in building names, only with declarative rather than
normative force.  In statements, I think only =i is used
this way.  Occasionally this or some other important use
of =i will show up in combination with one of the regular
=i's or =bi's, as either =i=i or =i=bi.

In any case, the =bi that shows up in a name may well
indicate the plurality of passivization that occurs in
Dakotan; but we cannot be sure that no other semantic
implications are involved.  A name ending in =bi might
mean e.g. "He-who-is-to-be-feared" as easily as
"He-who-is-feared".  Therefore, I don't think the
"name defense" is needed either, though it seems to be
true that only =bi occurs in names.


> 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.

John and Regina have both been arguing on the basis
of this assumption, that =i is a reflex of MVS *pi,
and that its existence in that form is a quirk unique
to OP.  Regina has suggested that Osage might have
borrowed =i from OP, or that =i might simply be a
speech variation of Osage =pi, to explain the =i forms
that show up in the La Flesche dictionary of Osage.
Against this, John points out the geographical
separation of Osage from OP, and the fact that both
modern Osage and a set of early ritual texts use =pi
for pluralizing; he suspects that La Flesche's Omaha
background may somehow have corrupted the dictionary.

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?

If they did, Osage presumably used the =pi particle
most commonly, unlike OP where =i was more frequent.
In the old ritual texts, =pi might by chance have
been the one used in the grammar of the ritual, and
replicated throughout because the grammar there was
always the same.  In modern times, =pi might indeed
be used for plural, in a grammar largely restructured
in a process similar to creolization over a hundred
years of close exposure to English.  In between,
Francis La Flesche may have caught some genuine Osage
=i forms.  His Omaha background may have made it
especially likely that he would.  Thinking as an
Omaha, he might have been trying to elicit =i forms:
"Could you say: 'MiNdse kHe oNgdhuza=i' ?" he might
have asked, trying to fill out his paradigm, and his
elderly informants would nod indulgently and reply,
"Yes, you could say that!", though the Osage meaning
might have been odd and notably different from what
he thought he was getting at.

I've looked at the short collection of Osage sayings
at the end of the dictionary, which I understand are
supposed to be basically correct, and not from
La Flesche.  There is one case in which =i appears,
in the tta=i tHe form which is common in Omaha, and
which in the context indicates a very certain future.
There are also two or three cases in which =bi is
used for what is singular in the translation.  Both
=azhi and =bazhi are used for the negative plural.
If this material is valid, it seems unlikely to me
that Osage =pi was simply a pluralizer at the time
it was collected.  It is certainly true, though, that
=bi (=pi) occurs in many places where we would find
=i in 19th century Omaha.

To me, it seems entirely probable that OP =bi and =i
have been distinct particles for a long time, most
likely back to proto-Dhegiha and possibly back to
MVS.  I agree that OP =bi and Osage =pi descend from
MVS *pi.  But I would like to ask John and other
comparativists what evidence causes us to be sure
that OP =i derives from the same particle.

Rory