Helmsbrecht Queries: Frequentative -ke

[Koontz John E]

Let me get a few more of my queries out of the way now, though Johannes is
away for a few weeks and won't see these until he gets back.  Actually, it
seems like he may not be the only one away for a few weeks ... I wish I
was, too, because it's hot as blazes here.

JH (2002. Nouns and Verbs in Hocank IJAL 68.1:9) mentions 'The verbal
suffix -ke (see Lipkind 1945:39), which functions as a frequentative
marker, also appears with nouns or, more accurately, derived words
designating a nominal concept, ...'

He gives as examples caan-niN'(k)-keres^-ge 'fawn' = 'deer-DIM-be
spotted-FREQ' and kiriki'ris-ge 'downy woodpecker' = 'be striped-FREQ', to
use his orthoraphy (rendered in ASCII) with dashes to clarify the
morphological structure. It would be normaly for -ke to be voiced in this
context, I believe.

Lipkind's examples of the -ke frequentative, are

(1) maNaNs^o'-ke-s^uNnuN-naN 'he often whittles' = '(he)
whittles-FREQ-CUST-DEC', in which people will no doubt recognize in
-s^uNnuN- (< *s^nuN) the cognate of that Dhegiha habitual -s^naN-

(2) haNke' waru'c^-ka-niN-naN 'he seldom eats' = 'NEG (he)
eats-FREQ-NEG-DEC', showing that the frequentative -ke is abluating (-kE)

However, I'd suggest that the -ge in the two examples shown is not the
frequentative, but a noun-forming suffix *-ka found elsewhere in
Mississippi Valley Siouan and paricularly common in the often
transparently derived names of animals.  Other examples abound in
Winnebago, of course, but it might be more to the point to show examples
elsewhere, e.g., Mandan raN'he 'back' vs. OP naN'kka < naNh-ka; Winnebago
ruuc^-ge' 'dove', OP dhi'tta < *rut-ka; Dakotan c^haNt(e'), OP naN'de vs.
Mandan raN't-ka, Winnebago naNaNc^-ge' ~ naNaNc^ 'heart'.

Sometimes the underlying root is clear, and the whole can be rendered,
perhaps, 'characterized by X', which is not inconsistant with the
frequentative reading, but the Winnebago form of the frequentative -kE, is
from earlier -hkE (since it's voiceless), while the noun-forming suffix is
*-ka, unaspirated.  The 'characterized by' gloss is often applied to a
somewhat similar suffix -ka in Dakotan, though that is perhaps more often
added to stative verbs.  Interesting, Rankin has shown that -kha endings
in Dakotan are often from *h-ka, i.e., -ka endings added to historically
h-final roots.  This also explains -ga and -kka endings in Omaha-Ponca,
i.e., they are from *-ka and *h-ka, or *ka added to V-final and h-final
roots.  It's not impossible that the Winnebago -ke frequentative is itself
a result of reanalysis of *h-ke < *h-ka as an independent formation.

By way of background, pretty nearly all *ka and *xa become ge and xe in
Winnebago (and Chiwere), whereupon Winnebago loses the final e, along with
other final -e, unless the e follows a cluster like *-hk or *s^k and so
on.  Thus, Winnebago ends up with -k vs. -C-ge vs. -ke (<-h-ka) for nouns
with this formant.  Some nouns just end in -k, of course.

Anyway, I'm not sure if we want to conclude that -ge final nouns in
Winnebago reflect the particularly Winnebago -ke frequentative, though
it's not an unreasonable hypothesis in Winnebago (if one omits the -k
final nouns).

Given Johannes' goal in this apper of looking for markers of nouns, it's
possible we might want to reexamine -ge in this context.  However, we have
to bear in mind that across Siouan *-ka is often added to stative verbs,
too.  The *-htaN-ka 'be big' set is a well known case, and often surfaces
without -ka.

In short, the problem of identifying nouns is an old one.

JEK