Companion Terms for 7 and 8 (Re: 'eight' some more)

[Koontz John E]

On Thu, 6 May 2004, ROOD DAVID S wrote:
> John, I have only been following this discussion a little bit, since
> it's final exam time.  This mentioned Wichita, however, and I remember
> that there was some theory about one of the Dhegiha languages borrowing
> Wichita [eight], so here's the Wichita data, if it's worth anything:
> 	The numbers 6,7,8 are built on 1,2,3 with a prefix kiyah- which
> doesn't seem to mean anything recoverable now (a homonym meaning 'who?,
> someone' is an unlikely source). (* is high pitch on preceding vowel)
> 	six: ki*yehess  ('one' is chi7ass; the 7ass is the numeral)
> 	seven: kiya*hwic ('two' is wic)
> 	eight kiya*tawha ('three' is tawha, final vowel voiceless)
> 	'nine' is 'one is missing' and ten is long but unanalyzeable.

Bob's comparing Osage hki'etopa to the Wichita 'eight' form.  Note that
topa is the Osage 'four' term.  Wichita tawha 'three' is the Wichita
version of the Caddoan 'three' form.  'Three' is one of Wally's
Macro-Siouan sets, as I recall it, and would be cognate with PS *raa'priN
(or, equivalently, *raa'wriN).  I don't recall the Proto-Caddoan version
of the form.

If Osage does borrow 'eight' from Wichita, it's a second or third case of
borrowed 'eight' terms in the general Midwestern area - Wichita > Osage,
Tutelo (or something Siouan) > Miami-Illinois, and (something Siouan) >
Ioway-Otoe.  It's almost as if 'eight' were tabooed in some way.  As Bob
points out in regard to the possibility of the IO form being borrowed, the
existing terms, attested or hypothetical by analogy with respect to
transparent formations for adjacent numerals or related languages are not
problematic in any morphological way.

On the other hand 'higher' digits (six to nine) seem particularly unstable
and/or transparently formed in Siouan languages.  The 'nine' term is
definitely widely borrowed regionally, too.  Maybe this is just a culture
area in which innovation and/or multiple sets of forms were common in this
area in the pre-contact period.  After contact base-ten numeration would
tend to be common as a result of European commercial influence.

It's interesting that along with Osage {hki'e}topa and the Wichita kiyah-
based higher digits we also have IO {gree}raapriN and PS *{kye}praN
(or *{kyee}wraN) 'ten'.  This almost looks like an areal "higher digit"
formant.

I also wonder about the *s^aak- element that is so common in Siouan higher
digits.  Could it be *s^aak < **kyaak?