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

[Koontz John E]

On Thu, 6 May 2004, Rory M Larson wrote:
> One thing that struck me at the time was that the word for 'hand' had
> been replaced in the (MVS) group: in Hidatsa and Biloxi, and presumably
> PS, 'hand' had been something like *s^aki or *s^ake.  In MVS, however,
> 'hand' was *naNpe, while the *s^aki/*s^ake term had moved on to mean
> 'nail', 'claw', 'hoof' or 'talon'. Am I remembering this right?

Right, as, of course, Bob has already confirmed.  (And I think I heard
once that somebody had observed from a general study that terms for parts
of the limbs tended to move outward along the limb, or perhaps it was the
reverse.)

I had, of course, been thinking something along these lines, though in my
case from having read Bob's paper at some point (or at least having seen
the handout), but I couldn't help noticing the other similarity, too.  I
suspect it's just a coincidence.  I notice that the kiyah- formant is
restricted to Wichita, even, within Caddoan, at least as far as I can make
out at Mark Rosenfelder's numeral collection site
(http://www.zompist.com/numbers.shtml).  For that matter, though IO gree-
in gree'raabriN 'eight' could be from either *kree- or *kyee-, I think the
similarity of the initial syllable there to PMV *kyee'praN 'ten' is
probably also coincidental, because a positional analysis involving *the
(*k + the > *kre, and Bob suggests *the < *re-he).  Only Dakotan preserves
*ky distinct from *kr, incidentally.

> Anyway, if 'hand' was originally *s^aki/*s^ake, might that not be the
> derivation of the *s^aak- we find in some of the higher order numbers?
> It would be an obvious choice for the 'five+' requirement.

Obviously Bob agrees, and I'll chime in and agree with both of you.  I
suspect Wolff and Matthews might even have noticed this, too, though I
don't recall a reference, off hand.

> The problem is that this is also the only "higher digit" formant that
> pairs with things that don't look anything like Siouan 'one', 'two' or
> 'three'. We would seem to need a "lower digit" counting system something
> like:
>
>   1 = *pe
>   2 = *owiN/*owaN
>   3 = *rog^aN/*yog^aN

> I don't suppose anybody knows of any numerical system comparable to
> that, either inside or outside of Siouan? Given how widespread the
> *s^aak- terms are, the original formation of these should be pretty old.

It would definitely be interesting to look for a numeral series like this,
and I did briefly look for it at Rosenfelder's site, without any real
luck, but no persistence either.  If we look within Siouan, and if we bear
in mind that the numerals often show little contrast between *Cp and *Cw,
even though the phonology isn't always regular, and that, in fact, we
don't reconstruct *Cw clusters except in a few unusual cases, it's
possible that *s^aakpe might be *s^aak-we, and that the *we might be
essentially wiN or waN 'one'.  Then *owiN could be "(one more) on one."
However, beyond that my imagination fails me.

I think Bob pursued the matter further, though not necessarily in this
direction.  As I recall it, his hypothesis is that the forms are not
*s^aak 'hand' + numeral, but *s^aak 'hand' + description of gesture made
with hand to indicate the numeral in question.  There are various hand
counting systems that involve using only one hand, and after doing 1-5
with fingers in some way making some differentiating gestures to handle
6-10.  I wish I knew of a survey of such systems - if one exists!