Re: [tied] Re: Patterns in Khoisan Numerals (was: Why borrow 'seven

On Sat, 18 Sep 2004 14:22:12 +0200, petusek
<petusek@...> wrote:

>Yes, Kartvelian "4" is very similar to IE "8", but various linguist
>reconstruct various proto-K forms:
>
>Schmidt (1962): *otjxw-
>Klimov (1964): *o(s1)tx(w)-
>Gamgrelidz & Ivanov (1984): *o(s^)tx(w)-
>Fähnrich & Sardshweladse (1995): *otxo-

Based on Geo. otx-, Megr. otx-, Laz otxo-, ontxo-, otxu-,
Svan wos^tx(w), wo:s^dxw.

>Blazek thinks:
>...
>The -s^t- cluster in Svan can be explained by analogy to the following
>numeral wo-xus^d, wo-xwis^d
>...

On the other hand, there are dozens of items where Svan
/s^t/ corresponds with Georgian /t/.

>The metathesis of IE *-k^t- > K *-tx- has analogies. For example, K *usx(o)
>"a bull for sacrifice" < IE *ukso:n- (Klimov)
>...
>But Manaster-Ramer (1995) reconstructs K *os1tx(o/w), seeing here a
>substitution of the satem variant of the IE *ok^toH "8"

Another item with a k^t / s't correspondence Alexis and me
discovered in private correspondence was Kartv. *das'tw-
"bear" (Geo. datw-, Megr. tunt-, Laz (m)tut-, Svan däs^dw,
des^dw, däs^t.w, das^t.w) ~ PIE *r.k^to-.

>(but the final -x-
>can also be explained by the alliteration with the following numeral
>*xu(s1)t-.
>...

From Semitic *xamis^t-, according to AMR.

Final -xw can also reflect *-h3 (probably /xW/) in *ok^toh3.

>If this is accepted, the semantic change "8" > "4" (Glen has already
>mentioned his idea of IE "8" being the dual of "4") ca be connected to the
>opposite semantic shift for K *arwa- "8" < Semitic *?arbaH- "4"
>...
>
>So I think, it is very plausible to expect Kartvelian "4" to be a loan form
>IE, too.

=======================
Miguel Carrasquer Vidal
mcv@...