Salmon (was: Renfrew's theory renamed as Vasco-Caucasian)

From: stlatos
Message: 49991
Date: 2007-09-19

--- In, Piotr Gasiorowski <gpiotr@...> wrote:
> On 2007-09-18 01:22, stlatos wrote:
> > To make sure I wasn't misremembering I went to
> > to find:

> > Etymology: From PIE lok´si- `salmon, salmon-trout' [: OHG lahs (m.),
> > Old English leax (m.), Old Norse lax (m.), all `salmon' (<
> > Proto-Germanic *lahsa-), Old Prussian lasasso (f.) `salmon' (<
> > *lok´sok´yeha-), Lithuanian la~šis (m.), Latvian lasis (m.) `salmon'
> > (< *lok´si-), Lithuanian lašiša` (f.) `id.' (< *lok´sik´yeha-),
> > Russian loso´s' `salmon,' Ossetic läsäg `brown trout' (< *lok´sok´o-)
> > (P:653; MA:497)] (cf. Schrader/Nehring, 1929:2). In Tocharian we see
> > the zero-grade *l,k´si- of a paradigm whose strong grade was *lok´si-
> > (cf. Krause, 1961). We need not assume that this basic word is a
> > borrowing from an unattested TchA *laks from PIE *lok´so- as does VW
> > (254-5). For a fuller treatment of the meaning and form of this etymon
> > in Indo-European, see Diebold (1976).

> > Some of this is obviously wrong, but what do you think about the
> > likelihood of borrowing?

> The bottom line is that the word was most likely borrowed as
> *luksu- from an IE (but non-Iranian) source, which only makes sense if
> the donor language had something like *lok(^)so- and inherited *o
had by
> that time been unrounded in the ancestor of Tocharian.

When I said "Some of this is obviously wrong" I meant that Ossetic
läsäg < *lok´sok´o- was impossible, having a separate et. for each
B-Sl form was unnecessary (haplology and analogy in Lith. make more
sense), etc.

I think many of these words come from a dim. *LakYs.ko+ or
*LakYs.ako+. So Proto-Romance *loksika > *lokka, *loska, etc.,
'loach'; met. caused by retro. in B-Sl:


This opens the possibility that TB comes from *LakYs.Lo+s
*LakYs.u+ which mixes > *LakYs.Lu+ then assimilates > *LukYs.Lu+ (to
make Lu-Lu in consecutive syl.).

I can't disprove the possibility of borrowing, but I don't know what
language it could possibly be from, or how likely it would be to be
borrowed at a stage when o>a from one that didn't have o>a or kY>c^, etc.