Re: [tied] root behind Gk kalamos, Latin culmus?

On 2007-01-10 15:48, Carl Edlund Anderson wrote:

> Can anyone comment on the etymology of or PIE root behind Gk kalamos and
> Latin culmus (both "reed", cognate with English haulm)?
>
> When I look at proposed etymologies, I see forms like k'olm-, k'alH2m-,
> and k'ol@..., and I can see the Greek form coming from kal- or the Latin
> from kol-, but am having trouble seeing how the Latin form could come
> from kal- or the Greek from kol-. The Germanic form must have been
> *xalm-, which also suggests k'al-, but whence that Latin /u/ (presumably
> from earlier /o/)?

There are also Balto-Slavic forms that might be derived from
*k^alh2-mo-/-mah2- or *k^olh2-mo-/-mah2-. The Germanic form is likewise
ambiguous. Latin points to *o, and Greek either to *a or to accented
*k^l.h2-. In the type of o-grade derivative represented by the tentative
form *k^olh2-mo- the *h2 should have been lost. This is all right for
Latin and Germanic, and for Balto-Slavic if the Saussurean effect makes
the laryngeal asyllabic RATHER THAN lost. Then perhaps we should
reconstruct something like

*k^ól(h2)-mo- (perhaps with restored *h2)
*k^l.h2-máh2- (collective)

with analogical levelling, e.g. *k^l.'h2-mo- > Gk. kálamos, *k^ólh2-mah2

> Slavic *solma

Piotr