At 7:06:48 PM on Sunday, March 16, 2008, Patrick Ryan wrote:
[...]
> The root for 'protect (probably, closer to 'challenge
> danger by advancing to the front')' is clearly *po:(H)-
> not *pa:(H)-.
[...]
> *pa:(H)-, 'herder, leader to pasture, provider'
What makes this clear? If there actually are two roots
here, Gk. <pô:u> and <poimé:n> 'a shepherd' would seem to
point the other way. Given the close semantic connections
among 'protect', 'herd', and 'nourish', it seems much easier
to see a single root *peh2- with o-grade *poh2-. (And in
the other direction I believe that Hittite /pahs-/ 'to
protect' tends to suggest *peh2-, not *peh3-.)
Brian