>>The prototype of Gk. polis, Lith,
pilis, Skt. pu:r is usually reconstructed as *p(o)lh1-(i-) 'hill-fort,
citadel'. The quality of the laryngeal is not quite certain, and there are
subtle internal difficulties (Greek and Lithuanian have an i-stem as opposed to
the Indic root noun, Greek has a mysterious dialectal variant with initial
<pt->). It is hard to decide if this word can be related to
*pelh1- and *polh1-u- (as "a place for many people"?).
I'd rather say: "as a place, surronded by
ramparts" (rampart - Lith. pylimas, Amer. Eng. fill). Rampart is made of ground,
filled up with ground. It is well known expression in Lithuanian folklore
"su-pilti pili:". The hill-forts (piliakalniai) are traditionally regarded as
artificial hills (though only ramparts are artificial in reality). The
territorial community with a centre in such pilis, polis - "a place, surronded
by ramparts" - is defined naturraly by the same name - Greek polis may testify
this. And thus pilis/polis may easily acquire meaning of wide, open territory,
occupied by a certain community (Slavic *polje). Thus I think there is no
difference between the origin of these roots:
>>*pelh1- 'fill'. The best-known
derivatives, apart from various verb stems, are *plh1-no- 'full' and
*polh1-u- 'much, many'. There are also numerous more complex formations based on
this root.
>>*pelh2- 'flat, open'. Lots of
derivatives, several subsenses, *pélh2-tu- 'open field' (Eng. field, German
feld), *plh2-no- 'flat' (Lat. pla:nus), *polh2-jo- (Slavic *polje), *plh2-m-ah2-
'palm (of one's hand)'.
"A
place, surronded by ramparts" also may explain relation between hill and the
mentioned root, as the castles, fortifications used to be aranged on
hills.