Everyone balks at an alternation *d-/*H- such as there supposedly is
in the PIE word for "two" *dw-/*Hw-. How odd that it occurs in Malayo-
Polynesian too:
*hipaR "opposite side of a river"
dipag "other side, opposite side" Mansaka
dehipag "the opposite side of
a canyon or valley" Manobo
difar "the other side, in the sense
of the side facing the speaker" Tiruray
'ifar "to cross over to the other side
(as of a river or street)"
se'ifar tamuk "to negotiate formally
the terms of a brideprice"
dipah "opposite bank of a river" Mukah
dipah "opposite bank of a river" Kayan (Baluy)
dipar "opposite side" Kelabit
dipah "either of the sides of a river" Uma Juman
"It thus seems likely that the dual divisions of
Proto-Malayo-Polynesian society were at least traditionally,
if not physically, associated with settlements on
either side of a river" (R. Blust)
in a word I think has relatives in both IE and AfrAs (and Proto-Bantu
to boot), as does "two".
http://www.angelfire.com/rant/tgpedersen/Opr.html
Torsten