On Sat, 13 Oct 2001 13:19:01 -0300, "João S. Lopes Filho"
<
jodan99@...> wrote:
>Thus,
>fidwor , *pwetwor-
>wolf < *wlpwos
>oven < *upwen-
>
>In other IE branches is there a dissimilation pw-w/w-pw > kw-w/w-kw ?
>Are there more examples of such process?
I guess there must be: labialized bilabials are rare phonemes in
general, and if for some reason they arise (e.g. after the loss of
/u/), they will tend to either disappear (/pw/ > /p/) or merge with
some phoneme that is more apt to be labialized , i.e. /k/. In
Armenian, /tw/ and /sw/ merged with /kw/ to become Classical Armenian
<k`>.
>Is Germanic the only IE branch where pw>p, instead of pw>kw?
That is my theory (there are also the cases where in Germanic we find
*b for reconstructed *ghw, such as G. "bitten" < *ghwedh-). There are
also cases in other branches where we find unexpected /p/ (e.g. Lat.
<lupus> "wolf" or Arm. <leard> (*leprt) "liver").