Re: [tied] again gW>b and getae

From: m_iacomi
Message: 16786
Date: 2002-11-18

In cybalist@..., Piotr Gasiorowski <piotr.gasiorowski@...> wrote:

> Who told you that any of these names derives from a protoform with
> *kW? Almost all of them seem to contain the root *{h2ep-} 'flowing
> water, river', which has PIE *p. This root, by the way, is best
> attested in three _Satem_ groups that never show /p/ for *kW:
> Indo-Aryan (OIA a:paH 'waters' < *h2op-es, dvi:pa- 'island' <
> *dwi-h2p-o-, etc.), Iranian (Av. a:fs^ < *h2o:p-s, etc.), and
> Baltic (OPr. ape). Latin and Celtic possibly have it in
> *h2ap-h(o)n- > *h2ab(o)n- 'river', and judging from the frequent
> occurrence of "ap-" hydronyms in the Balkan area, it was widespread
> in the ancient local languages as well. It should *not* be equated
> with Latin/Germanic *akWa:, which would not have produced such
> reflexes. If Iranian, Indo-Aryan and Baltic have *h2ap- and are
> Satem nevertheless, its presence in Thracian or Dacian surely does
> not suggest that they are not Satem.

That's crystal clear. In fact, C.D. Buck gives four (P)IE roots
meaning water in some sense:

1. *wedo:r-, *wodo:r-, *udor-
from which Gr. udor, Umbr. utur, Ir. usce, Goth. wato:, Lith. vanduo,
Lett. u:dens, Sl. voda, Skr. udan-, Hitt. watar, Alb. ujë, etc.

2. *akwa: or *ak'wa:- [Pokorny: akw'a:, &kw'a, e:kw/-]
from which Lat. aqua, Goth. ahwa, OE e:a; Toch. yok-, etc.
[elsewhere from Latin, mostly said of "running water", in words of
type `river']

3. *a:p- [Pokorny: *a:b-, *ab- ]
from which Skt. a:p-, ap-, mostly pl. a:pas (personified waters),
Av. a:p-, ap- (also frequently personified), O.Pers a:pi-, N. Pers.
a:b (`water'); elsewhere only on running (flowing) water, as in
Lith. upe, Lett. upe, Lat. amnis, Ir. abann, etc.

4. *wer-
from which Skt. va:r(i), Toch. war, etc.; also Lat. urina.

The interesting points are 2 and 3, which show clearly there is
no connection between -*kW- and p (b) in Satem languages which
continues obviously an original labial.
Ernout & Meillet are admitting the link between Lat. amnis and
Ir. abann, Gall. afon, O. Prus. ape, etc. -- that is the word is
a descendent of the 3rd PIE root. A Thracian or Dacian word "apa"
meaning `water' should obviously belong to this category too.

> Rom. apã could in theory continue Dacian, Thracian or Illyrian
> *apa: if such a form were attested (it is not, in fact; the Dacian
> word, for example, was apparently a reflex of *h2apos). However,
> the fact that it means just 'water' like Lat. aqua (not 'flowing
> water' or 'river' in particular), and that it is derivable from
> <aqua> via regular Romanian sound changes, compels us to regard
> it as a relative of <eau>, <agua>, <acqua> and other Romance
> 'water' words, not of <apos>.

Undoubtedly, Latin "aqua" should have given in Romanian "apã",
so there is little doubt about the latinity of the Romanian word.
Nevertheless, as one can see from the several different meanings
given by Buck, the semantic doesn't exclude a potential mixing of
the latin word with a probable substrate "*apa" which should have
given the same result in Romanian.


Marius Iacomi