--- In firstname.lastname@example.org
, "tgpedersen" <tgpedersen@...>
> How would either of you two's theory, O-infix (onfix?) and a-
> Epenthesis account for Latin <umbo>, Greek <omphalos>, English
> <navel>? Is the rule (whichever) applied only partially?
> Einzelsprachlich? Only in some (grammatical) cases?
This is a root *H3nebh- and one of Rix' main examples for what is
now known as "lex Rix". In Greek, all sequences of the structure
#HRC- with a syllabic sonant preceded by a laryngeal develop in a
way that is sensitive to the exact identity of the laryngeal,
yielding Greek eRC-/aRC-/oRC- form *H1RC-/*H2RC-/*H3RC- (in that
order). Therefore Gk. /omph-/ reflects *H3n.bh-, the zero-grade of
*H3nebh-/H3nobh- which is seen in navel.
An addition to Rix' Law is that the same tripe representation is
sen in Latin, but apparently only with nasal; therefore Latin /umb-/
from *omb- is also here from zero-grade *H3n.bh-.
None of the words for navel can be shown to contain the infix we
are talking about.
> BTW on the legitimacy of the extra vowels of Glen's dog *kawana-,
> *kawanasa- : I found Luwian <kuwana->.
You are not specifying what we are to do with this information. The
form is Hittite; Puhvel enters kuwan- 'dog' (used of "dog-man" in
cultic function, if I understand), the Luwian form of which is
attested as Hieroglyphic zuwan- (with *k^-). Since none of the
writing systems used for old Anatolian languages have initial
clusters, the form can stand for *kwan- or for *kuwan-; both can be
accomodated as IE *k^won- or as *k^uwon- (the latter as a Lindeman
variant). The form cannot be seen to contain any more vowels
than 'dog' does in other IE languages.