> I like the explanation for Greek, but what is this o-H2g^-o-s?
> What is the initial o-?
Well, while Jens tries to explain away a vowel with a consonant
(????!), I'd much rather explain the vowel as a vowel.
Like duh! :)
My theory is that these words with what appears to be a
prefixed *o- are in fact prothetic in order to avoid some
potentially yucky phonetics right after the event of syncope
in early Late IE. You see, when unstressed vowels get slashed
left and right, sometimes ugly things happen.
An example would have been *kWatWaxa (I think I just wrote
*kWetWaxe in a previous post but I forgot to update it to my
current phonetics, sorry!). It would have become *ktwax (yuck!),
but the accented a-prothesis kicked in, hence *aktwax > *aktwa:
> *okto:u "eight". A similar example is *osdos "perch". It once
was a genitival construct based on *set:- "sit", hence initially
*sat:-asa, becoming *sdas right after syncope (yikes!). So it
was tidied up with prothetics: *asdas > *osdos. Note that all
genitival constructs are misanalysed as thematic roots plus
nominative. Another really snazzy case is *oxwiom "egg" which
must surely come from *xawi- "bird" somehow but... how?? It
sure looks similar. Well it IS derived from "bird" thanks to
a-prothesis. We start with another genitival construct,
*xawai-ana, this time with *-ana (> *-om gen.pl.) to convey
collectivity. It is reduced by syncope to *xwian (blech!) and
is given prothesis to become *axwian > *oxwiom (traditionally
Jens consonant idea makes no sense to me so I hope everybody
loves my idea which only depends on automatic sound changes.
:) When Jens mentioned *oxgos, I became titulated since this
can also work with this hypothesis. It would derive from
*axgas, a workaround for the disastrous *xgas < MIE *xak:asa,
genitival construct of *xek:- "to lead". It must be remembered
that this prothetic rule occurs on nominal roots that become
divorced of the source of their derivation. So while *sat:asa
managed to become *(a)sdas even though it derived originally
from *sed-, the genitive of eLIE *pad (*po:d-s) remained
*pedas, not **pdas, because of paradigmatic strengthening.
In other words, *pedas was NEVER divorced from the source
of its derivation, *pad, thereby retaining a form of the root
vowel as it occured in the nominative. Or yet another way
of saying it is: *pedas was part of a larger declensional
system for *pad, while *(a)sdas was its own root with its
own declensional system.