On Wed, 11 Feb 2004, P&G wrote:
> > *o:k^u- 'swift',
>
> As always, I appreciate list members' patience with me on this.
>
> Are we saying two contradictory things at the same time:
> (a) after o- an initial laryngeal is deleted (ala Beekes 1969)
> (b) after o- an initial laryngeal explains the attested long initial o:-.
A laryngeal contiguous with the o-element is not deleted. The position of
the deletion has been determined by Martin Peters in his yellow book
(Untersuchungen zur Vertretung der idg. Laryngale im Griechischen, Wien
1980) as __Ro and oR__. In my view, this is correct only if the o is of
infixal origin.
The formula __Ro refers to the absence of prothetoc vowels in word with
the infixal -o-. A fine example is Gk. loigós 'ruin, death' vs. olígos
'feeble' (IE *loyg-ó-, but *H3lig-), another is Gk. moikhós 'adulterer'
vs. omeíkho: 'urinate' (*moyg^h-ó-s vs *H3méyg^h-o:). The formula oR__ of
course captures the type *pórnaH2 which has lost a laryngeal in r_n).
Jens