Typology and the phonetics of laryngeals

[Alexander S. Nikolaev]

> Jens said:

>> I believe the facts of IE are plain in themselves ... A voiced value of
>> /H3/ is demanded by *pi'be/o- 'drink';

> Steady on!   This word is far from clear, and is, I believe, the only
> evidence for voicing in H3.  Can we really construct our theories on one
> isolated unclear word?

The two other examples of H3-based voicing in PIE are
1) proto-Celtic word *abon 'river' < PIE *H2ep-H3on ~
'that, which has the running water'
(where *H3on is a "grundsprachliches Posessivsuffix")

2) notoriously known greek _ogdoos_ '8th' < PIE (H)ok^tH3u-H2o-s
(reconstruction according to Rix GrGr, 172)

The second example is even more suspicious than the first, because
in greek, especially in the system of numerals there're so many voicings,
which are hard to account for, cf. '7th' hebdomos.

However, i do believe, too, that in the terms of distinctive
features H3 should be characterized as [+voiced], (as well as
[+labialized]), given voice was distinctive for PIE at all :-), which
is a different issue.

		Best wishes,
			Alex Nikolaev