On 2008-03-13 14:59, alexandru_mg3 wrote:
> 1) first I doubt regarding the idea to put h3 outside from h1 and h2 =>
> this supposed 'a well known nature of PIE-s h1, h2, h3" => that for
> sure is not the case today ...
What do you mean -- today? By now, all PIE laryngeals have disappeared,
but if you yourself use different indexes for the three of them, they
were surely different consonants, and if you have a natural class
involving three phonemes, there will usually be another natural class to
which only two of them belong, thus ganging up against the third. In
Hittite, *h2 and *h3 (dorsal fricatives) are reflected as <h(h)> in many
positions, while *h1 (probably just a glottal glide) is always lost. If
*h3 was voiced (and there's some serious evidence that it was), then *h1
and *h2 might have patterned together as voiceless.
> 2) Anyway: you can check and see that my examples were limited to h1
> and h2 and in addition I quoted exactly she's own derivations (not to
> work with examples that she didn't give)
> But 'Just for the record:' barba:tus (with t- (sic!)) has an h2
> inside, isn't it? and is quite an old word viewing its Balto-Slavic
> cognates.
Eng. bearded could also be considered a cognate, but the formation is
too productive to be dateable. It's like arguing that Eng. was/were
can't be due to Verner's Law, since there are no other traces of VL in
the English verb system. In fact, its effects (easy to see in OE) have
been eliminated almost completely through the creation of analogical
forms (Mod.E choose/chose/chosen replacing OE ce:osan/ce:as/curon/coren.
If you want to find traces of the "Olsen preaspiration", you should look
for them among residual, partly obscured forms, such as Skt. ti:rtHá-
'passage' < *tl.h2-tó- (cf. Lith. tìltas 'bridge, plank').
Piotr