Grzegorz Jagodzinski wrote:
> Please consult the Lubotsky's article "Against a Proto-Indo-European phoneme
> *a", available on https://openaccess.leidenuniv.nl/handle/1887/2662. Of
> course, if no *a's were present in PIE, all the discussion on uvular or
> pharyngeal character of *k/*g/*gh is just groundless.
Strangely enough, Lubotsky doesn't even mention some of the most
commonly quoted roots and words with *a, such as *kap- and *kan-. Some
of the analyses (e.g. that of *k^aso-) strain all credulity. Some rather
plausible counterarguments are just dismissed without any discussion
(e.g., why can't Slavic *s^e^rU reflect Germanic *xaira-?), and the
problem of *laiwo- (one could add a few similar words) is noted but no
solution is offered.
> Of the previously mentioned by me, Lubotsky gives the following
> reconstructions: *bheH2g^- for bhajati / phagein,
An obvious mistake. IIr. *bHaga- guarantees a "plain" velar, so the Skt.
palatal in <bHajati> must have been generalised from *bHag-e-ti etc. I
don't understand at all how Lubotsky proposes to get rid of the alleged
laryngeal in this word. Not that I accept his rule that laryngeals were
lost before media + another consonant, but his example doesn't even fall
under it!
> *g^heH2n-s- ~ *g^hH2n-s- for goose.
If the word is onomatopoeic, and I think it is, something like
*g^Han(s)- (especially if the *g^H was a true velar once upon a time) is
preferable to a laryngeal-stuffed goose. The greylag goose goes more or
less like "aNg-aNg-aNg", which people may also hear as "gaN-gaN-gaN".
Piotr