On 2006-05-26 11:17, Mate Kapović wrote:
> How would you then explain for instance IE *gWlh1tós (Greek ble:tós) > Skr
> ud-gu:rn.ás "risen"? The influence of ud-?
No, /u/ predominates after velars if the influence of the context isn't
strong, but in order to prove that the u-colouring is due to a
labiovelar you'd have to show that
(a) *kr.hV > kirV, *kr.hC > ki:rC
(b) *kWr.hV > kurV, *kWr.hC > ku:rC
etc., AND NOT THE OTHER WAY OUT with at least statistically significant
consistency. Individual examples chosen to illustrate the change don't
carry much weight as long as you ignore the counterexamples. There is
some variation even after palatals, cf. RV ju:rn.á- vs. AV ji:rn.á- <
*g^r.h2-nó- 'worn out with age'.
Piotr