Re: [tied] PIE dorsals

It has been argued that different reflexes of the three dorsals can be found
in both Latin and Sanskrit, which, if right, rather does to death any
attempt at a two-way system in PIE.

For Latin, Peter Schrijver makes the claim that *k'e > Latin >ce<, whereas
plain *ke > ca. Examples of velar ke > ca are calidus < *kel, candeo <
*(s)kend, carro < *(s)kers, carpo < *(s)kerp scabo < *skebh, scando <
*skend. There are some apparent exceptions, e.g. cena < *kerts-na, celsus
< *kelh and a few others. This claim is still disputed.

For Sanskrit, if I remember rightly, the labiovelars are to be distinguished
from the plain velars by the reflex of syllabic r as in guru, which implies
a labiovelar, confirmed by developments in Greek /barys/ Latin gravis and
elsewhere.

Peter