Re: [tied] Armenian three series of velars?

>> Olsen suggests that *gh does palatalize like *gWh, based on
>> el~j^ewr "horn" < *(e)rghĂŠ:wr < *ghrĂŠ:wr.. But then it
>> doesn't in *gher(h)n.tiyo- > gerandi "sickle", so that's
>> unclear.

What does Birgit Olsen make of dustr < *dug^hter?
Is it correct that labiovelars fall with the palatals after /u/, whereas
velars do not?

Peter