From: Glen Gordon
>Obviously, the quality of the root vowel here was /o/. However, it cannotUgh. This is painfully simple. The rule is that the
>be the "same" /o/ (i.e. it's very likely to have a different origin) as the
>one that alternates with "zero-grade" /e/ (*w�dr ~
>*w�dns (< **wed�nos), *p�:ds ~ *p�ds (< **ped�s)).
>Aha. So it's the *s, not the *y, that makes itIndeed, I concur. I haven't seen a precise (and
>different from *bhewg-. To satisfy my curiosity,
>can you formulate a precise rule for what constitutes
>a complex root and what doesn't?
>If not, the alternative possibility remains thatLaryngicide? I think we need some pseudophonemicide
>the laryngeals vanished because of some laryngicide
>property of *R.