From: Miguel Carrasquer
>I don'tI don't understand what you say about *e:. As to *o:, as far as it's not
>see why *o must be long, both in respect to the IE vowel system as we know
>it and in terms of the overall evidence of daughter languages which affirm
>that it is short except in _exceptional_ circumstances like the Law we are
>discussing. To say that *o is long causes an unstoppable chain-reaction
>that forces us to further explain ad absurdum what instances of *o: could
>possibly be and why it's on a par with *e: even though *e: should be
>Maybe it's also worth mentioning that *o is obviouslyThen why is Skt. /u/ not lengthened in an open syllable?
>from *e in the respect of labiality and so this extra quality may have
>the reason for a labial-marked *o becoming /a:/ rather than an unmarked *e.
>In other words, a kind of shift from labial-marking to length-marking. An
>exchange of one quality for another.