On 2008-02-03 23:52, tgpedersen wrote:
> Hm. I assume that would be *sU-tráva? Why would that -U- not have been
> dropped?
Because in the middle of the 5th century it was pronounced as /u/, not
yet reduced to a schwa-like central vowel; actually, in an archaic
_noun_ with this prefix I'd even expect *soN- rather than *sU- (the
latter is primarily a verbal prefix), so a Latinate rendering of such a
word would have been *<sutraua> or *<santraua>. The root would have been
that of OCS -trovoN, travljoN 'consume' < *treuh- (Gk. trú:o: 'wear out').
Piotr