From: dgkilday57
Message: 71220
Date: 2013-05-24
>The Latvian word has a circumflex over the first vowel, which is therefore long and cannot represent zero-grade PIE *ud-. It escapes me how anyone posing as a lexicographer can neglect diacritics. The Lithuanian word also has a circumflex over the last vowel. It points to a protoform with nom. sg. *wondo:r, obl. sg. *wonden-, but other BS lgs. do not have the same inflection. For starters see Pokorny, IEW 79-80, and references.
> Wiktionary gives 'vanduo' as modern Lithuanian for 'water', and 'udens' as the corresponding Latvian word. In both etymologies the same Proto-Balto-Slavic root of *wondor is given.
>
> However, at least to me as an etymology hobbyist, the start "ud" of the Latvian word 'udens' suspiciously looks like the zero-grade form *ud- of the PIE root *wed-
>
> Could it be that the original *wondor was still ablauting in some way and had a zero-grade form, and then Lithuanian kept the o-grade form, whereas Latvian kept the zero-grade form?
>
> If yes, how this zero-grade form could have looked like? If no, how do you arrive at udens from *wondor?