Achrelos wrote:I know that w>v isn’t uncommon, at least in the Romance branch of languages, but I was under the impression that in these diphthongs the old upsilon was pronounced something like the French u, not like a w, as in the word ελεύθερα.
As far as I can tell, that wasn't true as the non-syllabic element of a diphthong. The Wikis transcribe it as /w/ or the equivalent /u̯/.
Achrelos wrote:It strikes me odd that unlike the Latin>Romance examples, upsilon retains its vowel status when appearing alone or in between other vowels and yet takes a consonant status in diphthongs that typically directly precede consonants and a consonantal w makes no sense (at least in my prior experience).
Four things:
1) If by "appearing alone", you mean "syllabic", that's not so odd. Syllabic vowels don't undergo fortition very easily.
2) It looks like w>v also occurred between vowels.
3) Before a consonant is a very reasonable place for fortition. Consonants are less sonorous than semivowels, so a semivowel followed by a consonant could assimilate by losing sonority.
4) /w/ can certainly be consonantal; it might pattern with either consonants or vowels or both depending on the language. Note how semivowels are also called semiconsonants.
Achrelos wrote:This is true even for prefixes of Greek origin, such as Spanish automóvil vs Greek αυτοκίνητο (aftokínito).
Most Romance borrowings are based on Ancient Greek, so pre-fortition.
N'hésite pas à corriger mes erreurs.