justplainpossum wrote:However, are you saying that what Pharr writes in 601 is indeed incorrect? That in compensative lengthening, epsilon should become eta (not ει) and omicron should become omega (and not ου)? If so, I'll correct it in my book.
Not at all. There are cases, like masc. acc. pl. of 2nd declension, where o turns to the (spurious)
diphthong ου (-ον-ς > -ους), and there are others, like noun stems ending in ν and οντ,
where o turns to ω (δαιμον* > δαίμων; γεροντ* > γέρων, but note dat. pl. γεροντ-σι* > (ντ drops) > γέρουσι).
With ε, there is λιμεν* > λιμήν, and there is ἐ-μεν-σα* > ἔμεινα (aorist of μένω; example from Smyth).