Problem of Ā/AU in Sindarin
I believe we have a consensus concerning the ineffectiveness of the I-affection on non-final O[<Ā/AU], judging by Rodyn. Yet there are two other examples which rather offer the same as E after Prestanneth:
Rhevain, pl. of Rhovan, from Srāban [HoME-XI/219]
Negyth, pl. of Nogoth, from ?N(a)ukotto [HoME-XI/388, PE17/45, ]
Rhevain is also given as Rhovain in PE17/18, suggesting perhaps that both forms are acceptable. However, Nogoth is "probably from some such form as nukotto", meaning its O comes from Ŭ, which does later develop to O>Œ>E when plural. Yet PE17/45 gives the same as " nogoth [< naugoth: naug, nuk- + augmentative suffix -oth], of which plural was negyth". This could mean that the phonetic developments Ā>Ǭ>AU>O took place before I-affection, at least in some cases or dialects.
Unless I am missing some other non-final O[<Ā/AU] word with an attested plural, the times that the plural of Nogoth is given as Nœgyth/Negyth (never "nogyth") outweighs Rodyn and Rhovain, with the latter having the alternative Rhevain. Thus, I believe it would be (at least a bit) safer to say that non-final O, be it from the ancient A/O/U, usually becomes E when plural. Any objections?
Björn Fromén Jun 01, 2017 (00:12)
Александр Запрягаев Jun 05, 2017 (10:59)
Ekin Gören Jun 05, 2017 (12:09)
This may suggest that, while the diphthong AU became OE when plural, the monophthong O from Ā/AU became Œ(>E).