S. [œi] became [ui] or [y]; [œi] > [ui|y]
One open question is how the diphthong [œi] developed in Sindarin, assuming it occurred at all. This diphthong is relatively common in the Noldorin of the 1930s (where it became [ei]), arising in the phonetic history of plural forms of words with a base vowel o, where the raising of [o] to [u] was usually inhibited in the final syllable of polysyllables:
- ✶ñgolodī > (ñ)gœlœidh (fronting but no raising) > N. geleidh or geledh plural of N. Golodh “Noldo” (Ety/ÑGOLOD, PE22/41).
However, [œi] was not a factor in the phonetic development of Sindarin plurals, because the o was generally raised to u and thereafter fronted to y.
- ✶ñgolodī [> *(ñ)goludhi (raising)] > gœlydh (fronting) > S. Gelydh plural of S. Golodh “Noldo” (S/238; WJ/364).
Nonetheless, the diphthong [œi] could theoretically arise in Sindarin in cases where the raising of o did not occur but fronting to œ still happened. This could occur in words with primitive forms like -uCya, -oCya or -oCye, since the non-final y/ı̯ would not cause raising but would still cause fronting. All of these would develop through -oCı̯(ǝ) > -œCı̯ > -œiC, though the case of -uCya first involves a-affection of u > o. There are a few Sindarin words that seem to have these primitive forms, but they show varying vowels in the resulting Sindarin forms:
- S. ruin “fiery red” vs. Q. runya (PM/366), presumably derived from *runyā.
- S. fuir “north” (VT42/20), possibly derived from *phoryā as suggested by David Salo (GS/255).
- S. eryn “wood, forest” < ✶oronyē (PE17/119).
For the primitive form of the second example, compare an earlier word from the The Etymologies: N. feir < fœir vs. ᴹQ. forya “right (hand)”, both derived from the root ᴹ√PHOR which was the basis for words having to do with “north” (Ety/PHOR). Based on examples like fuir, Roman Rausch suggested one possible phonetic development of [œi] was to [ui] in his article On the Diphthongs ei, ai in Noldorin and Sindarin (DEANS/§4.5, 2008). However, Bertrand Bellet pointed out that ruin and fuir might have instead developed from *rūnya and *phōryā > *phūryā, and so didn’t involve [œi] at all (Vowel Affection in Sindarin and Noldorin, VASN, 2005). Elsewhere another Quenya cognate of S. ruin is given as Q. rúnya “red flame” (SA/ruin), which supports but does not confirm Bellet’s position.
Bellet didn’t suggest a Sindarin development for [œi]. However, based on the example ✶oronyē [> *œrœni(e) > *œrœin] > eryn (PE17/119), a second possibility is that [œi] > [y] might be the normal Sindarin sound change; this example was published after Bellet’s article. There is a Noldorin example that might be a precursor to this phonetic development: ᴹ✶ronyō > N. rhŷn “chaser, hound of chase” (Ety/ROY¹). But eryn also has an alternate etymology as an archaic plural form of S. orn “tree” from primitive oronī (PE17/33), making this phonetic development ambiguous as well.
A third possibility is that [œi] might become either [y] or [ui], depending on other factors. It is perhaps notable that eryn is polysyllabic and ruin, fuir are monosyllables. There is a similar y/ui variation in Sindarin plurals between polysyllables and monosyllables with the vowel o. Compare the following plurals:
- S. eryd plural of S. orod “mountain” (PE17/89; RC/765).
- S. Gelydh plural of S. Golodh “Noldo” (S/238).
- S. thuin plural of S. thôn “pine” (PE17/81).
- S. thuil plural of S. thôl “helmet” (PE17/186).
Although these plurals did not involve [œi], the same forces that lead to y/ui variations in plurals involving o might have led to a similar variation in the development of [œi]: [œi] > [ui] when stressed and [œi] > [y] when unstressed.
Thus we have three possible developments: (1) [œi] > [ui] always, (2) [œi] > [y] always or (3) [œi] > [ui] or [y] conditionally based on patterns of word length and stress. I like this third theory best myself, since it is analogous to the plural developments. In practice, though, we don’t really have enough information to draw any definite conclusions, because it isn’t certain that any of the Sindarin examples given above actually involved the diphthong [œi].
Note that Roman Rausch suggested a fourth possibility, based on a varying list plurals for S. og(o)l “evil”: egl, eigl, eigil (PE17/149). The middle plural hints that perhaps [œi] > [ei] as in Noldorin, and Rausch suggests this might sometimes have occurred in Sindarin as well (DEANS/§4.5). However, the numerous plural forms make it clear Tolkien was vacillating among several possible outcomes. Therefore, I hesitate to draw any conclusions from this particular example, especially since none of these plurals resemble the regular Sindarin plural pattern (where we would expect to see *egyl).