In both Quenya and Sindarin, a stressed short wŏ became wa; this change did not apply to long wō. The most notable examples of this change was the (Sindarin) prefix wo- “together”, which became wa- and eventually developed into gwa- after initial [w] became [gw]. When unstressed, this prefix did not change its vowel and developed into go-, because later [wo] became [o] in Sindarin. A similar sound change occurred in Welsh (WGHC/§65v). Since this sound change occurred in both major branches of Eldarin, it likely occurred in Common Eldarin.

In Primitive Elvish the combination wo was fairly rare, so there aren’t many clear examples of this phonetic change, the best being the Sindarin prefix go-/gwa- noted above. However, Tolkien did explicitly describe this change a number of times:

• The form wŏ if stressed > wa in Eldarin (The Etymologies, 1930s, Ety/WŌ).
• When stressed the sequence wo was usually changed to wa [in Quenya]. A similar change occurs in Noldorin. (Outline of Phonetic Development [OP1], 1930s, PE19/53).
• When stressed the sequence wo was usually changed to wa [in Quenya]. A similar change also occurred in Sindarin ... the sequence wō remained unchanged. (Outline of Phonology [OP2], early 1950s, PE19/106).
• gwa- is regularly developed from *wo > *wa > gwa, when stressed in prehistoric Sindarin. (Quendi and Eldar, circa 1960, WJ/367).
• cf. wo to wa (notes on Quenya verb forms from the late 1960s, PE22/155).

The last note from the late 1960s is associated with a Quenya example of this sound change, where the verbal adjective suffix -oiti became -aiti when it appeared after w (PE22/155). A second Quenya example seems to indicate the opposite development), however ✶ñwa-lōth > Q. olos “inflorescence” (PE17/160), where ñwa- > ñwo- and then later the initial [ŋ] vanished and ultimately [wo] became [o] [@@@ discuss possible unstressed wa > wo]. A third potential example appears in The Etymologies of the 1930s:

• ᴹ✶nidwō > ON. nidwa, N. nedhw, ᴹQ. nirwa (Ety/NID).

Since we have final -wō > -wa in both Quenya and Noldorin [pre-Sindarin], this might be an example of this phonetic rule. However, the rule specifically applied to primitive short stressed wŏ, and this example is long wō (and quite possibly unstressed). Thus this last example could easily be some distinct phonetic development, as theorized by David Salo, who proposed that final -wō > -wā in the Common Eldarin period as a separate sound change (GS/§4.15).

As a final note, Tolkien mentioned that later S. gwa- became gwo- and then go- after stress shifted in Sindarin (WJ/368), indicating that wŏ́ > wa (or at least gwa- > gwo-) remained an active phonetic rule in Sindarin for some time. David Salo also recorded this later application of the phonetic rule (GS/§4.103).

Conceptual Development: The prefixal variation gwa-/go- dates all the way back to Gnomish, but as suggested in Roman Rausch’s Historical Phonology of Goldogrin (HPG/§1.4), in the earliest period this sound change seems to be reversed, in that unstressed wa > wo rather than stressed wo > wa:

[G.] go unaccented form < gwa-. [primitive] *ŋu̯a. cp. Q. ma- together, in one, etc. (GL/40)

This also seems to be true of the Early Noldorin of the 1920s:

[ᴱN.] go-. Originally unaccented form < gwá- = together, etc. (Q, T va-) (PE13/162)

Based on the Early Quenya and Early Telerin forms ma- and va-, this change to unstressed wo seems to apply only in Gnomish/Early Noldorin, not to the other languages.

By the 1930s, Tolkien revised the sound change to stressed wŏ́ > wa, and designated it as a Common Eldarin sound change that applied to both Quenya and Noldorin (Ety/WŌ, PE19/53). Base on notes appearing in the Outline of Phonology (OP2) given above, this seems to still be a Common Eldarin sound change in the 1950s (PE19/106).

GS/§4.15 GS/§4.77 GS/§4.103 WGHC/§65v @@@

HPG/§1.4, HPG/§2.8