Generally in Quenya, any isolated w developed into v, but this change occurred in two waves. The first wave was when original primitive w > ƀ [β] and occurred in Parmaquesta [PQ], as discussed in the entry on how [w] became [β] initially and between vowels. Any primitive gw survived this first wave, since at the time they remained spirantal ʒw [ɣw]. This entry discusses the second wave which occurred in Tarquesta [TQ]. After [ɣ] from [g] vanished, these remaining w derived from gw also developed into v, as described in the Outline of Phonology (OP2) from the 1950s:

gw became [ʒw], and then initially a spirantal frictional w (with slight friction both at the back and at the lips); but already in PQ this w was weakened. The weakening was later than the beginnings of the change of original w > v, and w < gw remained as such and distinct from w, v for some time: in earlier PQ it is not represented by the sign for v; but in TQ it had become identified with w, v (OP2: PE19/75-76).

In notes from the late 1950s or early 1960s, Tolkien indicated this was a Ñoldorin-only change:

In this respect EQ [Exilic Quenya] represented the recognition of sound-changes which had begun among the Noldor before the Exile and had already caused Noldorin Quenya to diverge from the language of the Vanyar. The principal of these were: ... the change of w to v (except in nw, tw, hw, kw, ngw) ... (PE17/129).

He did not specifically declare that the “w” in question was limited to the ones derived from gw, but that would make the most sense in light of his notes elsewhere.

As a further indication of the relative lateness of this change, Tolkien continued to spell many of these words with w instead of v (or in some cases, vacillated on the spelling):

• √GWAL > [g]walna > Q. walda “excited, wild” vs. S. balch “fierce, ferocious” (PE17/154).
• S. gwing vs. Q. wingë “foam” (PM/376; Ety/WIG), sometimes ving- as in Q. Vingilótë “Foam-flower” (S/246).
• S. gwend vs. Q. wendë “maiden” (PE17/191; Ety/WEN), sometimes vénde (VT44/10).
• N. gwilwileth vs. Q. wilwarin “butterfly” (MC/222; Ety/WIL), sometimes vilvarin (PE16/72; MR/166).

The Sindarin/Noldorin forms provide additional evidence that the etymology of these words involved gw, but they present a problem as well. If, in fact, these words began with gw in Primitive Elvish, then the proper Sindarin/Noldorin forms should show gw- > b-, as in S. balch above, but many of the examples show S. gw-. Perhaps the initial gw developed in these words after the Ancient Telerin sound shift (gw- > b-), but in a period before the second sundering when the Elves were still in contact with one another.

Other evidence for the relative lateness of this shift can be found in Tengwar usage. In The Lord of the Rings Appendix E, Tolkien said that the tengwa vilya “air” (n) originally has the pronunciation wilya (LotR/1123); however where it appeared in tengwar writing it was used to represent [w]. This word is almost certainly related to wilwarin “butterfly” with the same phonetic development gw > ʒw > w > v, and furthermore the last change was late enough that the tengwa n was still used to represent [w] in many contexts.

Furthermore, it is likely that tengwar Grade 6 containing n (vilya) was originally a set of voiced spirants: 6 y h n = ð ƀ ʒ ʒw ([ð] [β] [ɣ] [ɣw]): that is dental, bilabial, velar and labialized-velar voiced spirants. Ultimately these came to represent 6 = medial [r], y = [v] and n = [w], with h (anna) losing most of its original function and being repurposed to represent various vowel and semi-vowel sounds, such as [o] in Sindarin.

This begs the question of how [w] was represented in tengwar before these sound changes. Somewhat cryptically Tolkien said that “[tengwa] 35 [l] and 36 [.] were, when used as consonants, mostly applied to y and w respectively (LotR/1121).” This is never the case in any tengwar samples, where they are used exclusively to represent the final vowel in diphthongs, [ı̯] or [u̯] respectively, such as: au = .E. However, their use as consonants might describe their old [PQ] usage: perhaps these were the original representation of consonantal y and w.

If the tengwa . (úre) = [w] in early PQ, this function would have been largely unnecessary after primitive w merged with ƀ < b: the tengwa y (vala) could have been used to represent this new pronunciation. The tengwa n (wilya) = ʒw would have been the only remaining “w like” tengwa, and was probably repurposed to represent this sound in other combinations like lw = jn or tw = 1n. It seems to have retained this function even after ʒw > w > v. See for example, Tolkien’s writing on the Feanorian Alphabet from the 1940s where n was used for [w] (PE22/16, 43), though in this conceptual period the tengwa name was winge (PE22/22, 51). See also the Old Valinorian use where . was used for [w] (PE22/13 note #22, replacing deleted n).

Finally, Tolkien noted that tengwa #22 came to represent v in all cases: “Quenya did not possess dh, gh, ghw, and for v used letter 22” (LotR/1032). Thus it seems:

• In early PQ . (úre) and l (yanta) were used for the semi-vowels [w] and [j] as both consonants and in diphthongs.
• In late PQ after w > ƀ, n (ʒw) was gradually repurposed for consonantal [w], a trend aided by ʒw > w.
• Likewise . and l came to be used only for diphthongs.
• After the remaining w > v, n (vilya) was still used for [w] combined with other consonants.
• However, y (vala) was used to represent v derived from primitive b and w, including those derived from ancient gw.

All of this was aided by the fact that the most common w-clusters had dedicated tengwar, such as z = [kw] and b = [ŋw]. As such, these tengwar would not require any changes as the writing system switched from using . to using n as the sign for [w]. Note that a similar change occurred in tengwar usage for y [j]. The tengwa l (yanta) gradually came to be only used for diphthongs, and the combination hÍ (anna with palatalization tehta, originally ʒy) came to be used for [j] after ʒy > y.

Some of the above is speculative, but these conclusions fit the available evidence, and indicate the depth of Tolkien’s thinking on the interaction of various phonetic developments with the use of the tengwar writing system. Similar developments occur in real-world languages: the history of English “w” (double-“u”) has its own interesting complexities.

Conceptual Development: This sound change was not relevant in the 1930s, since ʒw > w before w > v, so that ʒw and w merged before it became a labial spirant. See the entry on how [w] became [β] initially and between vowels for further discussion.

Neo-Quenya: It is common in Neo-Quenya to avoid all these complexities simply by changing all initial w- to v- in (Neo) Quenya words, a practice followed in Eldamo as well. However, the late retention of w from gw is a useful excuse for retaining initial w in the Quenya words where it appears if you wish to mimic Tolkien’s own writing, where w was often retained. This could be representative of classical spelling, with either [w] or [v] being viable as pronunciations. As noted above, derivation from gw may not actually be the case for all such words, however, especially those that show initial gw- in Sindarin.