√GWAL “be stirred, excited; emotion, movement of feelings”
This root appeared in a couple different notes from the late 1950s, as {√GWAL} >> √WAL “emotion, movement of feelings” in one (PE17/189) and as √GWAL “be stirred, excited” in the other (PE17/154). However, its one Sindarin derivative balch “fierce, ferocious” (PE17/154) requires the primitive form be √GWAL, since it demonstrates the Ancient Telerin sound change whereby labialized velars became labials (gw > b). This is further supported by the prevalence of w-forms in its Quenya derivations, since primitive ✶gw- would have remained w- for longer in Quenya (PE19/75).
√(G)WAN “pale, fair”
A root appearing several times in notes written on or shortly before 1960 with a general meaning “pale, fair” (PE17/150, 154, 165, 189; WJ/383), as opposed to √BAN which was simply “beautiful”. It seems Tolkien introduced this root when he realized that “VAN cannot only = fair (blonde), since vanima is applied in LR to Arwen who was like Lúthien dark” (PE17/165). In the Quendi and Eldar essay from 1959-60 he used this root for a new etymology for the tribal name of the Vanyar referring to their fair, blond hair.
This root appeared as √GWAN, √WAN and √GWAY, though the note where the last of these appeared was marked through (PE17/154). With √GWAN the root would merge with √BAN in both Quenya and Sindarin: Q. vanya and S. bain “fair and beautiful” (PE17/154). But with √WAN the two would remain distinct in Sindarin, as in bain “beautiful” vs. gwain “fair haired” (PE17/150). I find the second paradigm more interesting, and thus recommend assuming the ancient root was √WAN for the purposes of Neo-Eldarin.
This root might be a restoration of an earlier separation of ᴱ√ɃANA versus ᴱ√WANA in the Elvish languages as Tolkien conceived of them in the 1910s, though the semantic divisions in the earlier conception were not the same. See the entry on √BAN for further discussion.
√GWEN “fair, beautiful; (probably originally) fresh, fair, unblemished (especially of beauty of youth)”
This root appeared in The Etymologies of the 1930s with derivatives having to do with freshness and greenness, as in ᴹQ. wenya “green, yellow-green, fresh” and Ilk. gwên “greenness”. In that document it was also the basis for N. bein “fair, beautiful” by way of blending with ᴹ√BAN “beautiful”. In The Etymologies Tolkien explicitly stated that ᴹ√GWEN was distinct from ᴹ√WEN(ED) “maiden”.
In later notes from 1959 this root reappeared with the gloss “fair, beautiful (probably originally fresh, fair, unblemished especially of beauty of youth)” (PE17/191). In this note √GWEN was this basis for “maiden” words, but Tolkien soon rejected this notion and reverted back to √WEN(ED); this may be a reflection of Tolkien’s ongoing vacillation between Q. wende versus vende as the proper Quenya word for “maiden”; wende (which seems to be his preference) would be better supported by an ancient root beginning with gw-, but that would also invalidate the Sindarin form gwend “maiden” (not **bend).
It is possible that the 1959 rejection of √GWEN was only for its role in “maiden” words, and that the 1930s sense “fresh” remains valid. That is what I would assume for the purposes of Neo-Eldarin.
ᴹ√GYEL “ringing, [ᴹ√] *cry of joy or triumph”
This root appeared in The Etymologies of the 1930s unglossed, but with derivatives like ᴹQ. yello “call, shout, cry of triumph” and N. gell “joy, triumph, (?victory)” (Ety/GYEL); the last gloss being unclear (EtyAC/GYEL). In The Etymologies, it replaced deleted ᴹ√GEL, also indicated by ᴹQ. ello >> yello. The root reappeared a couple decades later in a list of sound word as part of the set √GYEL, √ÑGYEL, √GYOL, √ÑGYOL collectively glossed “ringing” (PE17/138); see √ÑYEL for further discussion. For the purposes of Neo-Eldarin, I think it best to assume this root retained it 1930s meanings, which was probably something like “*cry of joy or triumph”.