At various points Tolkien indicated that initial palatalized dentals (those with an additional y-sound) lost their palatalization and became simple dentals in the Telerin language family. The clearest description of this change is from a discussion of the development of primitive initial combinations in the Comparative Tables of the 1930s:

[Initial combinations] ty, thy, dy, ny, ly only appear in Lindarin (Quenya), as ty, hy, y, ny, ly respectively. Telerin, Noldorin, Ilkorin, Danian have plain dentals to correspond (PE19/22).

There are a number of examples of roots beginning with labialized dentals in the Etymologies, and they all show this development in Noldorin, for example: ᴹ√DYEL > N. deleb “horrible” vs. ᴹQ. yelwa “loathsome” (Ety/DYEL); ᴹ√NYEL > N. nell “bell” vs. ᴹQ. nyelle (Ety/NYEL); ᴹ√tyalı̯ā- > N. telio “to play” vs. ᴹQ. tyal- (Ety/TYAL). The last of these examples demonstrates that the change did not occur medially, since the y-sound survived after medial l; other ya-verbs like N. nœdia- and N. egledhia- demonstrate the preservation of the y-sound for other medial dental+y combinations.

By the 1950s, however, Tolkien seems to have abandoned the possibility of initial palatalized dentals in Primitive Elvish. In the first version of Tengwesta Qenderinwa from the 1930s, palatalized dentals appear in the list of possible primitive initial groups (PE18/44) but they are missing from the second version of Tengwesta Qenderinwa from the 1950s (PE18/93-94). Some of the roots with labialized dentals were changed to labialized velars instead, for example: ᴹQ. tyulma “mast” < ᴹ✶tyulmā (Ety/TYUL) but with later primitive form ✶kyulmā (PE21/74); ᴹ√NYEL “ring” (Ety/NYEL) but later root form √Ñ(G)YEL (PE17/138, PE18/94). Since primitive combinations like [ky-] and [ty-] produced the same result (ty-) in Quenya, such root changes could have been made universally without affecting Quenya at all. Not so with Sindarin.

There are, in fact, a few remnants of the older concepts appearing in Sindarin, such as S. narn “tale” (S/198) vs. Q. nyárë (MR/143). If Tolkien had revised the root from √NYAR > √ÑYAR, the Sindarin form should have become **iarn, or possibly **garn if Tolkien re-derived it from strengthened [ŋgjarn-]. Furthermore, there seems to be no systematic attempt by Tolkien to revise old Noldorin forms that were originally derived from palatalized dentals when he changed Noldorin to Sindarin. It is unclear how he might have revised such forms, or if he ultimately would have contrived some way to preserve them.

Conceptual Development: In the earliest period, Primitive Elvish had only a single palatal series, which became labialized dentals in Early Qenya but became simple velars in Gnomish, such as: ᴱ√DYULU [IPA ɟulu] > G. gaul “burden” vs. ᴱQ. yulu- “to carry” (GL/38); ᴱ√TYAVA [IPA caβa] > G. caf- “to taste” vs. ᴱQ. tyava- (GL/24, QL/49). As such, the Gnomish palatal developments match the later Sindarin development of primitive labialized velars.

In the 1920s, however, Tolkien introduced the idea of primitive palatalized dentals distinct from palatalized velars, allowing for different phonetic developments (PE14/64-66), though the first version of this system was quite free-form and included many combinations Tolkien would ultimately reject. Lack of examples makes it hard to determine how Tolkien initially imagined such sounds would develop, but ᴱN. teilian “play; mockery” vs. ᴱQ. tyalien hints that he quickly adopted the Noldorin paradigm described above, and he definitely did so by the 1930s.

As noted above, the gradual abandonment of primitive initial palatalized dentals (possibly as early as the 1940s but definitely by the 1950s) rendered this sound change moot. There are also indications that even in the 1930s, Tolkien considered making palatalized dentals a Quenya-only innovation, completely absent from the other languages:

Dentals + y and w are often included here, but their case is not clear. In Qenya (and Lindarin) ty, hy, ny often appear as variations or modifications of t, s (of any origin), n [Tolkien’s sidenote: And so corresponding to T. and N. t, th, n, s]; while Q. also possessed the initial group ly (infrequent). These forms were probably not of CQ origin, but arose in L. and Q. after their change of k-y > t-y, n-y > ny, etc. There is no trace of such initials as ty- in T. or N. It is said tj, dj, nj etc. occur in certain Avarin dialects, but this is probably a quite independent development, though it may be of some antiquity (PE18/44).

Neo-Sindarin: For purposes of Neo-Eldarin, I think it is best to ignore Tolkien’s removal of initial palatalized dentals from Primitive Elvish. Removing these primitive sounds introduces problems with the etymologies of several well-established words, such as S. narn “tale” and N. telia- “to play”. If primitive initial palatalized dentals are retained in Neo-Eladrin, it is likely this phonetic change would have occurred in the Ancient Telerin period, as suggested by David Salo (GS/§4.30, 4.31).

GS/§4.30 WGHC/§100ii-1 [j vanished after initially consonants early] @@@