S. medial [j] became [i]; [-j-] > [-i-]
In Sindarin and Noldorin any surviving medial -y- [j] became i at some stage. This is a factor in Sindarin orthography, where i always represents the vowel [i] medially, but at the beginning of a word before another vowel represents the “y”-sound [j], because it is the only place where this sound can appear:
I initially before another vowel has the consonantal sound of y in you, yore in Sindarin only: as in Ioreth, Iarwain (LotR/1114).
Tolkien described this change in his notes on the Noldorin usage of the Feanorian Alphabet from the 1930s:
This change in use was due to the loss of all spirantal quality by ON y = [j] in Exhilic, and to its frequent conversion to vocalic [i] (PE22/32).
The clearest indication of this sound change are the Sindarin verbs ending in -ia, which generally correspond to primitive -yā and Q. -ya:
- ✶PHERE > S. feria- “make ready” vs. Q. ferya- (PE17/181).
- ᴹ✶baryā́- > N. beria- “to protect” vs. ᴹQ. varya- (Ety/BAR).
It’s easiest to assume this change took place before i-affection, since the resulting forms invariably show this sound mutation wherever it is appropriate. However, some attested developmental series hint that the actual set of sound changes were a bit more complex:
- ✶eryā > erı̯a > erı̯ > irı̯ > S. îr “*alone” (VT50/18).
- ✶lisyā > liχı̯ā > leχı̯ > leich > S. laich “sweet” (PE17/148).
These two examples indicate that y first became asyllabic ı̯, and may not have always developed into a full vowel, especially in cases of later i-intrusion. From a purely symbolic perspective, however, it is easiest to represent this change as [-j-] > [-i-]. Elaran suggested this approach to me in a private Discord chat on 2018-08-25.
Conceptual Development: In Early Noldorin, examples like ᴱN. leithia- “to release” (LB/154) and ᴱN. rhoidia- as well as i-intrusion in adjective forms like ᴱ✶mburyā́ > ᴱN. boir “hot, raging” (PE13/139) and ᴱ✶nindyā > ᴱN. nainn “blue” (PE13/150, 164) make it clear that this sound change was a part of the language’s phonological history in the 1920s.
There is not much clear evidence of this change in Gnomish of the 1910s, but since y only appears medially as vowel, it seems likely this sound change was present even at the earliest conceptual stages, despite the fact that Gnomish did not include the later system of i-affection that was so influential in Noldorin and Sindarin.