David Kiltz wrote:
> A first element _mor-_ (without final vowel) is also
> found in Q, e.g. _mormacil_.
> Maybe Tolkien later revised his idea of what form a first
> element _mor-_ originally had, and he really shifted to
In any case, _Moria_ seems to require a Sindarin _mor-_.
[Also _Morannon_, and no doubt other forms. CFH]
[I feel that I should interject here that we have to be careful not to treat even just all strictly Sindarin forms bearing on this matter as necessarily subject to the same phonological or morphological rules; far less so should we imagine Noldorin of the _Etymologies_ (or any other era) to be uniform with all Sindarin forms. There is, of course, the obvious fact that Tolkien coined these words at different times in his life, and thus at different points in his ever-shifting aesthetic and artistic conception of his languages, so there's no reason to think that his preferences in this matter were unchanged over even the later decades of his life. But the "internal" situation was also most likely variable as well, in at least two dimensions:
First, in the matter of compounding, it should be noted that compounds in "real" languages exhibit different phonological behaviors at the contact depending on the period of development in which they were formed. That is, compounds formed at a later period in the historical development of a language can and do behave differently than those formed at an earlier period; and there's no reason to think that Tolkien's languages -- which, after all, Tolkien was concerned to give the appearance of actual historical development -- differ from "real" languages in this regard.
Second, in the matter of consonant mutation (incl. lenition) or the lack thereof, it must again be reiterated that while these have phonologically-conditioned origins, once a pattern has been established as associated with a particular grammatical function (and the absence of mutation is as much a pattern for these purposes as is the presence), it could and did spread to other, grammatically-related forms in which historically the mutation would not otherwise have occurred. That is, we shouldn't necessarily be surprised to find mutation where the phonological situation doesn't support it, or its absence where we would expect it on phonological grounds.
Two examples spring to mind:
1) Tolkien notes of Gnomish _sithagong_ 'dragonfly' that it is a "new [compound] hence _-g-_" (PE11:68): that is, being a late composition, its second element _gong_ did not undergo lenition (to zero) as it would have in an historically old compound of (the ancestral forms of) the same elements.
2) Tolkien wrote in explaining N. _Eglador_ that "In old compounds the genetival element preceded (as in Quenya)" -- that is, obviously, in contrast with the situation in newer compounds (VT42:4).
It is clear that Tolkien himself did not imagine (or desire) that there be one and only one invariant rule for compound formation and consequent mutation throughout the history of his languages. We should be careful not to gloss over this feature of his languages in our studies. CFH]