I must highlight and commend the refreshingly thoughtful, descriptive
and linguistic response on Elfling today by David Salo to Tolkien's
own ideas on his own languages, in reply (and marked contrast) to the
dismissive response by Helge Fauskanger of the same.
In regard to the particular matter of the nature and use of _i_ in
Quenya as a complementizer, I would remind all parties of the
examination of the matter made 10 years ago (!) by Ivan Derzhanski in
_Vinyar Tengwar_ 38, in which he wrote e.g. (p. 17) that:
"the decomposition of _nai_, the marker of the Quenya periphrastic
'optative', into _nā_ + _i_ 'be it that' (R:68) suggests that _i_ is
in fact a complementiser akin to the English _that_ and introduces a
sentence rather than a verb phrase. The temporal connective _mennai_
'until' in the "Koivienéni Sentence" (KvnS, VT14:5) can be similarly
analysed as _menna_ + _i_ 'until the time/place/point that' (cf.
Italian _finché_ < _fin(o)_ + _che_). The KvnS also contains the
purpose construction _i erenekkoitanie_ 'that he might awake them',
which can be regarded as a sentence introduced by the complementiser
_i_, if the verb form can be shown to contain a subject marker. I
would suggest that the prefix _e-_, which is characteristically
present in all variants of the construction _i V-nie_ and absent in
all variants of the construction _na V_ in the "Koivienéni
Manuscript" (KvnM, VT28), may in fact be a subject marker. Its
distribution would then be explained by the fact that _i_
subcategorises for a sentence...."
As this shows, the use of _i_ as a complementizer 'that' has been
attested since long ago, nor was it regarded as problematic by
Tolkien, for it happily coexisted in his mind with the use of _i_ as
the definite article across many decades.