I agree in general, but "always" and Tolkien can be a difficult pairing.
And what we sometimes have in
The Letters of JRRT, for examples, are readers pointing out difficulties with
author-published ideas or statements. Yes, Tolkien usually looks for what I call internal explanations -- the idea being, not error-by-author, but seeming discrepancy because the translator has more material to draw from than the reader knows about. I often engage in this myself, in threads or in my head at times...
... and Tolkien even sometimes appears to treat "private draft material" (from his point of view) as if already published, and tries to find an internal answer.
But that said:
Quenta Silmarillion was still open to drastic revision, and if, in 1958, Tolkien thinks that Balrogs being Maiar might be problematic if they existed in the thousands -- despite the noted strengths of the First Age (or problematic for whatever reason) -- he is very free to make this revision.
Nothing about Balrog numbers had been published, and obviously JRRT is not bound to private writing, or even a given letter in my opinion. JRRT ultimately dropped his
long held idea about how Elves were reincarnated -- a change he was free to make given what had been published about this... interestingly perhaps, even here Tolkien "holds on" to the old reincarnation idea by noting not simply that it is false, but that it might be noted in the legendarium as a false Mannish idea. Thus it still will arguably find its way into print (internal in one sense), despite it being no longer true internally.
And I could use the same argument against me with respect to the Glorfindel case I referred to above, and I (the other me) would have to at least concede that the idea of Glorfindel defeating a Balrog does not appear in anything Tolkien himself published...
... if it had, my argument (the other me again) would have been arguably easier!
