I think, in any event, had Tolkien lived, I don't think he would have let the first cousin thing stand, it's inconceivable. I thought he did it by accident, but on purpose!? Did he ever give any reason at all why he wanted them to be closely related?
Not that I know of.
He was in total control over his legendarium, I don't know why he agonized over so much of this, he could do anything with his material. He could solve any problem with the stroke of a pen.
Well, perhaps you don't mean 'anything' in any case, but I wouldn't say anything -- or at least not so easily. Already published text is a different animal, and in The Problem Of Ros for instance, we see Tolkien forgetting a detail already in print (while writing the solution), then realizing that 'most of this fails' when he remembered Cair Andros.
Since there seem to be many versions of the genealogies of Celeborn and Galadriel, I think that Christopher would be justified in using the one that did the least violence to the characters of those two Elves.
Hmm, why not leave it as it is? If I recall correctly, Celeborn's lineage is not noted in The Silmarillion beyond 'kinsman of Thingol' which was published by JRRT and served well enough for The Lord of the Rings.
(...) Christopher did a great job of collating the Silmarillion, it wouldn't exist without him, but maybe future editions of that work should be amended by him, or another Tolkien, to reflect some of the most glaring inconsistencies to be found in what was then published and what was written after. The Hobbit was greatly changed.
But The Hobbit was changed by Tolkien himself, to fall in line better with The Lord of the Rings; and here we are dealing with consistency with respect to two author-published works.
What should go into a one volume version of The Silmarillion is a very complicated question; but for example, I would argue that the version of Galadriel and Celeborn's history chosen by CJRT was correct, as it is the version which best agrees with Tolkien-published material.
If on the other side of the coin, I might agree that Gil-galad's parentage could have been left ambiguous. CJRT later thought he probably should have taken that path (as he chose for The Children of Hurin later, incidentally)...
... but even then, if Gil-galad's parentage might have been a 'mistake' even according to CJRT, another question might be: how problematic it is within the tale itself -- as opposed to knowing that Gil-galad as a son of Fingon was only one of Tolkien's ideas, and not his last -- which is an external consideration.
Generally speaking, I would say that the approach for the 1977 Silmarillion was not necessarily to try to incorporate every latest note or text (if the latest note or text could actually be determined with respect to a given matter), but to strive for a self-consistent version -- a version consistent with already published work -- or as 'consistent' as could be achieved in any case.
The 1977 Silmarillion is a version for readers. Not 'the' version, as that could not be known, but a version, and constructed (like the recent Children of Hurin) so it could be experienced as it was meant to be experienced. It's obviously not a scholarly presentation of disparate or evolving sources and ideas, which would provide a very different experience for the reader -- as this is the second presentation of The Silmarillion, as found in The History of Middle-Earth volumes.
They could at least say that Celeborn was born in Aman. This would not harm the Silmarillion, and would make it more accurate.
This would make it less accurate in my opinion, as Tolkien published (twice) that Celeborn was Sindarin.
Celeborn the Sinda of Doriath is a good example of a choice by CJRT that I very much agree with!