Some notes that were scribbled down at Sidmouth in Devon in the late summer of 1938 (Biography, p. 187) on a page of doodles evidently represent my father's thoughts for the next stages of the story at this time:The Return of the Shadow, 214-215
Consultation. Over M[isty] M[ountains]. Down Great River to Mordor. Dark Tower. Beyond(?) which is the Fiery Hill.
Story of Gil-galad told by Elrond? Who is Trotter? Glorfindel tells of his ancestry in Gondolin.
... Very notable is "Glorfindel tells of his ancestry in Gondolin". Years later, long after the publication of The Lord of the Rings, my father gave a great deal of thought to the matter of Glorfindel, and at that time he wrote: "[The use of Glorfindel] in LotR is one of the cases of the somewhat random use of the names found in the older legends, now referred to as the Silm, which escaped reconsideration in the final published form of The Lord of the Rings." He came to the conclusion that Glorfindel of Gondolin, who fell to his death in combat with a Balrog after the sack of the city (II. 192-4, IV.145), and Glorfindel of Rivendell were one and the same: he was released from Mandos and returned to Middle-earth in the Second Age.
["Trotter" was the original name of the mysterious stranger later called "Strider" (who at this stage of the composition was a hobbit); II and IV refer to other volumes in the HoMe series.]
A number of reasons have been advanced for not taking this at face value. Since Christopher's report of Tolkien's conclusion was not part of the rough drafts, the question of whether rough drafts can be canonical does not arise in this case. The suggestion that lack of premeditation is grounds for rejection also seems inadequate, since many elements were introduced with little thought of future consequences yet later became important parts of the mythos.
It is true that we have no examples of any other elf journeying eastwards to Middle-earth during the Second Age (though some did visit Númenor), but this is not enough to disprove the possibility of Glorfindel's having done so. There were in fact no direct statements either way, which means that Tolkien could have established whatever background he wanted to any story he might have written. The previous lack of specific information on this matter was no constraint.
The strongest objection is that the way Christopher presents this inspires less confidence than it might because he doesn't provide any direct quotes -- rather, he merely describes a "conclusion" that his father eventually "came to". Evidently, Tolkien never actually wrote his conclusion down. The matter therefore reduces to a question of how much one trusts Christopher, and whether one supposes that he might attach too much importance to a casual statement. The majority of readers appear to accept that this was indeed a thoughtful conclusion that Tolkien reached only after long deliberation (we do know that he and Christopher discussed the matter of Middle-earth often). A significant minority continue to reject it.
In the last analysis, of course, certainty either way is impossible, since no evidence beyond the above exists. On the one hand, we can at least say that Tolkien apparently saw no objection to the idea that a re-embodied Glorfindel could have returned. On the other hand, the usual caveats concerning unpublished material are even stronger than usual in this case, since he not only might have changed his mind before publishing but also might have done so before he wrote the story, or while he wrote it (not an unusual occurrence). Still, there seems a good chance that he would have stuck to the one Glorfindel idea, since he seems not to have come to the decision lightly.