[The Squatter of Amon Rûdh]

Yet why did he still use a word which would suggest and support this kind of false allegory?

It only supports the false allegory in the same way that Moria supports the idea of a biblical allegory, or Gondor suggests that Tolkien preferred open-plan offices. Readers can plausibly connect Moria and Morīah or Gollum and Golem as well. Then, since we're choosing only to notice the first syllables of words, there's Fangorn and fang, even Aragorn and arrogant. Perhaps Tolkien left in those references because he didn't think they were there in the first place; because he couldn't second-guess every association that might be made by every reader.

I don't like to keep bringing up Moria and Morīah, but Tolkien's comments on the proposed link bear heavily on the instance under discussion. The two place-names are far more similar than nazgûl and Nazi, and more readers would probably make the connection if more people read the Bible.

I utterly repudiate any such significances and symbolisms. My mind does not work that way; and (in my view) you are led astray by a purely fortuitous similarity, more obvious in spelling than speech, which cannot be justified from the real intended significance of my story.

Letter #297, August 1967 (emphasis mine)

What Tolkien says here as clearly as he can is that he simply would not deliberately make associations between words on the sole basis of a similarity in written form. It even seems to me that he was as irritated by the assumption that he would as I would be by the assumption that my use of the word route in a set of directions was intended to insult those directed by implying rout. This example has added complexities, since as I speak English the two words sound completely different, but I know that they could be homophones to an American.

Perhaps it would be useful to point out that philologists are very pedantic about accents, pronunciation and spelling. It's a survival trait in a field which has many pitfalls for the unwary, a good number of which concern the similarity of written forms. To pick a field in which Tolkien had much early experience, when deriving a dictionary etymology, a completely inaccurate result can be reached by the mistaken association of two words in different languages which look similar on the page. Tolkien's early history with the New English Dictionary, if not his earlier training under one of the greatest English philologists of his era, would have beaten out of him the fuzzy thinking that grasps at obvious connexions at first sight. Time and again he shows us his abiding interest in the meaning of a word, which is only natural given that in mediæval texts one can encounter anything up to a dozen different spellings of a single word without its meaning or pronunciation varying once, a stylistic feature that even affects proper nouns.

What I am trying to point out is that Tolkien is highly unlikely to have noticed a similarity between Nazgûl and Nazi, even though many readers have made that connexion. He was so accustomed to careful and precise reading that he would have found it very difficult to make the sort of associations that less thoroughly trained readers find equally difficult to avoid. Moreover he was more accustomed than any of us to regard the sound, written form, meaning and history of a word as aspects of a whole, all of which must agree for an association to exist. He spoke disdainfully of "an age in which almost all auctorial manhandling of English is permitted (especially if disruptive) in the name of art or 'personal expression'": in short, he was a man obsessed with precision and detail, so that when he wanted to give a name a hidden meaning he simply translated it into another language or an earlier version of English. Since he would have demanded complete consistency with contemporary politics if he intended his Ringwraiths to have any conceptual link with Hitler, and since his entire intellectual history had been leading away from superficial correspondences of form, I don't think that Tolkien would even have noticed this supposed link unless someone pointed it out to him. For all his howlers, I notice that Mr. Rang never did.

I think it important to remember at all times when reading Tolkien that his side of English studies is very different from the study of literature even today. In Tolkien's youth, philology was considered a science, not an art; and all of the empirical, systematic instincts of Victorian academic method are embodied in its practice. Victorian philology was not a matter of opinion, but of fact; it was not a matter of metaphor or allusion but of laws and their application; and it was a matter not of free expression or association but of solid logical reasoning and painstaking research. Not only can I not imagine Tolkien spotting some of the associations that have occurred to his readers (I have to say that I didn't see the link under discussion either), but I can't imagine any other philologist of his day seeing it either. Philology demands a different style of reading from that of literary criticism, and for someone like Joseph Wright, Walter Skeat or Kenneth Sisam a connexion between two words not supported by sound, sense or narrative context would have been no connexion at all.