Who wrote The Silmarillion? It’s a question with a more complicated answer than it seems on the surface. Yes, of course, J.R.R. Tolkien wrote the book that I picked up from a Barnes & Noble fourteen years ago, that is now on my desk with its cover coming off and its corners rounded from being read so many times. But who, in the vast imagined world within its pages, is telling the story? The narrator of The Silmarillion is so distant as to be barely discernible at all; it is possible to believe he doesn’t exist at all. Indeed, in at least my first two readings, I did not think much of him. I assumed a distant, omniscient presence recounting what happened in plain, incontestable terms. Just the facts, ma’am.

The fact is, though, that J.R.R. Tolkien–the Silmarillion author whose name is on the cover–always imagined and constructed his stories not just as stories but as historiography: documents written by someone within the universe in which the history occurs. This complicates things: gone is the distance, the omniscience, and perhaps most importantly, the impression that the stories happened exactly as they are told.

This wouldn’t be a problem–in fact, would be quite simple, as most fiction has point of view that is biased or unreliable–but for the fact that this isn’t simple: This is Tolkien. So naturally, he had this idea that he wanted to write his stories as historiography, with a loremaster or chronicler who was himself a part of that history, but he couldn’t make up his mind who this person was. In fact, he changed his mind several times, reversals that are documented in The History of Middle-earth for fans and scholars to angst and argue over.

I’m going to make the case that the “Quenta Silmarillion” is part of the Elvish tradition. This is contrary to the belief of Christopher Tolkien and other Tolkien scholars, who assign it to the “Mannish”–namely Númenórean–tradition. (Ingwiel has an excellent discussion of the evidence for this approach.) I understand why they did this, but I think that if you look deeply at the texts and the evidence those texts provide, there is not much to support that the tradition originated with the Númenóreans. (I am willing to concede that Elvish texts may have passed through Númenórean hands on their way back to Elrond and, eventually, Bilbo, but I persist in believing they are nonetheless predominantly Elvish texts representing an Elvish point of view.)

The Idea of the “Mannish” Tradition

The Elvish loremaster Pengolodh was first introduced as the primary author of the “Quenta Silmarillion” prior to 1930, when he was assigned author of the Annals of Beleriand (HoMe IV). Pengolodh was a tenacious character: Texts written as late as 1960 were still being assigned to him. So what happened?

The idea of the “Mannish” loremaster was a late idea and introduced as part of the series of writings collected by Christopher Tolkien under the title Myths Transformed (HoMe X). In a text that CT dates to 1958, Tolkien writes:

It is now clear to me that in any case the Mythology must actually be a ‘Mannish’ affair. … The High Eldar living and being tutored by the demiurgic beings must have known, or at least their writers and loremasters must have known, the ‘truth’ (according to their measure of understanding). What we have in the Silmarillion etc. are traditions (especially personalized, and centred upon actors, such as Fëanor) handed on by Men in Númenor and later in Middle-earth (Arnor and Gondor); but already far back–from the first association of the Dúnedain and Elf-friends with the Eldar in Beleriand–blended and confused with their own Mannish myths and cosmic ideas. (Myths Transformed, “Text I,” emphasis in the original)

To summarize: in 1958, Tolkien began to deeply question whether a civilization as advanced as that of the Eldar–a civilization that also had access to the teachings of the Ainur, who knew firsthand the structure of the universe–would produce myths that included such components as a flat Earth and the “astronomically absurd business of the making of the Sun and Moon” (“Text I”). This led to some radical cosmological rearrangements in Myths Transformed–and the relatively overlooked decision to reimagine the Silmarillion histories from a Mortal rather than an Elvish perspective. In an undated text also presented in Myths Transformed, Tolkien again takes up this question and explains the method of textual transmission in greater detail:

It has to be remembered that the ‘mythology’ is represented as being two stages removed from a true record: it is based first upon Elvish records and lore about the Valar and their own dealings with them; and these have reached us (fragmentarily) only through relics of Númenórean (human) traditions, derived from the Eldar, in the earlier parts, though for later times supplemented by anthropocentric histories and tales. These, it is true, came down through the ‘Faithful’ and their descendants in Middle-earth, but could not altogether escape the darkening of the picture due to the hostility of the rebellious Númenóreans to the Valar. (“Text VII”)

“A leading consideration in the preparation of the text was the achievement of coherence and consistency,” Christopher Tolkien wrote in a note on The Valaquenta in The Later Quenta Silmarillion II (HoMe X), “and a fundamental problem was uncertainty as to the mode by which in my father’s later thought the ‘Lore of the Eldar’ had been transmitted.”

Christopher Tolkien tentatively dates The Later Quenta Silmarillion II (LQ2) to 1958. Along with the last set of annals, The Annals of Aman and The Grey Annals, this represents the final version of the Silmarillion that Tolkien produced. (See “Note on Dating” at the end of LQ2 in HoMe X.) Interesting about these texts–especially LQ2–is the fact that Tolkien removes all attributions Pengolodh. Mentions of Pengolodh are sprinkled throughout the Later Quenta Silmarillion I, which was written around 1951-52. When Tolkien “remoulded” LQ1 into LQ2, he removed Pengolodh. Also written around 1958? That first Myths Transformed text in which Tolkien asserts that his cosmology requires a Mortal and specifically bars an Eldarin loremaster.

And the Evidence of the Elvish, Part 1: The Creative Process

All this probably seems very simple. Tolkien was clear on his intentions. The Elvish tradition doesn’t work, in his opinion. Therefore, it must be Mortal. He even took out the Elvish loremaster from the oldest Silmarillion draft. Simple, right?

Never. Working with any of the Silmarillion material–including the published Silmarillion–is necessarily speculative. This is a posthumous text, unfinished and existing in many forms. I think Christopher Tolkien did an admirable job of making a published book out of the tangle of his father’s writings, but making that book required making decisions, as alluded to above, about how to decide what to include.

On this particular question, there are two approaches to making a decision on mode of transmission: There is Tolkien’s stated intention, and there are the texts themselves and what they show of the realization of that intention. CT, and many scholars, clearly prefer the first approach. Tolkien was clear on what he wanted, so that’s the way to read the texts.

I prefer the second.

Perhaps this is because I approach the welter of Silmarillion texts as a creative writer as well as a Tolkien scholar. My experiences as an author of fiction myself lead me to question whether the creative process lends itself to the kind of neat analysis that says, “The author stated his intention. Here we have our answer.” My experience tells me it is rarely that simple.

Below is what I imagine the creative process looks like for worldbuilding and constructing stories in that world, done in clipart and scribbles. For me, most of my work goes on in my mind: while driving to work, falling asleep at night, reading other authors’ work, washing dishes, daydreaming. Some of the thinking is intentional, other occurs because it’s where my mind wanders where I’m bored. Sometimes, thinking is sparked by an outside stimulus: an interview on the radio, a song, an image, a clip from a movie or TV. All of it goes into this tangle of thought constantly swirling in my mind, making and remaking my imagined world. Every now and then, an idea leaves my mind and takes concrete form as I write it down.

But only in limited instances do these ideas become finalized, incontrovertible–“canon,” if you will. Sometimes an idea won’t work and withers, unfinished. Other times, an idea is written down, only to dive back into the welter of thought in my mind for further reworking and reshaping–sometimes radically so.

Every author’s creative process is different, of course, but hold up The Tale of the Sun and the Moon from the Book of Lost Tales next to the story Tolkien writes in Myths Transformed and the two are radically different, showing what any Silmarillion fan can tell you: Just because Tolkien wrote it down doesn’t mean he meant it. Likewise, he went years at times without working on the legendarium, yet his letters and the progress we observe in the drafts show that he was always thinking about it. In short, his creative process, in this regard, seems a lot like mine.

Around 1958, we can say with some certainty that an idea crystallized from Tolkien’s thoughts about his legendarium that the mode of transmission had to be centered on Mortals, not Elves. The idea seems to have loomed large in his awareness–along with ancillary ideas about cosmology–to the extent that it appears to reflect in decisions he made in revising LQ2.

But does that mean it is definitive? That it is “canon”? Not necessarily. In fact, I’d argue that we have proof that this particular manifestation of an idea was one that was far from finalized but dove back into the swirl of thinking on worldbuilding to be reconsidered and reworked–and ultimately unrealized.

And the Evidence of the Elvish, Part 2: Point of View

Point of view is no small thing in a story. In fact, in all but a few cases, it is so essential that to change the point of view risks breaking the story in a way that changing other elements rarely does. It’s like painting. If you paint from the point of view of a peasant looking at a castle from the field where she labors, you cannot suddenly decide that the point of view is that of the princess looking out from the room in the castle where she spends most of her days, at least without redoing the painting entirely.

Likewise, one cannot take a story written from one point of view, then suddenly decide to change to a different point of view with any guarantee that the story will still work, much less make sense, without rewriting the story. In fact, in many cases, it will not.

Changing from an Elvish to a Númenórean point of view is not so simple as declaring, “Let it be!” and there it is. In the case of the “Quenta Silmarillion,” the Eldarin (specifically Gondolindrim) perspective is deeply embedded. Grundy’s post here is a good run-down of how the narrator’s affiliation with Gondolin is revealed, even if never stated, in the stories included as part of the “Quenta.” My article Attainable Vistas looks at some of the numerical data I’ve compiled that suggests a Gondolindrim perspective. At last year’s Tolkien at UVM Conference, I presented more of that data, as well as new evidence that even the narrator’s language in the “Quenta,” reveals the point of view of a loremaster from Gondolin. Tolkien didn’t put Pengolodh’s biography down on paper until the 1959-60 text Quendi and Eldar, but the texts suggest that Pengolodh’s identity was swirling in his mind many decades before that, and he wrote the “Quenta” with that point of view always in mind. Changing the point of view of such a story requires significant rewriting of the text. Do we have evidence that any of that rewriting–short of striking Pengolodh’s name from LQ2–occurred?

No, we do not.

In fact, we sometimes see the opposite.

And the Evidence of the Elvish, Part 3: The Texts

The Later Quenta Silmarillion II becomes a relevant text to examine here because 1) it was written at the time when we know Tolkien was thinking about the mode of transmission and 2) his striking of Pengolodh’s name from this version suggests he was beginning to act on his stated intention to revise the “Quenta” to reflect a Númenórean point of view. It is also an interesting text to study because it is a revision of LQ1, written about seven years earlier when the mode of transmission was, as far as I can tell, unreservedly Elvish.

Does the LQ2 contain other revisions toward a Númenórean mode? No, it is does not. In fact, it includes additions that, from a Mortal perspective, are suspect.

Laws and Customs among the Eldar. One of two major additions to LQ2 was Laws and Customs among the Eldar (L&C). L&C is explicitly attributed to Ælfwine, the Mortal man who was part of the mode of the transmission involving Pengolodh. Ælfwine is Anglo-Saxon, not Númenórean, but L&C is clearly written from the point of view of a Mortal commenting on Elves.

L&C opens with the sentence, “The Eldar grew in bodily form slower than Men, but in mind more swiftly.” This comparison immediately establishes a Mortal point of view different from that of the “Quenta” as a whole, where Mortals are usually but supporting actors in a drama enacted by Elves. The first two paragraphs continue this comparison and assume the distinct point of view of a Mortal. Later, in the section “Of Naming,” the narrator notes that the variety of names used by a single Elf “in the reading of their histories may to us seem bewildering,” again establishing a Mortal point of view (emphasis mine).

L&C is an example of what a text written from a Mortal point of view would look like. “Men are really only interested in Men and in Men’s ideas and visions,” Tolkien wrote in Myths Transformed, and L&C acknowledges this by bringing Elvish customs into the context of Mortal experience (“Text I”). This text shows that Tolkien did manipulate the point of view based on his narrator. (This won’t be shocking to any writer of fiction, and as an Anglo-Saxonist, Tolkien would have been aware of the impact of point of view on a historical text from that perspective as well.)

Christopher Tolkien dates L&C to the late 1950s (LQ2, “Note on Dating”). The argument could be made that Tolkien hadn’t yet begun the process of revising to incorporate a Númenórean narrator. However, of all of the texts in LQ2, L&C is the easiest to revise to change the point of view. It is already from a Mortal point of view! Simply change the attribution to Ælfwine to a suitably Númenórean name and you have a major chapter of LQ2 aligned with the Númenórean mode of transmission. It is a surface change on the order of removing attributions to Pengolodh, and the fact that Tolkien didn’t undertake it makes me question how seriously he truly undertook to revise the point of view.

The Statute of Finwë and Míriel. The Statute of Finwë and Míriel is the second major addition to LQ2. It is likewise dated to the late 1950s (LQ2, “Note on Dating”). If Tolkien wanted to write a text representing a Númenórean point of view, he couldn’t have done a worse job of it with the Statute of Finwë and Míriel. Here we have a text deeply concerned with eschatology: Elven eschatology.

The Númenóreans were also concerned with eschatology. You could even say the Númenóreans were obsessed with eschatology, and immortality in particular. Here’s a people, after all, annihilated because of their king’s obsession with an Oasis song proclaiming, “You and I are gonna live forever!” A Númenórean text that represents Elven eschatology with no commentary grounding it in a Mortal perspective (like Tolkien does with L&C) is almost impossible to fathom. A text that centers on immortality and the decision to forgo immortality would certainly excite commentary from a Númenórean loremaster. Revisions to represent a Númenórean point of view would have to address this chapter–but they don’t. Again, this suggests that Tolkien’s ideas about the mode of transmission weren’t as definitive as his writings in Myths Transformed suggest.

And all the rest …? Outside of L&C, the remainder of LQ2 includes nothing that suggests a Mortal point of view, even though L&C shows that Tolkien was capable, with the addition of a few words, of beginning to establish this. A convincing Númenórean text would have needed deep revisions, but surface changes–like the deletion of Pengolodh–set the course in the right direction.

Throughout the “Quenta,” Tolkien often uses formulas like “it is said,” “it is told,” and “it is sung” to indicate that the narrator is receiving information secondhand versus as an eyewitness account. As I discussed at the Tolkien at UVM Conference in 2017, these formulas are used primarily to refer to the Ainur and Mortals. The image to the right shows the data from one of my slides from this presentation. Adding these formulas is a rather easy way to signal that the information the narrator is reporting is at a distance from him–and LQ2 uses them when reporting on the actions of Ainur to which even an Eldarin narrator could not have borne witness–yet Tolkien did not make these revisions. Instead, this part of the Silmarillion (originally attributed to Rúmil, who would have been present for this history) is written in the style of an eyewitness account, even though a Númenórean loremaster would certainly find these chapters of history the most distant and unattainable. Yet nothing in the style in which these chapters are written suggest this.

There is one passage in particular in the chapter “Of the Silmarils and the Unrest of the Noldor” that, like the Statute of Finwë and Míriel, seems a ripe opportunity to make revisions to signal a Mortal narrator if Tolkien desired to do so:

In those days, moreover, though the Valar knew indeed of the coming of Men that were to be, the Elves as yet knew naught of it … but now a whisper went among the Elves that Manwë held them captive so that Men might come and supplant them in the dominions of Middle-earth. For the Valar saw that this weaker and short-lived race would be easily swayed by them. Alas! little have the Valar ever prevailed to sway the wills of Men; but many of the Noldor believed, or half-believed, these evil words. (emphasis mine)

It is hard to imagine a Mortal loremaster self-identifying as “weaker and short-lived” or offering up such an assessment without commentary. Yet this passage was brought over from LQ1–when the narrator was Elvish–without revision.

Another small passage again betrays the Elvish perspective: “But now the deeds of Fëanor could not be passed over, and the Valar were wroth; and dismayed also, perceiving that more was at work than the wilfulness of youth” (“Of the Silmarils,” emphasis mine). According to the timelines in the Annals of Aman, Fëanor was about three hundred years old at this point. For a Mortal narrator–even a long-lived Númenórean one–to describe him as a “youth” is difficult to fathom.

Now one might claim that perhaps Tolkien just didn’t get the chance to add material. It is one thing to remove Pengolodh but quite another to add content, even in a very brief form. However, he does add other content from Myths Transformed to LQ2: In the chapter “Of the Darkening of Valinor,” he adds a reference to the “dome of Varda” that alludes to his revised cosmology. He also adds volumes of other details to flesh out the narrative of LQ1. Yet he leaves the question of transmission untouched. Ironically, Christopher Tolkien refused to consider the revised Myths Transformed cosmology in making the published Silmarllion but stripped Pengolodh from the story on the strength of the purported revision to a Mortal tradition found also in Myths Transformed.

I find the opposite. Tolkien may have stated a desire to change the mode of transmission, but he didn’t actually do much to effect this, even when he had the chance to make surface changes to the text that would have set him on the path to deeper revisions. Therefore, I conclude that the “Quenta Silmarillion” should be read as a text written by an Eldarin loremaster and from an Eldarin point of view, with all that entails.