The Silmarillion: A Tale of Two Endings
In Tolkien’s Legendarium, Arda is the name of the world in which we still live and in which the major events concerning elves and men are said to have occurred many thousands of years ago (NOTE: Middle-earth is a continent on Arda, and Ëa is the entire universe, which includes Arda). The rebellion of the Vala Melkor and the subsequent wars between him and the faithful Valar resulted in the Marring of Arda. This raises an interesting question within Tolkien's legends: Will Arda ever be healed? The final paragraph of the published version of the Quenta Silmarillion leaves that question up in the air. It reads:
Here ends the SILMARILLION. If it has passed from the high and the beautiful to darkness and ruin, that was of old the fate of Arda Marred; and if any change shall come and the Marring be amended, Manwë and Varda may know; but they have not revealed it, and it is not declared in the dooms of Mandos (Published Silmarillion, p. 255).
In other words, according to this ending, whether Arda shall ever be healed is unknown, even to the Valar. Even Mandos, the Vala who pronounces what amount to prophecies, has not declared anything about Arda Healed. It appears the answer to our question is settled.
Appearances, however, can be deceiving.
Before proceeding, it is important to remember that J.R.R. Tolkien never completed the Silmarillion in his own lifetime. He left that project to his son Christopher Tolkien and authorized him to make the necessary decisions. The published Silmarillion was the result of his son’s enormous labors, but it required numerous editorial choices. In short, the ending we read in the published Silmarillion was not put there by J.R.R. Tolkien himself. It was his son's decision. (NOTE: Throughout the remainder of this brief essay, I will be using "Tolkien" to speak of J.R.R. Tolkien and "Christopher" to speak of Christopher Tolkien. This seems to be the most convenient and respectful way to distinguish the two men).
Tolkien died in 1973. After years of labor, Christopher published the Silmarillion in 1977. He then embarked on the herculean task of publishing and commenting on the vast number of manuscripts left behind by his father. These were the materials that were the sources for his work on the published Silmarillion. Some of these manuscripts date back to World War I and preserve the earliest forms of the tales that would develop into the Silmarillion. The publication of these manuscripts along with Christopher's editorial comments on them eventually filled twelve volumes, which go by the title The History of Middle-earth.
In volume 11 of this series, titled The War of the Jewels, Christopher says that the published Silmarillion is “a construction devised out of the existing materials” (p. x). He then adds: “Those materials are now made available . . . and with them a criticism of the ‘constructed’ Silmarillion becomes possible. I shall not enter into that question; although it will be apparent in this book that there are aspects of the work that I view with regret” (p. x).
I am not going to speculate on which parts of the published Silmarillion Christopher viewed with regret, but since he did say that the publication of the materials from which he constructed the published Silmarillion now makes criticism possible, I do want to look more closely at the conclusion he constructed and attempt to determine whether one decision in particular was the best one.
A Tale of Two Endings
The first thing that must be observed is that the conclusion to the published version of the Quenta Silmarillion was originally an ending of one Valaquenta manuscript from 1958. It reads:
Here ends The Valaquenta. If it has passed from the high and beautiful to darkness and ruin, that was of old the fate of Arda Marred; and if any change shall come and the Marring be amended, Manwë and Varda may know; but they have not revealed it, and it is not declared in the dooms of Mandos (Morgoth’s Ring, pp. 203–4).
Christopher took this Valaquenta ending and used it to conclude the published Silmarillion (Morgoth’s Ring, p. 204). Christopher’s only significant change was replacing the words “The Valaquenta” with “the SILMARILLION.”
Christopher’s choice to conclude the published Silmarillion with this passage from another work is interesting because this ending differs radically from earlier versions of the ending written by his father. Consider the final paragraphs of the 1936 version of the Quenta Silmarillion written by Tolkien. They read as follows:
§29 This was the doom of the Gods, when Fionwë and the sons of the Valar returned to Valmar and told of all the things that had been done. Thereafter the Hither Lands of Middle-earth should be for Mankind, the younger children of the world; but to the Elves, the Firstborn, alone should the gateways of the West stand ever open. And if the Elves would not come thither and tarried in the lands of Men, then they should slowly fade and fail. This is the most grievous of the fruits of the lies and works that Morgoth wrought, that the Eldalië should be sundered and estranged from Men. For a while other evils that he had devised or nurtured lived on, although he himself was taken away; and Orcs and Dragons, breeding again in dark places, became names of terror, and did evil deeds, as in sundry regions they still do; but ere the End all shall perish. But Morgoth himself the Gods thrust through the Door of Night into the Timeless Void, beyond the Walls of the World; and a guard is set for ever on that door, and Eärendel keeps watch upon the ramparts of the sky.
§30 Yet the lies that Melkor, the mighty and accursed, Morgoth Bauglir, the Power of Terror and of Hate, sowed in the hearts of Elves and Men are a seed that doth not die and cannot by the Gods be destroyed; and ever and anon it sprouts anew, and bears dark fruit even to these latest days. Some say also that Morgoth himself has at times crept back, secretly as a cloud that cannot be seen, and yet is venomous, surmounting the Walls, and visiting the world to encourage his servants and set on foot evil when all seems fair. But others say that this is the black shadow of Sauron, whom the Gnomes named Gorthû, who served Morgoth even in Valinor and came with him, and was the greatest and most evil of his underlings; and Sauron fled from the Great Battle and escaped, and he dwelt in dark places and perverted Men to his dreadful allegiance and his foul worship.
§31 Thus spake Mandos in prophecy, when the Gods sat in judgement in Valinor, and the rumour of his words was whispered among all the Elves of the West. When the world is old and the Powers grow weary, then Morgoth, seeing that the guard sleepeth, shall come back through the Door of Night out of the Timeless Void; and he shall destroy the Sun and Moon. But Eärendel shall descend upon him as a white and searing flame and drive him from the airs. Then shall the Last Battle be gathered on the fields of Valinor. In that day Tulkas shall strive with Morgoth, and on his right hand shall be Fionwë, and on his left Túrin Turambar, son of Húrin, coming from the halls of Mandos; and the black sword of Túrin shall deal unto Morgoth his death and final end; and so shall the children of Húrin and all Men be avenged.
§32 Thereafter shall Earth be broken and re-made, and the Silmarils shall be recovered out of Air and Earth and Sea; for Eärendel shall descend and surrender that flame which he hath had in keeping. Then Fëanor shall take the Three Jewels and bear them to Yavanna Palúrien; and she will break them and with their fire rekindle the Two Trees, and a great light shall come forth. And the Mountains of Valinor shall be levelled, so that the Light shall go out over all the world. In that light the Gods will grow young again, and the Elves awake and all their dead arise, and the purpose of Ilúvatar be fulfilled concerning them. But of Men in that day the prophecy of Mandos doth not speak, and no Man it names, save Túrin only, and to him a place is given among the sons of the Valar.
§33 Here endeth The Silmarillion: which is drawn out in brief from those songs and histories which are yet sung and told by the fading Elves, and (more clearly and fully) by the vanished Elves that dwell now upon the Lonely Isle, Tol Eressëa, whither few mariners of Men have ever come, save once or twice in a long age when some man of Eärendel’s race hath passed beyond the lands of mortal sight and seen the glimmer of the lamps upon the quays of Avallon, and smelt afar the undying flowers in the meads of Dorwinion. Of whom was Eriol one, that men named Ælfwine, and he alone returned and brought tidings of Cortirion to the Hither Lands.
This portion of the text of the Quenta Silmarillion (written around 1936) is found on pp. 332–334 of The Lost Road and Other Writings. As the final paragraph makes clear, it is still set within the original framework for these tales, a framework in which the Anglo-Saxon mariner Eriol/Ælfwine sails west and discovers the Lonely Island, Tol Eressëa, where he meets elves who tell him these stories. At the heart of this ending is what is called the Second Prophecy of Mandos. In this prophecy, Mandos reveals that Arda will ultimately be healed.
The ending containing the Second Prophecy of Mandos is obviously very different from the ending found in the published Silmarillion. Few have read the 1936 ending because it is published in a twelve-volume work that even the hardiest of Tolkien fans rarely complete.
This 1936 ending to the Silmarillion is consistent with the earlier versions of the story (See the 1926 “Earliest Silmarillion” in The Shaping of Middle-earth, pp, 40–41, and the 1930 Quenta Noldorinwa in The Shaping of Middle-Earth, pp. 163–165. Also, although the very earliest version of these tales, The Book of Lost Tales, written around 1916–1917, was never completed, there are hints in Tolkien’s notes that he already had this ending in mind even at that early date: see The Book of Lost Tales, Vol. II, pp. 281–282 on the defeat of Melko[r] by Fionwe and Túrin; and pp. 285–286 on the rekindling of the trees).
Obviously, the most significant part of the 1936 ending that is missing from the published ending is the “Second Prophecy of Mandos.” In this prophecy, Melkor is finally slain, by Túrin. Arda is re-made. The Silmarils are recovered. The Two Trees are re-kindled, and those elves who have died are raised. The only thing about which nothing is definitively said is the future fate of men. I’ll return to that issue below.
It seems clear that by the time of his death, Tolkien no longer wanted to use the Ælfwine journey to Tol Eressëa as the framework for the tales, so that final paragraph of the 1936 ending would have been removed or radically changed even if Tolkien had completed the Silmarillion himself. That, however, is not the decision that had the most significant impact on the published Silmarillion. It is not the most important question.
The most important question is whether Christopher made the best editorial decision by removing the Second Prophecy of Mandos concerning the Healing of Arda. Would his father have removed it had he lived long enough to complete the work? Again, we have to remember that Christopher was dealing with a massive number of manuscripts, some typed and some handwritten, that his father was ever revising throughout his life. At the time of his father’s death, the revisions were still incomplete, so Christopher had no definitive word on the ending to the Silmarillion. The fact that Christopher did the incredibly painstaking work required to publish any version of the Silmarillion is quite remarkable. The fact that he edited all of the materials he had at hand and published them in twelve large volumes is borderline miraculous.
In the remainder of this essay, I will look first at the reasons why Christopher chose to end the published Silmarillion in the way he did. I will then seek to determine whether there is counter-evidence that would have supported keeping an ending closer to that of the 1936 version – one that includes the Second Prophecy of Mandos about the Healing of Arda.
Reasons for Christopher Tolkien’s Editorial Decision
As far as I can tell based on my reading through The History of Middle-earth, Christopher Tolkien does not provide an explicit statement regarding why he chose to remove the “Second Prophecy of Mandos.” If it is there somewhere, I have missed it. What I have found are the following comments that seem relevant to the question:
In The War of the Jewels, Christopher notes that in one manuscript containing these concluding paragraphs, his father wrote an “X” in the margin next to one of the sentences (p. 247). It is not entirely clear to me that this “X” played any part in his decision to remove the entire Second Prophecy. Perhaps he thought the “X” meant delete all of it, but I’m skeptical about that possibility. I think it is more likely that he would have understood the "X" as referring to the specific sentences where it was placed in the margin. Even if that is the case, however, the meaning of the “X” remains ambiguous. It could be a notation Tolkien wrote to remind himself of something he intended to do, but it could mean “Delete” or “Rework” or any number of things.
More significant to Christopher’s decision, apparently, was the wording at the end of the version of the Valaquenta that he ended up using to conclude the Silmarillion. Again, those words were: “and if any change shall come and the Marring be amended, Manwë and Varda may know; but they have not revealed it, and it is not declared in the dooms of Mandos.” Those were words his father wrote in 1958 as the conclusion to one version of the Valaquenta. I think it is likely that Christopher removed the “Second Prophecy of Mandos” because he thought it was inconsistent with what his father wrote here. The Second Prophecy is either declared in the dooms of Mandos or it isn’t. It appears that Christopher thought he had to choose one ending or another.
That this is a likely explanation for his decision may be inferred from something Christopher wrote regarding another perceived inconsistency in his father’s works. He says:
A leading consideration in the preparation of the text was the achievement of coherence and consistency; 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. But I now think I attached too much importance to the aim of consistency, which may be present when not evident, and was too ready to deal with ‘difficulties’ simply by eliminating them (Morgoth’s Ring, pp. 204–5).
I would suggest that this is probably what happened with the Second Prophecy of Mandos. In a laudable attempt to achieve internal coherence and consistency, Christopher dealt with the difficulty caused by the Second Prophecy by simply eliminating it, even though the consistency for which he strived was arguably present but not evident to him. Interestingly, some of that actual desired consistency is related to the way the “Lore of the Eldar’ was transmitted. I will explain this last point below.
Reasons for Retaining the Second Prophecy of Mandos
Christopher, it appears, removed the Second Prophecy of Mandos about the Healing of Arda because he believed it was inconsistent with another passage his father had written during his revisions of the work in the late 1950s. Is there any evidence, however, that the Second Prophecy should have been retained in spite of the perceived difficulties?
I think there is sufficient evidence to conclude that the Second Prophecy of Mandos and the Healing of Arda is not only consistent within the Silmarillion but that it was clearly Tolkien’s intention for it to remain part of the conclusion of the Silmarillion. In what follows, I will merely summarize the evidence.
1. The ending with the Second Prophecy is consistent with Tolkien’s view of Fairy-stories. In his essay, “On Fairy Stories,” Tolkien writes:
But the ‘consolation of fairy-stories has another aspect than the imaginative satisfaction of ancient desires. Far more important is the Consolation of the Happy Ending. Almost I would venture to assert that all complete fairy-stories must have it. At least I would say that Tragedy is the true form of Drama, its highest function; but the opposite is true of Fairy-story. Since we do not appear to possess a word that expresses this opposite – I will call it Eucatastrophe. The eucatastrophic tale is the true form of fairy-tale, and its highest function (Tree and Leaf, p. 68).
Given that Tolkien devoted his entire adult life to constructing his own massive “fairy-story,” it is somewhat difficult to believe that he would want to leave out what he emphatically states “all complete fairy-stories must have” – namely, the Happy Ending. It is difficult to believe that he would want his fairy-story to not have the “true form” or the “highest function” of a fairy-story. The conclusion of the published Silmarillion leaves us with an agnostic answer about the possibility of any happy ending to the whole story. The ending with the Second Prophecy of Mandos brings everything to the kind of conclusion that Tolkien said all fairy-stories must have.
2. The ending with the Second Prophecy is consistent with what Tolkien wrote in his “Words, Phrases and Passages in the Lord of the Rings.” The evidence here is more general and indirect, but I think it is still worth considering. This document was written after the publication of The Lord of the Rings, but it was apparently never completed. It was edited by Christopher Gilson and published as Parma Eldalamberon in 2007 (See Hammond and Scull, The J.R.R. Tolkien Companion and Guide: Reader’s Guide, Part II, p. 1441). In a discussion of the reason for the creation of elves and men, Tolkien writes in this work:
These were said [in the legends of the elves] to have been an addition made by Eru himself after the Revelation to the primal spirits of the Great Design. They were not subject to the subcreative activities of the Valar, and one of the purposes of this addition was to provide the Valar with objects of love, as being in no way their own subject, but having a direct relationship with Eru Himself, like their own but different from it. They were, or were to be, thus ‘other’ than the Valar, independent creations of His love, and so objects for their reverence and true (entirely unselfregarding) love. Another purpose they had, which remained a mystery to the Valar, was to complete the Design by ‘healing’ the hurts which it suffered, and so ultimately not to recover ‘Arda Unmarred’ (that is the world as it would have been if Evil had never appeared), but the far greater thing ‘Arda Healed’ (Parma Eldalamberon, pp. 177–8).
Tolkien, here writing in the latter half of the 1950s or perhaps the early 1960s, is pointing to something beyond Arda Marred. He still has the idea of ‘Arda Healed’ as part of his thinking about the Legendarium. He hasn’t rejected the ultimate happy ending.
3. The ending with the Second Prophecy is consistent with what Tolkien wrote in his letters. In Letter 131 (to Milton Waldman, 1951), Tolkien provides a “brief” sketch of his legends. In this letter, at one point, he writes:
This legendarium ends with a vision of the end of the world, of its breaking and remaking, and the recovery of the Silmarilli and the ‘light before the Sun’ – after a final battle which owes, I suppose, more to the Norse vision of Ragnarök than to anything else, though it is not much like it.
This “vision of the end of the world” is the Second Prophecy of Mandos. Interestingly, Christopher added this particular letter to the Preface of the second edition of the published Silmarillion. The result is that Tolkien’s comment in the letter about the “vision of the end of the world" is now contained within the covers of the published Silmarillion, but the vision itself (the Second Prophecy of Mandos) remains absent from the published ending of the tales. What this means is that the second edition of the published Silmarillion now contains an inconsistency that would be resolved by including the Second Prophecy of Mandos.
4. The ending with the Second Prophecy is consistent with what is said elsewhere in the published Silmarillion. One of the major points made in the Ainulindalë is that Melkor/Morgoth’s attempts to create discord in the music will ultimately be overcome by Ilúvatar. This is found in the earliest versions of the Ainulindalë (See, e.g. The Lost Road, p. 158). It remains in the published version where Ilúvatar says to Melkor:
And thou, Melkor, shall see that no theme may be played that hath not its uttermost source in me, nor can any alter the music in my despite. For he that attempteth this shall prove but mine instrument in the devising of things more wonderful, which he himself hath not imagined (p. 17).
This passage early on in the Ainulindalë does not specifically mention the Second Prophecy of Mandos or Arda Healed, but it points toward the idea that Eru will overcome evil no matter what Melkor tries to do. The version of the Silmarillion that includes both this statement by Eru at the beginning and the Second Prophecy at the end perfectly frames all of the terrible tragedies that occur in between.
5. The ending with the Second Prophecy is consistent with the revisions Tolkien was making to the Legendarium later in his life. In his revisions of the later Quenta Silmarillion, after the publication of The Lord of the Rings, there is evidence that Tolkien intended to keep the Second Prophecy of Mandos. In The War of the Jewels, for example, Christopher himself notes that when he first commented (in The Lost Road) on the new subheadings that his father added to the text “I neglected however to mention there the introduction of a further subheading, The Second Prophecy of Mandos, at §31” (War of the Jewels, p. 247). In other words, it appears that even at this late stage in the revisions of the Quenta Silmarillion, Tolkien was planning to keep the Second Prophecy of Mandos. That is consistent with him introducing the subheading with that very title in his revised version.
Additionally, and perhaps more importantly, in the later Quenta Silmarillion, Tolkien included a section in which Manwë explicitly speaks of Arda Healed. After a discussion of the nature of death, Manwë says,
For Arda Unmarred hath two aspects or senses. The first is the Unmarred that they discern in the Marred, if their eyes are not dimmed, and yearn for, as we yearn for the Will of Eru: this is the ground upon which Hope is built. The second is the Unmarred that shall be: that is, to speak according to Time in which they have their being, the Arda Healed, which shall be greater and more fair than the first, because of the Marring: this is the Hope that sustaineth. It cometh not only from the yearning for the Will of Ilúvatar the Begetter (which by itself may lead those within Time to no more than regret), but also from trust in Eru the Lord everlasting, that he is good, and that his works shall all end in good. This the Marrer hath denied, and in this denial is the root of evil, and its end is in despair (Morgoth’s Ring, p. 245).
This is much more consistent with the ending of the Silamrillion that includes the Second Prophecy of Mandos than it is with the published ending that says “if any change shall come and the Marring be amended, Manwë and Varda may know; but they have not revealed it.” In these later revisions to the Quenta Silmarillion, Manwë does reveal it.
6. The ending with the Second Prophecy is consistent with what is said by characters in The Lord of the Rings. When Tom Bombadil drives the Barrow-wight away, he ends his song with a reference to a future time when “the world is mended” (Lord of the Rings, p. 142). When Galadriel says goodbye to Treebeard, she says they will not meet again in Middle-earth “nor until the lands that lie under the wave are lifted up again” (p. 981). This appears to be referring to the Healing of Arda at some unknown point in the future.
7. The ending with the Second Prophecy is consistent with the Athrabeth Finrod Ah Andreth. The Second Prophecy says, “But of Men in that day the prophecy of Mandos doth not speak, and no Man it names, save Túrin only, and to him a place is given among the sons of the Valar.” The Athrabeth is the record of a dialogue between an elf and a human woman. It includes, among many other topics, speculation by these two characters on the ultimate fate of men. As such, it provides a fascinating conclusion to the Silmarillion, or it would have had it been included in the work.
It appears very clear that Tolkien wanted the Athrabeth to be included as an Appendix to the Silmarillion. As Christopher notes, the manuscript was wrapped in a newspaper and on the wrapping, his father wrote the words: “Should be last item in an appendix.” Christopher indicates that what his father meant was an appendix to the Silmarillion (See Morgoth’s Ring, p. 329). Of course, as readers of the Silmarillion know, the Athrabeth is not found as an appendix to the published Silmarillion.
Had Christopher included the Athrabeth, however, not only would it have added an interesting perspective on the ultimate fate of men. It would have also provided a ready solution to the seeming inconsistency between the ending of the Valaquenta and the ending of the Silmarillion containing the Second Prophecy of Mandos. In Tolkien’s own commentary on the Athrabeth, he writes:
It is noteworthy that the Elves had no myths or legends dealing with the end of the world. The myth that appears at the end of the Silmarillion is of Númenórean origin; it is clearly made by Men, though Men acquainted with Elvish tradition (Morgoth’s Ring, p. 342)
When Tolkien speaks of the “myth that appears at the end of the Silmarillion” he is referring to the Second Prophecy of Mandos (See Morgoth’s Ring, p. 359, note 19). This quote by Tolkien is significant, first because at the time he completed the Athrabeth (around 1959 according to Christopher), he is still speaking of the Second Prophecy of Mandos as being part of the end of the Silmarillion. Recall that the apparent main reason Christopher removed the Second Prophecy was because it appeared inconsistent with how his father concluded a 1958 manuscript of the Valaquenta. But here in 1959, after writing that conclusion to the Valaquenta, Tolkien continues to speak of the Prophecy of Mandos as something appearing at the end of the Silmarillion. Apparently, he did not sense any inconsistency.
Second, this quote is significant because, in it, Tolkien distinguishes between elvish myths and Númenórean (human) myths. The elvish myths may lack things that human myths contain, or vice versa. They don’t have to be consistent in every respect. In fact, given human nature, it is likely that the Númenóreans would develop their own legends. So, by the time the various tales are gathered centuries or millennia later, one might have the Valaquenta legend ending in a way that is consistent with the elves having no myths or legends dealing with the end of the world. At the same time, the collections might also preserve Númenórean legends that contain the Second Prophecy of Mandos about the Healing of Arda. In other words, in a collection of elvish and human legends, the ending of the Valaquenta could easily co-exist with the ending of the Silmarillion containing the Second Prophecy of Mandos.
Whether Tolkien may have been toying with the idea of removing or changing the references to Túrin’s involvement with the final defeat of Melkor, I do not know. He probably planned to remove or change the references to Ælfwine. However, regardless of what he might have done with some of these secondary elements, the evidence is sufficient and clear that Tolkien had no intention to remove the Second Prophecy of Mandos. It seems clear that he intended to conclude these tales with a vision of the final battle that would result in Arda Healed. Furthermore, I believe that ending was absolutely fundamental to his Legendarium because it was fundamental to his concept of the nature and function of fairy-stories. The removal of that ending and the replacement of it with the more agnostic one we find in the published Silmarillion changes the entire tone of the stories.
Every fan of J.R.R. Tolkien owes his son Christopher Tolkien an enormous debt of gratitude. The fact that we have any version of the Silmarillion is only because of Christopher's devotion to his father's literary legacy and his intense labors. He devoted his life to making his father’s works available to the reading public. I have nothing but the deepest respect for him, and as a fan, I am incredibly thankful for his life’s work. It is only because Christopher himself indicated that there were things he regretted about the ‘constructed’ Silmarillion that I have ventured to say anything at all. I do not know whether the removal of the Second Prophecy of Mandos was one of the things Christopher later regretted. Based on my own study of the materials he himself has made available, it is one thing I regret and wish could be corrected (in a third edition of the published Silmarillion?). That is probably unlikely, but at least the Second Prophecy is now in print for those who are interested. And with Christopher's addition of Tolkien's letter to Milton Waldman in the Preface of the published Silmarillion, the Second Prophecy is now there "in spirit" if not in text.
UPDATE: OCTOBER 18, 2021 -- I had the wonderful opportunity of discussing this blog post with John Carswell on his Tolkien Road Podcast. The link to the podcast is here.