Tolkien Studies 1.1 (2004) 1-15
[Access article in PDF]

Light-elves, Dark-elves, and Others:

Tolkien's Elvish Problem

In chapter 15 of C. S. Lewis's 1938 novel Out of the Silent Planet, Elwin Ransom the philologist for the first time encounters a sorn, one of the tall, intellectual species that inhabits the highlands of Mars. They fall into a discussion of Oyarsa, the spiritual being who rules the planet, and Augray the sorn tells him that Oyarsa is an eldil. The eldila seem insubstantial to humans and Martians, Augray explains, but this is a mistake. The eldila can go through walls and doors not because they themselves are insubstantial but because to them our material world is insubstantial. "These things are not strange," says Augray, "though they are beyond our senses. But it is strange that the eldila never visit Thulcandra"—Thulcandra being "the silent planet" itself, Earth: "'Of that I am not certain,' said Ransom. It had dawned on him that the recurrent human tradition of bright, elusive people appearing on the earth—albs, devas, and the like—might after all have another explanation than the anthropologists had yet given."

What, one may well ask, are "albs" and "devas"? The second word presents no difficulties. If one looks it up in the Oxford English Dictionary, the sense given for "deva," entirely appropriately for the context above, is "'a bright, shining one'. . . a god, a divinity; one of the good spirits of Hindu mythology." All the OED has to offer for "alb," however, is that it is a tunic or ecclesiastical vestment, while "albs" does not occur at all.

Tolkien's connections with this passage are multiple. In the first place it is generally agreed that Elwin Ransom is an affectionate portrait of Tolkien himself. In the second place, the whole novel is now known to have grown out of the famous agreement by Tolkien and Lewis, in 1936, to write separate fictions, Lewis taking the theme of space-travel and Tolkien that of time-travel.1 Tolkien's contribution was never finished or published in his lifetime, seeing print eventually first as "The Lost Road" and then as "The Notion Club Papers," in volumes V and VIII respectively of "The History of Middle-earth."2 In both, the name Elwin, or forms of it such as Alwyn or Alboin, are significant.3 However, the immediate connection with the passage above is that "albs" is surely a word borrowed by Lewis from Tolkien, perhaps in conversation. *albs is in fact the unrecorded and hypothetical, or "reconstructed" Proto-Germanic [End Page 1] form of the word which descends into English as "elf," into Old English as ælf, into Old Norse as álfr, into Middle High German as alp, and so on. It then makes an entirely suitable match with "deva," being mythological, widespread, and bearing witness to a human attempt to label some phenomenon outside their normal comprehension. Only Tolkien is likely to have told Lewis such a thing. It would be entirely typical of Lewis, whose recorded remarks show several errors in Old English morphology, though he taught the subject at Magdalen College,4 to mis-hear it, and to assume the -s was a plural ending, so making "alb-s" (wrongly) parallel with "deva-s."

What the word and the passage show is that Tolkien had considered the whole problem of the variant forms of "elf" in Germanic languages, and presumably talked about it. It must have been a topic of Inkling conversation, one of several we can infer from cross-comparison of Lewis's, Tolkien's, Williams's, and Barfield's works (and possibly others as well). If Tolkien had considered the problem, we may again well ask what conclusions he had come to, and what further problems in the conflicting traditions of North-West Europe he would have encountered. The purpose of this essay is to suggest that it was indeed in these problems—even more than in the traditions—that Tolkien found inspiration for his fiction in the various versions of the Silmarillion, and eventually in sections of The Lord of the Rings.

The problems take a certain amount of explanation. One may begin with the thought, fundamental to the early investigators of comparative philology and mythology, that if a word existed in several "cognate," i.e., clearly related but nevertheless independent, forms in different languages, then the word and presumably the concept behind it must go back to a time before the languages separated from each other: the word and idea of "elf," then is quite literally immemorially old.5 But how does one then cope with the fact that the different linguistic and cultural traditions often seem to have quite different ideas of what the word means? Does this just mean that the word never did have any clear, agreed, stable referent (probably because the whole thing was pure fantasy, "just mythical," made-up from nothing)? Such an answer makes good sense, but was entirely unacceptable to Tolkien. This is the opinion of "the anthropologists" which Lewis's Ransom suddenly finds himself doubting.6 Or is it the case that we have not understood the data? That we need to think differently, as Augray the sorn tells Ransom he must rethink the idea of eldila? This was the view of Tolkien and the Inklings.

The data as regards elves had been known to investigators, at least in great part, since well before Tolkien's time.7 There are some ten words for "elf" in Old English, the male and female forms ælf and ælfen, and the compound words land-, dún-, feld-, munt-, sæ-, wæter-, wudu-, and possibly [End Page 2] berg—ælfen, or, more rarely, -ælf, i.e., "hill-, land-, field-, mountain-, sea-, water-, wood-," and once again "mountain-elf." These look promisingly precise and varied, but are in fact almost always glosses, words written in over a Latin text to translate a hard word in Latin, in this case and respectively to items four to nine in the list above castalides, moides, oreades, naiades, nymphae, and dryades. The simplest explanation is that an Anglo-Saxon translator long ago, stumped for an equivalent to "naiad, nymph, dryad," decided not unreasonably to solve all his problems at once and create "sea-elf, water-elf, wood-elf," etc. Meanwhile Anglo-Saxon medical or magic texts throw up another run of more interesting if more threatening compounds, such as ælfadl, wæterælfadl, ælfsiden, ælfsogoda, the names of "elf-diseases" like (it has been suggested) chicken-pox, dropsy, lunacy, epilepsy, anaemia.8 The last is a guess from ælfsogoda, "elf-sucking," and indicates that one way elves were thought to work their damage was by a kind of vampirism, while we also hear several times of "elf-shot" or ylfa gescot, which implies a belief (perhaps illustrated in one of these texts) in invisible disease-bearing darts. Elves also appear to have been associated with sexual temptation. Several charms associate the elves with nihtgengan, "night-walkers," with "temptations of the fiend" and with þam mannum þe deofol mid hæm, "the people the devil has sex with." It is not surprising that Anglo-Saxon elves are commonly called "malignant" by modern scholars.9 And yet it is a compliment for a woman to be called ælfsciene, "elf-beautiful," and Anglo-Saxons stubbornly continued to give their children names like Ælf-wine, Ælf-red, Ælf-stan, and so on, "Elf-friend, Elf-counsel, Elf-stone." Some of the names, like the common Alfred and the rare Elwin (as in Elwin Ransom), have remained in use to this day, though no longer with any sense of their meaning, and some of the beliefs about sexually alluring elves, elf-hills, and elf-changelings also lasted into the modern period.

The Scandinavian tradition is even more well-attested, though not as old, and on the face of it rather different. The álfar are mentioned thirty times in the poems of the Elder Edda, though in a rather restricted list of uses: usually they occur in association with either the Æsir, the pagan gods, or with the iötnar, the giants, as if to imply universality: "everyone knows it, elves and gods," "tell me its name among the elves, tell me its name among the giants," and so on. There are hints of meaning in the poems of the Elder Edda, as there are here and there in sagas. But the work which attracted most attention from the beginning of modern investigation, and which seemed closest to giving answers of the thoroughness and complexity which philologists demanded, was the Prose Edda of Snorri Sturluson, the nearest thing we have to a mythical handbook of pre-Christian belief.

Commentators often forget that Snorri was not writing a pagan text. He wrote his work in the 1230s, by which time Iceland had been Christian [End Page 3] for more than two centuries, and Snorri's own family had been Christian for six generations. He knew no more about what pagans really did, or really thought, than we would about the folk-beliefs of the eighteenth century. His work was in essence an attempt to explain poetic diction, the phrases used and allusions made in traditional poetry, but to do this he had to tell stories, often about the gods, giants, elves, dwarves, and other supernatural creatures of the pre-Christian world. The connected nature (and the literary power) of what Snorri wrote perhaps aroused unreal expectations in his first modern admirers, for what Snorri says about elves is hard to make out. He invariably uses álfr as a compound, one of these being Álfheim or "Elf-home." But every other time he uses álfr, he prefixes it with a word of color, ljós-, dökk-, or svart-, i.e., "light-elves," "dark-elves," "black-elves." A critical passage is this one:

Sá er einn staðr þar er kallaðr er Álfheimr. Þar byggvir fólk þat er ljósálfar heita, en dökkálfar búa niðri í jörðu, ok eru þeir ólíkir þeim sýnum en myklu ólíkari reyndum. Ljósálfar eru fegri en sól sýnum, en dökkálfar eru svartari en bik.

There is one place that is called Alfheim. There live the folk called light-elves, but dark-elves live down in the ground, and they are unlike them in appearance, and even more unlike them in nature. Light-elves are fairer than the sun to look at, but dark-elves are blacker than pitch.10

What Snorri says is clear and unequivocal, but it raises an immediate problem. "Dark-elves" (dökkálfar), he says, are "black" (svart). Surely that means that they are "black-elves" (svartálfar)? But everywhere else in Snorri's work, it is clear that when he says "black-elves" (svartálfar), he means "dwarves": Odin sends Skirnir í Svartálfaheim til dverga nokkurra, "to the home of the black-elves to certain dwarfs," and Loki too goes into Svartálfaheim where he too "comes across a dwarf." There is a simple explanation here, which is that while Snorri identifies four groups, light-elves, dark-elves, black-elves, and dwarves, there are really only two: the last three are just different names for the same group. The first group, meanwhile, are very like angels, or for that matter eldila—these are Lewis's "albs"—while the last group have been made to seem faintly diabolic, quite like the Anglo-Saxon elves of the medical textbooks, indeed. This line of thought has the blessing of being clear, and of not multiplying entities, but it was once again quite unacceptable to early investigators, including Tolkien: it meant, in effect, throwing away their best text, just as my suggestion about a baffled Anglo-Saxon translator above meant saying that dún-ælf and the rest were just "ghost-words," with no real meaning in Anglo-Saxon culture. Neither proposal has been popular, and Tolkien devoted considerable fictional energy to providing [End Page 4] more face-saving refutations to both.

It is not absolutely clear when Tolkien focused for the first time on what we may call the "elf-problem." When he did do so, though, it would be natural for him to look at what "the authorities" said, and entirely characteristic of him (as happens so often with Tolkien and the OED) then to found a theory on profound disagreement with scholarly opinion, and to make a determined attempt to protect the original sources, if necessary by explaining how they could have been mistaken. The author of Sir Gawain, after all, or perhaps the scribe who copied him, had made the same mistake as C. S. Lewis, taking a singular ending in -s to be a plural, writing wodwos for what should have been *wodwosen. It was the job of the true scholar, Tolkien thought—he exemplifies it frequently in his edition of the Old English Exodus and the "Finnsburg" poems—to rescue poems and myths from their careless or uncomprehending scribes and annotators. And this is what he tried to do, in my opinion, with the elves.

The original sources mentioned above had been known to scholars for centuries, if with very little original circulation. Snorri's Prose Edda, for instance, had been edited by the Dane Peter Resen (Resenius) as early as 1664, while the Old English medical texts and glosses had been discovered at various times up to the 1830s. The "elf-problem," however, did not surface until scholars began to ask themselves not just about the words, but about what they represented. And here two famous scholars, in particular, are likely to have attracted Tolkien's attention.

The first was the Dane, N.F.S. Grundtvig (1783-1872). There are several reasons why Tolkien might have paid careful attention to him. Nikolai Grundtvig was, for one thing, the first person in modern times to read Beowulf intelligently. (It was he, for instance, alone of the first seven reviewers of the first modern edition of the poem, Grímur Thorkelín's of 1815, who realized that the poem began with a funeral, not a Viking raid as the editor had thought.) He continued to be an active scholar for nearly sixty years after that, with particular interest in Beowulf, in Old English, and in Northern mythology. But even more importantly, Grundtvig did for Denmark what Tolkien would have liked to do for England: he gave it a history and a mythology founded on ancient sources, but released again into national life and national politics by his popular writings, his many songs and hymns, and his creation of the Grundtvig High Schools with their avowed aim of protecting national culture, primarily from German encroachment.11 Grundtvig in Denmark, Lönnrot of the Kalevala in Finland: if Tolkien ever had "role-models," they would be these.

Grundtvig's first book on mythology, Nordens Mytologi, was published in 1808, at which point works like Beowulf were still unpublished. Grundtvig rewrote the work as (different spelling) Nordens Mythologi in 1832, and in [End Page 5] this he turned his attention to "Vætter, Alfer, og Dværge," "Wights, Elves, and Dwarves"; he was, I believe, the first to note and be concerned about Snorri's inconsistencies in the Prose Edda, as noted above. His solution was to go part of the way toward the reductionist four-groups-down-to-two model outlined above, with one significant compromise. Light-elves were obviously angelic, and black-elves were evidently dwarves, but perhaps dark-elves were different from both:

Alfer var det gamle Nordens Engle, og Dværgene kun et Mellem-Slags af dem: hverken Lys-Alfer eller Mörk-Alfer, men saa at sige Skumrings-Alfer.

Elves were the angels of the ancient North, and dwarves only a middle grade of them: neither light-elves nor dark-elves, but so to speak elves of the twilight.12

The trouble with this otherwise neat solution, one might say, is that it puts black-elves in between the other two groups, where one might expect them to be a limiting term. But it does introduce the rather attractive idea of Skumrings-Alfer, "elves of the twilight."

Jacob Grimm's Deutsche Mythologie, the first edition of which was published in 1835, may have owed more to Grundtvig's pioneering work than Grimm was prepared to admit. The philological battle-lines were already drawn up—they were to become real battle-lines in the two Prusso-Danish wars over Schleswig-Holstein, or Slesvig-Holsten, in 1850-51 and 1864—with the Germans, and Grimm in particular, claiming that Scandinavian languages were really just a branch of "Germanic," with the Eddas and sagas in effect common intellectual property, and Scandinavian scholars replying furiously that Scandinavia had a right to cultural as well as political autonomy. It was a problem and an annoyance for Grimm that the Middle High German word for "elf" seemed to have been lost, to be replaced in modern German by a borrowing from English, Elfe, Elfen. Grimm dealt with this by deleting the latter from his Deutsches Wörterbuch or "German Dictionary" and inserting a modernized version of the former: Elb, Elbe. But he too was bothered by Snorri, though his solution was significantly worse than Grundtvig's, vague and indecisive. I give it below, in sections, in Grimm's German and in the translation of J. S. Stallybrass, with my own attempts to explain what he meant interpolated:

Man findet in dem Gegensatz der lichten und schwarzen elbe den dualismus, der auch in anderen mythologien zwischen guten und bösen, freundlichen und feindlichen, himlischen und höllischen geistern, zwischen engel des lichts und der finsternis aufgestellt wird.
(Grimm 1:368) [End Page 6]
Some have seen, in this antithesis of light and black elves, the same Dualism that other mythologies have set up between spirits good and bad, friendly and hostile, heavenly and hellish, between angels of light and of darkness.
(Stallybrass 2: 444-5)

Grimm is here, I think, contradicting Grundtvig without mentioning him. He feels that Grundtvig has abandoned Snorri's tripartite division too readily:

Sollten aber nicht drei arten nordischer genien anzunehmen sein: liosálfar, döckálfar, svartálfar?

But ought we not rather to assume three kinds of Norse genii, liosálfar, döckálfar, svartálfar?

The trouble with this is Snorri's statement above that dark-elves are black, which would lead to the first reduction, dark-elves = black-elves. But Grimm cannot accept this because he knows it would lead on to black-elves = dwarves. He therefore continues:

ich erkläre damit freilich Snorris satz "döckálfar eru svartari en bik" für irreleitend.

No doubt I am thereby pronouncing Snorri's statement fallacious: "dark-elves are blacker than pitch."

The easiest way out at this stage is to say, rather unconvincingly, that maybe Snorri was half-right, did not choose his words carefully, at any rate has to be overruled:

döckr scheint mir weniger das entschieden schwarze, als das trübe, finstere; nicht niger, sondern obscurus, fuscus, aquilus.

Döckr seems to me not so much downright black as dim, dingy; not niger but obscurus, fuscus, aquilus.

Grimm backs this up with a sentence about a reference to dwarves and a dwarf name that contain or resemble the word iarpr, "dark," which actually does not seem to help his case that dark-elves are different from black-elves and dwarves, but concludes that rejecting Snorri's one-off statement on the whole saves more trouble than it creates:

dann bliebe die gleichstellung der zwerge und schwarzelbe gültig, aber auch jener alteddische unterschied zwischen zwergen und dunkelelben gerechtfertigt. [End Page 7]

In that case the identity of dwarfs and black elves would still hold good, and at the same time the Old Eddic distinction between dwarfs and dark elves be justified.

Grimm then embarks on a lengthy search for other references in German story to tripartite color-systems, but ends abruptly, perhaps aware of his own inconclusiveness:

Festgehalten werden muss die identität der svartálfar und dvergar.

One thing we must not let go: the identity of svartálfar and dvergar.

Snorri can be trusted, then, when he says something Grimm is prepared to accept, but has to be ruled out when his statement is unwelcome.

I believe that Tolkien must have read this passage in the most familiar account of Northern mythology and was probably annoyed by it. However, along with Snorri and Grundtvig and the other Old English texts mentioned above, Grimm's argument does raise a whole sequence of problems which cry out for some better solution. I would list them as follows:

  1. What are light-elves and dark-elves, and what is the difference between them if it is not a matter of color?
  2. If it is not a matter of color, why does Snorri say that dark-elves are black?
  3. If dwarves are different from elves, as almost all early evidence agrees, then why call them black-elves?
  4. What are all these Old English groups, like wood-elves and sea-elves, and where do they fit in?
  5. Is there anything to be said for Grundtvig's idea that there may have been "elves of the twilight"?

Anyone familiar with The Silmarillion can see how clearly and incisively, if imaginatively, Tolkien was in the end to answer these questions. Did he have the questions, if not the answers, in mind from the beginning? He was to say of himself at one point, with reference to ents, "As usually with me, they grew rather out of their name, than the other way about" (Letters 313), and I would suggest that the same may be true of Tolkien's elves. One of the starting points of his whole developed mythology was this problem in nomenclature, this apparent contradiction in ancient texts and in one ancient text in particular, a problem made only more challenging by the groping attempts of earlier scholars to solve it. [End Page 8] However, as the twelve volumes of The History of Middle-earth have made abundantly clear, it was also characteristic of Tolkien to edge up on the solution to a problem through several or many stages of dissatisfaction.

The Book of Lost Tales thus does not, as far as I can see, contain the basic distinction later to be made between Light-elves and Dark-elves: such references as are indexed are to later stages of Tolkien's conception. There is however an interesting passage in Lost Tales I which suggests that Tolkien was already considering the terms, and was perhaps aware of Grundtvig's compromise solution quoted above. In "Gilfanon's Tale," just after the first mention of "Dark Elves," we are told of "a certain fay . . . Tu the wizard":

wandering about the world he found the . . . Elves and drew them to him and taught them many deep things, and he became as a mighty king among them, and their tales name him the Lord of Gloaming and all the fairies of his realm Hisildi or the twilight people.

The missing word in the phrase "the . . . Elves" above, Christopher Tolkien reports, could be either "dim" or "dun" (Lost Tales I 244). "Dun" would correspond to one of the Anglo-Saxon glossary words noted above, but "dim" is one of Grimm's suggestions, at least as translated by Stallybrass.13 Meanwhile "Gloaming" is a good translation of the first word in Grundtvig's phrase Skumrings-Alfer, but "twilight people" is used as well. Perhaps Tolkien had already rejected the concept "black-elves," looking on this as an uninformed variant on "dwarves," as it seems to be, but at this point had no explanation of "dark-elves" other than to say that they were only to be glimpsed at twilight. The index of Lost Tales II supports the suggestion that Tolkien was groping, for there one finds ten different groups of elves, but not yet "Light-elves." The tale of "The Fall of Gondolin" already has the character of Meglin (later Maeglin), son of Eöl, but very little is said of the latter other than "that tale of Isfin and Eöl may not here be told" (165). "The Lay of the Fall of Gondolin," included among the "Poems Early Abandoned" in The Lays of Beleriand, goes a little further in describing the capture of Isfin by Eöl: "that she ever since hath been / his mate in Doriath's forest, where she weepeth in the gloam; / for the Dark Elves were his kindred that wander without home" (146).

But though the idea of a White Lady glimpsed in the half-light was to remain through to The Silmarillion, there is no further advance on the dark/light distinction. Tolkien seems to have no clear idea of what a "dark-elf" is, in which, of course, he is in agreement with his predecessors; and the term "light-elves" is not used at all.

This last was to change with the writing of "The Earliest [End Page 9] Silmarillion," in the late 1920s, where we find (Shaping 13) the division of the Eldar into three groups, "Light-elves," "Deep-elves," and "Sea-elves," corresponding closely though not exactly to the Vanyar, Noldor, and Teleri of The Silmarillion. The real breakthrough comes, however, in the "Quenta" of 1930. Here we find that the Quendi, led by Ingwë, are "the Light-elves," the Noldoli, led by Finwë, are "the Deep-elves," and the Teleri, led by Elwë, are "the Sea-elves" (Shaping 85). A vital addition, though, is that "many of the elfin race were lost upon the long dark roads . . . and never came to Valinor, nor saw the light of the Two Trees. . . . The Dark-elves are they." One might note at this time the use of the invented Anglo-Saxon terms Léohtelfe, deorc-elf[e],14 in "The Earliest Annals of Valinor" (Shaping 286, 288), words which correspond exactly to Snorri's ljósálfar, dökkálfar. This decision to make the light/dark distinction not a matter of color, as Grimm had tacitly assumed, was a brilliant stroke, rather like Augray the sorn explaining the eldila. But one result was that it left Eöl, identified already as a Dark Elf, see above, without any clear mark of distinction. He is mentioned in both "The Earliest Silmarillion" and the "Quenta" as "the Dark-elf Eöl" (Shaping 34, 136, with variant spellings), but in both cases this could just mean that he is a Dark-elf, one of the Dark-elves: there is nothing particular to mark him out. His son Meglin, though, is picked out as "swart" (Shaping 141), a word that goes back to Lost Tales I (165), as if Tolkien had not yet quite abandoned hope of reconciling Snorri's dökkálfar and svartálfar —could Eöl be seen as "a" Dark-elf, but also "the" Swart-elf? This hint was never taken up, and indeed may never have been in Tolkien's mind, but as so often with Tolkien, it seems that for him to solve one problem was to generate another.

Tolkien was to develop his basic distinction between those who had and those who had not seen the Light of the Two Trees in "The Lhammas" and "The Quenta Silmarillion" (see Lost Road 197, 215), while some of his terminology became canonical in the familiar passage from chapter 8 of The Hobbit, published in 1937, about the Wood-elves: "more dangerous and less wise" than "the High Elves of the West," these latter further particularized as "the Light-elves and the Deep-elvesand the Sea-elves." As for the Wood-elves, they:

lingered in the twilight of our Sun and Moon, but loved best the stars; and they wandered in the great forests that grew tall in lands that are now lost. They dwelt most often by the edges of the woods, from which they would escape at times to hunt, or to ride and run over the open lands by moonlight or starlight, and after the coming of Men they took more and more to the gloaming and the dusk.15 [End Page 10]

They are, in other words, very much Skumrings-Alfer, twilight-elves.

At this stage, one might say, Tolkien had settled the first and fifth of the questions outlined above and made space for a solution to the fourth. The other two, however, remained quite obscure: why dark-elves might be black, as Snorri reported, and what if anything they had to do with dwarves. Both are nevertheless settled firmly and even convincingly by the re-organization of the story of Eöl, Dark Elf par excellence, in chapter 16 of the Silmarillion. It is astonishing how much of previous speculation is taken up and dealt with on pages 132-3 of that work.

We learn first that Eöl "was named the Dark Elf," and here it is his personal appellation, not just a generic description. The reason he is "the Dark Elf" is that he has left Doriath for Nan Elmoth, and "there he lived in deep shadow, loving the night and the twilight under the stars." He resents in particular the Noldor among the Light-elves, as usurpers, "but for the Dwarves he had more liking than any other of the Elvenfolk of old." From them he learns metalwork and devises a metal of his own. "He named it galvorn, for it was black and shining like jet, and he was clad in it whenever he went abroad." His son Maeglin is called (by his mother) Lómion, "Child of the Twilight." From these few sentences one could construct a story which would explain all that Snorri says, without corroborating it. It would not be true that there were three kinds of elf, for there were no "black-elves," no svartálfar at all. Just the same, in later story someone might well think there were, for while there were no "black-elves," there was an elf always dressed in black, whom someone might have labeled "the Black Elf." Similarly, this svartálfr was certainly not a dwarf, but was associated with them and shared some of their characteristics, like the fascination with metalwork. Again, in careless repetition "like" could become "the same as." Finally, there may be no such generic term as a Skumrings-Alf or "twilight elf," but if Maeglin is "Child of the Twilight," then his father might again, mistakenly, be heard as "the twilight," especially as that is the time he goes abroad. One may at this point see the force of Christopher Tolkien's repeated statements that the Silmarillion was seen all along by his father as a "compendium," which needs to be read from the point of view of someone looking back at events from a much later period.16 A text, to Tolkien Sr., was not just the words on the page one happened to be reading, it was also the whole history of how the words got there—a history, in many of the works he devoted his professional life to studying, of misunderstanding and downright error. One might paraphrase by saying that Tolkien (like Grimm) was prepared to say that Snorri Sturluson had just got it wrong. But unlike Grimm he insisted on providing a story to explain how Snorri got it wrong, and to make that explanation plausible and even natural.

In much the same way, Tolkien approached the oddly contradictory Anglo-Saxon accounts, where descriptions of malignant elves contrasted [End Page 11] with a seeming deep-rooted respect for them. In The Lord of the Rings he confronts this problem at least three times. The feeling that elves are dangerous is expressed first by Boromir, who does not want to enter the Golden Wood of Lothlórien, because "of that perilous land we have heard in Gondor, and it is said that few come out who once go in; and of that few none have escaped unscathed" (FR, II, vi, 353). Aragorn corrects Boromir, but does not entirely deny what he says. Boromir's feelings are then echoed by Éomer (TT, III, ii, 34-35), who uses "elvish" to mean "uncanny," and also believes the Lady of the Wood to be some kind of sorceress. This time Gimli corrects him. Just the same, though both men are misinformed, there is a basis for their fear and suspicion, as Sam Gamgee points out. When Faramir, wiser than his brother, nevertheless hints that Galadriel must be "perilously fair," Sam picks up the implied criticism and half-agrees with it: "I don't know about perilous. . . . It strikes me that folk takes their peril with them into Lórien, and finds it there because they've brought it. But perhaps you could call her perilous, because she's so strong in herself. You, you could dash yourself to pieces on her, like a ship on a rock; or drownd yourself, like a hobbit in a river. But neither rock nor river would be to blame" (TT, IV, v, 288). At the end of a long chain of transmission it might be agreed that to be ælfsciene like Galadriel would be an immense compliment, but at the same time that any association with elves might well be disastrous for ordinary people; the end of this chain is line 112 of Beowulf, eotenas ond ylfe ond orcneas, in which elves and orcs have become much the same thing.17 Tolkien put a very high value on his ancient texts, like Beowulf and the Prose Edda, but he knew they were the work of fallible mortals, and probably several generations away from what he would have regarded as authentic tradition.

What he meant to do, then, was to recover the authentic tradition which lay further back than any account we possess, the tradition which gave rise to Snorri and Beowulf and the Eddic poems and the Anglo-Saxon charms and all the other scraps of evidence, which however integrated them, resolved their contradictions, and explained the nature of their misunderstandings. The idea that there was some such authentic tradition is the thought that strikes Ransom/Tolkien in Lewis's story quoted at the start of this essay. It is possible, of course, that the whole idea is mistaken, and highly probable that even if there were to have been some original single integrated conception of "elves" or "devas" then, it is now beyond recall. Nevertheless, Tolkien's reconstructions are not only imaginative, they are also rigorous, controlled both by respect for evidence and awareness of the nature of the evidence. Philology was a hard science, not a soft science. This is one of the qualities which makes Tolkien's work inimitable. [End Page 12]


1. The best account of this is John D. Rateliff (199-218).

2. There is a full-length study of them by Verlyn Flieger, A Question of Time: J.R.R. Tolkien's Road to Faërie.

3. I discuss the origin and significance of the name in its variant forms in chapter 9 of The Road to Middle-earth, 3rd ed.

4. Lewis for instance wrote a piece in praise of Tolkien, the title of which began "Hwæt we holbytlan . . . ," clearly echoing the opening words of Beowulf, "Hwæt we Gar-Dena. . . ." But Gar-Dena is genitive plural. The genitive plural of holbytla would be, not holbytlan, but holbytlena. Lewis was extremely learned and an excellent Classicist, but he could not be called a philologist in Tolkien's sense of the word

5. This point is made explicitly by Max Müller in his essay "Modern Mythology" (1856). The essay is best known now for Müller's attempt to relate all myth to celestial phenomena, for his argument that myth is "a disease of language," and for the parody of the whole theory by R. F. Littledale, "The Oxford Solar Myth," in which the Rev. Littledale proved by Müller's own methods that Müller was himself a solar myth. Most of the essay, however, is a reasoned statement of the methods of comparative philology, before the proposal is made that a similar technique could be used to create comparative mythology. Both Müller's and Littledale's pieces can be found reprinted in Müller, Comparative Mythology: An Essay. Tolkien refers to Müller, while inverting the "disease of language" thesis, in "On Fairy-Stories."

6. It is not absolutely clear which anthropologists Lewis meant here, but probably not American structural or cultural anthropologists. He was probably thinking of post-Müllerian schools of thought like the followers of J. G. Frazer, or the "ritual" school of Jane Harrison. Lewis's essay "The Anthropological Approach" attacks later and minor members of these groups (301-11), and they appear in disguised form in his 1956 novel Till We Have Faces.

7. I discuss the data at much greater length in "Alias Oves Habeo: The Elves as a Category Problem." The essays in the collection to which it belongs discuss the accounts of various groups of Germanic non-humans, elves, dwarves, trolls, dragons, etc., but all contributors have been warned not to discuss Tolkien. The problem now is to imagine any solutions other than Tolkien's: a measure of his success.

8. See the valuable book by M. L. Cameron, Anglo-Saxon Medicine. As a [End Page 13] professor of biology, Cameron is able to talk about the recipes and their possible efficacy in a pragmatic way.

9. See, for instance, Nils Thun (378-96) and also Heather Stuart (313-20).

10. For the original, see Sturluson (19), translated in Faulkes (19-20).

11. For an account in English of Grundtvig's life and works, see Allchin.

12. Grundtvig, Nordens Mythologi (1832), 263, with my translation.

13. In context "dun-elves" sounds better, but in that case one wonders whether Tolkien could be playing on the two senses of the word, Old English dún-ælf, "mountain-elf," and modern English "dun," i.e., "dark."

14. The form deorc-elfa in Shaping (288) is another genitive plural.

15. The text given appeared first in the revised edition of 1966. Earlier versions have slightly different wording, and the "twilight" is "the twilight before the raising of the Sun and Moon" (Hammond and Anderson 32).

16. Christopher Tolkien makes the point in Lost Tales I: "To read The Silmarillion one must place oneself imaginatively at the time of the ending of the Third Age—within Middle-earth, looking back" (4). This is good advice, but the exercise becomes much easier if one has prior experience of the way texts and stories change over time.

17. The line is part of the introduction of the monster Grendel. The poet says that all the monster-species derive from the first murderer, Cain, and exemplifies them as "ettins and elves and (?) demon-corpses, and the giants, who fought against God for a long time." This is the most "hard-line" hostile statement made about elves in any ancient source, and must have caused Tolkien some thought, as it comes from a text he respected and valued greatly: it was often identified by early scholars as an interpolation, not the work of the original poet.

Works Cited

Allchin, A. M. N.F.S. Grundtvig: An Introduction to His Life and Work. Oakville, CT: Aarhus University Press, 1997.

Cameron, M. L. Anglo-Saxon Medicine. Cambridge: Cambridge University Press, 1993 [End Page 14]

Flieger, Verlyn. A Question of Time: J.R.R. Tolkien's Road to Faërie. Kent, Ohio and London: Kent State University Press, 1997.

Grimm, Jacob. Deutsche Mythologie, 4th ed. 3 vols. Berlin: Dümmler, 1875-78.

______. Grimm's Teutonic Mythology, translated by J. S. Stallybrass. 4 vols. London: George Bell and Sons, 1882-88.

Hammond, Wayne G. with Douglas A. Anderson. J.R.R. Tolkien: A Descriptive Bibliography. Winchester: St Paul's Bibliographies; New Castle, DE: Oak Knoll Books, 1993.

Lewis, C. S. "The Anthropological Approach." In Selected Literary Essays of C.S. Lewis, edited by Walter Hooper. Cambridge: Cambridge University Press, 1969.

Müller, Max. Comparative Mythology: An Essay, edited by A. Smythe Palmer. London: Routledge; New York: Dutton, 1977.

Rateliff, John D. "The Lost Road, The Dark Tower, and The Notion Club Papers: Tolkien and Lewis's Time Travel Triad." In Tolkien's Legendarium: Essays on The History of Middle-earth, edited by Verlyn Flieger and Carl F. Hostetter. Greenwood: Westport, CT, 2000.

Shippey, Tom. The Road to Middle-earth. 3rd ed. Boston: Houghton and Mifflin, 2003.

______. "Alias Oves Habeo: The Elves as a Category Problem." In The Shadow-walkers: Jacob Grimm's Mythology of the Monstrous, edited by Tom Shippey. Tempe: Arizona State University Press, forthcoming.

Sturluson, Snorri. Edda: Prologue and Gylfaginning, edited by Anthony Faulkes. Oxford: Clarendon Press, 1982.

Sturluson, Snorri. Edda, translated by Anthony Faulkes. London: Dent, 1987.

Stuart, Heather. "The Anglo-Saxon Elf." Studia Neophilologica 48 (1976): 313-20.

Thun, Nils. "The Malignant Elves," Studia Neophilologica 41 (1969): 378-96.