*This is a preprint of an article published in Tolkien Studies 4 (2007): 51-71, available online at: http://muse.jhu.edu/article/215125/pdf. At the time the article was published Tolkien’s The Fall of Arthur had not appeared in print yet, and The Lay of Aotrou and Itroun was only accessible via its original publication in the New Welsh Review. This is the second of two articles I wrote on Tolkien’s engagement with “things Celtic”. The first one can be found here.
1. Tolkien’s Celtic Library
After J.R.R. Tolkien’s death, a number of books from his personal collection ended up in two Oxford libraries. A small number are in the Bodleian Library, within the Tolkien manuscript collection, in the section “Tolkien E16.” A considerably larger number are to be found in the Library of the English Faculty. According to the library’s own classification system, the books are shelved in section V, which is described as “Tolkien’s Celtic Library.”
An initial reaction to this description might be surprise. Tolkien’s dislike for “things Celtic,” strongly expressed in his much-quoted 1937 letter to Stanley Unwin (Letters 26), is well known and could be taken as a definitive discouragement to research in Tolkien’s Celtic sources. It is only recently that scholarship has attempted a serious evaluation of the Celtic elements of Tolkien’s inspiration (see Burns; Fimi; Flieger Interrupted Music). Nevertheless, Tolkien’s “Celtic Library” holds exciting revelations, if only for its sheer size. Over three hundred books originally owned by Tolkien are held in the Bodleian and the English Faculty Library, of which approximately a third belong to the discipline of Celtic Studies. It is, of course, not easy to determine what percentage of the whole body of Tolkien’s books they comprise. It is known that the bulk of Tolkien’s books passed initially to his son Christopher, and that only a small part of these were donated to the two Oxford libraries mentioned above, while others were sold through an Oxford bookseller (Anderson “Personal”). Still, this data is both valuable and significant for Tolkien scholarship, especially in terms of his involvement with Celtic Studies. Tolkien’s “Celtic Library” at the English Faculty Library consists of books on Celtic languages (including Welsh, Old and Middle Irish, Gaelic, and Breton), and also an important number on Irish and Welsh medieval literature, together with translations, editions and even facsimiles of manuscripts of original texts. An example of how specialized this collection can be is the so-called “Mabinogion” from both the Red Book of Hergest and the White Book of Rhydderch in four editions: those by Rhŷs and Evans (1887), by Evans (1907), by Edwards (1921) and by Mühlhausen (1925), as well as its famous translation by Lady Guest (1913). Tolkien also owned a copy of Pedeir Keinc y Mabinogi, being a reduced reproduction of a part of the Rhŷs and Evans 1887 edition, bequeathed to him by his school friend G.B. Smith, who died in the Great War. This is the only Celtic book found within the Tolkien MS collection in the Bodleian library (Tolkien E16/20).
Many of the books in Tolkien’s “Celtic library” are dated by him, and it is notable that one third of them were bought between 1920 and 1926, most of them in 1922. Of course, that could be a reflection of the book-buying zeal of a young academic who finds himself in his first full-time job—in 1920 Tolkien was appointed Reader and four years later Professor in English Language at Leeds University, and many of his non-Celtic books are also dated within this period. But, revealingly, it was in 1922 that he started working with his colleague E. V. Gordon on the edition of the Middle English poem Sir Gawain and the Green Knight, which was finally published in 1925. During this period he also contributed to the “Philology: General Works” section of the Year’s Work in English Studies, for three consecutive years, presenting and reviewing academic works pertaining to philology. In his 1923 and 1924 articles he commented on publications of English place-names, including their Celtic elements, and he voiced his views on the ongoing debate on the adventus Saxonum and the role of the Celtic population of Britain (Tolkien “Philology 1923” 30-32; “Philology 1924” 58-59). His interest in Celtic Studies was, therefore, very much at the core of his academic work of this time.
The contents of Tolkien’s “Celtic Library” not only add to our knowledge of what he was familiar with in Celtic Studies; they can also occasionally offer insights into his sources for specific works. An example is “The Lay of Aotrou and Itroun.” This long poem was written during the nineteen-thirties but not published until 1945 in the Welsh Review. The tragic story of the childless lord and his disastrous dealings with a Corrigan has long been recognised as inspired by the legends of Brittany (Carpenter 167-68). Jessica Yates has discussed in detail the origins of Tolkien’s poem, and has contested Tom Shippey’s claim that Tolkien’s main source was “Le Seigneur Nann et la Fée” from Wimberly’s collection of English and Scottish ballads (1928). Yates argues that Tolkien could have equally started from “Aotrou Nann hag ar Corrigan” from Child’s collection English and Scottish Popular Ballads (1882) which Shippey also cites as one of the books that Tolkien certainly knew, although he does not associate it with “The Lay of Aotrou and Itroun” (65-66). Yates also refers to two translations of the folk poem mentioned by Wimberley (Keightley and Taylor) that Tolkien might have also known and Shippey, in his revised The Road to Middle-earth adds a note where he agrees that Keightley’s book could have also been a possible source of Tolkien’s poem (446). Yates, however, makes it clear that the ultimate source of the folk poem in both Child’s and Wimberley’s collection is named as Villemarqué’s Barzas-Breiz: Chants Populaires de la Bretagne. Villemarqué was the nineteenth-century folklorist who originally collected and recorded the poem. In an older article, Alexi Kondratiev had argued for Villemarqué as the source for Tolkien’s “Aotrou and Itroun,” although he admitted that there was no proof of Tolkien having read the Barzas-Breiz and thus concluded that Tolkien must have known the story from another source. Yates refers briefly to Kondratiev’s article and claims that the lack of proof does not mean that Tolkien did not read Villemarqué (66). Tolkien’s “Celtic Library” proves her right: Tolkien owned his own copy of Villemarqué’s Barzas-Breiz. He also owned the Lais of Marie de France, edited by Karl Warnke, which—according to Shippey—Tolkien imitated in “Aotrou and Itroun” (277). Both books, together with a few other volumes on Breton folklore and some books on the Breton language, are in “Tolkien’s Celtic Library” and most of them were bought between 1920-1922.
The Book of Lost Tales, the earliest version of Tolkien’s early nationalistic project for a “mythology for England,” included Celtic elements which I have discussed in detail before (Fimi). What I wish to demonstrate in this article is how Tolkien’s continuous involvement in Celtic Studies can account for an unbroken sequence of Celtic elements sneaking into Middle-earth, whether intentionally or not. In The Book of Lost Tales the story of the Tuatha Dé Danann already played an important role as an inspiration for the tragic story of the Gnomes’s departure from Valinor, and the whole framework of the Irish “Book of Invasions” was used as a model for the pseudo-historic “Seven Invasions of Luthany” (Fimi 161-64). In the next stage of the evolution of the Lost Tales, c. 1920, the narrative of “Ælfwine of England” emerged. The story of Ælfwine, like its predecessor, the story of Eriol, was conceived as an integral part of the “framework” for presenting the Lost Tales and associated the Lost Tales with England, and especially Anglo-Saxon England. However, Tolkien introduced an unexpected change to Ælfwine’s pedigree. Instead of being a pure Anglo-Saxon, Ælfwine is portrayed as being the son of Déor, a man “of English blood” and Éadgifu, “a maiden from the West, from Lionesse as some have named it since” (Lost Tales II 313). Now, Lionesse (or “Lyonesse,” as it is usually spelt) brings to mind the Arthurian legend; Cornish and Breton folklore about sunken lands, and also evokes a certain air of the medieval romances. Given his well-publicized rejection of “things Celtic,” could all of this sound more Un-Tolkienian? Nevertheless, the character of Éadgifu from Lionesse can serve as the first missing piece of the legendarium jigsaw-puzzle as a merging of Anglo-Saxon and Celtic traditions, rather than a favoring of the former and a rejection of the latter as Tolkien would have us believe. In the following parts of the article I will explore Tolkien’s work after the Book of Lost Tales, notably his writings between the 1920s and 1940s, and I will attempt to provide further evidence that supports the importance of his “Celtic Library” to his work. This library was not merely a marginal academic interest, but demonstrates his continuous fascination with the mythic and legendary past of Britain.
2. Time-travel Stories and the Blending of Celtic and Anglo-Saxon Traditions
Ælfwine’s Anglo-Saxon and Celtic ancestry is indeed an important change to the original conception of the Book of Lost Tales. Lyonesse is one of those enchanted places of Arthurian romance, usually associated with Tristan or Galahad. It has been claimed to have been originally Leonais in Brittany or Lothian in Scotland, but it is most often linked to a part of Cornwall which in legend was sunk under the sea. It is probable that Tolkien had Cornwall in mind as Lionesse, since in The Lost Road, the next of his writings in which Æelfwine appears, his wife (rather than his mother) is said to be from Cornwall (Lost Road 84-85).
The Lost Road began as part of an agreement between Tolkien and C. S. Lewis to write, respectively, a time-travel and a space-travel story (Letters 347, 378). In Lewis’s case the result was Out of the Silent Planet (1938). Tolkien’s very ambitious plan of time-travel, which remained unfinished, involved a series of fathers and a sons (always bearing names that could be translated as “Bliss-friend” and “Elf-friend”) re-living—by means of dreams—old myths and legends, concluding with the fall of Númenor. Tolkien only wrote two parts of the book: the “opening chapters,” concerning father and son of modern times, Alboin and Audoin, and the “Númenórean chapters,” concerning Elendil and Herendil of Númenor. Nevertheless, we have his notes and drafts of the “unwritten chapters” that would come between the “old” and the “modern” father-and-son stories. It is within these notes and drafts that one finds such Germanic myths and legends as the Lombard legend of Alboin, the semi-mythical semi-historical Norse story of the voyage to Vinland, and the legend of King Sheave. At the same time, the Anglo-Saxon poem The Seafarer would also become an integral part of one of the unfinished father-son stories, the “Anglo-Saxon” episode of the book, featuring Aelfwine and his son Eadwine. In the draft of that part, Ælfwine recites a slightly modified version of lines 36-38 and 44-46 of The Seafarer.
In The Seafarer a mariner recounts his wretched experiences at sea, but he also expresses his urge to live the seaman’s life. Tolkien’s translation of, and intervention in, these six lines reads:
The desire of my spirit urges me to journey forth over the flowing sea, that far hence across the hills of water and the whale’s country I may seek the land of strangers. No mind have I for harp, nor gift of ring, nor delight in women, nor joy in the world, nor concern with aught else save the rolling of the waves. (Lost Road 84)
It seems that the way Tolkien was planning to use the poem would be as an expression of Ælfwine’s desire to sail upon the western sea and find the “Straight Road,” the “Lost Road” that leads to Valinor and the Elves even after the world is “bent.” What is surprising is that he seems to associate Ælfwine’s voyage to the West with the immram genre of Irish tradition and specifically with the voyage of St. Brendan, which Tolkien was to use again in his later writings. Already in the outlines he was drafting for the Lost Road, among Northern stories with a shared Germanic background that Tolkien regarded as part of England’s past, he had included “a Tuatha-de-Danaan story, or Tir-nan-Óg,” and in another outline “the Irish legend of Tuatha-de-Danaan—and oldest man in the world” (Lost Road 77, 78).
The mention of Tir-nan-Óg, the otherworldly land of Irish tradition, is very significant here. The idea of the Western happy otherworld island and the geography and function of Valinor, are points of similarity between Tolkien’s mythos and Celtic legends that can hardly be missed. In the Lost Road, the title itself seems to refer to the lost road to the West, to Valinor where the Elves live. In other drafts and extracts of the Anglo-Saxon part of the story there are also references to the voyage of St. Brendan and of Maelduin and to the Insula Deliciarum (Lost Tales 80, 84-85). The location of the Celtic otherworld in overseas islands has been disputed in later scholarship (see Carey; Carney), and the generally accepted view today is that the concept is of ecclesiastical origin and goes back to the biblical earthly paradise (Dumville; Mac Mathúna 280-85).
However, when Tolkien was writing The Lost Road this hypothesis had still not been vigorously challenged. From 1924, Tolkien was already thinking of parallels between the Celtic otherworld and Valinor when he wrote a poem entitled “The Nameless Land.” In it he talks about a heavenly overseas island, which he compares to Tír-nan Óg and to the Christian Paradise. Although no name is given to the island, nor Elves are mentioned, he was probably thinking of Valinor, the blissful land of the Elves. The poem includes the following lines: “Such loveliness to look upon / no Bran nor Brendan ever won” (Lost Road 99), which refers to the voyage of St. Brendan and to the journey of Bran (“Immram Brain”), another account of a sea expedition involving otherworld islands, also belonging to the immram tradition (see Mac Mathúna).
Tolkien seems to have thought that there was an intriguing point of similarity between the Germanic legends of sea-voyages and the Irish immram tradition. The legend of King Sheave, especially as found in Beowulf, where the story is attached to Scyld, seems to have occupied a lot of his time and study (see Lost Road 92-98). In Beowulf, Scyld arrives as a child from the sea and his provenance remains mysterious. He becomes a renowned king of the Danes, and when he dies his body is placed in a ship, together with all of his treasures, and is given to the sea, sailing on its own accord to an unknown destination (see lines 26-52). In a quotation from an undated lecture by Tolkien on this subject, as given by his son Christopher, one can read that in Beowulf:
the poet is not explicit, and the idea was probably not fully formed in his mind—that Scyld went back to some mysterious land whence he had come. He came out of the Unknown beyond the Great Sea, and returned into It… In the last lines ‘Men can give no certain account of the havens where that ship was unladed’ we catch an echo of the ‘mood’ of pagan times in which ship-burial was practised. A mood in which the symbolism (what we should call the ritual) of a departure over the sea whose further shore was unknown; and an actual belief in a magical land or otherworld located ‘over the sea’, can hardly be distinguished. (Lost Road 95-96)
It seems that by incorporating in his legendarium the Celtic tradition, Tolkien was able to establish a pagan, pre-Christian otherworld that the Anglo-Saxons also knew, and that the poets of The Seafarer and Beowulf alluded to, a land where the real Elves were, a land that was central in his conception of “a mythology for England.” The Lost Road is the first instance in Tolkien’s writings in which the Irish tradition is clearly accepted as part of the same whole of Northern European myth and legend, and as part of what he was trying to “reconstruct” for England. Tolkien had started preparing an edition of The Seafarer with E. V. Gordon, which was left unfinished after he moved to Oxford. The notes they had kept were finally used by Gordon’s wife Ida in her 1960 edition of the poem (Gordon 17-18), and it is noteworthy that in the introductory remarks to that edition the Old English poem, along with other lyric-elegies, is compared with Celtic poetry of the same kind and is attributed to a “Celtic inspiration” (Gordon 17-18). Using The Seafarer in the same context with the story of St. Brendan and Tir-nan-Óg, and giving Ælfwine a wife from Cornwall—stating also that “the Welsh tongue is not strange to him” (Lost Road 84)—seems to confirm Tolkien’s acceptance of an unavoidable historical amalgamation of Celtic and Anglo-Saxon cultures.
Later on, Tolkien elaborated on some of the ideas that first appeared in The Lost Road and included them in The Notion Club Papers, another abortive time-travel book very much along the same lines as the former (Sauron 145-237). This was written between 1945-46, but this time the dreamers that travel to the past are Oxford dons that belong to a literary group, the “Notion Club,” which is very reminiscent of the Inklings, the group that C. S. Lewis and Tolkien himself belonged to. In this work the dreamers travel back in time in the realm of myth and legend, although this time they mainly find themselves in places from Tolkien’s own mythology, like Númenor and Valinor. However, during a gathering of the Notion Club, one of the members, Frankley, reads a poem he has written, entitled “The Death of St. Brendan” (Sauron 261-64). Here, the Saint is portrayed as narrating the most memorable of his legendary voyages to a younger monk, before he dies. The younger monk is already interrogating him about “islands by deep spells beguiled /where dwell the Elven-kind,” asking if he found “the road to Heaven or the Living Land” (Sauron 261). St. Brendan talks about an island where he and his companions saw a tree with white leaves. Suddenly the leaves of the tree fell and flew to the sky, and coming from them they heard:
a music not of bird,
not voice of man, nor angel’s voice;
but maybe there is a third
fair kindred in the world yet lingers
beyond the foundered land. (Sauron 263)
This incident is strictly modelled upon one of the episodes in the Navigatio Sancti Brendani abbatis (The Voyage of Saint Brendan the abbot), an early tenth-century Latin prose version of St. Brendan’s legend, written by an Irish author and imitating the immram genre of Irish medieval literature. In this episode the Saint comes to the “Paradise of Birds,” where he encounters a magnificent tree with white birds. He learns from one of them that they are spirits that fell as a result of Lucifer’s rebellion, not being part of his followers, but not part of the “faithful” either (O’Meara and Wooding 34, 36). Indeed, Tolkien has the members of the “Notion Club” discussing the sources of Frankley’s poem. The poet himself admits that: “I read the Navigatio Sancti Brendani, of course, once upon a time, years ago, as well as that early Anglo-French thing, Benedeit’s Vita” (Sauron 265), the latter being the early fourteenth century Anglo-Norman Voyage de Saint Brendan (see Burgess). Lowdham, another member of the Club, dismisses both sources as “rather dismal. Whatever merits they may have, any glimmer of a perception of what they are talking about is not one of them…” and he also adds for the latter source that “you won’t learn much about the West from that.” However, he refers to the actual episode of the tree and the birds in the Navigatio, by saying:
And the Tree in St. Brendan was covered with white birds that were fallen angels. The one really interesting idea in the whole thing, I thought: they were angels that lived in a kind of limbo, because they were only lesser spirits that followed Satan only as their feudal overlord, and had no real part, by will or design, in the Great Rebellion. But you make them a third fair race. (Sauron 265)
Indeed, in Frankley’s poem of St. Brendan’s voyages, the fallen angels have been transformed to Tolkien’s own Elves, and that makes the poem a good source of information about the “true West,” in contrast with the Navigatio. What should also be pointed out here is that in Irish folklore, in many occasions elves and fairies were considered to be actually fallen angels, a view that originated in an effort to impose biblical exegesis onto folk belief that was supposed to be a remnant of “pagan” religions of the past (Ó hógain 187-88).
Verlyn Flieger has added one more element of Celtic inspiration in the canvas of The Lost Road and The Notion Club Papers: that of the legends of sunken lands in Celtic sources. As a parallel to Plato’s Atlantis, she cites the Welsh Cantref of Gwaelod, the Breton Ker-Ys, the Irish Hy-Brasil, and the Cornish Lyonesse (Interrupted Music 125-30). Tolkien was certainly familiar with such legends, and indeed they might have played an important role in contextualizing his “Atlantis complex” within British (rather than Classical) tradition. The inclusion of the “new version of the Atlantis legend” in his mythology would, thus, be justified as an integral part of Britain’s past. The mention of Lyonesse or Lionesse brings to mind the Arthurian legend, and especially its French versions. The next two sections of this essay will explore Tolkien and the Arthurian legend as a continuation of the blending of traditions of the British Isles in Tolkien’s work.
3. The Continuing Fascination of the Arthurian Legend
Tolkien’s views on the Arthurian legend and the reasons why—for him—it did not qualify as “English” mythology were expressed in his 1951 letter to Milton Waldman (Letters 144). Flieger has discussed Tolkien’s charges against the Matter of Britain: its Christian elements, incoherence, lavish fantasy, and repetitiveness, and its “Britishness” rather than “Englishness” (“Matter of Britain”). Still, Tolkien’s fascination with the Arthurian legend persisted from his childhood to his mature years. Carpenter notes Tolkien’s enthusiasm for the Arthurian legend in his childhood (22). This enthusiasm might have been further enhanced by his school friend G. B. Smith and the latter’s keen interest in the Celtic—and particularly Welsh—tradition. Smith was an admirer of the “Mabinogion” tales, which he was re-reading while in the trenches of World War I, and of the Arthurian legend, particularly its Welsh origins rather than its more famous French versions. He was an aspiring poet, one of his longest poems, entitled “Glastonbury,” having an Arthurian subject. He also seemed to know the Welsh triads, he showed an interest for the Welsh language itself, and he esteemed the poetry of W.B. Yeats (Garth 7, 32, 55, 67, 82, 122, 195).
During 1925-27, Tolkien composed a story for his son Michael entitled “Roverandom,” which was published posthumously (1998). In that story Tolkien included as a character a white dragon, who turns out to be a familiar one in Arthurian tradition:
All the white dragons originally come from the moon, as you probably know; but this one had been to the world and back, so he had learned a thing or two. He fought the Red Dragon in Caerdragon in Merlin’s time, as you will find in all the more up-to-date history books; after which the other dragon was Very Red. Later he did lots more damage in the Three Islands, and went to live on the top of Snowdon for a time. People did not bother to climb up while that lasted—except for one man, and the dragon caught him drinking out of a bottle. That man finished in such a hurry that he left the bottle on the top, and his example has been followed by many people since. A long time since, and not until the dragon had flown off to Gwynfa, some time after King Arthur’s disappearance, at a time when dragons’ tails were esteemed a great delicacy by the Saxon Kings. (33)
Although humorous and light-hearted, this passage alludes to the famous story of the White and Red Dragons, symbolizing the Anglo-Saxons and the Welsh respectively, which is mostly known from Merlin’s prophecy in Geoffrey of Monmouth’s Historia Regum Britanniae (Book VII, Chapter III). Tolkien also alludes to the earlier version of the legend, recorded in the “Mabinogion” story “Lludd and Llefelys,” in which the two dragons end up at mount Snowdon in Wales.
At some point in the 1930s Tolkien started working on a long poem entitled “The Fall of Arthur.” The poem remained unfinished, and it has never been published or made available to consult in its entirety. As a result, the only noteworthy information about it can be traced in Tolkien’s own letters as well as in his authorized biography by Humphrey Carpenter, who had access to Tolkien’s manuscript. Carpenter describes the poem as “an individual rendering of the Morte d’Arthur,” and provides a very brief outline of the work, concentrating rather on the characters of Mordred and Guinever (Carpenter 168-69). Carpenter seems to be pointing to Malory’s Morte D’Arthur as an immediate source for Tolkien’s poem. However, Tolkien’s poem was written in alliterative metre, while Malory’s is in prose.
There are actually two very short fragments, of no more than forty-five lines in total, from the unpublished “Fall of Arthur” available to the researcher within Tolkien’s manuscripts at the Bodleian (Tolkien A 30/1, Folios 90-91). Although they constitute too small a sample for any valuable judgment of the poem, and some of the lines are very difficult to decipher due to Tolkien’s notoriously difficult handwriting, it is significant that Christopher Tolkien’s note on the Folio refers to the poem as “Morte Arthure.” This title would not point to Malory’s poem but rather to the Middle-English text known as the Alliterative Morte Arthure. This is a late fourteenth-century Arthurian romance, which, together with the almost contemporary Stanzaic Morte Arthur, forms the main English Arthurian tradition before Malory (Benson 2). Tolkien might have showed a preference for this poem rather than for Malory’s work because of it being part of the alliterative revival, a literary movement which began in the mid-fourteenth century and included such works as the anonymous Sir Gawain and the Green Knight, which Tolkien co-edited with his colleague E. V. Gordon and also translated into Modern English alliterative verse. Carl Phelpstead has discussed in detail the importance of the alliterative form for the Inklings, and its associations with Englishness and English identity, and has suggested that they attempted a “second alliterative revival” in the twentieth-century (457).
In addition, in his 1935 essay “The Alliterative Metre,” C.S. Lewis talks about “a return to our own ancient system, the alliterative line,” “against the long reign of foreign, syllabic metres in English.” He goes on to refer to the alliterative poetry of Auden, adding that “Professor Tolkien will soon, I hope, be ready to publish an alliterative poem” (Rehabilitations 119). The Tolkien poem that Lewis refers to can definitely be identified as the unfinished “Fall of Arthur” (Selected Literary Essays 15). It is known that the poem was read and praised by E. V. Gordon and R. W. Chambers, the latter describing it as: “great stuff—really heroic, quite apart from its value as showing how the Beowulf metre can be used in modern English” (Carpenter 168). In this context, Tolkien’s desire to finish the poem as late as 1955 (Letters 219) can be justified.
Apart from a tribute to the alliterative form, though, the poem could also be a tribute to the Arthur of the English, to the English Arthurian tradition. Unfortunately, Tolkien’s “Fall of Arthur” is not available to consult and compare with either Malory’s Morte d’Arthur, or the Alliterative Morte Arthure, but both works would qualify as depicting an English Arthur, as opposed to a Welsh, or even worse for Tolkien, a French one. In an article written around the same time when Tolkien’s Arthurian poem was being composed, E. V. Gordon published an article (together with E. Vinaver) on the Alliterative Morte Arthure, discussing the new light thrown on it by the discovery of the Winchester manuscript of Malory’s Morte Arthure. This discovery had caused excitement in the academic world and might have contributed to Tolkien’s Arthurian venture.
After the 1930s and “The Fall of Arthur” Tolkien’s next engagement with the Arthurian legend was in his unfinished work The Notion Club Papers, in which discussions held by the members of the Notion Club reveal once more Tolkien’s continuous fascination with the story of Arthur. In the record of Night 65, after a missing leaf in the account, Jeremy and Frankley start a conversation about myth and reality, about mythical and historical truth. Jeremy argues that there is truth in myths, although this would not necessarily be the conventional “scientific” truth we think of today. He also argues that “real details,” “facts.” like real historical personalities, are “caught up” in myth: the example he uses is that “There was a man called Arthur at the centre of the cycle” (Sauron 227). Frankley is quick to respond and reject the Arthurian romances as “real” but Jeremy insists that they might be “real” in a different way than “true past events” are. This argument about a historical Arthur being “caught up” in myth had been explored previously by Tolkien in his 1939 Andre Lang lecture “On Fairy-stories.” Talking about “the Pot of Soup, the Cauldron of Story,” and on how folk stories are born and developed, he writes:
It seems fairly plain that Arthur, once historical (but perhaps as such not of great importance), was also put into the Pot. There he was boiled for a long time, together with many other older figures and devices, of mythology and Faerie, and even some other stray bones of history (such as Alfred’s defence against the Danes), until he emerged as a King of Faerie. (MC 126)
This seems like a conclusive view, following the model of folklore theory. Still, the historicity of Arthur, and the lack of enough documents for the study of the origins of the Arthurian legend seemed to concern Tolkien later on, when he was writing the Notion Club Papers. The record of the Notion Club for Night 61 includes a dream scene that Ramer narrates to the rest of the members. This is the image of an old librarian looking through “a volume made up of various manuscript-fragments bound together, probably in the sixteenth century.” Ramer continues:
In the remembered bit of the dream I knew I had been able to read the page before he turned over, and that it was not English; but I could remember no more than that—except that I was delighted, or he was. Actually it was a leaf, a unique fragment of a MS. in very early Welsh, before Geoffrey, about the death of Arthur. (Sauron 192)
Even the hint of the existence of such a manuscript would not only be “of superlative importance in the study of the Arthurian legend” as Christopher Tolkien notes (Sauron 216), but would re-open heated debates about the historicity of Arthur and thorny issues about national identity and culture. Geoffrey of Monmouth’s Historia Regum Britannie has always been considered as the work that popularized Arthur and made him an internationally recognized hero, but also as a work that gave Arthur a mythical aura, thus obscuring the origins of the legend, which might point to the historical Arthur. Indeed, most academic work on the historicity of Arthur concentrates by default on the pre-Galfridian material.
Tolkien’s rejection of the Arthurian legend as an authentic part of England’s heritage must be taken with a pinch of salt, and the same is valid for his whole reaction to “things Celtic,” as I have argued elsewhere (Fimi 156-70). Tolkien’s Anglo-Saxonism was a major strand of his mythology, indeed instrumental in the creation of the Middle-earth saga, but after the Lost Tales he seems to have started thinking in more “British” terms. After the Anglo-Saxon material that Tolkien looked up to and used in his legendarium, the Middle English literature followed. Middle English literature was under the shadow of French cultural influence, which was becoming dominant in the rest of Europe at the time. The French form of the romance was introduced in England, and the Arthurian legend was re-introduced in its French guise, notably through Chrétien de Troyes’s romances. Some of the most important texts of that period were either translated or adapted from French sources. Tolkien held in great esteem such Middle English texts as Sir Gawain and the Green Knight and Sir Orfeo, two works that he had studied extensively. His own edition of Sir Gawain (with E. V. Gordon) recognizes the French influence on the poem (Tolkien and Gordon xii-v), while in his editorial notes to his “Middle English Version” of Sir Orfeo he seems to subscribe to the theory that the poem is a translation or adaptation of a lost Old French original (Hostetter 104). The Arthur of the English was a fusion of its Celtic origins, its French re-working, and its Middle-English context. Tolkien’s attempt for an Arthurian poem modelled—at least in terms of metre—upon the Alliterative Morte Arhure was an acknowledgment of the Arthurian legend as an integral part of England’s cultural and mythical heritage, whatever its historical origin.
4. Middle-English, French and Celtic: “The Lay of Leithian”
As briefly referred to above, in the early 1920s Tolkien gave his key character Ælfwine a mother from Lionesse, the romance country of Lyonesse of the Arthurian legend. In the period 1925-1927, before the Lost Road venture, Tolkien began a long poem on the story of Beren and Lúthien, a story already told in the Book of Lost Tales. The poem was called “The Lay of Leithian,” its extended title being “The Gest of Beren son of Barahir and Lúthien the Fay called Tinúviel the Nightingale, or the Lay of Leithian Release from Bondage.” The poem seems to have a Middle-English atmosphere. The word “gest” used in its title would allude to the French chansons de geste, mainly associated with the “matter of France.” However, the term has also been used for accounts of the Arthurian legend in France and Britain. As Christopher Tolkien notes, lines 2298-9 and 2348-9 of the Lay come more or less directly from lines 285-6 of Sir Orfeo and seem also to echo Tolkien’s translation of these lines (Lays 237, 238, 248). In addition to that, Tolkien also uses in the Lay two archaic words he also used in his translation of the Middle-English poem Pearl: “ruel-bone” for “some kind of ivory,” and “stared” for “shone,” the latter coming straight from the original (Lays 236, 266, 371; Tolkien Sir Gawain 92, 94). But probably the most striking French-cum-Arthurian reference in the Lay is the location of Thingol’s kingdom, which is initially called Broceliand, and in a later draft Broseliand (Lays 158, 159, 160).
The forest of Brocéliande is one of the most famous Arthurian locations associated with Brittany. It features in Chrétien de Troyes’s Yvain, it is mentioned in the writings of Giraldus Cambrensis, and its marvels are described by Wace, an Anglo-Norman poet who wrote the Geste des Bretuns, today better known as the Roman de Brut. The Roman de Brut was based on Geoffrey of Monmouth’s Historia and of course covered the story of Arthur, which was becoming more and more popular at that time. Tolkien’s term Broceliand/Broseliand underwent much consideration and change and ended up in the much more familiar Tolkienian name Beleriand, but if the poem is read with the reader substituting Beleriand for Broceliand/Broseliand, as it was initially, then its Arthurian/Celtic ambience becomes immediately discernible.
The most telling sign is the motif of the relationship of a fairy woman with a mortal man, which had always been a favorite Celtic theme. Indeed, such “fay” women are also frequently encountered in French Arthurian romances. In the Lay, both Lúthien and her mother Melian are referred to as “fays” (Lays 153, 172, 229), and their stories follow that motif: Lúthien is an elf or fairy who gives her love to a mortal man, while in the case of Thingol and Melian we have an analogue of the same idea: Thingol is an elf himself, but Melian is a Maia, a higher being than Thingol and closer to the Valar. Marjorie Burns has also noted Melian’s “Arthurian heritage,” remarking that the various names that Tolkien considered for her in The Book of Lost Tales, including Wendelin, Gwendeling, Gwenniel, Gwenethlin, Gwendhiling, Gwendelin, and Gwedhelin, have “a clear Arthurian ring” (196). Especially the initial g or w of the name brings to mind the names “Waynor” and “Gaynor,” used interchangeably for Guinevere in the alliterative Morte Arthure (see for example l. 84 and l. 233).
In a 1954 letter to Naomi Mitchison Tolkien seems to openly contradict his 1937 letter to Allen and Unwin, in which he refutes the characterisation of his works as “Celtic.” He writes:
The living language of the Western Elves (Sindarin or Greyelven) is the one usually met, especially in names. This is derived from an origin common to it and Quenya; but the changes have been deliberately devised to give it a linguistic character very like (though not identical with) British-Welsh: because that character is one that I find, in some linguistic moods, very attractive; and because it seems to fit the rather “Celtic” type of legends and stories told of its speakers. (Letters 176)
Which are these “‘Celtic’ type of legends” that the Grey Elves, the Sindar, seem to have? I would argue that it is exactly the story of the “Lay of Leithian,” the story of Beren and Lúthien, and the previous parallel story of Melian and Thingol. Indeed, this seems to be the main “legend” of the Sindar, who speak a language much like Welsh in its phonetic structure. Melian is referred to as a “fay” throughout the Book of Lost Tales (Lost Tales I 120; Lost Tales II 9, 10, 43, 76, 96, 233) and there are further Celtic elements in the original “Tale of Tinúviel,” elements that were later further developed in “The Lay of Leithian.” The original idea of the “Girdle of Melian” seems to be very much like the enchanted forests of the Arthurian romances, so that the name Broceliand/Broseliand used in the Lay seems to fit very well:
Hidden was his dwelling from the vision and knowledge of Melko by the magics of Gwendeling the fay, and she wove spells about the paths thereto that none but the Eldar might tread them easily, and so was the king secured from all dangers save it be treachery alone. (Lost Tales II 9; my emphasis)
Also, both the Tale and the Lay make use of a triad, which seems to work as a mnemonic device, exactly in the same way as the Welsh triads were used. By arranging names, place-names or storylines in groups of three, the bards could recall more easily a major part of the body of orally preserved Welsh repertoire of myths and legends (Bromwich lxv). Both in the Tale and the Lay, the triad concern Lúthien’s brother, Dairon, who was one of “the three most magic players of the Elves,” listing Tinfang Warble and “Ivárë who plays beside the sea” as the other two (Lost Tales II 10). The triad in the Lay runs thus:
Such players have there only been
thrice in all Elfinesse, I ween:
Tinfang Gelion who still the moon
enchants on summer nights of June
and kindles the pale firstling star;
and he who harps upon the far
forgotten beaches and dark shores
where western foam for ever roars,
Maglor whose voice is like the sea;
and Dairon, mightiest of the three. (Lays 174)
The use of triads in Tolkien’s legendarium is quite rare, and it cannot be accidental that two of them appear in the most famous legend of the Sindar.
Finally, some later evidence that Tolkien was still thinking of the story of Beren and Lúthien as Celtic/Arthurian, can be found in “The Etymologies,” written in the late 1930s, and in the map that accompanied the “Later Quenta Silmarillion,” written in the 1950s. In the “Etymologies,” under the stems GAT(H)-, GARAT- and THUR-, the word “Garthurian” appears, explained as “Fenced Realm,” and is used to refer to Doriath (Lost Road 358, 360, 393). In the map associated with the “Later Quenta Silmarillion,” we find the place-name “Garthurian” at the edges of Doriath (Morgoth 183). In his commentary on the map Christopher Tolkien adds his father note that “the Noldor often used the name Arthurien for Doriath, though this is but an alteration of the Sindarin Garthurian ‘hidden realm’” (Morgoth 189).
Tolkien’s linguistic invention was a serious philological venture, which simulated the process through which real languages develop. For example, when composing his Elvish languages, Tolkien would establish first the common Proto-Elvish root and then modify it to fit its development into words in Quenya and Sindarin, following the phonetic rules he had established for these languages. However, “inventing” the root first was not always the case. As Christopher Tolkien notes, already from the period of the Lost Tales, in some cases the word was “already there” and its etymology was worked out backwards (Lost Tales I 246). In the early documents of the “Qenya” and the “Gnomish” Lexicons, it seems that some sort of “historical punning” was present as Christopher Tolkien notes on his introduction. I would suggest that “Arthurien” and “Garthurien” for Doriath is such a “historical punning,” linking this Middle-earth location with its original conception as Broseliand/Broceliand, and alluding to the “Celtic” type of legends of its people.
5. Merging Traditions
In his 1955 O’Donnell lecture “English and Welsh,” Tolkien referred to the appeal of Welsh as a “native language,” noting that “for satisfaction and therefore for delight… we are still ‘British’ at heart” (MC 194). This acceptance of the term “British” is remarkable. As late as 1943 Tolkien still insisted on the distinction of English and British when talking about patriotism: “For I love England (not Great Britain and certainly not the British Commonwealth (grr!)” (Letters 65). However, the transition from English (which would have primarily meant Anglo-Saxon), to British (which would include the merging of different traditions, including Celtic and French, and would start with the Middle English period) was gradual in Tolkien’s mind and was also reflected as a gradual process in his creative writings. As discussed above, the Book of Lost Tales already included Celtic elements, and the Lost Road made an explicit declaration of the blending of Irish and Anglo-Saxon traditions. At the same time the Sindar were always associated with a language very similar to Welsh, and their most famous legend, the story of Beren and Lúthien, was, from early on, conceived as a “‘Celtic’ type of legend,” very close to medieval Arthurian romances.
Finally, Arthur as a historical and mythical figure was always present in Tolkien’s imagination, leading to the writing of one more unfinished work. It seems that by the end of his life Tolkien explicitly acknowledged what is true for many nation-states and their history: that purity of tradition is not a realistic part of the process of nation-building, and that significant merging of peoples, languages, and cultural elements occurs. Especially in the case of the British Isles, a long troubled history of invasions, conquests, and linguistic amalgamations created the modern state of the United Kingdom, and the mythology of Middle-earth, either consciously or not, reflects this process right from its original conception. Tolkien’s 1954 letter to Naomi Mitchison quoted above, together with his conclusion of his 1955 O’Donnell lecture, seem to be the reflections of a more “mature” Tolkien upon his own work, perhaps realising that his “mythology for England” eventually became a “mythology for Britain.”
 The collection also includes a separate edition of one of the “Mabinogion” tales, Peredur Ab Efrawc (Meyer).
 The role of Eriol and Ælfwine as mediators in the “Lost Tales” and their role in linking Tolkien’s mythology with England have been discussed by Verlyn Flieger (“Footsteps”) and Michael Drout.
 For an overview see The New Arthurian Encyclopedia (Lacy 287-88).
 Norma Roche has explored the significance of the story of St. Brendan and the idea of a Western Otherworld island in Tolkien’s work. Her article was published before the ninth volume of the History of Middle-earth was released (1992), which contained The Notion Club Papers, and so her discussion is restricted to the evidence in The Lost Road. More recently, Verlyn Flieger has also discussed Tolkien’s debt to the story of St. Brendan, including how the poem “Imram” is now recognised as part of The Notion Club Papers (Interrupted Music 130-34).
 The projects of academic collaboration, fruitful or not, between Tolkien and Gordon, including the edition of The Seafarer, have been discussed by Anderson (“Industrious”). Wilcox has also referred to
their collaboration on The Seafarer in her discussion of that poem’s impact on Tolkien’s literature.
 Apart from Pedeir Keinc y Mabinogi, Tolkien’s Celtic Library includes five more books bequeathed to him by G.B. Smith. These are An English-Welsh Pronouncing Dictionary (Spurrell), Hanes A Chan (Edwards 1908), Essai sur La Composition du Roman Gallois de Peredur (Williams), Gwaith Samuel Roberts (Roberts), and Gwaith Twm o’r Nant (Edwards 1909).
 Tolkien also refers to the story of “Lludd and Llefelys” in his O’Donnell lecture “English and Welsh” (MC 189). For a commentary on the extract quoted see also Scull’s and Hammond’s notes in their edition of Roverandom (101-3).
 For an introduction to the poem see Johnson.
 Phelpstead also refers to the alliterative verses contained in The Lord of the Rings, and claims that “The enormous popularity of this book means that these verses must be the most widely read alliterative poetry of the twentieth century, if not of any period” (444).
 For the forest of Brocéliande see The New Arthurian Encyclopedia (Lacy 55).
 It seems that Tolkien was undecided for a while whether Beren would be a man or an elf. Christopher Tolkien refers to the lost original of the “Tale of Tinúviel” where Beren was a man, in contrast to Beren being an elf in the extant version of the Tale (Lost Tales II 52, 139). However, from “The Lay of Leithien” and on Beren’s identity is securely fixed as that of a man.
 This process can be attested in the fact that when he attempted to record his languages in a “dictionary form” he often listed the Proto-Elvish roots alphabetically, with derived words under every entry. This is how, for example, the “Qenya Lexicon” and the “Etymologies” work. This account of Tolkien’s creation of languages is, of course, an over-simplified one that does not reflect the successive layers of revisions and alterations.
 For example, the root SAHA, meaning “be hot” produces (apart from saiwa, “hot” or sara “fiery”) the word Sahora, “the South.” This “historical punning” seems also to be true of the word Atalantie meaning “Downfall” in Quenya. Tolkien himself seems to be making a semi-humorous note on this “coincidence,” by writing: “It is a curious chance that the stem √talat used in Q[uenya] for ‘slipping, sliding, falling down’, of which atalantie is a normal (in Q) noun-formation, should so much resemble Atlantis” (Letters 347).
 For Tolkien’s conception of the English vs. British identity and the historical background of how these terms were used at the time see my discussion (Fimi 159-61).
Anderson, Douglas A. “‘An Industrious Little Devil’: E. V. Gordon as friend and collaborator with Tolkien.” In Tolkien the Medievalist, edited by Jane Chance. London and New York: Routledge, 2003.
Anderson, Douglas A. “Personal Library.” In The J. R. R. Tolkien Encyclopedia: Scholarship and Critical Assessment, ed. Michael D. C. Drout. New York: Routledge, 2006.
Benson, Larry D. “Introduction.” In King Arthur’s Death: The Middle English Stanzaic Morte Arthur and Alliterative Morte Arthure, edited by Larry D. Benson, revised by Edward E. Foster. Kalamazoo, Michigan: Medieval Institute Publications, 1994.
Bromwich, Rachel, ed. and trans. Trioedd Ynys Prydein: The Welsh Triads. Cardiff: University of Wales Press, 1961.
Burgess, Glyn S., trans. “The Anglo-Norman Version.” In The Voyage of Saint Brendan: Representative Versions of the Legend in English Translation, edited by W.R.J. Barron and Glyn S. Burgess. Exeter: University of Exeter Press, 2002.
Burns, Marjorie. Perilous Realms: Celtic and Norse in Tolkien’s Middle-earth. Toronto: University of Toronto Press, 2005.
Carey, J. “Time, Space and the Otherworld.” Proceedings of the Harvard Celtic Colloquium 7 (1982): 1-27.
Carney, J. “Review of Carl Selmer’s Navigatio Sancti Brendani Abbatis.” Medium Aevum 32 (1963): 42-43.
Carpenter, Humphrey. Tolkien: A Biography. London: Allen and Unwin, 1977.
Child, Francis James, ed. English and Scottish Popular Ballads. Boston: Houghton Mifflin, 1882.
Drout, Michael D. C. “A Mythology for Anglo-Saxon England.” In J.R.R. Tolkien and the Invention of Myth: A Reader, edited by Jane Chance. Lexington, KY: University Press of Kentucky, 2004.
Dumville, David N. “Echtrae and Immram: Some Problems of Definition.” Ériu 27 (1976): 73-94.
Edwards, J. M. Hanes A Chan. Newport: Southall, 1908.
———, ed. Mabinogion (o Lyfr Coch Hergest). Wrecsam: Hughes & Son, 1921.
Edwards, Thomas. Gwaith Twm o’r Nant. Llanuwchllyn: Ab Owen, 1909.
Evans, J. G., ed. White Book Mabinogion. Pwllheli: Evans, 1907.
Fimi, Dimitra. “‘Mad Elves’ and ‘Elusive Beauty’: Some Celtic Strands of Tolkien’s Mythology.” Folklore 117, no. 2 (2006):156-70.
Flieger, Verlyn. “J. R. R. Tolkien and the Matter of Britain.” Mythlore 23 (2000): 47-59.
_____. “The Footsteps of Ælfwine.” In Tolkien’s Legendarium: Essays on The History of Middle-earth, edited by Verlyn Flieger and Carl F. Hostetter. Westport, Connecticut; London: Greenwood Press, 2000.
_____. Interrupted Music: The Making of Tolkien’s Mythology. Kent, OH: Kent State University Press, 2005.
Garth, John. Tolkien and the Great War: The Threshold of Middle-earth. Boston: Houghton Mifflin, 2003.
Gordon, E. V. and Eugène Vinaver. “New light on the text of the alliterative ‘Morte Arthure.’” Medium Ævum 6 (1937): 81-98.
Gordon, I. L.. ed. The Seafarer. London: Methuen, 1960.
Guest, C. (Lady), trans. The Mabinogion. London: Dent, 1913.
Hersart de la Villemarqué, Théodore Claude Henri. Barzas-Breiz: Chants Populaires de la Bretagne, Tome I and II. Paris: Franck, 1846.
Hostetter, Carl. F. “Sir Orfeo: A Middle English Version By J.R.R. Tolkien.” Tolkien Studies 1 (2004): 85-123.
Johnson, Lesley. “The Alliterative Morte Arthure.” In The Arthur of the English: The Arthurian Legend in Medieval English Life and Literature, edited by W.R.J. Barron. Cardiff: University of Wales Press, 1999.
Keightley, Thomas. The Fairy Mythology, Illustrative of the Romance and Superstition of Various Countries. London: Bell & Daldy, 1873.
Kondratiev, Alexi. “New Myths for Old: The Legacy of Iolo Morgannwg and Hersard de le Villemarque.” Mythlore 36 (1984): 43-46.
Lacy, Norris J., ed. The New Arthurian Encyclopedia. London, Chicago: St. James Press, 1991.
Lewis, C. S. Rehabilitations and Other Essays. London: Oxford Univ. Press, 1939.
________. Out of the Silent Planet. London: Bodley Head, 1938.
________. Selected Literary Essays, ed. Walter Hooper. Cambridge: Cambridge Univ. Press, 1969.
Mac Mathúna, Séamus, ed. and trans. Immram Brain: Bran’s Journey to the Land of the Women. Tübingen: Niemeyer, 1985.
Meyer, K., ed. Peredur Ab Efrawc. Leipzig: Hirzel, 1887.
Mühlhausen, L., ed. Die vier Zweige des Mabinogi. Halle: Niemeyer, 1925.
Ó hógáin, Dáithí. Myth, Legend and Romance: An Encyclopaedia of the Irish FolkTradition. London: Ryan, 1990.
O’ Meara, John J. and Wooding, Jonathan M., trans. “The Latin Version.” In The Voyage of Saint Brendan: Representative Versions of the Legend in English Translation, edited by W.R.J. Barron and Glyn S. Burgess. Exeter: University of Exeter Press, 2002.
Phelpstead, Carl. “Auden and the Inklings: An Alliterative Revival.” Journal of English and Germanic Philology 23, no. 4 (2004): 433-57.
Rhŷs, J. and Evans J.G., eds. Red Book of Hergest, Vol. I. (Mabinogion). Oxford: Evans, 1887.
Roche, Norma. “Sailing West: Tolkien, the Saint Brendan Story, and the Idea of Paradise in the West.” Mythlore 66 (1991): 16-20.
Roberts, Samuel. Gwaith Samuel Roberts (S.R.), edited by Owen M. Edwards. lanuwchllyn: Ab Owen, 1906.
Shippey, T. A. J.R.R. Tolkien: Author of the Century. London: HarperCollins, 2001.
________. The Road to Middle-earth: Revised Edition. London: HarperCollins, 2005.
Spurrell, William. An English-Welsh Pronouncing Dictionary. Caerfyrddin: W. Spurrell, 1909.
Taylor, Tom, trans. Ballads and Songs of Brittany. London: Macmillan, 1865.
Tolkien, J.R.R. “Philology: General Works.” The Year’s Work in English Studies IV (1923): 20-37.
________. “Philology: General Works.” The Year’s Work in English Studies V (1924): 26-65.
________. “The Lay of Aotrou and Itroun.” The Welsh Review 4:4 (1945): 254-66.
________, trans. Sir Gawain and the Green Knight, Pearl, and Sir Orfeo. London: George Allen & Unwin, 1975.
________. Roverandom, edited by Christina Scull and Waynne G. Hammond. London: HarperCollins, 1998.
Tolkien, J.R.R. and Gordon, E. V., eds. Sir Gawain and the Green Knight. Oxford: Clarendon Press, 1925.
Warnke, Karl, ed. Die Lais der Marie de France. Halle: M. Niemeyer, 1900.
Wilcox, Miranda. “Exilic Imagining in The Seafarer and The Lord of the Rings.” In Tolkien the Medievalist, edited by Jane Chance. London; New York: Routledge, 2003.
Williams, Mary Rhiannon. Essai sur la Composition du Roman Gallois de Peredur. Paris : H. Champion, 1909.
Wimberly, Lowry Charles. Folklore in the English and Scottish Ballads. Chicago, IL: The University of Chicago Press, 1928.
Villemarqué (Hersart de la), Théodore Claude Henri. Barzas-Breiz: Chants Populaires de la Bretagne. Tome I and II. Paris: Franck, 1846.
Yates, Jessica. “The Source of ‘The Lay of Aotrou and Itroun.’” In Leaves from the Tree: J.R.R. Tolkien’s Shorter Fiction, edited by T. A. Shippey, et al. 63-71. London: Tolkien Society, 1991.