The History of the Hobbit

by John D. Rateliff

  . . . the necklace of Girion, Lord of Dale, made of five hundred emeralds green as grass, which he gave for the arming of his eldest son in a coat of dwarf-linked rings the like of which had never been made before, for it was wrought of pure silver to the power and strength of triple steel. But fairest of all was the great white gem, which the dwarves had found beneath the roots of the Mountain, the Heart of the Mountain, the Arkenstone of Thrain.13

  Obviously, since in the revised story this gem had never been owned by Girion, the old name ‘Gem of Girion’ had to be replaced. The choice of Arkenstone is significant, since in other writings Tolkien was making at the same time he was using a variant of this same name as a term for the Silmarils themselves, forging a link between the Jewels of Fëanor and the Arkenstone of Thrain in the legendarium.

  Thus, in the ‘(Earliest) Annals of Valinor’, the entry for the Valian Year 2500 (that is, the equivalent of 25,000 solar years from the time the Valar entered Arda), reads:

  About 2500 the Noldoli [Noldor] invented and began the fashioning of gems; and after a while Fëanor the smith, eldest son of Finwë chief of the Noldoli, devised the thrice-renowned Silmarils, concerning the fates of which these tales tell. They shone of their own light, being filled with the radiance of the Two Trees, the holy light of Valinor, blended to a marvellous fire (HME IV.265).

  This work is associated with the 1930 Quenta and only very slightly later in date – that is, contemporaneous with Tolkien’s work on The Hobbit (HME IV.262). And among the very earliest work Tolkien did on the Annals (ibid., page 281) was an Old English version by Ælfwine/Eriol, the frame narrator of The Book of Lost Tales, in which the entry given above is translated thusly:

  MMD Hér þurh searucræftas aþóhton and beworhton þá Nold-ielfe gimmas missenlice, 7 Féanor Noldena hláford worhte þá Silmarillas, þæt wæron Eorclanstánas (ibid. 282).

  Literally translated, this remarkable passage reads:

  [The Year] 2500. Here through cunning craft/artistic skill the Noldor elves devised (‘a-thought’) and created (‘be-worked’) many gems, & Fëanor the Noldor lord wrought the Silmarils, that were precious/holy stones [Eorclanstánas or ‘Arkenstones’].

  Furthermore, in a later draft of the same work in Mercian dialect, the fictional translator ‘Ælfwine of Ongulcynne’ (Elf-friend of England) lists the three parts that make up ‘The Silmarillion’ – The Annals of Valinor, the Annals of Beleriand, and the Quenta – noting ‘and þes þridda dæl man éac nemneð Silmarillion þæt is Eorclanstána gewyrd’, which translates as ‘and this third part is also named “Silmarillion”’; that is ‘[the] history/fate [of the] Precious/Holy Stones’ (HME IV.291). The equivalent Eorclanstána = Silmarils also appears in Ælfwine’s Old English translation of part of the ‘Annals of Beleriand’, which date from about the same time as the complementary ‘Annals of Valinor’:

  Morgoþ . . . genóm þá eorclanstánas Féanóres . . . ond þá eorclanstánas sette he on his isernan helme [‘Morgoth . . . stole the silmarils of Fëanor . . . and the silmarils he set in his iron crown’].

  —HME IV.338.

  The idea that the Arkenstone could be a Silmaril, or was at least somehow linked to the Silmarils in Tolkien’s mind, has additional support from the philological roots of the word. As Jacob Grimm pointed out back in 1844, there was little stone-lore in Teutonic mythology, but foremost among what he discovered he cites the ‘time-honoured myth’ of the holy iarkna-steinn of the Elder Edda, listing the Old English equivalent (eorcan-stân) and postulating Gothic (áirkna-stáins), and Old High German (erchan-stein) forms (Jacob Grimm, Teutonic Mythology, tr. Stallybrass [1883]; vol. III page 1217).14 Furthermore, within the Edda, the term is at one point applied to gems made by craft, not natural stones.15 In Old English, the most famous usage of eorcanstan occurs in Beowulf (line 1208a), where it appears under the variant spelling eorclanstanas; interestingly enough, it is used there to describe a wonderful jeweled necklace of gold and gems given to Beowulf by Queen Wealhtheow (Hrothgar’s consort).16 Tolkien’s source, however, may have lain not in Beowulf but in Cynewulf’s Christ – the same work from which he took the name ‘earendel’ (line 104) some twenty years earlier – where earcnanstan (‘precious/holy stone’) appears in line 1194 as a metaphor for Christ [cf. Gollancz’s edition, pages 100–101, and Cook’s edition, pages 45 & 200]. This is made somewhat more likely by the spelling found in Cynewulf’s poem, and the fact that words in Old English which began with eor- (eorl, eorth, eornoste) generally became ear- in modern English (earl, earth, earnest), whereas words beginning ear- typically became ar- (e.g., earc > ark [= Noah’s Ark]). Alternately, rather than modernization of the Old English, ‘Arkenstone’ as it appears in The Hobbit could represent an anglicization from the Old Norse. It is, after all, a dwarven stone and hence should have a dwarven name, and all the names Tolkien gives the dwarves in The Hobbit (this being several years before the creation of a distinctive Dwarven language, e.g., Khuzdul), from Fimbulfambi to Dain, are Old Norse.

  One element worth stressing that would link the Arkenstone even more closely with the Silmarils is the implication of the eorcan element in its name. Although usually translated simply as ‘precious’ (that is, highly valuable) and generally applied to any gemstone (e.g., at various times to topaz, opal, and pearl), Grimm stressed that the Gothic equivalent airkna meant rather ‘holy’, and was so used in the Gothic translation of the New Testament (the oldest surviving document in any Germanic language). G. H. Balg (A Comparative Glossary of the Gothic Language, 1887)17 goes further, linking airkns to the Greek argos or ’apyós (‘bright’) and Sanskrit arjuna (‘bright, pure’).18 The Silmarils are referred to over and over again in the legendarium as the ‘holy jewels’, who burn evil-doers (such as Melkor the Morgoth, Karkaras/Carcharoth, Maidros/Maedhros and Maglor) at the touch. We have no way of knowing if the Arkenstone shares this same power, since within our story it never comes into direct contact with any evil-doer (or, if we do assume it shares this characteristic with the other Silmarils, then its failure to scorch Bilbo is a testament to the integrity of his intentions and the rightness of his action in purloining it, concealing it from Thorin, and giving it to Bard as a hostage for the dragon-slayer’s due portion of the treasure). Certainly, although like Beren’s Silmaril the Arkenstone inspires fierce possessiveness in all who behold it, so that not even Bilbo can give it up without a pang, it seems nonetheless pure and innocent (again, like the Silmarils); no pejorative or sinister terms are ever employed in describing it (not even the obvious one within Tolkien’s moral lexicography, ‘precious’, a word never applied to the Arkenstone within The Hobbit). Like the Silmarils in the main branch of the legendarium, and unlike the One Ring in the sequel, the Arkenstone inspires greed but is not itself malicious in any way:

  ‘The Arkenstone! The Arkenstone!’ murmured Thorin in the dark, half dreaming with his chin upon his knees. ‘It was like a globe with a thousand facets; it shone like silver in the fire-light, like water in the sun, like snow under the stars, like rain upon the Moon!’

  —First Typescript, pasteover on typescript page 125 (1/1/62:11).

  The original description of the Gem of Girion as a bright, shining jewel, a globe with many facets (pp. 514–15), which shone of its own light yet catches and magnifies all light that falls on it (page 579), sounds remarkably like Tolkien’s descriptions of the Silmarils. Unfortunately we cannot compare them in detail, because for all their importance to the story Tolkien only rarely describes the Silmarils themselves, and then more in terms of their effect on the viewer than in appearance. For example, the earliest account of their creation (BLT I.128) lists the materials Fëanor assembled – the sheen of pearls, phosphorescence, lamp- and candle-light reflected through other gems, the half-colours of opals, and the all-important Light of the Two Trees – but aside from their radiance the only specific detail about their appearance is that ‘[he gave] all those magic lights a body to dwell in of such perfect glass as he alone could ma
ke’, implying that they were clear.19 Although we are told Fëanor started by acquiring ‘a great pearl’ we could not even tell from this account whether the Silmarils were smooth or faceted (the pearl cannot have provided the actual body for the first Silmaril, since he has only one such pearl yet makes three Silmarils before he runs out of materials). References in the alliterative poems to ‘fair enchanted globes of crystal’ (‘The Flight of the Noldor from Valinor’, lines 139b-140a; HME III.135) and ‘thrice-enchanted globes of light’ (‘The Lay of Leithian’, line 1642; HME III.212) imply a smooth sphere, but the descriptions of them in the 1926 ‘Sketch’ (HME IV.14) and 1930 Quenta (e.g., HME IV.88) are too cursory to provide any details beyond that they shine with their own inner light. Not until the 1937 Quenta Silmarillion, which of course postdates The Hobbit and hence the Arkenstone, do we learn that ‘all lights that fell upon them . . . they took and reflected in marvellous hues to which their own inner fire gave a surpassing loveliness’ (HME V.227), implying that they had facets that refracted incoming light. Compare Bilbo’s first sight of the Arkenstone as ‘a small globe . . . that shone of its own light within . . . cut and fashioned by the dwarves . . . it caught and splintered all light that it received’ (page 579; emphasis mine). It thus seems that some of the most characteristic features of a Silmaril’s appearance familiar to us from the published Silmarillion20 – that it magnifies incoming light and that it splinters this light like a magnificent prism – derive not from the direct line of descent (BLT > ‘Sketch’ + alliterative poems > 1930 Quenta + earliest Annals > 1937 QS + later Annals > later QS + final Annals) but first appeared in the description of the Arkenstone in The Hobbit and from there were imported back into what had been the ‘main line’ of the legendarium. This does not prove, of course, that the Arkenstone is a Silmaril, but it does show that not only was the description of the Gem of Girion based upon the Silmarils but that it in turn influenced the way the Silmarils were described henceforth.

  If however the Arkenstone is indeed a Silmaril, the question arises: which one? Is there any way to reconcile the presence of a Silmaril within the fabled Hoard of Thror with what is said of the Jewels of Fëanor elsewhere in the legendarium? It is out of the question that it might be a ‘fourth Silmaril’, since all accounts from The Book of Lost Tales onwards are unanimous that Fëanor made only three and could never repeat his achievement, but might one of the Three have found its way hence? The answer, just as with the identification of the Elvenking with Thingol Greycloak, is both ‘yes’ and ‘no’.

  The fate of the Silmarils is not addressed in The Book of Lost Tales, aside from the one rescued by Beren and Tinúviel and later incorporated within the Nauglafring, which according to Tolkien’s outlines for the unwritten ‘Tale of Eärendel’ was lost in the sea when Elwing drowned.21 Tolkien himself left the fate of the other two an open question, jotting ‘What became of the Silmarils after the capture of Melko?’ in his notebook, and Christopher Tolkien observes: ‘. . . the question is itself a testimony to the relatively minor importance of the jewels of Fëanor’ at the time (BLT II.259; see also BLT 1.156). The matter was not addressed until Tolkien came to write the 1926 ‘Sketch of the Mythology’, where Elwing’s Silmaril is still lost at sea, while of the two recovered from Morgoth’s crown one is stolen by Fëanor’s son Maglor the minstrel who, finding that the holy jewel burns him, ‘casts himself into a pit’ – presumably with the Silmaril, since the next sentence states ‘One Silmaril is now in the sea [i.e., Elwing’s], and one in the earth [i.e., Maglor’s]’ (HME IV.39). The third is claimed by Maidros the maimed, Fëanor’s last surviving son, but the Valar deny his right to it because of the Fëanoreans’ many evil deeds and grant it instead to Eärendel, who with Maidros’s aid finds Elwing and transforms her from a seabird back into her own form again: ‘thus it was that the last Silmaril came into the air’ (HME IV.41). As Christopher Tolkien observes (HME IV.201–2), this was shuffled about in the 1930 Quenta: first (QI) Elwing’s is still lost in the sea ‘whence it shall not return until the End’ (HME IV.150), while Maidros and Maglor seize the other two (IV.158). Both are scorched by the holy jewels: Maglor throws himself and his jewel ‘into a yawning gap filled with fire . . . and the jewel vanished into the bosom of the Earth’ (IV.159), while Maidros throws his to the ground and kills himself (IV.158) and his Silmaril is reclaimed by Fionwë the Valar’s herald; this text breaks off just before the jewel was presumably given to Eärendel (IV.164). The revised version of the 1930 Quenta (QII) has Elwing surviving and bringing her Silmaril to Eärendel (IV.153), who sails the night sky with the Silmaril on his brow (IV.164), while Maidros ‘cast himself into a gaping chasm filled with fire’, taking his Silmaril ‘into the bosom of the Earth’, and Maglor ‘cast [his] . . . into the sea’ (IV.162); this idea was retained, in much the same words, in the Conclusion to the 1937 Quenta Silmarillion (HME V.331) except that here instead of simply stating that the three Silmarils came to rest in ‘sea and earth and air’ (HME IV.40 & 165),22 as had been the case in the three earlier versions (1926 ‘Sketch’, 1930 Quenta, and revised 1930 Quenta), they now ‘found their long homes: one in the airs of heaven, and one in the fires of the heart of the world, and one in the deep waters’ (HME V.331), words that would be carried over verbatim into the published Silmarillion four decades later (Silm.254).

  Thus while to us, thirty years after the posthumous publication of The Silmarillion, it seems inevitable that the three jewels would be lost beyond recovery – in fact, an addition to the revised 1930 Quenta is explicit on this point, stating that ‘the Silmarils . . . could not be again found, unless the world was broken and re-made anew’ (HME IV.163) – in 1931–2 when Tolkien created the ‘Gem of Girion’ or in late 1932 when he was writing about the Arkenstone in the Third Phase text of The Hobbit, this was most definitely not the case. Despite the sense of finality in the passage just quoted, Tolkien had in fact at that point changed his mind four times in the previous fifteen years about the holy jewels’ fate, all in a series of unpublished works that remained in flux and were each to be replaced by a new version of the story. There is no way any observer at that time could have told that this one point would henceforward remain relatively fixed within the Silmarillion texts; the one constant had been that the story ended with all three of the jewels remote and inaccessible. Just as the sword of Turgon King of Gondolin had somehow survived the fall of his city and found its way through the ages into that troll-lair and hence Bladorthin/Gandalf’s hands, it is thus more than possible that Tolkien was playing in The Hobbit with the idea of having one of Fëanor’s wondrous Jewels re-appear, no doubt the one that had been thrown into a fiery chasm and lost deep within the earth – which is, after all, exactly where the dwarves find the Arkenstone, buried at the roots of an extinct volcano. As with his borrowings regarding Tinwelint’s quarrel with the dwarves in ‘The Nauglafring’ for the chapter about the wood-elves and their king’s ‘old quarrel’ with the dwarves, Tolkien drew on his legendarium without committing himself: it was a one-way borrowing in which elements from the 1930 Quenta and Early Annals found their way into The Hobbit but that ‘unofficial’ usage did not in turn force changes in what Tolkien was still thinking of as the main line of the legendarium. By avoiding the use of the word silmaril and instead using the ingenious and agreeable synonym Arkenstone (Eorcanstán), Tolkien got to draw on his rich homebrew mythology, which by the early 1930s had developed a remarkable depth and sophistication, without worrying what the effect of his new story would be on that mythology (and hence could blithely include such statements as ‘indeed there could not be two such gems, even in so marvellous a hoard, even in all the world’). It was probably this idea of one-way borrowing to which Tolkien referred when, on occasion, he denied that The Hobbit was part of his mythology (e.g., Letters pp. 215 & 346). Not until the publication, and success, of The Hobbit called for a sequel did the new side-line of Middle-earth’s story displace the old legend of the war against Morgoth as the main story of t
he legendarium and the events of The Hobbit and The Lord of the Rings require the older stories to be rewritten and revised with the published chronicles in mind.

  (iii)

  A Note on Cram

  The first mention of cram appears in the First Typescript (typescript page 133; 1/1/63:7); the line about its being made by the Lake-men is added there in ink at some point before the corresponding passage in the Second Typescript (1/1/44:8) was created, and it was taken up into the page proofs and published book. Douglas Anderson notes (DAA.300) that Tolkien gave ‘cram’ an Elvish derivation in ‘The Etymologies’ (which seems to have been mostly written in late 1937 and early 1938, or at least five years after this part of the First Typescript):

  KRAB- press. N[oldorin] cramb, cram cake of compressed flour or meal (often containing honey and milk) used on long journey.

  —HME V.365.

  This is however almost certainly an afterthought on Tolkien’s part, like the entry there regarding ‘Esgaroth’ (see the commentary following Chapter XIII on page 562). Not only would it be extraordinary for the Noldorin (Sindarin) and English words to be so similar in both form and meaning – the elvish meaning being due to the ingredients being pressed together and the Old English ancestor (crammian) of the modern-day familiar word ‘cram’ meaning to squeeze in or stuff, itself in turn deriving from an Indo-European root (*grem-) meaning ‘to press or compress’ – but Tolkien explicitly states in a draft passage for The Lord of the Rings that

  Cram was, as you may remember, a word in the language of the men of Dale and the Long-lake . . . Bilbo Baggins brought back the recipe – he used cram after he got home on some of his long and mysterious walks. Gandalf also took to using it on his perpetual journeys . . . (HME VI.177; emphasis mine).