The History of the Hobbit Read online

  (ii)

  A Note on the Text

  Edith has gone to bed and the house is in darkness when [Tolkien] gets home. He builds up the fire in the study stove and fills his pipe. He ought, he knows, to do some more work on his lecture notes for the next morning, but he cannot resist taking from a drawer the half-finished manuscript of a story that he is writing to amuse himself and his children. It is probably, he suspects, a waste of time; certainly if he is going to devote any attention to this sort of thing it ought to be to The Silmarillion. But something draws him back night after night to this amusing little tale – at least it seems to amuse the boys. He sits down at the desk, fits a new relief nib to his dip pen (which he prefers to a fountain pen), unscrews the ink bottle, takes a sheet of old examination paper (which still has a candidate’s essay on the Battle of Maldon on the back of it), and begins to write: ‘When Bilbo opened his eyes, he wondered if he had; for it was just as dark as with them shut. No one was anywhere near him. Just imagine his fright! . . .’

  We will leave him now. He will be at his desk until half past one, or two o’clock, or perhaps even later, with only the scratching of his pen to disturb the silence, while around him Northmoor Road sleeps.

  —Humphrey Carpenter, Tolkien: A Biography, pp. 120–21.

  The preceding passage from the chapter ‘Oxford Life’ in Carpenter’s biography concludes his fictional recreation of a typical ‘day in the life’ of J. R. R. Tolkien. While entertaining, it is by no means accurate as an account of The Hobbit’s composition. For one thing, the text Carpenter quotes is not that of the Ms. (see p. 153) but the published book (cf. The Annotated Hobbit p. [115]). Nor is the manuscript of The Hobbit written on the back of student exams, with the exception of a single page;11 I suspect Carpenter has gone astray here by confusing the manuscript of The Lord of the Rings, parts of which were drafted on any scraps of paper its author could lay his hands on during the wartime paper shortage, including many from students’ exams, with that of The Hobbit, which contains very little extraneous material. Finally, the idea that the book was written by burning the midnight oil, faithfully added to night after night after a long day’s academic chores, has no evidence to support it and a good deal against it. For one thing, Tolkien’s letters are full of references that make it clear that almost all his creative writing was done not in term-time but during his too-brief vacations between academic semesters, and indeed his son Christopher confirms (private communication) that this was his father’s usual pattern of composition.

  The physical appearance of the manuscript also argues for periodic bursts of rapid writing rather than the nightly diligence Carpenter projects. As Carpenter himself notes elsewhere,

  The manuscript of The Hobbit suggests that the actual writing of the main part of the story was done over a comparatively short period of time: the ink, paper, and handwriting style are consistent, the pages are numbered consecutively, and there are almost no chapter divisions. It would also appear that Tolkien wrote the story fluently and with little hesitation, for there are comparatively few erasures or revisions.

  —Carpenter, pp. 177–8.

  In fact, as we shall see, there are a great many changes made to the rough draft in the process of writing, and many more afterwards. Parts of the manuscript show signs of having been written in great haste, while other sections are careful fair copy. Nor does Carpenter’s suggestion account for the several sharp breaks that occur in the Ms. where the handwriting, names of characters, and paper all change. Large sections are consistent in writing style and the paper used, only to have no less than three sudden and marked changes in writing paper and handwriting, the first and last of which almost certainly mark the long hiatuses Tolkien describes in his letter to The Observer. In short, the situation is far more complicated, and also much more interesting, than Carpenter indicates.

  The present text is organized around the major breaks in the Ms., which occur midway through the first chapter (between typescript page 12 & manuscript page 13), just after what is now the beginning of Chapter IX (between manuscript pages 118 & 119), and about a third of the way through what is now Chapter XV (following manuscript page 167). The very first stage of writing that grew out of the scribbled line ‘In a hole in the ground . . .’, which I call the First Phase, is now represented by six surviving pages of manuscript (an incomplete draft corresponding roughly to pages 25–32 of the first edition or pages 45–54 of The Annotated Hobbit) and by the twelve-page typescript that replaced this earliest draft before the missing pages were lost. These I refer to as ‘The Pryftan Fragment’ and ‘The Bladorthin Typescript’, respectively, after the names of the dragon and wizard used in each.

  The Second Phase begins with manuscript page 13, which picks up exactly where page 12 of the Bladorthin Typescript had left off, completing its final sentence. Written on good-quality ‘foolscap’ paper, this comprises the main stage of Tolkien’s work on the book. Tolkien once admitted that ‘They say it is the first step that costs the effort. I do not find it so. I am sure I could write unlimited “first chapters”. I have indeed written many’ (JRRT to Charles Furth, 17th February 1938; Letters p. 29). The Second Phase marks the stage at which an intriguing opening developed into a nearly complete story. Given its length (over one hundred and fifty manuscript pages), it’s not surprising that this phase was interrupted several times, these points being marked by Tolkien’s pausing to draw up outlines or sketch out ‘plot notes’ of upcoming sections. These various interruptions are described in detail in the main text that follows; for now, we need only note the major break that occurred in the middle of the Second Phase, just at the point when Bilbo and the twelve remaining dwarves are ambushed and captured by the wood-elves, in what is now early in Chapter IX. Here Tolkien clearly paused for some months, because when he resumed he changed to a completely different type of writing paper, these being the unlined backs of lined sheets of writing paper probably extracted from the unused portion of students’ exam booklets. Thus, the Second Phase falls into two distinct parts: manuscript pages 13–118 on the good-quality ‘foolscap’ paper Tolkien favored (it also recurs as his paper of choice when writing The Lord of the Rings) and manuscript pages 119–67 on slightly poorer quality paper.

  The Third Phase, which saw the completion of the initial draft, can be divided into several stages like the phase that preceded it. First Tolkien returned to the beginning of the story and created the First Typescript, covering what is now Chapters I through XII and part of Chapter XIV. He then made a handwritten fair copy manuscript of Chapter XIII and inserted this into the typescript. Finally, and most importantly, he completed the story by the addition of another forty-five pages of very hastily written manuscript, again on the same good-quality paper as the bulk of the Second Phase. This final section, which starts in Chapter XIV (again completing a sentence left unfinished on the last page of the typescript as it then existed) and covers Chapters XV through the end of the book (i.e., Chapter XIX), was almost certainly written in December 1932 and January 1933. The resulting composite typescript/fair copy/manuscript, sometimes referred to by Tolkien as the ‘home manuscript’ (cf. JRRT to Susan Dagnall, letter of 4th January 1937; Letters p. 14), was then circulated among Tolkien’s friends over the next several years. Sometime in the summer of 193612 Tolkien was asked to submit The Hobbit to Allen & Unwin, so he at this time extended the First Typescript to include Chapter XIII, the rest of Chapter XIV, and Chapters XV through XIX to the end of the book.

  In addition to the First Typescript, there is also another copy of the completed story. For many years the processors at Marquette and also scholars consulting the original manuscripts were puzzled by the presence of a second typescript that in some ways seemed earlier than what I have called the First Typescript but in others was demonstrably later.13 Taum Santoski solved this problem by demonstrating that this text, which I call the Second Typescript, was made after the First Typescript and derives from it, but that it was rejected by To
lkien who then made the final layer of pre-submission revisions on the First Typescript instead, which thus became the ‘Typescript for Printers’ (i.e., the text from which the printers set the book). A clue within Carpenter’s biography makes it possible for us to reconstruct the story behind this second typescript’s creation, establish its relationship with the first typescript, and see the reason why it was ultimately rejected in favor of its predecessor.

  Since Tolkien had, characteristically, made many revisions to his typescript while he had been re-reading the entire story and preparing it for submission to the publisher, the desirability of a cleaner typescript would have become obvious, especially given Tolkien’s difficult handwriting. Tolkien himself had no time to undertake this onerous task, and so he set his son Michael to create a second typescript that would incorporate all the changes (mostly handwritten in black ink) on the original. According to Carpenter, Michael (then sixteen), had badly injured his right hand on broken glass and so did all his typing for the book one-handed (Carpenter, p. 180).14 Although Carpenter does not distinguish between the two typescripts, it is clear that the Second Typescript was not made by Tolkien himself but by an inexpert typist who often skipped or misread words, occasionally dropped lines, sometimes had difficulty in reading Tolkien’s handwriting, and generally produced a poor-quality text. As a daunting task undertaken by a dutiful son and apparently completed within a very short space of time, the Second Typescript speaks well of Michael’s filial piety, but as an accurate text of The Hobbit it is sadly lacking. Even when carefully corrected by Tolkien, it is still inferior to the by now rather battered First Typescript, which therefore became the copy Tolkien ultimately sent off to Allen & Unwin (on 3rd October 1936 according to Carpenter; see Letters p.14) and which thence went to the printers, Unwin Brothers.

  In the end, however, it is fortunate that the Second Typescript exists, because it enables us to date some of the changes Tolkien made to the work. Just as he revised the manuscript in two distinct stages (in ink at or soon after the time of composition, and in pencil later when preparing it to be superseded by the typescript), so too he revised the First Typescript in layers, and it is often not self-evident whether a given reading dates from the time when he was completing the tale (that is, corrections made in the course of typing or not long after) or several years later when he was preparing the text for submission to the publisher. However, comparison with the corresponding section of the Second Typescript often resolves the question: if a revision made in ink on the First Typescript is incorporated into the Second Typescript as first typed, then it belongs to the earlier layer of changes; if on the other hand it is written onto both typescripts then it is generally part of the later set of revisions. The issue is confused by two factors. First, Tolkien inked in corrections to set right Michael’s accidental omissions and errors. This led early processors at Marquette, seeing that these sections appeared as ink additions to one typescript (Michael’s) but as first typed in the other (Tolkien’s), to mistake these corrections for new additions to the text taken up in the other typescript and thus assume that Michael’s Typescript predated the ‘Typescript for Printer’. Second, even after he had rejected the Second Typescript as the current text, Tolkien continued to scrupulously enter corrections he made to the First Typescript onto the other rejected typescript as well. Thus, very late changes appear added to both. In effect, the Second Typescript became Tolkien’s safe copy, from which he could reconstruct the work if the final ‘Typescript for Printer’ were to become lost in the mail, be destroyed by an accident at the printer, or suffer some other misfortune.

  For the most part, while including all revisions to the manuscript page itself I have not recorded changes between the manuscript and the typescript(s), since these invariably move the story closer to its familiar published form, although I have, on occasion, noted just when some significant line or event entered into the tale between draft and publication (e.g., a rider, first typescript, second typescript, or page proofs). Similarly, I have only rarely noted changes made between the typescripts and page proofs, or on the page proofs themselves; anyone examining the three sets of page proofs15 now at Marquette will be deeply impressed by Tolkien’s close attention to detail, his ability to spot potential contradictions, and his gift (no doubt developed through years of practice with academic publications) of replacing a problematic passage with new text that takes up exactly the same amount of space, but to address every change made at every stage would call for a variorum edition – a worthy goal, but one beyond the scope of this book.

  With the material I have labelled the Fourth Phase, we enter into the post-publication history of The Hobbit. While the book was so successful that a sequel was called for almost at once, at several times in later years Tolkien returned to the original story and re-wrote parts of it to better suit his evolving conception of Middle-earth and the role which the story of Bilbo’s adventure played in it. The first and most important of these re-visionings is what I here call the Fourth Phase: his recasting of the encounter with Gollum in Chapter V to bring that character’s actions into line with what he had written about him in The Lord of the Rings (then unpublished and indeed still unfinished). This tour-de-force, perhaps the most famous scene Tolkien ever wrote, was drafted in 1944, sent to Allen & Unwin in 1947, and published as the ‘second edition’ of The Hobbit in 1951.

  Another significant piece of writing relating to The Hobbit is ‘The Quest of Erebor’, originally written as part of Appendix A of The Lord of the Rings in the early 1950s but in the event omitted from that work for reasons of space. This presents Bilbo’s story, particularly the opening chapter of the book, from Gandalf’s point of view and sets it firmly within the larger context of the war against Sauron. While a fascinating and relevant piece, I have not included it here because it is readily available elsewhere: different drafts or excerpts of it have been published in Unfinished Tales (pp. 321–36), The Annotated Hobbit (revised edition, pp. [367]–77), The War of the Ring (HME VIII, pp. 357–8), and The Peoples of Middle-earth (HME XII, pp. 281ff).

  This brings us to our final text, the 1960 Hobbit, representing the Fifth Phase of Tolkien’s work on the book. In this previously unpublished material, Tolkien returned to the concerns of ‘The Quest of Erebor’ and set out to re-write the entire Hobbit in the style of The Lord of the Rings. Although he wisely abandoned this new draft at the start of Chapter III, this fascinating glimpse into a radically different approach to the story helps us appreciate the story as it stands all the more, besides providing some interesting and hitherto unknown details about Bilbo’s itinerary. A few years later, when Tolkien was asked by his American publisher to revise both The Lord of the Rings and The Hobbit in order to assert the American copyright against the unauthorized edition of the former that had just been issued by Ace Books, he used a few of the changes he had contemplated in the 1960 Hobbit but for the most part refrained from any but minor changes to the established text. It might be argued that these constitute a ‘Sixth Phase’ of work on the book, but if so it would be the only one that was imposed on Tolkien from without rather than arose from within. Since the 1966 ‘third edition’ changes are both minor and very well documented by Douglas Anderson in The Annotated Hobbit I have not listed them all here and instead refer readers either to his excellent book or to Hammond’s definitive Descriptive Bibliography, pages 28–39.

  More information on each of these stages is contained in the head-note to each section of the text.

  (iii)

  The Plan of This Edition

  My presentation of the text is intended to distinguish as much as possible what Tolkien wrote from my own commentary and notes upon it. The format for each chapter is thus a brief headnote by me, followed by Tolkien’s text, often followed by a brief tailnote. Next come Text Notes (TN) discussing difficult readings, highlighting various changes or sequences of changes, and the like. After this comes my Commentary in the form of mini-essays on topics arising ou
t of that chapter, followed by Notes upon the commentary. Wherever possible, I have kept my own commentary and Tolkien’s texts typographically distinct.

  It must be stressed that there are no chapter divisions in the original manuscript, which flows as one continuous text with no more than the occasional skipped line to mark a change in scene or passage of time. My decision after much internal debate to follow Marquette’s lead, and also Christopher Tolkien’s practice at various points in The History of Middle-earth – that is, to insert chapter breaks where Tolkien himself later chose to make chapter divisions – comes as a result of my conviction that doing so greatly improves ease of reference, making it possible for those familiar with the published book to find any corresponding manuscript passage with relative ease. Nevertheless, these chapter breaks are an editorial contrivance and some readers may wish to ignore them, moving directly from the end of one ‘Chapter’ to the continuation of the text at the beginning of the next.

  Formatting

  It had been my original intent to record every brushstroke, cancellation, and addition to each manuscript page, so that in lieu of a facsimile reproduction this book could serve as a means by which scholars of Tolkien’s work could follow every step, letter by letter and line by line, of the process by which Tolkien created his work. However, over the long course of working with the manuscript for this edition I have been persuaded that such mechanical fidelity would produce only confusion and slowly come to the conclusion that an edition of a manuscript should be, well, edited. Accordingly, I have silently omitted minor changes (such as Tolkien’s own correction of miswritten or misspelled words) and sometimes slightly re-arranged material for clarity. I have also provided punctuation where necessary (mainly quotation marks and periods at the end of sentences), although I have kept this to a minimum in order to preserve the lightly-punctuated flow of the original. Changes in the manuscript by Tolkien himself are indicated by brackets; brackets have also been used in a few instances to mark missing words necessary for the sense that have been provided editorially. An arrow coming at the end of the bracketed passage [thus >