The Letters of J. R. R. Tolkien

The volume is a selection, not a complete correspondence. Humphrey Carpenter, who had already written the authorised biography, chose and annotated the letters; Christopher Tolkien assisted, and the result appeared in 1981, eight years after his father's death. The 354 items are numbered consecutively and arranged by date, running from Tolkien's youth to the last weeks of his life, so that the book reads as a chronological spine rather than a thematic arrangement. Many letters are given in extract only, with the editor's square-bracketed prose supplying the addressee, the occasion, and whatever context the fragment needs. That editorial layer is worth noticing: the reader is never quite alone with the writer, and the shape of the whole — what was included, what was cut, what was passed over — belongs to the editors rather than to Tolkien.

The recipients are various enough that the register shifts constantly. Some letters are to publishers and their readers, written with a professional eye on schedules, costs, and the awkward business of a book that would not stay the size it was meant to be; some are to his sons, especially Christopher during the war years, and these are domestic and unguarded; some are to strangers who had written in with questions about Elvish, about the Ring, about whether the story meant anything beyond itself. The longest and most quoted items belong to this last group, above all the account of his mythology drafted for Milton Waldman in an attempt to persuade a publisher to take the whole of it (letter 131). Elsewhere the tone is quite different — brief, tired, apologetic about the backlog of unanswered post — and the collection does not smooth this out.

Within the legendarium the book stands one step outside. It is not a narrative source and does not report events in Middle-earth on its own authority; what it documents is the author's thinking about his invented world — his intentions, his etymologies, his dissatisfactions, his objections to readings he thought wrong. This makes it indispensable for interpretation and unreliable as chronicle, and it is cited here accordingly: by letter number, never by page, since paginations diverge between editions and printings while the numbering does not. The letters sit at the K2 layer for the same reason the posthumous narratives do. The words are Tolkien's; the book is not.

The material collected in 1981 was not written as a book and never had a period of composition in the ordinary sense. The earliest item is dated October 1914, sent to Edith Bratt while Tolkien was still an Oxford undergraduate; the latest belongs to the last days of August 1973, written from Bournemouth shortly before the illness that killed him on 2 September. Between those points lie close to six decades of occasional writing done in the margins of other work — at night, in vacations, between tutorials and examination marking — and the physical character of the correspondence changes with the circumstances: hurried notes in a difficult hand, typed sheets when a typewriter was available, the thin wartime forms he used for letters sent to Christopher on air-force service abroad in 1944. After the success of The Lord of the Rings the volume of incoming post rose sharply, and a large part of what survives from the last two decades consists of replies to readers who had written with questions.

The book belongs to a specific stretch of posthumous editorial activity. Humphrey Carpenter had been given access to the family papers for the authorised biography published in 1977, the same year in which Christopher Tolkien brought out The Silmarillion; Unfinished Tales followed in 1980, and the letters in 1981. The editing was therefore done by someone already familiar with the archive and with the shape of the life, and the annotation reflects that: dates, addressees, and the surrounding events are supplied economically because the biographical groundwork had been laid elsewhere. First publication was by George Allen & Unwin in London and by Houghton Mifflin in Boston, the two houses that had handled Tolkien's books during his lifetime.

What could be printed depended on what had survived, and survival was uneven. Letters to publishers were preserved in office files and are consequently well represented from 1937 onwards; letters to family and friends came back only where the recipients had kept them; letters to strangers exist in the book chiefly because Tolkien had made carbons or kept his own drafts. That last habit accounts for some of the most substantial items in the collection. Several of the long expository letters were worked at over more than one sitting and in the end were never posted, so that the fullest surviving statement of a point is sometimes a draft the intended reader never saw — the extended discussion of Frodo at the Cracks of Doom is of this kind (letter 246). The retrospective accounts of how the books were written, such as the summary sent to W. H. Auden in 1955, likewise come from the years when Tolkien had begun to be asked such questions in quantity (letter 163).

The 1981 text was reprinted in hardback and paperback over the following decades without change of substance, and a German translation by Wolfgang Krege appeared in 1991 under the title J. R. R. Tolkien: Briefe. The one significant revision came in 2023, when HarperCollins issued a revised and expanded edition adding well over a hundred letters that had come to light or become available since Carpenter's work, together with corrections to the annotation. The expansion was arranged so as not to disturb the original sequence: the numbers of the 1981 selection were left as they stood and the new items inserted beside them under suffixed numbers. This is why citation by letter number remains the only reliable practice — it survives the 2023 revision as it survives the differences in pagination between the British and American settings — and why references here, including the long account of the mythology written for Milton Waldman (letter 131), are given in that form alone.

The book has almost no internal architecture beyond the numbered sequence itself. There are no parts, no books, no thematic divisions and no running headings; each item begins with its number, the addressee and a date, and the next follows immediately. Nothing here corresponds to the appendices, maps and genealogical tables that accompany the narrative volumes — the letters document a working life rather than a place, and no cartography would serve them. Such apparatus as exists sits at the two ends of the volume: an editorial introduction at the front, setting out how the selection was made, and at the back a body of notes keyed to individual letters together with a full index of persons, works and names. In a book without chapters the index carries nearly all of the navigational weight, and a reader looking for a topic rather than a date will normally start there.

The sequence is very unevenly filled, and knowing where the weight lies is the quickest orientation available. The first two dozen numbers cover almost a quarter of a century and are correspondingly thin; the sequence thickens abruptly in 1937 with the publication of The Hobbit and the publisher's request for a successor, and the letters of the later thirties are dominated by that negotiation and by the slow beginning of the new story. Among them stands the draft prepared for a German publisher who had asked whether Tolkien was of Aryan descent, which he answered with a philologist's precision and considerable coldness (letter 30). The war years belong largely to the letters sent to Christopher on service abroad; the early fifties to the protracted business of finding a publisher for both the new book and the older mythology, which produced the long summary written for Milton Waldman (letter 131). From 1954 onward the character of the sequence changes again as replies to readers come to predominate, and the last hundred or so numbers are mostly of that kind.

Because the arrangement is chronological, the substantial expository letters lie scattered rather than grouped, and the reader who wants them must know their numbers. The fullest statements about religion in The Lord of the Rings were written to Robert Murray, a Jesuit and a family friend, who had read the book in proof (letter 142). The most detailed retrospective account of how the work came to be written was sent to W. H. Auden (letter 163). Tolkien's objections to a proposed film treatment, itemised scene by scene, form one of the longest single items in the book and remain the clearest evidence of what he thought could and could not be altered in his story (letter 210). Questions of language, etymology and nomenclature recur throughout without ever being gathered into one place, as do his remarks on allegory and applicability, and on his own temperament and habits (letter 213). The notes at the back supply identifications where the extract alone would be opaque, but they do not group the letters by subject; the volume expects to be consulted by number.

The book presupposes everything else. It was assembled for readers who already knew The Lord of the Rings and, from 1977, The Silmarillion, and very little of it is intelligible without them: the letters argue, defend, correct and explain, and take the stories themselves as given. Read on its own the volume would be a thicket of half-identified names and unexplained objections. What it supplies that no narrative source can is the reasoning behind the invention — why a thing was made, what it was meant to carry, and what Tolkien thought had gone wrong when a reader understood it otherwise. The account drafted for Milton Waldman is the clearest instance: a connected description of the whole mythology in the author's own voice and from a single moment, set down at a time when almost none of that material had been printed and no reader outside his circle could have known what shape it had (letter 131).

Its authority is of a particular and limited kind. On what happened in Middle-earth the narrative works decide, and a letter that appears to contradict them is evidence of what Tolkien thought or intended at the moment of writing, not a revision of the story; on intention, etymology and interpretation the order reverses, and the letters are the primary witness with nothing else standing in their place. They are also dated documents, and the collection preserves rather than reconciles their disagreements: an answer given in the mid-fifties to a reader pressing him on the theology of his invented world need not agree with what he wrote a decade later, and several of the most careful discussions are drafts in which he can be seen changing his mind between one page and the next (letter 153). The long analysis of Frodo at the Cracks of Doom is of that argumentative sort — a defence of the ending against a reading he considered unjust, not a fresh statement of what occurred (letter 246).

The letters are not the only place Tolkien commented on his own work, and the parallel matters for weight. He did it in print himself, in the foreword written for the second edition of The Lord of the Rings, where he distinguished the allegory he disliked from the applicability he allowed the reader, and in the appendix on languages and translation, where the conceit governing names, calendars and scripts is set out at length. Those statements were published in his lifetime and form part of the book; they carry the K1 layer. The letters stand behind them, fuller and more combative, and frequently earlier in date, so that a point compressed into a paragraph of the appendices may be argued over several pages of correspondence. Where the two cover the same ground, the printed statement is the canonical form and the letter the explanation of how it was reached.

One further point of position is worth recording. Until the series of draft volumes began to appear in the eighties, this book was the principal published evidence for how the works had been composed, and for two decades the retrospective narratives it contains — chiefly the account sent to W. H. Auden — did duty as the standard history of the writing (letter 163). That function has since been taken over by the drafts themselves, which show the process rather than describe it. What has not been superseded is the interpretative material, and it is for that, rather than for its documentary value, that the volume is most often consulted here.

The most visible consequence of the book's appearance was a change in how Tolkien is quoted. Within a few years of 1981 the volume had become a standard point of reference in writing about him, and the convention of citing it by number rather than by page settled quickly, for the practical reason that the numbering is the only element the various printings and settings hold in common. Later reference works built on it directly: the chronology and the alphabetical guide compiled by Wayne G. Hammond and Christina Scull draw on the correspondence throughout, and the annotated companions to the narrative works use it to date decisions and to identify the point at which a given name or explanation first appears. Its documentary weight is therefore secondary as well as primary — it is consulted less often as a book to be read through than as a body of evidence to be checked at a particular number.

Individual items have circulated well beyond the volume that contains them. The draft answering a German publisher's enquiry about ancestry is quoted regularly outside Tolkien studies altogether, whenever his politics are at issue (letter 30). An extract from the long account of the mythology written for Milton Waldman has been printed as supplementary matter with later editions of The Silmarillion, so that many readers meet it without ever opening the letters (letter 131). The statements about religion in The Lord of the Rings and the defence of the ending at the Cracks of Doom recur in theological and critical discussion to the point of being the fixed reference for those questions.

In print the book has remained continuously available, in hardback and paperback, and it exists in German through Wolfgang Krege's translation. The revision of 2023 complicates the record only slightly: because the original numbers were kept and the added material given suffixed ones, citations made before and after that date remain mutually legible, and only a reference to newly added material requires the edition to be named as well.