Alex Lewis *

J. R. R. Tolkien wrote the paper On Fairy-stories for an Andrew Lang lecture given on 8 March 1939 at St Andrews University [1], although Carpenter suggests that the basis for the talk was meant to be a lecture given to undergraduates at Worcester College a year before (ref. 1, p. 190). Various writers [2,3] have described this talk as extremely significant because of the date of its composition — coming right at the start of the process of the writing of The Lord of the Rings. Some even go so far as to suggest that it set the pattern for how The Lord of the Rings would be written.

It has also been suggested that this paper contained Tolkien’s settled and final thoughts on various broader artistic matters, and have used it as a stick with which to beat potential illustrators of his works, as well as determining Tolkien’s low opinion of drama as an art form. Both suggestions require caution. Indeed, Flieger and Anderson [3] warn against making exactly that assumption, and that Tolkien revisited these questions when writing Smith of Wootton Major later in his life, and came to somewhat different conclusions. So, a kind of unsubstantiated rumour has built up around On Fairy-stories, and yet as far as I am aware, there has been no detailed analysis of the talk in terms of its changing content and context. Shippey [4] suggests that it was Tolkien’s ‘least successful if most discussed piece of argumentative prose’. One point of this paper is to address that view.

First: why was Tolkien asked to give the Andrew Lang lecture at all? He was, as an academic, little known. The timing of the lecture makes it very tempting to suggest that it was because Tolkien had recently had The Hobbit published (on 21 September 1937) — and that someone mistook that book for a ‘traditional fairy story’. But that would seem to be a rather simplistic explanation, which in fact turns out to be mistaken.

Rachel Hart (ref. 2, Ch. 1, pp. 1–2) explains that Tolkien was third choice as presenter of this paid lecture — £30, a considerable sum in 1939, when a week’s wages for a working man would have been anything from 10 shillings to £3, depending on his level of skills5. Indeed, working it out in those terms, it is astonishing to think that £30 would have been around half the average man’s annual pay, or £12,000 in today’s terms.

The first two people approached were Gilbert Murray, a very well-known Oxford academic; and Hugh Macmillan, who had been Lord Provost of Scotland and briefly Minister for Information in the opening months of the Second World War. Neither man could fulfil the engagement and so Tolkien was approached. He had meant to give a talk on fairy stories at Worcester College the year before, but had instead read out a version of Farmer Giles of Ham (ref. 1, p. 191), and so the lecture was something he could write and give. As it turned out, Murray gave the Andrew Lang lecture the following year, and Macmillan the year after that.

It is possible that Tolkien was suggested by his friend and former colleague from Pembroke College, R. G. Collingwood, who had since moved on to the Waynflete Chair at Magdalen College. One of Collingwood’s former pupils at Pembroke, T. M. Knox, was on the board charged with appointing Andrew Lang lecturers. Given Collingwood’s own interests [6], Collingwood himself might have been in the frame for the lectureship. But Collingwood was ill: he retired in 1941 and died a few years later.

Tolkien and Collingwood had most certainly collaborated to some degree on academic work [7]. Flieger and Anderson believe that Knox might have approached Tolkien directly (ref. 3, p. 123), but given the closeness of Knox and Collingwood, and the latter’s high profile in academia, it seems far more likely that it would have been Collingwood that Knox wanted initially to deliver the lecture and effectively had to settle for Tolkien in the end.

Multiple versions

The essay we now know as On Fairy-stories is best thought of as a development of four initially quite different papers [3]. The first two come from before the Second World War. The first, of which no certain record now exists, was partially written for Worcester College in 1938; the second (version 2), developed and given at St Andrews in 1939. The third version was published in Essays Presented To Charles Williams (Oxford, 1947), and which I shall call version 3 for the purposes of this paper. Finally, there is the greatly expanded version published in Tree and Leaf in 1966 (ref. 8), alongside the story Leaf By Niggle. I shall call this final iteration version 4.

It is important to emphasize that these are four quite different papers putting forward different views in many key areas. Also they cover varying if overlapping areas in and of themselves. In short, they reveal Tolkien’s own changing views over three decades.

One of the interesting things that changes dramatically is Tolkien’s attitude to Andrew Lang himself. It moves from a broadly supportive position in the first two versions to an increasingly negative one in the third, but mellowing by the fourth in 1966.

To turn up and give an Andrew Lang lecture at his Alma Mater and castigate the man would seem the height of folly, and yet that is what Tolkien appeared to be doing in print. However, versions 3 and 4 as published each run to around 60 pages, far too long to have been delivered in a one-hour lecture. So, what was missed out?

Manuscript A — one of the manuscripts discussed, for the first time, by Flieger and Anderson3 — runs to some 19 pages, including deletions and alterations. It is, in length, the likeliest to have been the basis for the Lang lecture itself. It is not the entire paper — the start and end are missing — but most of it is there. It is concerned with defining fairy stories, and is highly positive about the work and studies of Andrew Lang. It also matches most closely contemporary newspaper reports of the lecture (ref. 3, pp. 161–169). Here are some examples:

For me the standard, the unrivalled [books of fairy stories] are the twelve books of twelve colours by Andrew Lang and his wife.

(ref. 3, p. 176)

Origins and study of fairy stories:

In this question of which Lang was deeply interested and wrote brilliantly and originally. And others have of course followed.

(ref. 3, p. 179)

And again:

More interesting if origins are discussed is the question of relation of what Andrew Lang called the higher and lower mythologies (and of both to religion strictly so-called). The biographer of Andrew Lang held that he had ‘proved that folk lore was not the debris of a higher or literary mythology but the foundation on which that mythology rests.

(ref. 3, pp. 181–182)

To be sure, Tolkien does say that this is an inversion of the truth, but is criticising Max Muller the famous nineteenthcentury German philologist at that point and not Lang as he is in later versions.

Andrew Lang’s famous collections were of course a by-product of adult research into mythology and folk-lore, specially drawn off and adapted for ‘children.

(ref. 3, p. 187)

‘The adaptation of the Story of Sigurd (done by Andrew Lang himself from Morris’s translation of the Volsunga Saga) was my favourite without rival.’

(ref. 3, p. 188)

This is likely to be the paper from which Tolkien extracted the material for the Lang lecture.

Manuscript B is expanded, to some 34 pages if one discounts crossed out sections and reworkings (ref. 3, pp. 206– 251). Here the praise for Andrew Lang is more qualified, but still present, and a new element on the nature of magic is introduced. For instance, the above example where Max Muller is being criticised becomes:

Among the many interesting questions which an enquiry into origins raises is one we have already just glimpsed; the relation of what Andrew Lang called the higher and lower mythologies: which would now probably be called myth and folktale. The once dominant view (which he especially opposed) was that which derived all from nature-myths… That would seem to be the truth nearly upside down …

(ref. 3, p. 223)

So I will not say children have changed since Andrew Lang’s time. I will say that I wonder if they were ever like that.

(ref. 3, p. 234).

I believe that Manuscript B may be the workings that Tolkien began as something to send to St Andrews for inclusion in their publications — they were planning to publish all the Andrew Lang lectures given in one volume, and asked Tolkien for his lecture to include in there. When it appeared it did so without Tolkien’s contribution for he never sent any version for their inclusion in that volume of papers. He did, however, send a copy of Tree and Leaf to St Andrews in due course, with his apologies.

Then we have the 1947 Essays version [10]. There are some differences between the 1947 and the 1966 versions of the paper — primarily version 3 starts with the introduction of an Englishman in Scotland and adds or removes various lines here and there throughout the text — page 33 has an extra line: “Of this seriousness the medieval Sir Gawain and the Green Knight is an admirable example.” And this is removed in Version 4. Also removed on page 59 was: “All children’s books are on a strict judgement poor books. Books written entirely for children are poor even as children’s books.” This is a rather bald statement that would be difficult to defend. On page 73 a sentence is removed from Essays: “Andrew Lang is, I fear, an example of this.” In the 1966 version. It adds merely “As Lang said.”

The discussion of art is somewhat complex. Version 4 has: “We do not, or need not, despair of painting because all lines must either be straight or curved. The combinations may not be infinite (for we are not) but they are innumerable.” (p. 82) whereas version 3 is more extended from page 73 paragraph 2 through to page 74.

Version 3 has an extended note on page 78 which is cut short on version 4 (p. 72).

And finally the second half of Note G, which appears in version 3 (p. 88), is entirely missing from version 4 — and is nine and a half lines long. It does deal with matters of eugenics, and perhaps Tolkien felt that to be an unwise subject to bring up in 1966, whereas in 1947 it might still have been acceptable.

Finally we have Tree and Leaf [8], with its stinging attack on Andrew Lang as well as on Max Muller. Tolkien damns Lang with faint praise, and I shall give some examples. My own comments on Tolkien’s comments are given in square brackets after each quote.

Drayton’s Nymphidia is one ancestor of that long line of flowerfairies and fluttering sprites with antennae that I so disliked as a child. Andrew Lang had similar feelings. In the preface to the Lilac Fairy Book he refers to the tales of tiresome contemporary authors… (p. 30) [see below with reference to Voyage to Lilliput] The number of collections of fairy stories is now very great. In English none probably rival either the popularity, or the inclusiveness, or the general merits of the 12 books of 12 colours which we owe to Andrew Lang and his wife …. Most of its contents pass the test, more or less clearly. … but I note in passing that of thestories in this Blue Fairy Book none are primarily about fairies, few refer to them. (p. 33) [So in effect the Blue Fairy Book is bogus — it does not contain what it says in its title]

But what is to be said of the appearance in the Blue Fairy Book of A Voyage to Lilliput? I will say this: it is not a fairy story … I fear that it was included merely because Lilliputians are small … (p. 34) [Note that before this on page 30 he says that Lang disliked the smallness and triteness — so how come this story is included? It would seem Tolkien is accusing Lang of a lapse of taste or even academic decision]

Now The Monkey’s Heart is also plainly only a beast-fable. I suspect that its inclusion in a ‘Fairy book’ is due not primarily to its entertaining quality, but precisely to the monkey’s heart supposed to have been left behind in a bag. That was significant to Lang …. ( p. 37) [but I would imagine Lang and his wife included it for the former reason]

Andrew Lang’s Fairy Books are not perhaps lumber rooms. They are more like stalls at a rummage sale … His collections are largely a by product of his adult study of mythology and folk-lore; but they were made into and presented as books for children. Some of the reasons that Lang gave are worth considering …. They represent the young age of man true to his early loves and have his unblunted edge of belief, a fresh appetite for marvels … It seems clear that Lang uses belief in its ordinary sense … if so then I fear that Lang’s words, stripped of sentiment, can only imply that the teller of marvellous tales to children … does trade on their credulity … Children are capable of course of ‘willing suspension of disbelief ’ … but if they really liked [the tale] for itself they would not have to suspend disbelief: they would believe — in this sense.

Now if Lang had meant anything like this there might have been some truth in his words … [meaning there is none?] … And as for children of the present day, Lang’s description does not fit my own memories, or my experience of children. Lang may have been mistaken about the children he knew, but if he was not, then at any rate children differ considerably … and such generalizations which treat them as a class … are delusory (from pp. 51–54).

It is true that the age of childhood sentiment has produced some delightful books … but it has also produced a dreadful undergrowth of stories written or adapted to what was or is conceived to be the measure of children’s minds and needs … the imitations are often merely silly … or patronising or (deadliest of all) covertly sniggering, with an eye on the other grown-ups present. I will not accuse Andrew Lang of sniggering, but certainly he smiled to himself, and certainly too often he had an eye on the faces of other clever people over the heads of his child-audience … (p. 56)

I do not deny there is a truth in Andrew Lang’s words (sentimental though they may sound) … ‘he who would enter into the kingdom of faerie should have the heart of a little child.’… ‘For children are innocent and love justice; while most of us are wicked and naturally prefer mercy.’ Andrew Lang was confused on this point. Let us not divide the human race into Eloi and Morlocks: pretty children — ‘elves’ as the 18th century often idiotically called them … (p. 57)

[Talking of eucatastrophic events in fairy stories] Even modern fairy stories can produce this effect sometimes … It happens even in Andrew Lang’s own fairy story Prince Prigio, unsatisfactory in many ways as that is. Note 1 — this is characteristic of Lang’s wavering balance. [again, Tolkien accuses Lang of a lack of academic decisiveness as he had before]

In context

Thankfully because of the hard work and scholarship of Flieger and Anderson, we are now able to put the whole lecture of On Fairy-stories into context because we can work out when certain ideas came into the essay and others were removed or changed. Few researchers have pointed out that Tolkien was indeed hostile to Lang’s works in On Fairy-stories. Ruth Berman (ref. 11, p. 127) comes close, though she merely says: “Tolkien disliked much in Lang’s work, and was by no means a follower of Lang, especially in 1939, when he was trying to write The Lord of the Rings as a story for adults, and so particularly resented the assumption that fairy-tales were necessarily children’s literature.”

She then goes on merely to compare motifs in Lang’s Green Fairy Book with similar ones in Tolkien’s writing, which do not seem to prove anything much at all. Berman is wrong in several points — notably that Tolkien was antagonistic to Lang’s work in 1939 — see Manuscript A and B as discussed above — and that in 1939 he was writing The Lord of the Rings for adults — he was at that point merely writing a sequel to The Hobbit, which was meant as a children’s story, and was being drawn towards the darker material in his Silmarillion and other ‘serious’ writings [12].

But that is not the end of the matter — as Flieger points out, Tolkien’s views were changing still as he grew older, and the introduction to Smith of Wootton Major [13] introduced concepts markedly different to those in Tree and Leaf. Whereas Tolkien says that Elves are not involved in human affairs and not interested in them in versions 3 and 4 of On Fairy-stories, in Smith he says that Elves and Men share the world and may even have their destinies intertwined in some way.

So, why was the whole situation changing? Very likely it was due to the interactions between Tolkien and his fellow Inklings. Glyer14 shows how much the members of this loose writers’ group influenced one another and in what ways they did so. She points out that influence is not a simple thing and has many components including both positive and negative influence (resonators and antagonists). She comes out in a mid position between the early writers who claimed that the Inklings were some kind of ‘artistic movement’ and consciously integrated their works to some Christian end, and the opposite position which Humphrey Carpenter espoused that Tolkien was not influenced by anyone.

Glyer quotes correspondence from various sources (ref. 14, pp. 5, 34, 58 — the wager of a space and a time travel story between CSL and JRRT; ref. 14, pp. 73, 84, 88, 94, 116–119 — influence on The Lord of the Rings as it was being written). It seems pretty clear that this discussion group, debating society and writers circle that was the Inklings as a loose group of people did affect one another both in what they wrote and in their held views. Debates were ferocious and ideas had to be defended against others in the group. Under these circumstances, it is not a surprise that Tolkien’s views on Lang as one example would change markedly with time.

Alex Lewis is a gentleman and a scholar.

See page 7 for David Doughan’s review of the new edition of On Fairy-stories.


[1]. Carpenter, H. Tolkien, a Biography 191 (George Allen & Unwin, 1977).

[2]. Hart, T. & Khovaks, I. Tree of Tales (Baylor Univ. Press, 2007).

[3]. Flieger, V. & Anderson, D. A. On Fairy-stories (HarperCollins, 2008).

[4]. Shippey, T. A. The Road to Middle-earth 45 (HarperCollins, 1982).

[5]. Hart, R. A. Hours and Wages in the Depression: Engineering 1926–38 IZA DP No. 132 (2000).

[6]. Collingwood, R. G. The Philosophy of Enchantment (OUP, 2007).

[7]. Lacon, R. unpublished papers from personal correspondence.

[8]. Tolkien, J. R. R. Tree and Leaf (George Allen & Unwin, 1964).

[9]. Hammond & Scull in The Handbook to Tolkien (HarperCollins, 2006) — under reference to On Fairy-stories.

[10]. Essays Presented to Charles Williams (ed. C. S. Lewis), (OUP, 1947).

[11]. Berman, R. ‘Tolkien as a child of the Green Fairy Book’ Mythlore Winter 127–135 (2007).

[12]. Lewis, A. & Currie, E. Uncharted Realms of Tolkien (Medea Publishing, 2002).

[13]. Tolkien, J. R. R. Introduction to Smith of Wootton Major, Extended Edition, (ed. V. Flieger), (HarperCollins, 2005).

[14]. Glyer, D. P. The Company They Keep (Kent State Univ. Press, 2007).

* Republished from Mallorn, The Journal of the Tolkien Society, spring 2009