Cookie policy: on we use analytics cookies to understand how visitors use the site. The anonymous information they provide suggests improvements and alerts us to technical errors. For more information, see our cookies page, which also explains how to block or remove cookies.  Search T. E. Lawrence Studies

Research & Discussion

General biography

Rejected legend

Youth 1888-1914

War service 1914-1918

Diplomacy 1918-1922

Service years 1922-1935

Lawrence's personality

Writings and criticism

Lawrence and book production

Film, TV, radio

Book reviews

Bibliography & Collecting


Discussion list


Which text is better written?

Or go to Page 1 or Page 2

In some respects, the Oxford Text and the subscribers' abridgement are quite different. Each has relative strengths and weaknesses. To prefer one is not to condemn everything about the other.

A difference worth noting is that the Oxford Text was entirely Lawrence's work. That is not true of the subscribers' abridgement, read in proof by Charlotte Shaw, and in parts also by Bernard Shaw, E. M. Forster, Edward Garnett and others. In practice, most writers seek comments on their drafts. Often, as in Lawrence's case, they adopt suggested changes. The 1922 text is completely free from such influences. In the Castle Hill Press editions only the punctuation was lightly edited. The words are Lawrence's own.

As to which text is best, my views at the outset were conventional. I believed that, since the subscribers' abridgement was Lawrence's final version, it must be the better-written. The reason for publishing the fuller 1922 version would be its additional historical and biographical content. I also assumed that the subscribers' abridgement must have been the version that Lawrence himself preferred.

That was the position taken by Jonathan Cape and the other Seven Pillars publishers when they refused to publish the Oxford Text in the early 1980s. A. W. Lawrence, as literary executor, doubtless made similar assumptions in 1935 when, after his brother's unexpected death, he had to choose which text to publish.60 Yet, as I came to realise, these assumptions are debatable.

It is, of course, easy to find sentences in the subscribers' version that have been technically improved. Also, Lawrence clarified or omitted some opaque 'philosophical' passages. However, improving individual sentences does not necessarily improve a text as a whole - and in my judgment not all the changes were improvements.

While there was much praise in 1927 for the subscribers' Seven Pillars and its further abridgement, Revolt in the Desert, some of the comments were critical. Lawrence understood the criticisms and became less certain about which version was best.

An early hint of the problem came from Leonard Woolf in his review of Revolt. He wrote: 'every sentence begins again with a full breath and ends with a really full stop.'61

In August, Robert Graves sent to Karachi some draft chapters of his forthcoming popular biography Lawrence and the Arabs. In one of them he compared the two Seven Pillars texts: 'There is such a thing as a book being too well written, too much a part of literature . . . It should somehow, one feels, have been a little more casual, for the nervous strain of its ideal of faultlessness is almost oppressive . . . On the whole I prefer the earliest surviving version, the so-called Oxford text, to the final printed text.'62

Lawrence found this unsettling. Graves was a skilled writer. Lawrence had long admired his work and could not lightly dismiss his opinion. Moreover, Lawrence and the Arabs would be widely read and quoted. What might the subscribers think?

He wrote back: 'I've just heard from Edward Garnett, whose competence as a critic I vehemently admire, strongly maintaining the opposite view. He says that my revision has done wonders, in turning a shapeless story into a shapely one. I hope his view is right. It would be hard luck to do four years' work on a text [in reality, work "in my spare evenings" during three years], without bettering it: indeed I think it would be impossible, if the writer had any good in him.'63

Lawrence probably hoped that Graves would drop the comment, but Graves merely amended it. In the final version he wrote: 'On the whole I prefer the earliest surviving version, the so-called Oxford text, to the final printed book which was the version that I first read consecutively. This is a physical rather than a critical reaction. The earlier version is 330,000 words long instead of 280,000 [Lawrence's figures from Some Notes] and the greater looseness of the writing makes it easier to read. From a critical point of view no doubt the revised version is better. It is impossible that a man like Lawrence would spend four years on polishing the text without improving it, but the nervous rigor that the revised book gave me has seemingly dulled my critical judgment'64

Lawrence suggested to one correspondent that Graves was merely boasting that he had seen the Oxford Text; but privately he could not laugh it off. When Forster agreed to compare the two versions, he postponed his plan to destroy the six surviving Oxford Times copies.65 It was some years before Forster made the comparison. When his judgment came, it was politely equivocal: 'I had to admit that the sentences in the revision were more concise and showed a superior sense for the functions, and incidentally for the etymology, of the words employed in them. But the relation between the sentences seemed to me a little impaired: the correction, though logical, wasn't always easy.

'I know you will not agree with this. You will say (i) that it's impossible to take too much care over one's sentences (ii) that your defects as a prose writer spring from deeper causes, and extend more widely than I suggest. Still I thought I'd try to write it out. Your defect seems to me exactly as above defined - no larger. It would disappear if you would take less trouble of the "craftsmanship" type.'66

Forster may have said more in conversation, or in a letter I have not seen. Lawrence took no steps to destroy the Oxford Times copies. He put Seven Pillars behind him. For better or worse, the subscribers' abridgement had been issued. He had already said it would not be republished in his lifetime.

After Lawrence's death, when the subscribers' abridgement was published for general circulation, Forster's review left no doubt about his real opinion. He wrote: 'the Oxford is in the judgment of several critics even superior to the version offered now, and it is good news that a reprint of it may eventually be made.'67

The comments by Leonard Woolf, Robert Graves and E. M. Forster all point to the same problem: the language of the abridgement is overwrought. When revising, Lawrence polished the sentences individually. In doing so, he unwittingly altered or removed narrative links. So his changes tended to impede the flow rather than helping it. If you look at the text closely, the problems are not easy to spot. But try reading both versions aloud, as I did while checking the Castle Hill Press typesetting. It is soon obvious what the critics were talking about. The subscribers' abridgement is granular. The Oxford Text - despite its occasional flaws - speaks to the reader more fluently. Of course, some readers may prefer the abridgement. My point is this: there is no consensus to support my assumption, at the outset, that the abridgement is better written.

Lawrence's lopsided abridgement

Probably the most significant difference between the Oxford Text and the subscribers' abridgement is length. Taken together, the cuts Lawrence made add up to about 84,000 words. If the omitted passages were added to the current Penguin edition of Seven Pillars (which is the subscribers' abridgement) the book would be around 200 pages longer. Why did he cut so much?

I trained as a historian not a literary critic; but I often edit and reduce texts for publication. Among these was my authorised biography of Lawrence, which I abridged for a 'concise edition'.

Consciously or unconsciously, when making an abridgement you develop a set of principles. These draw your attention to passages you might omit. However, looking at Lawrence's treatment of the first and second halves of Seven Pillars, I could find no coherent approach, still less a set of literary principles. There is often little or no difference in type between content retained in one place and cut in another, while the cuts in the second half were far more ruthless than in the first. For example, some readers find the Introductory Book - consisting of background history and analysis - such heavy going that they abandon Seven Pillars without ever reaching the main story. It survived virtually intact. So did the accounts of early journeys, which could easily have been trimmed. Yet, on the evidence, if these sections had been in the second half of Seven Pillars they would have been shortened.

I returned to these puzzles in the spring of 2004, searching for clues in Lawrence's letters. As he worked on the Oxford text and the subscribers' abridgement between 1921 and 1926 I paid special attention to those years. What I found shows that his treatment of Seven Pillars was linked to his deteriorating mental balance.

Given his statements at the beginning of the project that the book would be shortened only slightly, my first object was to discover if and when his intention changed. The answers are clear, both from the text and from his letters. It was probably in the first weeks of 1925 that he decided to make heavy cuts in the part of the book he had yet to revise. At the time, the revision was about half way through.

Prior to that he was still working to the original plan. Thus in September 1924, eight months after beginning work, he wrote to a friend: 'The new edition of my book, to be ready next year probably, contains 90% of the matter of the private copy you have.'68 A fortnight after that, he wrote to Garnett: 'did you hear of my difficulty in cutting it down only 12%? I'm like a parent who has been converted to 4-toed children, but can't make up his mind which of the five is best off.69 As the table on Page 2 shows, the early Books of Seven Pillars that he had worked on during the preceding months were trimmed only slightly more than the original target.

The cuts from Book V onwards were of a different order; but why? One difficulty that appears in the correspondence was a problem with funding. Could that have led him to make the cuts, perhaps in order to reduce the remaining cost? By February 1925 the project was running well over budget. The bank was evidently concerned. However, Lawrence solved the problem almost immediately by selling rights to a popular abridgement, Revolt in the Desert. In the interim, the bank held his Pole Hill land as security. By March 1925 there was no further worry about money. In fact, Lawrence spent lavishly thereafter, adding illustrations. He later calculated that copies of the finished book had cost three times more than the subscription.

His real concern, as is clear from the letters, was time. He lamented that: 'The 7 Pillars goes so slowly'.70 The cause of delay was simple: it was the time Pike had to spend resetting and adjusting pages to satisfy Lawrence's growing obsession with typography. Pike's work was also hampered to an extent by temperament and domestic problems.

The succession of proofs slowed Lawrence as well, forcing him to switch his focus repeatedly from one part of the book to another, rather than concentrating on revision. In February 1925 he wrote to Buxton: 'I am cutting out much, changing some, and smoothing out many of the stylistic roughnesses. To your eye it may not seem difficult... but the most I can manage is a page a day, and any extra duty (such as we have sometimes) or a visitor, or an urgent letter from outside, breaks the current of my corrections, and makes their progress slower. It is really hard work, since it is grudged work. As you know I don't think much of the book as a piece of writing; its faults are due to my own insufficiency: and I'm trying in this revise . . . to cut out some of the rubbish, and polish off some of the roughness.'71

One way to speed up the project was to get Pike an assistant. In principle, each section of Seven Pillars should have been be printed once the last corrections were done. In practice, the printing had fallen behind. In May Lawrence hired an experienced pressman, Herbert Hodgson, leaving Pike free to work on typesetting.

But by then Lawrence had found a still more effective way to shorten the job: he axed the number of pages that remained to be done. Cutting the text saved time at every stage. While working on Book IV he wrote: 'My abridgement consists in cutting out every fifth word of the old text: when possible. If the fifth won't go out, the sixth probably will'.72 The result was a cut of 21%. In Book V the cuts reached 27.5%. Book VI escaped fairly lightly - he thought it the best-written section of Seven Pillars - but Book VII lost 29% and Book VIII 47.5%. The two final books each lost a third.

Early in 1925 Lawrence began to talk of finishing Seven Pillars by Christmas. It was an unrealistic target. Less than half the text had passed through its proofing stages during the first year - and much less had been printed. How could the rest, plus additional tasks such as maps and index and binding, be achieved in the second? Even with help, Pike could not have finished Seven Pillars so quickly. Nevertheless, that was the deadline Lawrence set. He told Hornby that he would need a pressman for about five or six months.

At that stage, nothing could be gained by reworking the early Books. Some were already printed, others in advanced proof. To interfere with them would lose time, not save it. So the new urgency could only affect the second half of Seven Pillars.

Why had time become so important? It is tempting to think that he was worried about the subscribers. Could he try their patience indefinitely? In June 1925 he sent them a printed letter, apologising for the delay. Hardly any dropped out. In the event, almost all would wait patiently until 1927 for their copy of Seven Pillars. They were not his main concern.

The real motive for his haste ran much deeper. It probably defies rational analysis and for that reason may never be fully explained. Revising Seven Pillars was an unfortunate kind of therapy for his troubled state of mind. The work reminded him of wartime experiences he needed to put behind him. Seven months into the task, he had written to a friend: 'Lately the job of proof-correcting has made the war-memories very vivid to me, so that they have been coming back as night-terrors to shorten my already few hours regular sleep.'73 At about that time Lawrence persuaded John Bruce to administer beatings, a masochistic disorder which both John Mack and I linked to his experience at Deraa in 1917.

In June 1925, Lawrence sent a letter to Edward Garnett in which he threatened to commit suicide. Earlier letters show that he had been considering this - perhaps fearing it - for some months. Thus he had written to Robin Buxton the previous November seeking to increase the bank's security: 'The idea in my mind was that I could assign to the Bank, in case of my death or disappearance, the right to publish an abridgement . . . and to apply the profits of such transaction to meeting any charges they had against either of my accounts. . . .

'Is this a possible document? It won't be necessary if I go on all right... but I might go off, just as easily. A burst front tyre, or weariness, or the other fate I'm always fearing. You know, Robin, I'm hardly sane at times.'74

A few days later he wrote in strange terms to Lionel Curtis, one of his closest friends. 'I think I've about done with the Tank Corps, since I now fit it smoothly.... And the next step must be considered with circumspection. That's the worst of having once been fool enough to have been "somebody". At times I fear that my news value may not end till the day after I'm dead.'75

In January 1925 Lawrence sent Curtis a still stranger letter: 'I'm here, probably . . . till the labour of this reprint is past: say November 1925. After that, if I feel as at present, I'll look for some safer employment... but it will be difficult to find . . . In the next stage there may be obeying of orders... no objection to that... but there must be no scope for voluntary effort supplementary to, or apart from, orders. That's where I fail here: the leisure lets me carry on another life when I'm off duty'.76 In what seems to have been an unrelated strand of thought, Lawrence was nearly desperate to transfer from the Tank Corps back into the ranks of the RAF, where he had served briefly in 1922. The strength of his feelings about this is hard to understand. In private letters he pestered Trenchard, the Chief of Air Staff, about it.

In mid-May 1925 he learned that his latest request for transfer had been refused. Two letters written on the 16th hint at what was to come. He wrote to Charlotte Shaw: 'Exactly what effect the disappearance of my last ambition will have upon my course I can't say yet, since for the rest of the year all my attention must be upon finishing the revise of the Seven Pillars. About Xmas I will have to make up this very veering and fickle mind, afresh.'77 To Robin Buxton he wrote: 'I'm most grateful to you and your bank for making this last two years possible: and your reliance on my not pegging out (deliberately that is) without settling all up, shan't be disappointed.'78

The crisis became obvious a month later, when Lawrence sent Edward Garnett his revised version of Book VI, containing the Deraa chapter. He wrote: 'Trenchard withdrew his objection to my rejoining the Air Force. I got seventh-heaven for two weeks: but then Sam Hoare came back from Mespot, and refused to entertain the idea. That, and the closer acquaintance with the Seven Pillars (which I now know better than anyone ever will) have together convinced me that I'm no bloody good on earth. So I'm going to quit: but in my usual comic fashion I'm going to finish the reprint and square up with Cape before I hop it!

'There is nothing like deliberation, order and regularity in these things. I shall bequeath you my notes on life in the recruits camp of the R.A.F. [later completed as The Mint] They will disappoint you.'79

The following week a further letter to Garnett revealed the extent of the cuts Lawrence now proposed for Book VIII: 'After that point, in the old text I put a "flat"... a dull area, describing a dull period, before the final effort against Damascus . . . Everybody who tried to read the book . . . said the "flat" was hopelessly long. I agree . . . The crucial question is to make brief the flat: while yet leaving it a very marked flat. In the Oxford text it covers two Books: VIII and IX. Eight is attached, and the summary of IX, to help what I hope will be your advice upon the curtailment . . . In my second revise about 10% more will come out, to clean up the ruins of the first cut. Don't worry about that... but tell me if the shortening is well and truly done. More than half has gone. I aim at shortening Book IX proportionately.'80

The claim that 'everybody' who had read the Oxford Text found the flat too long was untrue. (In Some Notes - a more public document - he wrote 'several of those who . . .') As mentioned earlier, in the surviving correspondence I have seen only Garnett had wished it shorter. It is surely significant that Lawrence now turned to him for advice, rather than to Forster or the Shaws.

The evidence cited thus far shows that Lawrence started cutting Seven Pillars heavily in the spring of 1925, at a time when he was gripped by some kind of psychological crisis. The link between them is confirmed beyond possible doubt in two further letters.

The first, to Charlotte Shaw, was written in July. By then, news of his suicide threat had reached senior politicians. To avoid a scandal, he would be allowed to rejoin the RAF. He wrote: 'This has made the world feel very funny. The first effect was like a sunset: something very quiet and slow: as if all the fuss and trouble of the day was over. Now I feel inclined to lie down and rest, as if there was never going to be any more voyaging. I suppose it is something like a ship getting into harbour at last. The impulse to get that book finished by Christmas is over. I may be living on for years now, and so why hurry it?'81 The second letter, written several months later, is to Francis Rodd: 'Mark me down for a further spell of quite happy existence. That also is an odd change. For I had made up my mind, in Bovington, to come to a natural end about Xmas, when the reprint of my book would have been finished.'82

By the time he learned of the transfer, his abridgement of Books V-X was nearly finished. It was too late to reconsider the cuts, though at the Shaws' request he did restore some passages. A letter to Pike, written a few weeks afterwards, comments: 'Mrs. Shaw has VIII, IX and X. We have been fighting over 8 and 9. They want them longer: I want them shorter.'83

The subscribers' Seven Pillars was finally sent to the binders at the end of 1926. Some Notes, circulated to subscribers the following spring, became the accepted account of its history. When the abridged Seven Pillars was published after Lawrence's death, Some Notes was included in the Preface. Until now, no one has questioned its accuracy.

The Oxford Times copies survived. A letter from November 1925 may explain why Lawrence had once been so keen to destroy them: 'The reprint differs, in many ways, from the "Oxford" text . . . I do not want to leave bibliophiles of the twenty-first century two variants, to spend useful hours comparing and cross-checking.'84

No indeed. If they did that, they might unravel the curious history of the subscribers' abridgement.


The creative drive that inspired and shaped Seven Pillars gave out in 1922. Its achievement was the Oxford Text. That was the version which Bernard Shaw described in a letter to the Prime Minister as 'a masterpiece, one of the few very best of its kind in the world'.85 He had tried repeatedly to persuade Lawrence to leave it alone: 'You have something to say; and you say it as accurately and vividly as you can: and when you have done that you do not go fooling with your statement with the notion that if you do it over again five or six times you will do it five or six times better.'86

It may be that the Oxford Text fell short of Lawrence's ambition. Yet meddling with it in his spare time - while tired, depressed and in a wholly different frame of mind - was never likely to improve it. Until the Oxford Text was published, few readers could judge it against the abridgement. Safe from comparison, the abridgement became a world classic, enjoying huge success for nearly seven decades. But is it much more than a literary curiosity? Yes, many sentences were improved; but overall, the tightened prose seems to hinder the narrative, not help it. The thousands of changes made to prettify the typesetting served no literary purpose - no purpose at all once the book had been re-set. In the rush to end the task, Lawrence left out much that is of interest. A literary scholar might choose to read both versions. But if you have time for only one, almost all the arguments point to the fuller text.

I will leave the last word to John Rodenbeck, a professor of English Literature who lived and worked for many years in the Middle East. Preferring the Oxford Text to the subscribers' abridgement, he wrote: 'A much nobler voice speaks here. Many puzzles are unravelled and many new avenues opened.'

Jeremy Wilson

Copyright 2004, 2006, Jeremy Wilson. Previously unpublished quotation from letters by T.E. Lawrence copyright 2004, The Seven Pillars of Wisdom Trust. John Rodenbeck quoted with his permission.

Seven Pillars of Wisdom, The Complete 1922 Text, Fordingbridge, Castle Hill Press, 1997, 2003 (revised one-volume edition with index).


60. When Lawrence died, Jonathan Cape immediately announced a reprint of Revolt in the Desert. This, however, was potentially damaging to prospects for publishing Seven Pillars. A. W. Lawrence as Literary Executor told Cape that if he reissued Revolt he would lose any chance of publishing Seven Pillars. Cape cancelled the new edition of Revolt (though it was serialised) and at astonishing speed produced an edition of Seven Pillars. Only the subscribers' abridgement could have been produced so rapidly, since an edition of the Oxford Text would have required months of editorial work.

61. Leonard Woolf, review of Revolt in the Desert in Nation and Atheneum, 19 March 1927.

62. T. E. Lawrence to his Biographer Robert Graves (London, Cassell, 1963) p.116.

63. Lawrence had written to Garnett, in a letter seeking reassurance about his work: 'I rejoice that you are going to read Seven Pillars: and I'll hope to have your critical opinion, when you end it. It matters to me, for I put months, years, of work into it after you said it was worth working at. You will find hardly a sentence of the Oxford text standing. If I'm any good at all at writing, the revised S.P. should betray it.' (This letter is dated 1 August 1927, but evidently sent earlier, since Garnett replied to it on 18 July. See DG p.533 and Garnett's letter LTEL pp.93-6.) As was to be expected, Garnett's reply was fulsomely positive. Heaping praise on the subscribers' Seven Pillars, he urged Lawrence to write more. Lavish praise was Garnett's technique for encouraging a diffident author - and Lawrence probably guessed that. He was nevertheless happy to offset Garnett's comments against those by Graves.

64. Robert Graves, Lawrence and the Arabs (London, Jonathan Cape, 1927), pp.407-8.

65. On 27 April 1927 Lawrence wrote to Forster about the Oxford Text: 'It is, I think and hope and believe, doomed to be destroyed in the fullness of time, upon your confirming my verdict. If not destroyed, it will be bestowed in some scholarly and shadowy place.' Transcript in TEL Papers.

66. E. M. Forster to T. E. Lawrence, 18 January 1931, LTEL p.74.

67. E. M. Forster, broadcast review of Seven Pillars of Wisdom, published in The Listener, 31 July 1935.

68. T. E. Lawrence to W. F. Stirling, 1 September 1924, transcript in author's possession.

69. T. E. Lawrence to Edward Garnett, 16 September 1924, transcript in TEL Papers.

70. T. E. Lawrence to Edward Garnett, 9 April 1925, DG p. 473.

71. T. E. Lawrence to Robin Buxton, 18 February 1925, Jesus College, Oxford.

72. T. E. Lawrence to Robin Buxton, 3 May 1924, Jesus College, Oxford.

73. T. E. Lawrence to Alan Dawnay 27 July 1924, MB p.271.

74. T. E. Lawrence to Robin Buxton, 25 November 1924, Jesus College, Oxford.

75. T. E. Lawrence to Lionel Curtis, 5 December 1924, All Souls College, Oxford.

76. T. E. Lawrence to Lionel Curtis, 2 January 1925, quoted in Wilson p.750.

77. T. E. Lawrence to Charlotte Shaw, 16 May 1925, Letters I p.134.

78. T. E. Lawrence to Robin Buxton, 16 May 1925, Jesus College, Oxford.

79. T. E. Lawrence to Edward Garnett, 13 June 1925, DG p.477.

80. T. E. Lawrence to Edward Garnett, 21 June 1925, transcript in TEL Papers.

81. T. E. Lawrence to Charlotte Shaw, 4 July 1925, Letters I p.141.

82. T. E. Lawrence to Francis Rodd, 3 November 1925, DG p.485.

83. T. E. Lawrence to Manning Pike, 21 October 1925, copy in private collection. Original thought to be in the Henry E. Huntington Library and Art Gallery.

84. T. E. Lawrence to Francis Rodd, 6 November 1925, transcript in TEL Papers.

85. Bernard Shaw to Stanley Baldwin, 31 May 1923, Letters I p.41.

86. G. B. Shaw to T. E. Lawrence, 4 January 1923, Letters I p.31.

Copyright, privacy, contact | Cookies help