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 Castle Hill Press


 Untitled 1


Military Report on the Sinai Peninsula
Towards 'An English Fourth'
Seven Pillars of Wisdom, The Complete 1922 Text
Paperback edition
'The Mint' and Later Writings About Service Life
Boats for the R.A.F. 1929-1935
reports and correspondence


The Forest Giant


Correspondence with Bernard and Charlotte Shaw
Correspondence with E. M. Forster and F.L. Lucas
More Correspondence with Writers
Correspondence with Edward and David Garnett
Correspondence with Henry Williamson
Translating the Bruce Rogers 'Odyssey'
Correspondence with the Political Elite 1922-1935


Check for programme updates on our News page

Thirty-nine (or so) steps: how we assemble a volume of T. E. Lawrence letters

Do you ever ask yourself how the book in your hands came into being? Possibly not: whatever the process, what you enjoy is the finished product.

Yet our subscribers do sometimes ask how we set about editing and producing these volumes. The answer is more complicated than most people expect.

"Why don't you write that up somewhere..."

So - strictly for those who are interested - here is a step-by-step outline of what we do.

PART I: From Start to Content Draft

1. Locating and collecting letter texts
2. Putting letter texts on disk
3. The working Chrono Files
4. How should the letters edition be organised?
5. Dividing-up volumes

6. Listing the volume content
7. Getting the best possible text
8. Designing the page
9. Typesetting the text
10. First check
11. Adding outline notes and queries
12. Reading the chrono files
13. Working through the editorial notes
14. Reading through
15. Out-of-house research
16. Rechecking the texts
17. The Content Draft
18. Looking for illustrations
19. Getting high-quality prints or scans

Next page:
PART II: From Content Draft to Finished Book


1. Locating and collecting letter texts

It was over forty years ago, in the late 1960s, that I began to think of editing a large-scale scholarly edition of T.E. Lawrence's correspondence. I discussed the idea with A.W. Lawrence. He saw a place for such an edition, but thought its time had not yet come. In the meantime he was keen to publish an updated general selection along the lines of David Garnett's Letters of T.E. Lawrence. In the early 1980s he asked me to edit that - but by then I was already working on the authorised biography and that had to take priority. In any case, some preliminary thinking only confirmed that editing a general letters selection did not really appeal to me. I suggested offering the project to Malcolm Brown.

The authorised biography had nevertheless given me a reason to collect texts of Lawrence's letters. Between 1975 and 1989 I gathered several thousand. Many were photocopies of originals; some were typed or handwritten copies; some were photocopies of earlier transcripts. It is fortunate that so many of the letters were copied in the late 1930s, in preparation for David Garnett's selection. While most of the originals are now in one or other of the major collections, some have disappeared or are inaccessible. Others may have been destroyed during the war.

2. Putting letter texts on disk

The letter texts I gathered arrived on different sizes of paper. As a collection they were awkward to use. I decided to store these primary source-copies in a safe place. Having been an "early adopter" of word-processing machines, I arranged for my secretary to use spare time keyboarding letters. Initially the transcripts, in the form of ASCII (.txt) files, went on 8-inch floppy disks (remember those?)

In the long run, having letter texts on disk has proved extremely useful. Nowadays they are on the hard-drive of my PC, so I can run word-searches for names or topics.

In 2006 we made some of this benefit available to the general public, by posting works and letters that had gone out of UK copyright on, where visitors can search them using Google.

3. The working chrono files

We used print-outs of the transcripts to build up a series of working loose-leaf files, where documents are arranged in chronological sequence. To these we added collateral material such as contemporary press reports, and also relevant material from wartime and diplomatic archives.

The chronological files became my key resource for the authorised biography. They are equally valuable for our Letters editions. Nicole and I add to them whenever we come across new material.

The files provide a remarkably detailed chronological autobiography. Some letters (notably to his biographers Graves and Liddell Hart) contain comments about earlier periods in his life. We often put copies of such comments at the relevant date.

4. How should the letters edition be organised?

When I began thinking about a scholarly edition of Lawrence's letters, I had to take some decisions. The first was fundamental. Should we arrange letters chronologically, as in our working files, or by individual correspondent?

The selections edited by David Garnett and Malcolm Brown show both the strengths and weaknesses of chronological arrangement.

A chronological selection can retain the feel of an "autobiography through letters" - which is what A. W. Lawrence intended. But if you include letters to different recipients with similar content you risk irritating readers - even though the different ways Lawrence wrote about a topic to this person or that may provide interesting clues about those relationships.

Garnett omitted letters with similar content, while Brown omitted repetitive paragraphs. Both techniques distort the impression a reader might gain of an individual relationship.

A more serious handicap in a general selection is lack of space. Garnett's selection, the longer of the two, contained perhaps 10-15% of Lawrence's surviving letters.

I find the fragmentation of Lawrence's correspondence in these general collections frustrating. He rarely met the people with whom he exchanged letters, so in a sense the letters "were" the relationship. As Robert Graves pointed out, the character of these friendships-by-post could differ markedly.

One evening in September 2004 I had dinner with John Mack and Malcolm Brown (it was the only occasion the three of us met in private). Among other things, John talked about the way Lawrence showed different sides of his personality to different people, sometimes to such an extent that he seemed to change personality. John did not think this behaviour exceptional - everyone does it to some extent - but it can be extremely revealing.

I decided that our letters edition must be based - where that made sense - on individual relationships. For anyone wishing to read the letters in date sequence, we would build a chronological list covering the content of all the volumes.

5. Dividing-up volumes

While most of the edition would focus on individual relationships, few of those produced enough letters to fill a volume. So how should we group the shorter correspondences?

Our thoughts on this have evolved, partly because what began as an edition of Lawrence's correspondence has ended up including works as well. Some volumes will include both works and letters, such as:

  • Boats for the RAF, which includes Lawrence's manual for the 200 Class Seaplane Tender
  • The letters volume for 1905-10, which will include Lawrence's university thesis
  • A volume now in preparation which will include both Lawrence's Odyssey translation and his letters about the project

The division into volumes as currently proposed is listed here.


6. Listing the volume content

Once we have decided what group of letters a volume will contain, we list everything suitable for inclusion that we already have or know of. This seems an obvious thing to do, but exactly what we include in the list has evolved over time.

We use an Excel spreadsheet, with rows for the letters to be included, and columns for:

  • The current text source
  • The copyright status
  • Copyright ownership

and to track:

  • Obtaining missing source texts or copies from better sources
  • Typing in a new text
  • First check from best available source
  • Second check from best available source

(In Part II an edited version of this spreadsheet will serve a another important purpose.)

7. Getting the best possible text

The next task is to obtain the best possible source - ideally a good copy of the original.

If we have a transcript but know where the original is, we try to get a photocopy or microfilm or scan.

Failing that, we may be able to see the original to check our text.

If we can't locate the original, or can't get access to it, we use the best text we have. Most Lawrence letters have been copied or transcribed at one time or another. Over time many of these copies have passed to accessible collections.

At this point I should say how grateful we are to the private collectors who give us access to original letters.

A few collectors worry that publication might reduce the value of a letter they believe is unpublished. Whatever sellers may claim when asking a high price for an "unpublished" letter, I don't think publication does affect value in any obvious way. Nothing will make a trivial letter important, while an important letter will always be desirable.

In reality, if "unpublished" status is perceived to add value, then that value is quite probably false. Lawrence died more than 75 years ago. No seller could guarantee that, in all that time, a letter has never been copied. If a copy or a transcript exists, a letter-text that is unpublished today could be published (from that copy) tomorrow. That's because owning an original letter does not confer copyright ownership of its content.

It is surely far more important that serious long-term interest in Lawrence continues and grows. That is the purpose of the work Nicole and I are doing. Castle Hill Press projects,, and the time I spend advising exhibitions and media-projects all point in the same direction. We aim to put reliable information about Lawrence where someone who is interested will find it. Without that interest, what would any Lawrence letter be worth?

8. Designing the page

Rather than assembling our books as word-processed documents, we build them from the outset as typeset text. That lets us spread the task of formatting book titles, letter-headings, signatures and so on. Also, the earlier we have the text in typeset form, the more opportunity there'll be to spot formatting errors.

Before we start typesetting we must take decisions about page format, typeface, interline spacing and so on. It's possible to change these later (and occasionally we do) but we prefer to have them right at the outset.

We have used only three page-sizes in our editions, but have varied the typography to suit the content. Much thought goes into these matters - and the issues involved are too complicated to cover here. In the long run, I hope to post design notes on each of our books, but there are some general design notes here.

9. Typesetting the text

We usually start by creating an empty typesetting file and (on the basis of our initial list) adding the central italic heading for each letter we plan to include.

Then we go through the letters we have on disk, adding letter-texts beneath their headings.

Next, we type-in any texts that we only have in paper copies.

That sometimes leaves us with a few blanks - letters we know exist, though we lack a source-copy. We set to work to obtain one.

10. First check

In principle, we check every letter in our printed text at least twice. We do this together, one reading aloud from the source while the other follows in the typeset text. The task demands far more concentration than you might expect. We find we can't do it successfully for more than three or four hours a day, broken up into several sessions.

11. Adding outline notes and queries

At this stage we have a division of labour. Nicole goes through the text looking for things that will definitely require a note (such as dates of people mentioned, and correct titles and publication details for books) and for other points where explanation could be helpful (such as references to contemporary events). Nicole adds draft notes in the typeset text that contain the required information.

To help with this task we have a fair collection of reference works and a wall of biographies and autobiographies covering the inter-war period. There are also Internet resources such as the catalogues of the British Library and the Bodleian Library, and the online archive of The Times. We use sources such as Wikipedia and special-interest websites, but with caution. Even for something as basic as a date of birth, websites can be unreliable.

Nicole's first pass inevitably leaves unanswered queries. Many of these will be resolved by the work I have been doing meanwhile.

12. Reading the chrono files

While Nicole works on notes, I read through our chrono files covering the same period as the volume (and often through to 1935, to catch anything retrospective). I look for:

  • Content in other letters that may narrow-down or even pinpoint the dates of undated letters in the volume
  • Content in other letters that clarifies activities referred to in the volume - for example, Lawrence may mention vaguely that he has been away for a few days, while other letters state where he has been
  • Content in other letters that helps explain obscure allusions in the volume - for example, Lawrence may mention a book he is reading without giving the title; but he may mention the title in a contemporary letter to someone else
  • Comments written by Lawrence, in letters to third parties, that fill out the narrative, as in Boats for the RAF
  • Comments written by Lawrence, in letters to third parties, that express opinions about the recipient(s) of letters in the volume. Such material can be extremely interesting, as in Correspondence with E.M. Forster and F.L. Lucas. Sometimes we have access to comment of this kind on the other side. For example I went to Cambridge to read copies of Forster's letters, notably to Siegfried Sassoon, looking for comments about Lawrence. In practice, though, the search for third-party comment cannot be exhaustive. If we know of material we can see, we try to see it.

13. Working through the editorial notes

The next stage is to put all this together. With the typeset text on a wide screen we can both read, we start at the beginning of the book and work through, letter by letter. As we do so we add headnotes, including material to be quoted from the chrono files - which we double-check as we go along. Re-reading each letter, we consider whether we have picked up everything that needs a note. Then we finalise the notes Nicole has researched, and complete any that can be answered from the chrono files.

There are invariably a few stubborn queries that need further research somewhere like the British Library or the Bodleian, or in a specialised collection; or we may need to ask someone working in another field. Nicole makes a list of these.

This is the stage where the book takes shape. As it does, we get a better idea of its likely length.

14. Reading through

By this time we will have run through one or two print-outs of the typesetting file. After a while these get so covered with working notes and comments that we need a clean copy.

When we finish the draft notes we make a new print-out which we both read. I often do that on train journeys to London, on my way to search for answers to the remaining queries.

15. Out-of-house research

It should by now be obvious that editing one of these volumes involves months of work. The stubborn queries, in particular, can be extremely time-consuming. An example that has stayed in my mind comes in Lawrence's letter to E.T. Leeds of 30 September 1913, where the final paragraph refers cryptically to "I.M.F.", "D.A.P." and "Wood-Bodicote". I and Ian Wood, my research assistant at the time, must have spent two days trying to discover what they were about. Finally - and without much optimism - I asked Ian to go to the Ashmolean Museum library to look through any academic papers Leeds might have written at the time. He found one that produced the complete answer (see T. E. Lawrence, Letters to E.T. Leeds, Whittington Press, 1988, p. 83, note 4).

We spend, typically, between two days and a week in libraries researching answers to stubborn queries. Even so, there are usually a handful that remain.

16. Rechecking the texts

Spread through these stages we do a second check on the letter texts, again reading the letters aloud. If an original is very difficult to read we often ask for another opinion.

17. The Content Draft

We now update the typeset file. There are usually quite a number of amendments following our read-through, and with luck we have resolved most of the stubborn queries. Any that remain are noted. The result is a reasonably complete text of the body of the book. At this stage the prelims still lack the introduction. Likewise, there is no end-matter (index, acknowledgements, etc).

We call this the Content Draft, and make four or five laser-printed proofs which we send to volunteer out-of-house readers for comment.

Although the text is formatted as printed pages, the typesetting of the Content Draft is not final. It is the modern equivalent of a galley proof. The page-endings are not yet fixed, so we can still add or remove content.

18. Find the illustrations

Rather than researching facsimiles and contemporary photographs to illustrate our books, it might be simpler to commission illustrations from an artist. In some collector's eyes, doing that would enhance our status as a fine press.

But we aren't publishing fiction. There are sometimes photographs we cannot leave out (for example, because they are discussed in letters). We don't think archive images and modern artistic images mix well, so we don't commission illustrations. When we do reproduce a drawing or painting (for instance, a portrait), it has to date from the same period as the text.

We usually need more illustrations for the 'specials' than for standard copies. Finding them may be simple - we have built up, over the years, a reference collection of Lawrence-related photographs. Sometimes, however, we need to search. That can be time-consuming. For Boats for the R.A.F. I hoped from the first to include some original design drawings, having seen a general-arrangement drawing of a 200 Class boat reproduced elsewhere. The Scott-Paine archive seemed a good place to look, but it is stored in a warehouse and public access is difficult. It was several months before I was able to see it. I then spent half a day sifting through the surviving drawings to choose something suitable.

19. Get high-quality prints or scans

Even if we have a print of a photograph, the quality may not be good enough for reproduction. There are many copy-prints of Lawrence photographs in circulation, and these rarely approach the quality of a print from the original negative. Obtaining high-quality copies or scans can involve long lead-times. Sometimes we order and compare versions of the same image from different sources, to see which is best. In the case of the boats volume, the large-format drawings had to go to a specialist for scanning, as part of a much larger batch. The entire process, from my first enquiry to receiving the scans, took ten months.

Copyright © Jeremy and Nicole Wilson, 2011

Related pages:

We accept payment by PayPal, Mastercard and Visa (via Paypal 'guest' service), or sterling cheque drawn on a UK bank