The form I use is narrative because that is what comes naturally to me. There is of course another equally important and valid form of history which is written for the purpose of putting the material and the author’s conclusions on the record. Such an author is less concerned with communicating than with establishing the facts. He is historian first and writer second, if at all, whereas I am a writer first whose subject is history, and whose purpose is communication. I am very conscious of the reader as a listener whose attention must be held if he is not to wander away. In my mind is a picture of Kipling’s itinerant storyteller of India, with his rice bowl, who tells tales of ancient romance and legend to a circle of villagers by firelight. If he sees figures drifting away from the edge of the circle in the darkness, and his audience thinning out, he knows his rice bowl will be meagerly filled. He must hold his listeners in order to eat. I feel just as urgent a connection with the reader.
As a form, narrative has an inherent validity because it is the key to the problem of causation. Events do not happen in categories—economic, intellectual, military—they happen in sequence: When they are arranged in sequence as strictly as possible, down to the week and day, sometimes even time of day, cause and effect which may have been previously obscure will come clear. However, it is not always possible to narrate everything in straight consecutive sequence because there are always times when events are taking place simultaneously in separate places. In August 1914 the developments leading to the Battle of the Frontiers on the Western front and to the Battle of Tannenberg on the Eastern front were unfolding at the same time, putting the narrator in a quandary. The same problem was present with Stilwell when the accelerating deterioration and the launching of the last Japanese offensive took place in China while he was leading the return campaign through Burma. To break off events in one place in order to take up what is happening elsewhere ruins dramatic tension and only accomplishes utter confusion in the mind of the reader—even though that’s the way things happen in reality. One has to manipulate reality just a little and carry events through to a natural climax on one scene before moving to the other.
In organization, however, if not always in the finished product, chronology remains the spine. When I started writing The Guns of August I planned to begin with the guns going off so people should not think this was yet another book about diplomatic origins—Sarajevo and All That. I had worked out an intricate arrangement of four chapters in which war opened in each country and was followed by an internal flashback in each to explain the background. It was as beautifully designed as a Bach fugue, but when I had finished these chapters my editor didn’t know what to make of them. On rereading them, neither did I. He suggested trying it chronologically. This was so simple that I had thought it inartistic, but when the flashbacks were lifted out and put first where they belonged, behold, the result read as simply and naturally as if it had been ordained. I have avoided razzle-dazzle arrangements ever since.
With each book, one encounters new problems of organization and presentation. Obviously the dual theme of Stilwell—the biography of a man and the relationship of two countries—was a major difficulty throughout, but it was my choice, and peculiar to this book, so I can’t generalize from it—except to say “never again.” Every time I started a new chapter I felt like Jacob wrestling with the angel all through the night. Although it was hard work, the dual theme was justified, I think, because the figure of Stilwell as a continuing focus supplies human interest and drama, while the over-all Sino-American relationship gives the subject importance.
The Chinese scene of the book was another problem. It meant, as I was aware all the way through, that the reader had no familiar frame of reference. If you write a book laid in Europe or America, you can count on the reader having a mental picture of the relative location of France and Germany, or of Texas and Alaska, or where the Rockies are, or the Great Lakes. Equally with people. Once introduced let us say to Francis Drake and Walter Raleigh or Robert Oppenheimer and Edward Teller, he will have no great difficulty in keeping them distinct, but what is he going to make of Sun Li-jen and Li Tsung-jen, two prominent persons in my book, or of Yen Hsi-shan and Wang Ching-wei and Wei Li-huang and Chang Tso-lin and Chang Tsung-chang and all those other triple monosyllables—not to mention the provinces: Kwangtung and Kwangsi, which adjoin each other, Kiangsu and Kiangsi, which do not, Honan and Hunan, Shensi and Shansi, and all the rest. I tried at first to avoid using these names, and to locate places in relation to the more familiar rivers and cities, but this soon proved impossible. China’s provinces can no more be avoided than America’s states.
Especially in an alien setting like China—but the rule should hold true for all historical writing—I try never to introduce a place name without locating it in relation to some place already mentioned, nor introduce a person without describing some attribute that will fix him in the reader’s mind. People and places must be given recognizable identities, otherwise the reader flounders in a sea of unknowns; he will miss the point of this or that and sooner or later, bored by incomprehension, will drift away.
The mere parading of names without taking the trouble to locate or personify them is either simple laziness on the part of the writer or else showing-off, in which case it is no trick; anyone can do it just as anyone can double the length of his bibliography if he has a mind to. I never can understand why historians who go in for this name-dropping make themselves great reputations. In D. W. Brogan’s France Under the Republic, for example, one can count thirty names to a page, all faceless. Michael Howard recently established himself as a leading military historian with a book on the Franco-Prussian War which one can open at random at any page and find sentences like the following: “The Emperor put Failly’s 5th Corps under his command and on 5th August while the divisions of 1st Corps concentrated around Froeschwiller and Felix Douay packed off Conseil Dumesnil’s division from 7th Corps by train from Belfort, Macmahon summoned Failly to bring his corps south through the Vosges.” In the next sentence we learn that Failly’s units were spread between Sarreguemines and Bitche and could not be moved until relieved by troops from Rohrbach. On the same page is a sketch map which shows none of these place names. I am sure Mr. Howard knows all there is to know about the Franco-Prussian War, and his book was highly praised, but it left this reader giddy. I did not gather from it a picture of the Battle of Froeschwiller but only how not to describe a battle.
Another difficulty peculiar to the Stilwell book, especially to the second half, was over-documentation. Besides Stilwell’s diaries and letters, bringing the scale of events down to a daily basis which I did not want, there was a mountainous mass of military and diplomatic records: messages, reports, memoranda, conference minutes, plus all the material of the China controversy—the White Papers, the Foreign Relations series, the interminable testimony before congressional investigating committees in thousand-page volumes. Ever since the advent of mechanical means of duplication there has been a multiplication of material that cannot be dealt with by less than teams of researchers. The twentieth century is likely to be the doom of the individual historian. (Actually, I do not really believe that. Though the doom seems logical, I believe somehow he will illogically survive.) Today we have the opposite problem from that of the researcher in ancient history who suffers from paucity of records and must work from coins, tombs, and artifacts. Beginning with Gutenberg, the sources expand. The nineteenth century is really the great period, with ample information of every kind, yet short of the over-supply of today.
With the appearance of the tape-recorder, a monster with the appetite of a tapeworm, we now have a new problem of what I call artificial survival. The effort needed to write a book, even of memoirs, requires discipline and perseverance which until now imposed a certain natural selection on what survived in print. But with all sorts of people being encouraged to ramble effortlessly and endlessly into a tape-recorder, prodded daily by an acolyte of Oral History, some veins of gold and a vast mas
s of trivia are being preserved which would otherwise have gone to dust. I should hastily add here that among the veins of gold two of the richest sources I found were two verbal interviews with General Marshall tape-recorded by Army historians in 1949. Marshall, however, was a summit figure worth recording.
As a result of over-documentation I was constantly struggling with the problems of scale in the Stilwell book. It was as if I had been a cartographer trying to draw a map on a scale of 100 miles to the inch while working from surveys detailed to a scale of one mile to the inch. Following in the track of the diary and the official documents, I would get caught up in some issue that was all-absorbing at the time, and spend days writing the developments from Tuesday to Friday when what I should have been doing was the over-all development from, say, May to November. I had to stop short and remind myself: What does this matter in the long perspective?
As a result pages went into the discard—for example, the Henry Wallace mission. Because he was Vice-President, Wallace’s visit and conversations with Chiang Kai-shek assumed enormous importance at the time and blew up a swirl of passions, intrigues, and, of course, prolific reports by everyone for miles around. The path of research widened out like the mouth of the Yangtse and the narration likewise in its wake. I had an uncomfortable feeling, however, that something was wrong. Then one day someone asked me what actually had been the significance of the Wallace mission and I heard myself answering, “None.” It had really had no effect on the course of events one way or another.
Because of all the quotable reports it spawned, this affair was a good example of the bewitching effect of diplomatic documents. An episode like the Wallace mission exercises the same effect as Everest on Mallory. You write it because it is there. Then it turns out not to mean anything. It would have been false to history to leave out the Wallace mission altogether, so I condensed it as much as I could, even at the cost of cutting a wonderful characterization of Wallace by a man who said, “Henry would cut off his right hand for the sake of an idea—and yours too for that matter.” I hated to let that go, but since Wallace no longer appeared as a personality, it no longer belonged.
The larger scale cannot be achieved by blithely skipping over whole episodes or chunks of time; it requires condensing, which is the hardest work I know, and selection, which is the most delicate. Selection is everything; it is the test of the historian. The end product, after all, consists of what the historian has chosen to put in, as well as chosen to leave out. Simply to put in everything is easy—and safe—and results in one of those 900-page jobs in which the writer has abdicated and left all the work to the reader.
Selection is the task of distinguishing the significant from the insignificant. It must be honest, that is, true to the circumstances, and fair, that is, truly representative of the whole, never loaded. It can be used to reveal large meaning in a small sample. As Robert Frost said, “The artist needs only a sample.” At Chiang Kai-shek’s residence the glimpse of secret-service boots peeking below red curtains, which I took from someone who was present, was a tiny selection that bespoke a whole atmosphere. Likewise the letters of Colonel Carlson to President Roosevelt (which, incidentally, have not before been printed) crystallized, I think, the American idealized view of China at the time.
One must resist the selection that does too much. By that I mean an item or incident which, by the fact of being made part of the narrative, appears representative and leaves the reader with an impression that may not be entirely justified. The author wields tremendous influence in this way which no one superintends but his own conscience.
I remember facing one such choice at the climax of the debacle in Burma when Stilwell was trying desperately to organize transport and food for the retreat before it collapsed into chaos. The Chinese general who was Chiang Kai-shek’s personal liaison officer could not be found because, as it happened, he was elsewhere engaged in organizing the retreat to China of a Rolls-Royce which he had delightedly acquired from the British Governor-General in trade for two jeeps. I intended to cap this incident with an aphorism I had picked up from the warlord years in the 1920s: “In Chinese warfare commanding officers have never been known to retire poor.” While that may have been reasonably true, it would have left American readers with the impression that all Chinese generals were venal—which is true only in American terms. I am not an authority on China, but I know enough to know that it would be quite false to write about China in the framework of Western values. So I took out the aphorism and the Rolls-Royce too. This illustrates the reasoning behind a negative selection.
I seem to be giving you chiefly examples of what I left out, and this reflects what was a constant struggle. I made a vow when I started that I would keep the finished book under 500 pages, and in the course of that effort I discarded or radically pruned everything I thought could be spared or that was not germane to my main theme. I missed my goal by 51 pages, but it was not for lack of trying.
Which brings me to another working principle: Do not argue the evidence in front of the reader. The author’s thought processes have no place in the narrative. One should resolve one’s doubts, examine conflicting evidence, and determine motives behind the scenes, and carry on any disputes with one’s sources in the reference notes, not in the text. For one thing, this keeps the author invisible and the less his presence is felt, the greater is the reader’s sense of immediacy to the events. For another thing, by eliminating discussion one establishes a tone of this-is-the-way-it-was which the reader quickly accepts. He does not want to be bothered by a lot of maybes and perhapses, on-the-one-hand and on-the-other-hand; he wants to follow along with the action, feeling confident this was the way things happened.
In order to identify with the period it is also essential to eliminate hindsight. I try not to refer to anything not known at the time. According to Emerson’s rule, every scripture is entitled to be read in the light of the circumstances that brought it forth. To understand the choices open to people of another time, one must limit oneself to what they knew; see the past in its own clothes, as it were, not in ours. To me this is an absolute, although I realize it is one that many historians would fiercely dispute. According to their view, History is properly the interpretation of past events in terms of their consequences, and in the light shed upon them by present knowledge and present values. The history of Kuomintang China, according to this school, is told in the light of the ultimate Communist triumph, although in fact no policy-maker of the 1930s ever seriously considered that within ten or fifteen years China would be ruled by the Communists. An account told in the light of Now must be false to the past, as I see it, whereas the other school maintains that the view from inside the past results in a false judgment for today. The difference is one of philosophic stance and is unlikely to be resolved.
In closing I may say that though I do not think of myself as a military historian, I agree on the need for military history, if only to bring home to the general public that conflict has been a central theme in the human story from pre-history to the present. Except for specialist studies, military history should be treated, I think, not as a separate category, but along with political, economic, and intellectual history, as part of a whole whose object is to exhibit what a given society was like at a given time. That object, it seems to me, should be the historian’s purpose. That is what I tried to achieve in The Proud Tower, which is the reason I like it the best of my books.
* * *
Address, National Archives Conference on Research in the Second World War, June 1971. Maryland Historian, Fall 1971.
The Houses of Research
TO A HISTORIAN libraries are food, shelter, and even muse. They are of two kinds: the library of published material—books, pamphlets, periodicals, etc.—and the archive of unpublished papers and documents. In the first category, one of the greatest is happily in my home town: the New York Public Library. In resources (not to mention problems) the NYPL has everything: every published work you need to c
onsult on virtually any subject, besides a lot more you do not know you need because you do not know they exist until you come across them by serendipity. In the course of research extending over twenty years on subjects stretching from the Phoenicians of the Bronze Age to the music of Richard Strauss to Americans in China, there were, as I remember, only two books I asked for that the Library lacked. One was in their catalogue but could not be located, and both they were able to borrow for me.
Since most of the work on Stilwell was done in unpublished papers and interviews, I did not spend as much time at the NYPL on this book as on my others; nevertheless, at 42nd Street I made an unexpected strike of the kind that brings the occasional rare thrill in research. In this case it was a full run on microfilm of the Sentinel, the weekly journal of the 15th Infantry stationed in Tientsin, to which Stilwell was attached in 1926–9. These were the crucial years when the Kuomintang under Chiang Kai-shek made its bid for control of China, but up to then I had found almost nothing on the views and attitudes of the American military on what was happening all around them. To my intense disappointment, after winding laboriously through the first reel, scanning every page, I found nothing of interest; the Sentinel might have been published at some regimental post in the heart of Kansas for all its notice of China. I was ready to send the box back, but decided as a matter of conscience to look at the second reel. There on the first page of the first issue was an article by Major Stilwell, the regiment’s recognized expert on Chinese affairs, inaugurating a series, no less, on the personalities and issues of the civil war! His articles continued to appear each week in the Sentinel for more than a year, providing me with my protagonist’s own judgment of events at a climactic time in which he shared.