What follows is based on the premise that the historian is a literary craftsman who composes in words, for the mind’s eye, a picture of the past. He works in the realm of imagination, for it is only in imagination that he can recapture people, places, and circumstances that, in their physical reality, are gone forever.

It hardly needs to be said that the image the historian presents should be as close as possible to the vanished actuality it pretends to represent. How does he go about meeting this requirement?

One way is by research. The historian combs through such records as the past has left: official documents, letters, eyewitness accounts, reports at second hand. In pursuing this research he is in a position similar to that of a jury in a criminal trial. He has to discriminate among the bits and pieces of evidence in terms of credibility, selecting those that are then combined in his imagination as one combines the pieces of a picture-puzzle, the objective being to end up with as complete and convincing a picture as possible of the people involved, of the circumstances, and of what actually happened.

In building up his mental picture the historian is faced, time and again, by the need to make choices among alternative possible interpretations, choices based on nothing more reliable than his own judgment of plausibility. Having to select among such alternatives, he will reject some because, in his own intuitive view, they are contrary to human nature as he understands it, or to the behavior to be expected of a certain nation or a certain individual. This kind of subjective judgment—of how human beings behave, individually or collectively—calls for the same quality of insight as marks the writer of such fiction as undertakes to hold the mirror up to human reality. So it is that the great historian (e. g., Thucydides) is great by virtue of his possession of an insight into human behavior like that of a great composer of fiction (e. g., Shakespeare).

All this is to say that I share the view of those who regard history as belonging to the domain of literary art, as allied to fiction.

This, however, is a view that makes many of us professors uneasy today, even where we do not reject it outright. For it seems to go counter to the view we are disposed to propagate among our students and the public, that history is a scientific discipline, no less so than chemistry.

The view of history as a science, which did not exist before the latter part of the 19th century, has its own history.¹ Before the 19th century, the writing of history was not an established profession under the jealous guardianship of the academic world. Thucydides, Caesar, Tacitus, Polybius, Livy, Eusebius, Froissart, Machiavelli, Guicciardini, Voltaire, Gibbon, Hume—none of these were accredited members of a community of historians that imposed its own norms of approach, of method, and of presentation. Consequently, their individual genius and their foibles alike had free play. Gibbon wrote as he pleased, for there was no body of professional peers to whose collective judgment his work had to be submitted, whether as a condition of publication or as a basis for criticism after publication had taken place.

No one can doubt the advantages of bringing the historian’s vocation under the discipline of university faculties, as happened in the 19th century. Standards of objectivity and scholarship, representing the scientific disposition, were then set up as a basis for condemning meretricious work and encouraging its opposite. I need not here elaborate on the recognized improvements this made in the practice of historiography. Alongside these improvements, however, were the standard evils that result from bringing the exercise of the mind under the discipline of an academic establishment that, inevitably, comes to constitute itself the guardian of orthodoxy.

The evil consequences of having a ruling establishment in any field of scholarship are magnified by the swelling of its membership. As the community of historians rises from ten to ten thousand, mediocrity becomes increasingly prevalent in it, and the consensus of mediocrity, identified with orthodoxy, becomes increasingly dominant, not to say tyrannical.² So the occasional individual who thinks for himself, thereby separating himself from the common mind, may find it impossible to make a career.

If one takes the position that historiography is a science, to be conducted according to scientific methods, then the question arises of where, in its practice, such methods can be applied. The answer is that they can be applied to the research that precedes the actual writing of history. At this point, however, one finds oneself embarrassed to say what, precisely, are the applicable methods to which one has attached such an aura of difficulty and importance.

The first and most obvious answer is that the researcher must evaluate his sources for reliability by every means possible, rejecting forgeries, allowing for bias. This is, indeed, a matter of the first importance, but one that has always been fairly obvious. It is hard to say just what other elements of scientific method are relevant—if one disallows, as all should but none do, the pretentious use of a pseudoscientific jargon. Some researchers have made play with methods of quantification, but one can hardly say that anything except a marginal role has been found for these methods in most historical research. (Variations over the years in the number of hectares under cultivation for cabbages, or changes in tax rolls, may be highly significant in a sociological interpretation of history.) For the rest, in the absence of anything better, scientific methods are associated with outward formalities in the presentation of the results of research. The scholarly apparatus of footnotes and bibliography assumes an importance as evidence of scientific rigor that it may or may not have. Again, however, this kind of thing, useful as it may be, cannot be regarded as central to the historian’s objective of presenting a picture of the past.

II

After the above ramble through its outskirts, I come at last to my theme itself. The scientific approach, in all fields, tends to bear on analysis rather than its opposite, synthesis. It tends to be reductionist rather than holistic. It is readier to take its subject apart, dividing it into its ever finer components, than to assemble the components in ever larger wholes. It prefers the short view to the long, the microscope and dissecting needle to the telescope.

The fact is that analysis, as opposed to synthesis, comes within the competence of anybody who enjoys moderate intelligence and the rudiments of an education. It requires no imagination to disassemble a watch, that being a task which any normal child would take a certain joy in accomplishing. However, if one spreads out the disassembled parts, and asks how they are related to one another, only the holistic imagination, which is the creative imagination, can provide the answer. It is harder to assemble a watch than to disassemble it, just as it is harder to put a picture-puzzle together than to take it apart. Any researcher of ordinary competence can deal with the question of the particular year in which Constantine the Great was born, weighing what evidence remains for A. D. 280 as opposed to A. D. 274; but how many can embrace in one view what Gibbon called “the decline and fall of the Roman empire”? How many can see within the frame of one picture the continuous development that leads across five centuries from Constantine to Charlemagne? These holistic challenges require wide-angle vision. They require the large, all-embracing, imaginative vision that distinguished Gibbon. Yet, without such a holistic approach, without the making of connections over time and space, without the imagination this entails, history is meaningless, a chaos of unrelated items. It is meaningless if every fact stands by itself. It is meaningless as a mere laundry-list of unconnected items. J. B. Bury (1861—1927), with his insistence on history as a science only, might indeed have reduced it to this if that had been possible. (So would Lord Acton, whose projected and understandably unaccomplished history of freedom remains the greatest book never written.)

Many years ago I undertook to achieve, by reading on my own, such a mastery of medieval history as was possible for me in the circumstances in which I found myself. The authority I consulted, with this in view, made it clear that I must, to begin with, read Gibbon’s Decline and Fall of the Roman Empire—but only (and this was imperative) in Bury’s edition, which corrected the errors of Gibbon that would otherwise mislead me. So I acquired Bury’s edition, which I have before me now as I write. Then I discovered that the contribution he had made, to save readers like me from grievous misconceptions, consisted almost entirely of such footnotes as one that, commenting on Gibbon’s passing reference to the use of aromatics “in religious worship and the pomp of funerals,” reads: “But the use of aromatic spices among the Romans was by no means confined to these purposes”; or the one which reveals that Gibbon’s reference to “Albim” should have been to “Albam.” Gibbon, referring to the vices of Carinus, who ascended the imperial throne in A. D. 284, tells us: “In the course of a few months, he successively married and divorced nine wives, most of whom he left pregnant. . . .” To this Bury adds, in a footnote: “The name of one of his wives, Magnia Urbica, is now known”— just that and nothing more. His appendices show a like devotion to isolated points of insignificant detail. (What is impressive is how few and inconsequential were the corrections or additions with which he found it necessary to repair the supposedly defective historiography of the master.) I don’t mean to discount completely the contribution Bury made, but it has no bearing at all on the large picture Gibbon presents, and the large picture is what counts.

Collingwood said of Bury: “History for him, in the true positivistic manner, consists of an assemblage of isolated facts, each capable of being ascertained or investigated without reference to the others. Thus he was able to accomplish the very strange feat of bringing Gibbon up to date by means of footnotes, adding to the aggregate of knowledge already contained in his pages the numerous facts that had been ascertained in the meantime, without suspecting that the very discovery of these facts resulted from an historical mentality so different from Gibbon’s own that the result was not unlike adding a saxophone obbligato to an Elizabethan madrigal.”

History is not simply a heap of details, any more than the cathedral of Chartres is a pile of stones. Just as the individual stones are not what count in Chartres, so the details are not what count in history.

Take another analogy, that of any landscape painting, say Jacob van Ruisdael’s “Wheatfields” in New York’s Metropolitan Museum. Imagine, now, a student of painting whose method of viewing is to approach the canvas with a jeweler’s magnifying glass and scrutinize it close up, detail by detail, beginning in the upper left-hand corner and going across to the right, then repeating the operation a centimeter lower— and so on until, in this fashion, he has covered the whole painting, coming to the end in the lower right-hand corner. Although he will indeed have seen the whole painting, and in far greater detail than the average observer, he will not have seen it as a whole—which is to say that he won’t really know it at all.

A Bury may be intimately acquainted with the whole of Gibbon, detail by detail, and still not know it because he has never seen it as a whole. He may be acquainted in like fashion with the whole of Roman history, and still not know it because he cannot see it as a whole.

Every historian who amounts to anything is, first of all, a man of vision. He is possessed by a grand, coherent vision of the past that he seeks to get down on paper, a vision that embodies his own insights into the human situation and into the wellsprings of human behavior. In composing his vision for presentation, he must properly engage in the most conscientious research into all its aspects, including the details, making whatever corrections that calls for. But the vision is still primary, the research secondary.

We professors, who are producing the new generations of historians, cannot teach our students how to evoke in themselves vision and insight. All we can teach them is how to do research (which is not hard) and, incidentally, how to present their work. This means that we tend to emphasize the secondary scientific aspect exclusively, automatically concentrating on analytical detail, like a teacher of painting who addresses himself to the technique required for the individual brushstrokes but overlooks composition. And this means that we come to think of history, and to teach our students to think of it, as no more than an analytical exercise for the accumulation of verified details.

III

The analytical approach goes from the small to the smaller. But history, below a certain scale in space and time, is no longer history. A history limited to one square meter in France, even over a period of many centuries, would not be history because the area is too small to comprehend selfcontained historical happenings. A history covering the whole world, but limited to five minutes, would not be history because the time is too short to comprehend selfcontained historical happenings. The local chronicles, of a town or a parish, such as are found in local libraries or archives, may be called history, but they are on such a small scale that they lack the large significance we commonly attach to the term, Great history is not without a certain grandeur and significance that requires it to cover a large scene over an ample span of time—as in Gibbon’s history, as in Bernard Pares’s History of Russia. And this implies the holistic vision, rather than the reductionist lack of vision in terms of which historical research is conducted scientifically.

Perhaps it is inevitable that in history, as in politics, we teach only what lends itself to teaching, what is teachable, omitting everything else, however transcendent its importance. What is teachable is research, which entails the scientific, reductionist approach. But we mislead our students if we teach this as representing the entirety or even the principal element of the historian’s vocation.