In September 2002 the British Prime Minister was considering making war upon another country, Iraq. When the House of Commons insisted upon being consulted about this, he put before them a lengthy document which, he claimed, was evidence of Iraq’s warlike intentions. A good deal of Parliamentary, as of public, opinion rejected this evidence. It was not enough to justify the attack that was being considered by the British government. Similarly, there are a number of notorious crimes (the Stephen Lawrence murder is an example) where the perpetrators are well known to the police, but they cannot be brought to justice because adequate evidence for prosecution is lacking.

So what is this thing called ‘evidence’? It is something that exists now before us; it is obvious, it is evident. In a law court objects - a weapon, a letter - are put before the jury as evidence of guilt. Why? Because they relate to the prisoner’s crime - or so the prosecutor says. They help the jury to decide about something that is not evident - that is, the prisoner’s guilt.

So it is with historical evidence: that which is evident helps us to decide about that which is not evident. Truths immediately obvious may be conveyed either by our senses (‘It is raining and my feet are wet’) or by our reason (‘The square on the hypotenuse equals the sum of the squares on the other two sides’). But historical truths are not obvious in either sense. Historical judgements, like legal and political and scientific judgements, are not immediately apparent. They have to be reached by some combination of sensory perception and rational thinking. Such judgements can range from the virtually certain to the utterly preposterous, but they can never be 100 per cent certain because they concern the past - which, by definition, is not present and not, therefore, evident.

Two things are worth remembering. One is that all the evidence there is for the past is now in existence. There can never be any more. For example, there is probably more of Roman Britain under the soil, still unexcavated, than has yet been found. This means that in the future our knowledge of Roman Britain can be greater than it is now. But the quantity of potential evidence can never increase. That would require the Romans to leave more behind in the future than they already have - a logical impossibility. But though we can never increase the evidence for the past, yet we can learn more by making better use of what remains. We may find more relics and we may become better at understanding what we have found.

The other point to remember is that almost anything can become historical evidence - that is, be used by historians to learn about the past. Post-modernist critics equate historical evidence with ‘texts’ on the fallacious assumption that all historical sources are documentary. We shall see how wrong this is in the next section. So we may conclude with two slogans for the thinking historian: all evidence is here; all here is evidence.

It is time to distinguish ‘relic’ from ‘evidence’. A relic is anything that remains from the past. Everything around us is actually a relic (except whatever has just come into existence). However, we do not usually call something a relic unless it reminds us of the past. Examples might be a garment that we bought when other styles were in fashion, or an old clay tobacco pipe left in the ground by a Victorian workman. These antiquated objects are not necessarily evidence. For that we have to ask (and answer) four questions. What is it? Of what is it evidence? For whom is it evidence? And by what means can it be interpreted as evidence?

Let us suppose that on a country walk we come across an old church and examine its doorway. Is it genuine thirteenth-century stonework or a modern restoration? (Q 1.) Is it evidence for church-building in the thirteenth century? Or in the nineteenth-century? Is it evidence of greater wealth or of greater spirituality at either of these times? And so on (Q 2). To the uninstructed it may not be evidence of any of these things - just stone round a doorway. A church architect can learn a lot from it, but probably he cannot tell whether it is evidence of wealth or spirituality. Yet it might well furnish such evidence to a social historian (Q 3). Finally, what meaning does the doorway have for the twenty-first century visitor in relation to his knowledge of the thirteenth (or nineteenth) century? The passer-by may remark the dog-tooth pattern and confidently ascribe it to the thirteenth century. He may be blind to the further details that lead the church architect to say it is a nineteenth-century restoration. Of these two interpretations one must be wrong. Whichever view we accept will have certain implications for our knowledge of one of these periods. In either case our belief should fit into the context of other beliefs about that church and that period. Therefore, an interpretation rests on the context of what the interpreter already knows (Q 4).

We see, then, that evidence forms the essential bridge between the present and the past. Sir Geoffrey Elton maintained that two questions only are fundamental to historical method: ‘Exactly what evidence is there? And what does it mean?’ These are indeed sufficient for historical knowledge or belief. But if the historian is to express his conclusions to others, or even to formulate them to himself, then there arise questions of language and literary form. Indeed, such questions can arise even earlier; namely in the written evidence itself. But questions of interpretation and narrative construction take us too far from our present concern with evidence.

It is essential to remember that historical knowledge is always uncertain, for its necessary twin bases - evidence and interpretation - are both quite fallible. The Oxford Companion to Philosophy defines evidence as:

‘That body of belief, often of an observational sort, which supports some less well-established hypothesis.’
It goes on:
‘Doubtless the wise man should apportion his belief to the evidence he has, but …’

Indeed, there are many ‘buts’, of which one is that in daily life we are virtually compelled to act on beliefs for which our evidence is inadequate (catching a train or plane, or buying groceries, for example). This is a comparatively modern attitude. Ancient philosophers saw a crucial distinction between Being and Becoming: that is, they believed that the world of Being is eternal and unchanging (like God and mathematics and logic), while the world of Becoming is volatile and uncertain, fickle and unreliable (like the weather). Thus, knowledge is possible only of the certain world of Being. Of the kaleidoscopic world of Becoming one can have nothing more than opinion. During the seventeenth century (roughly) such views were replaced by the modern outlook. This believes it possible to build up bodies of pretty reliable knowledge that can form a basis for action, even though they fall short of complete certainty. This is the foundation of two great achievements, modern science and modern history. These apparently solid bodies of knowledge are ever changing, ever enlarging and (on the whole) improving. But neither ever reaches absolute certainty. In all these respects it is wise to proportion our beliefs to our evidence. Nevertheless, in science, in law, in history and archaeology, evidence is fundamental.

One must never forget that a piece of evidence may have several interpretations, including some that have never occurred to you. Collingwood insisted that the historian may only assert ‘what the evidence before him obliges him to assert.’ (R.G. Collingwood, 1946, p 204.) But historians have often taken that obligation too readily and overlooked the possibility of other interpretations.

Before the Second World War virtually the only historical evidence available to a school student was the text book. Some 30 or 40 years ago enlightened teachers began to introduce their pupils to documents contemporary with the events studied, so that they could see how things seemed to people at the time. Thus, we have the two main divisions of historical evidence: primary sources, whose origin is contemporary with the period in question; and secondary sources, which are accounts written by later writers (sometimes much later).

Some suppose that primary sources, being strictly contemporary, are the unprocessed raw material of history. This can be misleading. The most important documents of the last five centuries are usually official papers, ordered and catalogued in state or local archives. The originals present many problems to the researcher - apart from those of access in a remote city or country. They include difficulties of handwriting and language. Scholarly works on European history can easily require a mastery of four, five or more languages. Unfortunately, the necessary work of the archivists can, at times, lead the researcher astray. Documents relating to a particular department, policy, person or project are filed together. Sometimes a revealing paper lacks one or more of the normal indications of origin: that is, name of sender, place of writing, date, name of addressee, subject or topic. The archivist has to guess and put it somewhere. If he or she is wrong, then pertinent information is thenceforth missing from another file, while misleading ‘facts’ are incorporated into the context in which it has been placed. Then again, the document may be false in some respect. It may be intended to deceive and so not emanate from the signatory, or not have been sent on the attached date, and so on. The researcher, who may not be a technical expert on inks, papers, handwriting, etc, has to make up his own mind about these things or seek what help he can find. For there are a number of ‘auxiliary sciences’ that the researcher may need. They include palaeography, diplomatics, legal forms, heraldry, place-names, linguistics, coins, seals and medals. We have remarked earlier that the first question to be asked of a possible piece of evidence is, What exactly is it that I have before me? This is not nearly so easy to establish as the novice may at first suppose, and if he is not himself skilled in such things he must enquire of the expert.

As well as being arranged by the archivist in boxes, bundles or files, official papers are frequently calendared. A ‘calendar’ is a printing of the documents from a single source (often a particular department) in strict chronological order. The texts are either printed in full, or (if too lengthy) shortened or summarised. State Papers of the Reign of Henry VIII, State Papers Venetian, Ordnances of the City of Bristol are examples of sixteenth-century English sources calendared in the nineteenth or twentieth centuries. They stand, as it were, between true primary and secondary sources. They are certainly contemporary in origin, but they are by no means unprocessed. They are subject to all the possible errors and misunderstandings described earlier. The biggest danger, however, comes not from the fallible archivist, transcriber or printer, but from what is not there at all. It is almost impossible to be sure that all the relevant documents have been assembled in one place. There is always the possibility that one or more papers have been omitted that would lead to a very different picture of the whole.

The historian can use only the evidence that has survived, but exactly what survives and what does not is often a matter of chance. Until 1947 the papers of the Government of India were kept in Delhi. When India was divided and Pakistan became a separate state, it was arranged that all the govermnent papers that dealt with that part of India which had now become Pakistan should be conveyed from Delhi to the new capital. But the railway system broke down amid the mutual Hindu-Muslim massacres that cost three-quarters of a million lives. Some time later it was among the wreckage of this disaster that the sacks of documents from Delhi were sought - and in many cases not found. Thus, the sources in Islamabad for the history of Pakistan are much less adequate than those for modern India which have remained in Delhi.

Nevertheless, for all their possible shortcomings, calendars of documents are invaluable to the researcher. They save him or her the time and trouble of reading thousands of documents that turn out to have little or no interest for that particular study. Several life-times of archival labour are saved.

Secondary sources are even more ambivalent. For the historian it is as true as of some other necessities that ‘you cannot live with them; you cannot live without them.’ Herbert Butterfield in Man on His Past, a study of historical scholarship, demonstrated how historians of the Seven Years’ War were long misled into following the erroneous assumptions of their predecessors. As A.J.P. Taylor remarked, ‘History does not repeat itself, historians repeat each other’. On the other hand, before the historian approaches a subject it is essential that all is read that has already been written on it. Only a wide reading of the secondary sources will enlighten a reader to the different ways in which the events can be seen. The historian must then consider whether he or she is likely to add anything useful before proceeding with the researches.

The use of different kinds of evidence goes with changes in the writing of history. Once the official archives had been opened, several generations of scholars trawled through them to write (mostly) political history. The emergence in the late nineteenth century of the study of economic history led to a widening of the sources needed - though some useful material was still available in the archives already worked for political history. That just meant other people reading the same documents for a different purpose. But mostly completely fresh sources had to be sought.

For the early Middle Ages we have sparse royal or ecclesiastical documents relating to lands or buildings given or sold, often with details of their economic organisation. Perhaps the most remarkable of these is the Domesday Book of 1086. For the later Middle Ages economic sources are both fuller and more varied. They are the consequence of two phenomena: the organisation of royal governments and the rapid development of towns and trade. They produced quantities of tax and customs returns at both royal and local levels. Ecclesiastical documents, mostly from cathedrals and monasteries, also multiplied. To these official documents can be added a steady flow from private sources, notably from banks and other business houses. The invention of printing and the expansion of Europe overseas (almost at the same time) turned the flow into a flood.

In the twentieth century social history grew and expanded with the social sciences, as did new specialities like historical demography or landscape history. Most interesting, though, has been the development of cultural history. This is because historians’ attention is here focused upon the inner lives of men and women - their loves and hates, fears and hopes, beliefs and assumptions. A classic example is Keith Thomas’s Religion and the Decline of Magic (1971). In this sort of work new problems of evidence arise. It is obviously more difficult to find reliable evidence of what men thought than of what they did or said. Fortunately, for the historian, the Inquisition kept scrupulously detailed records of trials. They have furnished the material for two illuminating studies of medieval and early modern beliefs: Emmanuel Le Roy Ladurie’s Montaillou (1978) and Carlo Ginzburg’s The Cheese and the Worms (1976).

Nowadays historians can make use, it seems, of almost anything that has survived from the past, including crop patterns, hedgerows, ruined and rejected machinery, folk-songs, myths, memories, blood-groups, speech patterns, parish registers, land surveys, account books, place-names, sunken wrecks, and even pollen counts for the history of climate.

To sum up, we have seen what evidence is, what things can constitute it, and some of the problems and difficulties that arise in its use. People often complain of a lack of evidence. And it is true that we could do with a good deal more, especially for ancient and early medieval history. If you want to experience at first hand some of the historian’s problems, just try to write a history of your family over the last 50 years, insisting that there should be one piece (preferably two pieces) of firm evidence for every statement that you make. Even given a lack of evidence, a great deal more can be done than is often supposed. One must remember that among the greatest limitations on evidence are lack of time and lack of imagination.

Finally we conclude, as we began, in the year 2002. In the Times Literary Supplement of 10 May 2002 Imogen Gassert drew attention to a new source for our understanding of the British Army in the First World War. That source was the reading habits of its soldiers - especially its officers. Gassert writes:

The notion that, after the carnage of the Somme and Passchendaele, they were thoroughly disillusioned with the conflict (in contrast to the belligerent spirit of those
at home) is flatly contradicted by their choice of reading
matter (p 17).

The evidence has been apparent and available for nearly a century. It consists of ‘commercial archives, personal and public records, and the literature - especially the novels - published while the war was going on’. (Ibid.) Apparently, our accepted picture of the experience of that war has been derived too much from anti-war writers like Siegfied Sassoon, Wilfred Owen and Robert Graves. It is clear that we shall have to re-assess our views of the inner lives of those great armies - notably as portrayed in Paul Fussell’s influential work, The Great War and Modern Memory (1977). The whole issue, too complex to go into here, will be of interest to those many teachers who run history or English courses on the literature of the First World War. Here two points may be made. One is the reminder that evidence can lie unseen right under our eyes. Yet until it is perceived as evidence, it is as much hidden as if it lay underground like those Roman remains. The other illustrates one of the most important and yet most difficult kinds of history - the history of the inside of men’s minds. Incidentally, Tolstoy dwells on this in the battle episodes of his classic, War and Peace. Thus, what has long been the central concern of novelists has now also become a major concern of historians. Questions of evidence are now even more intriguing.

Dr Michael Stanford, who has degrees in both History and Philosophy, is the author of The Nature of Historical Knowledge (1986), A Companion to the Study of History (1994) and An Introduction to the Philosophy of History (1998).