Establishing Evidentiary Satisfaction 79
tween the two would continue, and that they would continue for the same reasons. What Blainey pointed out, however, was that when Russia decisively beat Turkey in 1878, matters changed. Then the power differential became evident and, as a consequence, the situation stabilized because Turkey was no longer willing to risk open conflict with Russia. Since then, just as Blainey predicted, the two have fought only as part of more general conflicts, as in World War I. Analogy, thus, works only so long as the fundamental circumstances producing the similarities remain unchanged.
All the techniques we have discussed are, it should be clear, flawed. Each, no matter how "scientific" in form, requires historians to make assumptions, to fill in blanks, to make intuitive guesses, to reason from the specific to the general. This is, in the crudest sense, the essence of historical interpretation, and such interpretation is just that—interpretation. This does not mean, however, that historians do not hold each other accountable for the way they reason. No historian's work will survive if it is based on sloppy reasoning in the judgment of colleagues in the profession. Among the many fallacies of reasoning that are guaranteed not to pass this test of plausibility are the following: reasoning from the isolated case to a general rule (all Irishmen have red hair); confusing correlation with cause (taxes were cut and the GNP rose; hence the tax cut caused the economy to grow); the use of two unconnected and unproved facts to prove a third (the Star Wars defense system planned by the United States in the 1980s would have made the country impregnable to missile attack, and the huge investments the Soviets made in their military caused the nation's economic collapse; hence, Star Wars brought about the end of the Soviet empire); and the adduction of evidence that has no established relevance to the case (X is not a good parent; hence, he will not make a good president).
B. Establishing Evidentiary Satisfaction
At what point do historians (and their professional critics) think they have enough evidence to support their arguments? Given that no interpretation is ever fully secure, when is it "secure enough"? While perfect certainty is never achievable, there are gradations of plausibility—some kinds of evidence are better than others, some kinds of interpretations are easier to support. In the first two chapters of this book, we considered the credibility of individual sources: whether, for example, the source was intentionally or unintentionally created; whether it presents data of a so-
8o
Historical Interpretation
cial bookkeeping kind that have a certain reliability because they report patterns of social action; whether the source speaks from a personal voice, as does a diary; whether it purports to relay a political judgment; whether the source relates what people were supposed to have done (i.e., a normative law) or what they did (i.e., police records). Here we want to consider the issue of evidence more generally, isolating those factors that seem to most historians crucial in judging the quality of evidence. While we can offer beginning historians no formulas for deciding "how much" evidence is enough, we hope to provide some useful aids for making that decision.
To a large extent, the amount or quality of evidence required depends on the kind of event being studied. Nineteenth-century positivist historians thought that the problem of "true" or "false" in historical interpretation was a matter of deciding whether something actually happened ("true") or not ("false"). For Ranke, true facts would have included easily establishable events such as the birth and death of Chancellor Otto von Bismarck. But there are, it hardly need be said, few events of interest to historians that can pass this Ranke an test, not even the simplest "fact" of "what happened." Let us take, for example, a question that seems quite straightforward: what did Boris Yeltsin do on the morning of Monday, August 19, 1991, in order to telephone the White House in Washington from his home outside Moscow? Still harder to establish with certainty are the "facts" surrounding the KGB's decision not to prevent his calling. Even statistical "facts" are seldom as uncontroversial as the numbers imply. Exactly how many people died in the Nazi concentration camps, and how many among them were Jews? No numerical answer given to a question like this, no matter how plausible, can ever be absolutely proved. Even if historians were to gather all the archival material in the world about this matter, they would still not be able to answer that question with full certainty, for the records are simply not good enough for that level of precision, and those that once existed often do not still survive.
"Facts" that imply knowledge of an actor's intention are considerably harder to verify and will never pass the positivist test of "true" or "false." Why did Gorbachev, in 1991, continue to believe that the Communist Party would survive, even after the putsch and the dramatic confrontation with Yeltsin in the parliament? While we can be sure that he had certain objectives in mind, and a logic that could explain his actions, we can also be sure that what he would later claim to be his motives do not constitute the whole story. Our certainty about the patterns, the causes, and the effects of general processes will be even less secure. What forces were unleashed by the 1991 putsch in Moscow? The principal problem with a
Establishing Evidentiary Satisfaction
81
question like this is that the "forces" are many—social, political, economic, and ideological—and each observer will give different weight to each, will assess each differently. Finally, there are certain events of the past that can never even be subjected to this test of true/false. An essay by Jean-Paul Sartre, a novel by Ernest Hemingway, a poem by Maya Angelou, these are all "true" in the sense that they exist and that they exist as the "truth" of their text.
Historians also apply a kind of quantitative measure in assessing evidence, in general following the principle that the more evidence, the better. But historians never have just what they want or need. At one extreme is the historian limited to one source. Einhard's Life of Charlemagne is, for example, the only source scholars have about the private life of Europe's first emperor. Like many of the political biographies written today, this one is more hagiography than critical biography, and in the best of worlds historians might well refuse to use it as evidence about Charlemagne's life and his character. But historians, although conscious that they are prisoners of the unique source and bear all the risks that this involves, use it because it is all they have. At the other extreme are historians studying the recent past. They have a great many sources, and in many ways their problems are thus fewer. But even here there is no certainty. The historian studying the American war in Vietnam has plenty of "evidence," but there are still a great many issues that remain murky, even at the level of simple event.
Oddly, moreover, it is often the case that some of the sources treating the matters historians of the contemporary period are studying have not yet become available and that there is still the chance of getting new, absolutely crucial evidence. For example, the historian who is now, in 2001, studying the peace accord negotiated in 1993 among Israel, the PLO, and the Arab world might rightly suspect that useful information about this event, such as deals made between Shimon Peres and Mahmoud Abbas, still lies secret and might someday be released. Today, however, this information is still too delicate, too raw, for public opinion and cannot be permitted to surface.
In contrast, the classical historian's problem is typically scarcity of sources, not abundance. The paradox here is that because the scarcity is irremediable, this historian has to spend Iitde time searching for or worrying about undiscovered documents. Of course, this does not mean that ancient historians or any other historians of the distant past can imagine that all they need to know on a subject is collected in archives and neatly labeled, passively awaiting their arrival and their questions. Like all historians, these scholars must decide what kinds of sources
82
Historical Interpretation
could provide relevant information and must seek it out, often moving from archive to archive, from dossier to dossier. The crucial difference between them and someone working on the recent past is, then, not so much that they have few sources and the modern historian has many; it is simply that they are much less likely to come upon previously unknown sources.
Even when all the sources available agree, even when there are a lot of sources concerning an event, even then historians cannot be sure about the facts because it is often the case that the sources available represent the opinions, or the versions of the events, of the winning party in a contest or the dominant powers in a system. From the early to the high Middle Ages (roughly from the sixth century through the twelfth), for example, almost all the sources that survive were produced by ecclesiastical institutions—monasteries, the papacy, bishops, and their officials. None of these sources had very much good to say about merchants; none of them knew much about the ordinary life of secular people or thought to say much about it. Hence, what they reveal about any of these matters has to be taken with handfuls of salt. And it is possible, as some historians have recendy argued, that the profession's reliance on these sources for views of this half-millennium has radically distorted understanding of religious life in this period, luring historians into imagining that Europe was uniformly Christian and that people lived according to well-understood Christian precepts.7 Similarly, what historians know about workers' movements before the nineteenth century in European history comes to them almost entirely from records produced by their repressors or opponents. Conservative newspapers of the day unanimously portrayed the early leaders of socialist movements in Europe as dangerous and subversive; police files repeated the story. The socialists' own voices are extremely difficult to find; scholars have only a few personal letters and diaries, a few reports of pro to-socialist movements. The situation regarding the German peasants' revolt of 1525 is similar, for it too is described largely in the words of its enemies.
Another factor to consider is the historian's personal distance from the events, for that very much determines the way he or she uses the evidence and thus its "sufficiency." In general, we would expect that the closer a historian is to events, the better he or she can judge the available evidence; thus, he or she will require less evidence than would someone
7. For a recent study arguing that the Middle Ages was in many respects more pagan than Christian, see Ludo Mills, Heidense middekeuwen (Brussels, 1991; The Pagan Middle Ages [Rochester, N.Y., 1998]).
Establishing Evidentiary Satisfaction
83
studying a more distant period. Some commentators have even argued that only participants in an event can reliably tell its history. Elie Aron Cohen, a Dutch physician who was himself in the Nazi concentration camp in Auschwitz, published a study in 1979, Be negentien treinen van So-bibor (The Nineteen Trains of Sobibor), which investigated the events at Sobi-bor, one of the most notorious camps, where certain Jews assisted in the exterminations in exchange for their own lives. For Cohen, these were inexplicable events, comprehensible only to those who had experienced them. He concluded, "someone who had not experienced [such an] event, cannot write about it," thus constructing an impenetrable barrier between those who survived the camps and those who did not. In recent years some scholars have similarly argued that only women can write women's history or that only African Americans can teach African American history.
This barrier is too high. While it is certainly true that participants in an event, immediate observers of it, and scholars who share the social identity of the group being studied have a particular—and uniquely valuable—perspective, their knowledge and understanding of it are hardly perfect. The historian who is distant from the history he or she is relating is sometimes in a better position to comprehend certain aspects of that history, even to have better evidence about the events. Such scholars have another advantage as well in that they know the aftermath, and that knowledge helps them identify the significant features of the event. Sometimes information that contemporaries could not have known will become available to later generations—secret files will be opened, stories once too shameful or too dangerous to relate will become tellable, facts that no one at the time saw as relevant will be revealed as significant. In 1971, for example, the New York Times published the famous "Pentagon Papers" leaked by Daniel Ellsberg, which, among other things, showed that the decision to bomb North Vietnam had been made long before the Gulf of Tonkin attack. None of this was known in 1964, and the historian writing between 1964 and 1971 would have told a very different story about the war than would someone studying this event thereafter. Other barriers hinder historians who are personally involved in the history they are writing. They are less able to distance themselves from the politics of the events they study, more likely to inject their own views or wishes into the story they are relating. For such historians, it is often too easy to interpret the past in terms of the shared ideology rather than in terms of the ideas and circumstances that shaped the ideology of that past. Sometimes, indeed, the historian who is not in sympathy with her subject is in a posi-