FOR ONLY $129 YOU CAN TAKE an online course from IAP Career College, at the end of which you will receive a certificate designating you as a “lifestyle expert.”1 Relatedly, the actress Gwyneth Paltrow also designates herself as a lifestyle expert, though it is not entirely clear what makes her a lifestyle expert. Or for that matter what makes anyone a lifestyle expert. What do lifestyle experts know that the rest of us do not? Each of us, after all, has a lifestyle. And those of us who more or less like our lifestyles might even think we are, at the very least, expert about our own lifestyles.
It is easy to mock lifestyle experts, but their very existence suggests that there are all sorts of experts, most of whom are not, and do not purport to be, scientists of any kind. Although the forensics experts discussed on Chapter 10 vary in the scientific rigor of their methods, most of them at least purport to be doing something close to the science we explored in Chapter 9. Not so, however, with those whose expertise is in music appreciation, art authentication, body language, wine tasting, literary criticism, and various other endeavors of greater or lesser respectability. What remains to be addressed, therefore, is the question of expertise as evidence where the very nature of the expertise is far more elusive than it is in the case of scientists.2
Consider, for example, chicken sexing, a topic that most of us have not actually ever considered.3 But the practice really does exist, and it is genuinely important. Although it is not particularly difficult to distinguish a male from a female chicken once the chicks are more than a few weeks old, that turns out to be too late. Given the nature of the modern chicken-raising and egg industries, determining the sex of a chick almost immediately after hatching is crucial to the success of the industries and the individual companies involved in those industries. And so there has been the emergence of a cadre of expert chicken sexers.4
Two things are important about chicken sexing. One is that it works. Experienced expert sexers can apparently get it right about 98 percent of the time, although novices do barely better than 50 percent. In other words, novices are no better than random, but experts are a great deal better. And the second thing is that most of the experts cannot provide a very precise description of just what method they use to achieve this 98 percent success rate. And even if they could provide some description of their methods, it is unlikely that they could give a satisfactory analysis of why those methods work. In the same way, we suspect that even the most successful baseball pitchers do not have much understanding of the physics of curveballs, even though they know that the ball curves and know how to make it curve.
As we have seen, one form of evidence that occupies much of our evidentiary universe comes from the testimony and conclusions of experts. We know there is climate change because scientific experts, in ways that most of us cannot comprehend, have determined that there is and have announced that determination publicly. We know that vaccination works, and with few side effects, again because of the work of scientists, who then report their conclusions to the rest of us. And so too with the safety of GMOs, the flight control system of a lunar probe, the efficacy of radar, and, for most of us, even the physics of the internal combustion engine that moves the automobile taking us from one place to another. And we saw in Chapter 10 that many of the contemporary controversies surrounding forensic evidence, such as the validity of handwriting identification, focus on whether such evidence is scientific in the same way that climate science, rocket science, and brain science are scientific.
Other forms of expertise do not look like science and do not purport to be science. Chicken sexing, for example. And yet they work. The challenge, accordingly, is to consider the evidentiary status of such forms of expert judgment when they claim to provide us with information and thus of evidence. If a chicken sexer says that this chick is a male, is that evidence that the chick is indeed a male? And how good is that evidence? This is the question to be addressed, with the benefit of examples that may be less obscure and more useful to most of us than chicken sexing.
One of the best-known examples of qualitative, impressionistic, nonscientific, but nevertheless highly respectable expertise is art authentication. In January 2021 a painting entitled Young Man Holding a Roundel, by Sandro Botticelli, sold in a Sotheby’s online auction for $92.2 million dollars.5 That is a lot of money for a painting. Or for anything else, for that matter. And much of that amount was a function of the fact that this was by all accounts a genuine original Botticelli actually painted by Botticelli. Had it been a really good reproduction—one that accurately reproduced the colors and even the brushstrokes and the surface texture of a Botticelli, and that looked authentic to most laypeople and maybe even to some art professionals—its value would have been measured in the hundreds or maybe the thousands but not in the tens of millions. Even if it had been a genuine original oil painting, its value as “attributed to Botticelli, “studio of Botticelli, “circle of Botticelli,” “follower of Botticelli,” “manner of Botticelli,” or “after Botticelli” would have been but a small fraction of the price that was actually paid.6 Similarly, but with even less value, had it been a hand-painted actual oil painting that was a forgery, no matter how indistinguishable the forgery may have been from a genuine Botticelli.
The $92.2 million figure for this Botticelli may be extreme, but it illustrates the fact that a great deal of money, in addition to pure historical or academic interest, turns on the ability to authenticate a painting as being a genuine, original work of the artist who is claimed to have painted the painting. Or sculpted the sculpture. Or drawn the drawing. And therefore we have experts in art authentication. The question, then, is what makes them experts, and why their expert conclusions should have such great evidentiary weight.
Before turning directly to the question as just posed, a few clarifications are in order. First, and perhaps most importantly, we need to distinguish authentication from evaluation. There are art experts who purport to be able to explain why Claude Monet’s Water Lilies is a better painting than the same artist’s Woman with a Parasol, and why Monet is a more important and better artist than his fellow impressionist Armand Guillaumin. But these judgments are different from art authentication, for two reasons. First, these judgments are heavily infused with questions of taste. Monet being a better painter than Guillaumin is, say the experts, not simply a matter of taste in the same way that vanilla fudge ice cream being better than chocolate chip is a matter of taste. But taste and preference are relevant to comparing artists or works of an artist in a way that the authenticity of a work of art, at least in theory, is not. For authentication, there is a fact of the matter. Experts might differ in their conclusions about authenticity, but they agree that what they are attempting to determine is whether, in our earlier example, an actual person named Sandro Botticelli actually applied paint to create a particular painting on a particular canvas. Even when the evidence for what the fact of the matter is might be inconclusive, that evidence is marshaled for or against a conclusion about a fact that is not itself uncertain. Botticelli either painted this painting or he did not.
A good contrast to fact-focused art authentication is wine tasting. Although (some) wine experts can identify a wine’s grape variety, region of origin, and, sometimes, vintage year in a blind tasting, and although the existence of an actual fact of the matter about a wine’s origin and production makes wine identification similar to art authentication, most of what wine experts report, usually with the assistance of uninformative adjectives such as “shy,” “naive,” and “authoritative,” are their own tastes and preferences, albeit tastes and preferences informed by extensive experience.7 Still, the typical wine evaluation by a wine expert is not the identification of an actual fact like the painter of a painting or the sex of a chicken. And in that sense art authentication by art experts involves the attempts by the experts, on the basis of evidence, to identify some fact about the world.8
An expert judgment about the authenticity of a work of art, therefore, can be measured, at least in theory, against the actual authenticity of the work. Actual authenticity is the object of art authentication, and it is to be distinguished from the way in which the monetary value of a work of art may be a function of the authenticating judgment—the evidence—of the experts. The dollar value of a Monet compared to a Guillaumin may be at least somewhat a function of what the experts have said just because it is the experts that are saying it, but authentication is different. Perhaps an authentic Botticelli is valuable as authentic because all of the Botticelli experts agree that it is an authentic Botticelli. But if it turned out that the expert judgment was mistaken, as with the Han van Meegeren-created “Vermeer” discussed in Chapter 2, the value would all but disappear.9 For authentication, even if not for valuation, the expert judgment is evidence of something else—authenticity—and rarely, and never conclusively, an end in itself.
As a final qualification, we need to set aside issues about the distinction between authorship and authenticity. Especially these days, even the authentic artists may not be the ones who physically create the objects properly attributed to them. Prominent contemporary artists such as Sol Lewitt, Damien Hurst, and Jeff Koons often—perhaps typically—sign their names to objects fabricated solely by the assistants in their studios. But if the assistants create an object under Lewitt’s supervision, and Lewitt then signs that object or acknowledges in some other way that it is his, that object is then an authentic Lewitt even if we might have a philosophical debate about what it means for this to be a Lewitt. That is an interesting and important issue, but it is not an issue about evidence, and it is not our issue. Here we are concerned only with distinguishing the authentic Lewitt, or Botticelli, or Rembrandt, from copies, fakes, forgeries, and misattributions. And for this issue the question of evidence is of crucial importance.
These qualifications out of the way, we have arrived at the heart of the matter. We are interested in the truth of the factual proposition that an actual human being named Sandro Botticelli in or around 1480 painted the painting now known as Young Man Holding a Roundel. We are not interested in whether this is a better painting than a triptych by Francis Bacon that sold for more than $80 million earlier in 2020, or in whether the Botticelli was “really” worth $92.2 million dollars. All we care about is whether Botticelli painted it.
At this point the role of experts takes center stage.10 And here there are similarities to and differences from the kinds of expertise we have already considered. It would be nice if, for art authentication, we could do the same kinds of tests that the best of forensic identification experts and their organizations are conducting these days, even if they have long resisted it, and even if it took the Supreme Court ruling in Daubert v. Merrill Dow Pharmaceuticals, Inc. to get them to change their ways.11 Following what the best of the forensic scientists in the best of the forensic sciences are now doing, and following as well what real scientists in their laboratories have long done, we could ask Botticelli to give us, say, fifty paintings, each accompanied by his personal sworn and polygraph-tested testimony that he in fact painted them. And then we could commission a group of accomplished art forgers to paint fifty paintings in the style of Botticelli, and then ask the forgers to swear, again with state-of-the-art polygraph testing, that they and not Botticelli painted them. We would then give the hundred paintings, unidentified as to their authenticity but whose authenticity was known to the tester, to the alleged Botticelli experts and ask them to tell us which are the real Botticellis and which are the forgeries. And when all of this was done, we would have, both for each expert and for the community of experts, a measure of their reliability as identifiers of positives (painted by Botticelli) and negatives (not painted by Botticelli). This would yield the error rate of false positives and the possibly different error rate of false negatives.
Of course, this is a fantasy. Botticelli has been dead for a very long time. Moreover, he did not paint a sufficient number of paintings to provide a large enough sample size. And so forth. But if we think of this as a thought experiment, it tells us something precisely because of its fantastic character. It tells us that the expertise of the Botticelli expert is of a variety different from that of the forensic expert. In part the expertise of the art authenticator resembles the expertise of the physician whose confidence in her diagnoses and treatments comes from long experience. But even there the physician knows that the patients she diagnoses in a certain way and treats in a certain way get better. And she knows that some of her diagnoses have turned out to be mistaken, because the symptoms did not improve, although they did when she shifted to an alternative diagnosis. But this approach does not appear applicable to the Botticelli expert.
So what makes a Botticelli expert an expert? What kinds of evidence do art authentication experts use to reach their conclusions? And once a Botticelli expert has reached a conclusion, what is it that warrants the rest of us in treating that conclusion as evidence of a painting’s authenticity—as a genuine Botticelli?
A considerable amount of the art authenticator’s expertise the lies in an encyclopedic knowledge of the lives of particular artists at issue. Botticelli lived in certain places at certain times, and a painting of some location in a particular year during which Botticelli lived far away would be some evidence of the painting’s lack of authenticity. And so too if the painting represented a subject alien to what the experts know about the artist.
More significant is the question of style. The color we now know as Titian red is so named because Titian characteristically used it in his paintings, and so the presence or absence of that color would be evidence for or against its authenticity as a genuine Titian. The same is true for Vermeer’s distinctive blue and the very different but equally distinctive blue created by Yves Klein and produced by a process that Klein had patented. In addition, experts in the style of a particular artist know the form of the brushstrokes, what the brushstrokes say about the brushes used, and a great deal more. This great deal more, however, does not, for most experts, include the artist’s signature. That may seem surprising, but not so much on further reflection. Forging a Salvatore Dali painting, to take the example of a frequently forged artist, is quite difficult, even if it is far from impossible. But forging Dali’s signature, even as characteristic as it is, would be well within the capacity of even a novice forger.12 Once we get beyond the signature, there are so many characteristics of the style of a particular painter that the evidence relied on by the experts is substantially the evidence that is based on knowledge of these characteristics. Just as someone running out of a bank wearing a ski mask and carrying a bag is evidence that that person robbed the bank because people so clothed and so behaving are usually people who have robbed a bank, a particular style of brushstroke is evidence that the painting was by, say, Botticelli, because of the authenticator’s generalized knowledge that Botticelli usually used brushstrokes of this style and most other painters did not. And although this might hardly be conclusive, it is an example of the type of evidence that, combined with other evidence of characteristic style, will lead the experts to the conclusion that the work is authentic—or not.
Less crude, or at least less subjective, are the modern techniques of genuinely scientific inspection of a painting. X-ray analysis, carbon dating, fiber analysis of the canvas, infrared reflectography, mass spectrometry, chemical analysis of the paint, and much more can tell us when (and sometimes where) a painting was painted, and although the typical Botticelli experts may not know how to use such techniques, they know the people who do and where to find them.13 Moreover, microscopic examination can reveal, in ways that visual inspection might not, such things as whether the particular style of brushstrokes or paint application did or did not resemble the painter’s distinctive style. And there is a great deal more of this variety.
Several things are important about this kind of expertise. One is that it is subjective in the sense of involving imprecise judgment.14 A brushstroke style is likely to be more or less like Botticelli’s rather than clearly his or clearly not. Second, much of this is expertise is about knowledge rather than technique. Although the line between the two is fuzzy, there is a difference between being an expert on Botticelli and being an expert skier. Or even an expert painter. And third, and often most importantly, an attribution of authenticity is likely to be an aggregation of multiple, scattered, and uncertain pieces of evidence rather than being the product of one piece of evidence that is conclusive one way or another. Indeed, experts in art authentication, just like expert chicken sexers, sometimes cannot explain just what it is that leads them to their conclusions. For the expert art authenticator, these multiple factors are often accumulated and described under the mysterious label of “connoisseurship” and the equally mysterious “eye” of the expert.15
Once we appreciate both the subjectivity of individual items of evidence of authenticity and the way in which an attribution is the result of a subjective accumulation of multiple subjective determinations, we can also appreciate that the conclusion to be drawn from these multiple items of evidence is itself subjective. Fingerprint examiners have worked out, and have used until recently, detailed protocols for determining when there are sufficient points of similarity—typically between twelve and twenty, depending on the laboratory and the jurisdiction in which it is located—between two fingerprints to be able to declare that one fingerprint matches another.16 But few analogs exist in the world of art authentication. Moreover, some of the subjective determinations might have more evidentiary value than others, and so it is not merely a matter of counting the similarities and the differences. One Botticelli expert might attribute a painting to Botticelli on the basis of eight authenticating pieces of evidence and only one relatively insignificant outlier, but another might be satisfied neither with the number of items of positive evidence nor with the alleged insignificance of the outlier, thus leading that expert to be unwilling to declare the painting authentic. Moreover, the final verdict of an individual authenticator will also be a function of that authenticator’s own burden of proof. One authenticator might be looking for something close to proof beyond a reasonable doubt, while another would be satisfied with the equivalent of clear and convincing evidence.17
Given all of this subjectivity, it is not surprising that the experts will frequently disagree. And then the question for the “consumer” of an attribution—the museum, art historian, auction house, purchaser, editor of a catalogue raisonné, and so on—is what to make of the disagreement. Presumably the consumer will have an also subjective and largely sociological sense of who the best experts are and whose opinion is entitled to less evidentiary value.18 But there will also be judgments—again subjective and even assuming that the experts have equal evidentiary weight—about how much agreement is necessary and how much disagreement is fatal.
A good example of the latter comes from a recent dispute about the authenticity of a painting by the mid-twentieth-century modernist painter Florine Stettheimer. As reported by the New York Times, the painting was to be offered for sale by Skinner’s, a Boston-based high-end auction house.19 But although several Stettheimer experts had declared the work authentic, another Stettheimer expert concluded that neither the subject nor the style were compatible with Stettheimer, leading this latter expert to declare the painting inauthentic. Skinner’s then withdrew the painting from the sale, providing as a reason the “difference of scholarly opinion.” Given that, according to the Times, multiple Stettheimer experts were satisfied with the disputed painting’s authenticity, the implication is that, at least for Skinner’s, anything short of unanimity would be insufficient to satisfy them. Skinner’s, after all, has its own reputation to think of, and it is not implausible to suppose that they operate under their own version of the precautionary principle discussed in Chapter 3. But that leads us back to our recurring issue of the relationship between the amount of evidence required—the burden of proof—and the consequence of that requirement having, or having not, been met. The very fact of disagreement may have been disqualifying for Skinner, but perhaps it would not have been so for collectors deciding how to label the paintings on their walls, and perhaps even not for a museum. If subsequent evidence proves a museum wrong in its attribution and labeling, it simply takes the painting down, or relabels it according to the now-less-certain attribution.20 But if a prominent auction house (or gallery) sells a painting subsequently shown to be inauthentic, the reputational (and financial) consequences may be much more severe.
What emerges from all of this is the irreducibly subjective judgment involved in the conclusions of an art authenticator, and in the decisions by the consumers of such authenticators, about whether to accept an authentication, and for what purpose. But as Judge McKenna said with respect to handwriting analysis, subjective does not mean worthless. Perhaps art authentication, like handwriting identification, is imprecise, subjective, open to interpretation, vulnerable to disagreement, and undeniably unscientific. But evidence that is the product of such imprecise, subjective, interpretive, and unscientific judgments can still have value as evidence, even if that value is less than the evidentiary value of a properly designed controlled laboratory experiment. Moreover, as the example of art authentication is designed to illustrate, there can still be people who are experts in areas other than science and whose expertise is not scientific. Where and when such expert judgments are reliable—whether they should be treated as evidence by others, and, if so, of what weight—is typically the result of acceptance of those judgments over time. Treating acceptance as the measure of reliability, however, has much the same odor of circularity as it does for traditional forensic science. But just as with traditional forensic science, authentication evidence for which the assurance of validity is only that the judgments have survived the test of time and that they have not been falsified in an obvious way may still, in some contexts and for some purposes, be the best we can do. And often the best we can do is good enough.
The evidence that art authenticators use is evidence about the existence—or not—of some concrete historical event. Did a man named Sandro Botticelli paint Young Man Holding a Roundel in or around 1480? Did Johannes Vermeer or Han van Meegeren paint Christ at Emmaus?21 Once we see that controversies about the authenticity of works of art are historical controversies—What did a particular individual painter do or not do at a particular time and place in the past?—we enter into the much larger world of controversies about historical events, about the evidence deployed as weapons in those controversies, and about the conclusions of historians and others as evidence. Did Lee Harvey Oswald have assistance in assassinating President John F. Kennedy?22 Did Nicola Sacco and Bartolomeo Vanzetti murder a paymaster and a guard in an armed robbery of a shoe company in Braintree, Massachusetts, in 1920?23 When did the Maori first arrive in what is now New Zealand?24
As with our discussion of art authentication, a few preliminary clarifications and qualifications are in order. First, the discussion here will focus on what we might think of as questions of hard singular fact. “Singular” in the sense of one act or event and not an aggregation, and thus different from the also-factual claim that “industrialization accelerated rapidly in late nineteenth-century Britain.” And “hard” in the sense that a contemporaneous observer would have been able to say with a high degree of certainty that the event did or did not happen. By contrast, many historical questions and controversies are about factual issues that resist straightforward answers, and are substantially interpretive, such as questions about the causes of World War I or whether lead pipes led to the decline of the Roman Empire. Even interpretive questions presuppose certain facts that are then interpreted, and therefore issues about historical interpretation do involve determinations of the hard facts of the past. But to keep first things first, we first have to know what, concretely, happened—and for such questions the issues of evidence about hard singular facts are both preliminary and essential.
Historians are also concerned with issues that are more are more normatively evaluative. One prominent example is the truth or falsity of the controversial claim in the New York Times’s 1619 Project that the preservation of slavery was “one of the primary reasons” for the American Revolution, a claim that attracted objections from a group of distinguished and otherwise sympathetic historians.25 Again, the controversy was substantially about the interpretation of large themes in American history, even as those offering competing interpretations each premised their interpretations on more concrete facts. So too with historical questions that involve counterfactual speculation, such as the question whether the sinking of the Titanic could have been prevented if nearby ships had paid better attention, or whether less self-confident physicians could have saved President James Garfield after he was shot by a deranged assassin.26
Interpretive, evaluative, and counterfactual historical questions, which are very much the stock in trade of academic historians, undoubtedly involve evidence, but they also—especially the interpretive and evaluative questions—involve subjective judgments of importance and value. Other questions are more purely factual. Trivially, many people believe that George Washington’s false teeth were made out of wood, but there is a factual answer to the question whether they were, however elusive that answer might be to us now.27 The inconclusiveness of the evidence may make the fact of the matter elusive, or at least it did in the past, but there still is a fact of the matter that, in theory, makes some hypotheses about the makeup of Washington’s teeth right and others wrong. Somewhat more consequentially, we do not know whether Anne Boleyn had an adulterous relationship with her own brother, that being one of the charges that led to her beheading in 1536.28 And more consequential yet is the claim that there were no people living in what is now northern South Africa when the first Dutch settlers of that area—the Voortrekkers—headed northward from what is now Cape Town in the seventeenth century, a claim that was long part of the apartheid narrative and justification, and has long been resisted, with strong evidence, by anti-apartheid forces.29
All these questions, and countless more, involve the use of evidence. We cannot ask George Washington what his teeth were made of, nor interrogate Anne Boleyn about her sexual practices, nor literally observe the terrain of southern Africa in the seventeenth century to see if there were any people there. When we do not have the relatively primary information that comes from first-person accounts or observations—information that is itself not necessarily certain, as we saw in Chapter 8—we are forced to rely on evidence and inferences from often uncertain and conflicting evidence as we attempt to answer such straightforward factual questions.
Although some philosophers of history have argued that historical factual inquiry is different from other sorts of factual inquiry, it is hardly clear that this claim is correct, as other philosophers of history have insisted with equal vehemence.30 For one thing, we are accustomed to using evidence to reach conclusions across space, which may not be fundamentally different from doing so across time. I did not live in 1914, but nor have I ever been in Uzbekistan, and thus there seems little cause for supposing that knowledge, or at least my knowledge, of what occurred in 1914 should be thought of as different in kind from knowledge about Uzbekistan. In both cases I must rely on testimony, and perhaps on testimony through multiple iterations. That testimony might of course be mistaken, and multiple iterations may increase the possibility of error, but the evidence that we use to learn about spatially remote events is not fundamentally different from the kinds of evidence we use to learn about temporally remote ones.
Indeed, much the same applies to any form of inquiry other than direct observation. Even if we assume that direct observation and memory are infallible, which of course they are not, as we discussed in Chapter 8, most of our knowledge comes from some form of testimony. And so too does historical knowledge. We know about the medical treatment (mis)administered to a dying President Garfield because of what people who were there said, because of what the people who heard what they said then reported to others, and so on for the past almost 150 years. Historical inquiry, at least and especially about singular facts, is fundamentally a process of evaluating testimony and evaluating hearsay. But that is hardly unique to the events of the distant and not-so-distant past.
These days historians increasingly use a battery of modern and sophisticated methods, including computerized analysis of texts and large data sets and techniques of cutting-edge natural science such as carbon-dating of artifacts and chemical examination of fibers and even bodily emanations.31 And especially in the use of these and other scientific approaches, the techniques of the historian increasingly resemble even more at least some of the techniques of the art authenticator and the forensics expert. But, again, there is little about the pastness of the past that makes the evidence used by historians different in kind from the evidence that is used whenever we have no immediate access to the facts that interest us. Few of the methods of historical research—including reliance on testimony, the use of multiple hearsay, the examination of documents, and the analysis of physical objects, for example—are different from the way in which evidence is used in court. Trials, after all, also involve reconstructing events that are not happening in real time.32 And the evidence that courts use to make up for this epistemic distance is in many respects of the same variety that historians use to make up for epistemic distance, even if that distance is typically greater for the historian than it is for the judge and the jury.
At bottom, therefore, the building blocks of historical inquiry, even interpretive and evaluative history, are the singular facts that historians access and identify by using evidence. Some evidence might be strong and some might be weak, and conclusions from the evidence are typically the product of multiple items of evidence. Just like verdicts in court. Just like the factual conclusions of investigative journalists. And just like the everyday factual conclusions that we reach regarding contemporary acts and events.
There is one respect, however, in which historical inquiry is different, at least according to some influential commentators on historical methodology. Unlike everyday assessment of evidence by ordinary people in their daily lives, unlike most of the assessments of evidence in public policy settings, and even unlike the assessment of evidence by judges and juries, some of these influential voices about historical methods, often using the label “source criticism,” prescribe relatively specific rules for the weighting and prioritization of items of evidence.33 Physical objects are to be preferred to testimony, for example. Testimony is given greater weight if it is closer in time to the described event. Eyewitness testimony is to be preferred to other forms of testimony. And in some ways these “rules,” softer or harder depending on the author and on how they are used by the historian, resemble the “two witness” rule required by the US Constitution for prosecution of treason and the “two source” rule often employed in much of traditional print journalism.34
Although these prescriptions, and other like them, are not different in kind from what people normally do, and not that different from how judges and jurors normally evaluate evidence, what is noteworthy is the attempt to reduce these ordinary principles of evidentiary evaluation to explicit guidelines, or even rules. It is not surprising that these guidelines often come from commentators in countries whose legal systems follow the civil law and not the common-law tradition. For although common-law judges and juries are rarely in modern times given rules about how to evaluate or weight the items of evidence that are admitted for their consideration, the civil law, following its Roman law origins, is different. Civil-law principles of evidence often include just these kinds of rules of weighting and priority, rules that nowadays are alien to the common-law universe.35 And so just like civil-law judges (rarely are there juries in the civil-law world) are frequently expected to follow official rules for the evaluation of evidence, historians are being urged to do much the same thing, even if, obviously, less officially.
This brief excursion into the methodology of the expert historian has larger lessons. Perhaps what makes the historian an expert is knowledge (knowing what) about a particular area or period or person, as opposed to a distinctive approach (knowing how) not shared by others. Good historians use evidence the way good judges and good thinkers use evidence, and that is to be applauded. But when we think about why and when we should treat the judgments of historians as evidence for the truth of the conclusions they reach, we may find that the basis for treating those conclusions as evidence is more similar than we might have supposed to the basis for treating any other source of knowledge as evidence.