"Real historical writers probe factual uncertainties, but they do not invent convenient facts and they do not ignore inconvenient facts. People are entitled to their own opinions, but not to their own facts."
-- William Kelleher Storey, Writing History
When the Playboy interview with Michael Bellesiles appeared in January 2001, the Emory University history professor was riding high. He was basking in the heady glow of rave reviews and a media blitz hailing his book, Arming America: The Origins of a National Gun Culture, as a "tour de force" that "changes everything." Bellesiles claimed to have discovered that, contrary to accepted opinion, guns were scarce in early America and Americans were uninterested in owning them. "The notion that a well-armed public buttressed the American dream," he assured readers, "would have appeared harebrained to most Americans before the Civil War." It was all an "invented tradition," with historians joining "actively in the mythmaking." Reviewers of Arming America were quick to point out the "inescapable policy implications."
The Playboy interview ended with Bellesiles' challenge to the powerful gun lobby: "As for the NRA [National Rifle Association], when anyone talks about the history of guns in America, they're going to have to give me evidence -- facts, not folktales." Three months later, despite growing skepticism, Arming America was awarded Columbia University's prestigious Bancroft Prize for the best work of American history published in 2000. It seemed the question of what was fact, what fiction, had been settled.
Bellesiles had flung his challenge at a political organization, but unfortunately for him it was the scholarly community that picked it up. The evidence he had presented for his groundbreaking theory was investigated first by experts from a range of disciplines and political viewpoints; then by a special symposium in a learned journal; and finally, as a result of the disturbing findings, by the professor's university and an outside panel of scholars that it appointed. The results are now in: Bellesiles' arguments are based on wholesale misuse of evidence and, in some cases, no evidence at all. The "invented tradition" is fact, the professor's version a folk tale.
The same month that Playboy published its interview with Bellesiles, my review of Arming America appeared in reason, one of the first publications to expose his errors. (See "Concealed Weapons," January 2001.) It formed part of what was to swell into an astonishing stream of revelations, most of which were immediately branded by Bellesiles and his supporters as politically motivated or trivial. Scholars and reporters who asked about Bellesiles' sources were treated to continually changing stories about where his research was conducted and the locations of his materials. Their inquiries were greeted, by turns, with injured innocence or vicious invective. Bellesiles had, he said, accidently stepped "into a minefield."
Bellesiles maintains that his book has been subjected to unfair, unprecedented scrutiny. The problem, he told a British reporter last spring, is that there are "some topics which perhaps should not be addressed," that "this is a dangerous environment [in which] to talk about firearms." While Bellesiles admits to some errors in a handful of paragraphs, his response to the unfavorable report of Emory's investigative committee insists, "The overwhelming bulk of the evidence in support of this book's thesis remains unchallenged."
As we'll see, not only have virtually all aspects of his work been "challenged," but Bellesiles' critics have discovered wholesale and systematic misrepresentation of the historical evidence. If anything can be learned from this extraordinary episode, in which one of the most extravagantly praised scholarly books in many years has been exposed as one of the most fraudulent, it is the importance of maintaining rigorous intellectual standards even when they work against one's political preferences.
Before we consider the bizarre twists and turns of the Bellesiles saga, the case should be put in context. Although the Arming America story ranks as the worst scandal to hit the American historical profession in recent memory, while it was unfolding three distinguished historians were exposed for dishonesty of one sort or another.
First came Joseph Ellis, professor of history at Mount Holyoke College in Massachusetts, winner of the National Book Award for his biography of Thomas Jefferson and of a Pulitzer Prize for Founding Brothers: The Revolutionary Generation. In Ellis' case the integrity of his scholarship was not in question. Instead it was the stories of his experiences in the Vietnam War with which he regaled his students that proved his undoing. In June 2001 The Boston Globe revealed that Ellis' actual military experience consisted of ROTC at William and Mary College and teaching history at the U.S. Military Academy at West Point. Although he claimed to have served in Gen. William Westmoreland's Vietnam headquarters and as leader of a platoon that passed near My Lai shortly before the 1968 massacre there, military records show that he never left the East Coast. In August 2001, as critics began to find more and more discrepancies in Bellesiles' work, Ellis was suspended for a year without pay.
In January 2002 it was the turn of Stephen Ambrose, the best-selling author, co-author, or editor of an astonishing 35 books over some 40 years. The first revelation was that his latest book, The Wild Blue, contained extensive text from the work of others. Forbes.com writer Mark Lewis discovered that The Wild Blue contained passages from at least nine of the 28 books cited in the endnotes. These nine were the only books of the group that Lewis was able to check, but he posited that if a similar pattern were found in the other 19 books, "That would leave Ambrose open to the charge that he did not write Wild Blue so much as edit it."
Next, passages copied from the work of other scholars were found in five of Ambrose's other books. Ambrose remained unrepentant, insisting until his recent death that he had done nothing wrong, since he had acknowledged the various authors in his notes. Nevertheless, he presented their prose as his own, a clear case of plagiarism as the academy defines it and as every college freshman is expected to understand it. Some commentators have attributed Ambrose's copying to the pressure to churn out new books once his work began earning millions for himself and his publisher. With a small army of his children enlisted as research assistants, Ambrose & Co. had become something of a cottage industry. Yet similar "borrowing" has been found in books written before the historian's career took off, even as early as 1970. For 30 years Ambrose had gotten away with plagiarism. Whoever originally wrote the text, Ambrose's books continue to be sold under his name, earning money for his estate and his publisher, Simon and Schuster.
That same January, Doris Kearns Goodwin, author of successful biographies of Lyndon Johnson, various Kennedys, and Franklin and Eleanor Roosevelt, television regular and Harvard overseer, was found to have taken numerous passages in her 1987 book The Fitzgeralds from a book by Lynne McTaggart on Kathleen Kennedy. McTaggert had accepted a settlement from Kearns Goodwin and her publisher years before. Kearns Goodwin blamed the problem on sloppy note taking. She claimed she couldn't tell which of her handwritten notes were her own prose, which copied verbatim from her sources. To her credit, she never argued that copying was legitimate. She asked her publisher, also Simon and Schuster, to destroy its copies of the book and to publish a corrected edition, and she set her researchers to work combing the text for "borrowed" passages.
Kearns Goodwin, like Ambrose, has had defenders. In March 2002 she discussed her confusion over her notes in an address at the College of St. Catherine in Minnesota, and afterward the audience gave her a standing ovation. But a week later The Harvard Crimson called for her to resign as a Harvard overseer. An editorial on the front page of the student newspaper cited Harvard policy: "For students who have committed plagiarism any letter of recommendation written for or on behalf of Harvard College -- including letters to graduate schools, law schools, and medical schools -- will report that you were required to withdraw for academic dishonesty." "Why then," the editorial asked, "should an adult, much less a professional historian, continue in her position in the University without consequence?" There have been consequences. Kearns Goodwin is on leave from her post at PBS's The News Hour With Jim Lehrer, invitations for speaking engagements were withdrawn, and she has resigned from the Pulitzer Prize board (though she has been welcomed back to Meet the Press as an occasional commentator).
The cases of Ellis, Ambrose, and Kearns Goodwin are instances in which our most honored historians failed to live up to professional standards. Ellis' infraction seems to me the least serious. His scholarly work remains unquestioned, although personal integrity is expected in teaching no less than in scholarship. Kearns Goodwin plagiarized and is paying the penalty. Ambrose, by contrast, was guilty of wholesale plagiarism and was unrepentant. He made a fortune from the hard work of other scholars. Either his books should be removed from sale, or the actual writers should be compensated from the proceeds. But Michael Bellesiles' unprofessional conduct is of a much deeper dye than these. At least the plagiarized works presumably were accurate representations of the past.
Arming America, by contrast, repeatedly twists the truth to fit Bellesiles' thesis. Bellesiles begins with the European legacy of gun use, especially in England, where, he informs readers, the government "outlawed the use of guns by commoners." Such a policy might seem improbable once muskets were adopted for use by the English militia, which meant that large numbers of commoners must have had and used guns. But Bellesiles writes that the use of guns outside military practice was discouraged "and even forbidden," and that "all guns used by the militia were stored in government magazines."
Throughout the book, Bellesiles equates the presence of guns with violent crime, and he contends that what little violence did occur in England tended to take place "at public festivals, often between competing teams of Morris dancers." Bellesiles' assertion that these charming, white-costumed folk dancers, sporting bells at their knees, were responsible for most of the era's violence struck one British historian as Monty Pythonesque. The article in the English Bill of Rights of 1689 guaranteeing to English Protestants -- then 90 percent of the population -- "arms for their defence, suitable to their condition and as allowed by law," is spun by Bellesiles into proof the government "preferred to maintain tight control of guns."
From start to finish the records tell a different story: of widespread gun use by all classes, from nobles to laborers; laws exhorting men to practice with muskets; militia guns kept in the owners' homes; highways filled with robbers brandishing pistols; and, after 1689, a right to be armed for personal defense that English courts in the 18th century saw as belonging to all Protestants, and in the 19th century to all Englishmen.
On to America, where Bellesiles finds guns rare, expensive, highly regulated, inefficient, unwanted by civilians, and nearly useless for military purposes or hunting. He reports that colonists "often perceived the ax as the equal of a gun" and that there was a "complete failure" on their part to care for guns "or to learn their proper use." As a result of popular indifference and neglect, he says, all guns were made into "the property of the state, subject to storage in central storehouses," so thoroughly regulated that "no gun ever belonged unqualifiedly to an individual." As for hunting with guns, Bellesiles writes, "From the start hunting was an inessential luxury, associated either with the elite gentleman with too much time on his hands, or with the poorest fringes of civilization, if not outright savagery." Bellesiles dismisses the militia, which most historians acknowledge was not up to the standards of professional troops, as "little more than a political gesture," most members having no usable guns.
What is his evidence for these startling conclusions? To document the scarcity of guns, Bellesiles relies on his tabulation of 11,170 estate inventories from 40 counties over the years 1765 to 1859. From these he calculates that in colonial America 14.7 percent of adult white men owned firearms, 53 percent of which were either old or broken. On the frontier, he finds, only 14.2 percent of men owned guns. Not until 1849 to 1859, Bellesiles reports, did the percentages of households with guns increase, largely due to the successful advertising of Samuel Colt. For proof that guns were not used in hunting, he relies on a survey of some 80 travel accounts written between 1750 and 1860. His portrayal of the militia draws on an investigation of government records, letters, and other documents.
The Story Unravels
Here are some snapshots from the unraveling of Arming America:
Fall 2000. Even before the book is in print, having read both Bellesiles' 1996 article that formed the core of the book and advance press coverage, Northwestern University law professor James Lindgren asks to see Bellesiles' probate database. Bellesiles tells him there is no database, only yellow legal pads on which he made pencil ticks representing guns listed in the probate records. Unfortunately, Bellesiles reports, the pads were irreparably damaged in a May 2000 flood at Emory University.
As the book becomes available, scholars find serious mistakes in their own areas of expertise. I find English and colonial facts wrong. Clayton Cramer, the author of two books on gun history, finds extensive evidence of hunting with guns even in the accounts Bellesiles says he read. Discrepancies emerge in the book's claims about low homicide rates, the militia's preparedness, the range and reloading time of muskets, the supposed preference of Americans for the ax over the gun, laws requiring militia guns to be housed in government arsenals, and comments of the Founders.
Readers point out that Arming America lacks any specific information about the 11,170 probates examined, how many there were from any one county or any one time period, where Bellesiles examined them, and how he computed the national averages. Lindgren (a specialist in early American probates), Justin Heather, and other scholars eventually discover that Bellesiles' computations are mathematically impossible. To obtain his national gun ownership average of 14.7 percent, for example, the number of probates from frontier counties would have had to be improbably large.
The archivists at the federal archives in East Point, Georgia, where Bellesiles told Lindgren he did most of the probate work, insist they never had the probate records. Bellesiles changes his story, saying he crisscrossed the country for 10 years visiting individual county archives. He can't remember precisely which ones.
October and November 2000. Scholars examining the only set of probate records Bellesiles specifically cited, those for Providence, Rhode Island, find that nearly everything he says about them is mistaken. Bellesiles claimed the 186 inventories from early Providence were "all for property-owning adult males" and that, while guns appeared in 48 percent, more than half of them were "evaluated as old and of poor quality" and many were "state-owned." But three independent scholars find 17 of the estates were owned by women, 62 percent had guns, only 9 percent of the guns were described as old and of poor quality, and only one was state-owned.
December 2000. As debates about Arming America rage on discussion boards, Bellesiles claims he is getting threatening e-mails. He retreats from online debates and gets a secret e-mail address.
May and June 2001. In response to Bellesiles' claims that he has received death threats and that a fire was set against his office door, the governing councils of the American Historical Association, the Omohundro Institute of Early American History and Culture, and other organizations post resolutions condemning such harassment and defending civil exchange.
September 2001. One of the 40 counties from which Bellesiles said he drew data on gun ownership was San Francisco, where he claimed to have examined 1850s probates at the Superior Court. In response to inquiries from National Review's Melissa Seckora, the court's archivist reports that all 19th-century documents were destroyed in the great earthquake and fire of 1906. When questioned about this, Bellesiles says he must have seen these documents somewhere else, suggesting several other California libraries. None of them has the records either.
To support his contentions about the poor quality of guns, Bellesiles began posting on his Web site Vermont probate records listing guns described as old and broken. A scholar and a Boston Globe reporter examine the originals and find no such descriptions. Told of this, Bellesiles claims his Web site must have been hacked by someone who altered the documents. Emory University investigates but finds no evidence of hacking.
November 2001. In an unprecedented move, Emory demands that Bellesiles offer a "reasoned, measured, de-tailed, point-by-point response" to his critics. His reply, published by the Organization of American Historians in its online newsletter, instead complains of his ill treatment while addressing little of substance. Oddly, the reply refers to only 12 probates from San Francisco, in contrast to the "few hundred" he claimed to have examined when Seckora interviewed him.
January 2002. Bellesiles claims to have found the missing San Francisco probates at the Contra Costa County Historical Archive. He sends copies of these documents to reporters. The Contra Costa archivists point out that not one of these documents is from a San Francisco estate. Furthermore, they have no record that Bellesiles ever visited the collection prior to 2002.
The January 2002 issue of The William and Mary Quarterly carries a symposium in which three of the four scholars asked to examine Arming America, Gloria Main, Ira Gruber, and Randolph Roth, cite gross error and misuse of sources. Main, a professor of history at the University of Colorado at Boulder, finds Bellesiles' claim that probate inventories were complete "incredible." She says the amazing difference between Bellesiles' low figure for gun ownership in Maryland, 7 percent, and hers, 76 percent, "boggles the mind." Gruber, a professor of history at Rice University, concludes Bellesiles worked "to minimize the importance of guns, militia, and war in early America" using "a consistently biased reading of sources" and "careless uses of evidence and context." Roth, a professor of history at Ohio State University, finds Bellesiles' homicide information is "misleading or wrong" in every instance.
February 2002. Emory launches an internal review, later followed by appointment of an external panel to examine Bellesiles' work. By this point nearly all of Bellesiles' early boosters have gone silent.
The Victim Pose
Contrary to Bellesiles' pretense that he accidentally stepped into a political minefield, the introduction to Arming America is intentionally polemical. The book opens with a description of a toddler featured on the cover of Time magazine "dressed in camouflage and clutching a high-powered rifle." Despite a religious upbringing, the boy at age 11 joined a 13-year-old friend in shooting children at his Arkansas school. Bellesiles comments, "The temptation of a gun can trump a claim of faith in God and all dreams of childhood innocence."
he gun control movement's need for supportive scholarship is what gave Arming America so much visibility. The New York Times heralded Bellesiles' results five months before publication, although the reporter who lauded its findings complained to me that he had not even been able to see the galleys. Many early reviewers seemed so eager to extol the book that they suspended their critical faculties -- either that, or they were taken in by the trappings of scholarship or the wild approval of others they trusted. Politics propelled the book to dizzying heights.
But when critics began to point out the book's obvious flaws, Bellesiles, his publisher, and his supporters dismissed the criticisms as politically motivated. Columbia University's Bancroft Committee apparently did not bother to read the critical reviews before awarding Arming America the nation's most prestigious prize in American history. (In December, the trustees of Columbia University finally voted to rescind the honor, the first time they have had to take such action in the more than 50 years of the award's history. They have also asked that the $4,000 in prize money be returned.) We early critics were confronted with Bellesiles' complaints about hate mail and death threats, tales that prompted professional historical organizations to leap to his defense. To challenge Bellesiles' work was to risk being lumped together with "gun nuts" and extremists.
Bellesiles wrapped himself in the mantle of victimhood. But occasionally it slipped, and a different persona peered out. In an essay in the Chicago-Kent Law Review the year Arming America appeared, he portrayed scholars who found evidence of an individual right to be armed as "post-modernists" for whom "facts do not matter," partisans uninterested in "the hard and time-consuming task of archival research." He warned that "the next stop is Holocaust denial." When Clayton Cramer, an amateur historian, questioned some of Bellesiles' evidence in a polite letter to The Chronicle of Higher Education, Bellesiles lashed out:
"Clayton Cramer hates my research and has for some time....It is not my intention to give an introductory history lesson, but as a nonhistorian, Mr. Cramer may not appreciate that historians do not just chronicle the past."
In June 2002, when the soundness of his scholarship had become so suspect that the National Endowment for the Humanities stripped its name from his Newberry Library fellowship, Bellesiles told a British reporter that the move "reawakened the ghosts of McCarthyism." He described it as a "political decision that should send chills through academics everywhere and is clearly intended as a warning to any scholar who dares to work on a controversial topic."
Closure of a sort came on October 25, 2002, with the report of three distinguished historians -- Stanley Katz of Princeton University, Hanna Gray of the University of Chicago, and Laurel Thatcher Ulrich of Harvard -- charged by Emory University to investigate Bellesiles' use of probate records and militia counts. Their mission was surprisingly narrow, given all the doubts that had been raised about the veracity of Bellesiles' work. But they agreed that "the best that can be said of his work with the probate and militia records is that he is guilty of unprofessional and misleading work. Every aspect of his work in the probate records is deeply flawed."
The panelists found evaluating the table in which Bellesiles charted his probate findings "an exercise in frustration because it is almost impossible to tell where Bellesiles got his information." They reported that "we had the same question as Gloria Main [in her William and Mary Quarterly article]: 'Did no editors or referees ever ask that he supply this basic information?'" They found "egregious misrepresentation" in the construction of that table and evidence of "falsification" in, among other flaws, Bellesiles' silent omission of the years 1774-76 "precisely because they failed to show low numbers of guns." Applying Emory University's guidelines, the committee concluded there were "other serious deviations 'from accepted practices in carrying out or reporting results from research.'"
The committee members deemed Bellesiles' work with militia records "superficial and thesis-driven." They agreed with Ira Gruber, another of the scholars in the William and Mary Quarterly symposium, that Bellesiles' "efforts to minimize the importance of guns, militia, and war in early America...founder on a consistently biased reading of sources and on careless uses of evidence and context." Finally, they concluded that his professional scholarship "falls short" on every count and that "his scholarly integrity is seriously in question."
But there were two problems with the committee's charge that Bellesiles has been quick to capitalize on. First, it was to investigate only the use of probate and militia records, despite evidence that he misused many other sources. Second, the Emory standard that "intentional fabrication or falsification of research data" be found, a standard devised with the sciences in mind, was unrealistic, since much that was intentional could be written off as extreme bias or sloppy method.
Hence, although the three scholars could not find that Bellesiles' use of Vermont and Providence probates constituted "intentional fabrication or falsification," they were "seriously troubled by his scholarly conduct" and "sloppy scholarship," which "does not prove a deliberate attempt to mislead, however misleading the result." Even in the case of the nonexistent San Francisco data, they could not "prove" he "simply invented his California research, but neither do we have confidence that the Contra Costa inventories resolve the problem." Only in the case of the key probate table did they conclude there was deliberate falsification.
Trust, but Verify
Bellesiles resigned from his position at Emory after the report was issued, saying he could not "continue to teach in what I feel is a hostile environment," but he nonetheless has tried to minimize the significance of the committee's conclusions. "Obviously," Bellesiles told a reporter for the Chicago Tribune, "they were very angry at me." As for his mathematically impossible statistics, he shrugged, "I've never been good at math." In any event, he informed the Tribune, he wrote a book with 1,347 footnotes, and the panel found problems with only five of them.
Yet it clearly is not true that the essence of the book remains "unchallenged," as Bellesiles claimed in his response to the committee. Virtually every aspect of it -- including his conclusions about English weapon use, hunting, axes vs. guns, homicide rates, and the inefficiency of firearms -- has been shown to rely on faulty, at times nonexistent, evidence and biased research. This is a matter of fact, not interpretation, as Bellesiles would have it. Scholars should waste no more time on this discredited volume.
Checking footnotes is tedious, and everyone makes some mistakes in the course of a long, detailed work. But experts in a field, asked to review a manuscript, can ferret out those few monographs where the errors are wholesale and the facts misrepresented. The Bellesiles case underlines how essential it is that historians take their critical role seriously and always place it before their political inclinations. Our professional integrity demands it.
Indeed, Bellesiles' rise and fall demonstrate how damaging to any cause a dishonest work of scholarship can ultimately be. Haverford College historian Roger Lane, a gun control supporter who gave Arming America a laudatory review, now writes of Bellesiles: "I'm mad at the guy. He suckered me. It's entirely clear to me that he's made up a lot of these records. He's betrayed us. He's betrayed the cause."
University presses and other academic publishers routinely put scholarly works through a rigorous peer review. But that is not necessarily true of commercial publishers. They must be more scrupulous before foisting on the public a clever piece of fiction disguised as a work of history. And when a book proves to have been unscholarly, or plagiarized, it should be withdrawn. Knopf, to its credit, in January decided to stop selling Arming America. But Simon and Schuster still publishes Ambrose's plagiarized works.
The news media also play a key role. The New York Times and many other news outlets heralded Arming America, while their bias in favor of gun control has led them to ignore scholarly books uncongenial to the cause. In this instance most of the press ignored, or dismissed as politically motivated, growing skepticism about Bellesiles' evidence. The slowness of scholarly journals to review books makes the general press all the more essential if the public is to be alerted that something is amiss. Four reporters deserve special thanks for bringing out the facts: Kimberley Strassel of The Wall Street Journal, Melissa Seckora of National Review, David Mehegan of The Boston Globe, and David Skinner of The Weekly Standard. Thanks are also due Stacy McCain of The Washington Times and Robert Worth of The New York Times.
Those who wonder how Bellesiles thought he could get away with this forget that he nearly did. The book won a major prize. Court opinions began citing his findings. He told a reporter his only regret was that he delayed fighting back, that "for too long I thought this was only a tempest in a footnote."
All those scholars and reporters from across the political spectrum whose exhaustive investigative work exposed Arming America for the high-stakes fraud that it was deserve the gratitude of the historical profession and the American people. Scholarship relies on trust, and sometimes that trust is misplaced.