Did Another Bancroft Winner Have Trouble Counting?

The book was published by Alfred A. Knopf and won a Bancroft Prize. It was widely praised for its elegant prose and provocative interpretation of an important dimension of American culture."... this is a wonderful book --" gushed an editor of the Journal of American History,"-- incisive, far-seeing, and passionate." Its author was criticized, however, for reshaping original sources to make them sustain the book's thesis and offering tables of data that misrepresented reality. No. It is not Michael Bellesiles's Arming America. It is Christine Leigh Heyrman's Southern Cross: The Beginnings of the Bible Belt (New York: Alfred A. Knopf, 1997).[1]

Bellesiles and Heyrman first crossed paths in the History Department at the University of California, Irvine. There, as a member of its faculty in the mid-1980s, Heyrman directed Bellesiles's graduate work. [2]For a time, they shared a common publisher and editor, Knopf's Jane Garrett. That may explain why, at least superficially, their two books even look alike: no illustrations or bibliography, but lengthy endnotes and an appendix with tables of data to the rear, lending an aura of empirical authority. Together, the books suggest that award-winning work in history recognizes good writing and a striking thesis. It may have the appearance of statistical authority, but actually undervalue an accurate marshaling of evidence.

In April 1998, as Michael Bellesiles prepared to spend the next academic year writing Arming America at Stanford University's Humanities Center, Columbia University announced that his mentor, Christine Heyrman, had won the Bancroft Prize for her book, Southern Cross. In some ways a sophisticated book, it tracked the course of evangelical preachers as they overcame regional hostility and skepticism by increasingly accommodating a white male patriarchy. Heyrman had won history's gold ring and Bellesiles hoped to do the same. He should have paid close attention to reviews of her book, but the critical ones appeared in obscure places. During the summer, the Evangelical Studies Bulletin, published by the Institute for the Study of American Evangelicals at Wheaton College in Wheaton, Illinois, issued a review by Kurt Berends.

Berends was well qualified to review the book, having just earned his doctorate at Oxford University. His research had taken him through many of the same sources that Heyrman studied. Her book's"often persuasively argued" account of the triumph of evangelical Protestantism in the South, he said, was"imaginative, witty and compelling." Berends was critical of Heyrman's over-reliance on Methodist sources and a premature finding of evangelical hegemony in Southern culture. More to the point, he found that she manipulated her sources to sustain a thesis. [3]

Just as evangelical preachers tamed diabolical imagery in their preaching to achieve mastery, Southern Cross argued, they commonly led gullible listeners to believe that they possessed magical, shaman-like powers, capable of controlling even the weather. It was a theme Heyrman had found in Awash in a Sea of Faith by Jon Butler of Yale University, where she had earned her doctorate. [4]Heyrman recognized an interesting theme when she read about it and, apparently, determined to find it for her book on the South. That is thesis-driven research, comparable in part to Michael Bellesiles's more breathtaking example.

As evidence of Butler's magic/shaman theme in Southern evangelicalism, Heyrman cited the claim of John Early, a Methodist itinerant, who she said,"exulted to his diary in 1808 that when lightning flashed during one worship service, 'I immediately got to my knees and told the people I would ... pray that God in his great mercy would withhold the rain until we were done. The rain immediately blew over....'" [5]Berends remembered the passage otherwise from his own notes from John Early's diary. According to him, the Methodist preacher had written:"I immediately got to my knees and told the people I would, with the help of their prayers, pray that God in his great mercy would withhold the rain until we were done. The rain blew over for a season. About one hundred cam[e] forward to be prayed for and we were again frustrated by the rain which made our meeting break up sooner than it would have."[6]

Heyrman's editorial manipulation of Early's words as primary evidence of her argument was damaging, said Berends, because Early claimed no magical powers for himself. Whatever influence prayer might have on God's control of the weather was exercised by preacher and congregation together. Nor did the rain blow over completely. It blew over"for a season" and it soon returned to break up the meeting."Was Early really bragging about his personal power and ability to manipulate the weather," Berends asked,"when his first call for the rain to cease is made in cooperation with the congregation and the altar call -- the heart of the service and his moment in the spotlight -- was disrupted by a storm?" [7]

Kurt Berends's finding about the use of primary sources in Southern Cross calls for additional examination. Heyrman's evidence that evangelical preachers sought to wean white Southerners from practitioners of witchcraft comes from a book by a prominent Presbyterian pastor, Charles Colcock Jones, whom she mis-identified as a Baptist. [8] Referring to the slaves'"superstition," Jones wrote:"They believe in second-sight, in apparitions, charms, witchcraft, and in a kind of irresistible Satanic influence. The superstitions brought from Africa have not been wholly laid aside." From these words, Heyrman made the following point:"And in the evangelical campaign to weaken the hold of those wonder-workers on southern whites, there could have been few weapons more potent than the clergy's insistence that such credulity was confined to African Americans. As the Baptist minister Charles Colcock Jones suavely asserted in 1842, only blacks embraced the lore of 'second sight ... apparitions, charms, witchcraft, and ... a kind of irresistible Satanic influence .... superstitions brought from Africa [not] wholly laid aside.'"[9]

As the examples from both John Early's Diary and Charles Colcock Jones's book show, Heyrman's use of ellipses is idiosyncratic. Jones's text simply does not say that"superstitions" were limited to African Americans. He referred to those of the slaves, to be sure, but he does not say, as Heyrman twice insists, that white Southerners had no similar beliefs.[10]Most importantly, these two dubious quotations are the only evidence Heyrman offers of Butler's magic/shaman theme in early Southern evangelical Christianity.

Heyrman's thesis driven research also missed opportunities. Her colorful accounts of divine visions and demonic confrontations are drawn from the diaries and autobiographies of late eighteenth and early nineteenth century counter-cultural white Methodist and Baptist preachers. That literature would put the otherwise exotic divine visions and demonic confrontations of"The Confessions of Nat Turner" in context as another manifestation of that earlier tradition.[11]The religious inspiration of the South's last major slave rebellion makes Nat Turner an authentic heir to an earlier counter-cultural evangelical tradition. But that did not interest Heyrman, who wanted to explain to us that, by 1831, the South's white evangelicals had made their peace with the region's culture.

Yet, for scholars of religion in the American South, Heyrman's Southern Cross and its Bancroft Prize were an important benchmark for the field, granting it increased respectability. So, they launched a new e-journal, the Journal of Southern Religion, with -- not one -- but a symposium of three reviews of the book. One of its reviewers, Ann Taves of California's Claremont Graduate University, raised a second major criticism of Heyrman's Southern Cross. Its author had what Taves called a"boundary between North and South" which was"to say the least, rather fluid." She pointed to Heyrman's frequent references to the Methodist itinerant preacher, Benjamin Abbott. He grew up in Pennsylvania and New Jersey, began his preaching in New Jersey and spent his life on missionary circuits in New York, New Jersey, and Maryland. Another reviewer pointed to Heyrman's citations of the Illinois Methodist itinerant, Peter Cartwright. In her rendering of antebellum geography, Cartwright's birth in Kentucky made him no more and no less Southern than either Jefferson Davis or Abraham Lincoln. [12]

For Heyrman, the Allegheny Mountains have only southern exposures. She likes to report dramatic scenes and, twice within seven pages, a location"in the Allegheny Mountains" covers the fact they took place north of the Mason and Dixon Line. The Methodist itinerant, Valentine Cook, confronted the devil" in the Allegheny Mountains," she says. Old Satan tried to tempt Cook with prestigious pulpits -- not in Richmond, Charleston, and Savannah -- but in"New York, Philadelphia, and Baltimore." Heyrman did not tell her readers that the devilish temptations took place in six inches of snow outside of Uniontown, Pennsylvania. Similarly, the Methodist itinerant, Jeremiah Minter, met the devil and the Methodist bishop, Francis Asbury, says Heyrman,"in the Allegheny Mountains." Specifically, they met about two days by horseback out of the"red stone country" in the Monongahela River valley near Pittsburgh. [13]

Moreover, Taves pointed out, Heyrman's table V, which enumerated Southern Baptist, Methodist, and Presbyterian church membership by race in 1834-1836, counted Ohio, Indiana and Illinois as Southern states on the theory that they were largely settled by Southerners.[14]

Not since"Bleeding Kansas" had the Southern sense of manifest destiny ever been quite so bold. The Northwest Ordinance of 1787 banned slavery and, thus, most African Americans from those three states. The purpose of Heyrman's table V was to show distribution of Southern church membership by race and denomination, but the inclusion in her table of states where evangelical church membership was overwhelmingly white seriously over-represented white membership in Southern
Baptist, Methodist, and Presbyterian churches. [15]

The problem of Heyrman's addition of states from the Old Northwest begins with her table IV, which includes Ohio and Indiana, in a count of Southern evangelical church members by denomination and race in 1813. There is no evidence that Heyrman intended to deceive anyone with her tables IV and V, however. Her notes to them point out the inclusion of Ohio, Indiana, and Illinois data in her calculations of Southern church membership. Inflating figures for white membership in Southern evangelical churches by including data from states where they were overwhelmingly white, however, may seem to justify the inadequate attention which she gave to the role of African Americans in creating an evangelical South. There are additional reasons to doubt Heyrman's data. A train of authorities, from Charles Colcock Jones in 1842 to Edwin Scott Gaustad in 2001, suggest that Heyrman may have undercounted Afro-Baptists by half and her count of Afro-Methodists in tables IV and V certainly does not include any of those in Delaware and Maryland who organized the separate African Methodist Episcopal, Zion and the smaller African Union Methodist Protestant denominations. [16]

Moreover, even if one accepts Heyrman's geographical definition of the South and uses the data sources she cites, the numbers are simply not what her tables III, IV, and V indicate. In table III, which reports church membership by race and denomination in 1790, she undercounted white Methodists by some 7,000 and black Methodists by some 3,000, undercounts of 22 percent and 32 percent respectively. In table V, she overcounted Presbyterians by some 66,000, an overcount of 70 percent. [17]That overcount apparently caused her to overestimate Southern Presbyterian membership in her tables III and IV.[18]

Inexplicably, Heyrman counts all Presbyterians as white in all three tables. As early as 1790, Cyrus Gildersleeve's Presbyterian Midway Church in Liberty County, Georgia, alone had 49 African American members. In the periods covered by Heyrman's three tables, probably little more than 1 percent of Southern Presbyterians were black, but she seems not to know that black and white missionaries like John Chavis and Charles Colcock Jones built foundations for substantial Presbyterian congregations which, by 1860, were commonly 75 to 90 percent black in the Georgia and South Carolina low country.[19]Thus, beyond her count of states of the Old Northwest in tabulating Southern evangelical church membership, her undercounts, probably of Afro-Baptists and certainly of Afro-Methodists, Heyrman's overcount of Southern Presbyterians and arbitrarily declaring all of them white are additional ways in which her data are bleached.

Notes to the tables of data at the rear of Heyrman's book, however, suggest that their figures are so soft that a social scientist would simply chuckle at the mush. Church membership data in the period, as she says, are notoriously unreliable. Her endnote quotes a nineteenth century Methodist itinerant as observing that"The careless manner by which the preacher in charge too frequently arrives at the numbers in society cannot be too severely censured." [20] Using numbers of dubious reliability, Heyrman guessed that on average 10 percent of Methodist and Baptist church members in 1790, 20 percent of Methodist and Baptist church members in 1813, and 20 percent of Baptists and 18 percent of Methodists in 1834-1836 were African Americans. This is over a geographical range in which the African American population varied from less than 1 percent to 90 percent. The resulting numbers only look real. Racial identities were assigned to arbitrary percentages of digits and all the Presbyterians were white. Rarely has race been so obviously socially constructed. She might as well have drawn the Mason and Dixon Line at the Canadian border. Despite her claim to be doing so, Heyrman was not actually counting real church members. She was manipulating numerical abstractions. Her student claimed to be counting real guns and his charts were not qualified by extensive methodological notes. The difference is the difference between raising a social scientist's skeptical eyebrows and being held to a social scientist's standard.

Assessing a teacher's influence on a student is always difficult. Christine Heyrman is specifically not guilty by association with Michael Bellesiles, nor he with her. Yet, her attitude toward evidence is perhaps the most Southern thing about her book: it is cavalier. Her practice could shape the attitude of a student, particularly one so talented as Bellesiles. It has problems of the kind that his magnified. Southern Cross was a model for what his prize-winning work in history would be. Her connections with Knopf and the Bancroft Prize committees surely paved a way for his award. It is painful to re-read Arming America's acknowledgment of her influence."... my warmest thanks to Christine Heyrman, my graduate advisor, who taught me to check the sources for myself," he wrote."Good point." [21]

Appendix

1. Offered a summary of evidence cited in this article, Heyrman replied:

It's hard for me to assess your statistical findings from so a brief summary, but I certainly encourage you to write them up with all the usual scholarly apparatus and to submit the results to a refereed journal. (I'd like a copy, too.) What you say about Nat Turner is right on, but I promised myself that the book would end in the 1820s. [22]

I haven't seen Kurt Berends' review, so I don't know which interpretation of John Early you're referring to (he appears twice on p. 75). But I did check out the Charles C. Jones quotation and I'll stick by my interpretation -- his view is that various"superstitions" are peculiar to blacks and imports from Africa.[23]

I'd be the first to admit that my definition of the South is"expansive," but the migration patterns of the early national period (e.g., settlement of the Ohio valley mainly by southerners) justify that choice. [Christine Heyrman to Ralph E. Luker, 1 May 2003.]

2. Offered a transcription of the passage from John Early's diary cited by Kurt Berends and asked to compare it with her edited version of the document, Heyrman replied:

"Thanks for forwarding the Berends' comments -- I take his point, but I think Early's statement could bear either interpretation. All part of history's fascination, those ambiguities." (Heyrman to Luker, 1 May 2003.)

3. Subsequently, Heyrman responded on statistics:

I was perplexed by the divergence in our numbers on the Methodists in 1790 and the Presbyterians in 1834/36 so I dug out my notes. In the case of the Methodists in 1790, I inadvertantly [sic] left out Delaware, which I believe accounts for the discrepancy. [24] We're agreed on the importance of getting the numbers (and everything else) right, but I don't think that this undercount make much difference in my arguments. Factoring in Delaware for the 1790s, church membership among whites rises from 14.4% to about 16%; for blacks, it goes from 3.7% to about 5% - still a tiny minority. In fact, these higher percentages for 1790 lend support to my argument that evangelical expanion between 1790 and 1813 was surprisingly sluggish. ([Heyrman, Southern Cross,] p. 23).

As for the Presbyterians in the 1830s, have you consulted pp. 143 and 147 in Hayward? Many of these"offbrand" Presbyterians had their early strength in the South and the western country, and the Cumberland Presbyterians were quite numerous in those parts - which is how I came up with the figure of 160,000+ for all Presbyterians. [25]

Speaking of the Presbyterians, we can agree that Afro-Presbyterians were pretty sparse during the period covered by my tables, which is why I state on p. 262 that"virtually all" Presbyterians were white. While I had a few bits of evidence that there were Afro-Presbyterians (Midway Church, as you mention, a diary from the Northern Neck), none of it lent to conjecturing about numbers and percentages. [26]

I will notify UNC Press of these errors for 1790 and ask to correct the statistical tables in any future paperback editions. Thanks very much for bringing them to my notice, and again, good luck with your research.