The flawed scholarship of the book is further epitomized by a significant transcription error which appears in Appendix E, the text of Ellen Randolph Coolidge’s 1858 letter to Joseph Coolidge. In relevant part, the original letter as found in the Coolidge Letterbook, University of Virginia Library42—in clear handwriting—states the following about Jefferson’s rooms at Monticello:
His apartments had no private entrance not perfectly accessible and visible to all the household. No female domestic ever entered his chambers except at hours when he was known not to be there and none could have entered without being exposed to the public gaze.
As printed in the appendix to Professor Gordon-Reed’s book,43 however, the passage reads:
His apartments had no private entrance not perfectly accessible and visible to all the household. No female domestic ever entered his chambers except at hours when he was known not to be in the public gaze.
Even if we give Professor Gordon-Reed the benefit of the doubt and assume that omission of the crucial words—which obviously changes significantly the meaning of the sentence—was not a deliberate distortion of the evidence but rather an innocent transcription mistake, so critical an error casts doubt on the reliability of her work.
The TJMF (Monticello) Committee Report
Following release of the DNA study in the fall of 1998, Daniel P. Jordan, the president [1985–2008] of the Thomas Jefferson Memorial Foundation (TJMF)44—the institution that owns and operates Jefferson’s home Monticello—appointed a nine-person in-house research committee which was charged, in Jordan’s words, to “review, comprehensively and critically, all the evidence, scientific and otherwise,” including Dr. Foster’s DNA study, “relating to the relationship of Thomas Jefferson and Sally Hemings.” The Committee was chaired by Dianne Swann-Wright, Director of Special Programs at Monticello (including its Getting Word Oral History Project described below), and its members—described in its report as “including four Ph.D.’s and one medical doctor”—were all Monticello staff members. Although the Committee consulted with members of two other Monticello committees—the Advisory Committee for the International Center for Jefferson Studies and the Advisory Committee on African-American Interpretation—it is worth emphasizing that no scholar independent of Monticello had any input in the report.
Although the Committee had concluded its work by spring 1999, its report was not released until January 27, 2000. The report was immediately posted on the Internet, and Dan Jordan noted that within a week the Monticello website received nearly 60,000 “hits” a day, with some 3000 different individuals downloading the report. Two weeks later, after the television airing of the CBS miniseries Sally Hemings: An American Scandal, Jordan noted that the hits “maxed out” Monticello’s system, with as many as 900,000 in one day. Although he dismissed the CBS miniseries as “ridiculous as history,” “a soap opera,” and “strictly Hollywood,” Jordan acknowledged that “it certainly did encourage an interest in the story.” He added, “Anything that encourages and raises the consciousness of the American people about history and race is a good thing.”45
What was not mentioned in the TJMF’s press conference and not acknowledged on its website until about three months later, on March 23, 2000, was that one of the members of the Monticello Committee—White McKenzie (Ken) Wallenborn, M.D. (the “medical doctor” identified in the committee’s description)—had dissented stridently from the Committee’s report. Noting several areas of disagreement with the majority’s report, Dr. Wallenborn in his minority report (dated April 12, 1999) concluded that “[t]here is historical evidence of more or less equal statu[r]e on both sides of this issue that prevent a definitive answer as to Thomas Jefferson’s paternity of Sally Hemings’ son Eston Hemings or for that matter the other four of her children.” He urged the TJMF to continue to regard the paternity question as an open one. In an essay published subsequent to the release of his minority report, Dr. Wallenborn has charged that the Monticello Committee—and particularly its chair, Dianne Swann-Wright, and Lucia (Cinder) Stanton (Shannon Senior Research Historian at Monticello), whom he identified as the principal author of the Committee’s final report—“had already reached their conclusions” at the start of their deliberations. According to Dr. Wallenborn’s account, the Committee followed “the same tactic” that Professor Annette Gordon-Reed employed in her book, of ignoring or dismissing as problematic “most of the evidence that would exonerate Mr. Jefferson.”46 Equally troubling is Dr. Wallenborn’s statement that Dianne Swann-Wright failed to share his dissenting report with other members of the committee. Indeed, he notes that it was not shared with the interpretive staff at Monticello nor with the TJMF Board of Trustees until he began circulating it after the January 26, 2000 press conference.47
Dr. Wallenborn’s criticisms of the Monticello Committee appear to be well-founded. Upon close reading, its final report is far from being the “scholarly, meticulous, and thorough” analysis Dan Jordan claims it is. Its general conclusion, that Thomas Jefferson fathered one, if not all, of Sally Hemings’ children, fails to be adequately supported by the evidence gathered by the Committee and summarized in its findings.
Indeed, a fundamental problem with the Committee report is the apparent absence of any methodology for evaluating or weighing evidence. When the report concludes, specifically, that the “currently available documentary and statistical evidence, indicates a high probability that Thomas Jefferson fathered Eston Hemings, and that he was most likely the father of all six of Sally Hemings children,” it offers no standard by which the conclusory terms high probability or most likely can be objectively measured. Generally speaking, the Committee report seems to rest this conclusion on just a few pieces of evidence—the results of Dr. Foster’s DNA tests, Madison Hemings’ 1873 “memoir,” and the “Monte Carlo” statistical study conducted by Committee member Fraser Nieman—plus one critical, but unsupported, assumption: that all of Sally Hemings’ children were fathered by just one man. This single father postulate rests on the flimsiest of evidence: the naming of the Hemings siblings’ children after one another, which supposedly demonstrates the “closeness” of the family (and thus, it is assumed, Sally Hemings’ monogamy), and the claim of an absence of evidence that Sally Hemings was not monogamous (a false claim in light of the Edmund Bacon evidence, which the Committee discounts, as noted below). The only documentary evidence which the Committee can cite in support of its conclusion that Jefferson “most likely” fathered Sally Hemings’ children other than Eston is the Madison Hemings 1873 interview.48
Another fundamental flaw in the Committee’s report is the problem of bias and conflict of interest. Since 1993 the TJMF has been conducting an oral history research project called “Getting Word,” to locate the descendants of Monticello’s African-American community and to record and preserve their stories and histories. The project has interviewed over 100 people, including 22 descendants of Madison Hemings and four descendants of Eston Hemings. The very fact that Monticello staff members have been involved in this project makes it difficult for an in-house research committee to objectively evaluate oral history evidence. The problems of bias in favor of oral history evidence generally—and selective bias in favor of those particular families interviewed through the Getting Word project—were compounded by the fact that the chair of the ad hoc research committee was Dianne Swann-Wright, director of Special Programs at Monticello, who had been employed to work on the project since its inception (and her arrival at Monticello) in 1993. Given the intimate involvement of Dr. Swann-Wright and other Committee members with the people interviewed for the Getting Word project, it is not surprising that the Committee report heavily relies on the 1873 Madison Hemings story and the oral tradition among his descendants as the key evidence in support of the Jefferson paternity thesis.
As noted above, oral tradition evidence has a general problem of unreliability. The Committee report is flawed not only because it relies heavily on oral tradition evidenc
e, but that it relies on it selectively, taking seriously only that oral tradition that fits with the story of Jefferson’s paternity. The bias is evident in the report, where it infers from the seriousness of the Madison Hemings’ descendants’ “history” that it is true and therefore ought to be treated on par with documentary and other evidence. “In a climate of disbelief and hostility,” the report notes, “they continued to tell their children and grandchildren of their descent from Thomas Jefferson, often at significant times in their lives. … ”49 On the same page of the report, however, the Committee notes that the oral history of the Eston Hemings descendants claimed descent from Jefferson’s “uncle”—an oral tradition which apparently was taken just as seriously by this line of the family, until publication of Fawn Brodie’s Intimate History prompted family members to change the story—but the Committee dismisses the earlier tradition among Eston Hemings’ descendants as “altered to protect their passing into the white world.”50 The change of the Eston Hemings family oral tradition following publication of the Brodie book is acknowledged by family members. “We’re just learning—from some of our cousins—stories we weren’t able to hear,” one family member said.51
Significantly, the Committee report also concluded that Thomas Woodson was not the son of Thomas Jefferson, and indeed that there was no documentary evidence linking him even to Monticello and Sally Hemings.52 The significance of this is twofold. First, it acknowledges the falsity of a core allegation of both the original 1802 Callender story and the 1873 story attributed to Madison Hemings: the notion that Jefferson’s sexual relationship with Sally Hemings began in France, and that she bore him a son soon after their return to the United States. As the Committee report finds, there is no evidence of any child being born to Sally Hemings prior to 1795. Second, the findings regarding Thomas Woodson starkly reveal the inherent unreliability of oral tradition as evidence. The Woodson descendants just as fervently believed that their ancestor was the son of Thomas Jefferson, and the Committee found that “the longstanding oral history warrants inclusion of information” about Woodson despite the absence of documentation to connect him to Sally Hemings and Monticello, let alone to Thomas Jefferson.53
There is one other oral tradition, of course, which was summarily rejected by the Committee. Beginning with the direct testimony of Jefferson’s grandchildren, Thomas Jefferson Randolph and Ellen Randolph Coolidge, the oral tradition in the family descended from Jefferson’s daughter Martha Jefferson Randolph has identified one of the Carr brothers, Peter or Samuel (Jefferson’s nephews by his sister Martha) as the father of Sally Hemings’ children. Although that tradition apparently too was taken just as seriously as the tradition of Hemings descendants—and although it is arguably far more reliable, for it was based on the testimony of eyewitnesses to the events in question—the report essentially dismisses Carr paternity by pointing to the DNA test results on Eston Hemings’ descendant and assuming that Sally Hemings’ children were all fathered by the same man.
The Committee’s bias is evident also in the double standard it employs in weighing evidence. For example, the published account of Monticello overseer Edmund Bacon, which identified another, unnamed man as the father of Harriet Hemings, is dismissed as having “problems of chronology,” noting that Bacon was not employed at Monticello until five years after Harriet’s birth.54 But this ignores the real possibility that Bacon resided at Monticello as early as 1800 and also assumes that Bacon was describing an event he witnessed prior to Harriet’s birth when indeed he might have concluded that the man he saw some years later was the father of her children. However, immediately following this curt dismissal of Bacon’s account, the Committee report states that Israel Jefferson’s 1873 interview “corroborated Madison Hemings’s claim of Jefferson paternity”—even though Israel Jefferson’s account, besides the many problems noted above, also has a real “chronology problem” of its own: Israel was only eight years old at the time of the birth of Sally Hemings’ youngest child, Eston!55
Important pieces of evidence that question the Jefferson paternity thesis are either ignored or blithely dismissed by the Committee’s report. For example, Jefferson’s own denial of the Callender allegations, in an 1805 letter written to a member of his administration, is dismissed as “ambiguous”56—an assessment that fails to take into account its clear historical context (as discussed below). The account of former household slave Isaac Jefferson, who mentioned and described Sally Hemings in his memoir, is omitted from the Committee report, even though the fact that Isaac did not so much as hint that there was any special relationship between Jefferson and Sally Hemings is powerful evidence questioning the paternity thesis.57
Other problems, though relatively minor, in the Committee’s report reveal that it was far less meticulously written than one would expect it to be. For example, although the report does include a facsimile of the 1858 Ellen Randolph Coolidge letter, it follows it with the flawed transcription as found in Appendix E of Professor Gordon-Reed’s book. The draft of Committee member Fraser Neiman’s article, “Coincidence or Causal Connection? The Relationship between Thomas Jefferson’s Visits to Monticello and Sally Hemings’s Conceptions”—which was going to print in The William & Mary Quarterly in January 2000, just as the Committee report was released—contains a typographical error which distorts the DNA study in a significant way. The molecular geneticists tested “male-line descendants of Thomas Jefferson” (emphasis added), the article states, when of course it was not Thomas but Field Jefferson, Thomas Jefferson’s paternal uncle, whose descendants were tested.
The article by Mr. Nieman, who is director of archeology at Monticello, has far more serious problems than this embarrassing typographical error. As the Scholars Commission report notes, none of us was impressed by this so-called “Monte Carlo” statistical study. The Monte Carlo approach estimates the probability of a given outcome by comparing it to a very large number of random outcomes generated by a simulation model. Nieman’s study rested on two unsupported postulates: that there could only be a single father for all of Sally Hemings’ children, and that rival candidates to Thomas Jefferson would have had to arrive and depart on the exact same days he did. Here, the assumption of random behavior makes little sense, because the visits to Monticello of the other candidates for paternity—Jefferson’s friends and relatives (including his brother Randolph, Randolph Jefferson’s sons, and the Carr brothers)—were not random occurrences; they certainly would have been far more likely to occur after Jefferson’s return to Monticello from extended absences in Washington or elsewhere. The final impression one gets of the Nieman study is of a simulation whose parameters were deliberately set to “get” Thomas Jefferson as the father of Sally Hemings’ children.58
As Dr. Wallenborn has noted, Neiman’s statistical study “cries out for valid comparative studies of the other Jefferson males who might have fathered Eston, and in the absence of these comparisons, the results are inconclusive.”59 As the Scholars Commission report notes, the circumstantial case that Eston Hemings was fathered by Randolph Jefferson, Thomas Jefferson’s younger brother, is “many times stronger” than the case against Jefferson himself. Significantly, even the Monticello Committee report notes documentary evidence that Randolph Jefferson visited Monticello in August 1807, a probable conception time for Eston Hemings (this evidence consisting of a letter from Jefferson to his brother, inviting him to visit Monticello while Randolph’s twin sister, Anna Marks, was then visiting)—but the Committee rejects this evidence because no corroborating evidence has been found to indicate that Randolph did in fact visit at this time.60 The Committee report also notes that Randolph Jefferson’s sons Thomas, in 1800, and Robert Lewis, in 1807, “may well have been at Monticello during the conception periods of Harriet and Eston Hemings.”61
One final, critical assumption made both in the Neiman study and in the Monticello Committee report as a whole is the assumption that Sally Hemings was continuously present at Monticello. As
the Scholars Commission concludes, however, that assumption may be problematic. Sadly, we simply do not know enough about Sally Hemings—even her duties at Monticello—to conclude that she would have remained at Monticello rather than travel to, say, Poplar Forest, at some of the probable times of her conceptions. The question is particularly important given biographer Henry Randall’s intriguing reference to “well known circumstances” that prove Martha Jefferson Randolph’s denial of the charge that Jefferson fathered Sally Hemings’ children. As documented by Randall, Jefferson’s daughter “directed her sons’ attention to the fact that Mr. Jefferson and Sally Hemings could not have met—were far distant from each other—for fifteen months prior to the birth” of the child who supposedly most resembled Jefferson. Almost everyone has assumed that Mrs. Randolph was referring to Jefferson’s absence from Monticello at that time, but she may very well have been referring to Sally Hemings’.62 This intriguing possibility is yet another matter that cries out for additional research.
The Implausibility of the Story
Some people have suggested that the logical principle of Occam’s razor, which states essentially that the simplest of competing theories be preferred to the more complex, when applied to this controversy would make Thomas Jefferson the father of Sally Hemings’ children, or at least of Eston Hemings. But this is a gross misapplication of the principle. Indeed, as Professor Banning points out in his paper, “Occam’s Razor tells us to prefer the simplest theory only when the simplest theory seems equally true.”63 Here, the simple theory advanced by proponents of the Jefferson-Hemings story—that Sally Hemings’ children all had the same father, and that he was Thomas Jefferson—seems plausible only if one ignores all the many facts that strongly suggest against Jefferson’s paternity. Both Professor Banning and Professor Turner have identified a large number of relevant facts, either ignored or unjustifiably downplayed by Professor Gordon-Reed and the Monticello Report, which raises serious questions about the Jefferson paternity thesis. I agree fully with these points and would like here to emphasize two additional matters which, to me, make the thesis extremely implausible.
The Jefferson-Hemings Controversy Page 53