Monticello Affirms Thomas Jefferson Fathered Children with Sally Hemings
A Statement by the Thomas Jefferson Foundation
The issue of Jefferson’s paternity has been the subject of controversy for at least two centuries, ranging from contemporary newspaper articles in 1802 (when Jefferson was President) to scholarly debate well into the 1990s. It is now the Thomas Jefferson Foundation’s view that the issue is a settled historical matter.
A considerable body of evidence stretching from 1802 to 1873 (and beyond) describes Thomas Jefferson as the father of Sally Hemings’s children. It was corroborated by the findings of the Y-chromosome haplotype DNA study conducted by Dr. Eugene Foster and published in the scientific journal Nature in November 1998. The DNA study did prove paternity of a Jefferson family member and corroborated the ample documentary and oral history evidence. Other evidence supports Thomas Jefferson’s paternity as well, including his presence at Monticello during Sally Hemings’s likely windows of conception, the names of Hemings’s surviving children, and the fact that all of her children were granted freedom – they were either allowed to leave the plantation, or legally emancipated in Jefferson’s will, a unique occurrence among Monticello’s enslaved families. The summary of the most important evidence proving Jefferson’s paternity is listed below.1
Madison Hemings provided an account of his mother’s life that was published in an Ohio newspaper in 1873. The basic outline of Madison Hemings’s account, including his mother’s “treaty” with Jefferson and the freedom granted to him and his siblings, was well known to his community before it was published. His narrative is the most important extant evidence and much of the corroborating evidence supports the outline of his narrative.
The Foster et al. (1998) DNA study revealed that male-line descendants of Eston Hemings (a son of Sally Hemings) and male-line descendants of Field Jefferson’s father (who was Thomas Jefferson's grandfather), shared the same Y-chromosome haplotype. This demonstrates that Eston's father was a Jefferson male. This result not only corroborates Madison’s account in the Pike County Republican, it definitively refutes the claims by Jefferson grandchildren, including Ellen Randolph Coolidge and her brother Thomas Jefferson Randolph, that either Peter or Samuel Carr (they could not agree on which one) was the father of Sally Hemings's children.
Madison Hemings was described by a U.S. census taker as the son of Thomas Jefferson in 1870.
Israel Gillette Jefferson, formerly enslaved at Monticello, corroborated Madison Hemings’s claim in the same newspaper, referring to Sally Hemings as Thomas Jefferson’s “concubine.”
Eston Hemings changed his racial identity to white and his surname to Jefferson after moving from Ohio to Wisconsin in 1852. Newspaper accounts in Chillicothe, Ohio, in 1887 and 1902 recalled that Eston resembled Thomas Jefferson.
The two oldest surviving children of Sally Hemings, Beverly Hemings (a male) and Harriet Hemings, were both allowed to leave Monticello without pursuit and were described as “run away” in Jefferson’s inventory of enslaved families. In an 1858 letter to her husband Joseph Coolidge, Jefferson’s granddaughter, Ellen Wayles Randolph Coolidge, (while denying Jefferson’s paternity) described Sally Hemings’s children as “all fair and all set free at my grandfather’s death, or had been suffered to absent themselves permanently before he died."
Jefferson's records of his travels and the birthdays of Sally Hemings’s children reveal that he was present at Monticello during the estimated dates of conception for all six of Hemings's documented offspring. Statistical modeling shows the likelihood of this coincidence for any other male (if we assume that Thomas Jefferson is not the father) as 1 percent, or 1 chance in 100 — strong evidence of Thomas Jefferson’s paternity.2
Oral tradition connecting the Hemings and Jefferson families was transmitted among the descendants of both Madison Hemings and Eston Hemings over many generations. Madison Hemings calls Jefferson his “father” in his 1873 recollections, a fact repeated by his descendants. Eston Hemings’s descendants altered their family history to state that they were related to one of Thomas Jefferson’s relatives in order to hide Eston Hemings’s decision to change his racial identity when he moved to Wisconsin.
Jefferson freed all four surviving Hemings children (in accordance with the terms of his negotiation with Sally Hemings, as reported by her son Madison). He did not grant freedom to any other enslaved nuclear family.
The names of Sally Hemings’s four surviving children — William Beverly Hemings, Harriet Hemings, James Madison Hemings, and Thomas Eston Hemings — suggest family ties to Thomas Jefferson. Annette Gordon-Reed outlines these naming connections in her book, Thomas Jefferson & Sally Hemings: An American Controversy (1997). A man named William Beverly accompanied Jefferson’s father on an expedition through Virginia in 1746, and he was connected to Jefferson’s mother’s family by blood and marriage. There were multiple Harriets in the Randolph family, including a sister and a niece of Thomas Mann Randolph, Jefferson’s son-in-law. Madison Hemings was named at the request of Dolley Madison, whose husband, James Madison, was one of Jefferson’s close friends. Historian and biographer Fawn Brodie offered two possible explanations for Eston Hemings’s name: Eston was the birthplace of Jefferson’s maternal ancestor, William Randolph, in Yorkshire, England. Thomas Eston Randolph was also a first cousin of Jefferson; Jefferson described their two families as being “almost as one.”3 Furthermore, it was convention for Jefferson to be involved in the naming of family members. His children with Martha Jefferson were given the names of his sisters and mother, and he personally named each of his grandchildren.4
Why Remove the Qualifiers?
As the Thomas Jefferson Foundation began planning The Life of Sally Hemings, an exhibit that relies on the account left by her son, Madison Hemings, it became apparent that it was time to reexamine how to characterize Jefferson’s paternity. For nearly twenty years, the most complete summary of evidence has remained the report authored by the Foundation in January 2000. While there are some who disagree, the Foundation’s scholarly advisors and the larger community of academic historians who specialize in early American history have concurred for many years that the evidence is sufficiently strong to state that Thomas Jefferson fathered at least six children with Sally Hemings.
In the new exhibit exploring the life of Sally Hemings, her choices, and her connection to Thomas Jefferson, as well as in updates to our related online materials and print publications, the Foundation will henceforth assert what the evidence indicates and eliminate qualifying language related to the paternity of Eston Hemings as well as that related to Sally Hemings’s three other surviving children, whose descendants were not part of the 1998 DNA study. While it remains possible, though increasingly unlikely, that a more comprehensive documentary and genetic assemblage of evidence could emerge to support a different conclusion, no plausible alternative with the same array of evidence has surfaced in two decades.
2. Bayes' theorem allows us to measure just how strong. To take advantage of it, we need to be willing to summarize the strength of evidence that Jefferson was the father, based on other evidence (say the DNA result and Madison's testimony), as a "prior" probability. Bayes' theorem allows us to rationally update this prior probability, using the 1 percent likelihood, to yield a posterior probability that Jefferson was the father of all six children. Given a prior probability of 50%, Bayes' theorem yields a posterior probability of 99%: 99 chances out of 100 that Jefferson was the father of all six children.
3. Fawn M. Brodie, Thomas Jefferson: An Intimate History(New York: Norton, 1974).
4. Annette Gordon-Reed, Thomas Jefferson & Sally Hemings: An American Controversy (Charlottesville: University of Virginia Press, 1999) pp. 196-201.