Within the last three years, the old story that Thomas Jefferson fathered several children by his slave Sally Hemings—a claim that most Jefferson scholars had earlier considered so implausible that nearly all of them rejected it without a truly rigorous investigation—has gained new credibility and extensive national publicity. In 1997, law professor Annette Gordon-Reed reviewed the evidence and concluded that the case for Jefferson’s paternity was much stronger than scholars had supposed. In 1999, DNA tests proved compatible with the possibility that Jefferson had fathered Eston Hemings, Sally’s youngest son. The DNA report (very misleadingly titled), a conference held at the University of Virginia, a volume of essays resulting from that conference, a forum in the William and Mary Quarterly, a bad movie, a dreadful TV miniseries, and a report by the staff of the Thomas Jefferson Memorial Foundation, which manages Monticello—all accepting the likelihood of Jefferson’s paternity—have publicized the accusation far and wide.

The case for Jefferson’s paternity of the Hemings’ children is strong enough that we can readily understand why so many able scholars have come to think it likely. Their argument stands basically on five substantial pillars:

  • Monticello was populated by numerous mixed-race slaves, some of whom resembled Jefferson, who, moreover, was privately and publicly accused during his lifetime of being the father of Sally Hemings’s children and sometimes of other slaves as well;
  • Madison Hemings, another of Sally’s sons, said that he and his siblings were Jefferson’s children (and his only slave children) in a report which accords in much of its substance with other sources. According to this interview, Thomas and Sally initiated an affair while they were together in Paris from 1787 to 1789. Sally became pregnant and agreed to return to the United States after they entered into a “treaty” in which Jefferson promised “extraordinary privileges” for Sally and freedom for her children when they reached age 21;
  • Sally Hemings conceived all her children when Jefferson was at Monticello and no children when he was not;
  • DNA tests show that a descendant of Eston Hemings carries a Jefferson gene; and
  • DNA tests are incompatible with the possibility that Eston Hemings was fathered by Peter or Samuel Carr, Jefferson’s nephews, the sons of Dabney Carr and Jefferson’s sister Martha and the most plausible alternatives to Jefferson himself as a father.

The case for Jefferson’s paternity is strong enough that the burden of persuasion has shifted to the skeptics. Nevertheless, it is by no means true that no reasonable person can continue to doubt. It seems to me quite evident that unconscious bias and, in some cases, professed political objectives are as on the other.

There are remarkably few uncontested or uncontestable facts in this matter. It is very much a problem of weighing and interpreting the thin and disputable evidence we have—and, ideally, of being wary of our own predispositions and examining this record with as much dispassion as we can possibly muster. Let me run again through those five main pillars:

  • A supposed resemblance between Thomas Jefferson and some of Sally Hemings’s children (or other Monticello slaves) is hardly evidence of a very substantial kind. A resemblance is often seen by some observers and denied by others.
  • Madison Hemings’s “Memoir” is actually a reporter’s account of an interview with him, first published in 1873, in which we cannot tell with certainty which words and statements came from Madison Hemings and which may have been changed by the reporter—an account, moreover, in which both the interviewee and the reporter might or might not have been telling the truth. While many of its details can be corroborated from other sources, some can be disproved, and some suggest that Hemings or the interviewer was leaning on previously published accusations. Obviously, Madison Hemings could not have known first-hand who was the father of Sally’s children or whether all of them were fathered by the same man. It is not even certain that he had the story of the affair and the “treaty” from Sally herself, who left no reported statement that it was true. The “Memoir” does not unquestionably outweigh the countervailing testimony of a former overseer at Monticello and several members of Jefferson’s family. Neither is Madison’s story consistent with the oral tradition in the family of his brother Eston, who apparently passed down a claim that he was the son of a close relative of Jefferson himself. We can well doubt that Sally Hemings became pregnant in France and would have entered into the improbable “treaty” that Madison described. No one has solved the mystery of the baby, supposedly conceived in France, that Madison said was born in 1790. It is not clear that Thomas Woodson, said to be that baby in another oral tradition, was Sally Hemings’s son, and it now seems certain (based on DNA-testing of Woodson’s descendants) that, if he was, he was not the son of Thomas Jefferson, too.
  • The facts about Sally Hemings’s conceptions—at least about the conceptions after 1789—are among the strongest elements in the case for Thomas Jefferson’s paternity. It is curious that Jefferson’s grandson, Thomas Jefferson Randolph, and the historian Henry Randall both believed that they had independently confirmed Martha Randolph’s insistence that the child who most resembled Jefferson (presumably Eston) could not have been his, since he and Sally were “far distant” from one another for fifteen months and, indeed, for most of the specific period during which Eston would have been conceived. We have no reason, either, to believe that all these other people were simply lying.
  • The DNA results certainly do not prove Thomas Jefferson’s paternity of any of the Hemings’ children. These results were widely misreported as having done that, and this misperception may well have spread indelibly among the public. But every knowledgeable authority, including the scientists who conducted the tests, has denied that this is what was found and even that it was possible to make such findings from the sort of tests that were done. In plain words, they showed that a descendent of one of Sally Hemings’s children carries a Jefferson gene, not a Carr one, and that neither the Carrs nor the Jeffersons are related to the Woodsons.
  • Although they implicate a Jefferson, not a Carr, as Eston Hemings’s father, the DNA results cannot show that Thomas Jefferson was any more likely to have been Eston’s father than any of Thomas’s male-line relatives who might have had relations with Sally Hemings at the relevant times. To me, the absence of a wholly plausible alternative to Jefferson as a father is another of the strongest elements in the pro-paternity case. Thomas’s younger brother, Randolph Jefferson, was probably at Monticello at the right time, perhaps with some of his sons; and Randolph, unlike Thomas, was known to dance and play his fiddle with the slaves. But Randolph or one of his sons seems so unlikely to have fathered all of Sally Hemings’s children that respondents will continue to protest that this is a grasping at straws by Jefferson’s defenders. Peter or Samuel Carr, the nephews who were believed to be Sally’s lovers by Jefferson’s grandchildren and who were reported to have confessed it to him and another witness by Thomas Jefferson Randolph, might still, to be sure, have fathered all of Sally’s children except Eston. But this would not explain why Sally became pregnant only when Thomas was at home. So far as we know, one or both of the Carr brothers was usually at Monticello or quite nearby from 1794 to 1808.

The simplest explanation of most of the salient facts—that Jefferson fathered all the children born between 1795 and 1808—seems sound only when we focus so intently on the points I have reviewed thus far. I remain a skeptic principally because of a long list of highly implausible things we have to believe in order to accept the Jefferson-Hemings story in anything like the terms advanced by Madison Hemings and assumed to be likely in much recent work.

* * *

We are asked to believe this U.S. minister to France, a man with ready access to some of the most beautiful and accomplished women in Europe, initiated an affair with a 15 or 16-year-old slave girl, whom Abigail Adams had recently described as more in need of care than the 8-year-old she had attended. This girl was the personal servant (and likely something of a confidant) of Jefferson’s two daughters—an individual, that is, whose discretion the accomplished politician and diplomat could not possibly have trusted. Although it may well be that Sally lived, during much of her time in Paris, in the cross-town convent where Patsy and Polly were being schooled, we are to assume that Thomas and Sally carried on their affair in the crowded two-bedroom townhouse where Jefferson lived—and did so without arousing suspicion of David Humphreys, who slept in one of the bedrooms, or of anyone else who was there. She became pregnant with Jefferson’s child, the story continues, entered into an agreement with him, and (with Jefferson taking care that she would have a berth convenient to his daughters) sailed back to the U.S. in this condition with him, his two girls, and her brother James. The baby, if it existed, either died soon after birth, leaving no trace other than Madison Hemings’s statement, or became the elusive, unrecorded 12-year-old slave named “Tom,” mentioned in James Callender’s infamous 1802 newspaper accusation, who, if he later took the name Tom Woodson, was not Thomas Jefferson’s child.

According to this story, Jefferson would continue in a monogamous and fertile relationship with Sally for nearly 20 years, ultimately fathering five or six more children (the first of whom, however, was not born until 1795). During these twenty years, he was content, as was she, to confine the relationship to the times when he was at Monticello, although he took other slaves with him wherever he went and as many as a dozen to the White House. On these terms, he continued in the relationship until at least age 64, when Eston Hemings was conceived, five years after he had been publicly accused of a relationship with Sally and while he was contemplating his second presidential term. He carried it on, all this while, while constantly surrounded by visitors and by a large white family, none of whom—and least of all the daughters who would have known Sally best—ever had the least suspicion that he was involved with any of his slaves or ever saw the slightest indication that he was closer to Sally than to any other servant.

Indeed, the grandchildren who grew up at Monticello and managed it during Jefferson’s last years did not merely say that any such relationship was wholly unsuspected—never a touch or a word or a glance—they said it was simply impossible in this particular house. “His apartment,” his granddaughter told her husband, “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.” In fact, apart from Madison Hemings, no one who ever lived at Monticello and none of the uncounted visitors who stayed there overnight ever said that he was involved with Sally—not even Sally herself, though she lived in practical freedom in Charlottesville for ten years after his death.

It is possible, of course, that everyone except Madison Hemings was lying or covering up or engaged in psychological “denial.” Jefferson’s family had an interest in protecting his reputation, much as Madison Hemings had an interest in claiming descent from a famous man. I see no reason to think that any of these people were deliberately making things up. What of eye-witnesses who had no obvious interest in the matter either way? Former household slave Isaac Jefferson mentioned Sally Hemings in later years; but did not so much as hint that there was any special relationship between her and Jefferson. And in another interview, Edmund Bacon, who was overseer at Monticello when Eston was conceived and may have worked there for years before, raised the subject of the accusations against his employer.

He freed one girl some years before he died, and there was a great deal of talk about it. She was nearly as white as anybody, and very beautiful. People said he freed her because she was his own daughter. She was not his daughter; she was ______’s daughter. I know that. I have seen him come out of her mother’s room many a mourning, when I went up to Monticello very early.

The girl was certainly Harriet Hemings, Sally’s daughter. The father was named by Bacon but protected by the reporter, a preacher in Kentucky.

All of Sally Hemings’s children who lived to adulthood did achieve their freedom, either de facto or de jure, and it is often said that they were the only nuclear family of Monticello who did. However, contrary to the terms of the “treaty” as Madison Hemings described it, Sally Hemings did not receive extraordinary privileges at Monticello. Jefferson fed, clothed, and treated Sally Hemings pretty much indistinguishably from his other household servants, recorded her life and childbirths in much the same way, and left her as part of the estate. By Madison Hemings’s own account, moreover, Jefferson showed no particular affection for her children and reared them much as he did other household slaves. Can we believe that Jefferson, always concerned for contemporary and historical regard, would not only have continued an affair with Sally long after it became a public scandal—and still without arousing his family’s suspicion—but would have been so brazen about it as to have a slave whose resemblance to him was absolutely startling serve his foreign guests at dinner?

* * *

Strange things happen where sex is concerned, not to mention slavery and race. Racial mixing between masters and servants happened all the time in the Old South. It does not tax us highly to imagine how perfectly ordinary it was, or what a range of conduct it encompassed: from boys and girls coming into adolescence together in the familiarity of a plantation and playing games more adult, to deep affection, to tyranny in which it was as easy for a master so inclined to order a woman to his bed as it was to order her to do the laundry or tend the crop. As Jefferson himself told us, “the man must be a prodigy who can retain his manners and morals undepraved” within such a system.

Was Jefferson such a man? Did he, though, a slaveholder, retain his morals relatively undepraved? The honest answer, to my mind, is that we simply do not know and may never possess the evidence that could answer the question absolutely. Parts of the argument for his paternity of Sally Hemings’s children are hard to overcome. But this does not seem any harder, to my mind, than swallowing the lengthy list of radical implausibilities we have to swallow to accept that the story is true. It’s a close call; a hard case, as the lawyers would say. But there are surely grounds for strong, continuing doubts. Count me, then, in the minority who seem to think that the proper verdict, at this moment, has to be “not proven.”